[PATCH v3 4/5] KVM: arm64: Mask out filtered events in PCMEID{0,1}_EL1

[Marc Zyngier]

Hi Eric,

On 2020-09-09 18:43, Auger Eric wrote:
> Hi Marc,
> 
> On 9/8/20 9:58 AM, Marc Zyngier wrote:
>> As we can now hide events from the guest, let's also adjust its view 
>> of
>> PCMEID{0,1}_EL1 so that it can figure out why some common events are 
>> not
>> counting as they should.
> Referring to my previous comment should we filter the cycle counter 
> out?
>> 
>> The astute user can still look into the TRM for their CPU and find out
>> they've been cheated, though. Nobody's perfect.
>> 
>> Signed-off-by: Marc Zyngier <maz at kernel.org>
>> ---
>>  arch/arm64/kvm/pmu-emul.c | 29 +++++++++++++++++++++++++++++
>>  arch/arm64/kvm/sys_regs.c |  5 +----
>>  include/kvm/arm_pmu.h     |  5 +++++
>>  3 files changed, 35 insertions(+), 4 deletions(-)
>> 
>> diff --git a/arch/arm64/kvm/pmu-emul.c b/arch/arm64/kvm/pmu-emul.c
>> index 67a731bafbc9..0458860bade2 100644
>> --- a/arch/arm64/kvm/pmu-emul.c
>> +++ b/arch/arm64/kvm/pmu-emul.c
>> @@ -765,6 +765,35 @@ static int kvm_pmu_probe_pmuver(void)
>>  	return pmuver;
>>  }
>> 
>> +u64 kvm_pmu_get_pmceid(struct kvm_vcpu *vcpu, bool pmceid1)
>> +{
>> +	unsigned long *bmap = vcpu->kvm->arch.pmu_filter;
>> +	u64 val, mask = 0;
>> +	int base, i;
>> +
>> +	if (!pmceid1) {
>> +		val = read_sysreg(pmceid0_el0);
>> +		base = 0;
>> +	} else {
>> +		val = read_sysreg(pmceid1_el0);
>> +		base = 32;
>> +	}
>> +
>> +	if (!bmap)
>> +		return val;
>> +
>> +	for (i = 0; i < 32; i += 8) {
> s/32/4?

I don't think so, see below.

> 
> Thanks
> 
> Eric
>> +		u64 byte;
>> +
>> +		byte = bitmap_get_value8(bmap, base + i);
>> +		mask |= byte << i;

For each iteration of the loop, we read a byte from the bitmap
(hence the += 8 above), and orr it into the mask. This makes 4
iteration of the loop.

Or am I missing your point entirely?

Thanks,

         M.
-- 
Jazz is not dead. It just smells funny...