Re: [syzbot] [io-uring?] KASAN: null-ptr-deref Read in io_sqe_buffer_register

[Jens Axboe <axboe@kernel.dk>]

On 9/4/25 9:36 AM, syzbot wrote:
> Hello,
> 
> syzbot found the following issue on:
> 
> HEAD commit:    4ac65880ebca Add linux-next specific files for 20250904
> git tree:       linux-next
> console output: https://syzkaller.appspot.com/x/log.txt?x=1785fe62580000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=fbc16d9faf3a88a4
> dashboard link: https://syzkaller.appspot.com/bug?extid=1ab243d3eebb2aabf4a4
> compiler:       Debian clang version 20.1.8 (++20250708063551+0c9f909b7976-1~exp1~20250708183702.136), Debian LLD 20.1.8
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=13f23e62580000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=12cb6312580000
> 
> Downloadable assets:
> disk image: https://storage.googleapis.com/syzbot-assets/36645a51612c/disk-4ac65880.raw.xz
> vmlinux: https://storage.googleapis.com/syzbot-assets/bba80d634bef/vmlinux-4ac65880.xz
> kernel image: https://storage.googleapis.com/syzbot-assets/e58dd70dfd0f/bzImage-4ac65880.xz
> 
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+1ab243d3eebb2aabf4a4@syzkaller.appspotmail.com
> 
> ==================================================================
> BUG: KASAN: null-ptr-deref in instrument_atomic_read include/linux/instrumented.h:68 [inline]
> BUG: KASAN: null-ptr-deref in _test_bit include/asm-generic/bitops/instrumented-non-atomic.h:141 [inline]
> BUG: KASAN: null-ptr-deref in PageCompound include/linux/page-flags.h:331 [inline]
> BUG: KASAN: null-ptr-deref in io_buffer_account_pin io_uring/rsrc.c:668 [inline]
> BUG: KASAN: null-ptr-deref in io_sqe_buffer_register+0x369/0x20a0 io_uring/rsrc.c:817
> Read of size 8 at addr 0000000000000000 by task syz.0.17/6020
> 
> CPU: 0 UID: 0 PID: 6020 Comm: syz.0.17 Not tainted syzkaller #0 PREEMPT(full) 
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/12/2025
> Call Trace:
>  <TASK>
>  dump_stack_lvl+0x189/0x250 lib/dump_stack.c:120
>  kasan_report+0x118/0x150 mm/kasan/report.c:595
>  check_region_inline mm/kasan/generic.c:-1 [inline]
>  kasan_check_range+0x2b0/0x2c0 mm/kasan/generic.c:200
>  instrument_atomic_read include/linux/instrumented.h:68 [inline]
>  _test_bit include/asm-generic/bitops/instrumented-non-atomic.h:141 [inline]
>  PageCompound include/linux/page-flags.h:331 [inline]
>  io_buffer_account_pin io_uring/rsrc.c:668 [inline]
>  io_sqe_buffer_register+0x369/0x20a0 io_uring/rsrc.c:817
>  __io_sqe_buffers_update io_uring/rsrc.c:322 [inline]
>  __io_register_rsrc_update+0x55e/0x11b0 io_uring/rsrc.c:360
>  io_register_rsrc_update+0x196/0x1a0 io_uring/rsrc.c:391
>  __io_uring_register io_uring/register.c:736 [inline]
>  __do_sys_io_uring_register io_uring/register.c:926 [inline]
>  __se_sys_io_uring_register+0x795/0x11b0 io_uring/register.c:903
>  do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
>  do_syscall_64+0xfa/0xfa0 arch/x86/entry/syscall_64.c:94
>  entry_SYSCALL_64_after_hwframe+0x77/0x7f
> RIP: 0033:0x7f99b1f8ebe9
> Code: ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 a8 ff ff ff f7 d8 64 89 01 48
> RSP: 002b:00007f99b2d88038 EFLAGS: 00000246 ORIG_RAX: 00000000000001ab
> RAX: ffffffffffffffda RBX: 00007f99b21c5fa0 RCX: 00007f99b1f8ebe9
> RDX: 00002000000003c0 RSI: 0000000000000010 RDI: 0000000000000003
> RBP: 00007f99b2011e19 R08: 0000000000000000 R09: 0000000000000000
> R10: 0000000000000020 R11: 0000000000000246 R12: 0000000000000000
> R13: 00007f99b21c6038 R14: 00007f99b21c5fa0 R15: 00007ffeadfa5958
>  </TASK>
> ==================================================================

This is from the mm-unstable changes in linux-next, adding David as I
ran a quick bisect and it said:

da6b34293ff8dbb78f8b9278c9a492925bbf1f87 is the first bad commit
commit da6b34293ff8dbb78f8b9278c9a492925bbf1f87
Author: David Hildenbrand <david@redhat.com>
Date:   Mon Sep 1 17:03:40 2025 +0200

    mm/gup: remove record_subpages()
    
    We can just cleanup the code by calculating the #refs earlier, so we can
    just inline what remains of record_subpages().
    
    Calculate the number of references/pages ahead of times, and record them
    only once all our tests passed.
    
    Link: https://lkml.kernel.org/r/20250901150359.867252-20-david@redhat.com
    Signed-off-by: David Hildenbrand <david@redhat.com>
    Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

I won't personally have time to look into this until after the weekend,
but as it's linux-next specific, not a huge deal right now.

Note that there's also a similar report, which is the same thing:

https://lore.kernel.org/all/68b9d130.a00a0220.eb3d.0008.GAE@google.com/

which I marked as dupe of this one.

-- 
Jens Axboe