Re: [PATCH 1/8] KVM: arm64: Move AArch32 exceptions over to AArch64 sysregs

[Marc Zyngier <maz@kernel.org>]

On 2020-11-03 18:29, James Morse wrote:
> Hi Marc,
> 
> On 02/11/2020 19:16, Marc Zyngier wrote:
>> The use of the AArch32-specific accessors have always been a bit
>> annoying on 64bit, and it is time for a change.
>> 
>> Let's move the AArch32 exception injection over to the AArch64 
>> encoding,
>> which requires us to split the two halves of FAR_EL1 into DFAR and 
>> IFAR.
>> This enables us to drop the preempt_disable() games on VHE, and to 
>> kill
>> the last user of the vcpu_cp15() macro.
> 
> Hurrah!
> 
> 
>> diff --git a/arch/arm64/kvm/inject_fault.c 
>> b/arch/arm64/kvm/inject_fault.c
>> index e2a2e48ca371..975f65ba6a8b 100644
>> --- a/arch/arm64/kvm/inject_fault.c
>> +++ b/arch/arm64/kvm/inject_fault.c
>> @@ -100,39 +81,36 @@ static void inject_undef32(struct kvm_vcpu *vcpu)
>>   * Modelled after TakeDataAbortException() and 
>> TakePrefetchAbortException
>>   * pseudocode.
>>   */
>> -static void inject_abt32(struct kvm_vcpu *vcpu, bool is_pabt,
>> -			 unsigned long addr)
>> +static void inject_abt32(struct kvm_vcpu *vcpu, bool is_pabt, u32 
>> addr)
>>  {
>> -	u32 *far, *fsr;
>> -	bool is_lpae;
>> -	bool loaded;
>> +	u64 far;
>> +	u32 fsr;
> 
> 
>> +	/* Give the guest an IMPLEMENTATION DEFINED exception */
>> +	if (__vcpu_sys_reg(vcpu, TCR_EL1) & TTBCR_EAE) {
> 
> With VHE, won't __vcpu_sys_reg() read the potentially stale copy in
> the sys_reg array?
> 
> vcpu_read_sys_reg()?

Of course you are right. Now fixed.

Thanks,

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