Re: [Qemu-devel] [kvm-unit-tests RFC 05/15] arm/arm64: GICv3: add cpu count

[Auger Eric]

Hi Andre, Drew,

On 06/12/2016 10:32, Andre Przywara wrote:
> Hi,
> 
> On 06/12/16 09:29, Andrew Jones wrote:
>> On Mon, Dec 05, 2016 at 10:46:36PM +0100, Eric Auger wrote:
>>> Add a new cpu_count field in gicv3_data indicating the
>>> number of redistributors. This will be useful for enumeration
>>> of their resources such as LPI pending tables.
>>
>> I'm fine with the additional state, but just curious, will it
>> ever be possible for gicv3.cpu_count != nr_cpus?
> 
> If not you are in trouble, so that should in fact be one test.
> 
> Which brings me to my comment ...
> 
>>>
>>> Signed-off-by: Eric Auger 
>>> ---
>>>  lib/arm/asm/gic-v3.h | 1 +
>>>  lib/arm/gic-v3.c | 2 ++
>>>  2 files changed, 3 insertions(+)
>>>
>>> diff --git a/lib/arm/asm/gic-v3.h b/lib/arm/asm/gic-v3.h
>>> index ed330af..039b7c2 100644
>>> --- a/lib/arm/asm/gic-v3.h
>>> +++ b/lib/arm/asm/gic-v3.h
>>> @@ -58,6 +58,7 @@ struct gicv3_data {
>>> void *dist_base;
>>> void *redist_base[NR_CPUS];
>>> unsigned int irq_nr;
>>> +   unsigned int cpu_count;
> 
> Should that be called "nr_redists" or the like?
> Since this is what it counts in the code below.
> Later we can then compare this with nr_cpus to check for a match.

I fully agree with you suggestion.

Thanks

Eric
> 
> Cheers,
> Andre.
> 
>>>  };
>>>  extern struct gicv3_data gicv3_data;
>>>  
>>> diff --git a/lib/arm/gic-v3.c b/lib/arm/gic-v3.c
>>> index 6246221..9921f4d 100644
>>> --- a/lib/arm/gic-v3.c
>>> +++ b/lib/arm/gic-v3.c
>>> @@ -12,12 +12,14 @@ void gicv3_set_redist_base(size_t stride)
>>> void *ptr = gicv3_data.redist_base[0];
>>> u64 typer;
>>>  
>>> +   gicv3_data.cpu_count = 0;
>>> do {
>>> typer = gicv3_read_typer(ptr + GICR_TYPER);
>>> if ((typer >> 32) == aff) {
>>> gicv3_redist_base() = ptr;
>>> return;
>>> }
>>> +   gicv3_data.cpu_count++;
>>> ptr += stride; /* skip RD_base, SGI_base, etc. */
>>> } while (!(typer & GICR_TYPER_LAST));
>>>  
>>> -- 
>>> 2.5.5
>>>
>>>
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majord...@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
___
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

Re: [Qemu-devel] [kvm-unit-tests RFC 05/15] arm/arm64: GICv3: add cpu count

[Andre Przywara]

Hi,

On 06/12/16 09:29, Andrew Jones wrote:
> On Mon, Dec 05, 2016 at 10:46:36PM +0100, Eric Auger wrote:
>> Add a new cpu_count field in gicv3_data indicating the
>> number of redistributors. This will be useful for enumeration
>> of their resources such as LPI pending tables.
> 
> I'm fine with the additional state, but just curious, will it
> ever be possible for gicv3.cpu_count != nr_cpus?

If not you are in trouble, so that should in fact be one test.

Which brings me to my comment ...

>>
>> Signed-off-by: Eric Auger 
>> ---
>>  lib/arm/asm/gic-v3.h | 1 +
>>  lib/arm/gic-v3.c | 2 ++
>>  2 files changed, 3 insertions(+)
>>
>> diff --git a/lib/arm/asm/gic-v3.h b/lib/arm/asm/gic-v3.h
>> index ed330af..039b7c2 100644
>> --- a/lib/arm/asm/gic-v3.h
>> +++ b/lib/arm/asm/gic-v3.h
>> @@ -58,6 +58,7 @@ struct gicv3_data {
>>  void *dist_base;
>>  void *redist_base[NR_CPUS];
>>  unsigned int irq_nr;
>> +unsigned int cpu_count;

Should that be called "nr_redists" or the like?
Since this is what it counts in the code below.
Later we can then compare this with nr_cpus to check for a match.

Cheers,
Andre.

>>  };
>>  extern struct gicv3_data gicv3_data;
>>  
>> diff --git a/lib/arm/gic-v3.c b/lib/arm/gic-v3.c
>> index 6246221..9921f4d 100644
>> --- a/lib/arm/gic-v3.c
>> +++ b/lib/arm/gic-v3.c
>> @@ -12,12 +12,14 @@ void gicv3_set_redist_base(size_t stride)
>>  void *ptr = gicv3_data.redist_base[0];
>>  u64 typer;
>>  
>> +gicv3_data.cpu_count = 0;
>>  do {
>>  typer = gicv3_read_typer(ptr + GICR_TYPER);
>>  if ((typer >> 32) == aff) {
>>  gicv3_redist_base() = ptr;
>>  return;
>>  }
>> +gicv3_data.cpu_count++;
>>  ptr += stride; /* skip RD_base, SGI_base, etc. */
>>  } while (!(typer & GICR_TYPER_LAST));
>>  
>> -- 
>> 2.5.5
>>
>>
___
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

Re: [Qemu-devel] [kvm-unit-tests RFC 05/15] arm/arm64: GICv3: add cpu count

[Andrew Jones]

On Mon, Dec 05, 2016 at 10:46:36PM +0100, Eric Auger wrote:
> Add a new cpu_count field in gicv3_data indicating the
> number of redistributors. This will be useful for enumeration
> of their resources such as LPI pending tables.

I'm fine with the additional state, but just curious, will it
ever be possible for gicv3.cpu_count != nr_cpus?

> 
> Signed-off-by: Eric Auger 
> ---
>  lib/arm/asm/gic-v3.h | 1 +
>  lib/arm/gic-v3.c | 2 ++
>  2 files changed, 3 insertions(+)
> 
> diff --git a/lib/arm/asm/gic-v3.h b/lib/arm/asm/gic-v3.h
> index ed330af..039b7c2 100644
> --- a/lib/arm/asm/gic-v3.h
> +++ b/lib/arm/asm/gic-v3.h
> @@ -58,6 +58,7 @@ struct gicv3_data {
>   void *dist_base;
>   void *redist_base[NR_CPUS];
>   unsigned int irq_nr;
> + unsigned int cpu_count;
>  };
>  extern struct gicv3_data gicv3_data;
>  
> diff --git a/lib/arm/gic-v3.c b/lib/arm/gic-v3.c
> index 6246221..9921f4d 100644
> --- a/lib/arm/gic-v3.c
> +++ b/lib/arm/gic-v3.c
> @@ -12,12 +12,14 @@ void gicv3_set_redist_base(size_t stride)
>   void *ptr = gicv3_data.redist_base[0];
>   u64 typer;
>  
> + gicv3_data.cpu_count = 0;
>   do {
>   typer = gicv3_read_typer(ptr + GICR_TYPER);
>   if ((typer >> 32) == aff) {
>   gicv3_redist_base() = ptr;
>   return;
>   }
> + gicv3_data.cpu_count++;
>   ptr += stride; /* skip RD_base, SGI_base, etc. */
>   } while (!(typer & GICR_TYPER_LAST));
>  
> -- 
> 2.5.5
> 
> 
___
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm