Re: [RFC PATCH v3 00/12] Unify SMP stop generic logic to common code

From: Cristian Marussi
Date: Mon Jan 13 2020 - 12:05:53 EST


Hi Will

On 13/01/2020 16:40, Will Deacon wrote:
> On Thu, Dec 19, 2019 at 12:18:53PM +0000, Cristian Marussi wrote:
>> the logic underlying SMP stop and kexec crash procedures, beside containing
>> some arch-specific bits, is mostly generic and common across all archs:
>> despite this fact, such logic is now scattered across all architectures and
>> on some of them is flawed, in such a way that, under some specific
>> conditions, you can end up with a CPU left still running after a panic and
>> possibly lost across a subsequent kexec crash reboot. [1]
>
> Is this still the case even after 20bb759a66be ("panic: ensure preemption is
> disabled during panic()")?
>

v3 is based on 5.5-rc2 which seems to include 20bb759a66be, and when I tested before
re-posting a few weeks ago it was still failing as usual, i.e. kernel still alive after panic.
[but please be aware that to reproduce it, you need to have only one core online and another one
panicing while starting up (while still marked offline)]

Thanks

Cristian