[syzbot] BUG: unable to handle kernel NULL pointer dereference in folio_mark_dirty

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: 58e1100fdc59 MAINTAINERS: co-maintain random.c
git tree: upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=1362881eb00000
kernel config: https://syzkaller.appspot.com/x/.config?x=e9ea28d2c3c2c389
dashboard link: https://syzkaller.appspot.com/bug?extid=7cd473c2cac13fd2dd72
compiler: Debian clang version 11.0.1-2, GNU ld (GNU Binutils for Debian) 2.35.2

Unfortunately, I don't have any reproducer for this issue yet.

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+7cd473...@syzkaller.appspotmail.com

BUG: kernel NULL pointer dereference, address: 0000000000000000
#PF: supervisor instruction fetch in kernel mode
#PF: error_code(0x0010) - not-present page
PGD 70764067 P4D 70764067 PUD 0
Oops: 0010 [#1] PREEMPT SMP KASAN
CPU: 1 PID: 6541 Comm: syz-executor.3 Not tainted 5.16.0-rc3-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:0x0
Code: Unable to access opcode bytes at RIP 0xffffffffffffffd6.
RSP: 0018:ffffc900027ff7f8 EFLAGS: 00010246
RAX: 1ffffffff14fef03 RBX: ffffffff8a7f7818 RCX: ffff88801b40d700
RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffea0002790ec0
RBP: dffffc0000000000 R08: ffffffff81b0fa16 R09: fffff940004f21d9
R10: fffff940004f21d9 R11: 0000000000000000 R12: ffff88806c11c7b0
R13: 0000000000000000 R14: 1ffffd40004f21d9 R15: ffffea0002790ec0
FS: 0000555557165400(0000) GS:ffff8880b9b00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ffffffffffffffd6 CR3: 0000000030d85000 CR4: 00000000003526e0
Call Trace:
<TASK>
folio_mark_dirty+0x136/0x270 mm/page-writeback.c:2639
f2fs_update_meta_page+0x4b/0x380 fs/f2fs/segment.c:2485
do_checkpoint fs/f2fs/checkpoint.c:1513 [inline]
f2fs_write_checkpoint+0x31ad/0x5430 fs/f2fs/checkpoint.c:1674
f2fs_issue_checkpoint+0x361/0x4e0
sync_filesystem+0x19c/0x1f0 fs/sync.c:63
generic_shutdown_super+0x6b/0x300 fs/super.c:448
kill_block_super+0x79/0xd0 fs/super.c:1397
kill_f2fs_super+0x2f9/0x3c0 fs/f2fs/super.c:4478
deactivate_locked_super+0xa7/0xf0 fs/super.c:335
cleanup_mnt+0x462/0x510 fs/namespace.c:1137
task_work_run+0x146/0x1c0 kernel/task_work.c:164
tracehook_notify_resume include/linux/tracehook.h:189 [inline]
exit_to_user_mode_loop kernel/entry/common.c:175 [inline]
exit_to_user_mode_prepare+0x209/0x220 kernel/entry/common.c:207
__syscall_exit_to_user_mode_work kernel/entry/common.c:289 [inline]
syscall_exit_to_user_mode+0x2e/0x70 kernel/entry/common.c:300
do_syscall_64+0x53/0xd0 arch/x86/entry/common.c:86
entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x7f6cfdd59f57
Code: ff ff ff f7 d8 64 89 01 48 83 c8 ff c3 66 0f 1f 44 00 00 31 f6 e9 09 00 00 00 66 0f 1f 84 00 00 00 00 00 b8 a6 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 bc ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007fffcbddcad8 EFLAGS: 00000246 ORIG_RAX: 00000000000000a6
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00007f6cfdd59f57
RDX: 00007fffcbddcbac RSI: 000000000000000a RDI: 00007fffcbddcba0
RBP: 00007fffcbddcba0 R08: 00000000ffffffff R09: 00007fffcbddc970
R10: 0000555557166903 R11: 0000000000000246 R12: 00007f6cfddb2105
R13: 00007fffcbdddc60 R14: 0000555557166810 R15: 00007fffcbdddca0
</TASK>
Modules linked in:
CR2: 0000000000000000
---[ end trace 08eda5a5e35b48a0 ]---
RIP: 0010:0x0
Code: Unable to access opcode bytes at RIP 0xffffffffffffffd6.
RSP: 0018:ffffc900027ff7f8 EFLAGS: 00010246
RAX: 1ffffffff14fef03 RBX: ffffffff8a7f7818 RCX: ffff88801b40d700
RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffea0002790ec0
RBP: dffffc0000000000 R08: ffffffff81b0fa16 R09: fffff940004f21d9
R10: fffff940004f21d9 R11: 0000000000000000 R12: ffff88806c11c7b0
R13: 0000000000000000 R14: 1ffffd40004f21d9 R15: ffffea0002790ec0
FS: 0000555557165400(0000) GS:ffff8880b9b00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ffffffffffffffd6 CR3: 0000000030d85000 CR4: 00000000003526e0


---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

[Andrew Morton]

On Sat, 04 Dec 2021 01:55:17 -0800 syzbot <syzbot+7cd473...@syzkaller.appspotmail.com> wrote:

> Hello,
>
> syzbot found the following issue on:
>
> HEAD commit: 58e1100fdc59 MAINTAINERS: co-maintain random.c
> git tree: upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=1362881eb00000
> kernel config: https://syzkaller.appspot.com/x/.config?x=e9ea28d2c3c2c389
> dashboard link: https://syzkaller.appspot.com/bug?extid=7cd473c2cac13fd2dd72
> compiler: Debian clang version 11.0.1-2, GNU ld (GNU Binutils for Debian) 2.35.2
>
> Unfortunately, I don't have any reproducer for this issue yet.
>
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+7cd473...@syzkaller.appspotmail.com
>
> BUG: kernel NULL pointer dereference, address: 0000000000000000

cc linux-f2...@lists.sourceforge.net

And willy, who might help with diagnosing this. But it does seem that
f2fs got itself a NULL page* then put it in places where it shouldn't have.

[Matthew Wilcox]

On Mon, Dec 06, 2021 at 05:56:31PM -0800, Andrew Morton wrote:
> > BUG: kernel NULL pointer dereference, address: 0000000000000000
>
> cc linux-f2...@lists.sourceforge.net
>
> And willy, who might help with diagnosing this. But it does seem that
> f2fs got itself a NULL page* then put it in places where it shouldn't have.

I'm surprised it got that far.

void f2fs_update_meta_page(struct f2fs_sb_info *sbi,
void *src, block_t blk_addr)
{
struct page *page = f2fs_grab_meta_page(sbi, blk_addr);

memcpy(page_address(page), src, PAGE_SIZE);
set_page_dirty(page);
f2fs_put_page(page, 1);
}

How does page_address(NULL) not crash first? Or return an address that
can be the target of a memcpy()?

[Matthew Wilcox]

On Mon, Dec 06, 2021 at 05:56:31PM -0800, Andrew Morton wrote:

> On Sat, 04 Dec 2021 01:55:17 -0800 syzbot <syzbot+7cd473...@syzkaller.appspotmail.com> wrote:
>
> > Hello,
> >
> > syzbot found the following issue on:
> >
> > HEAD commit: 58e1100fdc59 MAINTAINERS: co-maintain random.c
> > git tree: upstream
> > console output: https://syzkaller.appspot.com/x/log.txt?x=1362881eb00000
> > kernel config: https://syzkaller.appspot.com/x/.config?x=e9ea28d2c3c2c389
> > dashboard link: https://syzkaller.appspot.com/bug?extid=7cd473c2cac13fd2dd72
> > compiler: Debian clang version 11.0.1-2, GNU ld (GNU Binutils for Debian) 2.35.2
> >
> > Unfortunately, I don't have any reproducer for this issue yet.
> >
> > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > Reported-by: syzbot+7cd473...@syzkaller.appspotmail.com
> >
> > BUG: kernel NULL pointer dereference, address: 0000000000000000
>
> cc linux-f2...@lists.sourceforge.net
>
> And willy, who might help with diagnosing this. But it does seem that
> f2fs got itself a NULL page* then put it in places where it shouldn't have.

Oh -- it's not a NULL data pointer. It's a NULL instruction pointer.

> > #PF: supervisor instruction fetch in kernel mode
> > #PF: error_code(0x0010) - not-present page
> > PGD 70764067 P4D 70764067 PUD 0
> > Oops: 0010 [#1] PREEMPT SMP KASAN
> > CPU: 1 PID: 6541 Comm: syz-executor.3 Not tainted 5.16.0-rc3-syzkaller #0
> > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
> > RIP: 0010:0x0
> > Code: Unable to access opcode bytes at RIP 0xffffffffffffffd6.
> > RSP: 0018:ffffc900027ff7f8 EFLAGS: 00010246
> > RAX: 1ffffffff14fef03 RBX: ffffffff8a7f7818 RCX: ffff88801b40d700
> > RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffea0002790ec0
> > RBP: dffffc0000000000 R08: ffffffff81b0fa16 R09: fffff940004f21d9
> > R10: fffff940004f21d9 R11: 0000000000000000 R12: ffff88806c11c7b0
> > R13: 0000000000000000 R14: 1ffffd40004f21d9 R15: ffffea0002790ec0
> > FS: 0000555557165400(0000) GS:ffff8880b9b00000(0000) knlGS:0000000000000000
> > CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > CR2: ffffffffffffffd6 CR3: 0000000030d85000 CR4: 00000000003526e0
> > Call Trace:
> > <TASK>
> > folio_mark_dirty+0x136/0x270 mm/page-writeback.c:2639

if (likely(mapping)) {
...
if (folio_test_reclaim(folio))
folio_clear_reclaim(folio);
return mapping->a_ops->set_page_dirty(&folio->page);

how do we get to a NULL ->set_page_dirty for a metadata page's
mapping->a_ops? This is definitely an f2fs expert question.

[Jaegeuk Kim]

I can't find anything in f2fs, since that page was got by f2fs_grab_meta_page
along with grab_cache_page() that we never unlocked it.

40 struct page *f2fs_grab_meta_page(struct f2fs_sb_info *sbi, pgoff_t index)
41 {
42 struct address_space *mapping = META_MAPPING(sbi);
43 struct page *page;
44 repeat:
45 page = f2fs_grab_cache_page(mapping, index, false);

-> grab_cache_page(mapping, index);

46 if (!page) {
47 cond_resched();
48 goto repeat;
49 }
50 f2fs_wait_on_page_writeback(page, META, true, true);
51 if (!PageUptodate(page))
52 SetPageUptodate(page);
53 return page;
54 }


Suspecting something in folio wrt folio_mapping()?

81 bool set_page_dirty(struct page *page)
82 {
83 return folio_mark_dirty(page_folio(page));
84 }

> _______________________________________________
> Linux-f2fs-devel mailing list
> Linux-f2...@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

[Matthew Wilcox]

... huh? How could folio_mapping() be getting this wrong?
page_folio() does the same thing as compound_head() -- as far as I know
you don't use compound pages for f2fs metadata, so this basically just
casts the page to a struct folio.

folio_mapping() is just like the old page_mapping() (see commit
2f52578f9c64). Unless you've done something like set the swapcache
bit on your metadata page, it's just going to return folio->mapping
(ie the same as page->mapping).

[Jaegeuk Kim]

Dunno.

> page_folio() does the same thing as compound_head() -- as far as I know
> you don't use compound pages for f2fs metadata, so this basically just
> casts the page to a struct folio.
>
> folio_mapping() is just like the old page_mapping() (see commit
> 2f52578f9c64). Unless you've done something like set the swapcache
> bit on your metadata page, it's just going to return folio->mapping
> (ie the same as page->mapping).

Hmm, I've never seen this call stack before, so simply started to suspect
folio.

[Chao Yu]

I'm afraid this is a f2fs bug... :(

folio wasn't merged at the first report time (5.14-rc2).

https://syzkaller.appspot.com/bug?extid=07ff38c9c93ca170de07

I doubt the direct reason of panic may be the same as the one of bug
reported in bugzilla:

https://bugzilla.kernel.org/show_bug.cgi?id=215231

However, I still can't figure out why meta_inode's a_ops will change
to f2fs_meta_aops.

Thanks,

[syzbot]

Auto-closing this bug as obsolete.
Crashes did not happen for a while, no reproducer and no activity.