On Wed, 2007-12-12 at 11:12 -0800, Russell Leidich wrote:
> Thanks for the idea, Subrata. Unfortunately, my code doesn't lend
> itself well to a test harness, as it uses PCI IO to get most of its
> work done. To test it would require either faking that IO, or
> stopping fans in order to create thermal throttling interrupts.
OK. That should be fine. In future when you think of writing testcases
for anything in Linux, please consider giving it to LTP. Your
contribution will be highly valuable.
> On Dec 12, 2007 12:20 AM, Subrata Modak <subrata@...> wrote:
> > Hi,
> > Do you have test cases for this which you would like to share with LTP.
> > Even if you have them in Raw C, we can work towards porting them to LTP.
> > You can also check out the following links for more info, and can always
> > ping me for anything:
> > http://ltp.sourceforge.net/documentation/how-to/ltp.php?
> > Regards--
> > Subrata
> > (LTP Maintainer)
> > On Tue, 2007-12-11 at 10:44 -0800, Russell Leidich wrote:
> > > This patch adds thermal interrupt logging support for AMD Barcelona to
> > > 2.6.18. In the style of the existing Intel implementation, thermal
> > > interrupts are logged as machine checks. Interrupts are logged only
> > > when throttling is activated, i.e. when the package gets too hot.
> > >
> > > Files affected:
> > >
> > > 2.6.18/arch/x86-64/kernel/Makefile
> > > 2.6.18/arch/x86-64/kernel/mce_amd.c
> > > 2.6.18/arch/x86-64/kernel/mce_intel.c
> > > 2.6.18/arch/x86-64/kernel/mce_thermal.c (new)
> > > 2.6.18/include/asm-x86-64/kernel/mce.h
> > >
> > > Signed-off-by: Russell Leidich <rml@...>
> > >
> > > --
> > > Russell Leidich