On Jan 8, 2007, at 7:18 PM, William Cohen wrote:
Philip Mucci wrote:
Hi folks,
Hapy Holidays. In regards to this subject, I sent a message asking
if we could standardize the PLM levels to actually mean something.
In this way, we don't have to change code for every processor
depending on the PLM mapping. In my private code, I have already
changed MIPS to work in accordance with the others, thus allowing
the oprofile code to work.
Does this definition also include masks for hypervisor/non-
hypervisor? This is going to be coming up on processors.
The MIPS one does. It includes hypervisor and interrupt PLM modes.
BTW, I have oprofile and perfmon working on i386 again with lots
of patches. Interested?
OProfile has kernel and user space masks. How does this differ from
the current implementation on OProfile? Maybe the easies thing
would be to see the patches.
Sorry, I mispoke. The patches are more oriented around working with
the latest code base of perfmon2. As far as the PLM bits, I left
oprofile/perfmon as is. I just had to change MIPS libpfm.
Phil
_______________________________________________
perfmon mailing list
[email protected]
http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
