Re: [syzbot] BUG: sleeping function called from invalid context in copy_huge_page

[Yang Shi]

On Tue, Oct 12, 2021 at 7:03 AM syzbot
<syzbot+aae069be1de40fb11825@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> Hello,
>
> syzbot found the following issue on:
>
> HEAD commit:    1da38549dd64 Merge tag 'nfsd-5.15-3' of git://git.kernel.o..
> git tree:       upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=14379148b00000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=76f7496a8217e5ec
> dashboard link: https://syzkaller.appspot.com/bug?extid=aae069be1de40fb11825
> compiler:       gcc (Debian 10.2.1-6) 10.2.1 20210110, 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+aae069be1de40fb11825@xxxxxxxxxxxxxxxxxxxxxxxxx
>
> BUG: sleeping function called from invalid context at mm/util.c:758
> in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 30700, name: syz-executor.2
> 2 locks held by syz-executor.2/30700:
>  #0: ffff88806ee498a8 (&mm->mmap_lock#2){++++}-{3:3}, at: mmap_write_lock include/linux/mmap_lock.h:71 [inline]
>  #0: ffff88806ee498a8 (&mm->mmap_lock#2){++++}-{3:3}, at: do_mbind+0x25d/0xeb0 mm/mempolicy.c:1314
>  #1: ffff888145989e18 (&mapping->private_lock){+.+.}-{2:2}, at: spin_lock include/linux/spinlock.h:363 [inline]
>  #1: ffff888145989e18 (&mapping->private_lock){+.+.}-{2:2}, at: __buffer_migrate_page+0x3af/0xca0 mm/migrate.c:723
> Preemption disabled at:
> [<0000000000000000>] 0x0
> CPU: 1 PID: 30700 Comm: syz-executor.2 Not tainted 5.15.0-rc4-syzkaller #0
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
> Call Trace:
>  __dump_stack lib/dump_stack.c:88 [inline]
>  dump_stack_lvl+0xcd/0x134 lib/dump_stack.c:106
>  ___might_sleep.cold+0x1f3/0x239 kernel/sched/core.c:9538
>  copy_huge_page+0x126/0x360 mm/util.c:758
>  migrate_page_copy+0xfc/0x340 mm/migrate.c:619
>  __buffer_migrate_page+0x8cb/0xca0 mm/migrate.c:758

It seems like this one has the similar root cause with
https://lore.kernel.org/lkml/CACkBjsYwLYLRmX8GpsDpMthagWOjWWrNxqY6ZLNQVr6yx+f5vA@xxxxxxxxxxxxxx/.
The THP is collapsed for page cache from raw block device. Then the
THP got migrated by calling buffer_migrate_page_norefs() in this BUG
report, which takes mapping->private_lock and is used by raw block
device.

So skipping the non-regular file in khugepaged
(https://lore.kernel.org/linux-mm/a07564a3-b2fc-9ffe-3ace-3f276075ea5c@xxxxxxxxxx/)
seems like a proper fix.

>  move_to_new_page+0x339/0xef0 mm/migrate.c:905
>  __unmap_and_move mm/migrate.c:1070 [inline]
>  unmap_and_move mm/migrate.c:1211 [inline]
>  migrate_pages+0xfc5/0x39e0 mm/migrate.c:1488
>  do_mbind+0xbc7/0xeb0 mm/mempolicy.c:1340
>  kernel_mbind mm/mempolicy.c:1483 [inline]
>  __do_sys_mbind mm/mempolicy.c:1490 [inline]
>  __se_sys_mbind mm/mempolicy.c:1486 [inline]
>  __x64_sys_mbind+0x233/0x2b0 mm/mempolicy.c:1486
>  do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>  do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
>  entry_SYSCALL_64_after_hwframe+0x44/0xae
> RIP: 0033:0x7f408623f8d9
> 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 bc ff ff ff f7 d8 64 89 01 48
> RSP: 002b:00007f40837b6188 EFLAGS: 00000246 ORIG_RAX: 00000000000000ed
> RAX: ffffffffffffffda RBX: 00007f4086343f60 RCX: 00007f408623f8d9
> RDX: 0000000000000000 RSI: 0000000000c00000 RDI: 0000000020012000
> RBP: 00007f4086299cb4 R08: 0000000000000000 R09: 0000010000000002
> R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
> R13: 00007fff13d3589f R14: 00007f40837b6300 R15: 0000000000022000
>
>
> ---
> 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 syzkaller@xxxxxxxxxxxxxxxx.
>
> syzbot will keep track of this issue. See:
> https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
>