FYI, Arch got 5.4.1 this morning, and it looks like the issue is still
present.
On 11/20/19 12:34 AM, Ryusuke Konishi wrote:
> As the result of bisection, it turned out that commit
> f4bdb2697ccc9cecf1a9de86905c309ad901da4c on 5.3.y
> ("mm/filemap.c: don't initiate writeback if mapping has no dirty pages")
> triggers the crash.
>
> This patch was back ported from upstream kernel 5.4-rc1, and
> applied to 4.19.84 as well.
>
> Will look into the change further and why it doesn't hit latest
> kernels (5.4-rc8, etc).
>
> Regards,
> Ryusuke Konishi
>
>
>
>
>
>
> On Wed, Nov 20, 2019 at 12:46 AM Ryusuke Konishi
> <konishi.ryusuke@xxxxxxxxx> wrote:
>>
>> Yeah, this looks a regression of the 5.3.11 stable kernel.
>>
>> I could reproduce the crash with 5.3.11.
>>
>> 5.3, 5.3.10 and 5.4-rc8 all worked fine.
>>
>> The regression would be identified by bisecting commits
>> between kernel 5.3.10 and 5.3.11.
>>
>> Thanks,
>> Ryusuke Konishi
>>
>> 2019年11月19日(火) 23:40 <linux-nilfs-ml.via.forwarder@xxxxxxxxxxxx>:
>>>
>>> I believe I have the same issue as Walton Hoops.
>>>
>>> I do not have a serial console, and all my filesystems are NILFS2,
>>>
>>> so I do not currently have an easy means to capture the crash.
>>>
>>> However, I experienced the same Kernel BUG on upgrading to kernel 5.11
>>>
>>> from 5.10. Reverting back to 5.10 results in a clean boot.
>>>
>>> The actual kernels I use are the Ubuntu 'mainline kernels'
>>>
>>> 5.3.10-050310-generic works
>>>
>>> 5.3.11-050311-generic crashes with Kernel BUG.
>>>
>>> So I suspect a change between those two versions has triggered the problem.
>>>
>>>
