Re: [syzbot] BUG: soft lockup in call_usermodehelper_exec_async

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, 17 Sept 2021 at 09:30, syzbot
<syzbot+c3963f4f025fa5dfb78b@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> Hello,
>
> syzbot found the following issue on:
>
> HEAD commit:    8cae8cd89f05 seq_file: disallow extremely large seq buffer..
> git tree:       upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=135dbd12300000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=548fbb1e84c1808e
> dashboard link: https://syzkaller.appspot.com/bug?extid=c3963f4f025fa5dfb78b
> compiler:       aarch64-linux-gnu-gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.1
> userspace arch: arm
>
> 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+c3963f4f025fa5dfb78b@xxxxxxxxxxxxxxxxxxxxxxxxx
>
> watchdog: BUG: soft lockup - CPU#1 stuck for 22s! [kworker/u4:5:21352]
> Modules linked in:
> irq event stamp: 196
> hardirqs last  enabled at (195): [<ffff80001457eb88>] exit_to_kernel_mode+0x38/0x230 arch/arm64/kernel/entry-common.c:71
> hardirqs last disabled at (196): [<ffff80001457eb40>] enter_el1_irq_or_nmi+0x10/0x20 arch/arm64/kernel/entry-common.c:113
> softirqs last  enabled at (116): [<ffff8000100109a4>] _stext+0x9a4/0x1050
> softirqs last disabled at (51): [<ffff800010161d30>] do_softirq_own_stack include/asm-generic/softirq_stack.h:10 [inline]
> softirqs last disabled at (51): [<ffff800010161d30>] invoke_softirq kernel/softirq.c:439 [inline]
> softirqs last disabled at (51): [<ffff800010161d30>] __irq_exit_rcu+0x470/0x520 kernel/softirq.c:636
> CPU: 1 PID: 21352 Comm: kworker/u4:5 Not tainted 5.14.0-rc2-syzkaller-00001-g8cae8cd89f05 #0
> Hardware name: linux,dummy-virt (DT)
> pstate: 00000005 (nzcv daif -PAN -UAO -TCO BTYPE=--)
> pc : clear_page+0x14/0x28 arch/arm64/lib/clear_page.S:23
> lr : clear_highpage include/linux/highmem.h:184 [inline]
> lr : kernel_init_free_pages.part.0+0x6c/0x17c mm/page_alloc.c:1283
> sp : ffff800018c67380
> x29: ffff800018c67380 x28: 0000000000000000 x27: ffff00007fc11ec0
> x26: ffff00000fe7aef8 x25: ffff600001fcf5df x24: 00000000000014b8
> x23: ffff00000fe79a40 x22: fffffc0000171340 x21: ffff800015706c40
> x20: dfff800000000000 x19: fffffc0000171300 x18: ffff00006aa70bc8
> x17: ffff800055045000 x16: ffff800018598000 x15: ffff000010dd0600
> x14: 1ffff0000318ce3a x13: 0000000000000014 x12: ffff7f800002e267
> x11: 1fffff800002e266 x10: ffff7f800002e266 x9 : 0000000000000000
> x8 : ffff600000b89a00 x7 : 0000000000000000 x6 : 000000000000003f
> x5 : 0000000000000040 x4 : 1ffff0000304db70 x3 : 1fffe00001fcf349
> x2 : 0000000000000004 x1 : 0000000000000040 x0 : ffff000005c4ca40
> Call trace:
>  clear_page+0x14/0x28 arch/arm64/lib/clear_page.S:21
>  kernel_init_free_pages mm/page_alloc.c:1280 [inline]
>  post_alloc_hook+0x1ac/0x25c mm/page_alloc.c:2423
>  prep_new_page mm/page_alloc.c:2433 [inline]
>  get_page_from_freelist+0x1840/0x2310 mm/page_alloc.c:4166
>  __alloc_pages+0x1a8/0x2234 mm/page_alloc.c:5388
>  alloc_pages_vma+0xb4/0x4dc mm/mempolicy.c:2208
>  alloc_zeroed_user_highpage_movable+0x9c/0xd0 arch/arm64/mm/fault.c:939
>  do_anonymous_page mm/memory.c:3767 [inline]
>  handle_pte_fault mm/memory.c:4547 [inline]
>  __handle_mm_fault+0xcbc/0x25e4 mm/memory.c:4684
>  handle_mm_fault+0x1dc/0x4ec mm/memory.c:4782
>  faultin_page mm/gup.c:951 [inline]
>  __get_user_pages+0x3e4/0x744 mm/gup.c:1173
>  __get_user_pages_locked mm/gup.c:1359 [inline]
>  __get_user_pages_remote+0x10c/0x500 mm/gup.c:1865
>  get_user_pages_remote+0x28/0x40 mm/gup.c:1938
>  get_arg_page+0x9c/0x1bc fs/exec.c:220
>  copy_string_kernel+0x170/0x294 fs/exec.c:631
>  kernel_execve+0x1ac/0x380 fs/exec.c:1961
>  call_usermodehelper_exec_async+0x260/0x460 kernel/umh.c:112
>  ret_from_fork+0x10/0x18 arch/arm64/kernel/entry.S:782

+linux-arm-kernels

This is caused by broken CONFIG_CMDLINE support on arm64, for context see:
https://lore.kernel.org/lkml/CACT4Y+a4S-oXsjgwDh3SmERqKFF1QbapvX6NiSpn51KRtqvTiQ@xxxxxxxxxxxxxx/

syzbot should use 150s stall timeout, but it got's default 20s which
produces false positive flakes.

#syz dup: BUG: soft lockup in handle_mm_fault (2)




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux