Date: Fri, 24 Dec 2010 17:16:40 +0100 From: Stefan Farfeleder <stefanf@FreeBSD.org> To: Alexander Motin <mav@FreeBSD.org> Cc: freebsd-current@freebsd.org Subject: Re: eventtimers hiccups Message-ID: <20101224161640.GD2642@mole.fafoe.narf.at> In-Reply-To: <4D14B30E.3090609@FreeBSD.org> References: <20101224143124.GC2642@mole.fafoe.narf.at> <4D14B30E.3090609@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Dec 24, 2010 at 04:49:50PM +0200, Alexander Motin wrote:
>
> Have you tried to look on what happens with your HPET interrupts during
> the problem? Is it timer hardware/driver problem, or something else?
>
> Have you tried to use LAPIC timer? - it has no race window between start
> and completion that may potentially stop HPET timer in some situations.
>
> Have you tried to enable kern.eventtimer.idletick?
>
> The best diagnostic would be to get KTR dump at the moment when problem
> begins. You should build kernel with
> options KTR
>
> options ALQ
>
> options KTR_ALQ
>
> options KTR_COMPILE=(KTR_SPARE2)
>
> options KTR_ENTRIES=131072
>
> options KTR_MASK=(KTR_SPARE2)
>
> and as soon as problem begins (before logs wrapped) you should run
> `ktrdump -c -o dump`.
>
> Also you may try in file acpi_hpet.c change line
> if (fdiv < 5000) {
> to the
> if (fdiv < 50000) {
> .
Hi Alexander,
thanks for your suggestions. I'll try them out and report then. It might
take some time though.
Stefan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20101224161640.GD2642>