RE: [RESEND PATCH v3 5/9] EDAC: Don't load chipset-specific edac drivers when ghes_edac is preferred

[Justin He]

> -----Original Message-----
> From: Kani, Toshi <toshi.kani@xxxxxxx>
> Sent: Thursday, August 25, 2022 7:04 AM
> To: Justin He <Justin.He@xxxxxxx>; Len Brown <lenb@xxxxxxxxxx>; James
> Morse <James.Morse@xxxxxxx>; Tony Luck <tony.luck@xxxxxxxxx>; Borislav
> Petkov <bp@xxxxxxxxx>; Mauro Carvalho Chehab <mchehab@xxxxxxxxxx>;
> Robert Richter <rric@xxxxxxxxxx>; Robert Moore <robert.moore@xxxxxxxxx>;
> Qiuxu Zhuo <qiuxu.zhuo@xxxxxxxxx>; Yazen Ghannam
> <yazen.ghannam@xxxxxxx>; Jonathan Corbet <corbet@xxxxxxx>; Jan
> Luebbe <jlu@xxxxxxxxxxxxxx>; Khuong Dinh
> <khuong@xxxxxxxxxxxxxxxxxxxxxx>
> Cc: Ard Biesheuvel <ardb@xxxxxxxxxx>; linux-acpi@xxxxxxxxxxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx; linux-edac@xxxxxxxxxxxxxxx; devel@xxxxxxxxxx;
> Rafael J . Wysocki <rafael@xxxxxxxxxx>; Shuai Xue
> <xueshuai@xxxxxxxxxxxxxxxxx>; Jarkko Sakkinen <jarkko@xxxxxxxxxx>;
> linux-efi@xxxxxxxxxxxxxxx; nd <nd@xxxxxxx>; Paul E. McKenney
> <paulmck@xxxxxxxxxx>; Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>;
> Neeraj Upadhyay <quic_neeraju@xxxxxxxxxxx>; Randy Dunlap
> <rdunlap@xxxxxxxxxxxxx>; Damien Le Moal
> <damien.lemoal@xxxxxxxxxxxxxxxxxx>; Muchun Song
> <songmuchun@xxxxxxxxxxxxx>; linux-doc@xxxxxxxxxxxxxxx
> Subject: RE: [RESEND PATCH v3 5/9] EDAC: Don't load chipset-specific edac
> drivers when ghes_edac is preferred
> 
> On Monday, August 22, 2022 9:41 AM, Jia He wrote:
> > diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c index
> > e17e0ee8f842..327386f3cf33 100644
> > --- a/drivers/acpi/apei/ghes.c
> > +++ b/drivers/acpi/apei/ghes.c
> > @@ -1537,16 +1537,25 @@ static struct acpi_platform_list plat_list[] = {
> >  	{ } /* End */
> >  };
> >
> > -struct list_head *ghes_get_devices(void)
> > +bool ghes_edac_preferred(void)
> >  {
> >  	int idx = -1;
> >
> >  	if (IS_ENABLED(CONFIG_X86)) {
> >  		idx = acpi_match_platform_list(plat_list);
> >  		if (idx < 0 && !ghes_edac_force)
> > -			return NULL;
> > +			return false;
> >  	}
> >
> > +	return true;
> > +}
> > +EXPORT_SYMBOL_GPL(ghes_edac_preferred);
> > +
> > +struct list_head *ghes_get_devices(void) {
> > +	if (!ghes_edac_preferred())
> > +		return NULL;
> > +
> >  	return &ghes_devs;
> >  }
> 
> ghes_get_devices() changing multiple times in the series is
> confusing to me.   Can you simply introduce ghes_get_devices()
> and ghes_preferred() in the right state in a patch?  Perhaps, patch #2, #5, #6
> can collapse to introduce the two funcs?
> 

My purpose was to make it easy for review. I am ok to merge these patches into one.

> The rest of patch #5 adding the call to ghes_edac_preferred() into other edac
> drivers can remain as a separate patch.

Okay, I assume you mean to merge all of the ghes_edac_preferred() checking for intel and
Arm edac drivers into 1 separate patch, am I understanding well?


--
Cheers,
Justin (Jia He)