Date: Mon, 21 Jun 2010 23:46:11 +0300 From: Alexander Motin <mav@FreeBSD.org> To: Kostik Belousov <kostikbel@gmail.com> Cc: svn-src-head@FreeBSD.org, svn-src-all@FreeBSD.org, src-committers@FreeBSD.org Subject: Re: svn commit: r209371 - in head/sys: amd64/amd64 amd64/include conf dev/acpica i386/i386 i386/include isa kern pc98/cbus sys x86/isa x86/x86 Message-ID: <4C1FCF93.9000703@FreeBSD.org> In-Reply-To: <20100621203419.GK13238@deviant.kiev.zoral.com.ua> References: <201006202133.o5KLXTG1023067@svn.freebsd.org> <20100621195838.GJ13238@deviant.kiev.zoral.com.ua> <4C1FC6C1.5060800@FreeBSD.org> <4C1FC6E1.4080001@FreeBSD.org> <20100621203419.GK13238@deviant.kiev.zoral.com.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
Kostik Belousov wrote: > On Mon, Jun 21, 2010 at 11:09:05PM +0300, Alexander Motin wrote: >> Alexander Motin wrote: >>> Kostik Belousov wrote: >>>> On Sun, Jun 20, 2010 at 09:33:29PM +0000, Alexander Motin wrote: >>>>> Author: mav >>>>> Date: Sun Jun 20 21:33:29 2010 >>>>> New Revision: 209371 >>>>> URL: http://svn.freebsd.org/changeset/base/209371 >>>>> >>>>> Log: >>>>> Implement new event timers infrastructure. It provides unified APIs for >>>>> writing event timer drivers, for choosing best possible drivers by machine >>>>> independent code and for operating them to supply kernel with hardclock(), >>>>> statclock() and profclock() events in unified fashion on various hardware. >>>> This broke QEMU for me. I cannot boot FreeBSD guest under QEMU anymore. >>>> QEMU (not FreeBSD kernel) panics with >>>> qemu: level-triggered hpet not supported >>>> message. >>> According to specification, it is not optional. No more cookies! > Is there a way to disable hpet at all ? It would be fine with me. > Does hint below supposed to do this ? Yes. Hint below disables HPET event timers, leaving only time counter part, almost same as it was before. >>>> Setting kern.eventtimer.timer1 to LAPIC or i8254, and timer2 to NONE >>>> does not help. >>> Try `hint.apic.0.clock=0` and send me verbose dmesg, I'll try to make >>> workaround if possible. >> Sorry, 'hint.hpet.0.clock=0'. >> > Setting the hint results in > panic: lock (time lock) sleep mutex does not match earlier (spin mutex) lock, > with the backtrace > enroll() > witness_init() > lock_init() > mtx_init() > initclocks() > mi_startup() > (transribed by hand). I am surprised that none of my systems fired it. Try this: --- subr_witness.c.prev 2010-06-17 14:07:36.000000000 +0300 +++ subr_witness.c 2010-06-21 23:43:14.000000000 +0300 @@ -495,6 +495,7 @@ static struct witness_order_list_entry o #ifdef HWPMC_HOOKS { "pmc-sleep", &lock_class_mtx_sleep }, #endif + { "time lock", &lock_class_mtx_sleep }, { NULL, NULL }, /* * Sockets @@ -654,7 +655,6 @@ static struct witness_order_list_entry o { "callout", &lock_class_mtx_spin }, { "entropy harvest mutex", &lock_class_mtx_spin }, { "syscons video lock", &lock_class_mtx_spin }, - { "time lock", &lock_class_mtx_spin }, #ifdef SMP { "smp rendezvous", &lock_class_mtx_spin }, #endif -- Alexander Motin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4C1FCF93.9000703>