linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Babu Moger <babu.moger@amd.com>
To: joro@8bytes.org, tglx@linutronix.de, mingo@redhat.com,
	hpa@zytor.com, x86@kernel.org, pbonzini@redhat.com,
	rkrcmar@redhat.com
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org, babu.moger@amd.com
Subject: [PATCH v2 0/5] KVM: x86: Implement Pause Loop Exit for SVM
Date: Fri, 16 Mar 2018 16:37:21 -0400	[thread overview]
Message-ID: <1521232646-79580-1-git-send-email-babu.moger@amd.com> (raw)

This series adds PLE(pause loop exit) logic from VMX to SVM.

We have noticed considerable reduction in number of VMEXITS due to pause
interceptions after these changes. Here are the numbers on one guest with
32 vcpus on AMD EPYC system. We have used boot parameter idle=poll to
simulate extensive pauses on the guest.

Here are VMEXITS in 10 seconds interval.

#VMEXITS(in 10s) 	Before the change  After the change
Pauses			810199			504
Total			882184			325415

v2:
Handled most of the Radim's comments. Here are the changes.
1. Fixed the module parameters to unsigned variants
1. Kept the default pause_filter_count = 3000 for AMD.
2. Removed the ple_window_actual_max parameter. Added a check
   in __grow_ple_window to avoid overflow.
3. Fixed the naming conventions for SVM module parameters.
   They are now called as pause_filter_count, pause_filter_thresh,
   pause_filter_shrink and pause_filter_grow.
4. Taken care of intercept setting and clearing.
5. Fixed few more text changes.

v1:
Initial RFC version

Babu Moger (5):
  KVM: VMX: Fix the module parameters for vmx
  KVM: VMX: Remove ple_window_actual_max
  KVM: VMX: Bring the common code to header file
  KVM: SVM: Add pause filter threshold
  KVM: SVM: Implement pause loop exit logic in SVM

 arch/x86/include/asm/svm.h |   3 +-
 arch/x86/kvm/svm.c         | 108 ++++++++++++++++++++++++++++++++++++++++++++-
 arch/x86/kvm/vmx.c         |  82 +++++++---------------------------
 arch/x86/kvm/x86.h         |  37 ++++++++++++++++
 4 files changed, 161 insertions(+), 69 deletions(-)

-- 
1.8.3.1

             reply	other threads:[~2018-03-16 20:37 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-16 20:37 Babu Moger [this message]
2018-03-16 20:37 ` [PATCH v2 1/5] KVM: VMX: Fix the module parameters for vmx Babu Moger
2018-03-16 20:37 ` [PATCH v2 2/5] KVM: VMX: Remove ple_window_actual_max Babu Moger
2018-03-28 20:24   ` Radim Krčmář
2018-03-16 20:37 ` [PATCH v2 3/5] KVM: VMX: Bring the common code to header file Babu Moger
2018-03-28 20:26   ` Radim Krčmář
2018-03-28 21:12     ` Moger, Babu
2018-03-16 20:37 ` [PATCH v2 4/5] KVM: SVM: Add pause filter threshold Babu Moger
2018-03-16 20:37 ` [PATCH v2 5/5] KVM: SVM: Implement pause loop exit logic in SVM Babu Moger
2018-03-28 20:30   ` Radim Krčmář
2018-03-28 21:14     ` Moger, Babu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1521232646-79580-1-git-send-email-babu.moger@amd.com \
    --to=babu.moger@amd.com \
    --cc=hpa@zytor.com \
    --cc=joro@8bytes.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=rkrcmar@redhat.com \
    --cc=tglx@linutronix.de \
    --cc=x86@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).