From mboxrd@z Thu Jan 1 00:00:00 1970 From: Shashidhar Patil Subject: Re: [PATCH 0/9] pci-assign: 64bit MMIO + better MSI-X table support Date: Sun, 12 Feb 2012 22:00:15 +0530 Message-ID: References: <20120128142104.25681.93072.stgit@bling.home> <1328549358.3074.108.camel@bling.home> <1328808186.2954.28.camel@ul30vt> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: kvm@vger.kernel.org, Jan Kiszka To: Alex Williamson Return-path: Received: from mail-ww0-f44.google.com ([74.125.82.44]:63822 "EHLO mail-ww0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755468Ab2BLQaR convert rfc822-to-8bit (ORCPT ); Sun, 12 Feb 2012 11:30:17 -0500 Received: by wgbdt10 with SMTP id dt10so4182919wgb.1 for ; Sun, 12 Feb 2012 08:30:15 -0800 (PST) In-Reply-To: <1328808186.2954.28.camel@ul30vt> Sender: kvm-owner@vger.kernel.org List-ID: HI Alex, I tried both FreeBSD stock and Intel official freebsd driver for 825= 99. I tried both FreeBSD 9.0 and FreeBSD8.2. No luck with traffic. -Shashidhar On Thu, Feb 9, 2012 at 10:53 PM, Alex Williamson wrote: > On Thu, 2012-02-09 at 21:53 +0530, Shashidhar Patil wrote: >> Hi Alex, >> =A0 =A0 I tested your code changes for two problems >> 1. MSIX vectors allocation for FreeBSD Guest >> =A0 =A0 Tested the allocation of MSI-X allocation happening for both >> FreeBSD-9.0 (also tested 8.2) and Linux guest. >> =A0 =A0 =A0Also tried multiple restart of the guests to see if the h= osts >> still maintains those allocations. With Linux >> =A0 =A0 =A0as guest I could pass traffic. Though interrupt allocatio= n worked >> the FreeBSD could to pass traffic. The FreeBSD >> =A0 =A0 =A0driver seems to be broken(atleast when running in guest). > > Which device/driver? =A0I was able to use an 82576, both physical dev= ice > and sr-iov virtual function) with the freebsd 9.0 igb driver. > >> 2. Guest MSI-X smp_affinity fix. >> =A0 =A0 =A0The smp_affinity configured in guest is honoured. The tra= ffic >> distribution is uniform. When smp affinity is configured >> =A0 =A0 =A0in host and guest properly this fix provides multi queue = and >> hence multi core traffic distribution in both host and >> =A0 =A0 =A0guest. The result is good packet throughput. >> >> Thanks for the the much needed fixes. > > Thanks for testing! > > Alex > >> -Shashidhar >> >> On Mon, Feb 6, 2012 at 10:59 PM, Alex Williamson >> wrote: >> > On Mon, 2012-02-06 at 21:25 +0530, Shashidhar Patil wrote: >> >> HI Alex, >> >> =A0 =A0 =A0I can give it a try. Please send me all patches as one= file if possible. >> >> I was about to report yet another problem with Guest MSI-X smp af= finity >> >> not being honoured by KVM while injecting interrupts. Its again i= n the context >> >> of 82599 device (tried with Linux as guest). I had a hunch that t= he MSI-X vector >> >> data which gets modified when SMP affinity is not tracked by the = qemu_kvm. >> >> So the interrupt injection was done in load balancing fashion. Be= cause of this >> >> interrupt LB in guest VCPUs I saw that packets for all the queues= were processed >> >> by one VCPU instead 2 or 4 available. >> >> Anyway I will test the the MSI-X affinity fix also and get back w= ith results. >> >> And I hope 82599 send/receive works. Thanks for your help. >> > >> > I pushed the v2 version to github, so you can grab it from here fo= r >> > testing: >> > >> > git://github.com/awilliam/qemu-kvm.git >> > >> > Branch pci-assign-msix. =A0Thanks, >> > >> > Alex >> > > > >