RE: mmotm 2022-03-30-13-01 uploaded (drivers/platform/x86/amd-pmc.o)

[Limonciello, Mario]

[Public]



> -----Original Message-----
> From: Hans de Goede <hdegoede@xxxxxxxxxx>
> Sent: Friday, April 1, 2022 10:01
> To: Limonciello, Mario <Mario.Limonciello@xxxxxxx>; Randy Dunlap
> <rdunlap@xxxxxxxxxxxxx>; Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>;
> broonie@xxxxxxxxxx; mhocko@xxxxxxx; sfr@xxxxxxxxxxxxxxxx; linux-
> next@xxxxxxxxxxxxxxx; linux-fsdevel@xxxxxxxxxxxxxxx; linux-mm@xxxxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx; mm-commits@xxxxxxxxxxxxxxx; platform-
> driver-x86@xxxxxxxxxxxxxxx
> Subject: Re: mmotm 2022-03-30-13-01 uploaded (drivers/platform/x86/amd-
> pmc.o)
> 
> Hi Mario,
> 
> On 3/31/22 20:41, Limonciello, Mario wrote:
> > [AMD Official Use Only]
> >
> >> -----Original Message-----
> >> From: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
> >> Sent: Thursday, March 31, 2022 10:51
> >> To: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>; broonie@xxxxxxxxxx;
> >> mhocko@xxxxxxx; sfr@xxxxxxxxxxxxxxxx; linux-next@xxxxxxxxxxxxxxx;
> linux-
> >> fsdevel@xxxxxxxxxxxxxxx; linux-mm@xxxxxxxxx; linux-
> >> kernel@xxxxxxxxxxxxxxx; mm-commits@xxxxxxxxxxxxxxx; platform-driver-
> >> x86@xxxxxxxxxxxxxxx; Limonciello, Mario <Mario.Limonciello@xxxxxxx>
> >> Subject: Re: mmotm 2022-03-30-13-01 uploaded
> (drivers/platform/x86/amd-
> >> pmc.o)
> >>
> >>
> >>
> >> On 3/30/22 13:01, Andrew Morton wrote:
> >>> The mm-of-the-moment snapshot 2022-03-30-13-01 has been uploaded
> to
> >>>
> >>>
> >>
> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fww
> >>
> w.ozlabs.org%2F~akpm%2Fmmotm%2F&amp;data=04%7C01%7Cmario.limo
> >>
> nciello%40amd.com%7Caa3aae02b7b6437c46ea08da132e4222%7C3dd8961fe
> >>
> 4884e608e11a82d994e183d%7C0%7C0%7C637843386674652995%7CUnknown
> >>
> %7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1ha
> >>
> WwiLCJXVCI6Mn0%3D%7C3000&amp;sdata=WpfXOyTRBgvqCj3iZ%2BJjXVTb
> >> V%2FUWDP4ds5XtDfa5bPc%3D&amp;reserved=0
> >>>
> >>> mmotm-readme.txt says
> >>>
> >>> README for mm-of-the-moment:
> >>>
> >>>
> >>
> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fww
> >>
> w.ozlabs.org%2F~akpm%2Fmmotm%2F&amp;data=04%7C01%7Cmario.limo
> >>
> nciello%40amd.com%7Caa3aae02b7b6437c46ea08da132e4222%7C3dd8961fe
> >>
> 4884e608e11a82d994e183d%7C0%7C0%7C637843386674652995%7CUnknown
> >>
> %7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1ha
> >>
> WwiLCJXVCI6Mn0%3D%7C3000&amp;sdata=WpfXOyTRBgvqCj3iZ%2BJjXVTb
> >> V%2FUWDP4ds5XtDfa5bPc%3D&amp;reserved=0
> >>>
> >>> This is a snapshot of my -mm patch queue.  Uploaded at random
> hopefully
> >>> more than once a week.
> >>>
> >>> You will need quilt to apply these patches to the latest Linus release (5.x
> >>> or 5.x-rcY).  The series file is in broken-out.tar.gz and is duplicated in
> >>>
> >>
> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fozla
> >>
> bs.org%2F~akpm%2Fmmotm%2Fseries&amp;data=04%7C01%7Cmario.limon
> >>
> ciello%40amd.com%7Caa3aae02b7b6437c46ea08da132e4222%7C3dd8961fe4
> >>
> 884e608e11a82d994e183d%7C0%7C0%7C637843386674652995%7CUnknown
> >>
> %7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1ha
> >>
> WwiLCJXVCI6Mn0%3D%7C3000&amp;sdata=adnYHchdMvieQ0lCqQ1er1jhHA
> >> UMomgFpaWBo%2BMnINk%3D&amp;reserved=0
> >>>
> >>
> >> on x86_64:
> >> when CONFIG_SUSPEND is not set:
> >>
> >> drivers/platform/x86/amd-pmc.o: in function `amd_pmc_remove':
> >> amd-pmc.c:(.text+0x11d): undefined reference to
> >> `acpi_unregister_lps0_dev'
> >> ld: drivers/platform/x86/amd-pmc.o: in function `amd_pmc_probe':
> >> amd-pmc.c:(.text+0x20be): undefined reference to
> `acpi_register_lps0_dev'
> >>
> >>
> >>
> >> --
> >> ~Randy
> >
> > AFAICT you're missing 20e1d6402a71dba7ad2b81f332a3c14c7d3b939b.
> 
> That is unlikely since the whole series got merged through
> the pdx86/for-next branch, so either some other tree
> has the whole series or none of it.
> 
> Also note that Randy wrote:
> 
> "when CONFIG_SUSPEND is not set"
> 
> The problem is that all of drivers/acpi/x86/s2idle.c is
> wrapped by #ifdef CONFIG_SUSPEND ... #endif
> 

Ah thanks.

> Since AFAIK the whole amd-pmc driver's whole purpose
> is to allow s2idle to work properly. I think this can
> be fixed by simply doing:
> 
> --- a/drivers/platform/x86/Kconfig
> +++ b/drivers/platform/x86/Kconfig
> @@ -198,7 +198,7 @@ config ACER_WMI
> 
>  config AMD_PMC
>  	tristate "AMD SoC PMC driver"
> -	depends on ACPI && PCI && RTC_CLASS
> +	depends on ACPI && SUSPEND && PCI && RTC_CLASS
>  	help
>  	  The driver provides support for AMD Power Management
> Controller
>  	  primarily responsible for S2Idle transactions that are driven from
> 
> 
> If you agree, please submit a patch with this change and I'll queue it
> up in my fixes branch.
> 
> Regards,
> 
> Hans

That's the primary reason, but you can technically also fetch an STB and
read idle mask from debugfs even without debugfs set.  I will weigh out
both options and send something to fix this then.