Re: [PATCH v3 3/7] acpi/ghes: Support GPIO error source.

[Jonathan Cameron via <qemu-devel@nongnu.org>]

On Thu, 1 Aug 2024 14:56:37 +0200
Mauro Carvalho Chehab <mchehab+huawei@kernel.org> wrote:

> Em Tue, 30 Jul 2024 10:40:28 +0200
> Igor Mammedov <imammedo@redhat.com> escreveu:
> 
> > > diff --git a/include/hw/acpi/ghes.h b/include/hw/acpi/ghes.h
> > > index 674f6958e905..4f1ab1a73a06 100644
> > > --- a/include/hw/acpi/ghes.h
> > > +++ b/include/hw/acpi/ghes.h
> > > @@ -58,6 +58,7 @@ enum AcpiGhesNotifyType {
> > >  
> > >  enum {
> > >      ACPI_HEST_SRC_ID_SEA = 0,
> > > +    ACPI_HEST_SRC_ID_GPIO = 1,    
> > is it defined by some spec, or just a made up number?  
> 
> I don't know. Maybe Jonathan or Shiju knows better, as the original patch
> came from them, but I didn't find any parts of the ACPI spec defining the
> values for source ID.

> 
> Checking at build_ghes_v2() implementation, this is used on two places:
> 
> 1. as GHESv2 source ID:
>     /*
>      * Type:
>      * Generic Hardware Error Source version 2(GHESv2 - Type 10)
>      */
>     build_append_int_noprefix(table_data, ACPI_GHES_SOURCE_GENERIC_ERROR_V2, 2);
>     /* Source Id */
>     build_append_int_noprefix(table_data, source_id, 2);
>     /* Related Source Id */
>     build_append_int_noprefix(table_data, 0xffff, 2);
> 
> as an address offset:
> 
>     address_offset = table_data->len;
>     /* Error Status Address */
>     build_append_gas(table_data, AML_AS_SYSTEM_MEMORY, 0x40, 0,
>                      4 /* QWord access */, 0);
>     bios_linker_loader_add_pointer(linker, ACPI_BUILD_TABLE_FILE,
>         address_offset + GAS_ADDR_OFFSET, sizeof(uint64_t),
>         ACPI_GHES_ERRORS_FW_CFG_FILE, source_id * sizeof(uint64_t));
> 
> So, if I had to guess, I'd say that this was made up, in a way that
> the size of the table will fit just two sources, starting from zero.
> 
> So, I'll change the code to just:
> 
> 	enum {
>             ACPI_HEST_SRC_ID_SEA = 0,
>             ACPI_HEST_SRC_ID_GPIO, 

LGTM.  The naming is perhaps not ideal but the scheme predates my
involvement so I'm not sure of the reasoning.  Could change it
to QEMU_ACPI...
to make it really really clear these aren't an ACPI spec thing, but
may not be worth it.

J

> 	    /* future ids go here */
> 	    ACPI_HEST_SRC_ID_RESERVED,
> 	};
> 
> To remove the false impression that this could be originated from the
> spec.
> 
> Thanks,
> Mauro
>