On Fri, 2023-02-03 at 11:02 -0500, Bruce Ashfield wrote:
> On Fri, Feb 3, 2023 at 10:47 AM Richard Purdie
> <richard.pur...@linuxfoundation.org> wrote:
> >
> > The pmu-events.c file is generated by a python script making os.scandir()
> > calls. The return value is "order on disk" which can cary between machines.
> >
> > Add in a sed to fix the perf source to sort this data which makes
> > the pmu-events.c file deterministic.
>
> Looks good to me. The perf recipe is the great collector of sed manipulations
> :)
>
> >
> > We should try and upstream this change but we'll need to sed for varying
> > kernel versions. We should also try and get the perf source being added
> > to the perf-devsrc package so when failures like this occur, diffoscope
> > is much more helpful!
>
> I can do this, if you haven't started on it. I can't say that i know
> exactly why it
> isn't already there, but it can't be that hard to figure out :)
I haven't looked at submission upstream. I did have a quick look at the
sources issue and realised:
diff --git a/meta/recipes-kernel/perf/perf.bb b/meta/recipes-kernel/perf/perf.bb
index 1dff39a17e4..0a3179f18be 100644
--- a/meta/recipes-kernel/perf/perf.bb
+++ b/meta/recipes-kernel/perf/perf.bb
@@ -361,5 +361,5 @@ FILES:${PN}-python = " \
FILES:${PN}-perl = "${libexecdir}/perf-core/scripts/perl"
-INHIBIT_PACKAGE_DEBUG_SPLIT="1"
+#INHIBIT_PACKAGE_DEBUG_SPLIT="1"
DEBUG_OPTIMIZATION:append = " -Wno-error=maybe-uninitialized"
which makes them appear. Does anyone remember why we have that and if
we can remove it? :)
Cheers,
Richard
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#176729):
https://lists.openembedded.org/g/openembedded-core/message/176729
Mute This Topic: https://lists.openembedded.org/mt/96725319/21656
Group Owner: openembedded-core+ow...@lists.openembedded.org
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
