Stephane Eranian wrote: > Yes, I realized that a while back and never got around to fixing it. > > Please try the (mostly untested) attached patch against 2.5.67 but it > should apply to older versions as well. Thank you for the patch, testing on a 2.5.64 kernel (I don't have 2.5.67 yet :-) ) with and without preempt. Hardware : tiger (4x Itanium2) > Let me know how it goes, especially on SMP. For now I would say it doesn't work :-( I get the same results from realfeel4 than before. However it works better (in a point of view) than before because now pfm_ovfl_block_reset() is called and sends a SIGPROF. But this is done only at the very end of realfeel4, according to the logs I think it's when sending the SIGALRM. Eric > PS: there is a dedicated mailing-list for perfmon. Sorry, I hadn't noticed, I'm now cross-posting.Received on Fri Apr 11 07:34:47 2003
This archive was generated by hypermail 2.1.8 : 2005-08-02 09:20:13 EST