as usual you were correct and i had screwed something up pmd structure i was reading. why is there not a pfmk_stop limitation for writing the registers but there exists one for reading. this seems fundamentally backwards to me?
thanks again -dwn On Wed, 2006-07-19 at 10:21 -0700, Stephane Eranian wrote: > Dave, > > On Wed, Jul 19, 2006 at 10:09:33AM -0400, Dave Nellans wrote: > > is there a method of resetting a given ctr to 0 that has less overhead > > than the whole process of setting up a context/desc/counter and pulling > > it all back down? > > > Of course there is, you just need to call pfmk_write_pmds() for this counter. > But again, rmember that you must be on the right CPU. You do not need to > issue pfmk_stop() to do this. > > > i though just stopping counting, writing the pmd, and starting counting > > might get the job done... but it didn't. there has to be a way to do > > this with less overhead. > > > -- Dave Nellans [EMAIL PROTECTED] http://www.nellans.org/~dwn/
signature.asc
Description: This is a digitally signed message part
_______________________________________________ perfmon mailing list [email protected] http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
