Re: [PATCH v4 6/8] KVM: PPC: Ultravisor: Restrict LDBAR access

[Claudio Carvalho <cclaudio@linux.ibm.com>]

On 7/11/19 9:57 AM, Michael Ellerman wrote:
> Claudio Carvalho <cclaudio@linux.ibm.com> writes:
>> When the ultravisor firmware is available, it takes control over the
>> LDBAR register. In this case, thread-imc updates and save/restore
>> operations on the LDBAR register are handled by ultravisor.
> Please roll up the replies to Alexey's question about LDBAR into the
> change log.
>
>> diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
>> index f9b2620fbecd..cffb365d9d02 100644
>> --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
>> +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
>> @@ -375,8 +375,10 @@ BEGIN_FTR_SECTION
>>  	mtspr	SPRN_RPR, r0
>>  	ld	r0, KVM_SPLIT_PMMAR(r6)
>>  	mtspr	SPRN_PMMAR, r0
>> +BEGIN_FW_FTR_SECTION_NESTED(70)
>>  	ld	r0, KVM_SPLIT_LDBAR(r6)
>>  	mtspr	SPRN_LDBAR, r0
>> +END_FW_FTR_SECTION_NESTED(FW_FEATURE_ULTRAVISOR, 0, 70)
> That's in Power8 code isn't it? Which will never have an ultravisor.

IIUC, it might be executed in Power9 as well, but I can double check that.


>
>> diff --git a/arch/powerpc/platforms/powernv/opal-imc.c b/arch/powerpc/platforms/powernv/opal-imc.c
>> index 1b6932890a73..5fe2d4526cbc 100644
>> --- a/arch/powerpc/platforms/powernv/opal-imc.c
>> +++ b/arch/powerpc/platforms/powernv/opal-imc.c
>> @@ -254,6 +254,10 @@ static int opal_imc_counters_probe(struct platform_device *pdev)
>>  	bool core_imc_reg = false, thread_imc_reg = false;
>>  	u32 type;
>>  
>> +	/* Disable IMC devices, when Ultravisor is enabled. */
>> +	if (firmware_has_feature(FW_FEATURE_ULTRAVISOR))
>> +		return -EACCES;
> I don't mind taking this change. But at the same time should the IMC
> stuff just be omitted from the device tree when we're in ultravisor mode?

Yes. Maddy said that he will patch skiboot to remove the IMC nodes if
ultravisor is present.

I added this check just to protect the kernel in case skiboot is not in the
right level for some
reason.

Thanks,
Claudio



>
> cheers
>