From: "jiale yao" <19888972804@163.com>
To: "Mark Fasheh" <mark@fasheh.com>,
"Joel Becker" <jlbec@evilplan.org>,
"Joseph Qi" <joseph.qi@linux.alibaba.com>,
"Andrew Morton" <akpm@linux-foundation.org>,
"ZhengYuan Huang" <gality369@gmail.com>,
ocfs2-devel@lists.linux.dev, linux-kernel@vger.kernel.org
Subject: Re:[PATCH] ocfs2: validate filecheck inode slots
Date: Sat, 26 Sep 2026 21:59:22 +0800 (CST) [thread overview]
Message-ID: <41659397.1e4f.1a0de0370ab.Coremail.19888972804@163.com> (raw)
In-Reply-To: <20260926135354.3420193-1-yaojiale02@163.com>
At 2026-09-26 21:53:53, "Jiale Yao" <yaojiale02@163.com> wrote:
>The online filecheck path reads inodes with
>ocfs2_filecheck_validate_inode_block() instead of
>ocfs2_validate_inode_block(). It does not check the slot fields that
>are used as indices into arrays sized by osb->max_slots.
>
>A corrupted dinode can set OCFS2_ORPHANED_FL or
>OCFS2_DIO_ORPHANED_FL while carrying an out-of-range orphan slot. An
>out-of-range i_suballoc_slot can also bypass the normal validator
>through the filecheck path. These values can later be used to index
>the slot-local system inode array.
Reproduced on 7.3-rc4 (x86_64 QEMU, KASAN):
[ 45.942783] BUG: KASAN: slab-out-of-bounds in ocfs2_evict_inode+0x28f7/0x3ac0
[ 45.942818] Read of size 4 at addr ffff8881
[ 45.942827] Call Trace:
[ 45.942829] dump_stack_lvl+0x2d/0x70
[ 45.942833] print_report+0x175/0x7d0
[ 45.942836] kasan_report+0x139/0x170
[ 45.942839] ocfs2_evict_inode+0x28f7/0x3ac0
[ 45.942840] evict+0x344/0x6e0
[ 45.942842] ? iput+0x45d/0x730
[ 45.942843] ocfs2_filecheck_attr_store+0x96a/0xd30
[ 45.942846] ocfs2_filecheck_store+0x5d/0x90
[ 45.942847] kernfs_fop_write_iter+0x262/0x370
[ 45.942849] vfs_write+0x933/0xcb0
[ 45.942852] ksys_write+0xf2/0x1b0
[ 45.942853] do_syscall_64+0x15b/0x420
[ 45.942856] entry_SYSCALL_64_after_hwframe
...
[ 45.942867] Allocated by task 81:
[ 45.942870] __kasan_kmalloc+0x72/0x90
[ 45.942871] __kmalloc_noprof+0x1d2/0x470
[ 45.942875] ocfs2_fill_super+0x195b/0x59f0s, ...)
[ 45.942885] The buggy address is located 4 bytes to the right of
[ 45.942885] allocated 4-byte region [ffff888105f04a40, ffff888105f04a44)
[ 45.942910] Oops: general protection fault, probably for non-canonical
address ... (the pointer read past inode)
>Mirror the normal validator's slot checks in the filecheck validator.
>Reject invalid slots in the repair path as well, since the correct
>slot cannot be recovered.
>
>Fixes: bb88131c9831 ("ocfs2: validate DIO orphan slot during inode read")
>Signed-off-by: Jiale Yao <yaojiale02@163.com>
>---
> fs/ocfs2/inode.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 45 insertions(+)
>
>diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
>index 180107a11046..ab82cf9146ef 100644
>--- a/fs/ocfs2/inode.c
>+++ b/fs/ocfs2/inode.c
>@@ -1754,6 +1754,44 @@ int ocfs2_validate_inode_block(struct super_block *sb,
> return rc;
> }
>
>+/*
>+ * Validate the slot fields used as indices into arrays sized by
>+ * osb->max_slots. The filecheck path uses this validator instead of
>+ * ocfs2_validate_inode_block(), so it must enforce the same bounds.
>+ */
>+static int ocfs2_filecheck_validate_slots(struct super_block *sb,
>+ struct ocfs2_dinode *di,
>+ unsigned long long blkno)
>+{
>+ struct ocfs2_super *osb = OCFS2_SB(sb);
>+
>+ if (le16_to_cpu(di->i_suballoc_slot) != (u16)OCFS2_INVALID_SLOT &&
>+ (u32)le16_to_cpu(di->i_suballoc_slot) > osb->max_slots - 1) {
>+ mlog(ML_ERROR,
>+ "Filecheck: invalid dinode #%llu: suballoc slot %u\n",
>+ blkno, le16_to_cpu(di->i_suballoc_slot));
>+ return -OCFS2_FILECHECK_ERR_INVALIDINO;
>+ }
>+
>+ if ((le32_to_cpu(di->i_flags) & OCFS2_ORPHANED_FL) &&
>+ le16_to_cpu(di->i_orphaned_slot) >= osb->max_slots) {
>+ mlog(ML_ERROR,
>+ "Filecheck: invalid dinode #%llu: orphaned slot %u\n",
>+ blkno, le16_to_cpu(di->i_orphaned_slot));
>+ return -OCFS2_FILECHECK_ERR_INVALIDINO;
>+ }
>+
>+ if ((le32_to_cpu(di->i_flags) & OCFS2_DIO_ORPHANED_FL) &&
>+ le16_to_cpu(di->i_dio_orphaned_slot) >= osb->max_slots) {
>+ mlog(ML_ERROR,
>+ "Filecheck: invalid dinode #%llu: DIO orphaned slot %u\n",
>+ blkno, le16_to_cpu(di->i_dio_orphaned_slot));
>+ return -OCFS2_FILECHECK_ERR_INVALIDINO;
>+ }
>+
>+ return 0;
>+}
>+
> static int ocfs2_filecheck_validate_inode_block(struct super_block *sb,
> struct buffer_head *bh)
> {
>@@ -1835,6 +1873,10 @@ static int ocfs2_filecheck_validate_inode_block(struct super_block *sb,
> goto bail;
> }
>
>+ rc = ocfs2_filecheck_validate_slots(sb, di, bh->b_blocknr);
>+ if (rc)
>+ goto bail;
>+
> if (ocfs2_dinode_has_size_without_clusters(sb, di)) {
> if (S_ISDIR(le16_to_cpu(di->i_mode)))
> mlog(ML_ERROR,
>@@ -1893,6 +1935,9 @@ static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
> return -OCFS2_FILECHECK_ERR_VALIDFLAG;
> }
>
>+ if (ocfs2_filecheck_validate_slots(sb, di, bh->b_blocknr))
>+ return -OCFS2_FILECHECK_ERR_INVALIDINO;
>+
> if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
> di->i_blkno = cpu_to_le64(bh->b_blocknr);
> changed = 1;
>--
>2.34.1
next prev parent reply other threads:[~2026-09-26 14:00 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-26 13:53 [PATCH] " Jiale Yao
2026-09-26 13:59 ` jiale yao [this message]
2026-09-28 9:58 ` Joseph Qi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=41659397.1e4f.1a0de0370ab.Coremail.19888972804@163.com \
--to=19888972804@163.com \
--cc=akpm@linux-foundation.org \
--cc=gality369@gmail.com \
--cc=jlbec@evilplan.org \
--cc=joseph.qi@linux.alibaba.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mark@fasheh.com \
--cc=ocfs2-devel@lists.linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®