Re: [perfmon2] [announce] Performance Counters for Linux, v6

2009-01-26 Thread stephane eranian
Hi, Corey brings up an interesting problem which I wanted to comment on. The current proposal hinges on the idea that by interpreting a single value the kernel can understand what the user wants to measure. For instance, if I pass type=0, then the kernel understands I want to measure CPU_CYCLES.

Re: [perfmon2] [announce] Performance Counters for Linux, v6

2009-01-26 Thread Ingo Molnar
* stephane eranian wrote: > Hi, > > Corey brings up an interesting problem which I wanted to comment on. > > The current proposal hinges on the idea that by interpreting a single > value the kernel can understand what the user wants to measure. For > instance, if I pass type=0, then the ker

[perfmon2] Build issues: perfmon- 090121 & linux kernel src 2.6.28

2009-01-26 Thread Kanaujia, Vishal (STSD)
Hi Stephan, We are trying to build Linux kernel(2.6.28) with latest perfmon(perfmon-new-base-090121). We are doing it on Red Hat Enterprise Linux ES release 4. We configured the kernel with h/w performance monitor support. While building the kernel, we are getting following error: .. ...

Re: [perfmon2] Build issues: perfmon- 090121 & linux kernel src 2.6.28

2009-01-26 Thread Kanaujia, Vishal (STSD)
Hi Stephan, Thanks for sharing the tip. It's working now. Regards, Vishal > -Original Message- > From: stephane eranian [mailto:eran...@googlemail.com] > Sent: Friday, January 23, 2009 6:01 PM > To: Kanaujia, Vishal (STSD) > Cc: perfmon2-devel@lists.sourceforge.net > Subject: Re: Build

Re: [perfmon2] [announce] Performance Counters for Linux, v6

2009-01-26 Thread stephane eranian
On Mon, Jan 26, 2009 at 4:17 PM, Ingo Molnar wrote: > > * stephane eranian wrote: > >> Hi, >> >> Corey brings up an interesting problem which I wanted to comment on. >> >> The current proposal hinges on the idea that by interpreting a single >> value the kernel can understand what the user wants

Re: [perfmon2] [announce] Performance Counters for Linux, v6

2009-01-26 Thread Corey Ashford
Ingo Molnar wrote: > * stephane eranian wrote: > >> Hi, >> >> Corey brings up an interesting problem which I wanted to comment on. >> >> The current proposal hinges on the idea that by interpreting a single >> value the kernel can understand what the user wants to measure. For >> instance, if

Re: [perfmon2] [announce] Performance Counters for Linux, v6

2009-01-26 Thread Luck, Tony
> - or the PMU capability is expressed as a special counter type (if it's > useful enough) - and then either the write() method or ioctl is extended > to express attributes we want to set/change while a counter is running. The product of: {exotic PMU modes} * {creative performance meas

Re: [perfmon2] [announce] Performance Counters for Linux, v6

2009-01-26 Thread Ingo Molnar
* Corey Ashford wrote: > Ingo Molnar wrote: >> * stephane eranian wrote: >> >>> Hi, >>> >>> Corey brings up an interesting problem which I wanted to comment on. >>> >>> The current proposal hinges on the idea that by interpreting a single >>> value the kernel can understand what the user want

Re: [perfmon2] [announce] Performance Counters for Linux, v6

2009-01-26 Thread Corey Ashford
Ingo Molnar wrote: > * Corey Ashford wrote: > >> Ingo Molnar wrote: >>> * stephane eranian wrote: >>> Hi, Corey brings up an interesting problem which I wanted to comment on. The current proposal hinges on the idea that by interpreting a single value the kernel ca

Re: [perfmon2] [announce] Performance Counters for Linux, v6

2009-01-26 Thread Ingo Molnar
* Luck, Tony wrote: > > - or the PMU capability is expressed as a special counter type (if it's > > useful enough) - and then either the write() method or ioctl is extended > > to express attributes we want to set/change while a counter is running. > > The product of: > {exotic PMU