Will, On Thu, Jan 18, 2007 at 03:50:50PM -0500, William Cohen wrote: > >I got OProfile to work on perfmon v2.3 all IA-64 and X86. I need to > >clean the user level code a little bit. I would appreciate if you could > >take a look at it. > > > >There is one thing I am not so sure about at this point: how does OProfile > >handle unavailable counters when it assigns event -> counters, > >take the configuration with the NMI watchdog, for instance. > > > >Do you know anything about this? > > Yes, I developed a patch to handle this for OProfile to work with some > registers being reserved for watchdog. However, I seemed to have forgotten > to post the patch. I have attached the patch to this email. > The way we have the kernel OProfile setup when using perfmon does conflict with what you arwe doing in this patch. When used on top of perfmon, we do not populate counter entries in /dev/oprofile, thus your code will find that no counter is available.
If we want to maintain backward compatibility for event names and event tables, then we need to stick with we have today at that level. In my current patch for the daemon, I map OProfile counter to Perfmon counter only AFTER the assignment is made. It is hard to fail there if we detect that a perfmon counter is not available. The list of available counter must be passed to the event assignment code. It looks like we would have to build the OProfile list of available based on the Perfmon available PMC/PMD. So we would need bi-directional mapping. That is fairly easy to do on all X86/Itanium processors except for P4. Any other idea? -- -Stephane _______________________________________________ perfmon mailing list [email protected] http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
