[PATCH 4/8] KVM: arm64: Map AArch32 cp14 register to AArch64 sysregs

[Marc Zyngier]

On 2020-11-03 18:29, James Morse wrote:
> Hi Marc,
> 
> On 02/11/2020 19:16, Marc Zyngier wrote:
>> Similarly to what has been done on the cp15 front, repaint the
>> debug registers to use their AArch64 counterparts. This results
>> in some simplification as we can remove the 32bit-specific
>> accessors.
> 
>> diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
>> index 137818793a4a..c41e7ca60c8c 100644
>> --- a/arch/arm64/kvm/sys_regs.c
>> +++ b/arch/arm64/kvm/sys_regs.c
>> @@ -361,26 +361,30 @@ static bool trap_debug_regs(struct kvm_vcpu 
>> *vcpu,
>> -#define DBGBXVR(n)							\
>> -	{ Op1( 0), CRn( 1), CRm((n)), Op2( 1), trap_xvr, NULL, n }
>> +#define DBG_BCR_BVR_WCR_WVR(n)						      \
>> +	/* DBGBVRn */							      \
>> +	{ AA32(LO), Op1( 0), CRn( 0), CRm((n)), Op2( 4), trap_bvr, NULL, n 
>> }, \
> 
> Just to check I understand what is going on here: This BVR AA32(LO) is
> needed because the
> dbg_bvr array is shared with the DBGBXVR registers...
> 
> 
>> +	/* DBGBCRn */							      \
>> +	{ AA32(LO), Op1( 0), CRn( 0), CRm((n)), Op2( 5), trap_bcr, NULL, n 
>> }, \
>> +	/* DBGWVRn */							      \
>> +	{ AA32(LO), Op1( 0), CRn( 0), CRm((n)), Op2( 6), trap_wvr, NULL, n 
>> }, \
>> +	/* DBGWCRn */							      \
>> +	{ AA32(LO), Op1( 0), CRn( 0), CRm((n)), Op2( 7), trap_wcr, NULL, n }
> 
> ... these don't have an alias, but its harmless.

This is a bug-for-bug translation of the original code. I guess I'll 
drop
that altogether.

> 
> [...]
> 
>> @@ -1931,7 +1896,9 @@ static const struct sys_reg_desc cp15_regs[] = {
>>  	/* DFSR */
>>  	{ Op1( 0), CRn( 5), CRm( 0), Op2( 0), access_vm_reg, NULL, ESR_EL1 
>> },
>>  	{ Op1( 0), CRn( 5), CRm( 0), Op2( 1), access_vm_reg, NULL, 
>> IFSR32_EL2 },
>> +	/* ADFSR */
>>  	{ Op1( 0), CRn( 5), CRm( 1), Op2( 0), access_vm_reg, NULL, AFSR0_EL1 
>> },
>> +	/* AIFSR */
>>  	{ Op1( 0), CRn( 5), CRm( 1), Op2( 1), access_vm_reg, NULL, AFSR1_EL1 
>> },
>>  	/* DFAR */
>>  	{ AA32(LO), Op1( 0), CRn( 6), CRm( 0), Op2( 0), access_vm_reg, NULL, 
>> FAR_EL1 },
> 
> I guess these were meant for the previous patch.

Yup, I'll move that.

Thanks,

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