From: hpa@zytor.com (H. Peter Anvin) Date: Wed, 13 Dec 2006 22:45:19 +0000 Subject: [lm-sensors] [RFC] new MSR r/w functions per CPU Message-Id: <4580827F.8080703@zytor.com> List-Id: References: <45807469.6040609@assembler.cz> <20061213221026.GF2418@redhat.com> <45807C88.6060807@zytor.com> <20061213222616.GJ2418@redhat.com> In-Reply-To: <20061213222616.GJ2418@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Dave Jones , "H. Peter Anvin" , Rudolf Marek , norsk5@xmission.com, lkml , LM Sensors , bluesmoke-devel@lists.sourceforge.net Dave Jones wrote: > > Can you explain this a little further? I'm fairly certain > there are places in the kernel already doing this (or similar). > In fact, I cut-n-pasted most of the above from similar code in the > powernow-k8 driver. What exactly can we deadlock on? > I wanted to change the MSR driver to do the above, and Alan Cox objected that with realtime priority routines and/or user set affinity, that code might never be executed, so I retained the IPI-based code (which executes at the target processor at interrupt priority.) -hpa