On 05/10, Chao Yu wrote:
> On 2024/5/10 11:36, Jaegeuk Kim wrote:
> > On 05/10, Chao Yu wrote:
> > > On 2024/5/9 23:52, Jaegeuk Kim wrote:
> > > > On 05/06, Chao Yu wrote:
> > > > > syzbot reports a f2fs bug as below:
> > > > >
> > > > > ------------[ cut here ]------------
> > > > > kernel BUG at fs/f2fs/inline.c:258!
> > > > > CPU: 1 PID: 34 Comm: kworker/u8:2 Not tainted
> > > > > 6.9.0-rc6-syzkaller-00012-g9e4bc4bcae01 #0
> > > > > RIP: 0010:f2fs_write_inline_data+0x781/0x790 fs/f2fs/inline.c:258
> > > > > Call Trace:
> > > > > f2fs_write_single_data_page+0xb65/0x1d60 fs/f2fs/data.c:2834
> > > > > f2fs_write_cache_pages fs/f2fs/data.c:3133 [inline]
> > > > > __f2fs_write_data_pages fs/f2fs/data.c:3288 [inline]
> > > > > f2fs_write_data_pages+0x1efe/0x3a90 fs/f2fs/data.c:3315
> > > > > do_writepages+0x35b/0x870 mm/page-writeback.c:2612
> > > > > __writeback_single_inode+0x165/0x10b0 fs/fs-writeback.c:1650
> > > > > writeback_sb_inodes+0x905/0x1260 fs/fs-writeback.c:1941
> > > > > wb_writeback+0x457/0xce0 fs/fs-writeback.c:2117
> > > > > wb_do_writeback fs/fs-writeback.c:2264 [inline]
> > > > > wb_workfn+0x410/0x1090 fs/fs-writeback.c:2304
> > > > > process_one_work kernel/workqueue.c:3254 [inline]
> > > > > process_scheduled_works+0xa12/0x17c0 kernel/workqueue.c:3335
> > > > > worker_thread+0x86d/0xd70 kernel/workqueue.c:3416
> > > > > kthread+0x2f2/0x390 kernel/kthread.c:388
> > > > > ret_from_fork+0x4d/0x80 arch/x86/kernel/process.c:147
> > > > > ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
> > > > >
> > > > > The root cause is: inline_data inode can be fuzzed, so that there may
> > > > > be valid blkaddr in its direct node, once f2fs triggers background GC
> > > > > to migrate the block, it will hit f2fs_bug_on() during dirty page
> > > > > writeback.
> > > > >
> > > > > Let's add sanity check on i_nid field for inline_data inode,
> > > > > meanwhile,
> > > > > forbid to migrate inline_data inode's data block to fix this issue.
> > > > >
> > > > > Reported-by: syzbot+848062ba19c8782ca...@syzkaller.appspotmail.com
> > > > > Closes:
> > > > > https://lore.kernel.org/linux-f2fs-devel/000000000000d103ce06174d7...@google.com
> > > > > Signed-off-by: Chao Yu <c...@kernel.org>
> > > > > ---
> > > > > fs/f2fs/f2fs.h | 2 +-
> > > > > fs/f2fs/gc.c | 6 ++++++
> > > > > fs/f2fs/inline.c | 17 ++++++++++++++++-
> > > > > fs/f2fs/inode.c | 2 +-
> > > > > 4 files changed, 24 insertions(+), 3 deletions(-)
> > > > >
> > > > > diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> > > > > index fced2b7652f4..c876813b5532 100644
> > > > > --- a/fs/f2fs/f2fs.h
> > > > > +++ b/fs/f2fs/f2fs.h
> > > > > @@ -4146,7 +4146,7 @@ extern struct kmem_cache *f2fs_inode_entry_slab;
> > > > > * inline.c
> > > > > */
> > > > > bool f2fs_may_inline_data(struct inode *inode);
> > > > > -bool f2fs_sanity_check_inline_data(struct inode *inode);
> > > > > +bool f2fs_sanity_check_inline_data(struct inode *inode, struct page
> > > > > *ipage);
> > > > > bool f2fs_may_inline_dentry(struct inode *inode);
> > > > > void f2fs_do_read_inline_data(struct page *page, struct page
> > > > > *ipage);
> > > > > void f2fs_truncate_inline_inode(struct inode *inode,
> > > > > diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
> > > > > index e86c7f01539a..041957750478 100644
> > > > > --- a/fs/f2fs/gc.c
> > > > > +++ b/fs/f2fs/gc.c
> > > > > @@ -1563,6 +1563,12 @@ static int gc_data_segment(struct f2fs_sb_info
> > > > > *sbi, struct f2fs_summary *sum,
> > > > > continue;
> > > > > }
> > > > > + if (f2fs_has_inline_data(inode)) {
> > > > > + iput(inode);
> > > > > + set_sbi_flag(sbi, SBI_NEED_FSCK);
> > > > > + continue;
> > > >
> > > > Any race condtion to get this as false alarm?
> > >
> > > Since there is no reproducer for the bug, I doubt it was caused by
> > > metadata
> > > fuzzing, something like this:
> > >
> > > - inline inode has one valid blkaddr in i_addr or in dnode reference by
> > > i_nid;
> > > - SIT/SSA entry of the block is valid;
> > > - background GC migrates the block;
> > > - kworker writeback it, and trigger the bug_on().
> >
> > Wasn't detected by sanity_check_inode?
>
> I fuzzed non-inline inode w/ below metadata fields:
> - i_blocks = 1
> - i_size = 2048
> - i_inline |= 0x02
>
> sanity_check_inode() doesn't complain.
I mean, the below sanity_check_inode() can cover the fuzzed case? I'm wondering
whether we really need to check it in the gc path.
>
> Thanks,
>
> >
> > >
> > > Thoughts?
> > >
> > > Thanks,
> > >
> > > >
> > > > > + }
> > > > > +
> > > > > err = f2fs_gc_pinned_control(inode, gc_type,
> > > > > segno);
> > > > > if (err == -EAGAIN) {
> > > > > iput(inode);
> > > > > diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c
> > > > > index ac00423f117b..067600fed3d4 100644
> > > > > --- a/fs/f2fs/inline.c
> > > > > +++ b/fs/f2fs/inline.c
> > > > > @@ -33,11 +33,26 @@ bool f2fs_may_inline_data(struct inode *inode)
> > > > > return !f2fs_post_read_required(inode);
> > > > > }
> > > > > -bool f2fs_sanity_check_inline_data(struct inode *inode)
> > > > > +static bool has_node_blocks(struct inode *inode, struct page *ipage)
> > > > > +{
> > > > > + struct f2fs_inode *ri = F2FS_INODE(ipage);
> > > > > + int i;
> > > > > +
> > > > > + for (i = 0; i < DEF_NIDS_PER_INODE; i++) {
> > > > > + if (ri->i_nid[i])
> > > > > + return true;
> > > > > + }
> > > > > + return false;
> > > > > +}
> > > > > +
> > > > > +bool f2fs_sanity_check_inline_data(struct inode *inode, struct page
> > > > > *ipage)
> > > > > {
> > > > > if (!f2fs_has_inline_data(inode))
> > > > > return false;
> > > > > + if (has_node_blocks(inode, ipage))
> > > > > + return false;
> > > > > +
> > > > > if (!support_inline_data(inode))
> > > > > return true;
> > > > > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
> > > > > index c26effdce9aa..1423cd27a477 100644
> > > > > --- a/fs/f2fs/inode.c
> > > > > +++ b/fs/f2fs/inode.c
> > > > > @@ -343,7 +343,7 @@ static bool sanity_check_inode(struct inode
> > > > > *inode, struct page *node_page)
> > > > > }
> > > > > }
> > > > > - if (f2fs_sanity_check_inline_data(inode)) {
> > > > > + if (f2fs_sanity_check_inline_data(inode, node_page)) {
> > > > > f2fs_warn(sbi, "%s: inode (ino=%lx, mode=%u) should not
> > > > > have inline_data, run fsck to fix",
> > > > > __func__, inode->i_ino, inode->i_mode);
> > > > > return false;
> > > > > --
> > > > > 2.40.1
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
