Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 19 Mar 2008 12:38:10 -0700
From:      "Kevin Oberman" <oberman@es.net>
To:        John Baldwin <jhb@freebsd.org>
Cc:        Sam Leffler <sam@freebsd.org>, Igor Mozolevsky <igor@hybrid-lab.co.uk>, freebsd-current@freebsd.org
Subject:   Re: Compiling 6.3->7: lapic frequency madness 
Message-ID:  <20080319193810.BE81145019@ptavv.es.net>
In-Reply-To: Your message of "Wed, 19 Mar 2008 09:19:59 EDT." <200803190919.59783.jhb@freebsd.org> 

next in thread | previous in thread | raw e-mail | index | archive | help
--==_Exmh_1205955490_93241P
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

> From: John Baldwin <jhb@freebsd.org>
> Date: Wed, 19 Mar 2008 09:19:59 -0400
> Sender: owner-freebsd-current@freebsd.org
> 
> On Tuesday 18 March 2008 04:24:33 pm Sam Leffler wrote:
> > Igor Mozolevsky wrote:
> > > On 18/03/2008, John Baldwin <jhb@freebsd.org> wrote:
> > >> On Tuesday 18 March 2008 11:49:29 am Igor Mozolevsky wrote:
> > >>  > On 18/03/2008, John Baldwin <jhb@freebsd.org> wrote:
> > >>  > > First, which one is broken?  Have you tried a RELENG_7_0 kernel and
> > >>  > > does
> > >>
> > >>  it do
> > >>
> > >>  > >  the same as a release CD kernel?
> > >>  >
> > >>  > Tried both ISO image and RELENG_7_0, they both work ok, but RELENG_7
> > >>
> > >>  doesn't...
> > >>
> > >>
> > >> Ok, and it is cpufreq related?  Which cpufreq drivers are attaching to
> > >> your CPU?
> > >
> > > Well, I wouldn't say I'm sure it's 100% cpufreq related, but removing
> > > it from the kernel fixes the problem... It's also possible that
> > > cpufreq is interfering with something else... Anyhow, I've got both
> > > est and p4tcc attaching.
> >
> > I'm having lots of problems with cpu freq throttling on t4x laptops.
> > The system gets very sluggish and when I check the frequency it's been
> > throttled down.  This happens even on ac (haven't checked what powerd is
> > doing then).
> 
> My laptop actually wigs out if it throttles the CPU down too slow.  What 
> happens to me is that an ACPI GPE triggers and it takes a long time to 
> execute at 100Mhz and I actually end up with a sort of GPE storm where the 
> CPU spends all its time handling the GPE interrupts. :-/  I worked around it 
> by using the 'debug.cpufreq.lowest' (or whatever it's called, it's something 
> like that) tunable to prevent the CPU from going slower than about 400 Mhz 
> and since then CPU throttling has worked fine on my laptop w/o issues.

I can run the CPU all the way down to 100 MHz without problems. I
disable P4TCC in loader.conf. EST seems to handle low speeds just
fine. I really don't see the advantage of manual setting of the TCC
when EST is available.

My loader.conf contains these items I consider T43 specific:
hint.apic.0.disabled=1
hint.psm.0.flags="0x2000"
hint.p4tcc.0.disabled="1"
hw.ata.atapi.dma="1"
hw.pci.do_power_nodriver="2"
hw.psm.synaptics_support="1"
acpi_ibm_load="YES"
snd_ich_load="YES"
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman@es.net			Phone: +1 510 486-8634
Key fingerprint:059B 2DDF 031C 9BA3 14A4  EADA 927D EBB3 987B 3751

--==_Exmh_1205955490_93241P
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.8 (FreeBSD)
Comment: Exmh version 2.5 06/03/2002

iD8DBQFH4Wuikn3rs5h7N1ERAmRPAJ0VuRz9ofXVv2eC2RWYNK7xAU2tjQCdFi0/
LAJ3+wBC5pEpnspaUmG8NB0=
=1oFO
-----END PGP SIGNATURE-----

--==_Exmh_1205955490_93241P--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20080319193810.BE81145019>