linux-kernel - Re: [PATCH] KVM: SVM: Modify AVIC GATag to support max number of 512 vCPUs

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:   Mon, 6 Feb 2023 10:52:22 +0700
From:   "Suthikulpanit, Suravee" <suravee.suthikulpanit@....com>
To:     linux-kernel@...r.kernel.org, kvm@...r.kernel.org
Cc:     seanjc@...gle.com, alejandro.j.jimenez@...cle.com,
        mlevitsk@...hat.com, pbonzini@...hat.com, jon.grimm@....com,
        vasant.hegde@....com, kishon.vijayabraham@....com
Subject: Re: [PATCH] KVM: SVM: Modify AVIC GATag to support max number of 512
 vCPUs

Hi Sean / Paolo,

Do you have any concerns or questions regarding this patch?

Thanks,
Suravee

On 1/17/2023 5:08 PM, Suravee Suthikulpanit wrote:
> AVIC GATag is managed by the SVM driver, and is used by the IOMMU driver
> to program the AMD IOMMU IRTE to provide reference back to SVM in case
> the IOMMU cannot inject interrupt to the non-running vCPU. In such case,
> the IOMMU hardware notify the IOMMU driver by creating GALog entry with
> the corresponded GATag. The IOMMU driver processes the GALog entry and
> notifies SVM to schedule in the target vCPU.
> 
> Currently, SVM uses 8-bit vCPU ID and 24-bit VM ID to encode 32-bit GATag.
> Since x2AVIC supports upto 512 vCPUs, it requires to use at least 9-bit
> vCPU ID.
> 
> Therefore, modify the GATag enconding to use the number of bits required
> to support the maximum vCPUs.
> 
> Reported-by: Alejandro Jimenez <alejandro.j.jimenez@...cle.com>
> Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@....com>
> ---
>   arch/x86/include/asm/svm.h | 3 ++-
>   arch/x86/kvm/svm/avic.c    | 9 ++++-----
>   2 files changed, 6 insertions(+), 6 deletions(-)
> 
> diff --git a/arch/x86/include/asm/svm.h b/arch/x86/include/asm/svm.h
> index 0361626841bc..6738faf155e4 100644
> --- a/arch/x86/include/asm/svm.h
> +++ b/arch/x86/include/asm/svm.h
> @@ -256,7 +256,8 @@ enum avic_ipi_failure_cause {
>   	AVIC_IPI_FAILURE_INVALID_BACKING_PAGE,
>   };
>   
> -#define AVIC_PHYSICAL_MAX_INDEX_MASK	GENMASK_ULL(9, 0)
> +#define AVIC_PHYSICAL_MAX_INDEX_BITS	9
> +#define AVIC_PHYSICAL_MAX_INDEX_MASK	GENMASK_ULL(AVIC_PHYSICAL_MAX_INDEX_BITS, 0)
>   
>   /*
>    * For AVIC, the max index allowed for physical APIC ID
> diff --git a/arch/x86/kvm/svm/avic.c b/arch/x86/kvm/svm/avic.c
> index dd0e41d454a7..b1f8f51bbbd9 100644
> --- a/arch/x86/kvm/svm/avic.c
> +++ b/arch/x86/kvm/svm/avic.c
> @@ -28,16 +28,15 @@
>   #include "svm.h"
>   
>   /* AVIC GATAG is encoded using VM and VCPU IDs */
> -#define AVIC_VCPU_ID_BITS		8
> -#define AVIC_VCPU_ID_MASK		((1 << AVIC_VCPU_ID_BITS) - 1)
> +#define AVIC_VCPU_ID_MASK		((1 << AVIC_PHYSICAL_MAX_INDEX_BITS) - 1)
>   
> -#define AVIC_VM_ID_BITS			24
> +#define AVIC_VM_ID_BITS			(32 - AVIC_PHYSICAL_MAX_INDEX_BITS)
>   #define AVIC_VM_ID_NR			(1 << AVIC_VM_ID_BITS)
>   #define AVIC_VM_ID_MASK			((1 << AVIC_VM_ID_BITS) - 1)
>   
> -#define AVIC_GATAG(x, y)		(((x & AVIC_VM_ID_MASK) << AVIC_VCPU_ID_BITS) | \
> +#define AVIC_GATAG(x, y)		(((x & AVIC_VM_ID_MASK) << AVIC_PHYSICAL_MAX_INDEX_BITS) | \
>   						(y & AVIC_VCPU_ID_MASK))
> -#define AVIC_GATAG_TO_VMID(x)		((x >> AVIC_VCPU_ID_BITS) & AVIC_VM_ID_MASK)
> +#define AVIC_GATAG_TO_VMID(x)		((x >> AVIC_PHYSICAL_MAX_INDEX_BITS) & AVIC_VM_ID_MASK)
>   #define AVIC_GATAG_TO_VCPUID(x)		(x & AVIC_VCPU_ID_MASK)
>   
>   static bool force_avic;