Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Guenter Roeck]

On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
> Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
> introduced a workaround for MWAIT for a specific x86 system.
> 
> Move the code outside of acpi_processor.c to acpi/x86/ directory for
> consistency and rename the functions associated with it, so their names
> start with "acpi_proc_quirk_" to make the goal obvious.
> 
> No intentional functional impact.
> 

Except for:

ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'

which breaks all ia64 builds.

Time to retire that architecture yet ? No one but me seems to even
build test it.

Guenter

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Rafael J. Wysocki]

On Tue, Aug 29, 2023 at 3:44 PM Guenter Roeck <linux@roeck-us.net> wrote:
>
> On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
> > Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
> > introduced a workaround for MWAIT for a specific x86 system.
> >
> > Move the code outside of acpi_processor.c to acpi/x86/ directory for
> > consistency and rename the functions associated with it, so their names
> > start with "acpi_proc_quirk_" to make the goal obvious.
> >
> > No intentional functional impact.
> >
>
> Except for:
>
> ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
> acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
> ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
> processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'
>
> which breaks all ia64 builds.
>
> Time to retire that architecture yet ? No one but me seems to even
> build test it.

Including 0-day it seems.  This had been in linux-next for several weeks.

Michal, can you have a look at this please?

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Wilczynski, Michal]

On 8/29/2023 3:54 PM, Rafael J. Wysocki wrote:
> On Tue, Aug 29, 2023 at 3:44 PM Guenter Roeck <linux@roeck-us.net> wrote:
>> On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
>>> Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
>>> introduced a workaround for MWAIT for a specific x86 system.
>>>
>>> Move the code outside of acpi_processor.c to acpi/x86/ directory for
>>> consistency and rename the functions associated with it, so their names
>>> start with "acpi_proc_quirk_" to make the goal obvious.
>>>
>>> No intentional functional impact.
>>>
>> Except for:
>>
>> ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
>> acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
>> ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
>> processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'
>>
>> which breaks all ia64 builds.
>>
>> Time to retire that architecture yet ? No one but me seems to even
>> build test it.
> Including 0-day it seems.  This had been in linux-next for several weeks.
>
> Michal, can you have a look at this please?

Hi,
I'll take a look and get back to you with a fix,

Regards,
Michał

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Rafael J. Wysocki]

[-- Attachment #1: Type: text/plain, Size: 1558 bytes --]

On Tue, Aug 29, 2023 at 3:58 PM Wilczynski, Michal
<michal.wilczynski@intel.com> wrote:
>
>
>
> On 8/29/2023 3:54 PM, Rafael J. Wysocki wrote:
> > On Tue, Aug 29, 2023 at 3:44 PM Guenter Roeck <linux@roeck-us.net> wrote:
> >> On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
> >>> Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
> >>> introduced a workaround for MWAIT for a specific x86 system.
> >>>
> >>> Move the code outside of acpi_processor.c to acpi/x86/ directory for
> >>> consistency and rename the functions associated with it, so their names
> >>> start with "acpi_proc_quirk_" to make the goal obvious.
> >>>
> >>> No intentional functional impact.
> >>>
> >> Except for:
> >>
> >> ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
> >> acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
> >> ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
> >> processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'
> >>
> >> which breaks all ia64 builds.
> >>
> >> Time to retire that architecture yet ? No one but me seems to even
> >> build test it.
> > Including 0-day it seems.  This had been in linux-next for several weeks.
> >
> > Michal, can you have a look at this please?
>
> Hi,
> I'll take a look and get back to you with a fix,

Actually, if I'm not mistaken, the attached patch should be sufficient.

[-- Attachment #2: acpi-processor-fix-ia64-buil.patch --]
[-- Type: text/x-patch, Size: 736 bytes --]

---
 drivers/acpi/internal.h |   14 ++++----------
 1 file changed, 4 insertions(+), 10 deletions(-)

Index: linux-pm/drivers/acpi/internal.h
===================================================================
--- linux-pm.orig/drivers/acpi/internal.h
+++ linux-pm/drivers/acpi/internal.h
@@ -148,8 +148,11 @@ int acpi_wakeup_device_init(void);
 #ifdef CONFIG_ARCH_MIGHT_HAVE_ACPI_PDC
 void acpi_early_processor_control_setup(void);
 void acpi_early_processor_set_pdc(void);
-
+#ifdef CONFIG_X86
 void acpi_proc_quirk_mwait_check(void);
+#else
+static inline void acpi_proc_quirk_mwait_check(void) {}
+#endif
 bool processor_physically_present(acpi_handle handle);
 #else
 static inline void acpi_early_processor_control_setup(void) {}

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Wilczynski, Michal]

On 8/29/2023 4:03 PM, Rafael J. Wysocki wrote:
> On Tue, Aug 29, 2023 at 3:58 PM Wilczynski, Michal
> <michal.wilczynski@intel.com> wrote:
>>
>>
>> On 8/29/2023 3:54 PM, Rafael J. Wysocki wrote:
>>> On Tue, Aug 29, 2023 at 3:44 PM Guenter Roeck <linux@roeck-us.net> wrote:
>>>> On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
>>>>> Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
>>>>> introduced a workaround for MWAIT for a specific x86 system.
>>>>>
>>>>> Move the code outside of acpi_processor.c to acpi/x86/ directory for
>>>>> consistency and rename the functions associated with it, so their names
>>>>> start with "acpi_proc_quirk_" to make the goal obvious.
>>>>>
>>>>> No intentional functional impact.
>>>>>
>>>> Except for:
>>>>
>>>> ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
>>>> acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
>>>> ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
>>>> processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'
>>>>
>>>> which breaks all ia64 builds.
>>>>
>>>> Time to retire that architecture yet ? No one but me seems to even
>>>> build test it.
>>> Including 0-day it seems.  This had been in linux-next for several weeks.
>>>
>>> Michal, can you have a look at this please?
>> Hi,
>> I'll take a look and get back to you with a fix,
> Actually, if I'm not mistaken, the attached patch should be sufficient.

Exactly, adding this empty stub will make sure there is no linker error, this function
is relevant only for x86 anyway.

If ia64 support for 0-day was turned off then I think it was rather recently, cause I've
seen it working I think back in May.

To be honest I'm not sure what is being done in such cases ? Will you send a fix to Linus
directly, or should I prepare a patch and sent in on the list ??

Thanks !
Michał

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Rafael J. Wysocki]

On Tue, Aug 29, 2023 at 4:21 PM Wilczynski, Michal
<michal.wilczynski@intel.com> wrote:
>
>
>
> On 8/29/2023 4:03 PM, Rafael J. Wysocki wrote:
> > On Tue, Aug 29, 2023 at 3:58 PM Wilczynski, Michal
> > <michal.wilczynski@intel.com> wrote:
> >>
> >>
> >> On 8/29/2023 3:54 PM, Rafael J. Wysocki wrote:
> >>> On Tue, Aug 29, 2023 at 3:44 PM Guenter Roeck <linux@roeck-us.net> wrote:
> >>>> On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
> >>>>> Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
> >>>>> introduced a workaround for MWAIT for a specific x86 system.
> >>>>>
> >>>>> Move the code outside of acpi_processor.c to acpi/x86/ directory for
> >>>>> consistency and rename the functions associated with it, so their names
> >>>>> start with "acpi_proc_quirk_" to make the goal obvious.
> >>>>>
> >>>>> No intentional functional impact.
> >>>>>
> >>>> Except for:
> >>>>
> >>>> ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
> >>>> acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
> >>>> ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
> >>>> processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'
> >>>>
> >>>> which breaks all ia64 builds.
> >>>>
> >>>> Time to retire that architecture yet ? No one but me seems to even
> >>>> build test it.
> >>> Including 0-day it seems.  This had been in linux-next for several weeks.
> >>>
> >>> Michal, can you have a look at this please?
> >> Hi,
> >> I'll take a look and get back to you with a fix,
> > Actually, if I'm not mistaken, the attached patch should be sufficient.
>
> Exactly, adding this empty stub will make sure there is no linker error, this function
> is relevant only for x86 anyway.
>
> If ia64 support for 0-day was turned off then I think it was rather recently, cause I've
> seen it working I think back in May.
>
> To be honest I'm not sure what is being done in such cases ? Will you send a fix to Linus
> directly, or should I prepare a patch and sent in on the list ??

I'll add a changelog to the fix, post it and merge it directly, so you
don't need to do anything (other than ACKing the fix if you will).

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Wilczynski, Michal]

On 8/29/2023 4:25 PM, Rafael J. Wysocki wrote:
> On Tue, Aug 29, 2023 at 4:21 PM Wilczynski, Michal
> <michal.wilczynski@intel.com> wrote:
>>
>>
>> On 8/29/2023 4:03 PM, Rafael J. Wysocki wrote:
>>> On Tue, Aug 29, 2023 at 3:58 PM Wilczynski, Michal
>>> <michal.wilczynski@intel.com> wrote:
>>>>
>>>> On 8/29/2023 3:54 PM, Rafael J. Wysocki wrote:
>>>>> On Tue, Aug 29, 2023 at 3:44 PM Guenter Roeck <linux@roeck-us.net> wrote:
>>>>>> On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
>>>>>>> Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
>>>>>>> introduced a workaround for MWAIT for a specific x86 system.
>>>>>>>
>>>>>>> Move the code outside of acpi_processor.c to acpi/x86/ directory for
>>>>>>> consistency and rename the functions associated with it, so their names
>>>>>>> start with "acpi_proc_quirk_" to make the goal obvious.
>>>>>>>
>>>>>>> No intentional functional impact.
>>>>>>>
>>>>>> Except for:
>>>>>>
>>>>>> ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
>>>>>> acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
>>>>>> ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
>>>>>> processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'
>>>>>>
>>>>>> which breaks all ia64 builds.
>>>>>>
>>>>>> Time to retire that architecture yet ? No one but me seems to even
>>>>>> build test it.
>>>>> Including 0-day it seems.  This had been in linux-next for several weeks.
>>>>>
>>>>> Michal, can you have a look at this please?
>>>> Hi,
>>>> I'll take a look and get back to you with a fix,
>>> Actually, if I'm not mistaken, the attached patch should be sufficient.
>> Exactly, adding this empty stub will make sure there is no linker error, this function
>> is relevant only for x86 anyway.
>>
>> If ia64 support for 0-day was turned off then I think it was rather recently, cause I've
>> seen it working I think back in May.
>>
>> To be honest I'm not sure what is being done in such cases ? Will you send a fix to Linus
>> directly, or should I prepare a patch and sent in on the list ??
> I'll add a changelog to the fix, post it and merge it directly, so you
> don't need to do anything (other than ACKing the fix if you will).

OK, thanks !

For the fix:
Acked-by: Michal Wilczynski <michal.wilczynski@intel.com>

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Guenter Roeck]

On Tue, Aug 29, 2023 at 04:38:33PM +0200, Wilczynski, Michal wrote:
> 
[ ... ]
> For the fix:
> Acked-by: Michal Wilczynski <michal.wilczynski@intel.com>
> 

Or just drop it. From Ard Biesheuvel's e-mail sent to 0-day:

    > Hello all,
    >
    > The Linux community is currently in the process of deprecating the
    > Itanium IA-64 architecture, which no longer has a maintainer or any
    > actual users.
    >
    > We aim to remove it entirely from the Linux tree before the end of the
    > year, but one of the first steps is to stop build testing it - there
    > is really no point any longer in spending any cycles on this, and it
    > is definitely undesirable to inform developers by email if their
    > changes result in any build issues on IA-64.
    >
    > So please remove IA64 from the set of architectures that are build
    > tested by LKP.

This is from the 0-day commit log disabling ia64 build tests. I have now
disabled ia64 in my build tests as well.

Sorry for the noise.

Guenter

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Rafael J. Wysocki]

On Tue, Aug 29, 2023 at 4:53 PM Guenter Roeck <linux@roeck-us.net> wrote:
>
> On Tue, Aug 29, 2023 at 04:38:33PM +0200, Wilczynski, Michal wrote:
> >
> [ ... ]
> > For the fix:
> > Acked-by: Michal Wilczynski <michal.wilczynski@intel.com>
> >
>
> Or just drop it. From Ard Biesheuvel's e-mail sent to 0-day:

Yeah.

>     > Hello all,
>     >
>     > The Linux community is currently in the process of deprecating the
>     > Itanium IA-64 architecture, which no longer has a maintainer or any
>     > actual users.
>     >
>     > We aim to remove it entirely from the Linux tree before the end of the
>     > year, but one of the first steps is to stop build testing it - there
>     > is really no point any longer in spending any cycles on this, and it
>     > is definitely undesirable to inform developers by email if their
>     > changes result in any build issues on IA-64.
>     >
>     > So please remove IA64 from the set of architectures that are build
>     > tested by LKP.
>
> This is from the 0-day commit log disabling ia64 build tests. I have now
> disabled ia64 in my build tests as well.
>
> Sorry for the noise.

No worries, thanks!

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Andy Shevchenko]

On Tue, Aug 29, 2023 at 07:53:21AM -0700, Guenter Roeck wrote:
> On Tue, Aug 29, 2023 at 04:38:33PM +0200, Wilczynski, Michal wrote:

[ ... ]

> Or just drop it. From Ard Biesheuvel's e-mail sent to 0-day:
> 
>     > Hello all,
>     >
>     > The Linux community is currently in the process of deprecating the
>     > Itanium IA-64 architecture, which no longer has a maintainer or any
>     > actual users.
>     >
>     > We aim to remove it entirely from the Linux tree before the end of the
>     > year, but one of the first steps is to stop build testing it - there
>     > is really no point any longer in spending any cycles on this, and it
>     > is definitely undesirable to inform developers by email if their
>     > changes result in any build issues on IA-64.
>     >
>     > So please remove IA64 from the set of architectures that are build
>     > tested by LKP.
> 
> This is from the 0-day commit log disabling ia64 build tests. I have now
> disabled ia64 in my build tests as well.

Oh, nice! Thank you for digging this up!

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Andy Shevchenko]

On Tue, Aug 29, 2023 at 03:54:18PM +0200, Rafael J. Wysocki wrote:
> On Tue, Aug 29, 2023 at 3:44 PM Guenter Roeck <linux@roeck-us.net> wrote:
> > On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
> > > Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
> > > introduced a workaround for MWAIT for a specific x86 system.
> > >
> > > Move the code outside of acpi_processor.c to acpi/x86/ directory for
> > > consistency and rename the functions associated with it, so their names
> > > start with "acpi_proc_quirk_" to make the goal obvious.
> > >
> > > No intentional functional impact.
> >
> > Except for:
> >
> > ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
> > acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
> > ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
> > processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'
> >
> > which breaks all ia64 builds.
> >
> > Time to retire that architecture yet ? No one but me seems to even
> > build test it.
> 
> Including 0-day it seems.  This had been in linux-next for several weeks.

IIRC someone explicitly asked to shut down 0-day on ia64. Maybe I'm mistaken?
I would suggest to ask internally Dave Hansen et al for this.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Guenter Roeck]

On 8/29/23 07:02, Andy Shevchenko wrote:
> On Tue, Aug 29, 2023 at 03:54:18PM +0200, Rafael J. Wysocki wrote:
>> On Tue, Aug 29, 2023 at 3:44 PM Guenter Roeck <linux@roeck-us.net> wrote:
>>> On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
>>>> Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
>>>> introduced a workaround for MWAIT for a specific x86 system.
>>>>
>>>> Move the code outside of acpi_processor.c to acpi/x86/ directory for
>>>> consistency and rename the functions associated with it, so their names
>>>> start with "acpi_proc_quirk_" to make the goal obvious.
>>>>
>>>> No intentional functional impact.
>>>
>>> Except for:
>>>
>>> ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
>>> acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
>>> ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
>>> processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'
>>>
>>> which breaks all ia64 builds.
>>>
>>> Time to retire that architecture yet ? No one but me seems to even
>>> build test it.
>>
>> Including 0-day it seems.  This had been in linux-next for several weeks.
> 
> IIRC someone explicitly asked to shut down 0-day on ia64. Maybe I'm mistaken?
> I would suggest to ask internally Dave Hansen et al for this.
> 

Actually, that is correct. Build testing it was removed from 0-day on request
from Ard Biesheuvel in preparation of its removal from the Linux kernel.

I'll stop build testing it as well.

Thanks,
Guenter

Re: [PATCH] ACPI: processor: Move MWAIT quirk out of acpi_processor.c

[Guenter Roeck]

On 8/29/23 06:54, Rafael J. Wysocki wrote:
> On Tue, Aug 29, 2023 at 3:44 PM Guenter Roeck <linux@roeck-us.net> wrote:
>>
>> On Mon, Jul 10, 2023 at 05:03:29PM +0300, Michal Wilczynski wrote:
>>> Commit 2a2a64714d9c ("ACPI: Disable MWAIT via DMI on broken Compal board")
>>> introduced a workaround for MWAIT for a specific x86 system.
>>>
>>> Move the code outside of acpi_processor.c to acpi/x86/ directory for
>>> consistency and rename the functions associated with it, so their names
>>> start with "acpi_proc_quirk_" to make the goal obvious.
>>>
>>> No intentional functional impact.
>>>
>>
>> Except for:
>>
>> ia64-linux-ld: drivers/acpi/acpi_processor.o: in function `acpi_early_processor_control_setup':
>> acpi_processor.c:(.init.text+0x712): undefined reference to `acpi_proc_quirk_mwait_check'
>> ia64-linux-ld: drivers/acpi/processor_pdc.o: in function `acpi_early_processor_set_pdc':
>> processor_pdc.c:(.init.text+0x72): undefined reference to `acpi_proc_quirk_mwait_check'
>>
>> which breaks all ia64 builds.
>>
>> Time to retire that architecture yet ? No one but me seems to even
>> build test it.
> 
> Including 0-day it seems.  This had been in linux-next for several weeks.
> 

Not sure if 0-day is running on linux-next nowadays. There are other
build failures in there which don't seem to be reported either.

Unfortunately, looking into build failures (and crashes) in linux-next
would be a full-time job. I see those failures, but I just don't have the
time to analyze and report them all.

Guenter