Re: [syzbot] WARNING in split_huge_page_to_list (2)

From: Yang Shi
Date: Tue Aug 23 2022 - 21:36:50 EST


On Tue, Aug 23, 2022 at 3:08 PM syzbot
<syzbot+07a218429c8d19b1fb25@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> Hello,
>
> syzbot found the following issue on:
>
> HEAD commit: 680fb5b009e8 Merge tag 'arm64-upstream' into for-kernelci
> git tree: git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git for-kernelci
> console output: https://syzkaller.appspot.com/x/log.txt?x=10ef7dc3080000
> kernel config: https://syzkaller.appspot.com/x/.config?x=4557ad2600fc45f4
> dashboard link: https://syzkaller.appspot.com/bug?extid=07a218429c8d19b1fb25
> compiler: Debian clang version 13.0.1-++20220126092033+75e33f71c2da-1~exp1~20220126212112.63, GNU ld (GNU Binutils for Debian) 2.35.2
> userspace arch: arm64
> syz repro: https://syzkaller.appspot.com/x/repro.syz?x=17993aa5080000
> C reproducer: https://syzkaller.appspot.com/x/repro.c?x=16521a3d080000
>
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+07a218429c8d19b1fb25@xxxxxxxxxxxxxxxxxxxxxxxxx
>
> raw: 05ffc00000010801 0000000000000000 dead000000000122 0000000000000000
> raw: 0000000000000000 0000000000000000 00000002ffffffff 0000000000000000
> page dumped because: VM_WARN_ON_ONCE_PAGE(is_hzp)
> ------------[ cut here ]------------
> WARNING: CPU: 1 PID: 3025 at mm/huge_memory.c:2555 split_huge_page_to_list+0x310/0xe2c mm/huge_memory.c:2555
> Modules linked in:
> CPU: 1 PID: 3025 Comm: syz-executor201 Not tainted 5.19.0-rc8-syzkaller-01618-g680fb5b009e8 #0
> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 06/20/2022
> pstate: 60400005 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
> pc : split_huge_page_to_list+0x310/0xe2c mm/huge_memory.c:2555
> lr : split_huge_page_to_list+0x308/0xe2c mm/huge_memory.c:2555
> sp : ffff80001321bba0
> x29: ffff80001321bbf0 x28: 0000000000400000 x27: 0000000000000001
> x26: ffff80000ceda000 x25: 05ffc00000000000 x24: ffff80000d8c1000
> x23: fffffc0003300000 x22: ffff0001ff012d40 x21: fffffc0003304400
> x20: 0000000000000000 x19: fffffc0003300000 x18: 00000000000000c0
> x17: ffff80000dcfd698 x16: ffff80000db3b658 x15: ffff0000c6708000
> x14: 0000000000000000 x13: 00000000ffffffff x12: ffff0000c6708000
> x11: ff808000081bf324 x10: 0000000000000000 x9 : 461f0eca467c0e00
> x8 : 0000000000000001 x7 : ffff8000081937c0 x6 : 0000000000000000
> x5 : 0000000000000080 x4 : 0000000000000001 x3 : 0000000000000000
> x2 : 0000000000000000 x1 : 0000000100000000 x0 : 0000000000000031
> Call trace:
> split_huge_page_to_list+0x310/0xe2c mm/huge_memory.c:2555
> split_huge_page include/linux/huge_mm.h:206 [inline]
> try_to_split_thp_page mm/memory-failure.c:1488 [inline]
> memory_failure+0x510/0x14cc mm/memory-failure.c:1870
> madvise_inject_error+0x260/0x334 mm/madvise.c:1115

MADV_HWPOISON is called on a huge zero page. The warning was raised
for this case intentionally (it was a BUG before), see commit
478d134e9506c7e9bfe2830ed03dd85e97966313 ("mm/huge_memory: do not
overkill when splitting huge_zero_page").

> do_madvise+0x258/0x5ac mm/madvise.c:1398
> __do_sys_madvise mm/madvise.c:1423 [inline]
> __se_sys_madvise mm/madvise.c:1421 [inline]
> __arm64_sys_madvise+0x2c/0x40 mm/madvise.c:1421
> __invoke_syscall arch/arm64/kernel/syscall.c:38 [inline]
> invoke_syscall arch/arm64/kernel/syscall.c:52 [inline]
> el0_svc_common+0x138/0x220 arch/arm64/kernel/syscall.c:142
> do_el0_svc+0x48/0x154 arch/arm64/kernel/syscall.c:206
> el0_svc+0x58/0x14c arch/arm64/kernel/entry-common.c:624
> el0t_64_sync_handler+0x84/0xf0 arch/arm64/kernel/entry-common.c:642
> el0t_64_sync+0x18c/0x190
> irq event stamp: 11566
> hardirqs last enabled at (11565): [<ffff8000081bd590>] __up_console_sem+0xb0/0xfc kernel/printk/printk.c:264
> hardirqs last disabled at (11566): [<ffff80000bfa4314>] el1_dbg+0x24/0x5c arch/arm64/kernel/entry-common.c:395
> softirqs last enabled at (11536): [<ffff8000080102e4>] _stext+0x2e4/0x37c
> softirqs last disabled at (11527): [<ffff800008101e20>] do_softirq_own_stack include/asm-generic/softirq_stack.h:10 [inline]
> softirqs last disabled at (11527): [<ffff800008101e20>] invoke_softirq+0x70/0xbc kernel/softirq.c:452
> ---[ end trace 0000000000000000 ]---
> Memory Failure: 0x10c110: thp split failed
> Memory failure: 0x10c110: recovery action for unsplit thp: Ignored
>
>
> ---
> 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.
> syzbot can test patches for this issue, for details see:
> https://goo.gl/tpsmEJ#testing-patches
>