From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dominik Brodowski Subject: Re: powernow-k8: support acpi Date: Tue, 6 Apr 2004 16:55:02 +0200 Sender: cpufreq-bounces@www.linux.org.uk Message-ID: <20040406145502.GA7307@dominikbrodowski.de> References: <99F2150714F93F448942F9A9F112634C1163C8D7@txexmtae.amd.com> <20040321185417.GA7969@dominikbrodowski.de> <20040326122931.GA321@elf.ucw.cz> <20040401235635.GA25744@redhat.com> <20040402162447.GP19001@poupinou.org> <20040402163902.GC5688@redhat.com> <20040402170926.GA289@elf.ucw.cz> <20040406130730.GA5133@redhat.com> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="jRHKVT23PllUwdXP" Return-path: Content-Disposition: inline In-Reply-To: <20040406130730.GA5133@redhat.com> List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: cpufreq-bounces+glkc-cpufreq=gmane.org@www.linux.org.uk To: Dave Jones Cc: Andrew Morton , Bruno Ducrot , mark.langsdorf@amd.com, Cpufreq mailing list , Pavel Machek , paul.devriendt@amd.com --jRHKVT23PllUwdXP Content-Type: text/plain; charset=us-ascii Content-Disposition: inline On Tue, Apr 06, 2004 at 02:07:30PM +0100, Dave Jones wrote: > On Fri, Apr 02, 2004 at 07:09:26PM +0200, Pavel Machek wrote: > > > > I'm about to disappear for a day or two. I'll look into integrating whats > > > outstanding on Sunday. Pavel, if you have time between now and then > > > to do a diff of whats remaining between your tree and the driver I posted > > > this afternoon, that would make my life a little easier. > > > > Here it is: (I was not able to locate your changes, so it is possible > > I'm reverting some of those; but few extra inlines should not hurt > > that bad :-). > > Ok, the merge-from-hell is complete (split up into 23 csets). > The only diffs between your tree and mine should now be trivial > whitespace changes, and some small bits that looked bogus. Just found a small bug which would leave the preempt count negative(!) in special cases. Patch is attached. > You can grab what I have in my tree pending pushing to Linus at > http://www.codemonkey.org.uk/projects/cpufreq/pending.diff Will http://www.codemonkey.org.uk/projects/bitkeeper/cpufreq/ be a permanent institution, or is it just available for the moment? > One thing outstanding that probably does need fixing. Andi Kleen > mentioned on IRC yesterday .. > > found a funny bug in the powernow-k8 acpi changes > when acpi is disabled processor.ko won't load, but the new module needs symbols from that > i added a hack to keep it loaded anyways > > Be nice to get that fixed, and merged before these changes get pushed to mainline. Hm. Keeping processor.ko loaded seems to be an ugly hack to me. However, I can't see a way (yet) how powernow-k8.ko can rely on processor.ko's objects but not depend on it if these are not neeeded (psb case). Dominik --jRHKVT23PllUwdXP Content-Type: text/plain; charset=us-ascii Content-Disposition: attachment; filename="cpufreq-2.6.5-fix-jump" diff -ruN linux-original/arch/i386/kernel/cpu/cpufreq/powernow-k8.c linux/arch/i386/kernel/cpu/cpufreq/powernow-k8.c --- linux-original/arch/i386/kernel/cpu/cpufreq/powernow-k8.c 2004-04-06 16:41:19.907636632 +0200 +++ linux/arch/i386/kernel/cpu/cpufreq/powernow-k8.c 2004-04-06 16:50:54.428296216 +0200 @@ -848,7 +848,7 @@ if (smp_processor_id() != pol->cpu) { printk(KERN_ERR "limiting to cpu %u failed\n", pol->cpu); - goto err_out; + goto sched_out; } /* from this point, do not exit without restoring preempt and cpu */ @@ -892,6 +892,7 @@ err_out: preempt_enable_no_resched(); +sched_out: set_cpus_allowed(current, oldmask); schedule(); --jRHKVT23PllUwdXP Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Cpufreq mailing list Cpufreq@www.linux.org.uk http://www.linux.org.uk/mailman/listinfo/cpufreq --jRHKVT23PllUwdXP--