[no subject]

From: YONETANI Tomokazu <qhwt+dfly@les.ath.cx>
Subject: Re: cvs commit: src/sys/boot/pc32/libi386 biosacpi.c src/sys/conf         acpi.mk files src/sys/dev/acpica5 Makefile Makefile.inc         acdragonfly.h acpi.c acpi_acad.c acpi_button.c acpi_cmbat.c         acpi_cpu.c acpi_ec.c acpi_lid.c acpi_resource.c acpi_thermal.c ...
Date: Fri, 19 Jan 2007 11:40:21 +0900
BestServHost: crater.dragonflybsd.org
List-Post: <mailto:commits@crater.dragonflybsd.org>
List-Subscribe: <mailto:commits-request@crater.dragonflybsd.org?body=subscribe>
List-Unsubscribe: <mailto:commits-request@crater.dragonflybsd.org?body=unsubscribe>
List-Help: <mailto:commits-request@crater.dragonflybsd.org?body=help>
List-Owner: <mailto:owner-commits@crater.dragonflybsd.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
References: <200701171731.l0HHVKth003337@crater.dragonflybsd.org> <45ae8a49$0$832$415eb37d@crater_reader.dragonflybsd.org> <20070117221805.GA781@grazer.bsd.krakow.pl> <20070118001145.GG19888@les.ath.cx> <45af1645$0$830$415eb37d@crater_reader.dragonflybsd.org> <20070118112935.GF883@les.ath.cx> <45af8aa4$0$833$415eb37d@crater_reader.dragonflybsd.org> <20070118165739.GG883@les.ath.cx> <45afba59$0$829$415eb37d@crater_reader.dragonflybsd.org> <45afc326$0$833$415eb37d@crater_reader.dragonflybsd.org>

In-Reply-To: <45afc326$0$833$415eb37d@crater_reader.dragonflybsd.org>
Sender: commits-errors@crater.dragonflybsd.org
Errors-To: commits-errors@crater.dragonflybsd.org
Lines: 29
NNTP-Posting-Host: 216.240.41.25
X-Trace: 1169174749 crater_reader.dragonflybsd.org 831 216.240.41.25
Xref: crater_reader.dragonflybsd.org dragonfly.commits:14008

On Thu, Jan 18, 2007 at 06:57:42PM +0000, Johannes Hofmann wrote:
> > Just an update. It seems that after a while it does not return from 
> > 
> >    Status = WalkState->DescendingCallback (WalkState, Op); 
> > 
> > in interpreter/parser/psloop.c:AcpiPsBuildNamedOp.
> > 
> > BTW, is there a way to enable more verbose logging, so that all
> > ACPI_FUNCTION_TRACE_PTR() calls would give some output. That way
> > I could avoid adding all those kprintfs.

Sorry, you could have added ACPI_ALL_COMPONENTS to debug.acpi.layer
to print function traces in ACPI-CA code.

> Uff, now I've got it.
> It hangs in OsdCache.c:AcpiOsAcquireObject
> 
> Object = objcache_get(Cache->cache, M_WAITOK);

Do you get the same hang if you change the malloc flag from M_WAITOK
to M_NOWAIT or M_INTWAIT?

I observed a lock up on an SMP kernel, but I think it's been fixed in
kern_objcache.c rev1.16, so maybe it's unreleated.
Can also you try rebuilding acpi.ko with `ACPI_USE_LOCAL_CACHE=yes'
defined?  If that works, maybe we need to turn on ACPI_USE_LOCAL_CACHE
for the release unless we find what's going wrong.

Cheers.