Date: Mon, 05 Nov 2007 06:51:36 +0200 From: Ian FREISLICH <ianf@clue.co.za> To: David Malone <dwmalone@maths.tcd.ie> Cc: current@freebsd.org Subject: Re: setitimer() odd timing Message-ID: <E1Iotw0-0000Sh-Vg@clue.co.za> In-Reply-To: Message from David Malone <dwmalone@maths.tcd.ie> of "Fri, 02 Nov 2007 15:55:35 GMT." <20071102155535.GA88062@walton.maths.tcd.ie>
next in thread | previous in thread | raw e-mail | index | archive | help
David Malone wrote: > On Thu, Nov 01, 2007 at 03:38:11PM +0200, Ian FREISLICH wrote: > > I'm not sure where to start looking. If someone points me at what > > needs instrumenting, I'll see if I can narrow the problem space. > > I'd have a look at the output of "vmstat -i" and compare it to the > working machines. It might also be worth checking what timecounter > hardware is in use: "sysctl kern.timecounter.hardware". Well on other SMP machines that work fine, the cpu? time interrupt is 2000Hz. This one is 724 after the most recent boot. It changes every boot. kern.timecounter.choice: TSC(-100) ACPI-fast(1000) HPET(900) i8254(0) dummy(-1000000) kern.timecounter.hardware: ACPI-fast dev.acpi_timer.0.%desc: 24-bit timer at 3.579545MHz Changing the counter to HPET doesn't change things at all. I suspect it's got to do with whatever generates the CPU timer interrupts. Ian -- Ian Freislich
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E1Iotw0-0000Sh-Vg>