linux-kernel - Re: [PATCH v2] x86/sev: Add support for allowing zero SEV ASIDs.

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] [day] [month] [year] [list]
Message-ID: <76b1660e-4b14-4485-9511-8a96b0f666f7@amd.com>
Date: Wed, 3 Jan 2024 23:35:12 -0600
From: "Kalra, Ashish" <ashish.kalra@....com>
To: Tom Lendacky <thomas.lendacky@....com>, pbonzini@...hat.com
Cc: seanjc@...gle.com, tglx@...utronix.de, mingo@...hat.com, bp@...en8.de,
 dave.hansen@...ux.intel.com, x86@...nel.org, hpa@...or.com,
 kvm@...r.kernel.org, linux-kernel@...r.kernel.org, joro@...tes.org
Subject: Re: [PATCH v2] x86/sev: Add support for allowing zero SEV ASIDs.

On 1/3/2024 11:28 PM, Tom Lendacky wrote:

> On 1/3/24 20:46, Ashish Kalra wrote:
>> From: Ashish Kalra <ashish.kalra@....com>
>>
>> Some BIOSes allow the end user to set the minimum SEV ASID value
>> (CPUID 0x8000001F_EDX) to be greater than the maximum number of
>> encrypted guests, or maximum SEV ASID value (CPUID 0x8000001F_ECX)
>> in order to dedicate all the SEV ASIDs to SEV-ES or SEV-SNP.
>>
>> The SEV support, as coded, does not handle the case where the minimum
>> SEV ASID value can be greater than the maximum SEV ASID value.
>> As a result, the following confusing message is issued:
>>
>> [   30.715724] kvm_amd: SEV enabled (ASIDs 1007 - 1006)
>>
>> Fix the support to properly handle this case.
>>
>> Fixes: 916391a2d1dc ("KVM: SVM: Add support for SEV-ES capability in 
>> KVM")
>> Suggested-by: Sean Christopherson <seanjc@...gle.com>
>> Signed-off-by: Ashish Kalra <ashish.kalra@....com>
>> Cc: stable@...r.kernel.org
>> ---
>>   arch/x86/kvm/svm/sev.c | 41 +++++++++++++++++++++++++----------------
>>   1 file changed, 25 insertions(+), 16 deletions(-)
>>
>> diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
>> index 4900c078045a..651d671ff8ae 100644
>> --- a/arch/x86/kvm/svm/sev.c
>> +++ b/arch/x86/kvm/svm/sev.c
>> @@ -143,8 +143,21 @@ static void sev_misc_cg_uncharge(struct 
>> kvm_sev_info *sev)
>>     static int sev_asid_new(struct kvm_sev_info *sev)
>>   {
>> -    int asid, min_asid, max_asid, ret;
>> +    /*
>> +     * SEV-enabled guests must use asid from min_sev_asid to 
>> max_sev_asid.
>> +     * SEV-ES-enabled guest can use from 1 to min_sev_asid - 1.
>> +     * Note: min ASID can end up larger than the max if basic SEV 
>> support is
>> +     * effectively disabled by disallowing use of ASIDs for SEV guests.
>> +     */
>> +    unsigned int min_asid = sev->es_active ? 1 : min_sev_asid;
>> +    unsigned int max_asid = sev->es_active ? min_sev_asid - 1 : 
>> max_sev_asid;
>> +    unsigned int asid;
>> +
>
> Remove this blank line.
>
>>       bool retry = true;
>> +    int ret;
>> +
>> +    if (min_asid > max_asid)
>> +        return -ENOTTY;
>>         WARN_ON(sev->misc_cg);
>>       sev->misc_cg = get_current_misc_cg();
>> @@ -157,12 +170,6 @@ static int sev_asid_new(struct kvm_sev_info *sev)
>>         mutex_lock(&sev_bitmap_lock);
>>   -    /*
>> -     * SEV-enabled guests must use asid from min_sev_asid to 
>> max_sev_asid.
>> -     * SEV-ES-enabled guest can use from 1 to min_sev_asid - 1.
>> -     */
>> -    min_asid = sev->es_active ? 1 : min_sev_asid;
>> -    max_asid = sev->es_active ? min_sev_asid - 1 : max_sev_asid;
>>   again:
>>       asid = find_next_zero_bit(sev_asid_bitmap, max_asid + 1, 
>> min_asid);
>>       if (asid > max_asid) {
>> @@ -246,21 +253,20 @@ static void sev_unbind_asid(struct kvm *kvm, 
>> unsigned int handle)
>>   static int sev_guest_init(struct kvm *kvm, struct kvm_sev_cmd *argp)
>>   {
>>       struct kvm_sev_info *sev = &to_kvm_svm(kvm)->sev_info;
>> -    int asid, ret;
>> +    int ret;
>>         if (kvm->created_vcpus)
>>           return -EINVAL;
>>   -    ret = -EBUSY;
>>       if (unlikely(sev->active))
>> -        return ret;
>> +        return -EINVAL;
>>         sev->active = true;
>>       sev->es_active = argp->id == KVM_SEV_ES_INIT;
>> -    asid = sev_asid_new(sev);
>> -    if (asid < 0)
>> +    ret = sev_asid_new(sev);
>> +    if (ret < 0)
>>           goto e_no_asid;
>> -    sev->asid = asid;
>> +    sev->asid = ret;
>>         ret = sev_platform_init(&argp->error);
>>       if (ret)
>> @@ -2229,8 +2235,10 @@ void __init sev_hardware_setup(void)
>>           goto out;
>>       }
>>   -    sev_asid_count = max_sev_asid - min_sev_asid + 1;
>> -    WARN_ON_ONCE(misc_cg_set_capacity(MISC_CG_RES_SEV, 
>> sev_asid_count));
>> +    if (min_sev_asid > max_sev_asid) {
>
> Shouldn't this be: if (min_sev_asid <= max_sev_asid) ?
>
> You only want to do the misc_cg_set_capactity() call if you can have 
> SEV guests.
>
>
Yes, it should be, will fix it and post v3.

Thanks, Ashish

>> +        sev_asid_count = max_sev_asid - min_sev_asid + 1;
>> +        WARN_ON_ONCE(misc_cg_set_capacity(MISC_CG_RES_SEV, 
>> sev_asid_count));
>> +    }
>>       sev_supported = true;
>>         /* SEV-ES support requested? */
>> @@ -2261,7 +2269,8 @@ void __init sev_hardware_setup(void)
>>   out:
>>       if (boot_cpu_has(X86_FEATURE_SEV))
>>           pr_info("SEV %s (ASIDs %u - %u)\n",
>> -            sev_supported ? "enabled" : "disabled",
>> +            sev_supported ? (min_sev_asid <= max_sev_asid ? 
>> "enabled" : "unusable")
>> +            : "disabled",
>>               min_sev_asid, max_sev_asid);
>>       if (boot_cpu_has(X86_FEATURE_SEV_ES))
>>           pr_info("SEV-ES %s (ASIDs %u - %u)\n",