Re: [PATCH v3 3/3] arm64/kvm: Fine grain _EL2 system registers list that affect nested virtualization

From: Marc Zyngier <maz@kernel.org>
To: eric.auger@redhat.com
Cc: Miguel Luis <miguel.luis@oracle.com>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	Oliver Upton <oliver.upton@linux.dev>,
	James Morse <james.morse@arm.com>,
	Suzuki K Poulose <suzuki.poulose@arm.com>,
	Zenghui Yu <yuzenghui@huawei.com>,
	Jing Zhang <jingzhangos@google.com>,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, kvmarm@lists.linux.dev
Subject: Re: [PATCH v3 3/3] arm64/kvm: Fine grain _EL2 system registers list that affect nested virtualization
Date: Thu, 12 Oct 2023 21:36:33 +0100	[thread overview]
Message-ID: <e6f3002c10848e911c4bfee3a1d472aa@kernel.org> (raw)
In-Reply-To: <7df53e6b-9141-de85-b7a3-b9eb092ef7b4@redhat.com>

On 2023-10-12 16:22, Eric Auger wrote:
> Hi Miguel,
> 
> On 10/11/23 20:01, Miguel Luis wrote:
>> Implement a fine grained approach in the _EL2 sysreg ranges.
>> 
>> Fixes: d0fc0a2519a6 ("KVM: arm64: nv: Add trap forwarding for 
>> HCR_EL2")
>> Signed-off-by: Miguel Luis <miguel.luis@oracle.com>
>> ---
>>  arch/arm64/kvm/emulate-nested.c | 88 
>> ++++++++++++++++++++++++++++++---
>>  1 file changed, 82 insertions(+), 6 deletions(-)
>> 
>> diff --git a/arch/arm64/kvm/emulate-nested.c 
>> b/arch/arm64/kvm/emulate-nested.c
>> index 9ced1bf0c2b7..3af49e130ee6 100644
>> --- a/arch/arm64/kvm/emulate-nested.c
>> +++ b/arch/arm64/kvm/emulate-nested.c
>> @@ -648,15 +648,91 @@ static const struct encoding_to_trap_config 
>> encoding_to_cgt[] __initconst = {
>>  	SR_TRAP(SYS_APGAKEYLO_EL1,	CGT_HCR_APK),
>>  	SR_TRAP(SYS_APGAKEYHI_EL1,	CGT_HCR_APK),
>>  	/* All _EL2 registers */
>> -	SR_RANGE_TRAP(sys_reg(3, 4, 0, 0, 0),
>> -		      sys_reg(3, 4, 3, 15, 7), CGT_HCR_NV),
>> +	SR_TRAP(SYS_VPIDR_EL2,		CGT_HCR_NV),
> I think you miss DBGVCR32_EL2

I don't think this register should be part of this list. We don't
support AArch32 with NV, and the spec says that DBGVCR32_EL2 UNDEFs
when EL1 doesn't support AArch32.

So the change that needs to happen is to inject an UNDEF when trapping
a DBGVCR32_EL2and not forward the trap to the guest.

Thanks,

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