linux-kernel - Re: [PATCH][KVM] Add support for Pause Filtering to AMD SVM

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:	Tue, 5 May 2009 18:05:44 +0200
From:	Bert Wesarg <bert.wesarg@...glemail.com>
To:	Mark Langsdorf <mark.langsdorf@....com>
Cc:	joerg.roedel@....com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH][KVM] Add support for Pause Filtering to AMD SVM

On Tue, May 5, 2009 at 16:09, Mark Langsdorf <mark.langsdorf@....com> wrote:
> commit 6f15c833f56267baf5abdd0fbc90a81489573053
> Author: Mark Langsdorf <mlangsdo@...pnow.amd.com>
> Date:   Mon May 4 15:02:38 2009 -0500
>
>    New AMD processors will support the Pause Filter Feature.
>    This feature creates a new field in the VMCB called Pause
>    Filter Count.  If Pause Filter Count is greater than 0 and
>    ntercepting PAUSEs is enabled, the processor will increment
>    an internal counter when a PAUSE instruction occurs instead
>    of intercepting.  When the internal counter reaches the
>    Pause Filter Count value, a PAUSE intercept will occur.
>
>    This feature can be used to detect contended spinlocks,
>    especially when the lock holding VCPU is not scheduled.
>    Rescheduling another VCPU prevents the VCPU seeking the
>    lock from wasting its quantum by spinning idly.
>
>    Experimental results show that most spinlocks are held
>    for less than 1000 PAUSE cycles or more than a few
>    thousand.  Default the Pause Filter Counter to 3000 to
>    detect the contended spinlocks.
>
>    Processor support for this feature is indicated by a CPUID
>    bit.
>
>    On a 24 core system running 4 guests each with 16 VCPUs,
>    this patch improved overall performance of each guest's
>    32 job kernbench by approximately 1%.  Further performance
>    improvement may be possible with a more sophisticated
>    yield algorithm.
>
>    -Mark Langsdorf
>    Operating System Research Center
>    AMD
>
>    Signed-off-by: Mark Langsdorf <mark.langsdorf@....com>
>
> diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
> index ef43a18..14dab13 100644
> --- a/arch/x86/kvm/svm.c
> +++ b/arch/x86/kvm/svm.c
> @@ -575,6 +576,12 @@ static void init_vmcb(struct vcpu_svm *svm)
>
>        svm->nested_vmcb = 0;
>        svm->vcpu.arch.hflags = HF_GIF_MASK;
> +
> +       if (svm_has(SVM_FEATURE_PAUSE_FILTER)) {
> +               control->pause_filter_count = 5000;
The commit message says something about 3000.

> +               control->intercept |= (1ULL << INTERCEPT_PAUSE);
> +       }
> +
>  }
>
>  static int svm_vcpu_reset(struct kvm_vcpu *vcpu)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/