[PATCH] powerpc: fix hardware PMU exception bug on PowerVM compatibility mode systems

[PATCH] powerpc: fix hardware PMU exception bug on PowerVM compatibility mode systems

[Desnes A. Nunes do Rosario]

PowerVM systems running compatibility mode on a few Power8 revisions are
still vulnerable to the hardware defect that loses PMU exceptions arriving
prior to a context switch.

The software fix for this issue is enabled through the CPU_FTR_PMAO_BUG
cpu_feature bit, nevertheless this bit also needs to be set for PowerVM
compatibility mode systems.

Fixes: 68f2f0d431d9ea4 ("powerpc: Add a cpu feature CPU_FTR_PMAO_BUG")
Signed-off-by: Desnes A. Nunes do Rosario <desnesn@linux.ibm.com>
---
 arch/powerpc/kernel/cputable.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/powerpc/kernel/cputable.c b/arch/powerpc/kernel/cputable.c
index e745abc5457a..5bfef6263dfb 100644
--- a/arch/powerpc/kernel/cputable.c
+++ b/arch/powerpc/kernel/cputable.c
@@ -2198,6 +2198,8 @@ static struct cpu_spec * __init setup_cpu_spec(unsigned long offset,
 		if (old.oprofile_cpu_type != NULL) {
 			t->oprofile_cpu_type = old.oprofile_cpu_type;
 			t->oprofile_type = old.oprofile_type;
+			if (old.cpu_features & CPU_FTR_PMAO_BUG)
+				t->cpu_features |= CPU_FTR_PMAO_BUG;
 		}
 	}
 
-- 
2.21.1

Re: [PATCH] powerpc: fix hardware PMU exception bug on PowerVM compatibility mode systems

[Leonardo Bras]

[-- Attachment #1: Type: text/plain, Size: 1509 bytes --]

Hello Desnes, thanks for the patch.

On Thu, 2020-02-13 at 21:19 -0300, Desnes A. Nunes do Rosario wrote:
> PowerVM systems running compatibility mode on a few Power8 revisions are
> still vulnerable to the hardware defect that loses PMU exceptions arriving
> prior to a context switch.
> 
> The software fix for this issue is enabled through the CPU_FTR_PMAO_BUG
> cpu_feature bit, nevertheless this bit also needs to be set for PowerVM
> compatibility mode systems.
> 
> Fixes: 68f2f0d431d9ea4 ("powerpc: Add a cpu feature CPU_FTR_PMAO_BUG")
> Signed-off-by: Desnes A. Nunes do Rosario <desnesn@linux.ibm.com>
> ---
>  arch/powerpc/kernel/cputable.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/arch/powerpc/kernel/cputable.c b/arch/powerpc/kernel/cputable.c
> index e745abc5457a..5bfef6263dfb 100644
> --- a/arch/powerpc/kernel/cputable.c
> +++ b/arch/powerpc/kernel/cputable.c
> @@ -2198,6 +2198,8 @@ static struct cpu_spec * __init setup_cpu_spec(unsigned long offset,
>  		if (old.oprofile_cpu_type != NULL) {
>  			t->oprofile_cpu_type = old.oprofile_cpu_type;
>  			t->oprofile_type = old.oprofile_type;
> +			if (old.cpu_features & CPU_FTR_PMAO_BUG)
> +				t->cpu_features |= CPU_FTR_PMAO_BUG;
>  		}
>  	}
>  

What are your thoughts about doing:
        t->cpu_features |=  old.cpu_features & CPU_FTR_PMAO_BUG;

Also, I would recommend adding a short comment on top of the added
lines explaining why it is needed. 

Best regards,
Leonardo Bras

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH] powerpc: fix hardware PMU exception bug on PowerVM compatibility mode systems

[Desnes Augusto Nunes do Rosario]

Hello Leonardo,

On 2/15/20 2:39 AM, Leonardo Bras wrote:
> Hello Desnes, thanks for the patch.
>
> "Desnes A. Nunes do Rosario" <desnesn@linux.ibm.com> writes:
>
>> PowerVM systems running compatibility mode on a few Power8 revisions are
>> still vulnerable to the hardware defect that loses PMU exceptions arriving
>> prior to a context switch.
>>
>> The software fix for this issue is enabled through the CPU_FTR_PMAO_BUG
>> cpu_feature bit, nevertheless this bit also needs to be set for PowerVM
>> compatibility mode systems.
>>
>> Fixes: 68f2f0d431d9ea4 ("powerpc: Add a cpu feature CPU_FTR_PMAO_BUG")
>> Signed-off-by: Desnes A. Nunes do Rosario <desnesn@linux.ibm.com>
>> ---
>>   arch/powerpc/kernel/cputable.c | 2 ++
>>   1 file changed, 2 insertions(+)
>>
>> diff --git a/arch/powerpc/kernel/cputable.c b/arch/powerpc/kernel/cputable.c
>> index e745abc5457a..5bfef6263dfb 100644
>> --- a/arch/powerpc/kernel/cputable.c
>> +++ b/arch/powerpc/kernel/cputable.c
>> @@ -2198,6 +2198,8 @@ static struct cpu_spec * __init setup_cpu_spec(unsigned long offset,
>>   		if (old.oprofile_cpu_type != NULL) {
>>   			t->oprofile_cpu_type = old.oprofile_cpu_type;
>>   			t->oprofile_type = old.oprofile_type;
>> +			if (old.cpu_features & CPU_FTR_PMAO_BUG)
>> +				t->cpu_features |= CPU_FTR_PMAO_BUG;
> What are your thoughts about doing:
> 	t->cpu_features |=  old.cpu_features & CPU_FTR_PMAO_BUG;
Looks good to me.
> Also, I would recommend adding a short comment on top of the added
> lines explaining why it is needed.
Sure, I'll send a second version.

Thanks for the review.
>
> Best regards,
> Leonardo Bras
>
-- 
Desnes A. Nunes do Rosario
--------------------------------
Advisory Software Engineer - IBM
Virtual Onsite Engineer - Red Hat

Re: [PATCH] powerpc: fix hardware PMU exception bug on PowerVM compatibility mode systems

[Leonardo Bras]

[-- Attachment #1: Type: text/plain, Size: 1475 bytes --]

Hello Desnes, thanks for the patch.

"Desnes A. Nunes do Rosario" <desnesn@linux.ibm.com> writes:

> PowerVM systems running compatibility mode on a few Power8 revisions are
> still vulnerable to the hardware defect that loses PMU exceptions arriving
> prior to a context switch.
>
> The software fix for this issue is enabled through the CPU_FTR_PMAO_BUG
> cpu_feature bit, nevertheless this bit also needs to be set for PowerVM
> compatibility mode systems.
>
> Fixes: 68f2f0d431d9ea4 ("powerpc: Add a cpu feature CPU_FTR_PMAO_BUG")
> Signed-off-by: Desnes A. Nunes do Rosario <desnesn@linux.ibm.com>
> ---
>  arch/powerpc/kernel/cputable.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/arch/powerpc/kernel/cputable.c b/arch/powerpc/kernel/cputable.c
> index e745abc5457a..5bfef6263dfb 100644
> --- a/arch/powerpc/kernel/cputable.c
> +++ b/arch/powerpc/kernel/cputable.c
> @@ -2198,6 +2198,8 @@ static struct cpu_spec * __init setup_cpu_spec(unsigned long offset,
>  		if (old.oprofile_cpu_type != NULL) {
>  			t->oprofile_cpu_type = old.oprofile_cpu_type;
>  			t->oprofile_type = old.oprofile_type;
> +			if (old.cpu_features & CPU_FTR_PMAO_BUG)
> +				t->cpu_features |= CPU_FTR_PMAO_BUG;

What are your thoughts about doing:
	t->cpu_features |=  old.cpu_features & CPU_FTR_PMAO_BUG;

Also, I would recommend adding a short comment on top of the added
lines explaining why it is needed. 

Best regards,
Leonardo Bras


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 833 bytes --]