On Wed, Aug 28, 2024 at 10:02 AM Andi Kleen [off-list ref] wrote:

On Wed, Aug 28, 2024 at 03:07:14PM +0100, James Clark wrote:

quoted

I rebased this one and made some other fixes so that I could test it,
so I thought I'd repost it here in case it's helpful. I also added a
new test.

But for the testing it all looks ok.

There is one small difference where it now hides _all_ <not supported>
events, when previously it would only hide some selected subset of
events like "stalled-cycles-frontend". I think this is now more
consistent across platforms because, for example, Apple M only has
cycles and instructions, and the rest of the default events would
always show as <not supported> there.

I'm not a big fan of hiding all of <unsupported>, when they are explicitly
specified on the command line they should be absolutely shown.

I do have tools that specify events on the command line and expect
the same order/events in the output. This might well cause breakage.

The patch series doesn't do this, it just doesn't display unsupported
events that came from the set of default events - default evsels have
a "skippable" flag set. In patch 3 James has added to
should_skip_zero_counter in stat-display.c:

+       /*
+        * Skip unsupported default events when not verbose. (default events
+        * are all marked 'skippable').
+        */
+       if (verbose == 0 && counter->skippable && !counter->supported)
+               return true;

Thanks,
Ian