From 5ca60b86f57a4d9648f68418a725b3a7de2816b0 Mon Sep 17 00:00:00 2001 From: Gautham Ananthakrishna Date: Wed, 18 Sep 2024 06:38:44 +0000 Subject: ocfs2: reserve space for inline xattr before attaching reflink tree One of our customers reported a crash and a corrupted ocfs2 filesystem. The crash was due to the detection of corruption. Upon troubleshooting, the fsck -fn output showed the below corruption [EXTENT_LIST_FREE] Extent list in owner 33080590 claims 230 as the next free chain record, but fsck believes the largest valid value is 227. Clamp the next record value? n The stat output from the debugfs.ocfs2 showed the following corruption where the "Next Free Rec:" had overshot the "Count:" in the root metadata block. Inode: 33080590 Mode: 0640 Generation: 2619713622 (0x9c25a856) FS Generation: 904309833 (0x35e6ac49) CRC32: 00000000 ECC: 0000 Type: Regular Attr: 0x0 Flags: Valid Dynamic Features: (0x16) HasXattr InlineXattr Refcounted Extended Attributes Block: 0 Extended Attributes Inline Size: 256 User: 0 (root) Group: 0 (root) Size: 281320357888 Links: 1 Clusters: 141738 ctime: 0x66911b56 0x316edcb8 -- Fri Jul 12 06:02:30.829349048 2024 atime: 0x66911d6b 0x7f7a28d -- Fri Jul 12 06:11:23.133669517 2024 mtime: 0x66911b56 0x12ed75d7 -- Fri Jul 12 06:02:30.317552087 2024 dtime: 0x0 -- Wed Dec 31 17:00:00 1969 Refcount Block: 2777346 Last Extblk: 2886943 Orphan Slot: 0 Sub Alloc Slot: 0 Sub Alloc Bit: 14 Tree Depth: 1 Count: 227 Next Free Rec: 230 ## Offset Clusters Block# 0 0 2310 2776351 1 2310 2139 2777375 2 4449 1221 2778399 3 5670 731 2779423 4 6401 566 2780447 ....... .... ....... ....... .... ....... The issue was in the reflink workfow while reserving space for inline xattr. The problematic function is ocfs2_reflink_xattr_inline(). By the time this function is called the reflink tree is already recreated at the destination inode from the source inode. At this point, this function reserves space for inline xattrs at the destination inode without even checking if there is space at the root metadata block. It simply reduces the l_count from 243 to 227 thereby making space of 256 bytes for inline xattr whereas the inode already has extents beyond this index (in this case up to 230), thereby causing corruption. The fix for this is to reserve space for inline metadata at the destination inode before the reflink tree gets recreated. The customer has verified the fix. Link: https://lkml.kernel.org/r/20240918063844.1830332-1-gautham.ananthakrishna@oracle.com Fixes: ef962df057aa ("ocfs2: xattr: fix inlined xattr reflink") Signed-off-by: Gautham Ananthakrishna Reviewed-by: Joseph Qi Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Changwei Ge Cc: Gang He Cc: Jun Piao Cc: Signed-off-by: Andrew Morton --- fs/ocfs2/refcounttree.c | 26 ++++++++++++++++++++++++-- fs/ocfs2/xattr.c | 11 +---------- 2 files changed, 25 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c index 4f85508538fc..004393b13c0a 100644 --- a/fs/ocfs2/refcounttree.c +++ b/fs/ocfs2/refcounttree.c @@ -25,6 +25,7 @@ #include "namei.h" #include "ocfs2_trace.h" #include "file.h" +#include "symlink.h" #include #include @@ -4148,8 +4149,9 @@ static int __ocfs2_reflink(struct dentry *old_dentry, int ret; struct inode *inode = d_inode(old_dentry); struct buffer_head *new_bh = NULL; + struct ocfs2_inode_info *oi = OCFS2_I(inode); - if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_SYSTEM_FILE) { + if (oi->ip_flags & OCFS2_INODE_SYSTEM_FILE) { ret = -EINVAL; mlog_errno(ret); goto out; @@ -4175,6 +4177,26 @@ static int __ocfs2_reflink(struct dentry *old_dentry, goto out_unlock; } + if ((oi->ip_dyn_features & OCFS2_HAS_XATTR_FL) && + (oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL)) { + /* + * Adjust extent record count to reserve space for extended attribute. + * Inline data count had been adjusted in ocfs2_duplicate_inline_data(). + */ + struct ocfs2_inode_info *new_oi = OCFS2_I(new_inode); + + if (!(new_oi->ip_dyn_features & OCFS2_INLINE_DATA_FL) && + !(ocfs2_inode_is_fast_symlink(new_inode))) { + struct ocfs2_dinode *new_di = (struct ocfs2_dinode *)new_bh->b_data; + struct ocfs2_dinode *old_di = (struct ocfs2_dinode *)old_bh->b_data; + struct ocfs2_extent_list *el = &new_di->id2.i_list; + int inline_size = le16_to_cpu(old_di->i_xattr_inline_size); + + le16_add_cpu(&el->l_count, -(inline_size / + sizeof(struct ocfs2_extent_rec))); + } + } + ret = ocfs2_create_reflink_node(inode, old_bh, new_inode, new_bh, preserve); if (ret) { @@ -4182,7 +4204,7 @@ static int __ocfs2_reflink(struct dentry *old_dentry, goto inode_unlock; } - if (OCFS2_I(inode)->ip_dyn_features & OCFS2_HAS_XATTR_FL) { + if (oi->ip_dyn_features & OCFS2_HAS_XATTR_FL) { ret = ocfs2_reflink_xattrs(inode, old_bh, new_inode, new_bh, preserve); diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index 0e58a5ce539e..dd0a05365e79 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -6511,16 +6511,7 @@ static int ocfs2_reflink_xattr_inline(struct ocfs2_xattr_reflink *args) } new_oi = OCFS2_I(args->new_inode); - /* - * Adjust extent record count to reserve space for extended attribute. - * Inline data count had been adjusted in ocfs2_duplicate_inline_data(). - */ - if (!(new_oi->ip_dyn_features & OCFS2_INLINE_DATA_FL) && - !(ocfs2_inode_is_fast_symlink(args->new_inode))) { - struct ocfs2_extent_list *el = &new_di->id2.i_list; - le16_add_cpu(&el->l_count, -(inline_size / - sizeof(struct ocfs2_extent_rec))); - } + spin_lock(&new_oi->ip_lock); new_oi->ip_dyn_features |= OCFS2_HAS_XATTR_FL | OCFS2_INLINE_XATTR_FL; new_di->i_dyn_features = cpu_to_le16(new_oi->ip_dyn_features); -- cgit v1.2.3 From 7bf1823e010e8db2fb649c790bd1b449a75f52d8 Mon Sep 17 00:00:00 2001 From: Mohammed Anees Date: Tue, 24 Sep 2024 09:32:57 +0000 Subject: ocfs2: fix deadlock in ocfs2_get_system_file_inode syzbot has found a possible deadlock in ocfs2_get_system_file_inode [1]. The scenario is depicted here, CPU0 CPU1 lock(&ocfs2_file_ip_alloc_sem_key); lock(&osb->system_file_mutex); lock(&ocfs2_file_ip_alloc_sem_key); lock(&osb->system_file_mutex); The function calls which could lead to this are: CPU0 ocfs2_mknod - lock(&ocfs2_file_ip_alloc_sem_key); . . . ocfs2_get_system_file_inode - lock(&osb->system_file_mutex); CPU1 - ocfs2_fill_super - lock(&osb->system_file_mutex); . . . ocfs2_read_virt_blocks - lock(&ocfs2_file_ip_alloc_sem_key); This issue can be resolved by making the down_read -> down_read_try in the ocfs2_read_virt_blocks. [1] https://syzkaller.appspot.com/bug?extid=e0055ea09f1f5e6fabdd Link: https://lkml.kernel.org/r/20240924093257.7181-1-pvmohammedanees2003@gmail.com Signed-off-by: Mohammed Anees Reviewed-by: Joseph Qi Reported-by: Closes: https://syzkaller.appspot.com/bug?extid=e0055ea09f1f5e6fabdd Tested-by: syzbot+e0055ea09f1f5e6fabdd@syzkaller.appspotmail.com Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Changwei Ge Cc: Gang He Cc: Jun Piao Cc: Signed-off-by: Andrew Morton --- fs/ocfs2/extent_map.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c index 70a768b623cf..f7672472fa82 100644 --- a/fs/ocfs2/extent_map.c +++ b/fs/ocfs2/extent_map.c @@ -973,7 +973,13 @@ int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr, } while (done < nr) { - down_read(&OCFS2_I(inode)->ip_alloc_sem); + if (!down_read_trylock(&OCFS2_I(inode)->ip_alloc_sem)) { + rc = -EAGAIN; + mlog(ML_ERROR, + "Inode #%llu ip_alloc_sem is temporarily unavailable\n", + (unsigned long long)OCFS2_I(inode)->ip_blkno); + break; + } rc = ocfs2_extent_map_get_blocks(inode, v_block + done, &p_block, &p_count, NULL); up_read(&OCFS2_I(inode)->ip_alloc_sem); -- cgit v1.2.3 From 2af148ef8549a12f8025286b8825c2833ee6bcb8 Mon Sep 17 00:00:00 2001 From: Joseph Qi Date: Wed, 25 Sep 2024 17:06:00 +0800 Subject: ocfs2: fix uninit-value in ocfs2_get_block() syzbot reported an uninit-value BUG: BUG: KMSAN: uninit-value in ocfs2_get_block+0xed2/0x2710 fs/ocfs2/aops.c:159 ocfs2_get_block+0xed2/0x2710 fs/ocfs2/aops.c:159 do_mpage_readpage+0xc45/0x2780 fs/mpage.c:225 mpage_readahead+0x43f/0x840 fs/mpage.c:374 ocfs2_readahead+0x269/0x320 fs/ocfs2/aops.c:381 read_pages+0x193/0x1110 mm/readahead.c:160 page_cache_ra_unbounded+0x901/0x9f0 mm/readahead.c:273 do_page_cache_ra mm/readahead.c:303 [inline] force_page_cache_ra+0x3b1/0x4b0 mm/readahead.c:332 force_page_cache_readahead mm/internal.h:347 [inline] generic_fadvise+0x6b0/0xa90 mm/fadvise.c:106 vfs_fadvise mm/fadvise.c:185 [inline] ksys_fadvise64_64 mm/fadvise.c:199 [inline] __do_sys_fadvise64 mm/fadvise.c:214 [inline] __se_sys_fadvise64 mm/fadvise.c:212 [inline] __x64_sys_fadvise64+0x1fb/0x3a0 mm/fadvise.c:212 x64_sys_call+0xe11/0x3ba0 arch/x86/include/generated/asm/syscalls_64.h:222 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xcd/0x1e0 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x77/0x7f This is because when ocfs2_extent_map_get_blocks() fails, p_blkno is uninitialized. So the error log will trigger the above uninit-value access. The error log is out-of-date since get_blocks() was removed long time ago. And the error code will be logged in ocfs2_extent_map_get_blocks() once ocfs2_get_cluster() fails, so fix this by only logging inode and block. Link: https://syzkaller.appspot.com/bug?extid=9709e73bae885b05314b Link: https://lkml.kernel.org/r/20240925090600.3643376-1-joseph.qi@linux.alibaba.com Fixes: ccd979bdbce9 ("[PATCH] OCFS2: The Second Oracle Cluster Filesystem") Signed-off-by: Joseph Qi Reported-by: syzbot+9709e73bae885b05314b@syzkaller.appspotmail.com Tested-by: syzbot+9709e73bae885b05314b@syzkaller.appspotmail.com Cc: Heming Zhao Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Changwei Ge Cc: Gang He Cc: Jun Piao Cc: Signed-off-by: Andrew Morton --- fs/ocfs2/aops.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index 1fea43c33b6b..db72b3e924b3 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -156,9 +156,8 @@ int ocfs2_get_block(struct inode *inode, sector_t iblock, err = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno, &count, &ext_flags); if (err) { - mlog(ML_ERROR, "Error %d from get_blocks(0x%p, %llu, 1, " - "%llu, NULL)\n", err, inode, (unsigned long long)iblock, - (unsigned long long)p_blkno); + mlog(ML_ERROR, "get_blocks() failed, inode: 0x%p, " + "block: %llu\n", inode, (unsigned long long)iblock); goto bail; } -- cgit v1.2.3