From mboxrd@z Thu Jan 1 00:00:00 1970 From: Gregory Haskins Subject: Re: Host latency peaks due to kvm-intel Date: Fri, 24 Jul 2009 08:01:57 -0400 Message-ID: <4A69A2B5.4010303@gmail.com> References: <4A68A6E5.6010808@siemens.com> <4A68BD5D.1070302@gmail.com> <4A6981B0.3000008@siemens.com> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="------------enigD647FB58A9F3E3C160C097DF" Cc: kvm-devel , RT , Avi Kivity , "Yang, Sheng" To: Jan Kiszka Return-path: In-Reply-To: <4A6981B0.3000008@siemens.com> Sender: linux-rt-users-owner@vger.kernel.org List-Id: kvm.vger.kernel.org This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --------------enigD647FB58A9F3E3C160C097DF Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: quoted-printable Jan Kiszka wrote: > Gregory Haskins wrote: > =20 >> Jan Kiszka wrote: >> =20 >>> Hi, >>> >>> did anyone recently tried current KVM for Intel over some real-time >>> Linux? I'm seeing more than 500 us latency peaks on the host, >>> specifically during VM startup. This applies to both 2.6.29.6-rt23 an= d >>> Xenomai/I-pipe. For -rt, I both tried the included (patched) KVM modu= les >>> as well as kvm.git head with some additionally required -rt fixes. >>> Xenomai ran over a 2.6.30 kernel with my own KVM-enabler patch. >>> >>> Early instrumentation actually points to the guest exit itself: I add= ed >>> markers right before and after the assembly part of vmx_vcpu_run, and= >>> further instrumentation reports that the next host APIC tick should g= o >>> off right inside guest mode. But KVM leaves the switching part 500 us= >>> too late in that case - as if guest exit on external IRQs was disable= d. >>> >>> Will debug this further, but I'm also curious to hear other user >>> experiences. >>> >>> Jan >>> >>> =20 >>> =20 >> Hi Jan, >> Did you try to run with latency-tracer enabled? If not, this may >> pinpoint the source for you. >> =20 > > I did, see above. > =20 Ah, sorry. It wasn't clear what "instrumenation" was or if you felt it was definitively pinpointed. :P Regards, -Greg --------------enigD647FB58A9F3E3C160C097DF Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.11 (Darwin) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkpporUACgkQP5K2CMvXmqFeVACeKGPbwas0TUsoQ/R0r9f197C1 4W4Ani5SUPV9E1VPS/JM5sYMRL53Kr2l =c9w5 -----END PGP SIGNATURE----- --------------enigD647FB58A9F3E3C160C097DF--