From b9a49520679e98700d3d89689cc91c08a1c88c1d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sun, 19 Jan 2025 00:55:32 +0100 Subject: rcuref: Plug slowpath race in rcuref_put() Kernel test robot reported an "imbalanced put" in the rcuref_put() slow path, which turned out to be a false positive. Consider the following race: ref = 0 (via rcuref_init(ref, 1)) T1 T2 rcuref_put(ref) -> atomic_add_negative_release(-1, ref) # ref -> 0xffffffff -> rcuref_put_slowpath(ref) rcuref_get(ref) -> atomic_add_negative_relaxed(1, &ref->refcnt) -> return true; # ref -> 0 rcuref_put(ref) -> atomic_add_negative_release(-1, ref) # ref -> 0xffffffff -> rcuref_put_slowpath() -> cnt = atomic_read(&ref->refcnt); # cnt -> 0xffffffff / RCUREF_NOREF -> atomic_try_cmpxchg_release(&ref->refcnt, &cnt, RCUREF_DEAD)) # ref -> 0xe0000000 / RCUREF_DEAD -> return true -> cnt = atomic_read(&ref->refcnt); # cnt -> 0xe0000000 / RCUREF_DEAD -> if (cnt > RCUREF_RELEASED) # 0xe0000000 > 0xc0000000 -> WARN_ONCE(cnt >= RCUREF_RELEASED, "rcuref - imbalanced put()") The problem is the additional read in the slow path (after it decremented to RCUREF_NOREF) which can happen after the counter has been marked RCUREF_DEAD. Prevent this by reusing the return value of the decrement. Now every "final" put uses RCUREF_NOREF in the slow path and attempts the final cmpxchg() to RCUREF_DEAD. [ bigeasy: Add changelog ] Fixes: ee1ee6db07795 ("atomics: Provide rcuref - scalable reference counting") Reported-by: kernel test robot Debugged-by: Sebastian Andrzej Siewior Signed-off-by: Thomas Gleixner Signed-off-by: Sebastian Andrzej Siewior Signed-off-by: Thomas Gleixner Reviewed-by: Sebastian Andrzej Siewior Cc: stable@vger.kernel.org Closes: https://lore.kernel.org/oe-lkp/202412311453.9d7636a2-lkp@intel.com --- include/linux/rcuref.h | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/include/linux/rcuref.h b/include/linux/rcuref.h index 2c8bfd0f1b6b..6322d8c1c6b4 100644 --- a/include/linux/rcuref.h +++ b/include/linux/rcuref.h @@ -71,27 +71,30 @@ static inline __must_check bool rcuref_get(rcuref_t *ref) return rcuref_get_slowpath(ref); } -extern __must_check bool rcuref_put_slowpath(rcuref_t *ref); +extern __must_check bool rcuref_put_slowpath(rcuref_t *ref, unsigned int cnt); /* * Internal helper. Do not invoke directly. */ static __always_inline __must_check bool __rcuref_put(rcuref_t *ref) { + int cnt; + RCU_LOCKDEP_WARN(!rcu_read_lock_held() && preemptible(), "suspicious rcuref_put_rcusafe() usage"); /* * Unconditionally decrease the reference count. The saturation and * dead zones provide enough tolerance for this. */ - if (likely(!atomic_add_negative_release(-1, &ref->refcnt))) + cnt = atomic_sub_return_release(1, &ref->refcnt); + if (likely(cnt >= 0)) return false; /* * Handle the last reference drop and cases inside the saturation * and dead zones. */ - return rcuref_put_slowpath(ref); + return rcuref_put_slowpath(ref, cnt); } /** -- cgit v1.2.3 From 889c57066ceee5e9172232da0608a8ac053bb6e5 Mon Sep 17 00:00:00 2001 From: Ming Lei Date: Tue, 25 Feb 2025 10:21:41 +0800 Subject: block: make segment size limit workable for > 4K PAGE_SIZE Using PAGE_SIZE as a minimum expected DMA segment size in consideration of devices which have a max DMA segment size of < 64k when used on 64k PAGE_SIZE systems leads to devices not being able to probe such as eMMC and Exynos UFS controller [0] [1] you can end up with a probe failure as follows: WARNING: CPU: 2 PID: 397 at block/blk-settings.c:339 blk_validate_limits+0x364/0x3c0 Ensure we use min(max_seg_size, seg_boundary_mask + 1) as the new min segment size when max segment size is < PAGE_SIZE for 16k and 64k base page size systems. If anyone need to backport this patch, the following commits are depended: commit 6aeb4f836480 ("block: remove bio_add_pc_page") commit 02ee5d69e3ba ("block: remove blk_rq_bio_prep") commit b7175e24d6ac ("block: add a dma mapping iterator") Link: https://lore.kernel.org/linux-block/20230612203314.17820-1-bvanassche@acm.org/ # [0] Link: https://lore.kernel.org/linux-block/1d55e942-5150-de4c-3a02-c3d066f87028@acm.org/ # [1] Cc: Yi Zhang Cc: John Garry Cc: Keith Busch Tested-by: Paul Bunyan Reviewed-by: Daniel Gomez Reviewed-by: Luis Chamberlain Reviewed-by: Bart Van Assche Signed-off-by: Ming Lei Link: https://lore.kernel.org/r/20250225022141.2154581-1-ming.lei@redhat.com Signed-off-by: Jens Axboe --- block/blk-merge.c | 2 +- block/blk-settings.c | 14 +++++++++++--- block/blk.h | 9 +++++++-- include/linux/blkdev.h | 1 + 4 files changed, 20 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/block/blk-merge.c b/block/blk-merge.c index c7c85e10cf9c..1d1589c35297 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -329,7 +329,7 @@ int bio_split_rw_at(struct bio *bio, const struct queue_limits *lim, if (nsegs < lim->max_segments && bytes + bv.bv_len <= max_bytes && - bv.bv_offset + bv.bv_len <= PAGE_SIZE) { + bv.bv_offset + bv.bv_len <= lim->min_segment_size) { nsegs++; bytes += bv.bv_len; } else { diff --git a/block/blk-settings.c b/block/blk-settings.c index c44dadc35e1e..b9c6f0ec1c49 100644 --- a/block/blk-settings.c +++ b/block/blk-settings.c @@ -246,6 +246,7 @@ int blk_validate_limits(struct queue_limits *lim) { unsigned int max_hw_sectors; unsigned int logical_block_sectors; + unsigned long seg_size; int err; /* @@ -303,7 +304,7 @@ int blk_validate_limits(struct queue_limits *lim) max_hw_sectors = min_not_zero(lim->max_hw_sectors, lim->max_dev_sectors); if (lim->max_user_sectors) { - if (lim->max_user_sectors < PAGE_SIZE / SECTOR_SIZE) + if (lim->max_user_sectors < BLK_MIN_SEGMENT_SIZE / SECTOR_SIZE) return -EINVAL; lim->max_sectors = min(max_hw_sectors, lim->max_user_sectors); } else if (lim->io_opt > (BLK_DEF_MAX_SECTORS_CAP << SECTOR_SHIFT)) { @@ -341,7 +342,7 @@ int blk_validate_limits(struct queue_limits *lim) */ if (!lim->seg_boundary_mask) lim->seg_boundary_mask = BLK_SEG_BOUNDARY_MASK; - if (WARN_ON_ONCE(lim->seg_boundary_mask < PAGE_SIZE - 1)) + if (WARN_ON_ONCE(lim->seg_boundary_mask < BLK_MIN_SEGMENT_SIZE - 1)) return -EINVAL; /* @@ -362,10 +363,17 @@ int blk_validate_limits(struct queue_limits *lim) */ if (!lim->max_segment_size) lim->max_segment_size = BLK_MAX_SEGMENT_SIZE; - if (WARN_ON_ONCE(lim->max_segment_size < PAGE_SIZE)) + if (WARN_ON_ONCE(lim->max_segment_size < BLK_MIN_SEGMENT_SIZE)) return -EINVAL; } + /* setup min segment size for building new segment in fast path */ + if (lim->seg_boundary_mask > lim->max_segment_size - 1) + seg_size = lim->max_segment_size; + else + seg_size = lim->seg_boundary_mask + 1; + lim->min_segment_size = min_t(unsigned int, seg_size, PAGE_SIZE); + /* * We require drivers to at least do logical block aligned I/O, but * historically could not check for that due to the separate calls diff --git a/block/blk.h b/block/blk.h index 90fa5f28ccab..9cf9a0099416 100644 --- a/block/blk.h +++ b/block/blk.h @@ -14,6 +14,7 @@ struct elevator_type; #define BLK_DEV_MAX_SECTORS (LLONG_MAX >> 9) +#define BLK_MIN_SEGMENT_SIZE 4096 /* Max future timer expiry for timeouts */ #define BLK_MAX_TIMEOUT (5 * HZ) @@ -358,8 +359,12 @@ struct bio *bio_split_zone_append(struct bio *bio, static inline bool bio_may_need_split(struct bio *bio, const struct queue_limits *lim) { - return lim->chunk_sectors || bio->bi_vcnt != 1 || - bio->bi_io_vec->bv_len + bio->bi_io_vec->bv_offset > PAGE_SIZE; + if (lim->chunk_sectors) + return true; + if (bio->bi_vcnt != 1) + return true; + return bio->bi_io_vec->bv_len + bio->bi_io_vec->bv_offset > + lim->min_segment_size; } /** diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 248416ecd01c..58ff5aca83b6 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -367,6 +367,7 @@ struct queue_limits { unsigned int max_sectors; unsigned int max_user_sectors; unsigned int max_segment_size; + unsigned int min_segment_size; unsigned int physical_block_size; unsigned int logical_block_size; unsigned int alignment_offset; -- cgit v1.2.3 From 68f3ea7ee199ef77551e090dfef5a49046ea8443 Mon Sep 17 00:00:00 2001 From: Ard Biesheuvel Date: Fri, 21 Feb 2025 14:57:06 +0100 Subject: vmlinux.lds: Ensure that const vars with relocations are mapped R/O In the kernel, there are architectures (x86, arm64) that perform boot-time relocation (for KASLR) without relying on PIE codegen. In this case, all const global objects are emitted into .rodata, including const objects with fields that will be fixed up by the boot-time relocation code. This implies that .rodata (and .text in some cases) need to be writable at boot, but they will usually be mapped read-only as soon as the boot completes. When using PIE codegen, the compiler will emit const global objects into .data.rel.ro rather than .rodata if the object contains fields that need such fixups at boot-time. This permits the linker to annotate such regions as requiring read-write access only at load time, but not at execution time (in user space), while keeping .rodata truly const (in user space, this is important for reducing the CoW footprint of dynamic executables). This distinction does not matter for the kernel, but it does imply that const data will end up in writable memory if the .data.rel.ro sections are not treated in a special way, as they will end up in the writable .data segment by default. So emit .data.rel.ro into the .rodata segment. Cc: stable@vger.kernel.org Signed-off-by: Ard Biesheuvel Link: https://lore.kernel.org/r/20250221135704.431269-5-ardb+git@google.com Signed-off-by: Josh Poimboeuf --- include/asm-generic/vmlinux.lds.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index 54504013c749..337d3336e175 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -457,7 +457,7 @@ defined(CONFIG_AUTOFDO_CLANG) || defined(CONFIG_PROPELLER_CLANG) . = ALIGN((align)); \ .rodata : AT(ADDR(.rodata) - LOAD_OFFSET) { \ __start_rodata = .; \ - *(.rodata) *(.rodata.*) \ + *(.rodata) *(.rodata.*) *(.data.rel.ro*) \ SCHED_DATA \ RO_AFTER_INIT_DATA /* Read only after init */ \ . = ALIGN(8); \ -- cgit v1.2.3 From 73cfc53cc3b6380eccf013049574485f64cb83ca Mon Sep 17 00:00:00 2001 From: Ard Biesheuvel Date: Fri, 21 Feb 2025 14:57:07 +0100 Subject: objtool: Fix C jump table annotations for Clang A C jump table (such as the one used by the BPF interpreter) is a const global array of absolute code addresses, and this means that the actual values in the table may not be known until the kernel is booted (e.g., when using KASLR or when the kernel VA space is sized dynamically). When using PIE codegen, the compiler will default to placing such const global objects in .data.rel.ro (which is annotated as writable), rather than .rodata (which is annotated as read-only). As C jump tables are explicitly emitted into .rodata, this used to result in warnings for LoongArch builds (which uses PIE codegen for the entire kernel) like Warning: setting incorrect section attributes for .rodata..c_jump_table due to the fact that the explicitly specified .rodata section inherited the read-write annotation that the compiler uses for such objects when using PIE codegen. This warning was suppressed by explicitly adding the read-only annotation to the __attribute__((section(""))) string, by commit c5b1184decc8 ("compiler.h: specify correct attribute for .rodata..c_jump_table") Unfortunately, this hack does not work on Clang's integrated assembler, which happily interprets the appended section type and permission specifiers as part of the section name, which therefore no longer matches the hard-coded pattern '.rodata..c_jump_table' that objtool expects, causing it to emit a warning kernel/bpf/core.o: warning: objtool: ___bpf_prog_run+0x20: sibling call from callable instruction with modified stack frame Work around this, by emitting C jump tables into .data.rel.ro instead, which is treated as .rodata by the linker script for all builds, not just PIE based ones. Fixes: c5b1184decc8 ("compiler.h: specify correct attribute for .rodata..c_jump_table") Tested-by: Tiezhu Yang # on LoongArch Signed-off-by: Ard Biesheuvel Link: https://lore.kernel.org/r/20250221135704.431269-6-ardb+git@google.com Signed-off-by: Josh Poimboeuf --- include/linux/compiler.h | 2 +- tools/objtool/check.c | 7 ++++--- tools/objtool/include/objtool/special.h | 2 +- 3 files changed, 6 insertions(+), 5 deletions(-) (limited to 'include') diff --git a/include/linux/compiler.h b/include/linux/compiler.h index b087de2f3e94..0c25f3e429bb 100644 --- a/include/linux/compiler.h +++ b/include/linux/compiler.h @@ -110,7 +110,7 @@ void ftrace_likely_update(struct ftrace_likely_data *f, int val, /* Unreachable code */ #ifdef CONFIG_OBJTOOL /* Annotate a C jump table to allow objtool to follow the code flow */ -#define __annotate_jump_table __section(".rodata..c_jump_table,\"a\",@progbits #") +#define __annotate_jump_table __section(".data.rel.ro.c_jump_table") #else /* !CONFIG_OBJTOOL */ #define __annotate_jump_table #endif /* CONFIG_OBJTOOL */ diff --git a/tools/objtool/check.c b/tools/objtool/check.c index 497cb8dfb3eb..1b5a1b3ea7a9 100644 --- a/tools/objtool/check.c +++ b/tools/objtool/check.c @@ -2471,13 +2471,14 @@ static void mark_rodata(struct objtool_file *file) * * - .rodata: can contain GCC switch tables * - .rodata.: same, if -fdata-sections is being used - * - .rodata..c_jump_table: contains C annotated jump tables + * - .data.rel.ro.c_jump_table: contains C annotated jump tables * * .rodata.str1.* sections are ignored; they don't contain jump tables. */ for_each_sec(file, sec) { - if (!strncmp(sec->name, ".rodata", 7) && - !strstr(sec->name, ".str1.")) { + if ((!strncmp(sec->name, ".rodata", 7) && + !strstr(sec->name, ".str1.")) || + !strncmp(sec->name, ".data.rel.ro", 12)) { sec->rodata = true; found = true; } diff --git a/tools/objtool/include/objtool/special.h b/tools/objtool/include/objtool/special.h index e7ee7ffccefd..e049679bb17b 100644 --- a/tools/objtool/include/objtool/special.h +++ b/tools/objtool/include/objtool/special.h @@ -10,7 +10,7 @@ #include #include -#define C_JUMP_TABLE_SECTION ".rodata..c_jump_table" +#define C_JUMP_TABLE_SECTION ".data.rel.ro.c_jump_table" struct special_alt { struct list_head list; -- cgit v1.2.3 From a6aa36e957a1bfb5341986dec32d013d23228fe1 Mon Sep 17 00:00:00 2001 From: Damien Le Moal Date: Fri, 14 Feb 2025 13:14:34 +0900 Subject: block: Remove zone write plugs when handling native zone append writes For devices that natively support zone append operations, REQ_OP_ZONE_APPEND BIOs are not processed through zone write plugging and are immediately issued to the zoned device. This means that there is no write pointer offset tracking done for these operations and that a zone write plug is not necessary. However, when receiving a zone append BIO, we may already have a zone write plug for the target zone if that zone was previously partially written using regular write operations. In such case, since the write pointer offset of the zone write plug is not incremented by the amount of sectors appended to the zone, 2 issues arise: 1) we risk leaving the plug in the disk hash table if the zone is fully written using zone append or regular write operations, because the write pointer offset will never reach the "zone full" state. 2) Regular write operations that are issued after zone append operations will always be failed by blk_zone_wplug_prepare_bio() as the write pointer alignment check will fail, even if the user correctly accounted for the zone append operations and issued the regular writes with a correct sector. Avoid these issues by immediately removing the zone write plug of zones that are the target of zone append operations when blk_zone_plug_bio() is called. The new function blk_zone_wplug_handle_native_zone_append() implements this for devices that natively support zone append. The removal of the zone write plug using disk_remove_zone_wplug() requires aborting all plugged regular write using disk_zone_wplug_abort() as otherwise the plugged write BIOs would never be executed (with the plug removed, the completion path will never see again the zone write plug as disk_get_zone_wplug() will return NULL). Rate-limited warnings are added to blk_zone_wplug_handle_native_zone_append() and to disk_zone_wplug_abort() to signal this. Since blk_zone_wplug_handle_native_zone_append() is called in the hot path for operations that will not be plugged, disk_get_zone_wplug() is optimized under the assumption that a user issuing zone append operations is not at the same time issuing regular writes and that there are no hashed zone write plugs. The struct gendisk atomic counter nr_zone_wplugs is added to check this, with this counter incremented in disk_insert_zone_wplug() and decremented in disk_remove_zone_wplug(). To be consistent with this fix, we do not need to fill the zone write plug hash table with zone write plugs for zones that are partially written for a device that supports native zone append operations. So modify blk_revalidate_seq_zone() to return early to avoid allocating and inserting a zone write plug for partially written sequential zones if the device natively supports zone append. Reported-by: Jorgen Hansen Fixes: 9b1ce7f0c6f8 ("block: Implement zone append emulation") Cc: stable@vger.kernel.org Signed-off-by: Damien Le Moal Tested-by: Jorgen Hansen Link: https://lore.kernel.org/r/20250214041434.82564-1-dlemoal@kernel.org Signed-off-by: Jens Axboe --- block/blk-zoned.c | 76 +++++++++++++++++++++++++++++++++++++++++++++----- include/linux/blkdev.h | 7 +++-- 2 files changed, 73 insertions(+), 10 deletions(-) (limited to 'include') diff --git a/block/blk-zoned.c b/block/blk-zoned.c index 761ea662ddc3..0c77244a35c9 100644 --- a/block/blk-zoned.c +++ b/block/blk-zoned.c @@ -410,13 +410,14 @@ static bool disk_insert_zone_wplug(struct gendisk *disk, } } hlist_add_head_rcu(&zwplug->node, &disk->zone_wplugs_hash[idx]); + atomic_inc(&disk->nr_zone_wplugs); spin_unlock_irqrestore(&disk->zone_wplugs_lock, flags); return true; } -static struct blk_zone_wplug *disk_get_zone_wplug(struct gendisk *disk, - sector_t sector) +static struct blk_zone_wplug *disk_get_hashed_zone_wplug(struct gendisk *disk, + sector_t sector) { unsigned int zno = disk_zone_no(disk, sector); unsigned int idx = hash_32(zno, disk->zone_wplugs_hash_bits); @@ -437,6 +438,15 @@ static struct blk_zone_wplug *disk_get_zone_wplug(struct gendisk *disk, return NULL; } +static inline struct blk_zone_wplug *disk_get_zone_wplug(struct gendisk *disk, + sector_t sector) +{ + if (!atomic_read(&disk->nr_zone_wplugs)) + return NULL; + + return disk_get_hashed_zone_wplug(disk, sector); +} + static void disk_free_zone_wplug_rcu(struct rcu_head *rcu_head) { struct blk_zone_wplug *zwplug = @@ -503,6 +513,7 @@ static void disk_remove_zone_wplug(struct gendisk *disk, zwplug->flags |= BLK_ZONE_WPLUG_UNHASHED; spin_lock_irqsave(&disk->zone_wplugs_lock, flags); hlist_del_init_rcu(&zwplug->node); + atomic_dec(&disk->nr_zone_wplugs); spin_unlock_irqrestore(&disk->zone_wplugs_lock, flags); disk_put_zone_wplug(zwplug); } @@ -593,6 +604,11 @@ static void disk_zone_wplug_abort(struct blk_zone_wplug *zwplug) { struct bio *bio; + if (bio_list_empty(&zwplug->bio_list)) + return; + + pr_warn_ratelimited("%s: zone %u: Aborting plugged BIOs\n", + zwplug->disk->disk_name, zwplug->zone_no); while ((bio = bio_list_pop(&zwplug->bio_list))) blk_zone_wplug_bio_io_error(zwplug, bio); } @@ -1040,6 +1056,47 @@ plug: return true; } +static void blk_zone_wplug_handle_native_zone_append(struct bio *bio) +{ + struct gendisk *disk = bio->bi_bdev->bd_disk; + struct blk_zone_wplug *zwplug; + unsigned long flags; + + /* + * We have native support for zone append operations, so we are not + * going to handle @bio through plugging. However, we may already have a + * zone write plug for the target zone if that zone was previously + * partially written using regular writes. In such case, we risk leaving + * the plug in the disk hash table if the zone is fully written using + * zone append operations. Avoid this by removing the zone write plug. + */ + zwplug = disk_get_zone_wplug(disk, bio->bi_iter.bi_sector); + if (likely(!zwplug)) + return; + + spin_lock_irqsave(&zwplug->lock, flags); + + /* + * We are about to remove the zone write plug. But if the user + * (mistakenly) has issued regular writes together with native zone + * append, we must aborts the writes as otherwise the plugged BIOs would + * not be executed by the plug BIO work as disk_get_zone_wplug() will + * return NULL after the plug is removed. Aborting the plugged write + * BIOs is consistent with the fact that these writes will most likely + * fail anyway as there is no ordering guarantees between zone append + * operations and regular write operations. + */ + if (!bio_list_empty(&zwplug->bio_list)) { + pr_warn_ratelimited("%s: zone %u: Invalid mix of zone append and regular writes\n", + disk->disk_name, zwplug->zone_no); + disk_zone_wplug_abort(zwplug); + } + disk_remove_zone_wplug(disk, zwplug); + spin_unlock_irqrestore(&zwplug->lock, flags); + + disk_put_zone_wplug(zwplug); +} + /** * blk_zone_plug_bio - Handle a zone write BIO with zone write plugging * @bio: The BIO being submitted @@ -1096,8 +1153,10 @@ bool blk_zone_plug_bio(struct bio *bio, unsigned int nr_segs) */ switch (bio_op(bio)) { case REQ_OP_ZONE_APPEND: - if (!bdev_emulates_zone_append(bdev)) + if (!bdev_emulates_zone_append(bdev)) { + blk_zone_wplug_handle_native_zone_append(bio); return false; + } fallthrough; case REQ_OP_WRITE: case REQ_OP_WRITE_ZEROES: @@ -1284,6 +1343,7 @@ static int disk_alloc_zone_resources(struct gendisk *disk, { unsigned int i; + atomic_set(&disk->nr_zone_wplugs, 0); disk->zone_wplugs_hash_bits = min(ilog2(pool_size) + 1, BLK_ZONE_WPLUG_MAX_HASH_BITS); @@ -1338,6 +1398,7 @@ static void disk_destroy_zone_wplugs_hash_table(struct gendisk *disk) } } + WARN_ON_ONCE(atomic_read(&disk->nr_zone_wplugs)); kfree(disk->zone_wplugs_hash); disk->zone_wplugs_hash = NULL; disk->zone_wplugs_hash_bits = 0; @@ -1550,11 +1611,12 @@ static int blk_revalidate_seq_zone(struct blk_zone *zone, unsigned int idx, } /* - * We need to track the write pointer of all zones that are not - * empty nor full. So make sure we have a zone write plug for - * such zone if the device has a zone write plug hash table. + * If the device needs zone append emulation, we need to track the + * write pointer of all zones that are not empty nor full. So make sure + * we have a zone write plug for such zone if the device has a zone + * write plug hash table. */ - if (!disk->zone_wplugs_hash) + if (!queue_emulates_zone_append(disk->queue) || !disk->zone_wplugs_hash) return 0; disk_zone_wplug_sync_wp_offset(disk, zone); diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 58ff5aca83b6..d37751789bf5 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -196,10 +196,11 @@ struct gendisk { unsigned int zone_capacity; unsigned int last_zone_capacity; unsigned long __rcu *conv_zones_bitmap; - unsigned int zone_wplugs_hash_bits; - spinlock_t zone_wplugs_lock; + unsigned int zone_wplugs_hash_bits; + atomic_t nr_zone_wplugs; + spinlock_t zone_wplugs_lock; struct mempool_s *zone_wplugs_pool; - struct hlist_head *zone_wplugs_hash; + struct hlist_head *zone_wplugs_hash; struct workqueue_struct *zone_wplugs_wq; #endif /* CONFIG_BLK_DEV_ZONED */ -- cgit v1.2.3 From 02410ac72ac3707936c07ede66e94360d0d65319 Mon Sep 17 00:00:00 2001 From: Ryan Roberts Date: Wed, 26 Feb 2025 12:06:51 +0000 Subject: mm: hugetlb: Add huge page size param to huge_ptep_get_and_clear() In order to fix a bug, arm64 needs to be told the size of the huge page for which the huge_pte is being cleared in huge_ptep_get_and_clear(). Provide for this by adding an `unsigned long sz` parameter to the function. This follows the same pattern as huge_pte_clear() and set_huge_pte_at(). This commit makes the required interface modifications to the core mm as well as all arches that implement this function (arm64, loongarch, mips, parisc, powerpc, riscv, s390, sparc). The actual arm64 bug will be fixed in a separate commit. Cc: stable@vger.kernel.org Fixes: 66b3923a1a0f ("arm64: hugetlb: add support for PTE contiguous bit") Acked-by: David Hildenbrand Reviewed-by: Alexandre Ghiti # riscv Reviewed-by: Christophe Leroy Reviewed-by: Catalin Marinas Reviewed-by: Anshuman Khandual Signed-off-by: Ryan Roberts Acked-by: Alexander Gordeev # s390 Link: https://lore.kernel.org/r/20250226120656.2400136-2-ryan.roberts@arm.com Signed-off-by: Will Deacon --- arch/arm64/include/asm/hugetlb.h | 4 ++-- arch/arm64/mm/hugetlbpage.c | 8 +++++--- arch/loongarch/include/asm/hugetlb.h | 6 ++++-- arch/mips/include/asm/hugetlb.h | 6 ++++-- arch/parisc/include/asm/hugetlb.h | 2 +- arch/parisc/mm/hugetlbpage.c | 2 +- arch/powerpc/include/asm/hugetlb.h | 6 ++++-- arch/riscv/include/asm/hugetlb.h | 3 ++- arch/riscv/mm/hugetlbpage.c | 2 +- arch/s390/include/asm/hugetlb.h | 16 ++++++++++++---- arch/s390/mm/hugetlbpage.c | 4 ++-- arch/sparc/include/asm/hugetlb.h | 2 +- arch/sparc/mm/hugetlbpage.c | 2 +- include/asm-generic/hugetlb.h | 2 +- include/linux/hugetlb.h | 4 +++- mm/hugetlb.c | 4 ++-- 16 files changed, 46 insertions(+), 27 deletions(-) (limited to 'include') diff --git a/arch/arm64/include/asm/hugetlb.h b/arch/arm64/include/asm/hugetlb.h index c6dff3e69539..03db9cb21ace 100644 --- a/arch/arm64/include/asm/hugetlb.h +++ b/arch/arm64/include/asm/hugetlb.h @@ -42,8 +42,8 @@ extern int huge_ptep_set_access_flags(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep, pte_t pte, int dirty); #define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR -extern pte_t huge_ptep_get_and_clear(struct mm_struct *mm, - unsigned long addr, pte_t *ptep); +extern pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, + pte_t *ptep, unsigned long sz); #define __HAVE_ARCH_HUGE_PTEP_SET_WRPROTECT extern void huge_ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, pte_t *ptep); diff --git a/arch/arm64/mm/hugetlbpage.c b/arch/arm64/mm/hugetlbpage.c index 98a2a0e64e25..06db4649af91 100644 --- a/arch/arm64/mm/hugetlbpage.c +++ b/arch/arm64/mm/hugetlbpage.c @@ -396,8 +396,8 @@ void huge_pte_clear(struct mm_struct *mm, unsigned long addr, __pte_clear(mm, addr, ptep); } -pte_t huge_ptep_get_and_clear(struct mm_struct *mm, - unsigned long addr, pte_t *ptep) +pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, + pte_t *ptep, unsigned long sz) { int ncontig; size_t pgsize; @@ -549,6 +549,8 @@ bool __init arch_hugetlb_valid_size(unsigned long size) pte_t huge_ptep_modify_prot_start(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep) { + unsigned long psize = huge_page_size(hstate_vma(vma)); + if (alternative_has_cap_unlikely(ARM64_WORKAROUND_2645198)) { /* * Break-before-make (BBM) is required for all user space mappings @@ -558,7 +560,7 @@ pte_t huge_ptep_modify_prot_start(struct vm_area_struct *vma, unsigned long addr if (pte_user_exec(__ptep_get(ptep))) return huge_ptep_clear_flush(vma, addr, ptep); } - return huge_ptep_get_and_clear(vma->vm_mm, addr, ptep); + return huge_ptep_get_and_clear(vma->vm_mm, addr, ptep, psize); } void huge_ptep_modify_prot_commit(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep, diff --git a/arch/loongarch/include/asm/hugetlb.h b/arch/loongarch/include/asm/hugetlb.h index c8e4057734d0..4dc4b3e04225 100644 --- a/arch/loongarch/include/asm/hugetlb.h +++ b/arch/loongarch/include/asm/hugetlb.h @@ -36,7 +36,8 @@ static inline void huge_pte_clear(struct mm_struct *mm, unsigned long addr, #define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR static inline pte_t huge_ptep_get_and_clear(struct mm_struct *mm, - unsigned long addr, pte_t *ptep) + unsigned long addr, pte_t *ptep, + unsigned long sz) { pte_t clear; pte_t pte = ptep_get(ptep); @@ -51,8 +52,9 @@ static inline pte_t huge_ptep_clear_flush(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep) { pte_t pte; + unsigned long sz = huge_page_size(hstate_vma(vma)); - pte = huge_ptep_get_and_clear(vma->vm_mm, addr, ptep); + pte = huge_ptep_get_and_clear(vma->vm_mm, addr, ptep, sz); flush_tlb_page(vma, addr); return pte; } diff --git a/arch/mips/include/asm/hugetlb.h b/arch/mips/include/asm/hugetlb.h index d0a86ce83de9..fbc71ddcf0f6 100644 --- a/arch/mips/include/asm/hugetlb.h +++ b/arch/mips/include/asm/hugetlb.h @@ -27,7 +27,8 @@ static inline int prepare_hugepage_range(struct file *file, #define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR static inline pte_t huge_ptep_get_and_clear(struct mm_struct *mm, - unsigned long addr, pte_t *ptep) + unsigned long addr, pte_t *ptep, + unsigned long sz) { pte_t clear; pte_t pte = *ptep; @@ -42,13 +43,14 @@ static inline pte_t huge_ptep_clear_flush(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep) { pte_t pte; + unsigned long sz = huge_page_size(hstate_vma(vma)); /* * clear the huge pte entry firstly, so that the other smp threads will * not get old pte entry after finishing flush_tlb_page and before * setting new huge pte entry */ - pte = huge_ptep_get_and_clear(vma->vm_mm, addr, ptep); + pte = huge_ptep_get_and_clear(vma->vm_mm, addr, ptep, sz); flush_tlb_page(vma, addr); return pte; } diff --git a/arch/parisc/include/asm/hugetlb.h b/arch/parisc/include/asm/hugetlb.h index 5b3a5429f71b..21e9ace17739 100644 --- a/arch/parisc/include/asm/hugetlb.h +++ b/arch/parisc/include/asm/hugetlb.h @@ -10,7 +10,7 @@ void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, #define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, - pte_t *ptep); + pte_t *ptep, unsigned long sz); #define __HAVE_ARCH_HUGE_PTEP_CLEAR_FLUSH static inline pte_t huge_ptep_clear_flush(struct vm_area_struct *vma, diff --git a/arch/parisc/mm/hugetlbpage.c b/arch/parisc/mm/hugetlbpage.c index e9d18cf25b79..a94fe546d434 100644 --- a/arch/parisc/mm/hugetlbpage.c +++ b/arch/parisc/mm/hugetlbpage.c @@ -126,7 +126,7 @@ void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, - pte_t *ptep) + pte_t *ptep, unsigned long sz) { pte_t entry; diff --git a/arch/powerpc/include/asm/hugetlb.h b/arch/powerpc/include/asm/hugetlb.h index dad2e7980f24..86326587e58d 100644 --- a/arch/powerpc/include/asm/hugetlb.h +++ b/arch/powerpc/include/asm/hugetlb.h @@ -45,7 +45,8 @@ void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep, #define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR static inline pte_t huge_ptep_get_and_clear(struct mm_struct *mm, - unsigned long addr, pte_t *ptep) + unsigned long addr, pte_t *ptep, + unsigned long sz) { return __pte(pte_update(mm, addr, ptep, ~0UL, 0, 1)); } @@ -55,8 +56,9 @@ static inline pte_t huge_ptep_clear_flush(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep) { pte_t pte; + unsigned long sz = huge_page_size(hstate_vma(vma)); - pte = huge_ptep_get_and_clear(vma->vm_mm, addr, ptep); + pte = huge_ptep_get_and_clear(vma->vm_mm, addr, ptep, sz); flush_hugetlb_page(vma, addr); return pte; } diff --git a/arch/riscv/include/asm/hugetlb.h b/arch/riscv/include/asm/hugetlb.h index faf3624d8057..446126497768 100644 --- a/arch/riscv/include/asm/hugetlb.h +++ b/arch/riscv/include/asm/hugetlb.h @@ -28,7 +28,8 @@ void set_huge_pte_at(struct mm_struct *mm, #define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR pte_t huge_ptep_get_and_clear(struct mm_struct *mm, - unsigned long addr, pte_t *ptep); + unsigned long addr, pte_t *ptep, + unsigned long sz); #define __HAVE_ARCH_HUGE_PTEP_CLEAR_FLUSH pte_t huge_ptep_clear_flush(struct vm_area_struct *vma, diff --git a/arch/riscv/mm/hugetlbpage.c b/arch/riscv/mm/hugetlbpage.c index 42314f093922..b4a78a4b35cf 100644 --- a/arch/riscv/mm/hugetlbpage.c +++ b/arch/riscv/mm/hugetlbpage.c @@ -293,7 +293,7 @@ int huge_ptep_set_access_flags(struct vm_area_struct *vma, pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, - pte_t *ptep) + pte_t *ptep, unsigned long sz) { pte_t orig_pte = ptep_get(ptep); int pte_num; diff --git a/arch/s390/include/asm/hugetlb.h b/arch/s390/include/asm/hugetlb.h index 7c52acaf9f82..663e87220e89 100644 --- a/arch/s390/include/asm/hugetlb.h +++ b/arch/s390/include/asm/hugetlb.h @@ -25,8 +25,16 @@ void __set_huge_pte_at(struct mm_struct *mm, unsigned long addr, #define __HAVE_ARCH_HUGE_PTEP_GET pte_t huge_ptep_get(struct mm_struct *mm, unsigned long addr, pte_t *ptep); +pte_t __huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, + pte_t *ptep); + #define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR -pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep); +static inline pte_t huge_ptep_get_and_clear(struct mm_struct *mm, + unsigned long addr, pte_t *ptep, + unsigned long sz) +{ + return __huge_ptep_get_and_clear(mm, addr, ptep); +} static inline void arch_clear_hugetlb_flags(struct folio *folio) { @@ -48,7 +56,7 @@ static inline void huge_pte_clear(struct mm_struct *mm, unsigned long addr, static inline pte_t huge_ptep_clear_flush(struct vm_area_struct *vma, unsigned long address, pte_t *ptep) { - return huge_ptep_get_and_clear(vma->vm_mm, address, ptep); + return __huge_ptep_get_and_clear(vma->vm_mm, address, ptep); } #define __HAVE_ARCH_HUGE_PTEP_SET_ACCESS_FLAGS @@ -59,7 +67,7 @@ static inline int huge_ptep_set_access_flags(struct vm_area_struct *vma, int changed = !pte_same(huge_ptep_get(vma->vm_mm, addr, ptep), pte); if (changed) { - huge_ptep_get_and_clear(vma->vm_mm, addr, ptep); + __huge_ptep_get_and_clear(vma->vm_mm, addr, ptep); __set_huge_pte_at(vma->vm_mm, addr, ptep, pte); } return changed; @@ -69,7 +77,7 @@ static inline int huge_ptep_set_access_flags(struct vm_area_struct *vma, static inline void huge_ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, pte_t *ptep) { - pte_t pte = huge_ptep_get_and_clear(mm, addr, ptep); + pte_t pte = __huge_ptep_get_and_clear(mm, addr, ptep); __set_huge_pte_at(mm, addr, ptep, pte_wrprotect(pte)); } diff --git a/arch/s390/mm/hugetlbpage.c b/arch/s390/mm/hugetlbpage.c index d9ce199953de..2e568f175cd4 100644 --- a/arch/s390/mm/hugetlbpage.c +++ b/arch/s390/mm/hugetlbpage.c @@ -188,8 +188,8 @@ pte_t huge_ptep_get(struct mm_struct *mm, unsigned long addr, pte_t *ptep) return __rste_to_pte(pte_val(*ptep)); } -pte_t huge_ptep_get_and_clear(struct mm_struct *mm, - unsigned long addr, pte_t *ptep) +pte_t __huge_ptep_get_and_clear(struct mm_struct *mm, + unsigned long addr, pte_t *ptep) { pte_t pte = huge_ptep_get(mm, addr, ptep); pmd_t *pmdp = (pmd_t *) ptep; diff --git a/arch/sparc/include/asm/hugetlb.h b/arch/sparc/include/asm/hugetlb.h index c714ca6a05aa..e7a9cdd498dc 100644 --- a/arch/sparc/include/asm/hugetlb.h +++ b/arch/sparc/include/asm/hugetlb.h @@ -20,7 +20,7 @@ void __set_huge_pte_at(struct mm_struct *mm, unsigned long addr, #define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, - pte_t *ptep); + pte_t *ptep, unsigned long sz); #define __HAVE_ARCH_HUGE_PTEP_CLEAR_FLUSH static inline pte_t huge_ptep_clear_flush(struct vm_area_struct *vma, diff --git a/arch/sparc/mm/hugetlbpage.c b/arch/sparc/mm/hugetlbpage.c index eee601a0d2cf..80504148d8a5 100644 --- a/arch/sparc/mm/hugetlbpage.c +++ b/arch/sparc/mm/hugetlbpage.c @@ -260,7 +260,7 @@ void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, } pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, - pte_t *ptep) + pte_t *ptep, unsigned long sz) { unsigned int i, nptes, orig_shift, shift; unsigned long size; diff --git a/include/asm-generic/hugetlb.h b/include/asm-generic/hugetlb.h index f42133dae68e..2afc95bf1655 100644 --- a/include/asm-generic/hugetlb.h +++ b/include/asm-generic/hugetlb.h @@ -90,7 +90,7 @@ static inline void set_huge_pte_at(struct mm_struct *mm, unsigned long addr, #ifndef __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR static inline pte_t huge_ptep_get_and_clear(struct mm_struct *mm, - unsigned long addr, pte_t *ptep) + unsigned long addr, pte_t *ptep, unsigned long sz) { return ptep_get_and_clear(mm, addr, ptep); } diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index ec8c0ccc8f95..bf5f7256bd28 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -1004,7 +1004,9 @@ static inline void hugetlb_count_sub(long l, struct mm_struct *mm) static inline pte_t huge_ptep_modify_prot_start(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep) { - return huge_ptep_get_and_clear(vma->vm_mm, addr, ptep); + unsigned long psize = huge_page_size(hstate_vma(vma)); + + return huge_ptep_get_and_clear(vma->vm_mm, addr, ptep, psize); } #endif diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 65068671e460..de9d49e521c1 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -5447,7 +5447,7 @@ static void move_huge_pte(struct vm_area_struct *vma, unsigned long old_addr, if (src_ptl != dst_ptl) spin_lock_nested(src_ptl, SINGLE_DEPTH_NESTING); - pte = huge_ptep_get_and_clear(mm, old_addr, src_pte); + pte = huge_ptep_get_and_clear(mm, old_addr, src_pte, sz); if (need_clear_uffd_wp && pte_marker_uffd_wp(pte)) huge_pte_clear(mm, new_addr, dst_pte, sz); @@ -5622,7 +5622,7 @@ void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct *vma, set_vma_resv_flags(vma, HPAGE_RESV_UNMAPPED); } - pte = huge_ptep_get_and_clear(mm, address, ptep); + pte = huge_ptep_get_and_clear(mm, address, ptep, sz); tlb_remove_huge_tlb_entry(h, tlb, ptep, address); if (huge_pte_dirty(pte)) set_page_dirty(page); -- cgit v1.2.3 From 916b7f42b3b3b539a71c204a9b49fdc4ca92cd82 Mon Sep 17 00:00:00 2001 From: Keith Busch Date: Thu, 27 Feb 2025 15:06:31 -0800 Subject: kvm: retry nx_huge_page_recovery_thread creation A VMM may send a non-fatal signal to its threads, including vCPU tasks, at any time, and thus may signal vCPU tasks during KVM_RUN. If a vCPU task receives the signal while its trying to spawn the huge page recovery vhost task, then KVM_RUN will fail due to copy_process() returning -ERESTARTNOINTR. Rework call_once() to mark the call complete if and only if the called function succeeds, and plumb the function's true error code back to the call_once() invoker. This provides userspace with the correct, non-fatal error code so that the VMM doesn't terminate the VM on -ENOMEM, and allows subsequent KVM_RUN a succeed by virtue of retrying creation of the NX huge page task. Co-developed-by: Sean Christopherson Signed-off-by: Sean Christopherson [implemented the kvm user side] Signed-off-by: Keith Busch Message-ID: <20250227230631.303431-3-kbusch@meta.com> Signed-off-by: Paolo Bonzini --- arch/x86/kvm/mmu/mmu.c | 10 ++++------ include/linux/call_once.h | 47 ++++++++++++++++++++++++++++++++++------------- 2 files changed, 38 insertions(+), 19 deletions(-) (limited to 'include') diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c index 18ca1ea6dc24..8160870398b9 100644 --- a/arch/x86/kvm/mmu/mmu.c +++ b/arch/x86/kvm/mmu/mmu.c @@ -7460,7 +7460,7 @@ static bool kvm_nx_huge_page_recovery_worker(void *data) return true; } -static void kvm_mmu_start_lpage_recovery(struct once *once) +static int kvm_mmu_start_lpage_recovery(struct once *once) { struct kvm_arch *ka = container_of(once, struct kvm_arch, nx_once); struct kvm *kvm = container_of(ka, struct kvm, arch); @@ -7472,12 +7472,13 @@ static void kvm_mmu_start_lpage_recovery(struct once *once) kvm, "kvm-nx-lpage-recovery"); if (IS_ERR(nx_thread)) - return; + return PTR_ERR(nx_thread); vhost_task_start(nx_thread); /* Make the task visible only once it is fully started. */ WRITE_ONCE(kvm->arch.nx_huge_page_recovery_thread, nx_thread); + return 0; } int kvm_mmu_post_init_vm(struct kvm *kvm) @@ -7485,10 +7486,7 @@ int kvm_mmu_post_init_vm(struct kvm *kvm) if (nx_hugepage_mitigation_hard_disabled) return 0; - call_once(&kvm->arch.nx_once, kvm_mmu_start_lpage_recovery); - if (!kvm->arch.nx_huge_page_recovery_thread) - return -ENOMEM; - return 0; + return call_once(&kvm->arch.nx_once, kvm_mmu_start_lpage_recovery); } void kvm_mmu_pre_destroy_vm(struct kvm *kvm) diff --git a/include/linux/call_once.h b/include/linux/call_once.h index 6261aa0b3fb0..13cd6469e7e5 100644 --- a/include/linux/call_once.h +++ b/include/linux/call_once.h @@ -26,20 +26,41 @@ do { \ __once_init((once), #once, &__key); \ } while (0) -static inline void call_once(struct once *once, void (*cb)(struct once *)) +/* + * call_once - Ensure a function has been called exactly once + * + * @once: Tracking struct + * @cb: Function to be called + * + * If @once has never completed successfully before, call @cb and, if + * it returns a zero or positive value, mark @once as completed. Return + * the value returned by @cb + * + * If @once has completed succesfully before, return 0. + * + * The call to @cb is implicitly surrounded by a mutex, though for + * efficiency the * function avoids taking it after the first call. + */ +static inline int call_once(struct once *once, int (*cb)(struct once *)) { - /* Pairs with atomic_set_release() below. */ - if (atomic_read_acquire(&once->state) == ONCE_COMPLETED) - return; - - guard(mutex)(&once->lock); - WARN_ON(atomic_read(&once->state) == ONCE_RUNNING); - if (atomic_read(&once->state) != ONCE_NOT_STARTED) - return; - - atomic_set(&once->state, ONCE_RUNNING); - cb(once); - atomic_set_release(&once->state, ONCE_COMPLETED); + int r, state; + + /* Pairs with atomic_set_release() below. */ + if (atomic_read_acquire(&once->state) == ONCE_COMPLETED) + return 0; + + guard(mutex)(&once->lock); + state = atomic_read(&once->state); + if (unlikely(state != ONCE_NOT_STARTED)) + return WARN_ON_ONCE(state != ONCE_COMPLETED) ? -EINVAL : 0; + + atomic_set(&once->state, ONCE_RUNNING); + r = cb(once); + if (r < 0) + atomic_set(&once->state, ONCE_NOT_STARTED); + else + atomic_set_release(&once->state, ONCE_COMPLETED); + return r; } #endif /* _LINUX_CALL_ONCE_H */ -- cgit v1.2.3 From e04918dc594669068f5d59d567d08db531167188 Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Sun, 2 Mar 2025 15:18:24 +0800 Subject: cred: Fix RCU warnings in override/revert_creds Fix RCU warnings in override_creds and revert_creds by turning the RCU pointer into a normal pointer using rcu_replace_pointer. These warnings were previously private to the cred code, but due to the move into the header file they are now polluting unrelated subsystems. Fixes: 49dffdfde462 ("cred: Add a light version of override/revert_creds()") Signed-off-by: Herbert Xu Link: https://lore.kernel.org/r/Z8QGQGW0IaSklKG7@gondor.apana.org.au Signed-off-by: Christian Brauner --- include/linux/cred.h | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-) (limited to 'include') diff --git a/include/linux/cred.h b/include/linux/cred.h index 0c3c4b16b469..5658a3bfe803 100644 --- a/include/linux/cred.h +++ b/include/linux/cred.h @@ -172,18 +172,12 @@ static inline bool cap_ambient_invariant_ok(const struct cred *cred) static inline const struct cred *override_creds(const struct cred *override_cred) { - const struct cred *old = current->cred; - - rcu_assign_pointer(current->cred, override_cred); - return old; + return rcu_replace_pointer(current->cred, override_cred, 1); } static inline const struct cred *revert_creds(const struct cred *revert_cred) { - const struct cred *override_cred = current->cred; - - rcu_assign_pointer(current->cred, revert_cred); - return override_cred; + return rcu_replace_pointer(current->cred, revert_cred, 1); } /** -- cgit v1.2.3 From 3d252160b818045f3a152b13756f6f37ca34639d Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Tue, 4 Mar 2025 13:51:38 +0000 Subject: fs/pipe: Read pipe->{head,tail} atomically outside pipe->mutex pipe_readable(), pipe_writable(), and pipe_poll() can read "pipe->head" and "pipe->tail" outside of "pipe->mutex" critical section. When the head and the tail are read individually in that order, there is a window for interruption between the two reads in which both the head and the tail can be updated by concurrent readers and writers. One of the problematic scenarios observed with hackbench running multiple groups on a large server on a particular pipe inode is as follows: pipe->head = 36 pipe->tail = 36 hackbench-118762 [057] ..... 1029.550548: pipe_write: *wakes up: pipe not full* hackbench-118762 [057] ..... 1029.550548: pipe_write: head: 36 -> 37 [tail: 36] hackbench-118762 [057] ..... 1029.550548: pipe_write: *wake up next reader 118740* hackbench-118762 [057] ..... 1029.550548: pipe_write: *wake up next writer 118768* hackbench-118768 [206] ..... 1029.55055X: pipe_write: *writer wakes up* hackbench-118768 [206] ..... 1029.55055X: pipe_write: head = READ_ONCE(pipe->head) [37] ... CPU 206 interrupted (exact wakeup was not traced but 118768 did read head at 37 in traces) hackbench-118740 [057] ..... 1029.550558: pipe_read: *reader wakes up: pipe is not empty* hackbench-118740 [057] ..... 1029.550558: pipe_read: tail: 36 -> 37 [head = 37] hackbench-118740 [057] ..... 1029.550559: pipe_read: *pipe is empty; wakeup writer 118768* hackbench-118740 [057] ..... 1029.550559: pipe_read: *sleeps* hackbench-118766 [185] ..... 1029.550592: pipe_write: *New writer comes in* hackbench-118766 [185] ..... 1029.550592: pipe_write: head: 37 -> 38 [tail: 37] hackbench-118766 [185] ..... 1029.550592: pipe_write: *wakes up reader 118766* hackbench-118740 [185] ..... 1029.550598: pipe_read: *reader wakes up; pipe not empty* hackbench-118740 [185] ..... 1029.550599: pipe_read: tail: 37 -> 38 [head: 38] hackbench-118740 [185] ..... 1029.550599: pipe_read: *pipe is empty* hackbench-118740 [185] ..... 1029.550599: pipe_read: *reader sleeps; wakeup writer 118768* ... CPU 206 switches back to writer hackbench-118768 [206] ..... 1029.550601: pipe_write: tail = READ_ONCE(pipe->tail) [38] hackbench-118768 [206] ..... 1029.550601: pipe_write: pipe_full()? (u32)(37 - 38) >= 16? Yes hackbench-118768 [206] ..... 1029.550601: pipe_write: *writer goes back to sleep* [ Tasks 118740 and 118768 can then indefinitely wait on each other. ] The unsigned arithmetic in pipe_occupancy() wraps around when "pipe->tail > pipe->head" leading to pipe_full() returning true despite the pipe being empty. The case of genuine wraparound of "pipe->head" is handled since pipe buffer has data allowing readers to make progress until the pipe->tail wraps too after which the reader will wakeup a sleeping writer, however, mistaking the pipe to be full when it is in fact empty can lead to readers and writers waiting on each other indefinitely. This issue became more problematic and surfaced as a hang in hackbench after the optimization in commit aaec5a95d596 ("pipe_read: don't wake up the writer if the pipe is still full") significantly reduced the number of spurious wakeups of writers that had previously helped mask the issue. To avoid missing any updates between the reads of "pipe->head" and "pipe->write", unionize the two with a single unsigned long "pipe->head_tail" member that can be loaded atomically. Using "pipe->head_tail" to read the head and the tail ensures the lockless checks do not miss any updates to the head or the tail and since those two are only updated under "pipe->mutex", it ensures that the head is always ahead of, or equal to the tail resulting in correct calculations. [ prateek: commit log, testing on x86 platforms. ] Reported-and-debugged-by: Swapnil Sapkal Closes: https://lore.kernel.org/lkml/e813814e-7094-4673-bc69-731af065a0eb@amd.com/ Reported-by: Alexey Gladkov Closes: https://lore.kernel.org/all/Z8Wn0nTvevLRG_4m@example.org/ Fixes: 8cefc107ca54 ("pipe: Use head and tail pointers for the ring, not cursor and length") Tested-by: Swapnil Sapkal Reviewed-by: Oleg Nesterov Tested-by: Alexey Gladkov Signed-off-by: K Prateek Nayak Signed-off-by: Linus Torvalds --- fs/pipe.c | 19 ++++++++----------- include/linux/pipe_fs_i.h | 39 +++++++++++++++++++++++++++++++++++++-- 2 files changed, 45 insertions(+), 13 deletions(-) (limited to 'include') diff --git a/fs/pipe.c b/fs/pipe.c index ce1af7592780..e8e6698f3698 100644 --- a/fs/pipe.c +++ b/fs/pipe.c @@ -210,11 +210,10 @@ static const struct pipe_buf_operations anon_pipe_buf_ops = { /* Done while waiting without holding the pipe lock - thus the READ_ONCE() */ static inline bool pipe_readable(const struct pipe_inode_info *pipe) { - unsigned int head = READ_ONCE(pipe->head); - unsigned int tail = READ_ONCE(pipe->tail); + union pipe_index idx = { .head_tail = READ_ONCE(pipe->head_tail) }; unsigned int writers = READ_ONCE(pipe->writers); - return !pipe_empty(head, tail) || !writers; + return !pipe_empty(idx.head, idx.tail) || !writers; } static inline unsigned int pipe_update_tail(struct pipe_inode_info *pipe, @@ -417,11 +416,10 @@ static inline int is_packetized(struct file *file) /* Done while waiting without holding the pipe lock - thus the READ_ONCE() */ static inline bool pipe_writable(const struct pipe_inode_info *pipe) { - unsigned int head = READ_ONCE(pipe->head); - unsigned int tail = READ_ONCE(pipe->tail); + union pipe_index idx = { .head_tail = READ_ONCE(pipe->head_tail) }; unsigned int max_usage = READ_ONCE(pipe->max_usage); - return !pipe_full(head, tail, max_usage) || + return !pipe_full(idx.head, idx.tail, max_usage) || !READ_ONCE(pipe->readers); } @@ -659,7 +657,7 @@ pipe_poll(struct file *filp, poll_table *wait) { __poll_t mask; struct pipe_inode_info *pipe = filp->private_data; - unsigned int head, tail; + union pipe_index idx; /* Epoll has some historical nasty semantics, this enables them */ WRITE_ONCE(pipe->poll_usage, true); @@ -680,19 +678,18 @@ pipe_poll(struct file *filp, poll_table *wait) * if something changes and you got it wrong, the poll * table entry will wake you up and fix it. */ - head = READ_ONCE(pipe->head); - tail = READ_ONCE(pipe->tail); + idx.head_tail = READ_ONCE(pipe->head_tail); mask = 0; if (filp->f_mode & FMODE_READ) { - if (!pipe_empty(head, tail)) + if (!pipe_empty(idx.head, idx.tail)) mask |= EPOLLIN | EPOLLRDNORM; if (!pipe->writers && filp->f_pipe != pipe->w_counter) mask |= EPOLLHUP; } if (filp->f_mode & FMODE_WRITE) { - if (!pipe_full(head, tail, pipe->max_usage)) + if (!pipe_full(idx.head, idx.tail, pipe->max_usage)) mask |= EPOLLOUT | EPOLLWRNORM; /* * Most Unices do not set EPOLLERR for FIFOs but on Linux they diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h index 8ff23bf5a819..3cc4f8eab853 100644 --- a/include/linux/pipe_fs_i.h +++ b/include/linux/pipe_fs_i.h @@ -31,6 +31,33 @@ struct pipe_buffer { unsigned long private; }; +/* + * Really only alpha needs 32-bit fields, but + * might as well do it for 64-bit architectures + * since that's what we've historically done, + * and it makes 'head_tail' always be a simple + * 'unsigned long'. + */ +#ifdef CONFIG_64BIT +typedef unsigned int pipe_index_t; +#else +typedef unsigned short pipe_index_t; +#endif + +/* + * We have to declare this outside 'struct pipe_inode_info', + * but then we can't use 'union pipe_index' for an anonymous + * union, so we end up having to duplicate this declaration + * below. Annoying. + */ +union pipe_index { + unsigned long head_tail; + struct { + pipe_index_t head; + pipe_index_t tail; + }; +}; + /** * struct pipe_inode_info - a linux kernel pipe * @mutex: mutex protecting the whole thing @@ -58,8 +85,16 @@ struct pipe_buffer { struct pipe_inode_info { struct mutex mutex; wait_queue_head_t rd_wait, wr_wait; - unsigned int head; - unsigned int tail; + + /* This has to match the 'union pipe_index' above */ + union { + unsigned long head_tail; + struct { + pipe_index_t head; + pipe_index_t tail; + }; + }; + unsigned int max_usage; unsigned int ring_size; unsigned int nr_accounted; -- cgit v1.2.3 From 778b94d7ac17b5800aa857222911f09cc986b509 Mon Sep 17 00:00:00 2001 From: Mario Limonciello Date: Fri, 28 Feb 2025 11:01:53 -0600 Subject: ACPI: platform_profile: Add support for hidden choices MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When two drivers don't support all the same profiles the legacy interface only exports the common profiles. This causes problems for cases where one driver uses low-power but another uses quiet because the result is that neither is exported to sysfs. To allow two drivers to disagree, add support for "hidden choices". Hidden choices are platform profiles that a driver supports to be compatible with the platform profile of another driver. Fixes: 688834743d67 ("ACPI: platform_profile: Allow multiple handlers") Reported-by: Antheas Kapenekakis Closes: https://lore.kernel.org/platform-driver-x86/e64b771e-3255-42ad-9257-5b8fc6c24ac9@gmx.de/T/#mc068042dd29df36c16c8af92664860fc4763974b Signed-off-by: Mario Limonciello Tested-by: Antheas Kapenekakis Tested-by: Derek J. Clark Acked-by: Ilpo Järvinen Link: https://patch.msgid.link/20250228170155.2623386-2-superm1@kernel.org Signed-off-by: Rafael J. Wysocki --- drivers/acpi/platform_profile.c | 94 +++++++++++++++++++++++++++++++--------- include/linux/platform_profile.h | 3 ++ 2 files changed, 76 insertions(+), 21 deletions(-) (limited to 'include') diff --git a/drivers/acpi/platform_profile.c b/drivers/acpi/platform_profile.c index 2ad53cc6aae5..ef9444482db1 100644 --- a/drivers/acpi/platform_profile.c +++ b/drivers/acpi/platform_profile.c @@ -21,9 +21,15 @@ struct platform_profile_handler { struct device dev; int minor; unsigned long choices[BITS_TO_LONGS(PLATFORM_PROFILE_LAST)]; + unsigned long hidden_choices[BITS_TO_LONGS(PLATFORM_PROFILE_LAST)]; const struct platform_profile_ops *ops; }; +struct aggregate_choices_data { + unsigned long aggregate[BITS_TO_LONGS(PLATFORM_PROFILE_LAST)]; + int count; +}; + static const char * const profile_names[] = { [PLATFORM_PROFILE_LOW_POWER] = "low-power", [PLATFORM_PROFILE_COOL] = "cool", @@ -73,7 +79,7 @@ static int _store_class_profile(struct device *dev, void *data) lockdep_assert_held(&profile_lock); handler = to_pprof_handler(dev); - if (!test_bit(*bit, handler->choices)) + if (!test_bit(*bit, handler->choices) && !test_bit(*bit, handler->hidden_choices)) return -EOPNOTSUPP; return handler->ops->profile_set(dev, *bit); @@ -239,21 +245,44 @@ static const struct class platform_profile_class = { /** * _aggregate_choices - Aggregate the available profile choices * @dev: The device - * @data: The available profile choices + * @arg: struct aggregate_choices_data * * Return: 0 on success, -errno on failure */ -static int _aggregate_choices(struct device *dev, void *data) +static int _aggregate_choices(struct device *dev, void *arg) { + unsigned long tmp[BITS_TO_LONGS(PLATFORM_PROFILE_LAST)]; + struct aggregate_choices_data *data = arg; struct platform_profile_handler *handler; - unsigned long *aggregate = data; lockdep_assert_held(&profile_lock); handler = to_pprof_handler(dev); - if (test_bit(PLATFORM_PROFILE_LAST, aggregate)) - bitmap_copy(aggregate, handler->choices, PLATFORM_PROFILE_LAST); + bitmap_or(tmp, handler->choices, handler->hidden_choices, PLATFORM_PROFILE_LAST); + if (test_bit(PLATFORM_PROFILE_LAST, data->aggregate)) + bitmap_copy(data->aggregate, tmp, PLATFORM_PROFILE_LAST); else - bitmap_and(aggregate, handler->choices, aggregate, PLATFORM_PROFILE_LAST); + bitmap_and(data->aggregate, tmp, data->aggregate, PLATFORM_PROFILE_LAST); + data->count++; + + return 0; +} + +/** + * _remove_hidden_choices - Remove hidden choices from aggregate data + * @dev: The device + * @arg: struct aggregate_choices_data + * + * Return: 0 on success, -errno on failure + */ +static int _remove_hidden_choices(struct device *dev, void *arg) +{ + struct aggregate_choices_data *data = arg; + struct platform_profile_handler *handler; + + lockdep_assert_held(&profile_lock); + handler = to_pprof_handler(dev); + bitmap_andnot(data->aggregate, handler->choices, + handler->hidden_choices, PLATFORM_PROFILE_LAST); return 0; } @@ -270,22 +299,31 @@ static ssize_t platform_profile_choices_show(struct device *dev, struct device_attribute *attr, char *buf) { - unsigned long aggregate[BITS_TO_LONGS(PLATFORM_PROFILE_LAST)]; + struct aggregate_choices_data data = { + .aggregate = { [0 ... BITS_TO_LONGS(PLATFORM_PROFILE_LAST) - 1] = ~0UL }, + .count = 0, + }; int err; - set_bit(PLATFORM_PROFILE_LAST, aggregate); + set_bit(PLATFORM_PROFILE_LAST, data.aggregate); scoped_cond_guard(mutex_intr, return -ERESTARTSYS, &profile_lock) { err = class_for_each_device(&platform_profile_class, NULL, - aggregate, _aggregate_choices); + &data, _aggregate_choices); if (err) return err; + if (data.count == 1) { + err = class_for_each_device(&platform_profile_class, NULL, + &data, _remove_hidden_choices); + if (err) + return err; + } } /* no profile handler registered any more */ - if (bitmap_empty(aggregate, PLATFORM_PROFILE_LAST)) + if (bitmap_empty(data.aggregate, PLATFORM_PROFILE_LAST)) return -EINVAL; - return _commmon_choices_show(aggregate, buf); + return _commmon_choices_show(data.aggregate, buf); } /** @@ -373,7 +411,10 @@ static ssize_t platform_profile_store(struct device *dev, struct device_attribute *attr, const char *buf, size_t count) { - unsigned long choices[BITS_TO_LONGS(PLATFORM_PROFILE_LAST)]; + struct aggregate_choices_data data = { + .aggregate = { [0 ... BITS_TO_LONGS(PLATFORM_PROFILE_LAST) - 1] = ~0UL }, + .count = 0, + }; int ret; int i; @@ -381,13 +422,13 @@ static ssize_t platform_profile_store(struct device *dev, i = sysfs_match_string(profile_names, buf); if (i < 0 || i == PLATFORM_PROFILE_CUSTOM) return -EINVAL; - set_bit(PLATFORM_PROFILE_LAST, choices); + set_bit(PLATFORM_PROFILE_LAST, data.aggregate); scoped_cond_guard(mutex_intr, return -ERESTARTSYS, &profile_lock) { ret = class_for_each_device(&platform_profile_class, NULL, - choices, _aggregate_choices); + &data, _aggregate_choices); if (ret) return ret; - if (!test_bit(i, choices)) + if (!test_bit(i, data.aggregate)) return -EOPNOTSUPP; ret = class_for_each_device(&platform_profile_class, NULL, &i, @@ -453,12 +494,15 @@ EXPORT_SYMBOL_GPL(platform_profile_notify); */ int platform_profile_cycle(void) { + struct aggregate_choices_data data = { + .aggregate = { [0 ... BITS_TO_LONGS(PLATFORM_PROFILE_LAST) - 1] = ~0UL }, + .count = 0, + }; enum platform_profile_option next = PLATFORM_PROFILE_LAST; enum platform_profile_option profile = PLATFORM_PROFILE_LAST; - unsigned long choices[BITS_TO_LONGS(PLATFORM_PROFILE_LAST)]; int err; - set_bit(PLATFORM_PROFILE_LAST, choices); + set_bit(PLATFORM_PROFILE_LAST, data.aggregate); scoped_cond_guard(mutex_intr, return -ERESTARTSYS, &profile_lock) { err = class_for_each_device(&platform_profile_class, NULL, &profile, _aggregate_profiles); @@ -470,14 +514,14 @@ int platform_profile_cycle(void) return -EINVAL; err = class_for_each_device(&platform_profile_class, NULL, - choices, _aggregate_choices); + &data, _aggregate_choices); if (err) return err; /* never iterate into a custom if all drivers supported it */ - clear_bit(PLATFORM_PROFILE_CUSTOM, choices); + clear_bit(PLATFORM_PROFILE_CUSTOM, data.aggregate); - next = find_next_bit_wrap(choices, + next = find_next_bit_wrap(data.aggregate, PLATFORM_PROFILE_LAST, profile + 1); @@ -532,6 +576,14 @@ struct device *platform_profile_register(struct device *dev, const char *name, return ERR_PTR(-EINVAL); } + if (ops->hidden_choices) { + err = ops->hidden_choices(drvdata, pprof->hidden_choices); + if (err) { + dev_err(dev, "platform_profile hidden_choices failed\n"); + return ERR_PTR(err); + } + } + guard(mutex)(&profile_lock); /* create class interface for individual handler */ diff --git a/include/linux/platform_profile.h b/include/linux/platform_profile.h index 8ab5b0e8eb2c..8c9df7dadd5d 100644 --- a/include/linux/platform_profile.h +++ b/include/linux/platform_profile.h @@ -33,6 +33,8 @@ enum platform_profile_option { * @probe: Callback to setup choices available to the new class device. These * choices will only be enforced when setting a new profile, not when * getting the current one. + * @hidden_choices: Callback to setup choices that are not visible to the user + * but can be set by the driver. * @profile_get: Callback that will be called when showing the current platform * profile in sysfs. * @profile_set: Callback that will be called when storing a new platform @@ -40,6 +42,7 @@ enum platform_profile_option { */ struct platform_profile_ops { int (*probe)(void *drvdata, unsigned long *choices); + int (*hidden_choices)(void *drvdata, unsigned long *choices); int (*profile_get)(struct device *dev, enum platform_profile_option *profile); int (*profile_set)(struct device *dev, enum platform_profile_option profile); }; -- cgit v1.2.3 From 95d0d094ba26432ec467e2260f4bf553053f1f8f Mon Sep 17 00:00:00 2001 From: Qingfang Deng Date: Sat, 1 Mar 2025 21:55:16 +0800 Subject: ppp: use IFF_NO_QUEUE in virtual interfaces MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit For PPPoE, PPTP, and PPPoL2TP, the start_xmit() function directly forwards packets to the underlying network stack and never returns anything other than 1. So these interfaces do not require a qdisc, and the IFF_NO_QUEUE flag should be set. Introduces a direct_xmit flag in struct ppp_channel to indicate when IFF_NO_QUEUE should be applied. The flag is set in ppp_connect_channel() for relevant protocols. While at it, remove the usused latency member from struct ppp_channel. Signed-off-by: Qingfang Deng Reviewed-by: Toke Høiland-Jørgensen Link: https://patch.msgid.link/20250301135517.695809-1-dqfext@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ppp/ppp_generic.c | 4 ++++ drivers/net/ppp/pppoe.c | 1 + drivers/net/ppp/pptp.c | 1 + include/linux/ppp_channel.h | 3 +-- net/l2tp/l2tp_ppp.c | 1 + 5 files changed, 8 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c index 6220866258fc..815108c98b78 100644 --- a/drivers/net/ppp/ppp_generic.c +++ b/drivers/net/ppp/ppp_generic.c @@ -3493,6 +3493,10 @@ ppp_connect_channel(struct channel *pch, int unit) ret = -ENOTCONN; goto outl; } + if (pch->chan->direct_xmit) + ppp->dev->priv_flags |= IFF_NO_QUEUE; + else + ppp->dev->priv_flags &= ~IFF_NO_QUEUE; spin_unlock_bh(&pch->downl); if (pch->file.hdrlen > ppp->file.hdrlen) ppp->file.hdrlen = pch->file.hdrlen; diff --git a/drivers/net/ppp/pppoe.c b/drivers/net/ppp/pppoe.c index 2ea4f4890d23..68e631718ab0 100644 --- a/drivers/net/ppp/pppoe.c +++ b/drivers/net/ppp/pppoe.c @@ -693,6 +693,7 @@ static int pppoe_connect(struct socket *sock, struct sockaddr *uservaddr, po->chan.mtu = dev->mtu - sizeof(struct pppoe_hdr) - 2; po->chan.private = sk; po->chan.ops = &pppoe_chan_ops; + po->chan.direct_xmit = true; error = ppp_register_net_channel(dev_net(dev), &po->chan); if (error) { diff --git a/drivers/net/ppp/pptp.c b/drivers/net/ppp/pptp.c index 689687bd2574..5feaa70b5f47 100644 --- a/drivers/net/ppp/pptp.c +++ b/drivers/net/ppp/pptp.c @@ -465,6 +465,7 @@ static int pptp_connect(struct socket *sock, struct sockaddr *uservaddr, po->chan.mtu -= PPTP_HEADER_OVERHEAD; po->chan.hdrlen = 2 + sizeof(struct pptp_gre_header); + po->chan.direct_xmit = true; error = ppp_register_channel(&po->chan); if (error) { pr_err("PPTP: failed to register PPP channel (%d)\n", error); diff --git a/include/linux/ppp_channel.h b/include/linux/ppp_channel.h index 45e6e427ceb8..f73fbea0dbc2 100644 --- a/include/linux/ppp_channel.h +++ b/include/linux/ppp_channel.h @@ -42,8 +42,7 @@ struct ppp_channel { int hdrlen; /* amount of headroom channel needs */ void *ppp; /* opaque to channel */ int speed; /* transfer rate (bytes/second) */ - /* the following is not used at present */ - int latency; /* overhead time in milliseconds */ + bool direct_xmit; /* no qdisc, xmit directly */ }; #ifdef __KERNEL__ diff --git a/net/l2tp/l2tp_ppp.c b/net/l2tp/l2tp_ppp.c index 53baf2dd5d5d..fc5c2fd8f34c 100644 --- a/net/l2tp/l2tp_ppp.c +++ b/net/l2tp/l2tp_ppp.c @@ -806,6 +806,7 @@ static int pppol2tp_connect(struct socket *sock, struct sockaddr *uservaddr, po->chan.private = sk; po->chan.ops = &pppol2tp_chan_ops; po->chan.mtu = pppol2tp_tunnel_mtu(tunnel); + po->chan.direct_xmit = true; error = ppp_register_net_channel(sock_net(sk), &po->chan); if (error) { -- cgit v1.2.3 From d186f405fdf4229d0e9a52ba71662404b06cc002 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Sun, 2 Mar 2025 12:42:36 +0000 Subject: tcp: add RCU management to inet_bind_bucket Add RCU protection to inet_bind_bucket structure. - Add rcu_head field to the structure definition. - Use kfree_rcu() at destroy time, and remove inet_bind_bucket_destroy() first argument. - Use hlist_del_rcu() and hlist_add_head_rcu() methods. Signed-off-by: Eric Dumazet Reviewed-by: Jason Xing Reviewed-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20250302124237.3913746-4-edumazet@google.com Signed-off-by: Jakub Kicinski --- include/net/inet_hashtables.h | 4 ++-- net/ipv4/inet_connection_sock.c | 2 +- net/ipv4/inet_hashtables.c | 14 +++++++------- net/ipv4/inet_timewait_sock.c | 2 +- 4 files changed, 11 insertions(+), 11 deletions(-) (limited to 'include') diff --git a/include/net/inet_hashtables.h b/include/net/inet_hashtables.h index da818fb0205f..1061c4f536a6 100644 --- a/include/net/inet_hashtables.h +++ b/include/net/inet_hashtables.h @@ -89,6 +89,7 @@ struct inet_bind_bucket { bool fast_ipv6_only; struct hlist_node node; struct hlist_head bhash2; + struct rcu_head rcu; }; struct inet_bind2_bucket { @@ -226,8 +227,7 @@ struct inet_bind_bucket * inet_bind_bucket_create(struct kmem_cache *cachep, struct net *net, struct inet_bind_hashbucket *head, const unsigned short snum, int l3mdev); -void inet_bind_bucket_destroy(struct kmem_cache *cachep, - struct inet_bind_bucket *tb); +void inet_bind_bucket_destroy(struct inet_bind_bucket *tb); bool inet_bind_bucket_match(const struct inet_bind_bucket *tb, const struct net *net, unsigned short port, diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c index b4e514da22b6..e93c66034077 100644 --- a/net/ipv4/inet_connection_sock.c +++ b/net/ipv4/inet_connection_sock.c @@ -598,7 +598,7 @@ fail_unlock: if (bhash2_created) inet_bind2_bucket_destroy(hinfo->bind2_bucket_cachep, tb2); if (bhash_created) - inet_bind_bucket_destroy(hinfo->bind_bucket_cachep, tb); + inet_bind_bucket_destroy(tb); } if (head2_lock_acquired) spin_unlock(&head2->lock); diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index 46d39aa2199e..b737e13f8459 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -76,7 +76,7 @@ struct inet_bind_bucket *inet_bind_bucket_create(struct kmem_cache *cachep, tb->fastreuse = 0; tb->fastreuseport = 0; INIT_HLIST_HEAD(&tb->bhash2); - hlist_add_head(&tb->node, &head->chain); + hlist_add_head_rcu(&tb->node, &head->chain); } return tb; } @@ -84,11 +84,11 @@ struct inet_bind_bucket *inet_bind_bucket_create(struct kmem_cache *cachep, /* * Caller must hold hashbucket lock for this tb with local BH disabled */ -void inet_bind_bucket_destroy(struct kmem_cache *cachep, struct inet_bind_bucket *tb) +void inet_bind_bucket_destroy(struct inet_bind_bucket *tb) { if (hlist_empty(&tb->bhash2)) { - __hlist_del(&tb->node); - kmem_cache_free(cachep, tb); + hlist_del_rcu(&tb->node); + kfree_rcu(tb, rcu); } } @@ -201,7 +201,7 @@ static void __inet_put_port(struct sock *sk) } spin_unlock(&head2->lock); - inet_bind_bucket_destroy(hashinfo->bind_bucket_cachep, tb); + inet_bind_bucket_destroy(tb); spin_unlock(&head->lock); } @@ -285,7 +285,7 @@ bhash2_find: error: if (created_inet_bind_bucket) - inet_bind_bucket_destroy(table->bind_bucket_cachep, tb); + inet_bind_bucket_destroy(tb); spin_unlock(&head2->lock); spin_unlock(&head->lock); return -ENOMEM; @@ -1162,7 +1162,7 @@ error: spin_unlock(&head2->lock); if (tb_created) - inet_bind_bucket_destroy(hinfo->bind_bucket_cachep, tb); + inet_bind_bucket_destroy(tb); spin_unlock(&head->lock); if (tw) diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c index 337390ba85b4..aded4bf1bc16 100644 --- a/net/ipv4/inet_timewait_sock.c +++ b/net/ipv4/inet_timewait_sock.c @@ -39,7 +39,7 @@ void inet_twsk_bind_unhash(struct inet_timewait_sock *tw, tw->tw_tb = NULL; tw->tw_tb2 = NULL; inet_bind2_bucket_destroy(hashinfo->bind2_bucket_cachep, tb2); - inet_bind_bucket_destroy(hashinfo->bind_bucket_cachep, tb); + inet_bind_bucket_destroy(tb); __sock_put((struct sock *)tw); } -- cgit v1.2.3 From 86c2bc293b8130aec9fa504e953531a84a6eb9a6 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Sun, 2 Mar 2025 12:42:37 +0000 Subject: tcp: use RCU lookup in __inet_hash_connect() When __inet_hash_connect() has to try many 4-tuples before finding an available one, we see a high spinlock cost from the many spin_lock_bh(&head->lock) performed in its loop. This patch adds an RCU lookup to avoid the spinlock cost. check_established() gets a new @rcu_lookup argument. First reason is to not make any changes while head->lock is not held. Second reason is to not make this RCU lookup a second time after the spinlock has been acquired. Tested: Server: ulimit -n 40000; neper/tcp_crr -T 200 -F 30000 -6 --nolog Client: ulimit -n 40000; neper/tcp_crr -T 200 -F 30000 -6 --nolog -c -H server Before series: utime_start=0.288582 utime_end=1.548707 stime_start=20.637138 stime_end=2002.489845 num_transactions=484453 latency_min=0.156279245 latency_max=20.922042756 latency_mean=1.546521274 latency_stddev=3.936005194 num_samples=312537 throughput=47426.00 perf top on the client: 49.54% [kernel] [k] _raw_spin_lock 25.87% [kernel] [k] _raw_spin_lock_bh 5.97% [kernel] [k] queued_spin_lock_slowpath 5.67% [kernel] [k] __inet_hash_connect 3.53% [kernel] [k] __inet6_check_established 3.48% [kernel] [k] inet6_ehashfn 0.64% [kernel] [k] rcu_all_qs After this series: utime_start=0.271607 utime_end=3.847111 stime_start=18.407684 stime_end=1997.485557 num_transactions=1350742 latency_min=0.014131929 latency_max=17.895073144 latency_mean=0.505675853 # Nice reduction of latency metrics latency_stddev=2.125164772 num_samples=307884 throughput=139866.80 # 190 % increase perf top on client: 56.86% [kernel] [k] __inet6_check_established 17.96% [kernel] [k] __inet_hash_connect 13.88% [kernel] [k] inet6_ehashfn 2.52% [kernel] [k] rcu_all_qs 2.01% [kernel] [k] __cond_resched 0.41% [kernel] [k] _raw_spin_lock Signed-off-by: Eric Dumazet Reviewed-by: Jason Xing Tested-by: Jason Xing Reviewed-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20250302124237.3913746-5-edumazet@google.com Signed-off-by: Jakub Kicinski --- include/net/inet_hashtables.h | 3 ++- net/ipv4/inet_hashtables.c | 52 +++++++++++++++++++++++++++++-------------- net/ipv6/inet6_hashtables.c | 24 +++++++++++--------- 3 files changed, 50 insertions(+), 29 deletions(-) (limited to 'include') diff --git a/include/net/inet_hashtables.h b/include/net/inet_hashtables.h index 1061c4f536a6..f447d61d9598 100644 --- a/include/net/inet_hashtables.h +++ b/include/net/inet_hashtables.h @@ -529,7 +529,8 @@ int __inet_hash_connect(struct inet_timewait_death_row *death_row, struct sock *sk, u64 port_offset, int (*check_established)(struct inet_timewait_death_row *, struct sock *, __u16, - struct inet_timewait_sock **)); + struct inet_timewait_sock **, + bool rcu_lookup)); int inet_hash_connect(struct inet_timewait_death_row *death_row, struct sock *sk); diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index b737e13f8459..d1b5f45ee718 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -537,7 +537,8 @@ EXPORT_SYMBOL_GPL(__inet_lookup_established); /* called with local bh disabled */ static int __inet_check_established(struct inet_timewait_death_row *death_row, struct sock *sk, __u16 lport, - struct inet_timewait_sock **twp) + struct inet_timewait_sock **twp, + bool rcu_lookup) { struct inet_hashinfo *hinfo = death_row->hashinfo; struct inet_sock *inet = inet_sk(sk); @@ -556,17 +557,17 @@ static int __inet_check_established(struct inet_timewait_death_row *death_row, struct sock *sk2; spinlock_t *lock; - rcu_read_lock(); - sk_nulls_for_each(sk2, node, &head->chain) { - if (sk2->sk_hash != hash || - !inet_match(net, sk2, acookie, ports, dif, sdif)) - continue; - if (sk2->sk_state == TCP_TIME_WAIT) - break; - rcu_read_unlock(); - return -EADDRNOTAVAIL; + if (rcu_lookup) { + sk_nulls_for_each(sk2, node, &head->chain) { + if (sk2->sk_hash != hash || + !inet_match(net, sk2, acookie, ports, dif, sdif)) + continue; + if (sk2->sk_state == TCP_TIME_WAIT) + break; + return -EADDRNOTAVAIL; + } + return 0; } - rcu_read_unlock(); lock = inet_ehash_lockp(hinfo, hash); spin_lock(lock); @@ -1007,7 +1008,8 @@ static u32 *table_perturb; int __inet_hash_connect(struct inet_timewait_death_row *death_row, struct sock *sk, u64 port_offset, int (*check_established)(struct inet_timewait_death_row *, - struct sock *, __u16, struct inet_timewait_sock **)) + struct sock *, __u16, struct inet_timewait_sock **, + bool rcu_lookup)) { struct inet_hashinfo *hinfo = death_row->hashinfo; struct inet_bind_hashbucket *head, *head2; @@ -1025,7 +1027,7 @@ int __inet_hash_connect(struct inet_timewait_death_row *death_row, if (port) { local_bh_disable(); - ret = check_established(death_row, sk, port, NULL); + ret = check_established(death_row, sk, port, NULL, false); local_bh_enable(); return ret; } @@ -1061,6 +1063,21 @@ other_parity_scan: continue; head = &hinfo->bhash[inet_bhashfn(net, port, hinfo->bhash_size)]; + rcu_read_lock(); + hlist_for_each_entry_rcu(tb, &head->chain, node) { + if (!inet_bind_bucket_match(tb, net, port, l3mdev)) + continue; + if (tb->fastreuse >= 0 || tb->fastreuseport >= 0) { + rcu_read_unlock(); + goto next_port; + } + if (!check_established(death_row, sk, port, &tw, true)) + break; + rcu_read_unlock(); + goto next_port; + } + rcu_read_unlock(); + spin_lock_bh(&head->lock); /* Does not bother with rcv_saddr checks, because @@ -1070,12 +1087,12 @@ other_parity_scan: if (inet_bind_bucket_match(tb, net, port, l3mdev)) { if (tb->fastreuse >= 0 || tb->fastreuseport >= 0) - goto next_port; + goto next_port_unlock; WARN_ON(hlist_empty(&tb->bhash2)); if (!check_established(death_row, sk, - port, &tw)) + port, &tw, false)) goto ok; - goto next_port; + goto next_port_unlock; } } @@ -1089,8 +1106,9 @@ other_parity_scan: tb->fastreuse = -1; tb->fastreuseport = -1; goto ok; -next_port: +next_port_unlock: spin_unlock_bh(&head->lock); +next_port: cond_resched(); } diff --git a/net/ipv6/inet6_hashtables.c b/net/ipv6/inet6_hashtables.c index 3604a5cae5d2..9be315496459 100644 --- a/net/ipv6/inet6_hashtables.c +++ b/net/ipv6/inet6_hashtables.c @@ -263,7 +263,8 @@ EXPORT_SYMBOL_GPL(inet6_lookup); static int __inet6_check_established(struct inet_timewait_death_row *death_row, struct sock *sk, const __u16 lport, - struct inet_timewait_sock **twp) + struct inet_timewait_sock **twp, + bool rcu_lookup) { struct inet_hashinfo *hinfo = death_row->hashinfo; struct inet_sock *inet = inet_sk(sk); @@ -281,17 +282,18 @@ static int __inet6_check_established(struct inet_timewait_death_row *death_row, struct sock *sk2; spinlock_t *lock; - rcu_read_lock(); - sk_nulls_for_each(sk2, node, &head->chain) { - if (sk2->sk_hash != hash || - !inet6_match(net, sk2, saddr, daddr, ports, dif, sdif)) - continue; - if (sk2->sk_state == TCP_TIME_WAIT) - break; - rcu_read_unlock(); - return -EADDRNOTAVAIL; + if (rcu_lookup) { + sk_nulls_for_each(sk2, node, &head->chain) { + if (sk2->sk_hash != hash || + !inet6_match(net, sk2, saddr, daddr, + ports, dif, sdif)) + continue; + if (sk2->sk_state == TCP_TIME_WAIT) + break; + return -EADDRNOTAVAIL; + } + return 0; } - rcu_read_unlock(); lock = inet_ehash_lockp(hinfo, hash); spin_lock(lock); -- cgit v1.2.3 From e859d375d1694488015e6804bfeea527a0b25b9f Mon Sep 17 00:00:00 2001 From: Wojtek Wasko Date: Mon, 3 Mar 2025 18:13:43 +0200 Subject: posix-clock: Store file pointer in struct posix_clock_context File descriptor based pc_clock_*() operations of dynamic posix clocks have access to the file pointer and implement permission checks in the generic code before invoking the relevant dynamic clock callback. Character device operations (open, read, poll, ioctl) do not implement a generic permission control and the dynamic clock callbacks have no access to the file pointer to implement them. Extend struct posix_clock_context with a struct file pointer and initialize it in posix_clock_open(), so that all dynamic clock callbacks can access it. Acked-by: Richard Cochran Reviewed-by: Vadim Fedorenko Reviewed-by: Thomas Gleixner Signed-off-by: Wojtek Wasko Signed-off-by: David S. Miller --- include/linux/posix-clock.h | 6 +++++- kernel/time/posix-clock.c | 1 + 2 files changed, 6 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/posix-clock.h b/include/linux/posix-clock.h index ef8619f48920..a500d3160fe8 100644 --- a/include/linux/posix-clock.h +++ b/include/linux/posix-clock.h @@ -95,10 +95,13 @@ struct posix_clock { * struct posix_clock_context - represents clock file operations context * * @clk: Pointer to the clock + * @fp: Pointer to the file used to open the clock * @private_clkdata: Pointer to user data * * Drivers should use struct posix_clock_context during specific character - * device file operation methods to access the posix clock. + * device file operation methods to access the posix clock. In particular, + * the file pointer can be used to verify correct access mode for ioctl() + * calls. * * Drivers can store a private data structure during the open operation * if they have specific information that is required in other file @@ -106,6 +109,7 @@ struct posix_clock { */ struct posix_clock_context { struct posix_clock *clk; + struct file *fp; void *private_clkdata; }; diff --git a/kernel/time/posix-clock.c b/kernel/time/posix-clock.c index 1af0bb2cc45c..4e114e34a6e0 100644 --- a/kernel/time/posix-clock.c +++ b/kernel/time/posix-clock.c @@ -129,6 +129,7 @@ static int posix_clock_open(struct inode *inode, struct file *fp) goto out; } pccontext->clk = clk; + pccontext->fp = fp; if (clk->ops.open) { err = clk->ops.open(pccontext, fp->f_mode); if (err) { -- cgit v1.2.3 From c27c66afc449b80f3b4b84d123358c0248f2cf63 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Wed, 5 Mar 2025 07:08:09 -1000 Subject: fs/pipe: Fix pipe_occupancy() with 16-bit indexes The pipe_occupancy() logic implicitly relied on the natural unsigned modulo arithmetic in C, but that doesn't work for the new 'pipe_index_t' case, since any arithmetic will be done in 'int' (and here we had also made it 'unsigned int' due to the function call boundary). So make the modulo arithmetic explicit by casting the result to the proper type. Cc: Oleg Nesterov Cc: Mateusz Guzik Cc: Manfred Spraul Cc: Christian Brauner Cc: Swapnil Sapkal Cc: Alexey Gladkov Cc: K Prateek Nayak Link: https://lore.kernel.org/all/CAHk-=wjyHsGLx=rxg6PKYBNkPYAejgo7=CbyL3=HGLZLsAaJFQ@mail.gmail.com/ Fixes: 3d252160b818 ("fs/pipe: Read pipe->{head,tail} atomically outside pipe->mutex") Signed-off-by: Linus Torvalds --- include/linux/pipe_fs_i.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h index 3cc4f8eab853..1f013ed7577e 100644 --- a/include/linux/pipe_fs_i.h +++ b/include/linux/pipe_fs_i.h @@ -192,7 +192,7 @@ static inline bool pipe_empty(unsigned int head, unsigned int tail) */ static inline unsigned int pipe_occupancy(unsigned int head, unsigned int tail) { - return head - tail; + return (pipe_index_t)(head - tail); } /** -- cgit v1.2.3 From cfced12f5100e50d56bc587299393fd33c1169a9 Mon Sep 17 00:00:00 2001 From: K Prateek Nayak Date: Wed, 5 Mar 2025 11:23:01 +0000 Subject: include/linux/pipe_fs_i: Add htmldoc annotation for "head_tail" member Add htmldoc annotation for the newly introduced "head_tail" member describing it to be a union of the pipe_inode_info's @head and @tail members. Reported-by: Stephen Rothwell Closes: https://lore.kernel.org/lkml/20250305204609.5e64768e@canb.auug.org.au/ Fixes: 3d252160b818 ("fs/pipe: Read pipe->{head,tail} atomically outside pipe->mutex") Signed-off-by: K Prateek Nayak Signed-off-by: Linus Torvalds --- include/linux/pipe_fs_i.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h index 1f013ed7577e..05ccbc5d0129 100644 --- a/include/linux/pipe_fs_i.h +++ b/include/linux/pipe_fs_i.h @@ -65,6 +65,7 @@ union pipe_index { * @wr_wait: writer wait point in case of full pipe * @head: The point of buffer production * @tail: The point of buffer consumption + * @head_tail: unsigned long union of @head and @tail * @note_loss: The next read() should insert a data-lost message * @max_usage: The maximum number of slots that may be used in the ring * @ring_size: total number of buffers (should be a power of 2) -- cgit v1.2.3 From 0d2d0f3d93ddd6556f23c917d910becd9925ddeb Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Wed, 5 Mar 2025 07:35:40 -1000 Subject: fs/pipe: remove buggy and unused 'helper' function While looking for incorrect users of the pipe head/tail fields (see commit c27c66afc449: "fs/pipe: Fix pipe_occupancy() with 16-bit indexes"), I found a bug in pipe_discard_from() that looked entirely broken. However, the fix is trivial: this buggy function isn't actually called by anything, so let's just remove it ASAP. Signed-off-by: Linus Torvalds --- include/linux/pipe_fs_i.h | 9 --------- 1 file changed, 9 deletions(-) (limited to 'include') diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h index 05ccbc5d0129..e572e6fc4f81 100644 --- a/include/linux/pipe_fs_i.h +++ b/include/linux/pipe_fs_i.h @@ -281,15 +281,6 @@ static inline bool pipe_buf_try_steal(struct pipe_inode_info *pipe, return buf->ops->try_steal(pipe, buf); } -static inline void pipe_discard_from(struct pipe_inode_info *pipe, - unsigned int old_head) -{ - unsigned int mask = pipe->ring_size - 1; - - while (pipe->head > old_head) - pipe_buf_release(pipe, &pipe->bufs[--pipe->head & mask]); -} - /* Differs from PIPE_BUF in that PIPE_SIZE is the length of the actual memory allocation, whereas PIPE_BUF makes atomicity guarantees. */ #define PIPE_SIZE PAGE_SIZE -- cgit v1.2.3 From 7e2f7e25f6ffc229fd5be0488e1d8aac7bdd80e8 Mon Sep 17 00:00:00 2001 From: "David E. Box" Date: Thu, 27 Feb 2025 20:15:19 +0800 Subject: arch: x86: add IPC mailbox accessor function and add SoC register access MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Exports intel_pmc_ipc() for host access to the PMC IPC mailbox - Enables the host to access specific SoC registers through the PMC firmware using IPC commands. This access method is necessary for registers that are not available through direct Memory-Mapped I/O (MMIO), which is used for other accessible parts of the PMC. Signed-off-by: David E. Box Signed-off-by: Chao Qin Signed-off-by: Choong Yong Liang Acked-by: Ilpo Järvinen Link: https://patch.msgid.link/20250227121522.1802832-4-yong.liang.choong@linux.intel.com Signed-off-by: Jakub Kicinski --- MAINTAINERS | 1 + include/linux/platform_data/x86/intel_pmc_ipc.h | 94 +++++++++++++++++++++++++ 2 files changed, 95 insertions(+) create mode 100644 include/linux/platform_data/x86/intel_pmc_ipc.h (limited to 'include') diff --git a/MAINTAINERS b/MAINTAINERS index 7078199fcebf..2eccaff95ca2 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -11862,6 +11862,7 @@ L: platform-driver-x86@vger.kernel.org S: Maintained F: Documentation/ABI/testing/sysfs-platform-intel-pmc F: drivers/platform/x86/intel/pmc/ +F: linux/platform_data/x86/intel_pmc_ipc.h INTEL PMIC GPIO DRIVERS M: Andy Shevchenko diff --git a/include/linux/platform_data/x86/intel_pmc_ipc.h b/include/linux/platform_data/x86/intel_pmc_ipc.h new file mode 100644 index 000000000000..6e603a8c075f --- /dev/null +++ b/include/linux/platform_data/x86/intel_pmc_ipc.h @@ -0,0 +1,94 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Intel Core SoC Power Management Controller Header File + * + * Copyright (c) 2025, Intel Corporation. + * All Rights Reserved. + * + */ +#ifndef INTEL_PMC_IPC_H +#define INTEL_PMC_IPC_H +#include + +#define IPC_SOC_REGISTER_ACCESS 0xAA +#define IPC_SOC_SUB_CMD_READ 0x00 +#define IPC_SOC_SUB_CMD_WRITE 0x01 +#define PMC_IPCS_PARAM_COUNT 7 +#define VALID_IPC_RESPONSE 5 + +struct pmc_ipc_cmd { + u32 cmd; + u32 sub_cmd; + u32 size; + u32 wbuf[4]; +}; + +struct pmc_ipc_rbuf { + u32 buf[4]; +}; + +/** + * intel_pmc_ipc() - PMC IPC Mailbox accessor + * @ipc_cmd: Prepared input command to send + * @rbuf: Allocated array for returned IPC data + * + * Return: 0 on success. Non-zero on mailbox error + */ +static inline int intel_pmc_ipc(struct pmc_ipc_cmd *ipc_cmd, struct pmc_ipc_rbuf *rbuf) +{ + struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL }; + union acpi_object params[PMC_IPCS_PARAM_COUNT] = { + {.type = ACPI_TYPE_INTEGER,}, + {.type = ACPI_TYPE_INTEGER,}, + {.type = ACPI_TYPE_INTEGER,}, + {.type = ACPI_TYPE_INTEGER,}, + {.type = ACPI_TYPE_INTEGER,}, + {.type = ACPI_TYPE_INTEGER,}, + {.type = ACPI_TYPE_INTEGER,}, + }; + struct acpi_object_list arg_list = { PMC_IPCS_PARAM_COUNT, params }; + union acpi_object *obj; + int status; + + if (!ipc_cmd || !rbuf) + return -EINVAL; + + /* + * 0: IPC Command + * 1: IPC Sub Command + * 2: Size + * 3-6: Write Buffer for offset + */ + params[0].integer.value = ipc_cmd->cmd; + params[1].integer.value = ipc_cmd->sub_cmd; + params[2].integer.value = ipc_cmd->size; + params[3].integer.value = ipc_cmd->wbuf[0]; + params[4].integer.value = ipc_cmd->wbuf[1]; + params[5].integer.value = ipc_cmd->wbuf[2]; + params[6].integer.value = ipc_cmd->wbuf[3]; + + status = acpi_evaluate_object(NULL, "\\IPCS", &arg_list, &buffer); + if (ACPI_FAILURE(status)) + return -ENODEV; + + obj = buffer.pointer; + + if (obj && obj->type == ACPI_TYPE_PACKAGE && + obj->package.count == VALID_IPC_RESPONSE) { + const union acpi_object *objs = obj->package.elements; + + if ((u8)objs[0].integer.value != 0) + return -EINVAL; + + rbuf->buf[0] = objs[1].integer.value; + rbuf->buf[1] = objs[2].integer.value; + rbuf->buf[2] = objs[3].integer.value; + rbuf->buf[3] = objs[4].integer.value; + } else { + return -EINVAL; + } + + return 0; +} + +#endif /* INTEL_PMC_IPC_H */ -- cgit v1.2.3 From e654cfc718d451bdf6c1554efc945171239f03f5 Mon Sep 17 00:00:00 2001 From: Choong Yong Liang Date: Thu, 27 Feb 2025 20:15:20 +0800 Subject: net: stmmac: configure SerDes on mac_finish SerDes will configure according to the provided interface mode after finish a major reconfiguration of the interface mode. Reviewed-by: Russell King (Oracle) Signed-off-by: Choong Yong Liang Link: https://patch.msgid.link/20250227121522.1802832-5-yong.liang.choong@linux.intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 13 +++++++++++++ include/linux/stmmac.h | 4 ++++ 2 files changed, 17 insertions(+) (limited to 'include') diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c index aec230353ac4..1a928e21b286 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c @@ -1129,6 +1129,18 @@ static int stmmac_mac_enable_tx_lpi(struct phylink_config *config, u32 timer, return 0; } +static int stmmac_mac_finish(struct phylink_config *config, unsigned int mode, + phy_interface_t interface) +{ + struct net_device *ndev = to_net_dev(config->dev); + struct stmmac_priv *priv = netdev_priv(ndev); + + if (priv->plat->mac_finish) + priv->plat->mac_finish(ndev, priv->plat->bsp_priv, mode, interface); + + return 0; +} + static const struct phylink_mac_ops stmmac_phylink_mac_ops = { .mac_get_caps = stmmac_mac_get_caps, .mac_select_pcs = stmmac_mac_select_pcs, @@ -1137,6 +1149,7 @@ static const struct phylink_mac_ops stmmac_phylink_mac_ops = { .mac_link_up = stmmac_mac_link_up, .mac_disable_tx_lpi = stmmac_mac_disable_tx_lpi, .mac_enable_tx_lpi = stmmac_mac_enable_tx_lpi, + .mac_finish = stmmac_mac_finish, }; /** diff --git a/include/linux/stmmac.h b/include/linux/stmmac.h index cd0d1383df87..b6f03ab12595 100644 --- a/include/linux/stmmac.h +++ b/include/linux/stmmac.h @@ -239,6 +239,10 @@ struct plat_stmmacenet_data { int (*serdes_powerup)(struct net_device *ndev, void *priv); void (*serdes_powerdown)(struct net_device *ndev, void *priv); void (*speed_mode_2500)(struct net_device *ndev, void *priv); + int (*mac_finish)(struct net_device *ndev, + void *priv, + unsigned int mode, + phy_interface_t interface); void (*ptp_clk_freq_config)(struct stmmac_priv *priv); int (*init)(struct platform_device *pdev, void *priv); void (*exit)(struct platform_device *pdev, void *priv); -- cgit v1.2.3 From e7f984e925d23f8fd0469e344a8dc225e1a1b0ab Mon Sep 17 00:00:00 2001 From: Heiner Kallweit Date: Mon, 3 Mar 2025 21:18:46 +0100 Subject: net: phy: move PHY package related code from phy.h to phy_package.c Move PHY package related inline functions from phy.h to phy_package.c. While doing so remove locked versions phy_package_read() and phy_package_write() which have no user. Signed-off-by: Heiner Kallweit Link: https://patch.msgid.link/a4518379-7a5d-45f3-831c-b7fde6512c65@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/phy/bcm54140.c | 1 + drivers/net/phy/mscc/mscc_main.c | 2 + drivers/net/phy/phy-core.c | 1 + drivers/net/phy/phy_package.c | 61 +++++++++++++++++++++++++++ drivers/net/phy/phylib-internal.h | 2 + drivers/net/phy/phylib.h | 6 +++ include/linux/phy.h | 86 --------------------------------------- 7 files changed, 73 insertions(+), 86 deletions(-) (limited to 'include') diff --git a/drivers/net/phy/bcm54140.c b/drivers/net/phy/bcm54140.c index 7969345f6b35..a8edf45fd733 100644 --- a/drivers/net/phy/bcm54140.c +++ b/drivers/net/phy/bcm54140.c @@ -10,6 +10,7 @@ #include #include +#include "phylib.h" #include "bcm-phy-lib.h" /* RDB per-port registers diff --git a/drivers/net/phy/mscc/mscc_main.c b/drivers/net/phy/mscc/mscc_main.c index 19cf12ee8990..7ff975efd8e7 100644 --- a/drivers/net/phy/mscc/mscc_main.c +++ b/drivers/net/phy/mscc/mscc_main.c @@ -17,6 +17,8 @@ #include #include #include + +#include "../phylib.h" #include "mscc_serdes.h" #include "mscc.h" diff --git a/drivers/net/phy/phy-core.c b/drivers/net/phy/phy-core.c index b1c1670de23b..154d29be6293 100644 --- a/drivers/net/phy/phy-core.c +++ b/drivers/net/phy/phy-core.c @@ -6,6 +6,7 @@ #include #include +#include "phylib.h" #include "phylib-internal.h" /** diff --git a/drivers/net/phy/phy_package.c b/drivers/net/phy/phy_package.c index 873420e58f02..12c92d26ee0b 100644 --- a/drivers/net/phy/phy_package.c +++ b/drivers/net/phy/phy_package.c @@ -7,6 +7,7 @@ #include #include "phylib.h" +#include "phylib-internal.h" struct device_node *phy_package_get_node(struct phy_device *phydev) { @@ -20,6 +21,66 @@ void *phy_package_get_priv(struct phy_device *phydev) } EXPORT_SYMBOL_GPL(phy_package_get_priv); +int phy_package_address(struct phy_device *phydev, unsigned int addr_offset) +{ + struct phy_package_shared *shared = phydev->shared; + u8 base_addr = shared->base_addr; + + if (addr_offset >= PHY_MAX_ADDR - base_addr) + return -EIO; + + /* we know that addr will be in the range 0..31 and thus the + * implicit cast to a signed int is not a problem. + */ + return base_addr + addr_offset; +} + +int __phy_package_read(struct phy_device *phydev, unsigned int addr_offset, + u32 regnum) +{ + int addr = phy_package_address(phydev, addr_offset); + + if (addr < 0) + return addr; + + return __mdiobus_read(phydev->mdio.bus, addr, regnum); +} +EXPORT_SYMBOL_GPL(__phy_package_read); + +int __phy_package_write(struct phy_device *phydev, unsigned int addr_offset, + u32 regnum, u16 val) +{ + int addr = phy_package_address(phydev, addr_offset); + + if (addr < 0) + return addr; + + return __mdiobus_write(phydev->mdio.bus, addr, regnum, val); +} +EXPORT_SYMBOL_GPL(__phy_package_write); + +static bool __phy_package_set_once(struct phy_device *phydev, unsigned int b) +{ + struct phy_package_shared *shared = phydev->shared; + + if (!shared) + return false; + + return !test_and_set_bit(b, &shared->flags); +} + +bool phy_package_init_once(struct phy_device *phydev) +{ + return __phy_package_set_once(phydev, 0); +} +EXPORT_SYMBOL_GPL(phy_package_init_once); + +bool phy_package_probe_once(struct phy_device *phydev) +{ + return __phy_package_set_once(phydev, 1); +} +EXPORT_SYMBOL_GPL(phy_package_probe_once); + /** * phy_package_join - join a common PHY group * @phydev: target phy_device struct diff --git a/drivers/net/phy/phylib-internal.h b/drivers/net/phy/phylib-internal.h index dc9592c6bb8e..afac2bd15b50 100644 --- a/drivers/net/phy/phylib-internal.h +++ b/drivers/net/phy/phylib-internal.h @@ -20,6 +20,8 @@ void of_set_phy_timing_role(struct phy_device *phydev); int phy_speed_down_core(struct phy_device *phydev); void phy_check_downshift(struct phy_device *phydev); +int phy_package_address(struct phy_device *phydev, unsigned int addr_offset); + int genphy_c45_read_eee_adv(struct phy_device *phydev, unsigned long *adv); #endif /* __PHYLIB_INTERNAL_H */ diff --git a/drivers/net/phy/phylib.h b/drivers/net/phy/phylib.h index a42e1fc079cb..06c50d275b90 100644 --- a/drivers/net/phy/phylib.h +++ b/drivers/net/phy/phylib.h @@ -11,5 +11,11 @@ struct phy_device; struct device_node *phy_package_get_node(struct phy_device *phydev); void *phy_package_get_priv(struct phy_device *phydev); +int __phy_package_read(struct phy_device *phydev, unsigned int addr_offset, + u32 regnum); +int __phy_package_write(struct phy_device *phydev, unsigned int addr_offset, + u32 regnum, u16 val); +bool phy_package_init_once(struct phy_device *phydev); +bool phy_package_probe_once(struct phy_device *phydev); #endif /* __PHYLIB_H */ diff --git a/include/linux/phy.h b/include/linux/phy.h index 7bfbae51070a..2b12d1bef3cc 100644 --- a/include/linux/phy.h +++ b/include/linux/phy.h @@ -350,10 +350,6 @@ struct phy_package_shared { void *priv; }; -/* used as bit number in atomic bitops */ -#define PHY_SHARED_F_INIT_DONE 0 -#define PHY_SHARED_F_PROBE_DONE 1 - /** * struct mii_bus - Represents an MDIO bus * @@ -2149,67 +2145,6 @@ int __phy_hwtstamp_set(struct phy_device *phydev, struct kernel_hwtstamp_config *config, struct netlink_ext_ack *extack); -static inline int phy_package_address(struct phy_device *phydev, - unsigned int addr_offset) -{ - struct phy_package_shared *shared = phydev->shared; - u8 base_addr = shared->base_addr; - - if (addr_offset >= PHY_MAX_ADDR - base_addr) - return -EIO; - - /* we know that addr will be in the range 0..31 and thus the - * implicit cast to a signed int is not a problem. - */ - return base_addr + addr_offset; -} - -static inline int phy_package_read(struct phy_device *phydev, - unsigned int addr_offset, u32 regnum) -{ - int addr = phy_package_address(phydev, addr_offset); - - if (addr < 0) - return addr; - - return mdiobus_read(phydev->mdio.bus, addr, regnum); -} - -static inline int __phy_package_read(struct phy_device *phydev, - unsigned int addr_offset, u32 regnum) -{ - int addr = phy_package_address(phydev, addr_offset); - - if (addr < 0) - return addr; - - return __mdiobus_read(phydev->mdio.bus, addr, regnum); -} - -static inline int phy_package_write(struct phy_device *phydev, - unsigned int addr_offset, u32 regnum, - u16 val) -{ - int addr = phy_package_address(phydev, addr_offset); - - if (addr < 0) - return addr; - - return mdiobus_write(phydev->mdio.bus, addr, regnum, val); -} - -static inline int __phy_package_write(struct phy_device *phydev, - unsigned int addr_offset, u32 regnum, - u16 val) -{ - int addr = phy_package_address(phydev, addr_offset); - - if (addr < 0) - return addr; - - return __mdiobus_write(phydev->mdio.bus, addr, regnum, val); -} - int __phy_package_read_mmd(struct phy_device *phydev, unsigned int addr_offset, int devad, u32 regnum); @@ -2226,27 +2161,6 @@ int phy_package_write_mmd(struct phy_device *phydev, unsigned int addr_offset, int devad, u32 regnum, u16 val); -static inline bool __phy_package_set_once(struct phy_device *phydev, - unsigned int b) -{ - struct phy_package_shared *shared = phydev->shared; - - if (!shared) - return false; - - return !test_and_set_bit(b, &shared->flags); -} - -static inline bool phy_package_init_once(struct phy_device *phydev) -{ - return __phy_package_set_once(phydev, PHY_SHARED_F_INIT_DONE); -} - -static inline bool phy_package_probe_once(struct phy_device *phydev) -{ - return __phy_package_set_once(phydev, PHY_SHARED_F_PROBE_DONE); -} - extern const struct bus_type mdio_bus_type; struct mdio_board_info { -- cgit v1.2.3 From a4002849776948f257d564944e8fd7a727362ed9 Mon Sep 17 00:00:00 2001 From: Heiner Kallweit Date: Mon, 3 Mar 2025 21:19:25 +0100 Subject: net: phy: remove remaining PHY package related definitions from phy.h Move definition of struct phy_package_shared to phy_package.c, and move remaining PHY package related declarations from phy.h to phylib.h, thus making them accessible for PHY drivers only. Signed-off-by: Heiner Kallweit Link: https://patch.msgid.link/211e14b6-e2f8-43d7-b533-3628ec548456@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/phy/phy_package.c | 31 +++++++++++++++++++++++++++++++ drivers/net/phy/phylib.h | 7 +++++++ include/linux/phy.h | 38 -------------------------------------- 3 files changed, 38 insertions(+), 38 deletions(-) (limited to 'include') diff --git a/drivers/net/phy/phy_package.c b/drivers/net/phy/phy_package.c index 12c92d26ee0b..c738f76e8664 100644 --- a/drivers/net/phy/phy_package.c +++ b/drivers/net/phy/phy_package.c @@ -9,6 +9,37 @@ #include "phylib.h" #include "phylib-internal.h" +/** + * struct phy_package_shared - Shared information in PHY packages + * @base_addr: Base PHY address of PHY package used to combine PHYs + * in one package and for offset calculation of phy_package_read/write + * @np: Pointer to the Device Node if PHY package defined in DT + * @refcnt: Number of PHYs connected to this shared data + * @flags: Initialization of PHY package + * @priv_size: Size of the shared private data @priv + * @priv: Driver private data shared across a PHY package + * + * Represents a shared structure between different phydev's in the same + * package, for example a quad PHY. See phy_package_join() and + * phy_package_leave(). + */ +struct phy_package_shared { + u8 base_addr; + /* With PHY package defined in DT this points to the PHY package node */ + struct device_node *np; + refcount_t refcnt; + unsigned long flags; + size_t priv_size; + + /* private data pointer */ + /* note that this pointer is shared between different phydevs and + * the user has to take care of appropriate locking. It is allocated + * and freed automatically by phy_package_join() and + * phy_package_leave(). + */ + void *priv; +}; + struct device_node *phy_package_get_node(struct phy_device *phydev) { return phydev->shared->np; diff --git a/drivers/net/phy/phylib.h b/drivers/net/phy/phylib.h index 06c50d275b90..f0e499fed1f7 100644 --- a/drivers/net/phy/phylib.h +++ b/drivers/net/phy/phylib.h @@ -17,5 +17,12 @@ int __phy_package_write(struct phy_device *phydev, unsigned int addr_offset, u32 regnum, u16 val); bool phy_package_init_once(struct phy_device *phydev); bool phy_package_probe_once(struct phy_device *phydev); +int phy_package_join(struct phy_device *phydev, int base_addr, size_t priv_size); +int of_phy_package_join(struct phy_device *phydev, size_t priv_size); +void phy_package_leave(struct phy_device *phydev); +int devm_phy_package_join(struct device *dev, struct phy_device *phydev, + int base_addr, size_t priv_size); +int devm_of_phy_package_join(struct device *dev, struct phy_device *phydev, + size_t priv_size); #endif /* __PHYLIB_H */ diff --git a/include/linux/phy.h b/include/linux/phy.h index 2b12d1bef3cc..c4a6385faf41 100644 --- a/include/linux/phy.h +++ b/include/linux/phy.h @@ -319,37 +319,6 @@ struct mdio_bus_stats { struct u64_stats_sync syncp; }; -/** - * struct phy_package_shared - Shared information in PHY packages - * @base_addr: Base PHY address of PHY package used to combine PHYs - * in one package and for offset calculation of phy_package_read/write - * @np: Pointer to the Device Node if PHY package defined in DT - * @refcnt: Number of PHYs connected to this shared data - * @flags: Initialization of PHY package - * @priv_size: Size of the shared private data @priv - * @priv: Driver private data shared across a PHY package - * - * Represents a shared structure between different phydev's in the same - * package, for example a quad PHY. See phy_package_join() and - * phy_package_leave(). - */ -struct phy_package_shared { - u8 base_addr; - /* With PHY package defined in DT this points to the PHY package node */ - struct device_node *np; - refcount_t refcnt; - unsigned long flags; - size_t priv_size; - - /* private data pointer */ - /* note that this pointer is shared between different phydevs and - * the user has to take care of appropriate locking. It is allocated - * and freed automatically by phy_package_join() and - * phy_package_leave(). - */ - void *priv; -}; - /** * struct mii_bus - Represents an MDIO bus * @@ -2109,13 +2078,6 @@ int phy_ethtool_get_link_ksettings(struct net_device *ndev, int phy_ethtool_set_link_ksettings(struct net_device *ndev, const struct ethtool_link_ksettings *cmd); int phy_ethtool_nway_reset(struct net_device *ndev); -int phy_package_join(struct phy_device *phydev, int base_addr, size_t priv_size); -int of_phy_package_join(struct phy_device *phydev, size_t priv_size); -void phy_package_leave(struct phy_device *phydev); -int devm_phy_package_join(struct device *dev, struct phy_device *phydev, - int base_addr, size_t priv_size); -int devm_of_phy_package_join(struct device *dev, struct phy_device *phydev, - size_t priv_size); int __init mdio_bus_init(void); void mdio_bus_exit(void); -- cgit v1.2.3 From f130a0cc1b4ff1ef28a307428d40436032e2b66e Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 4 Mar 2025 12:59:18 +0000 Subject: inet: fix lwtunnel_valid_encap_type() lock imbalance After blamed commit rtm_to_fib_config() now calls lwtunnel_valid_encap_type{_attr}() without RTNL held, triggering an unlock balance in __rtnl_unlock, as reported by syzbot [1] IPv6 and rtm_to_nh_config() are not yet converted. Add a temporary @rtnl_is_held parameter to lwtunnel_valid_encap_type() and lwtunnel_valid_encap_type_attr(). While we are at it replace the two rcu_dereference() in lwtunnel_valid_encap_type() with more appropriate rcu_access_pointer(). [1] syz-executor245/5836 is trying to release lock (rtnl_mutex) at: [] __rtnl_unlock+0x6c/0xf0 net/core/rtnetlink.c:142 but there are no more locks to release! other info that might help us debug this: no locks held by syz-executor245/5836. stack backtrace: CPU: 0 UID: 0 PID: 5836 Comm: syz-executor245 Not tainted 6.14.0-rc4-syzkaller-00873-g3424291dd242 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 02/12/2025 Call Trace: __dump_stack lib/dump_stack.c:94 [inline] dump_stack_lvl+0x241/0x360 lib/dump_stack.c:120 print_unlock_imbalance_bug+0x25b/0x2d0 kernel/locking/lockdep.c:5289 __lock_release kernel/locking/lockdep.c:5518 [inline] lock_release+0x47e/0xa30 kernel/locking/lockdep.c:5872 __mutex_unlock_slowpath+0xec/0x800 kernel/locking/mutex.c:891 __rtnl_unlock+0x6c/0xf0 net/core/rtnetlink.c:142 lwtunnel_valid_encap_type+0x38a/0x5f0 net/core/lwtunnel.c:169 lwtunnel_valid_encap_type_attr+0x113/0x270 net/core/lwtunnel.c:209 rtm_to_fib_config+0x949/0x14e0 net/ipv4/fib_frontend.c:808 inet_rtm_newroute+0xf6/0x2a0 net/ipv4/fib_frontend.c:917 rtnetlink_rcv_msg+0x791/0xcf0 net/core/rtnetlink.c:6919 netlink_rcv_skb+0x206/0x480 net/netlink/af_netlink.c:2534 netlink_unicast_kernel net/netlink/af_netlink.c:1313 [inline] netlink_unicast+0x7f6/0x990 net/netlink/af_netlink.c:1339 netlink_sendmsg+0x8de/0xcb0 net/netlink/af_netlink.c:1883 sock_sendmsg_nosec net/socket.c:709 [inline] Fixes: 1dd2af7963e9 ("ipv4: fib: Convert RTM_NEWROUTE and RTM_DELROUTE to per-netns RTNL.") Reported-by: syzbot+3f18ef0f7df107a3f6a0@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/67c6f87a.050a0220.38b91b.0147.GAE@google.com/T/#u Signed-off-by: Eric Dumazet Reviewed-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20250304125918.2763514-1-edumazet@google.com Signed-off-by: Jakub Kicinski --- include/net/lwtunnel.h | 12 ++++++++---- net/core/lwtunnel.c | 23 ++++++++++++----------- net/ipv4/fib_frontend.c | 4 ++-- net/ipv4/nexthop.c | 3 ++- net/ipv6/route.c | 6 ++++-- 5 files changed, 28 insertions(+), 20 deletions(-) (limited to 'include') diff --git a/include/net/lwtunnel.h b/include/net/lwtunnel.h index 53bd2d02a4f0..39cd50300a18 100644 --- a/include/net/lwtunnel.h +++ b/include/net/lwtunnel.h @@ -116,9 +116,11 @@ int lwtunnel_encap_add_ops(const struct lwtunnel_encap_ops *op, int lwtunnel_encap_del_ops(const struct lwtunnel_encap_ops *op, unsigned int num); int lwtunnel_valid_encap_type(u16 encap_type, - struct netlink_ext_ack *extack); + struct netlink_ext_ack *extack, + bool rtnl_is_held); int lwtunnel_valid_encap_type_attr(struct nlattr *attr, int len, - struct netlink_ext_ack *extack); + struct netlink_ext_ack *extack, + bool rtnl_is_held); int lwtunnel_build_state(struct net *net, u16 encap_type, struct nlattr *encap, unsigned int family, const void *cfg, @@ -201,13 +203,15 @@ static inline int lwtunnel_encap_del_ops(const struct lwtunnel_encap_ops *op, } static inline int lwtunnel_valid_encap_type(u16 encap_type, - struct netlink_ext_ack *extack) + struct netlink_ext_ack *extack, + bool rtnl_is_held) { NL_SET_ERR_MSG(extack, "CONFIG_LWTUNNEL is not enabled in this kernel"); return -EOPNOTSUPP; } static inline int lwtunnel_valid_encap_type_attr(struct nlattr *attr, int len, - struct netlink_ext_ack *extack) + struct netlink_ext_ack *extack, + bool rtnl_is_held) { /* return 0 since we are not walking attr looking for * RTA_ENCAP_TYPE attribute on nexthops. diff --git a/net/core/lwtunnel.c b/net/core/lwtunnel.c index 711cd3b4347a..6d3833269c2b 100644 --- a/net/core/lwtunnel.c +++ b/net/core/lwtunnel.c @@ -147,7 +147,8 @@ int lwtunnel_build_state(struct net *net, u16 encap_type, } EXPORT_SYMBOL_GPL(lwtunnel_build_state); -int lwtunnel_valid_encap_type(u16 encap_type, struct netlink_ext_ack *extack) +int lwtunnel_valid_encap_type(u16 encap_type, struct netlink_ext_ack *extack, + bool rtnl_is_held) { const struct lwtunnel_encap_ops *ops; int ret = -EINVAL; @@ -158,21 +159,19 @@ int lwtunnel_valid_encap_type(u16 encap_type, struct netlink_ext_ack *extack) return ret; } - rcu_read_lock(); - ops = rcu_dereference(lwtun_encaps[encap_type]); - rcu_read_unlock(); + ops = rcu_access_pointer(lwtun_encaps[encap_type]); #ifdef CONFIG_MODULES if (!ops) { const char *encap_type_str = lwtunnel_encap_str(encap_type); if (encap_type_str) { - __rtnl_unlock(); + if (rtnl_is_held) + __rtnl_unlock(); request_module("rtnl-lwt-%s", encap_type_str); - rtnl_lock(); + if (rtnl_is_held) + rtnl_lock(); - rcu_read_lock(); - ops = rcu_dereference(lwtun_encaps[encap_type]); - rcu_read_unlock(); + ops = rcu_access_pointer(lwtun_encaps[encap_type]); } } #endif @@ -185,7 +184,8 @@ int lwtunnel_valid_encap_type(u16 encap_type, struct netlink_ext_ack *extack) EXPORT_SYMBOL_GPL(lwtunnel_valid_encap_type); int lwtunnel_valid_encap_type_attr(struct nlattr *attr, int remaining, - struct netlink_ext_ack *extack) + struct netlink_ext_ack *extack, + bool rtnl_is_held) { struct rtnexthop *rtnh = (struct rtnexthop *)attr; struct nlattr *nla_entype; @@ -207,7 +207,8 @@ int lwtunnel_valid_encap_type_attr(struct nlattr *attr, int remaining, encap_type = nla_get_u16(nla_entype); if (lwtunnel_valid_encap_type(encap_type, - extack) != 0) + extack, + rtnl_is_held) != 0) return -EOPNOTSUPP; } } diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c index 6de77415b5b3..3f4e629998fa 100644 --- a/net/ipv4/fib_frontend.c +++ b/net/ipv4/fib_frontend.c @@ -807,7 +807,7 @@ static int rtm_to_fib_config(struct net *net, struct sk_buff *skb, case RTA_MULTIPATH: err = lwtunnel_valid_encap_type_attr(nla_data(attr), nla_len(attr), - extack); + extack, false); if (err < 0) goto errout; cfg->fc_mp = nla_data(attr); @@ -825,7 +825,7 @@ static int rtm_to_fib_config(struct net *net, struct sk_buff *skb, case RTA_ENCAP_TYPE: cfg->fc_encap_type = nla_get_u16(attr); err = lwtunnel_valid_encap_type(cfg->fc_encap_type, - extack); + extack, false); if (err < 0) goto errout; break; diff --git a/net/ipv4/nexthop.c b/net/ipv4/nexthop.c index 09a3d73b45ba..01df7dd795f0 100644 --- a/net/ipv4/nexthop.c +++ b/net/ipv4/nexthop.c @@ -3187,7 +3187,8 @@ static int rtm_to_nh_config(struct net *net, struct sk_buff *skb, } cfg->nh_encap_type = nla_get_u16(tb[NHA_ENCAP_TYPE]); - err = lwtunnel_valid_encap_type(cfg->nh_encap_type, extack); + err = lwtunnel_valid_encap_type(cfg->nh_encap_type, + extack, true); if (err < 0) goto out; diff --git a/net/ipv6/route.c b/net/ipv6/route.c index ef2d23a1e3d5..fb2e99a56529 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -5128,7 +5128,8 @@ static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh, cfg->fc_mp_len = nla_len(tb[RTA_MULTIPATH]); err = lwtunnel_valid_encap_type_attr(cfg->fc_mp, - cfg->fc_mp_len, extack); + cfg->fc_mp_len, + extack, true); if (err < 0) goto errout; } @@ -5147,7 +5148,8 @@ static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh, if (tb[RTA_ENCAP_TYPE]) { cfg->fc_encap_type = nla_get_u16(tb[RTA_ENCAP_TYPE]); - err = lwtunnel_valid_encap_type(cfg->fc_encap_type, extack); + err = lwtunnel_valid_encap_type(cfg->fc_encap_type, + extack, true); if (err < 0) goto errout; } -- cgit v1.2.3 From 67bab13307c83fb742c2556b06cdc39dbad27f07 Mon Sep 17 00:00:00 2001 From: Ge Yang Date: Wed, 19 Feb 2025 11:46:44 +0800 Subject: mm/hugetlb: wait for hugetlb folios to be freed Since the introduction of commit c77c0a8ac4c52 ("mm/hugetlb: defer freeing of huge pages if in non-task context"), which supports deferring the freeing of hugetlb pages, the allocation of contiguous memory through cma_alloc() may fail probabilistically. In the CMA allocation process, if it is found that the CMA area is occupied by in-use hugetlb folios, these in-use hugetlb folios need to be migrated to another location. When there are no available hugetlb folios in the free hugetlb pool during the migration of in-use hugetlb folios, new folios are allocated from the buddy system. A temporary state is set on the newly allocated folio. Upon completion of the hugetlb folio migration, the temporary state is transferred from the new folios to the old folios. Normally, when the old folios with the temporary state are freed, it is directly released back to the buddy system. However, due to the deferred freeing of hugetlb pages, the PageBuddy() check fails, ultimately leading to the failure of cma_alloc(). Here is a simplified call trace illustrating the process: cma_alloc() ->__alloc_contig_migrate_range() // Migrate in-use hugetlb folios ->unmap_and_move_huge_page() ->folio_putback_hugetlb() // Free old folios ->test_pages_isolated() ->__test_page_isolated_in_pageblock() ->PageBuddy(page) // Check if the page is in buddy To resolve this issue, we have implemented a function named wait_for_freed_hugetlb_folios(). This function ensures that the hugetlb folios are properly released back to the buddy system after their migration is completed. By invoking wait_for_freed_hugetlb_folios() before calling PageBuddy(), we ensure that PageBuddy() will succeed. Link: https://lkml.kernel.org/r/1739936804-18199-1-git-send-email-yangge1116@126.com Fixes: c77c0a8ac4c5 ("mm/hugetlb: defer freeing of huge pages if in non-task context") Signed-off-by: Ge Yang Reviewed-by: Muchun Song Acked-by: David Hildenbrand Cc: Baolin Wang Cc: Barry Song <21cnbao@gmail.com> Cc: Oscar Salvador Cc: Signed-off-by: Andrew Morton --- include/linux/hugetlb.h | 5 +++++ mm/hugetlb.c | 8 ++++++++ mm/page_isolation.c | 10 ++++++++++ 3 files changed, 23 insertions(+) (limited to 'include') diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index ec8c0ccc8f95..dbe76d4f1bfc 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -682,6 +682,7 @@ struct huge_bootmem_page { int isolate_or_dissolve_huge_page(struct page *page, struct list_head *list); int replace_free_hugepage_folios(unsigned long start_pfn, unsigned long end_pfn); +void wait_for_freed_hugetlb_folios(void); struct folio *alloc_hugetlb_folio(struct vm_area_struct *vma, unsigned long addr, bool cow_from_owner); struct folio *alloc_hugetlb_folio_nodemask(struct hstate *h, int preferred_nid, @@ -1066,6 +1067,10 @@ static inline int replace_free_hugepage_folios(unsigned long start_pfn, return 0; } +static inline void wait_for_freed_hugetlb_folios(void) +{ +} + static inline struct folio *alloc_hugetlb_folio(struct vm_area_struct *vma, unsigned long addr, bool cow_from_owner) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 163190e89ea1..811b29f77abf 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -2943,6 +2943,14 @@ int replace_free_hugepage_folios(unsigned long start_pfn, unsigned long end_pfn) return ret; } +void wait_for_freed_hugetlb_folios(void) +{ + if (llist_empty(&hpage_freelist)) + return; + + flush_work(&free_hpage_work); +} + typedef enum { /* * For either 0/1: we checked the per-vma resv map, and one resv diff --git a/mm/page_isolation.c b/mm/page_isolation.c index c608e9d72865..a051a29e95ad 100644 --- a/mm/page_isolation.c +++ b/mm/page_isolation.c @@ -607,6 +607,16 @@ int test_pages_isolated(unsigned long start_pfn, unsigned long end_pfn, struct zone *zone; int ret; + /* + * Due to the deferred freeing of hugetlb folios, the hugepage folios may + * not immediately release to the buddy system. This can cause PageBuddy() + * to fail in __test_page_isolated_in_pageblock(). To ensure that the + * hugetlb folios are properly released back to the buddy system, we + * invoke the wait_for_freed_hugetlb_folios() function to wait for the + * release to complete. + */ + wait_for_freed_hugetlb_folios(); + /* * Note: pageblock_nr_pages != MAX_PAGE_ORDER. Then, chunks of free * pages are not aligned to pageblock_nr_pages. -- cgit v1.2.3 From ce6d9c1c2b5cc785016faa11b48b6cd317eb367e Mon Sep 17 00:00:00 2001 From: Mike Snitzer Date: Mon, 24 Feb 2025 21:20:02 -0500 Subject: NFS: fix nfs_release_folio() to not deadlock via kcompactd writeback Add PF_KCOMPACTD flag and current_is_kcompactd() helper to check for it so nfs_release_folio() can skip calling nfs_wb_folio() from kcompactd. Otherwise NFS can deadlock waiting for kcompactd enduced writeback which recurses back to NFS (which triggers writeback to NFSD via NFS loopback mount on the same host, NFSD blocks waiting for XFS's call to __filemap_get_folio): 6070.550357] INFO: task kcompactd0:58 blocked for more than 4435 seconds. {--- [58] "kcompactd0" [<0>] folio_wait_bit+0xe8/0x200 [<0>] folio_wait_writeback+0x2b/0x80 [<0>] nfs_wb_folio+0x80/0x1b0 [nfs] [<0>] nfs_release_folio+0x68/0x130 [nfs] [<0>] split_huge_page_to_list_to_order+0x362/0x840 [<0>] migrate_pages_batch+0x43d/0xb90 [<0>] migrate_pages_sync+0x9a/0x240 [<0>] migrate_pages+0x93c/0x9f0 [<0>] compact_zone+0x8e2/0x1030 [<0>] compact_node+0xdb/0x120 [<0>] kcompactd+0x121/0x2e0 [<0>] kthread+0xcf/0x100 [<0>] ret_from_fork+0x31/0x40 [<0>] ret_from_fork_asm+0x1a/0x30 ---} [akpm@linux-foundation.org: fix build] Link: https://lkml.kernel.org/r/20250225022002.26141-1-snitzer@kernel.org Fixes: 96780ca55e3c ("NFS: fix up nfs_release_folio() to try to release the page") Signed-off-by: Mike Snitzer Cc: Anna Schumaker Cc: Trond Myklebust Cc: Signed-off-by: Andrew Morton --- fs/nfs/file.c | 3 ++- include/linux/compaction.h | 5 +++++ include/linux/sched.h | 2 +- mm/compaction.c | 3 +++ 4 files changed, 11 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 1bb646752e46..033feeab8c34 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -29,6 +29,7 @@ #include #include #include +#include #include #include @@ -457,7 +458,7 @@ static bool nfs_release_folio(struct folio *folio, gfp_t gfp) /* If the private flag is set, then the folio is not freeable */ if (folio_test_private(folio)) { if ((current_gfp_context(gfp) & GFP_KERNEL) != GFP_KERNEL || - current_is_kswapd()) + current_is_kswapd() || current_is_kcompactd()) return false; if (nfs_wb_folio(folio->mapping->host, folio) < 0) return false; diff --git a/include/linux/compaction.h b/include/linux/compaction.h index e94776496049..7bf0c521db63 100644 --- a/include/linux/compaction.h +++ b/include/linux/compaction.h @@ -80,6 +80,11 @@ static inline unsigned long compact_gap(unsigned int order) return 2UL << order; } +static inline int current_is_kcompactd(void) +{ + return current->flags & PF_KCOMPACTD; +} + #ifdef CONFIG_COMPACTION extern unsigned int extfrag_for_order(struct zone *zone, unsigned int order); diff --git a/include/linux/sched.h b/include/linux/sched.h index 9632e3318e0d..9c15365a30c0 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1701,7 +1701,7 @@ extern struct pid *cad_pid; #define PF_USED_MATH 0x00002000 /* If unset the fpu must be initialized before use */ #define PF_USER_WORKER 0x00004000 /* Kernel thread cloned from userspace thread */ #define PF_NOFREEZE 0x00008000 /* This thread should not be frozen */ -#define PF__HOLE__00010000 0x00010000 +#define PF_KCOMPACTD 0x00010000 /* I am kcompactd */ #define PF_KSWAPD 0x00020000 /* I am kswapd */ #define PF_MEMALLOC_NOFS 0x00040000 /* All allocations inherit GFP_NOFS. See memalloc_nfs_save() */ #define PF_MEMALLOC_NOIO 0x00080000 /* All allocations inherit GFP_NOIO. See memalloc_noio_save() */ diff --git a/mm/compaction.c b/mm/compaction.c index 12ed8425fa17..a3203d97123e 100644 --- a/mm/compaction.c +++ b/mm/compaction.c @@ -3181,6 +3181,7 @@ static int kcompactd(void *p) long default_timeout = msecs_to_jiffies(HPAGE_FRAG_CHECK_INTERVAL_MSEC); long timeout = default_timeout; + current->flags |= PF_KCOMPACTD; set_freezable(); pgdat->kcompactd_max_order = 0; @@ -3237,6 +3238,8 @@ static int kcompactd(void *p) pgdat->proactive_compact_trigger = false; } + current->flags &= ~PF_KCOMPACTD; + return 0; } -- cgit v1.2.3 From c29564d8b46f64f5e6e6f1c9c02f7761b7b90963 Mon Sep 17 00:00:00 2001 From: Su Hui Date: Fri, 21 Feb 2025 15:16:25 +0800 Subject: include/linux/log2.h: mark is_power_of_2() with __always_inline When building kernel with randconfig, there is an error: In function `kvm_is_cr4_bit_set',inlined from `kvm_update_cpuid_runtime' at arch/x86/kvm/cpuid.c:310:9: include/linux/compiler_types.h:542:38: error: call to `__compiletime_assert_380' declared with attribute error: BUILD_BUG_ON failed: !is_power_of_2(cr4_bit). '!is_power_of_2(X86_CR4_OSXSAVE)' is False, but gcc treats is_power_of_2() as non-inline function and a compilation error happens. Fix this by marking is_power_of_2() with __always_inline. Link: https://lkml.kernel.org/r/20250221071624.1356899-1-suhui@nfschina.com Signed-off-by: Su Hui Cc: Binbin Wu Cc: Paolo Bonzini Cc: Sean Christopherson Signed-off-by: Andrew Morton --- include/linux/log2.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/log2.h b/include/linux/log2.h index 9f30d087a128..1366cb688a6d 100644 --- a/include/linux/log2.h +++ b/include/linux/log2.h @@ -41,7 +41,7 @@ int __ilog2_u64(u64 n) * *not* considered a power of two. * Return: true if @n is a power of 2, otherwise false. */ -static inline __attribute__((const)) +static __always_inline __attribute__((const)) bool is_power_of_2(unsigned long n) { return (n != 0 && ((n & (n - 1)) == 0)); -- cgit v1.2.3 From fb8286562ecfb585e26b033c5e32e6fb85efb0b3 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Thu, 6 Mar 2025 04:05:26 +0100 Subject: netfilter: nf_tables: make destruction work queue pernet The call to flush_work before tearing down a table from the netlink notifier was supposed to make sure that all earlier updates (e.g. rule add) that might reference that table have been processed. Unfortunately, flush_work() waits for the last queued instance. This could be an instance that is different from the one that we must wait for. This is because transactions are protected with a pernet mutex, but the work item is global, so holding the transaction mutex doesn't prevent another netns from queueing more work. Make the work item pernet so that flush_work() will wait for all transactions queued from this netns. A welcome side effect is that we no longer need to wait for transaction objects from foreign netns. The gc work queue is still global. This seems to be ok because nft_set structures are reference counted and each container structure owns a reference on the net namespace. The destroy_list is still protected by a global spinlock rather than pernet one but the hold time is very short anyway. v2: call cancel_work_sync before reaping the remaining tables (Pablo). Fixes: 9f6958ba2e90 ("netfilter: nf_tables: unconditionally flush pending work before notifier") Reported-by: syzbot+5d8c5789c8cb076b2c25@syzkaller.appspotmail.com Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- include/net/netfilter/nf_tables.h | 4 +++- net/netfilter/nf_tables_api.c | 24 ++++++++++++++---------- net/netfilter/nft_compat.c | 8 ++++---- 3 files changed, 21 insertions(+), 15 deletions(-) (limited to 'include') diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index 60d5dcdb289c..803d5f1601f9 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -1891,7 +1891,7 @@ void nft_chain_filter_fini(void); void __init nft_chain_route_init(void); void nft_chain_route_fini(void); -void nf_tables_trans_destroy_flush_work(void); +void nf_tables_trans_destroy_flush_work(struct net *net); int nf_msecs_to_jiffies64(const struct nlattr *nla, u64 *result); __be64 nf_jiffies64_to_msecs(u64 input); @@ -1905,6 +1905,7 @@ static inline int nft_request_module(struct net *net, const char *fmt, ...) { re struct nftables_pernet { struct list_head tables; struct list_head commit_list; + struct list_head destroy_list; struct list_head commit_set_list; struct list_head binding_list; struct list_head module_list; @@ -1915,6 +1916,7 @@ struct nftables_pernet { unsigned int base_seq; unsigned int gc_seq; u8 validate_state; + struct work_struct destroy_work; }; extern unsigned int nf_tables_net_id; diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index a34de9c17cf1..c2df81b7e950 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -34,7 +34,6 @@ unsigned int nf_tables_net_id __read_mostly; static LIST_HEAD(nf_tables_expressions); static LIST_HEAD(nf_tables_objects); static LIST_HEAD(nf_tables_flowtables); -static LIST_HEAD(nf_tables_destroy_list); static LIST_HEAD(nf_tables_gc_list); static DEFINE_SPINLOCK(nf_tables_destroy_list_lock); static DEFINE_SPINLOCK(nf_tables_gc_list_lock); @@ -125,7 +124,6 @@ static void nft_validate_state_update(struct nft_table *table, u8 new_validate_s table->validate_state = new_validate_state; } static void nf_tables_trans_destroy_work(struct work_struct *w); -static DECLARE_WORK(trans_destroy_work, nf_tables_trans_destroy_work); static void nft_trans_gc_work(struct work_struct *work); static DECLARE_WORK(trans_gc_work, nft_trans_gc_work); @@ -10006,11 +10004,12 @@ static void nft_commit_release(struct nft_trans *trans) static void nf_tables_trans_destroy_work(struct work_struct *w) { + struct nftables_pernet *nft_net = container_of(w, struct nftables_pernet, destroy_work); struct nft_trans *trans, *next; LIST_HEAD(head); spin_lock(&nf_tables_destroy_list_lock); - list_splice_init(&nf_tables_destroy_list, &head); + list_splice_init(&nft_net->destroy_list, &head); spin_unlock(&nf_tables_destroy_list_lock); if (list_empty(&head)) @@ -10024,9 +10023,11 @@ static void nf_tables_trans_destroy_work(struct work_struct *w) } } -void nf_tables_trans_destroy_flush_work(void) +void nf_tables_trans_destroy_flush_work(struct net *net) { - flush_work(&trans_destroy_work); + struct nftables_pernet *nft_net = nft_pernet(net); + + flush_work(&nft_net->destroy_work); } EXPORT_SYMBOL_GPL(nf_tables_trans_destroy_flush_work); @@ -10484,11 +10485,11 @@ static void nf_tables_commit_release(struct net *net) trans->put_net = true; spin_lock(&nf_tables_destroy_list_lock); - list_splice_tail_init(&nft_net->commit_list, &nf_tables_destroy_list); + list_splice_tail_init(&nft_net->commit_list, &nft_net->destroy_list); spin_unlock(&nf_tables_destroy_list_lock); nf_tables_module_autoload_cleanup(net); - schedule_work(&trans_destroy_work); + schedule_work(&nft_net->destroy_work); mutex_unlock(&nft_net->commit_mutex); } @@ -11853,7 +11854,7 @@ static int nft_rcv_nl_event(struct notifier_block *this, unsigned long event, gc_seq = nft_gc_seq_begin(nft_net); - nf_tables_trans_destroy_flush_work(); + nf_tables_trans_destroy_flush_work(net); again: list_for_each_entry(table, &nft_net->tables, list) { if (nft_table_has_owner(table) && @@ -11895,6 +11896,7 @@ static int __net_init nf_tables_init_net(struct net *net) INIT_LIST_HEAD(&nft_net->tables); INIT_LIST_HEAD(&nft_net->commit_list); + INIT_LIST_HEAD(&nft_net->destroy_list); INIT_LIST_HEAD(&nft_net->commit_set_list); INIT_LIST_HEAD(&nft_net->binding_list); INIT_LIST_HEAD(&nft_net->module_list); @@ -11903,6 +11905,7 @@ static int __net_init nf_tables_init_net(struct net *net) nft_net->base_seq = 1; nft_net->gc_seq = 0; nft_net->validate_state = NFT_VALIDATE_SKIP; + INIT_WORK(&nft_net->destroy_work, nf_tables_trans_destroy_work); return 0; } @@ -11931,14 +11934,17 @@ static void __net_exit nf_tables_exit_net(struct net *net) if (!list_empty(&nft_net->module_list)) nf_tables_module_autoload_cleanup(net); + cancel_work_sync(&nft_net->destroy_work); __nft_release_tables(net); nft_gc_seq_end(nft_net, gc_seq); mutex_unlock(&nft_net->commit_mutex); + WARN_ON_ONCE(!list_empty(&nft_net->tables)); WARN_ON_ONCE(!list_empty(&nft_net->module_list)); WARN_ON_ONCE(!list_empty(&nft_net->notify_list)); + WARN_ON_ONCE(!list_empty(&nft_net->destroy_list)); } static void nf_tables_exit_batch(struct list_head *net_exit_list) @@ -12029,10 +12035,8 @@ static void __exit nf_tables_module_exit(void) unregister_netdevice_notifier(&nf_tables_flowtable_notifier); nft_chain_filter_fini(); nft_chain_route_fini(); - nf_tables_trans_destroy_flush_work(); unregister_pernet_subsys(&nf_tables_net_ops); cancel_work_sync(&trans_gc_work); - cancel_work_sync(&trans_destroy_work); rcu_barrier(); rhltable_destroy(&nft_objname_ht); nf_tables_core_module_exit(); diff --git a/net/netfilter/nft_compat.c b/net/netfilter/nft_compat.c index 7ca4f0d21fe2..72711d62fddf 100644 --- a/net/netfilter/nft_compat.c +++ b/net/netfilter/nft_compat.c @@ -228,7 +228,7 @@ static int nft_parse_compat(const struct nlattr *attr, u16 *proto, bool *inv) return 0; } -static void nft_compat_wait_for_destructors(void) +static void nft_compat_wait_for_destructors(struct net *net) { /* xtables matches or targets can have side effects, e.g. * creation/destruction of /proc files. @@ -236,7 +236,7 @@ static void nft_compat_wait_for_destructors(void) * work queue. If we have pending invocations we thus * need to wait for those to finish. */ - nf_tables_trans_destroy_flush_work(); + nf_tables_trans_destroy_flush_work(net); } static int @@ -262,7 +262,7 @@ nft_target_init(const struct nft_ctx *ctx, const struct nft_expr *expr, nft_target_set_tgchk_param(&par, ctx, target, info, &e, proto, inv); - nft_compat_wait_for_destructors(); + nft_compat_wait_for_destructors(ctx->net); ret = xt_check_target(&par, size, proto, inv); if (ret < 0) { @@ -515,7 +515,7 @@ __nft_match_init(const struct nft_ctx *ctx, const struct nft_expr *expr, nft_match_set_mtchk_param(&par, ctx, match, info, &e, proto, inv); - nft_compat_wait_for_destructors(); + nft_compat_wait_for_destructors(ctx->net); return xt_check_match(&par, size, proto, inv); } -- cgit v1.2.3 From 74d42bdb3a4673b1c10d1f457184e4d3c9cb0196 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Thu, 6 Mar 2025 07:30:42 -1000 Subject: fs/pipe: express 'pipe_empty()' in terms of 'pipe_occupancy()' That's what 'pipe_full()' does, so it's more consistent. But more importantly it gets the type limits right when the pipe head and tail are no longer necessarily 'unsigned int'. Signed-off-by: Linus Torvalds --- include/linux/pipe_fs_i.h | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h index e572e6fc4f81..4d0a2267e6ef 100644 --- a/include/linux/pipe_fs_i.h +++ b/include/linux/pipe_fs_i.h @@ -177,23 +177,23 @@ static inline bool pipe_has_watch_queue(const struct pipe_inode_info *pipe) } /** - * pipe_empty - Return true if the pipe is empty + * pipe_occupancy - Return number of slots used in the pipe * @head: The pipe ring head pointer * @tail: The pipe ring tail pointer */ -static inline bool pipe_empty(unsigned int head, unsigned int tail) +static inline unsigned int pipe_occupancy(unsigned int head, unsigned int tail) { - return head == tail; + return (pipe_index_t)(head - tail); } /** - * pipe_occupancy - Return number of slots used in the pipe + * pipe_empty - Return true if the pipe is empty * @head: The pipe ring head pointer * @tail: The pipe ring tail pointer */ -static inline unsigned int pipe_occupancy(unsigned int head, unsigned int tail) +static inline bool pipe_empty(unsigned int head, unsigned int tail) { - return (pipe_index_t)(head - tail); + return !pipe_occupancy(head, tail); } /** -- cgit v1.2.3 From d4c22ec680c8db832ffc0b964c6008e65436cba8 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:19 -0800 Subject: net: hold netdev instance lock during ndo_open/ndo_stop For the drivers that use shaper API, switch to the mode where core stack holds the netdev lock. This affects two drivers: * iavf - already grabs netdev lock in ndo_open/ndo_stop, so mostly remove these * netdevsim - switch to _locked APIs to avoid deadlock iavf_close diff is a bit confusing, the existing call looks like this: iavf_close() { netdev_lock() .. netdev_unlock() wait_event_timeout(down_waitqueue) } I change it to the following: netdev_lock() iavf_close() { .. netdev_unlock() wait_event_timeout(down_waitqueue) netdev_lock() // reusing this lock call } netdev_unlock() Since I'm reusing existing netdev_lock call, so it looks like I only add netdev_unlock. Cc: Saeed Mahameed Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-2-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/iavf/iavf_main.c | 14 ++++++-------- drivers/net/netdevsim/netdev.c | 14 +++++++++----- include/linux/netdevice.h | 23 +++++++++++++++++++++++ net/core/dev.c | 12 ++++++++++++ net/core/dev.h | 6 ++++-- 5 files changed, 54 insertions(+), 15 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index 71f11f64b13d..9f4d223dffcf 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -4562,22 +4562,21 @@ static int iavf_open(struct net_device *netdev) struct iavf_adapter *adapter = netdev_priv(netdev); int err; + netdev_assert_locked(netdev); + if (adapter->flags & IAVF_FLAG_PF_COMMS_FAILED) { dev_err(&adapter->pdev->dev, "Unable to open device due to PF driver failure.\n"); return -EIO; } - netdev_lock(netdev); while (!mutex_trylock(&adapter->crit_lock)) { /* If we are in __IAVF_INIT_CONFIG_ADAPTER state the crit_lock * is already taken and iavf_open is called from an upper * device's notifier reacting on NETDEV_REGISTER event. * We have to leave here to avoid dead lock. */ - if (adapter->state == __IAVF_INIT_CONFIG_ADAPTER) { - netdev_unlock(netdev); + if (adapter->state == __IAVF_INIT_CONFIG_ADAPTER) return -EBUSY; - } usleep_range(500, 1000); } @@ -4626,7 +4625,6 @@ static int iavf_open(struct net_device *netdev) iavf_irq_enable(adapter, true); mutex_unlock(&adapter->crit_lock); - netdev_unlock(netdev); return 0; @@ -4639,7 +4637,6 @@ err_setup_tx: iavf_free_all_tx_resources(adapter); err_unlock: mutex_unlock(&adapter->crit_lock); - netdev_unlock(netdev); return err; } @@ -4661,12 +4658,12 @@ static int iavf_close(struct net_device *netdev) u64 aq_to_restore; int status; - netdev_lock(netdev); + netdev_assert_locked(netdev); + mutex_lock(&adapter->crit_lock); if (adapter->state <= __IAVF_DOWN_PENDING) { mutex_unlock(&adapter->crit_lock); - netdev_unlock(netdev); return 0; } @@ -4719,6 +4716,7 @@ static int iavf_close(struct net_device *netdev) if (!status) netdev_warn(netdev, "Device resources not yet released\n"); + netdev_lock(netdev); mutex_lock(&adapter->crit_lock); adapter->aq_required |= aq_to_restore; mutex_unlock(&adapter->crit_lock); diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c index a41dc79e9c2e..aaa3b58e2e3e 100644 --- a/drivers/net/netdevsim/netdev.c +++ b/drivers/net/netdevsim/netdev.c @@ -402,7 +402,7 @@ static int nsim_init_napi(struct netdevsim *ns) for (i = 0; i < dev->num_rx_queues; i++) { rq = ns->rq[i]; - netif_napi_add_config(dev, &rq->napi, nsim_poll, i); + netif_napi_add_config_locked(dev, &rq->napi, nsim_poll, i); } for (i = 0; i < dev->num_rx_queues; i++) { @@ -422,7 +422,7 @@ err_pp_destroy: } for (i = 0; i < dev->num_rx_queues; i++) - __netif_napi_del(&ns->rq[i]->napi); + __netif_napi_del_locked(&ns->rq[i]->napi); return err; } @@ -452,7 +452,7 @@ static void nsim_enable_napi(struct netdevsim *ns) struct nsim_rq *rq = ns->rq[i]; netif_queue_set_napi(dev, i, NETDEV_QUEUE_TYPE_RX, &rq->napi); - napi_enable(&rq->napi); + napi_enable_locked(&rq->napi); } } @@ -461,6 +461,8 @@ static int nsim_open(struct net_device *dev) struct netdevsim *ns = netdev_priv(dev); int err; + netdev_assert_locked(dev); + err = nsim_init_napi(ns); if (err) return err; @@ -478,8 +480,8 @@ static void nsim_del_napi(struct netdevsim *ns) for (i = 0; i < dev->num_rx_queues; i++) { struct nsim_rq *rq = ns->rq[i]; - napi_disable(&rq->napi); - __netif_napi_del(&rq->napi); + napi_disable_locked(&rq->napi); + __netif_napi_del_locked(&rq->napi); } synchronize_net(); @@ -494,6 +496,8 @@ static int nsim_stop(struct net_device *dev) struct netdevsim *ns = netdev_priv(dev); struct netdevsim *peer; + netdev_assert_locked(dev); + netif_carrier_off(dev); peer = rtnl_dereference(ns->peer); if (peer) diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 7ab86ec228b7..33066b155c84 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2753,6 +2753,29 @@ static inline void netdev_assert_locked_or_invisible(struct net_device *dev) netdev_assert_locked(dev); } +static inline bool netdev_need_ops_lock(struct net_device *dev) +{ + bool ret = false; + +#if IS_ENABLED(CONFIG_NET_SHAPER) + ret |= !!dev->netdev_ops->net_shaper_ops; +#endif + + return ret; +} + +static inline void netdev_lock_ops(struct net_device *dev) +{ + if (netdev_need_ops_lock(dev)) + netdev_lock(dev); +} + +static inline void netdev_unlock_ops(struct net_device *dev) +{ + if (netdev_need_ops_lock(dev)) + netdev_unlock(dev); +} + void netif_napi_set_irq_locked(struct napi_struct *napi, int irq); static inline void netif_napi_set_irq(struct napi_struct *napi, int irq) diff --git a/net/core/dev.c b/net/core/dev.c index 2dc705604509..7a327c782ea4 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1627,6 +1627,8 @@ static int __dev_open(struct net_device *dev, struct netlink_ext_ack *extack) if (ret) return ret; + netdev_lock_ops(dev); + set_bit(__LINK_STATE_START, &dev->state); if (ops->ndo_validate_addr) @@ -1646,6 +1648,8 @@ static int __dev_open(struct net_device *dev, struct netlink_ext_ack *extack) add_device_randomness(dev->dev_addr, dev->addr_len); } + netdev_unlock_ops(dev); + return ret; } @@ -1716,11 +1720,19 @@ static void __dev_close_many(struct list_head *head) * We allow it to be called even after a DETACH hot-plug * event. */ + + /* TODO: move the lock up before clearing __LINK_STATE_START. + * Generates spurious lockdep warning. + */ + netdev_lock_ops(dev); + if (ops->ndo_stop) ops->ndo_stop(dev); netif_set_up(dev, false); netpoll_poll_enable(dev); + + netdev_unlock_ops(dev); } } diff --git a/net/core/dev.h b/net/core/dev.h index caa13e431a6b..25bb9d6afbce 100644 --- a/net/core/dev.h +++ b/net/core/dev.h @@ -134,9 +134,11 @@ static inline void netif_set_up(struct net_device *dev, bool value) else dev->flags &= ~IFF_UP; - netdev_lock(dev); + if (!netdev_need_ops_lock(dev)) + netdev_lock(dev); dev->up = value; - netdev_unlock(dev); + if (!netdev_need_ops_lock(dev)) + netdev_unlock(dev); } static inline void netif_set_gso_max_size(struct net_device *dev, -- cgit v1.2.3 From c4f0f30b424e7258a777bcbcbf9006207da4854c Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:20 -0800 Subject: net: hold netdev instance lock during nft ndo_setup_tc Introduce new dev_setup_tc for nft ndo_setup_tc paths. Reviewed-by: Eric Dumazet Cc: Saeed Mahameed Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-3-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/iavf/iavf_main.c | 2 -- include/linux/netdevice.h | 2 ++ net/core/dev.c | 18 ++++++++++++++++++ net/netfilter/nf_flow_table_offload.c | 2 +- net/netfilter/nf_tables_offload.c | 2 +- 5 files changed, 22 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index 9f4d223dffcf..032e1a58af6f 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -3894,10 +3894,8 @@ exit: if (test_bit(__IAVF_IN_REMOVE_TASK, &adapter->crit_section)) return 0; - netdev_lock(netdev); netif_set_real_num_rx_queues(netdev, total_qps); netif_set_real_num_tx_queues(netdev, total_qps); - netdev_unlock(netdev); return ret; } diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 33066b155c84..69951eeb96d2 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -3353,6 +3353,8 @@ int dev_alloc_name(struct net_device *dev, const char *name); int dev_open(struct net_device *dev, struct netlink_ext_ack *extack); void dev_close(struct net_device *dev); void dev_close_many(struct list_head *head, bool unlink); +int dev_setup_tc(struct net_device *dev, enum tc_setup_type type, + void *type_data); void dev_disable_lro(struct net_device *dev); int dev_loopback_xmit(struct net *net, struct sock *sk, struct sk_buff *newskb); u16 dev_pick_tx_zero(struct net_device *dev, struct sk_buff *skb, diff --git a/net/core/dev.c b/net/core/dev.c index 7a327c782ea4..57af25683ea1 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1786,6 +1786,24 @@ void dev_close(struct net_device *dev) } EXPORT_SYMBOL(dev_close); +int dev_setup_tc(struct net_device *dev, enum tc_setup_type type, + void *type_data) +{ + const struct net_device_ops *ops = dev->netdev_ops; + int ret; + + ASSERT_RTNL(); + + if (!ops->ndo_setup_tc) + return -EOPNOTSUPP; + + netdev_lock_ops(dev); + ret = ops->ndo_setup_tc(dev, type, type_data); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL(dev_setup_tc); /** * dev_disable_lro - disable Large Receive Offload on a device diff --git a/net/netfilter/nf_flow_table_offload.c b/net/netfilter/nf_flow_table_offload.c index e06bc36f49fe..0ec4abded10d 100644 --- a/net/netfilter/nf_flow_table_offload.c +++ b/net/netfilter/nf_flow_table_offload.c @@ -1175,7 +1175,7 @@ static int nf_flow_table_offload_cmd(struct flow_block_offload *bo, nf_flow_table_block_offload_init(bo, dev_net(dev), cmd, flowtable, extack); down_write(&flowtable->flow_block_lock); - err = dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_FT, bo); + err = dev_setup_tc(dev, TC_SETUP_FT, bo); up_write(&flowtable->flow_block_lock); if (err < 0) return err; diff --git a/net/netfilter/nf_tables_offload.c b/net/netfilter/nf_tables_offload.c index 64675f1c7f29..b761899c143c 100644 --- a/net/netfilter/nf_tables_offload.c +++ b/net/netfilter/nf_tables_offload.c @@ -390,7 +390,7 @@ static int nft_block_offload_cmd(struct nft_base_chain *chain, nft_flow_block_offload_init(&bo, dev_net(dev), cmd, chain, &extack); - err = dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_BLOCK, &bo); + err = dev_setup_tc(dev, TC_SETUP_BLOCK, &bo); if (err < 0) return err; -- cgit v1.2.3 From cae03e5bdd9e0c8570506c50f1f234da40201732 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:23 -0800 Subject: net: hold netdev instance lock during queue operations For the drivers that use queue management API, switch to the mode where core stack holds the netdev instance lock. This affects the following drivers: - bnxt - gve - netdevsim Originally I locked only start/stop, but switched to holding the lock over all iterations to make them look atomic to the device (feels like it should be easier to reason about). Reviewed-by: Eric Dumazet Cc: Saeed Mahameed Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-6-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 16 ++++++++-------- drivers/net/ethernet/google/gve/gve_main.c | 12 ++++++++---- drivers/net/ethernet/google/gve/gve_utils.c | 6 +++--- drivers/net/netdevsim/netdev.c | 25 ++++++++++++++++--------- include/linux/netdevice.h | 2 +- net/core/netdev_rx_queue.c | 5 +++++ 6 files changed, 41 insertions(+), 25 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index 15c57a06ecaf..ead9fcaad6bd 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -11508,7 +11508,7 @@ static int bnxt_request_irq(struct bnxt *bp) if (rc) break; - netif_napi_set_irq(&bp->bnapi[i]->napi, irq->vector); + netif_napi_set_irq_locked(&bp->bnapi[i]->napi, irq->vector); irq->requested = 1; if (zalloc_cpumask_var(&irq->cpu_mask, GFP_KERNEL)) { @@ -11557,9 +11557,9 @@ static void bnxt_del_napi(struct bnxt *bp) for (i = 0; i < bp->cp_nr_rings; i++) { struct bnxt_napi *bnapi = bp->bnapi[i]; - __netif_napi_del(&bnapi->napi); + __netif_napi_del_locked(&bnapi->napi); } - /* We called __netif_napi_del(), we need + /* We called __netif_napi_del_locked(), we need * to respect an RCU grace period before freeing napi structures. */ synchronize_net(); @@ -11578,12 +11578,12 @@ static void bnxt_init_napi(struct bnxt *bp) cp_nr_rings--; for (i = 0; i < cp_nr_rings; i++) { bnapi = bp->bnapi[i]; - netif_napi_add_config(bp->dev, &bnapi->napi, poll_fn, - bnapi->index); + netif_napi_add_config_locked(bp->dev, &bnapi->napi, poll_fn, + bnapi->index); } if (BNXT_CHIP_TYPE_NITRO_A0(bp)) { bnapi = bp->bnapi[cp_nr_rings]; - netif_napi_add(bp->dev, &bnapi->napi, bnxt_poll_nitroa0); + netif_napi_add_locked(bp->dev, &bnapi->napi, bnxt_poll_nitroa0); } } @@ -11604,7 +11604,7 @@ static void bnxt_disable_napi(struct bnxt *bp) cpr->sw_stats->tx.tx_resets++; if (bnapi->in_reset) cpr->sw_stats->rx.rx_resets++; - napi_disable(&bnapi->napi); + napi_disable_locked(&bnapi->napi); } } @@ -11626,7 +11626,7 @@ static void bnxt_enable_napi(struct bnxt *bp) INIT_WORK(&cpr->dim.work, bnxt_dim_work); cpr->dim.mode = DIM_CQ_PERIOD_MODE_START_FROM_EQE; } - napi_enable(&bnapi->napi); + napi_enable_locked(&bnapi->napi); } } diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index 029be8342b7b..6dcdcaf518f4 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -1968,7 +1968,7 @@ static void gve_turndown(struct gve_priv *priv) netif_queue_set_napi(priv->dev, idx, NETDEV_QUEUE_TYPE_TX, NULL); - napi_disable(&block->napi); + napi_disable_locked(&block->napi); } for (idx = 0; idx < priv->rx_cfg.num_queues; idx++) { int ntfy_idx = gve_rx_idx_to_ntfy(priv, idx); @@ -1979,7 +1979,7 @@ static void gve_turndown(struct gve_priv *priv) netif_queue_set_napi(priv->dev, idx, NETDEV_QUEUE_TYPE_RX, NULL); - napi_disable(&block->napi); + napi_disable_locked(&block->napi); } /* Stop tx queues */ @@ -2009,7 +2009,7 @@ static void gve_turnup(struct gve_priv *priv) if (!gve_tx_was_added_to_block(priv, idx)) continue; - napi_enable(&block->napi); + napi_enable_locked(&block->napi); if (idx < priv->tx_cfg.num_queues) netif_queue_set_napi(priv->dev, idx, @@ -2037,7 +2037,7 @@ static void gve_turnup(struct gve_priv *priv) if (!gve_rx_was_added_to_block(priv, idx)) continue; - napi_enable(&block->napi); + napi_enable_locked(&block->napi); netif_queue_set_napi(priv->dev, idx, NETDEV_QUEUE_TYPE_RX, &block->napi); @@ -2887,6 +2887,7 @@ static int gve_suspend(struct pci_dev *pdev, pm_message_t state) priv->suspend_cnt++; rtnl_lock(); + netdev_lock(netdev); if (was_up && gve_close(priv->dev)) { /* If the dev was up, attempt to close, if close fails, reset */ gve_reset_and_teardown(priv, was_up); @@ -2895,6 +2896,7 @@ static int gve_suspend(struct pci_dev *pdev, pm_message_t state) gve_teardown_priv_resources(priv); } priv->up_before_suspend = was_up; + netdev_unlock(netdev); rtnl_unlock(); return 0; } @@ -2907,7 +2909,9 @@ static int gve_resume(struct pci_dev *pdev) priv->resume_cnt++; rtnl_lock(); + netdev_lock(netdev); err = gve_reset_recovery(priv, priv->up_before_suspend); + netdev_unlock(netdev); rtnl_unlock(); return err; } diff --git a/drivers/net/ethernet/google/gve/gve_utils.c b/drivers/net/ethernet/google/gve/gve_utils.c index 30fef100257e..ace9b8698021 100644 --- a/drivers/net/ethernet/google/gve/gve_utils.c +++ b/drivers/net/ethernet/google/gve/gve_utils.c @@ -110,13 +110,13 @@ void gve_add_napi(struct gve_priv *priv, int ntfy_idx, { struct gve_notify_block *block = &priv->ntfy_blocks[ntfy_idx]; - netif_napi_add(priv->dev, &block->napi, gve_poll); - netif_napi_set_irq(&block->napi, block->irq); + netif_napi_add_locked(priv->dev, &block->napi, gve_poll); + netif_napi_set_irq_locked(&block->napi, block->irq); } void gve_remove_napi(struct gve_priv *priv, int ntfy_idx) { struct gve_notify_block *block = &priv->ntfy_blocks[ntfy_idx]; - netif_napi_del(&block->napi); + netif_napi_del_locked(&block->napi); } diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c index aaa3b58e2e3e..54d03b0628d2 100644 --- a/drivers/net/netdevsim/netdev.c +++ b/drivers/net/netdevsim/netdev.c @@ -683,7 +683,8 @@ nsim_queue_mem_alloc(struct net_device *dev, void *per_queue_mem, int idx) goto err_free; if (!ns->rq_reset_mode) - netif_napi_add_config(dev, &qmem->rq->napi, nsim_poll, idx); + netif_napi_add_config_locked(dev, &qmem->rq->napi, nsim_poll, + idx); return 0; @@ -700,7 +701,7 @@ static void nsim_queue_mem_free(struct net_device *dev, void *per_queue_mem) page_pool_destroy(qmem->pp); if (qmem->rq) { if (!ns->rq_reset_mode) - netif_napi_del(&qmem->rq->napi); + netif_napi_del_locked(&qmem->rq->napi); page_pool_destroy(qmem->rq->page_pool); nsim_queue_free(qmem->rq); } @@ -712,9 +713,11 @@ nsim_queue_start(struct net_device *dev, void *per_queue_mem, int idx) struct nsim_queue_mem *qmem = per_queue_mem; struct netdevsim *ns = netdev_priv(dev); + netdev_assert_locked(dev); + if (ns->rq_reset_mode == 1) { ns->rq[idx]->page_pool = qmem->pp; - napi_enable(&ns->rq[idx]->napi); + napi_enable_locked(&ns->rq[idx]->napi); return 0; } @@ -722,15 +725,17 @@ nsim_queue_start(struct net_device *dev, void *per_queue_mem, int idx) * here we want to test various call orders. */ if (ns->rq_reset_mode == 2) { - netif_napi_del(&ns->rq[idx]->napi); - netif_napi_add_config(dev, &qmem->rq->napi, nsim_poll, idx); + netif_napi_del_locked(&ns->rq[idx]->napi); + netif_napi_add_config_locked(dev, &qmem->rq->napi, nsim_poll, + idx); } else if (ns->rq_reset_mode == 3) { - netif_napi_add_config(dev, &qmem->rq->napi, nsim_poll, idx); - netif_napi_del(&ns->rq[idx]->napi); + netif_napi_add_config_locked(dev, &qmem->rq->napi, nsim_poll, + idx); + netif_napi_del_locked(&ns->rq[idx]->napi); } ns->rq[idx] = qmem->rq; - napi_enable(&ns->rq[idx]->napi); + napi_enable_locked(&ns->rq[idx]->napi); return 0; } @@ -740,7 +745,9 @@ static int nsim_queue_stop(struct net_device *dev, void *per_queue_mem, int idx) struct nsim_queue_mem *qmem = per_queue_mem; struct netdevsim *ns = netdev_priv(dev); - napi_disable(&ns->rq[idx]->napi); + netdev_assert_locked(dev); + + napi_disable_locked(&ns->rq[idx]->napi); if (ns->rq_reset_mode == 1) { qmem->pp = ns->rq[idx]->page_pool; diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 69951eeb96d2..abda17b15950 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2755,7 +2755,7 @@ static inline void netdev_assert_locked_or_invisible(struct net_device *dev) static inline bool netdev_need_ops_lock(struct net_device *dev) { - bool ret = false; + bool ret = !!dev->queue_mgmt_ops; #if IS_ENABLED(CONFIG_NET_SHAPER) ret |= !!dev->netdev_ops->net_shaper_ops; diff --git a/net/core/netdev_rx_queue.c b/net/core/netdev_rx_queue.c index ddd54e1e7289..7419c41fd3cb 100644 --- a/net/core/netdev_rx_queue.c +++ b/net/core/netdev_rx_queue.c @@ -30,6 +30,8 @@ int netdev_rx_queue_restart(struct net_device *dev, unsigned int rxq_idx) goto err_free_new_mem; } + netdev_lock(dev); + err = qops->ndo_queue_mem_alloc(dev, new_mem, rxq_idx); if (err) goto err_free_old_mem; @@ -52,6 +54,8 @@ int netdev_rx_queue_restart(struct net_device *dev, unsigned int rxq_idx) qops->ndo_queue_mem_free(dev, old_mem); + netdev_unlock(dev); + kvfree(old_mem); kvfree(new_mem); @@ -76,6 +80,7 @@ err_free_new_queue_mem: qops->ndo_queue_mem_free(dev, new_mem); err_free_old_mem: + netdev_unlock(dev); kvfree(old_mem); err_free_new_mem: -- cgit v1.2.3 From 7e4d784f5810bba76c4593791028e13cce4af547 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:24 -0800 Subject: net: hold netdev instance lock during rtnetlink operations To preserve the atomicity, hold the lock while applying multiple attributes. The major issue with a full conversion to the instance lock are software nesting devices (bonding/team/vrf/etc). Those devices call into the core stack for their lower (potentially real hw) devices. To avoid explicitly wrapping all those places into instance lock/unlock, introduce new API boundaries: - (some) existing dev_xxx calls are now considered "external" (to drivers) APIs and they transparently grab the instance lock if needed (dev_api.c) - new netif_xxx calls are internal core stack API (naming is sketchy, I've tried netdev_xxx_locked per Jakub's suggestion, but it feels a bit verbose; but happy to get back to this naming scheme if this is the preference) This avoids touching most of the existing ioctl/sysfs/drivers paths. Note the special handling of ndo_xxx_slave operations: I exploit the fact that none of the drivers that call these functions need/use instance lock. At the same time, they use dev_xxx APIs, so the lower device has to be unlocked. Changes in unregister_netdevice_many_notify (to protect dev->state with instance lock) trigger lockdep - the loop over close_list (mostly from cleanup_net) introduces spurious ordering issues. netdev_lock_cmp_fn has a justification on why it's ok to suppress for now. Cc: Saeed Mahameed Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-7-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- include/linux/netdevice.h | 40 +++++++-- net/core/Makefile | 2 +- net/core/dev.c | 154 ++++++------------------------- net/core/dev.h | 13 ++- net/core/dev_api.c | 224 ++++++++++++++++++++++++++++++++++++++++++++++ net/core/rtnetlink.c | 46 +++++++--- 6 files changed, 329 insertions(+), 150 deletions(-) create mode 100644 net/core/dev_api.c (limited to 'include') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index abda17b15950..be3d09b61e95 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2620,16 +2620,35 @@ static inline void netdev_for_each_tx_queue(struct net_device *dev, f(dev, &dev->_tx[i], arg); } +static inline int netdev_lock_cmp_fn(const struct lockdep_map *a, + const struct lockdep_map *b) +{ + /* Only lower devices currently grab the instance lock, so no + * real ordering issues can occur. In the near future, only + * hardware devices will grab instance lock which also does not + * involve any ordering. Suppress lockdep ordering warnings + * until (if) we start grabbing instance lock on pure SW + * devices (bond/team/veth/etc). + */ + if (a == b) + return 0; + return -1; +} + #define netdev_lockdep_set_classes(dev) \ { \ static struct lock_class_key qdisc_tx_busylock_key; \ static struct lock_class_key qdisc_xmit_lock_key; \ static struct lock_class_key dev_addr_list_lock_key; \ + static struct lock_class_key dev_instance_lock_key; \ unsigned int i; \ \ (dev)->qdisc_tx_busylock = &qdisc_tx_busylock_key; \ lockdep_set_class(&(dev)->addr_list_lock, \ &dev_addr_list_lock_key); \ + lockdep_set_class(&(dev)->lock, \ + &dev_instance_lock_key); \ + lock_set_cmp_fn(&dev->lock, netdev_lock_cmp_fn, NULL); \ for (i = 0; i < (dev)->num_tx_queues; i++) \ lockdep_set_class(&(dev)->_tx[i]._xmit_lock, \ &qdisc_xmit_lock_key); \ @@ -2776,6 +2795,12 @@ static inline void netdev_unlock_ops(struct net_device *dev) netdev_unlock(dev); } +static inline void netdev_ops_assert_locked(struct net_device *dev) +{ + if (netdev_need_ops_lock(dev)) + lockdep_assert_held(&dev->lock); +} + void netif_napi_set_irq_locked(struct napi_struct *napi, int irq); static inline void netif_napi_set_irq(struct napi_struct *napi, int irq) @@ -3350,7 +3375,9 @@ struct net_device *dev_get_by_name_rcu(struct net *net, const char *name); struct net_device *__dev_get_by_name(struct net *net, const char *name); bool netdev_name_in_use(struct net *net, const char *name); int dev_alloc_name(struct net_device *dev, const char *name); +int netif_open(struct net_device *dev, struct netlink_ext_ack *extack); int dev_open(struct net_device *dev, struct netlink_ext_ack *extack); +void netif_close(struct net_device *dev); void dev_close(struct net_device *dev); void dev_close_many(struct list_head *head, bool unlink); int dev_setup_tc(struct net_device *dev, enum tc_setup_type type, @@ -4211,25 +4238,26 @@ int dev_ethtool(struct net *net, struct ifreq *ifr, void __user *userdata); unsigned int dev_get_flags(const struct net_device *); int __dev_change_flags(struct net_device *dev, unsigned int flags, struct netlink_ext_ack *extack); +int netif_change_flags(struct net_device *dev, unsigned int flags, + struct netlink_ext_ack *extack); int dev_change_flags(struct net_device *dev, unsigned int flags, struct netlink_ext_ack *extack); +int netif_set_alias(struct net_device *dev, const char *alias, size_t len); int dev_set_alias(struct net_device *, const char *, size_t); int dev_get_alias(const struct net_device *, char *, size_t); -int __dev_change_net_namespace(struct net_device *dev, struct net *net, +int netif_change_net_namespace(struct net_device *dev, struct net *net, const char *pat, int new_ifindex, struct netlink_ext_ack *extack); -static inline int dev_change_net_namespace(struct net_device *dev, struct net *net, - const char *pat) -{ - return __dev_change_net_namespace(dev, net, pat, 0, NULL); -} + const char *pat); int __dev_set_mtu(struct net_device *, int); int dev_set_mtu(struct net_device *, int); int dev_pre_changeaddr_notify(struct net_device *dev, const char *addr, struct netlink_ext_ack *extack); int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa, struct netlink_ext_ack *extack); +int netif_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, + struct netlink_ext_ack *extack); int dev_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, struct netlink_ext_ack *extack); int dev_get_mac_address(struct sockaddr *sa, struct net *net, char *dev_name); diff --git a/net/core/Makefile b/net/core/Makefile index d9326600e289..a10c3bd96798 100644 --- a/net/core/Makefile +++ b/net/core/Makefile @@ -9,7 +9,7 @@ obj-y := sock.o request_sock.o skbuff.o datagram.o stream.o scm.o \ obj-$(CONFIG_SYSCTL) += sysctl_net_core.o -obj-y += dev.o dev_addr_lists.o dst.o netevent.o \ +obj-y += dev.o dev_api.o dev_addr_lists.o dst.o netevent.o \ neighbour.o rtnetlink.o utils.o link_watch.o filter.o \ sock_diag.o dev_ioctl.o tso.o sock_reuseport.o \ fib_notifier.o xdp.o flow_offload.o gro.o \ diff --git a/net/core/dev.c b/net/core/dev.c index 57af25683ea1..39b214fb462e 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1371,15 +1371,7 @@ static int dev_get_valid_name(struct net *net, struct net_device *dev, return ret < 0 ? ret : 0; } -/** - * dev_change_name - change name of a device - * @dev: device - * @newname: name (or format string) must be at least IFNAMSIZ - * - * Change name of a device, can pass format strings "eth%d". - * for wildcarding. - */ -int dev_change_name(struct net_device *dev, const char *newname) +int netif_change_name(struct net_device *dev, const char *newname) { struct net *net = dev_net(dev); unsigned char old_assign_type; @@ -1449,15 +1441,7 @@ rollback: return err; } -/** - * dev_set_alias - change ifalias of a device - * @dev: device - * @alias: name up to IFALIASZ - * @len: limit of bytes to copy from info - * - * Set ifalias for a device, - */ -int dev_set_alias(struct net_device *dev, const char *alias, size_t len) +int netif_set_alias(struct net_device *dev, const char *alias, size_t len) { struct dev_ifalias *new_alias = NULL; @@ -1483,7 +1467,6 @@ int dev_set_alias(struct net_device *dev, const char *alias, size_t len) return len; } -EXPORT_SYMBOL(dev_set_alias); /** * dev_get_alias - get ifalias of a device @@ -1627,10 +1610,10 @@ static int __dev_open(struct net_device *dev, struct netlink_ext_ack *extack) if (ret) return ret; - netdev_lock_ops(dev); - set_bit(__LINK_STATE_START, &dev->state); + netdev_ops_assert_locked(dev); + if (ops->ndo_validate_addr) ret = ops->ndo_validate_addr(dev); @@ -1648,25 +1631,10 @@ static int __dev_open(struct net_device *dev, struct netlink_ext_ack *extack) add_device_randomness(dev->dev_addr, dev->addr_len); } - netdev_unlock_ops(dev); - return ret; } -/** - * dev_open - prepare an interface for use. - * @dev: device to open - * @extack: netlink extended ack - * - * Takes a device from down to up state. The device's private open - * function is invoked and then the multicast lists are loaded. Finally - * the device is moved into the up state and a %NETDEV_UP message is - * sent to the netdev notifier chain. - * - * Calling this function on an active interface is a nop. On a failure - * a negative errno code is returned. - */ -int dev_open(struct net_device *dev, struct netlink_ext_ack *extack) +int netif_open(struct net_device *dev, struct netlink_ext_ack *extack) { int ret; @@ -1682,7 +1650,6 @@ int dev_open(struct net_device *dev, struct netlink_ext_ack *extack) return ret; } -EXPORT_SYMBOL(dev_open); static void __dev_close_many(struct list_head *head) { @@ -1721,18 +1688,13 @@ static void __dev_close_many(struct list_head *head) * event. */ - /* TODO: move the lock up before clearing __LINK_STATE_START. - * Generates spurious lockdep warning. - */ - netdev_lock_ops(dev); + netdev_ops_assert_locked(dev); if (ops->ndo_stop) ops->ndo_stop(dev); netif_set_up(dev, false); netpoll_poll_enable(dev); - - netdev_unlock_ops(dev); } } @@ -1765,16 +1727,7 @@ void dev_close_many(struct list_head *head, bool unlink) } EXPORT_SYMBOL(dev_close_many); -/** - * dev_close - shutdown an interface. - * @dev: device to shutdown - * - * This function moves an active device into down state. A - * %NETDEV_GOING_DOWN is sent to the netdev notifier chain. The device - * is then deactivated and finally a %NETDEV_DOWN is sent to the notifier - * chain. - */ -void dev_close(struct net_device *dev) +void netif_close(struct net_device *dev) { if (dev->flags & IFF_UP) { LIST_HEAD(single); @@ -1784,7 +1737,6 @@ void dev_close(struct net_device *dev) list_del(&single); } } -EXPORT_SYMBOL(dev_close); int dev_setup_tc(struct net_device *dev, enum tc_setup_type type, void *type_data) @@ -9479,17 +9431,8 @@ void __dev_notify_flags(struct net_device *dev, unsigned int old_flags, } } -/** - * dev_change_flags - change device settings - * @dev: device - * @flags: device state flags - * @extack: netlink extended ack - * - * Change settings on device based state flags. The flags are - * in the userspace exported format. - */ -int dev_change_flags(struct net_device *dev, unsigned int flags, - struct netlink_ext_ack *extack) +int netif_change_flags(struct net_device *dev, unsigned int flags, + struct netlink_ext_ack *extack) { int ret; unsigned int changes, old_flags = dev->flags, old_gflags = dev->gflags; @@ -9502,7 +9445,6 @@ int dev_change_flags(struct net_device *dev, unsigned int flags, __dev_notify_flags(dev, old_flags, changes, 0, NULL); return ret; } -EXPORT_SYMBOL(dev_change_flags); int __dev_set_mtu(struct net_device *dev, int new_mtu) { @@ -9534,15 +9476,15 @@ int dev_validate_mtu(struct net_device *dev, int new_mtu, } /** - * dev_set_mtu_ext - Change maximum transfer unit + * netif_set_mtu_ext - Change maximum transfer unit * @dev: device * @new_mtu: new transfer unit * @extack: netlink extended ack * * Change the maximum transfer size of the network device. */ -int dev_set_mtu_ext(struct net_device *dev, int new_mtu, - struct netlink_ext_ack *extack) +int netif_set_mtu_ext(struct net_device *dev, int new_mtu, + struct netlink_ext_ack *extack) { int err, orig_mtu; @@ -9586,19 +9528,14 @@ int dev_set_mtu(struct net_device *dev, int new_mtu) int err; memset(&extack, 0, sizeof(extack)); - err = dev_set_mtu_ext(dev, new_mtu, &extack); + err = netif_set_mtu_ext(dev, new_mtu, &extack); if (err && extack._msg) net_err_ratelimited("%s: %s\n", dev->name, extack._msg); return err; } EXPORT_SYMBOL(dev_set_mtu); -/** - * dev_change_tx_queue_len - Change TX queue length of a netdevice - * @dev: device - * @new_len: new tx queue length - */ -int dev_change_tx_queue_len(struct net_device *dev, unsigned long new_len) +int netif_change_tx_queue_len(struct net_device *dev, unsigned long new_len) { unsigned int orig_len = dev->tx_queue_len; int res; @@ -9625,12 +9562,7 @@ err_rollback: return res; } -/** - * dev_set_group - Change group this device belongs to - * @dev: device - * @new_group: group this device should belong to - */ -void dev_set_group(struct net_device *dev, int new_group) +void netif_set_group(struct net_device *dev, int new_group) { dev->group = new_group; } @@ -9693,8 +9625,8 @@ EXPORT_SYMBOL(dev_set_mac_address); DECLARE_RWSEM(dev_addr_sem); -int dev_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, - struct netlink_ext_ack *extack) +int netif_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, + struct netlink_ext_ack *extack) { int ret; @@ -9703,7 +9635,6 @@ int dev_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, up_write(&dev_addr_sem); return ret; } -EXPORT_SYMBOL(dev_set_mac_address_user); int dev_get_mac_address(struct sockaddr *sa, struct net *net, char *dev_name) { @@ -9733,14 +9664,7 @@ unlock: } EXPORT_SYMBOL(dev_get_mac_address); -/** - * dev_change_carrier - Change device carrier - * @dev: device - * @new_carrier: new value - * - * Change device carrier - */ -int dev_change_carrier(struct net_device *dev, bool new_carrier) +int netif_change_carrier(struct net_device *dev, bool new_carrier) { const struct net_device_ops *ops = dev->netdev_ops; @@ -9851,13 +9775,7 @@ bool netdev_port_same_parent_id(struct net_device *a, struct net_device *b) } EXPORT_SYMBOL(netdev_port_same_parent_id); -/** - * dev_change_proto_down - set carrier according to proto_down. - * - * @dev: device - * @proto_down: new value - */ -int dev_change_proto_down(struct net_device *dev, bool proto_down) +int netif_change_proto_down(struct net_device *dev, bool proto_down) { if (!dev->change_proto_down) return -EOPNOTSUPP; @@ -9872,14 +9790,14 @@ int dev_change_proto_down(struct net_device *dev, bool proto_down) } /** - * dev_change_proto_down_reason - proto down reason + * netdev_change_proto_down_reason_locked - proto down reason * * @dev: device * @mask: proto down mask * @value: proto down value */ -void dev_change_proto_down_reason(struct net_device *dev, unsigned long mask, - u32 value) +void netdev_change_proto_down_reason_locked(struct net_device *dev, + unsigned long mask, u32 value) { u32 proto_down_reason; int b; @@ -10687,6 +10605,7 @@ int __netdev_update_features(struct net_device *dev) int err = -1; ASSERT_RTNL(); + netdev_ops_assert_locked(dev); features = netdev_get_wanted_features(dev); @@ -12036,11 +11955,14 @@ void unregister_netdevice_many_notify(struct list_head *head, } /* If device is running, close it first. */ - list_for_each_entry(dev, head, unreg_list) + list_for_each_entry(dev, head, unreg_list) { list_add_tail(&dev->close_list, &close_head); + netdev_lock_ops(dev); + } dev_close_many(&close_head, true); list_for_each_entry(dev, head, unreg_list) { + netdev_unlock_ops(dev); /* And unlink it from device chain. */ unlist_netdevice(dev); netdev_lock(dev); @@ -12153,24 +12075,7 @@ void unregister_netdev(struct net_device *dev) } EXPORT_SYMBOL(unregister_netdev); -/** - * __dev_change_net_namespace - move device to different nethost namespace - * @dev: device - * @net: network namespace - * @pat: If not NULL name pattern to try if the current device name - * is already taken in the destination network namespace. - * @new_ifindex: If not zero, specifies device index in the target - * namespace. - * @extack: netlink extended ack - * - * This function shuts down a device interface and moves it - * to a new network namespace. On success 0 is returned, on - * a failure a netagive errno code is returned. - * - * Callers must hold the rtnl semaphore. - */ - -int __dev_change_net_namespace(struct net_device *dev, struct net *net, +int netif_change_net_namespace(struct net_device *dev, struct net *net, const char *pat, int new_ifindex, struct netlink_ext_ack *extack) { @@ -12256,7 +12161,7 @@ int __dev_change_net_namespace(struct net_device *dev, struct net *net, */ /* If device is running close it first. */ - dev_close(dev); + netif_close(dev); /* And unlink it from device chain */ unlist_netdevice(dev); @@ -12338,7 +12243,6 @@ int __dev_change_net_namespace(struct net_device *dev, struct net *net, out: return err; } -EXPORT_SYMBOL_GPL(__dev_change_net_namespace); static int dev_cpu_dead(unsigned int oldcpu) { diff --git a/net/core/dev.h b/net/core/dev.h index 25bb9d6afbce..41b0831aba60 100644 --- a/net/core/dev.h +++ b/net/core/dev.h @@ -85,6 +85,7 @@ struct netdev_name_node { }; int netdev_get_name(struct net *net, char *name, int ifindex); +int netif_change_name(struct net_device *dev, const char *newname); int dev_change_name(struct net_device *dev, const char *newname); #define netdev_for_each_altname(dev, namenode) \ @@ -98,24 +99,28 @@ int netdev_name_node_alt_destroy(struct net_device *dev, const char *name); int dev_validate_mtu(struct net_device *dev, int mtu, struct netlink_ext_ack *extack); -int dev_set_mtu_ext(struct net_device *dev, int mtu, - struct netlink_ext_ack *extack); +int netif_set_mtu_ext(struct net_device *dev, int new_mtu, + struct netlink_ext_ack *extack); int dev_get_phys_port_id(struct net_device *dev, struct netdev_phys_item_id *ppid); int dev_get_phys_port_name(struct net_device *dev, char *name, size_t len); +int netif_change_proto_down(struct net_device *dev, bool proto_down); int dev_change_proto_down(struct net_device *dev, bool proto_down); -void dev_change_proto_down_reason(struct net_device *dev, unsigned long mask, - u32 value); +void netdev_change_proto_down_reason_locked(struct net_device *dev, + unsigned long mask, u32 value); typedef int (*bpf_op_t)(struct net_device *dev, struct netdev_bpf *bpf); int dev_change_xdp_fd(struct net_device *dev, struct netlink_ext_ack *extack, int fd, int expected_fd, u32 flags); +int netif_change_tx_queue_len(struct net_device *dev, unsigned long new_len); int dev_change_tx_queue_len(struct net_device *dev, unsigned long new_len); +void netif_set_group(struct net_device *dev, int new_group); void dev_set_group(struct net_device *dev, int new_group); +int netif_change_carrier(struct net_device *dev, bool new_carrier); int dev_change_carrier(struct net_device *dev, bool new_carrier); void __dev_set_rx_mode(struct net_device *dev); diff --git a/net/core/dev_api.c b/net/core/dev_api.c new file mode 100644 index 000000000000..21584f57e050 --- /dev/null +++ b/net/core/dev_api.c @@ -0,0 +1,224 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +#include + +#include "dev.h" + +/** + * dev_change_name() - change name of a device + * @dev: device + * @newname: name (or format string) must be at least IFNAMSIZ + * + * Change name of a device, can pass format strings "eth%d". + * for wildcarding. + * + * Return: 0 on success, -errno on failure. + */ +int dev_change_name(struct net_device *dev, const char *newname) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_change_name(dev, newname); + netdev_unlock_ops(dev); + + return ret; +} + +/** + * dev_set_alias() - change ifalias of a device + * @dev: device + * @alias: name up to IFALIASZ + * @len: limit of bytes to copy from info + * + * Set ifalias for a device. + * + * Return: 0 on success, -errno on failure. + */ +int dev_set_alias(struct net_device *dev, const char *alias, size_t len) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_set_alias(dev, alias, len); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL(dev_set_alias); + +/** + * dev_change_flags() - change device settings + * @dev: device + * @flags: device state flags + * @extack: netlink extended ack + * + * Change settings on device based state flags. The flags are + * in the userspace exported format. + * + * Return: 0 on success, -errno on failure. + */ +int dev_change_flags(struct net_device *dev, unsigned int flags, + struct netlink_ext_ack *extack) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_change_flags(dev, flags, extack); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL(dev_change_flags); + +/** + * dev_set_group() - change group this device belongs to + * @dev: device + * @new_group: group this device should belong to + */ +void dev_set_group(struct net_device *dev, int new_group) +{ + netdev_lock_ops(dev); + netif_set_group(dev, new_group); + netdev_unlock_ops(dev); +} + +int dev_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, + struct netlink_ext_ack *extack) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_set_mac_address_user(dev, sa, extack); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL(dev_set_mac_address_user); + +/** + * dev_change_net_namespace() - move device to different nethost namespace + * @dev: device + * @net: network namespace + * @pat: If not NULL name pattern to try if the current device name + * is already taken in the destination network namespace. + * + * This function shuts down a device interface and moves it + * to a new network namespace. On success 0 is returned, on + * a failure a netagive errno code is returned. + * + * Callers must hold the rtnl semaphore. + * + * Return: 0 on success, -errno on failure. + */ +int dev_change_net_namespace(struct net_device *dev, struct net *net, + const char *pat) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_change_net_namespace(dev, net, pat, 0, NULL); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL_GPL(dev_change_net_namespace); + +/** + * dev_change_carrier() - change device carrier + * @dev: device + * @new_carrier: new value + * + * Change device carrier + * + * Return: 0 on success, -errno on failure. + */ +int dev_change_carrier(struct net_device *dev, bool new_carrier) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_change_carrier(dev, new_carrier); + netdev_unlock_ops(dev); + + return ret; +} + +/** + * dev_change_tx_queue_len() - change TX queue length of a netdevice + * @dev: device + * @new_len: new tx queue length + * + * Return: 0 on success, -errno on failure. + */ +int dev_change_tx_queue_len(struct net_device *dev, unsigned long new_len) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_change_tx_queue_len(dev, new_len); + netdev_unlock_ops(dev); + + return ret; +} + +/** + * dev_change_proto_down() - set carrier according to proto_down + * @dev: device + * @proto_down: new value + * + * Return: 0 on success, -errno on failure. + */ +int dev_change_proto_down(struct net_device *dev, bool proto_down) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_change_proto_down(dev, proto_down); + netdev_unlock_ops(dev); + + return ret; +} + +/** + * dev_open() - prepare an interface for use + * @dev: device to open + * @extack: netlink extended ack + * + * Takes a device from down to up state. The device's private open + * function is invoked and then the multicast lists are loaded. Finally + * the device is moved into the up state and a %NETDEV_UP message is + * sent to the netdev notifier chain. + * + * Calling this function on an active interface is a nop. On a failure + * a negative errno code is returned. + * + * Return: 0 on success, -errno on failure. + */ +int dev_open(struct net_device *dev, struct netlink_ext_ack *extack) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_open(dev, extack); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL(dev_open); + +/** + * dev_close() - shutdown an interface + * @dev: device to shutdown + * + * This function moves an active device into down state. A + * %NETDEV_GOING_DOWN is sent to the netdev notifier chain. The device + * is then deactivated and finally a %NETDEV_DOWN is sent to the notifier + * chain. + */ +void dev_close(struct net_device *dev) +{ + netdev_lock_ops(dev); + netif_close(dev); + netdev_unlock_ops(dev); +} +EXPORT_SYMBOL(dev_close); diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index b4612d305970..9d539c9ce1a4 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -2912,12 +2912,19 @@ static int do_set_master(struct net_device *dev, int ifindex, const struct net_device_ops *ops; int err; + /* Release the lower lock, the upper is responsible for locking + * the lower if needed. None of the existing upper devices + * use netdev instance lock, so don't grab it. + */ + if (upper_dev) { if (upper_dev->ifindex == ifindex) return 0; ops = upper_dev->netdev_ops; if (ops->ndo_del_slave) { + netdev_unlock_ops(dev); err = ops->ndo_del_slave(upper_dev, dev); + netdev_lock_ops(dev); if (err) return err; } else { @@ -2931,7 +2938,9 @@ static int do_set_master(struct net_device *dev, int ifindex, return -EINVAL; ops = upper_dev->netdev_ops; if (ops->ndo_add_slave) { + netdev_unlock_ops(dev); err = ops->ndo_add_slave(upper_dev, dev, extack); + netdev_lock_ops(dev); if (err) return err; } else { @@ -2981,7 +2990,7 @@ static int do_set_proto_down(struct net_device *dev, if (pdreason[IFLA_PROTO_DOWN_REASON_MASK]) mask = nla_get_u32(pdreason[IFLA_PROTO_DOWN_REASON_MASK]); - dev_change_proto_down_reason(dev, mask, value); + netdev_change_proto_down_reason_locked(dev, mask, value); } if (nl_proto_down) { @@ -2992,8 +3001,7 @@ static int do_set_proto_down(struct net_device *dev, NL_SET_ERR_MSG(extack, "Cannot clear protodown, active reasons"); return -EBUSY; } - err = dev_change_proto_down(dev, - proto_down); + err = netif_change_proto_down(dev, proto_down); if (err) return err; } @@ -3013,6 +3021,8 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev, char ifname[IFNAMSIZ]; int err; + netdev_lock_ops(dev); + err = validate_linkmsg(dev, tb, extack); if (err < 0) goto errout; @@ -3028,7 +3038,8 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev, new_ifindex = nla_get_s32_default(tb[IFLA_NEW_IFINDEX], 0); - err = __dev_change_net_namespace(dev, tgt_net, pat, new_ifindex, extack); + err = netif_change_net_namespace(dev, tgt_net, pat, + new_ifindex, extack); if (err) goto errout; @@ -3078,7 +3089,7 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev, sa->sa_family = dev->type; memcpy(sa->sa_data, nla_data(tb[IFLA_ADDRESS]), dev->addr_len); - err = dev_set_mac_address_user(dev, sa, extack); + err = netif_set_mac_address_user(dev, sa, extack); kfree(sa); if (err) goto errout; @@ -3086,14 +3097,14 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev, } if (tb[IFLA_MTU]) { - err = dev_set_mtu_ext(dev, nla_get_u32(tb[IFLA_MTU]), extack); + err = netif_set_mtu_ext(dev, nla_get_u32(tb[IFLA_MTU]), extack); if (err < 0) goto errout; status |= DO_SETLINK_MODIFIED; } if (tb[IFLA_GROUP]) { - dev_set_group(dev, nla_get_u32(tb[IFLA_GROUP])); + netif_set_group(dev, nla_get_u32(tb[IFLA_GROUP])); status |= DO_SETLINK_NOTIFY; } @@ -3103,15 +3114,15 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev, * requested. */ if (ifm->ifi_index > 0 && ifname[0]) { - err = dev_change_name(dev, ifname); + err = netif_change_name(dev, ifname); if (err < 0) goto errout; status |= DO_SETLINK_MODIFIED; } if (tb[IFLA_IFALIAS]) { - err = dev_set_alias(dev, nla_data(tb[IFLA_IFALIAS]), - nla_len(tb[IFLA_IFALIAS])); + err = netif_set_alias(dev, nla_data(tb[IFLA_IFALIAS]), + nla_len(tb[IFLA_IFALIAS])); if (err < 0) goto errout; status |= DO_SETLINK_NOTIFY; @@ -3123,8 +3134,8 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev, } if (ifm->ifi_flags || ifm->ifi_change) { - err = dev_change_flags(dev, rtnl_dev_combine_flags(dev, ifm), - extack); + err = netif_change_flags(dev, rtnl_dev_combine_flags(dev, ifm), + extack); if (err < 0) goto errout; } @@ -3137,7 +3148,7 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev, } if (tb[IFLA_CARRIER]) { - err = dev_change_carrier(dev, nla_get_u8(tb[IFLA_CARRIER])); + err = netif_change_carrier(dev, nla_get_u8(tb[IFLA_CARRIER])); if (err) goto errout; status |= DO_SETLINK_MODIFIED; @@ -3146,7 +3157,7 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev, if (tb[IFLA_TXQLEN]) { unsigned int value = nla_get_u32(tb[IFLA_TXQLEN]); - err = dev_change_tx_queue_len(dev, value); + err = netif_change_tx_queue_len(dev, value); if (err) goto errout; status |= DO_SETLINK_MODIFIED; @@ -3377,6 +3388,8 @@ errout: dev->name); } + netdev_unlock_ops(dev); + return err; } @@ -3810,6 +3823,8 @@ static int rtnl_newlink_create(struct sk_buff *skb, struct ifinfomsg *ifm, goto out; } + netdev_lock_ops(dev); + err = rtnl_configure_link(dev, ifm, portid, nlh); if (err < 0) goto out_unregister; @@ -3818,9 +3833,12 @@ static int rtnl_newlink_create(struct sk_buff *skb, struct ifinfomsg *ifm, if (err) goto out_unregister; } + + netdev_unlock_ops(dev); out: return err; out_unregister: + netdev_unlock_ops(dev); if (ops->newlink) { LIST_HEAD(list_kill); -- cgit v1.2.3 From ffb7ed19ac0a9fa9ea79af1d7b42c03a10da98a5 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:25 -0800 Subject: net: hold netdev instance lock during ioctl operations Convert all ndo_eth_ioctl invocations to dev_eth_ioctl which does the locking. Reflow some of the dev_siocxxx to drop else clause. Cc: Saeed Mahameed Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-8-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- drivers/net/bonding/bond_main.c | 9 +++--- include/linux/netdevice.h | 3 ++ net/8021q/vlan_dev.c | 4 +-- net/core/dev.c | 4 +-- net/core/dev_api.c | 30 ++++++++++++++++++ net/core/dev_ioctl.c | 67 ++++++++++++++++++++++++----------------- 6 files changed, 80 insertions(+), 37 deletions(-) (limited to 'include') diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index fe7072336e8f..cd4afeb9ad42 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -855,7 +855,6 @@ static int bond_check_dev_link(struct bonding *bond, struct net_device *slave_dev, int reporting) { const struct net_device_ops *slave_ops = slave_dev->netdev_ops; - int (*ioctl)(struct net_device *, struct ifreq *, int); struct ifreq ifr; struct mii_ioctl_data *mii; @@ -871,8 +870,7 @@ static int bond_check_dev_link(struct bonding *bond, BMSR_LSTATUS : 0; /* Ethtool can't be used, fallback to MII ioctls. */ - ioctl = slave_ops->ndo_eth_ioctl; - if (ioctl) { + if (slave_ops->ndo_eth_ioctl) { /* TODO: set pointer to correct ioctl on a per team member * bases to make this more efficient. that is, once * we determine the correct ioctl, we will always @@ -888,9 +886,10 @@ static int bond_check_dev_link(struct bonding *bond, /* Yes, the mii is overlaid on the ifreq.ifr_ifru */ strscpy_pad(ifr.ifr_name, slave_dev->name, IFNAMSIZ); mii = if_mii(&ifr); - if (ioctl(slave_dev, &ifr, SIOCGMIIPHY) == 0) { + + if (dev_eth_ioctl(slave_dev, &ifr, SIOCGMIIPHY) == 0) { mii->reg_num = MII_BMSR; - if (ioctl(slave_dev, &ifr, SIOCGMIIREG) == 0) + if (dev_eth_ioctl(slave_dev, &ifr, SIOCGMIIREG) == 0) return mii->val_out & BMSR_LSTATUS; } } diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index be3d09b61e95..8d243c0ec39d 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -4229,6 +4229,8 @@ int put_user_ifreq(struct ifreq *ifr, void __user *arg); int dev_ioctl(struct net *net, unsigned int cmd, struct ifreq *ifr, void __user *data, bool *need_copyout); int dev_ifconf(struct net *net, struct ifconf __user *ifc); +int dev_eth_ioctl(struct net_device *dev, + struct ifreq *ifr, unsigned int cmd); int generic_hwtstamp_get_lower(struct net_device *dev, struct kernel_hwtstamp_config *kernel_cfg); int generic_hwtstamp_set_lower(struct net_device *dev, @@ -4251,6 +4253,7 @@ int netif_change_net_namespace(struct net_device *dev, struct net *net, int dev_change_net_namespace(struct net_device *dev, struct net *net, const char *pat); int __dev_set_mtu(struct net_device *, int); +int netif_set_mtu(struct net_device *dev, int new_mtu); int dev_set_mtu(struct net_device *, int); int dev_pre_changeaddr_notify(struct net_device *dev, const char *addr, struct netlink_ext_ack *extack); diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 91d134961357..ee3283400716 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -377,7 +377,6 @@ static int vlan_hwtstamp_set(struct net_device *dev, static int vlan_dev_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) { struct net_device *real_dev = vlan_dev_priv(dev)->real_dev; - const struct net_device_ops *ops = real_dev->netdev_ops; struct ifreq ifrr; int err = -EOPNOTSUPP; @@ -388,8 +387,7 @@ static int vlan_dev_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) case SIOCGMIIPHY: case SIOCGMIIREG: case SIOCSMIIREG: - if (netif_device_present(real_dev) && ops->ndo_eth_ioctl) - err = ops->ndo_eth_ioctl(real_dev, &ifrr, cmd); + err = dev_eth_ioctl(real_dev, &ifrr, cmd); break; } diff --git a/net/core/dev.c b/net/core/dev.c index 39b214fb462e..97a4fc0889d3 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -9522,7 +9522,7 @@ int netif_set_mtu_ext(struct net_device *dev, int new_mtu, return err; } -int dev_set_mtu(struct net_device *dev, int new_mtu) +int netif_set_mtu(struct net_device *dev, int new_mtu) { struct netlink_ext_ack extack; int err; @@ -9533,7 +9533,7 @@ int dev_set_mtu(struct net_device *dev, int new_mtu) net_err_ratelimited("%s: %s\n", dev->name, extack._msg); return err; } -EXPORT_SYMBOL(dev_set_mtu); +EXPORT_SYMBOL(netif_set_mtu); int netif_change_tx_queue_len(struct net_device *dev, unsigned long new_len) { diff --git a/net/core/dev_api.c b/net/core/dev_api.c index 21584f57e050..059413d9ef9d 100644 --- a/net/core/dev_api.c +++ b/net/core/dev_api.c @@ -222,3 +222,33 @@ void dev_close(struct net_device *dev) netdev_unlock_ops(dev); } EXPORT_SYMBOL(dev_close); + +int dev_eth_ioctl(struct net_device *dev, + struct ifreq *ifr, unsigned int cmd) +{ + const struct net_device_ops *ops = dev->netdev_ops; + int ret = -ENODEV; + + if (!ops->ndo_eth_ioctl) + return -EOPNOTSUPP; + + netdev_lock_ops(dev); + if (netif_device_present(dev)) + ret = ops->ndo_eth_ioctl(dev, ifr, cmd); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL(dev_eth_ioctl); + +int dev_set_mtu(struct net_device *dev, int new_mtu) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_set_mtu(dev, new_mtu); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL(dev_set_mtu); diff --git a/net/core/dev_ioctl.c b/net/core/dev_ioctl.c index 4c2098ac9d72..d9f350593121 100644 --- a/net/core/dev_ioctl.c +++ b/net/core/dev_ioctl.c @@ -110,7 +110,7 @@ static int dev_getifmap(struct net_device *dev, struct ifreq *ifr) return 0; } -static int dev_setifmap(struct net_device *dev, struct ifreq *ifr) +static int netif_setifmap(struct net_device *dev, struct ifreq *ifr) { struct compat_ifmap *cifmap = (struct compat_ifmap *)&ifr->ifr_map; @@ -240,20 +240,6 @@ int net_hwtstamp_validate(const struct kernel_hwtstamp_config *cfg) return 0; } -static int dev_eth_ioctl(struct net_device *dev, - struct ifreq *ifr, unsigned int cmd) -{ - const struct net_device_ops *ops = dev->netdev_ops; - - if (!ops->ndo_eth_ioctl) - return -EOPNOTSUPP; - - if (!netif_device_present(dev)) - return -ENODEV; - - return ops->ndo_eth_ioctl(dev, ifr, cmd); -} - /** * dev_get_hwtstamp_phylib() - Get hardware timestamping settings of NIC * or of attached phylib PHY @@ -305,7 +291,9 @@ static int dev_get_hwtstamp(struct net_device *dev, struct ifreq *ifr) return -ENODEV; kernel_cfg.ifr = ifr; + netdev_lock_ops(dev); err = dev_get_hwtstamp_phylib(dev, &kernel_cfg); + netdev_unlock_ops(dev); if (err) return err; @@ -429,7 +417,9 @@ static int dev_set_hwtstamp(struct net_device *dev, struct ifreq *ifr) if (!netif_device_present(dev)) return -ENODEV; + netdev_lock_ops(dev); err = dev_set_hwtstamp_phylib(dev, &kernel_cfg, &extack); + netdev_unlock_ops(dev); if (err) return err; @@ -504,10 +494,14 @@ static int dev_siocbond(struct net_device *dev, const struct net_device_ops *ops = dev->netdev_ops; if (ops->ndo_siocbond) { + int ret = -ENODEV; + + netdev_lock_ops(dev); if (netif_device_present(dev)) - return ops->ndo_siocbond(dev, ifr, cmd); - else - return -ENODEV; + ret = ops->ndo_siocbond(dev, ifr, cmd); + netdev_unlock_ops(dev); + + return ret; } return -EOPNOTSUPP; @@ -519,10 +513,14 @@ static int dev_siocdevprivate(struct net_device *dev, struct ifreq *ifr, const struct net_device_ops *ops = dev->netdev_ops; if (ops->ndo_siocdevprivate) { + int ret = -ENODEV; + + netdev_lock_ops(dev); if (netif_device_present(dev)) - return ops->ndo_siocdevprivate(dev, ifr, data, cmd); - else - return -ENODEV; + ret = ops->ndo_siocdevprivate(dev, ifr, data, cmd); + netdev_unlock_ops(dev); + + return ret; } return -EOPNOTSUPP; @@ -533,10 +531,14 @@ static int dev_siocwandev(struct net_device *dev, struct if_settings *ifs) const struct net_device_ops *ops = dev->netdev_ops; if (ops->ndo_siocwandev) { + int ret = -ENODEV; + + netdev_lock_ops(dev); if (netif_device_present(dev)) - return ops->ndo_siocwandev(dev, ifs); - else - return -ENODEV; + ret = ops->ndo_siocwandev(dev, ifs); + netdev_unlock_ops(dev); + + return ret; } return -EOPNOTSUPP; @@ -580,11 +582,16 @@ static int dev_ifsioc(struct net *net, struct ifreq *ifr, void __user *data, memcpy(dev->broadcast, ifr->ifr_hwaddr.sa_data, min(sizeof(ifr->ifr_hwaddr.sa_data_min), (size_t)dev->addr_len)); + netdev_lock_ops(dev); call_netdevice_notifiers(NETDEV_CHANGEADDR, dev); + netdev_unlock_ops(dev); return 0; case SIOCSIFMAP: - return dev_setifmap(dev, ifr); + netdev_lock_ops(dev); + err = netif_setifmap(dev, ifr); + netdev_unlock_ops(dev); + return err; case SIOCADDMULTI: if (!ops->ndo_set_rx_mode || @@ -592,7 +599,10 @@ static int dev_ifsioc(struct net *net, struct ifreq *ifr, void __user *data, return -EINVAL; if (!netif_device_present(dev)) return -ENODEV; - return dev_mc_add_global(dev, ifr->ifr_hwaddr.sa_data); + netdev_lock_ops(dev); + err = dev_mc_add_global(dev, ifr->ifr_hwaddr.sa_data); + netdev_unlock_ops(dev); + return err; case SIOCDELMULTI: if (!ops->ndo_set_rx_mode || @@ -600,7 +610,10 @@ static int dev_ifsioc(struct net *net, struct ifreq *ifr, void __user *data, return -EINVAL; if (!netif_device_present(dev)) return -ENODEV; - return dev_mc_del_global(dev, ifr->ifr_hwaddr.sa_data); + netdev_lock_ops(dev); + err = dev_mc_del_global(dev, ifr->ifr_hwaddr.sa_data); + netdev_unlock_ops(dev); + return err; case SIOCSIFTXQLEN: if (ifr->ifr_qlen < 0) -- cgit v1.2.3 From ad7c7b2172c388818a111455643491d75f535e90 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:26 -0800 Subject: net: hold netdev instance lock during sysfs operations Most of them are already covered by the converted dev_xxx APIs. Add the locking wrappers for the remaining ones. Cc: Saeed Mahameed Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-9-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- drivers/net/bonding/bond_main.c | 7 +++-- include/linux/netdevice.h | 4 +++ net/core/dev.c | 58 ++++++++---------------------------- net/core/dev_api.c | 65 +++++++++++++++++++++++++++++++++++++++++ net/core/net-sysfs.c | 2 ++ 5 files changed, 88 insertions(+), 48 deletions(-) (limited to 'include') diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index cd4afeb9ad42..cf0b02720dd8 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -2644,10 +2644,13 @@ static int __bond_release_one(struct net_device *bond_dev, dev_set_mac_address(slave_dev, (struct sockaddr *)&ss, NULL); } - if (unregister) + if (unregister) { + netdev_lock_ops(slave_dev); __dev_set_mtu(slave_dev, slave->original_mtu); - else + netdev_unlock_ops(slave_dev); + } else { dev_set_mtu(slave_dev, slave->original_mtu); + } if (!netif_is_bond_master(slave_dev)) slave_dev->priv_flags &= ~IFF_BONDING; diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 8d243c0ec39d..c61b12809588 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -3382,6 +3382,7 @@ void dev_close(struct net_device *dev); void dev_close_many(struct list_head *head, bool unlink); int dev_setup_tc(struct net_device *dev, enum tc_setup_type type, void *type_data); +void netif_disable_lro(struct net_device *dev); void dev_disable_lro(struct net_device *dev); int dev_loopback_xmit(struct net *net, struct sock *sk, struct sk_buff *newskb); u16 dev_pick_tx_zero(struct net_device *dev, struct sk_buff *skb, @@ -4257,6 +4258,8 @@ int netif_set_mtu(struct net_device *dev, int new_mtu); int dev_set_mtu(struct net_device *, int); int dev_pre_changeaddr_notify(struct net_device *dev, const char *addr, struct netlink_ext_ack *extack); +int netif_set_mac_address(struct net_device *dev, struct sockaddr *sa, + struct netlink_ext_ack *extack); int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa, struct netlink_ext_ack *extack); int netif_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, @@ -5016,6 +5019,7 @@ static inline void __dev_mc_unsync(struct net_device *dev, /* Functions used for secondary unicast and multicast support */ void dev_set_rx_mode(struct net_device *dev); int dev_set_promiscuity(struct net_device *dev, int inc); +int netif_set_allmulti(struct net_device *dev, int inc, bool notify); int dev_set_allmulti(struct net_device *dev, int inc); void netdev_state_change(struct net_device *dev); void __netdev_notify_peers(struct net_device *dev); diff --git a/net/core/dev.c b/net/core/dev.c index 97a4fc0889d3..121c0449f87f 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1757,15 +1757,7 @@ int dev_setup_tc(struct net_device *dev, enum tc_setup_type type, } EXPORT_SYMBOL(dev_setup_tc); -/** - * dev_disable_lro - disable Large Receive Offload on a device - * @dev: device - * - * Disable Large Receive Offload (LRO) on a net device. Must be - * called under RTNL. This is needed if received packets may be - * forwarded to another interface. - */ -void dev_disable_lro(struct net_device *dev) +void netif_disable_lro(struct net_device *dev) { struct net_device *lower_dev; struct list_head *iter; @@ -1776,10 +1768,12 @@ void dev_disable_lro(struct net_device *dev) if (unlikely(dev->features & NETIF_F_LRO)) netdev_WARN(dev, "failed to disable LRO!\n"); - netdev_for_each_lower_dev(dev, lower_dev, iter) - dev_disable_lro(lower_dev); + netdev_for_each_lower_dev(dev, lower_dev, iter) { + netdev_lock_ops(lower_dev); + netif_disable_lro(lower_dev); + netdev_unlock_ops(lower_dev); + } } -EXPORT_SYMBOL(dev_disable_lro); /** * dev_disable_gro_hw - disable HW Generic Receive Offload on a device @@ -6038,7 +6032,7 @@ static int generic_xdp_install(struct net_device *dev, struct netdev_bpf *xdp) static_branch_dec(&generic_xdp_needed_key); } else if (new && !old) { static_branch_inc(&generic_xdp_needed_key); - dev_disable_lro(dev); + netif_disable_lro(dev); dev_disable_gro_hw(dev); } break; @@ -9210,7 +9204,7 @@ int dev_set_promiscuity(struct net_device *dev, int inc) } EXPORT_SYMBOL(dev_set_promiscuity); -static int __dev_set_allmulti(struct net_device *dev, int inc, bool notify) +int netif_set_allmulti(struct net_device *dev, int inc, bool notify) { unsigned int old_flags = dev->flags, old_gflags = dev->gflags; unsigned int allmulti, flags; @@ -9245,25 +9239,6 @@ static int __dev_set_allmulti(struct net_device *dev, int inc, bool notify) return 0; } -/** - * dev_set_allmulti - update allmulti count on a device - * @dev: device - * @inc: modifier - * - * Add or remove reception of all multicast frames to a device. While the - * count in the device remains above zero the interface remains listening - * to all interfaces. Once it hits zero the device reverts back to normal - * filtering operation. A negative @inc value is used to drop the counter - * when releasing a resource needing all multicasts. - * Return 0 if successful or a negative errno code on error. - */ - -int dev_set_allmulti(struct net_device *dev, int inc) -{ - return __dev_set_allmulti(dev, inc, true); -} -EXPORT_SYMBOL(dev_set_allmulti); - /* * Upload unicast and multicast address lists to device and * configure RX filtering. When the device doesn't support unicast @@ -9396,7 +9371,7 @@ int __dev_change_flags(struct net_device *dev, unsigned int flags, int inc = (flags & IFF_ALLMULTI) ? 1 : -1; dev->gflags ^= IFF_ALLMULTI; - __dev_set_allmulti(dev, inc, false); + netif_set_allmulti(dev, inc, false); } return ret; @@ -9588,16 +9563,8 @@ int dev_pre_changeaddr_notify(struct net_device *dev, const char *addr, } EXPORT_SYMBOL(dev_pre_changeaddr_notify); -/** - * dev_set_mac_address - Change Media Access Control Address - * @dev: device - * @sa: new address - * @extack: netlink extended ack - * - * Change the hardware (MAC) address of the device - */ -int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa, - struct netlink_ext_ack *extack) +int netif_set_mac_address(struct net_device *dev, struct sockaddr *sa, + struct netlink_ext_ack *extack) { const struct net_device_ops *ops = dev->netdev_ops; int err; @@ -9621,7 +9588,6 @@ int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa, add_device_randomness(dev->dev_addr, dev->addr_len); return 0; } -EXPORT_SYMBOL(dev_set_mac_address); DECLARE_RWSEM(dev_addr_sem); @@ -9631,7 +9597,7 @@ int netif_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, int ret; down_write(&dev_addr_sem); - ret = dev_set_mac_address(dev, sa, extack); + ret = netif_set_mac_address(dev, sa, extack); up_write(&dev_addr_sem); return ret; } diff --git a/net/core/dev_api.c b/net/core/dev_api.c index 059413d9ef9d..7bd667b34b80 100644 --- a/net/core/dev_api.c +++ b/net/core/dev_api.c @@ -252,3 +252,68 @@ int dev_set_mtu(struct net_device *dev, int new_mtu) return ret; } EXPORT_SYMBOL(dev_set_mtu); + +/** + * dev_disable_lro() - disable Large Receive Offload on a device + * @dev: device + * + * Disable Large Receive Offload (LRO) on a net device. Must be + * called under RTNL. This is needed if received packets may be + * forwarded to another interface. + */ +void dev_disable_lro(struct net_device *dev) +{ + netdev_lock_ops(dev); + netif_disable_lro(dev); + netdev_unlock_ops(dev); +} +EXPORT_SYMBOL(dev_disable_lro); + +/** + * dev_set_allmulti() - update allmulti count on a device + * @dev: device + * @inc: modifier + * + * Add or remove reception of all multicast frames to a device. While the + * count in the device remains above zero the interface remains listening + * to all interfaces. Once it hits zero the device reverts back to normal + * filtering operation. A negative @inc value is used to drop the counter + * when releasing a resource needing all multicasts. + * + * Return: 0 on success, -errno on failure. + */ + +int dev_set_allmulti(struct net_device *dev, int inc) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_set_allmulti(dev, inc, true); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL(dev_set_allmulti); + +/** + * dev_set_mac_address() - change Media Access Control Address + * @dev: device + * @sa: new address + * @extack: netlink extended ack + * + * Change the hardware (MAC) address of the device + * + * Return: 0 on success, -errno on failure. + */ +int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa, + struct netlink_ext_ack *extack) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_set_mac_address(dev, sa, extack); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL(dev_set_mac_address); diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index 8d9dc048a548..27eea34d1b41 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -1483,8 +1483,10 @@ static ssize_t tx_maxrate_store(struct kobject *kobj, struct attribute *attr, return err; err = -EOPNOTSUPP; + netdev_lock_ops(dev); if (dev->netdev_ops->ndo_set_tx_maxrate) err = dev->netdev_ops->ndo_set_tx_maxrate(dev, index, rate); + netdev_unlock_ops(dev); if (!err) { queue->tx_maxrate = rate; -- cgit v1.2.3 From 97246d6d21c21fb4c5235770a21855e457096a96 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:27 -0800 Subject: net: hold netdev instance lock during ndo_bpf Cover the paths that come via bpf system call and XSK bind. Cc: Saeed Mahameed Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-10-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- include/linux/netdevice.h | 1 + kernel/bpf/offload.c | 6 ++++-- net/core/dev.c | 13 +++++++++++-- net/core/dev_api.c | 12 ++++++++++++ net/xdp/xsk.c | 3 +++ net/xdp/xsk_buff_pool.c | 2 ++ 6 files changed, 33 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index c61b12809588..ca9c09dab14e 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -4277,6 +4277,7 @@ struct sk_buff *dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev, int bpf_xdp_link_attach(const union bpf_attr *attr, struct bpf_prog *prog); u8 dev_xdp_prog_count(struct net_device *dev); +int netif_xdp_propagate(struct net_device *dev, struct netdev_bpf *bpf); int dev_xdp_propagate(struct net_device *dev, struct netdev_bpf *bpf); u8 dev_xdp_sb_prog_count(struct net_device *dev); u32 dev_xdp_prog_id(struct net_device *dev, enum bpf_xdp_mode mode); diff --git a/kernel/bpf/offload.c b/kernel/bpf/offload.c index 1a4fec330eaa..a10153c3be2d 100644 --- a/kernel/bpf/offload.c +++ b/kernel/bpf/offload.c @@ -528,10 +528,10 @@ struct bpf_map *bpf_map_offload_map_alloc(union bpf_attr *attr) return ERR_PTR(-ENOMEM); bpf_map_init_from_attr(&offmap->map, attr); - rtnl_lock(); - down_write(&bpf_devs_lock); offmap->netdev = __dev_get_by_index(net, attr->map_ifindex); + netdev_lock_ops(offmap->netdev); + down_write(&bpf_devs_lock); err = bpf_dev_offload_check(offmap->netdev); if (err) goto err_unlock; @@ -548,12 +548,14 @@ struct bpf_map *bpf_map_offload_map_alloc(union bpf_attr *attr) list_add_tail(&offmap->offloads, &ondev->maps); up_write(&bpf_devs_lock); + netdev_unlock_ops(offmap->netdev); rtnl_unlock(); return &offmap->map; err_unlock: up_write(&bpf_devs_lock); + netdev_unlock_ops(offmap->netdev); rtnl_unlock(); bpf_map_area_free(offmap); return ERR_PTR(err); diff --git a/net/core/dev.c b/net/core/dev.c index 121c0449f87f..404047d4d943 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -9852,7 +9852,7 @@ u8 dev_xdp_sb_prog_count(struct net_device *dev) return count; } -int dev_xdp_propagate(struct net_device *dev, struct netdev_bpf *bpf) +int netif_xdp_propagate(struct net_device *dev, struct netdev_bpf *bpf) { if (!dev->netdev_ops->ndo_bpf) return -EOPNOTSUPP; @@ -9872,7 +9872,6 @@ int dev_xdp_propagate(struct net_device *dev, struct netdev_bpf *bpf) return dev->netdev_ops->ndo_bpf(dev, bpf); } -EXPORT_SYMBOL_GPL(dev_xdp_propagate); u32 dev_xdp_prog_id(struct net_device *dev, enum bpf_xdp_mode mode) { @@ -9902,6 +9901,8 @@ static int dev_xdp_install(struct net_device *dev, enum bpf_xdp_mode mode, struct netdev_bpf xdp; int err; + netdev_ops_assert_locked(dev); + if (dev->cfg->hds_config == ETHTOOL_TCP_DATA_SPLIT_ENABLED && prog && !prog->aux->xdp_has_frags) { NL_SET_ERR_MSG(extack, "unable to install XDP to device using tcp-data-split"); @@ -10134,7 +10135,9 @@ static void bpf_xdp_link_release(struct bpf_link *link) * already NULL, in which case link was already auto-detached */ if (xdp_link->dev) { + netdev_lock_ops(xdp_link->dev); WARN_ON(dev_xdp_detach_link(xdp_link->dev, NULL, xdp_link)); + netdev_unlock_ops(xdp_link->dev); xdp_link->dev = NULL; } @@ -10216,10 +10219,12 @@ static int bpf_xdp_link_update(struct bpf_link *link, struct bpf_prog *new_prog, goto out_unlock; } + netdev_lock_ops(xdp_link->dev); mode = dev_xdp_mode(xdp_link->dev, xdp_link->flags); bpf_op = dev_xdp_bpf_op(xdp_link->dev, mode); err = dev_xdp_install(xdp_link->dev, mode, bpf_op, NULL, xdp_link->flags, new_prog); + netdev_unlock_ops(xdp_link->dev); if (err) goto out_unlock; @@ -11005,7 +11010,9 @@ int register_netdevice(struct net_device *dev) if (ret) goto err_uninit_notify; + netdev_lock_ops(dev); __netdev_update_features(dev); + netdev_unlock_ops(dev); /* * Default initial state at registry is that the @@ -11945,7 +11952,9 @@ void unregister_netdevice_many_notify(struct list_head *head, /* Shutdown queueing discipline. */ dev_shutdown(dev); dev_tcx_uninstall(dev); + netdev_lock_ops(dev); dev_xdp_uninstall(dev); + netdev_unlock_ops(dev); bpf_dev_bound_netdev_unregister(dev); dev_memory_provider_uninstall(dev); diff --git a/net/core/dev_api.c b/net/core/dev_api.c index 7bd667b34b80..98db990ce21c 100644 --- a/net/core/dev_api.c +++ b/net/core/dev_api.c @@ -317,3 +317,15 @@ int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa, return ret; } EXPORT_SYMBOL(dev_set_mac_address); + +int dev_xdp_propagate(struct net_device *dev, struct netdev_bpf *bpf) +{ + int ret; + + netdev_lock_ops(dev); + ret = netif_xdp_propagate(dev, bpf); + netdev_unlock_ops(dev); + + return ret; +} +EXPORT_SYMBOL_GPL(dev_xdp_propagate); diff --git a/net/xdp/xsk.c b/net/xdp/xsk.c index 84bf9f1d4bf2..f864e5d70b40 100644 --- a/net/xdp/xsk.c +++ b/net/xdp/xsk.c @@ -1181,6 +1181,8 @@ static int xsk_bind(struct socket *sock, struct sockaddr *addr, int addr_len) goto out_release; } + netdev_lock_ops(dev); + if (!xs->rx && !xs->tx) { err = -EINVAL; goto out_unlock; @@ -1315,6 +1317,7 @@ out_unlock: smp_wmb(); WRITE_ONCE(xs->state, XSK_BOUND); } + netdev_unlock_ops(dev); out_release: mutex_unlock(&xs->mutex); rtnl_unlock(); diff --git a/net/xdp/xsk_buff_pool.c b/net/xdp/xsk_buff_pool.c index c263fb7a68dc..0e6ca568fdee 100644 --- a/net/xdp/xsk_buff_pool.c +++ b/net/xdp/xsk_buff_pool.c @@ -1,5 +1,6 @@ // SPDX-License-Identifier: GPL-2.0 +#include #include #include #include @@ -219,6 +220,7 @@ int xp_assign_dev(struct xsk_buff_pool *pool, bpf.xsk.pool = pool; bpf.xsk.queue_id = queue_id; + netdev_ops_assert_locked(netdev); err = netdev->netdev_ops->ndo_bpf(netdev, &bpf); if (err) goto err_unreg_pool; -- cgit v1.2.3 From df43d8bf10316a7c3b1e47e3cc0057a54df4a5b8 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:29 -0800 Subject: net: replace dev_addr_sem with netdev instance lock Lockdep reports possible circular dependency in [0]. Instead of fixing the ordering, replace global dev_addr_sem with netdev instance lock. Most of the paths that set/get mac are RTNL protected. Two places where it's not, convert to explicit locking: - sysfs address_show - dev_get_mac_address via dev_ioctl 0: https://netdev-3.bots.linux.dev/vmksft-forwarding-dbg/results/993321/24-router-bridge-1d-lag-sh/stderr Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-12-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- drivers/net/tap.c | 2 +- drivers/net/tun.c | 2 +- include/linux/netdevice.h | 6 +----- net/core/dev.c | 52 ++++++++++++++++++++++++----------------------- net/core/dev.h | 3 +-- net/core/dev_api.c | 17 ++-------------- net/core/dev_ioctl.c | 2 +- net/core/net-sysfs.c | 7 ++----- net/core/rtnetlink.c | 6 +++++- 9 files changed, 41 insertions(+), 56 deletions(-) (limited to 'include') diff --git a/drivers/net/tap.c b/drivers/net/tap.c index d4ece538f1b2..4382f5e323b0 100644 --- a/drivers/net/tap.c +++ b/drivers/net/tap.c @@ -1017,7 +1017,7 @@ static long tap_ioctl(struct file *file, unsigned int cmd, rtnl_unlock(); return -ENOLINK; } - ret = dev_set_mac_address_user(tap->dev, &sa, NULL); + ret = dev_set_mac_address(tap->dev, &sa, NULL); tap_put_tap_dev(tap); rtnl_unlock(); return ret; diff --git a/drivers/net/tun.c b/drivers/net/tun.c index d8f4d3e996a7..1e645d5e225c 100644 --- a/drivers/net/tun.c +++ b/drivers/net/tun.c @@ -3175,7 +3175,7 @@ static long __tun_chr_ioctl(struct file *file, unsigned int cmd, case SIOCSIFHWADDR: /* Set hw address */ - ret = dev_set_mac_address_user(tun->dev, &ifr.ifr_hwaddr, NULL); + ret = dev_set_mac_address(tun->dev, &ifr.ifr_hwaddr, NULL); break; case TUNGETSNDBUF: diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index ca9c09dab14e..f3e6e6f27e22 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2492,7 +2492,7 @@ struct net_device { * * Protects: * @gro_flush_timeout, @napi_defer_hard_irqs, @napi_list, - * @net_shaper_hierarchy, @reg_state, @threaded + * @net_shaper_hierarchy, @reg_state, @threaded, @dev_addr * * Partially protects (writers must hold both @lock and rtnl_lock): * @up @@ -4262,10 +4262,6 @@ int netif_set_mac_address(struct net_device *dev, struct sockaddr *sa, struct netlink_ext_ack *extack); int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa, struct netlink_ext_ack *extack); -int netif_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, - struct netlink_ext_ack *extack); -int dev_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, - struct netlink_ext_ack *extack); int dev_get_mac_address(struct sockaddr *sa, struct net *net, char *dev_name); int dev_get_port_parent_id(struct net_device *dev, struct netdev_phys_item_id *ppid, bool recurse); diff --git a/net/core/dev.c b/net/core/dev.c index 404047d4d943..a0f75a1d1f5a 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1058,6 +1058,28 @@ struct net_device *netdev_get_by_index_lock(struct net *net, int ifindex) return __netdev_put_lock(dev); } +/** + * netdev_get_by_name_lock() - find a device by its name + * @net: the applicable net namespace + * @name: name of device + * + * Search for an interface by name. If a valid device + * with @name is found it will be returned with netdev->lock held. + * netdev_unlock() must be called to release it. + * + * Return: pointer to a device with lock held, NULL if not found. + */ +struct net_device *netdev_get_by_name_lock(struct net *net, const char *name) +{ + struct net_device *dev; + + dev = dev_get_by_name(net, name); + if (!dev) + return NULL; + + return __netdev_put_lock(dev); +} + struct net_device * netdev_xa_find_lock(struct net *net, struct net_device *dev, unsigned long *index) @@ -9589,44 +9611,24 @@ int netif_set_mac_address(struct net_device *dev, struct sockaddr *sa, return 0; } -DECLARE_RWSEM(dev_addr_sem); - -int netif_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, - struct netlink_ext_ack *extack) -{ - int ret; - - down_write(&dev_addr_sem); - ret = netif_set_mac_address(dev, sa, extack); - up_write(&dev_addr_sem); - return ret; -} - int dev_get_mac_address(struct sockaddr *sa, struct net *net, char *dev_name) { size_t size = sizeof(sa->sa_data_min); struct net_device *dev; - int ret = 0; - down_read(&dev_addr_sem); - rcu_read_lock(); + dev = netdev_get_by_name_lock(net, dev_name); + if (!dev) + return -ENODEV; - dev = dev_get_by_name_rcu(net, dev_name); - if (!dev) { - ret = -ENODEV; - goto unlock; - } if (!dev->addr_len) memset(sa->sa_data, 0, size); else memcpy(sa->sa_data, dev->dev_addr, min_t(size_t, size, dev->addr_len)); sa->sa_family = dev->type; + netdev_unlock(dev); -unlock: - rcu_read_unlock(); - up_read(&dev_addr_sem); - return ret; + return 0; } EXPORT_SYMBOL(dev_get_mac_address); diff --git a/net/core/dev.h b/net/core/dev.h index 41b0831aba60..b50ca645c086 100644 --- a/net/core/dev.h +++ b/net/core/dev.h @@ -28,6 +28,7 @@ netdev_napi_by_id_lock(struct net *net, unsigned int napi_id); struct net_device *dev_get_by_napi_id(unsigned int napi_id); struct net_device *netdev_get_by_index_lock(struct net *net, int ifindex); +struct net_device *netdev_get_by_name_lock(struct net *net, const char *name); struct net_device *__netdev_put_lock(struct net_device *dev); struct net_device * netdev_xa_find_lock(struct net *net, struct net_device *dev, @@ -69,8 +70,6 @@ extern int weight_p; extern int dev_weight_rx_bias; extern int dev_weight_tx_bias; -extern struct rw_semaphore dev_addr_sem; - /* rtnl helpers */ extern struct list_head net_todo_list; void netdev_run_todo(void); diff --git a/net/core/dev_api.c b/net/core/dev_api.c index 98db990ce21c..655a95fb7baa 100644 --- a/net/core/dev_api.c +++ b/net/core/dev_api.c @@ -82,19 +82,6 @@ void dev_set_group(struct net_device *dev, int new_group) netdev_unlock_ops(dev); } -int dev_set_mac_address_user(struct net_device *dev, struct sockaddr *sa, - struct netlink_ext_ack *extack) -{ - int ret; - - netdev_lock_ops(dev); - ret = netif_set_mac_address_user(dev, sa, extack); - netdev_unlock_ops(dev); - - return ret; -} -EXPORT_SYMBOL(dev_set_mac_address_user); - /** * dev_change_net_namespace() - move device to different nethost namespace * @dev: device @@ -310,9 +297,9 @@ int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa, { int ret; - netdev_lock_ops(dev); + netdev_lock(dev); ret = netif_set_mac_address(dev, sa, extack); - netdev_unlock_ops(dev); + netdev_unlock(dev); return ret; } diff --git a/net/core/dev_ioctl.c b/net/core/dev_ioctl.c index d9f350593121..296e52d1395d 100644 --- a/net/core/dev_ioctl.c +++ b/net/core/dev_ioctl.c @@ -574,7 +574,7 @@ static int dev_ifsioc(struct net *net, struct ifreq *ifr, void __user *data, case SIOCSIFHWADDR: if (dev->addr_len > sizeof(struct sockaddr)) return -EINVAL; - return dev_set_mac_address_user(dev, &ifr->ifr_hwaddr, NULL); + return dev_set_mac_address(dev, &ifr->ifr_hwaddr, NULL); case SIOCSIFHWBROADCAST: if (ifr->ifr_hwaddr.sa_family != dev->type) diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index 27eea34d1b41..02d1d40b47ae 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -262,14 +262,11 @@ static ssize_t address_show(struct device *dev, struct device_attribute *attr, struct net_device *ndev = to_net_dev(dev); ssize_t ret = -EINVAL; - down_read(&dev_addr_sem); - - rcu_read_lock(); + netdev_lock(ndev); if (dev_isalive(ndev)) ret = sysfs_format_mac(buf, ndev->dev_addr, ndev->addr_len); - rcu_read_unlock(); + netdev_unlock(ndev); - up_read(&dev_addr_sem); return ret; } static DEVICE_ATTR_RO(address); diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 9d539c9ce1a4..88a352b02bce 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -3089,7 +3089,11 @@ static int do_setlink(const struct sk_buff *skb, struct net_device *dev, sa->sa_family = dev->type; memcpy(sa->sa_data, nla_data(tb[IFLA_ADDRESS]), dev->addr_len); - err = netif_set_mac_address_user(dev, sa, extack); + if (!netdev_need_ops_lock(dev)) + netdev_lock(dev); + err = netif_set_mac_address(dev, sa, extack); + if (!netdev_need_ops_lock(dev)) + netdev_unlock(dev); kfree(sa); if (err) goto errout; -- cgit v1.2.3 From 605ef7aec0605bd5506b31591de278d652bd0096 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:30 -0800 Subject: net: add option to request netdev instance lock Currently only the drivers that implement shaper or queue APIs are grabbing instance lock. Add an explicit opt-in for the drivers that want to grab the lock without implementing the above APIs. There is a 3-byte hole after @up, use it: /* --- cacheline 47 boundary (3008 bytes) --- */ u32 napi_defer_hard_irqs; /* 3008 4 */ bool up; /* 3012 1 */ /* XXX 3 bytes hole, try to pack */ struct mutex lock; /* 3016 144 */ /* XXX last struct has 1 hole */ Cc: Saeed Mahameed Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-13-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- include/linux/netdevice.h | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index f3e6e6f27e22..adf201617b72 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2485,6 +2485,12 @@ struct net_device { */ bool up; + /** + * @request_ops_lock: request the core to run all @netdev_ops and + * @ethtool_ops under the @lock. + */ + bool request_ops_lock; + /** * @lock: netdev-scope lock, protects a small selection of fields. * Should always be taken using netdev_lock() / netdev_unlock() helpers. @@ -2774,7 +2780,7 @@ static inline void netdev_assert_locked_or_invisible(struct net_device *dev) static inline bool netdev_need_ops_lock(struct net_device *dev) { - bool ret = !!dev->queue_mgmt_ops; + bool ret = dev->request_ops_lock || !!dev->queue_mgmt_ops; #if IS_ENABLED(CONFIG_NET_SHAPER) ret |= !!dev->netdev_ops->net_shaper_ops; -- cgit v1.2.3 From cc34acd577f1a6ed805106bfcc9a262837dbd0da Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:31 -0800 Subject: docs: net: document new locking reality Also clarify ndo_get_stats (that read and write paths can run concurrently) and mention only RCU. Cc: Saeed Mahameed Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-14-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- Documentation/networking/netdevices.rst | 65 ++++++++++++++++++++++++++------- include/linux/netdevice.h | 4 ++ 2 files changed, 56 insertions(+), 13 deletions(-) (limited to 'include') diff --git a/Documentation/networking/netdevices.rst b/Documentation/networking/netdevices.rst index 1d37038e9fbe..d235a7364893 100644 --- a/Documentation/networking/netdevices.rst +++ b/Documentation/networking/netdevices.rst @@ -210,49 +210,55 @@ packets is preferred. struct net_device synchronization rules ======================================= ndo_open: - Synchronization: rtnl_lock() semaphore. + Synchronization: rtnl_lock() semaphore. In addition, netdev instance + lock if the driver implements queue management or shaper API. Context: process ndo_stop: - Synchronization: rtnl_lock() semaphore. + Synchronization: rtnl_lock() semaphore. In addition, netdev instance + lock if the driver implements queue management or shaper API. Context: process Note: netif_running() is guaranteed false ndo_do_ioctl: Synchronization: rtnl_lock() semaphore. - Context: process - This is only called by network subsystems internally, - not by user space calling ioctl as it was in before - linux-5.14. + This is only called by network subsystems internally, + not by user space calling ioctl as it was in before + linux-5.14. ndo_siocbond: - Synchronization: rtnl_lock() semaphore. + Synchronization: rtnl_lock() semaphore. In addition, netdev instance + lock if the driver implements queue management or shaper API. Context: process - Used by the bonding driver for the SIOCBOND family of - ioctl commands. + Used by the bonding driver for the SIOCBOND family of + ioctl commands. ndo_siocwandev: - Synchronization: rtnl_lock() semaphore. + Synchronization: rtnl_lock() semaphore. In addition, netdev instance + lock if the driver implements queue management or shaper API. Context: process Used by the drivers/net/wan framework to handle the SIOCWANDEV ioctl with the if_settings structure. ndo_siocdevprivate: - Synchronization: rtnl_lock() semaphore. + Synchronization: rtnl_lock() semaphore. In addition, netdev instance + lock if the driver implements queue management or shaper API. Context: process This is used to implement SIOCDEVPRIVATE ioctl helpers. These should not be added to new drivers, so don't use. ndo_eth_ioctl: - Synchronization: rtnl_lock() semaphore. + Synchronization: rtnl_lock() semaphore. In addition, netdev instance + lock if the driver implements queue management or shaper API. Context: process ndo_get_stats: - Synchronization: rtnl_lock() semaphore, or RCU. + Synchronization: RCU (can be called concurrently with the stats + update path). Context: atomic (can't sleep under RCU) ndo_start_xmit: @@ -284,6 +290,10 @@ ndo_set_rx_mode: Synchronization: netif_addr_lock spinlock. Context: BHs disabled +Most ndo callbacks not specified in the list above are running +under ``rtnl_lock``. In addition, netdev instance lock is taken as well if +the driver implements queue management or shaper API. + struct napi_struct synchronization rules ======================================== napi->poll: @@ -298,6 +308,35 @@ napi->poll: softirq will be called with interrupts disabled by netconsole. +struct netdev_queue_mgmt_ops synchronization rules +================================================== + +All queue management ndo callbacks are holding netdev instance lock. + +RTNL and netdev instance lock +============================= + +Historically, all networking control operations were protected by a single +global lock known as ``rtnl_lock``. There is an ongoing effort to replace this +global lock with separate locks for each network namespace. Additionally, +properties of individual netdev are increasingly protected by per-netdev locks. + +For device drivers that implement shaping or queue management APIs, all control +operations will be performed under the netdev instance lock. Currently, this +instance lock is acquired within the context of ``rtnl_lock``. The drivers +can also explicitly request instance lock to be acquired via +``request_ops_lock``. In the future, there will be an option for individual +drivers to opt out of using ``rtnl_lock`` and instead perform their control +operations directly under the netdev instance lock. + +Devices drivers are encouraged to rely on the instance lock where possible. + +For the (mostly software) drivers that need to interact with the core stack, +there are two sets of interfaces: ``dev_xxx`` and ``netif_xxx`` (e.g., +``dev_set_mtu`` and ``netif_set_mtu``). The ``dev_xxx`` functions handle +acquiring the instance lock themselves, while the ``netif_xxx`` functions +assume that the driver has already acquired the instance lock. + NETDEV_INTERNAL symbol namespace ================================ diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index adf201617b72..2f8560a354ba 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2505,6 +2505,10 @@ struct net_device { * * Also protects some fields in struct napi_struct. * + * For the drivers that implement shaper or queue API, the scope + * of this lock is expanded to cover most ndo/queue/ethtool/sysfs + * operations. + * * Ordering: take after rtnl_lock. */ struct mutex lock; -- cgit v1.2.3 From 004b5008016a2cc37103bf8d9968573771cd311f Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Wed, 5 Mar 2025 08:37:32 -0800 Subject: eth: bnxt: remove most dependencies on RTNL Only devlink and sriov paths are grabbing rtnl explicitly. The rest is covered by netdev instance lock which the core now grabs, so there is no need to manage rtnl in most places anymore. On the core side we can now try to drop rtnl in some places (do_setlink for example) for the drivers that signal non-rtnl mode (TBD). Boot-tested and with `ethtool -L eth1 combined 24` to trigger reset. Cc: Saeed Mahameed Reviewed-by: Michael Chan Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20250305163732.2766420-15-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/broadcom/bnxt/bnxt.c | 117 +++++++++++----------- drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c | 9 ++ drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c | 6 ++ drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c | 16 +-- drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c | 18 ++-- drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c | 3 +- include/linux/netdevice.h | 5 + 7 files changed, 101 insertions(+), 73 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index ead9fcaad6bd..1a1e6da77777 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -5246,8 +5246,10 @@ static void bnxt_free_ntp_fltrs(struct bnxt *bp, bool all) { int i; - /* Under rtnl_lock and all our NAPIs have been disabled. It's - * safe to delete the hash table. + netdev_assert_locked(bp->dev); + + /* Under netdev instance lock and all our NAPIs have been disabled. + * It's safe to delete the hash table. */ for (i = 0; i < BNXT_NTP_FLTR_HASH_SIZE; i++) { struct hlist_head *head; @@ -12789,7 +12791,6 @@ open_err_free_mem: return rc; } -/* rtnl_lock held */ int bnxt_open_nic(struct bnxt *bp, bool irq_re_init, bool link_re_init) { int rc = 0; @@ -12805,9 +12806,9 @@ int bnxt_open_nic(struct bnxt *bp, bool irq_re_init, bool link_re_init) return rc; } -/* rtnl_lock held, open the NIC half way by allocating all resources, but - * NAPI, IRQ, and TX are not enabled. This is mainly used for offline - * self tests. +/* netdev instance lock held, open the NIC half way by allocating all + * resources, but NAPI, IRQ, and TX are not enabled. This is mainly used + * for offline self tests. */ int bnxt_half_open_nic(struct bnxt *bp) { @@ -12842,8 +12843,8 @@ half_open_err: return rc; } -/* rtnl_lock held, this call can only be made after a previous successful - * call to bnxt_half_open_nic(). +/* netdev instance lock held, this call can only be made after a previous + * successful call to bnxt_half_open_nic(). */ void bnxt_half_close_nic(struct bnxt *bp) { @@ -12952,10 +12953,11 @@ void bnxt_close_nic(struct bnxt *bp, bool irq_re_init, bool link_re_init) if (test_bit(BNXT_STATE_IN_FW_RESET, &bp->state)) { /* If we get here, it means firmware reset is in progress * while we are trying to close. We can safely proceed with - * the close because we are holding rtnl_lock(). Some firmware - * messages may fail as we proceed to close. We set the - * ABORT_ERR flag here so that the FW reset thread will later - * abort when it gets the rtnl_lock() and sees the flag. + * the close because we are holding netdev instance lock. + * Some firmware messages may fail as we proceed to close. + * We set the ABORT_ERR flag here so that the FW reset thread + * will later abort when it gets the netdev instance lock + * and sees the flag. */ netdev_warn(bp->dev, "FW reset in progress during close, FW reset will be aborted\n"); set_bit(BNXT_STATE_ABORT_ERR, &bp->state); @@ -13046,7 +13048,7 @@ static int bnxt_hwrm_port_phy_write(struct bnxt *bp, u16 phy_addr, u16 reg, return hwrm_req_send(bp, req); } -/* rtnl_lock held */ +/* netdev instance lock held */ static int bnxt_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) { struct mii_ioctl_data *mdio = if_mii(ifr); @@ -13965,30 +13967,31 @@ bnxt_restart_timer: mod_timer(&bp->timer, jiffies + bp->current_interval); } -static void bnxt_rtnl_lock_sp(struct bnxt *bp) +static void bnxt_lock_sp(struct bnxt *bp) { /* We are called from bnxt_sp_task which has BNXT_STATE_IN_SP_TASK * set. If the device is being closed, bnxt_close() may be holding - * rtnl() and waiting for BNXT_STATE_IN_SP_TASK to clear. So we - * must clear BNXT_STATE_IN_SP_TASK before holding rtnl(). + * netdev instance lock and waiting for BNXT_STATE_IN_SP_TASK to clear. + * So we must clear BNXT_STATE_IN_SP_TASK before holding netdev + * instance lock. */ clear_bit(BNXT_STATE_IN_SP_TASK, &bp->state); - rtnl_lock(); + netdev_lock(bp->dev); } -static void bnxt_rtnl_unlock_sp(struct bnxt *bp) +static void bnxt_unlock_sp(struct bnxt *bp) { set_bit(BNXT_STATE_IN_SP_TASK, &bp->state); - rtnl_unlock(); + netdev_unlock(bp->dev); } /* Only called from bnxt_sp_task() */ static void bnxt_reset(struct bnxt *bp, bool silent) { - bnxt_rtnl_lock_sp(bp); + bnxt_lock_sp(bp); if (test_bit(BNXT_STATE_OPEN, &bp->state)) bnxt_reset_task(bp, silent); - bnxt_rtnl_unlock_sp(bp); + bnxt_unlock_sp(bp); } /* Only called from bnxt_sp_task() */ @@ -13996,9 +13999,9 @@ static void bnxt_rx_ring_reset(struct bnxt *bp) { int i; - bnxt_rtnl_lock_sp(bp); + bnxt_lock_sp(bp); if (!test_bit(BNXT_STATE_OPEN, &bp->state)) { - bnxt_rtnl_unlock_sp(bp); + bnxt_unlock_sp(bp); return; } /* Disable and flush TPA before resetting the RX ring */ @@ -14037,7 +14040,7 @@ static void bnxt_rx_ring_reset(struct bnxt *bp) } if (bp->flags & BNXT_FLAG_TPA) bnxt_set_tpa(bp, true); - bnxt_rtnl_unlock_sp(bp); + bnxt_unlock_sp(bp); } static void bnxt_fw_fatal_close(struct bnxt *bp) @@ -14093,7 +14096,7 @@ static bool is_bnxt_fw_ok(struct bnxt *bp) return false; } -/* rtnl_lock is acquired before calling this function */ +/* netdev instance lock is acquired before calling this function */ static void bnxt_force_fw_reset(struct bnxt *bp) { struct bnxt_fw_health *fw_health = bp->fw_health; @@ -14136,9 +14139,9 @@ void bnxt_fw_exception(struct bnxt *bp) netdev_warn(bp->dev, "Detected firmware fatal condition, initiating reset\n"); set_bit(BNXT_STATE_FW_FATAL_COND, &bp->state); bnxt_ulp_stop(bp); - bnxt_rtnl_lock_sp(bp); + bnxt_lock_sp(bp); bnxt_force_fw_reset(bp); - bnxt_rtnl_unlock_sp(bp); + bnxt_unlock_sp(bp); } /* Returns the number of registered VFs, or 1 if VF configuration is pending, or @@ -14168,7 +14171,7 @@ static int bnxt_get_registered_vfs(struct bnxt *bp) void bnxt_fw_reset(struct bnxt *bp) { bnxt_ulp_stop(bp); - bnxt_rtnl_lock_sp(bp); + bnxt_lock_sp(bp); if (test_bit(BNXT_STATE_OPEN, &bp->state) && !test_bit(BNXT_STATE_IN_FW_RESET, &bp->state)) { struct bnxt_ptp_cfg *ptp = bp->ptp_cfg; @@ -14214,7 +14217,7 @@ void bnxt_fw_reset(struct bnxt *bp) bnxt_queue_fw_reset_work(bp, tmo); } fw_reset_exit: - bnxt_rtnl_unlock_sp(bp); + bnxt_unlock_sp(bp); } static void bnxt_chk_missed_irq(struct bnxt *bp) @@ -14413,7 +14416,7 @@ static void bnxt_sp_task(struct work_struct *work) static void _bnxt_get_max_rings(struct bnxt *bp, int *max_rx, int *max_tx, int *max_cp); -/* Under rtnl_lock */ +/* Under netdev instance lock */ int bnxt_check_rings(struct bnxt *bp, int tx, int rx, bool sh, int tcs, int tx_xdp) { @@ -14841,10 +14844,10 @@ static void bnxt_fw_reset_task(struct work_struct *work) return; } bp->fw_reset_timestamp = jiffies; - rtnl_lock(); + netdev_lock(bp->dev); if (test_bit(BNXT_STATE_ABORT_ERR, &bp->state)) { bnxt_fw_reset_abort(bp, rc); - rtnl_unlock(); + netdev_unlock(bp->dev); goto ulp_start; } bnxt_fw_reset_close(bp); @@ -14855,7 +14858,7 @@ static void bnxt_fw_reset_task(struct work_struct *work) bp->fw_reset_state = BNXT_FW_RESET_STATE_ENABLE_DEV; tmo = bp->fw_reset_min_dsecs * HZ / 10; } - rtnl_unlock(); + netdev_unlock(bp->dev); bnxt_queue_fw_reset_work(bp, tmo); return; } @@ -14929,7 +14932,7 @@ static void bnxt_fw_reset_task(struct work_struct *work) bp->fw_reset_state = BNXT_FW_RESET_STATE_OPENING; fallthrough; case BNXT_FW_RESET_STATE_OPENING: - while (!rtnl_trylock()) { + while (!netdev_trylock(bp->dev)) { bnxt_queue_fw_reset_work(bp, HZ / 10); return; } @@ -14937,7 +14940,7 @@ static void bnxt_fw_reset_task(struct work_struct *work) if (rc) { netdev_err(bp->dev, "bnxt_open() failed during FW reset\n"); bnxt_fw_reset_abort(bp, rc); - rtnl_unlock(); + netdev_unlock(bp->dev); goto ulp_start; } @@ -14956,13 +14959,13 @@ static void bnxt_fw_reset_task(struct work_struct *work) bnxt_dl_health_fw_recovery_done(bp); bnxt_dl_health_fw_status_update(bp, true); } - rtnl_unlock(); + netdev_unlock(bp->dev); bnxt_ulp_start(bp, 0); bnxt_reenable_sriov(bp); - rtnl_lock(); + netdev_lock(bp->dev); bnxt_vf_reps_alloc(bp); bnxt_vf_reps_open(bp); - rtnl_unlock(); + netdev_unlock(bp->dev); break; } return; @@ -14975,9 +14978,9 @@ fw_reset_abort_status: netdev_err(bp->dev, "fw_health_status 0x%x\n", sts); } fw_reset_abort: - rtnl_lock(); + netdev_lock(bp->dev); bnxt_fw_reset_abort(bp, rc); - rtnl_unlock(); + netdev_unlock(bp->dev); ulp_start: bnxt_ulp_start(bp, rc); } @@ -15069,13 +15072,14 @@ init_err: return rc; } -/* rtnl_lock held */ static int bnxt_change_mac_addr(struct net_device *dev, void *p) { struct sockaddr *addr = p; struct bnxt *bp = netdev_priv(dev); int rc = 0; + netdev_assert_locked(dev); + if (!is_valid_ether_addr(addr->sa_data)) return -EADDRNOTAVAIL; @@ -15096,11 +15100,12 @@ static int bnxt_change_mac_addr(struct net_device *dev, void *p) return rc; } -/* rtnl_lock held */ static int bnxt_change_mtu(struct net_device *dev, int new_mtu) { struct bnxt *bp = netdev_priv(dev); + netdev_assert_locked(dev); + if (netif_running(dev)) bnxt_close_nic(bp, true, false); @@ -16257,7 +16262,7 @@ int bnxt_restore_pf_fw_resources(struct bnxt *bp) { int rc; - ASSERT_RTNL(); + netdev_ops_assert_locked(bp->dev); bnxt_hwrm_func_qcaps(bp); if (netif_running(bp->dev)) @@ -16657,7 +16662,7 @@ static void bnxt_shutdown(struct pci_dev *pdev) if (!dev) return; - rtnl_lock(); + netdev_lock(dev); bp = netdev_priv(dev); if (!bp) goto shutdown_exit; @@ -16675,7 +16680,7 @@ static void bnxt_shutdown(struct pci_dev *pdev) } shutdown_exit: - rtnl_unlock(); + netdev_unlock(dev); } #ifdef CONFIG_PM_SLEEP @@ -16687,7 +16692,7 @@ static int bnxt_suspend(struct device *device) bnxt_ulp_stop(bp); - rtnl_lock(); + netdev_lock(dev); if (netif_running(dev)) { netif_device_detach(dev); rc = bnxt_close(dev); @@ -16696,7 +16701,7 @@ static int bnxt_suspend(struct device *device) bnxt_ptp_clear(bp); pci_disable_device(bp->pdev); bnxt_free_ctx_mem(bp, false); - rtnl_unlock(); + netdev_unlock(dev); return rc; } @@ -16706,7 +16711,7 @@ static int bnxt_resume(struct device *device) struct bnxt *bp = netdev_priv(dev); int rc = 0; - rtnl_lock(); + netdev_lock(dev); rc = pci_enable_device(bp->pdev); if (rc) { netdev_err(dev, "Cannot re-enable PCI device during resume, err = %d\n", @@ -16749,7 +16754,7 @@ static int bnxt_resume(struct device *device) } resume_exit: - rtnl_unlock(); + netdev_unlock(bp->dev); bnxt_ulp_start(bp, rc); if (!rc) bnxt_reenable_sriov(bp); @@ -16784,7 +16789,7 @@ static pci_ers_result_t bnxt_io_error_detected(struct pci_dev *pdev, bnxt_ulp_stop(bp); - rtnl_lock(); + netdev_lock(netdev); netif_device_detach(netdev); if (test_and_set_bit(BNXT_STATE_IN_FW_RESET, &bp->state)) { @@ -16793,7 +16798,7 @@ static pci_ers_result_t bnxt_io_error_detected(struct pci_dev *pdev, } if (abort || state == pci_channel_io_perm_failure) { - rtnl_unlock(); + netdev_unlock(netdev); return PCI_ERS_RESULT_DISCONNECT; } @@ -16812,7 +16817,7 @@ static pci_ers_result_t bnxt_io_error_detected(struct pci_dev *pdev, if (pci_is_enabled(pdev)) pci_disable_device(pdev); bnxt_free_ctx_mem(bp, false); - rtnl_unlock(); + netdev_unlock(netdev); /* Request a slot slot reset. */ return PCI_ERS_RESULT_NEED_RESET; @@ -16842,7 +16847,7 @@ static pci_ers_result_t bnxt_io_slot_reset(struct pci_dev *pdev) test_bit(BNXT_STATE_PCI_CHANNEL_IO_FROZEN, &bp->state)) msleep(900); - rtnl_lock(); + netdev_lock(netdev); if (pci_enable_device(pdev)) { dev_err(&pdev->dev, @@ -16897,7 +16902,7 @@ static pci_ers_result_t bnxt_io_slot_reset(struct pci_dev *pdev) reset_exit: clear_bit(BNXT_STATE_IN_FW_RESET, &bp->state); bnxt_clear_reservations(bp, true); - rtnl_unlock(); + netdev_unlock(netdev); return result; } @@ -16916,7 +16921,7 @@ static void bnxt_io_resume(struct pci_dev *pdev) int err; netdev_info(bp->dev, "PCI Slot Resume\n"); - rtnl_lock(); + netdev_lock(netdev); err = bnxt_hwrm_func_qcaps(bp); if (!err) { @@ -16929,7 +16934,7 @@ static void bnxt_io_resume(struct pci_dev *pdev) if (!err) netif_device_attach(netdev); - rtnl_unlock(); + netdev_unlock(netdev); bnxt_ulp_start(bp, err); if (!err) bnxt_reenable_sriov(bp); diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c index ef8288fd68f4..b06fcddfc81c 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c @@ -439,14 +439,17 @@ static int bnxt_dl_reload_down(struct devlink *dl, bool netns_change, case DEVLINK_RELOAD_ACTION_DRIVER_REINIT: { bnxt_ulp_stop(bp); rtnl_lock(); + netdev_lock(bp->dev); if (bnxt_sriov_cfg(bp)) { NL_SET_ERR_MSG_MOD(extack, "reload is unsupported while VFs are allocated or being configured"); + netdev_unlock(bp->dev); rtnl_unlock(); bnxt_ulp_start(bp, 0); return -EOPNOTSUPP; } if (bp->dev->reg_state == NETREG_UNREGISTERED) { + netdev_unlock(bp->dev); rtnl_unlock(); bnxt_ulp_start(bp, 0); return -ENODEV; @@ -459,6 +462,7 @@ static int bnxt_dl_reload_down(struct devlink *dl, bool netns_change, NL_SET_ERR_MSG_MOD(extack, "Failed to deregister"); if (netif_running(bp->dev)) dev_close(bp->dev); + netdev_unlock(bp->dev); rtnl_unlock(); break; } @@ -479,7 +483,9 @@ static int bnxt_dl_reload_down(struct devlink *dl, bool netns_change, return -EPERM; } rtnl_lock(); + netdev_lock(bp->dev); if (bp->dev->reg_state == NETREG_UNREGISTERED) { + netdev_unlock(bp->dev); rtnl_unlock(); return -ENODEV; } @@ -493,6 +499,7 @@ static int bnxt_dl_reload_down(struct devlink *dl, bool netns_change, if (rc) { NL_SET_ERR_MSG_MOD(extack, "Failed to activate firmware"); clear_bit(BNXT_STATE_FW_ACTIVATE, &bp->state); + netdev_unlock(bp->dev); rtnl_unlock(); } break; @@ -568,7 +575,9 @@ static int bnxt_dl_reload_up(struct devlink *dl, enum devlink_reload_action acti } *actions_performed |= BIT(action); } else if (netif_running(bp->dev)) { + netdev_lock(bp->dev); dev_close(bp->dev); + netdev_unlock(bp->dev); } rtnl_unlock(); if (action == DEVLINK_RELOAD_ACTION_DRIVER_REINIT) diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c index 12b6ed51fd88..5ddddd89052f 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_sriov.c @@ -946,7 +946,9 @@ void bnxt_sriov_disable(struct bnxt *bp) /* Reclaim all resources for the PF. */ rtnl_lock(); + netdev_lock(bp->dev); bnxt_restore_pf_fw_resources(bp); + netdev_unlock(bp->dev); rtnl_unlock(); } @@ -956,17 +958,21 @@ int bnxt_sriov_configure(struct pci_dev *pdev, int num_vfs) struct bnxt *bp = netdev_priv(dev); rtnl_lock(); + netdev_lock(dev); if (!netif_running(dev)) { netdev_warn(dev, "Reject SRIOV config request since if is down!\n"); + netdev_unlock(dev); rtnl_unlock(); return 0; } if (test_bit(BNXT_STATE_IN_FW_RESET, &bp->state)) { netdev_warn(dev, "Reject SRIOV config request when FW reset is in progress\n"); + netdev_unlock(dev); rtnl_unlock(); return 0; } bp->sriov_cfg = true; + netdev_unlock(dev); rtnl_unlock(); if (pci_vfs_assigned(bp->pdev)) { diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c index e4a7f37036ed..a8e930d5dbb0 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c @@ -112,7 +112,7 @@ int bnxt_register_dev(struct bnxt_en_dev *edev, struct bnxt_ulp *ulp; int rc = 0; - rtnl_lock(); + netdev_lock(dev); mutex_lock(&edev->en_dev_lock); if (!bp->irq_tbl) { rc = -ENODEV; @@ -138,7 +138,7 @@ int bnxt_register_dev(struct bnxt_en_dev *edev, edev->flags |= BNXT_EN_FLAG_MSIX_REQUESTED; exit: mutex_unlock(&edev->en_dev_lock); - rtnl_unlock(); + netdev_unlock(dev); return rc; } EXPORT_SYMBOL(bnxt_register_dev); @@ -151,7 +151,7 @@ void bnxt_unregister_dev(struct bnxt_en_dev *edev) int i = 0; ulp = edev->ulp_tbl; - rtnl_lock(); + netdev_lock(dev); mutex_lock(&edev->en_dev_lock); if (ulp->msix_requested) edev->flags &= ~BNXT_EN_FLAG_MSIX_REQUESTED; @@ -169,7 +169,7 @@ void bnxt_unregister_dev(struct bnxt_en_dev *edev) i++; } mutex_unlock(&edev->en_dev_lock); - rtnl_unlock(); + netdev_unlock(dev); return; } EXPORT_SYMBOL(bnxt_unregister_dev); @@ -309,12 +309,14 @@ void bnxt_ulp_irq_stop(struct bnxt *bp) if (!ulp->msix_requested) return; - ops = rtnl_dereference(ulp->ulp_ops); + netdev_lock(bp->dev); + ops = rcu_dereference(ulp->ulp_ops); if (!ops || !ops->ulp_irq_stop) return; if (test_bit(BNXT_STATE_FW_RESET_DET, &bp->state)) reset = true; ops->ulp_irq_stop(ulp->handle, reset); + netdev_unlock(bp->dev); } } @@ -333,7 +335,8 @@ void bnxt_ulp_irq_restart(struct bnxt *bp, int err) if (!ulp->msix_requested) return; - ops = rtnl_dereference(ulp->ulp_ops); + netdev_lock(bp->dev); + ops = rcu_dereference(ulp->ulp_ops); if (!ops || !ops->ulp_irq_restart) return; @@ -345,6 +348,7 @@ void bnxt_ulp_irq_restart(struct bnxt *bp, int err) bnxt_fill_msix_vecs(bp, ent); } ops->ulp_irq_restart(ulp->handle, ent); + netdev_unlock(bp->dev); kfree(ent); } } diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c index 1467b94a6427..619f0844e778 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c @@ -257,8 +257,7 @@ bool bnxt_dev_is_vf_rep(struct net_device *dev) /* Called when the parent PF interface is closed: * As the mode transition from SWITCHDEV to LEGACY - * happens under the rtnl_lock() this routine is safe - * under the rtnl_lock() + * happens under the netdev instance lock this routine is safe */ void bnxt_vf_reps_close(struct bnxt *bp) { @@ -278,8 +277,7 @@ void bnxt_vf_reps_close(struct bnxt *bp) /* Called when the parent PF interface is opened (re-opened): * As the mode transition from SWITCHDEV to LEGACY - * happen under the rtnl_lock() this routine is safe - * under the rtnl_lock() + * happen under the netdev instance lock this routine is safe */ void bnxt_vf_reps_open(struct bnxt *bp) { @@ -348,7 +346,7 @@ void bnxt_vf_reps_destroy(struct bnxt *bp) /* Ensure that parent PF's and VF-reps' RX/TX has been quiesced * before proceeding with VF-rep cleanup. */ - rtnl_lock(); + netdev_lock(bp->dev); if (netif_running(bp->dev)) { bnxt_close_nic(bp, false, false); closed = true; @@ -365,10 +363,10 @@ void bnxt_vf_reps_destroy(struct bnxt *bp) bnxt_open_nic(bp, false, false); bp->eswitch_mode = DEVLINK_ESWITCH_MODE_SWITCHDEV; } - rtnl_unlock(); + netdev_unlock(bp->dev); - /* Need to call vf_reps_destroy() outside of rntl_lock - * as unregister_netdev takes rtnl_lock + /* Need to call vf_reps_destroy() outside of netdev instance lock + * as unregister_netdev takes it */ __bnxt_vf_reps_destroy(bp); } @@ -376,7 +374,7 @@ void bnxt_vf_reps_destroy(struct bnxt *bp) /* Free the VF-Reps in firmware, during firmware hot-reset processing. * Note that the VF-Rep netdevs are still active (not unregistered) during * this process. As the mode transition from SWITCHDEV to LEGACY happens - * under the rtnl_lock() this routine is safe under the rtnl_lock(). + * under the netdev instance lock this routine is safe. */ void bnxt_vf_reps_free(struct bnxt *bp) { @@ -413,7 +411,7 @@ static int bnxt_alloc_vf_rep(struct bnxt *bp, struct bnxt_vf_rep *vf_rep, /* Allocate the VF-Reps in firmware, during firmware hot-reset processing. * Note that the VF-Rep netdevs are still active (not unregistered) during * this process. As the mode transition from SWITCHDEV to LEGACY happens - * under the rtnl_lock() this routine is safe under the rtnl_lock(). + * under the netdev instance lock this routine is safe. */ int bnxt_vf_reps_alloc(struct bnxt *bp) { diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c index e6c64e4bd66c..0caf6e9bccb8 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c @@ -382,13 +382,14 @@ int bnxt_xdp_xmit(struct net_device *dev, int num_frames, return nxmit; } -/* Under rtnl_lock */ static int bnxt_xdp_set(struct bnxt *bp, struct bpf_prog *prog) { struct net_device *dev = bp->dev; int tx_xdp = 0, tx_cp, rc, tc; struct bpf_prog *old; + netdev_assert_locked(dev); + if (prog && !prog->aux->xdp_has_frags && bp->dev->mtu > BNXT_MAX_PAGE_MODE_MTU) { netdev_warn(dev, "MTU %d larger than %d without XDP frag support.\n", diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 2f8560a354ba..d206c9592b60 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2765,6 +2765,11 @@ static inline void netdev_lock(struct net_device *dev) mutex_lock(&dev->lock); } +static inline bool netdev_trylock(struct net_device *dev) +{ + return mutex_trylock(&dev->lock); +} + static inline void netdev_unlock(struct net_device *dev) { mutex_unlock(&dev->lock); -- cgit v1.2.3 From d4438ce68bf145aa1d7ed03ebf3b8ece337e3f64 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 5 Mar 2025 03:45:50 +0000 Subject: inet: call inet6_ehashfn() once from inet6_hash_connect() inet6_ehashfn() being called from __inet6_check_established() has a big impact on performance, as shown in the Tested section. After prior patch, we can compute the hash for port 0 from inet6_hash_connect(), and derive each hash in __inet_hash_connect() from this initial hash: hash(saddr, lport, daddr, dport) == hash(saddr, 0, daddr, dport) + lport Apply the same principle for __inet_check_established(), although inet_ehashfn() has a smaller cost. Tested: Server: ulimit -n 40000; neper/tcp_crr -T 200 -F 30000 -6 --nolog Client: ulimit -n 40000; neper/tcp_crr -T 200 -F 30000 -6 --nolog -c -H server Before this patch: utime_start=0.286131 utime_end=4.378886 stime_start=11.952556 stime_end=1991.655533 num_transactions=1446830 latency_min=0.001061085 latency_max=12.075275028 latency_mean=0.376375302 latency_stddev=1.361969596 num_samples=306383 throughput=151866.56 perf top: 50.01% [kernel] [k] __inet6_check_established 20.65% [kernel] [k] __inet_hash_connect 15.81% [kernel] [k] inet6_ehashfn 2.92% [kernel] [k] rcu_all_qs 2.34% [kernel] [k] __cond_resched 0.50% [kernel] [k] _raw_spin_lock 0.34% [kernel] [k] sched_balance_trigger 0.24% [kernel] [k] queued_spin_lock_slowpath After this patch: utime_start=0.315047 utime_end=9.257617 stime_start=7.041489 stime_end=1923.688387 num_transactions=3057968 latency_min=0.003041375 latency_max=7.056589232 latency_mean=0.141075048 # Better latency metrics latency_stddev=0.526900516 num_samples=312996 throughput=320677.21 # 111 % increase, and 229 % for the series perf top: inet6_ehashfn is no longer seen. 39.67% [kernel] [k] __inet_hash_connect 37.06% [kernel] [k] __inet6_check_established 4.79% [kernel] [k] rcu_all_qs 3.82% [kernel] [k] __cond_resched 1.76% [kernel] [k] sched_balance_domains 0.82% [kernel] [k] _raw_spin_lock 0.81% [kernel] [k] sched_balance_rq 0.81% [kernel] [k] sched_balance_trigger 0.76% [kernel] [k] queued_spin_lock_slowpath Signed-off-by: Eric Dumazet Reviewed-by: Kuniyuki Iwashima Tested-by: Jason Xing Reviewed-by: Jason Xing Link: https://patch.msgid.link/20250305034550.879255-3-edumazet@google.com Signed-off-by: Jakub Kicinski --- include/net/inet_hashtables.h | 4 +++- include/net/ip.h | 2 +- net/ipv4/inet_hashtables.c | 26 ++++++++++++++++++-------- net/ipv6/inet6_hashtables.c | 15 +++++++++++---- 4 files changed, 33 insertions(+), 14 deletions(-) (limited to 'include') diff --git a/include/net/inet_hashtables.h b/include/net/inet_hashtables.h index f447d61d9598..949641e92539 100644 --- a/include/net/inet_hashtables.h +++ b/include/net/inet_hashtables.h @@ -527,10 +527,12 @@ static inline void sk_rcv_saddr_set(struct sock *sk, __be32 addr) int __inet_hash_connect(struct inet_timewait_death_row *death_row, struct sock *sk, u64 port_offset, + u32 hash_port0, int (*check_established)(struct inet_timewait_death_row *, struct sock *, __u16, struct inet_timewait_sock **, - bool rcu_lookup)); + bool rcu_lookup, + u32 hash)); int inet_hash_connect(struct inet_timewait_death_row *death_row, struct sock *sk); diff --git a/include/net/ip.h b/include/net/ip.h index ce5e59957dd5..8a48ade24620 100644 --- a/include/net/ip.h +++ b/include/net/ip.h @@ -357,7 +357,7 @@ static inline void inet_get_local_port_range(const struct net *net, int *low, in bool inet_sk_get_local_port_range(const struct sock *sk, int *low, int *high); #ifdef CONFIG_SYSCTL -static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port) +static inline bool inet_is_local_reserved_port(const struct net *net, unsigned short port) { if (!net->ipv4.sysctl_local_reserved_ports) return false; diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index ccf69b534d2e..5bf163f756e9 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -538,7 +538,8 @@ EXPORT_SYMBOL_GPL(__inet_lookup_established); static int __inet_check_established(struct inet_timewait_death_row *death_row, struct sock *sk, __u16 lport, struct inet_timewait_sock **twp, - bool rcu_lookup) + bool rcu_lookup, + u32 hash) { struct inet_hashinfo *hinfo = death_row->hashinfo; struct inet_sock *inet = inet_sk(sk); @@ -549,8 +550,6 @@ static int __inet_check_established(struct inet_timewait_death_row *death_row, int sdif = l3mdev_master_ifindex_by_index(net, dif); INET_ADDR_COOKIE(acookie, saddr, daddr); const __portpair ports = INET_COMBINED_PORTS(inet->inet_dport, lport); - unsigned int hash = inet_ehashfn(net, daddr, lport, - saddr, inet->inet_dport); struct inet_ehash_bucket *head = inet_ehash_bucket(hinfo, hash); struct inet_timewait_sock *tw = NULL; const struct hlist_nulls_node *node; @@ -1007,9 +1006,10 @@ static u32 *table_perturb; int __inet_hash_connect(struct inet_timewait_death_row *death_row, struct sock *sk, u64 port_offset, + u32 hash_port0, int (*check_established)(struct inet_timewait_death_row *, struct sock *, __u16, struct inet_timewait_sock **, - bool rcu_lookup)) + bool rcu_lookup, u32 hash)) { struct inet_hashinfo *hinfo = death_row->hashinfo; struct inet_bind_hashbucket *head, *head2; @@ -1027,7 +1027,8 @@ int __inet_hash_connect(struct inet_timewait_death_row *death_row, if (port) { local_bh_disable(); - ret = check_established(death_row, sk, port, NULL, false); + ret = check_established(death_row, sk, port, NULL, false, + hash_port0 + port); local_bh_enable(); return ret; } @@ -1071,7 +1072,8 @@ other_parity_scan: rcu_read_unlock(); goto next_port; } - if (!check_established(death_row, sk, port, &tw, true)) + if (!check_established(death_row, sk, port, &tw, true, + hash_port0 + port)) break; rcu_read_unlock(); goto next_port; @@ -1090,7 +1092,8 @@ other_parity_scan: goto next_port_unlock; WARN_ON(hlist_empty(&tb->bhash2)); if (!check_established(death_row, sk, - port, &tw, false)) + port, &tw, false, + hash_port0 + port)) goto ok; goto next_port_unlock; } @@ -1197,11 +1200,18 @@ error: int inet_hash_connect(struct inet_timewait_death_row *death_row, struct sock *sk) { + const struct inet_sock *inet = inet_sk(sk); + const struct net *net = sock_net(sk); u64 port_offset = 0; + u32 hash_port0; if (!inet_sk(sk)->inet_num) port_offset = inet_sk_port_offset(sk); - return __inet_hash_connect(death_row, sk, port_offset, + + hash_port0 = inet_ehashfn(net, inet->inet_rcv_saddr, 0, + inet->inet_daddr, inet->inet_dport); + + return __inet_hash_connect(death_row, sk, port_offset, hash_port0, __inet_check_established); } EXPORT_SYMBOL_GPL(inet_hash_connect); diff --git a/net/ipv6/inet6_hashtables.c b/net/ipv6/inet6_hashtables.c index 3d95f1e75a11..76ee521189eb 100644 --- a/net/ipv6/inet6_hashtables.c +++ b/net/ipv6/inet6_hashtables.c @@ -264,7 +264,8 @@ EXPORT_SYMBOL_GPL(inet6_lookup); static int __inet6_check_established(struct inet_timewait_death_row *death_row, struct sock *sk, const __u16 lport, struct inet_timewait_sock **twp, - bool rcu_lookup) + bool rcu_lookup, + u32 hash) { struct inet_hashinfo *hinfo = death_row->hashinfo; struct inet_sock *inet = inet_sk(sk); @@ -274,8 +275,6 @@ static int __inet6_check_established(struct inet_timewait_death_row *death_row, struct net *net = sock_net(sk); const int sdif = l3mdev_master_ifindex_by_index(net, dif); const __portpair ports = INET_COMBINED_PORTS(inet->inet_dport, lport); - const unsigned int hash = inet6_ehashfn(net, daddr, lport, saddr, - inet->inet_dport); struct inet_ehash_bucket *head = inet_ehash_bucket(hinfo, hash); struct inet_timewait_sock *tw = NULL; const struct hlist_nulls_node *node; @@ -354,11 +353,19 @@ static u64 inet6_sk_port_offset(const struct sock *sk) int inet6_hash_connect(struct inet_timewait_death_row *death_row, struct sock *sk) { + const struct in6_addr *daddr = &sk->sk_v6_rcv_saddr; + const struct in6_addr *saddr = &sk->sk_v6_daddr; + const struct inet_sock *inet = inet_sk(sk); + const struct net *net = sock_net(sk); u64 port_offset = 0; + u32 hash_port0; if (!inet_sk(sk)->inet_num) port_offset = inet6_sk_port_offset(sk); - return __inet_hash_connect(death_row, sk, port_offset, + + hash_port0 = inet6_ehashfn(net, daddr, 0, saddr, inet->inet_dport); + + return __inet_hash_connect(death_row, sk, port_offset, hash_port0, __inet6_check_established); } EXPORT_SYMBOL_GPL(inet6_hash_connect); -- cgit v1.2.3 From e7112524e5e885181cc5ae4d258f33b9dbe0b907 Mon Sep 17 00:00:00 2001 From: Breno Leitao Date: Thu, 6 Mar 2025 08:27:51 -0800 Subject: block: Name the RQF flags enum Commit 5f89154e8e9e3445f9b59 ("block: Use enum to define RQF_x bit indexes") converted the RQF flags to an anonymous enum, which was a beneficial change. This patch goes one step further by naming the enum as "rqf_flags". This naming enables exporting these flags to BPF clients, eliminating the need to duplicate these flags in BPF code. Instead, BPF clients can now access the same kernel-side values through CO:RE (Compile Once, Run Everywhere), as shown in this example: rqf_stats = bpf_core_enum_value(enum rqf_flags, __RQF_STATS) Suggested-by: Yonghong Song Signed-off-by: Breno Leitao Link: https://lore.kernel.org/r/20250306-rqf_flags-v1-1-bbd64918b406@debian.org Signed-off-by: Jens Axboe --- include/linux/blk-mq.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index fa2a76cc2f73..71f4f0cc3dac 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -28,7 +28,7 @@ typedef enum rq_end_io_ret (rq_end_io_fn)(struct request *, blk_status_t); typedef __u32 __bitwise req_flags_t; /* Keep rqf_name[] in sync with the definitions below */ -enum { +enum rqf_flags { /* drive already may have started this one */ __RQF_STARTED, /* request for flush sequence */ -- cgit v1.2.3 From a2f61f1db85532e72fb8a3af51b06df94bb82912 Mon Sep 17 00:00:00 2001 From: Shahar Shitrit Date: Tue, 4 Mar 2025 18:06:15 +0200 Subject: net/mlx5: Relocate function declarations from port.h to mlx5_core.h The port header is a general file under include, yet it contains declarations for functions that are either not exported or exported but not used outside the mlx5_core driver. To enhance code organization, we move these declarations to mlx5_core.h, where they are more appropriately scoped. This refactor removes unnecessary exported symbols and prevents unexported functions from being inadvertently referenced outside of the mlx5_core driver. Signed-off-by: Shahar Shitrit Reviewed-by: Carolina Jubran Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/20250304160620.417580-2-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 85 ++++++++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/port.c | 20 ----- include/linux/mlx5/port.h | 85 +--------------------- 3 files changed, 86 insertions(+), 104 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index 6fef1005c469..6278b02105da 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -114,6 +114,21 @@ struct mlx5_cmd_alias_obj_create_attr { u8 access_key[ACCESS_KEY_LEN]; }; +struct mlx5_port_eth_proto { + u32 cap; + u32 admin; + u32 oper; +}; + +struct mlx5_module_eeprom_query_params { + u16 size; + u16 offset; + u16 i2c_address; + u32 page; + u32 bank; + u32 module_number; +}; + static inline void mlx5_printk(struct mlx5_core_dev *dev, int level, const char *format, ...) { struct device *device = dev->device; @@ -280,6 +295,76 @@ int mlx5_set_mtppse(struct mlx5_core_dev *mdev, u8 pin, u8 arm, u8 mode); struct mlx5_dm *mlx5_dm_create(struct mlx5_core_dev *dev); void mlx5_dm_cleanup(struct mlx5_core_dev *dev); +void mlx5_toggle_port_link(struct mlx5_core_dev *dev); +int mlx5_set_port_admin_status(struct mlx5_core_dev *dev, + enum mlx5_port_status status); +int mlx5_query_port_admin_status(struct mlx5_core_dev *dev, + enum mlx5_port_status *status); +int mlx5_set_port_beacon(struct mlx5_core_dev *dev, u16 beacon_duration); + +int mlx5_set_port_mtu(struct mlx5_core_dev *dev, u16 mtu, u8 port); +int mlx5_set_port_pause(struct mlx5_core_dev *dev, u32 rx_pause, u32 tx_pause); +int mlx5_query_port_pause(struct mlx5_core_dev *dev, + u32 *rx_pause, u32 *tx_pause); + +int mlx5_set_port_pfc(struct mlx5_core_dev *dev, u8 pfc_en_tx, u8 pfc_en_rx); +int mlx5_query_port_pfc(struct mlx5_core_dev *dev, u8 *pfc_en_tx, + u8 *pfc_en_rx); + +int mlx5_set_port_stall_watermark(struct mlx5_core_dev *dev, + u16 stall_critical_watermark, + u16 stall_minor_watermark); +int mlx5_query_port_stall_watermark(struct mlx5_core_dev *dev, + u16 *stall_critical_watermark, + u16 *stall_minor_watermark); + +int mlx5_max_tc(struct mlx5_core_dev *mdev); +int mlx5_set_port_prio_tc(struct mlx5_core_dev *mdev, u8 *prio_tc); +int mlx5_query_port_prio_tc(struct mlx5_core_dev *mdev, + u8 prio, u8 *tc); +int mlx5_set_port_tc_group(struct mlx5_core_dev *mdev, u8 *tc_group); +int mlx5_query_port_tc_group(struct mlx5_core_dev *mdev, + u8 tc, u8 *tc_group); +int mlx5_set_port_tc_bw_alloc(struct mlx5_core_dev *mdev, u8 *tc_bw); +int mlx5_query_port_tc_bw_alloc(struct mlx5_core_dev *mdev, + u8 tc, u8 *bw_pct); +int mlx5_modify_port_ets_rate_limit(struct mlx5_core_dev *mdev, + u8 *max_bw_value, + u8 *max_bw_unit); +int mlx5_query_port_ets_rate_limit(struct mlx5_core_dev *mdev, + u8 *max_bw_value, + u8 *max_bw_unit); +int mlx5_set_port_wol(struct mlx5_core_dev *mdev, u8 wol_mode); +int mlx5_query_port_wol(struct mlx5_core_dev *mdev, u8 *wol_mode); + +int mlx5_query_ports_check(struct mlx5_core_dev *mdev, u32 *out, int outlen); +int mlx5_set_ports_check(struct mlx5_core_dev *mdev, u32 *in, int inlen); +int mlx5_set_port_fcs(struct mlx5_core_dev *mdev, u8 enable); +void mlx5_query_port_fcs(struct mlx5_core_dev *mdev, bool *supported, + bool *enabled); +int mlx5_query_module_eeprom(struct mlx5_core_dev *dev, + u16 offset, u16 size, u8 *data); +int +mlx5_query_module_eeprom_by_page(struct mlx5_core_dev *dev, + struct mlx5_module_eeprom_query_params *params, + u8 *data); + +int mlx5_query_port_dcbx_param(struct mlx5_core_dev *mdev, u32 *out); +int mlx5_set_port_dcbx_param(struct mlx5_core_dev *mdev, u32 *in); +int mlx5_set_trust_state(struct mlx5_core_dev *mdev, u8 trust_state); +int mlx5_query_trust_state(struct mlx5_core_dev *mdev, u8 *trust_state); +int mlx5_set_dscp2prio(struct mlx5_core_dev *mdev, u8 dscp, u8 prio); +int mlx5_query_dscp2prio(struct mlx5_core_dev *mdev, u8 *dscp2prio); + +int mlx5_port_query_eth_proto(struct mlx5_core_dev *dev, u8 port, bool ext, + struct mlx5_port_eth_proto *eproto); +bool mlx5_ptys_ext_supported(struct mlx5_core_dev *mdev); +u32 mlx5_port_ptys2speed(struct mlx5_core_dev *mdev, u32 eth_proto_oper, + bool force_legacy); +u32 mlx5_port_speed2linkmodes(struct mlx5_core_dev *mdev, u32 speed, + bool force_legacy); +int mlx5_port_max_linkspeed(struct mlx5_core_dev *mdev, u32 *speed); + #define MLX5_PPS_CAP(mdev) (MLX5_CAP_GEN((mdev), pps) && \ MLX5_CAP_GEN((mdev), pps_modify) && \ MLX5_CAP_MCAM_FEATURE((mdev), mtpps_fs) && \ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/port.c b/drivers/net/ethernet/mellanox/mlx5/core/port.c index 3995df064101..c7d749e8e133 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/port.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/port.c @@ -196,7 +196,6 @@ void mlx5_toggle_port_link(struct mlx5_core_dev *dev) if (ps == MLX5_PORT_UP) mlx5_set_port_admin_status(dev, MLX5_PORT_UP); } -EXPORT_SYMBOL_GPL(mlx5_toggle_port_link); int mlx5_set_port_admin_status(struct mlx5_core_dev *dev, enum mlx5_port_status status) @@ -210,7 +209,6 @@ int mlx5_set_port_admin_status(struct mlx5_core_dev *dev, return mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out), MLX5_REG_PAOS, 0, 1); } -EXPORT_SYMBOL_GPL(mlx5_set_port_admin_status); int mlx5_query_port_admin_status(struct mlx5_core_dev *dev, enum mlx5_port_status *status) @@ -227,7 +225,6 @@ int mlx5_query_port_admin_status(struct mlx5_core_dev *dev, *status = MLX5_GET(paos_reg, out, admin_status); return 0; } -EXPORT_SYMBOL_GPL(mlx5_query_port_admin_status); static void mlx5_query_port_mtu(struct mlx5_core_dev *dev, u16 *admin_mtu, u16 *max_mtu, u16 *oper_mtu, u8 port) @@ -257,7 +254,6 @@ int mlx5_set_port_mtu(struct mlx5_core_dev *dev, u16 mtu, u8 port) return mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out), MLX5_REG_PMTU, 0, 1); } -EXPORT_SYMBOL_GPL(mlx5_set_port_mtu); void mlx5_query_port_max_mtu(struct mlx5_core_dev *dev, u16 *max_mtu, u8 port) @@ -447,7 +443,6 @@ int mlx5_query_module_eeprom(struct mlx5_core_dev *dev, return mlx5_query_mcia(dev, &query, data); } -EXPORT_SYMBOL_GPL(mlx5_query_module_eeprom); int mlx5_query_module_eeprom_by_page(struct mlx5_core_dev *dev, struct mlx5_module_eeprom_query_params *params, @@ -467,7 +462,6 @@ int mlx5_query_module_eeprom_by_page(struct mlx5_core_dev *dev, return mlx5_query_mcia(dev, params, data); } -EXPORT_SYMBOL_GPL(mlx5_query_module_eeprom_by_page); static int mlx5_query_port_pvlc(struct mlx5_core_dev *dev, u32 *pvlc, int pvlc_size, u8 local_port) @@ -518,7 +512,6 @@ int mlx5_set_port_pause(struct mlx5_core_dev *dev, u32 rx_pause, u32 tx_pause) return mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out), MLX5_REG_PFCC, 0, 1); } -EXPORT_SYMBOL_GPL(mlx5_set_port_pause); int mlx5_query_port_pause(struct mlx5_core_dev *dev, u32 *rx_pause, u32 *tx_pause) @@ -538,7 +531,6 @@ int mlx5_query_port_pause(struct mlx5_core_dev *dev, return 0; } -EXPORT_SYMBOL_GPL(mlx5_query_port_pause); int mlx5_set_port_stall_watermark(struct mlx5_core_dev *dev, u16 stall_critical_watermark, @@ -597,7 +589,6 @@ int mlx5_set_port_pfc(struct mlx5_core_dev *dev, u8 pfc_en_tx, u8 pfc_en_rx) return mlx5_core_access_reg(dev, in, sizeof(in), out, sizeof(out), MLX5_REG_PFCC, 0, 1); } -EXPORT_SYMBOL_GPL(mlx5_set_port_pfc); int mlx5_query_port_pfc(struct mlx5_core_dev *dev, u8 *pfc_en_tx, u8 *pfc_en_rx) { @@ -616,7 +607,6 @@ int mlx5_query_port_pfc(struct mlx5_core_dev *dev, u8 *pfc_en_tx, u8 *pfc_en_rx) return 0; } -EXPORT_SYMBOL_GPL(mlx5_query_port_pfc); int mlx5_max_tc(struct mlx5_core_dev *mdev) { @@ -667,7 +657,6 @@ int mlx5_set_port_prio_tc(struct mlx5_core_dev *mdev, u8 *prio_tc) return 0; } -EXPORT_SYMBOL_GPL(mlx5_set_port_prio_tc); int mlx5_query_port_prio_tc(struct mlx5_core_dev *mdev, u8 prio, u8 *tc) @@ -689,7 +678,6 @@ int mlx5_query_port_prio_tc(struct mlx5_core_dev *mdev, return err; } -EXPORT_SYMBOL_GPL(mlx5_query_port_prio_tc); static int mlx5_set_port_qetcr_reg(struct mlx5_core_dev *mdev, u32 *in, int inlen) @@ -728,7 +716,6 @@ int mlx5_set_port_tc_group(struct mlx5_core_dev *mdev, u8 *tc_group) return mlx5_set_port_qetcr_reg(mdev, in, sizeof(in)); } -EXPORT_SYMBOL_GPL(mlx5_set_port_tc_group); int mlx5_query_port_tc_group(struct mlx5_core_dev *mdev, u8 tc, u8 *tc_group) @@ -749,7 +736,6 @@ int mlx5_query_port_tc_group(struct mlx5_core_dev *mdev, return 0; } -EXPORT_SYMBOL_GPL(mlx5_query_port_tc_group); int mlx5_set_port_tc_bw_alloc(struct mlx5_core_dev *mdev, u8 *tc_bw) { @@ -763,7 +749,6 @@ int mlx5_set_port_tc_bw_alloc(struct mlx5_core_dev *mdev, u8 *tc_bw) return mlx5_set_port_qetcr_reg(mdev, in, sizeof(in)); } -EXPORT_SYMBOL_GPL(mlx5_set_port_tc_bw_alloc); int mlx5_query_port_tc_bw_alloc(struct mlx5_core_dev *mdev, u8 tc, u8 *bw_pct) @@ -784,7 +769,6 @@ int mlx5_query_port_tc_bw_alloc(struct mlx5_core_dev *mdev, return 0; } -EXPORT_SYMBOL_GPL(mlx5_query_port_tc_bw_alloc); int mlx5_modify_port_ets_rate_limit(struct mlx5_core_dev *mdev, u8 *max_bw_value, @@ -808,7 +792,6 @@ int mlx5_modify_port_ets_rate_limit(struct mlx5_core_dev *mdev, return mlx5_set_port_qetcr_reg(mdev, in, sizeof(in)); } -EXPORT_SYMBOL_GPL(mlx5_modify_port_ets_rate_limit); int mlx5_query_port_ets_rate_limit(struct mlx5_core_dev *mdev, u8 *max_bw_value, @@ -834,7 +817,6 @@ int mlx5_query_port_ets_rate_limit(struct mlx5_core_dev *mdev, return 0; } -EXPORT_SYMBOL_GPL(mlx5_query_port_ets_rate_limit); int mlx5_set_port_wol(struct mlx5_core_dev *mdev, u8 wol_mode) { @@ -845,7 +827,6 @@ int mlx5_set_port_wol(struct mlx5_core_dev *mdev, u8 wol_mode) MLX5_SET(set_wol_rol_in, in, wol_mode, wol_mode); return mlx5_cmd_exec_in(mdev, set_wol_rol, in); } -EXPORT_SYMBOL_GPL(mlx5_set_port_wol); int mlx5_query_port_wol(struct mlx5_core_dev *mdev, u8 *wol_mode) { @@ -860,7 +841,6 @@ int mlx5_query_port_wol(struct mlx5_core_dev *mdev, u8 *wol_mode) return err; } -EXPORT_SYMBOL_GPL(mlx5_query_port_wol); int mlx5_query_ports_check(struct mlx5_core_dev *mdev, u32 *out, int outlen) { diff --git a/include/linux/mlx5/port.h b/include/linux/mlx5/port.h index fd625e0dd869..58770b86f793 100644 --- a/include/linux/mlx5/port.h +++ b/include/linux/mlx5/port.h @@ -61,15 +61,6 @@ enum mlx5_an_status { #define MLX5_EEPROM_PAGE_LENGTH 256 #define MLX5_EEPROM_HIGH_PAGE_LENGTH 128 -struct mlx5_module_eeprom_query_params { - u16 size; - u16 offset; - u16 i2c_address; - u32 page; - u32 bank; - u32 module_number; -}; - enum mlx5e_link_mode { MLX5E_1000BASE_CX_SGMII = 0, MLX5E_1000BASE_KX = 1, @@ -145,12 +136,6 @@ enum mlx5_ptys_width { MLX5_PTYS_WIDTH_12X = 1 << 4, }; -struct mlx5_port_eth_proto { - u32 cap; - u32 admin; - u32 oper; -}; - #define MLX5E_PROT_MASK(link_mode) (1U << link_mode) #define MLX5_GET_ETH_PROTO(reg, out, ext, field) \ (ext ? MLX5_GET(reg, out, ext_##field) : \ @@ -163,14 +148,7 @@ int mlx5_query_port_ptys(struct mlx5_core_dev *dev, u32 *ptys, int mlx5_query_ib_port_oper(struct mlx5_core_dev *dev, u16 *link_width_oper, u16 *proto_oper, u8 local_port, u8 plane_index); -void mlx5_toggle_port_link(struct mlx5_core_dev *dev); -int mlx5_set_port_admin_status(struct mlx5_core_dev *dev, - enum mlx5_port_status status); -int mlx5_query_port_admin_status(struct mlx5_core_dev *dev, - enum mlx5_port_status *status); -int mlx5_set_port_beacon(struct mlx5_core_dev *dev, u16 beacon_duration); - -int mlx5_set_port_mtu(struct mlx5_core_dev *dev, u16 mtu, u8 port); + void mlx5_query_port_max_mtu(struct mlx5_core_dev *dev, u16 *max_mtu, u8 port); void mlx5_query_port_oper_mtu(struct mlx5_core_dev *dev, u16 *oper_mtu, u8 port); @@ -178,65 +156,4 @@ void mlx5_query_port_oper_mtu(struct mlx5_core_dev *dev, u16 *oper_mtu, int mlx5_query_port_vl_hw_cap(struct mlx5_core_dev *dev, u8 *vl_hw_cap, u8 local_port); -int mlx5_set_port_pause(struct mlx5_core_dev *dev, u32 rx_pause, u32 tx_pause); -int mlx5_query_port_pause(struct mlx5_core_dev *dev, - u32 *rx_pause, u32 *tx_pause); - -int mlx5_set_port_pfc(struct mlx5_core_dev *dev, u8 pfc_en_tx, u8 pfc_en_rx); -int mlx5_query_port_pfc(struct mlx5_core_dev *dev, u8 *pfc_en_tx, - u8 *pfc_en_rx); - -int mlx5_set_port_stall_watermark(struct mlx5_core_dev *dev, - u16 stall_critical_watermark, - u16 stall_minor_watermark); -int mlx5_query_port_stall_watermark(struct mlx5_core_dev *dev, - u16 *stall_critical_watermark, u16 *stall_minor_watermark); - -int mlx5_max_tc(struct mlx5_core_dev *mdev); - -int mlx5_set_port_prio_tc(struct mlx5_core_dev *mdev, u8 *prio_tc); -int mlx5_query_port_prio_tc(struct mlx5_core_dev *mdev, - u8 prio, u8 *tc); -int mlx5_set_port_tc_group(struct mlx5_core_dev *mdev, u8 *tc_group); -int mlx5_query_port_tc_group(struct mlx5_core_dev *mdev, - u8 tc, u8 *tc_group); -int mlx5_set_port_tc_bw_alloc(struct mlx5_core_dev *mdev, u8 *tc_bw); -int mlx5_query_port_tc_bw_alloc(struct mlx5_core_dev *mdev, - u8 tc, u8 *bw_pct); -int mlx5_modify_port_ets_rate_limit(struct mlx5_core_dev *mdev, - u8 *max_bw_value, - u8 *max_bw_unit); -int mlx5_query_port_ets_rate_limit(struct mlx5_core_dev *mdev, - u8 *max_bw_value, - u8 *max_bw_unit); -int mlx5_set_port_wol(struct mlx5_core_dev *mdev, u8 wol_mode); -int mlx5_query_port_wol(struct mlx5_core_dev *mdev, u8 *wol_mode); - -int mlx5_query_ports_check(struct mlx5_core_dev *mdev, u32 *out, int outlen); -int mlx5_set_ports_check(struct mlx5_core_dev *mdev, u32 *in, int inlen); -int mlx5_set_port_fcs(struct mlx5_core_dev *mdev, u8 enable); -void mlx5_query_port_fcs(struct mlx5_core_dev *mdev, bool *supported, - bool *enabled); -int mlx5_query_module_eeprom(struct mlx5_core_dev *dev, - u16 offset, u16 size, u8 *data); -int mlx5_query_module_eeprom_by_page(struct mlx5_core_dev *dev, - struct mlx5_module_eeprom_query_params *params, u8 *data); - -int mlx5_query_port_dcbx_param(struct mlx5_core_dev *mdev, u32 *out); -int mlx5_set_port_dcbx_param(struct mlx5_core_dev *mdev, u32 *in); - -int mlx5_set_trust_state(struct mlx5_core_dev *mdev, u8 trust_state); -int mlx5_query_trust_state(struct mlx5_core_dev *mdev, u8 *trust_state); -int mlx5_set_dscp2prio(struct mlx5_core_dev *mdev, u8 dscp, u8 prio); -int mlx5_query_dscp2prio(struct mlx5_core_dev *mdev, u8 *dscp2prio); - -int mlx5_port_query_eth_proto(struct mlx5_core_dev *dev, u8 port, bool ext, - struct mlx5_port_eth_proto *eproto); -bool mlx5_ptys_ext_supported(struct mlx5_core_dev *mdev); -u32 mlx5_port_ptys2speed(struct mlx5_core_dev *mdev, u32 eth_proto_oper, - bool force_legacy); -u32 mlx5_port_speed2linkmodes(struct mlx5_core_dev *mdev, u32 speed, - bool force_legacy); -int mlx5_port_max_linkspeed(struct mlx5_core_dev *mdev, u32 *speed); - #endif /* __MLX5_PORT_H__ */ -- cgit v1.2.3 From 00a7d39898c8010bfd5ff62af31ca5db34421b38 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Thu, 6 Mar 2025 18:25:35 -1000 Subject: fs/pipe: add simpler helpers for common cases The fix to atomically read the pipe head and tail state when not holding the pipe mutex has caused a number of headaches due to the size change of the involved types. It turns out that we don't have _that_ many places that access these fields directly and were affected, but we have more than we strictly should have, because our low-level helper functions have been designed to have intimate knowledge of how the pipes work. And as a result, that random noise of direct 'pipe->head' and 'pipe->tail' accesses makes it harder to pinpoint any actual potential problem spots remaining. For example, we didn't have a "is the pipe full" helper function, but instead had a "given these pipe buffer indexes and this pipe size, is the pipe full". That's because some low-level pipe code does actually want that much more complicated interface. But most other places literally just want a "is the pipe full" helper, and not having it meant that those places ended up being unnecessarily much too aware of this all. It would have been much better if only the very core pipe code that cared had been the one aware of this all. So let's fix it - better late than never. This just introduces the trivial wrappers for "is this pipe full or empty" and to get how many pipe buffers are used, so that instead of writing if (pipe_full(pipe->head, pipe->tail, pipe->max_usage)) the places that literally just want to know if a pipe is full can just say if (pipe_is_full(pipe)) instead. The existing trivial cases were converted with a 'sed' script. This cuts down on the places that access pipe->head and pipe->tail directly outside of the pipe code (and core splice code) quite a lot. The splice code in particular still revels in doing the direct low-level accesses, and the fuse fuse_dev_splice_write() code also seems a bit unnecessarily eager to go very low-level, but it's at least a bit better than it used to be. Signed-off-by: Linus Torvalds --- drivers/char/virtio_console.c | 4 ++-- fs/fuse/dev.c | 2 +- fs/pipe.c | 6 +++--- fs/splice.c | 20 ++++++++++---------- include/linux/pipe_fs_i.h | 27 +++++++++++++++++++++++++++ mm/filemap.c | 7 +++---- mm/shmem.c | 6 +++--- 7 files changed, 49 insertions(+), 23 deletions(-) (limited to 'include') diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 24442485e73e..18f92dd44d45 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -923,14 +923,14 @@ static ssize_t port_fops_splice_write(struct pipe_inode_info *pipe, pipe_lock(pipe); ret = 0; - if (pipe_empty(pipe->head, pipe->tail)) + if (pipe_is_empty(pipe)) goto error_out; ret = wait_port_writable(port, filp->f_flags & O_NONBLOCK); if (ret < 0) goto error_out; - occupancy = pipe_occupancy(pipe->head, pipe->tail); + occupancy = pipe_buf_usage(pipe); buf = alloc_buf(port->portdev->vdev, 0, occupancy); if (!buf) { diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 3c9caafca9e2..2c3a4d09e500 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -1457,7 +1457,7 @@ static ssize_t fuse_dev_splice_read(struct file *in, loff_t *ppos, if (ret < 0) goto out; - if (pipe_occupancy(pipe->head, pipe->tail) + cs.nr_segs > pipe->max_usage) { + if (pipe_buf_usage(pipe) + cs.nr_segs > pipe->max_usage) { ret = -EIO; goto out; } diff --git a/fs/pipe.c b/fs/pipe.c index 5c872775a6db..4d0799e4e719 100644 --- a/fs/pipe.c +++ b/fs/pipe.c @@ -394,7 +394,7 @@ pipe_read(struct kiocb *iocb, struct iov_iter *to) wake_next_reader = true; mutex_lock(&pipe->mutex); } - if (pipe_empty(pipe->head, pipe->tail)) + if (pipe_is_empty(pipe)) wake_next_reader = false; mutex_unlock(&pipe->mutex); @@ -577,11 +577,11 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from) kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN); wait_event_interruptible_exclusive(pipe->wr_wait, pipe_writable(pipe)); mutex_lock(&pipe->mutex); - was_empty = pipe_empty(pipe->head, pipe->tail); + was_empty = pipe_is_empty(pipe); wake_next_writer = true; } out: - if (pipe_full(pipe->head, pipe->tail, pipe->max_usage)) + if (pipe_is_full(pipe)) wake_next_writer = false; mutex_unlock(&pipe->mutex); diff --git a/fs/splice.c b/fs/splice.c index 28cfa63aa236..23fa5561b944 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -331,7 +331,7 @@ ssize_t copy_splice_read(struct file *in, loff_t *ppos, int i; /* Work out how much data we can actually add into the pipe */ - used = pipe_occupancy(pipe->head, pipe->tail); + used = pipe_buf_usage(pipe); npages = max_t(ssize_t, pipe->max_usage - used, 0); len = min_t(size_t, len, npages * PAGE_SIZE); npages = DIV_ROUND_UP(len, PAGE_SIZE); @@ -527,7 +527,7 @@ static int splice_from_pipe_next(struct pipe_inode_info *pipe, struct splice_des return -ERESTARTSYS; repeat: - while (pipe_empty(pipe->head, pipe->tail)) { + while (pipe_is_empty(pipe)) { if (!pipe->writers) return 0; @@ -820,7 +820,7 @@ ssize_t splice_to_socket(struct pipe_inode_info *pipe, struct file *out, if (signal_pending(current)) break; - while (pipe_empty(pipe->head, pipe->tail)) { + while (pipe_is_empty(pipe)) { ret = 0; if (!pipe->writers) goto out; @@ -968,7 +968,7 @@ static ssize_t do_splice_read(struct file *in, loff_t *ppos, return 0; /* Don't try to read more the pipe has space for. */ - p_space = pipe->max_usage - pipe_occupancy(pipe->head, pipe->tail); + p_space = pipe->max_usage - pipe_buf_usage(pipe); len = min_t(size_t, len, p_space << PAGE_SHIFT); if (unlikely(len > MAX_RW_COUNT)) @@ -1080,7 +1080,7 @@ ssize_t splice_direct_to_actor(struct file *in, struct splice_desc *sd, more = sd->flags & SPLICE_F_MORE; sd->flags |= SPLICE_F_MORE; - WARN_ON_ONCE(!pipe_empty(pipe->head, pipe->tail)); + WARN_ON_ONCE(!pipe_is_empty(pipe)); while (len) { size_t read_len; @@ -1268,7 +1268,7 @@ static int wait_for_space(struct pipe_inode_info *pipe, unsigned flags) send_sig(SIGPIPE, current, 0); return -EPIPE; } - if (!pipe_full(pipe->head, pipe->tail, pipe->max_usage)) + if (!pipe_is_full(pipe)) return 0; if (flags & SPLICE_F_NONBLOCK) return -EAGAIN; @@ -1652,13 +1652,13 @@ static int ipipe_prep(struct pipe_inode_info *pipe, unsigned int flags) * Check the pipe occupancy without the inode lock first. This function * is speculative anyways, so missing one is ok. */ - if (!pipe_empty(pipe->head, pipe->tail)) + if (!pipe_is_empty(pipe)) return 0; ret = 0; pipe_lock(pipe); - while (pipe_empty(pipe->head, pipe->tail)) { + while (pipe_is_empty(pipe)) { if (signal_pending(current)) { ret = -ERESTARTSYS; break; @@ -1688,13 +1688,13 @@ static int opipe_prep(struct pipe_inode_info *pipe, unsigned int flags) * Check pipe occupancy without the inode lock first. This function * is speculative anyways, so missing one is ok. */ - if (!pipe_full(pipe->head, pipe->tail, pipe->max_usage)) + if (!pipe_is_full(pipe)) return 0; ret = 0; pipe_lock(pipe); - while (pipe_full(pipe->head, pipe->tail, pipe->max_usage)) { + while (pipe_is_full(pipe)) { if (!pipe->readers) { send_sig(SIGPIPE, current, 0); ret = -EPIPE; diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h index 4d0a2267e6ef..b698758000f8 100644 --- a/include/linux/pipe_fs_i.h +++ b/include/linux/pipe_fs_i.h @@ -208,6 +208,33 @@ static inline bool pipe_full(unsigned int head, unsigned int tail, return pipe_occupancy(head, tail) >= limit; } +/** + * pipe_is_full - Return true if the pipe is full + * @pipe: the pipe + */ +static inline bool pipe_is_full(const struct pipe_inode_info *pipe) +{ + return pipe_full(pipe->head, pipe->tail, pipe->max_usage); +} + +/** + * pipe_is_empty - Return true if the pipe is empty + * @pipe: the pipe + */ +static inline bool pipe_is_empty(const struct pipe_inode_info *pipe) +{ + return pipe_empty(pipe->head, pipe->tail); +} + +/** + * pipe_buf_usage - Return how many pipe buffers are in use + * @pipe: the pipe + */ +static inline unsigned int pipe_buf_usage(const struct pipe_inode_info *pipe) +{ + return pipe_occupancy(pipe->head, pipe->tail); +} + /** * pipe_buf - Return the pipe buffer for the specified slot in the pipe ring * @pipe: The pipe to access diff --git a/mm/filemap.c b/mm/filemap.c index d4564a79eb35..2974691fdfad 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2897,8 +2897,7 @@ size_t splice_folio_into_pipe(struct pipe_inode_info *pipe, size = min(size, folio_size(folio) - offset); offset %= PAGE_SIZE; - while (spliced < size && - !pipe_full(pipe->head, pipe->tail, pipe->max_usage)) { + while (spliced < size && !pipe_is_full(pipe)) { struct pipe_buffer *buf = pipe_head_buf(pipe); size_t part = min_t(size_t, PAGE_SIZE - offset, size - spliced); @@ -2955,7 +2954,7 @@ ssize_t filemap_splice_read(struct file *in, loff_t *ppos, iocb.ki_pos = *ppos; /* Work out how much data we can actually add into the pipe */ - used = pipe_occupancy(pipe->head, pipe->tail); + used = pipe_buf_usage(pipe); npages = max_t(ssize_t, pipe->max_usage - used, 0); len = min_t(size_t, len, npages * PAGE_SIZE); @@ -3015,7 +3014,7 @@ ssize_t filemap_splice_read(struct file *in, loff_t *ppos, total_spliced += n; *ppos += n; in->f_ra.prev_pos = *ppos; - if (pipe_full(pipe->head, pipe->tail, pipe->max_usage)) + if (pipe_is_full(pipe)) goto out; } diff --git a/mm/shmem.c b/mm/shmem.c index 4ea6109a8043..20032a333d80 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -3487,7 +3487,7 @@ static size_t splice_zeropage_into_pipe(struct pipe_inode_info *pipe, size = min_t(size_t, size, PAGE_SIZE - offset); - if (!pipe_full(pipe->head, pipe->tail, pipe->max_usage)) { + if (!pipe_is_full(pipe)) { struct pipe_buffer *buf = pipe_head_buf(pipe); *buf = (struct pipe_buffer) { @@ -3514,7 +3514,7 @@ static ssize_t shmem_file_splice_read(struct file *in, loff_t *ppos, int error = 0; /* Work out how much data we can actually add into the pipe */ - used = pipe_occupancy(pipe->head, pipe->tail); + used = pipe_buf_usage(pipe); npages = max_t(ssize_t, pipe->max_usage - used, 0); len = min_t(size_t, len, npages * PAGE_SIZE); @@ -3601,7 +3601,7 @@ static ssize_t shmem_file_splice_read(struct file *in, loff_t *ppos, total_spliced += n; *ppos += n; in->f_ra.prev_pos = *ppos; - if (pipe_full(pipe->head, pipe->tail, pipe->max_usage)) + if (pipe_is_full(pipe)) break; cond_resched(); -- cgit v1.2.3 From ab6ab707a4d060a51c45fc13e3b2228d5f7c0b87 Mon Sep 17 00:00:00 2001 From: Luiz Augusto von Dentz Date: Tue, 4 Mar 2025 10:06:10 -0500 Subject: Revert "Bluetooth: hci_core: Fix sleeping function called from invalid context" This reverts commit 4d94f05558271654670d18c26c912da0c1c15549 which has problems (see [1]) and is no longer needed since 581dd2dc168f ("Bluetooth: hci_event: Fix using rcu_read_(un)lock while iterating") has reworked the code where the original bug has been found. [1] Link: https://lore.kernel.org/linux-bluetooth/877c55ci1r.wl-tiwai@suse.de/T/#t Fixes: 4d94f0555827 ("Bluetooth: hci_core: Fix sleeping function called from invalid context") Signed-off-by: Luiz Augusto von Dentz --- include/net/bluetooth/hci_core.h | 108 ++++++++++++++------------------------- net/bluetooth/hci_core.c | 10 ++-- net/bluetooth/iso.c | 6 --- net/bluetooth/l2cap_core.c | 12 ++--- net/bluetooth/rfcomm/core.c | 6 --- net/bluetooth/sco.c | 12 ++--- 6 files changed, 57 insertions(+), 97 deletions(-) (limited to 'include') diff --git a/include/net/bluetooth/hci_core.h b/include/net/bluetooth/hci_core.h index f756fac95488..6281063cbd8e 100644 --- a/include/net/bluetooth/hci_core.h +++ b/include/net/bluetooth/hci_core.h @@ -804,6 +804,7 @@ struct hci_conn_params { extern struct list_head hci_dev_list; extern struct list_head hci_cb_list; extern rwlock_t hci_dev_list_lock; +extern struct mutex hci_cb_list_lock; #define hci_dev_set_flag(hdev, nr) set_bit((nr), (hdev)->dev_flags) #define hci_dev_clear_flag(hdev, nr) clear_bit((nr), (hdev)->dev_flags) @@ -2010,47 +2011,24 @@ struct hci_cb { char *name; - bool (*match) (struct hci_conn *conn); void (*connect_cfm) (struct hci_conn *conn, __u8 status); void (*disconn_cfm) (struct hci_conn *conn, __u8 status); void (*security_cfm) (struct hci_conn *conn, __u8 status, - __u8 encrypt); + __u8 encrypt); void (*key_change_cfm) (struct hci_conn *conn, __u8 status); void (*role_switch_cfm) (struct hci_conn *conn, __u8 status, __u8 role); }; -static inline void hci_cb_lookup(struct hci_conn *conn, struct list_head *list) -{ - struct hci_cb *cb, *cpy; - - rcu_read_lock(); - list_for_each_entry_rcu(cb, &hci_cb_list, list) { - if (cb->match && cb->match(conn)) { - cpy = kmalloc(sizeof(*cpy), GFP_ATOMIC); - if (!cpy) - break; - - *cpy = *cb; - INIT_LIST_HEAD(&cpy->list); - list_add_rcu(&cpy->list, list); - } - } - rcu_read_unlock(); -} - static inline void hci_connect_cfm(struct hci_conn *conn, __u8 status) { - struct list_head list; - struct hci_cb *cb, *tmp; - - INIT_LIST_HEAD(&list); - hci_cb_lookup(conn, &list); + struct hci_cb *cb; - list_for_each_entry_safe(cb, tmp, &list, list) { + mutex_lock(&hci_cb_list_lock); + list_for_each_entry(cb, &hci_cb_list, list) { if (cb->connect_cfm) cb->connect_cfm(conn, status); - kfree(cb); } + mutex_unlock(&hci_cb_list_lock); if (conn->connect_cfm_cb) conn->connect_cfm_cb(conn, status); @@ -2058,43 +2036,22 @@ static inline void hci_connect_cfm(struct hci_conn *conn, __u8 status) static inline void hci_disconn_cfm(struct hci_conn *conn, __u8 reason) { - struct list_head list; - struct hci_cb *cb, *tmp; - - INIT_LIST_HEAD(&list); - hci_cb_lookup(conn, &list); + struct hci_cb *cb; - list_for_each_entry_safe(cb, tmp, &list, list) { + mutex_lock(&hci_cb_list_lock); + list_for_each_entry(cb, &hci_cb_list, list) { if (cb->disconn_cfm) cb->disconn_cfm(conn, reason); - kfree(cb); } + mutex_unlock(&hci_cb_list_lock); if (conn->disconn_cfm_cb) conn->disconn_cfm_cb(conn, reason); } -static inline void hci_security_cfm(struct hci_conn *conn, __u8 status, - __u8 encrypt) -{ - struct list_head list; - struct hci_cb *cb, *tmp; - - INIT_LIST_HEAD(&list); - hci_cb_lookup(conn, &list); - - list_for_each_entry_safe(cb, tmp, &list, list) { - if (cb->security_cfm) - cb->security_cfm(conn, status, encrypt); - kfree(cb); - } - - if (conn->security_cfm_cb) - conn->security_cfm_cb(conn, status); -} - static inline void hci_auth_cfm(struct hci_conn *conn, __u8 status) { + struct hci_cb *cb; __u8 encrypt; if (test_bit(HCI_CONN_ENCRYPT_PEND, &conn->flags)) @@ -2102,11 +2059,20 @@ static inline void hci_auth_cfm(struct hci_conn *conn, __u8 status) encrypt = test_bit(HCI_CONN_ENCRYPT, &conn->flags) ? 0x01 : 0x00; - hci_security_cfm(conn, status, encrypt); + mutex_lock(&hci_cb_list_lock); + list_for_each_entry(cb, &hci_cb_list, list) { + if (cb->security_cfm) + cb->security_cfm(conn, status, encrypt); + } + mutex_unlock(&hci_cb_list_lock); + + if (conn->security_cfm_cb) + conn->security_cfm_cb(conn, status); } static inline void hci_encrypt_cfm(struct hci_conn *conn, __u8 status) { + struct hci_cb *cb; __u8 encrypt; if (conn->state == BT_CONFIG) { @@ -2133,38 +2099,40 @@ static inline void hci_encrypt_cfm(struct hci_conn *conn, __u8 status) conn->sec_level = conn->pending_sec_level; } - hci_security_cfm(conn, status, encrypt); + mutex_lock(&hci_cb_list_lock); + list_for_each_entry(cb, &hci_cb_list, list) { + if (cb->security_cfm) + cb->security_cfm(conn, status, encrypt); + } + mutex_unlock(&hci_cb_list_lock); + + if (conn->security_cfm_cb) + conn->security_cfm_cb(conn, status); } static inline void hci_key_change_cfm(struct hci_conn *conn, __u8 status) { - struct list_head list; - struct hci_cb *cb, *tmp; - - INIT_LIST_HEAD(&list); - hci_cb_lookup(conn, &list); + struct hci_cb *cb; - list_for_each_entry_safe(cb, tmp, &list, list) { + mutex_lock(&hci_cb_list_lock); + list_for_each_entry(cb, &hci_cb_list, list) { if (cb->key_change_cfm) cb->key_change_cfm(conn, status); - kfree(cb); } + mutex_unlock(&hci_cb_list_lock); } static inline void hci_role_switch_cfm(struct hci_conn *conn, __u8 status, __u8 role) { - struct list_head list; - struct hci_cb *cb, *tmp; - - INIT_LIST_HEAD(&list); - hci_cb_lookup(conn, &list); + struct hci_cb *cb; - list_for_each_entry_safe(cb, tmp, &list, list) { + mutex_lock(&hci_cb_list_lock); + list_for_each_entry(cb, &hci_cb_list, list) { if (cb->role_switch_cfm) cb->role_switch_cfm(conn, status, role); - kfree(cb); } + mutex_unlock(&hci_cb_list_lock); } static inline bool hci_bdaddr_is_rpa(bdaddr_t *bdaddr, u8 addr_type) diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c index e7ec12437c8b..012fc107901a 100644 --- a/net/bluetooth/hci_core.c +++ b/net/bluetooth/hci_core.c @@ -57,6 +57,7 @@ DEFINE_RWLOCK(hci_dev_list_lock); /* HCI callback list */ LIST_HEAD(hci_cb_list); +DEFINE_MUTEX(hci_cb_list_lock); /* HCI ID Numbering */ static DEFINE_IDA(hci_index_ida); @@ -2972,7 +2973,9 @@ int hci_register_cb(struct hci_cb *cb) { BT_DBG("%p name %s", cb, cb->name); - list_add_tail_rcu(&cb->list, &hci_cb_list); + mutex_lock(&hci_cb_list_lock); + list_add_tail(&cb->list, &hci_cb_list); + mutex_unlock(&hci_cb_list_lock); return 0; } @@ -2982,8 +2985,9 @@ int hci_unregister_cb(struct hci_cb *cb) { BT_DBG("%p name %s", cb, cb->name); - list_del_rcu(&cb->list); - synchronize_rcu(); + mutex_lock(&hci_cb_list_lock); + list_del(&cb->list); + mutex_unlock(&hci_cb_list_lock); return 0; } diff --git a/net/bluetooth/iso.c b/net/bluetooth/iso.c index 44acddf58a0c..0cb52a3308ba 100644 --- a/net/bluetooth/iso.c +++ b/net/bluetooth/iso.c @@ -2187,11 +2187,6 @@ done: return HCI_LM_ACCEPT; } -static bool iso_match(struct hci_conn *hcon) -{ - return hcon->type == ISO_LINK || hcon->type == LE_LINK; -} - static void iso_connect_cfm(struct hci_conn *hcon, __u8 status) { if (hcon->type != ISO_LINK) { @@ -2373,7 +2368,6 @@ drop: static struct hci_cb iso_cb = { .name = "ISO", - .match = iso_match, .connect_cfm = iso_connect_cfm, .disconn_cfm = iso_disconn_cfm, }; diff --git a/net/bluetooth/l2cap_core.c b/net/bluetooth/l2cap_core.c index b22078b67972..c27ea70f71e1 100644 --- a/net/bluetooth/l2cap_core.c +++ b/net/bluetooth/l2cap_core.c @@ -7182,11 +7182,6 @@ static struct l2cap_chan *l2cap_global_fixed_chan(struct l2cap_chan *c, return NULL; } -static bool l2cap_match(struct hci_conn *hcon) -{ - return hcon->type == ACL_LINK || hcon->type == LE_LINK; -} - static void l2cap_connect_cfm(struct hci_conn *hcon, u8 status) { struct hci_dev *hdev = hcon->hdev; @@ -7194,6 +7189,9 @@ static void l2cap_connect_cfm(struct hci_conn *hcon, u8 status) struct l2cap_chan *pchan; u8 dst_type; + if (hcon->type != ACL_LINK && hcon->type != LE_LINK) + return; + BT_DBG("hcon %p bdaddr %pMR status %d", hcon, &hcon->dst, status); if (status) { @@ -7258,6 +7256,9 @@ int l2cap_disconn_ind(struct hci_conn *hcon) static void l2cap_disconn_cfm(struct hci_conn *hcon, u8 reason) { + if (hcon->type != ACL_LINK && hcon->type != LE_LINK) + return; + BT_DBG("hcon %p reason %d", hcon, reason); l2cap_conn_del(hcon, bt_to_errno(reason)); @@ -7565,7 +7566,6 @@ unlock: static struct hci_cb l2cap_cb = { .name = "L2CAP", - .match = l2cap_match, .connect_cfm = l2cap_connect_cfm, .disconn_cfm = l2cap_disconn_cfm, .security_cfm = l2cap_security_cfm, diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c index 4c56ca5a216c..ad5177e3a69b 100644 --- a/net/bluetooth/rfcomm/core.c +++ b/net/bluetooth/rfcomm/core.c @@ -2134,11 +2134,6 @@ static int rfcomm_run(void *unused) return 0; } -static bool rfcomm_match(struct hci_conn *hcon) -{ - return hcon->type == ACL_LINK; -} - static void rfcomm_security_cfm(struct hci_conn *conn, u8 status, u8 encrypt) { struct rfcomm_session *s; @@ -2185,7 +2180,6 @@ static void rfcomm_security_cfm(struct hci_conn *conn, u8 status, u8 encrypt) static struct hci_cb rfcomm_cb = { .name = "RFCOMM", - .match = rfcomm_match, .security_cfm = rfcomm_security_cfm }; diff --git a/net/bluetooth/sco.c b/net/bluetooth/sco.c index ed6846864ea9..5d1bc0d6aee0 100644 --- a/net/bluetooth/sco.c +++ b/net/bluetooth/sco.c @@ -1407,13 +1407,11 @@ int sco_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, __u8 *flags) return lm; } -static bool sco_match(struct hci_conn *hcon) -{ - return hcon->type == SCO_LINK || hcon->type == ESCO_LINK; -} - static void sco_connect_cfm(struct hci_conn *hcon, __u8 status) { + if (hcon->type != SCO_LINK && hcon->type != ESCO_LINK) + return; + BT_DBG("hcon %p bdaddr %pMR status %u", hcon, &hcon->dst, status); if (!status) { @@ -1430,6 +1428,9 @@ static void sco_connect_cfm(struct hci_conn *hcon, __u8 status) static void sco_disconn_cfm(struct hci_conn *hcon, __u8 reason) { + if (hcon->type != SCO_LINK && hcon->type != ESCO_LINK) + return; + BT_DBG("hcon %p reason %d", hcon, reason); sco_conn_del(hcon, bt_to_errno(reason)); @@ -1455,7 +1456,6 @@ drop: static struct hci_cb sco_cb = { .name = "SCO", - .match = sco_match, .connect_cfm = sco_connect_cfm, .disconn_cfm = sco_disconn_cfm, }; -- cgit v1.2.3 From c8be7018d47cfdf36ae6b5bedddca9fc99cd2f0b Mon Sep 17 00:00:00 2001 From: "Dr. David Alan Gilbert" Date: Thu, 6 Mar 2025 18:45:34 +0000 Subject: net: phylink: Remove unused phylink_init_eee phylink_init_eee() is currently unused. It was last added in 2019 by commit 86e58135bc4a ("net: phylink: add phylink_init_eee() helper") but it didn't actually wire a use up. It had previous been removed in 2017 by commit 939eae25d9a5 ("phylink: remove phylink_init_eee()"). Remove it again. Signed-off-by: Dr. David Alan Gilbert Reviewed-by: Russell King (Oracle) Link: https://patch.msgid.link/20250306184534.246152-1-linux@treblig.org Signed-off-by: Jakub Kicinski --- drivers/net/phy/phylink.c | 18 ------------------ include/linux/phylink.h | 1 - 2 files changed, 19 deletions(-) (limited to 'include') diff --git a/drivers/net/phy/phylink.c b/drivers/net/phy/phylink.c index a3f64b6d2d34..76b1a329607b 100644 --- a/drivers/net/phy/phylink.c +++ b/drivers/net/phy/phylink.c @@ -3157,24 +3157,6 @@ int phylink_get_eee_err(struct phylink *pl) } EXPORT_SYMBOL_GPL(phylink_get_eee_err); -/** - * phylink_init_eee() - init and check the EEE features - * @pl: a pointer to a &struct phylink returned from phylink_create() - * @clk_stop_enable: allow PHY to stop receive clock - * - * Must be called either with RTNL held or within mac_link_up() - */ -int phylink_init_eee(struct phylink *pl, bool clk_stop_enable) -{ - int ret = -EOPNOTSUPP; - - if (pl->phydev) - ret = phy_init_eee(pl->phydev, clk_stop_enable); - - return ret; -} -EXPORT_SYMBOL_GPL(phylink_init_eee); - /** * phylink_ethtool_get_eee() - read the energy efficient ethernet parameters * @pl: a pointer to a &struct phylink returned from phylink_create() diff --git a/include/linux/phylink.h b/include/linux/phylink.h index 08df65f6867a..c187267a15b6 100644 --- a/include/linux/phylink.h +++ b/include/linux/phylink.h @@ -714,7 +714,6 @@ void phylink_ethtool_get_pauseparam(struct phylink *, int phylink_ethtool_set_pauseparam(struct phylink *, struct ethtool_pauseparam *); int phylink_get_eee_err(struct phylink *); -int phylink_init_eee(struct phylink *, bool); int phylink_ethtool_get_eee(struct phylink *link, struct ethtool_keee *eee); int phylink_ethtool_set_eee(struct phylink *link, struct ethtool_keee *eee); int phylink_mii_ioctl(struct phylink *, struct ifreq *, int); -- cgit v1.2.3 From 0d7336f8f06d4a1a1e2c62624d086561e8490bb7 Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Thu, 6 Mar 2025 12:29:28 +0100 Subject: tcp: ulp: diag: more info without CAP_NET_ADMIN When introduced in commit 61723b393292 ("tcp: ulp: add functions to dump ulp-specific information"), the whole ULP diag info has been exported only if the requester had CAP_NET_ADMIN. It looks like not everything is sensitive, and some info can be exported to all users in order to ease the debugging from the userspace side without requiring additional capabilities. Each layer should then decide what can be exposed to everybody. The 'net_admin' boolean is then passed to the different layers. On kTLS side, it looks like there is nothing sensitive there: version, cipher type, tx/rx user config type, plus some flags. So, only some metadata about the configuration, no cryptographic info like keys, etc. Then, everything can be exported to all users. On MPTCP side, that's different. The MPTCP-related sequence numbers per subflow should certainly not be exposed to everybody. For example, the DSS mapping and ssn_offset would give all users on the system access to narrow ranges of values for the subflow TCP sequence numbers and MPTCP-level DSNs, and then ease packet injection. The TCP diag interface doesn't expose the TCP sequence numbers for TCP sockets, so best to do the same here. The rest -- token, IDs, flags -- can be exported to everybody. Acked-by: Mat Martineau Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20250306-net-next-tcp-ulp-diag-net-admin-v1-2-06afdd860fc9@kernel.org Signed-off-by: Jakub Kicinski --- include/net/tcp.h | 4 ++-- net/ipv4/tcp_diag.c | 8 ++++---- net/mptcp/diag.c | 42 ++++++++++++++++++++++++++---------------- net/tls/tls_main.c | 4 ++-- 4 files changed, 34 insertions(+), 24 deletions(-) (limited to 'include') diff --git a/include/net/tcp.h b/include/net/tcp.h index a9bc959fb102..7207c52b1fc9 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -2598,8 +2598,8 @@ struct tcp_ulp_ops { /* cleanup ulp */ void (*release)(struct sock *sk); /* diagnostic */ - int (*get_info)(struct sock *sk, struct sk_buff *skb); - size_t (*get_info_size)(const struct sock *sk); + int (*get_info)(struct sock *sk, struct sk_buff *skb, bool net_admin); + size_t (*get_info_size)(const struct sock *sk, bool net_admin); /* clone ulp */ void (*clone)(const struct request_sock *req, struct sock *newsk, const gfp_t priority); diff --git a/net/ipv4/tcp_diag.c b/net/ipv4/tcp_diag.c index d8bba37dbffd..45e174b8cd22 100644 --- a/net/ipv4/tcp_diag.c +++ b/net/ipv4/tcp_diag.c @@ -96,8 +96,8 @@ static int tcp_diag_put_ulp(struct sk_buff *skb, struct sock *sk, if (err) goto nla_failure; - if (net_admin && ulp_ops->get_info) - err = ulp_ops->get_info(sk, skb); + if (ulp_ops->get_info) + err = ulp_ops->get_info(sk, skb, net_admin); if (err) goto nla_failure; @@ -170,8 +170,8 @@ static size_t tcp_diag_get_aux_size(struct sock *sk, bool net_admin) if (ulp_ops) { size += nla_total_size(0) + nla_total_size(TCP_ULP_NAME_MAX); - if (net_admin && ulp_ops->get_info_size) - size += ulp_ops->get_info_size(sk); + if (ulp_ops->get_info_size) + size += ulp_ops->get_info_size(sk, net_admin); } } return size; diff --git a/net/mptcp/diag.c b/net/mptcp/diag.c index 02205f7994d7..70cf9ebce833 100644 --- a/net/mptcp/diag.c +++ b/net/mptcp/diag.c @@ -12,7 +12,7 @@ #include #include "protocol.h" -static int subflow_get_info(struct sock *sk, struct sk_buff *skb) +static int subflow_get_info(struct sock *sk, struct sk_buff *skb, bool net_admin) { struct mptcp_subflow_context *sf; struct nlattr *start; @@ -56,15 +56,6 @@ static int subflow_get_info(struct sock *sk, struct sk_buff *skb) if (nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_TOKEN_REM, sf->remote_token) || nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_TOKEN_LOC, sf->token) || - nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_RELWRITE_SEQ, - sf->rel_write_seq) || - nla_put_u64_64bit(skb, MPTCP_SUBFLOW_ATTR_MAP_SEQ, sf->map_seq, - MPTCP_SUBFLOW_ATTR_PAD) || - nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_MAP_SFSEQ, - sf->map_subflow_seq) || - nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_SSN_OFFSET, sf->ssn_offset) || - nla_put_u16(skb, MPTCP_SUBFLOW_ATTR_MAP_DATALEN, - sf->map_data_len) || nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_FLAGS, flags) || nla_put_u8(skb, MPTCP_SUBFLOW_ATTR_ID_REM, sf->remote_id) || nla_put_u8(skb, MPTCP_SUBFLOW_ATTR_ID_LOC, subflow_get_local_id(sf))) { @@ -72,6 +63,21 @@ static int subflow_get_info(struct sock *sk, struct sk_buff *skb) goto nla_failure; } + /* Only export seq related counters to user with CAP_NET_ADMIN */ + if (net_admin && + (nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_RELWRITE_SEQ, + sf->rel_write_seq) || + nla_put_u64_64bit(skb, MPTCP_SUBFLOW_ATTR_MAP_SEQ, sf->map_seq, + MPTCP_SUBFLOW_ATTR_PAD) || + nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_MAP_SFSEQ, + sf->map_subflow_seq) || + nla_put_u32(skb, MPTCP_SUBFLOW_ATTR_SSN_OFFSET, sf->ssn_offset) || + nla_put_u16(skb, MPTCP_SUBFLOW_ATTR_MAP_DATALEN, + sf->map_data_len))) { + err = -EMSGSIZE; + goto nla_failure; + } + rcu_read_unlock(); unlock_sock_fast(sk, slow); nla_nest_end(skb, start); @@ -84,22 +90,26 @@ nla_failure: return err; } -static size_t subflow_get_info_size(const struct sock *sk) +static size_t subflow_get_info_size(const struct sock *sk, bool net_admin) { size_t size = 0; size += nla_total_size(0) + /* INET_ULP_INFO_MPTCP */ nla_total_size(4) + /* MPTCP_SUBFLOW_ATTR_TOKEN_REM */ nla_total_size(4) + /* MPTCP_SUBFLOW_ATTR_TOKEN_LOC */ - nla_total_size(4) + /* MPTCP_SUBFLOW_ATTR_RELWRITE_SEQ */ - nla_total_size_64bit(8) + /* MPTCP_SUBFLOW_ATTR_MAP_SEQ */ - nla_total_size(4) + /* MPTCP_SUBFLOW_ATTR_MAP_SFSEQ */ - nla_total_size(4) + /* MPTCP_SUBFLOW_ATTR_SSN_OFFSET */ - nla_total_size(2) + /* MPTCP_SUBFLOW_ATTR_MAP_DATALEN */ nla_total_size(4) + /* MPTCP_SUBFLOW_ATTR_FLAGS */ nla_total_size(1) + /* MPTCP_SUBFLOW_ATTR_ID_REM */ nla_total_size(1) + /* MPTCP_SUBFLOW_ATTR_ID_LOC */ 0; + + if (net_admin) + size += nla_total_size(4) + /* MPTCP_SUBFLOW_ATTR_RELWRITE_SEQ */ + nla_total_size_64bit(8) + /* MPTCP_SUBFLOW_ATTR_MAP_SEQ */ + nla_total_size(4) + /* MPTCP_SUBFLOW_ATTR_MAP_SFSEQ */ + nla_total_size(4) + /* MPTCP_SUBFLOW_ATTR_SSN_OFFSET */ + nla_total_size(2) + /* MPTCP_SUBFLOW_ATTR_MAP_DATALEN */ + 0; + return size; } diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c index 99ca4465f702..cb86b0bf9a53 100644 --- a/net/tls/tls_main.c +++ b/net/tls/tls_main.c @@ -1057,7 +1057,7 @@ static u16 tls_user_config(struct tls_context *ctx, bool tx) return 0; } -static int tls_get_info(struct sock *sk, struct sk_buff *skb) +static int tls_get_info(struct sock *sk, struct sk_buff *skb, bool net_admin) { u16 version, cipher_type; struct tls_context *ctx; @@ -1115,7 +1115,7 @@ nla_failure: return err; } -static size_t tls_get_info_size(const struct sock *sk) +static size_t tls_get_info_size(const struct sock *sk, bool net_admin) { size_t size = 0; -- cgit v1.2.3 From 248f6571fd4c51531f7f8f07f186f7ae98a50afc Mon Sep 17 00:00:00 2001 From: Breno Leitao Date: Tue, 4 Mar 2025 07:50:41 -0800 Subject: netpoll: Optimize skb refilling on critical path netpoll tries to refill the skb queue on every packet send, independently if packets are being consumed from the pool or not. This was particularly problematic while being called from printk(), where the operation would be done while holding the console lock. Introduce a more intelligent approach to skb queue management. Instead of constantly attempting to refill the queue, the system now defers refilling to a work queue and only triggers the workqueue when a buffer is actually dequeued. This change significantly reduces operations with the lock held. Add a work_struct to the netpoll structure for asynchronous refilling, updating find_skb() to schedule refill work only when necessary (skb is dequeued). These changes have demonstrated a 15% reduction in time spent during netpoll_send_msg operations, especially when no SKBs are not consumed from consumed from pool. When SKBs are being dequeued, the improvement is even better, around 70%, mainly because refilling the SKB pool is now happening outside of the critical patch (with console_owner lock held). Signed-off-by: Breno Leitao Reviewed-by: Simon Horman Link: https://patch.msgid.link/20250304-netpoll_refill_v2-v1-1-06e2916a4642@debian.org Signed-off-by: Jakub Kicinski --- include/linux/netpoll.h | 1 + net/core/netpoll.c | 15 +++++++++++++-- 2 files changed, 14 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/netpoll.h b/include/linux/netpoll.h index f91e50a76efd..f6e8abe0b1f1 100644 --- a/include/linux/netpoll.h +++ b/include/linux/netpoll.h @@ -33,6 +33,7 @@ struct netpoll { u16 local_port, remote_port; u8 remote_mac[ETH_ALEN]; struct sk_buff_head skb_pool; + struct work_struct refill_wq; }; struct netpoll_info { diff --git a/net/core/netpoll.c b/net/core/netpoll.c index 62b4041aae1a..8a0df2b274a8 100644 --- a/net/core/netpoll.c +++ b/net/core/netpoll.c @@ -284,12 +284,13 @@ static struct sk_buff *find_skb(struct netpoll *np, int len, int reserve) struct sk_buff *skb; zap_completion_queue(); - refill_skbs(np); repeat: skb = alloc_skb(len, GFP_ATOMIC); - if (!skb) + if (!skb) { skb = skb_dequeue(&np->skb_pool); + schedule_work(&np->refill_wq); + } if (!skb) { if (++count < 10) { @@ -535,6 +536,7 @@ static void skb_pool_flush(struct netpoll *np) { struct sk_buff_head *skb_pool; + cancel_work_sync(&np->refill_wq); skb_pool = &np->skb_pool; skb_queue_purge_reason(skb_pool, SKB_CONSUMED); } @@ -621,6 +623,14 @@ int netpoll_parse_options(struct netpoll *np, char *opt) } EXPORT_SYMBOL(netpoll_parse_options); +static void refill_skbs_work_handler(struct work_struct *work) +{ + struct netpoll *np = + container_of(work, struct netpoll, refill_wq); + + refill_skbs(np); +} + int __netpoll_setup(struct netpoll *np, struct net_device *ndev) { struct netpoll_info *npinfo; @@ -666,6 +676,7 @@ int __netpoll_setup(struct netpoll *np, struct net_device *ndev) /* fill up the skb queue */ refill_skbs(np); + INIT_WORK(&np->refill_wq, refill_skbs_work_handler); /* last thing to do is link it to the net device structure */ rcu_assign_pointer(ndev->npinfo, npinfo); -- cgit v1.2.3 From 8ef890df4031121a94407c84659125cbccd3fdbe Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Fri, 7 Mar 2025 10:30:06 -0800 Subject: net: move misc netdev_lock flavors to a separate header Move the more esoteric helpers for netdev instance lock to a dedicated header. This avoids growing netdevice.h to infinity and makes rebuilding the kernel much faster (after touching the header with the helpers). The main netdev_lock() / netdev_unlock() functions are used in static inlines in netdevice.h and will probably be used most commonly, so keep them in netdevice.h. Acked-by: Stanislav Fomichev Link: https://patch.msgid.link/20250307183006.2312761-1-kuba@kernel.org Signed-off-by: Jakub Kicinski --- drivers/net/bonding/bond_main.c | 1 + drivers/net/dummy.c | 1 + drivers/net/ethernet/broadcom/bnxt/bnxt.c | 1 + drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c | 1 + drivers/net/ethernet/intel/iavf/iavf_main.c | 1 + drivers/net/ethernet/microsoft/mana/mana_en.c | 1 + drivers/net/geneve.c | 1 + drivers/net/hamradio/bpqether.c | 1 + drivers/net/hyperv/netvsc_drv.c | 1 + drivers/net/ipvlan/ipvlan_main.c | 1 + drivers/net/loopback.c | 1 + drivers/net/macsec.c | 1 + drivers/net/macvlan.c | 1 + drivers/net/netdevsim/netdev.c | 1 + drivers/net/ppp/ppp_generic.c | 1 + drivers/net/team/team_core.c | 1 + drivers/net/veth.c | 1 + drivers/net/vrf.c | 1 + drivers/net/vxlan/vxlan_core.c | 1 + include/linux/netdevice.h | 81 +----------------------- include/net/netdev_lock.h | 89 +++++++++++++++++++++++++++ kernel/bpf/offload.c | 1 + net/8021q/vlan_dev.c | 1 + net/bluetooth/6lowpan.c | 1 + net/bridge/br_device.c | 2 + net/core/dev.c | 1 + net/core/dev.h | 1 + net/core/dev_api.c | 2 + net/core/dev_ioctl.c | 1 + net/core/net-sysfs.c | 1 + net/core/rtnetlink.c | 1 + net/dsa/conduit.c | 1 + net/ethtool/cabletest.c | 1 + net/ethtool/cmis_fw_update.c | 1 + net/ethtool/features.c | 2 + net/ethtool/ioctl.c | 1 + net/ethtool/module.c | 1 + net/ethtool/netlink.c | 1 + net/ethtool/phy.c | 1 + net/ethtool/rss.c | 2 + net/ethtool/tsinfo.c | 1 + net/ieee802154/6lowpan/core.c | 1 + net/ipv4/ip_tunnel.c | 1 + net/ipv6/ip6_gre.c | 1 + net/ipv6/ip6_tunnel.c | 1 + net/ipv6/ip6_vti.c | 1 + net/ipv6/sit.c | 1 + net/l2tp/l2tp_eth.c | 1 + net/sched/sch_api.c | 1 + net/xdp/xsk.c | 1 + net/xdp/xsk_buff_pool.c | 1 + 51 files changed, 143 insertions(+), 80 deletions(-) create mode 100644 include/net/netdev_lock.h (limited to 'include') diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index cf0b02720dd8..6c95f478ab80 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -90,6 +90,7 @@ #include #endif #include +#include #include #include "bonding_priv.h" diff --git a/drivers/net/dummy.c b/drivers/net/dummy.c index 005d79975f3b..a4938c6a5ebb 100644 --- a/drivers/net/dummy.c +++ b/drivers/net/dummy.c @@ -38,6 +38,7 @@ #include #include #include +#include #include #include diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c index 1a1e6da77777..b09171110ec4 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c @@ -54,6 +54,7 @@ #include #include #include +#include #include #include #include diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c index 0caf6e9bccb8..cf2b3ad75c9b 100644 --- a/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #include "bnxt_hsi.h" #include "bnxt.h" diff --git a/drivers/net/ethernet/intel/iavf/iavf_main.c b/drivers/net/ethernet/intel/iavf/iavf_main.c index 032e1a58af6f..6d7ba4d67a19 100644 --- a/drivers/net/ethernet/intel/iavf/iavf_main.c +++ b/drivers/net/ethernet/intel/iavf/iavf_main.c @@ -2,6 +2,7 @@ /* Copyright(c) 2013 - 2018 Intel Corporation. */ #include +#include #include "iavf.h" #include "iavf_ptp.h" diff --git a/drivers/net/ethernet/microsoft/mana/mana_en.c b/drivers/net/ethernet/microsoft/mana/mana_en.c index 0411a1897f57..2d826077d38c 100644 --- a/drivers/net/ethernet/microsoft/mana/mana_en.c +++ b/drivers/net/ethernet/microsoft/mana/mana_en.c @@ -13,6 +13,7 @@ #include #include +#include #include #include diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c index 2c65f867fd31..66e38ce9cd1d 100644 --- a/drivers/net/geneve.c +++ b/drivers/net/geneve.c @@ -18,6 +18,7 @@ #include #include #include +#include #include #define GENEVE_NETDEV_VER "0.6" diff --git a/drivers/net/hamradio/bpqether.c b/drivers/net/hamradio/bpqether.c index f6b0bfbbc753..0e0fe32d2da4 100644 --- a/drivers/net/hamradio/bpqether.c +++ b/drivers/net/hamradio/bpqether.c @@ -77,6 +77,7 @@ #include #include +#include #include #include diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c index 9c6501bf27bd..c51b318b8a72 100644 --- a/drivers/net/hyperv/netvsc_drv.c +++ b/drivers/net/hyperv/netvsc_drv.c @@ -29,6 +29,7 @@ #include #include +#include #include #include #include diff --git a/drivers/net/ipvlan/ipvlan_main.c b/drivers/net/ipvlan/ipvlan_main.c index b56144ca2fde..0ed2fd833a5d 100644 --- a/drivers/net/ipvlan/ipvlan_main.c +++ b/drivers/net/ipvlan/ipvlan_main.c @@ -3,6 +3,7 @@ */ #include +#include #include "ipvlan.h" diff --git a/drivers/net/loopback.c b/drivers/net/loopback.c index 201fddcd3b1e..1fb6ce6843ad 100644 --- a/drivers/net/loopback.c +++ b/drivers/net/loopback.c @@ -54,6 +54,7 @@ #include #include #include +#include #include /* blackhole_netdev - a device used for dsts that are marked expired! diff --git a/drivers/net/macsec.c b/drivers/net/macsec.c index 4de5d63fd577..3d315e30ee47 100644 --- a/drivers/net/macsec.c +++ b/drivers/net/macsec.c @@ -19,6 +19,7 @@ #include #include #include +#include #include #include #include diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c index 4e9d54be887c..d0dfa6bca6cc 100644 --- a/drivers/net/macvlan.c +++ b/drivers/net/macvlan.c @@ -28,6 +28,7 @@ #include #include #include +#include #include #include #include diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c index 54d03b0628d2..d71fd2907cc8 100644 --- a/drivers/net/netdevsim/netdev.c +++ b/drivers/net/netdevsim/netdev.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #include #include diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c index ca77661688c0..53463767cc43 100644 --- a/drivers/net/ppp/ppp_generic.c +++ b/drivers/net/ppp/ppp_generic.c @@ -45,6 +45,7 @@ #include #include #include +#include #include #include #include diff --git a/drivers/net/team/team_core.c b/drivers/net/team/team_core.c index fb917560d0a2..d8fc0c79745d 100644 --- a/drivers/net/team/team_core.c +++ b/drivers/net/team/team_core.c @@ -23,6 +23,7 @@ #include #include #include +#include #include #include #include diff --git a/drivers/net/veth.c b/drivers/net/veth.c index 05f5eeef539f..7bb53961c0ea 100644 --- a/drivers/net/veth.c +++ b/drivers/net/veth.c @@ -17,6 +17,7 @@ #include #include +#include #include #include #include diff --git a/drivers/net/vrf.c b/drivers/net/vrf.c index 36cf6191335e..7168b33adadb 100644 --- a/drivers/net/vrf.c +++ b/drivers/net/vrf.c @@ -34,6 +34,7 @@ #include #include #include +#include #include #include #include diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c index 227d7f5a302a..8c49e903cb3a 100644 --- a/drivers/net/vxlan/vxlan_core.c +++ b/drivers/net/vxlan/vxlan_core.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #include #include diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index d206c9592b60..9a297757df7e 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2630,40 +2630,6 @@ static inline void netdev_for_each_tx_queue(struct net_device *dev, f(dev, &dev->_tx[i], arg); } -static inline int netdev_lock_cmp_fn(const struct lockdep_map *a, - const struct lockdep_map *b) -{ - /* Only lower devices currently grab the instance lock, so no - * real ordering issues can occur. In the near future, only - * hardware devices will grab instance lock which also does not - * involve any ordering. Suppress lockdep ordering warnings - * until (if) we start grabbing instance lock on pure SW - * devices (bond/team/veth/etc). - */ - if (a == b) - return 0; - return -1; -} - -#define netdev_lockdep_set_classes(dev) \ -{ \ - static struct lock_class_key qdisc_tx_busylock_key; \ - static struct lock_class_key qdisc_xmit_lock_key; \ - static struct lock_class_key dev_addr_list_lock_key; \ - static struct lock_class_key dev_instance_lock_key; \ - unsigned int i; \ - \ - (dev)->qdisc_tx_busylock = &qdisc_tx_busylock_key; \ - lockdep_set_class(&(dev)->addr_list_lock, \ - &dev_addr_list_lock_key); \ - lockdep_set_class(&(dev)->lock, \ - &dev_instance_lock_key); \ - lock_set_cmp_fn(&dev->lock, netdev_lock_cmp_fn, NULL); \ - for (i = 0; i < (dev)->num_tx_queues; i++) \ - lockdep_set_class(&(dev)->_tx[i]._xmit_lock, \ - &qdisc_xmit_lock_key); \ -} - u16 netdev_pick_tx(struct net_device *dev, struct sk_buff *skb, struct net_device *sb_dev); struct netdev_queue *netdev_core_pick_tx(struct net_device *dev, @@ -2765,56 +2731,11 @@ static inline void netdev_lock(struct net_device *dev) mutex_lock(&dev->lock); } -static inline bool netdev_trylock(struct net_device *dev) -{ - return mutex_trylock(&dev->lock); -} - static inline void netdev_unlock(struct net_device *dev) { mutex_unlock(&dev->lock); } - -static inline void netdev_assert_locked(struct net_device *dev) -{ - lockdep_assert_held(&dev->lock); -} - -static inline void netdev_assert_locked_or_invisible(struct net_device *dev) -{ - if (dev->reg_state == NETREG_REGISTERED || - dev->reg_state == NETREG_UNREGISTERING) - netdev_assert_locked(dev); -} - -static inline bool netdev_need_ops_lock(struct net_device *dev) -{ - bool ret = dev->request_ops_lock || !!dev->queue_mgmt_ops; - -#if IS_ENABLED(CONFIG_NET_SHAPER) - ret |= !!dev->netdev_ops->net_shaper_ops; -#endif - - return ret; -} - -static inline void netdev_lock_ops(struct net_device *dev) -{ - if (netdev_need_ops_lock(dev)) - netdev_lock(dev); -} - -static inline void netdev_unlock_ops(struct net_device *dev) -{ - if (netdev_need_ops_lock(dev)) - netdev_unlock(dev); -} - -static inline void netdev_ops_assert_locked(struct net_device *dev) -{ - if (netdev_need_ops_lock(dev)) - lockdep_assert_held(&dev->lock); -} +/* Additional netdev_lock()-related helpers are in net/netdev_lock.h */ void netif_napi_set_irq_locked(struct napi_struct *napi, int irq); diff --git a/include/net/netdev_lock.h b/include/net/netdev_lock.h new file mode 100644 index 000000000000..99631fbd7f54 --- /dev/null +++ b/include/net/netdev_lock.h @@ -0,0 +1,89 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ + +#ifndef _NET_NETDEV_LOCK_H +#define _NET_NETDEV_LOCK_H + +#include +#include + +static inline bool netdev_trylock(struct net_device *dev) +{ + return mutex_trylock(&dev->lock); +} + +static inline void netdev_assert_locked(struct net_device *dev) +{ + lockdep_assert_held(&dev->lock); +} + +static inline void netdev_assert_locked_or_invisible(struct net_device *dev) +{ + if (dev->reg_state == NETREG_REGISTERED || + dev->reg_state == NETREG_UNREGISTERING) + netdev_assert_locked(dev); +} + +static inline bool netdev_need_ops_lock(struct net_device *dev) +{ + bool ret = dev->request_ops_lock || !!dev->queue_mgmt_ops; + +#if IS_ENABLED(CONFIG_NET_SHAPER) + ret |= !!dev->netdev_ops->net_shaper_ops; +#endif + + return ret; +} + +static inline void netdev_lock_ops(struct net_device *dev) +{ + if (netdev_need_ops_lock(dev)) + netdev_lock(dev); +} + +static inline void netdev_unlock_ops(struct net_device *dev) +{ + if (netdev_need_ops_lock(dev)) + netdev_unlock(dev); +} + +static inline void netdev_ops_assert_locked(struct net_device *dev) +{ + if (netdev_need_ops_lock(dev)) + lockdep_assert_held(&dev->lock); +} + +static inline int netdev_lock_cmp_fn(const struct lockdep_map *a, + const struct lockdep_map *b) +{ + /* Only lower devices currently grab the instance lock, so no + * real ordering issues can occur. In the near future, only + * hardware devices will grab instance lock which also does not + * involve any ordering. Suppress lockdep ordering warnings + * until (if) we start grabbing instance lock on pure SW + * devices (bond/team/veth/etc). + */ + if (a == b) + return 0; + return -1; +} + +#define netdev_lockdep_set_classes(dev) \ +{ \ + static struct lock_class_key qdisc_tx_busylock_key; \ + static struct lock_class_key qdisc_xmit_lock_key; \ + static struct lock_class_key dev_addr_list_lock_key; \ + static struct lock_class_key dev_instance_lock_key; \ + unsigned int i; \ + \ + (dev)->qdisc_tx_busylock = &qdisc_tx_busylock_key; \ + lockdep_set_class(&(dev)->addr_list_lock, \ + &dev_addr_list_lock_key); \ + lockdep_set_class(&(dev)->lock, \ + &dev_instance_lock_key); \ + lock_set_cmp_fn(&dev->lock, netdev_lock_cmp_fn, NULL); \ + for (i = 0; i < (dev)->num_tx_queues; i++) \ + lockdep_set_class(&(dev)->_tx[i]._xmit_lock, \ + &qdisc_xmit_lock_key); \ +} + +#endif diff --git a/kernel/bpf/offload.c b/kernel/bpf/offload.c index 4f707cfe7f10..42ae8d595c2c 100644 --- a/kernel/bpf/offload.c +++ b/kernel/bpf/offload.c @@ -25,6 +25,7 @@ #include #include #include +#include #include /* Protects offdevs, members of bpf_offload_netdev and offload members diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index ee3283400716..770a4dcf7f63 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -27,6 +27,7 @@ #include #include #include +#include #include "vlan.h" #include "vlanproc.h" diff --git a/net/bluetooth/6lowpan.c b/net/bluetooth/6lowpan.c index 50cfec8ccac4..1298c8685bad 100644 --- a/net/bluetooth/6lowpan.c +++ b/net/bluetooth/6lowpan.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #include diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c index 9d8c72ed01ab..a818fdc22da9 100644 --- a/net/bridge/br_device.c +++ b/net/bridge/br_device.c @@ -16,6 +16,8 @@ #include #include +#include + #include "br_private.h" #define COMMON_FEATURES (NETIF_F_SG | NETIF_F_FRAGLIST | NETIF_F_HIGHDMA | \ diff --git a/net/core/dev.c b/net/core/dev.c index a0f75a1d1f5a..1cb134ff7327 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -156,6 +156,7 @@ #include #include #include +#include #include #include #include diff --git a/net/core/dev.h b/net/core/dev.h index b50ca645c086..0ddd3631acb0 100644 --- a/net/core/dev.h +++ b/net/core/dev.h @@ -6,6 +6,7 @@ #include #include #include +#include struct net; struct netlink_ext_ack; diff --git a/net/core/dev_api.c b/net/core/dev_api.c index 655a95fb7baa..1f0e24849bc6 100644 --- a/net/core/dev_api.c +++ b/net/core/dev_api.c @@ -1,5 +1,7 @@ // SPDX-License-Identifier: GPL-2.0-or-later + #include +#include #include "dev.h" diff --git a/net/core/dev_ioctl.c b/net/core/dev_ioctl.c index 296e52d1395d..5471cf4fc984 100644 --- a/net/core/dev_ioctl.c +++ b/net/core/dev_ioctl.c @@ -10,6 +10,7 @@ #include #include #include +#include #include #include "dev.h" diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index 02d1d40b47ae..529a0f721268 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -23,6 +23,7 @@ #include #include #include +#include #include #include diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 88a352b02bce..90597bf84e3d 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -53,6 +53,7 @@ #include #include #include +#include #include #if IS_ENABLED(CONFIG_IPV6) #include diff --git a/net/dsa/conduit.c b/net/dsa/conduit.c index f21bb2551bed..4ae255cfb23f 100644 --- a/net/dsa/conduit.c +++ b/net/dsa/conduit.c @@ -10,6 +10,7 @@ #include #include #include +#include #include "conduit.h" #include "dsa.h" diff --git a/net/ethtool/cabletest.c b/net/ethtool/cabletest.c index ddcba073321f..0364b8fb577b 100644 --- a/net/ethtool/cabletest.c +++ b/net/ethtool/cabletest.c @@ -2,6 +2,7 @@ #include #include +#include #include "netlink.h" #include "common.h" diff --git a/net/ethtool/cmis_fw_update.c b/net/ethtool/cmis_fw_update.c index 946830af3e7c..df5f344209c4 100644 --- a/net/ethtool/cmis_fw_update.c +++ b/net/ethtool/cmis_fw_update.c @@ -2,6 +2,7 @@ #include #include +#include #include "common.h" #include "module_fw.h" diff --git a/net/ethtool/features.c b/net/ethtool/features.c index ccffd64d5a87..f2217983be2b 100644 --- a/net/ethtool/features.c +++ b/net/ethtool/features.c @@ -1,5 +1,7 @@ // SPDX-License-Identifier: GPL-2.0-only +#include + #include "netlink.h" #include "common.h" #include "bitset.h" diff --git a/net/ethtool/ioctl.c b/net/ethtool/ioctl.c index 496a2774100c..221639407c72 100644 --- a/net/ethtool/ioctl.c +++ b/net/ethtool/ioctl.c @@ -31,6 +31,7 @@ #include #include #include +#include #include #include "common.h" diff --git a/net/ethtool/module.c b/net/ethtool/module.c index d3d2e135e45e..4d4e0a82579a 100644 --- a/net/ethtool/module.c +++ b/net/ethtool/module.c @@ -4,6 +4,7 @@ #include #include #include +#include #include "netlink.h" #include "common.h" diff --git a/net/ethtool/netlink.c b/net/ethtool/netlink.c index 70834947f474..a163d40c6431 100644 --- a/net/ethtool/netlink.c +++ b/net/ethtool/netlink.c @@ -1,5 +1,6 @@ // SPDX-License-Identifier: GPL-2.0-only +#include #include #include #include diff --git a/net/ethtool/phy.c b/net/ethtool/phy.c index 1a6b725d1f14..1f590e8d75ed 100644 --- a/net/ethtool/phy.c +++ b/net/ethtool/phy.c @@ -9,6 +9,7 @@ #include #include #include +#include struct phy_req_info { struct ethnl_req_info base; diff --git a/net/ethtool/rss.c b/net/ethtool/rss.c index ec41d1d7eefe..6d9b1769896b 100644 --- a/net/ethtool/rss.c +++ b/net/ethtool/rss.c @@ -1,5 +1,7 @@ // SPDX-License-Identifier: GPL-2.0-only +#include + #include "netlink.h" #include "common.h" diff --git a/net/ethtool/tsinfo.c b/net/ethtool/tsinfo.c index 73b6a89b8731..32204cca24da 100644 --- a/net/ethtool/tsinfo.c +++ b/net/ethtool/tsinfo.c @@ -4,6 +4,7 @@ #include #include #include +#include #include "netlink.h" #include "common.h" diff --git a/net/ieee802154/6lowpan/core.c b/net/ieee802154/6lowpan/core.c index 9a9da74b0a4f..018929563c6b 100644 --- a/net/ieee802154/6lowpan/core.c +++ b/net/ieee802154/6lowpan/core.c @@ -50,6 +50,7 @@ #include #include +#include #include "6lowpan_i.h" diff --git a/net/ipv4/ip_tunnel.c b/net/ipv4/ip_tunnel.c index 4b06dc7e04f2..1024f961ec9a 100644 --- a/net/ipv4/ip_tunnel.c +++ b/net/ipv4/ip_tunnel.c @@ -40,6 +40,7 @@ #include #include #include +#include #include #include #include diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c index c6ebb6a6d390..957ca98fa70f 100644 --- a/net/ipv6/ip6_gre.c +++ b/net/ipv6/ip6_gre.c @@ -43,6 +43,7 @@ #include #include #include +#include #include #include diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c index 170a6ac30889..a04dd1bb4b19 100644 --- a/net/ipv6/ip6_tunnel.c +++ b/net/ipv6/ip6_tunnel.c @@ -52,6 +52,7 @@ #include #include #include +#include #include #include diff --git a/net/ipv6/ip6_vti.c b/net/ipv6/ip6_vti.c index 83c055996fbb..09ec4b0ad7dc 100644 --- a/net/ipv6/ip6_vti.c +++ b/net/ipv6/ip6_vti.c @@ -45,6 +45,7 @@ #include #include #include +#include #include #define IP6_VTI_HASH_SIZE_SHIFT 5 diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c index 6f04703fe638..9a0f32acb750 100644 --- a/net/ipv6/sit.c +++ b/net/ipv6/sit.c @@ -51,6 +51,7 @@ #include #include #include +#include #include /* diff --git a/net/l2tp/l2tp_eth.c b/net/l2tp/l2tp_eth.c index e83691073496..cf0b66f4fb29 100644 --- a/net/l2tp/l2tp_eth.c +++ b/net/l2tp/l2tp_eth.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #include #include diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index f5101c2ffc66..abace7665cfe 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -26,6 +26,7 @@ #include #include +#include #include #include #include diff --git a/net/xdp/xsk.c b/net/xdp/xsk.c index f864e5d70b40..e5d104ce7b82 100644 --- a/net/xdp/xsk.c +++ b/net/xdp/xsk.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #include diff --git a/net/xdp/xsk_buff_pool.c b/net/xdp/xsk_buff_pool.c index 0e6ca568fdee..14716ad3d7bc 100644 --- a/net/xdp/xsk_buff_pool.c +++ b/net/xdp/xsk_buff_pool.c @@ -1,6 +1,7 @@ // SPDX-License-Identifier: GPL-2.0 #include +#include #include #include #include -- cgit v1.2.3 From f6f425f3d251c059d1251edd4f37024290d3efca Mon Sep 17 00:00:00 2001 From: Chiara Meiohas Date: Wed, 26 Feb 2025 15:01:05 +0200 Subject: net/mlx5: Add RDMA_CTRL HW capabilities Add RDMA_CTRL UCTX capabilities and add the RDMA_CTRL general object type in hca_cap_2. Reviewed-by: Moshe Shemesh Signed-off-by: Chiara Meiohas Link: https://patch.msgid.link/ef7eb24be9a6f247ab52e8b4480350072e5182f5.1740574103.git.leon@kernel.org Signed-off-by: Leon Romanovsky --- include/linux/mlx5/mlx5_ifc.h | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index cc2875e843f7..3b3d88ffcacc 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -1570,6 +1570,8 @@ enum { enum { MLX5_UCTX_CAP_RAW_TX = 1UL << 0, MLX5_UCTX_CAP_INTERNAL_DEV_RES = 1UL << 1, + MLX5_UCTX_CAP_RDMA_CTRL = 1UL << 3, + MLX5_UCTX_CAP_RDMA_CTRL_OTHER_VHCA = 1UL << 4, }; #define MLX5_FC_BULK_SIZE_FACTOR 128 @@ -2140,7 +2142,8 @@ struct mlx5_ifc_cmd_hca_cap_2_bits { u8 log_min_mkey_entity_size[0x5]; u8 reserved_at_1b0[0x10]; - u8 reserved_at_1c0[0x60]; + u8 general_obj_types_127_64[0x40]; + u8 reserved_at_200[0x20]; u8 reserved_at_220[0x1]; u8 sw_vhca_id_valid[0x1]; @@ -12494,6 +12497,10 @@ enum { MLX5_HCA_CAP_GENERAL_OBJECT_TYPES_FLOW_METER_ASO = BIT_ULL(0x24), }; +enum { + MLX5_HCA_CAP_2_GENERAL_OBJECT_TYPES_RDMA_CTRL = BIT_ULL(0x13), +}; + enum { MLX5_GENERAL_OBJECT_TYPES_ENCRYPTION_KEY = 0xc, MLX5_GENERAL_OBJECT_TYPES_IPSEC = 0x13, @@ -12501,6 +12508,7 @@ enum { MLX5_GENERAL_OBJECT_TYPES_FLOW_METER_ASO = 0x24, MLX5_GENERAL_OBJECT_TYPES_MACSEC = 0x27, MLX5_GENERAL_OBJECT_TYPES_INT_KEK = 0x47, + MLX5_GENERAL_OBJECT_TYPES_RDMA_CTRL = 0x53, MLX5_GENERAL_OBJECT_TYPES_FLOW_TABLE_ALIAS = 0xff15, }; -- cgit v1.2.3 From 0a34fad1bed45ff2245ab8c315bc3d4c6471af46 Mon Sep 17 00:00:00 2001 From: Chiara Meiohas Date: Wed, 26 Feb 2025 15:01:06 +0200 Subject: net/mlx5: Allow the throttle mechanism to be more dynamic Previously, throttle commands were identified and limited based on opcode. These commands were limited to half the command slots using a semaphore, and callback commands checked the opcode to determine semaphore release. To allow exceptions, we introduce a variable to indicate when the throttle lock is held. This allows scenarios where throttle commands are not limited. Callback functions use this variable to determine if the throttle semaphore needs to be released. This patch contains no functional changes. It's a preparation for the next patch. Signed-off-by: Chiara Meiohas Reviewed-by: Tariq Toukan Reviewed-by: Moshe Shemesh Link: https://patch.msgid.link/055d975edeb816ac4c0fd1e665c6157d11947d26.1740574103.git.leon@kernel.org Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/cmd.c | 43 +++++++++++++++++++-------- include/linux/mlx5/driver.h | 1 + 2 files changed, 32 insertions(+), 12 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c index e733b81e18a2..19c0c15c7e08 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c @@ -1881,7 +1881,7 @@ static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out, { struct mlx5_cmd_msg *inb, *outb; u16 opcode = in_to_opcode(in); - bool throttle_op; + bool throttle_locked = false; int pages_queue; gfp_t gfp; u8 token; @@ -1890,12 +1890,12 @@ static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out, if (mlx5_cmd_is_down(dev) || !opcode_allowed(&dev->cmd, opcode)) return -ENXIO; - throttle_op = mlx5_cmd_is_throttle_opcode(opcode); - if (throttle_op) { - if (callback) { - if (down_trylock(&dev->cmd.vars.throttle_sem)) - return -EBUSY; - } else { + if (!callback) { + /* The semaphore is already held for callback commands. It was + * acquired in mlx5_cmd_exec_cb() + */ + if (mlx5_cmd_is_throttle_opcode(opcode)) { + throttle_locked = true; down(&dev->cmd.vars.throttle_sem); } } @@ -1941,7 +1941,7 @@ out_out: out_in: free_msg(dev, inb); out_up: - if (throttle_op) + if (throttle_locked) up(&dev->cmd.vars.throttle_sem); return err; } @@ -2104,17 +2104,17 @@ static void mlx5_cmd_exec_cb_handler(int status, void *_work) struct mlx5_async_work *work = _work; struct mlx5_async_ctx *ctx; struct mlx5_core_dev *dev; - u16 opcode; + bool throttle_locked; ctx = work->ctx; dev = ctx->dev; - opcode = work->opcode; + throttle_locked = work->throttle_locked; status = cmd_status_err(dev, status, work->opcode, work->op_mod, work->out); work->user_callback(status, work); /* Can't access "work" from this point on. It could have been freed in * the callback. */ - if (mlx5_cmd_is_throttle_opcode(opcode)) + if (throttle_locked) up(&dev->cmd.vars.throttle_sem); if (atomic_dec_and_test(&ctx->num_inflight)) complete(&ctx->inflight_done); @@ -2131,11 +2131,30 @@ int mlx5_cmd_exec_cb(struct mlx5_async_ctx *ctx, void *in, int in_size, work->opcode = in_to_opcode(in); work->op_mod = MLX5_GET(mbox_in, in, op_mod); work->out = out; + work->throttle_locked = false; if (WARN_ON(!atomic_inc_not_zero(&ctx->num_inflight))) return -EIO; + + if (mlx5_cmd_is_throttle_opcode(in_to_opcode(in))) { + if (down_trylock(&ctx->dev->cmd.vars.throttle_sem)) { + ret = -EBUSY; + goto dec_num_inflight; + } + work->throttle_locked = true; + } + ret = cmd_exec(ctx->dev, in, in_size, out, out_size, mlx5_cmd_exec_cb_handler, work, false); - if (ret && atomic_dec_and_test(&ctx->num_inflight)) + if (ret) + goto sem_up; + + return 0; + +sem_up: + if (work->throttle_locked) + up(&ctx->dev->cmd.vars.throttle_sem); +dec_num_inflight: + if (atomic_dec_and_test(&ctx->num_inflight)) complete(&ctx->inflight_done); return ret; diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index af86097641b0..876d6b03a87a 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -989,6 +989,7 @@ struct mlx5_async_work { mlx5_async_cbk_t user_callback; u16 opcode; /* cmd opcode */ u16 op_mod; /* cmd op_mod */ + u8 throttle_locked:1; void *out; /* pointer to the cmd output buffer */ }; -- cgit v1.2.3 From f9deed0980fe29c57488f395528fe3d923c91b1f Mon Sep 17 00:00:00 2001 From: Chiara Meiohas Date: Wed, 26 Feb 2025 15:01:07 +0200 Subject: net/mlx5: Limit non-privileged commands Limit non-privileged UID commands to half of the available command slots when privileged UIDs are present. Privileged throttle commands will not be limited. Use an xarray to store privileged UIDs. Add insert and remove functions for privileged UIDs management. Non-user commands (with uid 0) are not limited. Signed-off-by: Chiara Meiohas Reviewed-by: Moshe Shemesh Reviewed-by: Tariq Toukan Link: https://patch.msgid.link/d2f3dd9a0dbad3c9f2b4bb0723837995e4e06de2.1740574103.git.leon@kernel.org Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/cmd.c | 85 ++++++++++++++++++++++++--- include/linux/mlx5/driver.h | 5 ++ 2 files changed, 82 insertions(+), 8 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c index 19c0c15c7e08..e53dbdc0a7a1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c @@ -94,6 +94,11 @@ static u16 in_to_opcode(void *in) return MLX5_GET(mbox_in, in, opcode); } +static u16 in_to_uid(void *in) +{ + return MLX5_GET(mbox_in, in, uid); +} + /* Returns true for opcodes that might be triggered very frequently and throttle * the command interface. Limit their command slots usage. */ @@ -823,7 +828,7 @@ static void cmd_status_print(struct mlx5_core_dev *dev, void *in, void *out) opcode = in_to_opcode(in); op_mod = MLX5_GET(mbox_in, in, op_mod); - uid = MLX5_GET(mbox_in, in, uid); + uid = in_to_uid(in); status = MLX5_GET(mbox_out, out, status); if (!uid && opcode != MLX5_CMD_OP_DESTROY_MKEY && @@ -1871,6 +1876,17 @@ static int is_manage_pages(void *in) return in_to_opcode(in) == MLX5_CMD_OP_MANAGE_PAGES; } +static bool mlx5_has_privileged_uid(struct mlx5_core_dev *dev) +{ + return !xa_empty(&dev->cmd.vars.privileged_uids); +} + +static bool mlx5_cmd_is_privileged_uid(struct mlx5_core_dev *dev, + u16 uid) +{ + return !!xa_load(&dev->cmd.vars.privileged_uids, uid); +} + /* Notes: * 1. Callback functions may not sleep * 2. Page queue commands do not support asynchrous completion @@ -1882,6 +1898,8 @@ static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out, struct mlx5_cmd_msg *inb, *outb; u16 opcode = in_to_opcode(in); bool throttle_locked = false; + bool unpriv_locked = false; + u16 uid = in_to_uid(in); int pages_queue; gfp_t gfp; u8 token; @@ -1894,7 +1912,12 @@ static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out, /* The semaphore is already held for callback commands. It was * acquired in mlx5_cmd_exec_cb() */ - if (mlx5_cmd_is_throttle_opcode(opcode)) { + if (uid && mlx5_has_privileged_uid(dev)) { + if (!mlx5_cmd_is_privileged_uid(dev, uid)) { + unpriv_locked = true; + down(&dev->cmd.vars.unprivileged_sem); + } + } else if (mlx5_cmd_is_throttle_opcode(opcode)) { throttle_locked = true; down(&dev->cmd.vars.throttle_sem); } @@ -1943,6 +1966,9 @@ out_in: out_up: if (throttle_locked) up(&dev->cmd.vars.throttle_sem); + if (unpriv_locked) + up(&dev->cmd.vars.unprivileged_sem); + return err; } @@ -2105,10 +2131,12 @@ static void mlx5_cmd_exec_cb_handler(int status, void *_work) struct mlx5_async_ctx *ctx; struct mlx5_core_dev *dev; bool throttle_locked; + bool unpriv_locked; ctx = work->ctx; dev = ctx->dev; throttle_locked = work->throttle_locked; + unpriv_locked = work->unpriv_locked; status = cmd_status_err(dev, status, work->opcode, work->op_mod, work->out); work->user_callback(status, work); /* Can't access "work" from this point on. It could have been freed in @@ -2116,6 +2144,8 @@ static void mlx5_cmd_exec_cb_handler(int status, void *_work) */ if (throttle_locked) up(&dev->cmd.vars.throttle_sem); + if (unpriv_locked) + up(&dev->cmd.vars.unprivileged_sem); if (atomic_dec_and_test(&ctx->num_inflight)) complete(&ctx->inflight_done); } @@ -2124,6 +2154,8 @@ int mlx5_cmd_exec_cb(struct mlx5_async_ctx *ctx, void *in, int in_size, void *out, int out_size, mlx5_async_cbk_t callback, struct mlx5_async_work *work) { + struct mlx5_core_dev *dev = ctx->dev; + u16 uid; int ret; work->ctx = ctx; @@ -2132,18 +2164,29 @@ int mlx5_cmd_exec_cb(struct mlx5_async_ctx *ctx, void *in, int in_size, work->op_mod = MLX5_GET(mbox_in, in, op_mod); work->out = out; work->throttle_locked = false; + work->unpriv_locked = false; + uid = in_to_uid(in); + if (WARN_ON(!atomic_inc_not_zero(&ctx->num_inflight))) return -EIO; - if (mlx5_cmd_is_throttle_opcode(in_to_opcode(in))) { - if (down_trylock(&ctx->dev->cmd.vars.throttle_sem)) { + if (uid && mlx5_has_privileged_uid(dev)) { + if (!mlx5_cmd_is_privileged_uid(dev, uid)) { + if (down_trylock(&dev->cmd.vars.unprivileged_sem)) { + ret = -EBUSY; + goto dec_num_inflight; + } + work->unpriv_locked = true; + } + } else if (mlx5_cmd_is_throttle_opcode(in_to_opcode(in))) { + if (down_trylock(&dev->cmd.vars.throttle_sem)) { ret = -EBUSY; goto dec_num_inflight; } work->throttle_locked = true; } - ret = cmd_exec(ctx->dev, in, in_size, out, out_size, + ret = cmd_exec(dev, in, in_size, out, out_size, mlx5_cmd_exec_cb_handler, work, false); if (ret) goto sem_up; @@ -2152,7 +2195,9 @@ int mlx5_cmd_exec_cb(struct mlx5_async_ctx *ctx, void *in, int in_size, sem_up: if (work->throttle_locked) - up(&ctx->dev->cmd.vars.throttle_sem); + up(&dev->cmd.vars.throttle_sem); + if (work->unpriv_locked) + up(&dev->cmd.vars.unprivileged_sem); dec_num_inflight: if (atomic_dec_and_test(&ctx->num_inflight)) complete(&ctx->inflight_done); @@ -2390,10 +2435,16 @@ int mlx5_cmd_enable(struct mlx5_core_dev *dev) sema_init(&cmd->vars.sem, cmd->vars.max_reg_cmds); sema_init(&cmd->vars.pages_sem, 1); sema_init(&cmd->vars.throttle_sem, DIV_ROUND_UP(cmd->vars.max_reg_cmds, 2)); + sema_init(&cmd->vars.unprivileged_sem, + DIV_ROUND_UP(cmd->vars.max_reg_cmds, 2)); + + xa_init(&cmd->vars.privileged_uids); cmd->pool = dma_pool_create("mlx5_cmd", mlx5_core_dma_dev(dev), size, align, 0); - if (!cmd->pool) - return -ENOMEM; + if (!cmd->pool) { + err = -ENOMEM; + goto err_destroy_xa; + } err = alloc_cmd_page(dev, cmd); if (err) @@ -2427,6 +2478,8 @@ err_cmd_page: free_cmd_page(dev, cmd); err_free_pool: dma_pool_destroy(cmd->pool); +err_destroy_xa: + xa_destroy(&dev->cmd.vars.privileged_uids); return err; } @@ -2439,6 +2492,7 @@ void mlx5_cmd_disable(struct mlx5_core_dev *dev) destroy_msg_cache(dev); free_cmd_page(dev, cmd); dma_pool_destroy(cmd->pool); + xa_destroy(&dev->cmd.vars.privileged_uids); } void mlx5_cmd_set_state(struct mlx5_core_dev *dev, @@ -2446,3 +2500,18 @@ void mlx5_cmd_set_state(struct mlx5_core_dev *dev, { dev->cmd.state = cmdif_state; } + +int mlx5_cmd_add_privileged_uid(struct mlx5_core_dev *dev, u16 uid) +{ + return xa_insert(&dev->cmd.vars.privileged_uids, uid, + xa_mk_value(uid), GFP_KERNEL); +} +EXPORT_SYMBOL(mlx5_cmd_add_privileged_uid); + +void mlx5_cmd_remove_privileged_uid(struct mlx5_core_dev *dev, u16 uid) +{ + void *data = xa_erase(&dev->cmd.vars.privileged_uids, uid); + + WARN(!data, "Privileged UID %u does not exist\n", uid); +} +EXPORT_SYMBOL(mlx5_cmd_remove_privileged_uid); diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 876d6b03a87a..4f593a61220d 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -305,6 +305,8 @@ struct mlx5_cmd { struct semaphore sem; struct semaphore pages_sem; struct semaphore throttle_sem; + struct semaphore unprivileged_sem; + struct xarray privileged_uids; } vars; enum mlx5_cmdif_state state; void *cmd_alloc_buf; @@ -990,6 +992,7 @@ struct mlx5_async_work { u16 opcode; /* cmd opcode */ u16 op_mod; /* cmd op_mod */ u8 throttle_locked:1; + u8 unpriv_locked:1; void *out; /* pointer to the cmd output buffer */ }; @@ -1020,6 +1023,8 @@ int mlx5_cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out, int mlx5_cmd_exec_polling(struct mlx5_core_dev *dev, void *in, int in_size, void *out, int out_size); bool mlx5_cmd_is_down(struct mlx5_core_dev *dev); +int mlx5_cmd_add_privileged_uid(struct mlx5_core_dev *dev, u16 uid); +void mlx5_cmd_remove_privileged_uid(struct mlx5_core_dev *dev, u16 uid); void mlx5_core_uplink_netdev_set(struct mlx5_core_dev *mdev, struct net_device *netdev); void mlx5_core_uplink_netdev_event_replay(struct mlx5_core_dev *mdev); -- cgit v1.2.3 From ab7d228c7e0d0efcac52b81f8514b43985747dc6 Mon Sep 17 00:00:00 2001 From: Patrisious Haddad Date: Wed, 26 Feb 2025 15:01:08 +0200 Subject: net/mlx5: Query ADV_RDMA capabilities Query ADV_RDMA capabilities which provide information for advanced RDMA related features. Signed-off-by: Patrisious Haddad Reviewed-by: Mark Bloch Link: https://patch.msgid.link/e3e6ede03ea31cd201078dcdd4e407608e4a5a87.1740574103.git.leon@kernel.org Signed-off-by: Leon Romanovsky --- drivers/net/ethernet/mellanox/mlx5/core/fw.c | 7 +++++ drivers/net/ethernet/mellanox/mlx5/core/main.c | 1 + include/linux/mlx5/device.h | 5 +++ include/linux/mlx5/mlx5_ifc.h | 42 +++++++++++++++++++++++++- 4 files changed, 54 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fw.c b/drivers/net/ethernet/mellanox/mlx5/core/fw.c index b253d1673398..57476487e31f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fw.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fw.c @@ -287,6 +287,13 @@ int mlx5_query_hca_caps(struct mlx5_core_dev *dev) return err; } + if (MLX5_CAP_GEN(dev, adv_rdma)) { + err = mlx5_core_get_caps_mode(dev, MLX5_CAP_ADV_RDMA, + HCA_CAP_OPMOD_GET_CUR); + if (err) + return err; + } + return 0; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index ec956c4bcebd..af0b677393d0 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -1795,6 +1795,7 @@ static const int types[] = { MLX5_CAP_ADV_VIRTUALIZATION, MLX5_CAP_CRYPTO, MLX5_CAP_SHAMPO, + MLX5_CAP_ADV_RDMA, }; static void mlx5_hca_caps_free(struct mlx5_core_dev *dev) diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h index fd37f4e54d76..0ae6d69c5221 100644 --- a/include/linux/mlx5/device.h +++ b/include/linux/mlx5/device.h @@ -1251,6 +1251,7 @@ enum mlx5_cap_type { MLX5_CAP_GENERAL_2 = 0x20, MLX5_CAP_PORT_SELECTION = 0x25, MLX5_CAP_ADV_VIRTUALIZATION = 0x26, + MLX5_CAP_ADV_RDMA = 0x28, /* NUM OF CAP Types */ MLX5_CAP_NUM }; @@ -1384,6 +1385,10 @@ enum mlx5_qcam_feature_groups { MLX5_GET(adv_virtualization_cap, \ mdev->caps.hca[MLX5_CAP_ADV_VIRTUALIZATION]->cur, cap) +#define MLX5_CAP_ADV_RDMA(mdev, cap) \ + MLX5_GET(adv_rdma_cap, \ + mdev->caps.hca[MLX5_CAP_ADV_RDMA]->cur, cap) + #define MLX5_CAP_FLOWTABLE_PORT_SELECTION(mdev, cap) \ MLX5_CAP_PORT_SELECTION(mdev, flow_table_properties_port_selection.cap) diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 3b3d88ffcacc..fea8af42f954 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -1993,7 +1993,9 @@ struct mlx5_ifc_cmd_hca_cap_bits { u8 max_geneve_tlv_options[0x8]; u8 reserved_at_568[0x3]; u8 max_geneve_tlv_option_data_len[0x5]; - u8 reserved_at_570[0x9]; + u8 reserved_at_570[0x1]; + u8 adv_rdma[0x1]; + u8 reserved_at_572[0x7]; u8 adv_virtualization[0x1]; u8 reserved_at_57a[0x6]; @@ -13076,6 +13078,44 @@ struct mlx5_ifc_load_vhca_state_out_bits { u8 reserved_at_40[0x40]; }; +struct mlx5_ifc_adv_rdma_cap_bits { + u8 rdma_transport_manager[0x1]; + u8 rdma_transport_manager_other_eswitch[0x1]; + u8 reserved_at_2[0x1e]; + + u8 rcx_type[0x8]; + u8 reserved_at_28[0x2]; + u8 ps_entry_log_max_value[0x6]; + u8 reserved_at_30[0x6]; + u8 qp_max_ps_num_entry[0xa]; + + u8 mp_max_num_queues[0x8]; + u8 ps_user_context_max_log_size[0x8]; + u8 message_based_qp_and_striding_wq[0x8]; + u8 reserved_at_58[0x8]; + + u8 max_receive_send_message_size_stride[0x10]; + u8 reserved_at_70[0x10]; + + u8 max_receive_send_message_size_byte[0x20]; + + u8 reserved_at_a0[0x160]; + + struct mlx5_ifc_flow_table_prop_layout_bits rdma_transport_rx_flow_table_properties; + + struct mlx5_ifc_flow_table_prop_layout_bits rdma_transport_tx_flow_table_properties; + + struct mlx5_ifc_flow_table_fields_supported_2_bits rdma_transport_rx_ft_field_support_2; + + struct mlx5_ifc_flow_table_fields_supported_2_bits rdma_transport_tx_ft_field_support_2; + + struct mlx5_ifc_flow_table_fields_supported_2_bits rdma_transport_rx_ft_field_bitmask_support_2; + + struct mlx5_ifc_flow_table_fields_supported_2_bits rdma_transport_tx_ft_field_bitmask_support_2; + + u8 reserved_at_800[0x3800]; +}; + struct mlx5_ifc_adv_virtualization_cap_bits { u8 reserved_at_0[0x3]; u8 pg_track_log_max_num[0x5]; -- cgit v1.2.3 From 15b103df80b25025040faa8f35164c2595977bdb Mon Sep 17 00:00:00 2001 From: Patrisious Haddad Date: Wed, 26 Feb 2025 15:01:09 +0200 Subject: net/mlx5: fs, add RDMA TRANSPORT steering domain support Add RX and TX RDMA_TRANSPORT flow table namespace, and the ability to create flow tables in those namespaces. The RDMA_TRANSPORT RX and TX are per vport. Packets will traverse through RDMA_TRANSPORT_RX after RDMA_RX and through RDMA_TRANSPORT_TX before RDMA_TX, ensuring proper control and management. RDMA_TRANSPORT domains are managed by the vport group manager. Signed-off-by: Patrisious Haddad Reviewed-by: Mark Bloch Link: https://patch.msgid.link/a6b550d9859a197eafa804b9a8d76916ca481da9.1740574103.git.leon@kernel.org Signed-off-by: Leon Romanovsky --- .../ethernet/mellanox/mlx5/core/esw/acl/helper.c | 2 +- .../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 6 +- drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c | 2 + drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 178 +++++++++++++++++++-- drivers/net/ethernet/mellanox/mlx5/core/fs_core.h | 12 +- include/linux/mlx5/device.h | 6 + include/linux/mlx5/fs.h | 10 +- 7 files changed, 195 insertions(+), 21 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/helper.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/helper.c index d599e50af346..3ce455c2535c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/helper.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/helper.c @@ -27,7 +27,7 @@ esw_acl_table_create(struct mlx5_eswitch *esw, struct mlx5_vport *vport, int ns, esw_debug(dev, "Create vport[%d] %s ACL table\n", vport_num, ns == MLX5_FLOW_NAMESPACE_ESW_INGRESS ? "ingress" : "egress"); - root_ns = mlx5_get_flow_vport_acl_namespace(dev, ns, vport->index); + root_ns = mlx5_get_flow_vport_namespace(dev, ns, vport->index); if (!root_ns) { esw_warn(dev, "Failed to get E-Switch root namespace for vport (%d)\n", vport_num); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 20cc01ceee8a..1ee5791573d8 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -2828,9 +2828,9 @@ static int esw_set_master_egress_rule(struct mlx5_core_dev *master, if (IS_ERR(vport)) return PTR_ERR(vport); - egress_ns = mlx5_get_flow_vport_acl_namespace(master, - MLX5_FLOW_NAMESPACE_ESW_EGRESS, - vport->index); + egress_ns = mlx5_get_flow_vport_namespace(master, + MLX5_FLOW_NAMESPACE_ESW_EGRESS, + vport->index); if (!egress_ns) return -EINVAL; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c index ae20c061e0fb..a47c29571f64 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c @@ -1142,6 +1142,8 @@ const struct mlx5_flow_cmds *mlx5_fs_cmd_get_default(enum fs_flow_table_type typ case FS_FT_RDMA_RX: case FS_FT_RDMA_TX: case FS_FT_PORT_SEL: + case FS_FT_RDMA_TRANSPORT_RX: + case FS_FT_RDMA_TRANSPORT_TX: return mlx5_fs_cmd_get_fw_cmds(); default: return mlx5_fs_cmd_get_stub_cmds(); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c index 22dc23d991d2..6163bc98d94a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c @@ -1456,7 +1456,7 @@ mlx5_create_auto_grouped_flow_table(struct mlx5_flow_namespace *ns, struct mlx5_flow_table *ft; int autogroups_max_fte; - ft = mlx5_create_flow_table(ns, ft_attr); + ft = mlx5_create_vport_flow_table(ns, ft_attr, ft_attr->vport); if (IS_ERR(ft)) return ft; @@ -2764,9 +2764,9 @@ struct mlx5_flow_namespace *mlx5_get_flow_namespace(struct mlx5_core_dev *dev, } EXPORT_SYMBOL(mlx5_get_flow_namespace); -struct mlx5_flow_namespace *mlx5_get_flow_vport_acl_namespace(struct mlx5_core_dev *dev, - enum mlx5_flow_namespace_type type, - int vport) +struct mlx5_flow_namespace * +mlx5_get_flow_vport_namespace(struct mlx5_core_dev *dev, + enum mlx5_flow_namespace_type type, int vport_idx) { struct mlx5_flow_steering *steering = dev->priv.steering; @@ -2775,25 +2775,43 @@ struct mlx5_flow_namespace *mlx5_get_flow_vport_acl_namespace(struct mlx5_core_d switch (type) { case MLX5_FLOW_NAMESPACE_ESW_EGRESS: - if (vport >= steering->esw_egress_acl_vports) + if (vport_idx >= steering->esw_egress_acl_vports) return NULL; if (steering->esw_egress_root_ns && - steering->esw_egress_root_ns[vport]) - return &steering->esw_egress_root_ns[vport]->ns; + steering->esw_egress_root_ns[vport_idx]) + return &steering->esw_egress_root_ns[vport_idx]->ns; else return NULL; case MLX5_FLOW_NAMESPACE_ESW_INGRESS: - if (vport >= steering->esw_ingress_acl_vports) + if (vport_idx >= steering->esw_ingress_acl_vports) return NULL; if (steering->esw_ingress_root_ns && - steering->esw_ingress_root_ns[vport]) - return &steering->esw_ingress_root_ns[vport]->ns; + steering->esw_ingress_root_ns[vport_idx]) + return &steering->esw_ingress_root_ns[vport_idx]->ns; + else + return NULL; + case MLX5_FLOW_NAMESPACE_RDMA_TRANSPORT_RX: + if (vport_idx >= steering->rdma_transport_rx_vports) + return NULL; + if (steering->rdma_transport_rx_root_ns && + steering->rdma_transport_rx_root_ns[vport_idx]) + return &steering->rdma_transport_rx_root_ns[vport_idx]->ns; + else + return NULL; + case MLX5_FLOW_NAMESPACE_RDMA_TRANSPORT_TX: + if (vport_idx >= steering->rdma_transport_tx_vports) + return NULL; + + if (steering->rdma_transport_tx_root_ns && + steering->rdma_transport_tx_root_ns[vport_idx]) + return &steering->rdma_transport_tx_root_ns[vport_idx]->ns; else return NULL; default: return NULL; } } +EXPORT_SYMBOL(mlx5_get_flow_vport_namespace); static struct fs_prio *_fs_create_prio(struct mlx5_flow_namespace *ns, unsigned int prio, @@ -3199,6 +3217,127 @@ out_err: return err; } +static int +init_rdma_transport_rx_root_ns_one(struct mlx5_flow_steering *steering, + int vport_idx) +{ + struct fs_prio *prio; + + steering->rdma_transport_rx_root_ns[vport_idx] = + create_root_ns(steering, FS_FT_RDMA_TRANSPORT_RX); + if (!steering->rdma_transport_rx_root_ns[vport_idx]) + return -ENOMEM; + + /* create 1 prio*/ + prio = fs_create_prio(&steering->rdma_transport_rx_root_ns[vport_idx]->ns, + MLX5_RDMA_TRANSPORT_BYPASS_PRIO, 1); + return PTR_ERR_OR_ZERO(prio); +} + +static int +init_rdma_transport_tx_root_ns_one(struct mlx5_flow_steering *steering, + int vport_idx) +{ + struct fs_prio *prio; + + steering->rdma_transport_tx_root_ns[vport_idx] = + create_root_ns(steering, FS_FT_RDMA_TRANSPORT_TX); + if (!steering->rdma_transport_tx_root_ns[vport_idx]) + return -ENOMEM; + + /* create 1 prio*/ + prio = fs_create_prio(&steering->rdma_transport_tx_root_ns[vport_idx]->ns, + MLX5_RDMA_TRANSPORT_BYPASS_PRIO, 1); + return PTR_ERR_OR_ZERO(prio); +} + +static int init_rdma_transport_rx_root_ns(struct mlx5_flow_steering *steering) +{ + struct mlx5_core_dev *dev = steering->dev; + int total_vports; + int err; + int i; + + /* In case eswitch not supported and working in legacy mode */ + total_vports = mlx5_eswitch_get_total_vports(dev) ?: 1; + + steering->rdma_transport_rx_root_ns = + kcalloc(total_vports, + sizeof(*steering->rdma_transport_rx_root_ns), + GFP_KERNEL); + if (!steering->rdma_transport_rx_root_ns) + return -ENOMEM; + + for (i = 0; i < total_vports; i++) { + err = init_rdma_transport_rx_root_ns_one(steering, i); + if (err) + goto cleanup_root_ns; + } + steering->rdma_transport_rx_vports = total_vports; + return 0; + +cleanup_root_ns: + while (i--) + cleanup_root_ns(steering->rdma_transport_rx_root_ns[i]); + kfree(steering->rdma_transport_rx_root_ns); + steering->rdma_transport_rx_root_ns = NULL; + return err; +} + +static int init_rdma_transport_tx_root_ns(struct mlx5_flow_steering *steering) +{ + struct mlx5_core_dev *dev = steering->dev; + int total_vports; + int err; + int i; + + /* In case eswitch not supported and working in legacy mode */ + total_vports = mlx5_eswitch_get_total_vports(dev) ?: 1; + + steering->rdma_transport_tx_root_ns = + kcalloc(total_vports, + sizeof(*steering->rdma_transport_tx_root_ns), + GFP_KERNEL); + if (!steering->rdma_transport_tx_root_ns) + return -ENOMEM; + + for (i = 0; i < total_vports; i++) { + err = init_rdma_transport_tx_root_ns_one(steering, i); + if (err) + goto cleanup_root_ns; + } + steering->rdma_transport_tx_vports = total_vports; + return 0; + +cleanup_root_ns: + while (i--) + cleanup_root_ns(steering->rdma_transport_tx_root_ns[i]); + kfree(steering->rdma_transport_tx_root_ns); + steering->rdma_transport_tx_root_ns = NULL; + return err; +} + +static void cleanup_rdma_transport_roots_ns(struct mlx5_flow_steering *steering) +{ + int i; + + if (steering->rdma_transport_rx_root_ns) { + for (i = 0; i < steering->rdma_transport_rx_vports; i++) + cleanup_root_ns(steering->rdma_transport_rx_root_ns[i]); + + kfree(steering->rdma_transport_rx_root_ns); + steering->rdma_transport_rx_root_ns = NULL; + } + + if (steering->rdma_transport_tx_root_ns) { + for (i = 0; i < steering->rdma_transport_tx_vports; i++) + cleanup_root_ns(steering->rdma_transport_tx_root_ns[i]); + + kfree(steering->rdma_transport_tx_root_ns); + steering->rdma_transport_tx_root_ns = NULL; + } +} + /* FT and tc chains are stored in the same array so we can re-use the * mlx5_get_fdb_sub_ns() and tc api for FT chains. * When creating a new ns for each chain store it in the first available slot. @@ -3631,6 +3770,7 @@ void mlx5_fs_core_cleanup(struct mlx5_core_dev *dev) cleanup_root_ns(steering->rdma_rx_root_ns); cleanup_root_ns(steering->rdma_tx_root_ns); cleanup_root_ns(steering->egress_root_ns); + cleanup_rdma_transport_roots_ns(steering); devl_params_unregister(priv_to_devlink(dev), mlx5_fs_params, ARRAY_SIZE(mlx5_fs_params)); @@ -3700,6 +3840,18 @@ int mlx5_fs_core_init(struct mlx5_core_dev *dev) goto err; } + if (MLX5_CAP_FLOWTABLE_RDMA_TRANSPORT_RX(dev, ft_support)) { + err = init_rdma_transport_rx_root_ns(steering); + if (err) + goto err; + } + + if (MLX5_CAP_FLOWTABLE_RDMA_TRANSPORT_TX(dev, ft_support)) { + err = init_rdma_transport_tx_root_ns(steering); + if (err) + goto err; + } + return 0; err: @@ -3850,8 +4002,10 @@ mlx5_get_root_namespace(struct mlx5_core_dev *dev, enum mlx5_flow_namespace_type struct mlx5_flow_namespace *ns; if (ns_type == MLX5_FLOW_NAMESPACE_ESW_EGRESS || - ns_type == MLX5_FLOW_NAMESPACE_ESW_INGRESS) - ns = mlx5_get_flow_vport_acl_namespace(dev, ns_type, 0); + ns_type == MLX5_FLOW_NAMESPACE_ESW_INGRESS || + ns_type == MLX5_FLOW_NAMESPACE_RDMA_TRANSPORT_TX || + ns_type == MLX5_FLOW_NAMESPACE_RDMA_TRANSPORT_RX) + ns = mlx5_get_flow_vport_namespace(dev, ns_type, 0); else ns = mlx5_get_flow_namespace(dev, ns_type); if (!ns) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h index 20837e526679..dbe747e93841 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h @@ -115,7 +115,9 @@ enum fs_flow_table_type { FS_FT_PORT_SEL = 0X9, FS_FT_FDB_RX = 0xa, FS_FT_FDB_TX = 0xb, - FS_FT_MAX_TYPE = FS_FT_FDB_TX, + FS_FT_RDMA_TRANSPORT_RX = 0xd, + FS_FT_RDMA_TRANSPORT_TX = 0xe, + FS_FT_MAX_TYPE = FS_FT_RDMA_TRANSPORT_TX, }; enum fs_flow_table_op_mod { @@ -158,6 +160,10 @@ struct mlx5_flow_steering { struct mlx5_flow_root_namespace *port_sel_root_ns; int esw_egress_acl_vports; int esw_ingress_acl_vports; + struct mlx5_flow_root_namespace **rdma_transport_rx_root_ns; + struct mlx5_flow_root_namespace **rdma_transport_tx_root_ns; + int rdma_transport_rx_vports; + int rdma_transport_tx_vports; }; struct fs_node { @@ -434,7 +440,9 @@ struct mlx5_flow_root_namespace *find_root(struct fs_node *node); (type == FS_FT_PORT_SEL) ? MLX5_CAP_FLOWTABLE_PORT_SELECTION(mdev, cap) : \ (type == FS_FT_FDB_RX) ? MLX5_CAP_ESW_FLOWTABLE_FDB(mdev, cap) : \ (type == FS_FT_FDB_TX) ? MLX5_CAP_ESW_FLOWTABLE_FDB(mdev, cap) : \ - (BUILD_BUG_ON_ZERO(FS_FT_FDB_TX != FS_FT_MAX_TYPE))\ + (type == FS_FT_RDMA_TRANSPORT_RX) ? MLX5_CAP_FLOWTABLE_RDMA_TRANSPORT_RX(mdev, cap) : \ + (type == FS_FT_RDMA_TRANSPORT_TX) ? MLX5_CAP_FLOWTABLE_RDMA_TRANSPORT_TX(mdev, cap) : \ + (BUILD_BUG_ON_ZERO(FS_FT_RDMA_TRANSPORT_TX != FS_FT_MAX_TYPE))\ ) #endif diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h index 0ae6d69c5221..8fe56d0362c6 100644 --- a/include/linux/mlx5/device.h +++ b/include/linux/mlx5/device.h @@ -1346,6 +1346,12 @@ enum mlx5_qcam_feature_groups { #define MLX5_CAP_FLOWTABLE_RDMA_TX(mdev, cap) \ MLX5_CAP_FLOWTABLE(mdev, flow_table_properties_nic_transmit_rdma.cap) +#define MLX5_CAP_FLOWTABLE_RDMA_TRANSPORT_RX(mdev, cap) \ + MLX5_CAP_ADV_RDMA(mdev, rdma_transport_rx_flow_table_properties.cap) + +#define MLX5_CAP_FLOWTABLE_RDMA_TRANSPORT_TX(mdev, cap) \ + MLX5_CAP_ADV_RDMA(mdev, rdma_transport_tx_flow_table_properties.cap) + #define MLX5_CAP_ESW_FLOWTABLE(mdev, cap) \ MLX5_GET(flow_table_eswitch_cap, \ mdev->caps.hca[MLX5_CAP_ESWITCH_FLOW_TABLE]->cur, cap) diff --git a/include/linux/mlx5/fs.h b/include/linux/mlx5/fs.h index 01cb72d68c23..fd62b2b1611d 100644 --- a/include/linux/mlx5/fs.h +++ b/include/linux/mlx5/fs.h @@ -40,6 +40,7 @@ #define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v) +#define MLX5_RDMA_TRANSPORT_BYPASS_PRIO 0 #define MLX5_FS_MAX_POOL_SIZE BIT(30) enum mlx5_flow_destination_type { @@ -110,6 +111,8 @@ enum mlx5_flow_namespace_type { MLX5_FLOW_NAMESPACE_RDMA_TX_IPSEC, MLX5_FLOW_NAMESPACE_RDMA_RX_MACSEC, MLX5_FLOW_NAMESPACE_RDMA_TX_MACSEC, + MLX5_FLOW_NAMESPACE_RDMA_TRANSPORT_RX, + MLX5_FLOW_NAMESPACE_RDMA_TRANSPORT_TX, }; enum { @@ -194,9 +197,9 @@ struct mlx5_flow_namespace * mlx5_get_flow_namespace(struct mlx5_core_dev *dev, enum mlx5_flow_namespace_type type); struct mlx5_flow_namespace * -mlx5_get_flow_vport_acl_namespace(struct mlx5_core_dev *dev, - enum mlx5_flow_namespace_type type, - int vport); +mlx5_get_flow_vport_namespace(struct mlx5_core_dev *dev, + enum mlx5_flow_namespace_type type, + int vport_idx); struct mlx5_flow_table_attr { int prio; @@ -204,6 +207,7 @@ struct mlx5_flow_table_attr { u32 level; u32 flags; u16 uid; + u16 vport; struct mlx5_flow_table *next_ft; struct { -- cgit v1.2.3 From f550694e88b7b13b647777f889e03e544d9db60c Mon Sep 17 00:00:00 2001 From: Yael Chemla Date: Sun, 9 Mar 2025 20:41:37 +0200 Subject: net/mlx5: Add IFC bits for PPCNT recovery counters group Add recovery counters group layout of PPCNT (Ports Performance Counters Register). This group counts recovery events per link. Also add the corresponding bit in PCAM to indicate this group is supported. Signed-off-by: Yael Chemla Reviewed-by: Cosmin Ratiu Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/1741545697-23041-1-git-send-email-tariqt@nvidia.com Signed-off-by: Leon Romanovsky --- include/linux/mlx5/device.h | 1 + include/linux/mlx5/mlx5_ifc.h | 11 ++++++++++- 2 files changed, 11 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h index 8fe56d0362c6..904804e995aa 100644 --- a/include/linux/mlx5/device.h +++ b/include/linux/mlx5/device.h @@ -1517,6 +1517,7 @@ enum { MLX5_PHYSICAL_LAYER_COUNTERS_GROUP = 0x12, MLX5_PER_TRAFFIC_CLASS_CONGESTION_GROUP = 0x13, MLX5_PHYSICAL_LAYER_STATISTICAL_GROUP = 0x16, + MLX5_PHYSICAL_LAYER_RECOVERY_GROUP = 0x1a, MLX5_INFINIBAND_PORT_COUNTERS_GROUP = 0x20, MLX5_INFINIBAND_EXTENDED_PORT_COUNTERS_GROUP = 0x21, }; diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index fea8af42f954..2c09df4ee574 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -2645,6 +2645,12 @@ struct mlx5_ifc_field_select_802_1qau_rp_bits { u8 field_select_8021qaurp[0x20]; }; +struct mlx5_ifc_phys_layer_recovery_cntrs_bits { + u8 total_successful_recovery_events[0x20]; + + u8 reserved_at_20[0x7a0]; +}; + struct mlx5_ifc_phys_layer_cntrs_bits { u8 time_since_last_clear_high[0x20]; @@ -4846,6 +4852,7 @@ union mlx5_ifc_eth_cntrs_grp_data_layout_auto_bits { struct mlx5_ifc_ib_ext_port_cntrs_grp_data_layout_bits ib_ext_port_cntrs_grp_data_layout; struct mlx5_ifc_phys_layer_cntrs_bits phys_layer_cntrs; struct mlx5_ifc_phys_layer_statistical_cntrs_bits phys_layer_statistical_cntrs; + struct mlx5_ifc_phys_layer_recovery_cntrs_bits phys_layer_recovery_cntrs; u8 reserved_at_0[0x7c0]; }; @@ -10584,7 +10591,9 @@ struct mlx5_ifc_mtutc_reg_bits { }; struct mlx5_ifc_pcam_enhanced_features_bits { - u8 reserved_at_0[0x1d]; + u8 reserved_at_0[0x10]; + u8 ppcnt_recovery_counters[0x1]; + u8 reserved_at_11[0xc]; u8 fec_200G_per_lane_in_pplm[0x1]; u8 reserved_at_1e[0x2a]; u8 fec_100G_per_lane_in_pplm[0x1]; -- cgit v1.2.3 From a18dfa9925b9ef6107ea3aa5814ca3c704d34a8a Mon Sep 17 00:00:00 2001 From: Willem de Bruijn Date: Thu, 6 Mar 2025 22:34:09 -0500 Subject: ipv6: save dontfrag in cork When spanning datagram construction over multiple send calls using MSG_MORE, per datagram settings are configured on the first send. That is when ip(6)_setup_cork stores these settings for subsequent use in __ip(6)_append_data and others. The only flag that escaped this was dontfrag. As a result, a datagram could be constructed with df=0 on the first sendmsg, but df=1 on a next. Which is what cmsg_ip.sh does in an upcoming MSG_MORE test in the "diff" scenario. Changing datagram conditions in the middle of constructing an skb makes this already complex code path even more convoluted. It is here unintentional. Bring this flag in line with expected sockopt/cmsg behavior. And stop passing ipc6 to __ip6_append_data, to avoid such issues in the future. This is already the case for __ip_append_data. inet6_cork had a 6 byte hole, so the 1B flag has no impact. Signed-off-by: Willem de Bruijn Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/20250307033620.411611-3-willemdebruijn.kernel@gmail.com Signed-off-by: Jakub Kicinski --- include/linux/ipv6.h | 1 + net/ipv6/ip6_output.c | 9 +++++---- 2 files changed, 6 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/include/linux/ipv6.h b/include/linux/ipv6.h index a6e2aadbb91b..5aeeed22f35b 100644 --- a/include/linux/ipv6.h +++ b/include/linux/ipv6.h @@ -207,6 +207,7 @@ struct inet6_cork { struct ipv6_txoptions *opt; u8 hop_limit; u8 tclass; + u8 dontfrag:1; }; /* struct ipv6_pinfo - ipv6 private area */ diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index d91da522c34e..581bc6289081 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -1386,6 +1386,7 @@ static int ip6_setup_cork(struct sock *sk, struct inet_cork_full *cork, } v6_cork->hop_limit = ipc6->hlimit; v6_cork->tclass = ipc6->tclass; + v6_cork->dontfrag = ipc6->dontfrag; if (rt->dst.flags & DST_XFRM_TUNNEL) mtu = READ_ONCE(np->pmtudisc) >= IPV6_PMTUDISC_PROBE ? READ_ONCE(rt->dst.dev->mtu) : dst_mtu(&rt->dst); @@ -1421,7 +1422,7 @@ static int __ip6_append_data(struct sock *sk, int getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb), void *from, size_t length, int transhdrlen, - unsigned int flags, struct ipcm6_cookie *ipc6) + unsigned int flags) { struct sk_buff *skb, *skb_prev = NULL; struct inet_cork *cork = &cork_full->base; @@ -1475,7 +1476,7 @@ static int __ip6_append_data(struct sock *sk, if (headersize + transhdrlen > mtu) goto emsgsize; - if (cork->length + length > mtu - headersize && ipc6->dontfrag && + if (cork->length + length > mtu - headersize && v6_cork->dontfrag && (sk->sk_protocol == IPPROTO_UDP || sk->sk_protocol == IPPROTO_ICMPV6 || sk->sk_protocol == IPPROTO_RAW)) { @@ -1855,7 +1856,7 @@ int ip6_append_data(struct sock *sk, return __ip6_append_data(sk, &sk->sk_write_queue, &inet->cork, &np->cork, sk_page_frag(sk), getfrag, - from, length, transhdrlen, flags, ipc6); + from, length, transhdrlen, flags); } EXPORT_SYMBOL_GPL(ip6_append_data); @@ -2058,7 +2059,7 @@ struct sk_buff *ip6_make_skb(struct sock *sk, err = __ip6_append_data(sk, &queue, cork, &v6_cork, ¤t->task_frag, getfrag, from, length + exthdrlen, transhdrlen + exthdrlen, - flags, ipc6); + flags); if (err) { __ip6_flush_pending_frames(sk, &queue, cork, &v6_cork); return ERR_PTR(err); -- cgit v1.2.3 From 0a13c1e0a449917b29c45d90701eededa69c99d3 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Fri, 7 Mar 2025 20:47:26 -0800 Subject: net: revert to lockless TC_SETUP_BLOCK and TC_SETUP_FT There is a couple of places from which we can arrive to ndo_setup_tc with TC_SETUP_BLOCK/TC_SETUP_FT: - netlink - netlink notifier - netdev notifier Locking netdev too deep in this call chain seems to be problematic (especially assuming some/all of the call_netdevice_notifiers NETDEV_UNREGISTER) might soon be running with the instance lock). Revert to lockless ndo_setup_tc for TC_SETUP_BLOCK/TC_SETUP_FT. NFT framework already takes care of most of the locking. Document the assumptions. ndo_setup_tc TC_SETUP_BLOCK nft_block_offload_cmd nft_chain_offload_cmd nft_flow_block_chain nft_flow_offload_chain nft_flow_rule_offload_abort nft_flow_rule_offload_commit nft_flow_rule_offload_commit nf_tables_commit nfnetlink_rcv_batch nfnetlink_rcv_skb_batch nfnetlink_rcv nft_offload_netdev_event NETDEV_UNREGISTER notifier ndo_setup_tc TC_SETUP_FT nf_flow_table_offload_cmd nf_flow_table_offload_setup nft_unregister_flowtable_hook nft_register_flowtable_net_hooks nft_flowtable_update nf_tables_newflowtable nfnetlink_rcv_batch (.call NFNL_CB_BATCH) nft_flowtable_update nf_tables_newflowtable nft_flowtable_event nf_tables_flowtable_event NETDEV_UNREGISTER notifier __nft_unregister_flowtable_net_hooks nft_unregister_flowtable_net_hooks nf_tables_commit nfnetlink_rcv_batch (.call NFNL_CB_BATCH) __nf_tables_abort nf_tables_abort nfnetlink_rcv_batch __nft_release_hook __nft_release_hooks nf_tables_pre_exit_net -> module unload nft_rcv_nl_event netlink_register_notifier (oh boy) nft_register_flowtable_net_hooks nft_flowtable_update nf_tables_newflowtable nf_tables_newflowtable Fixes: c4f0f30b424e ("net: hold netdev instance lock during nft ndo_setup_tc") Signed-off-by: Stanislav Fomichev Reported-by: syzbot+0afb4bcf91e5a1afdcad@syzkaller.appspotmail.com Reviewed-by: Simon Horman Link: https://patch.msgid.link/20250308044726.1193222-1-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- Documentation/networking/netdevices.rst | 6 ++++++ include/linux/netdevice.h | 2 -- net/core/dev.c | 19 ------------------- net/netfilter/nf_flow_table_offload.c | 2 +- net/netfilter/nf_tables_offload.c | 2 +- 5 files changed, 8 insertions(+), 23 deletions(-) (limited to 'include') diff --git a/Documentation/networking/netdevices.rst b/Documentation/networking/netdevices.rst index d235a7364893..ebb868f50ac2 100644 --- a/Documentation/networking/netdevices.rst +++ b/Documentation/networking/netdevices.rst @@ -290,6 +290,12 @@ ndo_set_rx_mode: Synchronization: netif_addr_lock spinlock. Context: BHs disabled +ndo_setup_tc: + ``TC_SETUP_BLOCK`` and ``TC_SETUP_FT`` are running under NFT locks + (i.e. no ``rtnl_lock`` and no device instance lock). The rest of + ``tc_setup_type`` types run under netdev instance lock if the driver + implements queue management or shaper API. + Most ndo callbacks not specified in the list above are running under ``rtnl_lock``. In addition, netdev instance lock is taken as well if the driver implements queue management or shaper API. diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 9a297757df7e..0dbfe069a6e3 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -3316,8 +3316,6 @@ int dev_open(struct net_device *dev, struct netlink_ext_ack *extack); void netif_close(struct net_device *dev); void dev_close(struct net_device *dev); void dev_close_many(struct list_head *head, bool unlink); -int dev_setup_tc(struct net_device *dev, enum tc_setup_type type, - void *type_data); void netif_disable_lro(struct net_device *dev); void dev_disable_lro(struct net_device *dev); int dev_loopback_xmit(struct net *net, struct sock *sk, struct sk_buff *newskb); diff --git a/net/core/dev.c b/net/core/dev.c index 1cb134ff7327..812134b71f05 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1761,25 +1761,6 @@ void netif_close(struct net_device *dev) } } -int dev_setup_tc(struct net_device *dev, enum tc_setup_type type, - void *type_data) -{ - const struct net_device_ops *ops = dev->netdev_ops; - int ret; - - ASSERT_RTNL(); - - if (!ops->ndo_setup_tc) - return -EOPNOTSUPP; - - netdev_lock_ops(dev); - ret = ops->ndo_setup_tc(dev, type, type_data); - netdev_unlock_ops(dev); - - return ret; -} -EXPORT_SYMBOL(dev_setup_tc); - void netif_disable_lro(struct net_device *dev) { struct net_device *lower_dev; diff --git a/net/netfilter/nf_flow_table_offload.c b/net/netfilter/nf_flow_table_offload.c index 0ec4abded10d..e06bc36f49fe 100644 --- a/net/netfilter/nf_flow_table_offload.c +++ b/net/netfilter/nf_flow_table_offload.c @@ -1175,7 +1175,7 @@ static int nf_flow_table_offload_cmd(struct flow_block_offload *bo, nf_flow_table_block_offload_init(bo, dev_net(dev), cmd, flowtable, extack); down_write(&flowtable->flow_block_lock); - err = dev_setup_tc(dev, TC_SETUP_FT, bo); + err = dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_FT, bo); up_write(&flowtable->flow_block_lock); if (err < 0) return err; diff --git a/net/netfilter/nf_tables_offload.c b/net/netfilter/nf_tables_offload.c index b761899c143c..64675f1c7f29 100644 --- a/net/netfilter/nf_tables_offload.c +++ b/net/netfilter/nf_tables_offload.c @@ -390,7 +390,7 @@ static int nft_block_offload_cmd(struct nft_base_chain *chain, nft_flow_block_offload_init(&bo, dev_net(dev), cmd, chain, &extack); - err = dev_setup_tc(dev, TC_SETUP_BLOCK, &bo); + err = dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_BLOCK, &bo); if (err < 0) return err; -- cgit v1.2.3 From b6b67141d6f1f736b17aca87e5ecb43b7c3a8205 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Tue, 11 Mar 2025 07:40:24 -0700 Subject: net: create netdev_nl_sock to wrap bindings list No functional changes. Next patches will add more granular locking to netdev_nl_sock. Signed-off-by: Stanislav Fomichev Reviewed-by: Mina Almasry Link: https://patch.msgid.link/20250311144026.4154277-2-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- Documentation/netlink/specs/netdev.yaml | 4 ++-- include/net/netdev_netlink.h | 11 +++++++++++ net/core/netdev-genl-gen.c | 4 ++-- net/core/netdev-genl-gen.h | 6 +++--- net/core/netdev-genl.c | 19 +++++++++---------- 5 files changed, 27 insertions(+), 17 deletions(-) create mode 100644 include/net/netdev_netlink.h (limited to 'include') diff --git a/Documentation/netlink/specs/netdev.yaml b/Documentation/netlink/specs/netdev.yaml index 36f1152bfac3..f5e0750ab71d 100644 --- a/Documentation/netlink/specs/netdev.yaml +++ b/Documentation/netlink/specs/netdev.yaml @@ -745,8 +745,8 @@ operations: - irq-suspend-timeout kernel-family: - headers: [ "linux/list.h"] - sock-priv: struct list_head + headers: [ "net/netdev_netlink.h"] + sock-priv: struct netdev_nl_sock mcast-groups: list: diff --git a/include/net/netdev_netlink.h b/include/net/netdev_netlink.h new file mode 100644 index 000000000000..1599573d35c9 --- /dev/null +++ b/include/net/netdev_netlink.h @@ -0,0 +1,11 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef __NET_NETDEV_NETLINK_H +#define __NET_NETDEV_NETLINK_H + +#include + +struct netdev_nl_sock { + struct list_head bindings; +}; + +#endif /* __NET_NETDEV_NETLINK_H */ diff --git a/net/core/netdev-genl-gen.c b/net/core/netdev-genl-gen.c index 996ac6a449eb..739f7b6506a6 100644 --- a/net/core/netdev-genl-gen.c +++ b/net/core/netdev-genl-gen.c @@ -9,7 +9,7 @@ #include "netdev-genl-gen.h" #include -#include +#include /* Integer value ranges */ static const struct netlink_range_validation netdev_a_page_pool_id_range = { @@ -217,7 +217,7 @@ struct genl_family netdev_nl_family __ro_after_init = { .n_split_ops = ARRAY_SIZE(netdev_nl_ops), .mcgrps = netdev_nl_mcgrps, .n_mcgrps = ARRAY_SIZE(netdev_nl_mcgrps), - .sock_priv_size = sizeof(struct list_head), + .sock_priv_size = sizeof(struct netdev_nl_sock), .sock_priv_init = __netdev_nl_sock_priv_init, .sock_priv_destroy = __netdev_nl_sock_priv_destroy, }; diff --git a/net/core/netdev-genl-gen.h b/net/core/netdev-genl-gen.h index e09dd7539ff2..17d39fd64c94 100644 --- a/net/core/netdev-genl-gen.h +++ b/net/core/netdev-genl-gen.h @@ -10,7 +10,7 @@ #include #include -#include +#include /* Common nested types */ extern const struct nla_policy netdev_page_pool_info_nl_policy[NETDEV_A_PAGE_POOL_IFINDEX + 1]; @@ -42,7 +42,7 @@ enum { extern struct genl_family netdev_nl_family; -void netdev_nl_sock_priv_init(struct list_head *priv); -void netdev_nl_sock_priv_destroy(struct list_head *priv); +void netdev_nl_sock_priv_init(struct netdev_nl_sock *priv); +void netdev_nl_sock_priv_destroy(struct netdev_nl_sock *priv); #endif /* _LINUX_NETDEV_GEN_H */ diff --git a/net/core/netdev-genl.c b/net/core/netdev-genl.c index 2b774183d31c..a219be90c739 100644 --- a/net/core/netdev-genl.c +++ b/net/core/netdev-genl.c @@ -829,8 +829,8 @@ int netdev_nl_bind_rx_doit(struct sk_buff *skb, struct genl_info *info) { struct nlattr *tb[ARRAY_SIZE(netdev_queue_id_nl_policy)]; struct net_devmem_dmabuf_binding *binding; - struct list_head *sock_binding_list; u32 ifindex, dmabuf_fd, rxq_idx; + struct netdev_nl_sock *priv; struct net_device *netdev; struct sk_buff *rsp; struct nlattr *attr; @@ -845,10 +845,9 @@ int netdev_nl_bind_rx_doit(struct sk_buff *skb, struct genl_info *info) ifindex = nla_get_u32(info->attrs[NETDEV_A_DEV_IFINDEX]); dmabuf_fd = nla_get_u32(info->attrs[NETDEV_A_DMABUF_FD]); - sock_binding_list = genl_sk_priv_get(&netdev_nl_family, - NETLINK_CB(skb).sk); - if (IS_ERR(sock_binding_list)) - return PTR_ERR(sock_binding_list); + priv = genl_sk_priv_get(&netdev_nl_family, NETLINK_CB(skb).sk); + if (IS_ERR(priv)) + return PTR_ERR(priv); rsp = genlmsg_new(GENLMSG_DEFAULT_SIZE, GFP_KERNEL); if (!rsp) @@ -909,7 +908,7 @@ int netdev_nl_bind_rx_doit(struct sk_buff *skb, struct genl_info *info) goto err_unbind; } - list_add(&binding->list, sock_binding_list); + list_add(&binding->list, &priv->bindings); nla_put_u32(rsp, NETDEV_A_DMABUF_ID, binding->id); genlmsg_end(rsp, hdr); @@ -931,17 +930,17 @@ err_genlmsg_free: return err; } -void netdev_nl_sock_priv_init(struct list_head *priv) +void netdev_nl_sock_priv_init(struct netdev_nl_sock *priv) { - INIT_LIST_HEAD(priv); + INIT_LIST_HEAD(&priv->bindings); } -void netdev_nl_sock_priv_destroy(struct list_head *priv) +void netdev_nl_sock_priv_destroy(struct netdev_nl_sock *priv) { struct net_devmem_dmabuf_binding *binding; struct net_devmem_dmabuf_binding *temp; - list_for_each_entry_safe(binding, temp, priv, list) { + list_for_each_entry_safe(binding, temp, &priv->bindings, list) { rtnl_lock(); net_devmem_unbind_dmabuf(binding); rtnl_unlock(); -- cgit v1.2.3 From 10eef096be25f3811ada2b43b108d1b8d8170001 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Tue, 11 Mar 2025 07:40:25 -0700 Subject: net: add granular lock for the netdev netlink socket As we move away from rtnl_lock for queue ops, introduce per-netdev_nl_sock lock. Signed-off-by: Stanislav Fomichev Reviewed-by: Mina Almasry Link: https://patch.msgid.link/20250311144026.4154277-3-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- include/net/netdev_netlink.h | 1 + net/core/netdev-genl.c | 6 ++++++ 2 files changed, 7 insertions(+) (limited to 'include') diff --git a/include/net/netdev_netlink.h b/include/net/netdev_netlink.h index 1599573d35c9..075962dbe743 100644 --- a/include/net/netdev_netlink.h +++ b/include/net/netdev_netlink.h @@ -5,6 +5,7 @@ #include struct netdev_nl_sock { + struct mutex lock; struct list_head bindings; }; diff --git a/net/core/netdev-genl.c b/net/core/netdev-genl.c index a219be90c739..63e10717efc5 100644 --- a/net/core/netdev-genl.c +++ b/net/core/netdev-genl.c @@ -859,6 +859,7 @@ int netdev_nl_bind_rx_doit(struct sk_buff *skb, struct genl_info *info) goto err_genlmsg_free; } + mutex_lock(&priv->lock); rtnl_lock(); netdev = __dev_get_by_index(genl_info_net(info), ifindex); @@ -918,6 +919,7 @@ int netdev_nl_bind_rx_doit(struct sk_buff *skb, struct genl_info *info) goto err_unbind; rtnl_unlock(); + mutex_unlock(&priv->lock); return 0; @@ -925,6 +927,7 @@ err_unbind: net_devmem_unbind_dmabuf(binding); err_unlock: rtnl_unlock(); + mutex_unlock(&priv->lock); err_genlmsg_free: nlmsg_free(rsp); return err; @@ -933,6 +936,7 @@ err_genlmsg_free: void netdev_nl_sock_priv_init(struct netdev_nl_sock *priv) { INIT_LIST_HEAD(&priv->bindings); + mutex_init(&priv->lock); } void netdev_nl_sock_priv_destroy(struct netdev_nl_sock *priv) @@ -940,11 +944,13 @@ void netdev_nl_sock_priv_destroy(struct netdev_nl_sock *priv) struct net_devmem_dmabuf_binding *binding; struct net_devmem_dmabuf_binding *temp; + mutex_lock(&priv->lock); list_for_each_entry_safe(binding, temp, &priv->bindings, list) { rtnl_lock(); net_devmem_unbind_dmabuf(binding); rtnl_unlock(); } + mutex_unlock(&priv->lock); } static int netdev_genl_netdevice_event(struct notifier_block *nb, -- cgit v1.2.3 From 0114a91da67263c56612ff19a6aa0992466e7756 Mon Sep 17 00:00:00 2001 From: Chia-Yu Chang Date: Wed, 5 Mar 2025 23:38:43 +0100 Subject: tcp: use BIT() macro in include/net/tcp.h MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use BIT() macro for TCP flags field and TCP congestion control flags that will be used by the congestion control algorithm. No functional changes. Signed-off-by: Chia-Yu Chang Reviewed-by: Ilpo Järvinen Reviewed-by: Eric Dumazet Signed-off-by: David S. Miller --- include/net/tcp.h | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) (limited to 'include') diff --git a/include/net/tcp.h b/include/net/tcp.h index 7207c52b1fc9..d2315d69e019 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -26,6 +26,7 @@ #include #include #include +#include #include #include @@ -934,14 +935,14 @@ static inline u32 tcp_rsk_tsval(const struct tcp_request_sock *treq) #define tcp_flag_byte(th) (((u_int8_t *)th)[13]) -#define TCPHDR_FIN 0x01 -#define TCPHDR_SYN 0x02 -#define TCPHDR_RST 0x04 -#define TCPHDR_PSH 0x08 -#define TCPHDR_ACK 0x10 -#define TCPHDR_URG 0x20 -#define TCPHDR_ECE 0x40 -#define TCPHDR_CWR 0x80 +#define TCPHDR_FIN BIT(0) +#define TCPHDR_SYN BIT(1) +#define TCPHDR_RST BIT(2) +#define TCPHDR_PSH BIT(3) +#define TCPHDR_ACK BIT(4) +#define TCPHDR_URG BIT(5) +#define TCPHDR_ECE BIT(6) +#define TCPHDR_CWR BIT(7) #define TCPHDR_SYN_ECN (TCPHDR_SYN | TCPHDR_ECE | TCPHDR_CWR) @@ -1132,9 +1133,9 @@ enum tcp_ca_ack_event_flags { #define TCP_CA_UNSPEC 0 /* Algorithm can be set on socket without CAP_NET_ADMIN privileges */ -#define TCP_CONG_NON_RESTRICTED 0x1 +#define TCP_CONG_NON_RESTRICTED BIT(0) /* Requires ECN/ECT set on all packets */ -#define TCP_CONG_NEEDS_ECN 0x2 +#define TCP_CONG_NEEDS_ECN BIT(1) #define TCP_CONG_MASK (TCP_CONG_NON_RESTRICTED | TCP_CONG_NEEDS_ECN) union tcp_cc_info; -- cgit v1.2.3 From 2c2f08d31d2f6cd5621b121d596986bd54da679a Mon Sep 17 00:00:00 2001 From: Ilpo Järvinen Date: Wed, 5 Mar 2025 23:38:44 +0100 Subject: tcp: extend TCP flags to allow AE bit/ACE field MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit With AccECN, there's one additional TCP flag to be used (AE) and ACE field that overloads the definition of AE, CWR, and ECE flags. As tcp_flags was previously only 1 byte, the byte-order stuff needs to be added to it's handling. Signed-off-by: Ilpo Järvinen Signed-off-by: Chia-Yu Chang Reviewed-by: Eric Dumazet Signed-off-by: David S. Miller --- include/net/tcp.h | 11 +++++++++-- include/uapi/linux/tcp.h | 9 ++++++--- net/ipv4/tcp_ipv4.c | 2 +- net/ipv4/tcp_output.c | 8 ++++---- net/ipv6/tcp_ipv6.c | 2 +- net/netfilter/nf_log_syslog.c | 8 +++++--- 6 files changed, 26 insertions(+), 14 deletions(-) (limited to 'include') diff --git a/include/net/tcp.h b/include/net/tcp.h index d2315d69e019..4e11bdc86470 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -943,7 +943,14 @@ static inline u32 tcp_rsk_tsval(const struct tcp_request_sock *treq) #define TCPHDR_URG BIT(5) #define TCPHDR_ECE BIT(6) #define TCPHDR_CWR BIT(7) - +#define TCPHDR_AE BIT(8) +#define TCPHDR_FLAGS_MASK (TCPHDR_FIN | TCPHDR_SYN | TCPHDR_RST | \ + TCPHDR_PSH | TCPHDR_ACK | TCPHDR_URG | \ + TCPHDR_ECE | TCPHDR_CWR | TCPHDR_AE) +#define tcp_flags_ntohs(th) (ntohs(*(__be16 *)&tcp_flag_word(th)) & \ + TCPHDR_FLAGS_MASK) + +#define TCPHDR_ACE (TCPHDR_ECE | TCPHDR_CWR | TCPHDR_AE) #define TCPHDR_SYN_ECN (TCPHDR_SYN | TCPHDR_ECE | TCPHDR_CWR) /* State flags for sacked in struct tcp_skb_cb */ @@ -978,7 +985,7 @@ struct tcp_skb_cb { u16 tcp_gso_size; }; }; - __u8 tcp_flags; /* TCP header flags. (tcp[13]) */ + __u16 tcp_flags; /* TCP header flags (tcp[12-13])*/ __u8 sacked; /* State flags for SACK. */ __u8 ip_dsfield; /* IPv4 tos or IPv6 dsfield */ diff --git a/include/uapi/linux/tcp.h b/include/uapi/linux/tcp.h index 32a27b4a5020..92a2e79222ea 100644 --- a/include/uapi/linux/tcp.h +++ b/include/uapi/linux/tcp.h @@ -28,7 +28,8 @@ struct tcphdr { __be32 seq; __be32 ack_seq; #if defined(__LITTLE_ENDIAN_BITFIELD) - __u16 res1:4, + __u16 ae:1, + res1:3, doff:4, fin:1, syn:1, @@ -40,7 +41,8 @@ struct tcphdr { cwr:1; #elif defined(__BIG_ENDIAN_BITFIELD) __u16 doff:4, - res1:4, + res1:3, + ae:1, cwr:1, ece:1, urg:1, @@ -70,6 +72,7 @@ union tcp_word_hdr { #define tcp_flag_word(tp) (((union tcp_word_hdr *)(tp))->words[3]) enum { + TCP_FLAG_AE = __constant_cpu_to_be32(0x01000000), TCP_FLAG_CWR = __constant_cpu_to_be32(0x00800000), TCP_FLAG_ECE = __constant_cpu_to_be32(0x00400000), TCP_FLAG_URG = __constant_cpu_to_be32(0x00200000), @@ -78,7 +81,7 @@ enum { TCP_FLAG_RST = __constant_cpu_to_be32(0x00040000), TCP_FLAG_SYN = __constant_cpu_to_be32(0x00020000), TCP_FLAG_FIN = __constant_cpu_to_be32(0x00010000), - TCP_RESERVED_BITS = __constant_cpu_to_be32(0x0F000000), + TCP_RESERVED_BITS = __constant_cpu_to_be32(0x0E000000), TCP_DATA_OFFSET = __constant_cpu_to_be32(0xF0000000) }; diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index d9405b012dff..fab684221bf7 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -2159,7 +2159,7 @@ static void tcp_v4_fill_cb(struct sk_buff *skb, const struct iphdr *iph, TCP_SKB_CB(skb)->end_seq = (TCP_SKB_CB(skb)->seq + th->syn + th->fin + skb->len - th->doff * 4); TCP_SKB_CB(skb)->ack_seq = ntohl(th->ack_seq); - TCP_SKB_CB(skb)->tcp_flags = tcp_flag_byte(th); + TCP_SKB_CB(skb)->tcp_flags = tcp_flags_ntohs(th); TCP_SKB_CB(skb)->ip_dsfield = ipv4_get_dsfield(iph); TCP_SKB_CB(skb)->sacked = 0; TCP_SKB_CB(skb)->has_rxtstamp = diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 24e56bf96747..efd3cb5e1ded 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -403,7 +403,7 @@ static void tcp_ecn_send(struct sock *sk, struct sk_buff *skb, /* Constructs common control bits of non-data skb. If SYN/FIN is present, * auto increment end seqno. */ -static void tcp_init_nondata_skb(struct sk_buff *skb, u32 seq, u8 flags) +static void tcp_init_nondata_skb(struct sk_buff *skb, u32 seq, u16 flags) { skb->ip_summed = CHECKSUM_PARTIAL; @@ -1395,7 +1395,7 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, th->seq = htonl(tcb->seq); th->ack_seq = htonl(rcv_nxt); *(((__be16 *)th) + 6) = htons(((tcp_header_size >> 2) << 12) | - tcb->tcp_flags); + (tcb->tcp_flags & TCPHDR_FLAGS_MASK)); th->check = 0; th->urg_ptr = 0; @@ -1616,8 +1616,8 @@ int tcp_fragment(struct sock *sk, enum tcp_queue tcp_queue, struct sk_buff *buff; int old_factor; long limit; + u16 flags; int nlen; - u8 flags; if (WARN_ON(len > skb->len)) return -EINVAL; @@ -2171,7 +2171,7 @@ static int tso_fragment(struct sock *sk, struct sk_buff *skb, unsigned int len, { int nlen = skb->len - len; struct sk_buff *buff; - u8 flags; + u16 flags; /* All of a TSO frame must be composed of paged data. */ DEBUG_NET_WARN_ON_ONCE(skb->len != skb->data_len); diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 85c4820bfe15..a2fcc317a88e 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1731,7 +1731,7 @@ static void tcp_v6_fill_cb(struct sk_buff *skb, const struct ipv6hdr *hdr, TCP_SKB_CB(skb)->end_seq = (TCP_SKB_CB(skb)->seq + th->syn + th->fin + skb->len - th->doff*4); TCP_SKB_CB(skb)->ack_seq = ntohl(th->ack_seq); - TCP_SKB_CB(skb)->tcp_flags = tcp_flag_byte(th); + TCP_SKB_CB(skb)->tcp_flags = tcp_flags_ntohs(th); TCP_SKB_CB(skb)->ip_dsfield = ipv6_get_dsfield(hdr); TCP_SKB_CB(skb)->sacked = 0; TCP_SKB_CB(skb)->has_rxtstamp = diff --git a/net/netfilter/nf_log_syslog.c b/net/netfilter/nf_log_syslog.c index 58402226045e..86d5fc5d28e3 100644 --- a/net/netfilter/nf_log_syslog.c +++ b/net/netfilter/nf_log_syslog.c @@ -216,7 +216,9 @@ nf_log_dump_tcp_header(struct nf_log_buf *m, /* Max length: 9 "RES=0x3C " */ nf_log_buf_add(m, "RES=0x%02x ", (u_int8_t)(ntohl(tcp_flag_word(th) & TCP_RESERVED_BITS) >> 22)); - /* Max length: 32 "CWR ECE URG ACK PSH RST SYN FIN " */ + /* Max length: 35 "AE CWR ECE URG ACK PSH RST SYN FIN " */ + if (th->ae) + nf_log_buf_add(m, "AE "); if (th->cwr) nf_log_buf_add(m, "CWR "); if (th->ece) @@ -516,7 +518,7 @@ dump_ipv4_packet(struct net *net, struct nf_log_buf *m, /* Proto Max log string length */ /* IP: 40+46+6+11+127 = 230 */ - /* TCP: 10+max(25,20+30+13+9+32+11+127) = 252 */ + /* TCP: 10+max(25,20+30+13+9+35+11+127) = 255 */ /* UDP: 10+max(25,20) = 35 */ /* UDPLITE: 14+max(25,20) = 39 */ /* ICMP: 11+max(25, 18+25+max(19,14,24+3+n+10,3+n+10)) = 91+n */ @@ -526,7 +528,7 @@ dump_ipv4_packet(struct net *net, struct nf_log_buf *m, /* (ICMP allows recursion one level deep) */ /* maxlen = IP + ICMP + IP + max(TCP,UDP,ICMP,unknown) */ - /* maxlen = 230+ 91 + 230 + 252 = 803 */ + /* maxlen = 230+ 91 + 230 + 255 = 806 */ } static noinline_for_stack void -- cgit v1.2.3 From 041fb11d518f5d25d323766bc0f59b09bba314bc Mon Sep 17 00:00:00 2001 From: Ilpo Järvinen Date: Wed, 5 Mar 2025 23:38:47 +0100 Subject: tcp: helpers for ECN mode handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Create helpers for TCP ECN modes. No functional changes. Signed-off-by: Ilpo Järvinen Signed-off-by: Chia-Yu Chang Reviewed-by: Eric Dumazet Signed-off-by: David S. Miller --- include/net/tcp.h | 44 ++++++++++++++++++++++++++++++++++++++++---- net/ipv4/tcp.c | 2 +- net/ipv4/tcp_dctcp.c | 2 +- net/ipv4/tcp_input.c | 14 +++++++------- net/ipv4/tcp_minisocks.c | 4 +++- net/ipv4/tcp_output.c | 6 +++--- 6 files changed, 55 insertions(+), 17 deletions(-) (limited to 'include') diff --git a/include/net/tcp.h b/include/net/tcp.h index 4e11bdc86470..5520f2d62dc0 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -374,10 +374,46 @@ static inline void tcp_dec_quickack_mode(struct sock *sk) } } -#define TCP_ECN_OK 1 -#define TCP_ECN_QUEUE_CWR 2 -#define TCP_ECN_DEMAND_CWR 4 -#define TCP_ECN_SEEN 8 +#define TCP_ECN_MODE_RFC3168 BIT(0) +#define TCP_ECN_QUEUE_CWR BIT(1) +#define TCP_ECN_DEMAND_CWR BIT(2) +#define TCP_ECN_SEEN BIT(3) +#define TCP_ECN_MODE_ACCECN BIT(4) + +#define TCP_ECN_DISABLED 0 +#define TCP_ECN_MODE_PENDING (TCP_ECN_MODE_RFC3168 | TCP_ECN_MODE_ACCECN) +#define TCP_ECN_MODE_ANY (TCP_ECN_MODE_RFC3168 | TCP_ECN_MODE_ACCECN) + +static inline bool tcp_ecn_mode_any(const struct tcp_sock *tp) +{ + return tp->ecn_flags & TCP_ECN_MODE_ANY; +} + +static inline bool tcp_ecn_mode_rfc3168(const struct tcp_sock *tp) +{ + return (tp->ecn_flags & TCP_ECN_MODE_ANY) == TCP_ECN_MODE_RFC3168; +} + +static inline bool tcp_ecn_mode_accecn(const struct tcp_sock *tp) +{ + return (tp->ecn_flags & TCP_ECN_MODE_ANY) == TCP_ECN_MODE_ACCECN; +} + +static inline bool tcp_ecn_disabled(const struct tcp_sock *tp) +{ + return !tcp_ecn_mode_any(tp); +} + +static inline bool tcp_ecn_mode_pending(const struct tcp_sock *tp) +{ + return (tp->ecn_flags & TCP_ECN_MODE_PENDING) == TCP_ECN_MODE_PENDING; +} + +static inline void tcp_ecn_mode_set(struct tcp_sock *tp, u8 mode) +{ + tp->ecn_flags &= ~TCP_ECN_MODE_ANY; + tp->ecn_flags |= mode; +} enum tcp_tw_status { TCP_TW_SUCCESS = 0, diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 46951e749308..989c3c3d8e75 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -4138,7 +4138,7 @@ void tcp_get_info(struct sock *sk, struct tcp_info *info) info->tcpi_rcv_wscale = tp->rx_opt.rcv_wscale; } - if (tp->ecn_flags & TCP_ECN_OK) + if (tcp_ecn_mode_any(tp)) info->tcpi_options |= TCPI_OPT_ECN; if (tp->ecn_flags & TCP_ECN_SEEN) info->tcpi_options |= TCPI_OPT_ECN_SEEN; diff --git a/net/ipv4/tcp_dctcp.c b/net/ipv4/tcp_dctcp.c index 8a45a4aea933..03abe0848420 100644 --- a/net/ipv4/tcp_dctcp.c +++ b/net/ipv4/tcp_dctcp.c @@ -90,7 +90,7 @@ __bpf_kfunc static void dctcp_init(struct sock *sk) { const struct tcp_sock *tp = tcp_sk(sk); - if ((tp->ecn_flags & TCP_ECN_OK) || + if (tcp_ecn_mode_any(tp) || (sk->sk_state == TCP_LISTEN || sk->sk_state == TCP_CLOSE)) { struct dctcp *ca = inet_csk_ca(sk); diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 769048b559e5..5c270cf96678 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -342,7 +342,7 @@ static bool tcp_in_quickack_mode(struct sock *sk) static void tcp_ecn_queue_cwr(struct tcp_sock *tp) { - if (tp->ecn_flags & TCP_ECN_OK) + if (tcp_ecn_mode_rfc3168(tp)) tp->ecn_flags |= TCP_ECN_QUEUE_CWR; } @@ -369,7 +369,7 @@ static void tcp_data_ecn_check(struct sock *sk, const struct sk_buff *skb) { struct tcp_sock *tp = tcp_sk(sk); - if (!(tcp_sk(sk)->ecn_flags & TCP_ECN_OK)) + if (tcp_ecn_disabled(tp)) return; switch (TCP_SKB_CB(skb)->ip_dsfield & INET_ECN_MASK) { @@ -402,19 +402,19 @@ static void tcp_data_ecn_check(struct sock *sk, const struct sk_buff *skb) static void tcp_ecn_rcv_synack(struct tcp_sock *tp, const struct tcphdr *th) { - if ((tp->ecn_flags & TCP_ECN_OK) && (!th->ece || th->cwr)) - tp->ecn_flags &= ~TCP_ECN_OK; + if (tcp_ecn_mode_rfc3168(tp) && (!th->ece || th->cwr)) + tcp_ecn_mode_set(tp, TCP_ECN_DISABLED); } static void tcp_ecn_rcv_syn(struct tcp_sock *tp, const struct tcphdr *th) { - if ((tp->ecn_flags & TCP_ECN_OK) && (!th->ece || !th->cwr)) - tp->ecn_flags &= ~TCP_ECN_OK; + if (tcp_ecn_mode_rfc3168(tp) && (!th->ece || !th->cwr)) + tcp_ecn_mode_set(tp, TCP_ECN_DISABLED); } static bool tcp_ecn_rcv_ecn_echo(const struct tcp_sock *tp, const struct tcphdr *th) { - if (th->ece && !th->syn && (tp->ecn_flags & TCP_ECN_OK)) + if (th->ece && !th->syn && tcp_ecn_mode_rfc3168(tp)) return true; return false; } diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index 3cb8f281186b..0ae24add155b 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -461,7 +461,9 @@ void tcp_openreq_init_rwin(struct request_sock *req, static void tcp_ecn_openreq_child(struct tcp_sock *tp, const struct request_sock *req) { - tp->ecn_flags = inet_rsk(req)->ecn_ok ? TCP_ECN_OK : 0; + tcp_ecn_mode_set(tp, inet_rsk(req)->ecn_ok ? + TCP_ECN_MODE_RFC3168 : + TCP_ECN_DISABLED); } void tcp_ca_openreq_child(struct sock *sk, const struct dst_entry *dst) diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index efd3cb5e1ded..fffbd3ee44ed 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -325,7 +325,7 @@ static void tcp_ecn_send_synack(struct sock *sk, struct sk_buff *skb) const struct tcp_sock *tp = tcp_sk(sk); TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_CWR; - if (!(tp->ecn_flags & TCP_ECN_OK)) + if (tcp_ecn_disabled(tp)) TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_ECE; else if (tcp_ca_needs_ecn(sk) || tcp_bpf_ca_needs_ecn(sk)) @@ -351,7 +351,7 @@ static void tcp_ecn_send_syn(struct sock *sk, struct sk_buff *skb) if (use_ecn) { TCP_SKB_CB(skb)->tcp_flags |= TCPHDR_ECE | TCPHDR_CWR; - tp->ecn_flags = TCP_ECN_OK; + tcp_ecn_mode_set(tp, TCP_ECN_MODE_RFC3168); if (tcp_ca_needs_ecn(sk) || bpf_needs_ecn) INET_ECN_xmit(sk); } @@ -381,7 +381,7 @@ static void tcp_ecn_send(struct sock *sk, struct sk_buff *skb, { struct tcp_sock *tp = tcp_sk(sk); - if (tp->ecn_flags & TCP_ECN_OK) { + if (tcp_ecn_mode_rfc3168(tp)) { /* Not-retransmitted data segment: set ECT and inject CWR. */ if (skb->len != tcp_header_len && !before(TCP_SKB_CB(skb)->seq, tp->snd_nxt)) { -- cgit v1.2.3 From 023af5a72ab161f2e661afb53e3b6a6901f6ba00 Mon Sep 17 00:00:00 2001 From: Ilpo Järvinen Date: Wed, 5 Mar 2025 23:38:48 +0100 Subject: gso: AccECN support MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Handling the CWR flag differs between RFC 3168 ECN and AccECN. With RFC 3168 ECN aware TSO (NETIF_F_TSO_ECN) CWR flag is cleared starting from 2nd segment which is incompatible how AccECN handles the CWR flag. Such super-segments are indicated by SKB_GSO_TCP_ECN. With AccECN, CWR flag (or more accurately, the ACE field that also includes ECE & AE flags) changes only when new packet(s) with CE mark arrives so the flag should not be changed within a super-skb. The new skb/feature flags are necessary to prevent such TSO engines corrupting AccECN ACE counters by clearing the CWR flag (if the CWR handling feature cannot be turned off). If NIC is completely unaware of RFC3168 ECN (doesn't support NETIF_F_TSO_ECN) or its TSO engine can be set to not touch CWR flag despite supporting also NETIF_F_TSO_ECN, TSO could be safely used with AccECN on such NIC. This should be evaluated per NIC basis (not done in this patch series for any NICs). For the cases, where TSO cannot keep its hands off the CWR flag, a GSO fallback is provided by this patch. Signed-off-by: Ilpo Järvinen Signed-off-by: Chia-Yu Chang Reviewed-by: Eric Dumazet Signed-off-by: David S. Miller --- include/linux/netdev_features.h | 8 +++++--- include/linux/netdevice.h | 2 ++ include/linux/skbuff.h | 2 ++ net/ethtool/common.c | 1 + net/ipv4/tcp_offload.c | 6 +++++- 5 files changed, 15 insertions(+), 4 deletions(-) (limited to 'include') diff --git a/include/linux/netdev_features.h b/include/linux/netdev_features.h index 11be70a7929f..7a01c518e573 100644 --- a/include/linux/netdev_features.h +++ b/include/linux/netdev_features.h @@ -53,12 +53,12 @@ enum { NETIF_F_GSO_UDP_BIT, /* ... UFO, deprecated except tuntap */ NETIF_F_GSO_UDP_L4_BIT, /* ... UDP payload GSO (not UFO) */ NETIF_F_GSO_FRAGLIST_BIT, /* ... Fraglist GSO */ + NETIF_F_GSO_ACCECN_BIT, /* TCP AccECN w/ TSO (no clear CWR) */ /**/NETIF_F_GSO_LAST = /* last bit, see GSO_MASK */ - NETIF_F_GSO_FRAGLIST_BIT, + NETIF_F_GSO_ACCECN_BIT, NETIF_F_FCOE_CRC_BIT, /* FCoE CRC32 */ NETIF_F_SCTP_CRC_BIT, /* SCTP checksum offload */ - __UNUSED_NETIF_F_37, NETIF_F_NTUPLE_BIT, /* N-tuple filters supported */ NETIF_F_RXHASH_BIT, /* Receive hashing offload */ NETIF_F_RXCSUM_BIT, /* Receive checksumming offload */ @@ -128,6 +128,7 @@ enum { #define NETIF_F_SG __NETIF_F(SG) #define NETIF_F_TSO6 __NETIF_F(TSO6) #define NETIF_F_TSO_ECN __NETIF_F(TSO_ECN) +#define NETIF_F_GSO_ACCECN __NETIF_F(GSO_ACCECN) #define NETIF_F_TSO __NETIF_F(TSO) #define NETIF_F_VLAN_CHALLENGED __NETIF_F(VLAN_CHALLENGED) #define NETIF_F_RXFCS __NETIF_F(RXFCS) @@ -210,7 +211,8 @@ static inline int find_next_netdev_feature(u64 feature, unsigned long start) NETIF_F_TSO_ECN | NETIF_F_TSO_MANGLEID) /* List of features with software fallbacks. */ -#define NETIF_F_GSO_SOFTWARE (NETIF_F_ALL_TSO | NETIF_F_GSO_SCTP | \ +#define NETIF_F_GSO_SOFTWARE (NETIF_F_ALL_TSO | \ + NETIF_F_GSO_ACCECN | NETIF_F_GSO_SCTP | \ NETIF_F_GSO_UDP_L4 | NETIF_F_GSO_FRAGLIST) /* diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 0dbfe069a6e3..67527243459b 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -5269,6 +5269,8 @@ static inline bool net_gso_ok(netdev_features_t features, int gso_type) BUILD_BUG_ON(SKB_GSO_UDP != (NETIF_F_GSO_UDP >> NETIF_F_GSO_SHIFT)); BUILD_BUG_ON(SKB_GSO_UDP_L4 != (NETIF_F_GSO_UDP_L4 >> NETIF_F_GSO_SHIFT)); BUILD_BUG_ON(SKB_GSO_FRAGLIST != (NETIF_F_GSO_FRAGLIST >> NETIF_F_GSO_SHIFT)); + BUILD_BUG_ON(SKB_GSO_TCP_ACCECN != + (NETIF_F_GSO_ACCECN >> NETIF_F_GSO_SHIFT)); return (features & feature) == feature; } diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 14517e95a46c..b8a1343d6785 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -708,6 +708,8 @@ enum { SKB_GSO_UDP_L4 = 1 << 17, SKB_GSO_FRAGLIST = 1 << 18, + + SKB_GSO_TCP_ACCECN = 1 << 19, }; #if BITS_PER_LONG > 32 diff --git a/net/ethtool/common.c b/net/ethtool/common.c index ac8b6107863e..1c6f2a2f1871 100644 --- a/net/ethtool/common.c +++ b/net/ethtool/common.c @@ -36,6 +36,7 @@ const char netdev_features_strings[NETDEV_FEATURE_COUNT][ETH_GSTRING_LEN] = { [NETIF_F_TSO_BIT] = "tx-tcp-segmentation", [NETIF_F_GSO_ROBUST_BIT] = "tx-gso-robust", [NETIF_F_TSO_ECN_BIT] = "tx-tcp-ecn-segmentation", + [NETIF_F_GSO_ACCECN_BIT] = "tx-tcp-accecn-segmentation", [NETIF_F_TSO_MANGLEID_BIT] = "tx-tcp-mangleid-segmentation", [NETIF_F_TSO6_BIT] = "tx-tcp6-segmentation", [NETIF_F_FSO_BIT] = "tx-fcoe-segmentation", diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c index e27f4ee2a0de..8c74fd69db2a 100644 --- a/net/ipv4/tcp_offload.c +++ b/net/ipv4/tcp_offload.c @@ -142,6 +142,7 @@ struct sk_buff *tcp_gso_segment(struct sk_buff *skb, struct sk_buff *gso_skb = skb; __sum16 newcheck; bool ooo_okay, copy_destructor; + bool ecn_cwr_mask; __wsum delta; th = tcp_hdr(skb); @@ -201,6 +202,8 @@ struct sk_buff *tcp_gso_segment(struct sk_buff *skb, newcheck = ~csum_fold(csum_add(csum_unfold(th->check), delta)); + ecn_cwr_mask = !!(skb_shinfo(gso_skb)->gso_type & SKB_GSO_TCP_ACCECN); + while (skb->next) { th->fin = th->psh = 0; th->check = newcheck; @@ -220,7 +223,8 @@ struct sk_buff *tcp_gso_segment(struct sk_buff *skb, th = tcp_hdr(skb); th->seq = htonl(seq); - th->cwr = 0; + + th->cwr &= ecn_cwr_mask; } /* Following permits TCP Small Queues to work well with GSO : -- cgit v1.2.3 From 4618e195f9251947a5f15f7e5533bcc3a19b93ef Mon Sep 17 00:00:00 2001 From: Ilpo Järvinen Date: Wed, 5 Mar 2025 23:38:51 +0100 Subject: tcp: add new TCP_TW_ACK_OOW state and allow ECN bits in TOS MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ECN bits in TOS are always cleared when sending in ACKs in TW. Clearing them is problematic for TCP flows that used Accurate ECN because ECN bits decide which service queue the packet is placed into (L4S vs Classic). Effectively, TW ACKs are always downgraded from L4S to Classic queue which might impact, e.g., delay the ACK will experience on the path compared with the other packets of the flow. Change the TW ACK sending code to differentiate: - In tcp_v4_send_reset(), commit ba9e04a7ddf4f ("ip: fix tos reflection in ack and reset packets") cleans ECN bits for TW reset and this is not affected. - In tcp_v4_timewait_ack(), ECN bits for all TW ACKs are cleaned. But now only ECN bits of ACKs for oow data or paws_reject are cleaned, and ECN bits of other ACKs will not be cleaned. - In tcp_v4_reqsk_send_ack(), commit 66b13d99d96a1 ("ipv4: tcp: fix TOS value in ACK messages sent from TIME_WAIT") did not clean ECN bits of ACKs for oow data or paws_reject. But now the ECN bits rae cleaned for these ACKs. Signed-off-by: Ilpo Järvinen Signed-off-by: Chia-Yu Chang Signed-off-by: David S. Miller --- include/net/tcp.h | 3 ++- net/ipv4/ip_output.c | 3 +-- net/ipv4/tcp_ipv4.c | 29 +++++++++++++++++++++++------ net/ipv4/tcp_minisocks.c | 2 +- net/ipv6/tcp_ipv6.c | 24 +++++++++++++++++------- 5 files changed, 44 insertions(+), 17 deletions(-) (limited to 'include') diff --git a/include/net/tcp.h b/include/net/tcp.h index 5520f2d62dc0..c6bdd3d63159 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -419,7 +419,8 @@ enum tcp_tw_status { TCP_TW_SUCCESS = 0, TCP_TW_RST = 1, TCP_TW_ACK = 2, - TCP_TW_SYN = 3 + TCP_TW_SYN = 3, + TCP_TW_ACK_OOW = 4 }; diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index ea7a260bec8a..6e18d7ec5062 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -75,7 +75,6 @@ #include #include #include -#include #include #include #include @@ -1640,7 +1639,7 @@ void ip_send_unicast_reply(struct sock *sk, const struct sock *orig_sk, if (IS_ERR(rt)) return; - inet_sk(sk)->tos = arg->tos & ~INET_ECN_MASK; + inet_sk(sk)->tos = arg->tos; sk->sk_protocol = ip_hdr(skb)->protocol; sk->sk_bound_dev_if = arg->bound_dev_if; diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 87f270ebc635..4fa4fbb0ad12 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -66,6 +66,7 @@ #include #include #include +#include #include #include #include @@ -887,7 +888,8 @@ static void tcp_v4_send_reset(const struct sock *sk, struct sk_buff *skb, BUILD_BUG_ON(offsetof(struct sock, sk_bound_dev_if) != offsetof(struct inet_timewait_sock, tw_bound_dev_if)); - arg.tos = ip_hdr(skb)->tos; + /* ECN bits of TW reset are cleared */ + arg.tos = ip_hdr(skb)->tos & ~INET_ECN_MASK; arg.uid = sock_net_uid(net, sk && sk_fullsock(sk) ? sk : NULL); local_bh_disable(); local_lock_nested_bh(&ipv4_tcp_sk.bh_lock); @@ -1033,11 +1035,21 @@ static void tcp_v4_send_ack(const struct sock *sk, local_bh_enable(); } -static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb) +static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb, + enum tcp_tw_status tw_status) { struct inet_timewait_sock *tw = inet_twsk(sk); struct tcp_timewait_sock *tcptw = tcp_twsk(sk); struct tcp_key key = {}; + u8 tos = tw->tw_tos; + + /* Cleaning only ECN bits of TW ACKs of oow data or is paws_reject, + * while not cleaning ECN bits of other TW ACKs to avoid these ACKs + * being placed in a different service queues (Classic rather than L4S) + */ + if (tw_status == TCP_TW_ACK_OOW) + tos &= ~INET_ECN_MASK; + #ifdef CONFIG_TCP_AO struct tcp_ao_info *ao_info; @@ -1081,7 +1093,7 @@ static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb) READ_ONCE(tcptw->tw_ts_recent), tw->tw_bound_dev_if, &key, tw->tw_transparent ? IP_REPLY_ARG_NOSRCCHECK : 0, - tw->tw_tos, + tos, tw->tw_txhash); inet_twsk_put(tw); @@ -1151,6 +1163,7 @@ static void tcp_v4_reqsk_send_ack(const struct sock *sk, struct sk_buff *skb, key.type = TCP_KEY_MD5; } + /* Cleaning ECN bits of TW ACKs of oow data or is paws_reject */ tcp_v4_send_ack(sk, skb, seq, tcp_rsk(req)->rcv_nxt, tcp_synack_window(req) >> inet_rsk(req)->rcv_wscale, @@ -1158,7 +1171,7 @@ static void tcp_v4_reqsk_send_ack(const struct sock *sk, struct sk_buff *skb, req->ts_recent, 0, &key, inet_rsk(req)->no_srccheck ? IP_REPLY_ARG_NOSRCCHECK : 0, - ip_hdr(skb)->tos, + ip_hdr(skb)->tos & ~INET_ECN_MASK, READ_ONCE(tcp_rsk(req)->txhash)); if (tcp_key_is_ao(&key)) kfree(key.traffic_key); @@ -2175,6 +2188,7 @@ int tcp_v4_rcv(struct sk_buff *skb) { struct net *net = dev_net_rcu(skb->dev); enum skb_drop_reason drop_reason; + enum tcp_tw_status tw_status; int sdif = inet_sdif(skb); int dif = inet_iif(skb); const struct iphdr *iph; @@ -2402,7 +2416,9 @@ do_time_wait: inet_twsk_put(inet_twsk(sk)); goto csum_error; } - switch (tcp_timewait_state_process(inet_twsk(sk), skb, th, &isn)) { + + tw_status = tcp_timewait_state_process(inet_twsk(sk), skb, th, &isn); + switch (tw_status) { case TCP_TW_SYN: { struct sock *sk2 = inet_lookup_listener(net, net->ipv4.tcp_death_row.hashinfo, @@ -2423,7 +2439,8 @@ do_time_wait: /* to ACK */ fallthrough; case TCP_TW_ACK: - tcp_v4_timewait_ack(sk, skb); + case TCP_TW_ACK_OOW: + tcp_v4_timewait_ack(sk, skb, tw_status); break; case TCP_TW_RST: tcp_v4_send_reset(sk, skb, SK_RST_REASON_TCP_TIMEWAIT_SOCKET); diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index 0ae24add155b..fb9349be36b8 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/net/ipv4/tcp_minisocks.c @@ -44,7 +44,7 @@ tcp_timewait_check_oow_rate_limit(struct inet_timewait_sock *tw, /* Send ACK. Note, we do not put the bucket, * it will be released by caller. */ - return TCP_TW_ACK; + return TCP_TW_ACK_OOW; } /* We are rate-limiting, so just release the tw sock and drop skb. */ diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index a2fcc317a88e..e182ee0a2330 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -999,7 +999,7 @@ static void tcp_v6_send_response(const struct sock *sk, struct sk_buff *skb, u32 if (!IS_ERR(dst)) { skb_dst_set(buff, dst); ip6_xmit(ctl_sk, buff, &fl6, fl6.flowi6_mark, NULL, - tclass & ~INET_ECN_MASK, priority); + tclass, priority); TCP_INC_STATS(net, TCP_MIB_OUTSEGS); if (rst) TCP_INC_STATS(net, TCP_MIB_OUTRSTS); @@ -1135,7 +1135,8 @@ static void tcp_v6_send_reset(const struct sock *sk, struct sk_buff *skb, trace_tcp_send_reset(sk, skb, reason); tcp_v6_send_response(sk, skb, seq, ack_seq, 0, 0, 0, oif, 1, - ipv6_get_dsfield(ipv6h), label, priority, txhash, + ipv6_get_dsfield(ipv6h) & ~INET_ECN_MASK, + label, priority, txhash, &key); #if defined(CONFIG_TCP_MD5SIG) || defined(CONFIG_TCP_AO) @@ -1155,11 +1156,16 @@ static void tcp_v6_send_ack(const struct sock *sk, struct sk_buff *skb, u32 seq, tclass, label, priority, txhash, key); } -static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb) +static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb, + enum tcp_tw_status tw_status) { struct inet_timewait_sock *tw = inet_twsk(sk); struct tcp_timewait_sock *tcptw = tcp_twsk(sk); + u8 tclass = tw->tw_tclass; struct tcp_key key = {}; + + if (tw_status == TCP_TW_ACK_OOW) + tclass &= ~INET_ECN_MASK; #ifdef CONFIG_TCP_AO struct tcp_ao_info *ao_info; @@ -1203,7 +1209,7 @@ static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb) tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale, tcp_tw_tsval(tcptw), READ_ONCE(tcptw->tw_ts_recent), tw->tw_bound_dev_if, - &key, tw->tw_tclass, cpu_to_be32(tw->tw_flowlabel), + &key, tclass, cpu_to_be32(tw->tw_flowlabel), tw->tw_priority, tw->tw_txhash); #ifdef CONFIG_TCP_AO @@ -1280,7 +1286,8 @@ static void tcp_v6_reqsk_send_ack(const struct sock *sk, struct sk_buff *skb, tcp_synack_window(req) >> inet_rsk(req)->rcv_wscale, tcp_rsk_tsval(tcp_rsk(req)), req->ts_recent, sk->sk_bound_dev_if, - &key, ipv6_get_dsfield(ipv6_hdr(skb)), 0, + &key, ipv6_get_dsfield(ipv6_hdr(skb)) & ~INET_ECN_MASK, + 0, READ_ONCE(sk->sk_priority), READ_ONCE(tcp_rsk(req)->txhash)); if (tcp_key_is_ao(&key)) @@ -1742,6 +1749,7 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb) { struct net *net = dev_net_rcu(skb->dev); enum skb_drop_reason drop_reason; + enum tcp_tw_status tw_status; int sdif = inet6_sdif(skb); int dif = inet6_iif(skb); const struct tcphdr *th; @@ -1962,7 +1970,8 @@ do_time_wait: goto csum_error; } - switch (tcp_timewait_state_process(inet_twsk(sk), skb, th, &isn)) { + tw_status = tcp_timewait_state_process(inet_twsk(sk), skb, th, &isn); + switch (tw_status) { case TCP_TW_SYN: { struct sock *sk2; @@ -1987,7 +1996,8 @@ do_time_wait: /* to ACK */ fallthrough; case TCP_TW_ACK: - tcp_v6_timewait_ack(sk, skb); + case TCP_TW_ACK_OOW: + tcp_v6_timewait_ack(sk, skb, tw_status); break; case TCP_TW_RST: tcp_v6_send_reset(sk, skb, SK_RST_REASON_TCP_TIMEWAIT_SOCKET); -- cgit v1.2.3 From 9866884ce8ef25338c5b33cbb97c2b5d92088528 Mon Sep 17 00:00:00 2001 From: Ilpo Järvinen Date: Wed, 5 Mar 2025 23:38:52 +0100 Subject: tcp: Pass flags to __tcp_send_ack MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Accurate ECN needs to send custom flags to handle IP-ECN field reflection during handshake. Signed-off-by: Ilpo Järvinen Signed-off-by: Chia-Yu Chang Reviewed-by: Eric Dumazet Signed-off-by: David S. Miller --- include/net/tcp.h | 2 +- net/ipv4/bpf_tcp_ca.c | 2 +- net/ipv4/tcp_dctcp.h | 2 +- net/ipv4/tcp_output.c | 6 +++--- 4 files changed, 6 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/include/net/tcp.h b/include/net/tcp.h index c6bdd3d63159..d08fbf90495d 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -707,7 +707,7 @@ void tcp_send_active_reset(struct sock *sk, gfp_t priority, enum sk_rst_reason reason); int tcp_send_synack(struct sock *); void tcp_push_one(struct sock *, unsigned int mss_now); -void __tcp_send_ack(struct sock *sk, u32 rcv_nxt); +void __tcp_send_ack(struct sock *sk, u32 rcv_nxt, u16 flags); void tcp_send_ack(struct sock *sk); void tcp_send_delayed_ack(struct sock *sk); void tcp_send_loss_probe(struct sock *sk); diff --git a/net/ipv4/bpf_tcp_ca.c b/net/ipv4/bpf_tcp_ca.c index 554804774628..e01492234b0b 100644 --- a/net/ipv4/bpf_tcp_ca.c +++ b/net/ipv4/bpf_tcp_ca.c @@ -121,7 +121,7 @@ static int bpf_tcp_ca_btf_struct_access(struct bpf_verifier_log *log, BPF_CALL_2(bpf_tcp_send_ack, struct tcp_sock *, tp, u32, rcv_nxt) { /* bpf_tcp_ca prog cannot have NULL tp */ - __tcp_send_ack((struct sock *)tp, rcv_nxt); + __tcp_send_ack((struct sock *)tp, rcv_nxt, 0); return 0; } diff --git a/net/ipv4/tcp_dctcp.h b/net/ipv4/tcp_dctcp.h index d69a77cbd0c7..4b0259111d81 100644 --- a/net/ipv4/tcp_dctcp.h +++ b/net/ipv4/tcp_dctcp.h @@ -28,7 +28,7 @@ static inline void dctcp_ece_ack_update(struct sock *sk, enum tcp_ca_event evt, */ if (inet_csk(sk)->icsk_ack.pending & ICSK_ACK_TIMER) { dctcp_ece_ack_cwr(sk, *ce_state); - __tcp_send_ack(sk, *prior_rcv_nxt); + __tcp_send_ack(sk, *prior_rcv_nxt, 0); } inet_csk(sk)->icsk_ack.pending |= ICSK_ACK_NOW; } diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index fffbd3ee44ed..e0a4e5432399 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -4240,7 +4240,7 @@ void tcp_send_delayed_ack(struct sock *sk) } /* This routine sends an ack and also updates the window. */ -void __tcp_send_ack(struct sock *sk, u32 rcv_nxt) +void __tcp_send_ack(struct sock *sk, u32 rcv_nxt, u16 flags) { struct sk_buff *buff; @@ -4269,7 +4269,7 @@ void __tcp_send_ack(struct sock *sk, u32 rcv_nxt) /* Reserve space for headers and prepare control bits. */ skb_reserve(buff, MAX_TCP_HEADER); - tcp_init_nondata_skb(buff, tcp_acceptable_seq(sk), TCPHDR_ACK); + tcp_init_nondata_skb(buff, tcp_acceptable_seq(sk), TCPHDR_ACK | flags); /* We do not want pure acks influencing TCP Small Queues or fq/pacing * too much. @@ -4284,7 +4284,7 @@ EXPORT_SYMBOL_GPL(__tcp_send_ack); void tcp_send_ack(struct sock *sk) { - __tcp_send_ack(sk, tcp_sk(sk)->rcv_nxt); + __tcp_send_ack(sk, tcp_sk(sk)->rcv_nxt, 0); } /* This routine sends a packet with an out of date sequence -- cgit v1.2.3 From 82d3639ef7dc54d5b5cb454d9a13005202d7a701 Mon Sep 17 00:00:00 2001 From: Moshe Shemesh Date: Sun, 9 Mar 2025 20:07:42 +0200 Subject: net/mlx5: fs, add support for flow meters HWS action Add support for HW Steering action of flow meter range. Flow meters range can use one HWS action for the whole range. Thus, share a cached HWS action among rules that use same flow meter object range. Hold refcount for each rule using the cached action. Signed-off-by: Moshe Shemesh Reviewed-by: Mark Bloch Signed-off-by: Tariq Toukan Reviewed-by: Jacob Keller Link: https://patch.msgid.link/1741543663-22123-3-git-send-email-tariqt@nvidia.com Signed-off-by: Paolo Abeni --- .../net/ethernet/mellanox/mlx5/core/en/tc/meter.c | 5 + .../net/ethernet/mellanox/mlx5/core/en/tc/meter.h | 13 +++ .../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 1 + .../mellanox/mlx5/core/steering/hws/fs_hws.c | 120 +++++++++++++++++++-- .../mellanox/mlx5/core/steering/hws/fs_hws.h | 5 + include/linux/mlx5/fs.h | 1 + 6 files changed, 135 insertions(+), 10 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.c index 8218c892b161..7819fb297280 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.c @@ -593,3 +593,8 @@ mlx5e_tc_meter_get_stats(struct mlx5e_flow_meter_handle *meter, *drops = packets2; *lastuse = max_t(u64, lastuse1, lastuse2); } + +int mlx5e_flow_meter_get_base_id(struct mlx5e_flow_meter_handle *meter) +{ + return meter->meters_obj->base_id; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.h index 9b795cd106bb..d6afb6556875 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.h @@ -72,4 +72,17 @@ void mlx5e_tc_meter_get_stats(struct mlx5e_flow_meter_handle *meter, u64 *bytes, u64 *packets, u64 *drops, u64 *lastuse); +#if IS_ENABLED(CONFIG_MLX5_CLS_ACT) + +int mlx5e_flow_meter_get_base_id(struct mlx5e_flow_meter_handle *meter); + +#else /* CONFIG_MLX5_CLS_ACT */ + +static inline int +mlx5e_flow_meter_get_base_id(struct mlx5e_flow_meter_handle *meter) +{ + return 0; +} +#endif /* CONFIG_MLX5_CLS_ACT */ + #endif /* __MLX5_EN_FLOW_METER_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 8987f17e485c..a6a8eea5980c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -648,6 +648,7 @@ esw_setup_meter(struct mlx5_flow_attr *attr, struct mlx5_flow_act *flow_act) meter = attr->meter_attr.meter; flow_act->exe_aso.type = attr->exe_aso_type; flow_act->exe_aso.object_id = meter->obj_id; + flow_act->exe_aso.base_id = mlx5e_flow_meter_get_base_id(meter); flow_act->exe_aso.flow_meter.meter_idx = meter->idx; flow_act->exe_aso.flow_meter.init_color = MLX5_FLOW_METER_COLOR_GREEN; /* use metadata reg 5 for packet color */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/fs_hws.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/fs_hws.c index 50e43d574e0a..edf7db1e6e9c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/fs_hws.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/fs_hws.c @@ -66,6 +66,7 @@ static int mlx5_fs_init_hws_actions_pool(struct mlx5_core_dev *dev, xa_init(&hws_pool->table_dests); xa_init(&hws_pool->vport_dests); xa_init(&hws_pool->vport_vhca_dests); + xa_init(&hws_pool->aso_meters); return 0; cleanup_insert_hdr: @@ -88,10 +89,14 @@ destroy_tag: static void mlx5_fs_cleanup_hws_actions_pool(struct mlx5_fs_hws_context *fs_ctx) { struct mlx5_fs_hws_actions_pool *hws_pool = &fs_ctx->hws_pool; + struct mlx5_fs_hws_data *fs_hws_data; struct mlx5hws_action *action; struct mlx5_fs_pool *pool; unsigned long i; + xa_for_each(&hws_pool->aso_meters, i, fs_hws_data) + kfree(fs_hws_data); + xa_destroy(&hws_pool->aso_meters); xa_for_each(&hws_pool->vport_vhca_dests, i, action) mlx5hws_action_destroy(action); xa_destroy(&hws_pool->vport_vhca_dests); @@ -459,6 +464,70 @@ mlx5_fs_create_dest_action_range(struct mlx5hws_context *ctx, flags); } +static struct mlx5_fs_hws_data * +mlx5_fs_get_cached_hws_data(struct xarray *cache_xa, unsigned long index) +{ + struct mlx5_fs_hws_data *fs_hws_data; + int err; + + xa_lock(cache_xa); + fs_hws_data = xa_load(cache_xa, index); + if (!fs_hws_data) { + fs_hws_data = kzalloc(sizeof(*fs_hws_data), GFP_ATOMIC); + if (!fs_hws_data) { + xa_unlock(cache_xa); + return NULL; + } + refcount_set(&fs_hws_data->hws_action_refcount, 0); + mutex_init(&fs_hws_data->lock); + err = __xa_insert(cache_xa, index, fs_hws_data, GFP_ATOMIC); + if (err) { + kfree(fs_hws_data); + xa_unlock(cache_xa); + return NULL; + } + } + xa_unlock(cache_xa); + + return fs_hws_data; +} + +static struct mlx5hws_action * +mlx5_fs_get_action_aso_meter(struct mlx5_fs_hws_context *fs_ctx, + struct mlx5_exe_aso *exe_aso) +{ + struct mlx5_fs_hws_create_action_ctx create_ctx; + struct mlx5hws_context *ctx = fs_ctx->hws_ctx; + struct mlx5_fs_hws_data *meter_hws_data; + u32 id = exe_aso->base_id; + struct xarray *meters_xa; + + meters_xa = &fs_ctx->hws_pool.aso_meters; + meter_hws_data = mlx5_fs_get_cached_hws_data(meters_xa, id); + if (!meter_hws_data) + return NULL; + + create_ctx.hws_ctx = ctx; + create_ctx.actions_type = MLX5HWS_ACTION_TYP_ASO_METER; + create_ctx.id = id; + create_ctx.return_reg_id = exe_aso->return_reg_id; + + return mlx5_fs_get_hws_action(meter_hws_data, &create_ctx); +} + +static void mlx5_fs_put_action_aso_meter(struct mlx5_fs_hws_context *fs_ctx, + struct mlx5_exe_aso *exe_aso) +{ + struct mlx5_fs_hws_data *meter_hws_data; + struct xarray *meters_xa; + + meters_xa = &fs_ctx->hws_pool.aso_meters; + meter_hws_data = xa_load(meters_xa, exe_aso->base_id); + if (!meter_hws_data) + return; + return mlx5_fs_put_hws_action(meter_hws_data); +} + static struct mlx5hws_action * mlx5_fs_create_action_dest_array(struct mlx5hws_context *ctx, struct mlx5hws_action_dest_attr *dests, @@ -528,6 +597,11 @@ mlx5_fs_create_hws_action(struct mlx5_fs_hws_create_action_ctx *create_ctx) case MLX5HWS_ACTION_TYP_CTR: return mlx5hws_action_create_counter(create_ctx->hws_ctx, create_ctx->id, flags); + case MLX5HWS_ACTION_TYP_ASO_METER: + return mlx5hws_action_create_aso_meter(create_ctx->hws_ctx, + create_ctx->id, + create_ctx->return_reg_id, + flags); default: return NULL; } @@ -576,26 +650,33 @@ void mlx5_fs_put_hws_action(struct mlx5_fs_hws_data *fs_hws_data) mutex_unlock(&fs_hws_data->lock); } -static void mlx5_fs_destroy_fs_action(struct mlx5_fs_hws_rule_action *fs_action) +static void mlx5_fs_destroy_fs_action(struct mlx5_flow_root_namespace *ns, + struct mlx5_fs_hws_rule_action *fs_action) { + struct mlx5_fs_hws_context *fs_ctx = &ns->fs_hws_context; + switch (mlx5hws_action_get_type(fs_action->action)) { case MLX5HWS_ACTION_TYP_CTR: mlx5_fc_put_hws_action(fs_action->counter); break; + case MLX5HWS_ACTION_TYP_ASO_METER: + mlx5_fs_put_action_aso_meter(fs_ctx, fs_action->exe_aso); + break; default: mlx5hws_action_destroy(fs_action->action); } } static void -mlx5_fs_destroy_fs_actions(struct mlx5_fs_hws_rule_action **fs_actions, +mlx5_fs_destroy_fs_actions(struct mlx5_flow_root_namespace *ns, + struct mlx5_fs_hws_rule_action **fs_actions, int *num_fs_actions) { int i; /* Free in reverse order to handle action dependencies */ for (i = *num_fs_actions - 1; i >= 0; i--) - mlx5_fs_destroy_fs_action(*fs_actions + i); + mlx5_fs_destroy_fs_action(ns, *fs_actions + i); *num_fs_actions = 0; kfree(*fs_actions); *fs_actions = NULL; @@ -792,8 +873,25 @@ static int mlx5_fs_fte_get_hws_actions(struct mlx5_flow_root_namespace *ns, } if (fte_action->action & MLX5_FLOW_CONTEXT_ACTION_EXECUTE_ASO) { - err = -EOPNOTSUPP; - goto free_actions; + if (fte_action->exe_aso.type != MLX5_EXE_ASO_FLOW_METER || + num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) { + err = -EOPNOTSUPP; + goto free_actions; + } + + tmp_action = mlx5_fs_get_action_aso_meter(fs_ctx, + &fte_action->exe_aso); + if (!tmp_action) { + err = -ENOMEM; + goto free_actions; + } + (*ractions)[num_actions].aso_meter.offset = + fte_action->exe_aso.flow_meter.meter_idx; + (*ractions)[num_actions].aso_meter.init_color = + fte_action->exe_aso.flow_meter.init_color; + (*ractions)[num_actions++].action = tmp_action; + fs_actions[num_fs_actions].action = tmp_action; + fs_actions[num_fs_actions++].exe_aso = &fte_action->exe_aso; } if (fte_action->action & MLX5_FLOW_CONTEXT_ACTION_DROP) { @@ -907,7 +1005,7 @@ static int mlx5_fs_fte_get_hws_actions(struct mlx5_flow_root_namespace *ns, return 0; free_actions: - mlx5_fs_destroy_fs_actions(&fs_actions, &num_fs_actions); + mlx5_fs_destroy_fs_actions(ns, &fs_actions, &num_fs_actions); free_dest_actions_alloc: kfree(dest_actions); free_fs_actions_alloc: @@ -957,7 +1055,7 @@ static int mlx5_cmd_hws_create_fte(struct mlx5_flow_root_namespace *ns, return 0; free_actions: - mlx5_fs_destroy_fs_actions(&fte->fs_hws_rule.hws_fs_actions, + mlx5_fs_destroy_fs_actions(ns, &fte->fs_hws_rule.hws_fs_actions, &fte->fs_hws_rule.num_fs_actions); out_err: mlx5_core_err(ns->dev, "Failed to create hws rule err(%d)\n", err); @@ -977,7 +1075,8 @@ static int mlx5_cmd_hws_delete_fte(struct mlx5_flow_root_namespace *ns, err = mlx5hws_bwc_rule_destroy(rule->bwc_rule); rule->bwc_rule = NULL; - mlx5_fs_destroy_fs_actions(&rule->hws_fs_actions, &rule->num_fs_actions); + mlx5_fs_destroy_fs_actions(ns, &rule->hws_fs_actions, + &rule->num_fs_actions); return err; } @@ -1015,11 +1114,12 @@ static int mlx5_cmd_hws_update_fte(struct mlx5_flow_root_namespace *ns, if (ret) goto restore_actions; - mlx5_fs_destroy_fs_actions(&saved_hws_fs_actions, &saved_num_fs_actions); + mlx5_fs_destroy_fs_actions(ns, &saved_hws_fs_actions, + &saved_num_fs_actions); return ret; restore_actions: - mlx5_fs_destroy_fs_actions(&fte->fs_hws_rule.hws_fs_actions, + mlx5_fs_destroy_fs_actions(ns, &fte->fs_hws_rule.hws_fs_actions, &fte->fs_hws_rule.num_fs_actions); fte->fs_hws_rule.hws_fs_actions = saved_hws_fs_actions; fte->fs_hws_rule.num_fs_actions = saved_num_fs_actions; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/fs_hws.h b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/fs_hws.h index 1c4b853cf88d..6970b1aa9540 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/fs_hws.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/fs_hws.h @@ -22,6 +22,7 @@ struct mlx5_fs_hws_actions_pool { struct xarray table_dests; struct xarray vport_vhca_dests; struct xarray vport_dests; + struct xarray aso_meters; }; struct mlx5_fs_hws_context { @@ -49,6 +50,7 @@ struct mlx5_fs_hws_rule_action { struct mlx5hws_action *action; union { struct mlx5_fc *counter; + struct mlx5_exe_aso *exe_aso; }; }; @@ -68,6 +70,9 @@ struct mlx5_fs_hws_create_action_ctx { enum mlx5hws_action_type actions_type; struct mlx5hws_context *hws_ctx; u32 id; + union { + u8 return_reg_id; + }; }; struct mlx5hws_action * diff --git a/include/linux/mlx5/fs.h b/include/linux/mlx5/fs.h index fd62b2b1611d..939e58c2f386 100644 --- a/include/linux/mlx5/fs.h +++ b/include/linux/mlx5/fs.h @@ -244,6 +244,7 @@ void mlx5_destroy_flow_group(struct mlx5_flow_group *fg); struct mlx5_exe_aso { u32 object_id; + int base_id; u8 type; u8 return_reg_id; union { -- cgit v1.2.3 From 43e2aa56aea2e292c209cfd77c3f2d8553fb66e9 Mon Sep 17 00:00:00 2001 From: Heiner Kallweit Date: Sun, 9 Mar 2025 21:04:14 +0100 Subject: net: phy: move PHY package MMD access function declarations from phy.h to phylib.h These functions are used by PHY drivers only, therefore move their declaration to phylib.h. Signed-off-by: Heiner Kallweit Reviewed-by: Simon Horman Link: https://patch.msgid.link/406c8a20-b62e-4ee3-b174-b566724a0876@gmail.com Signed-off-by: Paolo Abeni --- drivers/net/phy/phylib.h | 6 ++++++ include/linux/phy.h | 8 -------- 2 files changed, 6 insertions(+), 8 deletions(-) (limited to 'include') diff --git a/drivers/net/phy/phylib.h b/drivers/net/phy/phylib.h index f0e499fed1f7..c15484a805b3 100644 --- a/drivers/net/phy/phylib.h +++ b/drivers/net/phy/phylib.h @@ -15,6 +15,12 @@ int __phy_package_read(struct phy_device *phydev, unsigned int addr_offset, u32 regnum); int __phy_package_write(struct phy_device *phydev, unsigned int addr_offset, u32 regnum, u16 val); +int __phy_package_read_mmd(struct phy_device *phydev, + unsigned int addr_offset, int devad, + u32 regnum); +int __phy_package_write_mmd(struct phy_device *phydev, + unsigned int addr_offset, int devad, + u32 regnum, u16 val); bool phy_package_init_once(struct phy_device *phydev); bool phy_package_probe_once(struct phy_device *phydev); int phy_package_join(struct phy_device *phydev, int base_addr, size_t priv_size); diff --git a/include/linux/phy.h b/include/linux/phy.h index c4a6385faf41..fc028bab10b7 100644 --- a/include/linux/phy.h +++ b/include/linux/phy.h @@ -2107,18 +2107,10 @@ int __phy_hwtstamp_set(struct phy_device *phydev, struct kernel_hwtstamp_config *config, struct netlink_ext_ack *extack); -int __phy_package_read_mmd(struct phy_device *phydev, - unsigned int addr_offset, int devad, - u32 regnum); - int phy_package_read_mmd(struct phy_device *phydev, unsigned int addr_offset, int devad, u32 regnum); -int __phy_package_write_mmd(struct phy_device *phydev, - unsigned int addr_offset, int devad, - u32 regnum, u16 val); - int phy_package_write_mmd(struct phy_device *phydev, unsigned int addr_offset, int devad, u32 regnum, u16 val); -- cgit v1.2.3 From 8ea221b22172a2cc5d0edbfec4b34ef3fe8de167 Mon Sep 17 00:00:00 2001 From: Heiner Kallweit Date: Sun, 9 Mar 2025 21:05:08 +0100 Subject: net: phy: remove unused functions phy_package_[read|write]_mmd These functions have never had a user, so remove them. Signed-off-by: Heiner Kallweit Reviewed-by: Simon Horman Link: https://patch.msgid.link/5792e2cd-6f0a-4f7d-a5ef-b932f94d82f3@gmail.com Signed-off-by: Paolo Abeni --- drivers/net/phy/phy-core.c | 75 ---------------------------------------------- include/linux/phy.h | 8 ----- 2 files changed, 83 deletions(-) (limited to 'include') diff --git a/drivers/net/phy/phy-core.c b/drivers/net/phy/phy-core.c index 154d29be6293..31ed1ce3b62f 100644 --- a/drivers/net/phy/phy-core.c +++ b/drivers/net/phy/phy-core.c @@ -734,43 +734,6 @@ int __phy_package_read_mmd(struct phy_device *phydev, } EXPORT_SYMBOL(__phy_package_read_mmd); -/** - * phy_package_read_mmd - read MMD reg relative to PHY package base addr - * @phydev: The phy_device struct - * @addr_offset: The offset to be added to PHY package base_addr - * @devad: The MMD to read from - * @regnum: The register on the MMD to read - * - * Convenience helper for reading a register of an MMD on a given PHY - * using the PHY package base address. The base address is added to - * the addr_offset value. - * - * Same calling rules as for phy_read(); - * - * NOTE: It's assumed that the entire PHY package is either C22 or C45. - */ -int phy_package_read_mmd(struct phy_device *phydev, - unsigned int addr_offset, int devad, - u32 regnum) -{ - int addr = phy_package_address(phydev, addr_offset); - int val; - - if (addr < 0) - return addr; - - if (regnum > (u16)~0 || devad > 32) - return -EINVAL; - - phy_lock_mdio_bus(phydev); - val = mmd_phy_read(phydev->mdio.bus, addr, phydev->is_c45, devad, - regnum); - phy_unlock_mdio_bus(phydev); - - return val; -} -EXPORT_SYMBOL(phy_package_read_mmd); - /** * __phy_package_write_mmd - write MMD reg relative to PHY package base addr * @phydev: The phy_device struct @@ -804,44 +767,6 @@ int __phy_package_write_mmd(struct phy_device *phydev, } EXPORT_SYMBOL(__phy_package_write_mmd); -/** - * phy_package_write_mmd - write MMD reg relative to PHY package base addr - * @phydev: The phy_device struct - * @addr_offset: The offset to be added to PHY package base_addr - * @devad: The MMD to write to - * @regnum: The register on the MMD to write - * @val: value to write to @regnum - * - * Convenience helper for writing a register of an MMD on a given PHY - * using the PHY package base address. The base address is added to - * the addr_offset value. - * - * Same calling rules as for phy_write(); - * - * NOTE: It's assumed that the entire PHY package is either C22 or C45. - */ -int phy_package_write_mmd(struct phy_device *phydev, - unsigned int addr_offset, int devad, - u32 regnum, u16 val) -{ - int addr = phy_package_address(phydev, addr_offset); - int ret; - - if (addr < 0) - return addr; - - if (regnum > (u16)~0 || devad > 32) - return -EINVAL; - - phy_lock_mdio_bus(phydev); - ret = mmd_phy_write(phydev->mdio.bus, addr, phydev->is_c45, devad, - regnum, val); - phy_unlock_mdio_bus(phydev); - - return ret; -} -EXPORT_SYMBOL(phy_package_write_mmd); - /** * phy_modify_changed - Function for modifying a PHY register * @phydev: the phy_device struct diff --git a/include/linux/phy.h b/include/linux/phy.h index fc028bab10b7..61a8cb9d1247 100644 --- a/include/linux/phy.h +++ b/include/linux/phy.h @@ -2107,14 +2107,6 @@ int __phy_hwtstamp_set(struct phy_device *phydev, struct kernel_hwtstamp_config *config, struct netlink_ext_ack *extack); -int phy_package_read_mmd(struct phy_device *phydev, - unsigned int addr_offset, int devad, - u32 regnum); - -int phy_package_write_mmd(struct phy_device *phydev, - unsigned int addr_offset, int devad, - u32 regnum, u16 val); - extern const struct bus_type mdio_bus_type; struct mdio_board_info { -- cgit v1.2.3 From 35b862eac9099bb492229627c44e541633fb5938 Mon Sep 17 00:00:00 2001 From: "Russell King (Oracle)" Date: Mon, 10 Mar 2025 11:10:52 +0000 Subject: net: phylink: expand on .pcs_config() method documentation Expand on the requirements of the .pcs_config() method documentation, specifically mentioning that it should cause minimal disruption to an established link, and that it should return a positive non-zero value when requiring the .pcs_an_restart() method to be called. Signed-off-by: Russell King (Oracle) Link: https://patch.msgid.link/E1trb24-005oVq-Is@rmk-PC.armlinux.org.uk Signed-off-by: Paolo Abeni --- include/linux/phylink.h | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'include') diff --git a/include/linux/phylink.h b/include/linux/phylink.h index c187267a15b6..79876c84ae81 100644 --- a/include/linux/phylink.h +++ b/include/linux/phylink.h @@ -595,6 +595,14 @@ void pcs_get_state(struct phylink_pcs *pcs, unsigned int neg_mode, * The %neg_mode argument should be tested via the phylink_mode_*() family of * functions, or for PCS that set pcs->neg_mode true, should be tested * against the PHYLINK_PCS_NEG_* definitions. + * + * pcs_config() will be called when configuration of the PCS is required + * or when the advertisement is possibly updated. It must not unnecessarily + * disrupt an established link. + * + * When an autonegotiation restart is required for 802.3z modes, .pcs_config() + * should return a positive non-zero integer (e.g. 1) to indicate to phylink + * to call the pcs_an_restart() method. */ int pcs_config(struct phylink_pcs *pcs, unsigned int neg_mode, phy_interface_t interface, const unsigned long *advertising, -- cgit v1.2.3 From 79f88a584e35133359c394506b351a60230cf37b Mon Sep 17 00:00:00 2001 From: Maxime Chevallier Date: Fri, 7 Mar 2025 18:35:58 +0100 Subject: net: ethtool: Export the link_mode_params definitions link_mode_params contains a lookup table of all 802.3 link modes that are currently supported with structured data about each mode's speed, duplex, number of lanes and mediums. As a preparation for a port representation, export that table for the rest of the net stack to use. Signed-off-by: Maxime Chevallier Link: https://patch.msgid.link/20250307173611.129125-2-maxime.chevallier@bootlin.com Signed-off-by: Paolo Abeni --- include/linux/ethtool.h | 8 ++++++++ net/ethtool/common.c | 1 + net/ethtool/common.h | 7 ------- 3 files changed, 9 insertions(+), 7 deletions(-) (limited to 'include') diff --git a/include/linux/ethtool.h b/include/linux/ethtool.h index 7f222dccc7d1..8210ece94fa6 100644 --- a/include/linux/ethtool.h +++ b/include/linux/ethtool.h @@ -210,6 +210,14 @@ static inline u8 *ethtool_rxfh_context_key(struct ethtool_rxfh_context *ctx) void ethtool_rxfh_context_lost(struct net_device *dev, u32 context_id); +struct link_mode_info { + int speed; + u8 lanes; + u8 duplex; +}; + +extern const struct link_mode_info link_mode_params[]; + /* declare a link mode bitmap */ #define __ETHTOOL_DECLARE_LINK_MODE_MASK(name) \ DECLARE_BITMAP(name, __ETHTOOL_LINK_MODE_MASK_NBITS) diff --git a/net/ethtool/common.c b/net/ethtool/common.c index 1c6f2a2f1871..7e3c16856c1a 100644 --- a/net/ethtool/common.c +++ b/net/ethtool/common.c @@ -424,6 +424,7 @@ const struct link_mode_info link_mode_params[] = { __DEFINE_LINK_MODE_PARAMS(800000, VR4, Full), }; static_assert(ARRAY_SIZE(link_mode_params) == __ETHTOOL_LINK_MODE_MASK_NBITS); +EXPORT_SYMBOL_GPL(link_mode_params); const char netif_msg_class_names[][ETH_GSTRING_LEN] = { [NETIF_MSG_DRV_BIT] = "drv", diff --git a/net/ethtool/common.h b/net/ethtool/common.h index a1088c2441d0..b4683d286a5a 100644 --- a/net/ethtool/common.h +++ b/net/ethtool/common.h @@ -15,12 +15,6 @@ #define __SOF_TIMESTAMPING_CNT (const_ilog2(SOF_TIMESTAMPING_LAST) + 1) #define __HWTSTAMP_FLAG_CNT (const_ilog2(HWTSTAMP_FLAG_LAST) + 1) -struct link_mode_info { - int speed; - u8 lanes; - u8 duplex; -}; - struct genl_info; struct hwtstamp_provider_desc; @@ -33,7 +27,6 @@ tunable_strings[__ETHTOOL_TUNABLE_COUNT][ETH_GSTRING_LEN]; extern const char phy_tunable_strings[__ETHTOOL_PHY_TUNABLE_COUNT][ETH_GSTRING_LEN]; extern const char link_mode_names[][ETH_GSTRING_LEN]; -extern const struct link_mode_info link_mode_params[]; extern const char netif_msg_class_names[][ETH_GSTRING_LEN]; extern const char wol_mode_names[][ETH_GSTRING_LEN]; extern const char sof_timestamping_names[][ETH_GSTRING_LEN]; -- cgit v1.2.3 From 8c8c4a87933dd924f9fb56dfd35bae7e8f30a4b5 Mon Sep 17 00:00:00 2001 From: Maxime Chevallier Date: Fri, 7 Mar 2025 18:36:00 +0100 Subject: net: phy: phy_caps: Move phy_speeds to phy_caps Use the newly introduced link_capabilities array to derive the list of possible speeds when given a combination of linkmodes. As link_capabilities is indexed by speed, we don't have to iterate the whole phy_settings array. Signed-off-by: Maxime Chevallier Link: https://patch.msgid.link/20250307173611.129125-4-maxime.chevallier@bootlin.com Signed-off-by: Paolo Abeni --- drivers/net/phy/phy-caps.h | 3 +++ drivers/net/phy/phy-core.c | 15 --------------- drivers/net/phy/phy.c | 3 ++- drivers/net/phy/phy_caps.c | 33 +++++++++++++++++++++++++++++++++ include/linux/phy.h | 2 -- 5 files changed, 38 insertions(+), 18 deletions(-) (limited to 'include') diff --git a/drivers/net/phy/phy-caps.h b/drivers/net/phy/phy-caps.h index 6024f1d11a93..5a5f4ee3006b 100644 --- a/drivers/net/phy/phy-caps.h +++ b/drivers/net/phy/phy-caps.h @@ -40,4 +40,7 @@ struct link_capabilities { int phy_caps_init(void); +size_t phy_caps_speeds(unsigned int *speeds, size_t size, + unsigned long *linkmodes); + #endif /* __PHY_CAPS_H */ diff --git a/drivers/net/phy/phy-core.c b/drivers/net/phy/phy-core.c index 31ed1ce3b62f..102834b11de0 100644 --- a/drivers/net/phy/phy-core.c +++ b/drivers/net/phy/phy-core.c @@ -340,21 +340,6 @@ phy_lookup_setting(int speed, int duplex, const unsigned long *mask, bool exact) } EXPORT_SYMBOL_GPL(phy_lookup_setting); -size_t phy_speeds(unsigned int *speeds, size_t size, - unsigned long *mask) -{ - size_t count; - int i; - - for (i = 0, count = 0; i < ARRAY_SIZE(settings) && count < size; i++) - if (settings[i].bit < __ETHTOOL_LINK_MODE_MASK_NBITS && - test_bit(settings[i].bit, mask) && - (count == 0 || speeds[count - 1] != settings[i].speed)) - speeds[count++] = settings[i].speed; - - return count; -} - static void __set_linkmode_max_speed(u32 max_speed, unsigned long *addr) { const struct phy_setting *p; diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c index 16ffc00b419c..3128df03feda 100644 --- a/drivers/net/phy/phy.c +++ b/drivers/net/phy/phy.c @@ -37,6 +37,7 @@ #include #include "phylib-internal.h" +#include "phy-caps.h" #define PHY_STATE_TIME HZ @@ -245,7 +246,7 @@ unsigned int phy_supported_speeds(struct phy_device *phy, unsigned int *speeds, unsigned int size) { - return phy_speeds(speeds, size, phy->supported); + return phy_caps_speeds(speeds, size, phy->supported); } /** diff --git a/drivers/net/phy/phy_caps.c b/drivers/net/phy/phy_caps.c index 6cb18e216d97..8ce91257160f 100644 --- a/drivers/net/phy/phy_caps.c +++ b/drivers/net/phy/phy_caps.c @@ -57,6 +57,9 @@ static int speed_duplex_to_capa(int speed, unsigned int duplex) return -EINVAL; } +#define for_each_link_caps_asc_speed(cap) \ + for (cap = link_caps; cap < &link_caps[__LINK_CAPA_MAX]; cap++) + /** * phy_caps_init() - Initializes the link_caps array from the link_mode_params. * @@ -88,3 +91,33 @@ int phy_caps_init(void) return 0; } + +/** + * phy_caps_speeds() - Fill an array of supported SPEED_* values for given modes + * @speeds: Output array to store the speeds list into + * @size: Size of the output array + * @linkmodes: Linkmodes to get the speeds from + * + * Fills the speeds array with all possible speeds that can be achieved with + * the specified linkmodes. + * + * Returns: The number of speeds filled into the array. If the input array isn't + * big enough to store all speeds, fill it as much as possible. + */ +size_t phy_caps_speeds(unsigned int *speeds, size_t size, + unsigned long *linkmodes) +{ + struct link_capabilities *lcap; + size_t count = 0; + + for_each_link_caps_asc_speed(lcap) { + if (linkmode_intersects(lcap->linkmodes, linkmodes) && + (count == 0 || speeds[count - 1] != lcap->speed)) { + speeds[count++] = lcap->speed; + if (count >= size) + break; + } + } + + return count; +} diff --git a/include/linux/phy.h b/include/linux/phy.h index 61a8cb9d1247..83c50bb21939 100644 --- a/include/linux/phy.h +++ b/include/linux/phy.h @@ -1287,8 +1287,6 @@ struct phy_setting { const struct phy_setting * phy_lookup_setting(int speed, int duplex, const unsigned long *mask, bool exact); -size_t phy_speeds(unsigned int *speeds, size_t size, - unsigned long *mask); /** * phy_is_started - Convenience function to check whether PHY is started -- cgit v1.2.3 From ce60fef7feccb5c71d5b49e489d24db7d79c2ac7 Mon Sep 17 00:00:00 2001 From: Maxime Chevallier Date: Fri, 7 Mar 2025 18:36:07 +0100 Subject: net: phy: drop phy_settings and the associated lookup helpers The phy_settings array is no longer relevant as it has now been replaced by the link_caps array and associated phy_caps helpers. Signed-off-by: Maxime Chevallier Link: https://patch.msgid.link/20250307173611.129125-11-maxime.chevallier@bootlin.com Signed-off-by: Paolo Abeni --- drivers/net/phy/phy-core.c | 184 --------------------------------------------- include/linux/phy.h | 13 ---- 2 files changed, 197 deletions(-) (limited to 'include') diff --git a/drivers/net/phy/phy-core.c b/drivers/net/phy/phy-core.c index a1caa0f313ec..e177037f9110 100644 --- a/drivers/net/phy/phy-core.c +++ b/drivers/net/phy/phy-core.c @@ -157,190 +157,6 @@ int phy_interface_num_ports(phy_interface_t interface) } EXPORT_SYMBOL_GPL(phy_interface_num_ports); -/* A mapping of all SUPPORTED settings to speed/duplex. This table - * must be grouped by speed and sorted in descending match priority - * - iow, descending speed. - */ - -#define PHY_SETTING(s, d, b) { .speed = SPEED_ ## s, .duplex = DUPLEX_ ## d, \ - .bit = ETHTOOL_LINK_MODE_ ## b ## _BIT} - -static const struct phy_setting settings[] = { - /* 800G */ - PHY_SETTING( 800000, FULL, 800000baseCR8_Full ), - PHY_SETTING( 800000, FULL, 800000baseKR8_Full ), - PHY_SETTING( 800000, FULL, 800000baseDR8_Full ), - PHY_SETTING( 800000, FULL, 800000baseDR8_2_Full ), - PHY_SETTING( 800000, FULL, 800000baseSR8_Full ), - PHY_SETTING( 800000, FULL, 800000baseVR8_Full ), - PHY_SETTING( 800000, FULL, 800000baseCR4_Full ), - PHY_SETTING( 800000, FULL, 800000baseKR4_Full ), - PHY_SETTING( 800000, FULL, 800000baseDR4_Full ), - PHY_SETTING( 800000, FULL, 800000baseDR4_2_Full ), - PHY_SETTING( 800000, FULL, 800000baseSR4_Full ), - PHY_SETTING( 800000, FULL, 800000baseVR4_Full ), - /* 400G */ - PHY_SETTING( 400000, FULL, 400000baseCR8_Full ), - PHY_SETTING( 400000, FULL, 400000baseKR8_Full ), - PHY_SETTING( 400000, FULL, 400000baseLR8_ER8_FR8_Full ), - PHY_SETTING( 400000, FULL, 400000baseDR8_Full ), - PHY_SETTING( 400000, FULL, 400000baseSR8_Full ), - PHY_SETTING( 400000, FULL, 400000baseCR4_Full ), - PHY_SETTING( 400000, FULL, 400000baseKR4_Full ), - PHY_SETTING( 400000, FULL, 400000baseLR4_ER4_FR4_Full ), - PHY_SETTING( 400000, FULL, 400000baseDR4_Full ), - PHY_SETTING( 400000, FULL, 400000baseSR4_Full ), - PHY_SETTING( 400000, FULL, 400000baseCR2_Full ), - PHY_SETTING( 400000, FULL, 400000baseKR2_Full ), - PHY_SETTING( 400000, FULL, 400000baseDR2_Full ), - PHY_SETTING( 400000, FULL, 400000baseDR2_2_Full ), - PHY_SETTING( 400000, FULL, 400000baseSR2_Full ), - PHY_SETTING( 400000, FULL, 400000baseVR2_Full ), - /* 200G */ - PHY_SETTING( 200000, FULL, 200000baseCR4_Full ), - PHY_SETTING( 200000, FULL, 200000baseKR4_Full ), - PHY_SETTING( 200000, FULL, 200000baseLR4_ER4_FR4_Full ), - PHY_SETTING( 200000, FULL, 200000baseDR4_Full ), - PHY_SETTING( 200000, FULL, 200000baseSR4_Full ), - PHY_SETTING( 200000, FULL, 200000baseCR2_Full ), - PHY_SETTING( 200000, FULL, 200000baseKR2_Full ), - PHY_SETTING( 200000, FULL, 200000baseLR2_ER2_FR2_Full ), - PHY_SETTING( 200000, FULL, 200000baseDR2_Full ), - PHY_SETTING( 200000, FULL, 200000baseSR2_Full ), - PHY_SETTING( 200000, FULL, 200000baseCR_Full ), - PHY_SETTING( 200000, FULL, 200000baseKR_Full ), - PHY_SETTING( 200000, FULL, 200000baseDR_Full ), - PHY_SETTING( 200000, FULL, 200000baseDR_2_Full ), - PHY_SETTING( 200000, FULL, 200000baseSR_Full ), - PHY_SETTING( 200000, FULL, 200000baseVR_Full ), - /* 100G */ - PHY_SETTING( 100000, FULL, 100000baseCR4_Full ), - PHY_SETTING( 100000, FULL, 100000baseKR4_Full ), - PHY_SETTING( 100000, FULL, 100000baseLR4_ER4_Full ), - PHY_SETTING( 100000, FULL, 100000baseSR4_Full ), - PHY_SETTING( 100000, FULL, 100000baseCR2_Full ), - PHY_SETTING( 100000, FULL, 100000baseKR2_Full ), - PHY_SETTING( 100000, FULL, 100000baseLR2_ER2_FR2_Full ), - PHY_SETTING( 100000, FULL, 100000baseDR2_Full ), - PHY_SETTING( 100000, FULL, 100000baseSR2_Full ), - PHY_SETTING( 100000, FULL, 100000baseCR_Full ), - PHY_SETTING( 100000, FULL, 100000baseKR_Full ), - PHY_SETTING( 100000, FULL, 100000baseLR_ER_FR_Full ), - PHY_SETTING( 100000, FULL, 100000baseDR_Full ), - PHY_SETTING( 100000, FULL, 100000baseSR_Full ), - /* 56G */ - PHY_SETTING( 56000, FULL, 56000baseCR4_Full ), - PHY_SETTING( 56000, FULL, 56000baseKR4_Full ), - PHY_SETTING( 56000, FULL, 56000baseLR4_Full ), - PHY_SETTING( 56000, FULL, 56000baseSR4_Full ), - /* 50G */ - PHY_SETTING( 50000, FULL, 50000baseCR2_Full ), - PHY_SETTING( 50000, FULL, 50000baseKR2_Full ), - PHY_SETTING( 50000, FULL, 50000baseSR2_Full ), - PHY_SETTING( 50000, FULL, 50000baseCR_Full ), - PHY_SETTING( 50000, FULL, 50000baseKR_Full ), - PHY_SETTING( 50000, FULL, 50000baseLR_ER_FR_Full ), - PHY_SETTING( 50000, FULL, 50000baseDR_Full ), - PHY_SETTING( 50000, FULL, 50000baseSR_Full ), - /* 40G */ - PHY_SETTING( 40000, FULL, 40000baseCR4_Full ), - PHY_SETTING( 40000, FULL, 40000baseKR4_Full ), - PHY_SETTING( 40000, FULL, 40000baseLR4_Full ), - PHY_SETTING( 40000, FULL, 40000baseSR4_Full ), - /* 25G */ - PHY_SETTING( 25000, FULL, 25000baseCR_Full ), - PHY_SETTING( 25000, FULL, 25000baseKR_Full ), - PHY_SETTING( 25000, FULL, 25000baseSR_Full ), - /* 20G */ - PHY_SETTING( 20000, FULL, 20000baseKR2_Full ), - PHY_SETTING( 20000, FULL, 20000baseMLD2_Full ), - /* 10G */ - PHY_SETTING( 10000, FULL, 10000baseCR_Full ), - PHY_SETTING( 10000, FULL, 10000baseER_Full ), - PHY_SETTING( 10000, FULL, 10000baseKR_Full ), - PHY_SETTING( 10000, FULL, 10000baseKX4_Full ), - PHY_SETTING( 10000, FULL, 10000baseLR_Full ), - PHY_SETTING( 10000, FULL, 10000baseLRM_Full ), - PHY_SETTING( 10000, FULL, 10000baseR_FEC ), - PHY_SETTING( 10000, FULL, 10000baseSR_Full ), - PHY_SETTING( 10000, FULL, 10000baseT_Full ), - /* 5G */ - PHY_SETTING( 5000, FULL, 5000baseT_Full ), - /* 2.5G */ - PHY_SETTING( 2500, FULL, 2500baseT_Full ), - PHY_SETTING( 2500, FULL, 2500baseX_Full ), - /* 1G */ - PHY_SETTING( 1000, FULL, 1000baseT_Full ), - PHY_SETTING( 1000, HALF, 1000baseT_Half ), - PHY_SETTING( 1000, FULL, 1000baseT1_Full ), - PHY_SETTING( 1000, FULL, 1000baseX_Full ), - PHY_SETTING( 1000, FULL, 1000baseKX_Full ), - /* 100M */ - PHY_SETTING( 100, FULL, 100baseT_Full ), - PHY_SETTING( 100, FULL, 100baseT1_Full ), - PHY_SETTING( 100, HALF, 100baseT_Half ), - PHY_SETTING( 100, HALF, 100baseFX_Half ), - PHY_SETTING( 100, FULL, 100baseFX_Full ), - /* 10M */ - PHY_SETTING( 10, FULL, 10baseT_Full ), - PHY_SETTING( 10, HALF, 10baseT_Half ), - PHY_SETTING( 10, FULL, 10baseT1L_Full ), - PHY_SETTING( 10, FULL, 10baseT1S_Full ), - PHY_SETTING( 10, HALF, 10baseT1S_Half ), - PHY_SETTING( 10, HALF, 10baseT1S_P2MP_Half ), - PHY_SETTING( 10, FULL, 10baseT1BRR_Full ), -}; -#undef PHY_SETTING - -/** - * phy_lookup_setting - lookup a PHY setting - * @speed: speed to match - * @duplex: duplex to match - * @mask: allowed link modes - * @exact: an exact match is required - * - * Search the settings array for a setting that matches the speed and - * duplex, and which is supported. - * - * If @exact is unset, either an exact match or %NULL for no match will - * be returned. - * - * If @exact is set, an exact match, the fastest supported setting at - * or below the specified speed, the slowest supported setting, or if - * they all fail, %NULL will be returned. - */ -const struct phy_setting * -phy_lookup_setting(int speed, int duplex, const unsigned long *mask, bool exact) -{ - const struct phy_setting *p, *match = NULL, *last = NULL; - int i; - - for (i = 0, p = settings; i < ARRAY_SIZE(settings); i++, p++) { - if (p->bit < __ETHTOOL_LINK_MODE_MASK_NBITS && - test_bit(p->bit, mask)) { - last = p; - if (p->speed == speed && p->duplex == duplex) { - /* Exact match for speed and duplex */ - match = p; - break; - } else if (!exact) { - if (!match && p->speed <= speed) - /* Candidate */ - match = p; - - if (p->speed < speed) - break; - } - } - } - - if (!match && !exact) - match = last; - - return match; -} -EXPORT_SYMBOL_GPL(phy_lookup_setting); - static void __set_phy_supported(struct phy_device *phydev, u32 max_speed) { phy_caps_linkmode_max_speed(max_speed, phydev->supported); diff --git a/include/linux/phy.h b/include/linux/phy.h index 83c50bb21939..c24e1a565819 100644 --- a/include/linux/phy.h +++ b/include/linux/phy.h @@ -1275,19 +1275,6 @@ const char *phy_rate_matching_to_str(int rate_matching); int phy_interface_num_ports(phy_interface_t interface); -/* A structure for mapping a particular speed and duplex - * combination to a particular SUPPORTED and ADVERTISED value - */ -struct phy_setting { - u32 speed; - u8 duplex; - u8 bit; -}; - -const struct phy_setting * -phy_lookup_setting(int speed, int duplex, const unsigned long *mask, - bool exact); - /** * phy_is_started - Convenience function to check whether PHY is started * @phydev: The phy_device struct -- cgit v1.2.3