On Sat, Apr 18, 2015 at 07:39:27PM +0200, Ingo Molnar wrote: > > * Jiri Olsa <jo...@kernel.org> wrote: > > > hi, > > adding support to report error from event string parsing. > > Very nice! > > > This patchset contains support for standard parsing errors and more > > logic to recognize tracepoint and 'pmu//' terms, like: > > > > $ perf record -e 'sched:krava' ls > > invalid or unsupported event: 'sched:krava' > > \___ unknown tracepoint > > > > $ ./perf record -e 'cpu/even=0x1/' ls > > invalid or unsupported event: 'cpu/even=0x1/' > > \___ unknown term > > > > $ perf record -e cycles,cache-mises ls > > invalid or unsupported event: '..es,cache-mises' > > \___ parser error > > > > any feedback about the error string shape would be great ;-) > > So since we now know exactly what's going on, we might want to drop > the 'invalid or unsupported event' language as well, and make it > specific: > > $ ./perf record -e 'cpu/even=0x1/' ls > event syntax error: 'cpu/even=0x1/' > \___ unknown term > > > ?
ok > > Also, for the above error, could we easily list the valid terms? An > error like: > > $ ./perf record -e 'cpu/even=0x1/' ls > event syntax error: 'cpu/even=0x1/' > \___ unknown term > valid terms: "event", "raw". > > or so? we already carry list of all terms for given pmu, so it shouldn't be a problem jirka -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
