From 8abed8d62a24da6dd02de441fef94245c4915faf Mon Sep 17 00:00:00 2001 From: Yang Shi Date: Mon, 27 Jun 2022 13:44:58 -0700 Subject: mm: rmap: use the correct parameter name for DEFINE_PAGE_VMA_WALK The parameter used by DEFINE_PAGE_VMA_WALK is _page not page, fix the parameter name. It didn't cause any build error, it is probably because the only caller is write_protect_page() from ksm.c, which pass in page. Link: https://lkml.kernel.org/r/20220512174551.81279-1-shy828301@gmail.com Fixes: 2aff7a4755be ("mm: Convert page_vma_mapped_walk to work on PFNs") Signed-off-by: Yang Shi Reviewed-by: Muchun Song Reviewed-by: Matthew Wilcox (Oracle) Cc: Muchun Song Signed-off-by: Andrew Morton --- include/linux/rmap.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/rmap.h b/include/linux/rmap.h index 9ec23138e410..bf80adca980b 100644 --- a/include/linux/rmap.h +++ b/include/linux/rmap.h @@ -325,8 +325,8 @@ struct page_vma_mapped_walk { #define DEFINE_PAGE_VMA_WALK(name, _page, _vma, _address, _flags) \ struct page_vma_mapped_walk name = { \ .pfn = page_to_pfn(_page), \ - .nr_pages = compound_nr(page), \ - .pgoff = page_to_pgoff(page), \ + .nr_pages = compound_nr(_page), \ + .pgoff = page_to_pgoff(_page), \ .vma = _vma, \ .address = _address, \ .flags = _flags, \ -- cgit v1.2.3 From 295d3cb6a1f47315813c3314905a5958e432780f Mon Sep 17 00:00:00 2001 From: Yang Shi Date: Fri, 13 May 2022 12:17:05 -0700 Subject: mm/page_vma_mapped.c: check possible huge PMD map with transhuge_vma_suitable() IIUC page_vma_mapped_walk() checks if the vma is possibly huge PMD mapped with transparent_hugepage_active() and "pvmw->nr_pages >= HPAGE_PMD_NR". Actually pvmw->nr_pages is returned by compound_nr() or folio_nr_pages(), so the page should be THP as long as "pvmw->nr_pages >= HPAGE_PMD_NR". And it is guaranteed THP is allocated for valid VMA in the first place. But it may be not PMD mapped if the VMA is file VMA and it is not properly aligned. The transhuge_vma_suitable() is used to do such check, so replace transparent_hugepage_active() to it, which is too heavy and overkilling. Link: https://lkml.kernel.org/r/20220513191705.457775-1-shy828301@gmail.com Signed-off-by: Yang Shi Reviewed-by: Muchun Song Cc: Matthew Wilcox (Oracle) Signed-off-by: Andrew Morton --- include/linux/huge_mm.h | 8 ++++++-- mm/page_vma_mapped.c | 2 +- 2 files changed, 7 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index de29821231c9..648cb3ce7099 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -117,8 +117,10 @@ extern struct kobj_attribute shmem_enabled_attr; extern unsigned long transparent_hugepage_flags; static inline bool transhuge_vma_suitable(struct vm_area_struct *vma, - unsigned long haddr) + unsigned long addr) { + unsigned long haddr; + /* Don't have to check pgoff for anonymous vma */ if (!vma_is_anonymous(vma)) { if (!IS_ALIGNED((vma->vm_start >> PAGE_SHIFT) - vma->vm_pgoff, @@ -126,6 +128,8 @@ static inline bool transhuge_vma_suitable(struct vm_area_struct *vma, return false; } + haddr = addr & HPAGE_PMD_MASK; + if (haddr < vma->vm_start || haddr + HPAGE_PMD_SIZE > vma->vm_end) return false; return true; @@ -342,7 +346,7 @@ static inline bool transparent_hugepage_active(struct vm_area_struct *vma) } static inline bool transhuge_vma_suitable(struct vm_area_struct *vma, - unsigned long haddr) + unsigned long addr) { return false; } diff --git a/mm/page_vma_mapped.c b/mm/page_vma_mapped.c index c10f839fc410..e971a467fcdf 100644 --- a/mm/page_vma_mapped.c +++ b/mm/page_vma_mapped.c @@ -243,7 +243,7 @@ restart: * cleared *pmd but not decremented compound_mapcount(). */ if ((pvmw->flags & PVMW_SYNC) && - transparent_hugepage_active(vma) && + transhuge_vma_suitable(vma, pvmw->address) && (pvmw->nr_pages >= HPAGE_PMD_NR)) { spinlock_t *ptl = pmd_lock(mm, pvmw->pmd); -- cgit v1.2.3 From 80fa6a83d8f102fb8554a34d072a6ea0dab28ae6 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Mon, 27 Jun 2022 13:44:58 -0700 Subject: mm: discard __GFP_ATOMIC __GFP_ATOMIC serves little purpose. Its main effect is to set ALLOC_HARDER which adds a few little boosts to increase the chance of an allocation succeeding, one of which is to lower the water-mark at which it will succeed. It is *always* paired with __GFP_HIGH which sets ALLOC_HIGH which also adjusts this watermark. It is probable that other users of __GFP_HIGH should benefit from the other little bonuses that __GFP_ATOMIC gets. __GFP_ATOMIC also gives a warning if used with __GFP_DIRECT_RECLAIM. There is little point to this. We already get a might_sleep() warning if __GFP_DIRECT_RECLAIM is set. __GFP_ATOMIC allows the "watermark_boost" to be side-stepped. It is probable that testing ALLOC_HARDER is a better fit here. __GFP_ATOMIC is used by tegra-smmu.c to check if the allocation might sleep. This should test __GFP_DIRECT_RECLAIM instead. This patch: - removes __GFP_ATOMIC - causes __GFP_HIGH to set ALLOC_HARDER unless __GFP_NOMEMALLOC is set (as well as ALLOC_HIGH). - makes other adjustments as suggested by the above. The net result is not change to GFP_ATOMIC allocations. Other allocations that use __GFP_HIGH will benefit from a few different extra privileges. This affects: xen, dm, md, ntfs3 the vermillion frame buffer hibernation ksm swap all of which likely produce more benefit than cost if these selected allocation are more likely to succeed quickly. Link: https://lkml.kernel.org/r/163712397076.13692.4727608274002939094@noble.neil.brown.name Signed-off-by: NeilBrown Reviewed-by: Matthew Wilcox (Oracle) Cc: Michal Hocko Cc: Thierry Reding Cc: Mel Gorman Signed-off-by: Andrew Morton --- Documentation/mm/balance.rst | 2 +- drivers/iommu/tegra-smmu.c | 4 ++-- include/linux/gfp.h | 12 ++++-------- include/trace/events/mmflags.h | 1 - lib/test_printf.c | 8 ++++---- mm/internal.h | 2 +- mm/page_alloc.c | 16 ++++------------ tools/include/linux/gfp.h | 3 +-- tools/perf/builtin-kmem.c | 1 - 9 files changed, 17 insertions(+), 32 deletions(-) (limited to 'include/linux') diff --git a/Documentation/mm/balance.rst b/Documentation/mm/balance.rst index 6a1fadf3e173..e38e9d83c1c7 100644 --- a/Documentation/mm/balance.rst +++ b/Documentation/mm/balance.rst @@ -6,7 +6,7 @@ Memory Balancing Started Jan 2000 by Kanoj Sarcar -Memory balancing is needed for !__GFP_ATOMIC and !__GFP_KSWAPD_RECLAIM as +Memory balancing is needed for !__GFP_HIGH and !__GFP_KSWAPD_RECLAIM as well as for non __GFP_IO allocations. The first reason why a caller may avoid reclaim is that the caller can not diff --git a/drivers/iommu/tegra-smmu.c b/drivers/iommu/tegra-smmu.c index 1fea68e551f1..2f2b12033618 100644 --- a/drivers/iommu/tegra-smmu.c +++ b/drivers/iommu/tegra-smmu.c @@ -671,12 +671,12 @@ static struct page *as_get_pde_page(struct tegra_smmu_as *as, * allocate page in a sleeping context if GFP flags permit. Hence * spinlock needs to be unlocked and re-locked after allocation. */ - if (!(gfp & __GFP_ATOMIC)) + if (gfp & __GFP_DIRECT_RECLAIM) spin_unlock_irqrestore(&as->lock, *flags); page = alloc_page(gfp | __GFP_DMA | __GFP_ZERO); - if (!(gfp & __GFP_ATOMIC)) + if (gfp & __GFP_DIRECT_RECLAIM) spin_lock_irqsave(&as->lock, *flags); /* diff --git a/include/linux/gfp.h b/include/linux/gfp.h index 2d2ccae933c2..9a88cce23e17 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -39,7 +39,7 @@ struct vm_area_struct; #define ___GFP_IO 0x40u #define ___GFP_FS 0x80u #define ___GFP_ZERO 0x100u -#define ___GFP_ATOMIC 0x200u +/* 0x200u unused */ #define ___GFP_DIRECT_RECLAIM 0x400u #define ___GFP_KSWAPD_RECLAIM 0x800u #define ___GFP_WRITE 0x1000u @@ -124,11 +124,8 @@ struct vm_area_struct; * * %__GFP_HIGH indicates that the caller is high-priority and that granting * the request is necessary before the system can make forward progress. - * For example, creating an IO context to clean pages. - * - * %__GFP_ATOMIC indicates that the caller cannot reclaim or sleep and is - * high priority. Users are typically interrupt handlers. This may be - * used in conjunction with %__GFP_HIGH + * For example creating an IO context to clean pages and requests + * from atomic context. * * %__GFP_MEMALLOC allows access to all memory. This should only be used when * the caller guarantees the allocation will allow more memory to be freed @@ -143,7 +140,6 @@ struct vm_area_struct; * %__GFP_NOMEMALLOC is used to explicitly forbid access to emergency reserves. * This takes precedence over the %__GFP_MEMALLOC flag if both are set. */ -#define __GFP_ATOMIC ((__force gfp_t)___GFP_ATOMIC) #define __GFP_HIGH ((__force gfp_t)___GFP_HIGH) #define __GFP_MEMALLOC ((__force gfp_t)___GFP_MEMALLOC) #define __GFP_NOMEMALLOC ((__force gfp_t)___GFP_NOMEMALLOC) @@ -337,7 +333,7 @@ struct vm_area_struct; * version does not attempt reclaim/compaction at all and is by default used * in page fault path, while the non-light is used by khugepaged. */ -#define GFP_ATOMIC (__GFP_HIGH|__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) +#define GFP_ATOMIC (__GFP_HIGH|__GFP_KSWAPD_RECLAIM) #define GFP_KERNEL (__GFP_RECLAIM | __GFP_IO | __GFP_FS) #define GFP_KERNEL_ACCOUNT (GFP_KERNEL | __GFP_ACCOUNT) #define GFP_NOWAIT (__GFP_KSWAPD_RECLAIM) diff --git a/include/trace/events/mmflags.h b/include/trace/events/mmflags.h index e87cb2b80ed3..11524cda4a95 100644 --- a/include/trace/events/mmflags.h +++ b/include/trace/events/mmflags.h @@ -31,7 +31,6 @@ gfpflag_string(__GFP_HIGHMEM), \ gfpflag_string(GFP_DMA32), \ gfpflag_string(__GFP_HIGH), \ - gfpflag_string(__GFP_ATOMIC), \ gfpflag_string(__GFP_IO), \ gfpflag_string(__GFP_FS), \ gfpflag_string(__GFP_NOWARN), \ diff --git a/lib/test_printf.c b/lib/test_printf.c index 07309c45f327..8010de49b6c5 100644 --- a/lib/test_printf.c +++ b/lib/test_printf.c @@ -673,17 +673,17 @@ flags(void) gfp = GFP_ATOMIC|__GFP_DMA; test("GFP_ATOMIC|GFP_DMA", "%pGg", &gfp); - gfp = __GFP_ATOMIC; - test("__GFP_ATOMIC", "%pGg", &gfp); + gfp = __GFP_HIGH; + test("__GFP_HIGH", "%pGg", &gfp); /* Any flags not translated by the table should remain numeric */ gfp = ~__GFP_BITS_MASK; snprintf(cmp_buffer, BUF_SIZE, "%#lx", (unsigned long) gfp); test(cmp_buffer, "%pGg", &gfp); - snprintf(cmp_buffer, BUF_SIZE, "__GFP_ATOMIC|%#lx", + snprintf(cmp_buffer, BUF_SIZE, "__GFP_HIGH|%#lx", (unsigned long) gfp); - gfp |= __GFP_ATOMIC; + gfp |= __GFP_HIGH; test(cmp_buffer, "%pGg", &gfp); kfree(cmp_buffer); diff --git a/mm/internal.h b/mm/internal.h index c0f8fbe0445b..7cf12a15475b 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -24,7 +24,7 @@ struct folio_batch; #define GFP_RECLAIM_MASK (__GFP_RECLAIM|__GFP_HIGH|__GFP_IO|__GFP_FS|\ __GFP_NOWARN|__GFP_RETRY_MAYFAIL|__GFP_NOFAIL|\ __GFP_NORETRY|__GFP_MEMALLOC|__GFP_NOMEMALLOC|\ - __GFP_ATOMIC|__GFP_NOLOCKDEP) + __GFP_NOLOCKDEP) /* The GFP flags allowed during early boot */ #define GFP_BOOT_MASK (__GFP_BITS_MASK & ~(__GFP_RECLAIM|__GFP_IO|__GFP_FS)) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 81fadb266973..0a32aee47ddd 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -3980,12 +3980,12 @@ static inline bool zone_watermark_fast(struct zone *z, unsigned int order, free_pages)) return true; /* - * Ignore watermark boosting for GFP_ATOMIC order-0 allocations + * Ignore watermark boosting for GFP_HIGH order-0 allocations * when checking the min watermark. The min watermark is the * point where boosting is ignored so that kswapd is woken up * when below the low watermark. */ - if (unlikely(!order && (gfp_mask & __GFP_ATOMIC) && z->watermark_boost + if (unlikely(!order && (alloc_flags & ALLOC_HARDER) && z->watermark_boost && ((alloc_flags & ALLOC_WMARK_MASK) == WMARK_MIN))) { mark = z->_watermark[WMARK_MIN]; return __zone_watermark_ok(z, order, mark, highest_zoneidx, @@ -4720,12 +4720,12 @@ gfp_to_alloc_flags(gfp_t gfp_mask) * The caller may dip into page reserves a bit more if the caller * cannot run direct reclaim, or if the caller has realtime scheduling * policy or is asking for __GFP_HIGH memory. GFP_ATOMIC requests will - * set both ALLOC_HARDER (__GFP_ATOMIC) and ALLOC_HIGH (__GFP_HIGH). + * set both ALLOC_HARDER (unless __GFP_NOMEMALLOC) and ALLOC_HIGH. */ alloc_flags |= (__force int) (gfp_mask & (__GFP_HIGH | __GFP_KSWAPD_RECLAIM)); - if (gfp_mask & __GFP_ATOMIC) { + if (gfp_mask & __GFP_HIGH) { /* * Not worth trying to allocate harder for __GFP_NOMEMALLOC even * if it can't schedule. @@ -4918,14 +4918,6 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order, unsigned int cpuset_mems_cookie; int reserve_flags; - /* - * We also sanity check to catch abuse of atomic reserves being used by - * callers that are not in atomic context. - */ - if (WARN_ON_ONCE((gfp_mask & (__GFP_ATOMIC|__GFP_DIRECT_RECLAIM)) == - (__GFP_ATOMIC|__GFP_DIRECT_RECLAIM))) - gfp_mask &= ~__GFP_ATOMIC; - retry_cpuset: compaction_retries = 0; no_progress_loops = 0; diff --git a/tools/include/linux/gfp.h b/tools/include/linux/gfp.h index b238dbc9eb85..56eec4445bc9 100644 --- a/tools/include/linux/gfp.h +++ b/tools/include/linux/gfp.h @@ -12,7 +12,6 @@ #define __GFP_FS 0x80u #define __GFP_NOWARN 0x200u #define __GFP_ZERO 0x8000u -#define __GFP_ATOMIC 0x80000u #define __GFP_ACCOUNT 0x100000u #define __GFP_DIRECT_RECLAIM 0x400000u #define __GFP_KSWAPD_RECLAIM 0x2000000u @@ -20,7 +19,7 @@ #define __GFP_RECLAIM (__GFP_DIRECT_RECLAIM | __GFP_KSWAPD_RECLAIM) #define GFP_ZONEMASK 0x0fu -#define GFP_ATOMIC (__GFP_HIGH | __GFP_ATOMIC | __GFP_KSWAPD_RECLAIM) +#define GFP_ATOMIC (__GFP_HIGH | __GFP_KSWAPD_RECLAIM) #define GFP_KERNEL (__GFP_RECLAIM | __GFP_IO | __GFP_FS) #define GFP_NOWAIT (__GFP_KSWAPD_RECLAIM) diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index ebfab2ca1702..4a06d83f2ac5 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -640,7 +640,6 @@ static const struct { { "__GFP_HIGHMEM", "HM" }, { "GFP_DMA32", "D32" }, { "__GFP_HIGH", "H" }, - { "__GFP_ATOMIC", "_A" }, { "__GFP_IO", "I" }, { "__GFP_FS", "F" }, { "__GFP_NOWARN", "NWR" }, -- cgit v1.2.3 From df33742a4ad9a39e02608aca11c6eb1ddf3afbe0 Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:46:47 +0000 Subject: Maple Tree: add new data structure Patch series "Introducing the Maple Tree". The maple tree is an RCU-safe range based B-tree designed to use modern processor cache efficiently. There are a number of places in the kernel that a non-overlapping range-based tree would be beneficial, especially one with a simple interface. If you use an rbtree with other data structures to improve performance or an interval tree to track non-overlapping ranges, then this is for you. The tree has a branching factor of 10 for non-leaf nodes and 16 for leaf nodes. With the increased branching factor, it is significantly shorter than the rbtree so it has fewer cache misses. The removal of the linked list between subsequent entries also reduces the cache misses and the need to pull in the previous and next VMA during many tree alterations. The first user that is covered in this patch set is the vm_area_struct, where three data structures are replaced by the maple tree: the augmented rbtree, the vma cache, and the linked list of VMAs in the mm_struct. The long term goal is to reduce or remove the mmap_lock contention. The plan is to get to the point where we use the maple tree in RCU mode. Readers will not block for writers. A single write operation will be allowed at a time. A reader re-walks if stale data is encountered. VMAs would be RCU enabled and this mode would be entered once multiple tasks are using the mm_struct. Davidlor said : Yes I like the maple tree, and at this stage I don't think we can ask for : more from this series wrt the MM - albeit there seems to still be some : folks reporting breakage. Fundamentally I see Liam's work to (re)move : complexity out of the MM (not to say that the actual maple tree is not : complex) by consolidating the three complimentary data structures very : much worth it considering performance does not take a hit. This was very : much a turn off with the range locking approach, which worst case scenario : incurred in prohibitive overhead. Also as Liam and Matthew have : mentioned, RCU opens up a lot of nice performance opportunities, and in : addition academia[1] has shown outstanding scalability of address spaces : with the foundation of replacing the locked rbtree with RCU aware trees. A similar work has been discovered in the academic press https://pdos.csail.mit.edu/papers/rcuvm:asplos12.pdf Sheer coincidence. We designed our tree with the intention of solving the hardest problem first. Upon settling on a b-tree variant and a rough outline, we researched ranged based b-trees and RCU b-trees and did find that article. So it was nice to find reassurances that we were on the right path, but our design choice of using ranges made that paper unusable for us. This patch (of 69): The maple tree is an RCU-safe range based B-tree designed to use modern processor cache efficiently. There are a number of places in the kernel that a non-overlapping range-based tree would be beneficial, especially one with a simple interface. If you use an rbtree with other data structures to improve performance or an interval tree to track non-overlapping ranges, then this is for you. The tree has a branching factor of 10 for non-leaf nodes and 16 for leaf nodes. With the increased branching factor, it is significantly shorter than the rbtree so it has fewer cache misses. The removal of the linked list between subsequent entries also reduces the cache misses and the need to pull in the previous and next VMA during many tree alterations. The first user that is covered in this patch set is the vm_area_struct, where three data structures are replaced by the maple tree: the augmented rbtree, the vma cache, and the linked list of VMAs in the mm_struct. The long term goal is to reduce or remove the mmap_lock contention. The plan is to get to the point where we use the maple tree in RCU mode. Readers will not block for writers. A single write operation will be allowed at a time. A reader re-walks if stale data is encountered. VMAs would be RCU enabled and this mode would be entered once multiple tasks are using the mm_struct. Link: https://lkml.kernel.org/r/20220621204632.3370049-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220504010716.661115-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220504002554.654642-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220504010716.661115-3-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220511144304.1430851-2-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220517145913.3480729-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220517152209.3486724-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220519150304.1289636-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220607063834.7004-1-lukas.bulwahn@gmail.com Link: https://lkml.kernel.org/r/20220615141921.417598-2-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220615141921.417598-3-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220616011739.802669-3-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220615174213.738849-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220617134609.1771611-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-2-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Signed-off-by: Matthew Wilcox (Oracle) Tested-by: David Howells Tested-by: Sven Schnelle Signed-off-by: Lukas Bulwahn Cc: Catalin Marinas Cc: SeongJae Park Cc: Vlastimil Babka Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- Documentation/core-api/index.rst | 1 + Documentation/core-api/maple_tree.rst | 217 + MAINTAINERS | 12 + include/linux/maple_tree.h | 685 ++ include/trace/events/maple_tree.h | 123 + init/main.c | 2 + lib/Kconfig.debug | 15 + lib/Makefile | 3 +- lib/maple_tree.c | 7041 ++++++++++++++++++++ tools/testing/radix-tree/.gitignore | 2 + tools/testing/radix-tree/generated/autoconf.h | 1 + tools/testing/radix-tree/linux/maple_tree.h | 7 + tools/testing/radix-tree/maple.c | 59 + tools/testing/radix-tree/trace/events/maple_tree.h | 5 + 14 files changed, 8172 insertions(+), 1 deletion(-) create mode 100644 Documentation/core-api/maple_tree.rst create mode 100644 include/linux/maple_tree.h create mode 100644 include/trace/events/maple_tree.h create mode 100644 lib/maple_tree.c create mode 100644 tools/testing/radix-tree/linux/maple_tree.h create mode 100644 tools/testing/radix-tree/maple.c create mode 100644 tools/testing/radix-tree/trace/events/maple_tree.h (limited to 'include/linux') diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index 5b1188494bcd..86ae4dbca1f5 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -36,6 +36,7 @@ Library functionality that is used throughout the kernel. kref assoc_array xarray + maple_tree idr circular-buffers rbtree diff --git a/Documentation/core-api/maple_tree.rst b/Documentation/core-api/maple_tree.rst new file mode 100644 index 000000000000..45defcf15da7 --- /dev/null +++ b/Documentation/core-api/maple_tree.rst @@ -0,0 +1,217 @@ +.. SPDX-License-Identifier: GPL-2.0+ + + +========== +Maple Tree +========== + +:Author: Liam R. Howlett + +Overview +======== + +The Maple Tree is a B-Tree data type which is optimized for storing +non-overlapping ranges, including ranges of size 1. The tree was designed to +be simple to use and does not require a user written search method. It +supports iterating over a range of entries and going to the previous or next +entry in a cache-efficient manner. The tree can also be put into an RCU-safe +mode of operation which allows reading and writing concurrently. Writers must +synchronize on a lock, which can be the default spinlock, or the user can set +the lock to an external lock of a different type. + +The Maple Tree maintains a small memory footprint and was designed to use +modern processor cache efficiently. The majority of the users will be able to +use the normal API. An :ref:`maple-tree-advanced-api` exists for more complex +scenarios. The most important usage of the Maple Tree is the tracking of the +virtual memory areas. + +The Maple Tree can store values between ``0`` and ``ULONG_MAX``. The Maple +Tree reserves values with the bottom two bits set to '10' which are below 4096 +(ie 2, 6, 10 .. 4094) for internal use. If the entries may use reserved +entries then the users can convert the entries using xa_mk_value() and convert +them back by calling xa_to_value(). If the user needs to use a reserved +value, then the user can convert the value when using the +:ref:`maple-tree-advanced-api`, but are blocked by the normal API. + +The Maple Tree can also be configured to support searching for a gap of a given +size (or larger). + +Pre-allocating of nodes is also supported using the +:ref:`maple-tree-advanced-api`. This is useful for users who must guarantee a +successful store operation within a given +code segment when allocating cannot be done. Allocations of nodes are +relatively small at around 256 bytes. + +.. _maple-tree-normal-api: + +Normal API +========== + +Start by initialising a maple tree, either with DEFINE_MTREE() for statically +allocated maple trees or mt_init() for dynamically allocated ones. A +freshly-initialised maple tree contains a ``NULL`` pointer for the range ``0`` +- ``ULONG_MAX``. There are currently two types of maple trees supported: the +allocation tree and the regular tree. The regular tree has a higher branching +factor for internal nodes. The allocation tree has a lower branching factor +but allows the user to search for a gap of a given size or larger from either +``0`` upwards or ``ULONG_MAX`` down. An allocation tree can be used by +passing in the ``MT_FLAGS_ALLOC_RANGE`` flag when initialising the tree. + +You can then set entries using mtree_store() or mtree_store_range(). +mtree_store() will overwrite any entry with the new entry and return 0 on +success or an error code otherwise. mtree_store_range() works in the same way +but takes a range. mtree_load() is used to retrieve the entry stored at a +given index. You can use mtree_erase() to erase an entire range by only +knowing one value within that range, or mtree_store() call with an entry of +NULL may be used to partially erase a range or many ranges at once. + +If you want to only store a new entry to a range (or index) if that range is +currently ``NULL``, you can use mtree_insert_range() or mtree_insert() which +return -EEXIST if the range is not empty. + +You can search for an entry from an index upwards by using mt_find(). + +You can walk each entry within a range by calling mt_for_each(). You must +provide a temporary variable to store a cursor. If you want to walk each +element of the tree then ``0`` and ``ULONG_MAX`` may be used as the range. If +the caller is going to hold the lock for the duration of the walk then it is +worth looking at the mas_for_each() API in the :ref:`maple-tree-advanced-api` +section. + +Sometimes it is necessary to ensure the next call to store to a maple tree does +not allocate memory, please see :ref:`maple-tree-advanced-api` for this use case. + +Finally, you can remove all entries from a maple tree by calling +mtree_destroy(). If the maple tree entries are pointers, you may wish to free +the entries first. + +Allocating Nodes +---------------- + +The allocations are handled by the internal tree code. See +:ref:`maple-tree-advanced-alloc` for other options. + +Locking +------- + +You do not have to worry about locking. See :ref:`maple-tree-advanced-locks` +for other options. + +The Maple Tree uses RCU and an internal spinlock to synchronise access: + +Takes RCU read lock: + * mtree_load() + * mt_find() + * mt_for_each() + * mt_next() + * mt_prev() + +Takes ma_lock internally: + * mtree_store() + * mtree_store_range() + * mtree_insert() + * mtree_insert_range() + * mtree_erase() + * mtree_destroy() + * mt_set_in_rcu() + * mt_clear_in_rcu() + +If you want to take advantage of the internal lock to protect the data +structures that you are storing in the Maple Tree, you can call mtree_lock() +before calling mtree_load(), then take a reference count on the object you +have found before calling mtree_unlock(). This will prevent stores from +removing the object from the tree between looking up the object and +incrementing the refcount. You can also use RCU to avoid dereferencing +freed memory, but an explanation of that is beyond the scope of this +document. + +.. _maple-tree-advanced-api: + +Advanced API +============ + +The advanced API offers more flexibility and better performance at the +cost of an interface which can be harder to use and has fewer safeguards. +You must take care of your own locking while using the advanced API. +You can use the ma_lock, RCU or an external lock for protection. +You can mix advanced and normal operations on the same array, as long +as the locking is compatible. The :ref:`maple-tree-normal-api` is implemented +in terms of the advanced API. + +The advanced API is based around the ma_state, this is where the 'mas' +prefix originates. The ma_state struct keeps track of tree operations to make +life easier for both internal and external tree users. + +Initialising the maple tree is the same as in the :ref:`maple-tree-normal-api`. +Please see above. + +The maple state keeps track of the range start and end in mas->index and +mas->last, respectively. + +mas_walk() will walk the tree to the location of mas->index and set the +mas->index and mas->last according to the range for the entry. + +You can set entries using mas_store(). mas_store() will overwrite any entry +with the new entry and return the first existing entry that is overwritten. +The range is passed in as members of the maple state: index and last. + +You can use mas_erase() to erase an entire range by setting index and +last of the maple state to the desired range to erase. This will erase +the first range that is found in that range, set the maple state index +and last as the range that was erased and return the entry that existed +at that location. + +You can walk each entry within a range by using mas_for_each(). If you want +to walk each element of the tree then ``0`` and ``ULONG_MAX`` may be used as +the range. If the lock needs to be periodically dropped, see the locking +section mas_pause(). + +Using a maple state allows mas_next() and mas_prev() to function as if the +tree was a linked list. With such a high branching factor the amortized +performance penalty is outweighed by cache optimization. mas_next() will +return the next entry which occurs after the entry at index. mas_prev() +will return the previous entry which occurs before the entry at index. + +mas_find() will find the first entry which exists at or above index on +the first call, and the next entry from every subsequent calls. + +mas_find_rev() will find the fist entry which exists at or below the last on +the first call, and the previous entry from every subsequent calls. + +If the user needs to yield the lock during an operation, then the maple state +must be paused using mas_pause(). + +There are a few extra interfaces provided when using an allocation tree. +If you wish to search for a gap within a range, then mas_empty_area() +or mas_empty_area_rev() can be used. mas_empty_area() searches for a gap +starting at the lowest index given up to the maximum of the range. +mas_empty_area_rev() searches for a gap starting at the highest index given +and continues downward to the lower bound of the range. + +.. _maple-tree-advanced-alloc: + +Advanced Allocating Nodes +------------------------- + +Allocations are usually handled internally to the tree, however if allocations +need to occur before a write occurs then calling mas_expected_entries() will +allocate the worst-case number of needed nodes to insert the provided number of +ranges. This also causes the tree to enter mass insertion mode. Once +insertions are complete calling mas_destroy() on the maple state will free the +unused allocations. + +.. _maple-tree-advanced-locks: + +Advanced Locking +---------------- + +The maple tree uses a spinlock by default, but external locks can be used for +tree updates as well. To use an external lock, the tree must be initialized +with the ``MT_FLAGS_LOCK_EXTERN flag``, this is usually done with the +MTREE_INIT_EXT() #define, which takes an external lock as an argument. + +Functions and structures +======================== + +.. kernel-doc:: include/linux/maple_tree.h +.. kernel-doc:: lib/maple_tree.c diff --git a/MAINTAINERS b/MAINTAINERS index 55fb1daa9057..770fab396728 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -11826,6 +11826,18 @@ L: linux-man@vger.kernel.org S: Maintained W: http://www.kernel.org/doc/man-pages +MAPLE TREE +M: Liam R. Howlett +L: linux-mm@kvack.org +S: Supported +F: Documentation/core-api/maple_tree.rst +F: include/linux/maple_tree.h +F: include/trace/events/maple_tree.h +F: lib/maple_tree.c +F: lib/test_maple_tree.c +F: tools/testing/radix-tree/linux/maple_tree.h +F: tools/testing/radix-tree/maple.c + MARDUK (CREATOR CI40) DEVICE TREE SUPPORT M: Rahul Bedarkar L: linux-mips@vger.kernel.org diff --git a/include/linux/maple_tree.h b/include/linux/maple_tree.h new file mode 100644 index 000000000000..3c6642358904 --- /dev/null +++ b/include/linux/maple_tree.h @@ -0,0 +1,685 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +#ifndef _LINUX_MAPLE_TREE_H +#define _LINUX_MAPLE_TREE_H +/* + * Maple Tree - An RCU-safe adaptive tree for storing ranges + * Copyright (c) 2018-2022 Oracle + * Authors: Liam R. Howlett + * Matthew Wilcox + */ + +#include +#include +#include +/* #define CONFIG_MAPLE_RCU_DISABLED */ +/* #define CONFIG_DEBUG_MAPLE_TREE_VERBOSE */ + +/* + * Allocated nodes are mutable until they have been inserted into the tree, + * at which time they cannot change their type until they have been removed + * from the tree and an RCU grace period has passed. + * + * Removed nodes have their ->parent set to point to themselves. RCU readers + * check ->parent before relying on the value that they loaded from the + * slots array. This lets us reuse the slots array for the RCU head. + * + * Nodes in the tree point to their parent unless bit 0 is set. + */ +#if defined(CONFIG_64BIT) || defined(BUILD_VDSO32_64) +/* 64bit sizes */ +#define MAPLE_NODE_SLOTS 31 /* 256 bytes including ->parent */ +#define MAPLE_RANGE64_SLOTS 16 /* 256 bytes */ +#define MAPLE_ARANGE64_SLOTS 10 /* 240 bytes */ +#define MAPLE_ARANGE64_META_MAX 15 /* Out of range for metadata */ +#define MAPLE_ALLOC_SLOTS (MAPLE_NODE_SLOTS - 1) +#else +/* 32bit sizes */ +#define MAPLE_NODE_SLOTS 63 /* 256 bytes including ->parent */ +#define MAPLE_RANGE64_SLOTS 32 /* 256 bytes */ +#define MAPLE_ARANGE64_SLOTS 21 /* 240 bytes */ +#define MAPLE_ARANGE64_META_MAX 31 /* Out of range for metadata */ +#define MAPLE_ALLOC_SLOTS (MAPLE_NODE_SLOTS - 2) +#endif /* defined(CONFIG_64BIT) || defined(BUILD_VDSO32_64) */ + +#define MAPLE_NODE_MASK 255UL + +/* + * The node->parent of the root node has bit 0 set and the rest of the pointer + * is a pointer to the tree itself. No more bits are available in this pointer + * (on m68k, the data structure may only be 2-byte aligned). + * + * Internal non-root nodes can only have maple_range_* nodes as parents. The + * parent pointer is 256B aligned like all other tree nodes. When storing a 32 + * or 64 bit values, the offset can fit into 4 bits. The 16 bit values need an + * extra bit to store the offset. This extra bit comes from a reuse of the last + * bit in the node type. This is possible by using bit 1 to indicate if bit 2 + * is part of the type or the slot. + * + * Once the type is decided, the decision of an allocation range type or a range + * type is done by examining the immutable tree flag for the MAPLE_ALLOC_RANGE + * flag. + * + * Node types: + * 0x??1 = Root + * 0x?00 = 16 bit nodes + * 0x010 = 32 bit nodes + * 0x110 = 64 bit nodes + * + * Slot size and location in the parent pointer: + * type : slot location + * 0x??1 : Root + * 0x?00 : 16 bit values, type in 0-1, slot in 2-6 + * 0x010 : 32 bit values, type in 0-2, slot in 3-6 + * 0x110 : 64 bit values, type in 0-2, slot in 3-6 + */ +typedef struct maple_enode *maple_enode; /* encoded node */ +typedef struct maple_pnode *maple_pnode; /* parent node */ + +/* + * This metadata is used to optimize the gap updating code and in reverse + * searching for gaps or any other code that needs to find the end of the data. + */ +struct maple_metadata { + unsigned char end; + unsigned char gap; +}; + +/* + * Leaf nodes do not store pointers to nodes, they store user data. Users may + * store almost any bit pattern. As noted above, the optimisation of storing an + * entry at 0 in the root pointer cannot be done for data which have the bottom + * two bits set to '10'. We also reserve values with the bottom two bits set to + * '10' which are below 4096 (ie 2, 6, 10 .. 4094) for internal use. Some APIs + * return errnos as a negative errno shifted right by two bits and the bottom + * two bits set to '10', and while choosing to store these values in the array + * is not an error, it may lead to confusion if you're testing for an error with + * mas_is_err(). + * + * Non-leaf nodes store the type of the node pointed to (enum maple_type in bits + * 3-6), bit 2 is reserved. That leaves bits 0-1 unused for now. + * + * In regular B-Tree terms, pivots are called keys. The term pivot is used to + * indicate that the tree is specifying ranges, Pivots may appear in the + * subtree with an entry attached to the value whereas keys are unique to a + * specific position of a B-tree. Pivot values are inclusive of the slot with + * the same index. + */ + +struct maple_range_64 { + struct maple_pnode *parent; + unsigned long pivot[MAPLE_RANGE64_SLOTS - 1]; + union { + void __rcu *slot[MAPLE_RANGE64_SLOTS]; + struct { + void __rcu *pad[MAPLE_RANGE64_SLOTS - 1]; + struct maple_metadata meta; + }; + }; +}; + +/* + * At tree creation time, the user can specify that they're willing to trade off + * storing fewer entries in a tree in return for storing more information in + * each node. + * + * The maple tree supports recording the largest range of NULL entries available + * in this node, also called gaps. This optimises the tree for allocating a + * range. + */ +struct maple_arange_64 { + struct maple_pnode *parent; + unsigned long pivot[MAPLE_ARANGE64_SLOTS - 1]; + void __rcu *slot[MAPLE_ARANGE64_SLOTS]; + unsigned long gap[MAPLE_ARANGE64_SLOTS]; + struct maple_metadata meta; +}; + +struct maple_alloc { + unsigned long total; + unsigned char node_count; + unsigned int request_count; + struct maple_alloc *slot[MAPLE_ALLOC_SLOTS]; +}; + +struct maple_topiary { + struct maple_pnode *parent; + struct maple_enode *next; /* Overlaps the pivot */ +}; + +enum maple_type { + maple_dense, + maple_leaf_64, + maple_range_64, + maple_arange_64, +}; + + +/** + * DOC: Maple tree flags + * + * * MT_FLAGS_ALLOC_RANGE - Track gaps in this tree + * * MT_FLAGS_USE_RCU - Operate in RCU mode + * * MT_FLAGS_HEIGHT_OFFSET - The position of the tree height in the flags + * * MT_FLAGS_HEIGHT_MASK - The mask for the maple tree height value + * * MT_FLAGS_LOCK_MASK - How the mt_lock is used + * * MT_FLAGS_LOCK_IRQ - Acquired irq-safe + * * MT_FLAGS_LOCK_BH - Acquired bh-safe + * * MT_FLAGS_LOCK_EXTERN - mt_lock is not used + * + * MAPLE_HEIGHT_MAX The largest height that can be stored + */ +#define MT_FLAGS_ALLOC_RANGE 0x01 +#define MT_FLAGS_USE_RCU 0x02 +#define MT_FLAGS_HEIGHT_OFFSET 0x02 +#define MT_FLAGS_HEIGHT_MASK 0x7C +#define MT_FLAGS_LOCK_MASK 0x300 +#define MT_FLAGS_LOCK_IRQ 0x100 +#define MT_FLAGS_LOCK_BH 0x200 +#define MT_FLAGS_LOCK_EXTERN 0x300 + +#define MAPLE_HEIGHT_MAX 31 + + +#define MAPLE_NODE_TYPE_MASK 0x0F +#define MAPLE_NODE_TYPE_SHIFT 0x03 + +#define MAPLE_RESERVED_RANGE 4096 + +#ifdef CONFIG_LOCKDEP +typedef struct lockdep_map *lockdep_map_p; +#define mt_lock_is_held(mt) lock_is_held(mt->ma_external_lock) +#define mt_set_external_lock(mt, lock) \ + (mt)->ma_external_lock = &(lock)->dep_map +#else +typedef struct { /* nothing */ } lockdep_map_p; +#define mt_lock_is_held(mt) 1 +#define mt_set_external_lock(mt, lock) do { } while (0) +#endif + +/* + * If the tree contains a single entry at index 0, it is usually stored in + * tree->ma_root. To optimise for the page cache, an entry which ends in '00', + * '01' or '11' is stored in the root, but an entry which ends in '10' will be + * stored in a node. Bits 3-6 are used to store enum maple_type. + * + * The flags are used both to store some immutable information about this tree + * (set at tree creation time) and dynamic information set under the spinlock. + * + * Another use of flags are to indicate global states of the tree. This is the + * case with the MAPLE_USE_RCU flag, which indicates the tree is currently in + * RCU mode. This mode was added to allow the tree to reuse nodes instead of + * re-allocating and RCU freeing nodes when there is a single user. + */ +struct maple_tree { + union { + spinlock_t ma_lock; + lockdep_map_p ma_external_lock; + }; + void __rcu *ma_root; + unsigned int ma_flags; +}; + +/** + * MTREE_INIT() - Initialize a maple tree + * @name: The maple tree name + * @flags: The maple tree flags + * + */ +#define MTREE_INIT(name, flags) { \ + .ma_lock = __SPIN_LOCK_UNLOCKED(name.ma_lock), \ + .ma_flags = flags, \ + .ma_root = NULL, \ +} + +/** + * MTREE_INIT_EXT() - Initialize a maple tree with an external lock. + * @name: The tree name + * @flags: The maple tree flags + * @lock: The external lock + */ +#ifdef CONFIG_LOCKDEP +#define MTREE_INIT_EXT(name, flags, lock) { \ + .ma_external_lock = &(lock).dep_map, \ + .ma_flags = flags, \ + .ma_root = NULL, \ +} +#else +#define MTREE_INIT_EXT(name, flags, lock) MTREE_INIT(name, flags) +#endif + +#define DEFINE_MTREE(name) \ + struct maple_tree name = MTREE_INIT(name, 0) + +#define mtree_lock(mt) spin_lock((&(mt)->ma_lock)) +#define mtree_unlock(mt) spin_unlock((&(mt)->ma_lock)) + +/* + * The Maple Tree squeezes various bits in at various points which aren't + * necessarily obvious. Usually, this is done by observing that pointers are + * N-byte aligned and thus the bottom log_2(N) bits are available for use. We + * don't use the high bits of pointers to store additional information because + * we don't know what bits are unused on any given architecture. + * + * Nodes are 256 bytes in size and are also aligned to 256 bytes, giving us 8 + * low bits for our own purposes. Nodes are currently of 4 types: + * 1. Single pointer (Range is 0-0) + * 2. Non-leaf Allocation Range nodes + * 3. Non-leaf Range nodes + * 4. Leaf Range nodes All nodes consist of a number of node slots, + * pivots, and a parent pointer. + */ + +struct maple_node { + union { + struct { + struct maple_pnode *parent; + void __rcu *slot[MAPLE_NODE_SLOTS]; + }; + struct { + void *pad; + struct rcu_head rcu; + struct maple_enode *piv_parent; + unsigned char parent_slot; + enum maple_type type; + unsigned char slot_len; + unsigned int ma_flags; + }; + struct maple_range_64 mr64; + struct maple_arange_64 ma64; + struct maple_alloc alloc; + }; +}; + +/* + * More complicated stores can cause two nodes to become one or three and + * potentially alter the height of the tree. Either half of the tree may need + * to be rebalanced against the other. The ma_topiary struct is used to track + * which nodes have been 'cut' from the tree so that the change can be done + * safely at a later date. This is done to support RCU. + */ +struct ma_topiary { + struct maple_enode *head; + struct maple_enode *tail; + struct maple_tree *mtree; +}; + +void *mtree_load(struct maple_tree *mt, unsigned long index); + +int mtree_insert(struct maple_tree *mt, unsigned long index, + void *entry, gfp_t gfp); +int mtree_insert_range(struct maple_tree *mt, unsigned long first, + unsigned long last, void *entry, gfp_t gfp); +int mtree_alloc_range(struct maple_tree *mt, unsigned long *startp, + void *entry, unsigned long size, unsigned long min, + unsigned long max, gfp_t gfp); +int mtree_alloc_rrange(struct maple_tree *mt, unsigned long *startp, + void *entry, unsigned long size, unsigned long min, + unsigned long max, gfp_t gfp); + +int mtree_store_range(struct maple_tree *mt, unsigned long first, + unsigned long last, void *entry, gfp_t gfp); +int mtree_store(struct maple_tree *mt, unsigned long index, + void *entry, gfp_t gfp); +void *mtree_erase(struct maple_tree *mt, unsigned long index); + +void mtree_destroy(struct maple_tree *mt); +void __mt_destroy(struct maple_tree *mt); + +/** + * mtree_empty() - Determine if a tree has any present entries. + * @mt: Maple Tree. + * + * Context: Any context. + * Return: %true if the tree contains only NULL pointers. + */ +static inline bool mtree_empty(const struct maple_tree *mt) +{ + return mt->ma_root == NULL; +} + +/* Advanced API */ + +/* + * The maple state is defined in the struct ma_state and is used to keep track + * of information during operations, and even between operations when using the + * advanced API. + * + * If state->node has bit 0 set then it references a tree location which is not + * a node (eg the root). If bit 1 is set, the rest of the bits are a negative + * errno. Bit 2 (the 'unallocated slots' bit) is clear. Bits 3-6 indicate the + * node type. + * + * state->alloc either has a request number of nodes or an allocated node. If + * stat->alloc has a requested number of nodes, the first bit will be set (0x1) + * and the remaining bits are the value. If state->alloc is a node, then the + * node will be of type maple_alloc. maple_alloc has MAPLE_NODE_SLOTS - 1 for + * storing more allocated nodes, a total number of nodes allocated, and the + * node_count in this node. node_count is the number of allocated nodes in this + * node. The scaling beyond MAPLE_NODE_SLOTS - 1 is handled by storing further + * nodes into state->alloc->slot[0]'s node. Nodes are taken from state->alloc + * by removing a node from the state->alloc node until state->alloc->node_count + * is 1, when state->alloc is returned and the state->alloc->slot[0] is promoted + * to state->alloc. Nodes are pushed onto state->alloc by putting the current + * state->alloc into the pushed node's slot[0]. + * + * The state also contains the implied min/max of the state->node, the depth of + * this search, and the offset. The implied min/max are either from the parent + * node or are 0-oo for the root node. The depth is incremented or decremented + * every time a node is walked down or up. The offset is the slot/pivot of + * interest in the node - either for reading or writing. + * + * When returning a value the maple state index and last respectively contain + * the start and end of the range for the entry. Ranges are inclusive in the + * Maple Tree. + */ +struct ma_state { + struct maple_tree *tree; /* The tree we're operating in */ + unsigned long index; /* The index we're operating on - range start */ + unsigned long last; /* The last index we're operating on - range end */ + struct maple_enode *node; /* The node containing this entry */ + unsigned long min; /* The minimum index of this node - implied pivot min */ + unsigned long max; /* The maximum index of this node - implied pivot max */ + struct maple_alloc *alloc; /* Allocated nodes for this operation */ + unsigned char depth; /* depth of tree descent during write */ + unsigned char offset; + unsigned char mas_flags; +}; + +struct ma_wr_state { + struct ma_state *mas; + struct maple_node *node; /* Decoded mas->node */ + unsigned long r_min; /* range min */ + unsigned long r_max; /* range max */ + enum maple_type type; /* mas->node type */ + unsigned char offset_end; /* The offset where the write ends */ + unsigned char node_end; /* mas->node end */ + unsigned long *pivots; /* mas->node->pivots pointer */ + unsigned long end_piv; /* The pivot at the offset end */ + void __rcu **slots; /* mas->node->slots pointer */ + void *entry; /* The entry to write */ + void *content; /* The existing entry that is being overwritten */ +}; + +#define mas_lock(mas) spin_lock(&((mas)->tree->ma_lock)) +#define mas_unlock(mas) spin_unlock(&((mas)->tree->ma_lock)) + + +/* + * Special values for ma_state.node. + * MAS_START means we have not searched the tree. + * MAS_ROOT means we have searched the tree and the entry we found lives in + * the root of the tree (ie it has index 0, length 1 and is the only entry in + * the tree). + * MAS_NONE means we have searched the tree and there is no node in the + * tree for this entry. For example, we searched for index 1 in an empty + * tree. Or we have a tree which points to a full leaf node and we + * searched for an entry which is larger than can be contained in that + * leaf node. + * MA_ERROR represents an errno. After dropping the lock and attempting + * to resolve the error, the walk would have to be restarted from the + * top of the tree as the tree may have been modified. + */ +#define MAS_START ((struct maple_enode *)1UL) +#define MAS_ROOT ((struct maple_enode *)5UL) +#define MAS_NONE ((struct maple_enode *)9UL) +#define MAS_PAUSE ((struct maple_enode *)17UL) +#define MA_ERROR(err) \ + ((struct maple_enode *)(((unsigned long)err << 2) | 2UL)) + +#define MA_STATE(name, mt, first, end) \ + struct ma_state name = { \ + .tree = mt, \ + .index = first, \ + .last = end, \ + .node = MAS_START, \ + .min = 0, \ + .max = ULONG_MAX, \ + } + +#define MA_WR_STATE(name, ma_state, wr_entry) \ + struct ma_wr_state name = { \ + .mas = ma_state, \ + .content = NULL, \ + .entry = wr_entry, \ + } + +#define MA_TOPIARY(name, tree) \ + struct ma_topiary name = { \ + .head = NULL, \ + .tail = NULL, \ + .mtree = tree, \ + } + +void *mas_walk(struct ma_state *mas); +void *mas_store(struct ma_state *mas, void *entry); +void *mas_erase(struct ma_state *mas); +int mas_store_gfp(struct ma_state *mas, void *entry, gfp_t gfp); +void mas_store_prealloc(struct ma_state *mas, void *entry); +void *mas_find(struct ma_state *mas, unsigned long max); +void *mas_find_rev(struct ma_state *mas, unsigned long min); +int mas_preallocate(struct ma_state *mas, void *entry, gfp_t gfp); + +bool mas_nomem(struct ma_state *mas, gfp_t gfp); +void mas_pause(struct ma_state *mas); +void maple_tree_init(void); +void mas_destroy(struct ma_state *mas); +int mas_expected_entries(struct ma_state *mas, unsigned long nr_entries); + +void *mas_prev(struct ma_state *mas, unsigned long min); +void *mas_next(struct ma_state *mas, unsigned long max); + +int mas_empty_area(struct ma_state *mas, unsigned long min, unsigned long max, + unsigned long size); + +/* Checks if a mas has not found anything */ +static inline bool mas_is_none(struct ma_state *mas) +{ + return mas->node == MAS_NONE; +} + +/* Checks if a mas has been paused */ +static inline bool mas_is_paused(struct ma_state *mas) +{ + return mas->node == MAS_PAUSE; +} + +void mas_dup_tree(struct ma_state *oldmas, struct ma_state *mas); +void mas_dup_store(struct ma_state *mas, void *entry); + +/* + * This finds an empty area from the highest address to the lowest. + * AKA "Topdown" version, + */ +int mas_empty_area_rev(struct ma_state *mas, unsigned long min, + unsigned long max, unsigned long size); +/** + * mas_reset() - Reset a Maple Tree operation state. + * @mas: Maple Tree operation state. + * + * Resets the error or walk state of the @mas so future walks of the + * array will start from the root. Use this if you have dropped the + * lock and want to reuse the ma_state. + * + * Context: Any context. + */ +static inline void mas_reset(struct ma_state *mas) +{ + mas->node = MAS_START; +} + +/** + * mas_for_each() - Iterate over a range of the maple tree. + * @mas: Maple Tree operation state (maple_state) + * @entry: Entry retrieved from the tree + * @max: maximum index to retrieve from the tree + * + * When returned, mas->index and mas->last will hold the entire range for the + * entry. + * + * Note: may return the zero entry. + * + */ +#define mas_for_each(mas, entry, max) \ + while (((entry) = mas_find((mas), (max))) != NULL) + + +/** + * mas_set_range() - Set up Maple Tree operation state for a different index. + * @mas: Maple Tree operation state. + * @start: New start of range in the Maple Tree. + * @last: New end of range in the Maple Tree. + * + * Move the operation state to refer to a different range. This will + * have the effect of starting a walk from the top; see mas_next() + * to move to an adjacent index. + */ +static inline +void mas_set_range(struct ma_state *mas, unsigned long start, unsigned long last) +{ + mas->index = start; + mas->last = last; + mas->node = MAS_START; +} + +/** + * mas_set() - Set up Maple Tree operation state for a different index. + * @mas: Maple Tree operation state. + * @index: New index into the Maple Tree. + * + * Move the operation state to refer to a different index. This will + * have the effect of starting a walk from the top; see mas_next() + * to move to an adjacent index. + */ +static inline void mas_set(struct ma_state *mas, unsigned long index) +{ + + mas_set_range(mas, index, index); +} + +static inline bool mt_external_lock(const struct maple_tree *mt) +{ + return (mt->ma_flags & MT_FLAGS_LOCK_MASK) == MT_FLAGS_LOCK_EXTERN; +} + +/** + * mt_init_flags() - Initialise an empty maple tree with flags. + * @mt: Maple Tree + * @flags: maple tree flags. + * + * If you need to initialise a Maple Tree with special flags (eg, an + * allocation tree), use this function. + * + * Context: Any context. + */ +static inline void mt_init_flags(struct maple_tree *mt, unsigned int flags) +{ + mt->ma_flags = flags; + if (!mt_external_lock(mt)) + spin_lock_init(&mt->ma_lock); + rcu_assign_pointer(mt->ma_root, NULL); +} + +/** + * mt_init() - Initialise an empty maple tree. + * @mt: Maple Tree + * + * An empty Maple Tree. + * + * Context: Any context. + */ +static inline void mt_init(struct maple_tree *mt) +{ + mt_init_flags(mt, 0); +} + +static inline bool mt_in_rcu(struct maple_tree *mt) +{ +#ifdef CONFIG_MAPLE_RCU_DISABLED + return false; +#endif + return mt->ma_flags & MT_FLAGS_USE_RCU; +} + +/** + * mt_clear_in_rcu() - Switch the tree to non-RCU mode. + * @mt: The Maple Tree + */ +static inline void mt_clear_in_rcu(struct maple_tree *mt) +{ + if (!mt_in_rcu(mt)) + return; + + if (mt_external_lock(mt)) { + BUG_ON(!mt_lock_is_held(mt)); + mt->ma_flags &= ~MT_FLAGS_USE_RCU; + } else { + mtree_lock(mt); + mt->ma_flags &= ~MT_FLAGS_USE_RCU; + mtree_unlock(mt); + } +} + +/** + * mt_set_in_rcu() - Switch the tree to RCU safe mode. + * @mt: The Maple Tree + */ +static inline void mt_set_in_rcu(struct maple_tree *mt) +{ + if (mt_in_rcu(mt)) + return; + + if (mt_external_lock(mt)) { + BUG_ON(!mt_lock_is_held(mt)); + mt->ma_flags |= MT_FLAGS_USE_RCU; + } else { + mtree_lock(mt); + mt->ma_flags |= MT_FLAGS_USE_RCU; + mtree_unlock(mt); + } +} + +void *mt_find(struct maple_tree *mt, unsigned long *index, unsigned long max); +void *mt_find_after(struct maple_tree *mt, unsigned long *index, + unsigned long max); +void *mt_prev(struct maple_tree *mt, unsigned long index, unsigned long min); +void *mt_next(struct maple_tree *mt, unsigned long index, unsigned long max); + +/** + * mt_for_each - Iterate over each entry starting at index until max. + * @tree: The Maple Tree + * @entry: The current entry + * @index: The index to update to track the location in the tree + * @max: The maximum limit for @index + * + * Note: Will not return the zero entry. + */ +#define mt_for_each(tree, entry, index, max) \ + for (entry = mt_find(tree, &(index), max); \ + entry; entry = mt_find_after(tree, &(index), max)) + + +#ifdef CONFIG_DEBUG_MAPLE_TREE +extern atomic_t maple_tree_tests_run; +extern atomic_t maple_tree_tests_passed; + +void mt_dump(const struct maple_tree *mt); +void mt_validate(struct maple_tree *mt); +#define MT_BUG_ON(tree, x) do { \ + atomic_inc(&maple_tree_tests_run); \ + if (x) { \ + pr_info("BUG at %s:%d (%u)\n", \ + __func__, __LINE__, x); \ + mt_dump(tree); \ + pr_info("Pass: %u Run:%u\n", \ + atomic_read(&maple_tree_tests_passed), \ + atomic_read(&maple_tree_tests_run)); \ + dump_stack(); \ + } else { \ + atomic_inc(&maple_tree_tests_passed); \ + } \ +} while (0) +#else +#define MT_BUG_ON(tree, x) BUG_ON(x) +#endif /* CONFIG_DEBUG_MAPLE_TREE */ + +#endif /*_LINUX_MAPLE_TREE_H */ diff --git a/include/trace/events/maple_tree.h b/include/trace/events/maple_tree.h new file mode 100644 index 000000000000..2be403bdc2bd --- /dev/null +++ b/include/trace/events/maple_tree.h @@ -0,0 +1,123 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM maple_tree + +#if !defined(_TRACE_MM_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_MM_H + + +#include + +struct ma_state; + +TRACE_EVENT(ma_op, + + TP_PROTO(const char *fn, struct ma_state *mas), + + TP_ARGS(fn, mas), + + TP_STRUCT__entry( + __field(const char *, fn) + __field(unsigned long, min) + __field(unsigned long, max) + __field(unsigned long, index) + __field(unsigned long, last) + __field(void *, node) + ), + + TP_fast_assign( + __entry->fn = fn; + __entry->min = mas->min; + __entry->max = mas->max; + __entry->index = mas->index; + __entry->last = mas->last; + __entry->node = mas->node; + ), + + TP_printk("%s\tNode: %p (%lu %lu) range: %lu-%lu", + __entry->fn, + (void *) __entry->node, + (unsigned long) __entry->min, + (unsigned long) __entry->max, + (unsigned long) __entry->index, + (unsigned long) __entry->last + ) +) +TRACE_EVENT(ma_read, + + TP_PROTO(const char *fn, struct ma_state *mas), + + TP_ARGS(fn, mas), + + TP_STRUCT__entry( + __field(const char *, fn) + __field(unsigned long, min) + __field(unsigned long, max) + __field(unsigned long, index) + __field(unsigned long, last) + __field(void *, node) + ), + + TP_fast_assign( + __entry->fn = fn; + __entry->min = mas->min; + __entry->max = mas->max; + __entry->index = mas->index; + __entry->last = mas->last; + __entry->node = mas->node; + ), + + TP_printk("%s\tNode: %p (%lu %lu) range: %lu-%lu", + __entry->fn, + (void *) __entry->node, + (unsigned long) __entry->min, + (unsigned long) __entry->max, + (unsigned long) __entry->index, + (unsigned long) __entry->last + ) +) + +TRACE_EVENT(ma_write, + + TP_PROTO(const char *fn, struct ma_state *mas, unsigned long piv, + void *val), + + TP_ARGS(fn, mas, piv, val), + + TP_STRUCT__entry( + __field(const char *, fn) + __field(unsigned long, min) + __field(unsigned long, max) + __field(unsigned long, index) + __field(unsigned long, last) + __field(unsigned long, piv) + __field(void *, val) + __field(void *, node) + ), + + TP_fast_assign( + __entry->fn = fn; + __entry->min = mas->min; + __entry->max = mas->max; + __entry->index = mas->index; + __entry->last = mas->last; + __entry->piv = piv; + __entry->val = val; + __entry->node = mas->node; + ), + + TP_printk("%s\tNode %p (%lu %lu) range:%lu-%lu piv (%lu) val %p", + __entry->fn, + (void *) __entry->node, + (unsigned long) __entry->min, + (unsigned long) __entry->max, + (unsigned long) __entry->index, + (unsigned long) __entry->last, + (unsigned long) __entry->piv, + (void *) __entry->val + ) +) +#endif /* _TRACE_MM_H */ + +/* This part must be outside protection */ +#include diff --git a/init/main.c b/init/main.c index 0ee39cdcfcac..1edb1bfec640 100644 --- a/init/main.c +++ b/init/main.c @@ -116,6 +116,7 @@ static int kernel_init(void *); extern void init_IRQ(void); extern void radix_tree_init(void); +extern void maple_tree_init(void); /* * Debug helper: via this flag we know that we are in 'early bootup code' @@ -1001,6 +1002,7 @@ asmlinkage __visible void __init __no_sanitize_address start_kernel(void) "Interrupts were enabled *very* early, fixing it\n")) local_irq_disable(); radix_tree_init(); + maple_tree_init(); /* * Set up housekeeping before setting up workqueues to allow the unbound diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 2e24db4bff19..05c3858cdf41 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -812,6 +812,13 @@ config DEBUG_VM_VMACACHE can cause significant overhead, so only enable it in non-production environments. +config DEBUG_VM_MAPLE_TREE + bool "Debug VM maple trees" + depends on DEBUG_VM + select DEBUG_MAPLE_TREE + help + Enable VM maple tree debugging information and extra validations. + If unsure, say N. config DEBUG_VM_RB @@ -1627,6 +1634,14 @@ config BUG_ON_DATA_CORRUPTION If unsure, say N. +config DEBUG_MAPLE_TREE + bool "Debug maple trees" + depends on DEBUG_KERNEL + help + Enable maple tree debugging information and extra validations. + + If unsure, say N. + endmenu config DEBUG_CREDENTIALS diff --git a/lib/Makefile b/lib/Makefile index f99bf61f8bbc..2b845a5e4a80 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -29,7 +29,8 @@ endif lib-y := ctype.o string.o vsprintf.o cmdline.o \ rbtree.o radix-tree.o timerqueue.o xarray.o \ - idr.o extable.o sha1.o irq_regs.o argv_split.o \ + maple_tree.o idr.o extable.o \ + sha1.o irq_regs.o argv_split.o \ flex_proportions.o ratelimit.o show_mem.o \ is_single_threaded.o plist.o decompress.o kobject_uevent.o \ earlycpio.o seq_buf.o siphash.o dec_and_lock.o \ diff --git a/lib/maple_tree.c b/lib/maple_tree.c new file mode 100644 index 000000000000..a968948b3e3e --- /dev/null +++ b/lib/maple_tree.c @@ -0,0 +1,7041 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * Maple Tree implementation + * Copyright (c) 2018-2022 Oracle Corporation + * Authors: Liam R. Howlett + * Matthew Wilcox + */ + +/* + * DOC: Interesting implementation details of the Maple Tree + * + * Each node type has a number of slots for entries and a number of slots for + * pivots. In the case of dense nodes, the pivots are implied by the position + * and are simply the slot index + the minimum of the node. + * + * In regular B-Tree terms, pivots are called keys. The term pivot is used to + * indicate that the tree is specifying ranges, Pivots may appear in the + * subtree with an entry attached to the value where as keys are unique to a + * specific position of a B-tree. Pivot values are inclusive of the slot with + * the same index. + * + * + * The following illustrates the layout of a range64 nodes slots and pivots. + * + * + * Slots -> | 0 | 1 | 2 | ... | 12 | 13 | 14 | 15 | + * ┬ ┬ ┬ ┬ ┬ ┬ ┬ ┬ ┬ + * │ │ │ │ │ │ │ │ └─ Implied maximum + * │ │ │ │ │ │ │ └─ Pivot 14 + * │ │ │ │ │ │ └─ Pivot 13 + * │ │ │ │ │ └─ Pivot 12 + * │ │ │ │ └─ Pivot 11 + * │ │ │ └─ Pivot 2 + * │ │ └─ Pivot 1 + * │ └─ Pivot 0 + * └─ Implied minimum + * + * Slot contents: + * Internal (non-leaf) nodes contain pointers to other nodes. + * Leaf nodes contain entries. + * + * The location of interest is often referred to as an offset. All offsets have + * a slot, but the last offset has an implied pivot from the node above (or + * UINT_MAX for the root node. + * + * Ranges complicate certain write activities. When modifying any of + * the B-tree variants, it is known that one entry will either be added or + * deleted. When modifying the Maple Tree, one store operation may overwrite + * the entire data set, or one half of the tree, or the middle half of the tree. + * + */ + + +#include +#include +#include +#include +#include +#include +#include + +#define CREATE_TRACE_POINTS +#include + +#define MA_ROOT_PARENT 1 + +/* Maple state flags */ +#define MA_STATE_BULK 1 +#define MA_STATE_REBALANCE 2 + +#define ma_parent_ptr(x) ((struct maple_pnode *)(x)) +#define ma_mnode_ptr(x) ((struct maple_node *)(x)) +#define ma_enode_ptr(x) ((struct maple_enode *)(x)) +static struct kmem_cache *maple_node_cache; + +#ifdef CONFIG_DEBUG_MAPLE_TREE +static const unsigned long mt_max[] = { + [maple_dense] = MAPLE_NODE_SLOTS, + [maple_leaf_64] = ULONG_MAX, + [maple_range_64] = ULONG_MAX, + [maple_arange_64] = ULONG_MAX, +}; +#define mt_node_max(x) mt_max[mte_node_type(x)] +#endif + +static const unsigned char mt_slots[] = { + [maple_dense] = MAPLE_NODE_SLOTS, + [maple_leaf_64] = MAPLE_RANGE64_SLOTS, + [maple_range_64] = MAPLE_RANGE64_SLOTS, + [maple_arange_64] = MAPLE_ARANGE64_SLOTS, +}; +#define mt_slot_count(x) mt_slots[mte_node_type(x)] + +static const unsigned char mt_pivots[] = { + [maple_dense] = 0, + [maple_leaf_64] = MAPLE_RANGE64_SLOTS - 1, + [maple_range_64] = MAPLE_RANGE64_SLOTS - 1, + [maple_arange_64] = MAPLE_ARANGE64_SLOTS - 1, +}; +#define mt_pivot_count(x) mt_pivots[mte_node_type(x)] + +static const unsigned char mt_min_slots[] = { + [maple_dense] = MAPLE_NODE_SLOTS / 2, + [maple_leaf_64] = (MAPLE_RANGE64_SLOTS / 2) - 2, + [maple_range_64] = (MAPLE_RANGE64_SLOTS / 2) - 2, + [maple_arange_64] = (MAPLE_ARANGE64_SLOTS / 2) - 1, +}; +#define mt_min_slot_count(x) mt_min_slots[mte_node_type(x)] + +#define MAPLE_BIG_NODE_SLOTS (MAPLE_RANGE64_SLOTS * 2 + 2) +#define MAPLE_BIG_NODE_GAPS (MAPLE_ARANGE64_SLOTS * 2 + 1) + +struct maple_big_node { + struct maple_pnode *parent; + unsigned long pivot[MAPLE_BIG_NODE_SLOTS - 1]; + union { + struct maple_enode *slot[MAPLE_BIG_NODE_SLOTS]; + struct { + unsigned long padding[MAPLE_BIG_NODE_GAPS]; + unsigned long gap[MAPLE_BIG_NODE_GAPS]; + }; + }; + unsigned char b_end; + enum maple_type type; +}; + +/* + * The maple_subtree_state is used to build a tree to replace a segment of an + * existing tree in a more atomic way. Any walkers of the older tree will hit a + * dead node and restart on updates. + */ +struct maple_subtree_state { + struct ma_state *orig_l; /* Original left side of subtree */ + struct ma_state *orig_r; /* Original right side of subtree */ + struct ma_state *l; /* New left side of subtree */ + struct ma_state *m; /* New middle of subtree (rare) */ + struct ma_state *r; /* New right side of subtree */ + struct ma_topiary *free; /* nodes to be freed */ + struct ma_topiary *destroy; /* Nodes to be destroyed (walked and freed) */ + struct maple_big_node *bn; +}; + +/* Functions */ +static inline struct maple_node *mt_alloc_one(gfp_t gfp) +{ + return kmem_cache_alloc(maple_node_cache, gfp | __GFP_ZERO); +} + +static inline int mt_alloc_bulk(gfp_t gfp, size_t size, void **nodes) +{ + return kmem_cache_alloc_bulk(maple_node_cache, gfp | __GFP_ZERO, size, + nodes); +} + +static inline void mt_free_bulk(size_t size, void __rcu **nodes) +{ + kmem_cache_free_bulk(maple_node_cache, size, (void **)nodes); +} + +static void mt_free_rcu(struct rcu_head *head) +{ + struct maple_node *node = container_of(head, struct maple_node, rcu); + + kmem_cache_free(maple_node_cache, node); +} + +/* + * ma_free_rcu() - Use rcu callback to free a maple node + * @node: The node to free + * + * The maple tree uses the parent pointer to indicate this node is no longer in + * use and will be freed. + */ +static void ma_free_rcu(struct maple_node *node) +{ + node->parent = ma_parent_ptr(node); + call_rcu(&node->rcu, mt_free_rcu); +} + +static unsigned int mt_height(const struct maple_tree *mt) +{ + return (mt->ma_flags & MT_FLAGS_HEIGHT_MASK) >> MT_FLAGS_HEIGHT_OFFSET; +} + +static void mas_set_height(struct ma_state *mas) +{ + unsigned int new_flags = mas->tree->ma_flags; + + new_flags &= ~MT_FLAGS_HEIGHT_MASK; + BUG_ON(mas->depth > MAPLE_HEIGHT_MAX); + new_flags |= mas->depth << MT_FLAGS_HEIGHT_OFFSET; + mas->tree->ma_flags = new_flags; +} + +static unsigned int mas_mt_height(struct ma_state *mas) +{ + return mt_height(mas->tree); +} + +static inline enum maple_type mte_node_type(const struct maple_enode *entry) +{ + return ((unsigned long)entry >> MAPLE_NODE_TYPE_SHIFT) & + MAPLE_NODE_TYPE_MASK; +} + +static inline bool ma_is_dense(const enum maple_type type) +{ + return type < maple_leaf_64; +} + +static inline bool ma_is_leaf(const enum maple_type type) +{ + return type < maple_range_64; +} + +static inline bool mte_is_leaf(const struct maple_enode *entry) +{ + return ma_is_leaf(mte_node_type(entry)); +} + +/* + * We also reserve values with the bottom two bits set to '10' which are + * below 4096 + */ +static inline bool mt_is_reserved(const void *entry) +{ + return ((unsigned long)entry < MAPLE_RESERVED_RANGE) && + xa_is_internal(entry); +} + +static inline void mas_set_err(struct ma_state *mas, long err) +{ + mas->node = MA_ERROR(err); +} + +static inline bool mas_is_ptr(struct ma_state *mas) +{ + return mas->node == MAS_ROOT; +} + +static inline bool mas_is_start(struct ma_state *mas) +{ + return mas->node == MAS_START; +} + +static inline bool mas_is_err(struct ma_state *mas) +{ + return xa_is_err(mas->node); +} + +static inline bool mas_searchable(struct ma_state *mas) +{ + if (mas_is_none(mas)) + return false; + + if (mas_is_ptr(mas)) + return false; + + return true; +} + +static inline struct maple_node *mte_to_node(const struct maple_enode *entry) +{ + return (struct maple_node *)((unsigned long)entry & ~MAPLE_NODE_MASK); +} + +/* + * mte_to_mat() - Convert a maple encoded node to a maple topiary node. + * @entry: The maple encoded node + * + * Return: a maple topiary pointer + */ +static inline struct maple_topiary *mte_to_mat(const struct maple_enode *entry) +{ + return (struct maple_topiary *) + ((unsigned long)entry & ~MAPLE_NODE_MASK); +} + +/* + * mas_mn() - Get the maple state node. + * @mas: The maple state + * + * Return: the maple node (not encoded - bare pointer). + */ +static inline struct maple_node *mas_mn(const struct ma_state *mas) +{ + return mte_to_node(mas->node); +} + +/* + * mte_set_node_dead() - Set a maple encoded node as dead. + * @mn: The maple encoded node. + */ +static inline void mte_set_node_dead(struct maple_enode *mn) +{ + mte_to_node(mn)->parent = ma_parent_ptr(mte_to_node(mn)); + smp_wmb(); /* Needed for RCU */ +} + +/* Bit 1 indicates the root is a node */ +#define MAPLE_ROOT_NODE 0x02 +/* maple_type stored bit 3-6 */ +#define MAPLE_ENODE_TYPE_SHIFT 0x03 +/* Bit 2 means a NULL somewhere below */ +#define MAPLE_ENODE_NULL 0x04 + +static inline struct maple_enode *mt_mk_node(const struct maple_node *node, + enum maple_type type) +{ + return (void *)((unsigned long)node | + (type << MAPLE_ENODE_TYPE_SHIFT) | MAPLE_ENODE_NULL); +} + +static inline void *mte_mk_root(const struct maple_enode *node) +{ + return (void *)((unsigned long)node | MAPLE_ROOT_NODE); +} + +static inline void *mte_safe_root(const struct maple_enode *node) +{ + return (void *)((unsigned long)node & ~MAPLE_ROOT_NODE); +} + +static inline void mte_set_full(const struct maple_enode *node) +{ + node = (void *)((unsigned long)node & ~MAPLE_ENODE_NULL); +} + +static inline void mte_clear_full(const struct maple_enode *node) +{ + node = (void *)((unsigned long)node | MAPLE_ENODE_NULL); +} + +static inline bool ma_is_root(struct maple_node *node) +{ + return ((unsigned long)node->parent & MA_ROOT_PARENT); +} + +static inline bool mte_is_root(const struct maple_enode *node) +{ + return ma_is_root(mte_to_node(node)); +} + +static inline bool mas_is_root_limits(const struct ma_state *mas) +{ + return !mas->min && mas->max == ULONG_MAX; +} + +static inline bool mt_is_alloc(struct maple_tree *mt) +{ + return (mt->ma_flags & MT_FLAGS_ALLOC_RANGE); +} + +/* + * The Parent Pointer + * Excluding root, the parent pointer is 256B aligned like all other tree nodes. + * When storing a 32 or 64 bit values, the offset can fit into 5 bits. The 16 + * bit values need an extra bit to store the offset. This extra bit comes from + * a reuse of the last bit in the node type. This is possible by using bit 1 to + * indicate if bit 2 is part of the type or the slot. + * + * Note types: + * 0x??1 = Root + * 0x?00 = 16 bit nodes + * 0x010 = 32 bit nodes + * 0x110 = 64 bit nodes + * + * Slot size and alignment + * 0b??1 : Root + * 0b?00 : 16 bit values, type in 0-1, slot in 2-7 + * 0b010 : 32 bit values, type in 0-2, slot in 3-7 + * 0b110 : 64 bit values, type in 0-2, slot in 3-7 + */ + +#define MAPLE_PARENT_ROOT 0x01 + +#define MAPLE_PARENT_SLOT_SHIFT 0x03 +#define MAPLE_PARENT_SLOT_MASK 0xF8 + +#define MAPLE_PARENT_16B_SLOT_SHIFT 0x02 +#define MAPLE_PARENT_16B_SLOT_MASK 0xFC + +#define MAPLE_PARENT_RANGE64 0x06 +#define MAPLE_PARENT_RANGE32 0x04 +#define MAPLE_PARENT_NOT_RANGE16 0x02 + +/* + * mte_parent_shift() - Get the parent shift for the slot storage. + * @parent: The parent pointer cast as an unsigned long + * Return: The shift into that pointer to the star to of the slot + */ +static inline unsigned long mte_parent_shift(unsigned long parent) +{ + /* Note bit 1 == 0 means 16B */ + if (likely(parent & MAPLE_PARENT_NOT_RANGE16)) + return MAPLE_PARENT_SLOT_SHIFT; + + return MAPLE_PARENT_16B_SLOT_SHIFT; +} + +/* + * mte_parent_slot_mask() - Get the slot mask for the parent. + * @parent: The parent pointer cast as an unsigned long. + * Return: The slot mask for that parent. + */ +static inline unsigned long mte_parent_slot_mask(unsigned long parent) +{ + /* Note bit 1 == 0 means 16B */ + if (likely(parent & MAPLE_PARENT_NOT_RANGE16)) + return MAPLE_PARENT_SLOT_MASK; + + return MAPLE_PARENT_16B_SLOT_MASK; +} + +/* + * mas_parent_enum() - Return the maple_type of the parent from the stored + * parent type. + * @mas: The maple state + * @node: The maple_enode to extract the parent's enum + * Return: The node->parent maple_type + */ +static inline +enum maple_type mte_parent_enum(struct maple_enode *p_enode, + struct maple_tree *mt) +{ + unsigned long p_type; + + p_type = (unsigned long)p_enode; + if (p_type & MAPLE_PARENT_ROOT) + return 0; /* Validated in the caller. */ + + p_type &= MAPLE_NODE_MASK; + p_type = p_type & ~(MAPLE_PARENT_ROOT | mte_parent_slot_mask(p_type)); + + switch (p_type) { + case MAPLE_PARENT_RANGE64: /* or MAPLE_PARENT_ARANGE64 */ + if (mt_is_alloc(mt)) + return maple_arange_64; + return maple_range_64; + } + + return 0; +} + +static inline +enum maple_type mas_parent_enum(struct ma_state *mas, struct maple_enode *enode) +{ + return mte_parent_enum(ma_enode_ptr(mte_to_node(enode)->parent), mas->tree); +} + +/* + * mte_set_parent() - Set the parent node and encode the slot + * @enode: The encoded maple node. + * @parent: The encoded maple node that is the parent of @enode. + * @slot: The slot that @enode resides in @parent. + * + * Slot number is encoded in the enode->parent bit 3-6 or 2-6, depending on the + * parent type. + */ +static inline +void mte_set_parent(struct maple_enode *enode, const struct maple_enode *parent, + unsigned char slot) +{ + unsigned long val = (unsigned long) parent; + unsigned long shift; + unsigned long type; + enum maple_type p_type = mte_node_type(parent); + + BUG_ON(p_type == maple_dense); + BUG_ON(p_type == maple_leaf_64); + + switch (p_type) { + case maple_range_64: + case maple_arange_64: + shift = MAPLE_PARENT_SLOT_SHIFT; + type = MAPLE_PARENT_RANGE64; + break; + default: + case maple_dense: + case maple_leaf_64: + shift = type = 0; + break; + } + + val &= ~MAPLE_NODE_MASK; /* Clear all node metadata in parent */ + val |= (slot << shift) | type; + mte_to_node(enode)->parent = ma_parent_ptr(val); +} + +/* + * mte_parent_slot() - get the parent slot of @enode. + * @enode: The encoded maple node. + * + * Return: The slot in the parent node where @enode resides. + */ +static inline unsigned int mte_parent_slot(const struct maple_enode *enode) +{ + unsigned long val = (unsigned long) mte_to_node(enode)->parent; + + /* Root. */ + if (val & 1) + return 0; + + /* + * Okay to use MAPLE_PARENT_16B_SLOT_MASK as the last bit will be lost + * by shift if the parent shift is MAPLE_PARENT_SLOT_SHIFT + */ + return (val & MAPLE_PARENT_16B_SLOT_MASK) >> mte_parent_shift(val); +} + +/* + * mte_parent() - Get the parent of @node. + * @node: The encoded maple node. + * + * Return: The parent maple node. + */ +static inline struct maple_node *mte_parent(const struct maple_enode *enode) +{ + return (void *)((unsigned long) + (mte_to_node(enode)->parent) & ~MAPLE_NODE_MASK); +} + +/* + * ma_dead_node() - check if the @enode is dead. + * @enode: The encoded maple node + * + * Return: true if dead, false otherwise. + */ +static inline bool ma_dead_node(const struct maple_node *node) +{ + struct maple_node *parent = (void *)((unsigned long) + node->parent & ~MAPLE_NODE_MASK); + + return (parent == node); +} +/* + * mte_dead_node() - check if the @enode is dead. + * @enode: The encoded maple node + * + * Return: true if dead, false otherwise. + */ +static inline bool mte_dead_node(const struct maple_enode *enode) +{ + struct maple_node *parent, *node; + + node = mte_to_node(enode); + parent = mte_parent(enode); + return (parent == node); +} + +/* + * mas_allocated() - Get the number of nodes allocated in a maple state. + * @mas: The maple state + * + * The ma_state alloc member is overloaded to hold a pointer to the first + * allocated node or to the number of requested nodes to allocate. If bit 0 is + * set, then the alloc contains the number of requested nodes. If there is an + * allocated node, then the total allocated nodes is in that node. + * + * Return: The total number of nodes allocated + */ +static inline unsigned long mas_allocated(const struct ma_state *mas) +{ + if (!mas->alloc || ((unsigned long)mas->alloc & 0x1)) + return 0; + + return mas->alloc->total; +} + +/* + * mas_set_alloc_req() - Set the requested number of allocations. + * @mas: the maple state + * @count: the number of allocations. + * + * The requested number of allocations is either in the first allocated node, + * located in @mas->alloc->request_count, or directly in @mas->alloc if there is + * no allocated node. Set the request either in the node or do the necessary + * encoding to store in @mas->alloc directly. + */ +static inline void mas_set_alloc_req(struct ma_state *mas, unsigned long count) +{ + if (!mas->alloc || ((unsigned long)mas->alloc & 0x1)) { + if (!count) + mas->alloc = NULL; + else + mas->alloc = (struct maple_alloc *)(((count) << 1U) | 1U); + return; + } + + mas->alloc->request_count = count; +} + +/* + * mas_alloc_req() - get the requested number of allocations. + * @mas: The maple state + * + * The alloc count is either stored directly in @mas, or in + * @mas->alloc->request_count if there is at least one node allocated. Decode + * the request count if it's stored directly in @mas->alloc. + * + * Return: The allocation request count. + */ +static inline unsigned int mas_alloc_req(const struct ma_state *mas) +{ + if ((unsigned long)mas->alloc & 0x1) + return (unsigned long)(mas->alloc) >> 1; + else if (mas->alloc) + return mas->alloc->request_count; + return 0; +} + +/* + * ma_pivots() - Get a pointer to the maple node pivots. + * @node - the maple node + * @type - the node type + * + * Return: A pointer to the maple node pivots + */ +static inline unsigned long *ma_pivots(struct maple_node *node, + enum maple_type type) +{ + switch (type) { + case maple_arange_64: + return node->ma64.pivot; + case maple_range_64: + case maple_leaf_64: + return node->mr64.pivot; + case maple_dense: + return NULL; + } + return NULL; +} + +/* + * ma_gaps() - Get a pointer to the maple node gaps. + * @node - the maple node + * @type - the node type + * + * Return: A pointer to the maple node gaps + */ +static inline unsigned long *ma_gaps(struct maple_node *node, + enum maple_type type) +{ + switch (type) { + case maple_arange_64: + return node->ma64.gap; + case maple_range_64: + case maple_leaf_64: + case maple_dense: + return NULL; + } + return NULL; +} + +/* + * mte_pivot() - Get the pivot at @piv of the maple encoded node. + * @mn: The maple encoded node. + * @piv: The pivot. + * + * Return: the pivot at @piv of @mn. + */ +static inline unsigned long mte_pivot(const struct maple_enode *mn, + unsigned char piv) +{ + struct maple_node *node = mte_to_node(mn); + + if (piv >= mt_pivots[piv]) { + WARN_ON(1); + return 0; + } + switch (mte_node_type(mn)) { + case maple_arange_64: + return node->ma64.pivot[piv]; + case maple_range_64: + case maple_leaf_64: + return node->mr64.pivot[piv]; + case maple_dense: + return 0; + } + return 0; +} + +/* + * mas_safe_pivot() - get the pivot at @piv or mas->max. + * @mas: The maple state + * @pivots: The pointer to the maple node pivots + * @piv: The pivot to fetch + * @type: The maple node type + * + * Return: The pivot at @piv within the limit of the @pivots array, @mas->max + * otherwise. + */ +static inline unsigned long +mas_safe_pivot(const struct ma_state *mas, unsigned long *pivots, + unsigned char piv, enum maple_type type) +{ + if (piv >= mt_pivots[type]) + return mas->max; + + return pivots[piv]; +} + +/* + * mas_safe_min() - Return the minimum for a given offset. + * @mas: The maple state + * @pivots: The pointer to the maple node pivots + * @offset: The offset into the pivot array + * + * Return: The minimum range value that is contained in @offset. + */ +static inline unsigned long +mas_safe_min(struct ma_state *mas, unsigned long *pivots, unsigned char offset) +{ + if (likely(offset)) + return pivots[offset - 1] + 1; + + return mas->min; +} + +/* + * mas_logical_pivot() - Get the logical pivot of a given offset. + * @mas: The maple state + * @pivots: The pointer to the maple node pivots + * @offset: The offset into the pivot array + * @type: The maple node type + * + * When there is no value at a pivot (beyond the end of the data), then the + * pivot is actually @mas->max. + * + * Return: the logical pivot of a given @offset. + */ +static inline unsigned long +mas_logical_pivot(struct ma_state *mas, unsigned long *pivots, + unsigned char offset, enum maple_type type) +{ + unsigned long lpiv = mas_safe_pivot(mas, pivots, offset, type); + + if (likely(lpiv)) + return lpiv; + + if (likely(offset)) + return mas->max; + + return lpiv; +} + +/* + * mte_set_pivot() - Set a pivot to a value in an encoded maple node. + * @mn: The encoded maple node + * @piv: The pivot offset + * @val: The value of the pivot + */ +static inline void mte_set_pivot(struct maple_enode *mn, unsigned char piv, + unsigned long val) +{ + struct maple_node *node = mte_to_node(mn); + enum maple_type type = mte_node_type(mn); + + BUG_ON(piv >= mt_pivots[type]); + switch (type) { + default: + case maple_range_64: + case maple_leaf_64: + node->mr64.pivot[piv] = val; + break; + case maple_arange_64: + node->ma64.pivot[piv] = val; + break; + case maple_dense: + break; + } + +} + +/* + * ma_slots() - Get a pointer to the maple node slots. + * @mn: The maple node + * @mt: The maple node type + * + * Return: A pointer to the maple node slots + */ +static inline void __rcu **ma_slots(struct maple_node *mn, enum maple_type mt) +{ + switch (mt) { + default: + case maple_arange_64: + return mn->ma64.slot; + case maple_range_64: + case maple_leaf_64: + return mn->mr64.slot; + case maple_dense: + return mn->slot; + } +} + +static inline bool mt_locked(const struct maple_tree *mt) +{ + return mt_external_lock(mt) ? mt_lock_is_held(mt) : + lockdep_is_held(&mt->ma_lock); +} + +static inline void *mt_slot(const struct maple_tree *mt, + void __rcu **slots, unsigned char offset) +{ + return rcu_dereference_check(slots[offset], mt_locked(mt)); +} + +/* + * mas_slot_locked() - Get the slot value when holding the maple tree lock. + * @mas: The maple state + * @slots: The pointer to the slots + * @offset: The offset into the slots array to fetch + * + * Return: The entry stored in @slots at the @offset. + */ +static inline void *mas_slot_locked(struct ma_state *mas, void __rcu **slots, + unsigned char offset) +{ + return rcu_dereference_protected(slots[offset], mt_locked(mas->tree)); +} + +/* + * mas_slot() - Get the slot value when not holding the maple tree lock. + * @mas: The maple state + * @slots: The pointer to the slots + * @offset: The offset into the slots array to fetch + * + * Return: The entry stored in @slots at the @offset + */ +static inline void *mas_slot(struct ma_state *mas, void __rcu **slots, + unsigned char offset) +{ + return mt_slot(mas->tree, slots, offset); +} + +/* + * mas_root() - Get the maple tree root. + * @mas: The maple state. + * + * Return: The pointer to the root of the tree + */ +static inline void *mas_root(struct ma_state *mas) +{ + return rcu_dereference_check(mas->tree->ma_root, mt_locked(mas->tree)); +} + +static inline void *mt_root_locked(struct maple_tree *mt) +{ + return rcu_dereference_protected(mt->ma_root, mt_locked(mt)); +} + +/* + * mas_root_locked() - Get the maple tree root when holding the maple tree lock. + * @mas: The maple state. + * + * Return: The pointer to the root of the tree + */ +static inline void *mas_root_locked(struct ma_state *mas) +{ + return mt_root_locked(mas->tree); +} + +static inline struct maple_metadata *ma_meta(struct maple_node *mn, + enum maple_type mt) +{ + switch (mt) { + case maple_arange_64: + return &mn->ma64.meta; + default: + return &mn->mr64.meta; + } +} + +/* + * ma_set_meta() - Set the metadata information of a node. + * @mn: The maple node + * @mt: The maple node type + * @offset: The offset of the highest sub-gap in this node. + * @end: The end of the data in this node. + */ +static inline void ma_set_meta(struct maple_node *mn, enum maple_type mt, + unsigned char offset, unsigned char end) +{ + struct maple_metadata *meta = ma_meta(mn, mt); + + meta->gap = offset; + meta->end = end; +} + +/* + * ma_meta_end() - Get the data end of a node from the metadata + * @mn: The maple node + * @mt: The maple node type + */ +static inline unsigned char ma_meta_end(struct maple_node *mn, + enum maple_type mt) +{ + struct maple_metadata *meta = ma_meta(mn, mt); + + return meta->end; +} + +/* + * ma_meta_gap() - Get the largest gap location of a node from the metadata + * @mn: The maple node + * @mt: The maple node type + */ +static inline unsigned char ma_meta_gap(struct maple_node *mn, + enum maple_type mt) +{ + BUG_ON(mt != maple_arange_64); + + return mn->ma64.meta.gap; +} + +/* + * ma_set_meta_gap() - Set the largest gap location in a nodes metadata + * @mn: The maple node + * @mn: The maple node type + * @offset: The location of the largest gap. + */ +static inline void ma_set_meta_gap(struct maple_node *mn, enum maple_type mt, + unsigned char offset) +{ + + struct maple_metadata *meta = ma_meta(mn, mt); + + meta->gap = offset; +} + +/* + * mat_add() - Add a @dead_enode to the ma_topiary of a list of dead nodes. + * @mat - the ma_topiary, a linked list of dead nodes. + * @dead_enode - the node to be marked as dead and added to the tail of the list + * + * Add the @dead_enode to the linked list in @mat. + */ +static inline void mat_add(struct ma_topiary *mat, + struct maple_enode *dead_enode) +{ + mte_set_node_dead(dead_enode); + mte_to_mat(dead_enode)->next = NULL; + if (!mat->tail) { + mat->tail = mat->head = dead_enode; + return; + } + + mte_to_mat(mat->tail)->next = dead_enode; + mat->tail = dead_enode; +} + +static void mte_destroy_walk(struct maple_enode *, struct maple_tree *); +static inline void mas_free(struct ma_state *mas, struct maple_enode *used); + +/* + * mas_mat_free() - Free all nodes in a dead list. + * @mas - the maple state + * @mat - the ma_topiary linked list of dead nodes to free. + * + * Free walk a dead list. + */ +static void mas_mat_free(struct ma_state *mas, struct ma_topiary *mat) +{ + struct maple_enode *next; + + while (mat->head) { + next = mte_to_mat(mat->head)->next; + mas_free(mas, mat->head); + mat->head = next; + } +} + +/* + * mas_mat_destroy() - Free all nodes and subtrees in a dead list. + * @mas - the maple state + * @mat - the ma_topiary linked list of dead nodes to free. + * + * Destroy walk a dead list. + */ +static void mas_mat_destroy(struct ma_state *mas, struct ma_topiary *mat) +{ + struct maple_enode *next; + + while (mat->head) { + next = mte_to_mat(mat->head)->next; + mte_destroy_walk(mat->head, mat->mtree); + mat->head = next; + } +} +/* + * mas_descend() - Descend into the slot stored in the ma_state. + * @mas - the maple state. + * + * Note: Not RCU safe, only use in write side or debug code. + */ +static inline void mas_descend(struct ma_state *mas) +{ + enum maple_type type; + unsigned long *pivots; + struct maple_node *node; + void __rcu **slots; + + node = mas_mn(mas); + type = mte_node_type(mas->node); + pivots = ma_pivots(node, type); + slots = ma_slots(node, type); + + if (mas->offset) + mas->min = pivots[mas->offset - 1] + 1; + mas->max = mas_safe_pivot(mas, pivots, mas->offset, type); + mas->node = mas_slot(mas, slots, mas->offset); +} + +/* + * mte_set_gap() - Set a maple node gap. + * @mn: The encoded maple node + * @gap: The offset of the gap to set + * @val: The gap value + */ +static inline void mte_set_gap(const struct maple_enode *mn, + unsigned char gap, unsigned long val) +{ + switch (mte_node_type(mn)) { + default: + break; + case maple_arange_64: + mte_to_node(mn)->ma64.gap[gap] = val; + break; + } +} + +/* + * mas_ascend() - Walk up a level of the tree. + * @mas: The maple state + * + * Sets the @mas->max and @mas->min to the correct values when walking up. This + * may cause several levels of walking up to find the correct min and max. + * May find a dead node which will cause a premature return. + * Return: 1 on dead node, 0 otherwise + */ +static int mas_ascend(struct ma_state *mas) +{ + struct maple_enode *p_enode; /* parent enode. */ + struct maple_enode *a_enode; /* ancestor enode. */ + struct maple_node *a_node; /* ancestor node. */ + struct maple_node *p_node; /* parent node. */ + unsigned char a_slot; + enum maple_type a_type; + unsigned long min, max; + unsigned long *pivots; + unsigned char offset; + bool set_max = false, set_min = false; + + a_node = mas_mn(mas); + if (ma_is_root(a_node)) { + mas->offset = 0; + return 0; + } + + p_node = mte_parent(mas->node); + if (unlikely(a_node == p_node)) + return 1; + a_type = mas_parent_enum(mas, mas->node); + offset = mte_parent_slot(mas->node); + a_enode = mt_mk_node(p_node, a_type); + + /* Check to make sure all parent information is still accurate */ + if (p_node != mte_parent(mas->node)) + return 1; + + mas->node = a_enode; + mas->offset = offset; + + if (mte_is_root(a_enode)) { + mas->max = ULONG_MAX; + mas->min = 0; + return 0; + } + + min = 0; + max = ULONG_MAX; + do { + p_enode = a_enode; + a_type = mas_parent_enum(mas, p_enode); + a_node = mte_parent(p_enode); + a_slot = mte_parent_slot(p_enode); + pivots = ma_pivots(a_node, a_type); + a_enode = mt_mk_node(a_node, a_type); + + if (!set_min && a_slot) { + set_min = true; + min = pivots[a_slot - 1] + 1; + } + + if (!set_max && a_slot < mt_pivots[a_type]) { + set_max = true; + max = pivots[a_slot]; + } + + if (unlikely(ma_dead_node(a_node))) + return 1; + + if (unlikely(ma_is_root(a_node))) + break; + + } while (!set_min || !set_max); + + mas->max = max; + mas->min = min; + return 0; +} + +/* + * mas_pop_node() - Get a previously allocated maple node from the maple state. + * @mas: The maple state + * + * Return: A pointer to a maple node. + */ +static inline struct maple_node *mas_pop_node(struct ma_state *mas) +{ + struct maple_alloc *ret, *node = mas->alloc; + unsigned long total = mas_allocated(mas); + + /* nothing or a request pending. */ + if (unlikely(!total)) + return NULL; + + if (total == 1) { + /* single allocation in this ma_state */ + mas->alloc = NULL; + ret = node; + goto single_node; + } + + if (!node->node_count) { + /* Single allocation in this node. */ + mas->alloc = node->slot[0]; + node->slot[0] = NULL; + mas->alloc->total = node->total - 1; + ret = node; + goto new_head; + } + + node->total--; + ret = node->slot[node->node_count]; + node->slot[node->node_count--] = NULL; + +single_node: +new_head: + ret->total = 0; + ret->node_count = 0; + if (ret->request_count) { + mas_set_alloc_req(mas, ret->request_count + 1); + ret->request_count = 0; + } + return (struct maple_node *)ret; +} + +/* + * mas_push_node() - Push a node back on the maple state allocation. + * @mas: The maple state + * @used: The used maple node + * + * Stores the maple node back into @mas->alloc for reuse. Updates allocated and + * requested node count as necessary. + */ +static inline void mas_push_node(struct ma_state *mas, struct maple_node *used) +{ + struct maple_alloc *reuse = (struct maple_alloc *)used; + struct maple_alloc *head = mas->alloc; + unsigned long count; + unsigned int requested = mas_alloc_req(mas); + + memset(reuse, 0, sizeof(*reuse)); + count = mas_allocated(mas); + + if (count && (head->node_count < MAPLE_ALLOC_SLOTS - 1)) { + if (head->slot[0]) + head->node_count++; + head->slot[head->node_count] = reuse; + head->total++; + goto done; + } + + reuse->total = 1; + if ((head) && !((unsigned long)head & 0x1)) { + head->request_count = 0; + reuse->slot[0] = head; + reuse->total += head->total; + } + + mas->alloc = reuse; +done: + if (requested > 1) + mas_set_alloc_req(mas, requested - 1); +} + +/* + * mas_alloc_nodes() - Allocate nodes into a maple state + * @mas: The maple state + * @gfp: The GFP Flags + */ +static inline void mas_alloc_nodes(struct ma_state *mas, gfp_t gfp) +{ + struct maple_alloc *node; + struct maple_alloc **nodep = &mas->alloc; + unsigned long allocated = mas_allocated(mas); + unsigned long success = allocated; + unsigned int requested = mas_alloc_req(mas); + unsigned int count; + void **slots = NULL; + unsigned int max_req = 0; + + if (!requested) + return; + + mas_set_alloc_req(mas, 0); + if (!allocated || mas->alloc->node_count == MAPLE_ALLOC_SLOTS - 1) { + node = (struct maple_alloc *)mt_alloc_one(gfp); + if (!node) + goto nomem_one; + + if (allocated) + node->slot[0] = mas->alloc; + + success++; + mas->alloc = node; + requested--; + } + + node = mas->alloc; + while (requested) { + max_req = MAPLE_NODE_SLOTS - 1; + if (node->slot[0]) { + unsigned int offset = node->node_count + 1; + + slots = (void **)&node->slot[offset]; + max_req -= offset; + } else { + slots = (void **)&node->slot; + } + + max_req = min(requested, max_req); + count = mt_alloc_bulk(gfp, max_req, slots); + if (!count) + goto nomem_bulk; + + node->node_count += count; + /* zero indexed. */ + if (slots == (void **)&node->slot) + node->node_count--; + + success += count; + nodep = &node->slot[0]; + node = *nodep; + requested -= count; + } + mas->alloc->total = success; + return; + +nomem_bulk: + /* Clean up potential freed allocations on bulk failure */ + memset(slots, 0, max_req * sizeof(unsigned long)); +nomem_one: + mas_set_alloc_req(mas, requested); + if (mas->alloc && !(((unsigned long)mas->alloc & 0x1))) + mas->alloc->total = success; + mas_set_err(mas, -ENOMEM); + return; + +} + +/* + * mas_free() - Free an encoded maple node + * @mas: The maple state + * @used: The encoded maple node to free. + * + * Uses rcu free if necessary, pushes @used back on the maple state allocations + * otherwise. + */ +static inline void mas_free(struct ma_state *mas, struct maple_enode *used) +{ + struct maple_node *tmp = mte_to_node(used); + + if (mt_in_rcu(mas->tree)) + ma_free_rcu(tmp); + else + mas_push_node(mas, tmp); +} + +/* + * mas_node_count() - Check if enough nodes are allocated and request more if + * there is not enough nodes. + * @mas: The maple state + * @count: The number of nodes needed + */ +static void mas_node_count(struct ma_state *mas, int count) +{ + unsigned long allocated = mas_allocated(mas); + + if (allocated < count) { + mas_set_alloc_req(mas, count - allocated); + mas_alloc_nodes(mas, GFP_NOWAIT | __GFP_NOWARN); + } +} + +/* + * mas_start() - Sets up maple state for operations. + * @mas: The maple state. + * + * If mas->node == MAS_START, then set the min, max, depth, and offset to + * defaults. + * + * Return: + * - If mas->node is an error or not MAS_START, return NULL. + * - If it's an empty tree: NULL & mas->node == MAS_NONE + * - If it's a single entry: The entry & mas->node == MAS_ROOT + * - If it's a tree: NULL & mas->node == safe root node. + */ +static inline struct maple_enode *mas_start(struct ma_state *mas) +{ + if (likely(mas_is_start(mas))) { + struct maple_enode *root; + + mas->node = MAS_NONE; + mas->min = 0; + mas->max = ULONG_MAX; + mas->depth = 0; + mas->offset = 0; + + root = mas_root(mas); + /* Tree with nodes */ + if (likely(xa_is_node(root))) { + mas->node = mte_safe_root(root); + return NULL; + } + + /* empty tree */ + if (unlikely(!root)) { + mas->offset = MAPLE_NODE_SLOTS; + return NULL; + } + + /* Single entry tree */ + mas->node = MAS_ROOT; + mas->offset = MAPLE_NODE_SLOTS; + + /* Single entry tree. */ + if (mas->index > 0) + return NULL; + + return root; + } + + return NULL; +} + +/* + * ma_data_end() - Find the end of the data in a node. + * @node: The maple node + * @type: The maple node type + * @pivots: The array of pivots in the node + * @max: The maximum value in the node + * + * Uses metadata to find the end of the data when possible. + * Return: The zero indexed last slot with data (may be null). + */ +static inline unsigned char ma_data_end(struct maple_node *node, + enum maple_type type, + unsigned long *pivots, + unsigned long max) +{ + unsigned char offset; + + if (type == maple_arange_64) + return ma_meta_end(node, type); + + offset = mt_pivots[type] - 1; + if (likely(!pivots[offset])) + return ma_meta_end(node, type); + + if (likely(pivots[offset] == max)) + return offset; + + return mt_pivots[type]; +} + +/* + * mas_data_end() - Find the end of the data (slot). + * @mas: the maple state + * + * This method is optimized to check the metadata of a node if the node type + * supports data end metadata. + * + * Return: The zero indexed last slot with data (may be null). + */ +static inline unsigned char mas_data_end(struct ma_state *mas) +{ + enum maple_type type; + struct maple_node *node; + unsigned char offset; + unsigned long *pivots; + + type = mte_node_type(mas->node); + node = mas_mn(mas); + if (type == maple_arange_64) + return ma_meta_end(node, type); + + pivots = ma_pivots(node, type); + offset = mt_pivots[type] - 1; + if (likely(!pivots[offset])) + return ma_meta_end(node, type); + + if (likely(pivots[offset] == mas->max)) + return offset; + + return mt_pivots[type]; +} + +/* + * mas_leaf_max_gap() - Returns the largest gap in a leaf node + * @mas - the maple state + * + * Return: The maximum gap in the leaf. + */ +static unsigned long mas_leaf_max_gap(struct ma_state *mas) +{ + enum maple_type mt; + unsigned long pstart, gap, max_gap; + struct maple_node *mn; + unsigned long *pivots; + void __rcu **slots; + unsigned char i; + unsigned char max_piv; + + mt = mte_node_type(mas->node); + mn = mas_mn(mas); + slots = ma_slots(mn, mt); + max_gap = 0; + if (unlikely(ma_is_dense(mt))) { + gap = 0; + for (i = 0; i < mt_slots[mt]; i++) { + if (slots[i]) { + if (gap > max_gap) + max_gap = gap; + gap = 0; + } else { + gap++; + } + } + if (gap > max_gap) + max_gap = gap; + return max_gap; + } + + /* + * Check the first implied pivot optimizes the loop below and slot 1 may + * be skipped if there is a gap in slot 0. + */ + pivots = ma_pivots(mn, mt); + if (likely(!slots[0])) { + max_gap = pivots[0] - mas->min + 1; + i = 2; + } else { + i = 1; + } + + /* reduce max_piv as the special case is checked before the loop */ + max_piv = ma_data_end(mn, mt, pivots, mas->max) - 1; + /* + * Check end implied pivot which can only be a gap on the right most + * node. + */ + if (unlikely(mas->max == ULONG_MAX) && !slots[max_piv + 1]) { + gap = ULONG_MAX - pivots[max_piv]; + if (gap > max_gap) + max_gap = gap; + } + + for (; i <= max_piv; i++) { + /* data == no gap. */ + if (likely(slots[i])) + continue; + + pstart = pivots[i - 1]; + gap = pivots[i] - pstart; + if (gap > max_gap) + max_gap = gap; + + /* There cannot be two gaps in a row. */ + i++; + } + return max_gap; +} + +/* + * ma_max_gap() - Get the maximum gap in a maple node (non-leaf) + * @node: The maple node + * @gaps: The pointer to the gaps + * @mt: The maple node type + * @*off: Pointer to store the offset location of the gap. + * + * Uses the metadata data end to scan backwards across set gaps. + * + * Return: The maximum gap value + */ +static inline unsigned long +ma_max_gap(struct maple_node *node, unsigned long *gaps, enum maple_type mt, + unsigned char *off) +{ + unsigned char offset, i; + unsigned long max_gap = 0; + + i = offset = ma_meta_end(node, mt); + do { + if (gaps[i] > max_gap) { + max_gap = gaps[i]; + offset = i; + } + } while (i--); + + *off = offset; + return max_gap; +} + +/* + * mas_max_gap() - find the largest gap in a non-leaf node and set the slot. + * @mas: The maple state. + * + * If the metadata gap is set to MAPLE_ARANGE64_META_MAX, there is no gap. + * + * Return: The gap value. + */ +static inline unsigned long mas_max_gap(struct ma_state *mas) +{ + unsigned long *gaps; + unsigned char offset; + enum maple_type mt; + struct maple_node *node; + + mt = mte_node_type(mas->node); + if (ma_is_leaf(mt)) + return mas_leaf_max_gap(mas); + + node = mas_mn(mas); + offset = ma_meta_gap(node, mt); + if (offset == MAPLE_ARANGE64_META_MAX) + return 0; + + gaps = ma_gaps(node, mt); + return gaps[offset]; +} + +/* + * mas_parent_gap() - Set the parent gap and any gaps above, as needed + * @mas: The maple state + * @offset: The gap offset in the parent to set + * @new: The new gap value. + * + * Set the parent gap then continue to set the gap upwards, using the metadata + * of the parent to see if it is necessary to check the node above. + */ +static inline void mas_parent_gap(struct ma_state *mas, unsigned char offset, + unsigned long new) +{ + unsigned long meta_gap = 0; + struct maple_node *pnode; + struct maple_enode *penode; + unsigned long *pgaps; + unsigned char meta_offset; + enum maple_type pmt; + + pnode = mte_parent(mas->node); + pmt = mas_parent_enum(mas, mas->node); + penode = mt_mk_node(pnode, pmt); + pgaps = ma_gaps(pnode, pmt); + +ascend: + meta_offset = ma_meta_gap(pnode, pmt); + if (meta_offset == MAPLE_ARANGE64_META_MAX) + meta_gap = 0; + else + meta_gap = pgaps[meta_offset]; + + pgaps[offset] = new; + + if (meta_gap == new) + return; + + if (offset != meta_offset) { + if (meta_gap > new) + return; + + ma_set_meta_gap(pnode, pmt, offset); + } else if (new < meta_gap) { + meta_offset = 15; + new = ma_max_gap(pnode, pgaps, pmt, &meta_offset); + ma_set_meta_gap(pnode, pmt, meta_offset); + } + + if (ma_is_root(pnode)) + return; + + /* Go to the parent node. */ + pnode = mte_parent(penode); + pmt = mas_parent_enum(mas, penode); + pgaps = ma_gaps(pnode, pmt); + offset = mte_parent_slot(penode); + penode = mt_mk_node(pnode, pmt); + goto ascend; +} + +/* + * mas_update_gap() - Update a nodes gaps and propagate up if necessary. + * @mas - the maple state. + */ +static inline void mas_update_gap(struct ma_state *mas) +{ + unsigned char pslot; + unsigned long p_gap; + unsigned long max_gap; + + if (!mt_is_alloc(mas->tree)) + return; + + if (mte_is_root(mas->node)) + return; + + max_gap = mas_max_gap(mas); + + pslot = mte_parent_slot(mas->node); + p_gap = ma_gaps(mte_parent(mas->node), + mas_parent_enum(mas, mas->node))[pslot]; + + if (p_gap != max_gap) + mas_parent_gap(mas, pslot, max_gap); +} + +/* + * mas_adopt_children() - Set the parent pointer of all nodes in @parent to + * @parent with the slot encoded. + * @mas - the maple state (for the tree) + * @parent - the maple encoded node containing the children. + */ +static inline void mas_adopt_children(struct ma_state *mas, + struct maple_enode *parent) +{ + enum maple_type type = mte_node_type(parent); + struct maple_node *node = mas_mn(mas); + void __rcu **slots = ma_slots(node, type); + unsigned long *pivots = ma_pivots(node, type); + struct maple_enode *child; + unsigned char offset; + + offset = ma_data_end(node, type, pivots, mas->max); + do { + child = mas_slot_locked(mas, slots, offset); + mte_set_parent(child, parent, offset); + } while (offset--); +} + +/* + * mas_replace() - Replace a maple node in the tree with mas->node. Uses the + * parent encoding to locate the maple node in the tree. + * @mas - the ma_state to use for operations. + * @advanced - boolean to adopt the child nodes and free the old node (false) or + * leave the node (true) and handle the adoption and free elsewhere. + */ +static inline void mas_replace(struct ma_state *mas, bool advanced) + __must_hold(mas->tree->lock) +{ + struct maple_node *mn = mas_mn(mas); + struct maple_enode *old_enode; + unsigned char offset = 0; + void __rcu **slots = NULL; + + if (ma_is_root(mn)) { + old_enode = mas_root_locked(mas); + } else { + offset = mte_parent_slot(mas->node); + slots = ma_slots(mte_parent(mas->node), + mas_parent_enum(mas, mas->node)); + old_enode = mas_slot_locked(mas, slots, offset); + } + + if (!advanced && !mte_is_leaf(mas->node)) + mas_adopt_children(mas, mas->node); + + if (mte_is_root(mas->node)) { + mn->parent = ma_parent_ptr( + ((unsigned long)mas->tree | MA_ROOT_PARENT)); + rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); + mas_set_height(mas); + } else { + rcu_assign_pointer(slots[offset], mas->node); + } + + if (!advanced) + mas_free(mas, old_enode); +} + +/* + * mas_new_child() - Find the new child of a node. + * @mas: the maple state + * @child: the maple state to store the child. + */ +static inline bool mas_new_child(struct ma_state *mas, struct ma_state *child) + __must_hold(mas->tree->lock) +{ + enum maple_type mt; + unsigned char offset; + unsigned char end; + unsigned long *pivots; + struct maple_enode *entry; + struct maple_node *node; + void __rcu **slots; + + mt = mte_node_type(mas->node); + node = mas_mn(mas); + slots = ma_slots(node, mt); + pivots = ma_pivots(node, mt); + end = ma_data_end(node, mt, pivots, mas->max); + for (offset = mas->offset; offset <= end; offset++) { + entry = mas_slot_locked(mas, slots, offset); + if (mte_parent(entry) == node) { + *child = *mas; + mas->offset = offset + 1; + child->offset = offset; + mas_descend(child); + child->offset = 0; + return true; + } + } + return false; +} + +/* + * mab_shift_right() - Shift the data in mab right. Note, does not clean out the + * old data or set b_node->b_end. + * @b_node: the maple_big_node + * @shift: the shift count + */ +static inline void mab_shift_right(struct maple_big_node *b_node, + unsigned char shift) +{ + unsigned long size = b_node->b_end * sizeof(unsigned long); + + memmove(b_node->pivot + shift, b_node->pivot, size); + memmove(b_node->slot + shift, b_node->slot, size); + if (b_node->type == maple_arange_64) + memmove(b_node->gap + shift, b_node->gap, size); +} + +/* + * mab_middle_node() - Check if a middle node is needed (unlikely) + * @b_node: the maple_big_node that contains the data. + * @size: the amount of data in the b_node + * @split: the potential split location + * @slot_count: the size that can be stored in a single node being considered. + * + * Return: true if a middle node is required. + */ +static inline bool mab_middle_node(struct maple_big_node *b_node, int split, + unsigned char slot_count) +{ + unsigned char size = b_node->b_end; + + if (size >= 2 * slot_count) + return true; + + if (!b_node->slot[split] && (size >= 2 * slot_count - 1)) + return true; + + return false; +} + +/* + * mab_no_null_split() - ensure the split doesn't fall on a NULL + * @b_node: the maple_big_node with the data + * @split: the suggested split location + * @slot_count: the number of slots in the node being considered. + * + * Return: the split location. + */ +static inline int mab_no_null_split(struct maple_big_node *b_node, + unsigned char split, unsigned char slot_count) +{ + if (!b_node->slot[split]) { + /* + * If the split is less than the max slot && the right side will + * still be sufficient, then increment the split on NULL. + */ + if ((split < slot_count - 1) && + (b_node->b_end - split) > (mt_min_slots[b_node->type])) + split++; + else + split--; + } + return split; +} + +/* + * mab_calc_split() - Calculate the split location and if there needs to be two + * splits. + * @bn: The maple_big_node with the data + * @mid_split: The second split, if required. 0 otherwise. + * + * Return: The first split location. The middle split is set in @mid_split. + */ +static inline int mab_calc_split(struct ma_state *mas, + struct maple_big_node *bn, unsigned char *mid_split, unsigned long min) +{ + unsigned char b_end = bn->b_end; + int split = b_end / 2; /* Assume equal split. */ + unsigned char slot_min, slot_count = mt_slots[bn->type]; + + /* + * To support gap tracking, all NULL entries are kept together and a node cannot + * end on a NULL entry, with the exception of the left-most leaf. The + * limitation means that the split of a node must be checked for this condition + * and be able to put more data in one direction or the other. + */ + if (unlikely((mas->mas_flags & MA_STATE_BULK))) { + *mid_split = 0; + if (ma_is_leaf(bn->type)) + slot_min = 2; + else + return b_end - mt_min_slots[bn->type]; + + split = b_end - slot_min; + mas->mas_flags |= MA_STATE_REBALANCE; + if (!bn->slot[split]) + split--; + return split; + } + + /* + * Although extremely rare, it is possible to enter what is known as the 3-way + * split scenario. The 3-way split comes about by means of a store of a range + * that overwrites the end and beginning of two full nodes. The result is a set + * of entries that cannot be stored in 2 nodes. Sometimes, these two nodes can + * also be located in different parent nodes which are also full. This can + * carry upwards all the way to the root in the worst case. + */ + if (unlikely(mab_middle_node(bn, split, slot_count))) { + split = b_end / 3; + *mid_split = split * 2; + } else { + slot_min = mt_min_slots[bn->type]; + + *mid_split = 0; + /* + * Avoid having a range less than the slot count unless it + * causes one node to be deficient. + * NOTE: mt_min_slots is 1 based, b_end and split are zero. + */ + while (((bn->pivot[split] - min) < slot_count - 1) && + (split < slot_count - 1) && (b_end - split > slot_min)) + split++; + } + + /* Avoid ending a node on a NULL entry */ + split = mab_no_null_split(bn, split, slot_count); + if (!(*mid_split)) + return split; + + *mid_split = mab_no_null_split(bn, *mid_split, slot_count); + + return split; +} + +/* + * mas_mab_cp() - Copy data from a maple state inclusively to a maple_big_node + * and set @b_node->b_end to the next free slot. + * @mas: The maple state + * @mas_start: The starting slot to copy + * @mas_end: The end slot to copy (inclusively) + * @b_node: The maple_big_node to place the data + * @mab_start: The starting location in maple_big_node to store the data. + */ +static inline void mas_mab_cp(struct ma_state *mas, unsigned char mas_start, + unsigned char mas_end, struct maple_big_node *b_node, + unsigned char mab_start) +{ + enum maple_type mt; + struct maple_node *node; + void __rcu **slots; + unsigned long *pivots, *gaps; + int i = mas_start, j = mab_start; + unsigned char piv_end; + + node = mas_mn(mas); + mt = mte_node_type(mas->node); + pivots = ma_pivots(node, mt); + if (!i) { + b_node->pivot[j] = pivots[i++]; + if (unlikely(i > mas_end)) + goto complete; + j++; + } + + piv_end = min(mas_end, mt_pivots[mt]); + for (; i < piv_end; i++, j++) { + b_node->pivot[j] = pivots[i]; + if (unlikely(!b_node->pivot[j])) + break; + + if (unlikely(mas->max == b_node->pivot[j])) + goto complete; + } + + if (likely(i <= mas_end)) + b_node->pivot[j] = mas_safe_pivot(mas, pivots, i, mt); + +complete: + b_node->b_end = ++j; + j -= mab_start; + slots = ma_slots(node, mt); + memcpy(b_node->slot + mab_start, slots + mas_start, sizeof(void *) * j); + if (!ma_is_leaf(mt) && mt_is_alloc(mas->tree)) { + gaps = ma_gaps(node, mt); + memcpy(b_node->gap + mab_start, gaps + mas_start, + sizeof(unsigned long) * j); + } +} + +/* + * mas_leaf_set_meta() - Set the metadata of a leaf if possible. + * @mas: The maple state + * @node: The maple node + * @pivots: pointer to the maple node pivots + * @mt: The maple type + * @end: The assumed end + * + * Note, end may be incremented within this function but not modified at the + * source. This is fine since the metadata is the last thing to be stored in a + * node during a write. + */ +static inline void mas_leaf_set_meta(struct ma_state *mas, + struct maple_node *node, unsigned long *pivots, + enum maple_type mt, unsigned char end) +{ + /* There is no room for metadata already */ + if (mt_pivots[mt] <= end) + return; + + if (pivots[end] && pivots[end] < mas->max) + end++; + + if (end < mt_slots[mt] - 1) + ma_set_meta(node, mt, 0, end); +} + +/* + * mab_mas_cp() - Copy data from maple_big_node to a maple encoded node. + * @b_node: the maple_big_node that has the data + * @mab_start: the start location in @b_node. + * @mab_end: The end location in @b_node (inclusively) + * @mas: The maple state with the maple encoded node. + */ +static inline void mab_mas_cp(struct maple_big_node *b_node, + unsigned char mab_start, unsigned char mab_end, + struct ma_state *mas, bool new_max) +{ + int i, j = 0; + enum maple_type mt = mte_node_type(mas->node); + struct maple_node *node = mte_to_node(mas->node); + void __rcu **slots = ma_slots(node, mt); + unsigned long *pivots = ma_pivots(node, mt); + unsigned long *gaps = NULL; + unsigned char end; + + if (mab_end - mab_start > mt_pivots[mt]) + mab_end--; + + if (!pivots[mt_pivots[mt] - 1]) + slots[mt_pivots[mt]] = NULL; + + i = mab_start; + pivots[j++] = b_node->pivot[i++]; + do { + pivots[j++] = b_node->pivot[i++]; + } while (i <= mab_end && likely(b_node->pivot[i])); + + memcpy(slots, b_node->slot + mab_start, + sizeof(void *) * (i - mab_start)); + + if (new_max) + mas->max = b_node->pivot[i - 1]; + + end = j - 1; + if (likely(!ma_is_leaf(mt) && mt_is_alloc(mas->tree))) { + unsigned long max_gap = 0; + unsigned char offset = 15; + + gaps = ma_gaps(node, mt); + do { + gaps[--j] = b_node->gap[--i]; + if (gaps[j] > max_gap) { + offset = j; + max_gap = gaps[j]; + } + } while (j); + + ma_set_meta(node, mt, offset, end); + } else { + mas_leaf_set_meta(mas, node, pivots, mt, end); + } +} + +/* + * mas_descend_adopt() - Descend through a sub-tree and adopt children. + * @mas: the maple state with the maple encoded node of the sub-tree. + * + * Descend through a sub-tree and adopt children who do not have the correct + * parents set. Follow the parents which have the correct parents as they are + * the new entries which need to be followed to find other incorrectly set + * parents. + */ +static inline void mas_descend_adopt(struct ma_state *mas) +{ + struct ma_state list[3], next[3]; + int i, n; + + /* + * At each level there may be up to 3 correct parent pointers which indicates + * the new nodes which need to be walked to find any new nodes at a lower level. + */ + + for (i = 0; i < 3; i++) { + list[i] = *mas; + list[i].offset = 0; + next[i].offset = 0; + } + next[0] = *mas; + + while (!mte_is_leaf(list[0].node)) { + n = 0; + for (i = 0; i < 3; i++) { + if (mas_is_none(&list[i])) + continue; + + if (i && list[i-1].node == list[i].node) + continue; + + while ((n < 3) && (mas_new_child(&list[i], &next[n]))) + n++; + + mas_adopt_children(&list[i], list[i].node); + } + + while (n < 3) + next[n++].node = MAS_NONE; + + /* descend by setting the list to the children */ + for (i = 0; i < 3; i++) + list[i] = next[i]; + } +} + +/* + * mas_bulk_rebalance() - Rebalance the end of a tree after a bulk insert. + * @mas: The maple state + * @end: The maple node end + * @mt: The maple node type + */ +static inline void mas_bulk_rebalance(struct ma_state *mas, unsigned char end, + enum maple_type mt) +{ + if (!(mas->mas_flags & MA_STATE_BULK)) + return; + + if (mte_is_root(mas->node)) + return; + + if (end > mt_min_slots[mt]) { + mas->mas_flags &= ~MA_STATE_REBALANCE; + return; + } +} + +/* + * mas_store_b_node() - Store an @entry into the b_node while also copying the + * data from a maple encoded node. + * @wr_mas: the maple write state + * @b_node: the maple_big_node to fill with data + * @offset_end: the offset to end copying + * + * Return: The actual end of the data stored in @b_node + */ +static inline void mas_store_b_node(struct ma_wr_state *wr_mas, + struct maple_big_node *b_node, unsigned char offset_end) +{ + unsigned char slot; + unsigned char b_end; + /* Possible underflow of piv will wrap back to 0 before use. */ + unsigned long piv; + struct ma_state *mas = wr_mas->mas; + + b_node->type = wr_mas->type; + b_end = 0; + slot = mas->offset; + if (slot) { + /* Copy start data up to insert. */ + mas_mab_cp(mas, 0, slot - 1, b_node, 0); + b_end = b_node->b_end; + piv = b_node->pivot[b_end - 1]; + } else + piv = mas->min - 1; + + if (piv + 1 < mas->index) { + /* Handle range starting after old range */ + b_node->slot[b_end] = wr_mas->content; + if (!wr_mas->content) + b_node->gap[b_end] = mas->index - 1 - piv; + b_node->pivot[b_end++] = mas->index - 1; + } + + /* Store the new entry. */ + mas->offset = b_end; + b_node->slot[b_end] = wr_mas->entry; + b_node->pivot[b_end] = mas->last; + + /* Appended. */ + if (mas->last >= mas->max) + goto b_end; + + /* Handle new range ending before old range ends */ + piv = mas_logical_pivot(mas, wr_mas->pivots, offset_end, wr_mas->type); + if (piv > mas->last) { + if (piv == ULONG_MAX) + mas_bulk_rebalance(mas, b_node->b_end, wr_mas->type); + + if (offset_end != slot) + wr_mas->content = mas_slot_locked(mas, wr_mas->slots, + offset_end); + + b_node->slot[++b_end] = wr_mas->content; + if (!wr_mas->content) + b_node->gap[b_end] = piv - mas->last + 1; + b_node->pivot[b_end] = piv; + } + + slot = offset_end + 1; + if (slot > wr_mas->node_end) + goto b_end; + + /* Copy end data to the end of the node. */ + mas_mab_cp(mas, slot, wr_mas->node_end + 1, b_node, ++b_end); + b_node->b_end--; + return; + +b_end: + b_node->b_end = b_end; +} + +/* + * mas_prev_sibling() - Find the previous node with the same parent. + * @mas: the maple state + * + * Return: True if there is a previous sibling, false otherwise. + */ +static inline bool mas_prev_sibling(struct ma_state *mas) +{ + unsigned int p_slot = mte_parent_slot(mas->node); + + if (mte_is_root(mas->node)) + return false; + + if (!p_slot) + return false; + + mas_ascend(mas); + mas->offset = p_slot - 1; + mas_descend(mas); + return true; +} + +/* + * mas_next_sibling() - Find the next node with the same parent. + * @mas: the maple state + * + * Return: true if there is a next sibling, false otherwise. + */ +static inline bool mas_next_sibling(struct ma_state *mas) +{ + MA_STATE(parent, mas->tree, mas->index, mas->last); + + if (mte_is_root(mas->node)) + return false; + + parent = *mas; + mas_ascend(&parent); + parent.offset = mte_parent_slot(mas->node) + 1; + if (parent.offset > mas_data_end(&parent)) + return false; + + *mas = parent; + mas_descend(mas); + return true; +} + +/* + * mte_node_or_node() - Return the encoded node or MAS_NONE. + * @enode: The encoded maple node. + * + * Shorthand to avoid setting %NULLs in the tree or maple_subtree_state. + * + * Return: @enode or MAS_NONE + */ +static inline struct maple_enode *mte_node_or_none(struct maple_enode *enode) +{ + if (enode) + return enode; + + return ma_enode_ptr(MAS_NONE); +} + +/* + * mas_wr_node_walk() - Find the correct offset for the index in the @mas. + * @wr_mas: The maple write state + * + * Uses mas_slot_locked() and does not need to worry about dead nodes. + */ +static inline void mas_wr_node_walk(struct ma_wr_state *wr_mas) +{ + struct ma_state *mas = wr_mas->mas; + unsigned char count; + unsigned char offset; + unsigned long index, min, max; + + if (unlikely(ma_is_dense(wr_mas->type))) { + wr_mas->r_max = wr_mas->r_min = mas->index; + mas->offset = mas->index = mas->min; + return; + } + + wr_mas->node = mas_mn(wr_mas->mas); + wr_mas->pivots = ma_pivots(wr_mas->node, wr_mas->type); + count = wr_mas->node_end = ma_data_end(wr_mas->node, wr_mas->type, + wr_mas->pivots, mas->max); + offset = mas->offset; + min = mas_safe_min(mas, wr_mas->pivots, offset); + max = wr_mas->pivots[offset]; + if (unlikely(offset == count)) + goto max; /* may have been set to zero */ + + index = mas->index; + if (unlikely(index <= max)) + goto done; + + if (unlikely(!max && offset)) + goto max; + + min = max + 1; + while (++offset < count) { + max = wr_mas->pivots[offset]; + if (index <= max) + goto done; + else if (unlikely(!max)) + break; + + min = max + 1; + } + +max: + max = mas->max; +done: + wr_mas->r_max = max; + wr_mas->r_min = min; + wr_mas->offset_end = mas->offset = offset; +} + +/* + * mas_topiary_range() - Add a range of slots to the topiary. + * @mas: The maple state + * @destroy: The topiary to add the slots (usually destroy) + * @start: The starting slot inclusively + * @end: The end slot inclusively + */ +static inline void mas_topiary_range(struct ma_state *mas, + struct ma_topiary *destroy, unsigned char start, unsigned char end) +{ + void __rcu **slots; + unsigned offset; + + MT_BUG_ON(mas->tree, mte_is_leaf(mas->node)); + slots = ma_slots(mas_mn(mas), mte_node_type(mas->node)); + for (offset = start; offset <= end; offset++) { + struct maple_enode *enode = mas_slot_locked(mas, slots, offset); + + if (mte_dead_node(enode)) + continue; + + mat_add(destroy, enode); + } +} + +/* + * mast_topiary() - Add the portions of the tree to the removal list; either to + * be freed or discarded (destroy walk). + * @mast: The maple_subtree_state. + */ +static inline void mast_topiary(struct maple_subtree_state *mast) +{ + MA_WR_STATE(wr_mas, mast->orig_l, NULL); + unsigned char r_start, r_end; + unsigned char l_start, l_end; + void **l_slots, **r_slots; + + wr_mas.type = mte_node_type(mast->orig_l->node); + mast->orig_l->index = mast->orig_l->last; + mas_wr_node_walk(&wr_mas); + l_start = mast->orig_l->offset + 1; + l_end = mas_data_end(mast->orig_l); + r_start = 0; + r_end = mast->orig_r->offset; + + if (r_end) + r_end--; + + l_slots = ma_slots(mas_mn(mast->orig_l), + mte_node_type(mast->orig_l->node)); + + r_slots = ma_slots(mas_mn(mast->orig_r), + mte_node_type(mast->orig_r->node)); + + if ((l_start < l_end) && + mte_dead_node(mas_slot_locked(mast->orig_l, l_slots, l_start))) { + l_start++; + } + + if (mte_dead_node(mas_slot_locked(mast->orig_r, r_slots, r_end))) { + if (r_end) + r_end--; + } + + if ((l_start > r_end) && (mast->orig_l->node == mast->orig_r->node)) + return; + + /* At the node where left and right sides meet, add the parts between */ + if (mast->orig_l->node == mast->orig_r->node) { + return mas_topiary_range(mast->orig_l, mast->destroy, + l_start, r_end); + } + + /* mast->orig_r is different and consumed. */ + if (mte_is_leaf(mast->orig_r->node)) + return; + + if (mte_dead_node(mas_slot_locked(mast->orig_l, l_slots, l_end))) + l_end--; + + + if (l_start <= l_end) + mas_topiary_range(mast->orig_l, mast->destroy, l_start, l_end); + + if (mte_dead_node(mas_slot_locked(mast->orig_r, r_slots, r_start))) + r_start++; + + if (r_start <= r_end) + mas_topiary_range(mast->orig_r, mast->destroy, 0, r_end); +} + +/* + * mast_rebalance_next() - Rebalance against the next node + * @mast: The maple subtree state + * @old_r: The encoded maple node to the right (next node). + */ +static inline void mast_rebalance_next(struct maple_subtree_state *mast) +{ + unsigned char b_end = mast->bn->b_end; + + mas_mab_cp(mast->orig_r, 0, mt_slot_count(mast->orig_r->node), + mast->bn, b_end); + mast->orig_r->last = mast->orig_r->max; +} + +/* + * mast_rebalance_prev() - Rebalance against the previous node + * @mast: The maple subtree state + * @old_l: The encoded maple node to the left (previous node) + */ +static inline void mast_rebalance_prev(struct maple_subtree_state *mast) +{ + unsigned char end = mas_data_end(mast->orig_l) + 1; + unsigned char b_end = mast->bn->b_end; + + mab_shift_right(mast->bn, end); + mas_mab_cp(mast->orig_l, 0, end - 1, mast->bn, 0); + mast->l->min = mast->orig_l->min; + mast->orig_l->index = mast->orig_l->min; + mast->bn->b_end = end + b_end; + mast->l->offset += end; +} + +/* + * mast_spanning_rebalance() - Rebalance nodes with nearest neighbour favouring + * the node to the right. Checking the nodes to the right then the left at each + * level upwards until root is reached. Free and destroy as needed. + * Data is copied into the @mast->bn. + * @mast: The maple_subtree_state. + */ +static inline +bool mast_spanning_rebalance(struct maple_subtree_state *mast) +{ + struct ma_state r_tmp = *mast->orig_r; + struct ma_state l_tmp = *mast->orig_l; + struct maple_enode *ancestor = NULL; + unsigned char start, end; + unsigned char depth = 0; + + r_tmp = *mast->orig_r; + l_tmp = *mast->orig_l; + do { + mas_ascend(mast->orig_r); + mas_ascend(mast->orig_l); + depth++; + if (!ancestor && + (mast->orig_r->node == mast->orig_l->node)) { + ancestor = mast->orig_r->node; + end = mast->orig_r->offset - 1; + start = mast->orig_l->offset + 1; + } + + if (mast->orig_r->offset < mas_data_end(mast->orig_r)) { + if (!ancestor) { + ancestor = mast->orig_r->node; + start = 0; + } + + mast->orig_r->offset++; + do { + mas_descend(mast->orig_r); + mast->orig_r->offset = 0; + depth--; + } while (depth); + + mast_rebalance_next(mast); + do { + unsigned char l_off = 0; + struct maple_enode *child = r_tmp.node; + + mas_ascend(&r_tmp); + if (ancestor == r_tmp.node) + l_off = start; + + if (r_tmp.offset) + r_tmp.offset--; + + if (l_off < r_tmp.offset) + mas_topiary_range(&r_tmp, mast->destroy, + l_off, r_tmp.offset); + + if (l_tmp.node != child) + mat_add(mast->free, child); + + } while (r_tmp.node != ancestor); + + *mast->orig_l = l_tmp; + return true; + + } else if (mast->orig_l->offset != 0) { + if (!ancestor) { + ancestor = mast->orig_l->node; + end = mas_data_end(mast->orig_l); + } + + mast->orig_l->offset--; + do { + mas_descend(mast->orig_l); + mast->orig_l->offset = + mas_data_end(mast->orig_l); + depth--; + } while (depth); + + mast_rebalance_prev(mast); + do { + unsigned char r_off; + struct maple_enode *child = l_tmp.node; + + mas_ascend(&l_tmp); + if (ancestor == l_tmp.node) + r_off = end; + else + r_off = mas_data_end(&l_tmp); + + if (l_tmp.offset < r_off) + l_tmp.offset++; + + if (l_tmp.offset < r_off) + mas_topiary_range(&l_tmp, mast->destroy, + l_tmp.offset, r_off); + + if (r_tmp.node != child) + mat_add(mast->free, child); + + } while (l_tmp.node != ancestor); + + *mast->orig_r = r_tmp; + return true; + } + } while (!mte_is_root(mast->orig_r->node)); + + *mast->orig_r = r_tmp; + *mast->orig_l = l_tmp; + return false; +} + +/* + * mast_ascend_free() - Add current original maple state nodes to the free list + * and ascend. + * @mast: the maple subtree state. + * + * Ascend the original left and right sides and add the previous nodes to the + * free list. Set the slots to point to the correct location in the new nodes. + */ +static inline void +mast_ascend_free(struct maple_subtree_state *mast) +{ + MA_WR_STATE(wr_mas, mast->orig_r, NULL); + struct maple_enode *left = mast->orig_l->node; + struct maple_enode *right = mast->orig_r->node; + + mas_ascend(mast->orig_l); + mas_ascend(mast->orig_r); + mat_add(mast->free, left); + + if (left != right) + mat_add(mast->free, right); + + mast->orig_r->offset = 0; + mast->orig_r->index = mast->r->max; + /* last should be larger than or equal to index */ + if (mast->orig_r->last < mast->orig_r->index) + mast->orig_r->last = mast->orig_r->index; + /* + * The node may not contain the value so set slot to ensure all + * of the nodes contents are freed or destroyed. + */ + wr_mas.type = mte_node_type(mast->orig_r->node); + mas_wr_node_walk(&wr_mas); + /* Set up the left side of things */ + mast->orig_l->offset = 0; + mast->orig_l->index = mast->l->min; + wr_mas.mas = mast->orig_l; + wr_mas.type = mte_node_type(mast->orig_l->node); + mas_wr_node_walk(&wr_mas); + + mast->bn->type = wr_mas.type; +} + +/* + * mas_new_ma_node() - Create and return a new maple node. Helper function. + * @mas: the maple state with the allocations. + * @b_node: the maple_big_node with the type encoding. + * + * Use the node type from the maple_big_node to allocate a new node from the + * ma_state. This function exists mainly for code readability. + * + * Return: A new maple encoded node + */ +static inline struct maple_enode +*mas_new_ma_node(struct ma_state *mas, struct maple_big_node *b_node) +{ + return mt_mk_node(ma_mnode_ptr(mas_pop_node(mas)), b_node->type); +} + +/* + * mas_mab_to_node() - Set up right and middle nodes + * + * @mas: the maple state that contains the allocations. + * @b_node: the node which contains the data. + * @left: The pointer which will have the left node + * @right: The pointer which may have the right node + * @middle: the pointer which may have the middle node (rare) + * @mid_split: the split location for the middle node + * + * Return: the split of left. + */ +static inline unsigned char mas_mab_to_node(struct ma_state *mas, + struct maple_big_node *b_node, struct maple_enode **left, + struct maple_enode **right, struct maple_enode **middle, + unsigned char *mid_split, unsigned long min) +{ + unsigned char split = 0; + unsigned char slot_count = mt_slots[b_node->type]; + + *left = mas_new_ma_node(mas, b_node); + *right = NULL; + *middle = NULL; + *mid_split = 0; + + if (b_node->b_end < slot_count) { + split = b_node->b_end; + } else { + split = mab_calc_split(mas, b_node, mid_split, min); + *right = mas_new_ma_node(mas, b_node); + } + + if (*mid_split) + *middle = mas_new_ma_node(mas, b_node); + + return split; + +} + +/* + * mab_set_b_end() - Add entry to b_node at b_node->b_end and increment the end + * pointer. + * @b_node - the big node to add the entry + * @mas - the maple state to get the pivot (mas->max) + * @entry - the entry to add, if NULL nothing happens. + */ +static inline void mab_set_b_end(struct maple_big_node *b_node, + struct ma_state *mas, + void *entry) +{ + if (!entry) + return; + + b_node->slot[b_node->b_end] = entry; + if (mt_is_alloc(mas->tree)) + b_node->gap[b_node->b_end] = mas_max_gap(mas); + b_node->pivot[b_node->b_end++] = mas->max; +} + +/* + * mas_set_split_parent() - combine_then_separate helper function. Sets the parent + * of @mas->node to either @left or @right, depending on @slot and @split + * + * @mas - the maple state with the node that needs a parent + * @left - possible parent 1 + * @right - possible parent 2 + * @slot - the slot the mas->node was placed + * @split - the split location between @left and @right + */ +static inline void mas_set_split_parent(struct ma_state *mas, + struct maple_enode *left, + struct maple_enode *right, + unsigned char *slot, unsigned char split) +{ + if (mas_is_none(mas)) + return; + + if ((*slot) <= split) + mte_set_parent(mas->node, left, *slot); + else if (right) + mte_set_parent(mas->node, right, (*slot) - split - 1); + + (*slot)++; +} + +/* + * mte_mid_split_check() - Check if the next node passes the mid-split + * @**l: Pointer to left encoded maple node. + * @**m: Pointer to middle encoded maple node. + * @**r: Pointer to right encoded maple node. + * @slot: The offset + * @*split: The split location. + * @mid_split: The middle split. + */ +static inline void mte_mid_split_check(struct maple_enode **l, + struct maple_enode **r, + struct maple_enode *right, + unsigned char slot, + unsigned char *split, + unsigned char mid_split) +{ + if (*r == right) + return; + + if (slot < mid_split) + return; + + *l = *r; + *r = right; + *split = mid_split; +} + +/* + * mast_set_split_parents() - Helper function to set three nodes parents. Slot + * is taken from @mast->l. + * @mast - the maple subtree state + * @left - the left node + * @right - the right node + * @split - the split location. + */ +static inline void mast_set_split_parents(struct maple_subtree_state *mast, + struct maple_enode *left, + struct maple_enode *middle, + struct maple_enode *right, + unsigned char split, + unsigned char mid_split) +{ + unsigned char slot; + struct maple_enode *l = left; + struct maple_enode *r = right; + + if (mas_is_none(mast->l)) + return; + + if (middle) + r = middle; + + slot = mast->l->offset; + + mte_mid_split_check(&l, &r, right, slot, &split, mid_split); + mas_set_split_parent(mast->l, l, r, &slot, split); + + mte_mid_split_check(&l, &r, right, slot, &split, mid_split); + mas_set_split_parent(mast->m, l, r, &slot, split); + + mte_mid_split_check(&l, &r, right, slot, &split, mid_split); + mas_set_split_parent(mast->r, l, r, &slot, split); +} + +/* + * mas_wmb_replace() - Write memory barrier and replace + * @mas: The maple state + * @free: the maple topiary list of nodes to free + * @destroy: The maple topiary list of nodes to destroy (walk and free) + * + * Updates gap as necessary. + */ +static inline void mas_wmb_replace(struct ma_state *mas, + struct ma_topiary *free, + struct ma_topiary *destroy) +{ + /* All nodes must see old data as dead prior to replacing that data */ + smp_wmb(); /* Needed for RCU */ + + /* Insert the new data in the tree */ + mas_replace(mas, true); + + if (!mte_is_leaf(mas->node)) + mas_descend_adopt(mas); + + mas_mat_free(mas, free); + + if (destroy) + mas_mat_destroy(mas, destroy); + + if (mte_is_leaf(mas->node)) + return; + + mas_update_gap(mas); +} + +/* + * mast_new_root() - Set a new tree root during subtree creation + * @mast: The maple subtree state + * @mas: The maple state + */ +static inline void mast_new_root(struct maple_subtree_state *mast, + struct ma_state *mas) +{ + mas_mn(mast->l)->parent = + ma_parent_ptr(((unsigned long)mas->tree | MA_ROOT_PARENT)); + if (!mte_dead_node(mast->orig_l->node) && + !mte_is_root(mast->orig_l->node)) { + do { + mast_ascend_free(mast); + mast_topiary(mast); + } while (!mte_is_root(mast->orig_l->node)); + } + if ((mast->orig_l->node != mas->node) && + (mast->l->depth > mas_mt_height(mas))) { + mat_add(mast->free, mas->node); + } +} + +/* + * mast_cp_to_nodes() - Copy data out to nodes. + * @mast: The maple subtree state + * @left: The left encoded maple node + * @middle: The middle encoded maple node + * @right: The right encoded maple node + * @split: The location to split between left and (middle ? middle : right) + * @mid_split: The location to split between middle and right. + */ +static inline void mast_cp_to_nodes(struct maple_subtree_state *mast, + struct maple_enode *left, struct maple_enode *middle, + struct maple_enode *right, unsigned char split, unsigned char mid_split) +{ + bool new_lmax = true; + + mast->l->node = mte_node_or_none(left); + mast->m->node = mte_node_or_none(middle); + mast->r->node = mte_node_or_none(right); + + mast->l->min = mast->orig_l->min; + if (split == mast->bn->b_end) { + mast->l->max = mast->orig_r->max; + new_lmax = false; + } + + mab_mas_cp(mast->bn, 0, split, mast->l, new_lmax); + + if (middle) { + mab_mas_cp(mast->bn, 1 + split, mid_split, mast->m, true); + mast->m->min = mast->bn->pivot[split] + 1; + split = mid_split; + } + + mast->r->max = mast->orig_r->max; + if (right) { + mab_mas_cp(mast->bn, 1 + split, mast->bn->b_end, mast->r, false); + mast->r->min = mast->bn->pivot[split] + 1; + } +} + +/* + * mast_combine_cp_left - Copy in the original left side of the tree into the + * combined data set in the maple subtree state big node. + * @mast: The maple subtree state + */ +static inline void mast_combine_cp_left(struct maple_subtree_state *mast) +{ + unsigned char l_slot = mast->orig_l->offset; + + if (!l_slot) + return; + + mas_mab_cp(mast->orig_l, 0, l_slot - 1, mast->bn, 0); +} + +/* + * mast_combine_cp_right: Copy in the original right side of the tree into the + * combined data set in the maple subtree state big node. + * @mast: The maple subtree state + */ +static inline void mast_combine_cp_right(struct maple_subtree_state *mast) +{ + if (mast->bn->pivot[mast->bn->b_end - 1] >= mast->orig_r->max) + return; + + mas_mab_cp(mast->orig_r, mast->orig_r->offset + 1, + mt_slot_count(mast->orig_r->node), mast->bn, + mast->bn->b_end); + mast->orig_r->last = mast->orig_r->max; +} + +/* + * mast_sufficient: Check if the maple subtree state has enough data in the big + * node to create at least one sufficient node + * @mast: the maple subtree state + */ +static inline bool mast_sufficient(struct maple_subtree_state *mast) +{ + if (mast->bn->b_end > mt_min_slot_count(mast->orig_l->node)) + return true; + + return false; +} + +/* + * mast_overflow: Check if there is too much data in the subtree state for a + * single node. + * @mast: The maple subtree state + */ +static inline bool mast_overflow(struct maple_subtree_state *mast) +{ + if (mast->bn->b_end >= mt_slot_count(mast->orig_l->node)) + return true; + + return false; +} + +static inline void *mtree_range_walk(struct ma_state *mas) +{ + unsigned long *pivots; + unsigned char offset; + struct maple_node *node; + struct maple_enode *next, *last; + enum maple_type type; + void __rcu **slots; + unsigned char end; + unsigned long max, min; + unsigned long prev_max, prev_min; + + last = next = mas->node; + prev_min = min = mas->min; + max = mas->max; + do { + offset = 0; + last = next; + node = mte_to_node(next); + type = mte_node_type(next); + pivots = ma_pivots(node, type); + end = ma_data_end(node, type, pivots, max); + if (unlikely(ma_dead_node(node))) + goto dead_node; + + if (pivots[offset] >= mas->index) { + prev_max = max; + prev_min = min; + max = pivots[offset]; + goto next; + } + + do { + offset++; + } while ((offset < end) && (pivots[offset] < mas->index)); + + prev_min = min; + min = pivots[offset - 1] + 1; + prev_max = max; + if (likely(offset < end && pivots[offset])) + max = pivots[offset]; + +next: + slots = ma_slots(node, type); + next = mt_slot(mas->tree, slots, offset); + if (unlikely(ma_dead_node(node))) + goto dead_node; + } while (!ma_is_leaf(type)); + + mas->offset = offset; + mas->index = min; + mas->last = max; + mas->min = prev_min; + mas->max = prev_max; + mas->node = last; + return (void *) next; + +dead_node: + mas_reset(mas); + return NULL; +} + +/* + * mas_spanning_rebalance() - Rebalance across two nodes which may not be peers. + * @mas: The starting maple state + * @mast: The maple_subtree_state, keeps track of 4 maple states. + * @count: The estimated count of iterations needed. + * + * Follow the tree upwards from @l_mas and @r_mas for @count, or until the root + * is hit. First @b_node is split into two entries which are inserted into the + * next iteration of the loop. @b_node is returned populated with the final + * iteration. @mas is used to obtain allocations. orig_l_mas keeps track of the + * nodes that will remain active by using orig_l_mas->index and orig_l_mas->last + * to account of what has been copied into the new sub-tree. The update of + * orig_l_mas->last is used in mas_consume to find the slots that will need to + * be either freed or destroyed. orig_l_mas->depth keeps track of the height of + * the new sub-tree in case the sub-tree becomes the full tree. + * + * Return: the number of elements in b_node during the last loop. + */ +static int mas_spanning_rebalance(struct ma_state *mas, + struct maple_subtree_state *mast, unsigned char count) +{ + unsigned char split, mid_split; + unsigned char slot = 0; + struct maple_enode *left = NULL, *middle = NULL, *right = NULL; + + MA_STATE(l_mas, mas->tree, mas->index, mas->index); + MA_STATE(r_mas, mas->tree, mas->index, mas->last); + MA_STATE(m_mas, mas->tree, mas->index, mas->index); + MA_TOPIARY(free, mas->tree); + MA_TOPIARY(destroy, mas->tree); + + /* + * The tree needs to be rebalanced and leaves need to be kept at the same level. + * Rebalancing is done by use of the ``struct maple_topiary``. + */ + mast->l = &l_mas; + mast->m = &m_mas; + mast->r = &r_mas; + mast->free = &free; + mast->destroy = &destroy; + l_mas.node = r_mas.node = m_mas.node = MAS_NONE; + if (!mas_is_root_limits(mast->orig_l) && + unlikely(mast->bn->b_end <= mt_min_slots[mast->bn->type])) + mast_spanning_rebalance(mast); + + mast->orig_l->depth = 0; + + /* + * Each level of the tree is examined and balanced, pushing data to the left or + * right, or rebalancing against left or right nodes is employed to avoid + * rippling up the tree to limit the amount of churn. Once a new sub-section of + * the tree is created, there may be a mix of new and old nodes. The old nodes + * will have the incorrect parent pointers and currently be in two trees: the + * original tree and the partially new tree. To remedy the parent pointers in + * the old tree, the new data is swapped into the active tree and a walk down + * the tree is performed and the parent pointers are updated. + * See mas_descend_adopt() for more information.. + */ + while (count--) { + mast->bn->b_end--; + mast->bn->type = mte_node_type(mast->orig_l->node); + split = mas_mab_to_node(mas, mast->bn, &left, &right, &middle, + &mid_split, mast->orig_l->min); + mast_set_split_parents(mast, left, middle, right, split, + mid_split); + mast_cp_to_nodes(mast, left, middle, right, split, mid_split); + + /* + * Copy data from next level in the tree to mast->bn from next + * iteration + */ + memset(mast->bn, 0, sizeof(struct maple_big_node)); + mast->bn->type = mte_node_type(left); + mast->orig_l->depth++; + + /* Root already stored in l->node. */ + if (mas_is_root_limits(mast->l)) + goto new_root; + + mast_ascend_free(mast); + mast_combine_cp_left(mast); + l_mas.offset = mast->bn->b_end; + mab_set_b_end(mast->bn, &l_mas, left); + mab_set_b_end(mast->bn, &m_mas, middle); + mab_set_b_end(mast->bn, &r_mas, right); + + /* Copy anything necessary out of the right node. */ + mast_combine_cp_right(mast); + mast_topiary(mast); + mast->orig_l->last = mast->orig_l->max; + + if (mast_sufficient(mast)) + continue; + + if (mast_overflow(mast)) + continue; + + /* May be a new root stored in mast->bn */ + if (mas_is_root_limits(mast->orig_l)) + break; + + if (!mast_spanning_rebalance(mast)) + break; + + /* rebalancing from other nodes may require another loop. */ + if (!count) + count++; + } + + l_mas.node = mt_mk_node(ma_mnode_ptr(mas_pop_node(mas)), + mte_node_type(mast->orig_l->node)); + mast->orig_l->depth++; + mab_mas_cp(mast->bn, 0, mt_slots[mast->bn->type] - 1, &l_mas, true); + mte_set_parent(left, l_mas.node, slot); + if (middle) + mte_set_parent(middle, l_mas.node, ++slot); + + if (right) + mte_set_parent(right, l_mas.node, ++slot); + + if (mas_is_root_limits(mast->l)) { +new_root: + mast_new_root(mast, mas); + } else { + mas_mn(&l_mas)->parent = mas_mn(mast->orig_l)->parent; + } + + if (!mte_dead_node(mast->orig_l->node)) + mat_add(&free, mast->orig_l->node); + + mas->depth = mast->orig_l->depth; + *mast->orig_l = l_mas; + mte_set_node_dead(mas->node); + + /* Set up mas for insertion. */ + mast->orig_l->depth = mas->depth; + mast->orig_l->alloc = mas->alloc; + *mas = *mast->orig_l; + mas_wmb_replace(mas, &free, &destroy); + mtree_range_walk(mas); + return mast->bn->b_end; +} + +/* + * mas_rebalance() - Rebalance a given node. + * @mas: The maple state + * @b_node: The big maple node. + * + * Rebalance two nodes into a single node or two new nodes that are sufficient. + * Continue upwards until tree is sufficient. + * + * Return: the number of elements in b_node during the last loop. + */ +static inline int mas_rebalance(struct ma_state *mas, + struct maple_big_node *b_node) +{ + char empty_count = mas_mt_height(mas); + struct maple_subtree_state mast; + unsigned char shift, b_end = ++b_node->b_end; + + MA_STATE(l_mas, mas->tree, mas->index, mas->last); + MA_STATE(r_mas, mas->tree, mas->index, mas->last); + + trace_ma_op(__func__, mas); + + /* + * Rebalancing occurs if a node is insufficient. Data is rebalanced + * against the node to the right if it exists, otherwise the node to the + * left of this node is rebalanced against this node. If rebalancing + * causes just one node to be produced instead of two, then the parent + * is also examined and rebalanced if it is insufficient. Every level + * tries to combine the data in the same way. If one node contains the + * entire range of the tree, then that node is used as a new root node. + */ + mas_node_count(mas, 1 + empty_count * 3); + if (mas_is_err(mas)) + return 0; + + mast.orig_l = &l_mas; + mast.orig_r = &r_mas; + mast.bn = b_node; + mast.bn->type = mte_node_type(mas->node); + + l_mas = r_mas = *mas; + + if (mas_next_sibling(&r_mas)) { + mas_mab_cp(&r_mas, 0, mt_slot_count(r_mas.node), b_node, b_end); + r_mas.last = r_mas.index = r_mas.max; + } else { + mas_prev_sibling(&l_mas); + shift = mas_data_end(&l_mas) + 1; + mab_shift_right(b_node, shift); + mas->offset += shift; + mas_mab_cp(&l_mas, 0, shift - 1, b_node, 0); + b_node->b_end = shift + b_end; + l_mas.index = l_mas.last = l_mas.min; + } + + return mas_spanning_rebalance(mas, &mast, empty_count); +} + +/* + * mas_destroy_rebalance() - Rebalance left-most node while destroying the maple + * state. + * @mas: The maple state + * @end: The end of the left-most node. + * + * During a mass-insert event (such as forking), it may be necessary to + * rebalance the left-most node when it is not sufficient. + */ +static inline void mas_destroy_rebalance(struct ma_state *mas, unsigned char end) +{ + enum maple_type mt = mte_node_type(mas->node); + struct maple_node reuse, *newnode, *parent, *new_left, *left, *node; + struct maple_enode *eparent; + unsigned char offset, tmp, split = mt_slots[mt] / 2; + void __rcu **l_slots, **slots; + unsigned long *l_pivs, *pivs, gap; + bool in_rcu = mt_in_rcu(mas->tree); + + MA_STATE(l_mas, mas->tree, mas->index, mas->last); + + l_mas = *mas; + mas_prev_sibling(&l_mas); + + /* set up node. */ + if (in_rcu) { + /* Allocate for both left and right as well as parent. */ + mas_node_count(mas, 3); + if (mas_is_err(mas)) + return; + + newnode = mas_pop_node(mas); + } else { + newnode = &reuse; + } + + node = mas_mn(mas); + newnode->parent = node->parent; + slots = ma_slots(newnode, mt); + pivs = ma_pivots(newnode, mt); + left = mas_mn(&l_mas); + l_slots = ma_slots(left, mt); + l_pivs = ma_pivots(left, mt); + if (!l_slots[split]) + split++; + tmp = mas_data_end(&l_mas) - split; + + memcpy(slots, l_slots + split + 1, sizeof(void *) * tmp); + memcpy(pivs, l_pivs + split + 1, sizeof(unsigned long) * tmp); + pivs[tmp] = l_mas.max; + memcpy(slots + tmp, ma_slots(node, mt), sizeof(void *) * end); + memcpy(pivs + tmp, ma_pivots(node, mt), sizeof(unsigned long) * end); + + l_mas.max = l_pivs[split]; + mas->min = l_mas.max + 1; + eparent = mt_mk_node(mte_parent(l_mas.node), + mas_parent_enum(&l_mas, l_mas.node)); + tmp += end; + if (!in_rcu) { + unsigned char max_p = mt_pivots[mt]; + unsigned char max_s = mt_slots[mt]; + + if (tmp < max_p) + memset(pivs + tmp, 0, + sizeof(unsigned long *) * (max_p - tmp)); + + if (tmp < mt_slots[mt]) + memset(slots + tmp, 0, sizeof(void *) * (max_s - tmp)); + + memcpy(node, newnode, sizeof(struct maple_node)); + ma_set_meta(node, mt, 0, tmp - 1); + mte_set_pivot(eparent, mte_parent_slot(l_mas.node), + l_pivs[split]); + + /* Remove data from l_pivs. */ + tmp = split + 1; + memset(l_pivs + tmp, 0, sizeof(unsigned long) * (max_p - tmp)); + memset(l_slots + tmp, 0, sizeof(void *) * (max_s - tmp)); + ma_set_meta(left, mt, 0, split); + + goto done; + } + + /* RCU requires replacing both l_mas, mas, and parent. */ + mas->node = mt_mk_node(newnode, mt); + ma_set_meta(newnode, mt, 0, tmp); + + new_left = mas_pop_node(mas); + new_left->parent = left->parent; + mt = mte_node_type(l_mas.node); + slots = ma_slots(new_left, mt); + pivs = ma_pivots(new_left, mt); + memcpy(slots, l_slots, sizeof(void *) * split); + memcpy(pivs, l_pivs, sizeof(unsigned long) * split); + ma_set_meta(new_left, mt, 0, split); + l_mas.node = mt_mk_node(new_left, mt); + + /* replace parent. */ + offset = mte_parent_slot(mas->node); + mt = mas_parent_enum(&l_mas, l_mas.node); + parent = mas_pop_node(mas); + slots = ma_slots(parent, mt); + pivs = ma_pivots(parent, mt); + memcpy(parent, mte_to_node(eparent), sizeof(struct maple_node)); + rcu_assign_pointer(slots[offset], mas->node); + rcu_assign_pointer(slots[offset - 1], l_mas.node); + pivs[offset - 1] = l_mas.max; + eparent = mt_mk_node(parent, mt); +done: + gap = mas_leaf_max_gap(mas); + mte_set_gap(eparent, mte_parent_slot(mas->node), gap); + gap = mas_leaf_max_gap(&l_mas); + mte_set_gap(eparent, mte_parent_slot(l_mas.node), gap); + mas_ascend(mas); + + if (in_rcu) + mas_replace(mas, false); + + mas_update_gap(mas); +} + +/* + * mas_split_final_node() - Split the final node in a subtree operation. + * @mast: the maple subtree state + * @mas: The maple state + * @height: The height of the tree in case it's a new root. + */ +static inline bool mas_split_final_node(struct maple_subtree_state *mast, + struct ma_state *mas, int height) +{ + struct maple_enode *ancestor; + + if (mte_is_root(mas->node)) { + if (mt_is_alloc(mas->tree)) + mast->bn->type = maple_arange_64; + else + mast->bn->type = maple_range_64; + mas->depth = height; + } + /* + * Only a single node is used here, could be root. + * The Big_node data should just fit in a single node. + */ + ancestor = mas_new_ma_node(mas, mast->bn); + mte_set_parent(mast->l->node, ancestor, mast->l->offset); + mte_set_parent(mast->r->node, ancestor, mast->r->offset); + mte_to_node(ancestor)->parent = mas_mn(mas)->parent; + + mast->l->node = ancestor; + mab_mas_cp(mast->bn, 0, mt_slots[mast->bn->type] - 1, mast->l, true); + mas->offset = mast->bn->b_end - 1; + return true; +} + +/* + * mast_fill_bnode() - Copy data into the big node in the subtree state + * @mast: The maple subtree state + * @mas: the maple state + * @skip: The number of entries to skip for new nodes insertion. + */ +static inline void mast_fill_bnode(struct maple_subtree_state *mast, + struct ma_state *mas, + unsigned char skip) +{ + bool cp = true; + struct maple_enode *old = mas->node; + unsigned char split; + + memset(mast->bn->gap, 0, sizeof(unsigned long) * ARRAY_SIZE(mast->bn->gap)); + memset(mast->bn->slot, 0, sizeof(unsigned long) * ARRAY_SIZE(mast->bn->slot)); + memset(mast->bn->pivot, 0, sizeof(unsigned long) * ARRAY_SIZE(mast->bn->pivot)); + mast->bn->b_end = 0; + + if (mte_is_root(mas->node)) { + cp = false; + } else { + mas_ascend(mas); + mat_add(mast->free, old); + mas->offset = mte_parent_slot(mas->node); + } + + if (cp && mast->l->offset) + mas_mab_cp(mas, 0, mast->l->offset - 1, mast->bn, 0); + + split = mast->bn->b_end; + mab_set_b_end(mast->bn, mast->l, mast->l->node); + mast->r->offset = mast->bn->b_end; + mab_set_b_end(mast->bn, mast->r, mast->r->node); + if (mast->bn->pivot[mast->bn->b_end - 1] == mas->max) + cp = false; + + if (cp) + mas_mab_cp(mas, split + skip, mt_slot_count(mas->node) - 1, + mast->bn, mast->bn->b_end); + + mast->bn->b_end--; + mast->bn->type = mte_node_type(mas->node); +} + +/* + * mast_split_data() - Split the data in the subtree state big node into regular + * nodes. + * @mast: The maple subtree state + * @mas: The maple state + * @split: The location to split the big node + */ +static inline void mast_split_data(struct maple_subtree_state *mast, + struct ma_state *mas, unsigned char split) +{ + unsigned char p_slot; + + mab_mas_cp(mast->bn, 0, split, mast->l, true); + mte_set_pivot(mast->r->node, 0, mast->r->max); + mab_mas_cp(mast->bn, split + 1, mast->bn->b_end, mast->r, false); + mast->l->offset = mte_parent_slot(mas->node); + mast->l->max = mast->bn->pivot[split]; + mast->r->min = mast->l->max + 1; + if (mte_is_leaf(mas->node)) + return; + + p_slot = mast->orig_l->offset; + mas_set_split_parent(mast->orig_l, mast->l->node, mast->r->node, + &p_slot, split); + mas_set_split_parent(mast->orig_r, mast->l->node, mast->r->node, + &p_slot, split); +} + +/* + * mas_push_data() - Instead of splitting a node, it is beneficial to push the + * data to the right or left node if there is room. + * @mas: The maple state + * @height: The current height of the maple state + * @mast: The maple subtree state + * @left: Push left or not. + * + * Keeping the height of the tree low means faster lookups. + * + * Return: True if pushed, false otherwise. + */ +static inline bool mas_push_data(struct ma_state *mas, int height, + struct maple_subtree_state *mast, bool left) +{ + unsigned char slot_total = mast->bn->b_end; + unsigned char end, space, split; + + MA_STATE(tmp_mas, mas->tree, mas->index, mas->last); + tmp_mas = *mas; + tmp_mas.depth = mast->l->depth; + + if (left && !mas_prev_sibling(&tmp_mas)) + return false; + else if (!left && !mas_next_sibling(&tmp_mas)) + return false; + + end = mas_data_end(&tmp_mas); + slot_total += end; + space = 2 * mt_slot_count(mas->node) - 2; + /* -2 instead of -1 to ensure there isn't a triple split */ + if (ma_is_leaf(mast->bn->type)) + space--; + + if (mas->max == ULONG_MAX) + space--; + + if (slot_total >= space) + return false; + + /* Get the data; Fill mast->bn */ + mast->bn->b_end++; + if (left) { + mab_shift_right(mast->bn, end + 1); + mas_mab_cp(&tmp_mas, 0, end, mast->bn, 0); + mast->bn->b_end = slot_total + 1; + } else { + mas_mab_cp(&tmp_mas, 0, end, mast->bn, mast->bn->b_end); + } + + /* Configure mast for splitting of mast->bn */ + split = mt_slots[mast->bn->type] - 2; + if (left) { + /* Switch mas to prev node */ + mat_add(mast->free, mas->node); + *mas = tmp_mas; + /* Start using mast->l for the left side. */ + tmp_mas.node = mast->l->node; + *mast->l = tmp_mas; + } else { + mat_add(mast->free, tmp_mas.node); + tmp_mas.node = mast->r->node; + *mast->r = tmp_mas; + split = slot_total - split; + } + split = mab_no_null_split(mast->bn, split, mt_slots[mast->bn->type]); + /* Update parent slot for split calculation. */ + if (left) + mast->orig_l->offset += end + 1; + + mast_split_data(mast, mas, split); + mast_fill_bnode(mast, mas, 2); + mas_split_final_node(mast, mas, height + 1); + return true; +} + +/* + * mas_split() - Split data that is too big for one node into two. + * @mas: The maple state + * @b_node: The maple big node + * Return: 1 on success, 0 on failure. + */ +static int mas_split(struct ma_state *mas, struct maple_big_node *b_node) +{ + + struct maple_subtree_state mast; + int height = 0; + unsigned char mid_split, split = 0; + + /* + * Splitting is handled differently from any other B-tree; the Maple + * Tree splits upwards. Splitting up means that the split operation + * occurs when the walk of the tree hits the leaves and not on the way + * down. The reason for splitting up is that it is impossible to know + * how much space will be needed until the leaf is (or leaves are) + * reached. Since overwriting data is allowed and a range could + * overwrite more than one range or result in changing one entry into 3 + * entries, it is impossible to know if a split is required until the + * data is examined. + * + * Splitting is a balancing act between keeping allocations to a minimum + * and avoiding a 'jitter' event where a tree is expanded to make room + * for an entry followed by a contraction when the entry is removed. To + * accomplish the balance, there are empty slots remaining in both left + * and right nodes after a split. + */ + MA_STATE(l_mas, mas->tree, mas->index, mas->last); + MA_STATE(r_mas, mas->tree, mas->index, mas->last); + MA_STATE(prev_l_mas, mas->tree, mas->index, mas->last); + MA_STATE(prev_r_mas, mas->tree, mas->index, mas->last); + MA_TOPIARY(mat, mas->tree); + + trace_ma_op(__func__, mas); + mas->depth = mas_mt_height(mas); + /* Allocation failures will happen early. */ + mas_node_count(mas, 1 + mas->depth * 2); + if (mas_is_err(mas)) + return 0; + + mast.l = &l_mas; + mast.r = &r_mas; + mast.orig_l = &prev_l_mas; + mast.orig_r = &prev_r_mas; + mast.free = &mat; + mast.bn = b_node; + + while (height++ <= mas->depth) { + if (mt_slots[b_node->type] > b_node->b_end) { + mas_split_final_node(&mast, mas, height); + break; + } + + l_mas = r_mas = *mas; + l_mas.node = mas_new_ma_node(mas, b_node); + r_mas.node = mas_new_ma_node(mas, b_node); + /* + * Another way that 'jitter' is avoided is to terminate a split up early if the + * left or right node has space to spare. This is referred to as "pushing left" + * or "pushing right" and is similar to the B* tree, except the nodes left or + * right can rarely be reused due to RCU, but the ripple upwards is halted which + * is a significant savings. + */ + /* Try to push left. */ + if (mas_push_data(mas, height, &mast, true)) + break; + + /* Try to push right. */ + if (mas_push_data(mas, height, &mast, false)) + break; + + split = mab_calc_split(mas, b_node, &mid_split, prev_l_mas.min); + mast_split_data(&mast, mas, split); + /* + * Usually correct, mab_mas_cp in the above call overwrites + * r->max. + */ + mast.r->max = mas->max; + mast_fill_bnode(&mast, mas, 1); + prev_l_mas = *mast.l; + prev_r_mas = *mast.r; + } + + /* Set the original node as dead */ + mat_add(mast.free, mas->node); + mas->node = l_mas.node; + mas_wmb_replace(mas, mast.free, NULL); + mtree_range_walk(mas); + return 1; +} + +/* + * mas_reuse_node() - Reuse the node to store the data. + * @wr_mas: The maple write state + * @bn: The maple big node + * @end: The end of the data. + * + * Will always return false in RCU mode. + * + * Return: True if node was reused, false otherwise. + */ +static inline bool mas_reuse_node(struct ma_wr_state *wr_mas, + struct maple_big_node *bn, unsigned char end) +{ + /* Need to be rcu safe. */ + if (mt_in_rcu(wr_mas->mas->tree)) + return false; + + if (end > bn->b_end) { + int clear = mt_slots[wr_mas->type] - bn->b_end; + + memset(wr_mas->slots + bn->b_end, 0, sizeof(void *) * clear--); + memset(wr_mas->pivots + bn->b_end, 0, sizeof(void *) * clear); + } + mab_mas_cp(bn, 0, bn->b_end, wr_mas->mas, false); + return true; +} + +/* + * mas_commit_b_node() - Commit the big node into the tree. + * @wr_mas: The maple write state + * @b_node: The maple big node + * @end: The end of the data. + */ +static inline int mas_commit_b_node(struct ma_wr_state *wr_mas, + struct maple_big_node *b_node, unsigned char end) +{ + struct maple_node *node; + unsigned char b_end = b_node->b_end; + enum maple_type b_type = b_node->type; + + if ((b_end < mt_min_slots[b_type]) && + (!mte_is_root(wr_mas->mas->node)) && + (mas_mt_height(wr_mas->mas) > 1)) + return mas_rebalance(wr_mas->mas, b_node); + + if (b_end >= mt_slots[b_type]) + return mas_split(wr_mas->mas, b_node); + + if (mas_reuse_node(wr_mas, b_node, end)) + goto reuse_node; + + mas_node_count(wr_mas->mas, 1); + if (mas_is_err(wr_mas->mas)) + return 0; + + node = mas_pop_node(wr_mas->mas); + node->parent = mas_mn(wr_mas->mas)->parent; + wr_mas->mas->node = mt_mk_node(node, b_type); + mab_mas_cp(b_node, 0, b_end, wr_mas->mas, true); + + mas_replace(wr_mas->mas, false); +reuse_node: + mas_update_gap(wr_mas->mas); + return 1; +} + +/* + * mas_root_expand() - Expand a root to a node + * @mas: The maple state + * @entry: The entry to store into the tree + */ +static inline int mas_root_expand(struct ma_state *mas, void *entry) +{ + void *contents = mas_root_locked(mas); + enum maple_type type = maple_leaf_64; + struct maple_node *node; + void __rcu **slots; + unsigned long *pivots; + int slot = 0; + + mas_node_count(mas, 1); + if (unlikely(mas_is_err(mas))) + return 0; + + node = mas_pop_node(mas); + pivots = ma_pivots(node, type); + slots = ma_slots(node, type); + node->parent = ma_parent_ptr( + ((unsigned long)mas->tree | MA_ROOT_PARENT)); + mas->node = mt_mk_node(node, type); + + if (mas->index) { + if (contents) { + rcu_assign_pointer(slots[slot], contents); + if (likely(mas->index > 1)) + slot++; + } + pivots[slot++] = mas->index - 1; + } + + rcu_assign_pointer(slots[slot], entry); + mas->offset = slot; + pivots[slot] = mas->last; + if (mas->last != ULONG_MAX) + slot++; + mas->depth = 1; + mas_set_height(mas); + + /* swap the new root into the tree */ + rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); + ma_set_meta(node, maple_leaf_64, 0, slot); + return slot; +} + +static inline void mas_store_root(struct ma_state *mas, void *entry) +{ + if (likely((mas->last != 0) || (mas->index != 0))) + mas_root_expand(mas, entry); + else if (((unsigned long) (entry) & 3) == 2) + mas_root_expand(mas, entry); + else { + rcu_assign_pointer(mas->tree->ma_root, entry); + mas->node = MAS_START; + } +} + +/* + * mas_is_span_wr() - Check if the write needs to be treated as a write that + * spans the node. + * @mas: The maple state + * @piv: The pivot value being written + * @type: The maple node type + * @entry: The data to write + * + * Spanning writes are writes that start in one node and end in another OR if + * the write of a %NULL will cause the node to end with a %NULL. + * + * Return: True if this is a spanning write, false otherwise. + */ +static bool mas_is_span_wr(struct ma_wr_state *wr_mas) +{ + unsigned long max; + unsigned long last = wr_mas->mas->last; + unsigned long piv = wr_mas->r_max; + enum maple_type type = wr_mas->type; + void *entry = wr_mas->entry; + + /* Contained in this pivot */ + if (piv > last) + return false; + + max = wr_mas->mas->max; + if (unlikely(ma_is_leaf(type))) { + /* Fits in the node, but may span slots. */ + if (last < max) + return false; + + /* Writes to the end of the node but not null. */ + if ((last == max) && entry) + return false; + + /* + * Writing ULONG_MAX is not a spanning write regardless of the + * value being written as long as the range fits in the node. + */ + if ((last == ULONG_MAX) && (last == max)) + return false; + } else if (piv == last) { + if (entry) + return false; + + /* Detect spanning store wr walk */ + if (last == ULONG_MAX) + return false; + } + + trace_ma_write(__func__, wr_mas->mas, piv, entry); + + return true; +} + +static inline void mas_wr_walk_descend(struct ma_wr_state *wr_mas) +{ + wr_mas->mas->depth++; + wr_mas->type = mte_node_type(wr_mas->mas->node); + mas_wr_node_walk(wr_mas); + wr_mas->slots = ma_slots(wr_mas->node, wr_mas->type); +} + +static inline void mas_wr_walk_traverse(struct ma_wr_state *wr_mas) +{ + wr_mas->mas->max = wr_mas->r_max; + wr_mas->mas->min = wr_mas->r_min; + wr_mas->mas->node = wr_mas->content; + wr_mas->mas->offset = 0; +} +/* + * mas_wr_walk() - Walk the tree for a write. + * @wr_mas: The maple write state + * + * Uses mas_slot_locked() and does not need to worry about dead nodes. + * + * Return: True if it's contained in a node, false on spanning write. + */ +static bool mas_wr_walk(struct ma_wr_state *wr_mas) +{ + struct ma_state *mas = wr_mas->mas; + + while (true) { + mas_wr_walk_descend(wr_mas); + if (unlikely(mas_is_span_wr(wr_mas))) + return false; + + wr_mas->content = mas_slot_locked(mas, wr_mas->slots, + mas->offset); + if (ma_is_leaf(wr_mas->type)) + return true; + + mas_wr_walk_traverse(wr_mas); + } + + return true; +} + +static bool mas_wr_walk_index(struct ma_wr_state *wr_mas) +{ + struct ma_state *mas = wr_mas->mas; + + while (true) { + mas_wr_walk_descend(wr_mas); + wr_mas->content = mas_slot_locked(mas, wr_mas->slots, + mas->offset); + if (ma_is_leaf(wr_mas->type)) + return true; + mas_wr_walk_traverse(wr_mas); + + } + return true; +} +/* + * mas_extend_spanning_null() - Extend a store of a %NULL to include surrounding %NULLs. + * @l_wr_mas: The left maple write state + * @r_wr_mas: The right maple write state + */ +static inline void mas_extend_spanning_null(struct ma_wr_state *l_wr_mas, + struct ma_wr_state *r_wr_mas) +{ + struct ma_state *r_mas = r_wr_mas->mas; + struct ma_state *l_mas = l_wr_mas->mas; + unsigned char l_slot; + + l_slot = l_mas->offset; + if (!l_wr_mas->content) + l_mas->index = l_wr_mas->r_min; + + if ((l_mas->index == l_wr_mas->r_min) && + (l_slot && + !mas_slot_locked(l_mas, l_wr_mas->slots, l_slot - 1))) { + if (l_slot > 1) + l_mas->index = l_wr_mas->pivots[l_slot - 2] + 1; + else + l_mas->index = l_mas->min; + + l_mas->offset = l_slot - 1; + } + + if (!r_wr_mas->content) { + if (r_mas->last < r_wr_mas->r_max) + r_mas->last = r_wr_mas->r_max; + r_mas->offset++; + } else if ((r_mas->last == r_wr_mas->r_max) && + (r_mas->last < r_mas->max) && + !mas_slot_locked(r_mas, r_wr_mas->slots, r_mas->offset + 1)) { + r_mas->last = mas_safe_pivot(r_mas, r_wr_mas->pivots, + r_wr_mas->type, r_mas->offset + 1); + r_mas->offset++; + } +} + +static inline void *mas_state_walk(struct ma_state *mas) +{ + void *entry; + + entry = mas_start(mas); + if (mas_is_none(mas)) + return NULL; + + if (mas_is_ptr(mas)) + return entry; + + return mtree_range_walk(mas); +} + +/* + * mtree_lookup_walk() - Internal quick lookup that does not keep maple state up + * to date. + * + * @mas: The maple state. + * + * Note: Leaves mas in undesirable state. + * Return: The entry for @mas->index or %NULL on dead node. + */ +static inline void *mtree_lookup_walk(struct ma_state *mas) +{ + unsigned long *pivots; + unsigned char offset; + struct maple_node *node; + struct maple_enode *next; + enum maple_type type; + void __rcu **slots; + unsigned char end; + unsigned long max; + + next = mas->node; + max = ULONG_MAX; + do { + offset = 0; + node = mte_to_node(next); + type = mte_node_type(next); + pivots = ma_pivots(node, type); + end = ma_data_end(node, type, pivots, max); + if (unlikely(ma_dead_node(node))) + goto dead_node; + + if (pivots[offset] >= mas->index) + goto next; + + do { + offset++; + } while ((offset < end) && (pivots[offset] < mas->index)); + + if (likely(offset > end)) + max = pivots[offset]; + +next: + slots = ma_slots(node, type); + next = mt_slot(mas->tree, slots, offset); + if (unlikely(ma_dead_node(node))) + goto dead_node; + } while (!ma_is_leaf(type)); + + return (void *) next; + +dead_node: + mas_reset(mas); + return NULL; +} + +/* + * mas_new_root() - Create a new root node that only contains the entry passed + * in. + * @mas: The maple state + * @entry: The entry to store. + * + * Only valid when the index == 0 and the last == ULONG_MAX + * + * Return 0 on error, 1 on success. + */ +static inline int mas_new_root(struct ma_state *mas, void *entry) +{ + struct maple_enode *root = mas_root_locked(mas); + enum maple_type type = maple_leaf_64; + struct maple_node *node; + void __rcu **slots; + unsigned long *pivots; + + if (!entry && !mas->index && mas->last == ULONG_MAX) { + mas->depth = 0; + mas_set_height(mas); + rcu_assign_pointer(mas->tree->ma_root, entry); + mas->node = MAS_START; + goto done; + } + + mas_node_count(mas, 1); + if (mas_is_err(mas)) + return 0; + + node = mas_pop_node(mas); + pivots = ma_pivots(node, type); + slots = ma_slots(node, type); + node->parent = ma_parent_ptr( + ((unsigned long)mas->tree | MA_ROOT_PARENT)); + mas->node = mt_mk_node(node, type); + rcu_assign_pointer(slots[0], entry); + pivots[0] = mas->last; + mas->depth = 1; + mas_set_height(mas); + rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node)); + +done: + if (xa_is_node(root)) + mte_destroy_walk(root, mas->tree); + + return 1; +} +/* + * mas_wr_spanning_store() - Create a subtree with the store operation completed + * and new nodes where necessary, then place the sub-tree in the actual tree. + * Note that mas is expected to point to the node which caused the store to + * span. + * @wr_mas: The maple write state + * + * Return: 0 on error, positive on success. + */ +static inline int mas_wr_spanning_store(struct ma_wr_state *wr_mas) +{ + struct maple_subtree_state mast; + struct maple_big_node b_node; + struct ma_state *mas; + unsigned char height; + + /* Left and Right side of spanning store */ + MA_STATE(l_mas, NULL, 0, 0); + MA_STATE(r_mas, NULL, 0, 0); + + MA_WR_STATE(r_wr_mas, &r_mas, wr_mas->entry); + MA_WR_STATE(l_wr_mas, &l_mas, wr_mas->entry); + + /* + * A store operation that spans multiple nodes is called a spanning + * store and is handled early in the store call stack by the function + * mas_is_span_wr(). When a spanning store is identified, the maple + * state is duplicated. The first maple state walks the left tree path + * to ``index``, the duplicate walks the right tree path to ``last``. + * The data in the two nodes are combined into a single node, two nodes, + * or possibly three nodes (see the 3-way split above). A ``NULL`` + * written to the last entry of a node is considered a spanning store as + * a rebalance is required for the operation to complete and an overflow + * of data may happen. + */ + mas = wr_mas->mas; + trace_ma_op(__func__, mas); + + if (unlikely(!mas->index && mas->last == ULONG_MAX)) + return mas_new_root(mas, wr_mas->entry); + /* + * Node rebalancing may occur due to this store, so there may be two new + * entries per level plus a new root. + */ + height = mas_mt_height(mas); + mas_node_count(mas, 1 + height * 3); + if (mas_is_err(mas)) + return 0; + + /* + * Set up right side. Need to get to the next offset after the spanning + * store to ensure it's not NULL and to combine both the next node and + * the node with the start together. + */ + r_mas = *mas; + /* Avoid overflow, walk to next slot in the tree. */ + if (r_mas.last + 1) + r_mas.last++; + + r_mas.index = r_mas.last; + mas_wr_walk_index(&r_wr_mas); + r_mas.last = r_mas.index = mas->last; + + /* Set up left side. */ + l_mas = *mas; + mas_wr_walk_index(&l_wr_mas); + + if (!wr_mas->entry) { + mas_extend_spanning_null(&l_wr_mas, &r_wr_mas); + mas->offset = l_mas.offset; + mas->index = l_mas.index; + mas->last = l_mas.last = r_mas.last; + } + + memset(&b_node, 0, sizeof(struct maple_big_node)); + /* Copy l_mas and store the value in b_node. */ + mas_store_b_node(&l_wr_mas, &b_node, l_wr_mas.node_end); + /* Copy r_mas into b_node. */ + if (r_mas.offset <= r_wr_mas.node_end) + mas_mab_cp(&r_mas, r_mas.offset, r_wr_mas.node_end, + &b_node, b_node.b_end + 1); + /* Stop spanning searches by searching for just index. */ + l_mas.index = l_mas.last = mas->index; + + mast.bn = &b_node; + mast.orig_l = &l_mas; + mast.orig_r = &r_mas; + /* Combine l_mas and r_mas and split them up evenly again. */ + return mas_spanning_rebalance(mas, &mast, height + 1); +} + +/* + * mas_wr_node_store() - Attempt to store the value in a node + * @wr_mas: The maple write state + * + * Attempts to reuse the node, but may allocate. + * + * Return: True if stored, false otherwise + */ +static inline bool mas_wr_node_store(struct ma_wr_state *wr_mas) +{ + struct ma_state *mas = wr_mas->mas; + void __rcu **dst_slots; + unsigned long *dst_pivots; + unsigned char dst_offset; + unsigned char new_end = wr_mas->node_end; + unsigned char offset; + unsigned char node_slots = mt_slots[wr_mas->type]; + struct maple_node reuse, *newnode; + unsigned char copy_size, max_piv = mt_pivots[wr_mas->type]; + bool in_rcu = mt_in_rcu(mas->tree); + + offset = mas->offset; + if (mas->last == wr_mas->r_max) { + /* runs right to the end of the node */ + if (mas->last == mas->max) + new_end = offset; + /* don't copy this offset */ + wr_mas->offset_end++; + } else if (mas->last < wr_mas->r_max) { + /* new range ends in this range */ + if (unlikely(wr_mas->r_max == ULONG_MAX)) + mas_bulk_rebalance(mas, wr_mas->node_end, wr_mas->type); + + new_end++; + } else { + if (wr_mas->end_piv == mas->last) + wr_mas->offset_end++; + + new_end -= wr_mas->offset_end - offset - 1; + } + + /* new range starts within a range */ + if (wr_mas->r_min < mas->index) + new_end++; + + /* Not enough room */ + if (new_end >= node_slots) + return false; + + /* Not enough data. */ + if (!mte_is_root(mas->node) && (new_end <= mt_min_slots[wr_mas->type]) && + !(mas->mas_flags & MA_STATE_BULK)) + return false; + + /* set up node. */ + if (in_rcu) { + mas_node_count(mas, 1); + if (mas_is_err(mas)) + return false; + + newnode = mas_pop_node(mas); + } else { + memset(&reuse, 0, sizeof(struct maple_node)); + newnode = &reuse; + } + + newnode->parent = mas_mn(mas)->parent; + dst_pivots = ma_pivots(newnode, wr_mas->type); + dst_slots = ma_slots(newnode, wr_mas->type); + /* Copy from start to insert point */ + memcpy(dst_pivots, wr_mas->pivots, sizeof(unsigned long) * (offset + 1)); + memcpy(dst_slots, wr_mas->slots, sizeof(void *) * (offset + 1)); + dst_offset = offset; + + /* Handle insert of new range starting after old range */ + if (wr_mas->r_min < mas->index) { + mas->offset++; + rcu_assign_pointer(dst_slots[dst_offset], wr_mas->content); + dst_pivots[dst_offset++] = mas->index - 1; + } + + /* Store the new entry and range end. */ + if (dst_offset < max_piv) + dst_pivots[dst_offset] = mas->last; + mas->offset = dst_offset; + rcu_assign_pointer(dst_slots[dst_offset++], wr_mas->entry); + + /* this range wrote to the end of the node. */ + if (wr_mas->offset_end > wr_mas->node_end) + goto done; + + /* Copy to the end of node if necessary. */ + copy_size = wr_mas->node_end - wr_mas->offset_end + 1; + memcpy(dst_slots + dst_offset, wr_mas->slots + wr_mas->offset_end, + sizeof(void *) * copy_size); + if (dst_offset < max_piv) { + if (copy_size > max_piv - dst_offset) + copy_size = max_piv - dst_offset; + memcpy(dst_pivots + dst_offset, wr_mas->pivots + wr_mas->offset_end, + sizeof(unsigned long) * copy_size); + } + +done: + if ((wr_mas->node_end == node_slots - 1) && (new_end < node_slots - 1)) + dst_pivots[new_end] = mas->max; + + mas_leaf_set_meta(mas, newnode, dst_pivots, maple_leaf_64, new_end); + if (in_rcu) { + mas->node = mt_mk_node(newnode, wr_mas->type); + mas_replace(mas, false); + } else { + memcpy(wr_mas->node, newnode, sizeof(struct maple_node)); + } + trace_ma_write(__func__, mas, 0, wr_mas->entry); + mas_update_gap(mas); + return true; +} + +/* + * mas_wr_slot_store: Attempt to store a value in a slot. + * @wr_mas: the maple write state + * + * Return: True if stored, false otherwise + */ +static inline bool mas_wr_slot_store(struct ma_wr_state *wr_mas) +{ + struct ma_state *mas = wr_mas->mas; + unsigned long lmax; /* Logical max. */ + unsigned char offset = mas->offset; + + if ((wr_mas->r_max > mas->last) && ((wr_mas->r_min != mas->index) || + (offset != wr_mas->node_end))) + return false; + + if (offset == wr_mas->node_end - 1) + lmax = mas->max; + else + lmax = wr_mas->pivots[offset + 1]; + + /* going to overwrite too many slots. */ + if (lmax < mas->last) + return false; + + if (wr_mas->r_min == mas->index) { + /* overwriting two or more ranges with one. */ + if (lmax == mas->last) + return false; + + /* Overwriting all of offset and a portion of offset + 1. */ + rcu_assign_pointer(wr_mas->slots[offset], wr_mas->entry); + wr_mas->pivots[offset] = mas->last; + goto done; + } + + /* Doesn't end on the next range end. */ + if (lmax != mas->last) + return false; + + /* Overwriting a portion of offset and all of offset + 1 */ + if ((offset + 1 < mt_pivots[wr_mas->type]) && + (wr_mas->entry || wr_mas->pivots[offset + 1])) + wr_mas->pivots[offset + 1] = mas->last; + + rcu_assign_pointer(wr_mas->slots[offset + 1], wr_mas->entry); + wr_mas->pivots[offset] = mas->index - 1; + mas->offset++; /* Keep mas accurate. */ + +done: + trace_ma_write(__func__, mas, 0, wr_mas->entry); + mas_update_gap(mas); + return true; +} + +static inline void mas_wr_end_piv(struct ma_wr_state *wr_mas) +{ + while ((wr_mas->mas->last > wr_mas->end_piv) && + (wr_mas->offset_end < wr_mas->node_end)) + wr_mas->end_piv = wr_mas->pivots[++wr_mas->offset_end]; + + if (wr_mas->mas->last > wr_mas->end_piv) + wr_mas->end_piv = wr_mas->mas->max; +} + +static inline void mas_wr_extend_null(struct ma_wr_state *wr_mas) +{ + struct ma_state *mas = wr_mas->mas; + + if (mas->last < wr_mas->end_piv && !wr_mas->slots[wr_mas->offset_end]) + mas->last = wr_mas->end_piv; + + /* Check next slot(s) if we are overwriting the end */ + if ((mas->last == wr_mas->end_piv) && + (wr_mas->node_end != wr_mas->offset_end) && + !wr_mas->slots[wr_mas->offset_end + 1]) { + wr_mas->offset_end++; + if (wr_mas->offset_end == wr_mas->node_end) + mas->last = mas->max; + else + mas->last = wr_mas->pivots[wr_mas->offset_end]; + wr_mas->end_piv = mas->last; + } + + if (!wr_mas->content) { + /* If this one is null, the next and prev are not */ + mas->index = wr_mas->r_min; + } else { + /* Check prev slot if we are overwriting the start */ + if (mas->index == wr_mas->r_min && mas->offset && + !wr_mas->slots[mas->offset - 1]) { + mas->offset--; + wr_mas->r_min = mas->index = + mas_safe_min(mas, wr_mas->pivots, mas->offset); + wr_mas->r_max = wr_mas->pivots[mas->offset]; + } + } +} + +static inline bool mas_wr_append(struct ma_wr_state *wr_mas) +{ + unsigned char end = wr_mas->node_end; + unsigned char new_end = end + 1; + struct ma_state *mas = wr_mas->mas; + unsigned char node_pivots = mt_pivots[wr_mas->type]; + + if ((mas->index != wr_mas->r_min) && (mas->last == wr_mas->r_max)) { + if (new_end < node_pivots) + wr_mas->pivots[new_end] = wr_mas->pivots[end]; + + if (new_end < node_pivots) + ma_set_meta(wr_mas->node, maple_leaf_64, 0, new_end); + + rcu_assign_pointer(wr_mas->slots[new_end], wr_mas->entry); + mas->offset = new_end; + wr_mas->pivots[end] = mas->index - 1; + + return true; + } + + if ((mas->index == wr_mas->r_min) && (mas->last < wr_mas->r_max)) { + if (new_end < node_pivots) + wr_mas->pivots[new_end] = wr_mas->pivots[end]; + + rcu_assign_pointer(wr_mas->slots[new_end], wr_mas->content); + if (new_end < node_pivots) + ma_set_meta(wr_mas->node, maple_leaf_64, 0, new_end); + + wr_mas->pivots[end] = mas->last; + rcu_assign_pointer(wr_mas->slots[end], wr_mas->entry); + return true; + } + + return false; +} + +static inline void mas_wr_modify(struct ma_wr_state *wr_mas) +{ + unsigned char node_slots; + unsigned char node_size; + struct ma_state *mas = wr_mas->mas; + struct maple_big_node b_node; + + /* Direct replacement */ + if (wr_mas->r_min == mas->index && wr_mas->r_max == mas->last) { + rcu_assign_pointer(wr_mas->slots[mas->offset], wr_mas->entry); + if (!!wr_mas->entry ^ !!wr_mas->content) + mas_update_gap(mas); + return; + } + + /* Attempt to append */ + node_slots = mt_slots[wr_mas->type]; + node_size = wr_mas->node_end - wr_mas->offset_end + mas->offset + 2; + if (mas->max == ULONG_MAX) + node_size++; + + /* slot and node store will not fit, go to the slow path */ + if (unlikely(node_size >= node_slots)) + goto slow_path; + + if (wr_mas->entry && (wr_mas->node_end < node_slots - 1) && + (mas->offset == wr_mas->node_end) && mas_wr_append(wr_mas)) { + if (!wr_mas->content || !wr_mas->entry) + mas_update_gap(mas); + return; + } + + if ((wr_mas->offset_end - mas->offset <= 1) && mas_wr_slot_store(wr_mas)) + return; + else if (mas_wr_node_store(wr_mas)) + return; + + if (mas_is_err(mas)) + return; + +slow_path: + memset(&b_node, 0, sizeof(struct maple_big_node)); + mas_store_b_node(wr_mas, &b_node, wr_mas->offset_end); + trace_ma_write(__func__, mas, 0, wr_mas->entry); + mas_commit_b_node(wr_mas, &b_node, wr_mas->node_end); +} + +/* + * mas_wr_store_entry() - Internal call to store a value + * @mas: The maple state + * @entry: The entry to store. + * + * Return: The contents that was stored at the index. + */ +static inline void *mas_wr_store_entry(struct ma_wr_state *wr_mas) +{ + struct ma_state *mas = wr_mas->mas; + + wr_mas->content = mas_start(mas); + if (mas_is_none(mas) || mas_is_ptr(mas)) { + mas_store_root(mas, wr_mas->entry); + return wr_mas->content; + } + + if (unlikely(!mas_wr_walk(wr_mas))) { + mas_wr_spanning_store(wr_mas); + return wr_mas->content; + } + + /* At this point, we are at the leaf node that needs to be altered. */ + wr_mas->end_piv = wr_mas->r_max; + mas_wr_end_piv(wr_mas); + + if (!wr_mas->entry) + mas_wr_extend_null(wr_mas); + + /* New root for a single pointer */ + if (unlikely(!mas->index && mas->last == ULONG_MAX)) { + mas_new_root(mas, wr_mas->entry); + return wr_mas->content; + } + + mas_wr_modify(wr_mas); + return wr_mas->content; +} + +/** + * mas_insert() - Internal call to insert a value + * @mas: The maple state + * @entry: The entry to store + * + * Return: %NULL or the contents that already exists at the requested index + * otherwise. The maple state needs to be checked for error conditions. + */ +static inline void *mas_insert(struct ma_state *mas, void *entry) +{ + MA_WR_STATE(wr_mas, mas, entry); + + /* + * Inserting a new range inserts either 0, 1, or 2 pivots within the + * tree. If the insert fits exactly into an existing gap with a value + * of NULL, then the slot only needs to be written with the new value. + * If the range being inserted is adjacent to another range, then only a + * single pivot needs to be inserted (as well as writing the entry). If + * the new range is within a gap but does not touch any other ranges, + * then two pivots need to be inserted: the start - 1, and the end. As + * usual, the entry must be written. Most operations require a new node + * to be allocated and replace an existing node to ensure RCU safety, + * when in RCU mode. The exception to requiring a newly allocated node + * is when inserting at the end of a node (appending). When done + * carefully, appending can reuse the node in place. + */ + wr_mas.content = mas_start(mas); + if (wr_mas.content) + goto exists; + + if (mas_is_none(mas) || mas_is_ptr(mas)) { + mas_store_root(mas, entry); + return NULL; + } + + /* spanning writes always overwrite something */ + if (!mas_wr_walk(&wr_mas)) + goto exists; + + /* At this point, we are at the leaf node that needs to be altered. */ + wr_mas.offset_end = mas->offset; + wr_mas.end_piv = wr_mas.r_max; + + if (wr_mas.content || (mas->last > wr_mas.r_max)) + goto exists; + + if (!entry) + return NULL; + + mas_wr_modify(&wr_mas); + return wr_mas.content; + +exists: + mas_set_err(mas, -EEXIST); + return wr_mas.content; + +} + +/* + * mas_prev_node() - Find the prev non-null entry at the same level in the + * tree. The prev value will be mas->node[mas->offset] or MAS_NONE. + * @mas: The maple state + * @min: The lower limit to search + * + * The prev node value will be mas->node[mas->offset] or MAS_NONE. + * Return: 1 if the node is dead, 0 otherwise. + */ +static inline int mas_prev_node(struct ma_state *mas, unsigned long min) +{ + enum maple_type mt; + int offset, level; + void __rcu **slots; + struct maple_node *node; + struct maple_enode *enode; + unsigned long *pivots; + + if (mas_is_none(mas)) + return 0; + + level = 0; + do { + node = mas_mn(mas); + if (ma_is_root(node)) + goto no_entry; + + /* Walk up. */ + if (unlikely(mas_ascend(mas))) + return 1; + offset = mas->offset; + level++; + } while (!offset); + + offset--; + mt = mte_node_type(mas->node); + node = mas_mn(mas); + slots = ma_slots(node, mt); + pivots = ma_pivots(node, mt); + mas->max = pivots[offset]; + if (offset) + mas->min = pivots[offset - 1] + 1; + if (unlikely(ma_dead_node(node))) + return 1; + + if (mas->max < min) + goto no_entry_min; + + while (level > 1) { + level--; + enode = mas_slot(mas, slots, offset); + if (unlikely(ma_dead_node(node))) + return 1; + + mas->node = enode; + mt = mte_node_type(mas->node); + node = mas_mn(mas); + slots = ma_slots(node, mt); + pivots = ma_pivots(node, mt); + offset = ma_data_end(node, mt, pivots, mas->max); + if (offset) + mas->min = pivots[offset - 1] + 1; + + if (offset < mt_pivots[mt]) + mas->max = pivots[offset]; + + if (mas->max < min) + goto no_entry; + } + + mas->node = mas_slot(mas, slots, offset); + if (unlikely(ma_dead_node(node))) + return 1; + + mas->offset = mas_data_end(mas); + if (unlikely(mte_dead_node(mas->node))) + return 1; + + return 0; + +no_entry_min: + mas->offset = offset; + if (offset) + mas->min = pivots[offset - 1] + 1; +no_entry: + if (unlikely(ma_dead_node(node))) + return 1; + + mas->node = MAS_NONE; + return 0; +} + +/* + * mas_next_node() - Get the next node at the same level in the tree. + * @mas: The maple state + * @max: The maximum pivot value to check. + * + * The next value will be mas->node[mas->offset] or MAS_NONE. + * Return: 1 on dead node, 0 otherwise. + */ +static inline int mas_next_node(struct ma_state *mas, struct maple_node *node, + unsigned long max) +{ + unsigned long min, pivot; + unsigned long *pivots; + struct maple_enode *enode; + int level = 0; + unsigned char offset; + enum maple_type mt; + void __rcu **slots; + + if (mas->max >= max) + goto no_entry; + + level = 0; + do { + if (ma_is_root(node)) + goto no_entry; + + min = mas->max + 1; + if (min > max) + goto no_entry; + + if (unlikely(mas_ascend(mas))) + return 1; + + offset = mas->offset; + level++; + node = mas_mn(mas); + mt = mte_node_type(mas->node); + pivots = ma_pivots(node, mt); + } while (unlikely(offset == ma_data_end(node, mt, pivots, mas->max))); + + slots = ma_slots(node, mt); + pivot = mas_safe_pivot(mas, pivots, ++offset, mt); + while (unlikely(level > 1)) { + /* Descend, if necessary */ + enode = mas_slot(mas, slots, offset); + if (unlikely(ma_dead_node(node))) + return 1; + + mas->node = enode; + level--; + node = mas_mn(mas); + mt = mte_node_type(mas->node); + slots = ma_slots(node, mt); + pivots = ma_pivots(node, mt); + offset = 0; + pivot = pivots[0]; + } + + enode = mas_slot(mas, slots, offset); + if (unlikely(ma_dead_node(node))) + return 1; + + mas->node = enode; + mas->min = min; + mas->max = pivot; + return 0; + +no_entry: + if (unlikely(ma_dead_node(node))) + return 1; + + mas->node = MAS_NONE; + return 0; +} + +/* + * mas_next_nentry() - Get the next node entry + * @mas: The maple state + * @max: The maximum value to check + * @*range_start: Pointer to store the start of the range. + * + * Sets @mas->offset to the offset of the next node entry, @mas->last to the + * pivot of the entry. + * + * Return: The next entry, %NULL otherwise + */ +static inline void *mas_next_nentry(struct ma_state *mas, + struct maple_node *node, unsigned long max, enum maple_type type) +{ + unsigned char count; + unsigned long pivot; + unsigned long *pivots; + void __rcu **slots; + void *entry; + + if (mas->last == mas->max) { + mas->index = mas->max; + return NULL; + } + + pivots = ma_pivots(node, type); + slots = ma_slots(node, type); + mas->index = mas_safe_min(mas, pivots, mas->offset); + if (ma_dead_node(node)) + return NULL; + + if (mas->index > max) + return NULL; + + count = ma_data_end(node, type, pivots, mas->max); + if (mas->offset > count) + return NULL; + + while (mas->offset < count) { + pivot = pivots[mas->offset]; + entry = mas_slot(mas, slots, mas->offset); + if (ma_dead_node(node)) + return NULL; + + if (entry) + goto found; + + if (pivot >= max) + return NULL; + + mas->index = pivot + 1; + mas->offset++; + } + + if (mas->index > mas->max) { + mas->index = mas->last; + return NULL; + } + + pivot = mas_safe_pivot(mas, pivots, mas->offset, type); + entry = mas_slot(mas, slots, mas->offset); + if (ma_dead_node(node)) + return NULL; + + if (!pivot) + return NULL; + + if (!entry) + return NULL; + +found: + mas->last = pivot; + return entry; +} + +static inline void mas_rewalk(struct ma_state *mas, unsigned long index) +{ + +retry: + mas_set(mas, index); + mas_state_walk(mas); + if (mas_is_start(mas)) + goto retry; + + return; + +} + +/* + * mas_next_entry() - Internal function to get the next entry. + * @mas: The maple state + * @limit: The maximum range start. + * + * Set the @mas->node to the next entry and the range_start to + * the beginning value for the entry. Does not check beyond @limit. + * Sets @mas->index and @mas->last to the limit if it is hit. + * Restarts on dead nodes. + * + * Return: the next entry or %NULL. + */ +static inline void *mas_next_entry(struct ma_state *mas, unsigned long limit) +{ + void *entry = NULL; + struct maple_enode *prev_node; + struct maple_node *node; + unsigned char offset; + unsigned long last; + enum maple_type mt; + + last = mas->last; +retry: + offset = mas->offset; + prev_node = mas->node; + node = mas_mn(mas); + mt = mte_node_type(mas->node); + mas->offset++; + if (unlikely(mas->offset >= mt_slots[mt])) { + mas->offset = mt_slots[mt] - 1; + goto next_node; + } + + while (!mas_is_none(mas)) { + entry = mas_next_nentry(mas, node, limit, mt); + if (unlikely(ma_dead_node(node))) { + mas_rewalk(mas, last); + goto retry; + } + + if (likely(entry)) + return entry; + + if (unlikely((mas->index > limit))) + break; + +next_node: + prev_node = mas->node; + offset = mas->offset; + if (unlikely(mas_next_node(mas, node, limit))) { + mas_rewalk(mas, last); + goto retry; + } + mas->offset = 0; + node = mas_mn(mas); + mt = mte_node_type(mas->node); + } + + mas->index = mas->last = limit; + mas->offset = offset; + mas->node = prev_node; + return NULL; +} + +/* + * mas_prev_nentry() - Get the previous node entry. + * @mas: The maple state. + * @limit: The lower limit to check for a value. + * + * Return: the entry, %NULL otherwise. + */ +static inline void *mas_prev_nentry(struct ma_state *mas, unsigned long limit, + unsigned long index) +{ + unsigned long pivot, min; + unsigned char offset; + struct maple_node *mn; + enum maple_type mt; + unsigned long *pivots; + void __rcu **slots; + void *entry; + +retry: + if (!mas->offset) + return NULL; + + mn = mas_mn(mas); + mt = mte_node_type(mas->node); + offset = mas->offset - 1; + if (offset >= mt_slots[mt]) + offset = mt_slots[mt] - 1; + + slots = ma_slots(mn, mt); + pivots = ma_pivots(mn, mt); + if (offset == mt_pivots[mt]) + pivot = mas->max; + else + pivot = pivots[offset]; + + if (unlikely(ma_dead_node(mn))) { + mas_rewalk(mas, index); + goto retry; + } + + while (offset && ((!mas_slot(mas, slots, offset) && pivot >= limit) || + !pivot)) + pivot = pivots[--offset]; + + min = mas_safe_min(mas, pivots, offset); + entry = mas_slot(mas, slots, offset); + if (unlikely(ma_dead_node(mn))) { + mas_rewalk(mas, index); + goto retry; + } + + if (likely(entry)) { + mas->offset = offset; + mas->last = pivot; + mas->index = min; + } + return entry; +} + +static inline void *mas_prev_entry(struct ma_state *mas, unsigned long min) +{ + void *entry; + +retry: + while (likely(!mas_is_none(mas))) { + entry = mas_prev_nentry(mas, min, mas->index); + if (unlikely(mas->last < min)) + goto not_found; + + if (likely(entry)) + return entry; + + if (unlikely(mas_prev_node(mas, min))) { + mas_rewalk(mas, mas->index); + goto retry; + } + + mas->offset++; + } + + mas->offset--; +not_found: + mas->index = mas->last = min; + return NULL; +} + +/* + * mas_rev_awalk() - Internal function. Reverse allocation walk. Find the + * highest gap address of a given size in a given node and descend. + * @mas: The maple state + * @size: The needed size. + * + * Return: True if found in a leaf, false otherwise. + * + */ +static bool mas_rev_awalk(struct ma_state *mas, unsigned long size) +{ + enum maple_type type = mte_node_type(mas->node); + struct maple_node *node = mas_mn(mas); + unsigned long *pivots, *gaps; + void __rcu **slots; + unsigned long gap = 0; + unsigned long max, min, index; + unsigned char offset; + + if (unlikely(mas_is_err(mas))) + return true; + + if (ma_is_dense(type)) { + /* dense nodes. */ + mas->offset = (unsigned char)(mas->index - mas->min); + return true; + } + + pivots = ma_pivots(node, type); + slots = ma_slots(node, type); + gaps = ma_gaps(node, type); + offset = mas->offset; + min = mas_safe_min(mas, pivots, offset); + /* Skip out of bounds. */ + while (mas->last < min) + min = mas_safe_min(mas, pivots, --offset); + + max = mas_safe_pivot(mas, pivots, offset, type); + index = mas->index; + while (index <= max) { + gap = 0; + if (gaps) + gap = gaps[offset]; + else if (!mas_slot(mas, slots, offset)) + gap = max - min + 1; + + if (gap) { + if ((size <= gap) && (size <= mas->last - min + 1)) + break; + + if (!gaps) { + /* Skip the next slot, it cannot be a gap. */ + if (offset < 2) + goto ascend; + + offset -= 2; + max = pivots[offset]; + min = mas_safe_min(mas, pivots, offset); + continue; + } + } + + if (!offset) + goto ascend; + + offset--; + max = min - 1; + min = mas_safe_min(mas, pivots, offset); + } + + if (unlikely(index > max)) { + mas_set_err(mas, -EBUSY); + return false; + } + + if (unlikely(ma_is_leaf(type))) { + mas->offset = offset; + mas->min = min; + mas->max = min + gap - 1; + return true; + } + + /* descend, only happens under lock. */ + mas->node = mas_slot(mas, slots, offset); + mas->min = min; + mas->max = max; + mas->offset = mas_data_end(mas); + return false; + +ascend: + if (mte_is_root(mas->node)) + mas_set_err(mas, -EBUSY); + + return false; +} + +static inline bool mas_anode_descend(struct ma_state *mas, unsigned long size) +{ + enum maple_type type = mte_node_type(mas->node); + unsigned long pivot, min, gap = 0; + unsigned char count, offset; + unsigned long *gaps = NULL, *pivots = ma_pivots(mas_mn(mas), type); + void __rcu **slots = ma_slots(mas_mn(mas), type); + bool found = false; + + if (ma_is_dense(type)) { + mas->offset = (unsigned char)(mas->index - mas->min); + return true; + } + + gaps = ma_gaps(mte_to_node(mas->node), type); + offset = mas->offset; + count = mt_slots[type]; + min = mas_safe_min(mas, pivots, offset); + for (; offset < count; offset++) { + pivot = mas_safe_pivot(mas, pivots, offset, type); + if (offset && !pivot) + break; + + /* Not within lower bounds */ + if (mas->index > pivot) + goto next_slot; + + if (gaps) + gap = gaps[offset]; + else if (!mas_slot(mas, slots, offset)) + gap = min(pivot, mas->last) - max(mas->index, min) + 1; + else + goto next_slot; + + if (gap >= size) { + if (ma_is_leaf(type)) { + found = true; + goto done; + } + if (mas->index <= pivot) { + mas->node = mas_slot(mas, slots, offset); + mas->min = min; + mas->max = pivot; + offset = 0; + type = mte_node_type(mas->node); + count = mt_slots[type]; + break; + } + } +next_slot: + min = pivot + 1; + if (mas->last < min) { + mas_set_err(mas, -EBUSY); + return true; + } + } + + if (mte_is_root(mas->node)) + found = true; +done: + mas->offset = offset; + return found; +} + +/** + * mas_walk() - Search for @mas->index in the tree. + * @mas: The maple state. + * + * mas->index and mas->last will be set to the range if there is a value. If + * mas->node is MAS_NONE, reset to MAS_START. + * + * Return: the entry at the location or %NULL. + */ +void *mas_walk(struct ma_state *mas) +{ + void *entry; + +retry: + entry = mas_state_walk(mas); + if (mas_is_start(mas)) + goto retry; + + if (mas_is_ptr(mas)) { + if (!mas->index) { + mas->last = 0; + } else { + mas->index = 1; + mas->last = ULONG_MAX; + } + return entry; + } + + if (mas_is_none(mas)) { + mas->index = 0; + mas->last = ULONG_MAX; + } + + return entry; +} + +static inline bool mas_rewind_node(struct ma_state *mas) +{ + unsigned char slot; + + do { + if (mte_is_root(mas->node)) { + slot = mas->offset; + if (!slot) + return false; + } else { + mas_ascend(mas); + slot = mas->offset; + } + } while (!slot); + + mas->offset = --slot; + return true; +} + +/* + * mas_skip_node() - Internal function. Skip over a node. + * @mas: The maple state. + * + * Return: true if there is another node, false otherwise. + */ +static inline bool mas_skip_node(struct ma_state *mas) +{ + unsigned char slot, slot_count; + unsigned long *pivots; + enum maple_type mt; + + mt = mte_node_type(mas->node); + slot_count = mt_slots[mt] - 1; + do { + if (mte_is_root(mas->node)) { + slot = mas->offset; + if (slot > slot_count) { + mas_set_err(mas, -EBUSY); + return false; + } + } else { + mas_ascend(mas); + slot = mas->offset; + mt = mte_node_type(mas->node); + slot_count = mt_slots[mt] - 1; + } + } while (slot > slot_count); + + mas->offset = ++slot; + pivots = ma_pivots(mas_mn(mas), mt); + if (slot > 0) + mas->min = pivots[slot - 1] + 1; + + if (slot <= slot_count) + mas->max = pivots[slot]; + + return true; +} + +/* + * mas_awalk() - Allocation walk. Search from low address to high, for a gap of + * @size + * @mas: The maple state + * @size: The size of the gap required + * + * Search between @mas->index and @mas->last for a gap of @size. + */ +static inline void mas_awalk(struct ma_state *mas, unsigned long size) +{ + struct maple_enode *last = NULL; + + /* + * There are 4 options: + * go to child (descend) + * go back to parent (ascend) + * no gap found. (return, slot == MAPLE_NODE_SLOTS) + * found the gap. (return, slot != MAPLE_NODE_SLOTS) + */ + while (!mas_is_err(mas) && !mas_anode_descend(mas, size)) { + if (last == mas->node) + mas_skip_node(mas); + else + last = mas->node; + } +} + +/* + * mas_fill_gap() - Fill a located gap with @entry. + * @mas: The maple state + * @entry: The value to store + * @slot: The offset into the node to store the @entry + * @size: The size of the entry + * @index: The start location + */ +static inline void mas_fill_gap(struct ma_state *mas, void *entry, + unsigned char slot, unsigned long size, unsigned long *index) +{ + MA_WR_STATE(wr_mas, mas, entry); + unsigned char pslot = mte_parent_slot(mas->node); + struct maple_enode *mn = mas->node; + unsigned long *pivots; + enum maple_type ptype; + /* + * mas->index is the start address for the search + * which may no longer be needed. + * mas->last is the end address for the search + */ + + *index = mas->index; + mas->last = mas->index + size - 1; + + /* + * It is possible that using mas->max and mas->min to correctly + * calculate the index and last will cause an issue in the gap + * calculation, so fix the ma_state here + */ + mas_ascend(mas); + ptype = mte_node_type(mas->node); + pivots = ma_pivots(mas_mn(mas), ptype); + mas->max = mas_safe_pivot(mas, pivots, pslot, ptype); + mas->min = mas_safe_min(mas, pivots, pslot); + mas->node = mn; + mas->offset = slot; + mas_wr_store_entry(&wr_mas); +} + +/* + * mas_sparse_area() - Internal function. Return upper or lower limit when + * searching for a gap in an empty tree. + * @mas: The maple state + * @min: the minimum range + * @max: The maximum range + * @size: The size of the gap + * @fwd: Searching forward or back + */ +static inline void mas_sparse_area(struct ma_state *mas, unsigned long min, + unsigned long max, unsigned long size, bool fwd) +{ + unsigned long start = 0; + + if (!unlikely(mas_is_none(mas))) + start++; + /* mas_is_ptr */ + + if (start < min) + start = min; + + if (fwd) { + mas->index = start; + mas->last = start + size - 1; + return; + } + + mas->index = max; +} + +/* + * mas_empty_area() - Get the lowest address within the range that is + * sufficient for the size requested. + * @mas: The maple state + * @min: The lowest value of the range + * @max: The highest value of the range + * @size: The size needed + */ +int mas_empty_area(struct ma_state *mas, unsigned long min, + unsigned long max, unsigned long size) +{ + unsigned char offset; + unsigned long *pivots; + enum maple_type mt; + + if (mas_is_start(mas)) + mas_start(mas); + else if (mas->offset >= 2) + mas->offset -= 2; + else if (!mas_skip_node(mas)) + return -EBUSY; + + /* Empty set */ + if (mas_is_none(mas) || mas_is_ptr(mas)) { + mas_sparse_area(mas, min, max, size, true); + return 0; + } + + /* The start of the window can only be within these values */ + mas->index = min; + mas->last = max; + mas_awalk(mas, size); + + if (unlikely(mas_is_err(mas))) + return xa_err(mas->node); + + offset = mas->offset; + if (unlikely(offset == MAPLE_NODE_SLOTS)) + return -EBUSY; + + mt = mte_node_type(mas->node); + pivots = ma_pivots(mas_mn(mas), mt); + if (offset) + mas->min = pivots[offset - 1] + 1; + + if (offset < mt_pivots[mt]) + mas->max = pivots[offset]; + + if (mas->index < mas->min) + mas->index = mas->min; + + mas->last = mas->index + size - 1; + return 0; +} + +/* + * mas_empty_area_rev() - Get the highest address within the range that is + * sufficient for the size requested. + * @mas: The maple state + * @min: The lowest value of the range + * @max: The highest value of the range + * @size: The size needed + */ +int mas_empty_area_rev(struct ma_state *mas, unsigned long min, + unsigned long max, unsigned long size) +{ + struct maple_enode *last = mas->node; + + if (mas_is_start(mas)) { + mas_start(mas); + mas->offset = mas_data_end(mas); + } else if (mas->offset >= 2) { + mas->offset -= 2; + } else if (!mas_rewind_node(mas)) { + return -EBUSY; + } + + /* Empty set. */ + if (mas_is_none(mas) || mas_is_ptr(mas)) { + mas_sparse_area(mas, min, max, size, false); + return 0; + } + + /* The start of the window can only be within these values. */ + mas->index = min; + mas->last = max; + + while (!mas_rev_awalk(mas, size)) { + if (last == mas->node) { + if (!mas_rewind_node(mas)) + return -EBUSY; + } else { + last = mas->node; + } + } + + if (unlikely(mas->offset == MAPLE_NODE_SLOTS)) + return -EBUSY; + + /* + * mas_rev_awalk() has set mas->min and mas->max to the gap values. If + * the maximum is outside the window we are searching, then use the last + * location in the search. + * mas->max and mas->min is the range of the gap. + * mas->index and mas->last are currently set to the search range. + */ + + /* Trim the upper limit to the max. */ + if (mas->max <= mas->last) + mas->last = mas->max; + + mas->index = mas->last - size + 1; + return 0; +} + +static inline int mas_alloc(struct ma_state *mas, void *entry, + unsigned long size, unsigned long *index) +{ + unsigned long min; + + mas_start(mas); + if (mas_is_none(mas) || mas_is_ptr(mas)) { + mas_root_expand(mas, entry); + if (mas_is_err(mas)) + return xa_err(mas->node); + + if (!mas->index) + return mte_pivot(mas->node, 0); + return mte_pivot(mas->node, 1); + } + + /* Must be walking a tree. */ + mas_awalk(mas, size); + if (mas_is_err(mas)) + return xa_err(mas->node); + + if (mas->offset == MAPLE_NODE_SLOTS) + goto no_gap; + + /* + * At this point, mas->node points to the right node and we have an + * offset that has a sufficient gap. + */ + min = mas->min; + if (mas->offset) + min = mte_pivot(mas->node, mas->offset - 1) + 1; + + if (mas->index < min) + mas->index = min; + + mas_fill_gap(mas, entry, mas->offset, size, index); + return 0; + +no_gap: + return -EBUSY; +} + +static inline int mas_rev_alloc(struct ma_state *mas, unsigned long min, + unsigned long max, void *entry, + unsigned long size, unsigned long *index) +{ + int ret = 0; + + ret = mas_empty_area_rev(mas, min, max, size); + if (ret) + return ret; + + if (mas_is_err(mas)) + return xa_err(mas->node); + + if (mas->offset == MAPLE_NODE_SLOTS) + goto no_gap; + + mas_fill_gap(mas, entry, mas->offset, size, index); + return 0; + +no_gap: + return -EBUSY; +} + +/* + * mas_dead_leaves() - Mark all leaves of a node as dead. + * @mas: The maple state + * @slots: Pointer to the slot array + * + * Must hold the write lock. + * + * Return: The number of leaves marked as dead. + */ +static inline +unsigned char mas_dead_leaves(struct ma_state *mas, void __rcu **slots) +{ + struct maple_node *node; + enum maple_type type; + void *entry; + int offset; + + for (offset = 0; offset < mt_slot_count(mas->node); offset++) { + entry = mas_slot_locked(mas, slots, offset); + type = mte_node_type(entry); + node = mte_to_node(entry); + /* Use both node and type to catch LE & BE metadata */ + if (!node || !type) + break; + + mte_set_node_dead(entry); + smp_wmb(); /* Needed for RCU */ + node->type = type; + rcu_assign_pointer(slots[offset], node); + } + + return offset; +} + +static void __rcu **mas_dead_walk(struct ma_state *mas, unsigned char offset) +{ + struct maple_node *node, *next; + void __rcu **slots = NULL; + + next = mas_mn(mas); + do { + mas->node = ma_enode_ptr(next); + node = mas_mn(mas); + slots = ma_slots(node, node->type); + next = mas_slot_locked(mas, slots, offset); + offset = 0; + } while (!ma_is_leaf(next->type)); + + return slots; +} + +static void mt_free_walk(struct rcu_head *head) +{ + void __rcu **slots; + struct maple_node *node, *start; + struct maple_tree mt; + unsigned char offset; + enum maple_type type; + MA_STATE(mas, &mt, 0, 0); + + node = container_of(head, struct maple_node, rcu); + + if (ma_is_leaf(node->type)) + goto free_leaf; + + mt_init_flags(&mt, node->ma_flags); + mas_lock(&mas); + start = node; + mas.node = mt_mk_node(node, node->type); + slots = mas_dead_walk(&mas, 0); + node = mas_mn(&mas); + do { + mt_free_bulk(node->slot_len, slots); + offset = node->parent_slot + 1; + mas.node = node->piv_parent; + if (mas_mn(&mas) == node) + goto start_slots_free; + + type = mte_node_type(mas.node); + slots = ma_slots(mte_to_node(mas.node), type); + if ((offset < mt_slots[type]) && (slots[offset])) + slots = mas_dead_walk(&mas, offset); + + node = mas_mn(&mas); + } while ((node != start) || (node->slot_len < offset)); + + slots = ma_slots(node, node->type); + mt_free_bulk(node->slot_len, slots); + +start_slots_free: +free_leaf: + mt_free_rcu(&node->rcu); +} + +static inline void __rcu **mas_destroy_descend(struct ma_state *mas, + struct maple_enode *prev, unsigned char offset) +{ + struct maple_node *node; + struct maple_enode *next = mas->node; + void __rcu **slots = NULL; + + do { + mas->node = next; + node = mas_mn(mas); + slots = ma_slots(node, mte_node_type(mas->node)); + next = mas_slot_locked(mas, slots, 0); + if ((mte_dead_node(next))) + next = mas_slot_locked(mas, slots, 1); + + mte_set_node_dead(mas->node); + node->type = mte_node_type(mas->node); + node->piv_parent = prev; + node->parent_slot = offset; + offset = 0; + prev = mas->node; + } while (!mte_is_leaf(next)); + + return slots; +} + +static void mt_destroy_walk(struct maple_enode *enode, unsigned char ma_flags, + bool free) +{ + void __rcu **slots; + struct maple_node *node = mte_to_node(enode); + struct maple_enode *start; + struct maple_tree mt; + + MA_STATE(mas, &mt, 0, 0); + + if (mte_is_leaf(enode)) + goto free_leaf; + + mt_init_flags(&mt, ma_flags); + mas_lock(&mas); + + mas.node = start = enode; + slots = mas_destroy_descend(&mas, start, 0); + node = mas_mn(&mas); + do { + enum maple_type type; + unsigned char offset; + + node->slot_len = mas_dead_leaves(&mas, slots); + if (free) + mt_free_bulk(node->slot_len, slots); + offset = node->parent_slot + 1; + mas.node = node->piv_parent; + if (mas_mn(&mas) == node) + goto start_slots_free; + + type = mte_node_type(mas.node); + slots = ma_slots(mte_to_node(mas.node), type); + if ((offset < mt_slots[type]) && mte_node_type(slots[offset]) && + mte_to_node(slots[offset])) { + struct maple_enode *parent = mas.node; + + mas.node = mas_slot_locked(&mas, slots, offset); + slots = mas_destroy_descend(&mas, parent, offset); + } + node = mas_mn(&mas); + } while (start != mas.node); + + node = mas_mn(&mas); + node->slot_len = mas_dead_leaves(&mas, slots); + if (free) + mt_free_bulk(node->slot_len, slots); + +start_slots_free: + mas_unlock(&mas); + +free_leaf: + if (free) + mt_free_rcu(&node->rcu); +} + +/* + * mte_destroy_walk() - Free a tree or sub-tree. + * @enode - the encoded maple node (maple_enode) to start + * @mn - the tree to free - needed for node types. + * + * Must hold the write lock. + */ +static inline void mte_destroy_walk(struct maple_enode *enode, + struct maple_tree *mt) +{ + struct maple_node *node = mte_to_node(enode); + + if (mt_in_rcu(mt)) { + mt_destroy_walk(enode, mt->ma_flags, false); + call_rcu(&node->rcu, mt_free_walk); + } else { + mt_destroy_walk(enode, mt->ma_flags, true); + } +} + +static void mas_wr_store_setup(struct ma_wr_state *wr_mas) +{ + if (!mas_is_start(wr_mas->mas)) { + if (mas_is_none(wr_mas->mas)) { + mas_reset(wr_mas->mas); + } else { + wr_mas->r_max = wr_mas->mas->max; + wr_mas->type = mte_node_type(wr_mas->mas->node); + if (mas_is_span_wr(wr_mas)) + mas_reset(wr_mas->mas); + } + } + +} + +/* Interface */ + +/** + * mas_store() - Store an @entry. + * @mas: The maple state. + * @entry: The entry to store. + * + * The @mas->index and @mas->last is used to set the range for the @entry. + * Note: The @mas should have pre-allocated entries to ensure there is memory to + * store the entry. Please see mas_expected_entries()/mas_destroy() for more details. + * + * Return: the first entry between mas->index and mas->last or %NULL. + */ +void *mas_store(struct ma_state *mas, void *entry) +{ + MA_WR_STATE(wr_mas, mas, entry); + + trace_ma_write(__func__, mas, 0, entry); +#ifdef CONFIG_DEBUG_MAPLE_TREE + if (mas->index > mas->last) + pr_err("Error %lu > %lu %p\n", mas->index, mas->last, entry); + MT_BUG_ON(mas->tree, mas->index > mas->last); + if (mas->index > mas->last) { + mas_set_err(mas, -EINVAL); + return NULL; + } + +#endif + + /* + * Storing is the same operation as insert with the added caveat that it + * can overwrite entries. Although this seems simple enough, one may + * want to examine what happens if a single store operation was to + * overwrite multiple entries within a self-balancing B-Tree. + */ + mas_wr_store_setup(&wr_mas); + mas_wr_store_entry(&wr_mas); + return wr_mas.content; +} + +/** + * mas_store_gfp() - Store a value into the tree. + * @mas: The maple state + * @entry: The entry to store + * @gfp: The GFP_FLAGS to use for allocations if necessary. + * + * Return: 0 on success, -EINVAL on invalid request, -ENOMEM if memory could not + * be allocated. + */ +int mas_store_gfp(struct ma_state *mas, void *entry, gfp_t gfp) +{ + MA_WR_STATE(wr_mas, mas, entry); + + mas_wr_store_setup(&wr_mas); + trace_ma_write(__func__, mas, 0, entry); +retry: + mas_wr_store_entry(&wr_mas); + if (unlikely(mas_nomem(mas, gfp))) + goto retry; + + if (unlikely(mas_is_err(mas))) + return xa_err(mas->node); + + return 0; +} + +/** + * mas_store_prealloc() - Store a value into the tree using memory + * preallocated in the maple state. + * @mas: The maple state + * @entry: The entry to store. + */ +void mas_store_prealloc(struct ma_state *mas, void *entry) +{ + MA_WR_STATE(wr_mas, mas, entry); + + mas_wr_store_setup(&wr_mas); + trace_ma_write(__func__, mas, 0, entry); + mas_wr_store_entry(&wr_mas); + BUG_ON(mas_is_err(mas)); + mas_destroy(mas); +} + +/** + * mas_preallocate() - Preallocate enough nodes for a store operation + * @mas: The maple state + * @entry: The entry that will be stored + * @gfp: The GFP_FLAGS to use for allocations. + * + * Return: 0 on success, -ENOMEM if memory could not be allocated. + */ +int mas_preallocate(struct ma_state *mas, void *entry, gfp_t gfp) +{ + int ret; + + mas_set_alloc_req(mas, 1 + mas_mt_height(mas) * 3); + mas_alloc_nodes(mas, gfp); + if (likely(!mas_is_err(mas))) + return 0; + + mas_set_alloc_req(mas, 0); + mas_destroy(mas); + ret = xa_err(mas->node); + mas->node = MAS_START; + return ret; +} + +/* + * mas_expected_entries() - Set the expected number of entries that will be inserted. + * @mas: The maple state + * @nr_entries: The number of expected entries. + * + * This will attempt to pre-allocate enough nodes to store the expected number + * of entries. The allocations will occur using the bulk allocator interface + * for speed. Please call mas_destroy() on the @mas after inserting the entries + * to ensure any unused nodes are freed. + * + * Return: 0 on success, -ENOMEM if memory could not be allocated. + */ +int mas_expected_entries(struct ma_state *mas, unsigned long nr_entries) +{ + int nonleaf_cap = MAPLE_ARANGE64_SLOTS - 2; + struct maple_enode *enode = mas->node; + int nr_nodes; + int ret; + + /* + * Sometimes it is necessary to duplicate a tree to a new tree, such as + * forking a process and duplicating the VMAs from one tree to a new + * tree. When such a situation arises, it is known that the new tree is + * not going to be used until the entire tree is populated. For + * performance reasons, it is best to use a bulk load with RCU disabled. + * This allows for optimistic splitting that favours the left and reuse + * of nodes during the operation. + */ + + /* Optimize splitting for bulk insert in-order */ + mas->mas_flags |= MA_STATE_BULK; + + /* + * Avoid overflow, assume a gap between each entry and a trailing null. + * If this is wrong, it just means allocation can happen during + * insertion of entries. + */ + nr_nodes = max(nr_entries, nr_entries * 2 + 1); + + if (!mt_is_alloc(mas->tree)) + nonleaf_cap = MAPLE_RANGE64_SLOTS - 2; + + /* Leaves */ + nr_nodes = DIV_ROUND_UP(nr_nodes, MAPLE_RANGE64_SLOTS - 1); + /* Internal nodes */ + nr_nodes += DIV_ROUND_UP(nr_nodes, nonleaf_cap); + mas_node_count(mas, nr_nodes); + + if (!mas_is_err(mas)) + return 0; + + ret = xa_err(mas->node); + mas->node = enode; + return ret; + +} + +/* + * mas_destroy() - destroy a maple state. + * @mas: The maple state + * + * Upon completion, check the left-most node and rebalance against the node to + * the right if necessary. Frees any allocated nodes associated with this maple + * state. + */ +void mas_destroy(struct ma_state *mas) +{ + struct maple_alloc *node; + + /* + * When using mas_for_each() to insert an expected number of elements, + * it is possible that the number inserted is less than the expected + * number. To fix an invalid final node, a check is performed here to + * rebalance the previous node with the final node. + */ + if (mas->mas_flags & MA_STATE_REBALANCE) { + unsigned char end; + + if (mas_is_start(mas)) + mas_start(mas); + + mtree_range_walk(mas); + end = mas_data_end(mas) + 1; + if (end < mt_min_slot_count(mas->node) - 1) + mas_destroy_rebalance(mas, end); + + mas->mas_flags &= ~MA_STATE_REBALANCE; + } + mas->mas_flags &= ~MA_STATE_BULK; + + while (mas->alloc && !((unsigned long)mas->alloc & 0x1)) { + node = mas->alloc; + mas->alloc = node->slot[0]; + if (node->node_count > 0) + mt_free_bulk(node->node_count, + (void __rcu **)&node->slot[1]); + kmem_cache_free(maple_node_cache, node); + } + mas->alloc = NULL; +} + +/** + * mas_next() - Get the next entry. + * @mas: The maple state + * @max: The maximum index to check. + * + * Returns the next entry after @mas->index. + * Must hold rcu_read_lock or the write lock. + * Can return the zero entry. + * + * Return: The next entry or %NULL + */ +void *mas_next(struct ma_state *mas, unsigned long max) +{ + if (mas_is_none(mas) || mas_is_paused(mas)) + mas->node = MAS_START; + + if (mas_is_start(mas)) + mas_walk(mas); /* Retries on dead nodes handled by mas_walk */ + + if (mas_is_ptr(mas)) { + if (!mas->index) { + mas->index = 1; + mas->last = ULONG_MAX; + } + return NULL; + } + + if (mas->last == ULONG_MAX) + return NULL; + + /* Retries on dead nodes handled by mas_next_entry */ + return mas_next_entry(mas, max); +} +EXPORT_SYMBOL_GPL(mas_next); + +/** + * mt_next() - get the next value in the maple tree + * @mt: The maple tree + * @index: The start index + * @max: The maximum index to check + * + * Return: The entry at @index or higher, or %NULL if nothing is found. + */ +void *mt_next(struct maple_tree *mt, unsigned long index, unsigned long max) +{ + void *entry = NULL; + MA_STATE(mas, mt, index, index); + + rcu_read_lock(); + entry = mas_next(&mas, max); + rcu_read_unlock(); + return entry; +} +EXPORT_SYMBOL_GPL(mt_next); + +/** + * mas_prev() - Get the previous entry + * @mas: The maple state + * @min: The minimum value to check. + * + * Must hold rcu_read_lock or the write lock. + * Will reset mas to MAS_START if the node is MAS_NONE. Will stop on not + * searchable nodes. + * + * Return: the previous value or %NULL. + */ +void *mas_prev(struct ma_state *mas, unsigned long min) +{ + if (!mas->index) { + /* Nothing comes before 0 */ + mas->last = 0; + return NULL; + } + + if (unlikely(mas_is_ptr(mas))) + return NULL; + + if (mas_is_none(mas) || mas_is_paused(mas)) + mas->node = MAS_START; + + if (mas_is_start(mas)) { + mas_walk(mas); + if (!mas->index) + return NULL; + } + + if (mas_is_ptr(mas)) { + if (!mas->index) { + mas->last = 0; + return NULL; + } + + mas->index = mas->last = 0; + return mas_root_locked(mas); + } + return mas_prev_entry(mas, min); +} +EXPORT_SYMBOL_GPL(mas_prev); + +/** + * mt_prev() - get the previous value in the maple tree + * @mt: The maple tree + * @index: The start index + * @min: The minimum index to check + * + * Return: The entry at @index or lower, or %NULL if nothing is found. + */ +void *mt_prev(struct maple_tree *mt, unsigned long index, unsigned long min) +{ + void *entry = NULL; + MA_STATE(mas, mt, index, index); + + rcu_read_lock(); + entry = mas_prev(&mas, min); + rcu_read_unlock(); + return entry; +} +EXPORT_SYMBOL_GPL(mt_prev); + +/** + * mas_pause() - Pause a mas_find/mas_for_each to drop the lock. + * @mas: The maple state to pause + * + * Some users need to pause a walk and drop the lock they're holding in + * order to yield to a higher priority thread or carry out an operation + * on an entry. Those users should call this function before they drop + * the lock. It resets the @mas to be suitable for the next iteration + * of the loop after the user has reacquired the lock. If most entries + * found during a walk require you to call mas_pause(), the mt_for_each() + * iterator may be more appropriate. + * + */ +void mas_pause(struct ma_state *mas) +{ + mas->node = MAS_PAUSE; +} +EXPORT_SYMBOL_GPL(mas_pause); + +/** + * mas_find() - On the first call, find the entry at or after mas->index up to + * %max. Otherwise, find the entry after mas->index. + * @mas: The maple state + * @max: The maximum value to check. + * + * Must hold rcu_read_lock or the write lock. + * If an entry exists, last and index are updated accordingly. + * May set @mas->node to MAS_NONE. + * + * Return: The entry or %NULL. + */ +void *mas_find(struct ma_state *mas, unsigned long max) +{ + if (unlikely(mas_is_paused(mas))) { + if (unlikely(mas->last == ULONG_MAX)) { + mas->node = MAS_NONE; + return NULL; + } + mas->node = MAS_START; + mas->index = ++mas->last; + } + + if (unlikely(mas_is_start(mas))) { + /* First run or continue */ + void *entry; + + if (mas->index > max) + return NULL; + + entry = mas_walk(mas); + if (entry) + return entry; + } + + if (unlikely(!mas_searchable(mas))) + return NULL; + + /* Retries on dead nodes handled by mas_next_entry */ + return mas_next_entry(mas, max); +} + +/** + * mas_find_rev: On the first call, find the first non-null entry at or below + * mas->index down to %min. Otherwise find the first non-null entry below + * mas->index down to %min. + * @mas: The maple state + * @min: The minimum value to check. + * + * Must hold rcu_read_lock or the write lock. + * If an entry exists, last and index are updated accordingly. + * May set @mas->node to MAS_NONE. + * + * Return: The entry or %NULL. + */ +void *mas_find_rev(struct ma_state *mas, unsigned long min) +{ + if (unlikely(mas_is_paused(mas))) { + if (unlikely(mas->last == ULONG_MAX)) { + mas->node = MAS_NONE; + return NULL; + } + mas->node = MAS_START; + mas->last = --mas->index; + } + + if (unlikely(mas_is_start(mas))) { + /* First run or continue */ + void *entry; + + if (mas->index < min) + return NULL; + + entry = mas_walk(mas); + if (entry) + return entry; + } + + if (unlikely(!mas_searchable(mas))) + return NULL; + + if (mas->index < min) + return NULL; + + /* Retries on dead nodes handled by mas_next_entry */ + return mas_prev_entry(mas, min); +} +EXPORT_SYMBOL_GPL(mas_find); + +/** + * mas_erase() - Find the range in which index resides and erase the entire + * range. + * @mas: The maple state + * + * Must hold the write lock. + * Searches for @mas->index, sets @mas->index and @mas->last to the range and + * erases that range. + * + * Return: the entry that was erased or %NULL, @mas->index and @mas->last are updated. + */ +void *mas_erase(struct ma_state *mas) +{ + void *entry; + MA_WR_STATE(wr_mas, mas, NULL); + + if (mas_is_none(mas) || mas_is_paused(mas)) + mas->node = MAS_START; + + /* Retry unnecessary when holding the write lock. */ + entry = mas_state_walk(mas); + if (!entry) + return NULL; + +write_retry: + /* Must reset to ensure spanning writes of last slot are detected */ + mas_reset(mas); + mas_wr_store_setup(&wr_mas); + mas_wr_store_entry(&wr_mas); + if (mas_nomem(mas, GFP_KERNEL)) + goto write_retry; + + return entry; +} +EXPORT_SYMBOL_GPL(mas_erase); + +/** + * mas_nomem() - Check if there was an error allocating and do the allocation + * if necessary If there are allocations, then free them. + * @mas: The maple state + * @gfp: The GFP_FLAGS to use for allocations + * Return: true on allocation, false otherwise. + */ +bool mas_nomem(struct ma_state *mas, gfp_t gfp) + __must_hold(mas->tree->lock) +{ + if (likely(mas->node != MA_ERROR(-ENOMEM))) { + mas_destroy(mas); + return false; + } + + if (gfpflags_allow_blocking(gfp) && !mt_external_lock(mas->tree)) { + mtree_unlock(mas->tree); + mas_alloc_nodes(mas, gfp); + mtree_lock(mas->tree); + } else { + mas_alloc_nodes(mas, gfp); + } + + if (!mas_allocated(mas)) + return false; + + mas->node = MAS_START; + return true; +} + +void __init maple_tree_init(void) +{ + maple_node_cache = kmem_cache_create("maple_node", + sizeof(struct maple_node), sizeof(struct maple_node), + SLAB_PANIC, NULL); +} + +/** + * mtree_load() - Load a value stored in a maple tree + * @mt: The maple tree + * @index: The index to load + * + * Return: the entry or %NULL + */ +void *mtree_load(struct maple_tree *mt, unsigned long index) +{ + MA_STATE(mas, mt, index, index); + void *entry; + + trace_ma_read(__func__, &mas); + rcu_read_lock(); +retry: + entry = mas_start(&mas); + if (unlikely(mas_is_none(&mas))) + goto unlock; + + if (unlikely(mas_is_ptr(&mas))) { + if (index) + entry = NULL; + + goto unlock; + } + + entry = mtree_lookup_walk(&mas); + if (!entry && unlikely(mas_is_start(&mas))) + goto retry; +unlock: + rcu_read_unlock(); + if (xa_is_zero(entry)) + return NULL; + + return entry; +} +EXPORT_SYMBOL(mtree_load); + +/** + * mtree_store_range() - Store an entry at a given range. + * @mt: The maple tree + * @index: The start of the range + * @last: The end of the range + * @entry: The entry to store + * @gfp: The GFP_FLAGS to use for allocations + * + * Return: 0 on success, -EINVAL on invalid request, -ENOMEM if memory could not + * be allocated. + */ +int mtree_store_range(struct maple_tree *mt, unsigned long index, + unsigned long last, void *entry, gfp_t gfp) +{ + MA_STATE(mas, mt, index, last); + MA_WR_STATE(wr_mas, &mas, entry); + + trace_ma_write(__func__, &mas, 0, entry); + if (WARN_ON_ONCE(xa_is_advanced(entry))) + return -EINVAL; + + if (index > last) + return -EINVAL; + + mtree_lock(mt); +retry: + mas_wr_store_entry(&wr_mas); + if (mas_nomem(&mas, gfp)) + goto retry; + + mtree_unlock(mt); + if (mas_is_err(&mas)) + return xa_err(mas.node); + + return 0; +} +EXPORT_SYMBOL(mtree_store_range); + +/** + * mtree_store() - Store an entry at a given index. + * @mt: The maple tree + * @index: The index to store the value + * @entry: The entry to store + * @gfp: The GFP_FLAGS to use for allocations + * + * Return: 0 on success, -EINVAL on invalid request, -ENOMEM if memory could not + * be allocated. + */ +int mtree_store(struct maple_tree *mt, unsigned long index, void *entry, + gfp_t gfp) +{ + return mtree_store_range(mt, index, index, entry, gfp); +} +EXPORT_SYMBOL(mtree_store); + +/** + * mtree_insert_range() - Insert an entry at a give range if there is no value. + * @mt: The maple tree + * @first: The start of the range + * @last: The end of the range + * @entry: The entry to store + * @gfp: The GFP_FLAGS to use for allocations. + * + * Return: 0 on success, -EEXISTS if the range is occupied, -EINVAL on invalid + * request, -ENOMEM if memory could not be allocated. + */ +int mtree_insert_range(struct maple_tree *mt, unsigned long first, + unsigned long last, void *entry, gfp_t gfp) +{ + MA_STATE(ms, mt, first, last); + + if (WARN_ON_ONCE(xa_is_advanced(entry))) + return -EINVAL; + + if (first > last) + return -EINVAL; + + mtree_lock(mt); +retry: + mas_insert(&ms, entry); + if (mas_nomem(&ms, gfp)) + goto retry; + + mtree_unlock(mt); + if (mas_is_err(&ms)) + return xa_err(ms.node); + + return 0; +} +EXPORT_SYMBOL(mtree_insert_range); + +/** + * mtree_insert() - Insert an entry at a give index if there is no value. + * @mt: The maple tree + * @index : The index to store the value + * @entry: The entry to store + * @gfp: The FGP_FLAGS to use for allocations. + * + * Return: 0 on success, -EEXISTS if the range is occupied, -EINVAL on invalid + * request, -ENOMEM if memory could not be allocated. + */ +int mtree_insert(struct maple_tree *mt, unsigned long index, void *entry, + gfp_t gfp) +{ + return mtree_insert_range(mt, index, index, entry, gfp); +} +EXPORT_SYMBOL(mtree_insert); + +int mtree_alloc_range(struct maple_tree *mt, unsigned long *startp, + void *entry, unsigned long size, unsigned long min, + unsigned long max, gfp_t gfp) +{ + int ret = 0; + + MA_STATE(mas, mt, min, max - size); + if (!mt_is_alloc(mt)) + return -EINVAL; + + if (WARN_ON_ONCE(mt_is_reserved(entry))) + return -EINVAL; + + if (min > max) + return -EINVAL; + + if (max < size) + return -EINVAL; + + if (!size) + return -EINVAL; + + mtree_lock(mt); +retry: + mas.offset = 0; + mas.index = min; + mas.last = max - size; + ret = mas_alloc(&mas, entry, size, startp); + if (mas_nomem(&mas, gfp)) + goto retry; + + mtree_unlock(mt); + return ret; +} +EXPORT_SYMBOL(mtree_alloc_range); + +int mtree_alloc_rrange(struct maple_tree *mt, unsigned long *startp, + void *entry, unsigned long size, unsigned long min, + unsigned long max, gfp_t gfp) +{ + int ret = 0; + + MA_STATE(mas, mt, min, max - size); + if (!mt_is_alloc(mt)) + return -EINVAL; + + if (WARN_ON_ONCE(mt_is_reserved(entry))) + return -EINVAL; + + if (min >= max) + return -EINVAL; + + if (max < size - 1) + return -EINVAL; + + if (!size) + return -EINVAL; + + mtree_lock(mt); +retry: + ret = mas_rev_alloc(&mas, min, max, entry, size, startp); + if (mas_nomem(&mas, gfp)) + goto retry; + + mtree_unlock(mt); + return ret; +} +EXPORT_SYMBOL(mtree_alloc_rrange); + +/** + * mtree_erase() - Find an index and erase the entire range. + * @mt: The maple tree + * @index: The index to erase + * + * Erasing is the same as a walk to an entry then a store of a NULL to that + * ENTIRE range. In fact, it is implemented as such using the advanced API. + * + * Return: The entry stored at the @index or %NULL + */ +void *mtree_erase(struct maple_tree *mt, unsigned long index) +{ + void *entry = NULL; + + MA_STATE(mas, mt, index, index); + trace_ma_op(__func__, &mas); + + mtree_lock(mt); + entry = mas_erase(&mas); + mtree_unlock(mt); + + return entry; +} +EXPORT_SYMBOL(mtree_erase); + +/** + * __mt_destroy() - Walk and free all nodes of a locked maple tree. + * @mt: The maple tree + * + * Note: Does not handle locking. + */ +void __mt_destroy(struct maple_tree *mt) +{ + void *root = mt_root_locked(mt); + + rcu_assign_pointer(mt->ma_root, NULL); + if (xa_is_node(root)) + mte_destroy_walk(root, mt); + + mt->ma_flags = 0; +} +EXPORT_SYMBOL_GPL(__mt_destroy); + +/** + * mtree_destroy() - Destroy a maple tree + * @mt: The maple tree + * + * Frees all resources used by the tree. Handles locking. + */ +void mtree_destroy(struct maple_tree *mt) +{ + mtree_lock(mt); + __mt_destroy(mt); + mtree_unlock(mt); +} +EXPORT_SYMBOL(mtree_destroy); + +/** + * mt_find() - Search from the start up until an entry is found. + * @mt: The maple tree + * @index: Pointer which contains the start location of the search + * @max: The maximum value to check + * + * Handles locking. @index will be incremented to one beyond the range. + * + * Return: The entry at or after the @index or %NULL + */ +void *mt_find(struct maple_tree *mt, unsigned long *index, unsigned long max) +{ + MA_STATE(mas, mt, *index, *index); + void *entry; +#ifdef CONFIG_DEBUG_MAPLE_TREE + unsigned long copy = *index; +#endif + + trace_ma_read(__func__, &mas); + + if ((*index) > max) + return NULL; + + rcu_read_lock(); +retry: + entry = mas_state_walk(&mas); + if (mas_is_start(&mas)) + goto retry; + + if (unlikely(xa_is_zero(entry))) + entry = NULL; + + if (entry) + goto unlock; + + while (mas_searchable(&mas) && (mas.index < max)) { + entry = mas_next_entry(&mas, max); + if (likely(entry && !xa_is_zero(entry))) + break; + } + + if (unlikely(xa_is_zero(entry))) + entry = NULL; +unlock: + rcu_read_unlock(); + if (likely(entry)) { + *index = mas.last + 1; +#ifdef CONFIG_DEBUG_MAPLE_TREE + if ((*index) && (*index) <= copy) + pr_err("index not increased! %lx <= %lx\n", + *index, copy); + MT_BUG_ON(mt, (*index) && ((*index) <= copy)); +#endif + } + + return entry; +} +EXPORT_SYMBOL(mt_find); + +/** + * mt_find_after() - Search from the start up until an entry is found. + * @mt: The maple tree + * @index: Pointer which contains the start location of the search + * @max: The maximum value to check + * + * Handles locking, detects wrapping on index == 0 + * + * Return: The entry at or after the @index or %NULL + */ +void *mt_find_after(struct maple_tree *mt, unsigned long *index, + unsigned long max) +{ + if (!(*index)) + return NULL; + + return mt_find(mt, index, max); +} +EXPORT_SYMBOL(mt_find_after); + +#ifdef CONFIG_DEBUG_MAPLE_TREE +atomic_t maple_tree_tests_run; +EXPORT_SYMBOL_GPL(maple_tree_tests_run); +atomic_t maple_tree_tests_passed; +EXPORT_SYMBOL_GPL(maple_tree_tests_passed); + +#ifndef __KERNEL__ +extern void kmem_cache_set_non_kernel(struct kmem_cache *, unsigned int); +void mt_set_non_kernel(unsigned int val) +{ + kmem_cache_set_non_kernel(maple_node_cache, val); +} + +extern unsigned long kmem_cache_get_alloc(struct kmem_cache *); +unsigned long mt_get_alloc_size(void) +{ + return kmem_cache_get_alloc(maple_node_cache); +} + +extern void kmem_cache_zero_nr_tallocated(struct kmem_cache *); +void mt_zero_nr_tallocated(void) +{ + kmem_cache_zero_nr_tallocated(maple_node_cache); +} + +extern unsigned int kmem_cache_nr_tallocated(struct kmem_cache *); +unsigned int mt_nr_tallocated(void) +{ + return kmem_cache_nr_tallocated(maple_node_cache); +} + +extern unsigned int kmem_cache_nr_allocated(struct kmem_cache *); +unsigned int mt_nr_allocated(void) +{ + return kmem_cache_nr_allocated(maple_node_cache); +} + +/* + * mas_dead_node() - Check if the maple state is pointing to a dead node. + * @mas: The maple state + * @index: The index to restore in @mas. + * + * Used in test code. + * Return: 1 if @mas has been reset to MAS_START, 0 otherwise. + */ +static inline int mas_dead_node(struct ma_state *mas, unsigned long index) +{ + if (unlikely(!mas_searchable(mas) || mas_is_start(mas))) + return 0; + + if (likely(!mte_dead_node(mas->node))) + return 0; + + mas_rewalk(mas, index); + return 1; +} +#endif /* not defined __KERNEL__ */ + +/* + * mas_get_slot() - Get the entry in the maple state node stored at @offset. + * @mas: The maple state + * @offset: The offset into the slot array to fetch. + * + * Return: The entry stored at @offset. + */ +static inline struct maple_enode *mas_get_slot(struct ma_state *mas, + unsigned char offset) +{ + return mas_slot(mas, ma_slots(mas_mn(mas), mte_node_type(mas->node)), + offset); +} + + +/* + * mas_first_entry() - Go the first leaf and find the first entry. + * @mas: the maple state. + * @limit: the maximum index to check. + * @*r_start: Pointer to set to the range start. + * + * Sets mas->offset to the offset of the entry, r_start to the range minimum. + * + * Return: The first entry or MAS_NONE. + */ +static inline void *mas_first_entry(struct ma_state *mas, struct maple_node *mn, + unsigned long limit, enum maple_type mt) + +{ + unsigned long max; + unsigned long *pivots; + void __rcu **slots; + void *entry = NULL; + + mas->index = mas->min; + if (mas->index > limit) + goto none; + + max = mas->max; + mas->offset = 0; + while (likely(!ma_is_leaf(mt))) { + slots = ma_slots(mn, mt); + pivots = ma_pivots(mn, mt); + max = pivots[0]; + entry = mas_slot(mas, slots, 0); + if (unlikely(ma_dead_node(mn))) + return NULL; + mas->node = entry; + mn = mas_mn(mas); + mt = mte_node_type(mas->node); + } + + mas->max = max; + slots = ma_slots(mn, mt); + entry = mas_slot(mas, slots, 0); + if (unlikely(ma_dead_node(mn))) + return NULL; + + /* Slot 0 or 1 must be set */ + if (mas->index > limit) + goto none; + + if (likely(entry)) + return entry; + + pivots = ma_pivots(mn, mt); + mas->index = pivots[0] + 1; + mas->offset = 1; + entry = mas_slot(mas, slots, 1); + if (unlikely(ma_dead_node(mn))) + return NULL; + + if (mas->index > limit) + goto none; + + if (likely(entry)) + return entry; + +none: + if (likely(!ma_dead_node(mn))) + mas->node = MAS_NONE; + return NULL; +} + +/* Depth first search, post-order */ +static void mas_dfs_postorder(struct ma_state *mas, unsigned long max) +{ + + struct maple_enode *p = MAS_NONE, *mn = mas->node; + unsigned long p_min, p_max; + + mas_next_node(mas, mas_mn(mas), max); + if (!mas_is_none(mas)) + return; + + if (mte_is_root(mn)) + return; + + mas->node = mn; + mas_ascend(mas); + while (mas->node != MAS_NONE) { + p = mas->node; + p_min = mas->min; + p_max = mas->max; + mas_prev_node(mas, 0); + } + + if (p == MAS_NONE) + return; + + mas->node = p; + mas->max = p_max; + mas->min = p_min; +} + +/* Tree validations */ +static void mt_dump_node(const struct maple_tree *mt, void *entry, + unsigned long min, unsigned long max, unsigned int depth); +static void mt_dump_range(unsigned long min, unsigned long max, + unsigned int depth) +{ + static const char spaces[] = " "; + + if (min == max) + pr_info("%.*s%lu: ", depth * 2, spaces, min); + else + pr_info("%.*s%lu-%lu: ", depth * 2, spaces, min, max); +} + +static void mt_dump_entry(void *entry, unsigned long min, unsigned long max, + unsigned int depth) +{ + mt_dump_range(min, max, depth); + + if (xa_is_value(entry)) + pr_cont("value %ld (0x%lx) [%p]\n", xa_to_value(entry), + xa_to_value(entry), entry); + else if (xa_is_zero(entry)) + pr_cont("zero (%ld)\n", xa_to_internal(entry)); + else if (mt_is_reserved(entry)) + pr_cont("UNKNOWN ENTRY (%p)\n", entry); + else + pr_cont("%p\n", entry); +} + +static void mt_dump_range64(const struct maple_tree *mt, void *entry, + unsigned long min, unsigned long max, unsigned int depth) +{ + struct maple_range_64 *node = &mte_to_node(entry)->mr64; + bool leaf = mte_is_leaf(entry); + unsigned long first = min; + int i; + + pr_cont(" contents: "); + for (i = 0; i < MAPLE_RANGE64_SLOTS - 1; i++) + pr_cont("%p %lu ", node->slot[i], node->pivot[i]); + pr_cont("%p\n", node->slot[i]); + for (i = 0; i < MAPLE_RANGE64_SLOTS; i++) { + unsigned long last = max; + + if (i < (MAPLE_RANGE64_SLOTS - 1)) + last = node->pivot[i]; + else if (!node->slot[i] && max != mt_max[mte_node_type(entry)]) + break; + if (last == 0 && i > 0) + break; + if (leaf) + mt_dump_entry(mt_slot(mt, node->slot, i), + first, last, depth + 1); + else if (node->slot[i]) + mt_dump_node(mt, mt_slot(mt, node->slot, i), + first, last, depth + 1); + + if (last == max) + break; + if (last > max) { + pr_err("node %p last (%lu) > max (%lu) at pivot %d!\n", + node, last, max, i); + break; + } + first = last + 1; + } +} + +static void mt_dump_arange64(const struct maple_tree *mt, void *entry, + unsigned long min, unsigned long max, unsigned int depth) +{ + struct maple_arange_64 *node = &mte_to_node(entry)->ma64; + bool leaf = mte_is_leaf(entry); + unsigned long first = min; + int i; + + pr_cont(" contents: "); + for (i = 0; i < MAPLE_ARANGE64_SLOTS; i++) + pr_cont("%lu ", node->gap[i]); + pr_cont("| %02X %02X| ", node->meta.end, node->meta.gap); + for (i = 0; i < MAPLE_ARANGE64_SLOTS - 1; i++) + pr_cont("%p %lu ", node->slot[i], node->pivot[i]); + pr_cont("%p\n", node->slot[i]); + for (i = 0; i < MAPLE_ARANGE64_SLOTS; i++) { + unsigned long last = max; + + if (i < (MAPLE_ARANGE64_SLOTS - 1)) + last = node->pivot[i]; + else if (!node->slot[i]) + break; + if (last == 0 && i > 0) + break; + if (leaf) + mt_dump_entry(mt_slot(mt, node->slot, i), + first, last, depth + 1); + else if (node->slot[i]) + mt_dump_node(mt, mt_slot(mt, node->slot, i), + first, last, depth + 1); + + if (last == max) + break; + if (last > max) { + pr_err("node %p last (%lu) > max (%lu) at pivot %d!\n", + node, last, max, i); + break; + } + first = last + 1; + } +} + +static void mt_dump_node(const struct maple_tree *mt, void *entry, + unsigned long min, unsigned long max, unsigned int depth) +{ + struct maple_node *node = mte_to_node(entry); + unsigned int type = mte_node_type(entry); + unsigned int i; + + mt_dump_range(min, max, depth); + + pr_cont("node %p depth %d type %d parent %p", node, depth, type, + node ? node->parent : NULL); + switch (type) { + case maple_dense: + pr_cont("\n"); + for (i = 0; i < MAPLE_NODE_SLOTS; i++) { + if (min + i > max) + pr_cont("OUT OF RANGE: "); + mt_dump_entry(mt_slot(mt, node->slot, i), + min + i, min + i, depth); + } + break; + case maple_leaf_64: + case maple_range_64: + mt_dump_range64(mt, entry, min, max, depth); + break; + case maple_arange_64: + mt_dump_arange64(mt, entry, min, max, depth); + break; + + default: + pr_cont(" UNKNOWN TYPE\n"); + } +} + +void mt_dump(const struct maple_tree *mt) +{ + void *entry = rcu_dereference_check(mt->ma_root, mt_locked(mt)); + + pr_info("maple_tree(%p) flags %X, height %u root %p\n", + mt, mt->ma_flags, mt_height(mt), entry); + if (!xa_is_node(entry)) + mt_dump_entry(entry, 0, 0, 0); + else if (entry) + mt_dump_node(mt, entry, 0, mt_max[mte_node_type(entry)], 0); +} + +/* + * Calculate the maximum gap in a node and check if that's what is reported in + * the parent (unless root). + */ +static void mas_validate_gaps(struct ma_state *mas) +{ + struct maple_enode *mte = mas->node; + struct maple_node *p_mn; + unsigned long gap = 0, max_gap = 0; + unsigned long p_end, p_start = mas->min; + unsigned char p_slot; + unsigned long *gaps = NULL; + unsigned long *pivots = ma_pivots(mte_to_node(mte), mte_node_type(mte)); + int i; + + if (ma_is_dense(mte_node_type(mte))) { + for (i = 0; i < mt_slot_count(mte); i++) { + if (mas_get_slot(mas, i)) { + if (gap > max_gap) + max_gap = gap; + gap = 0; + continue; + } + gap++; + } + goto counted; + } + + gaps = ma_gaps(mte_to_node(mte), mte_node_type(mte)); + for (i = 0; i < mt_slot_count(mte); i++) { + p_end = mas_logical_pivot(mas, pivots, i, mte_node_type(mte)); + + if (!gaps) { + if (mas_get_slot(mas, i)) { + gap = 0; + goto not_empty; + } + + gap += p_end - p_start + 1; + } else { + void *entry = mas_get_slot(mas, i); + + gap = gaps[i]; + if (!entry) { + if (gap != p_end - p_start + 1) { + pr_err("%p[%u] -> %p %lu != %lu - %lu + 1\n", + mas_mn(mas), i, + mas_get_slot(mas, i), gap, + p_end, p_start); + mt_dump(mas->tree); + + MT_BUG_ON(mas->tree, + gap != p_end - p_start + 1); + } + } else { + if (gap > p_end - p_start + 1) { + pr_err("%p[%u] %lu >= %lu - %lu + 1 (%lu)\n", + mas_mn(mas), i, gap, p_end, p_start, + p_end - p_start + 1); + MT_BUG_ON(mas->tree, + gap > p_end - p_start + 1); + } + } + } + + if (gap > max_gap) + max_gap = gap; +not_empty: + p_start = p_end + 1; + if (p_end >= mas->max) + break; + } + +counted: + if (mte_is_root(mte)) + return; + + p_slot = mte_parent_slot(mas->node); + p_mn = mte_parent(mte); + MT_BUG_ON(mas->tree, max_gap > mas->max); + if (ma_gaps(p_mn, mas_parent_enum(mas, mte))[p_slot] != max_gap) { + pr_err("gap %p[%u] != %lu\n", p_mn, p_slot, max_gap); + mt_dump(mas->tree); + } + + MT_BUG_ON(mas->tree, + ma_gaps(p_mn, mas_parent_enum(mas, mte))[p_slot] != max_gap); +} + +static void mas_validate_parent_slot(struct ma_state *mas) +{ + struct maple_node *parent; + struct maple_enode *node; + enum maple_type p_type = mas_parent_enum(mas, mas->node); + unsigned char p_slot = mte_parent_slot(mas->node); + void __rcu **slots; + int i; + + if (mte_is_root(mas->node)) + return; + + parent = mte_parent(mas->node); + slots = ma_slots(parent, p_type); + MT_BUG_ON(mas->tree, mas_mn(mas) == parent); + + /* Check prev/next parent slot for duplicate node entry */ + + for (i = 0; i < mt_slots[p_type]; i++) { + node = mas_slot(mas, slots, i); + if (i == p_slot) { + if (node != mas->node) + pr_err("parent %p[%u] does not have %p\n", + parent, i, mas_mn(mas)); + MT_BUG_ON(mas->tree, node != mas->node); + } else if (node == mas->node) { + pr_err("Invalid child %p at parent %p[%u] p_slot %u\n", + mas_mn(mas), parent, i, p_slot); + MT_BUG_ON(mas->tree, node == mas->node); + } + } +} + +static void mas_validate_child_slot(struct ma_state *mas) +{ + enum maple_type type = mte_node_type(mas->node); + void __rcu **slots = ma_slots(mte_to_node(mas->node), type); + unsigned long *pivots = ma_pivots(mte_to_node(mas->node), type); + struct maple_enode *child; + unsigned char i; + + if (mte_is_leaf(mas->node)) + return; + + for (i = 0; i < mt_slots[type]; i++) { + child = mas_slot(mas, slots, i); + if (!pivots[i] || pivots[i] == mas->max) + break; + + if (!child) + break; + + if (mte_parent_slot(child) != i) { + pr_err("Slot error at %p[%u]: child %p has pslot %u\n", + mas_mn(mas), i, mte_to_node(child), + mte_parent_slot(child)); + MT_BUG_ON(mas->tree, 1); + } + + if (mte_parent(child) != mte_to_node(mas->node)) { + pr_err("child %p has parent %p not %p\n", + mte_to_node(child), mte_parent(child), + mte_to_node(mas->node)); + MT_BUG_ON(mas->tree, 1); + } + } +} + +/* + * Validate all pivots are within mas->min and mas->max. + */ +static void mas_validate_limits(struct ma_state *mas) +{ + int i; + unsigned long prev_piv = 0; + void __rcu **slots = ma_slots(mte_to_node(mas->node), + mte_node_type(mas->node)); + + /* all limits are fine here. */ + if (mte_is_root(mas->node)) + return; + + for (i = 0; i < mt_slot_count(mas->node); i++) { + enum maple_type type = mte_node_type(mas->node); + unsigned long *pivots = ma_pivots(mas_mn(mas), type); + unsigned long piv = mas_safe_pivot(mas, pivots, type, i); + + if (!piv) + break; + + if (!mte_is_leaf(mas->node)) { + void *entry = mas_slot(mas, slots, i); + + if (!entry) + pr_err("%p[%u] cannot be null\n", + mas_mn(mas), i); + + MT_BUG_ON(mas->tree, !entry); + } + + if (prev_piv > piv) { + pr_err("%p[%u] piv %lu < prev_piv %lu\n", + mas_mn(mas), i, piv, prev_piv); + MT_BUG_ON(mas->tree, piv < prev_piv); + } + + if (piv < mas->min) { + pr_err("%p[%u] %lu < %lu\n", mas_mn(mas), i, + piv, mas->min); + MT_BUG_ON(mas->tree, piv < mas->min); + } + if (piv > mas->max) { + pr_err("%p[%u] %lu > %lu\n", mas_mn(mas), i, + piv, mas->max); + MT_BUG_ON(mas->tree, piv > mas->max); + } + prev_piv = piv; + if (piv == mas->max) + break; + } +} + +static void mt_validate_nulls(struct maple_tree *mt) +{ + void *entry, *last = (void *)1; + unsigned char offset = 0; + void __rcu **slots; + MA_STATE(mas, mt, 0, 0); + + mas_start(&mas); + if (mas_is_none(&mas) || (mas.node == MAS_ROOT)) + return; + + while (!mte_is_leaf(mas.node)) + mas_descend(&mas); + + slots = ma_slots(mte_to_node(mas.node), mte_node_type(mas.node)); + do { + entry = mas_slot(&mas, slots, offset); + if (!last && !entry) { + pr_err("Sequential nulls end at %p[%u]\n", + mas_mn(&mas), offset); + } + MT_BUG_ON(mt, !last && !entry); + last = entry; + if (offset == mas_data_end(&mas)) { + mas_next_node(&mas, mas_mn(&mas), ULONG_MAX); + if (mas_is_none(&mas)) + return; + offset = 0; + slots = ma_slots(mte_to_node(mas.node), + mte_node_type(mas.node)); + } else + offset++; + + } while (!mas_is_none(&mas)); +} + +/* + * validate a maple tree by checking: + * 1. The limits (pivots are within mas->min to mas->max) + * 2. The gap is correctly set in the parents + */ +void mt_validate(struct maple_tree *mt) +{ + unsigned char end; + + MA_STATE(mas, mt, 0, 0); + rcu_read_lock(); + mas_start(&mas); + if (!mas_searchable(&mas)) + goto done; + + mas_first_entry(&mas, mas_mn(&mas), ULONG_MAX, mte_node_type(mas.node)); + while (!mas_is_none(&mas)) { + MT_BUG_ON(mas.tree, mte_dead_node(mas.node)); + if (!mte_is_root(mas.node)) { + end = mas_data_end(&mas); + if ((end < mt_min_slot_count(mas.node)) && + (mas.max != ULONG_MAX)) { + pr_err("Invalid size %u of %p\n", end, + mas_mn(&mas)); + MT_BUG_ON(mas.tree, 1); + } + + } + mas_validate_parent_slot(&mas); + mas_validate_child_slot(&mas); + mas_validate_limits(&mas); + if (mt_is_alloc(mt)) + mas_validate_gaps(&mas); + mas_dfs_postorder(&mas, ULONG_MAX); + } + mt_validate_nulls(mt); +done: + rcu_read_unlock(); + +} + +#endif /* CONFIG_DEBUG_MAPLE_TREE */ diff --git a/tools/testing/radix-tree/.gitignore b/tools/testing/radix-tree/.gitignore index d971516401e6..c901d96dd013 100644 --- a/tools/testing/radix-tree/.gitignore +++ b/tools/testing/radix-tree/.gitignore @@ -6,3 +6,5 @@ main multiorder radix-tree.c xarray +maple +ma_xa_benchmark diff --git a/tools/testing/radix-tree/generated/autoconf.h b/tools/testing/radix-tree/generated/autoconf.h index 2218b3cc184e..e7da80350236 100644 --- a/tools/testing/radix-tree/generated/autoconf.h +++ b/tools/testing/radix-tree/generated/autoconf.h @@ -1 +1,2 @@ #define CONFIG_XARRAY_MULTI 1 +#define CONFIG_64BIT 1 diff --git a/tools/testing/radix-tree/linux/maple_tree.h b/tools/testing/radix-tree/linux/maple_tree.h new file mode 100644 index 000000000000..7d8d1f445b89 --- /dev/null +++ b/tools/testing/radix-tree/linux/maple_tree.h @@ -0,0 +1,7 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +#define atomic_t int32_t +#include "../../../../include/linux/maple_tree.h" +#define atomic_inc(x) uatomic_inc(x) +#define atomic_read(x) uatomic_read(x) +#define atomic_set(x, y) do {} while (0) +#define U8_MAX UCHAR_MAX diff --git a/tools/testing/radix-tree/maple.c b/tools/testing/radix-tree/maple.c new file mode 100644 index 000000000000..35082671928a --- /dev/null +++ b/tools/testing/radix-tree/maple.c @@ -0,0 +1,59 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * maple_tree.c: Userspace shim for maple tree test-suite + * Copyright (c) 2018 Liam R. Howlett + */ + +#define CONFIG_DEBUG_MAPLE_TREE +#define CONFIG_MAPLE_SEARCH +#include "test.h" + +#define module_init(x) +#define module_exit(x) +#define MODULE_AUTHOR(x) +#define MODULE_LICENSE(x) +#define dump_stack() assert(0) + +#include "../../../lib/maple_tree.c" +#undef CONFIG_DEBUG_MAPLE_TREE +#include "../../../lib/test_maple_tree.c" + +void farmer_tests(void) +{ + struct maple_node *node; + DEFINE_MTREE(tree); + + mt_dump(&tree); + + tree.ma_root = xa_mk_value(0); + mt_dump(&tree); + + node = mt_alloc_one(GFP_KERNEL); + node->parent = (void *)((unsigned long)(&tree) | 1); + node->slot[0] = xa_mk_value(0); + node->slot[1] = xa_mk_value(1); + node->mr64.pivot[0] = 0; + node->mr64.pivot[1] = 1; + node->mr64.pivot[2] = 0; + tree.ma_root = mt_mk_node(node, maple_leaf_64); + mt_dump(&tree); + + ma_free_rcu(node); +} + +void maple_tree_tests(void) +{ + farmer_tests(); + maple_tree_seed(); + maple_tree_harvest(); +} + +int __weak main(void) +{ + maple_tree_init(); + maple_tree_tests(); + rcu_barrier(); + if (nr_allocated) + printf("nr_allocated = %d\n", nr_allocated); + return 0; +} diff --git a/tools/testing/radix-tree/trace/events/maple_tree.h b/tools/testing/radix-tree/trace/events/maple_tree.h new file mode 100644 index 000000000000..97d0e1ddcf08 --- /dev/null +++ b/tools/testing/radix-tree/trace/events/maple_tree.h @@ -0,0 +1,5 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ + +#define trace_ma_op(a, b) do {} while (0) +#define trace_ma_read(a, b) do {} while (0) +#define trace_ma_write(a, b, c, d) do {} while (0) -- cgit v1.2.3 From 2ee236fe53a8e2ab54679c74e8a1fb77e55b29bb Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:46:53 +0000 Subject: mm: start tracking VMAs with maple tree Start tracking the VMAs with the new maple tree structure in parallel with the rb_tree. Add debug and trace events for maple tree operations and duplicate the rb_tree that is created on forks into the maple tree. The maple tree is added to the mm_struct including the mm_init struct, added support in required mm/mmap functions, added tracking in kernel/fork for process forking, and used to find the unmapped_area and checked against what the rbtree finds. This also moves the mmap_lock() in exit_mmap() since the oom reaper call does walk the VMAs. Otherwise lockdep will be unhappy if oom happens. Link: https://lkml.kernel.org/r/20220504010716.661115-10-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-9-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Signed-off-by: Matthew Wilcox (Oracle) Cc: Catalin Marinas Cc: David Howells Cc: SeongJae Park Cc: Vlastimil Babka Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- arch/x86/kernel/tboot.c | 1 + drivers/firmware/efi/efi.c | 1 + include/linux/mm.h | 5 + include/linux/mm_types.h | 3 + include/trace/events/mmap.h | 73 ++++++++++ kernel/fork.c | 20 ++- mm/init-mm.c | 2 + mm/mmap.c | 320 ++++++++++++++++++++++++++++++++++++++++---- mm/nommu.c | 13 ++ 9 files changed, 412 insertions(+), 26 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/tboot.c b/arch/x86/kernel/tboot.c index 0c1154a1c403..71c54ad3868a 100644 --- a/arch/x86/kernel/tboot.c +++ b/arch/x86/kernel/tboot.c @@ -97,6 +97,7 @@ void __init tboot_probe(void) static pgd_t *tboot_pg_dir; static struct mm_struct tboot_mm = { .mm_rb = RB_ROOT, + .mm_mt = MTREE_INIT_EXT(mm_mt, MM_MT_FLAGS, tboot_mm.mmap_lock), .pgd = swapper_pg_dir, .mm_users = ATOMIC_INIT(2), .mm_count = ATOMIC_INIT(1), diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c index 860534bcfdac..1eddef189d68 100644 --- a/drivers/firmware/efi/efi.c +++ b/drivers/firmware/efi/efi.c @@ -58,6 +58,7 @@ static unsigned long __initdata rt_prop = EFI_INVALID_TABLE_ADDR; struct mm_struct efi_mm = { .mm_rb = RB_ROOT, + .mm_mt = MTREE_INIT_EXT(mm_mt, MM_MT_FLAGS, efi_mm.mmap_lock), .mm_users = ATOMIC_INIT(2), .mm_count = ATOMIC_INIT(1), .write_protect_seq = SEQCNT_ZERO(efi_mm.write_protect_seq), diff --git a/include/linux/mm.h b/include/linux/mm.h index cf3d0d673f6b..adc963765d95 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2581,6 +2581,8 @@ extern bool arch_has_descending_max_zone_pfns(void); /* nommu.c */ extern atomic_long_t mmap_pages_allocated; extern int nommu_shrink_inode_mappings(struct inode *, size_t, size_t); +/* mmap.c */ +void vma_mas_store(struct vm_area_struct *vma, struct ma_state *mas); /* interval_tree.c */ void vma_interval_tree_insert(struct vm_area_struct *node, @@ -2644,6 +2646,9 @@ extern struct vm_area_struct *copy_vma(struct vm_area_struct **, bool *need_rmap_locks); extern void exit_mmap(struct mm_struct *); +void vma_mas_store(struct vm_area_struct *vma, struct ma_state *mas); +void vma_mas_remove(struct vm_area_struct *vma, struct ma_state *mas); + static inline int check_data_rlimit(unsigned long rlim, unsigned long new, unsigned long start, diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 6b961a29bf26..e810aaca6c04 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -9,6 +9,7 @@ #include #include #include +#include #include #include #include @@ -481,6 +482,7 @@ struct kioctx_table; struct mm_struct { struct { struct vm_area_struct *mmap; /* list of VMAs */ + struct maple_tree mm_mt; struct rb_root mm_rb; u64 vmacache_seqnum; /* per-thread vmacache */ #ifdef CONFIG_MMU @@ -676,6 +678,7 @@ struct mm_struct { unsigned long cpu_bitmap[]; }; +#define MM_MT_FLAGS (MT_FLAGS_ALLOC_RANGE | MT_FLAGS_LOCK_EXTERN) extern struct mm_struct init_mm; /* Pointer magic because the dynamic array size confuses some compilers. */ diff --git a/include/trace/events/mmap.h b/include/trace/events/mmap.h index 4661f7ba07c0..216de5f03621 100644 --- a/include/trace/events/mmap.h +++ b/include/trace/events/mmap.h @@ -42,6 +42,79 @@ TRACE_EVENT(vm_unmapped_area, __entry->low_limit, __entry->high_limit, __entry->align_mask, __entry->align_offset) ); + +TRACE_EVENT(vma_mas_szero, + TP_PROTO(struct maple_tree *mt, unsigned long start, + unsigned long end), + + TP_ARGS(mt, start, end), + + TP_STRUCT__entry( + __field(struct maple_tree *, mt) + __field(unsigned long, start) + __field(unsigned long, end) + ), + + TP_fast_assign( + __entry->mt = mt; + __entry->start = start; + __entry->end = end; + ), + + TP_printk("mt_mod %p, (NULL), SNULL, %lu, %lu,", + __entry->mt, + (unsigned long) __entry->start, + (unsigned long) __entry->end + ) +); + +TRACE_EVENT(vma_store, + TP_PROTO(struct maple_tree *mt, struct vm_area_struct *vma), + + TP_ARGS(mt, vma), + + TP_STRUCT__entry( + __field(struct maple_tree *, mt) + __field(struct vm_area_struct *, vma) + __field(unsigned long, vm_start) + __field(unsigned long, vm_end) + ), + + TP_fast_assign( + __entry->mt = mt; + __entry->vma = vma; + __entry->vm_start = vma->vm_start; + __entry->vm_end = vma->vm_end - 1; + ), + + TP_printk("mt_mod %p, (%p), STORE, %lu, %lu,", + __entry->mt, __entry->vma, + (unsigned long) __entry->vm_start, + (unsigned long) __entry->vm_end + ) +); + + +TRACE_EVENT(exit_mmap, + TP_PROTO(struct mm_struct *mm), + + TP_ARGS(mm), + + TP_STRUCT__entry( + __field(struct mm_struct *, mm) + __field(struct maple_tree *, mt) + ), + + TP_fast_assign( + __entry->mm = mm; + __entry->mt = &mm->mm_mt; + ), + + TP_printk("mt_mod %p, DESTROY\n", + __entry->mt + ) +); + #endif /* This part must be outside protection */ diff --git a/kernel/fork.c b/kernel/fork.c index 9d44f2d46c69..1840da0732f6 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -585,6 +585,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, int retval; unsigned long charge; LIST_HEAD(uf); + MA_STATE(mas, &mm->mm_mt, 0, 0); uprobe_start_dup_mmap(); if (mmap_write_lock_killable(oldmm)) { @@ -614,6 +615,10 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, goto out; khugepaged_fork(mm, oldmm); + retval = mas_expected_entries(&mas, oldmm->map_count); + if (retval) + goto out; + prev = NULL; for (mpnt = oldmm->mmap; mpnt; mpnt = mpnt->vm_next) { struct file *file; @@ -629,7 +634,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, */ if (fatal_signal_pending(current)) { retval = -EINTR; - goto out; + goto loop_out; } if (mpnt->vm_flags & VM_ACCOUNT) { unsigned long len = vma_pages(mpnt); @@ -694,6 +699,11 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, rb_link = &tmp->vm_rb.rb_right; rb_parent = &tmp->vm_rb; + /* Link the vma into the MT */ + mas.index = tmp->vm_start; + mas.last = tmp->vm_end - 1; + mas_store(&mas, tmp); + mm->map_count++; if (!(tmp->vm_flags & VM_WIPEONFORK)) retval = copy_page_range(tmp, mpnt); @@ -702,10 +712,12 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, tmp->vm_ops->open(tmp); if (retval) - goto out; + goto loop_out; } /* a new mm has just been created */ retval = arch_dup_mmap(oldmm, mm); +loop_out: + mas_destroy(&mas); out: mmap_write_unlock(mm); flush_tlb_mm(oldmm); @@ -721,7 +733,7 @@ fail_nomem_policy: fail_nomem: retval = -ENOMEM; vm_unacct_memory(charge); - goto out; + goto loop_out; } static inline int mm_alloc_pgd(struct mm_struct *mm) @@ -1111,6 +1123,8 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, { mm->mmap = NULL; mm->mm_rb = RB_ROOT; + mt_init_flags(&mm->mm_mt, MM_MT_FLAGS); + mt_set_external_lock(&mm->mm_mt, &mm->mmap_lock); mm->vmacache_seqnum = 0; atomic_set(&mm->mm_users, 1); atomic_set(&mm->mm_count, 1); diff --git a/mm/init-mm.c b/mm/init-mm.c index fbe7844d0912..b912b0f2eced 100644 --- a/mm/init-mm.c +++ b/mm/init-mm.c @@ -1,6 +1,7 @@ // SPDX-License-Identifier: GPL-2.0 #include #include +#include #include #include #include @@ -29,6 +30,7 @@ */ struct mm_struct init_mm = { .mm_rb = RB_ROOT, + .mm_mt = MTREE_INIT_EXT(mm_mt, MM_MT_FLAGS, init_mm.mmap_lock), .pgd = swapper_pg_dir, .mm_users = ATOMIC_INIT(2), .mm_count = ATOMIC_INIT(1), diff --git a/mm/mmap.c b/mm/mmap.c index c14d7286a379..79654a627baf 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -381,7 +381,70 @@ static int browse_rb(struct mm_struct *mm) } return bug ? -1 : i; } +#if defined(CONFIG_DEBUG_VM_MAPLE_TREE) +extern void mt_validate(struct maple_tree *mt); +extern void mt_dump(const struct maple_tree *mt); +/* Validate the maple tree */ +static void validate_mm_mt(struct mm_struct *mm) +{ + struct maple_tree *mt = &mm->mm_mt; + struct vm_area_struct *vma_mt, *vma = mm->mmap; + + MA_STATE(mas, mt, 0, 0); + mas_for_each(&mas, vma_mt, ULONG_MAX) { + if (xa_is_zero(vma_mt)) + continue; + + if (!vma) + break; + + if ((vma != vma_mt) || + (vma->vm_start != vma_mt->vm_start) || + (vma->vm_end != vma_mt->vm_end) || + (vma->vm_start != mas.index) || + (vma->vm_end - 1 != mas.last)) { + pr_emerg("issue in %s\n", current->comm); + dump_stack(); +#ifdef CONFIG_DEBUG_VM + dump_vma(vma_mt); + pr_emerg("and next in rb\n"); + dump_vma(vma->vm_next); +#endif + pr_emerg("mt piv: %px %lu - %lu\n", vma_mt, + mas.index, mas.last); + pr_emerg("mt vma: %px %lu - %lu\n", vma_mt, + vma_mt->vm_start, vma_mt->vm_end); + pr_emerg("rb vma: %px %lu - %lu\n", vma, + vma->vm_start, vma->vm_end); + pr_emerg("rb->next = %px %lu - %lu\n", vma->vm_next, + vma->vm_next->vm_start, vma->vm_next->vm_end); + + mt_dump(mas.tree); + if (vma_mt->vm_end != mas.last + 1) { + pr_err("vma: %px vma_mt %lu-%lu\tmt %lu-%lu\n", + mm, vma_mt->vm_start, vma_mt->vm_end, + mas.index, mas.last); + mt_dump(mas.tree); + } + VM_BUG_ON_MM(vma_mt->vm_end != mas.last + 1, mm); + if (vma_mt->vm_start != mas.index) { + pr_err("vma: %px vma_mt %px %lu - %lu doesn't match\n", + mm, vma_mt, vma_mt->vm_start, vma_mt->vm_end); + mt_dump(mas.tree); + } + VM_BUG_ON_MM(vma_mt->vm_start != mas.index, mm); + } + VM_BUG_ON(vma != vma_mt); + vma = vma->vm_next; + + } + VM_BUG_ON(vma); + mt_validate(&mm->mm_mt); +} +#else +#define validate_mm_mt(root) do { } while (0) +#endif static void validate_mm_rb(struct rb_root *root, struct vm_area_struct *ignore) { struct rb_node *nd; @@ -436,6 +499,7 @@ static void validate_mm(struct mm_struct *mm) } #else #define validate_mm_rb(root, ignore) do { } while (0) +#define validate_mm_mt(root) do { } while (0) #define validate_mm(mm) do { } while (0) #endif @@ -680,6 +744,56 @@ static void __vma_link_file(struct vm_area_struct *vma) } } +/* + * vma_mas_store() - Store a VMA in the maple tree. + * @vma: The vm_area_struct + * @mas: The maple state + * + * Efficient way to store a VMA in the maple tree when the @mas has already + * walked to the correct location. + * + * Note: the end address is inclusive in the maple tree. + */ +void vma_mas_store(struct vm_area_struct *vma, struct ma_state *mas) +{ + trace_vma_store(mas->tree, vma); + mas_set_range(mas, vma->vm_start, vma->vm_end - 1); + mas_store_prealloc(mas, vma); +} + +/* + * vma_mas_remove() - Remove a VMA from the maple tree. + * @vma: The vm_area_struct + * @mas: The maple state + * + * Efficient way to remove a VMA from the maple tree when the @mas has already + * been established and points to the correct location. + * Note: the end address is inclusive in the maple tree. + */ +void vma_mas_remove(struct vm_area_struct *vma, struct ma_state *mas) +{ + trace_vma_mas_szero(mas->tree, vma->vm_start, vma->vm_end - 1); + mas->index = vma->vm_start; + mas->last = vma->vm_end - 1; + mas_store_prealloc(mas, NULL); +} + +/* + * vma_mas_szero() - Set a given range to zero. Used when modifying a + * vm_area_struct start or end. + * + * @mm: The struct_mm + * @start: The start address to zero + * @end: The end address to zero. + */ +static inline void vma_mas_szero(struct ma_state *mas, unsigned long start, + unsigned long end) +{ + trace_vma_mas_szero(mas->tree, start, end - 1); + mas_set_range(mas, start, end - 1); + mas_store_prealloc(mas, NULL); +} + static void __vma_link(struct mm_struct *mm, struct vm_area_struct *vma, struct vm_area_struct *prev, struct rb_node **rb_link, @@ -689,17 +803,22 @@ __vma_link(struct mm_struct *mm, struct vm_area_struct *vma, __vma_link_rb(mm, vma, rb_link, rb_parent); } -static void vma_link(struct mm_struct *mm, struct vm_area_struct *vma, +static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma, struct vm_area_struct *prev, struct rb_node **rb_link, struct rb_node *rb_parent) { + MA_STATE(mas, &mm->mm_mt, 0, 0); struct address_space *mapping = NULL; + if (mas_preallocate(&mas, vma, GFP_KERNEL)) + return -ENOMEM; + if (vma->vm_file) { mapping = vma->vm_file->f_mapping; i_mmap_lock_write(mapping); } + vma_mas_store(vma, &mas); __vma_link(mm, vma, prev, rb_link, rb_parent); __vma_link_file(vma); @@ -708,13 +827,15 @@ static void vma_link(struct mm_struct *mm, struct vm_area_struct *vma, mm->map_count++; validate_mm(mm); + return 0; } /* * Helper for vma_adjust() in the split_vma insert case: insert a vma into the * mm's list and rbtree. It has already been inserted into the interval tree. */ -static void __insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma) +static void __insert_vm_struct(struct mm_struct *mm, struct ma_state *mas, + struct vm_area_struct *vma) { struct vm_area_struct *prev; struct rb_node **rb_link, *rb_parent; @@ -722,7 +843,10 @@ static void __insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma) if (find_vma_links(mm, vma->vm_start, vma->vm_end, &prev, &rb_link, &rb_parent)) BUG(); - __vma_link(mm, vma, prev, rb_link, rb_parent); + + vma_mas_store(vma, mas); + __vma_link_list(mm, vma, prev); + __vma_link_rb(mm, vma, rb_link, rb_parent); mm->map_count++; } @@ -756,10 +880,13 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, bool start_changed = false, end_changed = false; long adjust_next = 0; int remove_next = 0; + MA_STATE(mas, &mm->mm_mt, 0, 0); + struct vm_area_struct *exporter = NULL, *importer = NULL; - if (next && !insert) { - struct vm_area_struct *exporter = NULL, *importer = NULL; + validate_mm(mm); + validate_mm_mt(mm); + if (next && !insert) { if (end >= next->vm_end) { /* * vma expands, overlapping all the next, and @@ -842,6 +969,12 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, again: vma_adjust_trans_huge(orig_vma, start, end, adjust_next); + if (mas_preallocate(&mas, vma, GFP_KERNEL)) { + if (exporter && exporter->anon_vma) + unlink_anon_vmas(importer); + return -ENOMEM; + } + if (file) { mapping = file->f_mapping; root = &mapping->i_mmap; @@ -882,17 +1015,28 @@ again: } if (start != vma->vm_start) { + unsigned long old_start = vma->vm_start; vma->vm_start = start; + if (old_start < start) + vma_mas_szero(&mas, old_start, start); start_changed = true; } if (end != vma->vm_end) { + unsigned long old_end = vma->vm_end; vma->vm_end = end; + if (old_end > end) + vma_mas_szero(&mas, end, old_end); end_changed = true; } + + if (end_changed || start_changed) + vma_mas_store(vma, &mas); + vma->vm_pgoff = pgoff; if (adjust_next) { next->vm_start += adjust_next; next->vm_pgoff += adjust_next >> PAGE_SHIFT; + vma_mas_store(next, &mas); } if (file) { @@ -906,6 +1050,8 @@ again: /* * vma_merge has merged next into vma, and needs * us to remove next before dropping the locks. + * Since we have expanded over this vma, the maple tree will + * have overwritten by storing the value */ if (remove_next != 3) __vma_unlink(mm, next, next); @@ -928,7 +1074,7 @@ again: * us to insert it before dropping the locks * (it may either follow vma or precede it). */ - __insert_vm_struct(mm, insert); + __insert_vm_struct(mm, &mas, insert); } else { if (start_changed) vma_gap_update(vma); @@ -1025,6 +1171,7 @@ again: uprobe_mmap(insert); validate_mm(mm); + validate_mm_mt(mm); return 0; } @@ -1178,6 +1325,7 @@ struct vm_area_struct *vma_merge(struct mm_struct *mm, struct vm_area_struct *area, *next; int err; + validate_mm_mt(mm); /* * We later require that vma->vm_flags == vm_flags, * so this tests vma->vm_flags & VM_SPECIAL, too. @@ -1253,6 +1401,7 @@ struct vm_area_struct *vma_merge(struct mm_struct *mm, khugepaged_enter_vma(area, vm_flags); return area; } + validate_mm_mt(mm); return NULL; } @@ -1732,6 +1881,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr, struct rb_node **rb_link, *rb_parent; unsigned long charged = 0; + validate_mm_mt(mm); /* Check against address space limit. */ if (!may_expand_vm(mm, vm_flags, len >> PAGE_SHIFT)) { unsigned long nr_pages; @@ -1846,7 +1996,13 @@ unsigned long mmap_region(struct file *file, unsigned long addr, goto free_vma; } - vma_link(mm, vma, prev, rb_link, rb_parent); + if (vma_link(mm, vma, prev, rb_link, rb_parent)) { + error = -ENOMEM; + if (file) + goto unmap_and_free_vma; + else + goto free_vma; + } /* * vma_merge() calls khugepaged_enter_vma() either, the below @@ -1886,6 +2042,7 @@ out: vma_set_page_prot(vma); + validate_mm_mt(mm); return addr; unmap_and_free_vma: @@ -1902,6 +2059,7 @@ free_vma: unacct_error: if (charged) vm_unacct_memory(charged); + validate_mm_mt(mm); return error; } @@ -1918,12 +2076,19 @@ static unsigned long unmapped_area(struct vm_unmapped_area_info *info) struct mm_struct *mm = current->mm; struct vm_area_struct *vma; unsigned long length, low_limit, high_limit, gap_start, gap_end; + unsigned long gap; + MA_STATE(mas, &mm->mm_mt, 0, 0); /* Adjust search length to account for worst case alignment overhead */ length = info->length + info->align_mask; if (length < info->length) return -ENOMEM; + mas_empty_area(&mas, info->low_limit, info->high_limit - 1, + length); + gap = mas.index; + gap += (info->align_offset - gap) & info->align_mask; + /* Adjust search limits by the desired length */ if (info->high_limit < length) return -ENOMEM; @@ -2005,20 +2170,31 @@ found: VM_BUG_ON(gap_start + info->length > info->high_limit); VM_BUG_ON(gap_start + info->length > gap_end); + + VM_BUG_ON(gap != gap_start); return gap_start; } static unsigned long unmapped_area_topdown(struct vm_unmapped_area_info *info) { struct mm_struct *mm = current->mm; - struct vm_area_struct *vma; + struct vm_area_struct *vma = NULL; unsigned long length, low_limit, high_limit, gap_start, gap_end; + unsigned long gap; + + MA_STATE(mas, &mm->mm_mt, 0, 0); + validate_mm_mt(mm); /* Adjust search length to account for worst case alignment overhead */ length = info->length + info->align_mask; if (length < info->length) return -ENOMEM; + mas_empty_area_rev(&mas, info->low_limit, info->high_limit - 1, + length); + gap = mas.last + 1 - info->length; + gap -= (gap - info->align_offset) & info->align_mask; + /* * Adjust search limits by the desired length. * See implementation comment at top of unmapped_area(). @@ -2104,6 +2280,32 @@ found_highest: VM_BUG_ON(gap_end < info->low_limit); VM_BUG_ON(gap_end < gap_start); + + if (gap != gap_end) { + pr_err("%s: %px Gap was found: mt %lu gap_end %lu\n", __func__, + mm, gap, gap_end); + pr_err("window was %lu - %lu size %lu\n", info->high_limit, + info->low_limit, length); + pr_err("mas.min %lu max %lu mas.last %lu\n", mas.min, mas.max, + mas.last); + pr_err("mas.index %lu align mask %lu offset %lu\n", mas.index, + info->align_mask, info->align_offset); + pr_err("rb_find_vma find on %lu => %px (%px)\n", mas.index, + find_vma(mm, mas.index), vma); +#if defined(CONFIG_DEBUG_VM_MAPLE_TREE) + mt_dump(&mm->mm_mt); +#endif + { + struct vm_area_struct *dv = mm->mmap; + + while (dv) { + printk("vma %px %lu-%lu\n", dv, dv->vm_start, dv->vm_end); + dv = dv->vm_next; + } + } + VM_BUG_ON(gap != gap_end); + } + return gap_end; } @@ -2326,7 +2528,6 @@ struct vm_area_struct *find_vma(struct mm_struct *mm, unsigned long addr) vmacache_update(addr, vma); return vma; } - EXPORT_SYMBOL(find_vma); /* @@ -2399,7 +2600,9 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) struct vm_area_struct *next; unsigned long gap_addr; int error = 0; + MA_STATE(mas, &mm->mm_mt, 0, 0); + validate_mm_mt(mm); if (!(vma->vm_flags & VM_GROWSUP)) return -EFAULT; @@ -2423,9 +2626,14 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) /* Check that both stack segments have the same anon_vma? */ } + if (mas_preallocate(&mas, vma, GFP_KERNEL)) + return -ENOMEM; + /* We must make sure the anon_vma is allocated. */ - if (unlikely(anon_vma_prepare(vma))) + if (unlikely(anon_vma_prepare(vma))) { + mas_destroy(&mas); return -ENOMEM; + } /* * vma->vm_start/vm_end cannot change under us because the caller @@ -2462,6 +2670,8 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) vm_stat_account(mm, vma->vm_flags, grow); anon_vma_interval_tree_pre_update_vma(vma); vma->vm_end = address; + /* Overwrite old entry in mtree. */ + vma_mas_store(vma, &mas); anon_vma_interval_tree_post_update_vma(vma); if (vma->vm_next) vma_gap_update(vma->vm_next); @@ -2476,6 +2686,8 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) anon_vma_unlock_write(vma->anon_vma); khugepaged_enter_vma(vma, vma->vm_flags); validate_mm(mm); + validate_mm_mt(mm); + mas_destroy(&mas); return error; } #endif /* CONFIG_STACK_GROWSUP || CONFIG_IA64 */ @@ -2489,7 +2701,9 @@ int expand_downwards(struct vm_area_struct *vma, struct mm_struct *mm = vma->vm_mm; struct vm_area_struct *prev; int error = 0; + MA_STATE(mas, &mm->mm_mt, 0, 0); + validate_mm(mm); address &= PAGE_MASK; if (address < mmap_min_addr) return -EPERM; @@ -2503,9 +2717,14 @@ int expand_downwards(struct vm_area_struct *vma, return -ENOMEM; } + if (mas_preallocate(&mas, vma, GFP_KERNEL)) + return -ENOMEM; + /* We must make sure the anon_vma is allocated. */ - if (unlikely(anon_vma_prepare(vma))) + if (unlikely(anon_vma_prepare(vma))) { + mas_destroy(&mas); return -ENOMEM; + } /* * vma->vm_start/vm_end cannot change under us because the caller @@ -2543,6 +2762,8 @@ int expand_downwards(struct vm_area_struct *vma, anon_vma_interval_tree_pre_update_vma(vma); vma->vm_start = address; vma->vm_pgoff -= grow; + /* Overwrite old entry in mtree. */ + vma_mas_store(vma, &mas); anon_vma_interval_tree_post_update_vma(vma); vma_gap_update(vma); spin_unlock(&mm->page_table_lock); @@ -2554,6 +2775,7 @@ int expand_downwards(struct vm_area_struct *vma, anon_vma_unlock_write(vma->anon_vma); khugepaged_enter_vma(vma, vma->vm_flags); validate_mm(mm); + mas_destroy(&mas); return error; } @@ -2676,14 +2898,17 @@ static void unmap_region(struct mm_struct *mm, * vma list as we go.. */ static bool -detach_vmas_to_be_unmapped(struct mm_struct *mm, struct vm_area_struct *vma, - struct vm_area_struct *prev, unsigned long end) +detach_vmas_to_be_unmapped(struct mm_struct *mm, struct ma_state *mas, + struct vm_area_struct *vma, struct vm_area_struct *prev, + unsigned long end) { struct vm_area_struct **insertion_point; struct vm_area_struct *tail_vma = NULL; insertion_point = (prev ? &prev->vm_next : &mm->mmap); vma->vm_prev = NULL; + mas_set_range(mas, vma->vm_start, end - 1); + mas_store_prealloc(mas, NULL); do { vma_rb_erase(vma, &mm->mm_rb); if (vma->vm_flags & VM_LOCKED) @@ -2724,6 +2949,7 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma, { struct vm_area_struct *new; int err; + validate_mm_mt(mm); if (vma->vm_ops && vma->vm_ops->may_split) { err = vma->vm_ops->may_split(vma, addr); @@ -2776,6 +3002,7 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma, mpol_put(vma_policy(new)); out_free_vma: vm_area_free(new); + validate_mm_mt(mm); return err; } @@ -2802,6 +3029,8 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, { unsigned long end; struct vm_area_struct *vma, *prev, *last; + int error = -ENOMEM; + MA_STATE(mas, &mm->mm_mt, 0, 0); if ((offset_in_page(start)) || start > TASK_SIZE || len > TASK_SIZE-start) return -EINVAL; @@ -2822,6 +3051,9 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, vma = find_vma_intersection(mm, start, end); if (!vma) return 0; + + if (mas_preallocate(&mas, vma, GFP_KERNEL)) + return -ENOMEM; prev = vma->vm_prev; /* @@ -2832,7 +3064,6 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, * places tmp vma above, and higher split_vma places tmp vma below. */ if (start > vma->vm_start) { - int error; /* * Make sure that map_count on return from munmap() will @@ -2840,20 +3071,20 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, * its limit temporarily, to help free resources as expected. */ if (end < vma->vm_end && mm->map_count >= sysctl_max_map_count) - return -ENOMEM; + goto map_count_exceeded; error = __split_vma(mm, vma, start, 0); if (error) - return error; + goto split_failed; prev = vma; } /* Does it split the last one? */ last = find_vma(mm, end); if (last && end > last->vm_start) { - int error = __split_vma(mm, last, end, 1); + error = __split_vma(mm, last, end, 1); if (error) - return error; + goto split_failed; } vma = vma_next(mm, prev); @@ -2867,13 +3098,13 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, * split, despite we could. This is unlikely enough * failure that it's not worth optimizing it for. */ - int error = userfaultfd_unmap_prep(vma, start, end, uf); + error = userfaultfd_unmap_prep(vma, start, end, uf); if (error) - return error; + goto userfaultfd_error; } /* Detach vmas from rbtree */ - if (!detach_vmas_to_be_unmapped(mm, vma, prev, end)) + if (!detach_vmas_to_be_unmapped(mm, &mas, vma, prev, end)) downgrade = false; if (downgrade) @@ -2885,6 +3116,12 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, remove_vma_list(mm, vma); return downgrade ? 1 : 0; + +map_count_exceeded: +split_failed: +userfaultfd_error: + mas_destroy(&mas); + return error; } int do_munmap(struct mm_struct *mm, unsigned long start, size_t len, @@ -3024,6 +3261,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla pgoff_t pgoff = addr >> PAGE_SHIFT; int error; unsigned long mapped_addr; + validate_mm_mt(mm); /* Until we need other flags, refuse anything except VM_EXEC. */ if ((flags & (~VM_EXEC)) != 0) @@ -3073,7 +3311,9 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla vma->vm_pgoff = pgoff; vma->vm_flags = flags; vma->vm_page_prot = vm_get_page_prot(flags); - vma_link(mm, vma, prev, rb_link, rb_parent); + if(vma_link(mm, vma, prev, rb_link, rb_parent)) + goto no_vma_link; + out: perf_event_mmap(vma); mm->total_vm += len >> PAGE_SHIFT; @@ -3081,7 +3321,12 @@ out: if (flags & VM_LOCKED) mm->locked_vm += (len >> PAGE_SHIFT); vma->vm_flags |= VM_SOFTDIRTY; + validate_mm_mt(mm); return 0; + +no_vma_link: + vm_area_free(vma); + return -ENOMEM; } int vm_brk_flags(unsigned long addr, unsigned long request, unsigned long flags) @@ -3170,6 +3415,9 @@ void exit_mmap(struct mm_struct *mm) vma = remove_vma(vma); cond_resched(); } + + trace_exit_mmap(mm); + __mt_destroy(&mm->mm_mt); mm->mmap = NULL; mmap_write_unlock(mm); vm_unacct_memory(nr_accounted); @@ -3183,10 +3431,25 @@ int insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma) { struct vm_area_struct *prev; struct rb_node **rb_link, *rb_parent; + unsigned long start = vma->vm_start; + struct vm_area_struct *overlap = NULL; if (find_vma_links(mm, vma->vm_start, vma->vm_end, &prev, &rb_link, &rb_parent)) return -ENOMEM; + + overlap = mt_find(&mm->mm_mt, &start, vma->vm_end - 1); + if (overlap) { + + pr_err("Found vma ending at %lu\n", start - 1); + pr_err("vma : %lu => %lu-%lu\n", (unsigned long)overlap, + overlap->vm_start, overlap->vm_end - 1); +#if defined(CONFIG_DEBUG_VM_MAPLE_TREE) + mt_dump(&mm->mm_mt); +#endif + BUG(); + } + if ((vma->vm_flags & VM_ACCOUNT) && security_vm_enough_memory_mm(mm, vma_pages(vma))) return -ENOMEM; @@ -3208,7 +3471,9 @@ int insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma) vma->vm_pgoff = vma->vm_start >> PAGE_SHIFT; } - vma_link(mm, vma, prev, rb_link, rb_parent); + if (vma_link(mm, vma, prev, rb_link, rb_parent)) + return -ENOMEM; + return 0; } @@ -3226,7 +3491,9 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap, struct vm_area_struct *new_vma, *prev; struct rb_node **rb_link, *rb_parent; bool faulted_in_anon_vma = true; + unsigned long index = addr; + validate_mm_mt(mm); /* * If anonymous vma has not yet been faulted, update new pgoff * to match new location, to increase its chance of merging. @@ -3238,6 +3505,8 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap, if (find_vma_links(mm, addr, addr + len, &prev, &rb_link, &rb_parent)) return NULL; /* should never get here */ + if (mt_find(&mm->mm_mt, &index, addr+len - 1)) + BUG(); new_vma = vma_merge(mm, prev, addr, addr + len, vma->vm_flags, vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma), vma->vm_userfaultfd_ctx, anon_vma_name(vma)); @@ -3281,6 +3550,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap, vma_link(mm, new_vma, prev, rb_link, rb_parent); *need_rmap_locks = false; } + validate_mm_mt(mm); return new_vma; out_free_mempol: @@ -3288,6 +3558,7 @@ out_free_mempol: out_free_vma: vm_area_free(new_vma); out: + validate_mm_mt(mm); return NULL; } @@ -3424,6 +3695,7 @@ static struct vm_area_struct *__install_special_mapping( int ret; struct vm_area_struct *vma; + validate_mm_mt(mm); vma = vm_area_alloc(mm); if (unlikely(vma == NULL)) return ERR_PTR(-ENOMEM); @@ -3446,10 +3718,12 @@ static struct vm_area_struct *__install_special_mapping( perf_event_mmap(vma); + validate_mm_mt(mm); return vma; out: vm_area_free(vma); + validate_mm_mt(mm); return ERR_PTR(ret); } diff --git a/mm/nommu.c b/mm/nommu.c index 9d7afc2d959e..5af0b050eba8 100644 --- a/mm/nommu.c +++ b/mm/nommu.c @@ -545,6 +545,19 @@ static void put_nommu_region(struct vm_region *region) __put_nommu_region(region); } +void vma_mas_store(struct vm_area_struct *vma, struct ma_state *mas) +{ + mas_set_range(mas, vma->vm_start, vma->vm_end - 1); + mas_store_prealloc(mas, vma); +} + +void vma_mas_remove(struct vm_area_struct *vma, struct ma_state *mas) +{ + mas->index = vma->vm_start; + mas->last = vma->vm_end - 1; + mas_store_prealloc(mas, NULL); +} + /* * add a VMA into a process's mm_struct in the appropriate place in the list * and tree and add to the address space's page tree also if not an anonymous -- cgit v1.2.3 From 985ab5390e3d29c368e9f2401219f530b1965aa8 Mon Sep 17 00:00:00 2001 From: "Matthew Wilcox (Oracle)" Date: Tue, 21 Jun 2022 20:46:54 +0000 Subject: mm: add VMA iterator This thin layer of abstraction over the maple tree state is for iterating over VMAs. You can go forwards, go backwards or ask where the iterator is. Rename the existing vma_next() to __vma_next() -- it will be removed by the end of this series. Link: https://lkml.kernel.org/r/20220504010716.661115-11-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-10-Liam.Howlett@oracle.com Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Liam R. Howlett Acked-by: Vlastimil Babka Reviewed-by: David Hildenbrand Cc: Catalin Marinas Cc: David Howells Cc: SeongJae Park Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- include/linux/mm.h | 31 +++++++++++++++++++++++++++++++ include/linux/mm_types.h | 21 +++++++++++++++++++++ mm/mmap.c | 10 +++++----- 3 files changed, 57 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index adc963765d95..510289e6a447 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -658,6 +658,37 @@ static inline bool vma_is_accessible(struct vm_area_struct *vma) return vma->vm_flags & VM_ACCESS_FLAGS; } +static inline +struct vm_area_struct *vma_find(struct vma_iterator *vmi, unsigned long max) +{ + return mas_find(&vmi->mas, max); +} + +static inline struct vm_area_struct *vma_next(struct vma_iterator *vmi) +{ + /* + * Uses vma_find() to get the first VMA when the iterator starts. + * Calling mas_next() could skip the first entry. + */ + return vma_find(vmi, ULONG_MAX); +} + +static inline struct vm_area_struct *vma_prev(struct vma_iterator *vmi) +{ + return mas_prev(&vmi->mas, 0); +} + +static inline unsigned long vma_iter_addr(struct vma_iterator *vmi) +{ + return vmi->mas.index; +} + +#define for_each_vma(vmi, vma) while ((vma = vma_next(&(vmi))) != NULL) + +/* The MM code likes to work with exclusive end addresses */ +#define for_each_vma_range(vmi, vma, end) \ + while ((vma = vma_find(&(vmi), (end) - 1)) != NULL) + #ifdef CONFIG_SHMEM /* * The vma_is_shmem is not inline because it is used only by slow diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index e810aaca6c04..a0e4b638d6b4 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -696,6 +696,27 @@ static inline cpumask_t *mm_cpumask(struct mm_struct *mm) return (struct cpumask *)&mm->cpu_bitmap; } +struct vma_iterator { + struct ma_state mas; +}; + +#define VMA_ITERATOR(name, mm, addr) \ + struct vma_iterator name = { \ + .mas = { \ + .tree = &mm->mm_mt, \ + .index = addr, \ + .node = MAS_START, \ + }, \ + } + +static inline void vma_iter_init(struct vma_iterator *vmi, + struct mm_struct *mm, unsigned long addr) +{ + vmi->mas.tree = &mm->mm_mt; + vmi->mas.index = addr; + vmi->mas.node = MAS_START; +} + struct mmu_gather; extern void tlb_gather_mmu(struct mmu_gather *tlb, struct mm_struct *mm); extern void tlb_gather_mmu_fullmm(struct mmu_gather *tlb, struct mm_struct *mm); diff --git a/mm/mmap.c b/mm/mmap.c index 79654a627baf..9b0273e7d41e 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -632,7 +632,7 @@ static int find_vma_links(struct mm_struct *mm, unsigned long addr, } /* - * vma_next() - Get the next VMA. + * __vma_next() - Get the next VMA. * @mm: The mm_struct. * @vma: The current vma. * @@ -640,7 +640,7 @@ static int find_vma_links(struct mm_struct *mm, unsigned long addr, * * Returns: The next VMA after @vma. */ -static inline struct vm_area_struct *vma_next(struct mm_struct *mm, +static inline struct vm_area_struct *__vma_next(struct mm_struct *mm, struct vm_area_struct *vma) { if (!vma) @@ -1333,7 +1333,7 @@ struct vm_area_struct *vma_merge(struct mm_struct *mm, if (vm_flags & VM_SPECIAL) return NULL; - next = vma_next(mm, prev); + next = __vma_next(mm, prev); area = next; if (area && area->vm_end == end) /* cases 6, 7, 8 */ next = next->vm_next; @@ -2881,7 +2881,7 @@ static void unmap_region(struct mm_struct *mm, struct vm_area_struct *vma, struct vm_area_struct *prev, unsigned long start, unsigned long end) { - struct vm_area_struct *next = vma_next(mm, prev); + struct vm_area_struct *next = __vma_next(mm, prev); struct mmu_gather tlb; lru_add_drain(); @@ -3086,7 +3086,7 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, if (error) goto split_failed; } - vma = vma_next(mm, prev); + vma = __vma_next(mm, prev); if (unlikely(uf)) { /* -- cgit v1.2.3 From 9caf5c761da60146c42c5fb8da6db9034741254b Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:46:55 +0000 Subject: kernel/fork: use maple tree for dup_mmap() during forking The maple tree was already tracking VMAs in this function by an earlier commit, but the rbtree iterator was being used to iterate the list. Change the iterator to use a maple tree native iterator and switch to the maple tree advanced API to avoid multiple walks of the tree during insert operations. Unexport the now-unused vma_store() function. For performance reasons we bulk allocate the maple tree nodes. The node calculations are done internally to the tree and use the VMA count and assume the worst-case node requirements. The VM_DONT_COPY flag does not allow for the most efficient copy method of the tree and so a bulk loading algorithm is used. Link: https://lkml.kernel.org/r/20220504010716.661115-16-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-15-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Signed-off-by: Matthew Wilcox (Oracle) Acked-by: Vlastimil Babka Cc: Catalin Marinas Cc: David Howells Cc: SeongJae Park Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- include/linux/mm.h | 2 -- kernel/fork.c | 12 +++++++++--- 2 files changed, 9 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 510289e6a447..4753138e4c98 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2612,8 +2612,6 @@ extern bool arch_has_descending_max_zone_pfns(void); /* nommu.c */ extern atomic_long_t mmap_pages_allocated; extern int nommu_shrink_inode_mappings(struct inode *, size_t, size_t); -/* mmap.c */ -void vma_mas_store(struct vm_area_struct *vma, struct ma_state *mas); /* interval_tree.c */ void vma_interval_tree_insert(struct vm_area_struct *node, diff --git a/kernel/fork.c b/kernel/fork.c index 1840da0732f6..4ecca79556eb 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -583,9 +583,10 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, struct vm_area_struct *mpnt, *tmp, *prev, **pprev; struct rb_node **rb_link, *rb_parent; int retval; - unsigned long charge; - LIST_HEAD(uf); + unsigned long charge = 0; + MA_STATE(old_mas, &oldmm->mm_mt, 0, 0); MA_STATE(mas, &mm->mm_mt, 0, 0); + LIST_HEAD(uf); uprobe_start_dup_mmap(); if (mmap_write_lock_killable(oldmm)) { @@ -620,7 +621,12 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, goto out; prev = NULL; - for (mpnt = oldmm->mmap; mpnt; mpnt = mpnt->vm_next) { + + retval = mas_expected_entries(&mas, oldmm->map_count); + if (retval) + goto out; + + mas_for_each(&old_mas, mpnt, ULONG_MAX) { struct file *file; if (mpnt->vm_flags & VM_DONTCOPY) { -- cgit v1.2.3 From 30cad3debecefcc89ae4959800be7d54af4e992c Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:46:56 +0000 Subject: mm: remove rb tree. Remove the RB tree and start using the maple tree for vm_area_struct tracking. Drop validate_mm() calls in expand_upwards() and expand_downwards() as the lock is not held. Link: https://lkml.kernel.org/r/20220504011345.662299-2-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-18-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Cc: Catalin Marinas Cc: David Howells Cc: "Matthew Wilcox (Oracle)" Cc: SeongJae Park Cc: Vlastimil Babka Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- arch/x86/kernel/tboot.c | 1 - drivers/firmware/efi/efi.c | 1 - include/linux/mm.h | 2 - include/linux/mm_types.h | 14 -- kernel/fork.c | 8 - mm/init-mm.c | 2 - mm/mmap.c | 509 ++++++++++----------------------------------- mm/nommu.c | 87 +++----- mm/util.c | 10 +- 9 files changed, 144 insertions(+), 490 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/tboot.c b/arch/x86/kernel/tboot.c index 71c54ad3868a..3b388330a106 100644 --- a/arch/x86/kernel/tboot.c +++ b/arch/x86/kernel/tboot.c @@ -96,7 +96,6 @@ void __init tboot_probe(void) static pgd_t *tboot_pg_dir; static struct mm_struct tboot_mm = { - .mm_rb = RB_ROOT, .mm_mt = MTREE_INIT_EXT(mm_mt, MM_MT_FLAGS, tboot_mm.mmap_lock), .pgd = swapper_pg_dir, .mm_users = ATOMIC_INIT(2), diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c index 1eddef189d68..07677fde00af 100644 --- a/drivers/firmware/efi/efi.c +++ b/drivers/firmware/efi/efi.c @@ -57,7 +57,6 @@ static unsigned long __initdata mem_reserve = EFI_INVALID_TABLE_ADDR; static unsigned long __initdata rt_prop = EFI_INVALID_TABLE_ADDR; struct mm_struct efi_mm = { - .mm_rb = RB_ROOT, .mm_mt = MTREE_INIT_EXT(mm_mt, MM_MT_FLAGS, efi_mm.mmap_lock), .mm_users = ATOMIC_INIT(2), .mm_count = ATOMIC_INIT(1), diff --git a/include/linux/mm.h b/include/linux/mm.h index 4753138e4c98..600361547e9b 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2667,8 +2667,6 @@ extern int __split_vma(struct mm_struct *, struct vm_area_struct *, extern int split_vma(struct mm_struct *, struct vm_area_struct *, unsigned long addr, int new_below); extern int insert_vm_struct(struct mm_struct *, struct vm_area_struct *); -extern void __vma_link_rb(struct mm_struct *, struct vm_area_struct *, - struct rb_node **, struct rb_node *); extern void unlink_file_vma(struct vm_area_struct *); extern struct vm_area_struct *copy_vma(struct vm_area_struct **, unsigned long addr, unsigned long len, pgoff_t pgoff, diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index a0e4b638d6b4..0e226d390fca 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -405,19 +405,6 @@ struct vm_area_struct { /* linked list of VM areas per task, sorted by address */ struct vm_area_struct *vm_next, *vm_prev; - - struct rb_node vm_rb; - - /* - * Largest free memory gap in bytes to the left of this VMA. - * Either between this VMA and vma->vm_prev, or between one of the - * VMAs below us in the VMA rbtree and its ->vm_prev. This helps - * get_unmapped_area find a free area of the right size. - */ - unsigned long rb_subtree_gap; - - /* Second cache line starts here. */ - struct mm_struct *vm_mm; /* The address space we belong to. */ /* @@ -483,7 +470,6 @@ struct mm_struct { struct { struct vm_area_struct *mmap; /* list of VMAs */ struct maple_tree mm_mt; - struct rb_root mm_rb; u64 vmacache_seqnum; /* per-thread vmacache */ #ifdef CONFIG_MMU unsigned long (*get_unmapped_area) (struct file *filp, diff --git a/kernel/fork.c b/kernel/fork.c index 4ecca79556eb..db420a1e93c7 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -581,7 +581,6 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) { struct vm_area_struct *mpnt, *tmp, *prev, **pprev; - struct rb_node **rb_link, *rb_parent; int retval; unsigned long charge = 0; MA_STATE(old_mas, &oldmm->mm_mt, 0, 0); @@ -608,8 +607,6 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, mm->exec_vm = oldmm->exec_vm; mm->stack_vm = oldmm->stack_vm; - rb_link = &mm->mm_rb.rb_node; - rb_parent = NULL; pprev = &mm->mmap; retval = ksm_fork(mm, oldmm); if (retval) @@ -701,10 +698,6 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, tmp->vm_prev = prev; prev = tmp; - __vma_link_rb(mm, tmp, rb_link, rb_parent); - rb_link = &tmp->vm_rb.rb_right; - rb_parent = &tmp->vm_rb; - /* Link the vma into the MT */ mas.index = tmp->vm_start; mas.last = tmp->vm_end - 1; @@ -1128,7 +1121,6 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, struct user_namespace *user_ns) { mm->mmap = NULL; - mm->mm_rb = RB_ROOT; mt_init_flags(&mm->mm_mt, MM_MT_FLAGS); mt_set_external_lock(&mm->mm_mt, &mm->mmap_lock); mm->vmacache_seqnum = 0; diff --git a/mm/init-mm.c b/mm/init-mm.c index b912b0f2eced..c9327abb771c 100644 --- a/mm/init-mm.c +++ b/mm/init-mm.c @@ -1,6 +1,5 @@ // SPDX-License-Identifier: GPL-2.0 #include -#include #include #include #include @@ -29,7 +28,6 @@ * and size this cpu_bitmask to NR_CPUS. */ struct mm_struct init_mm = { - .mm_rb = RB_ROOT, .mm_mt = MTREE_INIT_EXT(mm_mt, MM_MT_FLAGS, init_mm.mmap_lock), .pgd = swapper_pg_dir, .mm_users = ATOMIC_INIT(2), diff --git a/mm/mmap.c b/mm/mmap.c index 1235d7824ca4..61ee6eb01db0 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -39,7 +39,6 @@ #include #include #include -#include #include #include #include @@ -294,93 +293,6 @@ out: return origbrk; } -static inline unsigned long vma_compute_gap(struct vm_area_struct *vma) -{ - unsigned long gap, prev_end; - - /* - * Note: in the rare case of a VM_GROWSDOWN above a VM_GROWSUP, we - * allow two stack_guard_gaps between them here, and when choosing - * an unmapped area; whereas when expanding we only require one. - * That's a little inconsistent, but keeps the code here simpler. - */ - gap = vm_start_gap(vma); - if (vma->vm_prev) { - prev_end = vm_end_gap(vma->vm_prev); - if (gap > prev_end) - gap -= prev_end; - else - gap = 0; - } - return gap; -} - -#ifdef CONFIG_DEBUG_VM_RB -static unsigned long vma_compute_subtree_gap(struct vm_area_struct *vma) -{ - unsigned long max = vma_compute_gap(vma), subtree_gap; - if (vma->vm_rb.rb_left) { - subtree_gap = rb_entry(vma->vm_rb.rb_left, - struct vm_area_struct, vm_rb)->rb_subtree_gap; - if (subtree_gap > max) - max = subtree_gap; - } - if (vma->vm_rb.rb_right) { - subtree_gap = rb_entry(vma->vm_rb.rb_right, - struct vm_area_struct, vm_rb)->rb_subtree_gap; - if (subtree_gap > max) - max = subtree_gap; - } - return max; -} - -static int browse_rb(struct mm_struct *mm) -{ - struct rb_root *root = &mm->mm_rb; - int i = 0, j, bug = 0; - struct rb_node *nd, *pn = NULL; - unsigned long prev = 0, pend = 0; - - for (nd = rb_first(root); nd; nd = rb_next(nd)) { - struct vm_area_struct *vma; - vma = rb_entry(nd, struct vm_area_struct, vm_rb); - if (vma->vm_start < prev) { - pr_emerg("vm_start %lx < prev %lx\n", - vma->vm_start, prev); - bug = 1; - } - if (vma->vm_start < pend) { - pr_emerg("vm_start %lx < pend %lx\n", - vma->vm_start, pend); - bug = 1; - } - if (vma->vm_start > vma->vm_end) { - pr_emerg("vm_start %lx > vm_end %lx\n", - vma->vm_start, vma->vm_end); - bug = 1; - } - spin_lock(&mm->page_table_lock); - if (vma->rb_subtree_gap != vma_compute_subtree_gap(vma)) { - pr_emerg("free gap %lx, correct %lx\n", - vma->rb_subtree_gap, - vma_compute_subtree_gap(vma)); - bug = 1; - } - spin_unlock(&mm->page_table_lock); - i++; - pn = nd; - prev = vma->vm_start; - pend = vma->vm_end; - } - j = 0; - for (nd = pn; nd; nd = rb_prev(nd)) - j++; - if (i != j) { - pr_emerg("backwards %d, forwards %d\n", j, i); - bug = 1; - } - return bug ? -1 : i; -} #if defined(CONFIG_DEBUG_VM_MAPLE_TREE) extern void mt_validate(struct maple_tree *mt); extern void mt_dump(const struct maple_tree *mt); @@ -406,19 +318,25 @@ static void validate_mm_mt(struct mm_struct *mm) (vma->vm_end - 1 != mas.last)) { pr_emerg("issue in %s\n", current->comm); dump_stack(); -#ifdef CONFIG_DEBUG_VM dump_vma(vma_mt); - pr_emerg("and next in rb\n"); + pr_emerg("and vm_next\n"); dump_vma(vma->vm_next); -#endif pr_emerg("mt piv: %px %lu - %lu\n", vma_mt, mas.index, mas.last); pr_emerg("mt vma: %px %lu - %lu\n", vma_mt, vma_mt->vm_start, vma_mt->vm_end); - pr_emerg("rb vma: %px %lu - %lu\n", vma, + if (vma->vm_prev) { + pr_emerg("ll prev: %px %lu - %lu\n", + vma->vm_prev, vma->vm_prev->vm_start, + vma->vm_prev->vm_end); + } + pr_emerg("ll vma: %px %lu - %lu\n", vma, vma->vm_start, vma->vm_end); - pr_emerg("rb->next = %px %lu - %lu\n", vma->vm_next, - vma->vm_next->vm_start, vma->vm_next->vm_end); + if (vma->vm_next) { + pr_emerg("ll next: %px %lu - %lu\n", + vma->vm_next, vma->vm_next->vm_start, + vma->vm_next->vm_end); + } mt_dump(mas.tree); if (vma_mt->vm_end != mas.last + 1) { @@ -442,21 +360,6 @@ static void validate_mm_mt(struct mm_struct *mm) VM_BUG_ON(vma); mt_validate(&mm->mm_mt); } -#else -#define validate_mm_mt(root) do { } while (0) -#endif -static void validate_mm_rb(struct rb_root *root, struct vm_area_struct *ignore) -{ - struct rb_node *nd; - - for (nd = rb_first(root); nd; nd = rb_next(nd)) { - struct vm_area_struct *vma; - vma = rb_entry(nd, struct vm_area_struct, vm_rb); - VM_BUG_ON_VMA(vma != ignore && - vma->rb_subtree_gap != vma_compute_subtree_gap(vma), - vma); - } -} static void validate_mm(struct mm_struct *mm) { @@ -465,7 +368,10 @@ static void validate_mm(struct mm_struct *mm) unsigned long highest_address = 0; struct vm_area_struct *vma = mm->mmap; + validate_mm_mt(mm); + while (vma) { +#ifdef CONFIG_DEBUG_VM_RB struct anon_vma *anon_vma = vma->anon_vma; struct anon_vma_chain *avc; @@ -475,6 +381,7 @@ static void validate_mm(struct mm_struct *mm) anon_vma_interval_tree_verify(avc); anon_vma_unlock_read(anon_vma); } +#endif highest_address = vm_end_gap(vma); vma = vma->vm_next; @@ -489,80 +396,13 @@ static void validate_mm(struct mm_struct *mm) mm->highest_vm_end, highest_address); bug = 1; } - i = browse_rb(mm); - if (i != mm->map_count) { - if (i != -1) - pr_emerg("map_count %d rb %d\n", mm->map_count, i); - bug = 1; - } VM_BUG_ON_MM(bug, mm); } -#else -#define validate_mm_rb(root, ignore) do { } while (0) + +#else /* !CONFIG_DEBUG_VM_MAPLE_TREE */ #define validate_mm_mt(root) do { } while (0) #define validate_mm(mm) do { } while (0) -#endif - -RB_DECLARE_CALLBACKS_MAX(static, vma_gap_callbacks, - struct vm_area_struct, vm_rb, - unsigned long, rb_subtree_gap, vma_compute_gap) - -/* - * Update augmented rbtree rb_subtree_gap values after vma->vm_start or - * vma->vm_prev->vm_end values changed, without modifying the vma's position - * in the rbtree. - */ -static void vma_gap_update(struct vm_area_struct *vma) -{ - /* - * As it turns out, RB_DECLARE_CALLBACKS_MAX() already created - * a callback function that does exactly what we want. - */ - vma_gap_callbacks_propagate(&vma->vm_rb, NULL); -} - -static inline void vma_rb_insert(struct vm_area_struct *vma, - struct rb_root *root) -{ - /* All rb_subtree_gap values must be consistent prior to insertion */ - validate_mm_rb(root, NULL); - - rb_insert_augmented(&vma->vm_rb, root, &vma_gap_callbacks); -} - -static void __vma_rb_erase(struct vm_area_struct *vma, struct rb_root *root) -{ - /* - * Note rb_erase_augmented is a fairly large inline function, - * so make sure we instantiate it only once with our desired - * augmented rbtree callbacks. - */ - rb_erase_augmented(&vma->vm_rb, root, &vma_gap_callbacks); -} - -static __always_inline void vma_rb_erase_ignore(struct vm_area_struct *vma, - struct rb_root *root, - struct vm_area_struct *ignore) -{ - /* - * All rb_subtree_gap values must be consistent prior to erase, - * with the possible exception of - * - * a. the "next" vma being erased if next->vm_start was reduced in - * __vma_adjust() -> __vma_unlink() - * b. the vma being erased in detach_vmas_to_be_unmapped() -> - * vma_rb_erase() - */ - validate_mm_rb(root, ignore); - - __vma_rb_erase(vma, root); -} - -static __always_inline void vma_rb_erase(struct vm_area_struct *vma, - struct rb_root *root) -{ - vma_rb_erase_ignore(vma, root, vma); -} +#endif /* CONFIG_DEBUG_VM_MAPLE_TREE */ /* * vma has some anon_vma assigned, and is already inserted on that @@ -596,39 +436,26 @@ anon_vma_interval_tree_post_update_vma(struct vm_area_struct *vma) anon_vma_interval_tree_insert(avc, &avc->anon_vma->rb_root); } -static int find_vma_links(struct mm_struct *mm, unsigned long addr, - unsigned long end, struct vm_area_struct **pprev, - struct rb_node ***rb_link, struct rb_node **rb_parent) +/* + * range_has_overlap() - Check the @start - @end range for overlapping VMAs and + * sets up a pointer to the previous VMA + * @mm: the mm struct + * @start: the start address of the range + * @end: the end address of the range + * @pprev: the pointer to the pointer of the previous VMA + * + * Returns: True if there is an overlapping VMA, false otherwise + */ +static inline +bool range_has_overlap(struct mm_struct *mm, unsigned long start, + unsigned long end, struct vm_area_struct **pprev) { - struct rb_node **__rb_link, *__rb_parent, *rb_prev; - - mmap_assert_locked(mm); - __rb_link = &mm->mm_rb.rb_node; - rb_prev = __rb_parent = NULL; - - while (*__rb_link) { - struct vm_area_struct *vma_tmp; + struct vm_area_struct *existing; - __rb_parent = *__rb_link; - vma_tmp = rb_entry(__rb_parent, struct vm_area_struct, vm_rb); - - if (vma_tmp->vm_end > addr) { - /* Fail if an existing vma overlaps the area */ - if (vma_tmp->vm_start < end) - return -ENOMEM; - __rb_link = &__rb_parent->rb_left; - } else { - rb_prev = __rb_parent; - __rb_link = &__rb_parent->rb_right; - } - } - - *pprev = NULL; - if (rb_prev) - *pprev = rb_entry(rb_prev, struct vm_area_struct, vm_rb); - *rb_link = __rb_link; - *rb_parent = __rb_parent; - return 0; + MA_STATE(mas, &mm->mm_mt, start, start); + existing = mas_find(&mas, end - 1); + *pprev = mas_prev(&mas, 0); + return existing ? true : false; } /* @@ -655,8 +482,6 @@ static inline struct vm_area_struct *__vma_next(struct mm_struct *mm, * @start: The start of the range. * @len: The length of the range. * @pprev: pointer to the pointer that will be set to previous vm_area_struct - * @rb_link: the rb_node - * @rb_parent: the parent rb_node * * Find all the vm_area_struct that overlap from @start to * @end and munmap them. Set @pprev to the previous vm_area_struct. @@ -665,14 +490,11 @@ static inline struct vm_area_struct *__vma_next(struct mm_struct *mm, */ static inline int munmap_vma_range(struct mm_struct *mm, unsigned long start, unsigned long len, - struct vm_area_struct **pprev, struct rb_node ***link, - struct rb_node **parent, struct list_head *uf) + struct vm_area_struct **pprev, struct list_head *uf) { - - while (find_vma_links(mm, start, start + len, pprev, link, parent)) + while (range_has_overlap(mm, start, start + len, pprev)) if (do_munmap(mm, start, len, uf)) return -ENOMEM; - return 0; } @@ -693,30 +515,6 @@ static unsigned long count_vma_pages_range(struct mm_struct *mm, return nr_pages; } -void __vma_link_rb(struct mm_struct *mm, struct vm_area_struct *vma, - struct rb_node **rb_link, struct rb_node *rb_parent) -{ - /* Update tracking information for the gap following the new vma. */ - if (vma->vm_next) - vma_gap_update(vma->vm_next); - else - mm->highest_vm_end = vm_end_gap(vma); - - /* - * vma->vm_prev wasn't known when we followed the rbtree to find the - * correct insertion point for that vma. As a result, we could not - * update the vma vm_rb parents rb_subtree_gap values on the way down. - * So, we first insert the vma with a zero rb_subtree_gap value - * (to be consistent with what we did on the way down), and then - * immediately update the gap to the correct value. Finally we - * rebalance the rbtree after all augmented values have been set. - */ - rb_link_node(&vma->vm_rb, rb_parent, rb_link); - vma->rb_subtree_gap = 0; - vma_gap_update(vma); - vma_rb_insert(vma, &mm->mm_rb); -} - static void __vma_link_file(struct vm_area_struct *vma) { struct file *file; @@ -784,18 +582,8 @@ static inline void vma_mas_szero(struct ma_state *mas, unsigned long start, mas_store_prealloc(mas, NULL); } -static void -__vma_link(struct mm_struct *mm, struct vm_area_struct *vma, - struct vm_area_struct *prev, struct rb_node **rb_link, - struct rb_node *rb_parent) -{ - __vma_link_list(mm, vma, prev); - __vma_link_rb(mm, vma, rb_link, rb_parent); -} - static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma, - struct vm_area_struct *prev, struct rb_node **rb_link, - struct rb_node *rb_parent) + struct vm_area_struct *prev) { MA_STATE(mas, &mm->mm_mt, 0, 0); struct address_space *mapping = NULL; @@ -809,7 +597,7 @@ static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma, } vma_mas_store(vma, &mas); - __vma_link(mm, vma, prev, rb_link, rb_parent); + __vma_link_list(mm, vma, prev); __vma_link_file(vma); if (mapping) @@ -822,34 +610,20 @@ static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma, /* * Helper for vma_adjust() in the split_vma insert case: insert a vma into the - * mm's list and rbtree. It has already been inserted into the interval tree. + * mm's list and the mm tree. It has already been inserted into the interval tree. */ static void __insert_vm_struct(struct mm_struct *mm, struct ma_state *mas, struct vm_area_struct *vma) { struct vm_area_struct *prev; - struct rb_node **rb_link, *rb_parent; - - if (find_vma_links(mm, vma->vm_start, vma->vm_end, - &prev, &rb_link, &rb_parent)) - BUG(); + mas_set(mas, vma->vm_start); + prev = mas_prev(mas, 0); vma_mas_store(vma, mas); __vma_link_list(mm, vma, prev); - __vma_link_rb(mm, vma, rb_link, rb_parent); mm->map_count++; } -static __always_inline void __vma_unlink(struct mm_struct *mm, - struct vm_area_struct *vma, - struct vm_area_struct *ignore) -{ - vma_rb_erase_ignore(vma, &mm->mm_rb, ignore); - __vma_unlink_list(mm, vma); - /* Kill the cache */ - vmacache_invalidate(mm); -} - /* * We cannot adjust vm_start, vm_end, vm_pgoff fields of a vma that * is already present in an i_mmap tree without adjusting the tree. @@ -862,20 +636,18 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, struct vm_area_struct *expand) { struct mm_struct *mm = vma->vm_mm; - struct vm_area_struct *next = vma->vm_next, *orig_vma = vma; + struct vm_area_struct *next_next, *next = find_vma(mm, vma->vm_end); + struct vm_area_struct *orig_vma = vma; struct address_space *mapping = NULL; struct rb_root_cached *root = NULL; struct anon_vma *anon_vma = NULL; struct file *file = vma->vm_file; - bool start_changed = false, end_changed = false; + bool vma_changed = false; long adjust_next = 0; int remove_next = 0; MA_STATE(mas, &mm->mm_mt, 0, 0); struct vm_area_struct *exporter = NULL, *importer = NULL; - validate_mm(mm); - validate_mm_mt(mm); - if (next && !insert) { if (end >= next->vm_end) { /* @@ -905,8 +677,9 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, * remove_next == 1 is case 1 or 7. */ remove_next = 1 + (end > next->vm_end); + next_next = find_vma(mm, next->vm_end); VM_WARN_ON(remove_next == 2 && - end != next->vm_next->vm_end); + end != next_next->vm_end); /* trim end to next, for case 6 first pass */ end = next->vm_end; } @@ -1005,21 +778,21 @@ again: } if (start != vma->vm_start) { - unsigned long old_start = vma->vm_start; + if (vma->vm_start < start) + vma_mas_szero(&mas, vma->vm_start, start); + vma_changed = true; vma->vm_start = start; - if (old_start < start) - vma_mas_szero(&mas, old_start, start); - start_changed = true; } if (end != vma->vm_end) { - unsigned long old_end = vma->vm_end; + if (vma->vm_end > end) + vma_mas_szero(&mas, end, vma->vm_end); + vma_changed = true; vma->vm_end = end; - if (old_end > end) - vma_mas_szero(&mas, end, old_end); - end_changed = true; + if (!next) + mm->highest_vm_end = vm_end_gap(vma); } - if (end_changed || start_changed) + if (vma_changed) vma_mas_store(vma, &mas); vma->vm_pgoff = pgoff; @@ -1037,25 +810,9 @@ again: } if (remove_next) { - /* - * vma_merge has merged next into vma, and needs - * us to remove next before dropping the locks. - * Since we have expanded over this vma, the maple tree will - * have overwritten by storing the value - */ - if (remove_next != 3) - __vma_unlink(mm, next, next); - else - /* - * vma is not before next if they've been - * swapped. - * - * pre-swap() next->vm_start was reduced so - * tell validate_mm_rb to ignore pre-swap() - * "next" (which is stored in post-swap() - * "vma"). - */ - __vma_unlink(mm, next, vma); + __vma_unlink_list(mm, next); + /* Kill the cache */ + vmacache_invalidate(mm); if (file) __remove_shared_vm_struct(next, file, mapping); } else if (insert) { @@ -1065,15 +822,6 @@ again: * (it may either follow vma or precede it). */ __insert_vm_struct(mm, &mas, insert); - } else { - if (start_changed) - vma_gap_update(vma); - if (end_changed) { - if (!next) - mm->highest_vm_end = vm_end_gap(vma); - else if (!adjust_next) - vma_gap_update(next); - } } if (anon_vma) { @@ -1100,7 +848,9 @@ again: anon_vma_merge(vma, next); mm->map_count--; mpol_put(vma_policy(next)); + BUG_ON(vma->vm_end < next->vm_end); vm_area_free(next); + /* * In mprotect's case 6 (see comments on vma_merge), * we must remove another next too. It would clutter @@ -1113,7 +863,7 @@ again: * "next->vm_prev->vm_end" changed and the * "vma->vm_next" gap must be updated. */ - next = vma->vm_next; + next = next_next; } else { /* * For the scope of the comment "next" and @@ -1128,13 +878,11 @@ again: next = vma; } if (remove_next == 2) { + mas_reset(&mas); remove_next = 1; end = next->vm_end; goto again; - } - else if (next) - vma_gap_update(next); - else { + } else if (!next) { /* * If remove_next == 2 we obviously can't * reach this path. @@ -1161,8 +909,6 @@ again: uprobe_mmap(insert); validate_mm(mm); - validate_mm_mt(mm); - return 0; } @@ -1315,7 +1061,6 @@ struct vm_area_struct *vma_merge(struct mm_struct *mm, struct vm_area_struct *area, *next; int err; - validate_mm_mt(mm); /* * We later require that vma->vm_flags == vm_flags, * so this tests vma->vm_flags & VM_SPECIAL, too. @@ -1391,7 +1136,6 @@ struct vm_area_struct *vma_merge(struct mm_struct *mm, khugepaged_enter_vma(area, vm_flags); return area; } - validate_mm_mt(mm); return NULL; } @@ -1561,6 +1305,7 @@ unsigned long do_mmap(struct file *file, unsigned long addr, vm_flags_t vm_flags; int pkey = 0; + validate_mm(mm); *populate = 0; if (!len) @@ -1868,10 +1613,8 @@ unsigned long mmap_region(struct file *file, unsigned long addr, struct mm_struct *mm = current->mm; struct vm_area_struct *vma, *prev, *merge; int error; - struct rb_node **rb_link, *rb_parent; unsigned long charged = 0; - validate_mm_mt(mm); /* Check against address space limit. */ if (!may_expand_vm(mm, vm_flags, len >> PAGE_SHIFT)) { unsigned long nr_pages; @@ -1887,8 +1630,8 @@ unsigned long mmap_region(struct file *file, unsigned long addr, return -ENOMEM; } - /* Clear old maps, set up prev, rb_link, rb_parent, and uf */ - if (munmap_vma_range(mm, addr, len, &prev, &rb_link, &rb_parent, uf)) + /* Clear old maps, set up prev and uf */ + if (munmap_vma_range(mm, addr, len, &prev, uf)) return -ENOMEM; /* * Private writable mapping: check memory availability @@ -1986,7 +1729,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr, goto free_vma; } - if (vma_link(mm, vma, prev, rb_link, rb_parent)) { + if (vma_link(mm, vma, prev)) { error = -ENOMEM; if (file) goto unmap_and_free_vma; @@ -2032,7 +1775,6 @@ out: vma_set_page_prot(vma); - validate_mm_mt(mm); return addr; unmap_and_free_vma: @@ -2049,7 +1791,6 @@ free_vma: unacct_error: if (charged) vm_unacct_memory(charged); - validate_mm_mt(mm); return error; } @@ -2402,7 +2143,6 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) int error = 0; MA_STATE(mas, &mm->mm_mt, 0, 0); - validate_mm_mt(mm); if (!(vma->vm_flags & VM_GROWSUP)) return -EFAULT; @@ -2454,15 +2194,13 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) error = acct_stack_growth(vma, size, grow); if (!error) { /* - * vma_gap_update() doesn't support concurrent - * updates, but we only hold a shared mmap_lock - * lock here, so we need to protect against - * concurrent vma expansions. - * anon_vma_lock_write() doesn't help here, as - * we don't guarantee that all growable vmas - * in a mm share the same root anon vma. - * So, we reuse mm->page_table_lock to guard - * against concurrent vma expansions. + * We only hold a shared mmap_lock lock here, so + * we need to protect against concurrent vma + * expansions. anon_vma_lock_write() doesn't + * help here, as we don't guarantee that all + * growable vmas in a mm share the same root + * anon vma. So, we reuse mm->page_table_lock + * to guard against concurrent vma expansions. */ spin_lock(&mm->page_table_lock); if (vma->vm_flags & VM_LOCKED) @@ -2473,9 +2211,7 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) /* Overwrite old entry in mtree. */ vma_mas_store(vma, &mas); anon_vma_interval_tree_post_update_vma(vma); - if (vma->vm_next) - vma_gap_update(vma->vm_next); - else + if (!vma->vm_next) mm->highest_vm_end = vm_end_gap(vma); spin_unlock(&mm->page_table_lock); @@ -2485,8 +2221,6 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) } anon_vma_unlock_write(vma->anon_vma); khugepaged_enter_vma(vma, vma->vm_flags); - validate_mm(mm); - validate_mm_mt(mm); mas_destroy(&mas); return error; } @@ -2495,15 +2229,13 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) /* * vma is the first one with address < vma->vm_start. Have to extend vma. */ -int expand_downwards(struct vm_area_struct *vma, - unsigned long address) +int expand_downwards(struct vm_area_struct *vma, unsigned long address) { struct mm_struct *mm = vma->vm_mm; struct vm_area_struct *prev; int error = 0; MA_STATE(mas, &mm->mm_mt, 0, 0); - validate_mm(mm); address &= PAGE_MASK; if (address < mmap_min_addr) return -EPERM; @@ -2545,15 +2277,13 @@ int expand_downwards(struct vm_area_struct *vma, error = acct_stack_growth(vma, size, grow); if (!error) { /* - * vma_gap_update() doesn't support concurrent - * updates, but we only hold a shared mmap_lock - * lock here, so we need to protect against - * concurrent vma expansions. - * anon_vma_lock_write() doesn't help here, as - * we don't guarantee that all growable vmas - * in a mm share the same root anon vma. - * So, we reuse mm->page_table_lock to guard - * against concurrent vma expansions. + * We only hold a shared mmap_lock lock here, so + * we need to protect against concurrent vma + * expansions. anon_vma_lock_write() doesn't + * help here, as we don't guarantee that all + * growable vmas in a mm share the same root + * anon vma. So, we reuse mm->page_table_lock + * to guard against concurrent vma expansions. */ spin_lock(&mm->page_table_lock); if (vma->vm_flags & VM_LOCKED) @@ -2565,7 +2295,6 @@ int expand_downwards(struct vm_area_struct *vma, /* Overwrite old entry in mtree. */ vma_mas_store(vma, &mas); anon_vma_interval_tree_post_update_vma(vma); - vma_gap_update(vma); spin_unlock(&mm->page_table_lock); perf_event_mmap(vma); @@ -2574,7 +2303,6 @@ int expand_downwards(struct vm_area_struct *vma, } anon_vma_unlock_write(vma->anon_vma); khugepaged_enter_vma(vma, vma->vm_flags); - validate_mm(mm); mas_destroy(&mas); return error; } @@ -2707,10 +2435,8 @@ detach_vmas_to_be_unmapped(struct mm_struct *mm, struct ma_state *mas, insertion_point = (prev ? &prev->vm_next : &mm->mmap); vma->vm_prev = NULL; - mas_set_range(mas, vma->vm_start, end - 1); - mas_store_prealloc(mas, NULL); + vma_mas_szero(mas, vma->vm_start, end); do { - vma_rb_erase(vma, &mm->mm_rb); if (vma->vm_flags & VM_LOCKED) mm->locked_vm -= vma_pages(vma); mm->map_count--; @@ -2718,10 +2444,9 @@ detach_vmas_to_be_unmapped(struct mm_struct *mm, struct ma_state *mas, vma = vma->vm_next; } while (vma && vma->vm_start < end); *insertion_point = vma; - if (vma) { + if (vma) vma->vm_prev = prev; - vma_gap_update(vma); - } else + else mm->highest_vm_end = prev ? vm_end_gap(prev) : 0; tail_vma->vm_next = NULL; @@ -2840,11 +2565,7 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, if (len == 0) return -EINVAL; - /* - * arch_unmap() might do unmaps itself. It must be called - * and finish any rbtree manipulation before this code - * runs and also starts to manipulate the rbtree. - */ + /* arch_unmap() might do unmaps itself. */ arch_unmap(mm, start, end); /* Find the first overlapping VMA where start < vma->vm_end */ @@ -2855,6 +2576,11 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, if (mas_preallocate(&mas, vma, GFP_KERNEL)) return -ENOMEM; prev = vma->vm_prev; + /* we have start < vma->vm_end */ + + /* if it doesn't overlap, we have nothing.. */ + if (vma->vm_start >= end) + return 0; /* * If we need to split any vma, do it now to save pain later. @@ -2915,6 +2641,8 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, /* Fix up all other VM information */ remove_vma_list(mm, vma); + + validate_mm(mm); return downgrade ? 1 : 0; map_count_exceeded: @@ -3053,11 +2781,11 @@ out: * anonymous maps. eventually we may be able to do some * brk-specific accounting here. */ -static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long flags, struct list_head *uf) +static int do_brk_flags(unsigned long addr, unsigned long len, + unsigned long flags, struct list_head *uf) { struct mm_struct *mm = current->mm; struct vm_area_struct *vma, *prev; - struct rb_node **rb_link, *rb_parent; pgoff_t pgoff = addr >> PAGE_SHIFT; int error; unsigned long mapped_addr; @@ -3076,8 +2804,8 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla if (error) return error; - /* Clear old maps, set up prev, rb_link, rb_parent, and uf */ - if (munmap_vma_range(mm, addr, len, &prev, &rb_link, &rb_parent, uf)) + /* Clear old maps, set up prev and uf */ + if (munmap_vma_range(mm, addr, len, &prev, uf)) return -ENOMEM; /* Check against address space limits *after* clearing old maps... */ @@ -3111,7 +2839,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla vma->vm_pgoff = pgoff; vma->vm_flags = flags; vma->vm_page_prot = vm_get_page_prot(flags); - if(vma_link(mm, vma, prev, rb_link, rb_parent)) + if(vma_link(mm, vma, prev)) goto no_vma_link; out: @@ -3230,26 +2958,10 @@ void exit_mmap(struct mm_struct *mm) int insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma) { struct vm_area_struct *prev; - struct rb_node **rb_link, *rb_parent; - unsigned long start = vma->vm_start; - struct vm_area_struct *overlap = NULL; - if (find_vma_links(mm, vma->vm_start, vma->vm_end, - &prev, &rb_link, &rb_parent)) + if (range_has_overlap(mm, vma->vm_start, vma->vm_end, &prev)) return -ENOMEM; - overlap = mt_find(&mm->mm_mt, &start, vma->vm_end - 1); - if (overlap) { - - pr_err("Found vma ending at %lu\n", start - 1); - pr_err("vma : %lu => %lu-%lu\n", (unsigned long)overlap, - overlap->vm_start, overlap->vm_end - 1); -#if defined(CONFIG_DEBUG_VM_MAPLE_TREE) - mt_dump(&mm->mm_mt); -#endif - BUG(); - } - if ((vma->vm_flags & VM_ACCOUNT) && security_vm_enough_memory_mm(mm, vma_pages(vma))) return -ENOMEM; @@ -3271,7 +2983,7 @@ int insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma) vma->vm_pgoff = vma->vm_start >> PAGE_SHIFT; } - if (vma_link(mm, vma, prev, rb_link, rb_parent)) + if (vma_link(mm, vma, prev)) return -ENOMEM; return 0; @@ -3289,9 +3001,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap, unsigned long vma_start = vma->vm_start; struct mm_struct *mm = vma->vm_mm; struct vm_area_struct *new_vma, *prev; - struct rb_node **rb_link, *rb_parent; bool faulted_in_anon_vma = true; - unsigned long index = addr; validate_mm_mt(mm); /* @@ -3303,10 +3013,9 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap, faulted_in_anon_vma = false; } - if (find_vma_links(mm, addr, addr + len, &prev, &rb_link, &rb_parent)) + if (range_has_overlap(mm, addr, addr + len, &prev)) return NULL; /* should never get here */ - if (mt_find(&mm->mm_mt, &index, addr+len - 1)) - BUG(); + new_vma = vma_merge(mm, prev, addr, addr + len, vma->vm_flags, vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma), vma->vm_userfaultfd_ctx, anon_vma_name(vma)); @@ -3347,12 +3056,16 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap, get_file(new_vma->vm_file); if (new_vma->vm_ops && new_vma->vm_ops->open) new_vma->vm_ops->open(new_vma); - vma_link(mm, new_vma, prev, rb_link, rb_parent); + if (vma_link(mm, new_vma, prev)) + goto out_vma_link; *need_rmap_locks = false; } validate_mm_mt(mm); return new_vma; +out_vma_link: + if (new_vma->vm_ops && new_vma->vm_ops->close) + new_vma->vm_ops->close(new_vma); out_free_mempol: mpol_put(vma_policy(new_vma)); out_free_vma: diff --git a/mm/nommu.c b/mm/nommu.c index 5af0b050eba8..f2031f865dbb 100644 --- a/mm/nommu.c +++ b/mm/nommu.c @@ -566,9 +566,9 @@ void vma_mas_remove(struct vm_area_struct *vma, struct ma_state *mas) */ static void add_vma_to_mm(struct mm_struct *mm, struct vm_area_struct *vma) { - struct vm_area_struct *pvma, *prev; struct address_space *mapping; - struct rb_node **p, *parent, *rb_prev; + struct vm_area_struct *prev; + MA_STATE(mas, &mm->mm_mt, vma->vm_start, vma->vm_end); BUG_ON(!vma->vm_region); @@ -586,42 +586,10 @@ static void add_vma_to_mm(struct mm_struct *mm, struct vm_area_struct *vma) i_mmap_unlock_write(mapping); } + prev = mas_prev(&mas, 0); + mas_reset(&mas); /* add the VMA to the tree */ - parent = rb_prev = NULL; - p = &mm->mm_rb.rb_node; - while (*p) { - parent = *p; - pvma = rb_entry(parent, struct vm_area_struct, vm_rb); - - /* sort by: start addr, end addr, VMA struct addr in that order - * (the latter is necessary as we may get identical VMAs) */ - if (vma->vm_start < pvma->vm_start) - p = &(*p)->rb_left; - else if (vma->vm_start > pvma->vm_start) { - rb_prev = parent; - p = &(*p)->rb_right; - } else if (vma->vm_end < pvma->vm_end) - p = &(*p)->rb_left; - else if (vma->vm_end > pvma->vm_end) { - rb_prev = parent; - p = &(*p)->rb_right; - } else if (vma < pvma) - p = &(*p)->rb_left; - else if (vma > pvma) { - rb_prev = parent; - p = &(*p)->rb_right; - } else - BUG(); - } - - rb_link_node(&vma->vm_rb, parent, p); - rb_insert_color(&vma->vm_rb, &mm->mm_rb); - - /* add VMA to the VMA list also */ - prev = NULL; - if (rb_prev) - prev = rb_entry(rb_prev, struct vm_area_struct, vm_rb); - + vma_mas_store(vma, &mas); __vma_link_list(mm, vma, prev); } @@ -634,6 +602,7 @@ static void delete_vma_from_mm(struct vm_area_struct *vma) struct address_space *mapping; struct mm_struct *mm = vma->vm_mm; struct task_struct *curr = current; + MA_STATE(mas, &vma->vm_mm->mm_mt, 0, 0); mm->map_count--; for (i = 0; i < VMACACHE_SIZE; i++) { @@ -656,8 +625,7 @@ static void delete_vma_from_mm(struct vm_area_struct *vma) } /* remove from the MM's tree and list */ - rb_erase(&vma->vm_rb, &mm->mm_rb); - + vma_mas_remove(vma, &mas); __vma_unlink_list(mm, vma); } @@ -681,24 +649,19 @@ static void delete_vma(struct mm_struct *mm, struct vm_area_struct *vma) struct vm_area_struct *find_vma(struct mm_struct *mm, unsigned long addr) { struct vm_area_struct *vma; + MA_STATE(mas, &mm->mm_mt, addr, addr); /* check the cache first */ vma = vmacache_find(mm, addr); if (likely(vma)) return vma; - /* trawl the list (there may be multiple mappings in which addr - * resides) */ - for (vma = mm->mmap; vma; vma = vma->vm_next) { - if (vma->vm_start > addr) - return NULL; - if (vma->vm_end > addr) { - vmacache_update(addr, vma); - return vma; - } - } + vma = mas_walk(&mas); - return NULL; + if (vma) + vmacache_update(addr, vma); + + return vma; } EXPORT_SYMBOL(find_vma); @@ -730,26 +693,23 @@ static struct vm_area_struct *find_vma_exact(struct mm_struct *mm, { struct vm_area_struct *vma; unsigned long end = addr + len; + MA_STATE(mas, &mm->mm_mt, addr, addr); /* check the cache first */ vma = vmacache_find_exact(mm, addr, end); if (vma) return vma; - /* trawl the list (there may be multiple mappings in which addr - * resides) */ - for (vma = mm->mmap; vma; vma = vma->vm_next) { - if (vma->vm_start < addr) - continue; - if (vma->vm_start > addr) - return NULL; - if (vma->vm_end == end) { - vmacache_update(addr, vma); - return vma; - } - } + vma = mas_walk(&mas); + if (!vma) + return NULL; + if (vma->vm_start != addr) + return NULL; + if (vma->vm_end != end) + return NULL; - return NULL; + vmacache_update(addr, vma); + return vma; } /* @@ -1546,6 +1506,7 @@ void exit_mmap(struct mm_struct *mm) delete_vma(mm, vma); cond_resched(); } + __mt_destroy(&mm->mm_mt); } int vm_brk(unsigned long addr, unsigned long len) diff --git a/mm/util.c b/mm/util.c index 5df8f2db7ca9..812365cbdd19 100644 --- a/mm/util.c +++ b/mm/util.c @@ -288,6 +288,8 @@ void __vma_link_list(struct mm_struct *mm, struct vm_area_struct *vma, vma->vm_next = next; if (next) next->vm_prev = vma; + else + mm->highest_vm_end = vm_end_gap(vma); } void __vma_unlink_list(struct mm_struct *mm, struct vm_area_struct *vma) @@ -300,8 +302,14 @@ void __vma_unlink_list(struct mm_struct *mm, struct vm_area_struct *vma) prev->vm_next = next; else mm->mmap = next; - if (next) + if (next) { next->vm_prev = prev; + } else { + if (prev) + mm->highest_vm_end = vm_end_gap(prev); + else + mm->highest_vm_end = 0; + } } /* Check if the vma is being used as a stack by this task */ -- cgit v1.2.3 From a116a414b69410fd58c2b0b95c2a29236f96704b Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:46:57 +0000 Subject: mm: optimize find_exact_vma() to use vma_lookup() Use vma_lookup() to walk the tree to the start value requested. If the vma at the start does not match, then the answer is NULL and there is no need to look at the next vma the way that find_vma() would. Link: https://lkml.kernel.org/r/20220504011345.662299-5-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-21-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Reviewed-by: Vlastimil Babka Cc: Catalin Marinas Cc: David Howells Cc: "Matthew Wilcox (Oracle)" Cc: SeongJae Park Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- include/linux/mm.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 600361547e9b..1163c98e877e 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2863,7 +2863,7 @@ static inline unsigned long vma_pages(struct vm_area_struct *vma) static inline struct vm_area_struct *find_exact_vma(struct mm_struct *mm, unsigned long vm_start, unsigned long vm_end) { - struct vm_area_struct *vma = find_vma(mm, vm_start); + struct vm_area_struct *vma = vma_lookup(mm, vm_start); if (vma && (vma->vm_start != vm_start || vma->vm_end != vm_end)) vma = NULL; -- cgit v1.2.3 From 650d89b18eb6436e6977e7ab80d0e0b7e832dc87 Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:46:58 +0000 Subject: mm: use maple tree operations for find_vma_intersection() Move find_vma_intersection() to mmap.c and change implementation to maple tree. When searching for a vma within a range, it is easier to use the maple tree interface. Exported find_vma_intersection() for kvm module. Link: https://lkml.kernel.org/r/20220504011345.662299-8-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-24-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Cc: Catalin Marinas Cc: David Howells Cc: "Matthew Wilcox (Oracle)" Cc: SeongJae Park Cc: Vlastimil Babka Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- include/linux/mm.h | 22 ++++------------------ mm/mmap.c | 29 +++++++++++++++++++++++++++++ mm/nommu.c | 11 +++++++++++ 3 files changed, 44 insertions(+), 18 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 1163c98e877e..f3c19292ea16 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2791,26 +2791,12 @@ extern struct vm_area_struct * find_vma(struct mm_struct * mm, unsigned long add extern struct vm_area_struct * find_vma_prev(struct mm_struct * mm, unsigned long addr, struct vm_area_struct **pprev); -/** - * find_vma_intersection() - Look up the first VMA which intersects the interval - * @mm: The process address space. - * @start_addr: The inclusive start user address. - * @end_addr: The exclusive end user address. - * - * Returns: The first VMA within the provided range, %NULL otherwise. Assumes - * start_addr < end_addr. +/* + * Look up the first VMA which intersects the interval [start_addr, end_addr) + * NULL if none. Assume start_addr < end_addr. */ -static inline struct vm_area_struct *find_vma_intersection(struct mm_struct *mm, - unsigned long start_addr, - unsigned long end_addr) -{ - struct vm_area_struct *vma = find_vma(mm, start_addr); - - if (vma && end_addr <= vma->vm_start) - vma = NULL; - return vma; -} + unsigned long start_addr, unsigned long end_addr); /** * vma_lookup() - Find a VMA at a specific address diff --git a/mm/mmap.c b/mm/mmap.c index 64d2dd8e1a9d..4b42d0028b39 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -2092,6 +2092,35 @@ get_unmapped_area(struct file *file, unsigned long addr, unsigned long len, EXPORT_SYMBOL(get_unmapped_area); +/** + * find_vma_intersection() - Look up the first VMA which intersects the interval + * @mm: The process address space. + * @start_addr: The inclusive start user address. + * @end_addr: The exclusive end user address. + * + * Returns: The first VMA within the provided range, %NULL otherwise. Assumes + * start_addr < end_addr. + */ +struct vm_area_struct *find_vma_intersection(struct mm_struct *mm, + unsigned long start_addr, + unsigned long end_addr) +{ + struct vm_area_struct *vma; + unsigned long index = start_addr; + + mmap_assert_locked(mm); + /* Check the cache first. */ + vma = vmacache_find(mm, start_addr); + if (likely(vma)) + return vma; + + vma = mt_find(&mm->mm_mt, &index, end_addr - 1); + if (vma) + vmacache_update(start_addr, vma); + return vma; +} +EXPORT_SYMBOL(find_vma_intersection); + /** * find_vma() - Find the VMA for a given address, or the next vma. * @mm: The mm_struct to check diff --git a/mm/nommu.c b/mm/nommu.c index f2031f865dbb..b098c0251137 100644 --- a/mm/nommu.c +++ b/mm/nommu.c @@ -642,6 +642,17 @@ static void delete_vma(struct mm_struct *mm, struct vm_area_struct *vma) vm_area_free(vma); } +struct vm_area_struct *find_vma_intersection(struct mm_struct *mm, + unsigned long start_addr, + unsigned long end_addr) +{ + unsigned long index = start_addr; + + mmap_assert_locked(mm); + return mt_find(&mm->mm_mt, &index, end_addr - 1); +} +EXPORT_SYMBOL(find_vma_intersection); + /* * look up the first VMA in which addr resides, NULL if none * - should be called with mm->mmap_lock at least held readlocked -- cgit v1.2.3 From 5f6e0cb1cb8e32620ae22bff9bbf39bfdd8ab268 Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:46:59 +0000 Subject: mm: remove vmacache By using the maple tree and the maple tree state, the vmacache is no longer beneficial and is complicating the VMA code. Remove the vmacache to reduce the work in keeping it up to date and code complexity. Link: https://lkml.kernel.org/r/20220504011345.662299-10-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-26-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Acked-by: Vlastimil Babka Cc: Catalin Marinas Cc: David Howells Cc: "Matthew Wilcox (Oracle)" Cc: SeongJae Park Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- fs/exec.c | 3 -- fs/proc/task_mmu.c | 1 - include/linux/mm_types.h | 1 - include/linux/mm_types_task.h | 12 ----- include/linux/sched.h | 1 - include/linux/vm_event_item.h | 4 -- include/linux/vmacache.h | 28 ---------- include/linux/vmstat.h | 6 --- kernel/debug/debug_core.c | 12 ----- kernel/fork.c | 5 -- lib/Kconfig.debug | 8 --- mm/Makefile | 2 +- mm/debug.c | 4 +- mm/mmap.c | 32 +----------- mm/nommu.c | 37 ++----------- mm/vmacache.c | 117 ------------------------------------------ mm/vmstat.c | 4 -- 17 files changed, 9 insertions(+), 268 deletions(-) delete mode 100644 include/linux/vmacache.h delete mode 100644 mm/vmacache.c (limited to 'include/linux') diff --git a/fs/exec.c b/fs/exec.c index 0989fb8472a1..b97afa682ffe 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -28,7 +28,6 @@ #include #include #include -#include #include #include #include @@ -1027,8 +1026,6 @@ static int exec_mmap(struct mm_struct *mm) activate_mm(active_mm, mm); if (IS_ENABLED(CONFIG_ARCH_WANT_IRQS_OFF_ACTIVATE_MM)) local_irq_enable(); - tsk->mm->vmacache_seqnum = 0; - vmacache_flush(tsk); task_unlock(tsk); if (old_mm) { mmap_read_unlock(old_mm); diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 2d04e3470d4c..616913bb1fa7 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -1,6 +1,5 @@ // SPDX-License-Identifier: GPL-2.0 #include -#include #include #include #include diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 0e226d390fca..040bfc207cb1 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -470,7 +470,6 @@ struct mm_struct { struct { struct vm_area_struct *mmap; /* list of VMAs */ struct maple_tree mm_mt; - u64 vmacache_seqnum; /* per-thread vmacache */ #ifdef CONFIG_MMU unsigned long (*get_unmapped_area) (struct file *filp, unsigned long addr, unsigned long len, diff --git a/include/linux/mm_types_task.h b/include/linux/mm_types_task.h index c1bc6731125c..0bb4b6da9993 100644 --- a/include/linux/mm_types_task.h +++ b/include/linux/mm_types_task.h @@ -24,18 +24,6 @@ IS_ENABLED(CONFIG_ARCH_ENABLE_SPLIT_PMD_PTLOCK)) #define ALLOC_SPLIT_PTLOCKS (SPINLOCK_SIZE > BITS_PER_LONG/8) -/* - * The per task VMA cache array: - */ -#define VMACACHE_BITS 2 -#define VMACACHE_SIZE (1U << VMACACHE_BITS) -#define VMACACHE_MASK (VMACACHE_SIZE - 1) - -struct vmacache { - u64 seqnum; - struct vm_area_struct *vmas[VMACACHE_SIZE]; -}; - /* * When updating this, please also update struct resident_page_types[] in * kernel/fork.c diff --git a/include/linux/sched.h b/include/linux/sched.h index c46f3a63b758..dc131048d46a 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -859,7 +859,6 @@ struct task_struct { struct mm_struct *active_mm; /* Per-thread vma caching: */ - struct vmacache vmacache; #ifdef SPLIT_RSS_COUNTING struct task_rss_stat rss_stat; diff --git a/include/linux/vm_event_item.h b/include/linux/vm_event_item.h index 404024486fa5..1ce8fadb2b1c 100644 --- a/include/linux/vm_event_item.h +++ b/include/linux/vm_event_item.h @@ -122,10 +122,6 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT, NR_TLB_LOCAL_FLUSH_ALL, NR_TLB_LOCAL_FLUSH_ONE, #endif /* CONFIG_DEBUG_TLBFLUSH */ -#ifdef CONFIG_DEBUG_VM_VMACACHE - VMACACHE_FIND_CALLS, - VMACACHE_FIND_HITS, -#endif #ifdef CONFIG_SWAP SWAP_RA, SWAP_RA_HIT, diff --git a/include/linux/vmacache.h b/include/linux/vmacache.h deleted file mode 100644 index 6fce268a4588..000000000000 --- a/include/linux/vmacache.h +++ /dev/null @@ -1,28 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0 */ -#ifndef __LINUX_VMACACHE_H -#define __LINUX_VMACACHE_H - -#include -#include - -static inline void vmacache_flush(struct task_struct *tsk) -{ - memset(tsk->vmacache.vmas, 0, sizeof(tsk->vmacache.vmas)); -} - -extern void vmacache_update(unsigned long addr, struct vm_area_struct *newvma); -extern struct vm_area_struct *vmacache_find(struct mm_struct *mm, - unsigned long addr); - -#ifndef CONFIG_MMU -extern struct vm_area_struct *vmacache_find_exact(struct mm_struct *mm, - unsigned long start, - unsigned long end); -#endif - -static inline void vmacache_invalidate(struct mm_struct *mm) -{ - mm->vmacache_seqnum++; -} - -#endif /* __LINUX_VMACACHE_H */ diff --git a/include/linux/vmstat.h b/include/linux/vmstat.h index bfe38869498d..19cf5b6892ce 100644 --- a/include/linux/vmstat.h +++ b/include/linux/vmstat.h @@ -125,12 +125,6 @@ static inline void vm_events_fold_cpu(int cpu) #define count_vm_tlb_events(x, y) do { (void)(y); } while (0) #endif -#ifdef CONFIG_DEBUG_VM_VMACACHE -#define count_vm_vmacache_event(x) count_vm_event(x) -#else -#define count_vm_vmacache_event(x) do {} while (0) -#endif - #define __count_zid_vm_events(item, zid, delta) \ __count_vm_events(item##_NORMAL - ZONE_NORMAL + zid, delta) diff --git a/kernel/debug/debug_core.c b/kernel/debug/debug_core.c index 7beceb447211..d5e9ccde3ab8 100644 --- a/kernel/debug/debug_core.c +++ b/kernel/debug/debug_core.c @@ -50,7 +50,6 @@ #include #include #include -#include #include #include #include @@ -283,17 +282,6 @@ static void kgdb_flush_swbreak_addr(unsigned long addr) if (!CACHE_FLUSH_IS_SAFE) return; - if (current->mm) { - int i; - - for (i = 0; i < VMACACHE_SIZE; i++) { - if (!current->vmacache.vmas[i]) - continue; - flush_cache_range(current->vmacache.vmas[i], - addr, addr + BREAK_INSTR_SIZE); - } - } - /* Force flush instruction cache if it was outside the mm */ flush_icache_range(addr, addr + BREAK_INSTR_SIZE); } diff --git a/kernel/fork.c b/kernel/fork.c index db420a1e93c7..a653b259c321 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -43,7 +43,6 @@ #include #include #include -#include #include #include #include @@ -1123,7 +1122,6 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, mm->mmap = NULL; mt_init_flags(&mm->mm_mt, MM_MT_FLAGS); mt_set_external_lock(&mm->mm_mt, &mm->mmap_lock); - mm->vmacache_seqnum = 0; atomic_set(&mm->mm_users, 1); atomic_set(&mm->mm_count, 1); seqcount_init(&mm->write_protect_seq); @@ -1578,9 +1576,6 @@ static int copy_mm(unsigned long clone_flags, struct task_struct *tsk) if (!oldmm) return 0; - /* initialize the new vmacache entries */ - vmacache_flush(tsk); - if (clone_flags & CLONE_VM) { mmget(oldmm); mm = oldmm; diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 05c3858cdf41..6b84d0e7c3ad 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -804,14 +804,6 @@ config DEBUG_VM If unsure, say N. -config DEBUG_VM_VMACACHE - bool "Debug VMA caching" - depends on DEBUG_VM - help - Enable this to turn on VMA caching debug information. Doing so - can cause significant overhead, so only enable it in non-production - environments. - config DEBUG_VM_MAPLE_TREE bool "Debug VM maple trees" depends on DEBUG_VM diff --git a/mm/Makefile b/mm/Makefile index 6f9ffa968a1a..298c9991ab75 100644 --- a/mm/Makefile +++ b/mm/Makefile @@ -52,7 +52,7 @@ obj-y := filemap.o mempool.o oom_kill.o fadvise.o \ readahead.o swap.o truncate.o vmscan.o shmem.o \ util.o mmzone.o vmstat.o backing-dev.o \ mm_init.o percpu.o slab_common.o \ - compaction.o vmacache.o \ + compaction.o \ interval_tree.o list_lru.o workingset.o \ debug.o gup.o mmap_lock.o $(mmu-y) diff --git a/mm/debug.c b/mm/debug.c index bef329bf28f0..2d625ca0e326 100644 --- a/mm/debug.c +++ b/mm/debug.c @@ -155,7 +155,7 @@ EXPORT_SYMBOL(dump_vma); void dump_mm(const struct mm_struct *mm) { - pr_emerg("mm %px mmap %px seqnum %llu task_size %lu\n" + pr_emerg("mm %px mmap %px task_size %lu\n" #ifdef CONFIG_MMU "get_unmapped_area %px\n" #endif @@ -183,7 +183,7 @@ void dump_mm(const struct mm_struct *mm) "tlb_flush_pending %d\n" "def_flags: %#lx(%pGv)\n", - mm, mm->mmap, (long long) mm->vmacache_seqnum, mm->task_size, + mm, mm->mmap, mm->task_size, #ifdef CONFIG_MMU mm->get_unmapped_area, #endif diff --git a/mm/mmap.c b/mm/mmap.c index 4197663696b5..f35c4237ee81 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -14,7 +14,6 @@ #include #include #include -#include #include #include #include @@ -724,9 +723,6 @@ inline int vma_expand(struct ma_state *mas, struct vm_area_struct *vma, /* Remove from mm linked list - also updates highest_vm_end */ __vma_unlink_list(mm, next); - /* Kill the cache */ - vmacache_invalidate(mm); - if (file) __remove_shared_vm_struct(next, file, mapping); @@ -966,8 +962,6 @@ again: if (remove_next) { __vma_unlink_list(mm, next); - /* Kill the cache */ - vmacache_invalidate(mm); if (file) __remove_shared_vm_struct(next, file, mapping); } else if (insert) { @@ -2265,19 +2259,10 @@ struct vm_area_struct *find_vma_intersection(struct mm_struct *mm, unsigned long start_addr, unsigned long end_addr) { - struct vm_area_struct *vma; unsigned long index = start_addr; mmap_assert_locked(mm); - /* Check the cache first. */ - vma = vmacache_find(mm, start_addr); - if (likely(vma)) - return vma; - - vma = mt_find(&mm->mm_mt, &index, end_addr - 1); - if (vma) - vmacache_update(start_addr, vma); - return vma; + return mt_find(&mm->mm_mt, &index, end_addr - 1); } EXPORT_SYMBOL(find_vma_intersection); @@ -2291,19 +2276,10 @@ EXPORT_SYMBOL(find_vma_intersection); */ struct vm_area_struct *find_vma(struct mm_struct *mm, unsigned long addr) { - struct vm_area_struct *vma; unsigned long index = addr; mmap_assert_locked(mm); - /* Check the cache first. */ - vma = vmacache_find(mm, addr); - if (likely(vma)) - return vma; - - vma = mt_find(&mm->mm_mt, &index, ULONG_MAX); - if (vma) - vmacache_update(addr, vma); - return vma; + return mt_find(&mm->mm_mt, &index, ULONG_MAX); } EXPORT_SYMBOL(find_vma); @@ -2693,9 +2669,6 @@ detach_vmas_to_be_unmapped(struct mm_struct *mm, struct ma_state *mas, mm->highest_vm_end = prev ? vm_end_gap(prev) : 0; tail_vma->vm_next = NULL; - /* Kill the cache */ - vmacache_invalidate(mm); - /* * Do not downgrade mmap_lock if we are next to VM_GROWSDOWN or * VM_GROWSUP VMA. Such VMAs can change their size under @@ -3075,7 +3048,6 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma, vma_mas_remove(&unmap, mas); - vmacache_invalidate(vma->vm_mm); if (vma->anon_vma) { anon_vma_interval_tree_post_update_vma(vma); anon_vma_unlock_write(vma->anon_vma); diff --git a/mm/nommu.c b/mm/nommu.c index b098c0251137..4ae252a785dc 100644 --- a/mm/nommu.c +++ b/mm/nommu.c @@ -19,7 +19,6 @@ #include #include #include -#include #include #include #include @@ -598,23 +597,12 @@ static void add_vma_to_mm(struct mm_struct *mm, struct vm_area_struct *vma) */ static void delete_vma_from_mm(struct vm_area_struct *vma) { - int i; - struct address_space *mapping; - struct mm_struct *mm = vma->vm_mm; - struct task_struct *curr = current; MA_STATE(mas, &vma->vm_mm->mm_mt, 0, 0); - mm->map_count--; - for (i = 0; i < VMACACHE_SIZE; i++) { - /* if the vma is cached, invalidate the entire cache */ - if (curr->vmacache.vmas[i] == vma) { - vmacache_invalidate(mm); - break; - } - } - + vma->vm_mm->map_count--; /* remove the VMA from the mapping */ if (vma->vm_file) { + struct address_space *mapping; mapping = vma->vm_file->f_mapping; i_mmap_lock_write(mapping); @@ -626,7 +614,7 @@ static void delete_vma_from_mm(struct vm_area_struct *vma) /* remove from the MM's tree and list */ vma_mas_remove(vma, &mas); - __vma_unlink_list(mm, vma); + __vma_unlink_list(vma->vm_mm, vma); } /* @@ -659,20 +647,9 @@ EXPORT_SYMBOL(find_vma_intersection); */ struct vm_area_struct *find_vma(struct mm_struct *mm, unsigned long addr) { - struct vm_area_struct *vma; MA_STATE(mas, &mm->mm_mt, addr, addr); - /* check the cache first */ - vma = vmacache_find(mm, addr); - if (likely(vma)) - return vma; - - vma = mas_walk(&mas); - - if (vma) - vmacache_update(addr, vma); - - return vma; + return mas_walk(&mas); } EXPORT_SYMBOL(find_vma); @@ -706,11 +683,6 @@ static struct vm_area_struct *find_vma_exact(struct mm_struct *mm, unsigned long end = addr + len; MA_STATE(mas, &mm->mm_mt, addr, addr); - /* check the cache first */ - vma = vmacache_find_exact(mm, addr, end); - if (vma) - return vma; - vma = mas_walk(&mas); if (!vma) return NULL; @@ -719,7 +691,6 @@ static struct vm_area_struct *find_vma_exact(struct mm_struct *mm, if (vma->vm_end != end) return NULL; - vmacache_update(addr, vma); return vma; } diff --git a/mm/vmacache.c b/mm/vmacache.c deleted file mode 100644 index 01a6e6688ec1..000000000000 --- a/mm/vmacache.c +++ /dev/null @@ -1,117 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 -/* - * Copyright (C) 2014 Davidlohr Bueso. - */ -#include -#include -#include -#include - -/* - * Hash based on the pmd of addr if configured with MMU, which provides a good - * hit rate for workloads with spatial locality. Otherwise, use pages. - */ -#ifdef CONFIG_MMU -#define VMACACHE_SHIFT PMD_SHIFT -#else -#define VMACACHE_SHIFT PAGE_SHIFT -#endif -#define VMACACHE_HASH(addr) ((addr >> VMACACHE_SHIFT) & VMACACHE_MASK) - -/* - * This task may be accessing a foreign mm via (for example) - * get_user_pages()->find_vma(). The vmacache is task-local and this - * task's vmacache pertains to a different mm (ie, its own). There is - * nothing we can do here. - * - * Also handle the case where a kernel thread has adopted this mm via - * kthread_use_mm(). That kernel thread's vmacache is not applicable to this mm. - */ -static inline bool vmacache_valid_mm(struct mm_struct *mm) -{ - return current->mm == mm && !(current->flags & PF_KTHREAD); -} - -void vmacache_update(unsigned long addr, struct vm_area_struct *newvma) -{ - if (vmacache_valid_mm(newvma->vm_mm)) - current->vmacache.vmas[VMACACHE_HASH(addr)] = newvma; -} - -static bool vmacache_valid(struct mm_struct *mm) -{ - struct task_struct *curr; - - if (!vmacache_valid_mm(mm)) - return false; - - curr = current; - if (mm->vmacache_seqnum != curr->vmacache.seqnum) { - /* - * First attempt will always be invalid, initialize - * the new cache for this task here. - */ - curr->vmacache.seqnum = mm->vmacache_seqnum; - vmacache_flush(curr); - return false; - } - return true; -} - -struct vm_area_struct *vmacache_find(struct mm_struct *mm, unsigned long addr) -{ - int idx = VMACACHE_HASH(addr); - int i; - - count_vm_vmacache_event(VMACACHE_FIND_CALLS); - - if (!vmacache_valid(mm)) - return NULL; - - for (i = 0; i < VMACACHE_SIZE; i++) { - struct vm_area_struct *vma = current->vmacache.vmas[idx]; - - if (vma) { -#ifdef CONFIG_DEBUG_VM_VMACACHE - if (WARN_ON_ONCE(vma->vm_mm != mm)) - break; -#endif - if (vma->vm_start <= addr && vma->vm_end > addr) { - count_vm_vmacache_event(VMACACHE_FIND_HITS); - return vma; - } - } - if (++idx == VMACACHE_SIZE) - idx = 0; - } - - return NULL; -} - -#ifndef CONFIG_MMU -struct vm_area_struct *vmacache_find_exact(struct mm_struct *mm, - unsigned long start, - unsigned long end) -{ - int idx = VMACACHE_HASH(start); - int i; - - count_vm_vmacache_event(VMACACHE_FIND_CALLS); - - if (!vmacache_valid(mm)) - return NULL; - - for (i = 0; i < VMACACHE_SIZE; i++) { - struct vm_area_struct *vma = current->vmacache.vmas[idx]; - - if (vma && vma->vm_start == start && vma->vm_end == end) { - count_vm_vmacache_event(VMACACHE_FIND_HITS); - return vma; - } - if (++idx == VMACACHE_SIZE) - idx = 0; - } - - return NULL; -} -#endif diff --git a/mm/vmstat.c b/mm/vmstat.c index 373d2730fcf2..da7e389cf33c 100644 --- a/mm/vmstat.c +++ b/mm/vmstat.c @@ -1382,10 +1382,6 @@ const char * const vmstat_text[] = { "nr_tlb_local_flush_one", #endif /* CONFIG_DEBUG_TLBFLUSH */ -#ifdef CONFIG_DEBUG_VM_VMACACHE - "vmacache_find_calls", - "vmacache_find_hits", -#endif #ifdef CONFIG_SWAP "swap_ra", "swap_ra_hit", -- cgit v1.2.3 From b47818e2afa6330838002a1b08fbfec01f34563d Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:46:59 +0000 Subject: mm: convert vma_lookup() to use mtree_load() Unlike the rbtree, the Maple Tree will return a NULL if there's nothing at a particular address. Since the previous commit dropped the vmacache, it is now possible to consult the tree directly. Link: https://lkml.kernel.org/r/20220504011345.662299-11-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-27-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Signed-off-by: Matthew Wilcox (Oracle) Acked-by: Vlastimil Babka Cc: Catalin Marinas Cc: David Howells Cc: SeongJae Park Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- include/linux/mm.h | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index f3c19292ea16..31141f97df3e 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2808,12 +2808,7 @@ struct vm_area_struct *find_vma_intersection(struct mm_struct *mm, static inline struct vm_area_struct *vma_lookup(struct mm_struct *mm, unsigned long addr) { - struct vm_area_struct *vma = find_vma(mm, addr); - - if (vma && addr < vma->vm_start) - vma = NULL; - - return vma; + return mtree_load(&mm->mm_mt, addr); } static inline unsigned long vm_start_gap(struct vm_area_struct *vma) -- cgit v1.2.3 From 392e35bc64f10697f255c7ed6548cd5a480f37d0 Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:47:01 +0000 Subject: mm/mmap: reorganize munmap to use maple states Remove __do_munmap() in favour of do_munmap(), do_mas_munmap(), and do_mas_align_munmap(). do_munmap() is a wrapper to create a maple state for any callers that have not been converted to the maple tree. do_mas_munmap() takes a maple state to mumap a range. This is just a small function which checks for error conditions and aligns the end of the range. do_mas_align_munmap() uses the aligned range to mumap a range. do_mas_align_munmap() starts with the first VMA in the range, then finds the last VMA in the range. Both start and end are split if necessary. Then the VMAs are removed from the linked list and the mm mlock count is updated at the same time. Followed by a single tree operation of overwriting the area in with a NULL. Finally, the detached list is unmapped and freed. By reorganizing the munmap calls as outlined, it is now possible to avoid extra work of aligning pre-aligned callers which are known to be safe, avoid extra VMA lookups or tree walks for modifications. detach_vmas_to_be_unmapped() is no longer used, so drop this code. vm_brk_flags() can just call the do_mas_munmap() as it checks for intersecting VMAs directly. Link: https://lkml.kernel.org/r/20220504011345.662299-13-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-29-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Cc: Catalin Marinas Cc: David Howells Cc: "Matthew Wilcox (Oracle)" Cc: SeongJae Park Cc: Vlastimil Babka Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- include/linux/mm.h | 5 +- mm/mmap.c | 231 +++++++++++++++++++++++++++++++++-------------------- mm/mremap.c | 17 ++-- 3 files changed, 158 insertions(+), 95 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 31141f97df3e..1da66b1bff04 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2723,8 +2723,9 @@ extern unsigned long mmap_region(struct file *file, unsigned long addr, extern unsigned long do_mmap(struct file *file, unsigned long addr, unsigned long len, unsigned long prot, unsigned long flags, unsigned long pgoff, unsigned long *populate, struct list_head *uf); -extern int __do_munmap(struct mm_struct *, unsigned long, size_t, - struct list_head *uf, bool downgrade); +extern int do_mas_munmap(struct ma_state *mas, struct mm_struct *mm, + unsigned long start, size_t len, struct list_head *uf, + bool downgrade); extern int do_munmap(struct mm_struct *, unsigned long, size_t, struct list_head *uf); extern int do_madvise(struct mm_struct *mm, unsigned long start, size_t len_in, int behavior); diff --git a/mm/mmap.c b/mm/mmap.c index 2503c00df794..b2a071abf6c7 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -2395,47 +2395,6 @@ static void unmap_region(struct mm_struct *mm, tlb_finish_mmu(&tlb); } -/* - * Create a list of vma's touched by the unmap, removing them from the mm's - * vma list as we go.. - */ -static bool -detach_vmas_to_be_unmapped(struct mm_struct *mm, struct ma_state *mas, - struct vm_area_struct *vma, struct vm_area_struct *prev, - unsigned long end) -{ - struct vm_area_struct **insertion_point; - struct vm_area_struct *tail_vma = NULL; - - insertion_point = (prev ? &prev->vm_next : &mm->mmap); - vma->vm_prev = NULL; - vma_mas_szero(mas, vma->vm_start, end); - do { - if (vma->vm_flags & VM_LOCKED) - mm->locked_vm -= vma_pages(vma); - mm->map_count--; - tail_vma = vma; - vma = vma->vm_next; - } while (vma && vma->vm_start < end); - *insertion_point = vma; - if (vma) - vma->vm_prev = prev; - else - mm->highest_vm_end = prev ? vm_end_gap(prev) : 0; - tail_vma->vm_next = NULL; - - /* - * Do not downgrade mmap_lock if we are next to VM_GROWSDOWN or - * VM_GROWSUP VMA. Such VMAs can change their size under - * down_read(mmap_lock) and collide with the VMA we are about to unmap. - */ - if (vma && (vma->vm_flags & VM_GROWSDOWN)) - return false; - if (prev && (prev->vm_flags & VM_GROWSUP)) - return false; - return true; -} - /* * __split_vma() bypasses sysctl_max_map_count checking. We use this where it * has already been checked or doesn't make sense to fail. @@ -2515,40 +2474,51 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma, return __split_vma(mm, vma, addr, new_below); } -/* Munmap is split into 2 main parts -- this part which finds - * what needs doing, and the areas themselves, which do the - * work. This now handles partial unmappings. - * Jeremy Fitzhardinge - */ -int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, - struct list_head *uf, bool downgrade) +static inline int +unlock_range(struct vm_area_struct *start, struct vm_area_struct **tail, + unsigned long limit) { - unsigned long end; - struct vm_area_struct *vma, *prev, *last; - int error = -ENOMEM; - MA_STATE(mas, &mm->mm_mt, 0, 0); + struct mm_struct *mm = start->vm_mm; + struct vm_area_struct *tmp = start; + int count = 0; - if ((offset_in_page(start)) || start > TASK_SIZE || len > TASK_SIZE-start) - return -EINVAL; + while (tmp && tmp->vm_start < limit) { + *tail = tmp; + count++; + if (tmp->vm_flags & VM_LOCKED) + mm->locked_vm -= vma_pages(tmp); - len = PAGE_ALIGN(len); - end = start + len; - if (len == 0) - return -EINVAL; + tmp = tmp->vm_next; + } - /* arch_unmap() might do unmaps itself. */ - arch_unmap(mm, start, end); + return count; +} - /* Find the first overlapping VMA where start < vma->vm_end */ - vma = find_vma_intersection(mm, start, end); - if (!vma) - return 0; +/* + * do_mas_align_munmap() - munmap the aligned region from @start to @end. + * @mas: The maple_state, ideally set up to alter the correct tree location. + * @vma: The starting vm_area_struct + * @mm: The mm_struct + * @start: The aligned start address to munmap. + * @end: The aligned end address to munmap. + * @uf: The userfaultfd list_head + * @downgrade: Set to true to attempt a write downgrade of the mmap_sem + * + * If @downgrade is true, check return code for potential release of the lock. + */ +static int +do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma, + struct mm_struct *mm, unsigned long start, + unsigned long end, struct list_head *uf, bool downgrade) +{ + struct vm_area_struct *prev, *last; + int error = -ENOMEM; + /* we have start < vma->vm_end */ - if (mas_preallocate(&mas, vma, GFP_KERNEL)) + if (mas_preallocate(mas, vma, GFP_KERNEL)) return -ENOMEM; - prev = vma->vm_prev; - /* we have start < vma->vm_end */ + mas->last = end - 1; /* * If we need to split any vma, do it now to save pain later. * @@ -2569,17 +2539,31 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, error = __split_vma(mm, vma, start, 0); if (error) goto split_failed; + prev = vma; + vma = __vma_next(mm, prev); + mas->index = start; + mas_reset(mas); + } else { + prev = vma->vm_prev; } + if (vma->vm_end >= end) + last = vma; + else + last = find_vma_intersection(mm, end - 1, end); + /* Does it split the last one? */ - last = find_vma(mm, end); - if (last && end > last->vm_start) { + if (last && end < last->vm_end) { error = __split_vma(mm, last, end, 1); + if (error) goto split_failed; + + if (vma == last) + vma = __vma_next(mm, prev); + mas_reset(mas); } - vma = __vma_next(mm, prev); if (unlikely(uf)) { /* @@ -2592,16 +2576,46 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, * failure that it's not worth optimizing it for. */ error = userfaultfd_unmap_prep(vma, start, end, uf); + if (error) goto userfaultfd_error; } - /* Detach vmas from rbtree */ - if (!detach_vmas_to_be_unmapped(mm, &mas, vma, prev, end)) - downgrade = false; + /* + * unlock any mlock()ed ranges before detaching vmas, count the number + * of VMAs to be dropped, and return the tail entry of the affected + * area. + */ + mm->map_count -= unlock_range(vma, &last, end); + /* Drop removed area from the tree */ + mas_store_prealloc(mas, NULL); - if (downgrade) - mmap_write_downgrade(mm); + /* Detach vmas from the MM linked list */ + vma->vm_prev = NULL; + if (prev) + prev->vm_next = last->vm_next; + else + mm->mmap = last->vm_next; + + if (last->vm_next) { + last->vm_next->vm_prev = prev; + last->vm_next = NULL; + } else + mm->highest_vm_end = prev ? vm_end_gap(prev) : 0; + + /* + * Do not downgrade mmap_lock if we are next to VM_GROWSDOWN or + * VM_GROWSUP VMA. Such VMAs can change their size under + * down_read(mmap_lock) and collide with the VMA we are about to unmap. + */ + if (downgrade) { + if (last && (last->vm_flags & VM_GROWSDOWN)) + downgrade = false; + else if (prev && (prev->vm_flags & VM_GROWSUP)) + downgrade = false; + else + mmap_write_downgrade(mm); + } unmap_region(mm, vma, prev, start, end); @@ -2615,14 +2629,63 @@ int __do_munmap(struct mm_struct *mm, unsigned long start, size_t len, map_count_exceeded: split_failed: userfaultfd_error: - mas_destroy(&mas); + mas_destroy(mas); return error; } +/* + * do_mas_munmap() - munmap a given range. + * @mas: The maple state + * @mm: The mm_struct + * @start: The start address to munmap + * @len: The length of the range to munmap + * @uf: The userfaultfd list_head + * @downgrade: set to true if the user wants to attempt to write_downgrade the + * mmap_sem + * + * This function takes a @mas that is either pointing to the previous VMA or set + * to MA_START and sets it up to remove the mapping(s). The @len will be + * aligned and any arch_unmap work will be preformed. + * + * Returns: -EINVAL on failure, 1 on success and unlock, 0 otherwise. + */ +int do_mas_munmap(struct ma_state *mas, struct mm_struct *mm, + unsigned long start, size_t len, struct list_head *uf, + bool downgrade) +{ + unsigned long end; + struct vm_area_struct *vma; + + if ((offset_in_page(start)) || start > TASK_SIZE || len > TASK_SIZE-start) + return -EINVAL; + + end = start + PAGE_ALIGN(len); + if (end == start) + return -EINVAL; + + /* arch_unmap() might do unmaps itself. */ + arch_unmap(mm, start, end); + + /* Find the first overlapping VMA */ + vma = mas_find(mas, end - 1); + if (!vma) + return 0; + + return do_mas_align_munmap(mas, vma, mm, start, end, uf, downgrade); +} + +/* do_munmap() - Wrapper function for non-maple tree aware do_munmap() calls. + * @mm: The mm_struct + * @start: The start address to munmap + * @len: The length to be munmapped. + * @uf: The userfaultfd list_head + */ int do_munmap(struct mm_struct *mm, unsigned long start, size_t len, struct list_head *uf) { - return __do_munmap(mm, start, len, uf, false); + MA_STATE(mas, &mm->mm_mt, start, start); + + return do_mas_munmap(&mas, mm, start, len, uf, false); } unsigned long mmap_region(struct file *file, unsigned long addr, @@ -2656,7 +2719,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr, } /* Unmap any existing mapping in the area */ - if (do_munmap(mm, addr, len, uf)) + if (do_mas_munmap(&mas, mm, addr, len, uf, false)) return -ENOMEM; /* @@ -2876,11 +2939,12 @@ static int __vm_munmap(unsigned long start, size_t len, bool downgrade) int ret; struct mm_struct *mm = current->mm; LIST_HEAD(uf); + MA_STATE(mas, &mm->mm_mt, start, start); if (mmap_write_lock_killable(mm)) return -EINTR; - ret = __do_munmap(mm, start, len, &uf, downgrade); + ret = do_mas_munmap(&mas, mm, start, len, &uf, downgrade); /* * Returning 1 indicates mmap_lock is downgraded. * But 1 is not legal return value of vm_munmap() and munmap(), reset @@ -3015,10 +3079,7 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma, if (likely((vma->vm_end < oldbrk) || ((vma->vm_start == newbrk) && (vma->vm_end == oldbrk)))) { /* remove entire mapping(s) */ - mas_set(mas, newbrk); - if (vma->vm_start != newbrk) - mas_reset(mas); /* cause a re-walk for the first overlap. */ - ret = __do_munmap(mm, newbrk, oldbrk - newbrk, uf, true); + ret = do_mas_munmap(mas, mm, newbrk, oldbrk-newbrk, uf, true); goto munmap_full_vma; } @@ -3199,9 +3260,7 @@ int vm_brk_flags(unsigned long addr, unsigned long request, unsigned long flags) if (ret) goto limits_failed; - if (find_vma_intersection(mm, addr, addr + len)) - ret = do_munmap(mm, addr, len, &uf); - + ret = do_mas_munmap(&mas, mm, addr, len, &uf, 0); if (ret) goto munmap_failed; diff --git a/mm/mremap.c b/mm/mremap.c index b522cd0259a0..e0fba9004246 100644 --- a/mm/mremap.c +++ b/mm/mremap.c @@ -975,20 +975,23 @@ SYSCALL_DEFINE5(mremap, unsigned long, addr, unsigned long, old_len, /* * Always allow a shrinking remap: that just unmaps * the unnecessary pages.. - * __do_munmap does all the needed commit accounting, and + * do_mas_munmap does all the needed commit accounting, and * downgrades mmap_lock to read if so directed. */ if (old_len >= new_len) { int retval; + MA_STATE(mas, &mm->mm_mt, addr + new_len, addr + new_len); - retval = __do_munmap(mm, addr+new_len, old_len - new_len, - &uf_unmap, true); - if (retval < 0 && old_len != new_len) { - ret = retval; - goto out; + retval = do_mas_munmap(&mas, mm, addr + new_len, + old_len - new_len, &uf_unmap, true); /* Returning 1 indicates mmap_lock is downgraded to read. */ - } else if (retval == 1) + if (retval == 1) { downgraded = true; + } else if (retval < 0 && old_len != new_len) { + ret = retval; + goto out; + } + ret = addr; goto out; } -- cgit v1.2.3 From 55574e0b6ac1a3e06d3c6d77c1c444d73fe46ed1 Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:47:07 +0000 Subject: userfaultfd: use maple tree iterator to iterate VMAs Don't use the mm_struct linked list or the vma->vm_next in prep for removal. Link: https://lkml.kernel.org/r/20220504011345.662299-29-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220615164150.652376-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-45-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Cc: Catalin Marinas Cc: David Howells Cc: "Matthew Wilcox (Oracle)" Cc: SeongJae Park Cc: Vlastimil Babka Cc: Will Deacon Cc: Davidlohr Bueso Signed-off-by: Andrew Morton --- fs/userfaultfd.c | 62 +++++++++++++++++++++++++++++-------------- include/linux/userfaultfd_k.h | 7 +++-- mm/mmap.c | 7 ++--- 3 files changed, 47 insertions(+), 29 deletions(-) (limited to 'include/linux') diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c index e943370107d0..fe6f283d26d5 100644 --- a/fs/userfaultfd.c +++ b/fs/userfaultfd.c @@ -613,14 +613,16 @@ static void userfaultfd_event_wait_completion(struct userfaultfd_ctx *ctx, if (release_new_ctx) { struct vm_area_struct *vma; struct mm_struct *mm = release_new_ctx->mm; + VMA_ITERATOR(vmi, mm, 0); /* the various vma->vm_userfaultfd_ctx still points to it */ mmap_write_lock(mm); - for (vma = mm->mmap; vma; vma = vma->vm_next) + for_each_vma(vmi, vma) { if (vma->vm_userfaultfd_ctx.ctx == release_new_ctx) { vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX; vma->vm_flags &= ~__VM_UFFD_FLAGS; } + } mmap_write_unlock(mm); userfaultfd_ctx_put(release_new_ctx); @@ -801,11 +803,13 @@ static bool has_unmap_ctx(struct userfaultfd_ctx *ctx, struct list_head *unmaps, return false; } -int userfaultfd_unmap_prep(struct vm_area_struct *vma, - unsigned long start, unsigned long end, - struct list_head *unmaps) +int userfaultfd_unmap_prep(struct mm_struct *mm, unsigned long start, + unsigned long end, struct list_head *unmaps) { - for ( ; vma && vma->vm_start < end; vma = vma->vm_next) { + VMA_ITERATOR(vmi, mm, start); + struct vm_area_struct *vma; + + for_each_vma_range(vmi, vma, end) { struct userfaultfd_unmap_ctx *unmap_ctx; struct userfaultfd_ctx *ctx = vma->vm_userfaultfd_ctx.ctx; @@ -855,6 +859,7 @@ static int userfaultfd_release(struct inode *inode, struct file *file) /* len == 0 means wake all */ struct userfaultfd_wake_range range = { .len = 0, }; unsigned long new_flags; + MA_STATE(mas, &mm->mm_mt, 0, 0); WRITE_ONCE(ctx->released, true); @@ -871,7 +876,7 @@ static int userfaultfd_release(struct inode *inode, struct file *file) */ mmap_write_lock(mm); prev = NULL; - for (vma = mm->mmap; vma; vma = vma->vm_next) { + mas_for_each(&mas, vma, ULONG_MAX) { cond_resched(); BUG_ON(!!vma->vm_userfaultfd_ctx.ctx ^ !!(vma->vm_flags & __VM_UFFD_FLAGS)); @@ -885,10 +890,13 @@ static int userfaultfd_release(struct inode *inode, struct file *file) vma->vm_file, vma->vm_pgoff, vma_policy(vma), NULL_VM_UFFD_CTX, anon_vma_name(vma)); - if (prev) + if (prev) { + mas_pause(&mas); vma = prev; - else + } else { prev = vma; + } + vma->vm_flags = new_flags; vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX; } @@ -1270,6 +1278,7 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx, bool found; bool basic_ioctls; unsigned long start, end, vma_end; + MA_STATE(mas, &mm->mm_mt, 0, 0); user_uffdio_register = (struct uffdio_register __user *) arg; @@ -1312,7 +1321,8 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx, goto out; mmap_write_lock(mm); - vma = find_vma_prev(mm, start, &prev); + mas_set(&mas, start); + vma = mas_find(&mas, ULONG_MAX); if (!vma) goto out_unlock; @@ -1337,7 +1347,7 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx, */ found = false; basic_ioctls = false; - for (cur = vma; cur && cur->vm_start < end; cur = cur->vm_next) { + for (cur = vma; cur; cur = mas_next(&mas, end - 1)) { cond_resched(); BUG_ON(!!cur->vm_userfaultfd_ctx.ctx ^ @@ -1397,8 +1407,10 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx, } BUG_ON(!found); - if (vma->vm_start < start) - prev = vma; + mas_set(&mas, start); + prev = mas_prev(&mas, 0); + if (prev != vma) + mas_next(&mas, ULONG_MAX); ret = 0; do { @@ -1428,6 +1440,8 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx, ((struct vm_userfaultfd_ctx){ ctx }), anon_vma_name(vma)); if (prev) { + /* vma_merge() invalidated the mas */ + mas_pause(&mas); vma = prev; goto next; } @@ -1435,11 +1449,15 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx, ret = split_vma(mm, vma, start, 1); if (ret) break; + /* split_vma() invalidated the mas */ + mas_pause(&mas); } if (vma->vm_end > end) { ret = split_vma(mm, vma, end, 0); if (ret) break; + /* split_vma() invalidated the mas */ + mas_pause(&mas); } next: /* @@ -1456,8 +1474,8 @@ static int userfaultfd_register(struct userfaultfd_ctx *ctx, skip: prev = vma; start = vma->vm_end; - vma = vma->vm_next; - } while (vma && vma->vm_start < end); + vma = mas_next(&mas, end - 1); + } while (vma); out_unlock: mmap_write_unlock(mm); mmput(mm); @@ -1501,6 +1519,7 @@ static int userfaultfd_unregister(struct userfaultfd_ctx *ctx, bool found; unsigned long start, end, vma_end; const void __user *buf = (void __user *)arg; + MA_STATE(mas, &mm->mm_mt, 0, 0); ret = -EFAULT; if (copy_from_user(&uffdio_unregister, buf, sizeof(uffdio_unregister))) @@ -1519,7 +1538,8 @@ static int userfaultfd_unregister(struct userfaultfd_ctx *ctx, goto out; mmap_write_lock(mm); - vma = find_vma_prev(mm, start, &prev); + mas_set(&mas, start); + vma = mas_find(&mas, ULONG_MAX); if (!vma) goto out_unlock; @@ -1544,7 +1564,7 @@ static int userfaultfd_unregister(struct userfaultfd_ctx *ctx, */ found = false; ret = -EINVAL; - for (cur = vma; cur && cur->vm_start < end; cur = cur->vm_next) { + for (cur = vma; cur; cur = mas_next(&mas, end - 1)) { cond_resched(); BUG_ON(!!cur->vm_userfaultfd_ctx.ctx ^ @@ -1564,8 +1584,10 @@ static int userfaultfd_unregister(struct userfaultfd_ctx *ctx, } BUG_ON(!found); - if (vma->vm_start < start) - prev = vma; + mas_set(&mas, start); + prev = mas_prev(&mas, 0); + if (prev != vma) + mas_next(&mas, ULONG_MAX); ret = 0; do { @@ -1630,8 +1652,8 @@ static int userfaultfd_unregister(struct userfaultfd_ctx *ctx, skip: prev = vma; start = vma->vm_end; - vma = vma->vm_next; - } while (vma && vma->vm_start < end); + vma = mas_next(&mas, end - 1); + } while (vma); out_unlock: mmap_write_unlock(mm); mmput(mm); diff --git a/include/linux/userfaultfd_k.h b/include/linux/userfaultfd_k.h index 732b522bacb7..eee374c29c85 100644 --- a/include/linux/userfaultfd_k.h +++ b/include/linux/userfaultfd_k.h @@ -173,9 +173,8 @@ extern bool userfaultfd_remove(struct vm_area_struct *vma, unsigned long start, unsigned long end); -extern int userfaultfd_unmap_prep(struct vm_area_struct *vma, - unsigned long start, unsigned long end, - struct list_head *uf); +extern int userfaultfd_unmap_prep(struct mm_struct *mm, unsigned long start, + unsigned long end, struct list_head *uf); extern void userfaultfd_unmap_complete(struct mm_struct *mm, struct list_head *uf); @@ -256,7 +255,7 @@ static inline bool userfaultfd_remove(struct vm_area_struct *vma, return true; } -static inline int userfaultfd_unmap_prep(struct vm_area_struct *vma, +static inline int userfaultfd_unmap_prep(struct mm_struct *mm, unsigned long start, unsigned long end, struct list_head *uf) { diff --git a/mm/mmap.c b/mm/mmap.c index 28d614aa42fe..07a9e1b0caee 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -2575,7 +2575,7 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma, * split, despite we could. This is unlikely enough * failure that it's not worth optimizing it for. */ - error = userfaultfd_unmap_prep(vma, start, end, uf); + error = userfaultfd_unmap_prep(mm, start, end, uf); if (error) goto userfaultfd_error; @@ -3083,10 +3083,7 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma, goto munmap_full_vma; } - vma_init(&unmap, mm); - unmap.vm_start = newbrk; - unmap.vm_end = oldbrk; - ret = userfaultfd_unmap_prep(&unmap, newbrk, oldbrk, uf); + ret = userfaultfd_unmap_prep(mm, newbrk, oldbrk, uf); if (ret) return ret; ret = 1; -- cgit v1.2.3 From e23cb70bcbef5c189790042d1d6d89e33308c386 Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Tue, 21 Jun 2022 20:47:14 +0000 Subject: mm: remove the vma linked list Replace any vm_next use with vma_find(). Update free_pgtables(), unmap_vmas(), and zap_page_range() to use the maple tree. Use the new free_pgtables() and unmap_vmas() in do_mas_align_munmap(). At the same time, alter the loop to be more compact. Now that free_pgtables() and unmap_vmas() take a maple tree as an argument, rearrange do_mas_align_munmap() to use the new tree to hold the vmas to remove. Remove __vma_link_list() and __vma_unlink_list() as they are exclusively used to update the linked list Drop linked list update from __insert_vm_struct(). Rework validation of tree as it was depending on the linked list. Link: https://lkml.kernel.org/r/20220504011345.662299-52-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220513141548.2019143-1-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20220621204632.3370049-68-Liam.Howlett@oracle.com Signed-off-by: Liam R. Howlett Cc: Catalin Marinas Cc: David Howells Cc: "Matthew Wilcox (Oracle)" Cc: SeongJae Park Cc: Vlastimil Babka Cc: Will Deacon Cc: Davidlohr Bueso Cc: Hulk Robot Cc: Yang Yingliang Signed-off-by: Andrew Morton --- include/linux/mm.h | 5 +- include/linux/mm_types.h | 4 - kernel/fork.c | 19 +- mm/debug.c | 14 +- mm/internal.h | 8 +- mm/memory.c | 33 +++- mm/mmap.c | 447 ++++++++++++++++++++--------------------------- mm/nommu.c | 6 - mm/util.c | 40 ----- 9 files changed, 229 insertions(+), 347 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 1da66b1bff04..ffde0619710f 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1873,8 +1873,9 @@ void zap_vma_ptes(struct vm_area_struct *vma, unsigned long address, unsigned long size); void zap_page_range(struct vm_area_struct *vma, unsigned long address, unsigned long size); -void unmap_vmas(struct mmu_gather *tlb, struct vm_area_struct *start_vma, - unsigned long start, unsigned long end); +void unmap_vmas(struct mmu_gather *tlb, struct maple_tree *mt, + struct vm_area_struct *start_vma, unsigned long start, + unsigned long end); struct mmu_notifier_range; diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 040bfc207cb1..5cc2cb396ca7 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -403,8 +403,6 @@ struct vm_area_struct { unsigned long vm_end; /* The first byte after our end address within vm_mm. */ - /* linked list of VM areas per task, sorted by address */ - struct vm_area_struct *vm_next, *vm_prev; struct mm_struct *vm_mm; /* The address space we belong to. */ /* @@ -468,7 +466,6 @@ struct vm_area_struct { struct kioctx_table; struct mm_struct { struct { - struct vm_area_struct *mmap; /* list of VMAs */ struct maple_tree mm_mt; #ifdef CONFIG_MMU unsigned long (*get_unmapped_area) (struct file *filp, @@ -483,7 +480,6 @@ struct mm_struct { unsigned long mmap_compat_legacy_base; #endif unsigned long task_size; /* size of task vm space */ - unsigned long highest_vm_end; /* highest vma end address */ pgd_t * pgd; #ifdef CONFIG_MEMBARRIER diff --git a/kernel/fork.c b/kernel/fork.c index 7fcefef4c84d..889f318db3c1 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -474,7 +474,6 @@ struct vm_area_struct *vm_area_dup(struct vm_area_struct *orig) */ *new = data_race(*orig); INIT_LIST_HEAD(&new->anon_vma_chain); - new->vm_next = new->vm_prev = NULL; dup_anon_vma_name(orig, new); } return new; @@ -579,7 +578,7 @@ static void dup_mm_exe_file(struct mm_struct *mm, struct mm_struct *oldmm) static __latent_entropy int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) { - struct vm_area_struct *mpnt, *tmp, *prev, **pprev; + struct vm_area_struct *mpnt, *tmp; int retval; unsigned long charge = 0; MA_STATE(old_mas, &oldmm->mm_mt, 0, 0); @@ -606,18 +605,11 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, mm->exec_vm = oldmm->exec_vm; mm->stack_vm = oldmm->stack_vm; - pprev = &mm->mmap; retval = ksm_fork(mm, oldmm); if (retval) goto out; khugepaged_fork(mm, oldmm); - retval = mas_expected_entries(&mas, oldmm->map_count); - if (retval) - goto out; - - prev = NULL; - retval = mas_expected_entries(&mas, oldmm->map_count); if (retval) goto out; @@ -689,14 +681,6 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, if (is_vm_hugetlb_page(tmp)) reset_vma_resv_huge_pages(tmp); - /* - * Link in the new vma and copy the page table entries. - */ - *pprev = tmp; - pprev = &tmp->vm_next; - tmp->vm_prev = prev; - prev = tmp; - /* Link the vma into the MT */ mas.index = tmp->vm_start; mas.last = tmp->vm_end - 1; @@ -1119,7 +1103,6 @@ static void mm_init_uprobes_state(struct mm_struct *mm) static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, struct user_namespace *user_ns) { - mm->mmap = NULL; mt_init_flags(&mm->mm_mt, MM_MT_FLAGS); mt_set_external_lock(&mm->mm_mt, &mm->mmap_lock); atomic_set(&mm->mm_users, 1); diff --git a/mm/debug.c b/mm/debug.c index 2d625ca0e326..0fd15ba70d16 100644 --- a/mm/debug.c +++ b/mm/debug.c @@ -139,13 +139,11 @@ EXPORT_SYMBOL(dump_page); void dump_vma(const struct vm_area_struct *vma) { - pr_emerg("vma %px start %px end %px\n" - "next %px prev %px mm %px\n" + pr_emerg("vma %px start %px end %px mm %px\n" "prot %lx anon_vma %px vm_ops %px\n" "pgoff %lx file %px private_data %px\n" "flags: %#lx(%pGv)\n", - vma, (void *)vma->vm_start, (void *)vma->vm_end, vma->vm_next, - vma->vm_prev, vma->vm_mm, + vma, (void *)vma->vm_start, (void *)vma->vm_end, vma->vm_mm, (unsigned long)pgprot_val(vma->vm_page_prot), vma->anon_vma, vma->vm_ops, vma->vm_pgoff, vma->vm_file, vma->vm_private_data, @@ -155,11 +153,11 @@ EXPORT_SYMBOL(dump_vma); void dump_mm(const struct mm_struct *mm) { - pr_emerg("mm %px mmap %px task_size %lu\n" + pr_emerg("mm %px task_size %lu\n" #ifdef CONFIG_MMU "get_unmapped_area %px\n" #endif - "mmap_base %lu mmap_legacy_base %lu highest_vm_end %lu\n" + "mmap_base %lu mmap_legacy_base %lu\n" "pgd %px mm_users %d mm_count %d pgtables_bytes %lu map_count %d\n" "hiwater_rss %lx hiwater_vm %lx total_vm %lx locked_vm %lx\n" "pinned_vm %llx data_vm %lx exec_vm %lx stack_vm %lx\n" @@ -183,11 +181,11 @@ void dump_mm(const struct mm_struct *mm) "tlb_flush_pending %d\n" "def_flags: %#lx(%pGv)\n", - mm, mm->mmap, mm->task_size, + mm, mm->task_size, #ifdef CONFIG_MMU mm->get_unmapped_area, #endif - mm->mmap_base, mm->mmap_legacy_base, mm->highest_vm_end, + mm->mmap_base, mm->mmap_legacy_base, mm->pgd, atomic_read(&mm->mm_users), atomic_read(&mm->mm_count), mm_pgtables_bytes(mm), diff --git a/mm/internal.h b/mm/internal.h index 7cf12a15475b..eba79d1d3b06 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -84,8 +84,9 @@ void folio_rotate_reclaimable(struct folio *folio); bool __folio_end_writeback(struct folio *folio); void deactivate_file_folio(struct folio *folio); -void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *start_vma, - unsigned long floor, unsigned long ceiling); +void free_pgtables(struct mmu_gather *tlb, struct maple_tree *mt, + struct vm_area_struct *start_vma, unsigned long floor, + unsigned long ceiling); void pmd_install(struct mm_struct *mm, pmd_t *pmd, pgtable_t *pte); struct zap_details; @@ -479,9 +480,6 @@ static inline bool is_data_mapping(vm_flags_t flags) } /* mm/util.c */ -void __vma_link_list(struct mm_struct *mm, struct vm_area_struct *vma, - struct vm_area_struct *prev); -void __vma_unlink_list(struct mm_struct *mm, struct vm_area_struct *vma); struct anon_vma *folio_anon_vma(struct folio *folio); #ifdef CONFIG_MMU diff --git a/mm/memory.c b/mm/memory.c index ed45dbf9c31e..fdfc119f181d 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -402,12 +402,21 @@ void free_pgd_range(struct mmu_gather *tlb, } while (pgd++, addr = next, addr != end); } -void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma, - unsigned long floor, unsigned long ceiling) +void free_pgtables(struct mmu_gather *tlb, struct maple_tree *mt, + struct vm_area_struct *vma, unsigned long floor, + unsigned long ceiling) { - while (vma) { - struct vm_area_struct *next = vma->vm_next; + MA_STATE(mas, mt, vma->vm_end, vma->vm_end); + + do { unsigned long addr = vma->vm_start; + struct vm_area_struct *next; + + /* + * Note: USER_PGTABLES_CEILING may be passed as ceiling and may + * be 0. This will underflow and is okay. + */ + next = mas_find(&mas, ceiling - 1); /* * Hide vma from rmap and truncate_pagecache before freeing @@ -426,7 +435,7 @@ void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma, while (next && next->vm_start <= vma->vm_end + PMD_SIZE && !is_vm_hugetlb_page(next)) { vma = next; - next = vma->vm_next; + next = mas_find(&mas, ceiling - 1); unlink_anon_vmas(vma); unlink_file_vma(vma); } @@ -434,7 +443,7 @@ void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma, floor, next ? next->vm_start : ceiling); } vma = next; - } + } while (vma); } void pmd_install(struct mm_struct *mm, pmd_t *pmd, pgtable_t *pte) @@ -1705,7 +1714,7 @@ static void unmap_single_vma(struct mmu_gather *tlb, * ensure that any thus-far unmapped pages are flushed before unmap_vmas() * drops the lock and schedules. */ -void unmap_vmas(struct mmu_gather *tlb, +void unmap_vmas(struct mmu_gather *tlb, struct maple_tree *mt, struct vm_area_struct *vma, unsigned long start_addr, unsigned long end_addr) { @@ -1715,12 +1724,14 @@ void unmap_vmas(struct mmu_gather *tlb, /* Careful - we need to zap private pages too! */ .even_cows = true, }; + MA_STATE(mas, mt, vma->vm_end, vma->vm_end); mmu_notifier_range_init(&range, MMU_NOTIFY_UNMAP, 0, vma, vma->vm_mm, start_addr, end_addr); mmu_notifier_invalidate_range_start(&range); - for ( ; vma && vma->vm_start < end_addr; vma = vma->vm_next) + do { unmap_single_vma(tlb, vma, start_addr, end_addr, &details); + } while ((vma = mas_find(&mas, end_addr - 1)) != NULL); mmu_notifier_invalidate_range_end(&range); } @@ -1735,8 +1746,11 @@ void unmap_vmas(struct mmu_gather *tlb, void zap_page_range(struct vm_area_struct *vma, unsigned long start, unsigned long size) { + struct maple_tree *mt = &vma->vm_mm->mm_mt; + unsigned long end = start + size; struct mmu_notifier_range range; struct mmu_gather tlb; + MA_STATE(mas, mt, vma->vm_end, vma->vm_end); lru_add_drain(); mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, @@ -1744,8 +1758,9 @@ void zap_page_range(struct vm_area_struct *vma, unsigned long start, tlb_gather_mmu(&tlb, vma->vm_mm); update_hiwater_rss(vma->vm_mm); mmu_notifier_invalidate_range_start(&range); - for ( ; vma && vma->vm_start < range.end; vma = vma->vm_next) + do { unmap_single_vma(&tlb, vma, start, range.end, NULL); + } while ((vma = mas_find(&mas, end - 1)) != NULL); mmu_notifier_invalidate_range_end(&range); tlb_finish_mmu(&tlb); } diff --git a/mm/mmap.c b/mm/mmap.c index 07a9e1b0caee..e17345eec39d 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -75,9 +75,10 @@ int mmap_rnd_compat_bits __read_mostly = CONFIG_ARCH_MMAP_RND_COMPAT_BITS; static bool ignore_rlimit_data; core_param(ignore_rlimit_data, ignore_rlimit_data, bool, 0644); -static void unmap_region(struct mm_struct *mm, +static void unmap_region(struct mm_struct *mm, struct maple_tree *mt, struct vm_area_struct *vma, struct vm_area_struct *prev, - unsigned long start, unsigned long end); + struct vm_area_struct *next, unsigned long start, + unsigned long end); /* description of effects of mapping type and prot in current implementation. * this is due to the limited x86 page protection hardware. The expected @@ -177,12 +178,10 @@ void unlink_file_vma(struct vm_area_struct *vma) } /* - * Close a vm structure and free it, returning the next. + * Close a vm structure and free it. */ -static struct vm_area_struct *remove_vma(struct vm_area_struct *vma) +static void remove_vma(struct vm_area_struct *vma) { - struct vm_area_struct *next = vma->vm_next; - might_sleep(); if (vma->vm_ops && vma->vm_ops->close) vma->vm_ops->close(vma); @@ -190,7 +189,6 @@ static struct vm_area_struct *remove_vma(struct vm_area_struct *vma) fput(vma->vm_file); mpol_put(vma_policy(vma)); vm_area_free(vma); - return next; } /* @@ -215,8 +213,7 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma, unsigned long newbrk, unsigned long oldbrk, struct list_head *uf); static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *brkvma, - unsigned long addr, unsigned long request, - unsigned long flags); + unsigned long addr, unsigned long request, unsigned long flags); SYSCALL_DEFINE1(brk, unsigned long, brk) { unsigned long newbrk, oldbrk, origbrk; @@ -285,7 +282,6 @@ SYSCALL_DEFINE1(brk, unsigned long, brk) * before calling do_brk_munmap(). */ mm->brk = brk; - mas.last = oldbrk - 1; ret = do_brk_munmap(&mas, brkvma, newbrk, oldbrk, &uf); if (ret == 1) { downgraded = true; @@ -340,42 +336,20 @@ extern void mt_dump(const struct maple_tree *mt); static void validate_mm_mt(struct mm_struct *mm) { struct maple_tree *mt = &mm->mm_mt; - struct vm_area_struct *vma_mt, *vma = mm->mmap; + struct vm_area_struct *vma_mt; MA_STATE(mas, mt, 0, 0); - mas_for_each(&mas, vma_mt, ULONG_MAX) { - if (xa_is_zero(vma_mt)) - continue; - - if (!vma) - break; - if ((vma != vma_mt) || - (vma->vm_start != vma_mt->vm_start) || - (vma->vm_end != vma_mt->vm_end) || - (vma->vm_start != mas.index) || - (vma->vm_end - 1 != mas.last)) { + mas_for_each(&mas, vma_mt, ULONG_MAX) { + if ((vma_mt->vm_start != mas.index) || + (vma_mt->vm_end - 1 != mas.last)) { pr_emerg("issue in %s\n", current->comm); dump_stack(); dump_vma(vma_mt); - pr_emerg("and vm_next\n"); - dump_vma(vma->vm_next); pr_emerg("mt piv: %px %lu - %lu\n", vma_mt, mas.index, mas.last); pr_emerg("mt vma: %px %lu - %lu\n", vma_mt, vma_mt->vm_start, vma_mt->vm_end); - if (vma->vm_prev) { - pr_emerg("ll prev: %px %lu - %lu\n", - vma->vm_prev, vma->vm_prev->vm_start, - vma->vm_prev->vm_end); - } - pr_emerg("ll vma: %px %lu - %lu\n", vma, - vma->vm_start, vma->vm_end); - if (vma->vm_next) { - pr_emerg("ll next: %px %lu - %lu\n", - vma->vm_next, vma->vm_next->vm_start, - vma->vm_next->vm_end); - } mt_dump(mas.tree); if (vma_mt->vm_end != mas.last + 1) { @@ -392,11 +366,7 @@ static void validate_mm_mt(struct mm_struct *mm) } VM_BUG_ON_MM(vma_mt->vm_start != mas.index, mm); } - VM_BUG_ON(vma != vma_mt); - vma = vma->vm_next; - } - VM_BUG_ON(vma); mt_validate(&mm->mm_mt); } @@ -404,12 +374,12 @@ static void validate_mm(struct mm_struct *mm) { int bug = 0; int i = 0; - unsigned long highest_address = 0; - struct vm_area_struct *vma = mm->mmap; + struct vm_area_struct *vma; + MA_STATE(mas, &mm->mm_mt, 0, 0); validate_mm_mt(mm); - while (vma) { + mas_for_each(&mas, vma, ULONG_MAX) { #ifdef CONFIG_DEBUG_VM_RB struct anon_vma *anon_vma = vma->anon_vma; struct anon_vma_chain *avc; @@ -421,18 +391,10 @@ static void validate_mm(struct mm_struct *mm) anon_vma_unlock_read(anon_vma); } #endif - - highest_address = vm_end_gap(vma); - vma = vma->vm_next; i++; } if (i != mm->map_count) { - pr_emerg("map_count %d vm_next %d\n", mm->map_count, i); - bug = 1; - } - if (highest_address != mm->highest_vm_end) { - pr_emerg("mm->highest_vm_end %lx, found %lx\n", - mm->highest_vm_end, highest_address); + pr_emerg("map_count %d mas_for_each %d\n", mm->map_count, i); bug = 1; } VM_BUG_ON_MM(bug, mm); @@ -492,29 +454,13 @@ bool range_has_overlap(struct mm_struct *mm, unsigned long start, struct vm_area_struct *existing; MA_STATE(mas, &mm->mm_mt, start, start); + rcu_read_lock(); existing = mas_find(&mas, end - 1); *pprev = mas_prev(&mas, 0); + rcu_read_unlock(); return existing ? true : false; } -/* - * __vma_next() - Get the next VMA. - * @mm: The mm_struct. - * @vma: The current vma. - * - * If @vma is NULL, return the first vma in the mm. - * - * Returns: The next VMA after @vma. - */ -static inline struct vm_area_struct *__vma_next(struct mm_struct *mm, - struct vm_area_struct *vma) -{ - if (!vma) - return mm->mmap; - - return vma->vm_next; -} - static unsigned long count_vma_pages_range(struct mm_struct *mm, unsigned long addr, unsigned long end) { @@ -599,8 +545,7 @@ static inline void vma_mas_szero(struct ma_state *mas, unsigned long start, mas_store_prealloc(mas, NULL); } -static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma, - struct vm_area_struct *prev) +static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma) { MA_STATE(mas, &mm->mm_mt, 0, 0); struct address_space *mapping = NULL; @@ -614,7 +559,6 @@ static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma, } vma_mas_store(vma, &mas); - __vma_link_list(mm, vma, prev); __vma_link_file(vma); if (mapping) @@ -625,22 +569,6 @@ static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma, return 0; } -/* - * Helper for vma_adjust() in the split_vma insert case: insert a vma into the - * mm's list and the mm tree. It has already been inserted into the interval tree. - */ -static void __insert_vm_struct(struct mm_struct *mm, struct ma_state *mas, - struct vm_area_struct *vma, unsigned long location) -{ - struct vm_area_struct *prev; - - mas_set(mas, location); - prev = mas_prev(mas, 0); - vma_mas_store(vma, mas); - __vma_link_list(mm, vma, prev); - mm->map_count++; -} - /* * vma_expand - Expand an existing VMA * @@ -719,15 +647,8 @@ inline int vma_expand(struct ma_state *mas, struct vm_area_struct *vma, } /* Expanding over the next vma */ - if (remove_next) { - /* Remove from mm linked list - also updates highest_vm_end */ - __vma_unlink_list(mm, next); - - if (file) - __remove_shared_vm_struct(next, file, mapping); - - } else if (!next) { - mm->highest_vm_end = vm_end_gap(vma); + if (remove_next && file) { + __remove_shared_vm_struct(next, file, mapping); } if (file) { @@ -785,7 +706,6 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, int remove_next = 0; MA_STATE(mas, &mm->mm_mt, 0, 0); struct vm_area_struct *exporter = NULL, *importer = NULL; - unsigned long ll_prev = vma->vm_start; /* linked list prev. */ if (next && !insert) { if (end >= next->vm_end) { @@ -831,7 +751,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, * next, if the vma overlaps with it. */ if (remove_next == 2 && !next->anon_vma) - exporter = next->vm_next; + exporter = find_vma(mm, next->vm_end); } else if (end > next->vm_start) { /* @@ -930,17 +850,14 @@ again: if (vma->vm_end > end) { if (!insert || (insert->vm_start != end)) { vma_mas_szero(&mas, end, vma->vm_end); + mas_reset(&mas); VM_WARN_ON(insert && insert->vm_end < vma->vm_end); - } else if (insert->vm_start == end) { - ll_prev = vma->vm_end; } } else { vma_changed = true; } vma->vm_end = end; - if (!next) - mm->highest_vm_end = vm_end_gap(vma); } if (vma_changed) @@ -960,17 +877,17 @@ again: flush_dcache_mmap_unlock(mapping); } - if (remove_next) { - __vma_unlink_list(mm, next); - if (file) - __remove_shared_vm_struct(next, file, mapping); + if (remove_next && file) { + __remove_shared_vm_struct(next, file, mapping); } else if (insert) { /* * split_vma has split insert from vma, and needs * us to insert it before dropping the locks * (it may either follow vma or precede it). */ - __insert_vm_struct(mm, &mas, insert, ll_prev); + mas_reset(&mas); + vma_mas_store(insert, &mas); + mm->map_count++; } if (anon_vma) { @@ -1009,8 +926,8 @@ again: /* * If "next" was removed and vma->vm_end was * expanded (up) over it, in turn - * "next->vm_prev->vm_end" changed and the - * "vma->vm_next" gap must be updated. + * "next->prev->vm_end" changed and the + * "vma->next" gap must be updated. */ next = next_next; } else { @@ -1031,34 +948,15 @@ again: remove_next = 1; end = next->vm_end; goto again; - } else if (!next) { - /* - * If remove_next == 2 we obviously can't - * reach this path. - * - * If remove_next == 3 we can't reach this - * path because pre-swap() next is always not - * NULL. pre-swap() "next" is not being - * removed and its next->vm_end is not altered - * (and furthermore "end" already matches - * next->vm_end in remove_next == 3). - * - * We reach this only in the remove_next == 1 - * case if the "next" vma that was removed was - * the highest vma of the mm. However in such - * case next->vm_end == "end" and the extended - * "vma" has vma->vm_end == next->vm_end so - * mm->highest_vm_end doesn't need any update - * in remove_next == 1 case. - */ - VM_WARN_ON(mm->highest_vm_end != vm_end_gap(vma)); } } - if (insert && file) + if (insert && file) { uprobe_mmap(insert); + } mas_destroy(&mas); validate_mm(mm); + return 0; } @@ -1218,10 +1116,10 @@ struct vm_area_struct *vma_merge(struct mm_struct *mm, if (vm_flags & VM_SPECIAL) return NULL; - next = __vma_next(mm, prev); + next = find_vma(mm, prev ? prev->vm_end : 0); area = next; if (area && area->vm_end == end) /* cases 6, 7, 8 */ - next = next->vm_next; + next = find_vma(mm, next->vm_end); /* verify some invariant that must be enforced by the caller */ VM_WARN_ON(prev && addr <= prev->vm_start); @@ -1355,18 +1253,24 @@ static struct anon_vma *reusable_anon_vma(struct vm_area_struct *old, struct vm_ */ struct anon_vma *find_mergeable_anon_vma(struct vm_area_struct *vma) { + MA_STATE(mas, &vma->vm_mm->mm_mt, vma->vm_end, vma->vm_end); struct anon_vma *anon_vma = NULL; + struct vm_area_struct *prev, *next; /* Try next first. */ - if (vma->vm_next) { - anon_vma = reusable_anon_vma(vma->vm_next, vma, vma->vm_next); + next = mas_walk(&mas); + if (next) { + anon_vma = reusable_anon_vma(next, vma, next); if (anon_vma) return anon_vma; } + prev = mas_prev(&mas, 0); + VM_BUG_ON_VMA(prev != vma, vma); + prev = mas_prev(&mas, 0); /* Try prev next. */ - if (vma->vm_prev) - anon_vma = reusable_anon_vma(vma->vm_prev, vma->vm_prev, vma); + if (prev) + anon_vma = reusable_anon_vma(prev, prev, vma); /* * We might reach here with anon_vma == NULL if we can't find @@ -2133,8 +2037,8 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) if (gap_addr < address || gap_addr > TASK_SIZE) gap_addr = TASK_SIZE; - next = vma->vm_next; - if (next && next->vm_start < gap_addr && vma_is_accessible(next)) { + next = find_vma_intersection(mm, vma->vm_end, gap_addr); + if (next && vma_is_accessible(next)) { if (!(next->vm_flags & VM_GROWSUP)) return -ENOMEM; /* Check that both stack segments have the same anon_vma? */ @@ -2185,8 +2089,6 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) /* Overwrite old entry in mtree. */ vma_mas_store(vma, &mas); anon_vma_interval_tree_post_update_vma(vma); - if (!vma->vm_next) - mm->highest_vm_end = vm_end_gap(vma); spin_unlock(&mm->page_table_lock); perf_event_mmap(vma); @@ -2206,16 +2108,16 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) int expand_downwards(struct vm_area_struct *vma, unsigned long address) { struct mm_struct *mm = vma->vm_mm; + MA_STATE(mas, &mm->mm_mt, vma->vm_start, vma->vm_start); struct vm_area_struct *prev; int error = 0; - MA_STATE(mas, &mm->mm_mt, 0, 0); address &= PAGE_MASK; if (address < mmap_min_addr) return -EPERM; /* Enforce stack_guard_gap */ - prev = vma->vm_prev; + prev = mas_prev(&mas, 0); /* Check that both stack segments have the same anon_vma? */ if (prev && !(prev->vm_flags & VM_GROWSDOWN) && vma_is_accessible(prev)) { @@ -2351,25 +2253,26 @@ find_extend_vma(struct mm_struct *mm, unsigned long addr) EXPORT_SYMBOL_GPL(find_extend_vma); /* - * Ok - we have the memory areas we should free on the vma list, - * so release them, and do the vma updates. + * Ok - we have the memory areas we should free on a maple tree so release them, + * and do the vma updates. * * Called with the mm semaphore held. */ -static void remove_vma_list(struct mm_struct *mm, struct vm_area_struct *vma) +static inline void remove_mt(struct mm_struct *mm, struct ma_state *mas) { unsigned long nr_accounted = 0; + struct vm_area_struct *vma; /* Update high watermark before we lower total_vm */ update_hiwater_vm(mm); - do { + mas_for_each(mas, vma, ULONG_MAX) { long nrpages = vma_pages(vma); if (vma->vm_flags & VM_ACCOUNT) nr_accounted += nrpages; vm_stat_account(mm, vma->vm_flags, -nrpages); - vma = remove_vma(vma); - } while (vma); + remove_vma(vma); + } vm_unacct_memory(nr_accounted); validate_mm(mm); } @@ -2379,18 +2282,18 @@ static void remove_vma_list(struct mm_struct *mm, struct vm_area_struct *vma) * * Called with the mm semaphore held. */ -static void unmap_region(struct mm_struct *mm, +static void unmap_region(struct mm_struct *mm, struct maple_tree *mt, struct vm_area_struct *vma, struct vm_area_struct *prev, + struct vm_area_struct *next, unsigned long start, unsigned long end) { - struct vm_area_struct *next = __vma_next(mm, prev); struct mmu_gather tlb; lru_add_drain(); tlb_gather_mmu(&tlb, mm); update_hiwater_rss(mm); - unmap_vmas(&tlb, vma, start, end); - free_pgtables(&tlb, vma, prev ? prev->vm_end : FIRST_USER_ADDRESS, + unmap_vmas(&tlb, mt, vma, start, end); + free_pgtables(&tlb, mt, vma, prev ? prev->vm_end : FIRST_USER_ADDRESS, next ? next->vm_start : USER_PGTABLES_CEILING); tlb_finish_mmu(&tlb); } @@ -2474,24 +2377,17 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma, return __split_vma(mm, vma, addr, new_below); } -static inline int -unlock_range(struct vm_area_struct *start, struct vm_area_struct **tail, - unsigned long limit) +static inline int munmap_sidetree(struct vm_area_struct *vma, + struct ma_state *mas_detach) { - struct mm_struct *mm = start->vm_mm; - struct vm_area_struct *tmp = start; - int count = 0; - - while (tmp && tmp->vm_start < limit) { - *tail = tmp; - count++; - if (tmp->vm_flags & VM_LOCKED) - mm->locked_vm -= vma_pages(tmp); + mas_set_range(mas_detach, vma->vm_start, vma->vm_end - 1); + if (mas_store_gfp(mas_detach, vma, GFP_KERNEL)) + return -ENOMEM; - tmp = tmp->vm_next; - } + if (vma->vm_flags & VM_LOCKED) + vma->vm_mm->locked_vm -= vma_pages(vma); - return count; + return 0; } /* @@ -2511,9 +2407,13 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma, struct mm_struct *mm, unsigned long start, unsigned long end, struct list_head *uf, bool downgrade) { - struct vm_area_struct *prev, *last; + struct vm_area_struct *prev, *next = NULL; + struct maple_tree mt_detach; + int count = 0; int error = -ENOMEM; - /* we have start < vma->vm_end */ + MA_STATE(mas_detach, &mt_detach, 0, 0); + mt_init_flags(&mt_detach, MT_FLAGS_LOCK_EXTERN); + mt_set_external_lock(&mt_detach, &mm->mmap_lock); if (mas_preallocate(mas, vma, GFP_KERNEL)) return -ENOMEM; @@ -2526,6 +2426,8 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma, * unmapped vm_area_struct will remain in use: so lower split_vma * places tmp vma above, and higher split_vma places tmp vma below. */ + + /* Does it split the first one? */ if (start > vma->vm_start) { /* @@ -2536,35 +2438,58 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma, if (end < vma->vm_end && mm->map_count >= sysctl_max_map_count) goto map_count_exceeded; + /* + * mas_pause() is not needed since mas->index needs to be set + * differently than vma->vm_end anyways. + */ error = __split_vma(mm, vma, start, 0); if (error) - goto split_failed; + goto start_split_failed; - prev = vma; - vma = __vma_next(mm, prev); - mas->index = start; - mas_reset(mas); - } else { - prev = vma->vm_prev; + mas_set(mas, start); + vma = mas_walk(mas); } - if (vma->vm_end >= end) - last = vma; - else - last = find_vma_intersection(mm, end - 1, end); + prev = mas_prev(mas, 0); + if (unlikely((!prev))) + mas_set(mas, start); - /* Does it split the last one? */ - if (last && end < last->vm_end) { - error = __split_vma(mm, last, end, 1); + /* + * Detach a range of VMAs from the mm. Using next as a temp variable as + * it is always overwritten. + */ + mas_for_each(mas, next, end - 1) { + /* Does it split the end? */ + if (next->vm_end > end) { + struct vm_area_struct *split; - if (error) - goto split_failed; + error = __split_vma(mm, next, end, 1); + if (error) + goto end_split_failed; + + mas_set(mas, end); + split = mas_prev(mas, 0); + if (munmap_sidetree(split, &mas_detach)) + goto munmap_sidetree_failed; - if (vma == last) - vma = __vma_next(mm, prev); - mas_reset(mas); + count++; + if (vma == next) + vma = split; + break; + } + if (munmap_sidetree(next, &mas_detach)) + goto munmap_sidetree_failed; + + count++; +#ifdef CONFIG_DEBUG_VM_MAPLE_TREE + BUG_ON(next->vm_start < start); + BUG_ON(next->vm_start > end); +#endif } + if (!next) + next = mas_next(mas, ULONG_MAX); + if (unlikely(uf)) { /* * If userfaultfd_unmap_prep returns an error the vmas @@ -2581,35 +2506,36 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma, goto userfaultfd_error; } - /* - * unlock any mlock()ed ranges before detaching vmas, count the number - * of VMAs to be dropped, and return the tail entry of the affected - * area. - */ - mm->map_count -= unlock_range(vma, &last, end); - /* Drop removed area from the tree */ + /* Point of no return */ + mas_set_range(mas, start, end - 1); +#if defined(CONFIG_DEBUG_VM_MAPLE_TREE) + /* Make sure no VMAs are about to be lost. */ + { + MA_STATE(test, &mt_detach, start, end - 1); + struct vm_area_struct *vma_mas, *vma_test; + int test_count = 0; + + rcu_read_lock(); + vma_test = mas_find(&test, end - 1); + mas_for_each(mas, vma_mas, end - 1) { + BUG_ON(vma_mas != vma_test); + test_count++; + vma_test = mas_next(&test, end - 1); + } + rcu_read_unlock(); + BUG_ON(count != test_count); + mas_set_range(mas, start, end - 1); + } +#endif mas_store_prealloc(mas, NULL); - - /* Detach vmas from the MM linked list */ - vma->vm_prev = NULL; - if (prev) - prev->vm_next = last->vm_next; - else - mm->mmap = last->vm_next; - - if (last->vm_next) { - last->vm_next->vm_prev = prev; - last->vm_next = NULL; - } else - mm->highest_vm_end = prev ? vm_end_gap(prev) : 0; - + mm->map_count -= count; /* * Do not downgrade mmap_lock if we are next to VM_GROWSDOWN or * VM_GROWSUP VMA. Such VMAs can change their size under * down_read(mmap_lock) and collide with the VMA we are about to unmap. */ if (downgrade) { - if (last && (last->vm_flags & VM_GROWSDOWN)) + if (next && (next->vm_flags & VM_GROWSDOWN)) downgrade = false; else if (prev && (prev->vm_flags & VM_GROWSUP)) downgrade = false; @@ -2617,18 +2543,22 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma, mmap_write_downgrade(mm); } - unmap_region(mm, vma, prev, start, end); - - /* Fix up all other VM information */ - remove_vma_list(mm, vma); + unmap_region(mm, &mt_detach, vma, prev, next, start, end); + /* Statistics and freeing VMAs */ + mas_set(&mas_detach, start); + remove_mt(mm, &mas_detach); + __mt_destroy(&mt_detach); validate_mm(mm); return downgrade ? 1 : 0; -map_count_exceeded: -split_failed: userfaultfd_error: +munmap_sidetree_failed: +end_split_failed: + __mt_destroy(&mt_detach); +start_split_failed: +map_count_exceeded: mas_destroy(mas); return error; } @@ -2863,7 +2793,6 @@ cannot_expand: i_mmap_lock_write(vma->vm_file->f_mapping); vma_mas_store(vma, &mas); - __vma_link_list(mm, vma, prev); mm->map_count++; if (vma->vm_file) { if (vma->vm_flags & VM_SHARED) @@ -2921,7 +2850,7 @@ unmap_and_free_vma: vma->vm_file = NULL; /* Undo any partial mapping done by a device driver. */ - unmap_region(mm, vma, prev, vma->vm_start, vma->vm_end); + unmap_region(mm, mas.tree, vma, prev, next, vma->vm_start, vma->vm_end); charged = 0; if (vm_flags & VM_SHARED) mapping_unmap_writable(file->f_mapping); @@ -3010,11 +2939,12 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size, goto out; if (start + size > vma->vm_end) { - struct vm_area_struct *next; + VMA_ITERATOR(vmi, mm, vma->vm_end); + struct vm_area_struct *next, *prev = vma; - for (next = vma->vm_next; next; next = next->vm_next) { + for_each_vma_range(vmi, next, start + size) { /* hole between vmas ? */ - if (next->vm_start != next->vm_prev->vm_end) + if (next->vm_start != prev->vm_end) goto out; if (next->vm_file != vma->vm_file) @@ -3023,8 +2953,7 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size, if (next->vm_flags != vma->vm_flags) goto out; - if (start + size <= next->vm_end) - break; + prev = next; } if (!next) @@ -3070,7 +2999,7 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma, struct list_head *uf) { struct mm_struct *mm = vma->vm_mm; - struct vm_area_struct unmap; + struct vm_area_struct unmap, *next; unsigned long unmap_pages; int ret; @@ -3086,6 +3015,7 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma, ret = userfaultfd_unmap_prep(mm, newbrk, oldbrk, uf); if (ret) return ret; + ret = 1; /* Change the oldbrk of vma to the newbrk of the munmap area */ @@ -3107,6 +3037,7 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma, vma_mas_remove(&unmap, mas); + vma->vm_end = newbrk; if (vma->anon_vma) { anon_vma_interval_tree_post_update_vma(vma); anon_vma_unlock_write(vma->anon_vma); @@ -3116,8 +3047,9 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma, if (vma->vm_flags & VM_LOCKED) mm->locked_vm -= unmap_pages; + next = mas_next(mas, ULONG_MAX); mmap_write_downgrade(mm); - unmap_region(mm, &unmap, vma, newbrk, oldbrk); + unmap_region(mm, mas->tree, &unmap, vma, next, newbrk, oldbrk); /* Statistics */ vm_stat_account(mm, vma->vm_flags, -unmap_pages); if (vma->vm_flags & VM_ACCOUNT) @@ -3141,11 +3073,9 @@ munmap_full_vma: * do some brk-specific accounting here. */ static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *vma, - unsigned long addr, unsigned long len, - unsigned long flags) + unsigned long addr, unsigned long len, unsigned long flags) { struct mm_struct *mm = current->mm; - struct vm_area_struct *prev = NULL; validate_mm_mt(mm); @@ -3189,7 +3119,6 @@ static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *vma, khugepaged_enter_vma(vma, flags); goto out; } - prev = vma; /* create a vma struct for an anonymous mapping */ vma = vm_area_alloc(mm); @@ -3207,12 +3136,6 @@ static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *vma, goto mas_store_fail; mm->map_count++; - - if (!prev) - prev = mas_prev(mas, 0); - - __vma_link_list(mm, vma, prev); - mm->map_count++; out: perf_event_mmap(vma); mm->total_vm += len >> PAGE_SHIFT; @@ -3220,7 +3143,7 @@ out: if (flags & VM_LOCKED) mm->locked_vm += (len >> PAGE_SHIFT); vma->vm_flags |= VM_SOFTDIRTY; - validate_mm_mt(mm); + validate_mm(mm); return 0; mas_store_fail: @@ -3294,6 +3217,8 @@ void exit_mmap(struct mm_struct *mm) struct mmu_gather tlb; struct vm_area_struct *vma; unsigned long nr_accounted = 0; + MA_STATE(mas, &mm->mm_mt, 0, 0); + int count = 0; /* mm's last user has gone, and its about to be pulled down */ mmu_notifier_release(mm); @@ -3318,7 +3243,7 @@ void exit_mmap(struct mm_struct *mm) mmap_write_lock(mm); arch_exit_mmap(mm); - vma = mm->mmap; + vma = mas_find(&mas, ULONG_MAX); if (!vma) { /* Can happen if dup_mmap() received an OOM */ mmap_write_unlock(mm); @@ -3329,22 +3254,29 @@ void exit_mmap(struct mm_struct *mm) flush_cache_mm(mm); tlb_gather_mmu_fullmm(&tlb, mm); /* update_hiwater_rss(mm) here? but nobody should be looking */ - /* Use -1 here to ensure all VMAs in the mm are unmapped */ - unmap_vmas(&tlb, vma, 0, -1); - free_pgtables(&tlb, vma, FIRST_USER_ADDRESS, USER_PGTABLES_CEILING); + /* Use ULONG_MAX here to ensure all VMAs in the mm are unmapped */ + unmap_vmas(&tlb, &mm->mm_mt, vma, 0, ULONG_MAX); + free_pgtables(&tlb, &mm->mm_mt, vma, FIRST_USER_ADDRESS, + USER_PGTABLES_CEILING); tlb_finish_mmu(&tlb); - /* Walk the list again, actually closing and freeing it. */ - while (vma) { + /* + * Walk the list again, actually closing and freeing it, with preemption + * enabled, without holding any MM locks besides the unreachable + * mmap_write_lock. + */ + do { if (vma->vm_flags & VM_ACCOUNT) nr_accounted += vma_pages(vma); - vma = remove_vma(vma); + remove_vma(vma); + count++; cond_resched(); - } + } while ((vma = mas_find(&mas, ULONG_MAX)) != NULL); + + BUG_ON(count != mm->map_count); trace_exit_mmap(mm); __mt_destroy(&mm->mm_mt); - mm->mmap = NULL; mmap_write_unlock(mm); vm_unacct_memory(nr_accounted); } @@ -3381,7 +3313,7 @@ int insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma) vma->vm_pgoff = vma->vm_start >> PAGE_SHIFT; } - if (vma_link(mm, vma, prev)) + if (vma_link(mm, vma)) return -ENOMEM; return 0; @@ -3411,7 +3343,8 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap, faulted_in_anon_vma = false; } - if (range_has_overlap(mm, addr, addr + len, &prev)) + new_vma = find_vma_prev(mm, addr, &prev); + if (new_vma->vm_start < addr + len) return NULL; /* should never get here */ new_vma = vma_merge(mm, prev, addr, addr + len, vma->vm_flags, @@ -3454,7 +3387,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap, get_file(new_vma->vm_file); if (new_vma->vm_ops && new_vma->vm_ops->open) new_vma->vm_ops->open(new_vma); - if (vma_link(mm, new_vma, prev)) + if (vma_link(mm, new_vma)) goto out_vma_link; *need_rmap_locks = false; } @@ -3759,12 +3692,13 @@ int mm_take_all_locks(struct mm_struct *mm) { struct vm_area_struct *vma; struct anon_vma_chain *avc; + MA_STATE(mas, &mm->mm_mt, 0, 0); mmap_assert_write_locked(mm); mutex_lock(&mm_all_locks_mutex); - for (vma = mm->mmap; vma; vma = vma->vm_next) { + mas_for_each(&mas, vma, ULONG_MAX) { if (signal_pending(current)) goto out_unlock; if (vma->vm_file && vma->vm_file->f_mapping && @@ -3772,7 +3706,8 @@ int mm_take_all_locks(struct mm_struct *mm) vm_lock_mapping(mm, vma->vm_file->f_mapping); } - for (vma = mm->mmap; vma; vma = vma->vm_next) { + mas_set(&mas, 0); + mas_for_each(&mas, vma, ULONG_MAX) { if (signal_pending(current)) goto out_unlock; if (vma->vm_file && vma->vm_file->f_mapping && @@ -3780,7 +3715,8 @@ int mm_take_all_locks(struct mm_struct *mm) vm_lock_mapping(mm, vma->vm_file->f_mapping); } - for (vma = mm->mmap; vma; vma = vma->vm_next) { + mas_set(&mas, 0); + mas_for_each(&mas, vma, ULONG_MAX) { if (signal_pending(current)) goto out_unlock; if (vma->anon_vma) @@ -3839,11 +3775,12 @@ void mm_drop_all_locks(struct mm_struct *mm) { struct vm_area_struct *vma; struct anon_vma_chain *avc; + MA_STATE(mas, &mm->mm_mt, 0, 0); mmap_assert_write_locked(mm); BUG_ON(!mutex_is_locked(&mm_all_locks_mutex)); - for (vma = mm->mmap; vma; vma = vma->vm_next) { + mas_for_each(&mas, vma, ULONG_MAX) { if (vma->anon_vma) list_for_each_entry(avc, &vma->anon_vma_chain, same_vma) vm_unlock_anon_vma(avc->anon_vma); diff --git a/mm/nommu.c b/mm/nommu.c index f6b187090d95..7b3fad611895 100644 --- a/mm/nommu.c +++ b/mm/nommu.c @@ -584,17 +584,12 @@ static void setup_vma_to_mm(struct vm_area_struct *vma, struct mm_struct *mm) static void mas_add_vma_to_mm(struct ma_state *mas, struct mm_struct *mm, struct vm_area_struct *vma) { - struct vm_area_struct *prev; - BUG_ON(!vma->vm_region); setup_vma_to_mm(vma, mm); - prev = mas_prev(mas, 0); - mas_reset(mas); /* add the VMA to the tree */ vma_mas_store(vma, mas); - __vma_link_list(mm, vma, prev); } /* @@ -647,7 +642,6 @@ static int delete_vma_from_mm(struct vm_area_struct *vma) /* remove from the MM's tree and list */ vma_mas_remove(vma, &mas); - __vma_unlink_list(vma->vm_mm, vma); return 0; } diff --git a/mm/util.c b/mm/util.c index 812365cbdd19..c831c9f8766b 100644 --- a/mm/util.c +++ b/mm/util.c @@ -272,46 +272,6 @@ void *memdup_user_nul(const void __user *src, size_t len) } EXPORT_SYMBOL(memdup_user_nul); -void __vma_link_list(struct mm_struct *mm, struct vm_area_struct *vma, - struct vm_area_struct *prev) -{ - struct vm_area_struct *next; - - vma->vm_prev = prev; - if (prev) { - next = prev->vm_next; - prev->vm_next = vma; - } else { - next = mm->mmap; - mm->mmap = vma; - } - vma->vm_next = next; - if (next) - next->vm_prev = vma; - else - mm->highest_vm_end = vm_end_gap(vma); -} - -void __vma_unlink_list(struct mm_struct *mm, struct vm_area_struct *vma) -{ - struct vm_area_struct *prev, *next; - - next = vma->vm_next; - prev = vma->vm_prev; - if (prev) - prev->vm_next = next; - else - mm->mmap = next; - if (next) { - next->vm_prev = prev; - } else { - if (prev) - mm->highest_vm_end = vm_end_gap(prev); - else - mm->highest_vm_end = 0; - } -} - /* Check if the vma is being used as a stack by this task */ int vma_is_stack_for_current(struct vm_area_struct *vma) { -- cgit v1.2.3 From 55e13fefa6110926e15b195644190bb997854bee Mon Sep 17 00:00:00 2001 From: Miaohe Lin Date: Mon, 30 May 2022 19:30:15 +0800 Subject: mm/migration: return errno when isolate_huge_page failed We might fail to isolate huge page due to e.g. the page is under migration which cleared HPageMigratable. We should return errno in this case rather than always return 1 which could confuse the user, i.e. the caller might think all of the memory is migrated while the hugetlb page is left behind. We make the prototype of isolate_huge_page consistent with isolate_lru_page as suggested by Huang Ying and rename isolate_huge_page to isolate_hugetlb as suggested by Muchun to improve the readability. Link: https://lkml.kernel.org/r/20220530113016.16663-4-linmiaohe@huawei.com Fixes: e8db67eb0ded ("mm: migrate: move_pages() supports thp migration") Signed-off-by: Miaohe Lin Suggested-by: Huang Ying Reported-by: kernel test robot (build error) Cc: Alistair Popple Cc: Christoph Hellwig Cc: Christoph Lameter Cc: David Hildenbrand Cc: David Howells Cc: Mike Kravetz Cc: Muchun Song Cc: Oscar Salvador Cc: Peter Xu Signed-off-by: Andrew Morton --- include/linux/hugetlb.h | 6 +++--- mm/gup.c | 2 +- mm/hugetlb.c | 11 +++++------ mm/memory-failure.c | 2 +- mm/memory_hotplug.c | 2 +- mm/mempolicy.c | 2 +- mm/migrate.c | 7 ++++--- 7 files changed, 16 insertions(+), 16 deletions(-) (limited to 'include/linux') diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index e4cff27d1198..756b66ff025e 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -170,7 +170,7 @@ bool hugetlb_reserve_pages(struct inode *inode, long from, long to, vm_flags_t vm_flags); long hugetlb_unreserve_pages(struct inode *inode, long start, long end, long freed); -bool isolate_huge_page(struct page *page, struct list_head *list); +int isolate_hugetlb(struct page *page, struct list_head *list); int get_hwpoison_huge_page(struct page *page, bool *hugetlb); int get_huge_page_for_hwpoison(unsigned long pfn, int flags); void putback_active_hugepage(struct page *page); @@ -376,9 +376,9 @@ static inline pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr, return NULL; } -static inline bool isolate_huge_page(struct page *page, struct list_head *list) +static inline int isolate_hugetlb(struct page *page, struct list_head *list) { - return false; + return -EBUSY; } static inline int get_hwpoison_huge_page(struct page *page, bool *hugetlb) diff --git a/mm/gup.c b/mm/gup.c index f0e5d68ea932..4ec77f68c6c1 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -1931,7 +1931,7 @@ static long check_and_migrate_movable_pages(unsigned long nr_pages, * Try to move out any movable page before pinning the range. */ if (folio_test_hugetlb(folio)) { - if (!isolate_huge_page(&folio->page, + if (isolate_hugetlb(&folio->page, &movable_page_list)) isolation_error_count++; continue; diff --git a/mm/hugetlb.c b/mm/hugetlb.c index b36a4ef87a2e..dd9a46ccb79c 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -2766,8 +2766,7 @@ retry: * Fail with -EBUSY if not possible. */ spin_unlock_irq(&hugetlb_lock); - if (!isolate_huge_page(old_page, list)) - ret = -EBUSY; + ret = isolate_hugetlb(old_page, list); spin_lock_irq(&hugetlb_lock); goto free_new; } else if (!HPageFreed(old_page)) { @@ -2843,7 +2842,7 @@ int isolate_or_dissolve_huge_page(struct page *page, struct list_head *list) if (hstate_is_gigantic(h)) return -ENOMEM; - if (page_count(head) && isolate_huge_page(head, list)) + if (page_count(head) && !isolate_hugetlb(head, list)) ret = 0; else if (!page_count(head)) ret = alloc_and_dissolve_huge_page(h, head, list); @@ -6960,15 +6959,15 @@ follow_huge_pgd(struct mm_struct *mm, unsigned long address, pgd_t *pgd, int fla return pte_page(*(pte_t *)pgd) + ((address & ~PGDIR_MASK) >> PAGE_SHIFT); } -bool isolate_huge_page(struct page *page, struct list_head *list) +int isolate_hugetlb(struct page *page, struct list_head *list) { - bool ret = true; + int ret = 0; spin_lock_irq(&hugetlb_lock); if (!PageHeadHuge(page) || !HPageMigratable(page) || !get_page_unless_zero(page)) { - ret = false; + ret = -EBUSY; goto unlock; } ClearHPageMigratable(page); diff --git a/mm/memory-failure.c b/mm/memory-failure.c index da39ec8afca8..845369f839e1 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -2178,7 +2178,7 @@ static bool isolate_page(struct page *page, struct list_head *pagelist) bool lru = PageLRU(page); if (PageHuge(page)) { - isolated = isolate_huge_page(page, pagelist); + isolated = !isolate_hugetlb(page, pagelist); } else { if (lru) isolated = !isolate_lru_page(page); diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c index 1f1a730c4499..84990a14d51a 100644 --- a/mm/memory_hotplug.c +++ b/mm/memory_hotplug.c @@ -1641,7 +1641,7 @@ do_migrate_range(unsigned long start_pfn, unsigned long end_pfn) if (PageHuge(page)) { pfn = page_to_pfn(head) + compound_nr(head) - 1; - isolate_huge_page(head, &source); + isolate_hugetlb(head, &source); continue; } else if (PageTransHuge(page)) pfn = page_to_pfn(head) + thp_nr_pages(page) - 1; diff --git a/mm/mempolicy.c b/mm/mempolicy.c index d5993b47bdb8..201b043e96f1 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -603,7 +603,7 @@ static int queue_pages_hugetlb(pte_t *pte, unsigned long hmask, /* With MPOL_MF_MOVE, we migrate only unshared hugepage. */ if (flags & (MPOL_MF_MOVE_ALL) || (flags & MPOL_MF_MOVE && page_mapcount(page) == 1)) { - if (!isolate_huge_page(page, qp->pagelist) && + if (isolate_hugetlb(page, qp->pagelist) && (flags & MPOL_MF_STRICT)) /* * Failed to isolate page but allow migrating pages diff --git a/mm/migrate.c b/mm/migrate.c index c83b3ae2e285..1d036dec1328 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -133,7 +133,7 @@ static void putback_movable_page(struct page *page) * * This function shall be used whenever the isolated pageset has been * built from lru, balloon, hugetlbfs page. See isolate_migratepages_range() - * and isolate_huge_page(). + * and isolate_hugetlb(). */ void putback_movable_pages(struct list_head *l) { @@ -1628,8 +1628,9 @@ static int add_page_for_migration(struct mm_struct *mm, unsigned long addr, if (PageHuge(page)) { if (PageHead(page)) { - isolate_huge_page(page, pagelist); - err = 1; + err = isolate_hugetlb(page, pagelist); + if (!err) + err = 1; } } else { struct page *head; -- cgit v1.2.3 From 8b4402b0bb09e8d162c153ab2c104a8627f04a42 Mon Sep 17 00:00:00 2001 From: Miaohe Lin Date: Mon, 30 May 2022 19:30:16 +0800 Subject: mm/migration: fix potential pte_unmap on an not mapped pte __migration_entry_wait and migration_entry_wait_on_locked assume pte is always mapped from caller. But this is not the case when it's called from migration_entry_wait_huge and follow_huge_pmd. Add a hugetlbfs variant that calls hugetlb_migration_entry_wait(ptep == NULL) to fix this issue. Link: https://lkml.kernel.org/r/20220530113016.16663-5-linmiaohe@huawei.com Fixes: 30dad30922cc ("mm: migration: add migrate_entry_wait_huge()") Signed-off-by: Miaohe Lin Suggested-by: David Hildenbrand Reviewed-by: David Hildenbrand Cc: Alistair Popple Cc: Christoph Hellwig Cc: Christoph Lameter Cc: David Howells Cc: Huang Ying Cc: kernel test robot Cc: Mike Kravetz Cc: Muchun Song Cc: Oscar Salvador Cc: Peter Xu Signed-off-by: Andrew Morton --- include/linux/swapops.h | 12 ++++++++---- mm/hugetlb.c | 4 ++-- mm/migrate.c | 23 +++++++++++++++++++---- 3 files changed, 29 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/include/linux/swapops.h b/include/linux/swapops.h index f24775b41880..bb7afd03a324 100644 --- a/include/linux/swapops.h +++ b/include/linux/swapops.h @@ -244,8 +244,10 @@ extern void __migration_entry_wait(struct mm_struct *mm, pte_t *ptep, spinlock_t *ptl); extern void migration_entry_wait(struct mm_struct *mm, pmd_t *pmd, unsigned long address); -extern void migration_entry_wait_huge(struct vm_area_struct *vma, - struct mm_struct *mm, pte_t *pte); +#ifdef CONFIG_HUGETLB_PAGE +extern void __migration_entry_wait_huge(pte_t *ptep, spinlock_t *ptl); +extern void migration_entry_wait_huge(struct vm_area_struct *vma, pte_t *pte); +#endif #else static inline swp_entry_t make_readable_migration_entry(pgoff_t offset) { @@ -271,8 +273,10 @@ static inline void __migration_entry_wait(struct mm_struct *mm, pte_t *ptep, spinlock_t *ptl) { } static inline void migration_entry_wait(struct mm_struct *mm, pmd_t *pmd, unsigned long address) { } -static inline void migration_entry_wait_huge(struct vm_area_struct *vma, - struct mm_struct *mm, pte_t *pte) { } +#ifdef CONFIG_HUGETLB_PAGE +static inline void __migration_entry_wait_huge(pte_t *ptep, spinlock_t *ptl) { } +static inline void migration_entry_wait_huge(struct vm_area_struct *vma, pte_t *pte) { } +#endif static inline int is_writable_migration_entry(swp_entry_t entry) { return 0; diff --git a/mm/hugetlb.c b/mm/hugetlb.c index dd9a46ccb79c..ed202d29ca46 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -5702,7 +5702,7 @@ vm_fault_t hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma, */ entry = huge_ptep_get(ptep); if (unlikely(is_hugetlb_entry_migration(entry))) { - migration_entry_wait_huge(vma, mm, ptep); + migration_entry_wait_huge(vma, ptep); return 0; } else if (unlikely(is_hugetlb_entry_hwpoisoned(entry))) return VM_FAULT_HWPOISON_LARGE | @@ -6927,7 +6927,7 @@ retry: } else { if (is_hugetlb_entry_migration(pte)) { spin_unlock(ptl); - __migration_entry_wait(mm, (pte_t *)pmd, ptl); + __migration_entry_wait_huge((pte_t *)pmd, ptl); goto retry; } /* diff --git a/mm/migrate.c b/mm/migrate.c index 1d036dec1328..7934eebf1689 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -315,13 +315,28 @@ void migration_entry_wait(struct mm_struct *mm, pmd_t *pmd, __migration_entry_wait(mm, ptep, ptl); } -void migration_entry_wait_huge(struct vm_area_struct *vma, - struct mm_struct *mm, pte_t *pte) +#ifdef CONFIG_HUGETLB_PAGE +void __migration_entry_wait_huge(pte_t *ptep, spinlock_t *ptl) { - spinlock_t *ptl = huge_pte_lockptr(hstate_vma(vma), mm, pte); - __migration_entry_wait(mm, pte, ptl); + pte_t pte; + + spin_lock(ptl); + pte = huge_ptep_get(ptep); + + if (unlikely(!is_hugetlb_entry_migration(pte))) + spin_unlock(ptl); + else + migration_entry_wait_on_locked(pte_to_swp_entry(pte), NULL, ptl); } +void migration_entry_wait_huge(struct vm_area_struct *vma, pte_t *pte) +{ + spinlock_t *ptl = huge_pte_lockptr(hstate_vma(vma), vma->vm_mm, pte); + + __migration_entry_wait_huge(pte, ptl); +} +#endif + #ifdef CONFIG_ARCH_ENABLE_THP_MIGRATION void pmd_migration_entry_wait(struct mm_struct *mm, pmd_t *pmd) { -- cgit v1.2.3 From 0ddbc2f1b0b7962352dce944097869f3abfde6f0 Mon Sep 17 00:00:00 2001 From: Alex Sierra Date: Tue, 31 May 2022 15:00:29 -0500 Subject: mm: add zone device coherent type memory support Patch series "Add MEMORY_DEVICE_COHERENT for coherent device memory mapping", v5. This patch series introduces MEMORY_DEVICE_COHERENT, a type of memory owned by a device that can be mapped into CPU page tables like MEMORY_DEVICE_GENERIC and can also be migrated like MEMORY_DEVICE_PRIVATE. This patch series is mostly self-contained except for a few places where it needs to update other subsystems to handle the new memory type. System stability and performance are not affected according to our ongoing testing, including xfstests. How it works: The system BIOS advertises the GPU device memory (aka VRAM) as SPM (special purpose memory) in the UEFI system address map. The amdgpu driver registers the memory with devmap as MEMORY_DEVICE_COHERENT using devm_memremap_pages. The initial user for this hardware page migration capability is the Frontier supercomputer project. This functionality is not AMD-specific. We expect other GPU vendors to find this functionality useful, and possibly other hardware types in the future. Our test nodes in the lab are similar to the Frontier configuration, with .5 TB of system memory plus 256 GB of device memory split across 4 GPUs, all in a single coherent address space. Page migration is expected to improve application efficiency significantly. We will report empirical results as they become available. Coherent device type pages at gup are now migrated back to system memory if they are being pinned long-term (FOLL_LONGTERM). The reason is, that long-term pinning would interfere with the device memory manager owning the device-coherent pages (e.g. evictions in TTM). These series incorporate Alistair Popple patches to do this migration from pin_user_pages() calls. hmm_gup_test has been added to hmm-test to test different get user pages calls. This series includes handling of device-managed anonymous pages returned by vm_normal_pages. Although they behave like normal pages for purposes of mapping in CPU page tables and for COW, they do not support LRU lists, NUMA migration or THP. We also introduce a FOLL_LRU flag that adds the same behaviour to follow_page and related APIs, to allow callers to specify that they expect to put pages on an LRU list. This patch (od 13): Device memory that is cache coherent from device and CPU point of view. This is used on platforms that have an advanced system bus (like CAPI or CXL). Any page of a process can be migrated to such memory. However, no one should be allowed to pin such memory so that it can always be evicted. [hch@lst.de: rebased ontop of the refcount changes, removed is_dev_private_or_coherent_page] Link: https://lkml.kernel.org/r/20220531200041.24904-1-alex.sierra@amd.com Link: https://lkml.kernel.org/r/20220531200041.24904-2-alex.sierra@amd.com Signed-off-by: Alex Sierra Signed-off-by: Christoph Hellwig Acked-by: Felix Kuehling Reviewed-by: Alistair Popple Cc: Jason Gunthorpe Cc: David Hildenbrand Cc: Ralph Campbell Cc: Jerome Glisse Cc: Matthew Wilcox Signed-off-by: Andrew Morton --- include/linux/memremap.h | 19 +++++++++++++++++++ mm/memcontrol.c | 7 ++++--- mm/memory-failure.c | 8 ++++++-- mm/memremap.c | 10 ++++++++++ mm/migrate_device.c | 16 +++++++--------- mm/rmap.c | 5 +++-- 6 files changed, 49 insertions(+), 16 deletions(-) (limited to 'include/linux') diff --git a/include/linux/memremap.h b/include/linux/memremap.h index 9f5ee49482de..738b41fa7c8f 100644 --- a/include/linux/memremap.h +++ b/include/linux/memremap.h @@ -41,6 +41,13 @@ struct vmem_altmap { * A more complete discussion of unaddressable memory may be found in * include/linux/hmm.h and Documentation/mm/hmm.rst. * + * MEMORY_DEVICE_COHERENT: + * Device memory that is cache coherent from device and CPU point of view. This + * is used on platforms that have an advanced system bus (like CAPI or CXL). A + * driver can hotplug the device memory using ZONE_DEVICE and with that memory + * type. Any page of a process can be migrated to such memory. However no one + * should be allowed to pin such memory so that it can always be evicted. + * * MEMORY_DEVICE_FS_DAX: * Host memory that has similar access semantics as System RAM i.e. DMA * coherent and supports page pinning. In support of coordinating page @@ -61,6 +68,7 @@ struct vmem_altmap { enum memory_type { /* 0 is reserved to catch uninitialized type fields */ MEMORY_DEVICE_PRIVATE = 1, + MEMORY_DEVICE_COHERENT, MEMORY_DEVICE_FS_DAX, MEMORY_DEVICE_GENERIC, MEMORY_DEVICE_PCI_P2PDMA, @@ -143,6 +151,17 @@ static inline bool folio_is_device_private(const struct folio *folio) return is_device_private_page(&folio->page); } +static inline bool is_device_coherent_page(const struct page *page) +{ + return is_zone_device_page(page) && + page->pgmap->type == MEMORY_DEVICE_COHERENT; +} + +static inline bool folio_is_device_coherent(const struct folio *folio) +{ + return is_device_coherent_page(&folio->page); +} + static inline bool is_pci_p2pdma_page(const struct page *page) { return IS_ENABLED(CONFIG_PCI_P2PDMA) && diff --git a/mm/memcontrol.c b/mm/memcontrol.c index d8e1b9ff72e6..370ef639110f 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -5693,8 +5693,8 @@ out: * 2(MC_TARGET_SWAP): if the swap entry corresponding to this pte is a * target for charge migration. if @target is not NULL, the entry is stored * in target->ent. - * 3(MC_TARGET_DEVICE): like MC_TARGET_PAGE but page is MEMORY_DEVICE_PRIVATE - * (so ZONE_DEVICE page and thus not on the lru). + * 3(MC_TARGET_DEVICE): like MC_TARGET_PAGE but page is device memory and + * thus not on the lru. * For now we such page is charge like a regular page would be as for all * intent and purposes it is just special memory taking the place of a * regular page. @@ -5732,7 +5732,8 @@ static enum mc_target_type get_mctgt_type(struct vm_area_struct *vma, */ if (page_memcg(page) == mc.from) { ret = MC_TARGET_PAGE; - if (is_device_private_page(page)) + if (is_device_private_page(page) || + is_device_coherent_page(page)) ret = MC_TARGET_DEVICE; if (target) target->page = page; diff --git a/mm/memory-failure.c b/mm/memory-failure.c index 845369f839e1..16cbaa7b92ad 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -1685,12 +1685,16 @@ static int memory_failure_dev_pagemap(unsigned long pfn, int flags, goto unlock; } - if (pgmap->type == MEMORY_DEVICE_PRIVATE) { + switch (pgmap->type) { + case MEMORY_DEVICE_PRIVATE: + case MEMORY_DEVICE_COHERENT: /* - * TODO: Handle HMM pages which may need coordination + * TODO: Handle device pages which may need coordination * with device-side memory. */ goto unlock; + default: + break; } /* diff --git a/mm/memremap.c b/mm/memremap.c index 8b5c8fd4ea8e..f0955785150f 100644 --- a/mm/memremap.c +++ b/mm/memremap.c @@ -315,6 +315,16 @@ void *memremap_pages(struct dev_pagemap *pgmap, int nid) return ERR_PTR(-EINVAL); } break; + case MEMORY_DEVICE_COHERENT: + if (!pgmap->ops->page_free) { + WARN(1, "Missing page_free method\n"); + return ERR_PTR(-EINVAL); + } + if (!pgmap->owner) { + WARN(1, "Missing owner\n"); + return ERR_PTR(-EINVAL); + } + break; case MEMORY_DEVICE_FS_DAX: if (IS_ENABLED(CONFIG_FS_DAX_LIMITED)) { WARN(1, "File system DAX not supported\n"); diff --git a/mm/migrate_device.c b/mm/migrate_device.c index 5052093d0262..a4847ad65da3 100644 --- a/mm/migrate_device.c +++ b/mm/migrate_device.c @@ -518,7 +518,7 @@ EXPORT_SYMBOL(migrate_vma_setup); * handle_pte_fault() * do_anonymous_page() * to map in an anonymous zero page but the struct page will be a ZONE_DEVICE - * private page. + * private or coherent page. */ static void migrate_vma_insert_page(struct migrate_vma *migrate, unsigned long addr, @@ -594,11 +594,8 @@ static void migrate_vma_insert_page(struct migrate_vma *migrate, page_to_pfn(page)); entry = swp_entry_to_pte(swp_entry); } else { - /* - * For now we only support migrating to un-addressable device - * memory. - */ - if (is_zone_device_page(page)) { + if (is_zone_device_page(page) && + !is_device_coherent_page(page)) { pr_warn_once("Unsupported ZONE_DEVICE page type.\n"); goto abort; } @@ -701,10 +698,11 @@ void migrate_vma_pages(struct migrate_vma *migrate) mapping = page_mapping(page); - if (is_device_private_page(newpage)) { + if (is_device_private_page(newpage) || + is_device_coherent_page(newpage)) { /* - * For now only support private anonymous when migrating - * to un-addressable device memory. + * For now only support anonymous memory migrating to + * device private or coherent memory. */ if (mapping) { migrate->src[i] &= ~MIGRATE_PFN_MIGRATE; diff --git a/mm/rmap.c b/mm/rmap.c index f22d7578ac02..6ec63ba4a86e 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -1972,7 +1972,7 @@ static bool try_to_migrate_one(struct folio *folio, struct vm_area_struct *vma, /* Update high watermark before we lower rss */ update_hiwater_rss(mm); - if (folio_is_zone_device(folio)) { + if (folio_is_device_private(folio)) { unsigned long pfn = folio_pfn(folio); swp_entry_t entry; pte_t swp_pte; @@ -2138,7 +2138,8 @@ void try_to_migrate(struct folio *folio, enum ttu_flags flags) TTU_SYNC))) return; - if (folio_is_zone_device(folio) && !folio_is_device_private(folio)) + if (folio_is_zone_device(folio) && + (!folio_is_device_private(folio) && !folio_is_device_coherent(folio))) return; /* -- cgit v1.2.3 From 286ef18e175138928d06f1226b1533239e8b0070 Mon Sep 17 00:00:00 2001 From: Alex Sierra Date: Tue, 31 May 2022 15:00:30 -0500 Subject: mm: handling Non-LRU pages returned by vm_normal_pages With DEVICE_COHERENT, we'll soon have vm_normal_pages() return device-managed anonymous pages that are not LRU pages. Although they behave like normal pages for purposes of mapping in CPU page, and for COW. They do not support LRU lists, NUMA migration or THP. We also introduced a FOLL_LRU flag that adds the same behaviour to follow_page and related APIs, to allow callers to specify that they expect to put pages on an LRU list. Link: https://lkml.kernel.org/r/20220531200041.24904-3-alex.sierra@amd.com Signed-off-by: Alex Sierra Acked-by: Felix Kuehling Reviewed-by: Alistair Popple Cc: Christoph Hellwig Cc: David Hildenbrand Cc: Jason Gunthorpe Cc: Jerome Glisse Cc: Matthew Wilcox Cc: Ralph Campbell Signed-off-by: Andrew Morton --- fs/proc/task_mmu.c | 2 +- include/linux/mm.h | 3 ++- mm/gup.c | 6 +++++- mm/huge_memory.c | 2 +- mm/khugepaged.c | 9 ++++++--- mm/ksm.c | 6 +++--- mm/madvise.c | 4 ++-- mm/memory.c | 9 ++++++++- mm/mempolicy.c | 2 +- mm/migrate.c | 4 ++-- mm/mlock.c | 2 +- mm/mprotect.c | 2 +- 12 files changed, 33 insertions(+), 18 deletions(-) (limited to 'include/linux') diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index f9c9abb50bb7..37ccb5c9f4f8 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -1800,7 +1800,7 @@ static struct page *can_gather_numa_stats(pte_t pte, struct vm_area_struct *vma, return NULL; page = vm_normal_page(vma, addr, pte); - if (!page) + if (!page || is_zone_device_page(page)) return NULL; if (PageReserved(page)) diff --git a/include/linux/mm.h b/include/linux/mm.h index ffde0619710f..445488e72bf0 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -601,7 +601,7 @@ struct vm_operations_struct { #endif /* * Called by vm_normal_page() for special PTEs to find the - * page for @addr. This is useful if the default behavior + * page for @addr. This is useful if the default behavior * (using pte_page()) would not find the correct page. */ struct page *(*find_special_page)(struct vm_area_struct *vma, @@ -2949,6 +2949,7 @@ struct page *follow_page(struct vm_area_struct *vma, unsigned long address, #define FOLL_NUMA 0x200 /* force NUMA hinting page fault */ #define FOLL_MIGRATION 0x400 /* wait for page to replace migration entry */ #define FOLL_TRIED 0x800 /* a retry, previous pass started an IO */ +#define FOLL_LRU 0x1000 /* return only LRU (anon or page cache) */ #define FOLL_REMOTE 0x2000 /* we are working on non-current tsk/mm */ #define FOLL_COW 0x4000 /* internal GUP flag */ #define FOLL_ANON 0x8000 /* don't do file mappings */ diff --git a/mm/gup.c b/mm/gup.c index 4ec77f68c6c1..67cc836eb1b5 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -532,7 +532,11 @@ retry: } page = vm_normal_page(vma, address, pte); - if (!page && pte_devmap(pte) && (flags & (FOLL_GET | FOLL_PIN))) { + if ((flags & FOLL_LRU) && ((page && is_zone_device_page(page)) || + (!page && pte_devmap(pte)))) { + page = ERR_PTR(-EEXIST); + goto out; + } else if (!page && pte_devmap(pte) && (flags & (FOLL_GET | FOLL_PIN))) { /* * Only return device mapping pages in the FOLL_GET or FOLL_PIN * case since they are only valid while holding the pgmap diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 906bc4a36df4..f71eeeda9533 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -2906,7 +2906,7 @@ static int split_huge_pages_pid(int pid, unsigned long vaddr_start, } /* FOLL_DUMP to ignore special (like zero) pages */ - page = follow_page(vma, addr, FOLL_GET | FOLL_DUMP); + page = follow_page(vma, addr, FOLL_GET | FOLL_DUMP | FOLL_LRU); if (IS_ERR(page)) continue; diff --git a/mm/khugepaged.c b/mm/khugepaged.c index 637bfecd6bf5..476d79360101 100644 --- a/mm/khugepaged.c +++ b/mm/khugepaged.c @@ -618,7 +618,7 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma, goto out; } page = vm_normal_page(vma, address, pteval); - if (unlikely(!page)) { + if (unlikely(!page) || unlikely(is_zone_device_page(page))) { result = SCAN_PAGE_NULL; goto out; } @@ -1267,7 +1267,7 @@ static int khugepaged_scan_pmd(struct mm_struct *mm, writable = true; page = vm_normal_page(vma, _address, pteval); - if (unlikely(!page)) { + if (unlikely(!page) || unlikely(is_zone_device_page(page))) { result = SCAN_PAGE_NULL; goto out_unmap; } @@ -1479,7 +1479,8 @@ void collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr) goto abort; page = vm_normal_page(vma, addr, *pte); - + if (WARN_ON_ONCE(page && is_zone_device_page(page))) + page = NULL; /* * Note that uprobe, debugger, or MAP_PRIVATE may change the * page table, but the new page will not be a subpage of hpage. @@ -1497,6 +1498,8 @@ void collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr) if (pte_none(*pte)) continue; page = vm_normal_page(vma, addr, *pte); + if (WARN_ON_ONCE(page && is_zone_device_page(page))) + goto abort; page_remove_rmap(page, vma, false); } diff --git a/mm/ksm.c b/mm/ksm.c index 73cbe2a628ed..6d0b779be010 100644 --- a/mm/ksm.c +++ b/mm/ksm.c @@ -474,7 +474,7 @@ static int break_ksm(struct vm_area_struct *vma, unsigned long addr) do { cond_resched(); page = follow_page(vma, addr, - FOLL_GET | FOLL_MIGRATION | FOLL_REMOTE); + FOLL_GET | FOLL_MIGRATION | FOLL_REMOTE | FOLL_LRU); if (IS_ERR_OR_NULL(page)) break; if (PageKsm(page)) @@ -559,7 +559,7 @@ static struct page *get_mergeable_page(struct rmap_item *rmap_item) if (!vma) goto out; - page = follow_page(vma, addr, FOLL_GET); + page = follow_page(vma, addr, FOLL_GET | FOLL_LRU); if (IS_ERR_OR_NULL(page)) goto out; if (PageAnon(page)) { @@ -2310,7 +2310,7 @@ next_mm: while (ksm_scan.address < vma->vm_end) { if (ksm_test_exit(mm)) break; - *page = follow_page(vma, ksm_scan.address, FOLL_GET); + *page = follow_page(vma, ksm_scan.address, FOLL_GET | FOLL_LRU); if (IS_ERR_OR_NULL(*page)) { ksm_scan.address += PAGE_SIZE; cond_resched(); diff --git a/mm/madvise.c b/mm/madvise.c index 12e8010c766c..7a8af04069b3 100644 --- a/mm/madvise.c +++ b/mm/madvise.c @@ -421,7 +421,7 @@ regular_page: continue; page = vm_normal_page(vma, addr, ptent); - if (!page) + if (!page || is_zone_device_page(page)) continue; /* @@ -639,7 +639,7 @@ static int madvise_free_pte_range(pmd_t *pmd, unsigned long addr, } page = vm_normal_page(vma, addr, ptent); - if (!page) + if (!page || is_zone_device_page(page)) continue; /* diff --git a/mm/memory.c b/mm/memory.c index fdfc119f181d..c905fabddeec 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -633,6 +633,13 @@ struct page *vm_normal_page(struct vm_area_struct *vma, unsigned long addr, if (is_zero_pfn(pfn)) return NULL; if (pte_devmap(pte)) +/* + * NOTE: New uers of ZONE_DEVICE will not set pte_devmap() and will have + * refcounts incremented on their struct pages when they are inserted into + * PTEs, thus they are safe to return here. Legacy ZONE_DEVICE pages that set + * pte_devmap() do not have refcounts. Example of legacy ZONE_DEVICE is + * MEMORY_DEVICE_FS_DAX type in pmem or virtio_fs drivers. + */ return NULL; print_bad_pte(vma, addr, pte, NULL); @@ -4708,7 +4715,7 @@ static vm_fault_t do_numa_page(struct vm_fault *vmf) pte = pte_modify(old_pte, vma->vm_page_prot); page = vm_normal_page(vma, vmf->address, pte); - if (!page) + if (!page || is_zone_device_page(page)) goto out_map; /* TODO: handle PTE-mapped THP */ diff --git a/mm/mempolicy.c b/mm/mempolicy.c index 201b043e96f1..b879a326bfb7 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -524,7 +524,7 @@ static int queue_pages_pte_range(pmd_t *pmd, unsigned long addr, if (!pte_present(*pte)) continue; page = vm_normal_page(vma, addr, *pte); - if (!page) + if (!page || is_zone_device_page(page)) continue; /* * vm_normal_page() filters out zero pages, but there might diff --git a/mm/migrate.c b/mm/migrate.c index 7934eebf1689..4cf6b57d3d8e 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -1623,7 +1623,7 @@ static int add_page_for_migration(struct mm_struct *mm, unsigned long addr, goto out; /* FOLL_DUMP to ignore special (like zero) pages */ - page = follow_page(vma, addr, FOLL_GET | FOLL_DUMP); + page = follow_page(vma, addr, FOLL_GET | FOLL_DUMP | FOLL_LRU); err = PTR_ERR(page); if (IS_ERR(page)) @@ -1815,7 +1815,7 @@ static void do_pages_stat_array(struct mm_struct *mm, unsigned long nr_pages, goto set_status; /* FOLL_DUMP to ignore special (like zero) pages */ - page = follow_page(vma, addr, FOLL_GET | FOLL_DUMP); + page = follow_page(vma, addr, FOLL_GET | FOLL_DUMP | FOLL_LRU); err = PTR_ERR(page); if (IS_ERR(page)) diff --git a/mm/mlock.c b/mm/mlock.c index c41604ba5197..43d19a1f28eb 100644 --- a/mm/mlock.c +++ b/mm/mlock.c @@ -333,7 +333,7 @@ static int mlock_pte_range(pmd_t *pmd, unsigned long addr, if (!pte_present(*pte)) continue; page = vm_normal_page(vma, addr, *pte); - if (!page) + if (!page || is_zone_device_page(page)) continue; if (PageTransCompound(page)) continue; diff --git a/mm/mprotect.c b/mm/mprotect.c index 27fb652151e4..f7b31657d1da 100644 --- a/mm/mprotect.c +++ b/mm/mprotect.c @@ -95,7 +95,7 @@ static unsigned long change_pte_range(struct mmu_gather *tlb, continue; page = vm_normal_page(vma, addr, oldpte); - if (!page || PageKsm(page)) + if (!page || is_zone_device_page(page) || PageKsm(page)) continue; /* Also skip shared copy-on-write pages */ -- cgit v1.2.3 From 190df2968d8d5c236e27647b04fdfcfa5af9ceb4 Mon Sep 17 00:00:00 2001 From: Alex Sierra Date: Tue, 31 May 2022 15:00:31 -0500 Subject: mm: add device coherent vma selection for memory migration This case is used to migrate pages from device memory, back to system memory. Device coherent type memory is cache coherent from device and CPU point of view. Link: https://lkml.kernel.org/r/20220531200041.24904-4-alex.sierra@amd.com Signed-off-by: Alex Sierra Signed-off-by: Christoph Hellwig Acked-by: Felix Kuehling Reviewed-by: Alistair Poppple Cc: David Hildenbrand Cc: Jason Gunthorpe Cc: Jerome Glisse Cc: Matthew Wilcox Cc: Ralph Campbell Signed-off-by: Andrew Morton --- include/linux/migrate.h | 1 + mm/migrate_device.c | 12 +++++++++--- 2 files changed, 10 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/include/linux/migrate.h b/include/linux/migrate.h index 069a89e847f3..b84908debe5c 100644 --- a/include/linux/migrate.h +++ b/include/linux/migrate.h @@ -148,6 +148,7 @@ static inline unsigned long migrate_pfn(unsigned long pfn) enum migrate_vma_direction { MIGRATE_VMA_SELECT_SYSTEM = 1 << 0, MIGRATE_VMA_SELECT_DEVICE_PRIVATE = 1 << 1, + MIGRATE_VMA_SELECT_DEVICE_COHERENT = 1 << 2, }; struct migrate_vma { diff --git a/mm/migrate_device.c b/mm/migrate_device.c index a4847ad65da3..18bc6483f63a 100644 --- a/mm/migrate_device.c +++ b/mm/migrate_device.c @@ -148,15 +148,21 @@ again: if (is_writable_device_private_entry(entry)) mpfn |= MIGRATE_PFN_WRITE; } else { - if (!(migrate->flags & MIGRATE_VMA_SELECT_SYSTEM)) - goto next; pfn = pte_pfn(pte); - if (is_zero_pfn(pfn)) { + if (is_zero_pfn(pfn) && + (migrate->flags & MIGRATE_VMA_SELECT_SYSTEM)) { mpfn = MIGRATE_PFN_MIGRATE; migrate->cpages++; goto next; } page = vm_normal_page(migrate->vma, addr, pte); + if (page && !is_zone_device_page(page) && + !(migrate->flags & MIGRATE_VMA_SELECT_SYSTEM)) + goto next; + else if (page && is_device_coherent_page(page) && + (!(migrate->flags & MIGRATE_VMA_SELECT_DEVICE_COHERENT) || + page->pgmap->owner != migrate->pgmap_owner)) + goto next; mpfn = migrate_pfn(pfn) | MIGRATE_PFN_MIGRATE; mpfn |= pte_write(pte) ? MIGRATE_PFN_WRITE : 0; } -- cgit v1.2.3 From e5cbd57126c5465c6c0019b4a37d566f3039dc2e Mon Sep 17 00:00:00 2001 From: SeongJae Park Date: Mon, 6 Jun 2022 18:23:06 +0000 Subject: mm/damon/{dbgfs,sysfs}: move target_has_pid() from dbgfs to damon.h The function for knowing if given monitoring context's targets will have pid or not is defined and used in dbgfs only. However, the logic is also needed for sysfs. This commit moves the code to damon.h and makes both dbgfs and sysfs to use it. Link: https://lkml.kernel.org/r/20220606182310.48781-3-sj@kernel.org Signed-off-by: SeongJae Park Cc: Jonathan Corbet Signed-off-by: Andrew Morton --- include/linux/damon.h | 6 ++++++ mm/damon/dbgfs.c | 15 +++++---------- mm/damon/sysfs.c | 8 +++----- 3 files changed, 14 insertions(+), 15 deletions(-) (limited to 'include/linux') diff --git a/include/linux/damon.h b/include/linux/damon.h index 2765c7d99beb..b9aae19fab3e 100644 --- a/include/linux/damon.h +++ b/include/linux/damon.h @@ -525,6 +525,12 @@ bool damon_is_registered_ops(enum damon_ops_id id); int damon_register_ops(struct damon_operations *ops); int damon_select_ops(struct damon_ctx *ctx, enum damon_ops_id id); +static inline bool damon_target_has_pid(const struct damon_ctx *ctx) +{ + return ctx->ops.id == DAMON_OPS_VADDR || ctx->ops.id == DAMON_OPS_FVADDR; +} + + int damon_start(struct damon_ctx **ctxs, int nr_ctxs, bool exclusive); int damon_stop(struct damon_ctx **ctxs, int nr_ctxs); diff --git a/mm/damon/dbgfs.c b/mm/damon/dbgfs.c index a0dab8b5e45f..5ae810927309 100644 --- a/mm/damon/dbgfs.c +++ b/mm/damon/dbgfs.c @@ -275,11 +275,6 @@ out: return ret; } -static inline bool target_has_pid(const struct damon_ctx *ctx) -{ - return ctx->ops.id == DAMON_OPS_VADDR; -} - static ssize_t sprint_target_ids(struct damon_ctx *ctx, char *buf, ssize_t len) { struct damon_target *t; @@ -288,7 +283,7 @@ static ssize_t sprint_target_ids(struct damon_ctx *ctx, char *buf, ssize_t len) int rc; damon_for_each_target(t, ctx) { - if (target_has_pid(ctx)) + if (damon_target_has_pid(ctx)) /* Show pid numbers to debugfs users */ id = pid_vnr(t->pid); else @@ -415,7 +410,7 @@ static int dbgfs_set_targets(struct damon_ctx *ctx, ssize_t nr_targets, struct damon_target *t, *next; damon_for_each_target_safe(t, next, ctx) { - if (target_has_pid(ctx)) + if (damon_target_has_pid(ctx)) put_pid(t->pid); damon_destroy_target(t); } @@ -425,11 +420,11 @@ static int dbgfs_set_targets(struct damon_ctx *ctx, ssize_t nr_targets, if (!t) { damon_for_each_target_safe(t, next, ctx) damon_destroy_target(t); - if (target_has_pid(ctx)) + if (damon_target_has_pid(ctx)) dbgfs_put_pids(pids, nr_targets); return -ENOMEM; } - if (target_has_pid(ctx)) + if (damon_target_has_pid(ctx)) t->pid = pids[i]; damon_add_target(ctx, t); } @@ -722,7 +717,7 @@ static void dbgfs_before_terminate(struct damon_ctx *ctx) { struct damon_target *t, *next; - if (!target_has_pid(ctx)) + if (!damon_target_has_pid(ctx)) return; mutex_lock(&ctx->kdamond_lock); diff --git a/mm/damon/sysfs.c b/mm/damon/sysfs.c index 09f9e8ca3d1f..8810e6abdb06 100644 --- a/mm/damon/sysfs.c +++ b/mm/damon/sysfs.c @@ -2136,8 +2136,7 @@ static void damon_sysfs_destroy_targets(struct damon_ctx *ctx) struct damon_target *t, *next; damon_for_each_target_safe(t, next, ctx) { - if (ctx->ops.id == DAMON_OPS_VADDR || - ctx->ops.id == DAMON_OPS_FVADDR) + if (damon_target_has_pid(ctx)) put_pid(t->pid); damon_destroy_target(t); } @@ -2181,8 +2180,7 @@ static int damon_sysfs_add_target(struct damon_sysfs_target *sys_target, if (!t) return -ENOMEM; - if (ctx->ops.id == DAMON_OPS_VADDR || - ctx->ops.id == DAMON_OPS_FVADDR) { + if (damon_target_has_pid(ctx)) { t->pid = find_get_pid(sys_target->pid); if (!t->pid) goto destroy_targets_out; @@ -2210,7 +2208,7 @@ static struct damon_target *damon_sysfs_existing_target( struct pid *pid; struct damon_target *t; - if (ctx->ops.id == DAMON_OPS_PADDR) { + if (!damon_target_has_pid(ctx)) { /* Up to only one target for paddr could exist */ damon_for_each_target(t, ctx) return t; -- cgit v1.2.3 From 1b11f6a82f0fcb42408a940245e47ae2a66a6bfe Mon Sep 17 00:00:00 2001 From: Andrey Konovalov Date: Thu, 9 Jun 2022 20:18:46 +0200 Subject: mm: introduce clear_highpage_kasan_tagged Add a clear_highpage_kasan_tagged() helper that does clear_highpage() on a page potentially tagged by KASAN. This helper is used by the following patch. Link: https://lkml.kernel.org/r/4471979b46b2c487787ddcd08b9dc5fedd1b6ffd.1654798516.git.andreyknvl@google.com Signed-off-by: Andrey Konovalov Cc: Alexander Potapenko Cc: Andrey Ryabinin Cc: Dmitry Vyukov Cc: Marco Elver Signed-off-by: Andrew Morton --- include/linux/highmem.h | 10 ++++++++++ mm/page_alloc.c | 8 ++------ 2 files changed, 12 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/highmem.h b/include/linux/highmem.h index 158a3a2907dc..177b07944640 100644 --- a/include/linux/highmem.h +++ b/include/linux/highmem.h @@ -243,6 +243,16 @@ static inline void clear_highpage(struct page *page) kunmap_local(kaddr); } +static inline void clear_highpage_kasan_tagged(struct page *page) +{ + u8 tag; + + tag = page_kasan_tag(page); + page_kasan_tag_reset(page); + clear_highpage(page); + page_kasan_tag_set(page, tag); +} + #ifndef __HAVE_ARCH_TAG_CLEAR_HIGHPAGE static inline void tag_clear_highpage(struct page *page) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index cf2619af0931..7bb359d3b9d4 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1302,12 +1302,8 @@ static void kernel_init_pages(struct page *page, int numpages) /* s390's use of memset() could override KASAN redzones. */ kasan_disable_current(); - for (i = 0; i < numpages; i++) { - u8 tag = page_kasan_tag(page + i); - page_kasan_tag_reset(page + i); - clear_highpage(page + i); - page_kasan_tag_set(page + i, tag); - } + for (i = 0; i < numpages; i++) + clear_highpage_kasan_tagged(page + i); kasan_enable_current(); } -- cgit v1.2.3 From 2f8077984564d94011d5e5f4d6c3e516f0f73f90 Mon Sep 17 00:00:00 2001 From: Roman Gushchin Date: Tue, 31 May 2022 20:22:22 -0700 Subject: mm: memcontrol: introduce mem_cgroup_ino() and mem_cgroup_get_from_ino() Patch series "mm: introduce shrinker debugfs interface", v5. The only existing debugging mechanism is a couple of tracepoints in do_shrink_slab(): mm_shrink_slab_start and mm_shrink_slab_end. They aren't covering everything though: shrinkers which report 0 objects will never show up, there is no support for memcg-aware shrinkers. Shrinkers are identified by their scan function, which is not always enough (e.g. hard to guess which super block's shrinker it is having only "super_cache_scan"). To provide a better visibility and debug options for memory shrinkers this patchset introduces a /sys/kernel/debug/shrinker interface, to some extent similar to /sys/kernel/slab. For each shrinker registered in the system a directory is created. As now, the directory will contain only a "scan" file, which allows to get the number of managed objects for each memory cgroup (for memcg-aware shrinkers) and each numa node (for numa-aware shrinkers on a numa machine). Other interfaces might be added in the future. To make debugging more pleasant, the patchset also names all shrinkers, so that debugfs entries can have meaningful names. This patch (of 5): Shrinker debugfs requires a way to represent memory cgroups without using full paths, both for displaying information and getting input from a user. Cgroup inode number is a perfect way, already used by bpf. This commit adds a couple of helper functions which will be used to handle memcg-aware shrinkers. Link: https://lkml.kernel.org/r/20220601032227.4076670-1-roman.gushchin@linux.dev Link: https://lkml.kernel.org/r/20220601032227.4076670-2-roman.gushchin@linux.dev Signed-off-by: Roman Gushchin Acked-by: Muchun Song Cc: Dave Chinner Cc: Kent Overstreet Cc: Hillf Danton Cc: Christophe JAILLET Cc: Roman Gushchin Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 21 +++++++++++++++++++++ mm/memcontrol.c | 23 +++++++++++++++++++++++ 2 files changed, 44 insertions(+) (limited to 'include/linux') diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 04f2f33607e9..4d31ce55b1c0 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -837,6 +837,15 @@ static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg) } struct mem_cgroup *mem_cgroup_from_id(unsigned short id); +#ifdef CONFIG_SHRINKER_DEBUG +static inline unsigned long mem_cgroup_ino(struct mem_cgroup *memcg) +{ + return memcg ? cgroup_ino(memcg->css.cgroup) : 0; +} + +struct mem_cgroup *mem_cgroup_get_from_ino(unsigned long ino); +#endif + static inline struct mem_cgroup *mem_cgroup_from_seq(struct seq_file *m) { return mem_cgroup_from_css(seq_css(m)); @@ -1343,6 +1352,18 @@ static inline struct mem_cgroup *mem_cgroup_from_id(unsigned short id) return NULL; } +#ifdef CONFIG_SHRINKER_DEBUG +static inline unsigned long mem_cgroup_ino(struct mem_cgroup *memcg) +{ + return 0; +} + +static inline struct mem_cgroup *mem_cgroup_get_from_ino(unsigned long ino) +{ + return NULL; +} +#endif + static inline struct mem_cgroup *mem_cgroup_from_seq(struct seq_file *m) { return NULL; diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 370ef639110f..163492b9efa9 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -5088,6 +5088,29 @@ struct mem_cgroup *mem_cgroup_from_id(unsigned short id) return idr_find(&mem_cgroup_idr, id); } +#ifdef CONFIG_SHRINKER_DEBUG +struct mem_cgroup *mem_cgroup_get_from_ino(unsigned long ino) +{ + struct cgroup *cgrp; + struct cgroup_subsys_state *css; + struct mem_cgroup *memcg; + + cgrp = cgroup_get_from_id(ino); + if (!cgrp) + return ERR_PTR(-ENOENT); + + css = cgroup_get_e_css(cgrp, &memory_cgrp_subsys); + if (css) + memcg = container_of(css, struct mem_cgroup, css); + else + memcg = ERR_PTR(-ENOENT); + + cgroup_put(cgrp); + + return memcg; +} +#endif + static int alloc_mem_cgroup_per_node_info(struct mem_cgroup *memcg, int node) { struct mem_cgroup_per_node *pn; -- cgit v1.2.3 From 4f65acafcb649754f75e6009a6b176b3caafb4b3 Mon Sep 17 00:00:00 2001 From: Roman Gushchin Date: Tue, 31 May 2022 20:22:23 -0700 Subject: mm: shrinkers: introduce debugfs interface for memory shrinkers This commit introduces the /sys/kernel/debug/shrinker debugfs interface which provides an ability to observe the state of individual kernel memory shrinkers. Because the feature adds some memory overhead (which shouldn't be large unless there is a huge amount of registered shrinkers), it's guarded by a config option (enabled by default). This commit introduces the "count" interface for each shrinker registered in the system. The output is in the following format: ... ... ... To reduce the size of output on machines with many thousands cgroups, if the total number of objects on all nodes is 0, the line is omitted. If the shrinker is not memcg-aware or CONFIG_MEMCG is off, 0 is printed as cgroup inode id. If the shrinker is not numa-aware, 0's are printed for all nodes except the first one. This commit gives debugfs entries simple numeric names, which are not very convenient. The following commit in the series will provide shrinkers with more meaningful names. Link: https://lkml.kernel.org/r/20220601032227.4076670-3-roman.gushchin@linux.dev Signed-off-by: Roman Gushchin Reviewed-by: Kent Overstreet Acked-by: Muchun Song Cc: Christophe JAILLET Cc: Dave Chinner Cc: Hillf Danton Signed-off-by: Andrew Morton --- include/linux/shrinker.h | 19 +++++- lib/Kconfig.debug | 9 +++ mm/Makefile | 1 + mm/shrinker_debug.c | 168 +++++++++++++++++++++++++++++++++++++++++++++++ mm/vmscan.c | 6 +- 5 files changed, 200 insertions(+), 3 deletions(-) create mode 100644 mm/shrinker_debug.c (limited to 'include/linux') diff --git a/include/linux/shrinker.h b/include/linux/shrinker.h index 76fbf92b04d9..2ced8149c513 100644 --- a/include/linux/shrinker.h +++ b/include/linux/shrinker.h @@ -72,6 +72,10 @@ struct shrinker { #ifdef CONFIG_MEMCG /* ID in shrinker_idr */ int id; +#endif +#ifdef CONFIG_SHRINKER_DEBUG + int debugfs_id; + struct dentry *debugfs_entry; #endif /* objs pending delete, per node */ atomic_long_t *nr_deferred; @@ -94,4 +98,17 @@ extern int register_shrinker(struct shrinker *shrinker); extern void unregister_shrinker(struct shrinker *shrinker); extern void free_prealloced_shrinker(struct shrinker *shrinker); extern void synchronize_shrinkers(void); -#endif + +#ifdef CONFIG_SHRINKER_DEBUG +extern int shrinker_debugfs_add(struct shrinker *shrinker); +extern void shrinker_debugfs_remove(struct shrinker *shrinker); +#else /* CONFIG_SHRINKER_DEBUG */ +static inline int shrinker_debugfs_add(struct shrinker *shrinker) +{ + return 0; +} +static inline void shrinker_debugfs_remove(struct shrinker *shrinker) +{ +} +#endif /* CONFIG_SHRINKER_DEBUG */ +#endif /* _LINUX_SHRINKER_H */ diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 6b84d0e7c3ad..fb328c095d5b 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -699,6 +699,15 @@ config DEBUG_OBJECTS_ENABLE_DEFAULT help Debug objects boot parameter default value +config SHRINKER_DEBUG + default y + bool "Enable shrinker debugging support" + depends on DEBUG_FS + help + Say Y to enable the shrinker debugfs interface which provides + visibility into the kernel memory shrinkers subsystem. + Disable it to avoid an extra memory footprint. + config HAVE_DEBUG_KMEMLEAK bool diff --git a/mm/Makefile b/mm/Makefile index 298c9991ab75..8083fa85a348 100644 --- a/mm/Makefile +++ b/mm/Makefile @@ -133,3 +133,4 @@ obj-$(CONFIG_PAGE_REPORTING) += page_reporting.o obj-$(CONFIG_IO_MAPPING) += io-mapping.o obj-$(CONFIG_HAVE_BOOTMEM_INFO_NODE) += bootmem_info.o obj-$(CONFIG_GENERIC_IOREMAP) += ioremap.o +obj-$(CONFIG_SHRINKER_DEBUG) += shrinker_debug.o diff --git a/mm/shrinker_debug.c b/mm/shrinker_debug.c new file mode 100644 index 000000000000..1a70556bd46c --- /dev/null +++ b/mm/shrinker_debug.c @@ -0,0 +1,168 @@ +// SPDX-License-Identifier: GPL-2.0 +#include +#include +#include +#include +#include +#include + +/* defined in vmscan.c */ +extern struct rw_semaphore shrinker_rwsem; +extern struct list_head shrinker_list; + +static DEFINE_IDA(shrinker_debugfs_ida); +static struct dentry *shrinker_debugfs_root; + +static unsigned long shrinker_count_objects(struct shrinker *shrinker, + struct mem_cgroup *memcg, + unsigned long *count_per_node) +{ + unsigned long nr, total = 0; + int nid; + + for_each_node(nid) { + if (nid == 0 || (shrinker->flags & SHRINKER_NUMA_AWARE)) { + struct shrink_control sc = { + .gfp_mask = GFP_KERNEL, + .nid = nid, + .memcg = memcg, + }; + + nr = shrinker->count_objects(shrinker, &sc); + if (nr == SHRINK_EMPTY) + nr = 0; + } else { + nr = 0; + } + + count_per_node[nid] = nr; + total += nr; + } + + return total; +} + +static int shrinker_debugfs_count_show(struct seq_file *m, void *v) +{ + struct shrinker *shrinker = m->private; + unsigned long *count_per_node; + struct mem_cgroup *memcg; + unsigned long total; + bool memcg_aware; + int ret, nid; + + count_per_node = kcalloc(nr_node_ids, sizeof(unsigned long), GFP_KERNEL); + if (!count_per_node) + return -ENOMEM; + + ret = down_read_killable(&shrinker_rwsem); + if (ret) { + kfree(count_per_node); + return ret; + } + rcu_read_lock(); + + memcg_aware = shrinker->flags & SHRINKER_MEMCG_AWARE; + + memcg = mem_cgroup_iter(NULL, NULL, NULL); + do { + if (memcg && !mem_cgroup_online(memcg)) + continue; + + total = shrinker_count_objects(shrinker, + memcg_aware ? memcg : NULL, + count_per_node); + if (total) { + seq_printf(m, "%lu", mem_cgroup_ino(memcg)); + for_each_node(nid) + seq_printf(m, " %lu", count_per_node[nid]); + seq_putc(m, '\n'); + } + + if (!memcg_aware) { + mem_cgroup_iter_break(NULL, memcg); + break; + } + + if (signal_pending(current)) { + mem_cgroup_iter_break(NULL, memcg); + ret = -EINTR; + break; + } + } while ((memcg = mem_cgroup_iter(NULL, memcg, NULL)) != NULL); + + rcu_read_unlock(); + up_read(&shrinker_rwsem); + + kfree(count_per_node); + return ret; +} +DEFINE_SHOW_ATTRIBUTE(shrinker_debugfs_count); + +int shrinker_debugfs_add(struct shrinker *shrinker) +{ + struct dentry *entry; + char buf[16]; + int id; + + lockdep_assert_held(&shrinker_rwsem); + + /* debugfs isn't initialized yet, add debugfs entries later. */ + if (!shrinker_debugfs_root) + return 0; + + id = ida_alloc(&shrinker_debugfs_ida, GFP_KERNEL); + if (id < 0) + return id; + shrinker->debugfs_id = id; + + snprintf(buf, sizeof(buf), "%d", id); + + /* create debugfs entry */ + entry = debugfs_create_dir(buf, shrinker_debugfs_root); + if (IS_ERR(entry)) { + ida_free(&shrinker_debugfs_ida, id); + return PTR_ERR(entry); + } + shrinker->debugfs_entry = entry; + + debugfs_create_file("count", 0220, entry, shrinker, + &shrinker_debugfs_count_fops); + return 0; +} + +void shrinker_debugfs_remove(struct shrinker *shrinker) +{ + lockdep_assert_held(&shrinker_rwsem); + + if (!shrinker->debugfs_entry) + return; + + debugfs_remove_recursive(shrinker->debugfs_entry); + ida_free(&shrinker_debugfs_ida, shrinker->debugfs_id); +} + +static int __init shrinker_debugfs_init(void) +{ + struct shrinker *shrinker; + struct dentry *dentry; + int ret = 0; + + dentry = debugfs_create_dir("shrinker", NULL); + if (IS_ERR(dentry)) + return PTR_ERR(dentry); + shrinker_debugfs_root = dentry; + + /* Create debugfs entries for shrinkers registered at boot */ + down_write(&shrinker_rwsem); + list_for_each_entry(shrinker, &shrinker_list, list) + if (!shrinker->debugfs_entry) { + ret = shrinker_debugfs_add(shrinker); + if (ret) + break; + } + up_write(&shrinker_rwsem); + + return ret; +} +late_initcall(shrinker_debugfs_init); diff --git a/mm/vmscan.c b/mm/vmscan.c index f7d9a683e3a7..91232beffa63 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -190,8 +190,8 @@ static void set_task_reclaim_state(struct task_struct *task, task->reclaim_state = rs; } -static LIST_HEAD(shrinker_list); -static DECLARE_RWSEM(shrinker_rwsem); +LIST_HEAD(shrinker_list); +DECLARE_RWSEM(shrinker_rwsem); #ifdef CONFIG_MEMCG static int shrinker_nr_max; @@ -650,6 +650,7 @@ void register_shrinker_prepared(struct shrinker *shrinker) down_write(&shrinker_rwsem); list_add_tail(&shrinker->list, &shrinker_list); shrinker->flags |= SHRINKER_REGISTERED; + WARN_ON_ONCE(shrinker_debugfs_add(shrinker)); up_write(&shrinker_rwsem); } @@ -677,6 +678,7 @@ void unregister_shrinker(struct shrinker *shrinker) shrinker->flags &= ~SHRINKER_REGISTERED; if (shrinker->flags & SHRINKER_MEMCG_AWARE) unregister_memcg_shrinker(shrinker); + shrinker_debugfs_remove(shrinker); up_write(&shrinker_rwsem); kfree(shrinker->nr_deferred); -- cgit v1.2.3 From bec0918551a79c3c6b63a493a80e35e8b402804f Mon Sep 17 00:00:00 2001 From: Roman Gushchin Date: Tue, 31 May 2022 20:22:24 -0700 Subject: mm: shrinkers: provide shrinkers with names Currently shrinkers are anonymous objects. For debugging purposes they can be identified by count/scan function names, but it's not always useful: e.g. for superblock's shrinkers it's nice to have at least an idea of to which superblock the shrinker belongs. This commit adds names to shrinkers. register_shrinker() and prealloc_shrinker() functions are extended to take a format and arguments to master a name. In some cases it's not possible to determine a good name at the time when a shrinker is allocated. For such cases shrinker_debugfs_rename() is provided. The expected format is: -[:]- For some shrinkers an instance can be encoded as (MAJOR:MINOR) pair. After this change the shrinker debugfs directory looks like: $ cd /sys/kernel/debug/shrinker/ $ ls dquota-cache-16 sb-devpts-28 sb-proc-47 sb-tmpfs-42 mm-shadow-18 sb-devtmpfs-5 sb-proc-48 sb-tmpfs-43 mm-zspool:zram0-34 sb-hugetlbfs-17 sb-pstore-31 sb-tmpfs-44 rcu-kfree-0 sb-hugetlbfs-33 sb-rootfs-2 sb-tmpfs-49 sb-aio-20 sb-iomem-12 sb-securityfs-6 sb-tracefs-13 sb-anon_inodefs-15 sb-mqueue-21 sb-selinuxfs-22 sb-xfs:vda1-36 sb-bdev-3 sb-nsfs-4 sb-sockfs-8 sb-zsmalloc-19 sb-bpf-32 sb-pipefs-14 sb-sysfs-26 thp-deferred_split-10 sb-btrfs:vda2-24 sb-proc-25 sb-tmpfs-1 thp-zero-9 sb-cgroup2-30 sb-proc-39 sb-tmpfs-27 xfs-buf:vda1-37 sb-configfs-23 sb-proc-41 sb-tmpfs-29 xfs-inodegc:vda1-38 sb-dax-11 sb-proc-45 sb-tmpfs-35 sb-debugfs-7 sb-proc-46 sb-tmpfs-40 Link: https://lkml.kernel.org/r/20220601032227.4076670-4-roman.gushchin@linux.dev Signed-off-by: Roman Gushchin Cc: Christophe JAILLET Cc: Dave Chinner Cc: Hillf Danton Cc: Kent Overstreet Cc: Muchun Song Signed-off-by: Andrew Morton --- arch/x86/kvm/mmu/mmu.c | 2 +- drivers/android/binder_alloc.c | 2 +- drivers/gpu/drm/i915/gem/i915_gem_shrinker.c | 3 +- drivers/gpu/drm/msm/msm_gem_shrinker.c | 2 +- drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c | 2 +- drivers/gpu/drm/ttm/ttm_pool.c | 2 +- drivers/md/bcache/btree.c | 2 +- drivers/md/dm-bufio.c | 3 +- drivers/md/dm-zoned-metadata.c | 4 +- drivers/md/raid5.c | 2 +- drivers/misc/vmw_balloon.c | 2 +- drivers/virtio/virtio_balloon.c | 2 +- drivers/xen/xenbus/xenbus_probe_backend.c | 2 +- fs/btrfs/super.c | 2 + fs/erofs/utils.c | 2 +- fs/ext4/extents_status.c | 3 +- fs/f2fs/super.c | 2 +- fs/gfs2/glock.c | 2 +- fs/gfs2/main.c | 2 +- fs/jbd2/journal.c | 3 +- fs/mbcache.c | 2 +- fs/nfs/nfs42xattr.c | 7 +-- fs/nfs/super.c | 2 +- fs/nfsd/filecache.c | 2 +- fs/nfsd/nfscache.c | 3 +- fs/quota/dquot.c | 2 +- fs/super.c | 6 ++- fs/ubifs/super.c | 2 +- fs/xfs/xfs_buf.c | 3 +- fs/xfs/xfs_icache.c | 2 +- fs/xfs/xfs_qm.c | 3 +- include/linux/shrinker.h | 12 ++++- kernel/rcu/tree.c | 2 +- mm/huge_memory.c | 4 +- mm/shrinker_debug.c | 47 ++++++++++++++++++- mm/vmscan.c | 58 ++++++++++++++++++++++-- mm/workingset.c | 2 +- mm/zsmalloc.c | 3 +- net/sunrpc/auth.c | 2 +- 39 files changed, 165 insertions(+), 45 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c index 17252f39bd7c..797d3286ecc1 100644 --- a/arch/x86/kvm/mmu/mmu.c +++ b/arch/x86/kvm/mmu/mmu.c @@ -6317,7 +6317,7 @@ int kvm_mmu_vendor_module_init(void) if (percpu_counter_init(&kvm_total_used_mmu_pages, 0, GFP_KERNEL)) goto out; - ret = register_shrinker(&mmu_shrinker); + ret = register_shrinker(&mmu_shrinker, "x86-mmu"); if (ret) goto out; diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c index 5a00a468a151..1014beb12802 100644 --- a/drivers/android/binder_alloc.c +++ b/drivers/android/binder_alloc.c @@ -1089,7 +1089,7 @@ int binder_alloc_shrinker_init(void) int ret = list_lru_init(&binder_alloc_lru); if (ret == 0) { - ret = register_shrinker(&binder_shrinker); + ret = register_shrinker(&binder_shrinker, "android-binder"); if (ret) list_lru_destroy(&binder_alloc_lru); } diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c index 6a6ff98a8746..e43577e03067 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c @@ -426,7 +426,8 @@ void i915_gem_driver_register__shrinker(struct drm_i915_private *i915) i915->mm.shrinker.count_objects = i915_gem_shrinker_count; i915->mm.shrinker.seeks = DEFAULT_SEEKS; i915->mm.shrinker.batch = 4096; - drm_WARN_ON(&i915->drm, register_shrinker(&i915->mm.shrinker)); + drm_WARN_ON(&i915->drm, register_shrinker(&i915->mm.shrinker, + "drm-i915_gem")); i915->mm.oom_notifier.notifier_call = i915_gem_shrinker_oom; drm_WARN_ON(&i915->drm, register_oom_notifier(&i915->mm.oom_notifier)); diff --git a/drivers/gpu/drm/msm/msm_gem_shrinker.c b/drivers/gpu/drm/msm/msm_gem_shrinker.c index 086dacf2f26a..26e84d2ea6ae 100644 --- a/drivers/gpu/drm/msm/msm_gem_shrinker.c +++ b/drivers/gpu/drm/msm/msm_gem_shrinker.c @@ -221,7 +221,7 @@ void msm_gem_shrinker_init(struct drm_device *dev) priv->shrinker.count_objects = msm_gem_shrinker_count; priv->shrinker.scan_objects = msm_gem_shrinker_scan; priv->shrinker.seeks = DEFAULT_SEEKS; - WARN_ON(register_shrinker(&priv->shrinker)); + WARN_ON(register_shrinker(&priv->shrinker, "drm-msm_gem")); priv->vmap_notifier.notifier_call = msm_gem_shrinker_vmap; WARN_ON(register_vmap_purge_notifier(&priv->vmap_notifier)); diff --git a/drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c b/drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c index 77e7cb6d1ae3..bf0170782f25 100644 --- a/drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c +++ b/drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c @@ -103,7 +103,7 @@ void panfrost_gem_shrinker_init(struct drm_device *dev) pfdev->shrinker.count_objects = panfrost_gem_shrinker_count; pfdev->shrinker.scan_objects = panfrost_gem_shrinker_scan; pfdev->shrinker.seeks = DEFAULT_SEEKS; - WARN_ON(register_shrinker(&pfdev->shrinker)); + WARN_ON(register_shrinker(&pfdev->shrinker, "drm-panfrost")); } /** diff --git a/drivers/gpu/drm/ttm/ttm_pool.c b/drivers/gpu/drm/ttm/ttm_pool.c index 1bba0a0ed3f9..21b61631f73a 100644 --- a/drivers/gpu/drm/ttm/ttm_pool.c +++ b/drivers/gpu/drm/ttm/ttm_pool.c @@ -722,7 +722,7 @@ int ttm_pool_mgr_init(unsigned long num_pages) mm_shrinker.count_objects = ttm_pool_shrinker_count; mm_shrinker.scan_objects = ttm_pool_shrinker_scan; mm_shrinker.seeks = 1; - return register_shrinker(&mm_shrinker); + return register_shrinker(&mm_shrinker, "drm-ttm_pool"); } /** diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c index e136d6edc1ed..147c493a989a 100644 --- a/drivers/md/bcache/btree.c +++ b/drivers/md/bcache/btree.c @@ -812,7 +812,7 @@ int bch_btree_cache_alloc(struct cache_set *c) c->shrink.seeks = 4; c->shrink.batch = c->btree_pages * 2; - if (register_shrinker(&c->shrink)) + if (register_shrinker(&c->shrink, "md-bcache:%pU", c->set_uuid)) pr_warn("bcache: %s: could not register shrinker\n", __func__); diff --git a/drivers/md/dm-bufio.c b/drivers/md/dm-bufio.c index 5ffa1dcf84cf..3ff571b20f14 100644 --- a/drivers/md/dm-bufio.c +++ b/drivers/md/dm-bufio.c @@ -1806,7 +1806,8 @@ struct dm_bufio_client *dm_bufio_client_create(struct block_device *bdev, unsign c->shrinker.scan_objects = dm_bufio_shrink_scan; c->shrinker.seeks = 1; c->shrinker.batch = 0; - r = register_shrinker(&c->shrinker); + r = register_shrinker(&c->shrinker, "md-%s:(%u:%u)", slab_name, + MAJOR(bdev->bd_dev), MINOR(bdev->bd_dev)); if (r) goto bad; diff --git a/drivers/md/dm-zoned-metadata.c b/drivers/md/dm-zoned-metadata.c index d1ea66114d14..46648f6100fb 100644 --- a/drivers/md/dm-zoned-metadata.c +++ b/drivers/md/dm-zoned-metadata.c @@ -2944,7 +2944,9 @@ int dmz_ctr_metadata(struct dmz_dev *dev, int num_dev, zmd->mblk_shrinker.seeks = DEFAULT_SEEKS; /* Metadata cache shrinker */ - ret = register_shrinker(&zmd->mblk_shrinker); + ret = register_shrinker(&zmd->mblk_shrinker, "md-meta:(%u:%u)", + MAJOR(dev->bdev->bd_dev), + MINOR(dev->bdev->bd_dev)); if (ret) { dmz_zmd_err(zmd, "Register metadata cache shrinker failed"); goto err; diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 5d09256d7f81..780ae66840b7 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -7414,7 +7414,7 @@ static struct r5conf *setup_conf(struct mddev *mddev) conf->shrinker.count_objects = raid5_cache_count; conf->shrinker.batch = 128; conf->shrinker.flags = 0; - ret = register_shrinker(&conf->shrinker); + ret = register_shrinker(&conf->shrinker, "md-raid5:%s", mdname(mddev)); if (ret) { pr_warn("md/raid:%s: couldn't register shrinker.\n", mdname(mddev)); diff --git a/drivers/misc/vmw_balloon.c b/drivers/misc/vmw_balloon.c index 086ce77d9074..c2d2fa114e65 100644 --- a/drivers/misc/vmw_balloon.c +++ b/drivers/misc/vmw_balloon.c @@ -1587,7 +1587,7 @@ static int vmballoon_register_shrinker(struct vmballoon *b) b->shrinker.count_objects = vmballoon_shrinker_count; b->shrinker.seeks = DEFAULT_SEEKS; - r = register_shrinker(&b->shrinker); + r = register_shrinker(&b->shrinker, "vmw-balloon"); if (r == 0) b->shrinker_registered = true; diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c index b9737da6c4dd..cba57b1f382f 100644 --- a/drivers/virtio/virtio_balloon.c +++ b/drivers/virtio/virtio_balloon.c @@ -875,7 +875,7 @@ static int virtio_balloon_register_shrinker(struct virtio_balloon *vb) vb->shrinker.count_objects = virtio_balloon_shrinker_count; vb->shrinker.seeks = DEFAULT_SEEKS; - return register_shrinker(&vb->shrinker); + return register_shrinker(&vb->shrinker, "virtio-balloon"); } static int virtballoon_probe(struct virtio_device *vdev) diff --git a/drivers/xen/xenbus/xenbus_probe_backend.c b/drivers/xen/xenbus/xenbus_probe_backend.c index 5abded97e1a7..9c09f89d8278 100644 --- a/drivers/xen/xenbus/xenbus_probe_backend.c +++ b/drivers/xen/xenbus/xenbus_probe_backend.c @@ -305,7 +305,7 @@ static int __init xenbus_probe_backend_init(void) register_xenstore_notifier(&xenstore_notifier); - if (register_shrinker(&backend_memory_shrinker)) + if (register_shrinker(&backend_memory_shrinker, "xen-backend")) pr_warn("shrinker registration failed\n"); return 0; diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 6627dd7875ee..eee3e96d877f 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -1815,6 +1815,8 @@ static struct dentry *btrfs_mount_root(struct file_system_type *fs_type, error = -EBUSY; } else { snprintf(s->s_id, sizeof(s->s_id), "%pg", bdev); + shrinker_debugfs_rename(&s->s_shrink, "sb-%s:%s", fs_type->name, + s->s_id); btrfs_sb(s)->bdev_holder = fs_type; if (!strstr(crc32c_impl(), "generic")) set_bit(BTRFS_FS_CSUM_IMPL_FAST, &fs_info->flags); diff --git a/fs/erofs/utils.c b/fs/erofs/utils.c index ec9a1d780dc1..46627cb69abe 100644 --- a/fs/erofs/utils.c +++ b/fs/erofs/utils.c @@ -282,7 +282,7 @@ static struct shrinker erofs_shrinker_info = { int __init erofs_init_shrinker(void) { - return register_shrinker(&erofs_shrinker_info); + return register_shrinker(&erofs_shrinker_info, "erofs-shrinker"); } void erofs_exit_shrinker(void) diff --git a/fs/ext4/extents_status.c b/fs/ext4/extents_status.c index 9a3a8996aacf..23167efda95e 100644 --- a/fs/ext4/extents_status.c +++ b/fs/ext4/extents_status.c @@ -1654,7 +1654,8 @@ int ext4_es_register_shrinker(struct ext4_sb_info *sbi) sbi->s_es_shrinker.scan_objects = ext4_es_scan; sbi->s_es_shrinker.count_objects = ext4_es_count; sbi->s_es_shrinker.seeks = DEFAULT_SEEKS; - err = register_shrinker(&sbi->s_es_shrinker); + err = register_shrinker(&sbi->s_es_shrinker, "ext4-es:%s", + sbi->s_sb->s_id); if (err) goto err4; diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 37221e94e5ef..bce02306f7a0 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -4579,7 +4579,7 @@ static int __init init_f2fs_fs(void) err = f2fs_init_sysfs(); if (err) goto free_garbage_collection_cache; - err = register_shrinker(&f2fs_shrinker_info); + err = register_shrinker(&f2fs_shrinker_info, "f2fs-shrinker"); if (err) goto free_sysfs; err = register_filesystem(&f2fs_fs_type); diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index c992d53013d3..dca842379cab 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -2533,7 +2533,7 @@ int __init gfs2_glock_init(void) return -ENOMEM; } - ret = register_shrinker(&glock_shrinker); + ret = register_shrinker(&glock_shrinker, "gfs2-glock"); if (ret) { destroy_workqueue(gfs2_delete_workqueue); destroy_workqueue(glock_workqueue); diff --git a/fs/gfs2/main.c b/fs/gfs2/main.c index 244187e3e70f..b66a3e1ec152 100644 --- a/fs/gfs2/main.c +++ b/fs/gfs2/main.c @@ -148,7 +148,7 @@ static int __init init_gfs2_fs(void) if (!gfs2_trans_cachep) goto fail_cachep8; - error = register_shrinker(&gfs2_qd_shrinker); + error = register_shrinker(&gfs2_qd_shrinker, "gfs2-qd"); if (error) goto fail_shrinker; diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index c0cbeeaec2d1..45e4655c8033 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -1418,7 +1418,8 @@ static journal_t *journal_init_common(struct block_device *bdev, if (percpu_counter_init(&journal->j_checkpoint_jh_count, 0, GFP_KERNEL)) goto err_cleanup; - if (register_shrinker(&journal->j_shrinker)) { + if (register_shrinker(&journal->j_shrinker, "jbd2-journal:(%u:%u)", + MAJOR(bdev->bd_dev), MINOR(bdev->bd_dev))) { percpu_counter_destroy(&journal->j_checkpoint_jh_count); goto err_cleanup; } diff --git a/fs/mbcache.c b/fs/mbcache.c index 97c54d3a2227..0b833da0a9a5 100644 --- a/fs/mbcache.c +++ b/fs/mbcache.c @@ -367,7 +367,7 @@ struct mb_cache *mb_cache_create(int bucket_bits) cache->c_shrink.count_objects = mb_cache_count; cache->c_shrink.scan_objects = mb_cache_scan; cache->c_shrink.seeks = DEFAULT_SEEKS; - if (register_shrinker(&cache->c_shrink)) { + if (register_shrinker(&cache->c_shrink, "mbcache-shrinker")) { kfree(cache->c_hash); kfree(cache); goto err_out; diff --git a/fs/nfs/nfs42xattr.c b/fs/nfs/nfs42xattr.c index e7b34f7e0614..a9bf09fdf2c3 100644 --- a/fs/nfs/nfs42xattr.c +++ b/fs/nfs/nfs42xattr.c @@ -1017,15 +1017,16 @@ int __init nfs4_xattr_cache_init(void) if (ret) goto out2; - ret = register_shrinker(&nfs4_xattr_cache_shrinker); + ret = register_shrinker(&nfs4_xattr_cache_shrinker, "nfs-xattr_cache"); if (ret) goto out1; - ret = register_shrinker(&nfs4_xattr_entry_shrinker); + ret = register_shrinker(&nfs4_xattr_entry_shrinker, "nfs-xattr_entry"); if (ret) goto out; - ret = register_shrinker(&nfs4_xattr_large_entry_shrinker); + ret = register_shrinker(&nfs4_xattr_large_entry_shrinker, + "nfs-xattr_large_entry"); if (!ret) return 0; diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 6ab5eeb000dc..82944e14fcea 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -149,7 +149,7 @@ int __init register_nfs_fs(void) ret = nfs_register_sysctl(); if (ret < 0) goto error_2; - ret = register_shrinker(&acl_shrinker); + ret = register_shrinker(&acl_shrinker, "nfs-acl"); if (ret < 0) goto error_3; #ifdef CONFIG_NFS_V4_2 diff --git a/fs/nfsd/filecache.c b/fs/nfsd/filecache.c index 9cb2d590c036..a605c0e39b09 100644 --- a/fs/nfsd/filecache.c +++ b/fs/nfsd/filecache.c @@ -670,7 +670,7 @@ nfsd_file_cache_init(void) goto out_err; } - ret = register_shrinker(&nfsd_file_shrinker); + ret = register_shrinker(&nfsd_file_shrinker, "nfsd-filecache"); if (ret) { pr_err("nfsd: failed to register nfsd_file_shrinker: %d\n", ret); goto out_lru; diff --git a/fs/nfsd/nfscache.c b/fs/nfsd/nfscache.c index 7da88bdc0d6c..9b31e1103e7b 100644 --- a/fs/nfsd/nfscache.c +++ b/fs/nfsd/nfscache.c @@ -176,7 +176,8 @@ int nfsd_reply_cache_init(struct nfsd_net *nn) nn->nfsd_reply_cache_shrinker.scan_objects = nfsd_reply_cache_scan; nn->nfsd_reply_cache_shrinker.count_objects = nfsd_reply_cache_count; nn->nfsd_reply_cache_shrinker.seeks = 1; - status = register_shrinker(&nn->nfsd_reply_cache_shrinker); + status = register_shrinker(&nn->nfsd_reply_cache_shrinker, + "nfsd-reply:%s", nn->nfsd_name); if (status) goto out_stats_destroy; diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c index 09d1307959d0..e0b659900e70 100644 --- a/fs/quota/dquot.c +++ b/fs/quota/dquot.c @@ -2995,7 +2995,7 @@ static int __init dquot_init(void) pr_info("VFS: Dquot-cache hash table entries: %ld (order %ld," " %ld bytes)\n", nr_hash, order, (PAGE_SIZE << order)); - if (register_shrinker(&dqcache_shrinker)) + if (register_shrinker(&dqcache_shrinker, "dquota-cache")) panic("Cannot register dquot shrinker"); return 0; diff --git a/fs/super.c b/fs/super.c index 60f57c7bc0a6..4fca6657f442 100644 --- a/fs/super.c +++ b/fs/super.c @@ -265,7 +265,7 @@ static struct super_block *alloc_super(struct file_system_type *type, int flags, s->s_shrink.count_objects = super_cache_count; s->s_shrink.batch = 1024; s->s_shrink.flags = SHRINKER_NUMA_AWARE | SHRINKER_MEMCG_AWARE; - if (prealloc_shrinker(&s->s_shrink)) + if (prealloc_shrinker(&s->s_shrink, "sb-%s", type->name)) goto fail; if (list_lru_init_memcg(&s->s_dentry_lru, &s->s_shrink)) goto fail; @@ -1288,6 +1288,8 @@ int get_tree_bdev(struct fs_context *fc, } else { s->s_mode = mode; snprintf(s->s_id, sizeof(s->s_id), "%pg", bdev); + shrinker_debugfs_rename(&s->s_shrink, "sb-%s:%s", + fc->fs_type->name, s->s_id); sb_set_blocksize(s, block_size(bdev)); error = fill_super(s, fc); if (error) { @@ -1363,6 +1365,8 @@ struct dentry *mount_bdev(struct file_system_type *fs_type, } else { s->s_mode = mode; snprintf(s->s_id, sizeof(s->s_id), "%pg", bdev); + shrinker_debugfs_rename(&s->s_shrink, "sb-%s:%s", + fs_type->name, s->s_id); sb_set_blocksize(s, block_size(bdev)); error = fill_super(s, data, flags & SB_SILENT ? 1 : 0); if (error) { diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 0978d01b0ea4..d0c9a09988bc 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -2430,7 +2430,7 @@ static int __init ubifs_init(void) if (!ubifs_inode_slab) return -ENOMEM; - err = register_shrinker(&ubifs_shrinker_info); + err = register_shrinker(&ubifs_shrinker_info, "ubifs-slab"); if (err) goto out_slab; diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index bf4e60871068..4aa9c9cf5b6e 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1986,7 +1986,8 @@ xfs_alloc_buftarg( btp->bt_shrinker.scan_objects = xfs_buftarg_shrink_scan; btp->bt_shrinker.seeks = DEFAULT_SEEKS; btp->bt_shrinker.flags = SHRINKER_NUMA_AWARE; - if (register_shrinker(&btp->bt_shrinker)) + if (register_shrinker(&btp->bt_shrinker, "xfs-buf:%s", + mp->m_super->s_id)) goto error_pcpu; return btp; diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index 5269354b1b69..a1941c8b8630 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -2201,5 +2201,5 @@ xfs_inodegc_register_shrinker( shrink->flags = SHRINKER_NONSLAB; shrink->batch = XFS_INODEGC_SHRINKER_BATCH; - return register_shrinker(shrink); + return register_shrinker(shrink, "xfs-inodegc:%s", mp->m_super->s_id); } diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index abf08bbf34a9..c31d57453ceb 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -677,7 +677,8 @@ xfs_qm_init_quotainfo( qinf->qi_shrinker.seeks = DEFAULT_SEEKS; qinf->qi_shrinker.flags = SHRINKER_NUMA_AWARE; - error = register_shrinker(&qinf->qi_shrinker); + error = register_shrinker(&qinf->qi_shrinker, "xfs-qm:%s", + mp->m_super->s_id); if (error) goto out_free_inos; diff --git a/include/linux/shrinker.h b/include/linux/shrinker.h index 2ced8149c513..64416f3e0a1f 100644 --- a/include/linux/shrinker.h +++ b/include/linux/shrinker.h @@ -75,6 +75,7 @@ struct shrinker { #endif #ifdef CONFIG_SHRINKER_DEBUG int debugfs_id; + const char *name; struct dentry *debugfs_entry; #endif /* objs pending delete, per node */ @@ -92,9 +93,9 @@ struct shrinker { */ #define SHRINKER_NONSLAB (1 << 3) -extern int prealloc_shrinker(struct shrinker *shrinker); +extern int prealloc_shrinker(struct shrinker *shrinker, const char *fmt, ...); extern void register_shrinker_prepared(struct shrinker *shrinker); -extern int register_shrinker(struct shrinker *shrinker); +extern int register_shrinker(struct shrinker *shrinker, const char *fmt, ...); extern void unregister_shrinker(struct shrinker *shrinker); extern void free_prealloced_shrinker(struct shrinker *shrinker); extern void synchronize_shrinkers(void); @@ -102,6 +103,8 @@ extern void synchronize_shrinkers(void); #ifdef CONFIG_SHRINKER_DEBUG extern int shrinker_debugfs_add(struct shrinker *shrinker); extern void shrinker_debugfs_remove(struct shrinker *shrinker); +extern int shrinker_debugfs_rename(struct shrinker *shrinker, + const char *fmt, ...); #else /* CONFIG_SHRINKER_DEBUG */ static inline int shrinker_debugfs_add(struct shrinker *shrinker) { @@ -110,5 +113,10 @@ static inline int shrinker_debugfs_add(struct shrinker *shrinker) static inline void shrinker_debugfs_remove(struct shrinker *shrinker) { } +static inline int shrinker_debugfs_rename(struct shrinker *shrinker, + const char *fmt, ...) +{ + return 0; +} #endif /* CONFIG_SHRINKER_DEBUG */ #endif /* _LINUX_SHRINKER_H */ diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c index c25ba442044a..4b3bf6ebb1eb 100644 --- a/kernel/rcu/tree.c +++ b/kernel/rcu/tree.c @@ -4884,7 +4884,7 @@ static void __init kfree_rcu_batch_init(void) INIT_DELAYED_WORK(&krcp->page_cache_work, fill_page_cache_func); krcp->initialized = true; } - if (register_shrinker(&kfree_rcu_shrinker)) + if (register_shrinker(&kfree_rcu_shrinker, "rcu-kfree")) pr_err("Failed to register kfree_rcu() shrinker!\n"); } diff --git a/mm/huge_memory.c b/mm/huge_memory.c index f71eeeda9533..d624b2544751 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -423,10 +423,10 @@ static int __init hugepage_init(void) if (err) goto err_slab; - err = register_shrinker(&huge_zero_page_shrinker); + err = register_shrinker(&huge_zero_page_shrinker, "thp-zero"); if (err) goto err_hzp_shrinker; - err = register_shrinker(&deferred_split_shrinker); + err = register_shrinker(&deferred_split_shrinker, "thp-deferred_split"); if (err) goto err_split_shrinker; diff --git a/mm/shrinker_debug.c b/mm/shrinker_debug.c index 1a70556bd46c..781ecbd3d608 100644 --- a/mm/shrinker_debug.c +++ b/mm/shrinker_debug.c @@ -102,7 +102,7 @@ DEFINE_SHOW_ATTRIBUTE(shrinker_debugfs_count); int shrinker_debugfs_add(struct shrinker *shrinker) { struct dentry *entry; - char buf[16]; + char buf[128]; int id; lockdep_assert_held(&shrinker_rwsem); @@ -116,7 +116,7 @@ int shrinker_debugfs_add(struct shrinker *shrinker) return id; shrinker->debugfs_id = id; - snprintf(buf, sizeof(buf), "%d", id); + snprintf(buf, sizeof(buf), "%s-%d", shrinker->name, id); /* create debugfs entry */ entry = debugfs_create_dir(buf, shrinker_debugfs_root); @@ -131,10 +131,53 @@ int shrinker_debugfs_add(struct shrinker *shrinker) return 0; } +int shrinker_debugfs_rename(struct shrinker *shrinker, const char *fmt, ...) +{ + struct dentry *entry; + char buf[128]; + const char *new, *old; + va_list ap; + int ret = 0; + + va_start(ap, fmt); + new = kvasprintf_const(GFP_KERNEL, fmt, ap); + va_end(ap); + + if (!new) + return -ENOMEM; + + down_write(&shrinker_rwsem); + + old = shrinker->name; + shrinker->name = new; + + if (shrinker->debugfs_entry) { + snprintf(buf, sizeof(buf), "%s-%d", shrinker->name, + shrinker->debugfs_id); + + entry = debugfs_rename(shrinker_debugfs_root, + shrinker->debugfs_entry, + shrinker_debugfs_root, buf); + if (IS_ERR(entry)) + ret = PTR_ERR(entry); + else + shrinker->debugfs_entry = entry; + } + + up_write(&shrinker_rwsem); + + kfree_const(old); + + return ret; +} +EXPORT_SYMBOL(shrinker_debugfs_rename); + void shrinker_debugfs_remove(struct shrinker *shrinker) { lockdep_assert_held(&shrinker_rwsem); + kfree_const(shrinker->name); + if (!shrinker->debugfs_entry) return; diff --git a/mm/vmscan.c b/mm/vmscan.c index 35dedff79eb4..97ac6c6c026d 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -608,7 +608,7 @@ static unsigned long lruvec_lru_size(struct lruvec *lruvec, enum lru_list lru, /* * Add a shrinker callback to be called from the vm. */ -int prealloc_shrinker(struct shrinker *shrinker) +static int __prealloc_shrinker(struct shrinker *shrinker) { unsigned int size; int err; @@ -632,8 +632,36 @@ int prealloc_shrinker(struct shrinker *shrinker) return 0; } +#ifdef CONFIG_SHRINKER_DEBUG +int prealloc_shrinker(struct shrinker *shrinker, const char *fmt, ...) +{ + va_list ap; + int err; + + va_start(ap, fmt); + shrinker->name = kvasprintf_const(GFP_KERNEL, fmt, ap); + va_end(ap); + if (!shrinker->name) + return -ENOMEM; + + err = __prealloc_shrinker(shrinker); + if (err) + kfree_const(shrinker->name); + + return err; +} +#else +int prealloc_shrinker(struct shrinker *shrinker, const char *fmt, ...) +{ + return __prealloc_shrinker(shrinker); +} +#endif + void free_prealloced_shrinker(struct shrinker *shrinker) { +#ifdef CONFIG_SHRINKER_DEBUG + kfree_const(shrinker->name); +#endif if (shrinker->flags & SHRINKER_MEMCG_AWARE) { down_write(&shrinker_rwsem); unregister_memcg_shrinker(shrinker); @@ -654,15 +682,39 @@ void register_shrinker_prepared(struct shrinker *shrinker) up_write(&shrinker_rwsem); } -int register_shrinker(struct shrinker *shrinker) +static int __register_shrinker(struct shrinker *shrinker) { - int err = prealloc_shrinker(shrinker); + int err = __prealloc_shrinker(shrinker); if (err) return err; register_shrinker_prepared(shrinker); return 0; } + +#ifdef CONFIG_SHRINKER_DEBUG +int register_shrinker(struct shrinker *shrinker, const char *fmt, ...) +{ + va_list ap; + int err; + + va_start(ap, fmt); + shrinker->name = kvasprintf_const(GFP_KERNEL, fmt, ap); + va_end(ap); + if (!shrinker->name) + return -ENOMEM; + + err = __register_shrinker(shrinker); + if (err) + kfree_const(shrinker->name); + return err; +} +#else +int register_shrinker(struct shrinker *shrinker, const char *fmt, ...) +{ + return __register_shrinker(shrinker); +} +#endif EXPORT_SYMBOL(register_shrinker); /* diff --git a/mm/workingset.c b/mm/workingset.c index 592569a8974c..a5e84862fc86 100644 --- a/mm/workingset.c +++ b/mm/workingset.c @@ -625,7 +625,7 @@ static int __init workingset_init(void) pr_info("workingset: timestamp_bits=%d max_order=%d bucket_order=%u\n", timestamp_bits, max_order, bucket_order); - ret = prealloc_shrinker(&workingset_shadow_shrinker); + ret = prealloc_shrinker(&workingset_shadow_shrinker, "mm-shadow"); if (ret) goto err; ret = __list_lru_init(&shadow_nodes, true, &shadow_nodes_key, diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c index 5d5fc04385b8..f24b71568e83 100644 --- a/mm/zsmalloc.c +++ b/mm/zsmalloc.c @@ -2217,7 +2217,8 @@ static int zs_register_shrinker(struct zs_pool *pool) pool->shrinker.batch = 0; pool->shrinker.seeks = DEFAULT_SEEKS; - return register_shrinker(&pool->shrinker); + return register_shrinker(&pool->shrinker, "mm-zspool:%s", + pool->name); } /** diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c index 682fcd24bf43..04e7b55fe0d9 100644 --- a/net/sunrpc/auth.c +++ b/net/sunrpc/auth.c @@ -874,7 +874,7 @@ int __init rpcauth_init_module(void) err = rpc_init_authunix(); if (err < 0) goto out1; - err = register_shrinker(&rpc_cred_shrinker); + err = register_shrinker(&rpc_cred_shrinker, "sunrpc_cred"); if (err < 0) goto out2; return 0; -- cgit v1.2.3 From 60cc9efaa8e7877568f0cbcf6270b5c858e15009 Mon Sep 17 00:00:00 2001 From: Suren Baghdasaryan Date: Tue, 31 May 2022 15:30:59 -0700 Subject: mm: drop oom code from exit_mmap The primary reason to invoke the oom reaper from the exit_mmap path used to be a prevention of an excessive oom killing if the oom victim exit races with the oom reaper (see [1] for more details). The invocation has moved around since then because of the interaction with the munlock logic but the underlying reason has remained the same (see [2]). Munlock code is no longer a problem since [3] and there shouldn't be any blocking operation before the memory is unmapped by exit_mmap so the oom reaper invocation can be dropped. The unmapping part can be done with the non-exclusive mmap_sem and the exclusive one is only required when page tables are freed. Remove the oom_reaper from exit_mmap which will make the code easier to read. This is really unlikely to make any observable difference although some microbenchmarks could benefit from one less branch that needs to be evaluated even though it almost never is true. [1] 212925802454 ("mm: oom: let oom_reap_task and exit_mmap run concurrently") [2] 27ae357fa82b ("mm, oom: fix concurrent munlock and oom reaper unmap, v3") [3] a213e5cf71cb ("mm/munlock: delete munlock_vma_pages_all(), allow oomreap") Link: https://lkml.kernel.org/r/20220531223100.510392-1-surenb@google.com Signed-off-by: Suren Baghdasaryan Acked-by: Michal Hocko Cc: Andrea Arcangeli Cc: Christian Brauner (Microsoft) Cc: Christoph Hellwig Cc: David Hildenbrand Cc: David Rientjes Cc: Jann Horn Cc: Johannes Weiner Cc: John Hubbard Cc: "Kirill A . Shutemov" Cc: Liam Howlett Cc: Matthew Wilcox Cc: Minchan Kim Cc: Oleg Nesterov Cc: Peter Xu Cc: Roman Gushchin Cc: Shakeel Butt Cc: Shuah Khan Signed-off-by: Andrew Morton --- include/linux/oom.h | 2 -- mm/mmap.c | 24 +++++++----------------- mm/oom_kill.c | 2 +- 3 files changed, 8 insertions(+), 20 deletions(-) (limited to 'include/linux') diff --git a/include/linux/oom.h b/include/linux/oom.h index 02d1e7bbd8cd..6cdde62b078b 100644 --- a/include/linux/oom.h +++ b/include/linux/oom.h @@ -106,8 +106,6 @@ static inline vm_fault_t check_stable_address_space(struct mm_struct *mm) return 0; } -bool __oom_reap_task_mm(struct mm_struct *mm); - long oom_badness(struct task_struct *p, unsigned long totalpages); diff --git a/mm/mmap.c b/mm/mmap.c index 5fd97a3cee6d..b357fb88706e 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -3196,23 +3196,6 @@ void exit_mmap(struct mm_struct *mm) /* mm's last user has gone, and its about to be pulled down */ mmu_notifier_release(mm); - if (unlikely(mm_is_oom_victim(mm))) { - /* - * Manually reap the mm to free as much memory as possible. - * Then, as the oom reaper does, set MMF_OOM_SKIP to disregard - * this mm from further consideration. Taking mm->mmap_lock for - * write after setting MMF_OOM_SKIP will guarantee that the oom - * reaper will not run on this mm again after mmap_lock is - * dropped. - * - * Nothing can be holding mm->mmap_lock here and the above call - * to mmu_notifier_release(mm) ensures mmu notifier callbacks in - * __oom_reap_task_mm() will not block. - */ - (void)__oom_reap_task_mm(mm); - set_bit(MMF_OOM_SKIP, &mm->flags); - } - mmap_write_lock(mm); arch_exit_mmap(mm); @@ -3229,6 +3212,13 @@ void exit_mmap(struct mm_struct *mm) /* update_hiwater_rss(mm) here? but nobody should be looking */ /* Use ULONG_MAX here to ensure all VMAs in the mm are unmapped */ unmap_vmas(&tlb, &mm->mm_mt, vma, 0, ULONG_MAX); + + /* + * Set MMF_OOM_SKIP to hide this task from the oom killer/reaper + * because the memory has been already freed. Do not bother checking + * mm_is_oom_victim because setting a bit unconditionally is cheaper. + */ + set_bit(MMF_OOM_SKIP, &mm->flags); free_pgtables(&tlb, &mm->mm_mt, vma, FIRST_USER_ADDRESS, USER_PGTABLES_CEILING); tlb_finish_mmu(&tlb); diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 3996301450e8..decb21474c6c 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -509,7 +509,7 @@ static DECLARE_WAIT_QUEUE_HEAD(oom_reaper_wait); static struct task_struct *oom_reaper_list; static DEFINE_SPINLOCK(oom_reaper_lock); -bool __oom_reap_task_mm(struct mm_struct *mm) +static bool __oom_reap_task_mm(struct mm_struct *mm) { struct vm_area_struct *vma; bool ret = true; -- cgit v1.2.3 From d6e900df72f46b9f6fa81159188baa3ed9d0e551 Mon Sep 17 00:00:00 2001 From: Suren Baghdasaryan Date: Tue, 31 May 2022 15:31:00 -0700 Subject: mm: delete unused MMF_OOM_VICTIM flag With the last usage of MMF_OOM_VICTIM in exit_mmap gone, this flag is now unused and can be removed. Link: https://lkml.kernel.org/r/20220531223100.510392-2-surenb@google.com Signed-off-by: Suren Baghdasaryan Acked-by: Michal Hocko Cc: David Rientjes Cc: Matthew Wilcox Cc: Johannes Weiner Cc: Roman Gushchin Cc: Minchan Kim Cc: "Kirill A . Shutemov" Cc: Andrea Arcangeli Cc: Christian Brauner (Microsoft) Cc: Christoph Hellwig Cc: Oleg Nesterov Cc: David Hildenbrand Cc: Jann Horn Cc: Shakeel Butt Cc: Peter Xu Cc: John Hubbard Cc: Shuah Khan Cc: Liam Howlett Signed-off-by: Andrew Morton --- include/linux/oom.h | 9 --------- include/linux/sched/coredump.h | 7 +++---- mm/oom_kill.c | 4 +--- 3 files changed, 4 insertions(+), 16 deletions(-) (limited to 'include/linux') diff --git a/include/linux/oom.h b/include/linux/oom.h index 6cdde62b078b..7d0c9c48a0c5 100644 --- a/include/linux/oom.h +++ b/include/linux/oom.h @@ -77,15 +77,6 @@ static inline bool tsk_is_oom_victim(struct task_struct * tsk) return tsk->signal->oom_mm; } -/* - * Use this helper if tsk->mm != mm and the victim mm needs a special - * handling. This is guaranteed to stay true after once set. - */ -static inline bool mm_is_oom_victim(struct mm_struct *mm) -{ - return test_bit(MMF_OOM_VICTIM, &mm->flags); -} - /* * Checks whether a page fault on the given mm is still reliable. * This is no longer true if the oom reaper started to reap the diff --git a/include/linux/sched/coredump.h b/include/linux/sched/coredump.h index 4d0a5be28b70..8270ad7ae14c 100644 --- a/include/linux/sched/coredump.h +++ b/include/linux/sched/coredump.h @@ -71,9 +71,8 @@ static inline int get_dumpable(struct mm_struct *mm) #define MMF_UNSTABLE 22 /* mm is unstable for copy_from_user */ #define MMF_HUGE_ZERO_PAGE 23 /* mm has ever used the global huge zero page */ #define MMF_DISABLE_THP 24 /* disable THP for all VMAs */ -#define MMF_OOM_VICTIM 25 /* mm is the oom victim */ -#define MMF_OOM_REAP_QUEUED 26 /* mm was queued for oom_reaper */ -#define MMF_MULTIPROCESS 27 /* mm is shared between processes */ +#define MMF_OOM_REAP_QUEUED 25 /* mm was queued for oom_reaper */ +#define MMF_MULTIPROCESS 26 /* mm is shared between processes */ /* * MMF_HAS_PINNED: Whether this mm has pinned any pages. This can be either * replaced in the future by mm.pinned_vm when it becomes stable, or grow into @@ -81,7 +80,7 @@ static inline int get_dumpable(struct mm_struct *mm) * pinned pages were unpinned later on, we'll still keep this bit set for the * lifecycle of this mm, just for simplicity. */ -#define MMF_HAS_PINNED 28 /* FOLL_PIN has run, never cleared */ +#define MMF_HAS_PINNED 27 /* FOLL_PIN has run, never cleared */ #define MMF_DISABLE_THP_MASK (1 << MMF_DISABLE_THP) #define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK |\ diff --git a/mm/oom_kill.c b/mm/oom_kill.c index decb21474c6c..35ec75cdfee2 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -765,10 +765,8 @@ static void mark_oom_victim(struct task_struct *tsk) return; /* oom_mm is bound to the signal struct life time. */ - if (!cmpxchg(&tsk->signal->oom_mm, NULL, mm)) { + if (!cmpxchg(&tsk->signal->oom_mm, NULL, mm)) mmgrab(tsk->signal->oom_mm); - set_bit(MMF_OOM_VICTIM, &mm->flags); - } /* * Make sure that the task is woken up from uninterruptible sleep -- cgit v1.2.3 From ee709c64ff2ffe9e3a12cec26153726006db8491 Mon Sep 17 00:00:00 2001 From: Shiyang Ruan Date: Fri, 3 Jun 2022 13:37:25 +0800 Subject: dax: introduce holder for dax_device Patch series "v14 fsdax-rmap + v11 fsdax-reflink", v2. The patchset fsdax-rmap is aimed to support shared pages tracking for fsdax. It moves owner tracking from dax_assocaite_entry() to pmem device driver, by introducing an interface ->memory_failure() for struct pagemap. This interface is called by memory_failure() in mm, and implemented by pmem device. Then call holder operations to find the filesystem which the corrupted data located in, and call filesystem handler to track files or metadata associated with this page. Finally we are able to try to fix the corrupted data in filesystem and do other necessary processing, such as killing processes who are using the files affected. The call trace is like this: memory_failure() |* fsdax case |------------ |pgmap->ops->memory_failure() => pmem_pgmap_memory_failure() | dax_holder_notify_failure() => | dax_device->holder_ops->notify_failure() => | - xfs_dax_notify_failure() | |* xfs_dax_notify_failure() | |-------------------------- | | xfs_rmap_query_range() | | xfs_dax_failure_fn() | | * corrupted on metadata | | try to recover data, call xfs_force_shutdown() | | * corrupted on file data | | try to recover data, call mf_dax_kill_procs() |* normal case |------------- |mf_generic_kill_procs() The patchset fsdax-reflink attempts to add CoW support for fsdax, and takes XFS, which has both reflink and fsdax features, as an example. One of the key mechanisms needed to be implemented in fsdax is CoW. Copy the data from srcmap before we actually write data to the destination iomap. And we just copy range in which data won't be changed. Another mechanism is range comparison. In page cache case, readpage() is used to load data on disk to page cache in order to be able to compare data. In fsdax case, readpage() does not work. So, we need another compare data with direct access support. With the two mechanisms implemented in fsdax, we are able to make reflink and fsdax work together in XFS. This patch (of 14): To easily track filesystem from a pmem device, we introduce a holder for dax_device structure, and also its operation. This holder is used to remember who is using this dax_device: - When it is the backend of a filesystem, the holder will be the instance of this filesystem. - When this pmem device is one of the targets in a mapped device, the holder will be this mapped device. In this case, the mapped device has its own dax_device and it will follow the first rule. So that we can finally track to the filesystem we needed. The holder and holder_ops will be set when filesystem is being mounted, or an target device is being activated. Link: https://lkml.kernel.org/r/20220603053738.1218681-1-ruansy.fnst@fujitsu.com Link: https://lkml.kernel.org/r/20220603053738.1218681-2-ruansy.fnst@fujitsu.com Signed-off-by: Shiyang Ruan Reviewed-by: Christoph Hellwig Reviewed-by: Dan Williams Reviewed-by: Darrick J. Wong Cc: Dave Chinner Cc: Jane Chu Cc: Goldwyn Rodrigues Cc: Al Viro Cc: Matthew Wilcox Cc: Naoya Horiguchi Cc: Miaohe Lin Cc: Dan Williams Cc: Goldwyn Rodrigues Cc: Ritesh Harjani Signed-off-by: Andrew Morton --- drivers/dax/super.c | 67 ++++++++++++++++++++++++++++++++++++++++++++++++++++- drivers/md/dm.c | 2 +- fs/erofs/super.c | 10 ++++---- fs/ext2/super.c | 7 +++--- fs/ext4/super.c | 9 +++---- fs/xfs/xfs_buf.c | 5 ++-- include/linux/dax.h | 33 +++++++++++++++++++------- 7 files changed, 110 insertions(+), 23 deletions(-) (limited to 'include/linux') diff --git a/drivers/dax/super.c b/drivers/dax/super.c index 50a08b2ec247..9b5e2a5eb0ae 100644 --- a/drivers/dax/super.c +++ b/drivers/dax/super.c @@ -22,6 +22,8 @@ * @private: dax driver private data * @flags: state and boolean properties * @ops: operations for this device + * @holder_data: holder of a dax_device: could be filesystem or mapped device + * @holder_ops: operations for the inner holder */ struct dax_device { struct inode inode; @@ -29,6 +31,8 @@ struct dax_device { void *private; unsigned long flags; const struct dax_operations *ops; + void *holder_data; + const struct dax_holder_operations *holder_ops; }; static dev_t dax_devt; @@ -71,8 +75,11 @@ EXPORT_SYMBOL_GPL(dax_remove_host); * fs_dax_get_by_bdev() - temporary lookup mechanism for filesystem-dax * @bdev: block device to find a dax_device for * @start_off: returns the byte offset into the dax_device that @bdev starts + * @holder: filesystem or mapped device inside the dax_device + * @ops: operations for the inner holder */ -struct dax_device *fs_dax_get_by_bdev(struct block_device *bdev, u64 *start_off) +struct dax_device *fs_dax_get_by_bdev(struct block_device *bdev, u64 *start_off, + void *holder, const struct dax_holder_operations *ops) { struct dax_device *dax_dev; u64 part_size; @@ -92,11 +99,26 @@ struct dax_device *fs_dax_get_by_bdev(struct block_device *bdev, u64 *start_off) dax_dev = xa_load(&dax_hosts, (unsigned long)bdev->bd_disk); if (!dax_dev || !dax_alive(dax_dev) || !igrab(&dax_dev->inode)) dax_dev = NULL; + else if (holder) { + if (!cmpxchg(&dax_dev->holder_data, NULL, holder)) + dax_dev->holder_ops = ops; + else + dax_dev = NULL; + } dax_read_unlock(id); return dax_dev; } EXPORT_SYMBOL_GPL(fs_dax_get_by_bdev); + +void fs_put_dax(struct dax_device *dax_dev, void *holder) +{ + if (dax_dev && holder && + cmpxchg(&dax_dev->holder_data, holder, NULL) == holder) + dax_dev->holder_ops = NULL; + put_dax(dax_dev); +} +EXPORT_SYMBOL_GPL(fs_put_dax); #endif /* CONFIG_BLOCK && CONFIG_FS_DAX */ enum dax_device_flags { @@ -204,6 +226,29 @@ size_t dax_recovery_write(struct dax_device *dax_dev, pgoff_t pgoff, } EXPORT_SYMBOL_GPL(dax_recovery_write); +int dax_holder_notify_failure(struct dax_device *dax_dev, u64 off, + u64 len, int mf_flags) +{ + int rc, id; + + id = dax_read_lock(); + if (!dax_alive(dax_dev)) { + rc = -ENXIO; + goto out; + } + + if (!dax_dev->holder_ops) { + rc = -EOPNOTSUPP; + goto out; + } + + rc = dax_dev->holder_ops->notify_failure(dax_dev, off, len, mf_flags); +out: + dax_read_unlock(id); + return rc; +} +EXPORT_SYMBOL_GPL(dax_holder_notify_failure); + #ifdef CONFIG_ARCH_HAS_PMEM_API void arch_wb_cache_pmem(void *addr, size_t size); void dax_flush(struct dax_device *dax_dev, void *addr, size_t size) @@ -277,8 +322,15 @@ void kill_dax(struct dax_device *dax_dev) if (!dax_dev) return; + if (dax_dev->holder_data != NULL) + dax_holder_notify_failure(dax_dev, 0, U64_MAX, 0); + clear_bit(DAXDEV_ALIVE, &dax_dev->flags); synchronize_srcu(&dax_srcu); + + /* clear holder data */ + dax_dev->holder_ops = NULL; + dax_dev->holder_data = NULL; } EXPORT_SYMBOL_GPL(kill_dax); @@ -420,6 +472,19 @@ void put_dax(struct dax_device *dax_dev) } EXPORT_SYMBOL_GPL(put_dax); +/** + * dax_holder() - obtain the holder of a dax device + * @dax_dev: a dax_device instance + + * Return: the holder's data which represents the holder if registered, + * otherwize NULL. + */ +void *dax_holder(struct dax_device *dax_dev) +{ + return dax_dev->holder_data; +} +EXPORT_SYMBOL_GPL(dax_holder); + /** * inode_dax: convert a public inode into its dax_dev * @inode: An inode with i_cdev pointing to a dax_dev diff --git a/drivers/md/dm.c b/drivers/md/dm.c index 2b75f1ef7386..0177a4ce9a18 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -758,7 +758,7 @@ static int open_table_device(struct table_device *td, dev_t dev, } td->dm_dev.bdev = bdev; - td->dm_dev.dax_dev = fs_dax_get_by_bdev(bdev, &part_off); + td->dm_dev.dax_dev = fs_dax_get_by_bdev(bdev, &part_off, NULL, NULL); return 0; } diff --git a/fs/erofs/super.c b/fs/erofs/super.c index 95addc5c9d34..3173debeaa5a 100644 --- a/fs/erofs/super.c +++ b/fs/erofs/super.c @@ -255,7 +255,8 @@ static int erofs_init_device(struct erofs_buf *buf, struct super_block *sb, if (IS_ERR(bdev)) return PTR_ERR(bdev); dif->bdev = bdev; - dif->dax_dev = fs_dax_get_by_bdev(bdev, &dif->dax_part_off); + dif->dax_dev = fs_dax_get_by_bdev(bdev, &dif->dax_part_off, + NULL, NULL); } dif->blocks = le32_to_cpu(dis->blocks); @@ -720,7 +721,8 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc) } sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev, - &sbi->dax_part_off); + &sbi->dax_part_off, + NULL, NULL); } err = erofs_read_superblock(sb); @@ -812,7 +814,7 @@ static int erofs_release_device_info(int id, void *ptr, void *data) { struct erofs_device_info *dif = ptr; - fs_put_dax(dif->dax_dev); + fs_put_dax(dif->dax_dev, NULL); if (dif->bdev) blkdev_put(dif->bdev, FMODE_READ | FMODE_EXCL); erofs_fscache_unregister_cookie(&dif->fscache); @@ -886,7 +888,7 @@ static void erofs_kill_sb(struct super_block *sb) return; erofs_free_dev_context(sbi->devs); - fs_put_dax(sbi->dax_dev); + fs_put_dax(sbi->dax_dev, NULL); erofs_fscache_unregister_cookie(&sbi->s_fscache); erofs_fscache_unregister_fs(sb); kfree(sbi->opt.fsid); diff --git a/fs/ext2/super.c b/fs/ext2/super.c index f6a19f6d9f6d..4638946251b9 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -171,7 +171,7 @@ static void ext2_put_super (struct super_block * sb) brelse (sbi->s_sbh); sb->s_fs_info = NULL; kfree(sbi->s_blockgroup_lock); - fs_put_dax(sbi->s_daxdev); + fs_put_dax(sbi->s_daxdev, NULL); kfree(sbi); } @@ -835,7 +835,8 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent) } sb->s_fs_info = sbi; sbi->s_sb_block = sb_block; - sbi->s_daxdev = fs_dax_get_by_bdev(sb->s_bdev, &sbi->s_dax_part_off); + sbi->s_daxdev = fs_dax_get_by_bdev(sb->s_bdev, &sbi->s_dax_part_off, + NULL, NULL); spin_lock_init(&sbi->s_lock); ret = -EINVAL; @@ -1204,7 +1205,7 @@ failed_mount_group_desc: failed_mount: brelse(bh); failed_sbi: - fs_put_dax(sbi->s_daxdev); + fs_put_dax(sbi->s_daxdev, NULL); sb->s_fs_info = NULL; kfree(sbi->s_blockgroup_lock); kfree(sbi); diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 845f2f8aee5f..1f8bf507ba5a 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -1307,7 +1307,7 @@ static void ext4_put_super(struct super_block *sb) if (sbi->s_chksum_driver) crypto_free_shash(sbi->s_chksum_driver); kfree(sbi->s_blockgroup_lock); - fs_put_dax(sbi->s_daxdev); + fs_put_dax(sbi->s_daxdev, NULL); fscrypt_free_dummy_policy(&sbi->s_dummy_enc_policy); #if IS_ENABLED(CONFIG_UNICODE) utf8_unload(sb->s_encoding); @@ -4272,7 +4272,7 @@ static void ext4_free_sbi(struct ext4_sb_info *sbi) return; kfree(sbi->s_blockgroup_lock); - fs_put_dax(sbi->s_daxdev); + fs_put_dax(sbi->s_daxdev, NULL); kfree(sbi); } @@ -4284,7 +4284,8 @@ static struct ext4_sb_info *ext4_alloc_sbi(struct super_block *sb) if (!sbi) return NULL; - sbi->s_daxdev = fs_dax_get_by_bdev(sb->s_bdev, &sbi->s_dax_part_off); + sbi->s_daxdev = fs_dax_get_by_bdev(sb->s_bdev, &sbi->s_dax_part_off, + NULL, NULL); sbi->s_blockgroup_lock = kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL); @@ -4296,7 +4297,7 @@ static struct ext4_sb_info *ext4_alloc_sbi(struct super_block *sb) sbi->s_sb = sb; return sbi; err_out: - fs_put_dax(sbi->s_daxdev); + fs_put_dax(sbi->s_daxdev, NULL); kfree(sbi); return NULL; } diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 4aa9c9cf5b6e..1ec2a7b6d44e 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1911,7 +1911,7 @@ xfs_free_buftarg( list_lru_destroy(&btp->bt_lru); blkdev_issue_flush(btp->bt_bdev); - fs_put_dax(btp->bt_daxdev); + fs_put_dax(btp->bt_daxdev, NULL); kmem_free(btp); } @@ -1964,7 +1964,8 @@ xfs_alloc_buftarg( btp->bt_mount = mp; btp->bt_dev = bdev->bd_dev; btp->bt_bdev = bdev; - btp->bt_daxdev = fs_dax_get_by_bdev(bdev, &btp->bt_dax_part_off); + btp->bt_daxdev = fs_dax_get_by_bdev(bdev, &btp->bt_dax_part_off, NULL, + NULL); /* * Buffer IO error rate limiting. Limit it to no more than 10 messages diff --git a/include/linux/dax.h b/include/linux/dax.h index e7b81634c52a..cf85fc36da5f 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -43,8 +43,21 @@ struct dax_operations { void *addr, size_t bytes, struct iov_iter *iter); }; +struct dax_holder_operations { + /* + * notify_failure - notify memory failure into inner holder device + * @dax_dev: the dax device which contains the holder + * @offset: offset on this dax device where memory failure occurs + * @len: length of this memory failure event + * @flags: action flags for memory failure handler + */ + int (*notify_failure)(struct dax_device *dax_dev, u64 offset, + u64 len, int mf_flags); +}; + #if IS_ENABLED(CONFIG_DAX) struct dax_device *alloc_dax(void *private, const struct dax_operations *ops); +void *dax_holder(struct dax_device *dax_dev); void put_dax(struct dax_device *dax_dev); void kill_dax(struct dax_device *dax_dev); void dax_write_cache(struct dax_device *dax_dev, bool wc); @@ -66,6 +79,10 @@ static inline bool daxdev_mapping_supported(struct vm_area_struct *vma, return dax_synchronous(dax_dev); } #else +static inline void *dax_holder(struct dax_device *dax_dev) +{ + return NULL; +} static inline struct dax_device *alloc_dax(void *private, const struct dax_operations *ops) { @@ -114,12 +131,9 @@ struct writeback_control; #if defined(CONFIG_BLOCK) && defined(CONFIG_FS_DAX) int dax_add_host(struct dax_device *dax_dev, struct gendisk *disk); void dax_remove_host(struct gendisk *disk); -struct dax_device *fs_dax_get_by_bdev(struct block_device *bdev, - u64 *start_off); -static inline void fs_put_dax(struct dax_device *dax_dev) -{ - put_dax(dax_dev); -} +struct dax_device *fs_dax_get_by_bdev(struct block_device *bdev, u64 *start_off, + void *holder, const struct dax_holder_operations *ops); +void fs_put_dax(struct dax_device *dax_dev, void *holder); #else static inline int dax_add_host(struct dax_device *dax_dev, struct gendisk *disk) { @@ -129,11 +143,12 @@ static inline void dax_remove_host(struct gendisk *disk) { } static inline struct dax_device *fs_dax_get_by_bdev(struct block_device *bdev, - u64 *start_off) + u64 *start_off, void *holder, + const struct dax_holder_operations *ops) { return NULL; } -static inline void fs_put_dax(struct dax_device *dax_dev) +static inline void fs_put_dax(struct dax_device *dax_dev, void *holder) { } #endif /* CONFIG_BLOCK && CONFIG_FS_DAX */ @@ -203,6 +218,8 @@ size_t dax_copy_to_iter(struct dax_device *dax_dev, pgoff_t pgoff, void *addr, size_t bytes, struct iov_iter *i); int dax_zero_page_range(struct dax_device *dax_dev, pgoff_t pgoff, size_t nr_pages); +int dax_holder_notify_failure(struct dax_device *dax_dev, u64 off, u64 len, + int mf_flags); void dax_flush(struct dax_device *dax_dev, void *addr, size_t size); ssize_t dax_iomap_rw(struct kiocb *iocb, struct iov_iter *iter, -- cgit v1.2.3 From 3396e82c576eb4320149e2f7b80ae8f4bc1c321c Mon Sep 17 00:00:00 2001 From: Shiyang Ruan Date: Fri, 3 Jun 2022 13:37:27 +0800 Subject: pagemap,pmem: introduce ->memory_failure() When memory-failure occurs, we call this function which is implemented by each kind of devices. For the fsdax case, pmem device driver implements it. Pmem device driver will find out the filesystem in which the corrupted page located in. With dax_holder notify support, we are able to notify the memory failure from pmem driver to upper layers. If there is something not support in the notify routine, memory_failure will fall back to the generic hanlder. Link: https://lkml.kernel.org/r/20220603053738.1218681-4-ruansy.fnst@fujitsu.com Signed-off-by: Shiyang Ruan Reviewed-by: Christoph Hellwig Reviewed-by: Dan Williams Reviewed-by: Darrick J. Wong Reviewed-by: Naoya Horiguchi Cc: Al Viro Cc: Dan Williams Cc: Dave Chinner Cc: Goldwyn Rodrigues Cc: Goldwyn Rodrigues Cc: Jane Chu Cc: Matthew Wilcox Cc: Miaohe Lin Cc: Ritesh Harjani Signed-off-by: Andrew Morton --- drivers/nvdimm/pmem.c | 17 +++++++++++++++++ include/linux/memremap.h | 12 ++++++++++++ mm/memory-failure.c | 14 ++++++++++++++ 3 files changed, 43 insertions(+) (limited to 'include/linux') diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index 629d10fcf53b..107c9cb3d57d 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -453,6 +453,21 @@ static void pmem_release_disk(void *__pmem) blk_cleanup_disk(pmem->disk); } +static int pmem_pagemap_memory_failure(struct dev_pagemap *pgmap, + unsigned long pfn, unsigned long nr_pages, int mf_flags) +{ + struct pmem_device *pmem = + container_of(pgmap, struct pmem_device, pgmap); + u64 offset = PFN_PHYS(pfn) - pmem->phys_addr - pmem->data_offset; + u64 len = nr_pages << PAGE_SHIFT; + + return dax_holder_notify_failure(pmem->dax_dev, offset, len, mf_flags); +} + +static const struct dev_pagemap_ops fsdax_pagemap_ops = { + .memory_failure = pmem_pagemap_memory_failure, +}; + static int pmem_attach_disk(struct device *dev, struct nd_namespace_common *ndns) { @@ -514,6 +529,7 @@ static int pmem_attach_disk(struct device *dev, pmem->pfn_flags = PFN_DEV; if (is_nd_pfn(dev)) { pmem->pgmap.type = MEMORY_DEVICE_FS_DAX; + pmem->pgmap.ops = &fsdax_pagemap_ops; addr = devm_memremap_pages(dev, &pmem->pgmap); pfn_sb = nd_pfn->pfn_sb; pmem->data_offset = le64_to_cpu(pfn_sb->dataoff); @@ -527,6 +543,7 @@ static int pmem_attach_disk(struct device *dev, pmem->pgmap.range.end = res->end; pmem->pgmap.nr_range = 1; pmem->pgmap.type = MEMORY_DEVICE_FS_DAX; + pmem->pgmap.ops = &fsdax_pagemap_ops; addr = devm_memremap_pages(dev, &pmem->pgmap); pmem->pfn_flags |= PFN_MAP; bb_range = pmem->pgmap.range; diff --git a/include/linux/memremap.h b/include/linux/memremap.h index 738b41fa7c8f..2e20b20a0543 100644 --- a/include/linux/memremap.h +++ b/include/linux/memremap.h @@ -87,6 +87,18 @@ struct dev_pagemap_ops { * the page back to a CPU accessible page. */ vm_fault_t (*migrate_to_ram)(struct vm_fault *vmf); + + /* + * Handle the memory failure happens on a range of pfns. Notify the + * processes who are using these pfns, and try to recover the data on + * them if necessary. The mf_flags is finally passed to the recover + * function through the whole notify routine. + * + * When this is not implemented, or it returns -EOPNOTSUPP, the caller + * will fall back to a common handler called mf_generic_kill_procs(). + */ + int (*memory_failure)(struct dev_pagemap *pgmap, unsigned long pfn, + unsigned long nr_pages, int mf_flags); }; #define PGMAP_ALTMAP_VALID (1 << 0) diff --git a/mm/memory-failure.c b/mm/memory-failure.c index 972282168503..f36b5a955900 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -1754,6 +1754,20 @@ static int memory_failure_dev_pagemap(unsigned long pfn, int flags, if (!pgmap_pfn_valid(pgmap, pfn)) goto out; + /* + * Call driver's implementation to handle the memory failure, otherwise + * fall back to generic handler. + */ + if (pgmap->ops->memory_failure) { + rc = pgmap->ops->memory_failure(pgmap, pfn, 1, flags); + /* + * Fall back to generic handler too if operation is not + * supported inside the driver/device/filesystem. + */ + if (rc != -EOPNOTSUPP) + goto out; + } + rc = mf_generic_kill_procs(pfn, flags, pgmap); out: /* drop pgmap ref acquired in caller */ -- cgit v1.2.3 From 191acf0b8926cdd68329adc3861ae06fdea14ccb Mon Sep 17 00:00:00 2001 From: Shiyang Ruan Date: Fri, 3 Jun 2022 13:37:28 +0800 Subject: fsdax: introduce dax_lock_mapping_entry() The current dax_lock_page() locks dax entry by obtaining mapping and index in page. To support 1-to-N RMAP in NVDIMM, we need a new function to lock a specific dax entry corresponding to this file's mapping,index. And output the page corresponding to the specific dax entry for caller use. Link: https://lkml.kernel.org/r/20220603053738.1218681-5-ruansy.fnst@fujitsu.com Signed-off-by: Shiyang Ruan Reviewed-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Cc: Al Viro Cc: Dan Williams Cc: Dan Williams Cc: Dave Chinner Cc: Goldwyn Rodrigues Cc: Goldwyn Rodrigues Cc: Jane Chu Cc: Matthew Wilcox Cc: Miaohe Lin Cc: Naoya Horiguchi Cc: Ritesh Harjani Signed-off-by: Andrew Morton --- fs/dax.c | 63 +++++++++++++++++++++++++++++++++++++++++++++++++++++ include/linux/dax.h | 15 +++++++++++++ 2 files changed, 78 insertions(+) (limited to 'include/linux') diff --git a/fs/dax.c b/fs/dax.c index 4155a6107fa1..65e44d78b3bb 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -455,6 +455,69 @@ void dax_unlock_page(struct page *page, dax_entry_t cookie) dax_unlock_entry(&xas, (void *)cookie); } +/* + * dax_lock_mapping_entry - Lock the DAX entry corresponding to a mapping + * @mapping: the file's mapping whose entry we want to lock + * @index: the offset within this file + * @page: output the dax page corresponding to this dax entry + * + * Return: A cookie to pass to dax_unlock_mapping_entry() or 0 if the entry + * could not be locked. + */ +dax_entry_t dax_lock_mapping_entry(struct address_space *mapping, pgoff_t index, + struct page **page) +{ + XA_STATE(xas, NULL, 0); + void *entry; + + rcu_read_lock(); + for (;;) { + entry = NULL; + if (!dax_mapping(mapping)) + break; + + xas.xa = &mapping->i_pages; + xas_lock_irq(&xas); + xas_set(&xas, index); + entry = xas_load(&xas); + if (dax_is_locked(entry)) { + rcu_read_unlock(); + wait_entry_unlocked(&xas, entry); + rcu_read_lock(); + continue; + } + if (!entry || + dax_is_zero_entry(entry) || dax_is_empty_entry(entry)) { + /* + * Because we are looking for entry from file's mapping + * and index, so the entry may not be inserted for now, + * or even a zero/empty entry. We don't think this is + * an error case. So, return a special value and do + * not output @page. + */ + entry = (void *)~0UL; + } else { + *page = pfn_to_page(dax_to_pfn(entry)); + dax_lock_entry(&xas, entry); + } + xas_unlock_irq(&xas); + break; + } + rcu_read_unlock(); + return (dax_entry_t)entry; +} + +void dax_unlock_mapping_entry(struct address_space *mapping, pgoff_t index, + dax_entry_t cookie) +{ + XA_STATE(xas, &mapping->i_pages, index); + + if (cookie == ~0UL) + return; + + dax_unlock_entry(&xas, (void *)cookie); +} + /* * Find page cache entry at given index. If it is a DAX entry, return it * with the entry locked. If the page cache doesn't contain an entry at diff --git a/include/linux/dax.h b/include/linux/dax.h index cf85fc36da5f..7116681b48c0 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -161,6 +161,10 @@ struct page *dax_layout_busy_page(struct address_space *mapping); struct page *dax_layout_busy_page_range(struct address_space *mapping, loff_t start, loff_t end); dax_entry_t dax_lock_page(struct page *page); void dax_unlock_page(struct page *page, dax_entry_t cookie); +dax_entry_t dax_lock_mapping_entry(struct address_space *mapping, + unsigned long index, struct page **page); +void dax_unlock_mapping_entry(struct address_space *mapping, + unsigned long index, dax_entry_t cookie); #else static inline struct page *dax_layout_busy_page(struct address_space *mapping) { @@ -188,6 +192,17 @@ static inline dax_entry_t dax_lock_page(struct page *page) static inline void dax_unlock_page(struct page *page, dax_entry_t cookie) { } + +static inline dax_entry_t dax_lock_mapping_entry(struct address_space *mapping, + unsigned long index, struct page **page) +{ + return 0; +} + +static inline void dax_unlock_mapping_entry(struct address_space *mapping, + unsigned long index, dax_entry_t cookie) +{ +} #endif int dax_zero_range(struct inode *inode, loff_t pos, loff_t len, bool *did_zero, -- cgit v1.2.3 From c2964da20594d4ba5279b083ad890a8ada1ca638 Mon Sep 17 00:00:00 2001 From: Shiyang Ruan Date: Fri, 3 Jun 2022 13:37:29 +0800 Subject: mm: introduce mf_dax_kill_procs() for fsdax case This new function is a variant of mf_generic_kill_procs that accepts a file, offset pair instead of a struct to support multiple files sharing a DAX mapping. It is intended to be called by the file systems as part of the memory_failure handler after the file system performed a reverse mapping from the storage address to the file and file offset. Link: https://lkml.kernel.org/r/20220603053738.1218681-6-ruansy.fnst@fujitsu.com Signed-off-by: Shiyang Ruan Reviewed-by: Dan Williams Reviewed-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Reviewed-by: Miaohe Lin Cc: Al Viro Cc: Dan Williams Cc: Dave Chinner Cc: Goldwyn Rodrigues Cc: Goldwyn Rodrigues Cc: Jane Chu Cc: Matthew Wilcox Cc: Naoya Horiguchi Cc: Ritesh Harjani Signed-off-by: Andrew Morton --- include/linux/mm.h | 2 ++ mm/memory-failure.c | 96 +++++++++++++++++++++++++++++++++++++++++++++++------ 2 files changed, 88 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 445488e72bf0..4444ae12d758 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -3250,6 +3250,8 @@ enum mf_flags { MF_UNPOISON = 1 << 4, MF_SW_SIMULATED = 1 << 5, }; +int mf_dax_kill_procs(struct address_space *mapping, pgoff_t index, + unsigned long count, int mf_flags); extern int memory_failure(unsigned long pfn, int flags); extern void memory_failure_queue(unsigned long pfn, int flags); extern void memory_failure_queue_kick(int cpu); diff --git a/mm/memory-failure.c b/mm/memory-failure.c index f36b5a955900..70317d88eae3 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -297,10 +297,9 @@ void shake_page(struct page *p) } EXPORT_SYMBOL_GPL(shake_page); -static unsigned long dev_pagemap_mapping_shift(struct page *page, - struct vm_area_struct *vma) +static unsigned long dev_pagemap_mapping_shift(struct vm_area_struct *vma, + unsigned long address) { - unsigned long address = vma_address(page, vma); unsigned long ret = 0; pgd_t *pgd; p4d_t *p4d; @@ -340,10 +339,14 @@ static unsigned long dev_pagemap_mapping_shift(struct page *page, /* * Schedule a process for later kill. * Uses GFP_ATOMIC allocations to avoid potential recursions in the VM. + * + * Notice: @fsdax_pgoff is used only when @p is a fsdax page. + * In other cases, such as anonymous and file-backend page, the address to be + * killed can be caculated by @p itself. */ static void add_to_kill(struct task_struct *tsk, struct page *p, - struct vm_area_struct *vma, - struct list_head *to_kill) + pgoff_t fsdax_pgoff, struct vm_area_struct *vma, + struct list_head *to_kill) { struct to_kill *tk; @@ -354,9 +357,15 @@ static void add_to_kill(struct task_struct *tsk, struct page *p, } tk->addr = page_address_in_vma(p, vma); - if (is_zone_device_page(p)) - tk->size_shift = dev_pagemap_mapping_shift(p, vma); - else + if (is_zone_device_page(p)) { + /* + * Since page->mapping is not used for fsdax, we need + * calculate the address based on the vma. + */ + if (p->pgmap->type == MEMORY_DEVICE_FS_DAX) + tk->addr = vma_pgoff_address(fsdax_pgoff, 1, vma); + tk->size_shift = dev_pagemap_mapping_shift(vma, tk->addr); + } else tk->size_shift = page_shift(compound_head(p)); /* @@ -505,7 +514,7 @@ static void collect_procs_anon(struct page *page, struct list_head *to_kill, if (!page_mapped_in_vma(page, vma)) continue; if (vma->vm_mm == t->mm) - add_to_kill(t, page, vma, to_kill); + add_to_kill(t, page, 0, vma, to_kill); } } read_unlock(&tasklist_lock); @@ -541,13 +550,41 @@ static void collect_procs_file(struct page *page, struct list_head *to_kill, * to be informed of all such data corruptions. */ if (vma->vm_mm == t->mm) - add_to_kill(t, page, vma, to_kill); + add_to_kill(t, page, 0, vma, to_kill); } } read_unlock(&tasklist_lock); i_mmap_unlock_read(mapping); } +#ifdef CONFIG_FS_DAX +/* + * Collect processes when the error hit a fsdax page. + */ +static void collect_procs_fsdax(struct page *page, + struct address_space *mapping, pgoff_t pgoff, + struct list_head *to_kill) +{ + struct vm_area_struct *vma; + struct task_struct *tsk; + + i_mmap_lock_read(mapping); + read_lock(&tasklist_lock); + for_each_process(tsk) { + struct task_struct *t = task_early_kill(tsk, true); + + if (!t) + continue; + vma_interval_tree_foreach(vma, &mapping->i_mmap, pgoff, pgoff) { + if (vma->vm_mm == t->mm) + add_to_kill(t, page, pgoff, vma, to_kill); + } + } + read_unlock(&tasklist_lock); + i_mmap_unlock_read(mapping); +} +#endif /* CONFIG_FS_DAX */ + /* * Collect the processes who have the corrupted page mapped to kill. */ @@ -1587,6 +1624,45 @@ unlock: return rc; } +#ifdef CONFIG_FS_DAX +/** + * mf_dax_kill_procs - Collect and kill processes who are using this file range + * @mapping: address_space of the file in use + * @index: start pgoff of the range within the file + * @count: length of the range, in unit of PAGE_SIZE + * @mf_flags: memory failure flags + */ +int mf_dax_kill_procs(struct address_space *mapping, pgoff_t index, + unsigned long count, int mf_flags) +{ + LIST_HEAD(to_kill); + dax_entry_t cookie; + struct page *page; + size_t end = index + count; + + mf_flags |= MF_ACTION_REQUIRED | MF_MUST_KILL; + + for (; index < end; index++) { + page = NULL; + cookie = dax_lock_mapping_entry(mapping, index, &page); + if (!cookie) + return -EBUSY; + if (!page) + goto unlock; + + SetPageHWPoison(page); + + collect_procs_fsdax(page, mapping, index, &to_kill); + unmap_and_kill(&to_kill, page_to_pfn(page), mapping, + index, mf_flags); +unlock: + dax_unlock_mapping_entry(mapping, index, cookie); + } + return 0; +} +EXPORT_SYMBOL_GPL(mf_dax_kill_procs); +#endif /* CONFIG_FS_DAX */ + /* * Called from hugetlb code with hugetlb_lock held. * -- cgit v1.2.3 From 7a68003869d74fd754b1a64f85bb559fdad422db Mon Sep 17 00:00:00 2001 From: Shiyang Ruan Date: Fri, 3 Jun 2022 13:37:31 +0800 Subject: fsdax: set a CoW flag when associate reflink mappings Introduce a PAGE_MAPPING_DAX_COW flag to support association with CoW file mappings. In this case, since the dax-rmap has already took the responsibility to look up for shared files by given dax page, the page->mapping is no longer to used for rmap but for marking that this dax page is shared. And to make sure disassociation works fine, we use page->index as refcount, and clear page->mapping to the initial state when page->index is decreased to 0. With the help of this new flag, it is able to distinguish normal case and CoW case, and keep the warning in normal case. Link: https://lkml.kernel.org/r/20220603053738.1218681-8-ruansy.fnst@fujitsu.com Signed-off-by: Shiyang Ruan Reviewed-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Cc: Al Viro Cc: Dan Williams Cc: Dan Williams Cc: Dave Chinner Cc: Goldwyn Rodrigues Cc: Goldwyn Rodrigues Cc: Jane Chu Cc: Matthew Wilcox Cc: Miaohe Lin Cc: Naoya Horiguchi Cc: Ritesh Harjani Signed-off-by: Andrew Morton --- fs/dax.c | 50 +++++++++++++++++++++++++++++++++++++--------- include/linux/page-flags.h | 6 ++++++ 2 files changed, 47 insertions(+), 9 deletions(-) (limited to 'include/linux') diff --git a/fs/dax.c b/fs/dax.c index 65e44d78b3bb..b59b864017ad 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -334,13 +334,35 @@ static unsigned long dax_end_pfn(void *entry) for (pfn = dax_to_pfn(entry); \ pfn < dax_end_pfn(entry); pfn++) +static inline bool dax_mapping_is_cow(struct address_space *mapping) +{ + return (unsigned long)mapping == PAGE_MAPPING_DAX_COW; +} + /* - * TODO: for reflink+dax we need a way to associate a single page with - * multiple address_space instances at different linear_page_index() - * offsets. + * Set the page->mapping with FS_DAX_MAPPING_COW flag, increase the refcount. + */ +static inline void dax_mapping_set_cow(struct page *page) +{ + if ((uintptr_t)page->mapping != PAGE_MAPPING_DAX_COW) { + /* + * Reset the index if the page was already mapped + * regularly before. + */ + if (page->mapping) + page->index = 1; + page->mapping = (void *)PAGE_MAPPING_DAX_COW; + } + page->index++; +} + +/* + * When it is called in dax_insert_entry(), the cow flag will indicate that + * whether this entry is shared by multiple files. If so, set the page->mapping + * FS_DAX_MAPPING_COW, and use page->index as refcount. */ static void dax_associate_entry(void *entry, struct address_space *mapping, - struct vm_area_struct *vma, unsigned long address) + struct vm_area_struct *vma, unsigned long address, bool cow) { unsigned long size = dax_entry_size(entry), pfn, index; int i = 0; @@ -352,9 +374,13 @@ static void dax_associate_entry(void *entry, struct address_space *mapping, for_each_mapped_pfn(entry, pfn) { struct page *page = pfn_to_page(pfn); - WARN_ON_ONCE(page->mapping); - page->mapping = mapping; - page->index = index + i++; + if (cow) { + dax_mapping_set_cow(page); + } else { + WARN_ON_ONCE(page->mapping); + page->mapping = mapping; + page->index = index + i++; + } } } @@ -370,7 +396,12 @@ static void dax_disassociate_entry(void *entry, struct address_space *mapping, struct page *page = pfn_to_page(pfn); WARN_ON_ONCE(trunc && page_ref_count(page) > 1); - WARN_ON_ONCE(page->mapping && page->mapping != mapping); + if (dax_mapping_is_cow(page->mapping)) { + /* keep the CoW flag if this page is still shared */ + if (page->index-- > 0) + continue; + } else + WARN_ON_ONCE(page->mapping && page->mapping != mapping); page->mapping = NULL; page->index = 0; } @@ -830,7 +861,8 @@ static void *dax_insert_entry(struct xa_state *xas, void *old; dax_disassociate_entry(entry, mapping, false); - dax_associate_entry(new_entry, mapping, vmf->vma, vmf->address); + dax_associate_entry(new_entry, mapping, vmf->vma, vmf->address, + false); /* * Only swap our new entry into the page cache if the current * entry is a zero page or an empty entry. If a normal PTE or diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h index e66f7aa3191d..a5263a21b72f 100644 --- a/include/linux/page-flags.h +++ b/include/linux/page-flags.h @@ -650,6 +650,12 @@ __PAGEFLAG(Reported, reported, PF_NO_COMPOUND) #define PAGE_MAPPING_KSM (PAGE_MAPPING_ANON | PAGE_MAPPING_MOVABLE) #define PAGE_MAPPING_FLAGS (PAGE_MAPPING_ANON | PAGE_MAPPING_MOVABLE) +/* + * Different with flags above, this flag is used only for fsdax mode. It + * indicates that this page->mapping is now under reflink case. + */ +#define PAGE_MAPPING_DAX_COW 0x1 + static __always_inline bool folio_mapping_flags(struct folio *folio) { return ((unsigned long)folio->mapping & PAGE_MAPPING_FLAGS) != 0; -- cgit v1.2.3 From 919b9f32db9316649e08d9074d1f59ad3159597a Mon Sep 17 00:00:00 2001 From: Shiyang Ruan Date: Fri, 3 Jun 2022 13:37:36 +0800 Subject: fsdax: dedup file range to use a compare function With dax we cannot deal with readpage() etc. So, we create a dax comparison function which is similar with vfs_dedupe_file_range_compare(). And introduce dax_remap_file_range_prep() for filesystem use. Link: https://lkml.kernel.org/r/20220603053738.1218681-13-ruansy.fnst@fujitsu.com Signed-off-by: Goldwyn Rodrigues Signed-off-by: Shiyang Ruan Reviewed-by: Darrick J. Wong Reviewed-by: Christoph Hellwig Cc: Al Viro Cc: Dan Williams Cc: Dan Williams Cc: Dave Chinner Cc: Goldwyn Rodrigues Cc: Jane Chu Cc: Matthew Wilcox Cc: Miaohe Lin Cc: Naoya Horiguchi Cc: Ritesh Harjani Signed-off-by: Andrew Morton --- fs/dax.c | 82 ++++++++++++++++++++++++++++++++++++++++++++++++++++ fs/remap_range.c | 31 ++++++++++++++++---- fs/xfs/xfs_reflink.c | 8 +++-- include/linux/dax.h | 8 +++++ include/linux/fs.h | 12 +++++--- 5 files changed, 130 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/fs/dax.c b/fs/dax.c index 0aab32300531..e0f9c4a0a0c1 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -1873,3 +1873,85 @@ vm_fault_t dax_finish_sync_fault(struct vm_fault *vmf, return dax_insert_pfn_mkwrite(vmf, pfn, order); } EXPORT_SYMBOL_GPL(dax_finish_sync_fault); + +static loff_t dax_range_compare_iter(struct iomap_iter *it_src, + struct iomap_iter *it_dest, u64 len, bool *same) +{ + const struct iomap *smap = &it_src->iomap; + const struct iomap *dmap = &it_dest->iomap; + loff_t pos1 = it_src->pos, pos2 = it_dest->pos; + void *saddr, *daddr; + int id, ret; + + len = min(len, min(smap->length, dmap->length)); + + if (smap->type == IOMAP_HOLE && dmap->type == IOMAP_HOLE) { + *same = true; + return len; + } + + if (smap->type == IOMAP_HOLE || dmap->type == IOMAP_HOLE) { + *same = false; + return 0; + } + + id = dax_read_lock(); + ret = dax_iomap_direct_access(smap, pos1, ALIGN(pos1 + len, PAGE_SIZE), + &saddr, NULL); + if (ret < 0) + goto out_unlock; + + ret = dax_iomap_direct_access(dmap, pos2, ALIGN(pos2 + len, PAGE_SIZE), + &daddr, NULL); + if (ret < 0) + goto out_unlock; + + *same = !memcmp(saddr, daddr, len); + if (!*same) + len = 0; + dax_read_unlock(id); + return len; + +out_unlock: + dax_read_unlock(id); + return -EIO; +} + +int dax_dedupe_file_range_compare(struct inode *src, loff_t srcoff, + struct inode *dst, loff_t dstoff, loff_t len, bool *same, + const struct iomap_ops *ops) +{ + struct iomap_iter src_iter = { + .inode = src, + .pos = srcoff, + .len = len, + .flags = IOMAP_DAX, + }; + struct iomap_iter dst_iter = { + .inode = dst, + .pos = dstoff, + .len = len, + .flags = IOMAP_DAX, + }; + int ret; + + while ((ret = iomap_iter(&src_iter, ops)) > 0) { + while ((ret = iomap_iter(&dst_iter, ops)) > 0) { + dst_iter.processed = dax_range_compare_iter(&src_iter, + &dst_iter, len, same); + } + if (ret <= 0) + src_iter.processed = ret; + } + return ret; +} + +int dax_remap_file_range_prep(struct file *file_in, loff_t pos_in, + struct file *file_out, loff_t pos_out, + loff_t *len, unsigned int remap_flags, + const struct iomap_ops *ops) +{ + return __generic_remap_file_range_prep(file_in, pos_in, file_out, + pos_out, len, remap_flags, ops); +} +EXPORT_SYMBOL_GPL(dax_remap_file_range_prep); diff --git a/fs/remap_range.c b/fs/remap_range.c index e112b5424cdb..231de627c1b9 100644 --- a/fs/remap_range.c +++ b/fs/remap_range.c @@ -14,6 +14,7 @@ #include #include #include +#include #include "internal.h" #include @@ -271,9 +272,11 @@ out_error: * If there's an error, then the usual negative error code is returned. * Otherwise returns 0 with *len set to the request length. */ -int generic_remap_file_range_prep(struct file *file_in, loff_t pos_in, - struct file *file_out, loff_t pos_out, - loff_t *len, unsigned int remap_flags) +int +__generic_remap_file_range_prep(struct file *file_in, loff_t pos_in, + struct file *file_out, loff_t pos_out, + loff_t *len, unsigned int remap_flags, + const struct iomap_ops *dax_read_ops) { struct inode *inode_in = file_inode(file_in); struct inode *inode_out = file_inode(file_out); @@ -333,8 +336,18 @@ int generic_remap_file_range_prep(struct file *file_in, loff_t pos_in, if (remap_flags & REMAP_FILE_DEDUP) { bool is_same = false; - ret = vfs_dedupe_file_range_compare(file_in, pos_in, - file_out, pos_out, *len, &is_same); + if (*len == 0) + return 0; + + if (!IS_DAX(inode_in)) + ret = vfs_dedupe_file_range_compare(file_in, pos_in, + file_out, pos_out, *len, &is_same); + else if (dax_read_ops) + ret = dax_dedupe_file_range_compare(inode_in, pos_in, + inode_out, pos_out, *len, &is_same, + dax_read_ops); + else + return -EINVAL; if (ret) return ret; if (!is_same) @@ -352,6 +365,14 @@ int generic_remap_file_range_prep(struct file *file_in, loff_t pos_in, return ret; } + +int generic_remap_file_range_prep(struct file *file_in, loff_t pos_in, + struct file *file_out, loff_t pos_out, + loff_t *len, unsigned int remap_flags) +{ + return __generic_remap_file_range_prep(file_in, pos_in, file_out, + pos_out, len, remap_flags, NULL); +} EXPORT_SYMBOL(generic_remap_file_range_prep); loff_t do_clone_file_range(struct file *file_in, loff_t pos_in, diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c index e7a7c00d93be..cbaf36d21020 100644 --- a/fs/xfs/xfs_reflink.c +++ b/fs/xfs/xfs_reflink.c @@ -1367,8 +1367,12 @@ xfs_reflink_remap_prep( if (IS_DAX(inode_in) || IS_DAX(inode_out)) goto out_unlock; - ret = generic_remap_file_range_prep(file_in, pos_in, file_out, pos_out, - len, remap_flags); + if (!IS_DAX(inode_in)) + ret = generic_remap_file_range_prep(file_in, pos_in, file_out, + pos_out, len, remap_flags); + else + ret = dax_remap_file_range_prep(file_in, pos_in, file_out, + pos_out, len, remap_flags, &xfs_read_iomap_ops); if (ret || *len == 0) goto out_unlock; diff --git a/include/linux/dax.h b/include/linux/dax.h index 7116681b48c0..ba985333e26b 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -246,6 +246,14 @@ vm_fault_t dax_finish_sync_fault(struct vm_fault *vmf, int dax_delete_mapping_entry(struct address_space *mapping, pgoff_t index); int dax_invalidate_mapping_entry_sync(struct address_space *mapping, pgoff_t index); +int dax_dedupe_file_range_compare(struct inode *src, loff_t srcoff, + struct inode *dest, loff_t destoff, + loff_t len, bool *is_same, + const struct iomap_ops *ops); +int dax_remap_file_range_prep(struct file *file_in, loff_t pos_in, + struct file *file_out, loff_t pos_out, + loff_t *len, unsigned int remap_flags, + const struct iomap_ops *ops); static inline bool dax_mapping(struct address_space *mapping) { return mapping->host && IS_DAX(mapping->host); diff --git a/include/linux/fs.h b/include/linux/fs.h index 9ad5e3520fae..134e9d7ad5d6 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -74,6 +74,7 @@ struct fsverity_operations; struct fs_context; struct fs_parameter_spec; struct fileattr; +struct iomap_ops; extern void __init inode_init(void); extern void __init inode_init_early(void); @@ -2070,10 +2071,13 @@ extern ssize_t vfs_copy_file_range(struct file *, loff_t , struct file *, extern ssize_t generic_copy_file_range(struct file *file_in, loff_t pos_in, struct file *file_out, loff_t pos_out, size_t len, unsigned int flags); -extern int generic_remap_file_range_prep(struct file *file_in, loff_t pos_in, - struct file *file_out, loff_t pos_out, - loff_t *count, - unsigned int remap_flags); +int __generic_remap_file_range_prep(struct file *file_in, loff_t pos_in, + struct file *file_out, loff_t pos_out, + loff_t *len, unsigned int remap_flags, + const struct iomap_ops *dax_read_ops); +int generic_remap_file_range_prep(struct file *file_in, loff_t pos_in, + struct file *file_out, loff_t pos_out, + loff_t *count, unsigned int remap_flags); extern loff_t do_clone_file_range(struct file *file_in, loff_t pos_in, struct file *file_out, loff_t pos_out, loff_t len, unsigned int remap_flags); -- cgit v1.2.3 From e6369794355535a6ede873c3633dfc91bf1932b9 Mon Sep 17 00:00:00 2001 From: SeongJae Park Date: Mon, 13 Jun 2022 19:22:56 +0000 Subject: mm/damon/schemes: add 'LRU_PRIO' DAMOS action This commit adds a new DAMOS action called 'LRU_PRIO' for the physical address space. The action prioritizes pages in the memory regions of the user-specified target access pattern on their LRU lists. This is hence supposed to be used for frequently accessed (hot) memory regions so that hot pages could be more likely protected under memory pressure. Internally, it simply calls 'mark_page_accessed()'. Link: https://lkml.kernel.org/r/20220613192301.8817-5-sj@kernel.org Signed-off-by: SeongJae Park Cc: Jonathan Corbet Signed-off-by: Andrew Morton --- include/linux/damon.h | 2 ++ mm/damon/ops-common.c | 42 ++++++++++++++++++++++++++++++++++++++++++ mm/damon/ops-common.h | 2 ++ mm/damon/paddr.c | 20 ++++++++++++++++++++ mm/damon/sysfs.c | 1 + 5 files changed, 67 insertions(+) (limited to 'include/linux') diff --git a/include/linux/damon.h b/include/linux/damon.h index b9aae19fab3e..4c64e03e94d8 100644 --- a/include/linux/damon.h +++ b/include/linux/damon.h @@ -86,6 +86,7 @@ struct damon_target { * @DAMOS_PAGEOUT: Call ``madvise()`` for the region with MADV_PAGEOUT. * @DAMOS_HUGEPAGE: Call ``madvise()`` for the region with MADV_HUGEPAGE. * @DAMOS_NOHUGEPAGE: Call ``madvise()`` for the region with MADV_NOHUGEPAGE. + * @DAMOS_LRU_PRIO: Prioritize the region on its LRU lists. * @DAMOS_STAT: Do nothing but count the stat. * @NR_DAMOS_ACTIONS: Total number of DAMOS actions */ @@ -95,6 +96,7 @@ enum damos_action { DAMOS_PAGEOUT, DAMOS_HUGEPAGE, DAMOS_NOHUGEPAGE, + DAMOS_LRU_PRIO, DAMOS_STAT, /* Do nothing but only record the stat */ NR_DAMOS_ACTIONS, }; diff --git a/mm/damon/ops-common.c b/mm/damon/ops-common.c index 10ef20b2003f..b1335de200e7 100644 --- a/mm/damon/ops-common.c +++ b/mm/damon/ops-common.c @@ -130,3 +130,45 @@ int damon_pageout_score(struct damon_ctx *c, struct damon_region *r, /* Return coldness of the region */ return DAMOS_MAX_SCORE - hotness; } + +int damon_hot_score(struct damon_ctx *c, struct damon_region *r, + struct damos *s) +{ + unsigned int max_nr_accesses; + int freq_subscore; + unsigned int age_in_sec; + int age_in_log, age_subscore; + unsigned int freq_weight = s->quota.weight_nr_accesses; + unsigned int age_weight = s->quota.weight_age; + int hotness; + + max_nr_accesses = c->aggr_interval / c->sample_interval; + freq_subscore = r->nr_accesses * DAMON_MAX_SUBSCORE / max_nr_accesses; + + age_in_sec = (unsigned long)r->age * c->aggr_interval / 1000000; + for (age_in_log = 0; age_in_log < DAMON_MAX_AGE_IN_LOG && age_in_sec; + age_in_log++, age_in_sec >>= 1) + ; + + /* If frequency is 0, higher age means it's colder */ + if (freq_subscore == 0) + age_in_log *= -1; + + /* + * Now age_in_log is in [-DAMON_MAX_AGE_IN_LOG, DAMON_MAX_AGE_IN_LOG]. + * Scale it to be in [0, 100] and set it as age subscore. + */ + age_in_log += DAMON_MAX_AGE_IN_LOG; + age_subscore = age_in_log * DAMON_MAX_SUBSCORE / + DAMON_MAX_AGE_IN_LOG / 2; + + hotness = (freq_weight * freq_subscore + age_weight * age_subscore); + if (freq_weight + age_weight) + hotness /= freq_weight + age_weight; + /* + * Transform it to fit in [0, DAMOS_MAX_SCORE] + */ + hotness = hotness * DAMOS_MAX_SCORE / DAMON_MAX_SUBSCORE; + + return hotness; +} diff --git a/mm/damon/ops-common.h b/mm/damon/ops-common.h index e790cb5f8fe0..52329ff361cd 100644 --- a/mm/damon/ops-common.h +++ b/mm/damon/ops-common.h @@ -14,3 +14,5 @@ void damon_pmdp_mkold(pmd_t *pmd, struct mm_struct *mm, unsigned long addr); int damon_pageout_score(struct damon_ctx *c, struct damon_region *r, struct damos *s); +int damon_hot_score(struct damon_ctx *c, struct damon_region *r, + struct damos *s); diff --git a/mm/damon/paddr.c b/mm/damon/paddr.c index 7bcd48066b43..f145b1d51e13 100644 --- a/mm/damon/paddr.c +++ b/mm/damon/paddr.c @@ -233,6 +233,22 @@ static unsigned long damon_pa_pageout(struct damon_region *r) return applied * PAGE_SIZE; } +static unsigned long damon_pa_mark_accessed(struct damon_region *r) +{ + unsigned long addr, applied = 0; + + for (addr = r->ar.start; addr < r->ar.end; addr += PAGE_SIZE) { + struct page *page = damon_get_page(PHYS_PFN(addr)); + + if (!page) + continue; + mark_page_accessed(page); + put_page(page); + applied++; + } + return applied * PAGE_SIZE; +} + static unsigned long damon_pa_apply_scheme(struct damon_ctx *ctx, struct damon_target *t, struct damon_region *r, struct damos *scheme) @@ -240,6 +256,8 @@ static unsigned long damon_pa_apply_scheme(struct damon_ctx *ctx, switch (scheme->action) { case DAMOS_PAGEOUT: return damon_pa_pageout(r); + case DAMOS_LRU_PRIO: + return damon_pa_mark_accessed(r); default: break; } @@ -253,6 +271,8 @@ static int damon_pa_scheme_score(struct damon_ctx *context, switch (scheme->action) { case DAMOS_PAGEOUT: return damon_pageout_score(context, r, scheme); + case DAMOS_LRU_PRIO: + return damon_hot_score(context, r, scheme); default: break; } diff --git a/mm/damon/sysfs.c b/mm/damon/sysfs.c index c35809c6087c..86c69f980927 100644 --- a/mm/damon/sysfs.c +++ b/mm/damon/sysfs.c @@ -762,6 +762,7 @@ static const char * const damon_sysfs_damos_action_strs[] = { "pageout", "hugepage", "nohugepage", + "lru_prio", "stat", }; -- cgit v1.2.3 From 50ff3001ae3e229c4540c8396cdcd9b4b21cb362 Mon Sep 17 00:00:00 2001 From: SeongJae Park Date: Mon, 13 Jun 2022 19:22:58 +0000 Subject: mm/damon/schemes: add 'LRU_DEPRIO' action This commit adds a new DAMON-based operation scheme action called 'LRU_DEPRIO' for physical address space. The action deprioritizes pages in the memory area of the target access pattern on their LRU lists. This is hence supposed to be used for rarely accessed (cold) memory regions so that cold pages could be more likely reclaimed first under memory pressure. Internally, it simply calls 'lru_deactivate()'. Using this with 'LRU_PRIO' action for hot pages, users can proactively sort LRU lists based on the access pattern. That is, it can make the LRU lists somewhat more trustworthy source of access temperature. As a result, efficiency of LRU-lists based mechanisms including the reclamation target selection could be improved. Link: https://lkml.kernel.org/r/20220613192301.8817-7-sj@kernel.org Signed-off-by: SeongJae Park Cc: Jonathan Corbet Signed-off-by: Andrew Morton --- include/linux/damon.h | 2 ++ mm/damon/paddr.c | 20 ++++++++++++++++++++ mm/damon/sysfs.c | 1 + 3 files changed, 23 insertions(+) (limited to 'include/linux') diff --git a/include/linux/damon.h b/include/linux/damon.h index 4c64e03e94d8..7b1f4a488230 100644 --- a/include/linux/damon.h +++ b/include/linux/damon.h @@ -87,6 +87,7 @@ struct damon_target { * @DAMOS_HUGEPAGE: Call ``madvise()`` for the region with MADV_HUGEPAGE. * @DAMOS_NOHUGEPAGE: Call ``madvise()`` for the region with MADV_NOHUGEPAGE. * @DAMOS_LRU_PRIO: Prioritize the region on its LRU lists. + * @DAMOS_LRU_DEPRIO: Deprioritize the region on its LRU lists. * @DAMOS_STAT: Do nothing but count the stat. * @NR_DAMOS_ACTIONS: Total number of DAMOS actions */ @@ -97,6 +98,7 @@ enum damos_action { DAMOS_HUGEPAGE, DAMOS_NOHUGEPAGE, DAMOS_LRU_PRIO, + DAMOS_LRU_DEPRIO, DAMOS_STAT, /* Do nothing but only record the stat */ NR_DAMOS_ACTIONS, }; diff --git a/mm/damon/paddr.c b/mm/damon/paddr.c index f145b1d51e13..dc131c6a5403 100644 --- a/mm/damon/paddr.c +++ b/mm/damon/paddr.c @@ -249,6 +249,22 @@ static unsigned long damon_pa_mark_accessed(struct damon_region *r) return applied * PAGE_SIZE; } +static unsigned long damon_pa_deactivate_pages(struct damon_region *r) +{ + unsigned long addr, applied = 0; + + for (addr = r->ar.start; addr < r->ar.end; addr += PAGE_SIZE) { + struct page *page = damon_get_page(PHYS_PFN(addr)); + + if (!page) + continue; + deactivate_page(page); + put_page(page); + applied++; + } + return applied * PAGE_SIZE; +} + static unsigned long damon_pa_apply_scheme(struct damon_ctx *ctx, struct damon_target *t, struct damon_region *r, struct damos *scheme) @@ -258,6 +274,8 @@ static unsigned long damon_pa_apply_scheme(struct damon_ctx *ctx, return damon_pa_pageout(r); case DAMOS_LRU_PRIO: return damon_pa_mark_accessed(r); + case DAMOS_LRU_DEPRIO: + return damon_pa_deactivate_pages(r); default: break; } @@ -273,6 +291,8 @@ static int damon_pa_scheme_score(struct damon_ctx *context, return damon_pageout_score(context, r, scheme); case DAMOS_LRU_PRIO: return damon_hot_score(context, r, scheme); + case DAMOS_LRU_DEPRIO: + return damon_pageout_score(context, r, scheme); default: break; } diff --git a/mm/damon/sysfs.c b/mm/damon/sysfs.c index 86c69f980927..7488e27c87c3 100644 --- a/mm/damon/sysfs.c +++ b/mm/damon/sysfs.c @@ -763,6 +763,7 @@ static const char * const damon_sysfs_damos_action_strs[] = { "hugepage", "nohugepage", "lru_prio", + "lru_deprio", "stat", }; -- cgit v1.2.3 From fb7db1cd137a12dbf84625e1efe737245d1c5e35 Mon Sep 17 00:00:00 2001 From: David Hildenbrand Date: Tue, 14 Jun 2022 11:36:29 +0200 Subject: mm/mprotect: try avoiding write faults for exclusive anonymous pages when changing protection Similar to our MM_CP_DIRTY_ACCT handling for shared, writable mappings, we can try mapping anonymous pages in a private writable mapping writable if they are exclusive, the PTE is already dirty, and no special handling applies. Mapping the anonymous page writable is essentially the same thing the write fault handler would do in this case. Special handling is required for uffd-wp and softdirty tracking, so take care of that properly. Also, leave PROT_NONE handling alone for now; in the future, we could similarly extend the logic in do_numa_page() or use pte_mk_savedwrite() here. While this improves mprotect(PROT_READ)+mprotect(PROT_READ|PROT_WRITE) performance, it should also be a valuable optimization for uffd-wp, when un-protecting. This has been previously suggested by Peter Collingbourne in [1], relevant in the context of the Scudo memory allocator, before we had PageAnonExclusive. This commit doesn't add the same handling for PMDs (i.e., anonymous THP, anonymous hugetlb); benchmark results from Andrea indicate that there are minor performance gains, so it's might still be valuable to streamline that logic for all anonymous pages in the future. As we now also set MM_CP_DIRTY_ACCT for private mappings, let's rename it to MM_CP_TRY_CHANGE_WRITABLE, to make it clearer what's actually happening. Micro-benchmark courtesy of Andrea: === #define _GNU_SOURCE #include #include #include #include #include #define SIZE (1024*1024*1024) int main(int argc, char *argv[]) { char *p; if (posix_memalign((void **)&p, sysconf(_SC_PAGESIZE)*512, SIZE)) perror("posix_memalign"), exit(1); if (madvise(p, SIZE, argc > 1 ? MADV_HUGEPAGE : MADV_NOHUGEPAGE)) perror("madvise"); explicit_bzero(p, SIZE); for (int loops = 0; loops < 40; loops++) { if (mprotect(p, SIZE, PROT_READ)) perror("mprotect"), exit(1); if (mprotect(p, SIZE, PROT_READ|PROT_WRITE)) perror("mprotect"), exit(1); explicit_bzero(p, SIZE); } } === Results on my Ryzen 9 3900X: Stock 10 runs (lower is better): AVG 6.398s, STDEV 0.043 Patched 10 runs (lower is better): AVG 3.780s, STDEV 0.026 === [1] https://lkml.kernel.org/r/20210429214801.2583336-1-pcc@google.com Link: https://lkml.kernel.org/r/20220614093629.76309-1-david@redhat.com Signed-off-by: David Hildenbrand Suggested-by: Peter Collingbourne Acked-by: Peter Xu Cc: Nadav Amit Cc: Dave Hansen Cc: Andrea Arcangeli Cc: Yang Shi Cc: Hugh Dickins Cc: Mel Gorman Signed-off-by: Andrew Morton --- include/linux/mm.h | 8 ++++-- mm/mprotect.c | 77 +++++++++++++++++++++++++++++++++++++++++++----------- 2 files changed, 68 insertions(+), 17 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 4444ae12d758..fb610184cf4e 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1994,8 +1994,12 @@ extern unsigned long move_page_tables(struct vm_area_struct *vma, * for now all the callers are only use one of the flags at the same * time. */ -/* Whether we should allow dirty bit accounting */ -#define MM_CP_DIRTY_ACCT (1UL << 0) +/* + * Whether we should manually check if we can map individual PTEs writable, + * because something (e.g., COW, uffd-wp) blocks that from happening for all + * PTEs automatically in a writable mapping. + */ +#define MM_CP_TRY_CHANGE_WRITABLE (1UL << 0) /* Whether this protection change is for NUMA hints */ #define MM_CP_PROT_NUMA (1UL << 1) /* Whether this change is for write protecting */ diff --git a/mm/mprotect.c b/mm/mprotect.c index f7b31657d1da..041beeb5fad6 100644 --- a/mm/mprotect.c +++ b/mm/mprotect.c @@ -38,6 +38,39 @@ #include "internal.h" +static inline bool can_change_pte_writable(struct vm_area_struct *vma, + unsigned long addr, pte_t pte) +{ + struct page *page; + + VM_BUG_ON(!(vma->vm_flags & VM_WRITE) || pte_write(pte)); + + if (pte_protnone(pte) || !pte_dirty(pte)) + return false; + + /* Do we need write faults for softdirty tracking? */ + if ((vma->vm_flags & VM_SOFTDIRTY) && !pte_soft_dirty(pte)) + return false; + + /* Do we need write faults for uffd-wp tracking? */ + if (userfaultfd_pte_wp(vma, pte)) + return false; + + if (!(vma->vm_flags & VM_SHARED)) { + /* + * We can only special-case on exclusive anonymous pages, + * because we know that our write-fault handler similarly would + * map them writable without any additional checks while holding + * the PT lock. + */ + page = vm_normal_page(vma, addr, pte); + if (!page || !PageAnon(page) || !PageAnonExclusive(page)) + return false; + } + + return true; +} + static unsigned long change_pte_range(struct mmu_gather *tlb, struct vm_area_struct *vma, pmd_t *pmd, unsigned long addr, unsigned long end, pgprot_t newprot, unsigned long cp_flags) @@ -46,7 +79,6 @@ static unsigned long change_pte_range(struct mmu_gather *tlb, spinlock_t *ptl; unsigned long pages = 0; int target_node = NUMA_NO_NODE; - bool dirty_accountable = cp_flags & MM_CP_DIRTY_ACCT; bool prot_numa = cp_flags & MM_CP_PROT_NUMA; bool uffd_wp = cp_flags & MM_CP_UFFD_WP; bool uffd_wp_resolve = cp_flags & MM_CP_UFFD_WP_RESOLVE; @@ -137,21 +169,27 @@ static unsigned long change_pte_range(struct mmu_gather *tlb, ptent = pte_wrprotect(ptent); ptent = pte_mkuffd_wp(ptent); } else if (uffd_wp_resolve) { - /* - * Leave the write bit to be handled - * by PF interrupt handler, then - * things like COW could be properly - * handled. - */ ptent = pte_clear_uffd_wp(ptent); } - /* Avoid taking write faults for known dirty pages */ - if (dirty_accountable && pte_dirty(ptent) && - (pte_soft_dirty(ptent) || - !(vma->vm_flags & VM_SOFTDIRTY))) { + /* + * In some writable, shared mappings, we might want + * to catch actual write access -- see + * vma_wants_writenotify(). + * + * In all writable, private mappings, we have to + * properly handle COW. + * + * In both cases, we can sometimes still change PTEs + * writable and avoid the write-fault handler, for + * example, if a PTE is already dirty and no other + * COW or special handling is required. + */ + if ((cp_flags & MM_CP_TRY_CHANGE_WRITABLE) && + !pte_write(ptent) && + can_change_pte_writable(vma, addr, ptent)) ptent = pte_mkwrite(ptent); - } + ptep_modify_prot_commit(vma, addr, pte, oldpte, ptent); if (pte_needs_flush(oldpte, ptent)) tlb_flush_pte_range(tlb, addr, PAGE_SIZE); @@ -505,9 +543,9 @@ mprotect_fixup(struct mmu_gather *tlb, struct vm_area_struct *vma, unsigned long oldflags = vma->vm_flags; long nrpages = (end - start) >> PAGE_SHIFT; unsigned long charged = 0; + bool try_change_writable; pgoff_t pgoff; int error; - int dirty_accountable = 0; if (newflags == oldflags) { *pprev = vma; @@ -583,11 +621,20 @@ success: * held in write mode. */ vma->vm_flags = newflags; - dirty_accountable = vma_wants_writenotify(vma, vma->vm_page_prot); + /* + * We want to check manually if we can change individual PTEs writable + * if we can't do that automatically for all PTEs in a mapping. For + * private mappings, that's always the case when we have write + * permissions as we properly have to handle COW. + */ + if (vma->vm_flags & VM_SHARED) + try_change_writable = vma_wants_writenotify(vma, vma->vm_page_prot); + else + try_change_writable = !!(vma->vm_flags & VM_WRITE); vma_set_page_prot(vma); change_protection(tlb, vma, start, end, vma->vm_page_prot, - dirty_accountable ? MM_CP_DIRTY_ACCT : 0); + try_change_writable ? MM_CP_TRY_CHANGE_WRITABLE : 0); /* * Private VM_LOCKED VMA becoming writable: trigger COW to avoid major -- cgit v1.2.3 From f491089d4ae5c44041be80c11dbd1fca1956c208 Mon Sep 17 00:00:00 2001 From: Yang Shi Date: Thu, 16 Jun 2022 10:48:35 -0700 Subject: mm: thp: consolidate vma size check to transhuge_vma_suitable There are couple of places that check whether the vma size is ok for THP or whether address fits, they are open coded and duplicate, use transhuge_vma_suitable() to do the job by passing in (vma->end - HPAGE_PMD_SIZE). Move vma size check into hugepage_vma_check(). This will make khugepaged_enter() is as same as khugepaged_enter_vma(). There is just one caller for khugepaged_enter(), replace it to khugepaged_enter_vma() and remove khugepaged_enter(). Link: https://lkml.kernel.org/r/20220616174840.1202070-3-shy828301@gmail.com Signed-off-by: Yang Shi Reviewed-by: Zach O'Keefe Cc: Kirill A. Shutemov Cc: Matthew Wilcox Cc: Miaohe Lin Cc: Vlastimil Babka Signed-off-by: Andrew Morton --- include/linux/huge_mm.h | 11 +++++++++++ include/linux/khugepaged.h | 14 -------------- mm/huge_memory.c | 2 +- mm/khugepaged.c | 19 ++++++------------- 4 files changed, 18 insertions(+), 28 deletions(-) (limited to 'include/linux') diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index 648cb3ce7099..8a5a8bfce0f5 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -116,6 +116,17 @@ extern struct kobj_attribute shmem_enabled_attr; extern unsigned long transparent_hugepage_flags; +/* + * Do the below checks: + * - For file vma, check if the linear page offset of vma is + * HPAGE_PMD_NR aligned within the file. The hugepage is + * guaranteed to be hugepage-aligned within the file, but we must + * check that the PMD-aligned addresses in the VMA map to + * PMD-aligned offsets within the file, else the hugepage will + * not be PMD-mappable. + * - For all vmas, check if the haddr is in an aligned HPAGE_PMD_SIZE + * area. + */ static inline bool transhuge_vma_suitable(struct vm_area_struct *vma, unsigned long addr) { diff --git a/include/linux/khugepaged.h b/include/linux/khugepaged.h index 392d34c3c59a..31ca8a7f78f4 100644 --- a/include/linux/khugepaged.h +++ b/include/linux/khugepaged.h @@ -51,16 +51,6 @@ static inline void khugepaged_exit(struct mm_struct *mm) if (test_bit(MMF_VM_HUGEPAGE, &mm->flags)) __khugepaged_exit(mm); } - -static inline void khugepaged_enter(struct vm_area_struct *vma, - unsigned long vm_flags) -{ - if (!test_bit(MMF_VM_HUGEPAGE, &vma->vm_mm->flags) && - khugepaged_enabled()) { - if (hugepage_vma_check(vma, vm_flags)) - __khugepaged_enter(vma->vm_mm); - } -} #else /* CONFIG_TRANSPARENT_HUGEPAGE */ static inline void khugepaged_fork(struct mm_struct *mm, struct mm_struct *oldmm) { @@ -68,10 +58,6 @@ static inline void khugepaged_fork(struct mm_struct *mm, struct mm_struct *oldmm static inline void khugepaged_exit(struct mm_struct *mm) { } -static inline void khugepaged_enter(struct vm_area_struct *vma, - unsigned long vm_flags) -{ -} static inline void khugepaged_enter_vma(struct vm_area_struct *vma, unsigned long vm_flags) { diff --git a/mm/huge_memory.c b/mm/huge_memory.c index d624b2544751..54735d4173b3 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -726,7 +726,7 @@ vm_fault_t do_huge_pmd_anonymous_page(struct vm_fault *vmf) return VM_FAULT_FALLBACK; if (unlikely(anon_vma_prepare(vma))) return VM_FAULT_OOM; - khugepaged_enter(vma, vma->vm_flags); + khugepaged_enter_vma(vma, vma->vm_flags); if (!(vmf->flags & FAULT_FLAG_WRITE) && !mm_forbids_zeropage(vma->vm_mm) && diff --git a/mm/khugepaged.c b/mm/khugepaged.c index b1dab94c0f1e..db0b334a7d1f 100644 --- a/mm/khugepaged.c +++ b/mm/khugepaged.c @@ -450,8 +450,8 @@ bool hugepage_vma_check(struct vm_area_struct *vma, if (vma_is_dax(vma)) return false; - if (vma->vm_file && !IS_ALIGNED((vma->vm_start >> PAGE_SHIFT) - - vma->vm_pgoff, HPAGE_PMD_NR)) + /* Check alignment for file vma and size for both file and anon vma */ + if (!transhuge_vma_suitable(vma, (vma->vm_end - HPAGE_PMD_SIZE))) return false; /* Enabled via shmem mount options or sysfs settings. */ @@ -512,9 +512,7 @@ void khugepaged_enter_vma(struct vm_area_struct *vma, unsigned long vm_flags) { if (!test_bit(MMF_VM_HUGEPAGE, &vma->vm_mm->flags) && - khugepaged_enabled() && - (((vma->vm_start + ~HPAGE_PMD_MASK) & HPAGE_PMD_MASK) < - (vma->vm_end & HPAGE_PMD_MASK))) { + khugepaged_enabled()) { if (hugepage_vma_check(vma, vm_flags)) __khugepaged_enter(vma->vm_mm); } @@ -950,7 +948,6 @@ static int hugepage_vma_revalidate(struct mm_struct *mm, unsigned long address, struct vm_area_struct **vmap) { struct vm_area_struct *vma; - unsigned long hstart, hend; if (unlikely(khugepaged_test_exit(mm))) return SCAN_ANY_PROCESS; @@ -959,9 +956,7 @@ static int hugepage_vma_revalidate(struct mm_struct *mm, unsigned long address, if (!vma) return SCAN_VMA_NULL; - hstart = (vma->vm_start + ~HPAGE_PMD_MASK) & HPAGE_PMD_MASK; - hend = vma->vm_end & HPAGE_PMD_MASK; - if (address < hstart || address + HPAGE_PMD_SIZE > hend) + if (!transhuge_vma_suitable(vma, address)) return SCAN_ADDRESS_RANGE; if (!hugepage_vma_check(vma, vma->vm_flags)) return SCAN_VMA_CHECK; @@ -2147,10 +2142,8 @@ skip: progress++; continue; } - hstart = (vma->vm_start + ~HPAGE_PMD_MASK) & HPAGE_PMD_MASK; - hend = vma->vm_end & HPAGE_PMD_MASK; - if (hstart >= hend) - goto skip; + hstart = round_up(vma->vm_start, HPAGE_PMD_SIZE); + hend = round_down(vma->vm_end, HPAGE_PMD_SIZE); if (khugepaged_scan.address > hend) goto skip; if (khugepaged_scan.address < hstart) -- cgit v1.2.3 From 2fbb1a28268ee9825d7edcd6caaf3457c56776e8 Mon Sep 17 00:00:00 2001 From: Yang Shi Date: Thu, 16 Jun 2022 10:48:37 -0700 Subject: mm: thp: kill transparent_hugepage_active() The transparent_hugepage_active() was introduced to show THP eligibility bit in smaps in proc, smaps is the only user. But it actually does the similar check as hugepage_vma_check() which is used by khugepaged. We definitely don't have to maintain two similar checks, so kill transparent_hugepage_active(). This patch also fixed the wrong behavior for VM_NO_KHUGEPAGED vmas. Also move hugepage_vma_check() to huge_memory.c and huge_mm.h since it is not only for khugepaged anymore. Link: https://lkml.kernel.org/r/20220616174840.1202070-5-shy828301@gmail.com Signed-off-by: Yang Shi Reviewed-by: Zach O'Keefe Cc: Kirill A. Shutemov Cc: Matthew Wilcox Cc: Miaohe Lin Cc: Vlastimil Babka Signed-off-by: Andrew Morton --- fs/proc/task_mmu.c | 2 +- include/linux/huge_mm.h | 16 +++++++++------ include/linux/khugepaged.h | 2 -- mm/huge_memory.c | 50 +++++++++++++++++++++++++++++++++++++--------- mm/khugepaged.c | 48 ++++---------------------------------------- 5 files changed, 56 insertions(+), 62 deletions(-) (limited to 'include/linux') diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 37ccb5c9f4f8..39a40ec181e7 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -863,7 +863,7 @@ static int show_smap(struct seq_file *m, void *v) __show_smap(m, &mss, false); seq_printf(m, "THPeligible: %d\n", - transparent_hugepage_active(vma)); + hugepage_vma_check(vma, vma->vm_flags, true)); if (arch_pkeys_enabled()) seq_printf(m, "ProtectionKey: %8u\n", vma_pkey(vma)); diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index 8a5a8bfce0f5..64487bcd0c7b 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -202,7 +202,9 @@ static inline bool file_thp_enabled(struct vm_area_struct *vma) !inode_is_open_for_write(inode) && S_ISREG(inode->i_mode); } -bool transparent_hugepage_active(struct vm_area_struct *vma); +bool hugepage_vma_check(struct vm_area_struct *vma, + unsigned long vm_flags, + bool smaps); #define transparent_hugepage_use_zero_page() \ (transparent_hugepage_flags & \ @@ -351,11 +353,6 @@ static inline bool __transparent_hugepage_enabled(struct vm_area_struct *vma) return false; } -static inline bool transparent_hugepage_active(struct vm_area_struct *vma) -{ - return false; -} - static inline bool transhuge_vma_suitable(struct vm_area_struct *vma, unsigned long addr) { @@ -368,6 +365,13 @@ static inline bool transhuge_vma_enabled(struct vm_area_struct *vma, return false; } +static inline bool hugepage_vma_check(struct vm_area_struct *vma, + unsigned long vm_flags, + bool smaps) +{ + return false; +} + static inline void prep_transhuge_page(struct page *page) {} #define transparent_hugepage_flags 0UL diff --git a/include/linux/khugepaged.h b/include/linux/khugepaged.h index 31ca8a7f78f4..ea5fd4c398f7 100644 --- a/include/linux/khugepaged.h +++ b/include/linux/khugepaged.h @@ -10,8 +10,6 @@ extern struct attribute_group khugepaged_attr_group; extern int khugepaged_init(void); extern void khugepaged_destroy(void); extern int start_stop_khugepaged(void); -extern bool hugepage_vma_check(struct vm_area_struct *vma, - unsigned long vm_flags); extern void __khugepaged_enter(struct mm_struct *mm); extern void __khugepaged_exit(struct mm_struct *mm); extern void khugepaged_enter_vma(struct vm_area_struct *vma, diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 54735d4173b3..6f539dc33091 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -69,21 +69,53 @@ static atomic_t huge_zero_refcount; struct page *huge_zero_page __read_mostly; unsigned long huge_zero_pfn __read_mostly = ~0UL; -bool transparent_hugepage_active(struct vm_area_struct *vma) +bool hugepage_vma_check(struct vm_area_struct *vma, + unsigned long vm_flags, + bool smaps) { - /* The addr is used to check if the vma size fits */ - unsigned long addr = (vma->vm_end & HPAGE_PMD_MASK) - HPAGE_PMD_SIZE; + if (!transhuge_vma_enabled(vma, vm_flags)) + return false; + + if (vm_flags & VM_NO_KHUGEPAGED) + return false; + + /* Don't run khugepaged against DAX vma */ + if (vma_is_dax(vma)) + return false; - if (!transhuge_vma_suitable(vma, addr)) + /* Check alignment for file vma and size for both file and anon vma */ + if (!transhuge_vma_suitable(vma, (vma->vm_end - HPAGE_PMD_SIZE))) return false; - if (vma_is_anonymous(vma)) - return __transparent_hugepage_enabled(vma); - if (vma_is_shmem(vma)) + + /* Enabled via shmem mount options or sysfs settings. */ + if (shmem_file(vma->vm_file)) return shmem_huge_enabled(vma); - if (transhuge_vma_enabled(vma, vma->vm_flags) && file_thp_enabled(vma)) + + if (!khugepaged_enabled()) + return false; + + /* THP settings require madvise. */ + if (!(vm_flags & VM_HUGEPAGE) && !khugepaged_always()) + return false; + + /* Only regular file is valid */ + if (file_thp_enabled(vma)) return true; - return false; + if (!vma_is_anonymous(vma)) + return false; + + if (vma_is_temporary_stack(vma)) + return false; + + /* + * THPeligible bit of smaps should show 1 for proper VMAs even + * though anon_vma is not initialized yet. + */ + if (!vma->anon_vma) + return smaps; + + return true; } static bool get_huge_zero_page(void) diff --git a/mm/khugepaged.c b/mm/khugepaged.c index 5baa394e34c8..3afd87f8c0b1 100644 --- a/mm/khugepaged.c +++ b/mm/khugepaged.c @@ -437,46 +437,6 @@ static inline int khugepaged_test_exit(struct mm_struct *mm) return atomic_read(&mm->mm_users) == 0; } -bool hugepage_vma_check(struct vm_area_struct *vma, - unsigned long vm_flags) -{ - if (!transhuge_vma_enabled(vma, vm_flags)) - return false; - - if (vm_flags & VM_NO_KHUGEPAGED) - return false; - - /* Don't run khugepaged against DAX vma */ - if (vma_is_dax(vma)) - return false; - - /* Check alignment for file vma and size for both file and anon vma */ - if (!transhuge_vma_suitable(vma, (vma->vm_end - HPAGE_PMD_SIZE))) - return false; - - /* Enabled via shmem mount options or sysfs settings. */ - if (shmem_file(vma->vm_file)) - return shmem_huge_enabled(vma); - - if (!khugepaged_enabled()) - return false; - - /* THP settings require madvise. */ - if (!(vm_flags & VM_HUGEPAGE) && !khugepaged_always()) - return false; - - /* Only regular file is valid */ - if (file_thp_enabled(vma)) - return true; - - if (!vma->anon_vma || !vma_is_anonymous(vma)) - return false; - if (vma_is_temporary_stack(vma)) - return false; - - return true; -} - void __khugepaged_enter(struct mm_struct *mm) { struct mm_slot *mm_slot; @@ -513,7 +473,7 @@ void khugepaged_enter_vma(struct vm_area_struct *vma, { if (!test_bit(MMF_VM_HUGEPAGE, &vma->vm_mm->flags) && khugepaged_enabled()) { - if (hugepage_vma_check(vma, vm_flags)) + if (hugepage_vma_check(vma, vm_flags, false)) __khugepaged_enter(vma->vm_mm); } } @@ -958,7 +918,7 @@ static int hugepage_vma_revalidate(struct mm_struct *mm, unsigned long address, if (!transhuge_vma_suitable(vma, address)) return SCAN_ADDRESS_RANGE; - if (!hugepage_vma_check(vma, vma->vm_flags)) + if (!hugepage_vma_check(vma, vma->vm_flags, false)) return SCAN_VMA_CHECK; /* * Anon VMA expected, the address may be unmapped then @@ -1448,7 +1408,7 @@ void collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr) * the valid THP. Add extra VM_HUGEPAGE so hugepage_vma_check() * will not fail the vma for missing VM_HUGEPAGE */ - if (!hugepage_vma_check(vma, vma->vm_flags | VM_HUGEPAGE)) + if (!hugepage_vma_check(vma, vma->vm_flags | VM_HUGEPAGE, false)) return; /* Keep pmd pgtable for uffd-wp; see comment in retract_page_tables() */ @@ -2143,7 +2103,7 @@ static unsigned int khugepaged_scan_mm_slot(unsigned int pages, progress++; break; } - if (!hugepage_vma_check(vma, vma->vm_flags)) { + if (!hugepage_vma_check(vma, vma->vm_flags, false)) { skip: progress++; continue; -- cgit v1.2.3 From 27fe4805820b6b559be445f8e35c8ba95c538269 Mon Sep 17 00:00:00 2001 From: Yang Shi Date: Thu, 16 Jun 2022 10:48:38 -0700 Subject: mm: thp: kill __transhuge_page_enabled() The page fault path checks THP eligibility with __transhuge_page_enabled() which does the similar thing as hugepage_vma_check(), so use hugepage_vma_check() instead. However page fault allows DAX and !anon_vma cases, so added a new flag, in_pf, to hugepage_vma_check() to make page fault work correctly. The in_pf flag is also used to skip shmem and file THP for page fault since shmem handles THP in its own shmem_fault() and file THP allocation on fault is not supported yet. Also remove hugepage_vma_enabled() since hugepage_vma_check() is the only caller now, it is not necessary to have a helper function. Link: https://lkml.kernel.org/r/20220616174840.1202070-6-shy828301@gmail.com Signed-off-by: Yang Shi Reviewed-by: Zach O'Keefe Cc: Kirill A. Shutemov Cc: Matthew Wilcox Cc: Miaohe Lin Cc: Vlastimil Babka Signed-off-by: Andrew Morton --- fs/proc/task_mmu.c | 2 +- include/linux/huge_mm.h | 57 ++----------------------------------------------- mm/huge_memory.c | 51 +++++++++++++++++++++++++++++++++---------- mm/khugepaged.c | 8 +++---- mm/memory.c | 7 ++++-- 5 files changed, 52 insertions(+), 73 deletions(-) (limited to 'include/linux') diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 39a40ec181e7..cef72e49acc5 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -863,7 +863,7 @@ static int show_smap(struct seq_file *m, void *v) __show_smap(m, &mss, false); seq_printf(m, "THPeligible: %d\n", - hugepage_vma_check(vma, vma->vm_flags, true)); + hugepage_vma_check(vma, vma->vm_flags, true, false)); if (arch_pkeys_enabled()) seq_printf(m, "ProtectionKey: %8u\n", vma_pkey(vma)); diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index 64487bcd0c7b..cd8a6c5d9fe5 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -146,48 +146,6 @@ static inline bool transhuge_vma_suitable(struct vm_area_struct *vma, return true; } -static inline bool transhuge_vma_enabled(struct vm_area_struct *vma, - unsigned long vm_flags) -{ - /* Explicitly disabled through madvise. */ - if ((vm_flags & VM_NOHUGEPAGE) || - test_bit(MMF_DISABLE_THP, &vma->vm_mm->flags)) - return false; - return true; -} - -/* - * to be used on vmas which are known to support THP. - * Use transparent_hugepage_active otherwise - */ -static inline bool __transparent_hugepage_enabled(struct vm_area_struct *vma) -{ - - /* - * If the hardware/firmware marked hugepage support disabled. - */ - if (transparent_hugepage_flags & (1 << TRANSPARENT_HUGEPAGE_NEVER_DAX)) - return false; - - if (!transhuge_vma_enabled(vma, vma->vm_flags)) - return false; - - if (vma_is_temporary_stack(vma)) - return false; - - if (transparent_hugepage_flags & (1 << TRANSPARENT_HUGEPAGE_FLAG)) - return true; - - if (vma_is_dax(vma)) - return true; - - if (transparent_hugepage_flags & - (1 << TRANSPARENT_HUGEPAGE_REQ_MADV_FLAG)) - return !!(vma->vm_flags & VM_HUGEPAGE); - - return false; -} - static inline bool file_thp_enabled(struct vm_area_struct *vma) { struct inode *inode; @@ -204,7 +162,7 @@ static inline bool file_thp_enabled(struct vm_area_struct *vma) bool hugepage_vma_check(struct vm_area_struct *vma, unsigned long vm_flags, - bool smaps); + bool smaps, bool in_pf); #define transparent_hugepage_use_zero_page() \ (transparent_hugepage_flags & \ @@ -348,26 +306,15 @@ static inline bool folio_test_pmd_mappable(struct folio *folio) return false; } -static inline bool __transparent_hugepage_enabled(struct vm_area_struct *vma) -{ - return false; -} - static inline bool transhuge_vma_suitable(struct vm_area_struct *vma, unsigned long addr) { return false; } -static inline bool transhuge_vma_enabled(struct vm_area_struct *vma, - unsigned long vm_flags) -{ - return false; -} - static inline bool hugepage_vma_check(struct vm_area_struct *vma, unsigned long vm_flags, - bool smaps) + bool smaps, bool in_pf) { return false; } diff --git a/mm/huge_memory.c b/mm/huge_memory.c index fd82ae260997..549a02137d13 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -71,27 +71,53 @@ unsigned long huge_zero_pfn __read_mostly = ~0UL; bool hugepage_vma_check(struct vm_area_struct *vma, unsigned long vm_flags, - bool smaps) + bool smaps, bool in_pf) { if (!vma->vm_mm) /* vdso */ return false; - if (!transhuge_vma_enabled(vma, vm_flags)) + /* + * Explicitly disabled through madvise or prctl, or some + * architectures may disable THP for some mappings, for + * example, s390 kvm. + * */ + if ((vm_flags & VM_NOHUGEPAGE) || + test_bit(MMF_DISABLE_THP, &vma->vm_mm->flags)) return false; - - if (vm_flags & VM_NO_KHUGEPAGED) + /* + * If the hardware/firmware marked hugepage support disabled. + */ + if (transparent_hugepage_flags & (1 << TRANSPARENT_HUGEPAGE_NEVER_DAX)) return false; - /* Don't run khugepaged against DAX vma */ + /* khugepaged doesn't collapse DAX vma, but page fault is fine. */ if (vma_is_dax(vma)) + return in_pf; + + /* + * Special VMA and hugetlb VMA. + * Must be checked after dax since some dax mappings may have + * VM_MIXEDMAP set. + */ + if (vm_flags & VM_NO_KHUGEPAGED) return false; - /* Check alignment for file vma and size for both file and anon vma */ - if (!transhuge_vma_suitable(vma, (vma->vm_end - HPAGE_PMD_SIZE))) + /* + * Check alignment for file vma and size for both file and anon vma. + * + * Skip the check for page fault. Huge fault does the check in fault + * handlers. And this check is not suitable for huge PUD fault. + */ + if (!in_pf && + !transhuge_vma_suitable(vma, (vma->vm_end - HPAGE_PMD_SIZE))) return false; - /* Enabled via shmem mount options or sysfs settings. */ - if (shmem_file(vma->vm_file)) + /* + * Enabled via shmem mount options or sysfs settings. + * Must be done before hugepage flags check since shmem has its + * own flags. + */ + if (!in_pf && shmem_file(vma->vm_file)) return shmem_huge_enabled(vma); if (!khugepaged_enabled()) @@ -102,7 +128,7 @@ bool hugepage_vma_check(struct vm_area_struct *vma, return false; /* Only regular file is valid */ - if (file_thp_enabled(vma)) + if (!in_pf && file_thp_enabled(vma)) return true; if (!vma_is_anonymous(vma)) @@ -114,9 +140,12 @@ bool hugepage_vma_check(struct vm_area_struct *vma, /* * THPeligible bit of smaps should show 1 for proper VMAs even * though anon_vma is not initialized yet. + * + * Allow page fault since anon_vma may be not initialized until + * the first page fault. */ if (!vma->anon_vma) - return smaps; + return (smaps || in_pf); return true; } diff --git a/mm/khugepaged.c b/mm/khugepaged.c index 3afd87f8c0b1..2a676f37c921 100644 --- a/mm/khugepaged.c +++ b/mm/khugepaged.c @@ -473,7 +473,7 @@ void khugepaged_enter_vma(struct vm_area_struct *vma, { if (!test_bit(MMF_VM_HUGEPAGE, &vma->vm_mm->flags) && khugepaged_enabled()) { - if (hugepage_vma_check(vma, vm_flags, false)) + if (hugepage_vma_check(vma, vm_flags, false, false)) __khugepaged_enter(vma->vm_mm); } } @@ -918,7 +918,7 @@ static int hugepage_vma_revalidate(struct mm_struct *mm, unsigned long address, if (!transhuge_vma_suitable(vma, address)) return SCAN_ADDRESS_RANGE; - if (!hugepage_vma_check(vma, vma->vm_flags, false)) + if (!hugepage_vma_check(vma, vma->vm_flags, false, false)) return SCAN_VMA_CHECK; /* * Anon VMA expected, the address may be unmapped then @@ -1408,7 +1408,7 @@ void collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr) * the valid THP. Add extra VM_HUGEPAGE so hugepage_vma_check() * will not fail the vma for missing VM_HUGEPAGE */ - if (!hugepage_vma_check(vma, vma->vm_flags | VM_HUGEPAGE, false)) + if (!hugepage_vma_check(vma, vma->vm_flags | VM_HUGEPAGE, false, false)) return; /* Keep pmd pgtable for uffd-wp; see comment in retract_page_tables() */ @@ -2103,7 +2103,7 @@ static unsigned int khugepaged_scan_mm_slot(unsigned int pages, progress++; break; } - if (!hugepage_vma_check(vma, vma->vm_flags, false)) { + if (!hugepage_vma_check(vma, vma->vm_flags, false, false)) { skip: progress++; continue; diff --git a/mm/memory.c b/mm/memory.c index cc3c8ed45d15..a8fd49e5bcb9 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -4986,6 +4986,7 @@ static vm_fault_t __handle_mm_fault(struct vm_area_struct *vma, .gfp_mask = __get_fault_gfp_mask(vma), }; struct mm_struct *mm = vma->vm_mm; + unsigned long vm_flags = vma->vm_flags; pgd_t *pgd; p4d_t *p4d; vm_fault_t ret; @@ -4999,7 +5000,8 @@ static vm_fault_t __handle_mm_fault(struct vm_area_struct *vma, if (!vmf.pud) return VM_FAULT_OOM; retry_pud: - if (pud_none(*vmf.pud) && __transparent_hugepage_enabled(vma)) { + if (pud_none(*vmf.pud) && + hugepage_vma_check(vma, vm_flags, false, true)) { ret = create_huge_pud(&vmf); if (!(ret & VM_FAULT_FALLBACK)) return ret; @@ -5032,7 +5034,8 @@ retry_pud: if (pud_trans_unstable(vmf.pud)) goto retry_pud; - if (pmd_none(*vmf.pmd) && __transparent_hugepage_enabled(vma)) { + if (pmd_none(*vmf.pmd) && + hugepage_vma_check(vma, vm_flags, false, true)) { ret = create_huge_pmd(&vmf); if (!(ret & VM_FAULT_FALLBACK)) return ret; -- cgit v1.2.3 From f080ec88ab21d6782de7804ac88c4bad01b1083e Mon Sep 17 00:00:00 2001 From: Yang Shi Date: Thu, 16 Jun 2022 10:48:39 -0700 Subject: mm: khugepaged: reorg some khugepaged helpers The khugepaged_{enabled|always|req_madv} are not khugepaged only anymore, move them to huge_mm.h and rename to hugepage_flags_xxx, and remove khugepaged_req_madv due to no users. Also move khugepaged_defrag to khugepaged.c since its only caller is in that file, it doesn't have to be in a header file. Link: https://lkml.kernel.org/r/20220616174840.1202070-7-shy828301@gmail.com Signed-off-by: Yang Shi Reviewed-by: Zach O'Keefe Cc: Kirill A. Shutemov Cc: Matthew Wilcox Cc: Miaohe Lin Cc: Vlastimil Babka Signed-off-by: Andrew Morton --- include/linux/huge_mm.h | 8 ++++++++ include/linux/khugepaged.h | 14 -------------- mm/huge_memory.c | 4 ++-- mm/khugepaged.c | 18 +++++++++++------- 4 files changed, 21 insertions(+), 23 deletions(-) (limited to 'include/linux') diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index cd8a6c5d9fe5..ae3d8e2fd9e2 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -116,6 +116,14 @@ extern struct kobj_attribute shmem_enabled_attr; extern unsigned long transparent_hugepage_flags; +#define hugepage_flags_enabled() \ + (transparent_hugepage_flags & \ + ((1<flags)) diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 549a02137d13..ea3243c81bad 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -120,11 +120,11 @@ bool hugepage_vma_check(struct vm_area_struct *vma, if (!in_pf && shmem_file(vma->vm_file)) return shmem_huge_enabled(vma); - if (!khugepaged_enabled()) + if (!hugepage_flags_enabled()) return false; /* THP settings require madvise. */ - if (!(vm_flags & VM_HUGEPAGE) && !khugepaged_always()) + if (!(vm_flags & VM_HUGEPAGE) && !hugepage_flags_always()) return false; /* Only regular file is valid */ diff --git a/mm/khugepaged.c b/mm/khugepaged.c index 2a676f37c921..d8ebb60aae36 100644 --- a/mm/khugepaged.c +++ b/mm/khugepaged.c @@ -472,7 +472,7 @@ void khugepaged_enter_vma(struct vm_area_struct *vma, unsigned long vm_flags) { if (!test_bit(MMF_VM_HUGEPAGE, &vma->vm_mm->flags) && - khugepaged_enabled()) { + hugepage_flags_enabled()) { if (hugepage_vma_check(vma, vm_flags, false, false)) __khugepaged_enter(vma->vm_mm); } @@ -763,6 +763,10 @@ static bool khugepaged_scan_abort(int nid) return false; } +#define khugepaged_defrag() \ + (transparent_hugepage_flags & \ + (1< Date: Fri, 17 Jun 2022 16:42:44 +0100 Subject: mm/vmscan: convert reclaim_clean_pages_from_list() to folios Patch series "nvert much of vmscan to folios" vmscan always operates on folios since it puts the pages on the LRU list. Switching all of these functions from pages to folios saves 1483 bytes of text from removing all the baggage around calling compound_page() and similar functions. This patch (of 5): This is a straightforward conversion which removes several hidden calls to compound_head, saving 330 bytes of kernel text. Link: https://lkml.kernel.org/r/20220617154248.700416-1-willy@infradead.org Link: https://lkml.kernel.org/r/20220617154248.700416-2-willy@infradead.org Signed-off-by: Matthew Wilcox (Oracle) Reviewed-by: Christoph Hellwig Signed-off-by: Andrew Morton --- include/linux/page-flags.h | 6 ++++++ mm/vmscan.c | 22 +++++++++++----------- 2 files changed, 17 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h index a5263a21b72f..109dda5a72a9 100644 --- a/include/linux/page-flags.h +++ b/include/linux/page-flags.h @@ -676,6 +676,12 @@ static __always_inline bool PageAnon(struct page *page) return folio_test_anon(page_folio(page)); } +static __always_inline bool __folio_test_movable(const struct folio *folio) +{ + return ((unsigned long)folio->mapping & PAGE_MAPPING_FLAGS) == + PAGE_MAPPING_MOVABLE; +} + static __always_inline int __PageMovable(struct page *page) { return ((unsigned long)page->mapping & PAGE_MAPPING_FLAGS) == diff --git a/mm/vmscan.c b/mm/vmscan.c index cd2201ba6d7c..439a7e849975 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -2047,7 +2047,7 @@ keep: } unsigned int reclaim_clean_pages_from_list(struct zone *zone, - struct list_head *page_list) + struct list_head *folio_list) { struct scan_control sc = { .gfp_mask = GFP_KERNEL, @@ -2055,16 +2055,16 @@ unsigned int reclaim_clean_pages_from_list(struct zone *zone, }; struct reclaim_stat stat; unsigned int nr_reclaimed; - struct page *page, *next; - LIST_HEAD(clean_pages); + struct folio *folio, *next; + LIST_HEAD(clean_folios); unsigned int noreclaim_flag; - list_for_each_entry_safe(page, next, page_list, lru) { - if (!PageHuge(page) && page_is_file_lru(page) && - !PageDirty(page) && !__PageMovable(page) && - !PageUnevictable(page)) { - ClearPageActive(page); - list_move(&page->lru, &clean_pages); + list_for_each_entry_safe(folio, next, folio_list, lru) { + if (!folio_test_hugetlb(folio) && folio_is_file_lru(folio) && + !folio_test_dirty(folio) && !__folio_test_movable(folio) && + !folio_test_unevictable(folio)) { + folio_clear_active(folio); + list_move(&folio->lru, &clean_folios); } } @@ -2075,11 +2075,11 @@ unsigned int reclaim_clean_pages_from_list(struct zone *zone, * change in the future. */ noreclaim_flag = memalloc_noreclaim_save(); - nr_reclaimed = shrink_page_list(&clean_pages, zone->zone_pgdat, &sc, + nr_reclaimed = shrink_page_list(&clean_folios, zone->zone_pgdat, &sc, &stat, true); memalloc_noreclaim_restore(noreclaim_flag); - list_splice(&clean_pages, page_list); + list_splice(&clean_folios, folio_list); mod_node_page_state(zone->zone_pgdat, NR_ISOLATED_FILE, -(long)nr_reclaimed); /* -- cgit v1.2.3 From 971164f0e606f69329e45229773d59a528e4ef3f Mon Sep 17 00:00:00 2001 From: "Matthew Wilcox (Oracle)" Date: Fri, 17 Jun 2022 18:49:59 +0100 Subject: mm: add folios_put() Patch series "Convert the swap code to be more folio-based". There's still more to do with the swap code, but this reaps a lot of the folio benefit. More than 4kB of kernel text saved (with the UEK7 kernel config). I don't know how much that's going to translate into CPU savings, but some of those compound_head() calls are on every page free, so it should be noticable. It might even be noticable just from an I-cache consumption perspective. This patch (of 22): This is just a wrapper around release_pages() for now. Place the prototype in mm.h along with folio_put() and folio_put_refs(). Link: https://lkml.kernel.org/r/20220617175020.717127-1-willy@infradead.org Link: https://lkml.kernel.org/r/20220617175020.717127-2-willy@infradead.org Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Andrew Morton --- include/linux/mm.h | 19 +++++++++++++++++++ include/linux/pagemap.h | 2 -- 2 files changed, 19 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index fb610184cf4e..c848169167d1 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1251,6 +1251,25 @@ static inline void folio_put_refs(struct folio *folio, int refs) __put_page(&folio->page); } +void release_pages(struct page **pages, int nr); + +/** + * folios_put - Decrement the reference count on an array of folios. + * @folios: The folios. + * @nr: How many folios there are. + * + * Like folio_put(), but for an array of folios. This is more efficient + * than writing the loop yourself as it will optimise the locks which + * need to be taken if the folios are freed. + * + * Context: May be called in process or interrupt context, but not in NMI + * context. May be called while holding a spinlock. + */ +static inline void folios_put(struct folio **folios, unsigned int nr) +{ + release_pages((struct page **)folios, nr); +} + static inline void put_page(struct page *page) { struct folio *folio = page_folio(page); diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h index ce96866fbec4..c399a9c5da7d 100644 --- a/include/linux/pagemap.h +++ b/include/linux/pagemap.h @@ -345,8 +345,6 @@ static inline void filemap_nr_thps_dec(struct address_space *mapping) #endif } -void release_pages(struct page **pages, int nr); - struct address_space *page_mapping(struct page *); struct address_space *folio_mapping(struct folio *); struct address_space *swapcache_mapping(struct folio *); -- cgit v1.2.3 From 61a72917f3f0dd88e93cc831de9a97ece7e8320d Mon Sep 17 00:00:00 2001 From: "Matthew Wilcox (Oracle)" Date: Fri, 17 Jun 2022 18:50:01 +0100 Subject: mm/swap: make __pagevec_lru_add static __pagevec_lru_add has no callers outside swap.c, so make it static, and move it to a more logical position in the file. Link: https://lkml.kernel.org/r/20220617175020.717127-4-willy@infradead.org Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Andrew Morton --- include/linux/pagevec.h | 1 - mm/swap.c | 126 ++++++++++++++++++++++++------------------------ 2 files changed, 63 insertions(+), 64 deletions(-) (limited to 'include/linux') diff --git a/include/linux/pagevec.h b/include/linux/pagevec.h index 67b1246f136b..b0e3540f3a4c 100644 --- a/include/linux/pagevec.h +++ b/include/linux/pagevec.h @@ -26,7 +26,6 @@ struct pagevec { }; void __pagevec_release(struct pagevec *pvec); -void __pagevec_lru_add(struct pagevec *pvec); unsigned pagevec_lookup_range(struct pagevec *pvec, struct address_space *mapping, pgoff_t *start, pgoff_t end); diff --git a/mm/swap.c b/mm/swap.c index 4265bee41bbd..cab77a5c64c7 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -228,6 +228,69 @@ static bool pagevec_add_and_need_flush(struct pagevec *pvec, struct page *page) typedef void (*move_fn_t)(struct lruvec *lruvec, struct folio *folio); +static void __pagevec_lru_add_fn(struct folio *folio, struct lruvec *lruvec) +{ + int was_unevictable = folio_test_clear_unevictable(folio); + long nr_pages = folio_nr_pages(folio); + + VM_BUG_ON_FOLIO(folio_test_lru(folio), folio); + + folio_set_lru(folio); + /* + * Is an smp_mb__after_atomic() still required here, before + * folio_evictable() tests PageMlocked, to rule out the possibility + * of stranding an evictable folio on an unevictable LRU? I think + * not, because __munlock_page() only clears PageMlocked while the LRU + * lock is held. + * + * (That is not true of __page_cache_release(), and not necessarily + * true of release_pages(): but those only clear PageMlocked after + * put_page_testzero() has excluded any other users of the page.) + */ + if (folio_evictable(folio)) { + if (was_unevictable) + __count_vm_events(UNEVICTABLE_PGRESCUED, nr_pages); + } else { + folio_clear_active(folio); + folio_set_unevictable(folio); + /* + * folio->mlock_count = !!folio_test_mlocked(folio)? + * But that leaves __mlock_page() in doubt whether another + * actor has already counted the mlock or not. Err on the + * safe side, underestimate, let page reclaim fix it, rather + * than leaving a page on the unevictable LRU indefinitely. + */ + folio->mlock_count = 0; + if (!was_unevictable) + __count_vm_events(UNEVICTABLE_PGCULLED, nr_pages); + } + + lruvec_add_folio(lruvec, folio); + trace_mm_lru_insertion(folio); +} + +/* + * Add the passed pages to the LRU, then drop the caller's refcount + * on them. Reinitialises the caller's pagevec. + */ +static void __pagevec_lru_add(struct pagevec *pvec) +{ + int i; + struct lruvec *lruvec = NULL; + unsigned long flags = 0; + + for (i = 0; i < pagevec_count(pvec); i++) { + struct folio *folio = page_folio(pvec->pages[i]); + + lruvec = folio_lruvec_relock_irqsave(folio, lruvec, &flags); + __pagevec_lru_add_fn(folio, lruvec); + } + if (lruvec) + unlock_page_lruvec_irqrestore(lruvec, flags); + release_pages(pvec->pages, pvec->nr); + pagevec_reinit(pvec); +} + static void folio_batch_move_lru(struct folio_batch *fbatch, move_fn_t move_fn) { int i; @@ -1036,69 +1099,6 @@ void __pagevec_release(struct pagevec *pvec) } EXPORT_SYMBOL(__pagevec_release); -static void __pagevec_lru_add_fn(struct folio *folio, struct lruvec *lruvec) -{ - int was_unevictable = folio_test_clear_unevictable(folio); - long nr_pages = folio_nr_pages(folio); - - VM_BUG_ON_FOLIO(folio_test_lru(folio), folio); - - folio_set_lru(folio); - /* - * Is an smp_mb__after_atomic() still required here, before - * folio_evictable() tests PageMlocked, to rule out the possibility - * of stranding an evictable folio on an unevictable LRU? I think - * not, because __munlock_page() only clears PageMlocked while the LRU - * lock is held. - * - * (That is not true of __page_cache_release(), and not necessarily - * true of release_pages(): but those only clear PageMlocked after - * put_page_testzero() has excluded any other users of the page.) - */ - if (folio_evictable(folio)) { - if (was_unevictable) - __count_vm_events(UNEVICTABLE_PGRESCUED, nr_pages); - } else { - folio_clear_active(folio); - folio_set_unevictable(folio); - /* - * folio->mlock_count = !!folio_test_mlocked(folio)? - * But that leaves __mlock_page() in doubt whether another - * actor has already counted the mlock or not. Err on the - * safe side, underestimate, let page reclaim fix it, rather - * than leaving a page on the unevictable LRU indefinitely. - */ - folio->mlock_count = 0; - if (!was_unevictable) - __count_vm_events(UNEVICTABLE_PGCULLED, nr_pages); - } - - lruvec_add_folio(lruvec, folio); - trace_mm_lru_insertion(folio); -} - -/* - * Add the passed pages to the LRU, then drop the caller's refcount - * on them. Reinitialises the caller's pagevec. - */ -void __pagevec_lru_add(struct pagevec *pvec) -{ - int i; - struct lruvec *lruvec = NULL; - unsigned long flags = 0; - - for (i = 0; i < pagevec_count(pvec); i++) { - struct folio *folio = page_folio(pvec->pages[i]); - - lruvec = folio_lruvec_relock_irqsave(folio, lruvec, &flags); - __pagevec_lru_add_fn(folio, lruvec); - } - if (lruvec) - unlock_page_lruvec_irqrestore(lruvec, flags); - release_pages(pvec->pages, pvec->nr); - pagevec_reinit(pvec); -} - /** * folio_batch_remove_exceptionals() - Prune non-folios from a batch. * @fbatch: The batch to prune -- cgit v1.2.3 From 8fd5420e16fbf44940927a0a833cffc91bd9f606 Mon Sep 17 00:00:00 2001 From: "Matthew Wilcox (Oracle)" Date: Fri, 17 Jun 2022 18:50:13 +0100 Subject: mm/swap: convert __put_page() to __folio_put() Saves 11 bytes of text by removing a check of PageTail. Link: https://lkml.kernel.org/r/20220617175020.717127-16-willy@infradead.org Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Andrew Morton --- include/linux/mm.h | 6 +++--- mm/swap.c | 14 +++++++------- net/core/page_pool.c | 2 +- 3 files changed, 11 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index c848169167d1..58dcab58a691 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -886,7 +886,7 @@ static inline struct folio *virt_to_folio(const void *x) return page_folio(page); } -void __put_page(struct page *page); +void __folio_put(struct folio *folio); void put_pages_list(struct list_head *pages); @@ -1228,7 +1228,7 @@ static inline __must_check bool try_get_page(struct page *page) static inline void folio_put(struct folio *folio) { if (folio_put_testzero(folio)) - __put_page(&folio->page); + __folio_put(folio); } /** @@ -1248,7 +1248,7 @@ static inline void folio_put(struct folio *folio) static inline void folio_put_refs(struct folio *folio, int refs) { if (folio_ref_sub_and_test(folio, refs)) - __put_page(&folio->page); + __folio_put(folio); } void release_pages(struct page **pages, int nr); diff --git a/mm/swap.c b/mm/swap.c index a5a91aec83da..d09e9ac53809 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -119,16 +119,16 @@ static void __put_compound_page(struct page *page) destroy_compound_page(page); } -void __put_page(struct page *page) +void __folio_put(struct folio *folio) { - if (unlikely(is_zone_device_page(page))) - free_zone_device_page(page); - else if (unlikely(PageCompound(page))) - __put_compound_page(page); + if (unlikely(folio_is_zone_device(folio))) + free_zone_device_page(&folio->page); + else if (unlikely(folio_test_large(folio))) + __put_compound_page(&folio->page); else - __put_single_page(page); + __put_single_page(&folio->page); } -EXPORT_SYMBOL(__put_page); +EXPORT_SYMBOL(__folio_put); /** * put_pages_list() - release a list of pages diff --git a/net/core/page_pool.c b/net/core/page_pool.c index f18e6e771993..db70e94c8df2 100644 --- a/net/core/page_pool.c +++ b/net/core/page_pool.c @@ -16,7 +16,7 @@ #include #include #include -#include /* for __put_page() */ +#include /* for put_page() */ #include #include -- cgit v1.2.3 From a73620b4dff62004856801bba299b694be66d90d Mon Sep 17 00:00:00 2001 From: "Matthew Wilcox (Oracle)" Date: Fri, 17 Jun 2022 18:50:17 +0100 Subject: mm: convert destroy_compound_page() to destroy_large_folio() All callers now have a folio, so push the folio->page conversion down to this function. Link: https://lkml.kernel.org/r/20220617175020.717127-20-willy@infradead.org Signed-off-by: Matthew Wilcox (Oracle) Signed-off-by: Andrew Morton --- include/linux/mm.h | 8 +++++--- mm/swap.c | 2 +- mm/vmscan.c | 4 ++-- 3 files changed, 8 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 58dcab58a691..9802d0c85931 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -923,10 +923,12 @@ static inline void set_compound_page_dtor(struct page *page, page[1].compound_dtor = compound_dtor; } -static inline void destroy_compound_page(struct page *page) +static inline void destroy_large_folio(struct folio *folio) { - VM_BUG_ON_PAGE(page[1].compound_dtor >= NR_COMPOUND_DTORS, page); - compound_page_dtors[page[1].compound_dtor](page); + enum compound_dtor_id dtor = folio_page(folio, 1)->compound_dtor; + + VM_BUG_ON_FOLIO(dtor >= NR_COMPOUND_DTORS, folio); + compound_page_dtors[dtor](&folio->page); } static inline int head_compound_pincount(struct page *head) diff --git a/mm/swap.c b/mm/swap.c index 5f6caa651599..1f563d857768 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -115,7 +115,7 @@ static void __folio_put_large(struct folio *folio) */ if (!folio_test_hugetlb(folio)) __page_cache_release(folio); - destroy_compound_page(&folio->page); + destroy_large_folio(folio); } void __folio_put(struct folio *folio) diff --git a/mm/vmscan.c b/mm/vmscan.c index 705ac3103aa0..77fc0b0fedb1 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -1985,7 +1985,7 @@ free_it: * appear not as the counts should be low */ if (unlikely(folio_test_large(folio))) - destroy_compound_page(&folio->page); + destroy_large_folio(folio); else list_add(&folio->lru, &free_pages); continue; @@ -2354,7 +2354,7 @@ static unsigned int move_pages_to_lru(struct lruvec *lruvec, if (unlikely(folio_test_large(folio))) { spin_unlock_irq(&lruvec->lru_lock); - destroy_compound_page(&folio->page); + destroy_large_folio(folio); spin_lock_irq(&lruvec->lru_lock); } else list_add(&folio->lru, &folios_to_free); -- cgit v1.2.3 From 96eb6b13cbd43d2c194c23bd834dcb2f0e109253 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Tue, 21 Jun 2022 15:24:15 -0700 Subject: mm-convert-destroy_compound_page-to-destroy_large_folio-fix uninline destroy_large_folio() to fix build issue Cc: "Matthew Wilcox (Oracle)" Cc: Stephen Rothwell Signed-off-by: Andrew Morton --- include/linux/mm.h | 8 +------- mm/page_alloc.c | 8 ++++++++ 2 files changed, 9 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm.h b/include/linux/mm.h index 9802d0c85931..cf746d37e504 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -923,13 +923,7 @@ static inline void set_compound_page_dtor(struct page *page, page[1].compound_dtor = compound_dtor; } -static inline void destroy_large_folio(struct folio *folio) -{ - enum compound_dtor_id dtor = folio_page(folio, 1)->compound_dtor; - - VM_BUG_ON_FOLIO(dtor >= NR_COMPOUND_DTORS, folio); - compound_page_dtors[dtor](&folio->page); -} +void destroy_large_folio(struct folio *folio); static inline int head_compound_pincount(struct page *head) { diff --git a/mm/page_alloc.c b/mm/page_alloc.c index e3c2f3223fc4..d7c72f43f1da 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -744,6 +744,14 @@ void prep_compound_page(struct page *page, unsigned int order) prep_compound_head(page, order); } +void destroy_large_folio(struct folio *folio) +{ + enum compound_dtor_id dtor = folio_page(folio, 1)->compound_dtor; + + VM_BUG_ON_FOLIO(dtor >= NR_COMPOUND_DTORS, folio); + compound_page_dtors[dtor](&folio->page); +} + #ifdef CONFIG_DEBUG_PAGEALLOC unsigned int _debug_guardpage_minorder; -- cgit v1.2.3 From 70eb3817e1596a4abbfe2c8fca31dff61e51b777 Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Fri, 17 Jun 2022 21:56:49 +0800 Subject: mm: memory_hotplug: enumerate all supported section flags Patch series "make hugetlb_optimize_vmemmap compatible with memmap_on_memory", v3. This series makes hugetlb_optimize_vmemmap compatible with memmap_on_memory. This patch (of 2): We are almost running out of section flags, only one bit is available in the worst case (powerpc with 256k pages). However, there are still some free bits (in ->section_mem_map) on other architectures (e.g. x86_64 has 10 bits available, arm64 has 8 bits available with worst case of 64K pages). We have hard coded those numbers in code, it is inconvenient to use those bits on other architectures except powerpc. So transfer those section flags to enumeration to make it easy to add new section flags in the future. Also, move SECTION_TAINT_ZONE_DEVICE into the scope of CONFIG_ZONE_DEVICE to save a bit on non-zone-device case. Link: https://lkml.kernel.org/r/20220617135650.74901-1-songmuchun@bytedance.com Link: https://lkml.kernel.org/r/20220617135650.74901-2-songmuchun@bytedance.com Signed-off-by: Muchun Song Reviewed-by: David Hildenbrand Cc: Jonathan Corbet Cc: Mike Kravetz Cc: Oscar Salvador Cc: Paul E. McKenney Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- include/linux/mmzone.h | 44 +++++++++++++++++++++++++++++++++++--------- mm/memory_hotplug.c | 6 ++++++ mm/sparse.c | 2 +- 3 files changed, 42 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index aab70355d64f..932843c6459b 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -1418,16 +1418,35 @@ extern size_t mem_section_usage_size(void); * (equal SECTION_SIZE_BITS - PAGE_SHIFT), and the * worst combination is powerpc with 256k pages, * which results in PFN_SECTION_SHIFT equal 6. - * To sum it up, at least 6 bits are available. + * To sum it up, at least 6 bits are available on all architectures. + * However, we can exceed 6 bits on some other architectures except + * powerpc (e.g. 15 bits are available on x86_64, 13 bits are available + * with the worst case of 64K pages on arm64) if we make sure the + * exceeded bit is not applicable to powerpc. */ -#define SECTION_MARKED_PRESENT (1UL<<0) -#define SECTION_HAS_MEM_MAP (1UL<<1) -#define SECTION_IS_ONLINE (1UL<<2) -#define SECTION_IS_EARLY (1UL<<3) -#define SECTION_TAINT_ZONE_DEVICE (1UL<<4) -#define SECTION_MAP_LAST_BIT (1UL<<5) -#define SECTION_MAP_MASK (~(SECTION_MAP_LAST_BIT-1)) -#define SECTION_NID_SHIFT 6 +enum { + SECTION_MARKED_PRESENT_BIT, + SECTION_HAS_MEM_MAP_BIT, + SECTION_IS_ONLINE_BIT, + SECTION_IS_EARLY_BIT, +#ifdef CONFIG_ZONE_DEVICE + SECTION_TAINT_ZONE_DEVICE_BIT, +#endif + SECTION_MAP_LAST_BIT, +}; + +enum { + SECTION_MARKED_PRESENT = BIT(SECTION_MARKED_PRESENT_BIT), + SECTION_HAS_MEM_MAP = BIT(SECTION_HAS_MEM_MAP_BIT), + SECTION_IS_ONLINE = BIT(SECTION_IS_ONLINE_BIT), + SECTION_IS_EARLY = BIT(SECTION_IS_EARLY_BIT), +#ifdef CONFIG_ZONE_DEVICE + SECTION_TAINT_ZONE_DEVICE = BIT(SECTION_TAINT_ZONE_DEVICE_BIT), +#endif +}; + +#define SECTION_MAP_MASK (~(BIT(SECTION_MAP_LAST_BIT) - 1)) +#define SECTION_NID_SHIFT SECTION_MAP_LAST_BIT static inline struct page *__section_mem_map_addr(struct mem_section *section) { @@ -1466,12 +1485,19 @@ static inline int online_section(struct mem_section *section) return (section && (section->section_mem_map & SECTION_IS_ONLINE)); } +#ifdef CONFIG_ZONE_DEVICE static inline int online_device_section(struct mem_section *section) { unsigned long flags = SECTION_IS_ONLINE | SECTION_TAINT_ZONE_DEVICE; return section && ((section->section_mem_map & flags) == flags); } +#else +static inline int online_device_section(struct mem_section *section) +{ + return 0; +} +#endif static inline int online_section_nr(unsigned long nr) { diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c index 84990a14d51a..a2a6d280054f 100644 --- a/mm/memory_hotplug.c +++ b/mm/memory_hotplug.c @@ -670,12 +670,18 @@ static void __meminit resize_pgdat_range(struct pglist_data *pgdat, unsigned lon } +#ifdef CONFIG_ZONE_DEVICE static void section_taint_zone_device(unsigned long pfn) { struct mem_section *ms = __pfn_to_section(pfn); ms->section_mem_map |= SECTION_TAINT_ZONE_DEVICE; } +#else +static inline void section_taint_zone_device(unsigned long pfn) +{ +} +#endif /* * Associate the pfn range with the given zone, initializing the memmaps diff --git a/mm/sparse.c b/mm/sparse.c index cb3bfae64036..e5a8a3a0edd7 100644 --- a/mm/sparse.c +++ b/mm/sparse.c @@ -281,7 +281,7 @@ static unsigned long sparse_encode_mem_map(struct page *mem_map, unsigned long p { unsigned long coded_mem_map = (unsigned long)(mem_map - (section_nr_to_pfn(pnum))); - BUILD_BUG_ON(SECTION_MAP_LAST_BIT > (1UL< PFN_SECTION_SHIFT); BUG_ON(coded_mem_map & ~SECTION_MAP_MASK); return coded_mem_map; } -- cgit v1.2.3 From 3a30769056578608d3685c65f0104cb2d407072f Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Mon, 20 Jun 2022 19:06:15 +0800 Subject: mm-memory_hotplug-enumerate-all-supported-section-flags-v5 replace enum with defines per David Link: https://lkml.kernel.org/r/20220620110616.12056-2-songmuchun@bytedance.com Signed-off-by: Muchun Song Cc: David Hildenbrand Cc: Jonathan Corbet Cc: Mike Kravetz Cc: Oscar Salvador Cc: Paul E. McKenney Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- include/linux/mmzone.h | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 932843c6459b..2b5757752333 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -1435,16 +1435,13 @@ enum { SECTION_MAP_LAST_BIT, }; -enum { - SECTION_MARKED_PRESENT = BIT(SECTION_MARKED_PRESENT_BIT), - SECTION_HAS_MEM_MAP = BIT(SECTION_HAS_MEM_MAP_BIT), - SECTION_IS_ONLINE = BIT(SECTION_IS_ONLINE_BIT), - SECTION_IS_EARLY = BIT(SECTION_IS_EARLY_BIT), +#define SECTION_MARKED_PRESENT BIT(SECTION_MARKED_PRESENT_BIT) +#define SECTION_HAS_MEM_MAP BIT(SECTION_HAS_MEM_MAP_BIT) +#define SECTION_IS_ONLINE BIT(SECTION_IS_ONLINE_BIT) +#define SECTION_IS_EARLY BIT(SECTION_IS_EARLY_BIT) #ifdef CONFIG_ZONE_DEVICE - SECTION_TAINT_ZONE_DEVICE = BIT(SECTION_TAINT_ZONE_DEVICE_BIT), +#define SECTION_TAINT_ZONE_DEVICE BIT(SECTION_TAINT_ZONE_DEVICE_BIT) #endif -}; - #define SECTION_MAP_MASK (~(BIT(SECTION_MAP_LAST_BIT) - 1)) #define SECTION_NID_SHIFT SECTION_MAP_LAST_BIT -- cgit v1.2.3 From 5694be584f2ca8089276e94c11786025ab476307 Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Fri, 17 Jun 2022 21:56:50 +0800 Subject: mm: memory_hotplug: make hugetlb_optimize_vmemmap compatible with memmap_on_memory For now, the feature of hugetlb_free_vmemmap is not compatible with the feature of memory_hotplug.memmap_on_memory, and hugetlb_free_vmemmap takes precedence over memory_hotplug.memmap_on_memory. However, someone wants to make memory_hotplug.memmap_on_memory takes precedence over hugetlb_free_vmemmap since memmap_on_memory makes it more likely to succeed memory hotplug in close-to-OOM situations. So the decision of making hugetlb_free_vmemmap take precedence is not wise and elegant. The proper approach is to have hugetlb_vmemmap.c do the check whether the section which the HugeTLB pages belong to can be optimized. If the section's vmemmap pages are allocated from the added memory block itself, hugetlb_free_vmemmap should refuse to optimize the vmemmap, otherwise, do the optimization. Then both kernel parameters are compatible. So this patch introduces VmemmapSelfHosted to mask any non-optimizable vmemmap pages. The hugetlb_vmemmap can use this flag to detect if a vmemmap page can be optimized. Link: https://lkml.kernel.org/r/20220617135650.74901-3-songmuchun@bytedance.com Signed-off-by: Muchun Song Co-developed-by: Oscar Salvador Signed-off-by: Oscar Salvador Acked-by: David Hildenbrand Cc: Jonathan Corbet Cc: Mike Kravetz Cc: Paul E. McKenney Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- Documentation/admin-guide/kernel-parameters.txt | 22 ++++++------ Documentation/admin-guide/sysctl/vm.rst | 5 ++- include/linux/memory_hotplug.h | 9 ----- include/linux/page-flags.h | 11 ++++++ mm/hugetlb_vmemmap.c | 47 ++++++++++++++++++++----- mm/memory_hotplug.c | 27 ++++++-------- 6 files changed, 74 insertions(+), 47 deletions(-) (limited to 'include/linux') diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt index 8c0ea6b6c6a9..2cacd4f8deb7 100644 --- a/Documentation/admin-guide/kernel-parameters.txt +++ b/Documentation/admin-guide/kernel-parameters.txt @@ -1722,9 +1722,11 @@ Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y, the default is on. - This is not compatible with memory_hotplug.memmap_on_memory. - If both parameters are enabled, hugetlb_free_vmemmap takes - precedence over memory_hotplug.memmap_on_memory. + Note that the vmemmap pages may be allocated from the added + memory block itself when memory_hotplug.memmap_on_memory is + enabled, those vmemmap pages cannot be optimized even if this + feature is enabled. Other vmemmap pages not allocated from + the added memory block itself do not be affected. hung_task_panic= [KNL] Should the hung task detector generate panics. @@ -3068,10 +3070,12 @@ [KNL,X86,ARM] Boolean flag to enable this feature. Format: {on | off (default)} When enabled, runtime hotplugged memory will - allocate its internal metadata (struct pages) - from the hotadded memory which will allow to - hotadd a lot of memory without requiring - additional memory to do so. + allocate its internal metadata (struct pages, + those vmemmap pages cannot be optimized even + if hugetlb_free_vmemmap is enabled) from the + hotadded memory which will allow to hotadd a + lot of memory without requiring additional + memory to do so. This feature is disabled by default because it has some implication on large (e.g. GB) allocations in some configurations (e.g. small @@ -3081,10 +3085,6 @@ Note that even when enabled, there are a few cases where the feature is not effective. - This is not compatible with hugetlb_free_vmemmap. If - both parameters are enabled, hugetlb_free_vmemmap takes - precedence over memory_hotplug.memmap_on_memory. - memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest Format: default : 0 diff --git a/Documentation/admin-guide/sysctl/vm.rst b/Documentation/admin-guide/sysctl/vm.rst index 9cbc88b8c5cf..5a810ea6dc77 100644 --- a/Documentation/admin-guide/sysctl/vm.rst +++ b/Documentation/admin-guide/sysctl/vm.rst @@ -565,9 +565,8 @@ See Documentation/admin-guide/mm/hugetlbpage.rst hugetlb_optimize_vmemmap ======================== -This knob is not available when memory_hotplug.memmap_on_memory (kernel parameter) -is configured or the size of 'struct page' (a structure defined in -include/linux/mm_types.h) is not power of two (an unusual system config could +This knob is not available when the size of 'struct page' (a structure defined +in include/linux/mm_types.h) is not power of two (an unusual system config could result in this). Enable (set to 1) or disable (set to 0) the feature of optimizing vmemmap pages diff --git a/include/linux/memory_hotplug.h b/include/linux/memory_hotplug.h index 20d7edf62a6a..e0b2209ab71c 100644 --- a/include/linux/memory_hotplug.h +++ b/include/linux/memory_hotplug.h @@ -351,13 +351,4 @@ void arch_remove_linear_mapping(u64 start, u64 size); extern bool mhp_supports_memmap_on_memory(unsigned long size); #endif /* CONFIG_MEMORY_HOTPLUG */ -#ifdef CONFIG_MHP_MEMMAP_ON_MEMORY -bool mhp_memmap_on_memory(void); -#else -static inline bool mhp_memmap_on_memory(void) -{ - return false; -} -#endif - #endif /* __LINUX_MEMORY_HOTPLUG_H */ diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h index 109dda5a72a9..f2ff65f1bf83 100644 --- a/include/linux/page-flags.h +++ b/include/linux/page-flags.h @@ -193,6 +193,11 @@ enum pageflags { /* Only valid for buddy pages. Used to track pages that are reported */ PG_reported = PG_uptodate, + +#ifdef CONFIG_MEMORY_HOTPLUG + /* For self-hosted memmap pages */ + PG_vmemmap_self_hosted = PG_owner_priv_1, +#endif }; #define PAGEFLAGS_MASK ((1UL << NR_PAGEFLAGS) - 1) @@ -628,6 +633,12 @@ PAGEFLAG_FALSE(SkipKASanPoison, skip_kasan_poison) */ __PAGEFLAG(Reported, reported, PF_NO_COMPOUND) +#ifdef CONFIG_MEMORY_HOTPLUG +PAGEFLAG(VmemmapSelfHosted, vmemmap_self_hosted, PF_ANY) +#else +PAGEFLAG_FALSE(VmemmapSelfHosted, vmemmap_self_hosted) +#endif + /* * On an anonymous page mapped into a user virtual memory area, * page->mapping points to its anon_vma, not to a struct address_space; diff --git a/mm/hugetlb_vmemmap.c b/mm/hugetlb_vmemmap.c index ba29c15c53d6..c05a336f9bfa 100644 --- a/mm/hugetlb_vmemmap.c +++ b/mm/hugetlb_vmemmap.c @@ -10,7 +10,6 @@ */ #define pr_fmt(fmt) "HugeTLB: " fmt -#include #include "hugetlb_vmemmap.h" /* @@ -97,18 +96,50 @@ int hugetlb_vmemmap_alloc(struct hstate *h, struct page *head) return ret; } +static unsigned int vmemmap_optimizable_pages(struct hstate *h, + struct page *head) +{ + struct mem_section *ms; + struct page *vmemmap_page; + unsigned long pfn = page_to_pfn(head); + + if (READ_ONCE(vmemmap_optimize_mode) == VMEMMAP_OPTIMIZE_OFF) + return 0; + + ms = __pfn_to_section(pfn); + vmemmap_page = sparse_decode_mem_map(ms->section_mem_map, + pfn_to_section_nr(pfn)); + /* + * Only the vmemmap pages' vmemmap may be marked as VmemmapSelfHosted. + * + * Due to HugeTLB alignment requirements, and the vmemmap pages being + * at the start of the hotplugged memory region. Checking any vmemmap + * page's vmemmap is fine. + * + * [ hotplugged memory ] + * [ vmemmap ][ usable memory ] + * ^ | | | + * +---+ | | + * ^ | | + * +--------+ | + * ^ | + * +-----------------+ + */ + if (PageVmemmapSelfHosted(vmemmap_page)) + return 0; + + return hugetlb_optimize_vmemmap_pages(h); +} + void hugetlb_vmemmap_free(struct hstate *h, struct page *head) { unsigned long vmemmap_addr = (unsigned long)head; unsigned long vmemmap_end, vmemmap_reuse, vmemmap_pages; - vmemmap_pages = hugetlb_optimize_vmemmap_pages(h); + vmemmap_pages = vmemmap_optimizable_pages(h, head); if (!vmemmap_pages) return; - if (READ_ONCE(vmemmap_optimize_mode) == VMEMMAP_OPTIMIZE_OFF) - return; - static_branch_inc(&hugetlb_optimize_vmemmap_key); vmemmap_addr += RESERVE_VMEMMAP_SIZE; @@ -199,10 +230,10 @@ static struct ctl_table hugetlb_vmemmap_sysctls[] = { static __init int hugetlb_vmemmap_sysctls_init(void) { /* - * If "memory_hotplug.memmap_on_memory" is enabled or "struct page" - * crosses page boundaries, the vmemmap pages cannot be optimized. + * If "struct page" crosses page boundaries, the vmemmap pages cannot + * be optimized. */ - if (!mhp_memmap_on_memory() && is_power_of_2(sizeof(struct page))) + if (is_power_of_2(sizeof(struct page))) register_sysctl_init("vm", hugetlb_vmemmap_sysctls); return 0; diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c index a2a6d280054f..99ecb2b3ff53 100644 --- a/mm/memory_hotplug.c +++ b/mm/memory_hotplug.c @@ -43,30 +43,22 @@ #include "shuffle.h" #ifdef CONFIG_MHP_MEMMAP_ON_MEMORY -static int memmap_on_memory_set(const char *val, const struct kernel_param *kp) -{ - if (hugetlb_optimize_vmemmap_enabled()) - return 0; - return param_set_bool(val, kp); -} - -static const struct kernel_param_ops memmap_on_memory_ops = { - .flags = KERNEL_PARAM_OPS_FL_NOARG, - .set = memmap_on_memory_set, - .get = param_get_bool, -}; - /* * memory_hotplug.memmap_on_memory parameter */ static bool memmap_on_memory __ro_after_init; -module_param_cb(memmap_on_memory, &memmap_on_memory_ops, &memmap_on_memory, 0444); +module_param(memmap_on_memory, bool, 0444); MODULE_PARM_DESC(memmap_on_memory, "Enable memmap on memory for memory hotplug"); -bool mhp_memmap_on_memory(void) +static inline bool mhp_memmap_on_memory(void) { return memmap_on_memory; } +#else +static inline bool mhp_memmap_on_memory(void) +{ + return false; +} #endif enum { @@ -1035,7 +1027,7 @@ int mhp_init_memmap_on_memory(unsigned long pfn, unsigned long nr_pages, struct zone *zone) { unsigned long end_pfn = pfn + nr_pages; - int ret; + int ret, i; ret = kasan_add_zero_shadow(__va(PFN_PHYS(pfn)), PFN_PHYS(nr_pages)); if (ret) @@ -1043,6 +1035,9 @@ int mhp_init_memmap_on_memory(unsigned long pfn, unsigned long nr_pages, move_pfn_range_to_zone(zone, pfn, nr_pages, NULL, MIGRATE_UNMOVABLE); + for (i = 0; i < nr_pages; i++) + SetPageVmemmapSelfHosted(pfn_to_page(pfn + i)); + /* * It might be that the vmemmap_pages fully span sections. If that is * the case, mark those sections online here as otherwise they will be -- cgit v1.2.3 From 2ec804f8aaa61f9617792faed963d62db4fd7af4 Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Tue, 21 Jun 2022 20:56:48 +0800 Subject: mm: memcontrol: remove dead code and comments MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Patch series "Use obj_cgroup APIs to charge the LRU pages", v6. With the following patchsets applied, all the kernel memory is charged with the new APIs of obj_cgroup: commit f2fe7b09a52b ("mm: memcg/slab: charge individual slab objects instead of pages") commit b4e0b68fbd9d ("mm: memcontrol: use obj_cgroup APIs to charge kmem pages") But user memory allocations (LRU pages) pinning memcgs for a long time - it exists at a larger scale and is causing recurring problems in the real world: page cache doesn't get reclaimed for a long time, or is used by the second, third, fourth, ... instance of the same job that was restarted into a new cgroup every time. Unreclaimable dying cgroups pile up, waste memory, and make page reclaim very inefficient. We can convert LRU pages and most other raw memcg pins to the objcg direction to fix this problem, and then the LRU pages will not pin the memcgs. This patchset aims to make the LRU pages to drop the reference to memory cgroup by using the APIs of obj_cgroup. Finally, we can see that the number of the dying cgroups will not increase if we run the following test script. #!/bin/bash dd if=/dev/zero of=temp bs=4096 count=1 cat /proc/cgroups | grep memory for i in {0..2000} do mkdir /sys/fs/cgroup/memory/test$i echo $$ > /sys/fs/cgroup/memory/test$i/cgroup.procs cat temp >> log echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir /sys/fs/cgroup/memory/test$i done cat /proc/cgroups | grep memory rm -f temp log This patch (of 11): Since no-hierarchy mode is deprecated after commit bef8620cd8e0 ("mm: memcg: deprecate the non-hierarchical mode") so parent_mem_cgroup() cannot return a NULL except root memcg, however, root memcg cannot be offline, so it is safe to drop the check of returned value of parent_mem_cgroup(). Remove those dead code. The comments in memcg_offline_kmem() above memcg_reparent_list_lrus() are out of date since commit 5abc1e37afa0 ("mm: list_lru: allocate list_lru_one only when needed") There is no ordering requirement between memcg_reparent_list_lrus() and memcg_reparent_objcgs(), so remove those outdated comments. Link: https://lkml.kernel.org/r/20220621125658.64935-1-songmuchun@bytedance.com Link: https://lkml.kernel.org/r/20220621125658.64935-2-songmuchun@bytedance.com Signed-off-by: Muchun Song Acked-by: Roman Gushchin Cc: Johannes Weiner Cc: Waiman Long Cc: Michal Hocko Cc: Shakeel Butt Cc: Xiongchun Duan Cc: Michal Koutný Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 3 +-- mm/memcontrol.c | 12 ------------ mm/vmscan.c | 6 +----- 3 files changed, 2 insertions(+), 19 deletions(-) (limited to 'include/linux') diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 4d31ce55b1c0..318d8880d62a 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -866,8 +866,7 @@ static inline struct mem_cgroup *lruvec_memcg(struct lruvec *lruvec) * parent_mem_cgroup - find the accounting parent of a memcg * @memcg: memcg whose parent to find * - * Returns the parent memcg, or NULL if this is the root or the memory - * controller is in legacy no-hierarchy mode. + * Returns the parent memcg, or NULL if this is the root. */ static inline struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg) { diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 163492b9efa9..fc706d6fc265 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -3684,17 +3684,7 @@ static void memcg_offline_kmem(struct mem_cgroup *memcg) return; parent = parent_mem_cgroup(memcg); - if (!parent) - parent = root_mem_cgroup; - memcg_reparent_objcgs(memcg, parent); - - /* - * After we have finished memcg_reparent_objcgs(), all list_lrus - * corresponding to this cgroup are guaranteed to remain empty. - * The ordering is imposed by list_lru_node->lock taken by - * memcg_reparent_list_lrus(). - */ memcg_reparent_list_lrus(memcg, parent); } #else @@ -7195,8 +7185,6 @@ static struct mem_cgroup *mem_cgroup_id_get_online(struct mem_cgroup *memcg) break; } memcg = parent_mem_cgroup(memcg); - if (!memcg) - memcg = root_mem_cgroup; } return memcg; } diff --git a/mm/vmscan.c b/mm/vmscan.c index 88fce64cfa96..b68b0216424d 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -409,13 +409,9 @@ void reparent_shrinker_deferred(struct mem_cgroup *memcg) { int i, nid; long nr; - struct mem_cgroup *parent; + struct mem_cgroup *parent = parent_mem_cgroup(memcg); struct shrinker_info *child_info, *parent_info; - parent = parent_mem_cgroup(memcg); - if (!parent) - parent = root_mem_cgroup; - /* Prevent from concurrent shrinker_info expand */ down_read(&shrinker_rwsem); for_each_node(nid) { -- cgit v1.2.3 From 9c8fd315a46b94667f45771256f850a5dede4414 Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Tue, 21 Jun 2022 20:56:49 +0800 Subject: mm: rename unlock_page_lruvec{_irq, _irqrestore} to lruvec_unlock{_irq, _irqrestore} MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It is weird to use folio_lruvec_lock() variants and unlock_page_lruvec() variants together, e.g. locking folio and unlocking page. So rename unlock_page_lruvec{_irq, _irqrestore} to lruvec_unlock{_irq, _irqrestore}. Link: https://lkml.kernel.org/r/20220621125658.64935-3-songmuchun@bytedance.com Signed-off-by: Muchun Song Acked-by: Roman Gushchin Cc: Johannes Weiner Cc: Michal Hocko Cc: Michal Koutný Cc: Shakeel Butt Cc: Waiman Long Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 10 +++++----- mm/compaction.c | 12 ++++++------ mm/huge_memory.c | 2 +- mm/mlock.c | 2 +- mm/swap.c | 14 +++++++------- mm/vmscan.c | 4 ++-- 6 files changed, 22 insertions(+), 22 deletions(-) (limited to 'include/linux') diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 318d8880d62a..d0c0da7cafb7 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -1579,17 +1579,17 @@ static inline struct lruvec *parent_lruvec(struct lruvec *lruvec) return mem_cgroup_lruvec(memcg, lruvec_pgdat(lruvec)); } -static inline void unlock_page_lruvec(struct lruvec *lruvec) +static inline void lruvec_unlock(struct lruvec *lruvec) { spin_unlock(&lruvec->lru_lock); } -static inline void unlock_page_lruvec_irq(struct lruvec *lruvec) +static inline void lruvec_unlock_irq(struct lruvec *lruvec) { spin_unlock_irq(&lruvec->lru_lock); } -static inline void unlock_page_lruvec_irqrestore(struct lruvec *lruvec, +static inline void lruvec_unlock_irqrestore(struct lruvec *lruvec, unsigned long flags) { spin_unlock_irqrestore(&lruvec->lru_lock, flags); @@ -1611,7 +1611,7 @@ static inline struct lruvec *folio_lruvec_relock_irq(struct folio *folio, if (folio_matches_lruvec(folio, locked_lruvec)) return locked_lruvec; - unlock_page_lruvec_irq(locked_lruvec); + lruvec_unlock_irq(locked_lruvec); } return folio_lruvec_lock_irq(folio); @@ -1625,7 +1625,7 @@ static inline struct lruvec *folio_lruvec_relock_irqsave(struct folio *folio, if (folio_matches_lruvec(folio, locked_lruvec)) return locked_lruvec; - unlock_page_lruvec_irqrestore(locked_lruvec, *flags); + lruvec_unlock_irqrestore(locked_lruvec, *flags); } return folio_lruvec_lock_irqsave(folio, flags); diff --git a/mm/compaction.c b/mm/compaction.c index 1f89b969c12b..46351a14eed2 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -864,7 +864,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn, */ if (!(low_pfn % COMPACT_CLUSTER_MAX)) { if (locked) { - unlock_page_lruvec_irqrestore(locked, flags); + lruvec_unlock_irqrestore(locked, flags); locked = NULL; } @@ -977,7 +977,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn, if (unlikely(__PageMovable(page)) && !PageIsolated(page)) { if (locked) { - unlock_page_lruvec_irqrestore(locked, flags); + lruvec_unlock_irqrestore(locked, flags); locked = NULL; } @@ -1060,7 +1060,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn, /* If we already hold the lock, we can skip some rechecking */ if (lruvec != locked) { if (locked) - unlock_page_lruvec_irqrestore(locked, flags); + lruvec_unlock_irqrestore(locked, flags); compact_lock_irqsave(&lruvec->lru_lock, &flags, cc); locked = lruvec; @@ -1119,7 +1119,7 @@ isolate_success_no_list: isolate_fail_put: /* Avoid potential deadlock in freeing page under lru_lock */ if (locked) { - unlock_page_lruvec_irqrestore(locked, flags); + lruvec_unlock_irqrestore(locked, flags); locked = NULL; } put_page(page); @@ -1135,7 +1135,7 @@ isolate_fail: */ if (nr_isolated) { if (locked) { - unlock_page_lruvec_irqrestore(locked, flags); + lruvec_unlock_irqrestore(locked, flags); locked = NULL; } putback_movable_pages(&cc->migratepages); @@ -1167,7 +1167,7 @@ isolate_fail: isolate_abort: if (locked) - unlock_page_lruvec_irqrestore(locked, flags); + lruvec_unlock_irqrestore(locked, flags); if (page) { SetPageLRU(page); put_page(page); diff --git a/mm/huge_memory.c b/mm/huge_memory.c index ea3243c81bad..05656438e95c 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -2519,7 +2519,7 @@ static void __split_huge_page(struct page *page, struct list_head *list, } ClearPageCompound(head); - unlock_page_lruvec(lruvec); + lruvec_unlock(lruvec); /* Caller disabled irqs, so they are still disabled here */ split_page_owner(head, nr); diff --git a/mm/mlock.c b/mm/mlock.c index 7032f6dd0ce1..d9039fb9c56b 100644 --- a/mm/mlock.c +++ b/mm/mlock.c @@ -205,7 +205,7 @@ static void mlock_pagevec(struct pagevec *pvec) } if (lruvec) - unlock_page_lruvec_irq(lruvec); + lruvec_unlock_irq(lruvec); release_pages(pvec->pages, pvec->nr); pagevec_reinit(pvec); } diff --git a/mm/swap.c b/mm/swap.c index 1f563d857768..127ef4db394f 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -86,7 +86,7 @@ static void __page_cache_release(struct folio *folio) lruvec = folio_lruvec_lock_irqsave(folio, &flags); lruvec_del_folio(lruvec, folio); __folio_clear_lru_flags(folio); - unlock_page_lruvec_irqrestore(lruvec, flags); + lruvec_unlock_irqrestore(lruvec, flags); } /* See comment on folio_test_mlocked in release_pages() */ if (unlikely(folio_test_mlocked(folio))) { @@ -249,7 +249,7 @@ static void folio_batch_move_lru(struct folio_batch *fbatch, move_fn_t move_fn) } if (lruvec) - unlock_page_lruvec_irqrestore(lruvec, flags); + lruvec_unlock_irqrestore(lruvec, flags); folios_put(fbatch->folios, folio_batch_count(fbatch)); folio_batch_init(fbatch); } @@ -392,7 +392,7 @@ static void folio_activate(struct folio *folio) if (folio_test_clear_lru(folio)) { lruvec = folio_lruvec_lock_irq(folio); folio_activate_fn(lruvec, folio); - unlock_page_lruvec_irq(lruvec); + lruvec_unlock_irq(lruvec); folio_set_lru(folio); } } @@ -948,7 +948,7 @@ void release_pages(struct page **pages, int nr) * same lruvec. The lock is held only if lruvec != NULL. */ if (lruvec && ++lock_batch == SWAP_CLUSTER_MAX) { - unlock_page_lruvec_irqrestore(lruvec, flags); + lruvec_unlock_irqrestore(lruvec, flags); lruvec = NULL; } @@ -957,7 +957,7 @@ void release_pages(struct page **pages, int nr) if (folio_is_zone_device(folio)) { if (lruvec) { - unlock_page_lruvec_irqrestore(lruvec, flags); + lruvec_unlock_irqrestore(lruvec, flags); lruvec = NULL; } if (put_devmap_managed_page(&folio->page)) @@ -972,7 +972,7 @@ void release_pages(struct page **pages, int nr) if (folio_test_large(folio)) { if (lruvec) { - unlock_page_lruvec_irqrestore(lruvec, flags); + lruvec_unlock_irqrestore(lruvec, flags); lruvec = NULL; } __folio_put_large(folio); @@ -1006,7 +1006,7 @@ void release_pages(struct page **pages, int nr) list_add(&folio->lru, &pages_to_free); } if (lruvec) - unlock_page_lruvec_irqrestore(lruvec, flags); + lruvec_unlock_irqrestore(lruvec, flags); mem_cgroup_uncharge_list(&pages_to_free); free_unref_page_list(&pages_to_free); diff --git a/mm/vmscan.c b/mm/vmscan.c index b68b0216424d..6a554712ef5d 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -2257,7 +2257,7 @@ int folio_isolate_lru(struct folio *folio) folio_get(folio); lruvec = folio_lruvec_lock_irq(folio); lruvec_del_folio(lruvec, folio); - unlock_page_lruvec_irq(lruvec); + lruvec_unlock_irq(lruvec); ret = 0; } @@ -4886,7 +4886,7 @@ void check_move_unevictable_pages(struct pagevec *pvec) if (lruvec) { __count_vm_events(UNEVICTABLE_PGRESCUED, pgrescued); __count_vm_events(UNEVICTABLE_PGSCANNED, pgscanned); - unlock_page_lruvec_irq(lruvec); + lruvec_unlock_irq(lruvec); } else if (pgscanned) { count_vm_events(UNEVICTABLE_PGSCANNED, pgscanned); } -- cgit v1.2.3 From 41fa167bb3362179afdd312e97aef1e20f7ec8b2 Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Tue, 21 Jun 2022 20:56:50 +0800 Subject: mm: memcontrol: prepare objcg API for non-kmem usage MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Pagecache pages are charged at the allocation time and holding a reference to the original memory cgroup until being reclaimed. Depending on the memory pressure, specific patterns of the page sharing between different cgroups and the cgroup creation and destruction rates, a large number of dying memory cgroups can be pinned by pagecache pages. It makes the page reclaim less efficient and wastes memory. We can convert LRU pages and most other raw memcg pins to the objcg direction to fix this problem, and then the page->memcg will always point to an object cgroup pointer. Therefore, the infrastructure of objcg no longer only serves CONFIG_MEMCG_KMEM. In this patch, we move the infrastructure of the objcg out of the scope of the CONFIG_MEMCG_KMEM so that the LRU pages can reuse it to charge pages. We know that the LRU pages are not accounted at the root level. But the page->memcg_data points to the root_mem_cgroup. So the page->memcg_data of the LRU pages always points to a valid pointer. But the root_mem_cgroup dose not have an object cgroup. If we use obj_cgroup APIs to charge the LRU pages, we should set the page->memcg_data to a root object cgroup. So we also allocate an object cgroup for the root_mem_cgroup. Link: https://lkml.kernel.org/r/20220621125658.64935-4-songmuchun@bytedance.com Signed-off-by: Muchun Song Acked-by: Johannes Weiner Reviewed-by: Michal Koutný Acked-by: Roman Gushchin Cc: Michal Hocko Cc: Shakeel Butt Cc: Waiman Long Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 2 +- mm/memcontrol.c | 56 +++++++++++++++++++++++++++------------------- 2 files changed, 34 insertions(+), 24 deletions(-) (limited to 'include/linux') diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index d0c0da7cafb7..111eda6ff1ce 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -321,10 +321,10 @@ struct mem_cgroup { #ifdef CONFIG_MEMCG_KMEM int kmemcg_id; +#endif struct obj_cgroup __rcu *objcg; /* list of inherited objcgs, protected by objcg_lock */ struct list_head objcg_list; -#endif MEMCG_PADDING(_pad2_); diff --git a/mm/memcontrol.c b/mm/memcontrol.c index fc706d6fc265..3c489651d312 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -252,9 +252,9 @@ struct mem_cgroup *vmpressure_to_memcg(struct vmpressure *vmpr) return container_of(vmpr, struct mem_cgroup, vmpressure); } -#ifdef CONFIG_MEMCG_KMEM static DEFINE_SPINLOCK(objcg_lock); +#ifdef CONFIG_MEMCG_KMEM bool mem_cgroup_kmem_disabled(void) { return cgroup_memory_nokmem; @@ -263,12 +263,10 @@ bool mem_cgroup_kmem_disabled(void) static void obj_cgroup_uncharge_pages(struct obj_cgroup *objcg, unsigned int nr_pages); -static void obj_cgroup_release(struct percpu_ref *ref) +static void obj_cgroup_release_bytes(struct obj_cgroup *objcg) { - struct obj_cgroup *objcg = container_of(ref, struct obj_cgroup, refcnt); unsigned int nr_bytes; unsigned int nr_pages; - unsigned long flags; /* * At this point all allocated objects are freed, and @@ -282,9 +280,9 @@ static void obj_cgroup_release(struct percpu_ref *ref) * 3) CPU1: a process from another memcg is allocating something, * the stock if flushed, * objcg->nr_charged_bytes = PAGE_SIZE - 92 - * 5) CPU0: we do release this object, + * 4) CPU0: we do release this object, * 92 bytes are added to stock->nr_bytes - * 6) CPU0: stock is flushed, + * 5) CPU0: stock is flushed, * 92 bytes are added to objcg->nr_charged_bytes * * In the result, nr_charged_bytes == PAGE_SIZE. @@ -296,6 +294,19 @@ static void obj_cgroup_release(struct percpu_ref *ref) if (nr_pages) obj_cgroup_uncharge_pages(objcg, nr_pages); +} +#else +static inline void obj_cgroup_release_bytes(struct obj_cgroup *objcg) +{ +} +#endif + +static void obj_cgroup_release(struct percpu_ref *ref) +{ + struct obj_cgroup *objcg = container_of(ref, struct obj_cgroup, refcnt); + unsigned long flags; + + obj_cgroup_release_bytes(objcg); spin_lock_irqsave(&objcg_lock, flags); list_del(&objcg->list); @@ -324,10 +335,10 @@ static struct obj_cgroup *obj_cgroup_alloc(void) return objcg; } -static void memcg_reparent_objcgs(struct mem_cgroup *memcg, - struct mem_cgroup *parent) +static void memcg_reparent_objcgs(struct mem_cgroup *memcg) { struct obj_cgroup *objcg, *iter; + struct mem_cgroup *parent = parent_mem_cgroup(memcg); objcg = rcu_replace_pointer(memcg->objcg, NULL, true); @@ -346,6 +357,7 @@ static void memcg_reparent_objcgs(struct mem_cgroup *memcg, percpu_ref_kill(&objcg->refcnt); } +#ifdef CONFIG_MEMCG_KMEM /* * A lot of the calls to the cache allocation functions are expected to be * inlined by the compiler. Since the calls to memcg_slab_pre_alloc_hook() are @@ -3651,21 +3663,12 @@ static u64 mem_cgroup_read_u64(struct cgroup_subsys_state *css, #ifdef CONFIG_MEMCG_KMEM static int memcg_online_kmem(struct mem_cgroup *memcg) { - struct obj_cgroup *objcg; - if (cgroup_memory_nokmem) return 0; if (unlikely(mem_cgroup_is_root(memcg))) return 0; - objcg = obj_cgroup_alloc(); - if (!objcg) - return -ENOMEM; - - objcg->memcg = memcg; - rcu_assign_pointer(memcg->objcg, objcg); - static_branch_enable(&memcg_kmem_enabled_key); memcg->kmemcg_id = memcg->id.id; @@ -3675,17 +3678,13 @@ static int memcg_online_kmem(struct mem_cgroup *memcg) static void memcg_offline_kmem(struct mem_cgroup *memcg) { - struct mem_cgroup *parent; - if (cgroup_memory_nokmem) return; if (unlikely(mem_cgroup_is_root(memcg))) return; - parent = parent_mem_cgroup(memcg); - memcg_reparent_objcgs(memcg, parent); - memcg_reparent_list_lrus(memcg, parent); + memcg_reparent_list_lrus(memcg, parent_mem_cgroup(memcg)); } #else static int memcg_online_kmem(struct mem_cgroup *memcg) @@ -5190,8 +5189,8 @@ static struct mem_cgroup *mem_cgroup_alloc(void) memcg->socket_pressure = jiffies; #ifdef CONFIG_MEMCG_KMEM memcg->kmemcg_id = -1; - INIT_LIST_HEAD(&memcg->objcg_list); #endif + INIT_LIST_HEAD(&memcg->objcg_list); #ifdef CONFIG_CGROUP_WRITEBACK INIT_LIST_HEAD(&memcg->cgwb_list); for (i = 0; i < MEMCG_CGWB_FRN_CNT; i++) @@ -5256,6 +5255,7 @@ mem_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) static int mem_cgroup_css_online(struct cgroup_subsys_state *css) { struct mem_cgroup *memcg = mem_cgroup_from_css(css); + struct obj_cgroup *objcg; if (memcg_online_kmem(memcg)) goto remove_id; @@ -5268,6 +5268,13 @@ static int mem_cgroup_css_online(struct cgroup_subsys_state *css) if (alloc_shrinker_info(memcg)) goto offline_kmem; + objcg = obj_cgroup_alloc(); + if (!objcg) + goto free_shrinker; + + objcg->memcg = memcg; + rcu_assign_pointer(memcg->objcg, objcg); + /* Online state pins memcg ID, memcg ID pins CSS */ refcount_set(&memcg->id.ref, 1); css_get(css); @@ -5276,6 +5283,8 @@ static int mem_cgroup_css_online(struct cgroup_subsys_state *css) queue_delayed_work(system_unbound_wq, &stats_flush_dwork, 2UL*HZ); return 0; +free_shrinker: + free_shrinker_info(memcg); offline_kmem: memcg_offline_kmem(memcg); remove_id: @@ -5303,6 +5312,7 @@ static void mem_cgroup_css_offline(struct cgroup_subsys_state *css) page_counter_set_min(&memcg->memory, 0); page_counter_set_low(&memcg->memory, 0); + memcg_reparent_objcgs(memcg); memcg_offline_kmem(memcg); reparent_shrinker_deferred(memcg); wb_memcg_offline(memcg); -- cgit v1.2.3 From a6561fb3a7488d95afca5eecd4adbf008fa5dc9c Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Tue, 21 Jun 2022 20:56:51 +0800 Subject: mm: memcontrol: make lruvec lock safe when LRU pages are reparented MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The diagram below shows how to make the folio lruvec lock safe when LRU pages are reparented. folio_lruvec_lock(folio) rcu_read_lock(); retry: lruvec = folio_lruvec(folio); // The folio is reparented at this time. spin_lock(&lruvec->lru_lock); if (unlikely(lruvec_memcg(lruvec) != folio_memcg(folio))) // Acquired the wrong lruvec lock and need to retry. // Because this folio is on the parent memcg lruvec list. spin_unlock(&lruvec->lru_lock); goto retry; // If we reach here, it means that folio_memcg(folio) is stable. memcg_reparent_objcgs(memcg) // lruvec belongs to memcg and lruvec_parent belongs to parent memcg. spin_lock(&lruvec->lru_lock); spin_lock(&lruvec_parent->lru_lock); // Move all the pages from the lruvec list to the parent lruvec list. spin_unlock(&lruvec_parent->lru_lock); spin_unlock(&lruvec->lru_lock); After we acquire the lruvec lock, we need to check whether the folio is reparented. If so, we need to reacquire the new lruvec lock. On the routine of the LRU pages reparenting, we will also acquire the lruvec lock (will be implemented in the later patch). So folio_memcg() cannot be changed when we hold the lruvec lock. Since lruvec_memcg(lruvec) is always equal to folio_memcg(folio) after we hold the lruvec lock, lruvec_memcg_debug() check is pointless. So remove it. This is a preparation for reparenting the LRU pages. Link: https://lkml.kernel.org/r/20220621125658.64935-5-songmuchun@bytedance.com Signed-off-by: Muchun Song Cc: Johannes Weiner Cc: Michal Hocko Cc: Michal Koutný Cc: Roman Gushchin Cc: Shakeel Butt Cc: Waiman Long Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 18 +++------------- mm/compaction.c | 27 +++++++++++++++++++---- mm/memcontrol.c | 53 ++++++++++++++++++++++++++-------------------- mm/swap.c | 5 +++++ 4 files changed, 61 insertions(+), 42 deletions(-) (limited to 'include/linux') diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 111eda6ff1ce..ff3106eca6f3 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -758,7 +758,9 @@ out: * folio_lruvec - return lruvec for isolating/putting an LRU folio * @folio: Pointer to the folio. * - * This function relies on folio->mem_cgroup being stable. + * The lruvec can be changed to its parent lruvec when the page reparented. + * The caller need to recheck if it cares about this changes (just like + * folio_lruvec_lock() does). */ static inline struct lruvec *folio_lruvec(struct folio *folio) { @@ -777,15 +779,6 @@ struct lruvec *folio_lruvec_lock_irq(struct folio *folio); struct lruvec *folio_lruvec_lock_irqsave(struct folio *folio, unsigned long *flags); -#ifdef CONFIG_DEBUG_VM -void lruvec_memcg_debug(struct lruvec *lruvec, struct folio *folio); -#else -static inline -void lruvec_memcg_debug(struct lruvec *lruvec, struct folio *folio) -{ -} -#endif - static inline struct mem_cgroup *mem_cgroup_from_css(struct cgroup_subsys_state *css){ return css ? container_of(css, struct mem_cgroup, css) : NULL; @@ -1260,11 +1253,6 @@ static inline struct lruvec *folio_lruvec(struct folio *folio) return &pgdat->__lruvec; } -static inline -void lruvec_memcg_debug(struct lruvec *lruvec, struct folio *folio) -{ -} - static inline struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg) { return NULL; diff --git a/mm/compaction.c b/mm/compaction.c index 46351a14eed2..fe49ac9aedd8 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -508,6 +508,25 @@ static bool compact_lock_irqsave(spinlock_t *lock, unsigned long *flags, return true; } +static struct lruvec * +compact_folio_lruvec_lock_irqsave(struct folio *folio, unsigned long *flags, + struct compact_control *cc) +{ + struct lruvec *lruvec; + + rcu_read_lock(); +retry: + lruvec = folio_lruvec(folio); + compact_lock_irqsave(&lruvec->lru_lock, flags, cc); + if (unlikely(lruvec_memcg(lruvec) != folio_memcg(folio))) { + spin_unlock_irqrestore(&lruvec->lru_lock, *flags); + goto retry; + } + rcu_read_unlock(); + + return lruvec; +} + /* * Compaction requires the taking of some coarse locks that are potentially * very heavily contended. The lock should be periodically unlocked to avoid @@ -834,6 +853,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn, /* Time to isolate some pages for migration */ for (; low_pfn < end_pfn; low_pfn++) { + struct folio *folio; if (skip_on_failure && low_pfn >= next_skip_pfn) { /* @@ -1055,18 +1075,17 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn, if (!TestClearPageLRU(page)) goto isolate_fail_put; - lruvec = folio_lruvec(page_folio(page)); + folio = page_folio(page); + lruvec = folio_lruvec(folio); /* If we already hold the lock, we can skip some rechecking */ if (lruvec != locked) { if (locked) lruvec_unlock_irqrestore(locked, flags); - compact_lock_irqsave(&lruvec->lru_lock, &flags, cc); + lruvec = compact_folio_lruvec_lock_irqsave(folio, &flags, cc); locked = lruvec; - lruvec_memcg_debug(lruvec, page_folio(page)); - /* Try get exclusive access under lock */ if (!skip_updated) { skip_updated = true; diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 3c489651d312..6f171480b2f2 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -1195,23 +1195,6 @@ int mem_cgroup_scan_tasks(struct mem_cgroup *memcg, return ret; } -#ifdef CONFIG_DEBUG_VM -void lruvec_memcg_debug(struct lruvec *lruvec, struct folio *folio) -{ - struct mem_cgroup *memcg; - - if (mem_cgroup_disabled()) - return; - - memcg = folio_memcg(folio); - - if (!memcg) - VM_BUG_ON_FOLIO(lruvec_memcg(lruvec) != root_mem_cgroup, folio); - else - VM_BUG_ON_FOLIO(lruvec_memcg(lruvec) != memcg, folio); -} -#endif - /** * folio_lruvec_lock - Lock the lruvec for a folio. * @folio: Pointer to the folio. @@ -1226,10 +1209,18 @@ void lruvec_memcg_debug(struct lruvec *lruvec, struct folio *folio) */ struct lruvec *folio_lruvec_lock(struct folio *folio) { - struct lruvec *lruvec = folio_lruvec(folio); + struct lruvec *lruvec; + rcu_read_lock(); +retry: + lruvec = folio_lruvec(folio); spin_lock(&lruvec->lru_lock); - lruvec_memcg_debug(lruvec, folio); + + if (unlikely(lruvec_memcg(lruvec) != folio_memcg(folio))) { + spin_unlock(&lruvec->lru_lock); + goto retry; + } + rcu_read_unlock(); return lruvec; } @@ -1249,10 +1240,18 @@ struct lruvec *folio_lruvec_lock(struct folio *folio) */ struct lruvec *folio_lruvec_lock_irq(struct folio *folio) { - struct lruvec *lruvec = folio_lruvec(folio); + struct lruvec *lruvec; + rcu_read_lock(); +retry: + lruvec = folio_lruvec(folio); spin_lock_irq(&lruvec->lru_lock); - lruvec_memcg_debug(lruvec, folio); + + if (unlikely(lruvec_memcg(lruvec) != folio_memcg(folio))) { + spin_unlock_irq(&lruvec->lru_lock); + goto retry; + } + rcu_read_unlock(); return lruvec; } @@ -1274,10 +1273,18 @@ struct lruvec *folio_lruvec_lock_irq(struct folio *folio) struct lruvec *folio_lruvec_lock_irqsave(struct folio *folio, unsigned long *flags) { - struct lruvec *lruvec = folio_lruvec(folio); + struct lruvec *lruvec; + rcu_read_lock(); +retry: + lruvec = folio_lruvec(folio); spin_lock_irqsave(&lruvec->lru_lock, *flags); - lruvec_memcg_debug(lruvec, folio); + + if (unlikely(lruvec_memcg(lruvec) != folio_memcg(folio))) { + spin_unlock_irqrestore(&lruvec->lru_lock, *flags); + goto retry; + } + rcu_read_unlock(); return lruvec; } diff --git a/mm/swap.c b/mm/swap.c index 127ef4db394f..987dcbd93ffa 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -337,6 +337,11 @@ void lru_note_cost(struct lruvec *lruvec, bool file, unsigned int nr_pages) void lru_note_cost_folio(struct folio *folio) { + WARN_ON_ONCE(!rcu_read_lock_held()); + /* + * The rcu read lock is held by the caller, so we do not need to + * care about the lruvec returned by folio_lruvec() being released. + */ lru_note_cost(folio_lruvec(folio), folio_is_file_lru(folio), folio_nr_pages(folio)); } -- cgit v1.2.3 From 3b8dd99412c81b7543bbb5599bf1c1709cd19f1f Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Tue, 21 Jun 2022 20:56:53 +0800 Subject: mm: thp: make split queue lock safe when LRU pages are reparented MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Similar to the lruvec lock, we use the same approach to make the split queue lock safe when LRU pages are reparented. Link: https://lkml.kernel.org/r/20220621125658.64935-7-songmuchun@bytedance.com Signed-off-by: Muchun Song Acked-by: Roman Gushchin Cc: Johannes Weiner Cc: Michal Hocko Cc: Michal Koutný Cc: Shakeel Butt Cc: Waiman Long Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 10 ++++ mm/huge_memory.c | 116 +++++++++++++++++++++++++++++++++++---------- 2 files changed, 100 insertions(+), 26 deletions(-) (limited to 'include/linux') diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index ff3106eca6f3..026b62b206b1 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -1691,6 +1691,11 @@ int alloc_shrinker_info(struct mem_cgroup *memcg); void free_shrinker_info(struct mem_cgroup *memcg); void set_shrinker_bit(struct mem_cgroup *memcg, int nid, int shrinker_id); void reparent_shrinker_deferred(struct mem_cgroup *memcg); + +static inline int shrinker_id(struct shrinker *shrinker) +{ + return shrinker->id; +} #else #define mem_cgroup_sockets_enabled 0 static inline void mem_cgroup_sk_alloc(struct sock *sk) { }; @@ -1704,6 +1709,11 @@ static inline void set_shrinker_bit(struct mem_cgroup *memcg, int nid, int shrinker_id) { } + +static inline int shrinker_id(struct shrinker *shrinker) +{ + return -1; +} #endif #ifdef CONFIG_MEMCG_KMEM diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 05656438e95c..b4cd5ea42aca 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -561,25 +561,90 @@ pmd_t maybe_pmd_mkwrite(pmd_t pmd, struct vm_area_struct *vma) } #ifdef CONFIG_MEMCG -static inline struct deferred_split *get_deferred_split_queue(struct page *page) +static inline struct mem_cgroup *folio_split_queue_memcg(struct folio *folio, + struct deferred_split *queue) { - struct mem_cgroup *memcg = page_memcg(compound_head(page)); - struct pglist_data *pgdat = NODE_DATA(page_to_nid(page)); + if (mem_cgroup_disabled()) + return NULL; + if (&NODE_DATA(folio_nid(folio))->deferred_split_queue == queue) + return NULL; + return container_of(queue, struct mem_cgroup, deferred_split_queue); +} - if (memcg) - return &memcg->deferred_split_queue; - else - return &pgdat->deferred_split_queue; +static inline struct deferred_split *folio_memcg_split_queue(struct folio *folio) +{ + struct mem_cgroup *memcg = folio_memcg(folio); + + return memcg ? &memcg->deferred_split_queue : NULL; } #else -static inline struct deferred_split *get_deferred_split_queue(struct page *page) +static inline struct mem_cgroup *folio_split_queue_memcg(struct folio *folio, + struct deferred_split *queue) { - struct pglist_data *pgdat = NODE_DATA(page_to_nid(page)); + return NULL; +} - return &pgdat->deferred_split_queue; +static inline struct deferred_split *folio_memcg_split_queue(struct folio *folio) +{ + return NULL; } #endif +static struct deferred_split *folio_split_queue(struct folio *folio) +{ + struct deferred_split *queue = folio_memcg_split_queue(folio); + + return queue ? : &NODE_DATA(folio_nid(folio))->deferred_split_queue; +} + +static struct deferred_split *folio_split_queue_lock(struct folio *folio) +{ + struct deferred_split *queue; + + rcu_read_lock(); +retry: + queue = folio_split_queue(folio); + spin_lock(&queue->split_queue_lock); + + if (unlikely(folio_split_queue_memcg(folio, queue) != folio_memcg(folio))) { + spin_unlock(&queue->split_queue_lock); + goto retry; + } + rcu_read_unlock(); + + return queue; +} + +static struct deferred_split * +folio_split_queue_lock_irqsave(struct folio *folio, unsigned long *flags) +{ + struct deferred_split *queue; + + rcu_read_lock(); +retry: + queue = folio_split_queue(folio); + spin_lock_irqsave(&queue->split_queue_lock, *flags); + + if (unlikely(folio_split_queue_memcg(folio, queue) != folio_memcg(folio))) { + spin_unlock_irqrestore(&queue->split_queue_lock, *flags); + goto retry; + } + rcu_read_unlock(); + + return queue; +} + +static inline void split_queue_unlock(struct deferred_split *queue) +{ + spin_unlock(&queue->split_queue_lock); +} + +static inline void split_queue_unlock_irqrestore(struct deferred_split *queue, + unsigned long flags) +{ + spin_unlock_irqrestore(&queue->split_queue_lock, flags); +} + void prep_transhuge_page(struct page *page) { /* @@ -2604,7 +2669,7 @@ int split_huge_page_to_list(struct page *page, struct list_head *list) { struct folio *folio = page_folio(page); struct page *head = &folio->page; - struct deferred_split *ds_queue = get_deferred_split_queue(head); + struct deferred_split *ds_queue; XA_STATE(xas, &head->mapping->i_pages, head->index); struct anon_vma *anon_vma = NULL; struct address_space *mapping = NULL; @@ -2696,13 +2761,13 @@ int split_huge_page_to_list(struct page *page, struct list_head *list) } /* Prevent deferred_split_scan() touching ->_refcount */ - spin_lock(&ds_queue->split_queue_lock); + ds_queue = folio_split_queue_lock(folio); if (page_ref_freeze(head, 1 + extra_pins)) { if (!list_empty(page_deferred_list(head))) { ds_queue->split_queue_len--; list_del(page_deferred_list(head)); } - spin_unlock(&ds_queue->split_queue_lock); + split_queue_unlock(ds_queue); if (mapping) { int nr = thp_nr_pages(head); @@ -2720,7 +2785,7 @@ int split_huge_page_to_list(struct page *page, struct list_head *list) __split_huge_page(page, list, end); ret = 0; } else { - spin_unlock(&ds_queue->split_queue_lock); + split_queue_unlock(ds_queue); fail: if (mapping) xas_unlock(&xas); @@ -2744,25 +2809,23 @@ out: void free_transhuge_page(struct page *page) { - struct deferred_split *ds_queue = get_deferred_split_queue(page); + struct deferred_split *ds_queue; unsigned long flags; - spin_lock_irqsave(&ds_queue->split_queue_lock, flags); + ds_queue = folio_split_queue_lock_irqsave(page_folio(page), &flags); if (!list_empty(page_deferred_list(page))) { ds_queue->split_queue_len--; list_del(page_deferred_list(page)); } - spin_unlock_irqrestore(&ds_queue->split_queue_lock, flags); + split_queue_unlock_irqrestore(ds_queue, flags); free_compound_page(page); } void deferred_split_huge_page(struct page *page) { - struct deferred_split *ds_queue = get_deferred_split_queue(page); -#ifdef CONFIG_MEMCG - struct mem_cgroup *memcg = page_memcg(compound_head(page)); -#endif + struct deferred_split *ds_queue; unsigned long flags; + struct folio *folio = page_folio(page); VM_BUG_ON_PAGE(!PageTransHuge(page), page); @@ -2779,18 +2842,19 @@ void deferred_split_huge_page(struct page *page) if (PageSwapCache(page)) return; - spin_lock_irqsave(&ds_queue->split_queue_lock, flags); + ds_queue = folio_split_queue_lock_irqsave(folio, &flags); if (list_empty(page_deferred_list(page))) { + struct mem_cgroup *memcg; + + memcg = folio_split_queue_memcg(folio, ds_queue); count_vm_event(THP_DEFERRED_SPLIT_PAGE); list_add_tail(page_deferred_list(page), &ds_queue->split_queue); ds_queue->split_queue_len++; -#ifdef CONFIG_MEMCG if (memcg) set_shrinker_bit(memcg, page_to_nid(page), - deferred_split_shrinker.id); -#endif + shrinker_id(&deferred_split_shrinker)); } - spin_unlock_irqrestore(&ds_queue->split_queue_lock, flags); + split_queue_unlock_irqrestore(ds_queue, flags); } static unsigned long deferred_split_count(struct shrinker *shrink, -- cgit v1.2.3 From 577affb1cd9a789f0b6a6707ca27dea6522c237b Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Tue, 21 Jun 2022 20:56:54 +0800 Subject: mm: memcontrol: make all the callers of {folio,page}_memcg() safe MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When we use objcg APIs to charge the LRU pages, the page will not hold a reference to the memcg associated with the page. So the caller of the {folio,page}_memcg() should hold an rcu read lock or obtain a reference to the memcg associated with the page to protect memcg from being released. So introduce get_mem_cgroup_from_{page,folio}() to obtain a reference to the memory cgroup associated with the page. In this patch, make all the callers hold an rcu read lock or obtain a reference to the memcg to protect memcg from being released when the LRU pages reparented. We do not need to adjust the callers of {folio,page}_memcg() during the whole process of mem_cgroup_move_task(). Because the cgroup migration and memory cgroup offlining are serialized by @cgroup_mutex. In this routine, the LRU pages cannot be reparented to its parent memory cgroup. So {folio,page}_memcg() is stable and cannot be released. This is a preparation for reparenting the LRU pages. Link: https://lkml.kernel.org/r/20220621125658.64935-8-songmuchun@bytedance.com Signed-off-by: Muchun Song Acked-by: Roman Gushchin Cc: Johannes Weiner Cc: Michal Hocko Cc: Michal Koutný Cc: Shakeel Butt Cc: Waiman Long Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- fs/buffer.c | 4 +-- fs/fs-writeback.c | 23 +++++++------- include/linux/memcontrol.h | 66 +++++++++++++++++++++++++++++++++----- include/trace/events/writeback.h | 5 +++ mm/memcontrol.c | 68 +++++++++++++++++++++++++++++----------- mm/migrate.c | 4 +++ mm/page_io.c | 5 +-- 7 files changed, 135 insertions(+), 40 deletions(-) (limited to 'include/linux') diff --git a/fs/buffer.c b/fs/buffer.c index 898c7f301b1b..04ec53f327e4 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -819,8 +819,7 @@ struct buffer_head *alloc_page_buffers(struct page *page, unsigned long size, if (retry) gfp |= __GFP_NOFAIL; - /* The page lock pins the memcg */ - memcg = page_memcg(page); + memcg = get_mem_cgroup_from_page(page); old_memcg = set_active_memcg(memcg); head = NULL; @@ -840,6 +839,7 @@ struct buffer_head *alloc_page_buffers(struct page *page, unsigned long size, set_bh_page(bh, page, offset); } out: + mem_cgroup_put(memcg); set_active_memcg(old_memcg); return head; /* diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 05221366a16d..1cbac56c810b 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -244,15 +244,13 @@ void __inode_attach_wb(struct inode *inode, struct page *page) if (inode_cgwb_enabled(inode)) { struct cgroup_subsys_state *memcg_css; - if (page) { - memcg_css = mem_cgroup_css_from_page(page); - wb = wb_get_create(bdi, memcg_css, GFP_ATOMIC); - } else { - /* must pin memcg_css, see wb_get_create() */ + /* must pin memcg_css, see wb_get_create() */ + if (page) + memcg_css = get_mem_cgroup_css_from_page(page); + else memcg_css = task_get_css(current, memory_cgrp_id); - wb = wb_get_create(bdi, memcg_css, GFP_ATOMIC); - css_put(memcg_css); - } + wb = wb_get_create(bdi, memcg_css, GFP_ATOMIC); + css_put(memcg_css); } if (!wb) @@ -869,16 +867,16 @@ void wbc_account_cgroup_owner(struct writeback_control *wbc, struct page *page, if (!wbc->wb || wbc->no_cgroup_owner) return; - css = mem_cgroup_css_from_page(page); + css = get_mem_cgroup_css_from_page(page); /* dead cgroups shouldn't contribute to inode ownership arbitration */ if (!(css->flags & CSS_ONLINE)) - return; + goto out; id = css->id; if (id == wbc->wb_id) { wbc->wb_bytes += bytes; - return; + goto out; } if (id == wbc->wb_lcand_id) @@ -891,6 +889,9 @@ void wbc_account_cgroup_owner(struct writeback_control *wbc, struct page *page, wbc->wb_tcand_bytes += bytes; else wbc->wb_tcand_bytes -= min(bytes, wbc->wb_tcand_bytes); + +out: + css_put(css); } EXPORT_SYMBOL_GPL(wbc_account_cgroup_owner); diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 026b62b206b1..a8bd4bb39502 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -379,7 +379,7 @@ static inline bool folio_memcg_kmem(struct folio *folio); * a valid memcg, but can be atomically swapped to the parent memcg. * * The caller must ensure that the returned memcg won't be released: - * e.g. acquire the rcu_read_lock or css_set_lock. + * e.g. acquire the rcu_read_lock or objcg_lock or cgroup_mutex. */ static inline struct mem_cgroup *obj_cgroup_memcg(struct obj_cgroup *objcg) { @@ -445,8 +445,8 @@ static inline struct obj_cgroup *__folio_objcg(struct folio *folio) * - lock_page_memcg() * - exclusive reference * - * For a kmem folio a caller should hold an rcu read lock to protect memcg - * associated with a kmem folio from being released. + * Note: The caller should hold an rcu read lock to protect memcg associated + * with a folio from being released. */ static inline struct mem_cgroup *folio_memcg(struct folio *folio) { @@ -455,12 +455,48 @@ static inline struct mem_cgroup *folio_memcg(struct folio *folio) return __folio_memcg(folio); } +/* + * page_memcg - Get the memory cgroup associated with a page. + * @page: Pointer to the page. + * + * See the cooments in folio_memcg(). + */ static inline struct mem_cgroup *page_memcg(struct page *page) { return folio_memcg(page_folio(page)); } -/** +/* + * get_mem_cgroup_from_folio - Obtain a reference on the memory cgroup + * associated with a folio. + * @folio: Pointer to the folio. + * + * Returns a pointer to the memory cgroup (and obtain a reference on it) + * associated with the folio, or NULL. This function assumes that the + * folio is known to have a proper memory cgroup pointer. It's not safe + * to call this function against some type of pages, e.g. slab pages or + * ex-slab pages. + */ +static inline struct mem_cgroup *get_mem_cgroup_from_folio(struct folio *folio) +{ + struct mem_cgroup *memcg; + + rcu_read_lock(); +retry: + memcg = folio_memcg(folio); + if (unlikely(memcg && !css_tryget(&memcg->css))) + goto retry; + rcu_read_unlock(); + + return memcg; +} + +static inline struct mem_cgroup *get_mem_cgroup_from_page(struct page *page) +{ + return get_mem_cgroup_from_folio(page_folio(page)); +} + +/* * folio_memcg_rcu - Locklessly get the memory cgroup associated with a folio. * @folio: Pointer to the folio. * @@ -888,7 +924,7 @@ static inline bool mm_match_cgroup(struct mm_struct *mm, return match; } -struct cgroup_subsys_state *mem_cgroup_css_from_page(struct page *page); +struct cgroup_subsys_state *get_mem_cgroup_css_from_page(struct page *page); ino_t page_cgroup_ino(struct page *page); static inline bool mem_cgroup_online(struct mem_cgroup *memcg) @@ -1058,19 +1094,25 @@ static inline void count_memcg_events(struct mem_cgroup *memcg, static inline void count_memcg_page_event(struct page *page, enum vm_event_item idx) { - struct mem_cgroup *memcg = page_memcg(page); + struct mem_cgroup *memcg; + rcu_read_lock(); + memcg = page_memcg(page); if (memcg) count_memcg_events(memcg, idx, 1); + rcu_read_unlock(); } static inline void count_memcg_folio_events(struct folio *folio, enum vm_event_item idx, unsigned long nr) { - struct mem_cgroup *memcg = folio_memcg(folio); + struct mem_cgroup *memcg; + rcu_read_lock(); + memcg = folio_memcg(folio); if (memcg) count_memcg_events(memcg, idx, nr); + rcu_read_unlock(); } static inline void count_memcg_event_mm(struct mm_struct *mm, @@ -1149,6 +1191,16 @@ static inline struct mem_cgroup *page_memcg(struct page *page) return NULL; } +static inline struct mem_cgroup *get_mem_cgroup_from_folio(struct folio *folio) +{ + return NULL; +} + +static inline struct mem_cgroup *get_mem_cgroup_from_page(struct page *page) +{ + return NULL; +} + static inline struct mem_cgroup *folio_memcg_rcu(struct folio *folio) { WARN_ON_ONCE(!rcu_read_lock_held()); diff --git a/include/trace/events/writeback.h b/include/trace/events/writeback.h index 86b2a82da546..cdb822339f13 100644 --- a/include/trace/events/writeback.h +++ b/include/trace/events/writeback.h @@ -258,6 +258,11 @@ TRACE_EVENT(track_foreign_dirty, __entry->ino = inode ? inode->i_ino : 0; __entry->memcg_id = wb->memcg_css->id; __entry->cgroup_ino = __trace_wb_assign_cgroup(wb); + /* + * TP_fast_assign() is under preemption disabled which can + * serve as an RCU read-side critical section so that the + * memcg returned by folio_memcg() cannot be freed. + */ __entry->page_cgroup_ino = cgroup_ino(folio_memcg(folio)->css.cgroup); ), diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 6f171480b2f2..346a954e190e 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -369,7 +369,7 @@ EXPORT_SYMBOL(memcg_kmem_enabled_key); #endif /** - * mem_cgroup_css_from_page - css of the memcg associated with a page + * get_mem_cgroup_css_from_page - get css of the memcg associated with a page * @page: page of interest * * If memcg is bound to the default hierarchy, css of the memcg associated @@ -379,13 +379,15 @@ EXPORT_SYMBOL(memcg_kmem_enabled_key); * If memcg is bound to a traditional hierarchy, the css of root_mem_cgroup * is returned. */ -struct cgroup_subsys_state *mem_cgroup_css_from_page(struct page *page) +struct cgroup_subsys_state *get_mem_cgroup_css_from_page(struct page *page) { struct mem_cgroup *memcg; - memcg = page_memcg(page); + if (!cgroup_subsys_on_dfl(memory_cgrp_subsys)) + return &root_mem_cgroup->css; - if (!memcg || !cgroup_subsys_on_dfl(memory_cgrp_subsys)) + memcg = get_mem_cgroup_from_page(page); + if (!memcg) memcg = root_mem_cgroup; return &memcg->css; @@ -768,13 +770,13 @@ void __mod_lruvec_state(struct lruvec *lruvec, enum node_stat_item idx, void __mod_lruvec_page_state(struct page *page, enum node_stat_item idx, int val) { - struct page *head = compound_head(page); /* rmap on tail pages */ + struct folio *folio = page_folio(page); /* rmap on tail pages */ struct mem_cgroup *memcg; pg_data_t *pgdat = page_pgdat(page); struct lruvec *lruvec; rcu_read_lock(); - memcg = page_memcg(head); + memcg = folio_memcg(folio); /* Untracked pages have no memcg, no lruvec. Update only the node */ if (!memcg) { rcu_read_unlock(); @@ -2056,7 +2058,9 @@ void folio_memcg_lock(struct folio *folio) * The RCU lock is held throughout the transaction. The fast * path can get away without acquiring the memcg->move_lock * because page moving starts with an RCU grace period. - */ + * + * The RCU lock also protects the memcg from being freed. + */ rcu_read_lock(); if (mem_cgroup_disabled()) @@ -3353,7 +3357,7 @@ void obj_cgroup_uncharge(struct obj_cgroup *objcg, size_t size) void split_page_memcg(struct page *head, unsigned int nr) { struct folio *folio = page_folio(head); - struct mem_cgroup *memcg = folio_memcg(folio); + struct mem_cgroup *memcg = get_mem_cgroup_from_folio(folio); int i; if (mem_cgroup_disabled() || !memcg) @@ -3366,6 +3370,8 @@ void split_page_memcg(struct page *head, unsigned int nr) obj_cgroup_get_many(__folio_objcg(folio), nr - 1); else css_get_many(&memcg->css, nr - 1); + + css_put(&memcg->css); } #ifdef CONFIG_MEMCG_SWAP @@ -4558,7 +4564,7 @@ void mem_cgroup_wb_stats(struct bdi_writeback *wb, unsigned long *pfilepages, void mem_cgroup_track_foreign_dirty_slowpath(struct folio *folio, struct bdi_writeback *wb) { - struct mem_cgroup *memcg = folio_memcg(folio); + struct mem_cgroup *memcg = get_mem_cgroup_from_folio(folio); struct memcg_cgwb_frn *frn; u64 now = get_jiffies_64(); u64 oldest_at = now; @@ -4605,6 +4611,7 @@ void mem_cgroup_track_foreign_dirty_slowpath(struct folio *folio, frn->memcg_id = wb->memcg_css->id; frn->at = now; } + css_put(&memcg->css); } /* issue foreign writeback flushes for recorded foreign dirtying events */ @@ -6167,6 +6174,14 @@ retry: atomic_dec(&mc.from->moving_account); } +/* + * The cgroup migration and memory cgroup offlining are serialized by + * @cgroup_mutex. If we reach here, it means that the LRU pages cannot + * be reparented to its parent memory cgroup. So during the whole process + * of mem_cgroup_move_task(), page_memcg(page) is stable. So we do not + * need to worry about the memcg (returned from page_memcg()) being + * released even if we do not hold an rcu read lock. + */ static void mem_cgroup_move_task(void) { if (mc.to) { @@ -7025,7 +7040,7 @@ void mem_cgroup_migrate(struct folio *old, struct folio *new) if (folio_memcg(new)) return; - memcg = folio_memcg(old); + memcg = get_mem_cgroup_from_folio(old); VM_WARN_ON_ONCE_FOLIO(!memcg, old); if (!memcg) return; @@ -7044,6 +7059,8 @@ void mem_cgroup_migrate(struct folio *old, struct folio *new) mem_cgroup_charge_statistics(memcg, nr_pages); memcg_check_events(memcg, folio_nid(new)); local_irq_restore(flags); + + css_put(&memcg->css); } DEFINE_STATIC_KEY_FALSE(memcg_sockets_enabled_key); @@ -7228,6 +7245,10 @@ void mem_cgroup_swapout(struct folio *folio, swp_entry_t entry) if (cgroup_subsys_on_dfl(memory_cgrp_subsys)) return; + /* + * Interrupts should be disabled by the caller (see the comments below), + * which can serve as RCU read-side critical sections. + */ memcg = folio_memcg(folio); VM_WARN_ON_ONCE_FOLIO(!memcg, folio); @@ -7289,19 +7310,21 @@ int __mem_cgroup_try_charge_swap(struct folio *folio, swp_entry_t entry) struct page_counter *counter; struct mem_cgroup *memcg; unsigned short oldid; + int ret = 0; if (!cgroup_subsys_on_dfl(memory_cgrp_subsys)) return 0; + rcu_read_lock(); memcg = folio_memcg(folio); VM_WARN_ON_ONCE_FOLIO(!memcg, folio); if (!memcg) - return 0; + goto out; if (!entry.val) { memcg_memory_event(memcg, MEMCG_SWAP_FAIL); - return 0; + goto out; } memcg = mem_cgroup_id_get_online(memcg); @@ -7311,7 +7334,8 @@ int __mem_cgroup_try_charge_swap(struct folio *folio, swp_entry_t entry) memcg_memory_event(memcg, MEMCG_SWAP_MAX); memcg_memory_event(memcg, MEMCG_SWAP_FAIL); mem_cgroup_id_put(memcg); - return -ENOMEM; + ret = -ENOMEM; + goto out; } /* Get references for the tail pages, too */ @@ -7320,8 +7344,10 @@ int __mem_cgroup_try_charge_swap(struct folio *folio, swp_entry_t entry) oldid = swap_cgroup_record(entry, mem_cgroup_id(memcg), nr_pages); VM_BUG_ON_FOLIO(oldid, folio); mod_memcg_state(memcg, MEMCG_SWAP, nr_pages); +out: + rcu_read_unlock(); - return 0; + return ret; } /** @@ -7366,6 +7392,7 @@ long mem_cgroup_get_nr_swap_pages(struct mem_cgroup *memcg) bool mem_cgroup_swap_full(struct page *page) { struct mem_cgroup *memcg; + bool ret = false; VM_BUG_ON_PAGE(!PageLocked(page), page); @@ -7374,19 +7401,24 @@ bool mem_cgroup_swap_full(struct page *page) if (cgroup_memory_noswap || !cgroup_subsys_on_dfl(memory_cgrp_subsys)) return false; + rcu_read_lock(); memcg = page_memcg(page); if (!memcg) - return false; + goto out; for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg)) { unsigned long usage = page_counter_read(&memcg->swap); if (usage * 2 >= READ_ONCE(memcg->swap.high) || - usage * 2 >= READ_ONCE(memcg->swap.max)) - return true; + usage * 2 >= READ_ONCE(memcg->swap.max)) { + ret = true; + goto out; + } } +out: + rcu_read_unlock(); - return false; + return ret; } static int __init setup_swap_account(char *s) diff --git a/mm/migrate.c b/mm/migrate.c index 4cf6b57d3d8e..4a9af5119502 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -451,6 +451,10 @@ int folio_migrate_mapping(struct address_space *mapping, struct lruvec *old_lruvec, *new_lruvec; struct mem_cgroup *memcg; + /* + * Irq is disabled, which can serve as RCU read-side critical + * sections. + */ memcg = folio_memcg(folio); old_lruvec = mem_cgroup_lruvec(memcg, oldzone->zone_pgdat); new_lruvec = mem_cgroup_lruvec(memcg, newzone->zone_pgdat); diff --git a/mm/page_io.c b/mm/page_io.c index 68318134dc92..f75ebbc95ee6 100644 --- a/mm/page_io.c +++ b/mm/page_io.c @@ -222,13 +222,14 @@ static void bio_associate_blkg_from_page(struct bio *bio, struct page *page) struct cgroup_subsys_state *css; struct mem_cgroup *memcg; + rcu_read_lock(); memcg = page_memcg(page); if (!memcg) - return; + goto out; - rcu_read_lock(); css = cgroup_e_css(memcg->css.cgroup, &io_cgrp_subsys); bio_associate_blkg_from_css(bio, css); +out: rcu_read_unlock(); } #else -- cgit v1.2.3 From c4d5d177467ca0038bd4522c7678c046e8b5e499 Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Tue, 21 Jun 2022 20:56:55 +0800 Subject: mm: memcontrol: introduce memcg_reparent_ops MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In the previous patch, we know how to make the lruvec lock safe when LRU pages are reparented. We should do something like following. memcg_reparent_objcgs(memcg) 1) lock // lruvec belongs to memcg and lruvec_parent belongs to parent memcg. spin_lock(&lruvec->lru_lock); spin_lock(&lruvec_parent->lru_lock); 2) relocate from current memcg to its parent // Move all the pages from the lruvec list to the parent lruvec list. 3) unlock spin_unlock(&lruvec_parent->lru_lock); spin_unlock(&lruvec->lru_lock); Apart from the page lruvec lock, the deferred split queue lock (THP only) also needs to do something similar. So we extract the necessary three steps in the memcg_reparent_objcgs(). memcg_reparent_objcgs(memcg) 1) lock memcg_reparent_ops->lock(memcg, parent); 2) relocate memcg_reparent_ops->relocate(memcg, reparent); 3) unlock memcg_reparent_ops->unlock(memcg, reparent); Now there are two different locks (e.g. lruvec lock and deferred split queue lock) need to use this infrastructure. In the next patch, we will use those APIs to make those locks safe when the LRU pages reparented. Link: https://lkml.kernel.org/r/20220621125658.64935-9-songmuchun@bytedance.com Signed-off-by: Muchun Song Cc: Johannes Weiner Cc: Michal Hocko Cc: Michal Koutný Cc: Roman Gushchin Cc: Shakeel Butt Cc: Waiman Long Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 20 +++++++++++++++ mm/memcontrol.c | 62 ++++++++++++++++++++++++++++++++++++---------- 2 files changed, 69 insertions(+), 13 deletions(-) (limited to 'include/linux') diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index a8bd4bb39502..63dbdef60cbd 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -353,6 +353,26 @@ struct mem_cgroup { struct mem_cgroup_per_node *nodeinfo[]; }; +struct memcg_reparent_ops { + /* + * Note that interrupt is disabled before calling those callbacks, + * so the interrupt should remain disabled when leaving those callbacks. + */ + void (*lock)(struct mem_cgroup *src, struct mem_cgroup *dst); + void (*relocate)(struct mem_cgroup *src, struct mem_cgroup *dst); + void (*unlock)(struct mem_cgroup *src, struct mem_cgroup *dst); +}; + +#define DEFINE_MEMCG_REPARENT_OPS(name) \ + const struct memcg_reparent_ops memcg_##name##_reparent_ops = { \ + .lock = name##_reparent_lock, \ + .relocate = name##_reparent_relocate, \ + .unlock = name##_reparent_unlock, \ + } + +#define DECLARE_MEMCG_REPARENT_OPS(name) \ + extern const struct memcg_reparent_ops memcg_##name##_reparent_ops + /* * size of first charge trial. "32" comes from vmscan.c's magic value. * TODO: maybe necessary to use big numbers in big irons. diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 346a954e190e..6ef3a264054e 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -335,24 +335,60 @@ static struct obj_cgroup *obj_cgroup_alloc(void) return objcg; } -static void memcg_reparent_objcgs(struct mem_cgroup *memcg) +static void objcg_reparent_lock(struct mem_cgroup *src, struct mem_cgroup *dst) +{ + spin_lock(&objcg_lock); +} + +static void objcg_reparent_relocate(struct mem_cgroup *src, struct mem_cgroup *dst) { struct obj_cgroup *objcg, *iter; - struct mem_cgroup *parent = parent_mem_cgroup(memcg); - objcg = rcu_replace_pointer(memcg->objcg, NULL, true); + objcg = rcu_replace_pointer(src->objcg, NULL, true); + /* 1) Ready to reparent active objcg. */ + list_add(&objcg->list, &src->objcg_list); + /* 2) Reparent active objcg and already reparented objcgs to dst. */ + list_for_each_entry(iter, &src->objcg_list, list) + WRITE_ONCE(iter->memcg, dst); + /* 3) Move already reparented objcgs to the dst's list */ + list_splice(&src->objcg_list, &dst->objcg_list); +} + +static void objcg_reparent_unlock(struct mem_cgroup *src, struct mem_cgroup *dst) +{ + spin_unlock(&objcg_lock); +} - spin_lock_irq(&objcg_lock); +static DEFINE_MEMCG_REPARENT_OPS(objcg); - /* 1) Ready to reparent active objcg. */ - list_add(&objcg->list, &memcg->objcg_list); - /* 2) Reparent active objcg and already reparented objcgs to parent. */ - list_for_each_entry(iter, &memcg->objcg_list, list) - WRITE_ONCE(iter->memcg, parent); - /* 3) Move already reparented objcgs to the parent's list */ - list_splice(&memcg->objcg_list, &parent->objcg_list); - - spin_unlock_irq(&objcg_lock); +static const struct memcg_reparent_ops *memcg_reparent_ops[] = { + &memcg_objcg_reparent_ops, +}; + +#define DEFINE_MEMCG_REPARENT_FUNC(phase) \ + static void memcg_reparent_##phase(struct mem_cgroup *src, \ + struct mem_cgroup *dst) \ + { \ + int i; \ + \ + for (i = 0; i < ARRAY_SIZE(memcg_reparent_ops); i++) \ + memcg_reparent_ops[i]->phase(src, dst); \ + } + +DEFINE_MEMCG_REPARENT_FUNC(lock) +DEFINE_MEMCG_REPARENT_FUNC(relocate) +DEFINE_MEMCG_REPARENT_FUNC(unlock) + +static void memcg_reparent_objcgs(struct mem_cgroup *src) +{ + struct mem_cgroup *dst = parent_mem_cgroup(src); + struct obj_cgroup *objcg = rcu_dereference_protected(src->objcg, true); + + local_irq_disable(); + memcg_reparent_lock(src, dst); + memcg_reparent_relocate(src, dst); + memcg_reparent_unlock(src, dst); + local_irq_enable(); percpu_ref_kill(&objcg->refcnt); } -- cgit v1.2.3 From 861a3e5bca5447d00c48c5e15267c83c81b76389 Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Tue, 21 Jun 2022 20:56:56 +0800 Subject: mm: memcontrol: use obj_cgroup APIs to charge the LRU pages MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We will reuse the obj_cgroup APIs to charge the LRU pages. Finally, page->memcg_data will have 2 different meanings. - For the slab pages, page->memcg_data points to an object cgroups vector. - For the kmem pages (exclude the slab pages) and the LRU pages, page->memcg_data points to an object cgroup. In this patch, we reuse obj_cgroup APIs to charge LRU pages. In the end, The page cache cannot prevent long-living objects from pinning the original memory cgroup in the memory. At the same time we also changed the rules of page and objcg or memcg binding stability. The new rules are as follows. For a page any of the following ensures page and objcg binding stability: - the page lock - LRU isolation - lock_page_memcg() - exclusive reference Based on the stable binding of page and objcg, for a page any of the following ensures page and memcg binding stability: - objcg_lock - cgroup_mutex - the lruvec lock - the split queue lock (only THP page) If the caller only want to ensure that the page counters of memcg are updated correctly, ensure that the binding stability of page and objcg is sufficient. Link: https://lkml.kernel.org/r/20220621125658.64935-10-songmuchun@bytedance.com Signed-off-by: Muchun Song Reviewed-by: Michal Koutný Acked-by: Roman Gushchin Cc: Johannes Weiner Cc: Michal Hocko Cc: Shakeel Butt Cc: Waiman Long Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 89 +++++-------- mm/huge_memory.c | 35 +++++ mm/memcontrol.c | 317 ++++++++++++++++++++++++++++++--------------- 3 files changed, 282 insertions(+), 159 deletions(-) (limited to 'include/linux') diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 63dbdef60cbd..744cde2b2368 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -392,8 +392,6 @@ enum page_memcg_data_flags { #define MEMCG_DATA_FLAGS_MASK (__NR_MEMCG_DATA_FLAGS - 1) -static inline bool folio_memcg_kmem(struct folio *folio); - /* * After the initialization objcg->memcg is always pointing at * a valid memcg, but can be atomically swapped to the parent memcg. @@ -407,43 +405,19 @@ static inline struct mem_cgroup *obj_cgroup_memcg(struct obj_cgroup *objcg) } /* - * __folio_memcg - Get the memory cgroup associated with a non-kmem folio - * @folio: Pointer to the folio. - * - * Returns a pointer to the memory cgroup associated with the folio, - * or NULL. This function assumes that the folio is known to have a - * proper memory cgroup pointer. It's not safe to call this function - * against some type of folios, e.g. slab folios or ex-slab folios or - * kmem folios. - */ -static inline struct mem_cgroup *__folio_memcg(struct folio *folio) -{ - unsigned long memcg_data = folio->memcg_data; - - VM_BUG_ON_FOLIO(folio_test_slab(folio), folio); - VM_BUG_ON_FOLIO(memcg_data & MEMCG_DATA_OBJCGS, folio); - VM_BUG_ON_FOLIO(memcg_data & MEMCG_DATA_KMEM, folio); - - return (struct mem_cgroup *)(memcg_data & ~MEMCG_DATA_FLAGS_MASK); -} - -/* - * __folio_objcg - get the object cgroup associated with a kmem folio. + * folio_objcg - get the object cgroup associated with a folio. * @folio: Pointer to the folio. * * Returns a pointer to the object cgroup associated with the folio, * or NULL. This function assumes that the folio is known to have a - * proper object cgroup pointer. It's not safe to call this function - * against some type of folios, e.g. slab folios or ex-slab folios or - * LRU folios. + * proper object cgroup pointer. */ -static inline struct obj_cgroup *__folio_objcg(struct folio *folio) +static inline struct obj_cgroup *folio_objcg(struct folio *folio) { unsigned long memcg_data = folio->memcg_data; VM_BUG_ON_FOLIO(folio_test_slab(folio), folio); VM_BUG_ON_FOLIO(memcg_data & MEMCG_DATA_OBJCGS, folio); - VM_BUG_ON_FOLIO(!(memcg_data & MEMCG_DATA_KMEM), folio); return (struct obj_cgroup *)(memcg_data & ~MEMCG_DATA_FLAGS_MASK); } @@ -457,22 +431,33 @@ static inline struct obj_cgroup *__folio_objcg(struct folio *folio) * proper memory cgroup pointer. It's not safe to call this function * against some type of folios, e.g. slab folios or ex-slab folios. * - * For a non-kmem folio any of the following ensures folio and memcg binding - * stability: + * For a folio any of the following ensures folio and objcg binding stability: * * - the folio lock * - LRU isolation * - lock_page_memcg() * - exclusive reference * + * Based on the stable binding of folio and objcg, for a folio any of the + * following ensures folio and memcg binding stability: + * + * - objcg_lock + * - cgroup_mutex + * - the lruvec lock + * - the split queue lock (only THP page) + * + * If the caller only want to ensure that the page counters of memcg are + * updated correctly, ensure that the binding stability of folio and objcg + * is sufficient. + * * Note: The caller should hold an rcu read lock to protect memcg associated * with a folio from being released. */ static inline struct mem_cgroup *folio_memcg(struct folio *folio) { - if (folio_memcg_kmem(folio)) - return obj_cgroup_memcg(__folio_objcg(folio)); - return __folio_memcg(folio); + struct obj_cgroup *objcg = folio_objcg(folio); + + return objcg ? obj_cgroup_memcg(objcg) : NULL; } /* @@ -496,6 +481,8 @@ static inline struct mem_cgroup *page_memcg(struct page *page) * folio is known to have a proper memory cgroup pointer. It's not safe * to call this function against some type of pages, e.g. slab pages or * ex-slab pages. + * + * The page and objcg or memcg binding rules can refer to folio_memcg(). */ static inline struct mem_cgroup *get_mem_cgroup_from_folio(struct folio *folio) { @@ -526,22 +513,20 @@ static inline struct mem_cgroup *get_mem_cgroup_from_page(struct page *page) * * Return: A pointer to the memory cgroup associated with the folio, * or NULL. + * + * The folio and objcg or memcg binding rules can refer to folio_memcg(). */ static inline struct mem_cgroup *folio_memcg_rcu(struct folio *folio) { unsigned long memcg_data = READ_ONCE(folio->memcg_data); + struct obj_cgroup *objcg; VM_BUG_ON_FOLIO(folio_test_slab(folio), folio); WARN_ON_ONCE(!rcu_read_lock_held()); - if (memcg_data & MEMCG_DATA_KMEM) { - struct obj_cgroup *objcg; + objcg = (void *)(memcg_data & ~MEMCG_DATA_FLAGS_MASK); - objcg = (void *)(memcg_data & ~MEMCG_DATA_FLAGS_MASK); - return obj_cgroup_memcg(objcg); - } - - return (struct mem_cgroup *)(memcg_data & ~MEMCG_DATA_FLAGS_MASK); + return objcg ? obj_cgroup_memcg(objcg) : NULL; } /* @@ -554,16 +539,10 @@ static inline struct mem_cgroup *folio_memcg_rcu(struct folio *folio) * has an associated memory cgroup pointer or an object cgroups vector or * an object cgroup. * - * For a non-kmem page any of the following ensures page and memcg binding - * stability: + * The page and objcg or memcg binding rules can refer to page_memcg(). * - * - the page lock - * - LRU isolation - * - lock_page_memcg() - * - exclusive reference - * - * For a kmem page a caller should hold an rcu read lock to protect memcg - * associated with a kmem page from being released. + * A caller should hold an rcu read lock to protect memcg associated with a + * page from being released. */ static inline struct mem_cgroup *page_memcg_check(struct page *page) { @@ -572,18 +551,14 @@ static inline struct mem_cgroup *page_memcg_check(struct page *page) * for slab pages, READ_ONCE() should be used here. */ unsigned long memcg_data = READ_ONCE(page->memcg_data); + struct obj_cgroup *objcg; if (memcg_data & MEMCG_DATA_OBJCGS) return NULL; - if (memcg_data & MEMCG_DATA_KMEM) { - struct obj_cgroup *objcg; - - objcg = (void *)(memcg_data & ~MEMCG_DATA_FLAGS_MASK); - return obj_cgroup_memcg(objcg); - } + objcg = (void *)(memcg_data & ~MEMCG_DATA_FLAGS_MASK); - return (struct mem_cgroup *)(memcg_data & ~MEMCG_DATA_FLAGS_MASK); + return objcg ? obj_cgroup_memcg(objcg) : NULL; } static inline struct mem_cgroup *get_mem_cgroup_from_objcg(struct obj_cgroup *objcg) diff --git a/mm/huge_memory.c b/mm/huge_memory.c index b4cd5ea42aca..7b7eedc6d5dd 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -561,6 +561,8 @@ pmd_t maybe_pmd_mkwrite(pmd_t pmd, struct vm_area_struct *vma) } #ifdef CONFIG_MEMCG +static struct shrinker deferred_split_shrinker; + static inline struct mem_cgroup *folio_split_queue_memcg(struct folio *folio, struct deferred_split *queue) { @@ -577,6 +579,39 @@ static inline struct deferred_split *folio_memcg_split_queue(struct folio *folio return memcg ? &memcg->deferred_split_queue : NULL; } + +static void thp_sq_reparent_lock(struct mem_cgroup *src, struct mem_cgroup *dst) +{ + spin_lock(&src->deferred_split_queue.split_queue_lock); + spin_lock_nested(&dst->deferred_split_queue.split_queue_lock, + SINGLE_DEPTH_NESTING); +} + +static void thp_sq_reparent_relocate(struct mem_cgroup *src, struct mem_cgroup *dst) +{ + int nid; + struct deferred_split *src_queue, *dst_queue; + + src_queue = &src->deferred_split_queue; + dst_queue = &dst->deferred_split_queue; + + if (!src_queue->split_queue_len) + return; + + list_splice_tail_init(&src_queue->split_queue, &dst_queue->split_queue); + dst_queue->split_queue_len += src_queue->split_queue_len; + src_queue->split_queue_len = 0; + + for_each_node(nid) + set_shrinker_bit(dst, nid, deferred_split_shrinker.id); +} + +static void thp_sq_reparent_unlock(struct mem_cgroup *src, struct mem_cgroup *dst) +{ + spin_unlock(&dst->deferred_split_queue.split_queue_lock); + spin_unlock(&src->deferred_split_queue.split_queue_lock); +} +DEFINE_MEMCG_REPARENT_OPS(thp_sq); #else static inline struct mem_cgroup *folio_split_queue_memcg(struct folio *folio, struct deferred_split *queue) diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 6ef3a264054e..803dbdf5f233 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -77,6 +77,7 @@ struct cgroup_subsys memory_cgrp_subsys __read_mostly; EXPORT_SYMBOL(memory_cgrp_subsys); struct mem_cgroup *root_mem_cgroup __read_mostly; +static struct obj_cgroup *root_obj_cgroup __read_mostly; /* Active memory cgroup to use from an interrupt context */ DEFINE_PER_CPU(struct mem_cgroup *, int_active_memcg); @@ -254,6 +255,11 @@ struct mem_cgroup *vmpressure_to_memcg(struct vmpressure *vmpr) static DEFINE_SPINLOCK(objcg_lock); +static inline bool obj_cgroup_is_root(struct obj_cgroup *objcg) +{ + return objcg == root_obj_cgroup; +} + #ifdef CONFIG_MEMCG_KMEM bool mem_cgroup_kmem_disabled(void) { @@ -361,8 +367,78 @@ static void objcg_reparent_unlock(struct mem_cgroup *src, struct mem_cgroup *dst static DEFINE_MEMCG_REPARENT_OPS(objcg); +static void lruvec_reparent_lock(struct mem_cgroup *src, struct mem_cgroup *dst) +{ + int nid, nest = 0; + + for_each_node(nid) { + spin_lock_nested(&mem_cgroup_lruvec(src, + NODE_DATA(nid))->lru_lock, nest++); + spin_lock_nested(&mem_cgroup_lruvec(dst, + NODE_DATA(nid))->lru_lock, nest++); + } +} + +static void lruvec_reparent_lru(struct lruvec *src, struct lruvec *dst, + enum lru_list lru) +{ + int zid; + struct mem_cgroup_per_node *mz_src, *mz_dst; + + mz_src = container_of(src, struct mem_cgroup_per_node, lruvec); + mz_dst = container_of(dst, struct mem_cgroup_per_node, lruvec); + + if (lru != LRU_UNEVICTABLE) + list_splice_tail_init(&src->lists[lru], &dst->lists[lru]); + + for (zid = 0; zid < MAX_NR_ZONES; zid++) { + mz_dst->lru_zone_size[zid][lru] += mz_src->lru_zone_size[zid][lru]; + mz_src->lru_zone_size[zid][lru] = 0; + } +} + +static void lruvec_reparent_relocate(struct mem_cgroup *src, struct mem_cgroup *dst) +{ + int nid; + + for_each_node(nid) { + enum lru_list lru; + struct lruvec *src_lruvec, *dst_lruvec; + + src_lruvec = mem_cgroup_lruvec(src, NODE_DATA(nid)); + dst_lruvec = mem_cgroup_lruvec(dst, NODE_DATA(nid)); + + dst_lruvec->anon_cost += src_lruvec->anon_cost; + dst_lruvec->file_cost += src_lruvec->file_cost; + + for_each_lru(lru) + lruvec_reparent_lru(src_lruvec, dst_lruvec, lru); + } +} + +static void lruvec_reparent_unlock(struct mem_cgroup *src, struct mem_cgroup *dst) +{ + int nid; + + for_each_node(nid) { + spin_unlock(&mem_cgroup_lruvec(dst, NODE_DATA(nid))->lru_lock); + spin_unlock(&mem_cgroup_lruvec(src, NODE_DATA(nid))->lru_lock); + } +} + +static DEFINE_MEMCG_REPARENT_OPS(lruvec); + +#ifdef CONFIG_TRANSPARENT_HUGEPAGE +DECLARE_MEMCG_REPARENT_OPS(thp_sq); +#endif + +/* The lock order depends on the order of elements in this array. */ static const struct memcg_reparent_ops *memcg_reparent_ops[] = { &memcg_objcg_reparent_ops, + &memcg_lruvec_reparent_ops, +#ifdef CONFIG_TRANSPARENT_HUGEPAGE + &memcg_thp_sq_reparent_ops, +#endif }; #define DEFINE_MEMCG_REPARENT_FUNC(phase) \ @@ -2825,18 +2901,33 @@ static inline void cancel_charge(struct mem_cgroup *memcg, unsigned int nr_pages page_counter_uncharge(&memcg->memsw, nr_pages); } -static void commit_charge(struct folio *folio, struct mem_cgroup *memcg) +static void commit_charge(struct folio *folio, struct obj_cgroup *objcg) { - VM_BUG_ON_FOLIO(folio_memcg(folio), folio); + VM_BUG_ON_FOLIO(folio_objcg(folio), folio); /* - * Any of the following ensures page's memcg stability: + * Any of the following ensures page's objcg stability: * * - the page lock * - LRU isolation * - lock_page_memcg() * - exclusive reference */ - folio->memcg_data = (unsigned long)memcg; + folio->memcg_data = (unsigned long)objcg; +} + +static struct obj_cgroup *get_obj_cgroup_from_memcg(struct mem_cgroup *memcg) +{ + struct obj_cgroup *objcg = NULL; + + rcu_read_lock(); + for (; memcg; memcg = parent_mem_cgroup(memcg)) { + objcg = rcu_dereference(memcg->objcg); + if (objcg && obj_cgroup_tryget(objcg)) + break; + } + rcu_read_unlock(); + + return objcg; } #ifdef CONFIG_MEMCG_KMEM @@ -2982,19 +3073,6 @@ struct mem_cgroup *mem_cgroup_from_slab_obj(void *p) return mem_cgroup_from_obj_folio(virt_to_folio(p), p); } -static struct obj_cgroup *__get_obj_cgroup_from_memcg(struct mem_cgroup *memcg) -{ - struct obj_cgroup *objcg = NULL; - - for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg)) { - objcg = rcu_dereference(memcg->objcg); - if (objcg && obj_cgroup_tryget(objcg)) - break; - objcg = NULL; - } - return objcg; -} - __always_inline struct obj_cgroup *get_obj_cgroup_from_current(void) { struct obj_cgroup *objcg = NULL; @@ -3008,7 +3086,16 @@ __always_inline struct obj_cgroup *get_obj_cgroup_from_current(void) memcg = active_memcg(); else memcg = mem_cgroup_from_task(current); - objcg = __get_obj_cgroup_from_memcg(memcg); + + if (mem_cgroup_is_root(memcg)) + goto out; + + objcg = get_obj_cgroup_from_memcg(memcg); + if (obj_cgroup_is_root(objcg)) { + obj_cgroup_put(objcg); + objcg = NULL; + } +out: rcu_read_unlock(); return objcg; } @@ -3020,20 +3107,10 @@ struct obj_cgroup *get_obj_cgroup_from_page(struct page *page) if (!memcg_kmem_enabled() || memcg_kmem_bypass()) return NULL; - if (PageMemcgKmem(page)) { - objcg = __folio_objcg(page_folio(page)); + objcg = folio_objcg(page_folio(page)); + if (objcg) obj_cgroup_get(objcg); - } else { - struct mem_cgroup *memcg; - rcu_read_lock(); - memcg = __folio_memcg(page_folio(page)); - if (memcg) - objcg = __get_obj_cgroup_from_memcg(memcg); - else - objcg = NULL; - rcu_read_unlock(); - } return objcg; } @@ -3128,13 +3205,13 @@ int __memcg_kmem_charge_page(struct page *page, gfp_t gfp, int order) void __memcg_kmem_uncharge_page(struct page *page, int order) { struct folio *folio = page_folio(page); - struct obj_cgroup *objcg; + struct obj_cgroup *objcg = folio_objcg(folio); unsigned int nr_pages = 1 << order; - if (!folio_memcg_kmem(folio)) + if (!objcg) return; - objcg = __folio_objcg(folio); + VM_BUG_ON_FOLIO(!folio_memcg_kmem(folio), folio); obj_cgroup_uncharge_pages(objcg, nr_pages); folio->memcg_data = 0; obj_cgroup_put(objcg); @@ -3388,26 +3465,21 @@ void obj_cgroup_uncharge(struct obj_cgroup *objcg, size_t size) #endif /* CONFIG_MEMCG_KMEM */ /* - * Because page_memcg(head) is not set on tails, set it now. + * Because page_objcg(head) is not set on tails, set it now. */ void split_page_memcg(struct page *head, unsigned int nr) { struct folio *folio = page_folio(head); - struct mem_cgroup *memcg = get_mem_cgroup_from_folio(folio); + struct obj_cgroup *objcg = folio_objcg(folio); int i; - if (mem_cgroup_disabled() || !memcg) + if (mem_cgroup_disabled() || !objcg) return; for (i = 1; i < nr; i++) folio_page(folio, i)->memcg_data = folio->memcg_data; - if (folio_memcg_kmem(folio)) - obj_cgroup_get_many(__folio_objcg(folio), nr - 1); - else - css_get_many(&memcg->css, nr - 1); - - css_put(&memcg->css); + obj_cgroup_get_many(objcg, nr - 1); } #ifdef CONFIG_MEMCG_SWAP @@ -5325,6 +5397,9 @@ static int mem_cgroup_css_online(struct cgroup_subsys_state *css) objcg->memcg = memcg; rcu_assign_pointer(memcg->objcg, objcg); + if (unlikely(mem_cgroup_is_root(memcg))) + root_obj_cgroup = objcg; + /* Online state pins memcg ID, memcg ID pins CSS */ refcount_set(&memcg->id.ref, 1); css_get(css); @@ -5729,10 +5804,12 @@ static int mem_cgroup_move_account(struct page *page, */ smp_mb(); - css_get(&to->css); - css_put(&from->css); + rcu_read_lock(); + obj_cgroup_get(rcu_dereference(to->objcg)); + obj_cgroup_put(rcu_dereference(from->objcg)); + rcu_read_unlock(); - folio->memcg_data = (unsigned long)to; + folio->memcg_data = (unsigned long)rcu_access_pointer(to->objcg); __folio_memcg_unlock(from); @@ -6208,6 +6285,42 @@ retry: walk_page_range(mc.mm, 0, ULONG_MAX, &charge_walk_ops, NULL); mmap_read_unlock(mc.mm); atomic_dec(&mc.from->moving_account); + + /* + * Moving its pages to another memcg is finished. Wait for already + * started RCU-only updates to finish to make sure that the caller + * of lock_page_memcg() can unlock the correct move_lock. The + * possible bad scenario would like: + * + * CPU0: CPU1: + * mem_cgroup_move_charge() + * walk_page_range() + * + * lock_page_memcg(page) + * memcg = folio_memcg() + * spin_lock_irqsave(&memcg->move_lock) + * memcg->move_lock_task = current + * + * atomic_dec(&mc.from->moving_account) + * + * mem_cgroup_css_offline() + * memcg_offline_kmem() + * memcg_reparent_objcgs() <== reparented + * + * unlock_page_memcg(page) + * memcg = folio_memcg() <== memcg has been changed + * if (memcg->move_lock_task == current) <== false + * spin_unlock_irqrestore(&memcg->move_lock) + * + * Once mem_cgroup_move_charge() returns (it means that the cgroup_mutex + * would be released soon), the page can be reparented to its parent + * memcg. When the unlock_page_memcg() is called for the page, we will + * miss unlock the move_lock. So using synchronize_rcu to wait for + * already started RCU-only updates to finish before this function + * returns (mem_cgroup_move_charge() and mem_cgroup_css_offline() are + * serialized by cgroup_mutex). + */ + synchronize_rcu(); } /* @@ -6822,21 +6935,26 @@ void mem_cgroup_calculate_protection(struct mem_cgroup *root, static int charge_memcg(struct folio *folio, struct mem_cgroup *memcg, gfp_t gfp) { + struct obj_cgroup *objcg; long nr_pages = folio_nr_pages(folio); - int ret; + int ret = 0; - ret = try_charge(memcg, gfp, nr_pages); + objcg = get_obj_cgroup_from_memcg(memcg); + /* Do not account at the root objcg level. */ + if (!obj_cgroup_is_root(objcg)) + ret = try_charge(memcg, gfp, nr_pages); if (ret) goto out; - css_get(&memcg->css); - commit_charge(folio, memcg); + obj_cgroup_get(objcg); + commit_charge(folio, objcg); local_irq_disable(); mem_cgroup_charge_statistics(memcg, nr_pages); memcg_check_events(memcg, folio_nid(folio)); local_irq_enable(); out: + obj_cgroup_put(objcg); return ret; } @@ -6922,7 +7040,7 @@ void mem_cgroup_swapin_uncharge_swap(swp_entry_t entry) } struct uncharge_gather { - struct mem_cgroup *memcg; + struct obj_cgroup *objcg; unsigned long nr_memory; unsigned long pgpgout; unsigned long nr_kmem; @@ -6937,63 +7055,56 @@ static inline void uncharge_gather_clear(struct uncharge_gather *ug) static void uncharge_batch(const struct uncharge_gather *ug) { unsigned long flags; + struct mem_cgroup *memcg; + rcu_read_lock(); + memcg = obj_cgroup_memcg(ug->objcg); if (ug->nr_memory) { - page_counter_uncharge(&ug->memcg->memory, ug->nr_memory); + page_counter_uncharge(&memcg->memory, ug->nr_memory); if (do_memsw_account()) - page_counter_uncharge(&ug->memcg->memsw, ug->nr_memory); + page_counter_uncharge(&memcg->memsw, ug->nr_memory); if (ug->nr_kmem) - memcg_account_kmem(ug->memcg, -ug->nr_kmem); - memcg_oom_recover(ug->memcg); + memcg_account_kmem(memcg, -ug->nr_kmem); + memcg_oom_recover(memcg); } local_irq_save(flags); - __count_memcg_events(ug->memcg, PGPGOUT, ug->pgpgout); - __this_cpu_add(ug->memcg->vmstats_percpu->nr_page_events, ug->nr_memory); - memcg_check_events(ug->memcg, ug->nid); + __count_memcg_events(memcg, PGPGOUT, ug->pgpgout); + __this_cpu_add(memcg->vmstats_percpu->nr_page_events, ug->nr_memory); + memcg_check_events(memcg, ug->nid); local_irq_restore(flags); + rcu_read_unlock(); /* drop reference from uncharge_folio */ - css_put(&ug->memcg->css); + obj_cgroup_put(ug->objcg); } static void uncharge_folio(struct folio *folio, struct uncharge_gather *ug) { long nr_pages; - struct mem_cgroup *memcg; struct obj_cgroup *objcg; VM_BUG_ON_FOLIO(folio_test_lru(folio), folio); /* * Nobody should be changing or seriously looking at - * folio memcg or objcg at this point, we have fully - * exclusive access to the folio. + * folio objcg at this point, we have fully exclusive + * access to the folio. */ - if (folio_memcg_kmem(folio)) { - objcg = __folio_objcg(folio); - /* - * This get matches the put at the end of the function and - * kmem pages do not hold memcg references anymore. - */ - memcg = get_mem_cgroup_from_objcg(objcg); - } else { - memcg = __folio_memcg(folio); - } - - if (!memcg) + objcg = folio_objcg(folio); + if (!objcg) return; - if (ug->memcg != memcg) { - if (ug->memcg) { + if (ug->objcg != objcg) { + if (ug->objcg) { uncharge_batch(ug); uncharge_gather_clear(ug); } - ug->memcg = memcg; + ug->objcg = objcg; ug->nid = folio_nid(folio); - /* pairs with css_put in uncharge_batch */ - css_get(&memcg->css); + /* pairs with obj_cgroup_put in uncharge_batch */ + obj_cgroup_get(objcg); } nr_pages = folio_nr_pages(folio); @@ -7001,19 +7112,15 @@ static void uncharge_folio(struct folio *folio, struct uncharge_gather *ug) if (folio_memcg_kmem(folio)) { ug->nr_memory += nr_pages; ug->nr_kmem += nr_pages; - - folio->memcg_data = 0; - obj_cgroup_put(objcg); } else { /* LRU pages aren't accounted at the root level */ - if (!mem_cgroup_is_root(memcg)) + if (!obj_cgroup_is_root(objcg)) ug->nr_memory += nr_pages; ug->pgpgout++; - - folio->memcg_data = 0; } - css_put(&memcg->css); + folio->memcg_data = 0; + obj_cgroup_put(objcg); } void __mem_cgroup_uncharge(struct folio *folio) @@ -7021,7 +7128,7 @@ void __mem_cgroup_uncharge(struct folio *folio) struct uncharge_gather ug; /* Don't touch folio->lru of any random page, pre-check: */ - if (!folio_memcg(folio)) + if (!folio_objcg(folio)) return; uncharge_gather_clear(&ug); @@ -7044,7 +7151,7 @@ void __mem_cgroup_uncharge_list(struct list_head *page_list) uncharge_gather_clear(&ug); list_for_each_entry(folio, page_list, lru) uncharge_folio(folio, &ug); - if (ug.memcg) + if (ug.objcg) uncharge_batch(&ug); } @@ -7061,6 +7168,7 @@ void __mem_cgroup_uncharge_list(struct list_head *page_list) void mem_cgroup_migrate(struct folio *old, struct folio *new) { struct mem_cgroup *memcg; + struct obj_cgroup *objcg; long nr_pages = folio_nr_pages(new); unsigned long flags; @@ -7073,30 +7181,33 @@ void mem_cgroup_migrate(struct folio *old, struct folio *new) return; /* Page cache replacement: new folio already charged? */ - if (folio_memcg(new)) + if (folio_objcg(new)) return; - memcg = get_mem_cgroup_from_folio(old); - VM_WARN_ON_ONCE_FOLIO(!memcg, old); - if (!memcg) + objcg = folio_objcg(old); + VM_WARN_ON_ONCE_FOLIO(!objcg, old); + if (!objcg) return; + rcu_read_lock(); + memcg = obj_cgroup_memcg(objcg); + /* Force-charge the new page. The old one will be freed soon */ - if (!mem_cgroup_is_root(memcg)) { + if (!obj_cgroup_is_root(objcg)) { page_counter_charge(&memcg->memory, nr_pages); if (do_memsw_account()) page_counter_charge(&memcg->memsw, nr_pages); } - css_get(&memcg->css); - commit_charge(new, memcg); + obj_cgroup_get(objcg); + commit_charge(new, objcg); local_irq_save(flags); mem_cgroup_charge_statistics(memcg, nr_pages); memcg_check_events(memcg, folio_nid(new)); local_irq_restore(flags); - css_put(&memcg->css); + rcu_read_unlock(); } DEFINE_STATIC_KEY_FALSE(memcg_sockets_enabled_key); @@ -7269,6 +7380,7 @@ static struct mem_cgroup *mem_cgroup_id_get_online(struct mem_cgroup *memcg) void mem_cgroup_swapout(struct folio *folio, swp_entry_t entry) { struct mem_cgroup *memcg, *swap_memcg; + struct obj_cgroup *objcg; unsigned int nr_entries; unsigned short oldid; @@ -7281,15 +7393,16 @@ void mem_cgroup_swapout(struct folio *folio, swp_entry_t entry) if (cgroup_subsys_on_dfl(memory_cgrp_subsys)) return; + objcg = folio_objcg(folio); + VM_WARN_ON_ONCE_FOLIO(!objcg, folio); + if (!objcg) + return; + /* * Interrupts should be disabled by the caller (see the comments below), * which can serve as RCU read-side critical sections. */ - memcg = folio_memcg(folio); - - VM_WARN_ON_ONCE_FOLIO(!memcg, folio); - if (!memcg) - return; + memcg = obj_cgroup_memcg(objcg); /* * In case the memcg owning these pages has been offlined and doesn't @@ -7308,7 +7421,7 @@ void mem_cgroup_swapout(struct folio *folio, swp_entry_t entry) folio->memcg_data = 0; - if (!mem_cgroup_is_root(memcg)) + if (!obj_cgroup_is_root(objcg)) page_counter_uncharge(&memcg->memory, nr_entries); if (!cgroup_memory_noswap && memcg != swap_memcg) { @@ -7328,7 +7441,7 @@ void mem_cgroup_swapout(struct folio *folio, swp_entry_t entry) memcg_stats_unlock(); memcg_check_events(memcg, folio_nid(folio)); - css_put(&memcg->css); + obj_cgroup_put(objcg); } /** -- cgit v1.2.3 From 1af63c7e2e633d3ee8280245a9a26314ed32c97a Mon Sep 17 00:00:00 2001 From: Muchun Song Date: Tue, 21 Jun 2022 20:56:57 +0800 Subject: mm: lru: add VM_WARN_ON_ONCE_FOLIO to lru maintenance function MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We need to make sure that the page is deleted from or added to the correct lruvec list. So add a VM_WARN_ON_ONCE_FOLIO() to catch invalid users. Then the VM_BUG_ON_PAGE() in move_pages_to_lru() could be removed since add_page_to_lru_list() will check that. Link: https://lkml.kernel.org/r/20220621125658.64935-11-songmuchun@bytedance.com Signed-off-by: Muchun Song Acked-by: Roman Gushchin Cc: Johannes Weiner Cc: Michal Hocko Cc: Michal Koutný Cc: Shakeel Butt Cc: Waiman Long Cc: Xiongchun Duan Signed-off-by: Andrew Morton --- include/linux/mm_inline.h | 6 ++++++ mm/vmscan.c | 1 - 2 files changed, 6 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/mm_inline.h b/include/linux/mm_inline.h index 7b25b53c474a..6585198b19e2 100644 --- a/include/linux/mm_inline.h +++ b/include/linux/mm_inline.h @@ -99,6 +99,8 @@ void lruvec_add_folio(struct lruvec *lruvec, struct folio *folio) { enum lru_list lru = folio_lru_list(folio); + VM_WARN_ON_ONCE_FOLIO(!folio_matches_lruvec(folio, lruvec), folio); + update_lru_size(lruvec, lru, folio_zonenum(folio), folio_nr_pages(folio)); if (lru != LRU_UNEVICTABLE) @@ -116,6 +118,8 @@ void lruvec_add_folio_tail(struct lruvec *lruvec, struct folio *folio) { enum lru_list lru = folio_lru_list(folio); + VM_WARN_ON_ONCE_FOLIO(!folio_matches_lruvec(folio, lruvec), folio); + update_lru_size(lruvec, lru, folio_zonenum(folio), folio_nr_pages(folio)); /* This is not expected to be used on LRU_UNEVICTABLE */ @@ -133,6 +137,8 @@ void lruvec_del_folio(struct lruvec *lruvec, struct folio *folio) { enum lru_list lru = folio_lru_list(folio); + VM_WARN_ON_ONCE_FOLIO(!folio_matches_lruvec(folio, lruvec), folio); + if (lru != LRU_UNEVICTABLE) list_del(&folio->lru); update_lru_size(lruvec, lru, folio_zonenum(folio), diff --git a/mm/vmscan.c b/mm/vmscan.c index 697656151431..51b1607c81e4 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -2361,7 +2361,6 @@ static unsigned int move_pages_to_lru(struct list_head *list) continue; } - VM_BUG_ON_FOLIO(!folio_matches_lruvec(folio, lruvec), folio); lruvec_add_folio(lruvec, folio); nr_pages = folio_nr_pages(folio); nr_moved += nr_pages; -- cgit v1.2.3 From 3cf5218e5c616b0c65ce4380b5b6ac49998f1138 Mon Sep 17 00:00:00 2001 From: Yun-Ze Li Date: Mon, 20 Jun 2022 07:15:16 +0000 Subject: mm, docs: fix comments that mention mem_hotplug_end() Comments that mention mem_hotplug_end() are confusing as there is no function called mem_hotplug_end(). Fix them by replacing all the occurences of mem_hotplug_end() in the comments with mem_hotplug_done(). Link: https://lkml.kernel.org/r/20220620071516.1286101-1-p76091292@gs.ncku.edu.tw Signed-off-by: Yun-Ze Li Cc: Souptick Joarder Signed-off-by: Andrew Morton --- include/linux/mmzone.h | 4 ++-- mm/compaction.c | 2 +- mm/vmscan.c | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 2b5757752333..35a2c8e18107 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -591,7 +591,7 @@ struct zone { * give them a chance of being in the same cacheline. * * Write access to present_pages at runtime should be protected by - * mem_hotplug_begin/end(). Any reader who can't tolerant drift of + * mem_hotplug_begin/done(). Any reader who can't tolerant drift of * present_pages should get_online_mems() to get a stable value. */ atomic_long_t managed_pages; @@ -870,7 +870,7 @@ typedef struct pglist_data { unsigned long nr_reclaim_start; /* nr pages written while throttled * when throttling started. */ struct task_struct *kswapd; /* Protected by - mem_hotplug_begin/end() */ + mem_hotplug_begin/done() */ int kswapd_order; enum zone_type kswapd_highest_zoneidx; diff --git a/mm/compaction.c b/mm/compaction.c index fe49ac9aedd8..8f30b863c63f 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -3030,7 +3030,7 @@ void kcompactd_run(int nid) /* * Called by memory hotplug when all memory in a node is offlined. Caller must - * hold mem_hotplug_begin/end(). + * be hold mem_hotplug_begin/done(). */ void kcompactd_stop(int nid) { diff --git a/mm/vmscan.c b/mm/vmscan.c index 51b1607c81e4..c47d7cafa4a8 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -4645,7 +4645,7 @@ void kswapd_run(int nid) /* * Called by memory hotplug when all memory in a node is offlined. Caller must - * hold mem_hotplug_begin/end(). + * be hold mem_hotplug_begin/done(). */ void kswapd_stop(int nid) { -- cgit v1.2.3 From 9d5d1c02c7b6bc556706fc0a63c6972bc38f3031 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Tue, 21 Jun 2022 17:05:38 -0700 Subject: mm-docs-fix-comments-that-mention-mem_hotplug_end-fix grammatical fixes Cc: Souptick Joarder Cc: Yun-Ze Li Signed-off-by: Andrew Morton --- include/linux/mmzone.h | 2 +- mm/compaction.c | 2 +- mm/vmscan.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 35a2c8e18107..735bf5b37949 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -592,7 +592,7 @@ struct zone { * * Write access to present_pages at runtime should be protected by * mem_hotplug_begin/done(). Any reader who can't tolerant drift of - * present_pages should get_online_mems() to get a stable value. + * present_pages should use get_online_mems() to get a stable value. */ atomic_long_t managed_pages; unsigned long spanned_pages; diff --git a/mm/compaction.c b/mm/compaction.c index 8f30b863c63f..2101ef2bf109 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -3030,7 +3030,7 @@ void kcompactd_run(int nid) /* * Called by memory hotplug when all memory in a node is offlined. Caller must - * be hold mem_hotplug_begin/done(). + * be holding mem_hotplug_begin/done(). */ void kcompactd_stop(int nid) { diff --git a/mm/vmscan.c b/mm/vmscan.c index c47d7cafa4a8..893b881711cb 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -4645,7 +4645,7 @@ void kswapd_run(int nid) /* * Called by memory hotplug when all memory in a node is offlined. Caller must - * be hold mem_hotplug_begin/done(). + * be holding mem_hotplug_begin/done(). */ void kswapd_stop(int nid) { -- cgit v1.2.3 From 80dad8a71596d47555efd7b3d82410832ad51dc0 Mon Sep 17 00:00:00 2001 From: Mike Kravetz Date: Tue, 21 Jun 2022 16:56:17 -0700 Subject: hugetlb: skip to end of PT page mapping when pte not present Patch series "hugetlb: speed up linear address scanning", v2. At unmap, fork and remap time hugetlb address ranges are linearly scanned. We can optimize these scans if the ranges are sparsely populated. Also, enable page table "Lazy copy" for hugetlb at fork. NOTE: Architectures not defining CONFIG_ARCH_WANT_GENERAL_HUGETLB need to add an arch specific version hugetlb_mask_last_page() to take advantage of sparse address scanning improvements. Baolin Wang added the routine for arm64. Other architectures which could be optimized are: ia64, mips, parisc, powerpc, s390, sh and sparc. This patch (of 4): HugeTLB address ranges are linearly scanned during fork, unmap and remap operations. If a non-present entry is encountered, the code currently continues to the next huge page aligned address. However, a non-present entry implies that the page table page for that entry is not present. Therefore, the linear scan can skip to the end of range mapped by the page table page. This can speed operations on large sparsely populated hugetlb mappings. Create a new routine hugetlb_mask_last_page() that will return an address mask. When the mask is ORed with an address, the result will be the address of the last huge page mapped by the associated page table page. Use this mask to update addresses in routines which linearly scan hugetlb address ranges when a non-present pte is encountered. hugetlb_mask_last_page is related to the implementation of huge_pte_offset as hugetlb_mask_last_page is called when huge_pte_offset returns NULL. This patch only provides a complete hugetlb_mask_last_page implementation when CONFIG_ARCH_WANT_GENERAL_HUGETLB is defined. Architectures which provide their own versions of huge_pte_offset can also provide their own version of hugetlb_mask_last_page. Link: https://lkml.kernel.org/r/20220621235620.291305-1-mike.kravetz@oracle.com Link: https://lkml.kernel.org/r/20220621235620.291305-2-mike.kravetz@oracle.com Signed-off-by: Mike Kravetz Tested-by: Baolin Wang Reviewed-by: Baolin Wang Acked-by: Muchun Song Reported-by: kernel test robot Cc: Michal Hocko Cc: Peter Xu Cc: Naoya Horiguchi Cc: James Houghton Cc: Mina Almasry Cc: "Aneesh Kumar K.V" Cc: Anshuman Khandual Cc: Paul Walmsley Cc: Christian Borntraeger Cc: Catalin Marinas Cc: Will Deacon Cc: Rolf Eike Beer Cc: David Hildenbrand Signed-off-by: Andrew Morton --- include/linux/hugetlb.h | 1 + mm/hugetlb.c | 56 ++++++++++++++++++++++++++++++++++++++++++++----- 2 files changed, 52 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index 756b66ff025e..bfae4efdd4a9 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -194,6 +194,7 @@ pte_t *huge_pte_alloc(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, unsigned long sz); pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr, unsigned long sz); +unsigned long hugetlb_mask_last_page(struct hstate *h); int huge_pmd_unshare(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long *addr, pte_t *ptep); void adjust_range_if_pmd_sharing_possible(struct vm_area_struct *vma, diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 65454896f174..b230233fe337 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -4727,6 +4727,7 @@ int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src, unsigned long npages = pages_per_huge_page(h); struct address_space *mapping = src_vma->vm_file->f_mapping; struct mmu_notifier_range range; + unsigned long last_addr_mask; int ret = 0; if (cow) { @@ -4746,11 +4747,14 @@ int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src, i_mmap_lock_read(mapping); } + last_addr_mask = hugetlb_mask_last_page(h); for (addr = src_vma->vm_start; addr < src_vma->vm_end; addr += sz) { spinlock_t *src_ptl, *dst_ptl; src_pte = huge_pte_offset(src, addr, sz); - if (!src_pte) + if (!src_pte) { + addr |= last_addr_mask; continue; + } dst_pte = huge_pte_alloc(dst, dst_vma, addr, sz); if (!dst_pte) { ret = -ENOMEM; @@ -4767,8 +4771,10 @@ int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src, * after taking the lock below. */ dst_entry = huge_ptep_get(dst_pte); - if ((dst_pte == src_pte) || !huge_pte_none(dst_entry)) + if ((dst_pte == src_pte) || !huge_pte_none(dst_entry)) { + addr |= last_addr_mask; continue; + } dst_ptl = huge_pte_lock(h, dst, dst_pte); src_ptl = huge_pte_lockptr(h, src, src_pte); @@ -4929,6 +4935,7 @@ int move_hugetlb_page_tables(struct vm_area_struct *vma, unsigned long sz = huge_page_size(h); struct mm_struct *mm = vma->vm_mm; unsigned long old_end = old_addr + len; + unsigned long last_addr_mask; unsigned long old_addr_copy; pte_t *src_pte, *dst_pte; struct mmu_notifier_range range; @@ -4944,12 +4951,16 @@ int move_hugetlb_page_tables(struct vm_area_struct *vma, flush_cache_range(vma, range.start, range.end); mmu_notifier_invalidate_range_start(&range); + last_addr_mask = hugetlb_mask_last_page(h); /* Prevent race with file truncation */ i_mmap_lock_write(mapping); for (; old_addr < old_end; old_addr += sz, new_addr += sz) { src_pte = huge_pte_offset(mm, old_addr, sz); - if (!src_pte) + if (!src_pte) { + old_addr |= last_addr_mask; + new_addr |= last_addr_mask; continue; + } if (huge_pte_none(huge_ptep_get(src_pte))) continue; @@ -4994,6 +5005,7 @@ static void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct struct hstate *h = hstate_vma(vma); unsigned long sz = huge_page_size(h); struct mmu_notifier_range range; + unsigned long last_addr_mask; bool force_flush = false; WARN_ON(!is_vm_hugetlb_page(vma)); @@ -5014,11 +5026,14 @@ static void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct end); adjust_range_if_pmd_sharing_possible(vma, &range.start, &range.end); mmu_notifier_invalidate_range_start(&range); + last_addr_mask = hugetlb_mask_last_page(h); address = start; for (; address < end; address += sz) { ptep = huge_pte_offset(mm, address, sz); - if (!ptep) + if (!ptep) { + address |= last_addr_mask; continue; + } ptl = huge_pte_lock(h, mm, ptep); if (huge_pmd_unshare(mm, vma, &address, ptep)) { @@ -6286,6 +6301,7 @@ unsigned long hugetlb_change_protection(struct vm_area_struct *vma, unsigned long pages = 0, psize = huge_page_size(h); bool shared_pmd = false; struct mmu_notifier_range range; + unsigned long last_addr_mask; bool uffd_wp = cp_flags & MM_CP_UFFD_WP; bool uffd_wp_resolve = cp_flags & MM_CP_UFFD_WP_RESOLVE; @@ -6302,12 +6318,15 @@ unsigned long hugetlb_change_protection(struct vm_area_struct *vma, flush_cache_range(vma, range.start, range.end); mmu_notifier_invalidate_range_start(&range); + last_addr_mask = hugetlb_mask_last_page(h); i_mmap_lock_write(vma->vm_file->f_mapping); for (; address < end; address += psize) { spinlock_t *ptl; ptep = huge_pte_offset(mm, address, psize); - if (!ptep) + if (!ptep) { + address |= last_addr_mask; continue; + } ptl = huge_pte_lock(h, mm, ptep); if (huge_pmd_unshare(mm, vma, &address, ptep)) { /* @@ -6858,6 +6877,33 @@ pte_t *huge_pte_offset(struct mm_struct *mm, return (pte_t *)pmd; } +/* + * Return a mask that can be used to update an address to the last huge + * page in a page table page mapping size. Used to skip non-present + * page table entries when linearly scanning address ranges. Architectures + * with unique huge page to page table relationships can define their own + * version of this routine. + */ +unsigned long hugetlb_mask_last_page(struct hstate *h) +{ + unsigned long hp_size = huge_page_size(h); + + if (hp_size == PUD_SIZE) + return P4D_SIZE - PUD_SIZE; + else if (hp_size == PMD_SIZE) + return PUD_SIZE - PMD_SIZE; + else + return 0UL; +} + +#else + +/* See description above. Architectures can provide their own version. */ +__weak unsigned long hugetlb_mask_last_page(struct hstate *h) +{ + return 0UL; +} + #endif /* CONFIG_ARCH_WANT_GENERAL_HUGETLB */ /* -- cgit v1.2.3 From c6ef882cb28fe67380c5bfc54965696a390eb2ad Mon Sep 17 00:00:00 2001 From: Mike Kravetz Date: Tue, 21 Jun 2022 16:56:19 -0700 Subject: hugetlb: do not update address in huge_pmd_unshare As an optimization for loops sequentially processing hugetlb address ranges, huge_pmd_unshare would update a passed address if it unshared a pmd. Updating a loop control variable outside the loop like this is generally a bad idea. These loops are now using hugetlb_mask_last_page to optimize scanning when non-present ptes are discovered. The same can be done when huge_pmd_unshare returns 1 indicating a pmd was unshared. Remove address update from huge_pmd_unshare. Change the passed argument type and update all callers. In loops sequentially processing addresses use hugetlb_mask_last_page to update address if pmd is unshared. Link: https://lkml.kernel.org/r/20220621235620.291305-4-mike.kravetz@oracle.com Signed-off-by: Mike Kravetz Acked-by: Muchun Song Reviewed-by: Baolin Wang Cc: "Aneesh Kumar K.V" Cc: Anshuman Khandual Cc: Catalin Marinas Cc: Christian Borntraeger Cc: David Hildenbrand Cc: James Houghton Cc: kernel test robot Cc: Michal Hocko Cc: Mina Almasry Cc: Naoya Horiguchi Cc: Paul Walmsley Cc: Peter Xu Cc: Rolf Eike Beer Cc: Will Deacon Cc: Stephen Rothwell Signed-off-by: Andrew Morton --- include/linux/hugetlb.h | 4 ++-- mm/hugetlb.c | 46 +++++++++++++++++++--------------------------- mm/rmap.c | 4 ++-- 3 files changed, 23 insertions(+), 31 deletions(-) (limited to 'include/linux') diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index bfae4efdd4a9..05c5317df380 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -196,7 +196,7 @@ pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr, unsigned long sz); unsigned long hugetlb_mask_last_page(struct hstate *h); int huge_pmd_unshare(struct mm_struct *mm, struct vm_area_struct *vma, - unsigned long *addr, pte_t *ptep); + unsigned long addr, pte_t *ptep); void adjust_range_if_pmd_sharing_possible(struct vm_area_struct *vma, unsigned long *start, unsigned long *end); struct page *follow_huge_addr(struct mm_struct *mm, unsigned long address, @@ -243,7 +243,7 @@ static inline struct address_space *hugetlb_page_mapping_lock_write( static inline int huge_pmd_unshare(struct mm_struct *mm, struct vm_area_struct *vma, - unsigned long *addr, pte_t *ptep) + unsigned long addr, pte_t *ptep) { return 0; } diff --git a/mm/hugetlb.c b/mm/hugetlb.c index b230233fe337..af388ba7e720 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -4936,7 +4936,6 @@ int move_hugetlb_page_tables(struct vm_area_struct *vma, struct mm_struct *mm = vma->vm_mm; unsigned long old_end = old_addr + len; unsigned long last_addr_mask; - unsigned long old_addr_copy; pte_t *src_pte, *dst_pte; struct mmu_notifier_range range; bool shared_pmd = false; @@ -4964,14 +4963,10 @@ int move_hugetlb_page_tables(struct vm_area_struct *vma, if (huge_pte_none(huge_ptep_get(src_pte))) continue; - /* old_addr arg to huge_pmd_unshare() is a pointer and so the - * arg may be modified. Pass a copy instead to preserve the - * value in old_addr. - */ - old_addr_copy = old_addr; - - if (huge_pmd_unshare(mm, vma, &old_addr_copy, src_pte)) { + if (huge_pmd_unshare(mm, vma, old_addr, src_pte)) { shared_pmd = true; + old_addr |= last_addr_mask; + new_addr |= last_addr_mask; continue; } @@ -5036,10 +5031,11 @@ static void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct } ptl = huge_pte_lock(h, mm, ptep); - if (huge_pmd_unshare(mm, vma, &address, ptep)) { + if (huge_pmd_unshare(mm, vma, address, ptep)) { spin_unlock(ptl); tlb_flush_pmd_range(tlb, address & PUD_MASK, PUD_SIZE); force_flush = true; + address |= last_addr_mask; continue; } @@ -6328,7 +6324,7 @@ unsigned long hugetlb_change_protection(struct vm_area_struct *vma, continue; } ptl = huge_pte_lock(h, mm, ptep); - if (huge_pmd_unshare(mm, vma, &address, ptep)) { + if (huge_pmd_unshare(mm, vma, address, ptep)) { /* * When uffd-wp is enabled on the vma, unshare * shouldn't happen at all. Warn about it if it @@ -6338,6 +6334,7 @@ unsigned long hugetlb_change_protection(struct vm_area_struct *vma, pages++; spin_unlock(ptl); shared_pmd = true; + address |= last_addr_mask; continue; } pte = huge_ptep_get(ptep); @@ -6761,11 +6758,11 @@ out: * 0 the underlying pte page is not shared, or it is the last user */ int huge_pmd_unshare(struct mm_struct *mm, struct vm_area_struct *vma, - unsigned long *addr, pte_t *ptep) + unsigned long addr, pte_t *ptep) { - pgd_t *pgd = pgd_offset(mm, *addr); - p4d_t *p4d = p4d_offset(pgd, *addr); - pud_t *pud = pud_offset(p4d, *addr); + pgd_t *pgd = pgd_offset(mm, addr); + p4d_t *p4d = p4d_offset(pgd, addr); + pud_t *pud = pud_offset(p4d, addr); i_mmap_assert_write_locked(vma->vm_file->f_mapping); BUG_ON(page_count(virt_to_page(ptep)) == 0); @@ -6775,14 +6772,6 @@ int huge_pmd_unshare(struct mm_struct *mm, struct vm_area_struct *vma, pud_clear(pud); put_page(virt_to_page(ptep)); mm_dec_nr_pmds(mm); - /* - * This update of passed address optimizes loops sequentially - * processing addresses in increments of huge page size (PMD_SIZE - * in this case). By clearing the pud, a PUD_SIZE area is unmapped. - * Update address to the 'last page' in the cleared area so that - * calling loop can move to first page past this area. - */ - *addr |= PUD_SIZE - PMD_SIZE; return 1; } @@ -6794,7 +6783,7 @@ pte_t *huge_pmd_share(struct mm_struct *mm, struct vm_area_struct *vma, } int huge_pmd_unshare(struct mm_struct *mm, struct vm_area_struct *vma, - unsigned long *addr, pte_t *ptep) + unsigned long addr, pte_t *ptep) { return 0; } @@ -6901,6 +6890,12 @@ unsigned long hugetlb_mask_last_page(struct hstate *h) /* See description above. Architectures can provide their own version. */ __weak unsigned long hugetlb_mask_last_page(struct hstate *h) { + unsigned long hp_size = huge_page_size(h); + +#ifdef CONFIG_ARCH_WANT_HUGE_PMD_SHARE + if (hp_size == PMD_SIZE) + return PUD_SIZE - PMD_SIZE; +#endif return 0UL; } @@ -7127,14 +7122,11 @@ void hugetlb_unshare_all_pmds(struct vm_area_struct *vma) mmu_notifier_invalidate_range_start(&range); i_mmap_lock_write(vma->vm_file->f_mapping); for (address = start; address < end; address += PUD_SIZE) { - unsigned long tmp = address; - ptep = huge_pte_offset(mm, address, sz); if (!ptep) continue; ptl = huge_pte_lock(h, mm, ptep); - /* We don't want 'address' to be changed */ - huge_pmd_unshare(mm, vma, &tmp, ptep); + huge_pmd_unshare(mm, vma, address, ptep); spin_unlock(ptl); } flush_hugetlb_tlb_range(vma, start, end); diff --git a/mm/rmap.c b/mm/rmap.c index a2933c186a5a..09ee423d59d1 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -1559,7 +1559,7 @@ static bool try_to_unmap_one(struct folio *folio, struct vm_area_struct *vma, * do this outside rmap routines. */ VM_BUG_ON(!anon && !(flags & TTU_RMAP_LOCKED)); - if (!anon && huge_pmd_unshare(mm, vma, &address, pvmw.pte)) { + if (!anon && huge_pmd_unshare(mm, vma, address, pvmw.pte)) { flush_tlb_range(vma, range.start, range.end); mmu_notifier_invalidate_range(mm, range.start, range.end); @@ -1937,7 +1937,7 @@ static bool try_to_migrate_one(struct folio *folio, struct vm_area_struct *vma, * do this outside rmap routines. */ VM_BUG_ON(!anon && !(flags & TTU_RMAP_LOCKED)); - if (!anon && huge_pmd_unshare(mm, vma, &address, pvmw.pte)) { + if (!anon && huge_pmd_unshare(mm, vma, address, pvmw.pte)) { flush_tlb_range(vma, range.start, range.end); mmu_notifier_invalidate_range(mm, range.start, range.end); -- cgit v1.2.3 From b308fe6ead26d0c44865023c3028ba3ae848b23f Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Fri, 24 Jun 2022 13:54:17 +0100 Subject: mm/page_alloc: add page->buddy_list and page->pcp_list Patch series "Drain remote per-cpu directly", v5. Some setups, notably NOHZ_FULL CPUs, may be running realtime or latency-sensitive applications that cannot tolerate interference due to per-cpu drain work queued by __drain_all_pages(). Introduce a new mechanism to remotely drain the per-cpu lists. It is made possible by remotely locking 'struct per_cpu_pages' new per-cpu spinlocks. This has two advantages, the time to drain is more predictable and other unrelated tasks are not interrupted. This series has the same intent as Nicolas' series "mm/page_alloc: Remote per-cpu lists drain support" -- avoid interference of a high priority task due to a workqueue item draining per-cpu page lists. While many workloads can tolerate a brief interruption, it may cause a real-time task running on a NOHZ_FULL CPU to miss a deadline and at minimum, the draining is non-deterministic. Currently an IRQ-safe local_lock protects the page allocator per-cpu lists. The local_lock on its own prevents migration and the IRQ disabling protects from corruption due to an interrupt arriving while a page allocation is in progress. This series adjusts the locking. A spinlock is added to struct per_cpu_pages to protect the list contents while local_lock_irq is ultimately replaced by just the spinlock in the final patch. This allows a remote CPU to safely. Follow-on work should allow the spin_lock_irqsave to be converted to spin_lock to avoid IRQs being disabled/enabled in most cases. The follow-on patch will be one kernel release later as it is relatively high risk and it'll make bisections more clear if there are any problems. Patch 1 is a cosmetic patch to clarify when page->lru is storing buddy pages and when it is storing per-cpu pages. Patch 2 shrinks per_cpu_pages to make room for a spin lock. Strictly speaking this is not necessary but it avoids per_cpu_pages consuming another cache line. Patch 3 is a preparation patch to avoid code duplication. Patch 4 is a minor correction. Patch 5 uses a spin_lock to protect the per_cpu_pages contents while still relying on local_lock to prevent migration, stabilise the pcp lookup and prevent IRQ reentrancy. Patch 6 remote drains per-cpu pages directly instead of using a workqueue. Patch 7 uses a normal spinlock instead of local_lock for remote draining This patch (of 7): The page allocator uses page->lru for storing pages on either buddy or PCP lists. Create page->buddy_list and page->pcp_list as a union with page->lru. This is simply to clarify what type of list a page is on in the page allocator. No functional change intended. [minchan@kernel.org: fix page lru fields in macros] Link: https://lkml.kernel.org/r/20220624125423.6126-2-mgorman@techsingularity.net Signed-off-by: Mel Gorman Tested-by: Minchan Kim Acked-by: Minchan Kim Reviewed-by: Nicolas Saenz Julienne Acked-by: Vlastimil Babka Cc: Marcelo Tosatti Cc: Michal Hocko Cc: Hugh Dickins Cc: Yu Zhao Cc: Marek Szyprowski Signed-off-by: Andrew Morton --- include/linux/mm_types.h | 5 +++++ mm/page_alloc.c | 24 ++++++++++++------------ 2 files changed, 17 insertions(+), 12 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 5cc2cb396ca7..acbd8d03e01e 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -88,6 +88,7 @@ struct page { */ union { struct list_head lru; + /* Or, for the Unevictable "LRU list" slot */ struct { /* Always even, to negate PageTail */ @@ -95,6 +96,10 @@ struct page { /* Count page's or folio's mlocks */ unsigned int mlock_count; }; + + /* Or, free page */ + struct list_head buddy_list; + struct list_head pcp_list; }; /* See page-flags.h for PAGE_MAPPING_FLAGS */ struct address_space *mapping; diff --git a/mm/page_alloc.c b/mm/page_alloc.c index d7c72f43f1da..fedc6e15d238 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -793,7 +793,7 @@ static inline bool set_page_guard(struct zone *zone, struct page *page, return false; __SetPageGuard(page); - INIT_LIST_HEAD(&page->lru); + INIT_LIST_HEAD(&page->buddy_list); set_page_private(page, order); /* Guard pages are not available for any usage */ __mod_zone_freepage_state(zone, -(1 << order), migratetype); @@ -936,7 +936,7 @@ static inline void add_to_free_list(struct page *page, struct zone *zone, { struct free_area *area = &zone->free_area[order]; - list_add(&page->lru, &area->free_list[migratetype]); + list_add(&page->buddy_list, &area->free_list[migratetype]); area->nr_free++; } @@ -946,7 +946,7 @@ static inline void add_to_free_list_tail(struct page *page, struct zone *zone, { struct free_area *area = &zone->free_area[order]; - list_add_tail(&page->lru, &area->free_list[migratetype]); + list_add_tail(&page->buddy_list, &area->free_list[migratetype]); area->nr_free++; } @@ -960,7 +960,7 @@ static inline void move_to_free_list(struct page *page, struct zone *zone, { struct free_area *area = &zone->free_area[order]; - list_move_tail(&page->lru, &area->free_list[migratetype]); + list_move_tail(&page->buddy_list, &area->free_list[migratetype]); } static inline void del_page_from_free_list(struct page *page, struct zone *zone, @@ -970,7 +970,7 @@ static inline void del_page_from_free_list(struct page *page, struct zone *zone, if (page_reported(page)) __ClearPageReported(page); - list_del(&page->lru); + list_del(&page->buddy_list); __ClearPageBuddy(page); set_page_private(page, 0); zone->free_area[order].nr_free--; @@ -1508,11 +1508,11 @@ static void free_pcppages_bulk(struct zone *zone, int count, do { int mt; - page = list_last_entry(list, struct page, lru); + page = list_last_entry(list, struct page, pcp_list); mt = get_pcppage_migratetype(page); /* must delete to avoid corrupting pcp list */ - list_del(&page->lru); + list_del(&page->pcp_list); count -= nr_pages; pcp->count -= nr_pages; @@ -3072,7 +3072,7 @@ static int rmqueue_bulk(struct zone *zone, unsigned int order, * for IO devices that can merge IO requests if the physical * pages are ordered properly. */ - list_add_tail(&page->lru, list); + list_add_tail(&page->pcp_list, list); allocated++; if (is_migrate_cma(get_pcppage_migratetype(page))) __mod_zone_page_state(zone, NR_FREE_CMA_PAGES, @@ -3322,7 +3322,7 @@ void mark_free_pages(struct zone *zone) for_each_migratetype_order(order, t) { list_for_each_entry(page, - &zone->free_area[order].free_list[t], lru) { + &zone->free_area[order].free_list[t], buddy_list) { unsigned long i; pfn = page_to_pfn(page); @@ -3411,7 +3411,7 @@ static void free_unref_page_commit(struct page *page, int migratetype, __count_vm_event(PGFREE); pcp = this_cpu_ptr(zone->per_cpu_pageset); pindex = order_to_pindex(migratetype, order); - list_add(&page->lru, &pcp->lists[pindex]); + list_add(&page->pcp_list, &pcp->lists[pindex]); pcp->count += 1 << order; /* @@ -3674,8 +3674,8 @@ struct page *__rmqueue_pcplist(struct zone *zone, unsigned int order, return NULL; } - page = list_first_entry(list, struct page, lru); - list_del(&page->lru); + page = list_first_entry(list, struct page, pcp_list); + list_del(&page->pcp_list); pcp->count -= 1 << order; } while (check_new_pcp(page, order)); -- cgit v1.2.3 From b9ce5c6a577772e43a20779984074d51d8845d87 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Fri, 24 Jun 2022 13:54:18 +0100 Subject: mm/page_alloc: use only one PCP list for THP-sized allocations The per_cpu_pages is cache-aligned on a standard x86-64 distribution configuration but a later patch will add a new field which would push the structure into the next cache line. Use only one list to store THP-sized pages on the per-cpu list. This assumes that the vast majority of THP-sized allocations are GFP_MOVABLE but even if it was another type, it would not contribute to serious fragmentation that potentially causes a later THP allocation failure. Align per_cpu_pages on the cacheline boundary to ensure there is no false cache sharing. After this patch, the structure sizing is; struct per_cpu_pages { int count; /* 0 4 */ int high; /* 4 4 */ int batch; /* 8 4 */ short int free_factor; /* 12 2 */ short int expire; /* 14 2 */ struct list_head lists[13]; /* 16 208 */ /* size: 256, cachelines: 4, members: 6 */ /* padding: 32 */ } __attribute__((__aligned__(64))); Link: https://lkml.kernel.org/r/20220624125423.6126-3-mgorman@techsingularity.net Signed-off-by: Mel Gorman Tested-by: Minchan Kim Acked-by: Minchan Kim Acked-by: Vlastimil Babka Cc: Hugh Dickins Cc: Marcelo Tosatti Cc: Marek Szyprowski Cc: Michal Hocko Cc: Nicolas Saenz Julienne Cc: Yu Zhao Signed-off-by: Andrew Morton --- include/linux/mmzone.h | 11 +++++++---- mm/page_alloc.c | 4 ++-- 2 files changed, 9 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 735bf5b37949..ce15c27dddb5 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -355,15 +355,18 @@ enum zone_watermarks { }; /* - * One per migratetype for each PAGE_ALLOC_COSTLY_ORDER plus one additional - * for pageblock size for THP if configured. + * One per migratetype for each PAGE_ALLOC_COSTLY_ORDER. One additional list + * for THP which will usually be GFP_MOVABLE. Even if it is another type, + * it should not contribute to serious fragmentation causing THP allocation + * failures. */ #ifdef CONFIG_TRANSPARENT_HUGEPAGE #define NR_PCP_THP 1 #else #define NR_PCP_THP 0 #endif -#define NR_PCP_LISTS (MIGRATE_PCPTYPES * (PAGE_ALLOC_COSTLY_ORDER + 1 + NR_PCP_THP)) +#define NR_LOWORDER_PCP_LISTS (MIGRATE_PCPTYPES * (PAGE_ALLOC_COSTLY_ORDER + 1)) +#define NR_PCP_LISTS (NR_LOWORDER_PCP_LISTS + NR_PCP_THP) /* * Shift to encode migratetype and order in the same integer, with order @@ -389,7 +392,7 @@ struct per_cpu_pages { /* Lists of pages, one per migrate type stored on the pcp-lists */ struct list_head lists[NR_PCP_LISTS]; -}; +} ____cacheline_aligned_in_smp; struct per_cpu_zonestat { #ifdef CONFIG_SMP diff --git a/mm/page_alloc.c b/mm/page_alloc.c index fedc6e15d238..cb4565f338c6 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -653,7 +653,7 @@ static inline unsigned int order_to_pindex(int migratetype, int order) #ifdef CONFIG_TRANSPARENT_HUGEPAGE if (order > PAGE_ALLOC_COSTLY_ORDER) { VM_BUG_ON(order != pageblock_order); - base = PAGE_ALLOC_COSTLY_ORDER + 1; + return NR_LOWORDER_PCP_LISTS; } #else VM_BUG_ON(order > PAGE_ALLOC_COSTLY_ORDER); @@ -667,7 +667,7 @@ static inline int pindex_to_order(unsigned int pindex) int order = pindex / MIGRATE_PCPTYPES; #ifdef CONFIG_TRANSPARENT_HUGEPAGE - if (order > PAGE_ALLOC_COSTLY_ORDER) + if (pindex == NR_LOWORDER_PCP_LISTS) order = pageblock_order; #else VM_BUG_ON(order > PAGE_ALLOC_COSTLY_ORDER); -- cgit v1.2.3 From c9743f2053f42862a74424b273db9bdca68fb892 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Fri, 24 Jun 2022 13:54:21 +0100 Subject: mm/page_alloc: protect PCP lists with a spinlock Currently the PCP lists are protected by using local_lock_irqsave to prevent migration and IRQ reentrancy but this is inconvenient. Remote draining of the lists is impossible and a workqueue is required and every task allocation/free must disable then enable interrupts which is expensive. As preparation for dealing with both of those problems, protect the lists with a spinlock. The IRQ-unsafe version of the lock is used because IRQs are already disabled by local_lock_irqsave. spin_trylock is used in preparation for a time when local_lock could be used instead of lock_lock_irqsave. The per_cpu_pages still fits within the same number of cache lines after this patch relative to before the series. struct per_cpu_pages { spinlock_t lock; /* 0 4 */ int count; /* 4 4 */ int high; /* 8 4 */ int batch; /* 12 4 */ short int free_factor; /* 16 2 */ short int expire; /* 18 2 */ /* XXX 4 bytes hole, try to pack */ struct list_head lists[13]; /* 24 208 */ /* size: 256, cachelines: 4, members: 7 */ /* sum members: 228, holes: 1, sum holes: 4 */ /* padding: 24 */ } __attribute__((__aligned__(64))); There is overhead in the fast path due to acquiring the spinlock even though the spinlock is per-cpu and uncontended in the common case. Page Fault Test (PFT) running on a 1-socket reported the following results on a 1 socket machine. 5.19.0-rc3 5.19.0-rc3 vanilla mm-pcpspinirq-v5r16 Hmean faults/sec-1 869275.7381 ( 0.00%) 874597.5167 * 0.61%* Hmean faults/sec-3 2370266.6681 ( 0.00%) 2379802.0362 * 0.40%* Hmean faults/sec-5 2701099.7019 ( 0.00%) 2664889.7003 * -1.34%* Hmean faults/sec-7 3517170.9157 ( 0.00%) 3491122.8242 * -0.74%* Hmean faults/sec-8 3965729.6187 ( 0.00%) 3939727.0243 * -0.66%* There is a small hit in the number of faults per second but given that the results are more stable, it's borderline noise. Link: https://lkml.kernel.org/r/20220624125423.6126-6-mgorman@techsingularity.net Signed-off-by: Mel Gorman Cc: Hugh Dickins Cc: Marcelo Tosatti Cc: Marek Szyprowski Cc: Michal Hocko Cc: Minchan Kim Cc: Nicolas Saenz Julienne Cc: Vlastimil Babka Cc: Yu Zhao Signed-off-by: Andrew Morton --- include/linux/mmzone.h | 1 + mm/page_alloc.c | 118 ++++++++++++++++++++++++++++++++++++++++--------- 2 files changed, 98 insertions(+), 21 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index ce15c27dddb5..ed3ebe5e9e50 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -382,6 +382,7 @@ enum zone_watermarks { /* Fields and list protected by pagesets local_lock in page_alloc.c */ struct per_cpu_pages { + spinlock_t lock; /* Protects lists field */ int count; /* number of pages in the list */ int high; /* high watermark, emptying needed */ int batch; /* chunk size for buddy add/remove */ diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 2a137cd121e3..8a07916aa274 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -133,6 +133,20 @@ static DEFINE_PER_CPU(struct pagesets, pagesets) = { .lock = INIT_LOCAL_LOCK(lock), }; +#if defined(CONFIG_SMP) || defined(CONFIG_PREEMPT_RT) +/* + * On SMP, spin_trylock is sufficient protection. + * On PREEMPT_RT, spin_trylock is equivalent on both SMP and UP. + */ +#define pcp_trylock_prepare(flags) do { } while (0) +#define pcp_trylock_finish(flag) do { } while (0) +#else + +/* UP spin_trylock always succeeds so disable IRQs to prevent re-entrancy. */ +#define pcp_trylock_prepare(flags) local_irq_save(flags) +#define pcp_trylock_finish(flags) local_irq_restore(flags) +#endif + #ifdef CONFIG_USE_PERCPU_NUMA_NODE_ID DEFINE_PER_CPU(int, numa_node); EXPORT_PER_CPU_SYMBOL(numa_node); @@ -3101,15 +3115,22 @@ static int rmqueue_bulk(struct zone *zone, unsigned int order, */ void drain_zone_pages(struct zone *zone, struct per_cpu_pages *pcp) { - unsigned long flags; int to_drain, batch; - local_lock_irqsave(&pagesets.lock, flags); batch = READ_ONCE(pcp->batch); to_drain = min(pcp->count, batch); - if (to_drain > 0) + if (to_drain > 0) { + unsigned long flags; + + /* + * free_pcppages_bulk expects IRQs disabled for zone->lock + * so even though pcp->lock is not intended to be IRQ-safe, + * it's needed in this context. + */ + spin_lock_irqsave(&pcp->lock, flags); free_pcppages_bulk(zone, to_drain, pcp, 0); - local_unlock_irqrestore(&pagesets.lock, flags); + spin_unlock_irqrestore(&pcp->lock, flags); + } } #endif @@ -3122,16 +3143,17 @@ void drain_zone_pages(struct zone *zone, struct per_cpu_pages *pcp) */ static void drain_pages_zone(unsigned int cpu, struct zone *zone) { - unsigned long flags; struct per_cpu_pages *pcp; - local_lock_irqsave(&pagesets.lock, flags); - pcp = per_cpu_ptr(zone->per_cpu_pageset, cpu); - if (pcp->count) - free_pcppages_bulk(zone, pcp->count, pcp, 0); + if (pcp->count) { + unsigned long flags; - local_unlock_irqrestore(&pagesets.lock, flags); + /* See drain_zone_pages on why this is disabling IRQs */ + spin_lock_irqsave(&pcp->lock, flags); + free_pcppages_bulk(zone, pcp->count, pcp, 0); + spin_unlock_irqrestore(&pcp->lock, flags); + } } /* @@ -3399,17 +3421,15 @@ static int nr_pcp_high(struct per_cpu_pages *pcp, struct zone *zone, return min(READ_ONCE(pcp->batch) << 2, high); } -static void free_unref_page_commit(struct page *page, int migratetype, +static void free_unref_page_commit(struct zone *zone, struct per_cpu_pages *pcp, + struct page *page, int migratetype, unsigned int order) { - struct zone *zone = page_zone(page); - struct per_cpu_pages *pcp; int high; int pindex; bool free_high; __count_vm_event(PGFREE); - pcp = this_cpu_ptr(zone->per_cpu_pageset); pindex = order_to_pindex(migratetype, order); list_add(&page->pcp_list, &pcp->lists[pindex]); pcp->count += 1 << order; @@ -3436,6 +3456,9 @@ static void free_unref_page_commit(struct page *page, int migratetype, void free_unref_page(struct page *page, unsigned int order) { unsigned long flags; + unsigned long __maybe_unused UP_flags; + struct per_cpu_pages *pcp; + struct zone *zone; unsigned long pfn = page_to_pfn(page); int migratetype; @@ -3459,7 +3482,16 @@ void free_unref_page(struct page *page, unsigned int order) } local_lock_irqsave(&pagesets.lock, flags); - free_unref_page_commit(page, migratetype, order); + zone = page_zone(page); + pcp_trylock_prepare(UP_flags); + pcp = this_cpu_ptr(zone->per_cpu_pageset); + if (spin_trylock(&pcp->lock)) { + free_unref_page_commit(zone, pcp, page, migratetype, order); + spin_unlock(&pcp->lock); + } else { + free_one_page(zone, page, pfn, order, migratetype, FPI_NONE); + } + pcp_trylock_finish(UP_flags); local_unlock_irqrestore(&pagesets.lock, flags); } @@ -3469,6 +3501,8 @@ void free_unref_page(struct page *page, unsigned int order) void free_unref_page_list(struct list_head *list) { struct page *page, *next; + struct per_cpu_pages *pcp = NULL; + struct zone *locked_zone = NULL; unsigned long flags; int batch_count = 0; int migratetype; @@ -3495,6 +3529,17 @@ void free_unref_page_list(struct list_head *list) local_lock_irqsave(&pagesets.lock, flags); list_for_each_entry_safe(page, next, list, lru) { + struct zone *zone = page_zone(page); + + /* Different zone, different pcp lock. */ + if (zone != locked_zone) { + if (pcp) + spin_unlock(&pcp->lock); + locked_zone = zone; + pcp = this_cpu_ptr(zone->per_cpu_pageset); + spin_lock(&pcp->lock); + } + /* * Non-isolated types over MIGRATE_PCPTYPES get added * to the MIGRATE_MOVABLE pcp list. @@ -3504,18 +3549,24 @@ void free_unref_page_list(struct list_head *list) migratetype = MIGRATE_MOVABLE; trace_mm_page_free_batched(page); - free_unref_page_commit(page, migratetype, 0); + free_unref_page_commit(zone, pcp, page, migratetype, 0); /* * Guard against excessive IRQ disabled times when we get * a large list of pages to free. */ if (++batch_count == SWAP_CLUSTER_MAX) { + spin_unlock(&pcp->lock); local_unlock_irqrestore(&pagesets.lock, flags); batch_count = 0; local_lock_irqsave(&pagesets.lock, flags); + pcp = this_cpu_ptr(locked_zone->per_cpu_pageset); + spin_lock(&pcp->lock); } } + + if (pcp) + spin_unlock(&pcp->lock); local_unlock_irqrestore(&pagesets.lock, flags); } @@ -3729,18 +3780,31 @@ static struct page *rmqueue_pcplist(struct zone *preferred_zone, struct list_head *list; struct page *page; unsigned long flags; + unsigned long __maybe_unused UP_flags; local_lock_irqsave(&pagesets.lock, flags); + /* + * spin_trylock may fail due to a parallel drain. In the future, the + * trylock will also protect against IRQ reentrancy. + */ + pcp = this_cpu_ptr(zone->per_cpu_pageset); + pcp_trylock_prepare(UP_flags); + if (!spin_trylock(&pcp->lock)) { + pcp_trylock_finish(UP_flags); + return NULL; + } + /* * On allocation, reduce the number of pages that are batch freed. * See nr_pcp_free() where free_factor is increased for subsequent * frees. */ - pcp = this_cpu_ptr(zone->per_cpu_pageset); pcp->free_factor >>= 1; list = &pcp->lists[order_to_pindex(migratetype, order)]; page = __rmqueue_pcplist(zone, order, migratetype, alloc_flags, pcp, list); + spin_unlock(&pcp->lock); + pcp_trylock_finish(UP_flags); local_unlock_irqrestore(&pagesets.lock, flags); if (page) { __count_zid_vm_events(PGALLOC, page_zonenum(page), 1); @@ -3775,7 +3839,8 @@ struct page *rmqueue(struct zone *preferred_zone, migratetype != MIGRATE_MOVABLE) { page = rmqueue_pcplist(preferred_zone, zone, order, gfp_flags, migratetype, alloc_flags); - goto out; + if (likely(page)) + goto out; } } @@ -5252,6 +5317,7 @@ unsigned long __alloc_pages_bulk(gfp_t gfp, int preferred_nid, { struct page *page; unsigned long flags; + unsigned long __maybe_unused UP_flags; struct zone *zone; struct zoneref *z; struct per_cpu_pages *pcp; @@ -5332,11 +5398,15 @@ unsigned long __alloc_pages_bulk(gfp_t gfp, int preferred_nid, if (unlikely(!zone)) goto failed; - /* Attempt the batch allocation */ + /* Is a parallel drain in progress? */ local_lock_irqsave(&pagesets.lock, flags); + pcp_trylock_prepare(UP_flags); pcp = this_cpu_ptr(zone->per_cpu_pageset); - pcp_list = &pcp->lists[order_to_pindex(ac.migratetype, 0)]; + if (!spin_trylock(&pcp->lock)) + goto failed_irq; + /* Attempt the batch allocation */ + pcp_list = &pcp->lists[order_to_pindex(ac.migratetype, 0)]; while (nr_populated < nr_pages) { /* Skip existing pages */ @@ -5349,8 +5419,10 @@ unsigned long __alloc_pages_bulk(gfp_t gfp, int preferred_nid, pcp, pcp_list); if (unlikely(!page)) { /* Try and allocate at least one page */ - if (!nr_account) + if (!nr_account) { + spin_unlock(&pcp->lock); goto failed_irq; + } break; } nr_account++; @@ -5363,6 +5435,8 @@ unsigned long __alloc_pages_bulk(gfp_t gfp, int preferred_nid, nr_populated++; } + spin_unlock(&pcp->lock); + pcp_trylock_finish(UP_flags); local_unlock_irqrestore(&pagesets.lock, flags); __count_zid_vm_events(PGALLOC, zone_idx(zone), nr_account); @@ -5372,6 +5446,7 @@ out: return nr_populated; failed_irq: + pcp_trylock_finish(UP_flags); local_unlock_irqrestore(&pagesets.lock, flags); failed: @@ -7017,6 +7092,7 @@ static void per_cpu_pages_init(struct per_cpu_pages *pcp, struct per_cpu_zonesta memset(pcp, 0, sizeof(*pcp)); memset(pzstats, 0, sizeof(*pzstats)); + spin_lock_init(&pcp->lock); for (pindex = 0; pindex < NR_PCP_LISTS; pindex++) INIT_LIST_HEAD(&pcp->lists[pindex]); -- cgit v1.2.3 From 5b207678e39bab6194709606609ae9c2984ccaae Mon Sep 17 00:00:00 2001 From: Kalesh Singh Date: Thu, 23 Jun 2022 15:06:07 -0700 Subject: procfs: add 'path' to /proc//fdinfo/ MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In order to identify the type of memory a process has pinned through its open fds, add the file path to fdinfo output. This allows identifying memory types based on common prefixes: e.g. "/memfd...", "/dmabuf...", "/dev/ashmem...". To be cautious, only expose the paths for anonymous inodes, and this also avoids printing path names with strange characters. Access to /proc//fdinfo is governed by PTRACE_MODE_READ_FSCREDS the same as /proc//maps which also exposes the file path of mappings; so the security permissions for accessing path is consistent with that of /proc//maps. Link: https://lkml.kernel.org/r/20220623220613.3014268-3-kaleshsingh@google.com Signed-off-by: Kalesh Singh Cc: Alexey Dobriyan Cc: Al Viro Cc: Christian König Cc: Christoph Anton Mitterer Cc: Christoph Hellwig Cc: Colin Cross Cc: David Laight Cc: Ioannis Ilkos Cc: Johannes Weiner Cc: Jonathan Corbet Cc: Paul Gortmaker Cc: Randy Dunlap Cc: Stephen Brennan Cc: Sumit Semwal Cc: Suren Baghdasaryan Cc: T.J. Mercier Signed-off-by: Andrew Morton --- Documentation/filesystems/proc.rst | 10 ++++++++++ fs/libfs.c | 9 +++++++++ fs/proc/fd.c | 13 +++++++++++-- include/linux/fs.h | 1 + 4 files changed, 31 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/Documentation/filesystems/proc.rst b/Documentation/filesystems/proc.rst index 640fe47586e3..47e95dbc820d 100644 --- a/Documentation/filesystems/proc.rst +++ b/Documentation/filesystems/proc.rst @@ -1912,6 +1912,9 @@ All locks associated with a file descriptor are shown in its fdinfo too:: lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF +Files with anonymous inodes have an additional 'path' field which represents +the anonymous file path. + The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags pair provide additional information particular to the objects they represent. @@ -1925,6 +1928,7 @@ Eventfd files mnt_id: 9 ino: 63107 size: 0 + path: anon_inode:[eventfd] eventfd-count: 5a where 'eventfd-count' is hex value of a counter. @@ -1939,6 +1943,7 @@ Signalfd files mnt_id: 9 ino: 63107 size: 0 + path: anon_inode:[signalfd] sigmask: 0000000000000200 where 'sigmask' is hex value of the signal mask associated @@ -1954,6 +1959,7 @@ Epoll files mnt_id: 9 ino: 63107 size: 0 + path: anon_inode:[eventpoll] tfd: 5 events: 1d data: ffffffffffffffff pos:0 ino:61af sdev:7 where 'tfd' is a target file descriptor number in decimal form, @@ -1973,6 +1979,7 @@ For inotify files the format is the following:: mnt_id: 9 ino: 63107 size: 0 + path: anon_inode:inotify inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d where 'wd' is a watch descriptor in decimal form, i.e. a target file @@ -1997,6 +2004,7 @@ For fanotify files the format is:: mnt_id: 9 ino: 63107 size: 0 + path: anon_inode:[fanotify] fanotify flags:10 event-flags:0 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4 @@ -2023,6 +2031,7 @@ Timerfd files mnt_id: 9 ino: 63107 size: 0 + path: anon_inode:[timerfd] clockid: 0 ticks: 0 settime flags: 01 @@ -2047,6 +2056,7 @@ DMA Buffer files mnt_id: 9 ino: 63107 size: 32768 + path: /dmabuf: count: 2 exp_name: system-heap diff --git a/fs/libfs.c b/fs/libfs.c index 31b0ddf01c31..6911749b4da7 100644 --- a/fs/libfs.c +++ b/fs/libfs.c @@ -1217,6 +1217,15 @@ void kfree_link(void *p) } EXPORT_SYMBOL(kfree_link); +static const struct address_space_operations anon_aops = { + .dirty_folio = noop_dirty_folio, +}; + +bool is_anon_inode(struct inode *inode) +{ + return inode->i_mapping->a_ops == &anon_aops; +} + struct inode *alloc_anon_inode(struct super_block *s) { static const struct address_space_operations anon_aops = { diff --git a/fs/proc/fd.c b/fs/proc/fd.c index 464bc3f55759..5bac79a2fa51 100644 --- a/fs/proc/fd.c +++ b/fs/proc/fd.c @@ -23,6 +23,7 @@ static int seq_show(struct seq_file *m, void *v) struct files_struct *files = NULL; int f_flags = 0, ret = -ENOENT; struct file *file = NULL; + struct inode *inode = NULL; struct task_struct *task; task = get_proc_task(m->private); @@ -54,11 +55,19 @@ static int seq_show(struct seq_file *m, void *v) if (ret) return ret; + inode = file_inode(file); + seq_printf(m, "pos:\t%lli\n", (long long)file->f_pos); seq_printf(m, "flags:\t0%o\n", f_flags); seq_printf(m, "mnt_id:\t%i\n", real_mount(file->f_path.mnt)->mnt_id); - seq_printf(m, "ino:\t%lu\n", file_inode(file)->i_ino); - seq_printf(m, "size:\t%lli\n", (long long)file_inode(file)->i_size); + seq_printf(m, "ino:\t%lu\n", inode->i_ino); + seq_printf(m, "size:\t%lli\n", (long long)inode->i_size); + + if (is_anon_inode(inode)) { + seq_puts(m, "path:\t"); + seq_file_path(m, file, "\n"); + seq_putc(m, '\n'); + } /* show_fd_locks() never deferences files so a stale value is safe */ show_fd_locks(m, file, files); diff --git a/include/linux/fs.h b/include/linux/fs.h index 134e9d7ad5d6..7132c6f955c1 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -3115,6 +3115,7 @@ extern void page_put_link(void *); extern int page_symlink(struct inode *inode, const char *symname, int len); extern const struct inode_operations page_symlink_inode_operations; extern void kfree_link(void *); +extern bool is_anon_inode(struct inode *inode); void generic_fillattr(struct user_namespace *, struct inode *, struct kstat *); void generic_fill_statx_attr(struct inode *inode, struct kstat *stat); extern int vfs_getattr_nosec(const struct path *, struct kstat *, u32, unsigned int); -- cgit v1.2.3 From d2547da81f09d0dc2375a4c32429c31718626c10 Mon Sep 17 00:00:00 2001 From: Qi Zheng Date: Sun, 26 Jun 2022 22:57:17 +0800 Subject: mm: hugetlb: kill set_huge_swap_pte_at() Commit e5251fd43007 ("mm/hugetlb: introduce set_huge_swap_pte_at() helper") add set_huge_swap_pte_at() to handle swap entries on architectures that support hugepages consisting of contiguous ptes. And currently the set_huge_swap_pte_at() is only overridden by arm64. set_huge_swap_pte_at() provide a sz parameter to help determine the number of entries to be updated. But in fact, all hugetlb swap entries contain pfn information, so we can find the corresponding folio through the pfn recorded in the swap entry, then the folio_size() is the number of entries that need to be updated. And considering that users will easily cause bugs by ignoring the difference between set_huge_swap_pte_at() and set_huge_pte_at(). Let's handle swap entries in set_huge_pte_at() and remove the set_huge_swap_pte_at(), then we can call set_huge_pte_at() anywhere, which simplifies our coding. Link: https://lkml.kernel.org/r/20220626145717.53572-1-zhengqi.arch@bytedance.com Signed-off-by: Qi Zheng Acked-by: Muchun Song Cc: Mike Kravetz Cc: Catalin Marinas Cc: Will Deacon Signed-off-by: Andrew Morton --- arch/arm64/include/asm/hugetlb.h | 3 --- arch/arm64/mm/hugetlbpage.c | 34 +++++++++++++++++----------------- include/linux/hugetlb.h | 13 ------------- mm/hugetlb.c | 8 +++----- mm/rmap.c | 11 +++-------- 5 files changed, 23 insertions(+), 46 deletions(-) (limited to 'include/linux') diff --git a/arch/arm64/include/asm/hugetlb.h b/arch/arm64/include/asm/hugetlb.h index 1fd2846dbefe..d20f5da2d76f 100644 --- a/arch/arm64/include/asm/hugetlb.h +++ b/arch/arm64/include/asm/hugetlb.h @@ -46,9 +46,6 @@ extern void huge_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep, unsigned long sz); #define __HAVE_ARCH_HUGE_PTEP_GET extern pte_t huge_ptep_get(pte_t *ptep); -extern void set_huge_swap_pte_at(struct mm_struct *mm, unsigned long addr, - pte_t *ptep, pte_t pte, unsigned long sz); -#define set_huge_swap_pte_at set_huge_swap_pte_at void __init arm64_hugetlb_cma_reserve(void); diff --git a/arch/arm64/mm/hugetlbpage.c b/arch/arm64/mm/hugetlbpage.c index c9e076683e5d..58b89b9d13e0 100644 --- a/arch/arm64/mm/hugetlbpage.c +++ b/arch/arm64/mm/hugetlbpage.c @@ -238,6 +238,13 @@ static void clear_flush(struct mm_struct *mm, flush_tlb_range(&vma, saddr, addr); } +static inline struct folio *hugetlb_swap_entry_to_folio(swp_entry_t entry) +{ + VM_BUG_ON(!is_migration_entry(entry) && !is_hwpoison_entry(entry)); + + return page_folio(pfn_to_page(swp_offset(entry))); +} + void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, pte_t pte) { @@ -247,11 +254,16 @@ void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, unsigned long pfn, dpfn; pgprot_t hugeprot; - /* - * Code needs to be expanded to handle huge swap and migration - * entries. Needed for HUGETLB and MEMORY_FAILURE. - */ - WARN_ON(!pte_present(pte)); + if (!pte_present(pte)) { + struct folio *folio; + + folio = hugetlb_swap_entry_to_folio(pte_to_swp_entry(pte)); + ncontig = num_contig_ptes(folio_size(folio), &pgsize); + + for (i = 0; i < ncontig; i++, ptep++) + set_pte_at(mm, addr, ptep, pte); + return; + } if (!pte_cont(pte)) { set_pte_at(mm, addr, ptep, pte); @@ -269,18 +281,6 @@ void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, set_pte_at(mm, addr, ptep, pfn_pte(pfn, hugeprot)); } -void set_huge_swap_pte_at(struct mm_struct *mm, unsigned long addr, - pte_t *ptep, pte_t pte, unsigned long sz) -{ - int i, ncontig; - size_t pgsize; - - ncontig = num_contig_ptes(sz, &pgsize); - - for (i = 0; i < ncontig; i++, ptep++) - set_pte(ptep, pte); -} - pte_t *huge_pte_alloc(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, unsigned long sz) { diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index 05c5317df380..75ee739d815b 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -904,14 +904,6 @@ static inline void hugetlb_count_sub(long l, struct mm_struct *mm) atomic_long_sub(l, &mm->hugetlb_usage); } -#ifndef set_huge_swap_pte_at -static inline void set_huge_swap_pte_at(struct mm_struct *mm, unsigned long addr, - pte_t *ptep, pte_t pte, unsigned long sz) -{ - set_huge_pte_at(mm, addr, ptep, pte); -} -#endif - #ifndef huge_ptep_modify_prot_start #define huge_ptep_modify_prot_start huge_ptep_modify_prot_start static inline pte_t huge_ptep_modify_prot_start(struct vm_area_struct *vma, @@ -1095,11 +1087,6 @@ static inline void hugetlb_count_sub(long l, struct mm_struct *mm) { } -static inline void set_huge_swap_pte_at(struct mm_struct *mm, unsigned long addr, - pte_t *ptep, pte_t pte, unsigned long sz) -{ -} - static inline pte_t huge_ptep_clear_flush(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep) { diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 0d802c67ae8b..24778a483511 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -4804,12 +4804,11 @@ again: entry = swp_entry_to_pte(swp_entry); if (userfaultfd_wp(src_vma) && uffd_wp) entry = huge_pte_mkuffd_wp(entry); - set_huge_swap_pte_at(src, addr, src_pte, - entry, sz); + set_huge_pte_at(src, addr, src_pte, entry); } if (!userfaultfd_wp(dst_vma) && uffd_wp) entry = huge_pte_clear_uffd_wp(entry); - set_huge_swap_pte_at(dst, addr, dst_pte, entry, sz); + set_huge_pte_at(dst, addr, dst_pte, entry); } else if (unlikely(is_pte_marker(entry))) { /* * We copy the pte marker only if the dst vma has @@ -6360,8 +6359,7 @@ unsigned long hugetlb_change_protection(struct vm_area_struct *vma, newpte = pte_swp_mkuffd_wp(newpte); else if (uffd_wp_resolve) newpte = pte_swp_clear_uffd_wp(newpte); - set_huge_swap_pte_at(mm, address, ptep, - newpte, psize); + set_huge_pte_at(mm, address, ptep, newpte); pages++; } spin_unlock(ptl); diff --git a/mm/rmap.c b/mm/rmap.c index 09ee423d59d1..edc06c52bc82 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -1618,9 +1618,7 @@ static bool try_to_unmap_one(struct folio *folio, struct vm_area_struct *vma, pteval = swp_entry_to_pte(make_hwpoison_entry(subpage)); if (folio_test_hugetlb(folio)) { hugetlb_count_sub(folio_nr_pages(folio), mm); - set_huge_swap_pte_at(mm, address, - pvmw.pte, pteval, - vma_mmu_pagesize(vma)); + set_huge_pte_at(mm, address, pvmw.pte, pteval); } else { dec_mm_counter(mm, mm_counter(&folio->page)); set_pte_at(mm, address, pvmw.pte, pteval); @@ -2011,9 +2009,7 @@ static bool try_to_migrate_one(struct folio *folio, struct vm_area_struct *vma, pteval = swp_entry_to_pte(make_hwpoison_entry(subpage)); if (folio_test_hugetlb(folio)) { hugetlb_count_sub(folio_nr_pages(folio), mm); - set_huge_swap_pte_at(mm, address, - pvmw.pte, pteval, - vma_mmu_pagesize(vma)); + set_huge_pte_at(mm, address, pvmw.pte, pteval); } else { dec_mm_counter(mm, mm_counter(&folio->page)); set_pte_at(mm, address, pvmw.pte, pteval); @@ -2081,8 +2077,7 @@ static bool try_to_migrate_one(struct folio *folio, struct vm_area_struct *vma, if (pte_uffd_wp(pteval)) swp_pte = pte_swp_mkuffd_wp(swp_pte); if (folio_test_hugetlb(folio)) - set_huge_swap_pte_at(mm, address, pvmw.pte, - swp_pte, vma_mmu_pagesize(vma)); + set_huge_pte_at(mm, address, pvmw.pte, swp_pte); else set_pte_at(mm, address, pvmw.pte, swp_pte); trace_set_migration_pte(address, pte_val(swp_pte), -- cgit v1.2.3 From c855ba8ed7befc7b7aecb17bde4770f46bb7e5e8 Mon Sep 17 00:00:00 2001 From: Miaohe Lin Date: Sat, 25 Jun 2022 17:28:16 +0800 Subject: mm/khugepaged: try to free transhuge swapcache when possible Transhuge swapcaches won't be freed in __collapse_huge_page_copy(). It's because release_pte_page() is not called for these pages and thus free_page_and_swap_cache can't grab the page lock. These pages won't be freed from swap cache even if we are the only user until next time reclaim. It shouldn't hurt indeed, but we could try to free these pages to save more memory for system. Link: https://lkml.kernel.org/r/20220625092816.4856-8-linmiaohe@huawei.com Signed-off-by: Miaohe Lin Cc: Alistair Popple Cc: Andrea Arcangeli Cc: David Hildenbrand Cc: David Howells Cc: Matthew Wilcox (Oracle) Cc: NeilBrown Cc: Peter Xu Cc: Suren Baghdasaryan Cc: Vlastimil Babka Cc: Yang Shi Cc: Zach O'Keefe Signed-off-by: Andrew Morton --- include/linux/swap.h | 5 +++++ mm/khugepaged.c | 7 ++++++- mm/swap.h | 5 ----- 3 files changed, 11 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/swap.h b/include/linux/swap.h index 95a5b7aa1ae9..6d11c51b2b62 100644 --- a/include/linux/swap.h +++ b/include/linux/swap.h @@ -455,6 +455,7 @@ static inline unsigned long total_swapcache_pages(void) return global_node_page_state(NR_SWAPCACHE); } +extern void free_swap_cache(struct page *page); extern void free_page_and_swap_cache(struct page *); extern void free_pages_and_swap_cache(struct page **, int); /* linux/mm/swapfile.c */ @@ -539,6 +540,10 @@ static inline void put_swap_device(struct swap_info_struct *si) /* used to sanity check ptes in zap_pte_range when CONFIG_SWAP=0 */ #define free_swap_and_cache(e) is_pfn_swap_entry(e) +static inline void free_swap_cache(struct page *page) +{ +} + static inline int add_swap_count_continuation(swp_entry_t swp, gfp_t gfp_mask) { return 0; diff --git a/mm/khugepaged.c b/mm/khugepaged.c index 6cb82a299eb2..cfe231c5958f 100644 --- a/mm/khugepaged.c +++ b/mm/khugepaged.c @@ -716,7 +716,12 @@ static void __collapse_huge_page_copy(pte_t *pte, struct page *page, list_for_each_entry_safe(src_page, tmp, compound_pagelist, lru) { list_del(&src_page->lru); - release_pte_page(src_page); + mod_node_page_state(page_pgdat(src_page), + NR_ISOLATED_ANON + page_is_file_lru(src_page), + -compound_nr(src_page)); + unlock_page(src_page); + free_swap_cache(src_page); + putback_lru_page(src_page); } } diff --git a/mm/swap.h b/mm/swap.h index fa0816af4712..17936e068c1c 100644 --- a/mm/swap.h +++ b/mm/swap.h @@ -41,7 +41,6 @@ void __delete_from_swap_cache(struct folio *folio, void delete_from_swap_cache(struct folio *folio); void clear_shadow_from_swap_cache(int type, unsigned long begin, unsigned long end); -void free_swap_cache(struct page *page); struct page *lookup_swap_cache(swp_entry_t entry, struct vm_area_struct *vma, unsigned long addr); @@ -81,10 +80,6 @@ static inline struct address_space *swap_address_space(swp_entry_t entry) return NULL; } -static inline void free_swap_cache(struct page *page) -{ -} - static inline void show_swap_cache_info(void) { } -- cgit v1.2.3 From 68fa702a5e67b712dd492415787ee84a9f64bd49 Mon Sep 17 00:00:00 2001 From: wuchi Date: Sun, 19 Jun 2022 15:46:41 +0800 Subject: net, lib/once: remove {net_}get_random_once_wait macro DO_ONCE(func, ...) will call func with spinlock which acquired by spin_lock_irqsave in __do_once_start. But the get_random_once_wait will sleep in get_random_bytes_wait -> wait_for_random_bytes. Fortunately, there is no place to use {net_}get_random_once_wait, so we could remove them simply. Link: https://lkml.kernel.org/r/20220619074641.40916-1-wuchi.zero@gmail.com Signed-off-by: wuchi Acked-by: Jakub Kicinski Cc: David S. Miller Cc: Eric Dumazet Cc: Paolo Abeni Signed-off-by: Andrew Morton --- include/linux/net.h | 2 -- include/linux/once.h | 2 -- 2 files changed, 4 deletions(-) (limited to 'include/linux') diff --git a/include/linux/net.h b/include/linux/net.h index 12093f4db50c..8613772a1f58 100644 --- a/include/linux/net.h +++ b/include/linux/net.h @@ -303,8 +303,6 @@ do { \ #define net_get_random_once(buf, nbytes) \ get_random_once((buf), (nbytes)) -#define net_get_random_once_wait(buf, nbytes) \ - get_random_once_wait((buf), (nbytes)) /* * E.g. XFS meta- & log-data is in slab pages, or bcache meta diff --git a/include/linux/once.h b/include/linux/once.h index f54523052bbc..b14d8b309d52 100644 --- a/include/linux/once.h +++ b/include/linux/once.h @@ -54,7 +54,5 @@ void __do_once_done(bool *done, struct static_key_true *once_key, #define get_random_once(buf, nbytes) \ DO_ONCE(get_random_bytes, (buf), (nbytes)) -#define get_random_once_wait(buf, nbytes) \ - DO_ONCE(get_random_bytes_wait, (buf), (nbytes)) \ #endif /* _LINUX_ONCE_H */ -- cgit v1.2.3 From 76a40218d6dc21ee2a91ecf007295349fc381a3e Mon Sep 17 00:00:00 2001 From: Uros Bizjak Date: Fri, 24 Jun 2022 16:14:12 +0200 Subject: compiler-gcc.h: remove ancient workaround for gcc PR 58670 The workaround for 'asm goto' miscompilation introduces a compiler barrier quirk that inhibits many useful compiler optimizations. For example, __try_cmpxchg_user compiles to: 11375: 41 8b 4d 00 mov 0x0(%r13),%ecx 11379: 41 8b 02 mov (%r10),%eax 1137c: f0 0f b1 0a lock cmpxchg %ecx,(%rdx) 11380: 0f 94 c2 sete %dl 11383: 84 d2 test %dl,%dl 11385: 75 c4 jne 1134b <...> 11387: 41 89 02 mov %eax,(%r10) where the barrier inhibits flags propagation from asm when compiled with gcc-12. When the mentioned quirk is removed, the following code is generated: 11553: 41 8b 4d 00 mov 0x0(%r13),%ecx 11557: 41 8b 02 mov (%r10),%eax 1155a: f0 0f b1 0a lock cmpxchg %ecx,(%rdx) 1155e: 74 c9 je 11529 <...> 11560: 41 89 02 mov %eax,(%r10) The refered compiler bug: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58670 was fixed for gcc-4.8.2. Current minimum required version of GCC is version 5.1 which has the above 'asm goto' miscompilation fixed, so remove the workaround. Link: https://lkml.kernel.org/r/20220624141412.72274-1-ubizjak@gmail.com Signed-off-by: Uros Bizjak Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: Thomas Gleixner Signed-off-by: Andrew Morton --- include/linux/compiler-gcc.h | 11 ----------- 1 file changed, 11 deletions(-) (limited to 'include/linux') diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h index a0c55eeaeaf1..9b157b71036f 100644 --- a/include/linux/compiler-gcc.h +++ b/include/linux/compiler-gcc.h @@ -66,17 +66,6 @@ __builtin_unreachable(); \ } while (0) -/* - * GCC 'asm goto' miscompiles certain code sequences: - * - * http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58670 - * - * Work it around via a compiler barrier quirk suggested by Jakub Jelinek. - * - * (asm goto is automatically volatile - the naming reflects this.) - */ -#define asm_volatile_goto(x...) do { asm goto(x); asm (""); } while (0) - #if defined(CONFIG_ARCH_USE_BUILTIN_BSWAP) #define __HAVE_BUILTIN_BSWAP32__ #define __HAVE_BUILTIN_BSWAP64__ -- cgit v1.2.3 From 931c199904c4db2267bc08582e597dfc74ce7223 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 24 Jun 2022 08:30:04 +0300 Subject: kfifo: fix kfifo_to_user() return type The kfifo_to_user() macro is supposed to return zero for success or negative error codes. Unfortunately, there is a signedness bug so it returns unsigned int. This only affects callers which try to save the result in ssize_t and as far as I can see the only place which does that is line6_hwdep_read(). TL;DR: s/_uint/_int/. Link: https://lkml.kernel.org/r/YrVL3OJVLlNhIMFs@kili Fixes: 144ecf310eb5 ("kfifo: fix kfifo_alloc() to return a signed int value") Signed-off-by: Dan Carpenter Cc: Stefani Seibold Cc: Randy Dunlap Signed-off-by: Andrew Morton --- include/linux/kfifo.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/kfifo.h b/include/linux/kfifo.h index 86249476b57f..0b35a41440ff 100644 --- a/include/linux/kfifo.h +++ b/include/linux/kfifo.h @@ -688,7 +688,7 @@ __kfifo_uint_must_check_helper( \ * writer, you don't need extra locking to use these macro. */ #define kfifo_to_user(fifo, to, len, copied) \ -__kfifo_uint_must_check_helper( \ +__kfifo_int_must_check_helper( \ ({ \ typeof((fifo) + 1) __tmp = (fifo); \ void __user *__to = (to); \ -- cgit v1.2.3