Messages in this thread | | | Date | Wed, 07 Aug 2013 16:24:23 -0400 | From | David Ahern <> | Subject | Re: [PATCH 4/5] perf kvm: option to print events that exceed a threshold |
| |
On 8/7/13 3:27 PM, Arnaldo Carvalho de Melo wrote: > Em Mon, Aug 05, 2013 at 09:41:36PM -0400, David Ahern escreveu: >> This is useful to spot high latency blips. It is normal for HLT >> reasons to have long exit times, so strip those from the threshold >> check. > > So 'perf trace' has a similar option: > > [acme@zoo linux]$ perf trace --help > <SNIP> > --duration <float> > show only events with duration > N.M ms > <SNIP> > > While you used: > >> + OPT_U64('T', "threshold", &kvm->threshold, >> + "show events other than HALT that take longer than threshold usecs"), > > I prefer 'duration', and haven't assigned a single letter to it, they > are scarce and we have shell long option completion, what do you think? > > The term was chosen, for 'perf trace' on the original patchset: > > http://lwn.net/Articles/415728/ > > - Arnaldo >
No preference really. I need to send out an update to the man page. I can update this patch as well and re-send.
David
| |