Re: [OOPS] Linux 5.14.19 crashes and burns!

From: Thorsten Leemhuis
Date: Thu Nov 18 2021 - 09:23:49 EST


On 18.11.21 15:04, Greg KH wrote:
> On Thu, Nov 18, 2021 at 10:29:14AM +0100, Greg KH wrote:
>> On Thu, Nov 18, 2021 at 10:15:23AM +0100, Thorsten Leemhuis wrote:
>>> On 18.11.21 09:57, Thorsten Leemhuis wrote:
>>>> Lo! CCing stable and regressions list. FWIW, I have a Fedora 34 VM that
>>>> stopped booting with 5.14.19 as well. More inline:
>>>>
>>>> On 18.11.21 01:55, Chris Rankin wrote:
>>>>>
>>>>> I have tried to boot a vanilla 5.14.19 kernel, but it crashes when
>>>>> "switching root" away from the initramfs. ("Unable to handle page
>>>>> fault...)
>>>>>
>>>>> Google's "copy text from image" feature has managed to scrape this
>>>>> information from my phone camera:
>>>>>
>>>>> 1tch Noo
>>>>> BUG: unable to handle page fault for address: ffffc980006cfeDS
>>>>> #PF: supervisor read access in kernel mode #PF: error_code (8x8888) -
>>>>> not-present page
>>>>> PGD 100000067 P4D 100000067 PUD 18885e867 PMD 104486867 PTE 8
>>>>> Oops: 0888 [#1] PREEMPT SMP PTI
>>>>> CPU: 6 PID: 1 Comm: systemd Tainted: G Hardware name: Gigabyte
>>>>> Technology Co., Ltd. EX58-UD3R/EX58-UD3R, BIOS FB 05/04/2009
>>>>> 5.14.19 #1
>>>>> RIP: 0010: __unwind_start+8xb5/8x15f
>>>>> Code: 48 8d 8d 88 88 88 88 48 89 e2 48 89 e8 48 89 4d 48 48 89 55 38
>>>>> 48 RSP: 0018:ffffc90088823bf0 EFLAGS: 00010006
>>>>> RAX: ffffc900006efde8 RBX: 0000000000000000 RCX: 08 RDX:
>>>>> ffffc900006efe18 RSI : ffff88818a9f6c80 RDI: ffffc
>>>>> RBP: ffffc90808823c18 R08: 00000000000001de R89: R10: ffff888107d20000
>>>>> R11: ffff888183b4ba88 R12: ffffc900006ef dell
>>>>> R13: ffff88810a9f73bc R14: ffff888103c58480 R15: FS: 00007f528fd19b40
>>>>> (0800) GS:ffff888343488808 (0888) knl6S:
>>>>> CS: 0818 DS: 0000 ES: 0000 CRO: 0888000088858833
>>>>> CR2: ffffc900006efe88 CR3: 088080818186a888 CR4: 8888
>>>>> Call Trace:
>>>>> <TASK>
>>>>> get_wchan+8x42/8x8f
>>>>> get_wchan+8x45/8x59
>>>>> do_task_stat+0x3ab/0x38
>>>>> proc_single_show+8x1e/8x68
>>>>> seq_read_iter+0x151/8x342 seq_read+8xf1/8x117
>>>>> uf's_read+Bxa3/8x183
>>>>> ksys_read+8x71/8xb9 do_syscal1_64+8x6d/8x88
>>>>> entry_SYSCALL_64_after_huframe+8x11/8xne
>>>>> RIP: 8033:8x7f529884f832
>>>>> Code: c0 e9 b2 fe ff ff 50 48 8d 3d ea 2e Bc 80 68 65 £9 01 00 0f 16 44 000
>>>>> RSP: 082b:00087ffed3b91f18 EFLAGS: 88888246 ORIG RAX:
>>>>> RAX: ffffffffffffffda RBX: 8888559462036658 RCX: 000071529084/83
>>>>> RDX: 0000000000000100 RSI: 8088559462c90b78 RDI:
>>>>> RBP: 000071529894a300 BAR:
>>>>> te fa
>>>>>
>>>>> And also this:
>>>>>
>>>>> do_syscall_64+0x6d/0x80
>>>>> entry_SYSCALL_64_after_huframe+0x44/0xae
>>>>> RIP: 8033:0x7f529884f832
>>>>> Code: c0 e9 b2 fe ff ff 50 48 8d 3d ea Ze Bc 80 e8 b5 f9 01 00 0f 1f
>>>>> 44 00 00 f3 Bf 1e fa 64 8b 84 25 18 88 RSP: 082b:00087ffed3b91f 18
>>>>> EFLAGS: 00000246 ORIG_RAX: 0000000000000000
>>>>> RAX: ffffffffffffffda RBX: 0000559462c36650 RCX: 000071529084f832 RDX:
>>>>> 0000000000000400 RSI: 0000559462c90b70 RDI: 0000000000000006
>>>>> RBP: 000071529094a3a0 R08: 0000000000000006 RO9: 0000000000000001 R18:
>>>>> 0000000000001000 R11: 0000000000000246 R12: 00007f528fd198f0
>>>>> R13: 0000000000000168 R14: 00007f52909497a0 R15: 00000000000001468
>>>>> </TASK>
>>>>> Modules linked in: ext4 crc32c_generic crc16 mbcache jbd2 sr_mod
>>>>> sd_mod cdrom hid_microsoft usbhid amdgpu uh ci drm_kms_helper ehci_hcd
>>>>> cf bf illrect syscopyarea cfbimgblt sysfillrect sysimgblt xhci pci
>>>>> xhci_hcd fb_sys_ mod usb_common drm_panel_orientation_quirks
>>>>> ipmi_devintf ipmi_msghandler msr sha256_ssse3 sha256_generic ipu CRZ:
>>>>> ffffc900006efe08
>>>>> --- end trace 9771b79967a8dd89 ]--- RIP: 8010:_unwind_start+8xb5/0x15f
>>>>> Code: 48 8d 8d 00 00 00 00 48 89 e2 48 89 e8 48 89 4d 48 48 89 55 38
>>>>> 48 89 45 40 eb 29 48 8b 86 98 Ba 80 00 RSP: 0818:ffffc90000023bf0
>>>>> EFLAGS: 00010006 RAX: ffffc900006efde0 RBX: 0000000000000000 RCX:
>>>>> 0000000000000000
>>>>> RDX: ffffc900006efe18 RSI: ffff88810a9f6c00 RDI: ffffc90000023c78 RBP:
>>>>> ffffc98800023c18 R08: 00000000000001de R09: 00000000005b20c6
>>>>> R10: ffff888107120000 R11: ffff888103b4ba80 R12: ffffc900006ef de
>>>>> R13: ffff88818a9f73bc R14: ffff888103c50480 R15: 0000000000000000 FS:
>>>>> 00007f528fd19b40(8000) GS:ffff888343180000 (0800)
>>>>> knlGS:0000000000000000 CS: 0810 DS: 0000 ES: 0000 CRO:
>>>>> 0000000080050033
>>>>> CR2: ffffc908806efe88 CR3: 000000010186a000 CR4: 00000000000006e0
>>>>> note: systemd [1] exited with preempt_count 1
>>>>> Kernel panic - not syncing: Attempted to kill init!
>>>>> exitcode=0x00000009 Kernel Offset: disabled
>>>>> --- [ end Kernel panic not syncing: Attempted to kill initf
>>>>> exitcode=0x00000009 1--- -
>>>>>
>>>>> I cannot capture the exact oops via any other means, although I can
>>>>> send the original camera pictures that I captured the text from, on
>>>>> request.
>>>>
>>>> By default I didn't get to see any messages,
>>>
>>> BTW: that was a stupid error on my side, I get them now.
>>>
>>>> the VM just hangs when
>>>> switching to root. Did anyone else already report or even track this
>>>> down already? Guess otherwise I need to take a closer look and maybe
>>>> start bisecting...
>>>
>>> On a quick looks above problems seems to be similar to the one already
>>> reported wrt to WCHAN:
>>> https://lore.kernel.org/stable/YZV02RCRVHIa144u@xxxxxxxxxxxxxxxxxxxx/
>>>
>>> Greg, is there any reason why you didn't drop them from 5.14.y as well,
>>> as you did for 5.15.y:
>>> https://lore.kernel.org/all/YZYLC9D6zpUneYtn@xxxxxxxxx/
>>>
>>> Should I try reverting these three and see if the situations improves?
>>
>> I will push out a new 5.14 release in a few hours with those commits
>> reverted, as yes, you are right, they should not be there if I have
>> dropped them from 5.15.y
>
> Should be resolved in 5.14.20.

Thx, it is, I just booted it successfully (my test machine just finished
compilation a minute after your mail arrived...)

Ciao, Thorsten

P.S.: Feel free to ignore the following lines, they are meant for
regzbot, my Linux kernel regression tracking bot

#regzbot fixed-by 1b86960dc451412e7029b267e9e87536fdd554cb

TWIMC: yes, the commits with the fix had a link to a msg in this thread,
but regzbot didn't automatically detect them as fixes for the
regression, as it right now expects a Link: to what it considers the
report of the tracked regression -- in this case that would have been
the msg that started this thread. Maybe regzbot should not be so strict
and accept Link: tags like these as well. Not sure yet; if something
like this happens frequently, I'll change regzbot. For now I just marked
the regression manually as fixed using above tag.