Hi all,
I didn't get a chance to comment on this one, but the below (mmaped)
reading of the data area was recently removed from perfmon in order
to increase the probability of getting the patch into the kernel.
While at this point, I think anything we can do to get this in is the
right thing, I would eventually like to see this feature added back
in. Fast and direct access to the full 64 bit virtualized counters on
systems that can read the physical counters in user mode is of the
utmost importance to many applications and tools that do 'first
person' monitoring. This is also one of the reasons the perfctr patch
continues to be very popular.
Regards,
Phil
On Oct 17, 2007, at 4:34 PM, Milena Milenkovic wrote:
By more accurate, we mean providing an option to exclude time spent
in interrupts from per-thread time.
By more efficient, we mean providing a way for user-space tools to
read a mapped data area
where perfmon would write the values of performance monitoring
counters at the last significant event (interrupt exit/dispatch)
for each thread.
_______________________________________________
perfmon mailing list
[email protected]
http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
