Re: queue-5.10: Panic on shutdown at platform_shutdown+0x9

[Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>]

Hello,

On 09/02/25 21:27, Chuck Lever wrote:
> On 2/7/25 10:10 AM, Greg KH wrote:
>> On Thu, Feb 06, 2025 at 01:31:42PM -0500, Chuck Lever wrote:
>>> Hi -
>>>
>>> For the past 3-4 days, NFSD CI runs on queue-5.10.y have been failing. I
>>> looked into it today, and the test guest fails to reboot because it
>>> panics during a reboot shutdown:
>>>
>>> [  146.793087] BUG: unable to handle page fault for address:
>>> ffffffffffffffe8
>>> [  146.793918] #PF: supervisor read access in kernel mode
>>> [  146.794544] #PF: error_code(0x0000) - not-present page
>>> [  146.795172] PGD 3d5c14067 P4D 3d5c15067 PUD 3d5c17067 PMD 0
>>> [  146.795865] Oops: 0000 [#1] SMP NOPTI
>>> [  146.796326] CPU: 3 PID: 1 Comm: systemd-shutdow Not tainted
>>> 5.10.234-g99349f441fe1 #1
>>> [  146.797256] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS
>>> 1.16.3-2.fc40 04/01/2014
>>> [  146.798267] RIP: 0010:platform_shutdown+0x9/0x20
>>> [  146.798838] Code: b7 46 08 c3 cc cc cc cc 31 c0 83 bf a8 02 00 00 ff
>>> 75 ec c3 cc cc cc cc 66 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 8b 47
>>> 68 <48> 8b 40 e8 48 85 c0 74 09 48 83 ef 10 ff e0 0f 1f 00 c3 cc cc cc
>>> [  146.801012] RSP: 0018:ff7f86f440013de0 EFLAGS: 00010246
>>> [  146.801651] RAX: 0000000000000000 RBX: ff4f0637469df418 RCX:
>>> 0000000000000000
>>> [  146.802500] RDX: 0000000000000001 RSI: ff4f0637469df418 RDI:
>>> ff4f0637469df410
>>> [  146.803350] RBP: ffffffffb2e79220 R08: ff4f0637469dd808 R09:
>>> ffffffffb2c5c698
>>> [  146.804203] R10: 0000000000000000 R11: 0000000000000000 R12:
>>> ff4f0637469df410
>>> [  146.805059] R13: ff4f0637469df490 R14: 00000000fee1dead R15:
>>> 0000000000000000
>>> [  146.805909] FS:  00007f4e7ecc6b80(0000) GS:ff4f063aafd80000(0000)
>>> knlGS:0000000000000000
>>> [  146.806866] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>> [  146.807558] CR2: ffffffffffffffe8 CR3: 000000010ecb2001 CR4:
>>> 0000000000771ee0
>>> [  146.808412] DR0: 0000000000000000 DR1: 0000000000000000 DR2:
>>> 0000000000000000
>>> [  146.809262] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7:
>>> 0000000000000400
>>> [  146.810109] PKRU: 55555554
>>> [  146.810460] Call Trace:
>>> [  146.810791]  ? __die_body.cold+0x1a/0x1f
>>> [  146.811282]  ? no_context.constprop.0+0xf8/0x2f0
>>> [  146.811854]  ? exc_page_fault+0xc5/0x150
>>> [  146.812342]  ? asm_exc_page_fault+0x1e/0x30
>>> [  146.812862]  ? platform_shutdown+0x9/0x20
>>> [  146.813362]  device_shutdown+0x158/0x1c0
>>> [  146.813853]  __do_sys_reboot.cold+0x2f/0x5b
>>> [  146.814370]  ? vfs_writev+0x9b/0x110
>>> [  146.814824]  ? do_writev+0x57/0xf0
>>> [  146.815254]  do_syscall_64+0x30/0x40
>>> [  146.815708]  entry_SYSCALL_64_after_hwframe+0x67/0xd1
>>>
>>> Let me know how to further assist.
>>
>> Bisect?
> 
> First bad commit:
> 
> commit a06b4817f3d20721ae729d8b353457ff9fe6ff9c
> Author:     Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> AuthorDate: Thu Nov 19 13:46:11 2020 +0100
> Commit:     Sasha Levin <sashal@kernel.org>
> CommitDate: Tue Feb 4 13:04:31 2025 -0500
> 
>      driver core: platform: use bus_type functions
> 
>      [ Upstream commit 9c30921fe7994907e0b3e0637b2c8c0fc4b5171f ]
> 
>      This works towards the goal mentioned in 2006 in commit 594c8281f905
>      ("[PATCH] Add bus_type probe, remove, shutdown methods.").
> 
>      The functions are moved to where the other bus_type functions are
>      defined and renamed to match the already established naming scheme.
> 
>      Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
>      Link:
> https://lore.kernel.org/r/20201119124611.2573057-3-u.kleine-koenig@pengutronix.de
>      Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>      Stable-dep-of: bf5821909eb9 ("mtd: hyperbus: hbmc-am654: fix an OF
> node reference leak")
>      Signed-off-by: Sasha Levin <sashal@kernel.org>
> 

While one option is to drop this, maybe we apply this below fix as well 
instead of dropping the above as it is pulled in as stable-dep-of for 
some other commit?

commit 46e85af0cc53f35584e00bb5db7db6893d0e16e5
Author: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Date:   Sun Dec 13 02:55:33 2020 +0300

     driver core: platform: don't oops in platform_shutdown() on unbound 
devices

     On shutdown the driver core calls the bus' shutdown callback also for
     unbound devices. A driver's shutdown callback however is only 
called for
     devices bound to this driver. Commit 9c30921fe799 ("driver core:
     platform: use bus_type functions") changed the platform bus from driver
     callbacks to bus callbacks, so the shutdown function must be 
prepared to
     be called without a driver. Add the corresponding check in the shutdown
     function.

     Fixes: 9c30921fe799 ("driver core: platform: use bus_type functions")
     Tested-by: Guenter Roeck <linux@roeck-us.net>
     Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
     Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
     Link: 
https://lore.kernel.org/r/20201212235533.247537-1-dmitry.baryshkov@linaro.org
     Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This commit talks about fixing an oops in platform_shutdown()

Thanks,
Harshit