linux-kernel - Re: [PATCH v3 2/6] KVM: arm64: maintain per VM value for CTR

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Wed, 29 May 2024 19:35:33 +0200
From: Eric Auger <eauger@...hat.com>
To: Sebastian Ott <sebott@...hat.com>
Cc: linux-arm-kernel@...ts.infradead.org, kvmarm@...ts.linux.dev,
 linux-kernel@...r.kernel.org, Marc Zyngier <maz@...nel.org>,
 Oliver Upton <oliver.upton@...ux.dev>, James Morse <james.morse@....com>,
 Suzuki K Poulose <suzuki.poulose@....com>,
 Catalin Marinas <catalin.marinas@....com>, Will Deacon <will@...nel.org>
Subject: Re: [PATCH v3 2/6] KVM: arm64: maintain per VM value for CTR_EL0



On 5/29/24 17:51, Sebastian Ott wrote:
> Hej Eric,
> 
> On Wed, 29 May 2024, Eric Auger wrote:
>>>  static int set_clidr(struct kvm_vcpu *vcpu, const struct
>>> sys_reg_desc *rd,
>>>                u64 val)
>>>  {
>>> -    u64 ctr_el0 = read_sanitised_ftr_reg(SYS_CTR_EL0);
>>>      u64 idc = !CLIDR_LOC(val) || (!CLIDR_LOUIS(val) &&
>>> !CLIDR_LOUU(val));
>>> +    u64 ctr_el0 = vcpu->kvm->arch.ctr_el0;
>>>
>>>      if ((val & CLIDR_EL1_RES0) || (!(ctr_el0 & CTR_EL0_IDC) && idc))
>>>          return -EINVAL;
>> nit: you may update the function doc comment which the extra handling of
>> CTR_EL0.
> 
> Hm, there's no extra handling of CTR_EL0 it just uses the emulated value.>
>>> @@ -3557,6 +3557,13 @@ void kvm_reset_sys_regs(struct kvm_vcpu *vcpu)
>>>      struct kvm *kvm = vcpu->kvm;
>>>      unsigned long i;
>>>
>>> +    if (!kvm_vcpu_initialized(vcpu))
>> at this stage of the reading, why is the above check needed?
> 
> To make sure that a later call to this function doesn't overwrite
> the value provided by userspace. (See e016333745c "KVM: arm64: Only
> reset vCPU-scoped feature ID regs once").
but isn't it overwritten through the .reset=reset_ctr() that is
populated in next patch?

Eric
> 
> Sebastian
>