On Tue, Nov 1, 2022 at 1:25 PM Igor Petruk <[email protected]> wrote:
>
> On Sun, Oct 30, 2022 at 11:04 AM Paul de Weerd <[email protected]> wrote:
> >
> > On Fri, Oct 28, 2022 at 01:15:08PM +0200, Sebastian Oswald wrote:
> > | >Glad to hear it worked.
> > | >
> > | >Hypothetically... I were to work on building a sysctl like:
> > | >
> > | >kern.acpi.mask_gpes=0x6f,0x42
> > | >
> > | >for people to enjoy normal OS operation before they get their
> > | >firmware fixes, would that be a good idea? Does it have a chance
> > | >to be accepted?
> > | >
> > | >My concern mainly is the interface. I don't know if a more
> > | >generic solution for all other interrupts need to be built,
> > | >not just ACPI GPE.
> > | >
> > | >Thanks,
> > | >Igor.
> > |
> > |
> > | This diff also solves the problem on my systems. Thanks!
> > |
> > | Regarding that sysctl: Given the large number of borked ACPI
> > | tables/implementations (and unwilling/incompetent vendors) out there,
> > | I think it would be benefical to have such masking capabilities.
> >
> > How would a regular user know which GPEs to mask?  This time, it's
> > 0x6f, but what will it be on the next poor implementation?
> >
> > I don't have an answer, but I think there's more to it than just
> > saying "mask these GPEs".
> >
> > Paul
> >
> > --
> > >++++++++[<++++++++++>-]<+++++++.>+++[<------>-]<.>+++[<+
> > +++++++++++>-]<.>++[<------------>-]<+.--------------.[-]
> >                  http://www.weirdnet.nl/
>
> This is a good point Paul, it is not trivial. In this thread we
> had to resort to adding printfs to the custom kernel build.
>
> However, next users can find this as a "known problem" of
> particular hardware in mailings lists, forums and at
> extremely low cost verify if it helps.
>
> Currently, due to lack of fix in the kernel or in the hardware,
> the OpenBSD user has to do the following:
>
> * Figure out that this is the root cause
> * Switch to current
> * Learn how to compile OpenBSD from sources
> * Get the sources, apply the patch, build the kernel,
>   the base system, xenocara
> * Maintain updates manually by refetching the source
>   code and re-applying the patch
> * Face the consequence of current source based
>   instability and packages being occasionaly out of sync
>
> If there was a mitigation (like in Linux and FreeBSD that
> users do use for imperfect hardware) - everyone could
> just use normal stable OpenBSD.
>
> I'd argue that the pain of figuring out what is the GPE and
> adding a config line is uncomparable to the cost of the current
> experience right now.
>
> For example due to the use of
> current, the application I am developing right now suddenly
> broke with "unable to allocate page guard" purely after
> another  kernel+base+xenocara rebuild (that gook many hours).
> I don't know where to even look for the problem.
> This ruins otherwise perfect OpenBSD experience.

On a related note, I've posted a question in the Intel
community forum, feel free to comment

https://community.intel.com/t5/Intel-NUCs/APCI-GPE-0x6F-Interrupt-Storm-under-OpenBSD/td-p/1426755

Thanks,
Igor.

Reply via email to