Messages in this thread | | | Date | Tue, 24 May 2011 23:37:35 +0200 | From | Ingo Molnar <> | Subject | Re: perf: regression -- missing /sys/devices/system/cpu/perf_events |
| |
* Vince Weaver <vweaver1@eecs.utk.edu> wrote:
> On Tue, 24 May 2011, Ingo Molnar wrote: > > > > So, what is wrong with the method Peter suggested: the presence of the perf > > syscall (it not returning -ENOSYS) is bona fide evidence that perf is > > available. > > it's just hard to do that from a shell script.
Yeah.
> also, running the perf syscall can be tricky if you have a new kernel but an > older set of header files that doesn't have the syscall number defined.
I suspect you could add some quick band-aid like:
#ifndef __NR_perf_event_open # ifdef __i386__ # define __NR_perf_event_open 336 # endif # ifdef __x86_64__ # define __NR_perf_event_open 298 # endif # ifdef __powerpc__ # define __NR_perf_event_open 319 # endif # ifdef __arm__ # define __NR_perf_event_open 364 # endif #endif
#ifndef __NR_perf_event_open # error Please add the __NR_perf_event_open definition for this architecture! #endif
This should cover 99% of the users - and fill in the table as people report build failures :)
Thanks,
Ingo
| |