From mboxrd@z Thu Jan 1 00:00:00 1970 From: Antoine Martin Subject: Re: idle load Date: Sat, 05 Apr 2008 09:57:22 +0100 Message-ID: <47F73EF2.6090104@nagafix.co.uk> References: <47ED8238.1010904@nagafix.co.uk> <1206823337.9594.50.camel@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Cc: kvm-devel@lists.sourceforge.net To: dor.laor@qumranet.com Return-path: In-Reply-To: <1206823337.9594.50.camel@localhost.localdomain> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: kvm-devel-bounces@lists.sourceforge.net Errors-To: kvm-devel-bounces@lists.sourceforge.net List-Id: kvm.vger.kernel.org -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Dor Laor wrote: | On Fri, 2008-03-28 at 23:41 +0000, Antoine Martin wrote: | Hi, | | Guests seem to be doing a lot of timer related polling, is there any way | to reduce that (which would reduce load on the host): | | |> It's weird we don't see any ioctl in the middle, are you sure kvm |> modules are loaded? Ooops, the guest I traced was started before loading them... Here is one that got started after (sorry for the line noise): select(5, [0 4], [], [], {0, 0}) = 0 (Timeout) select(0, [], NULL, NULL, {0, 0}) = 0 (Timeout) clock_gettime(CLOCK_MONOTONIC, {1685112, 356558791}) = 0 ioctl(5, 0x4008ae61, 0x7fff4f1caff0) = 0 ioctl(5, 0x4008ae61, 0x7fff4f1caff0) = 0 ioctl(5, 0x4008ae61, 0x7fff4f1caff0) = 0 ioctl(5, 0x4008ae61, 0x7fff4f1caff0) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 356664879}) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 356683016}) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 356703393}) = 0 timer_gettime(0x1, {it_interval={0, 0}, it_value={0, 0}}) = 0 timer_settime(0x1, 0, {it_interval={0, 0}, it_value={0, 19000000}}, NULL) = 0 ioctl(6, 0xae80, 0) = 0 ioctl(6, 0xae80, 0) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 356829961}) = 0 ioctl(5, 0x4008ae61, 0x7fff4f1cb160) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 356866505}) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 356886845}) = 0 timer_gettime(0x1, {it_interval={0, 0}, it_value={0, 18844990}}) = 0 ioctl(6, 0xae80, 0) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 356959171}) = 0 ioctl(6, 0xae80, 0) = 0 ioctl(6, 0xae80, 0) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 357036644}) = 0 ioctl(5, 0x4008ae61, 0x7fff4f1cb160) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 357072651}) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 357092946}) = 0 timer_gettime(0x1, {it_interval={0, 0}, it_value={0, 18639127}}) = 0 timer_settime(0x1, 0, {it_interval={0, 0}, it_value={0, 9158000}}, NULL) = 0 ioctl(6, 0xae80, 0) = 0 ioctl(6, 0xae80, 0) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 357235274}) = 0 ioctl(5, 0x4008ae61, 0x7fff4f1cb160) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 357271483}) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 357293651}) = 0 timer_gettime(0x1, {it_interval={0, 0}, it_value={0, 8985696}}) = 0 ioctl(6, 0xae80, 0) = 0 ioctl(6, 0xae80, 0) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 357386790}) = 0 ioctl(5, 0x4008ae61, 0x7fff4f1cb160) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 357422880}) = 0 clock_gettime(CLOCK_MONOTONIC, {1685112, 357442799}) = 0 | |> Another thing to do is run with -clock=[unix|rtc] What about -clock dynticks? Isn't that going to reduce unnecessary calls? Cheers Antoine -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEAREKAAYFAkf3PvIACgkQGK2zHPGK1rtVfgCcDqdt8zWS//HEUPHdVf9I+2t9 650AnjLM9HOOqyk/Rqsev4vQmXII37jr =1f+t -----END PGP SIGNATURE----- ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Register now and save $200. Hurry, offer ends at 11:59 p.m., Monday, April 7! Use priority code J8TLD2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone