On Thu, Sep 05, 2024 at 08:42:08AM +1000, Stephen Rothwell wrote:
> On Mon, 2 Sep 2024 10:51:21 +1000 Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx> wrote:
> > The following commits are also in the perf-current tree as different
> > commits (but the same patches):
> > 6236ebe07131 ("perf daemon: Fix the build on more 32-bit architectures")
> > 2518e13275ab ("perf python: Fix the build on 32-bit arm by including missing "util/sample.h"")
> > 74fd69a35cae ("perf lock contention: Fix spinlock and rwlock accounting")
> > 37e2a19c98bf ("perf test pmu: Set uninitialized PMU alias to null")
> > These are commits
> > 478e3c7ebbe7 ("perf daemon: Fix the build on more 32-bit architectures")
> > 4cb80590f12d ("perf python: include "util/sample.h"")
> > 60f47d2c58cb ("perf lock contention: Fix spinlock and rwlock accounting")
> > 387ad33e5410 ("perf test pmu: Set uninitialized PMU alias to null")
> > and this last one is causing an unnecessary conflict.
> These latter commits are now in Linus' tree.
I did the usual merge of torvalds/master into
perf-tools-next/perf-tools-next (devel) when perf-tools/perf-tools
(current/fixes) gets merged upstream and fixed up the trivial merge
conflict.
The end result is at the tmp.perf-tools-next branch while I do the full
container build tests, the merge commit has the explanation for the
resolution:
Merge remote-tracking branch 'torvalds/master' into perf-tools-next
To pick up fixes from perf-tools/perf-tools, some of which were also in
perf-tools-next but were then indentified as being more appropriate to
go sooner, to fix regressions in v6.11.
Resolve a simple merge conflict in tools/perf/tests/pmu.c where a more
future proof approach to initialize all fields of a struct was used in
perf-tools-next, the one that is going into v6.11 is enough for the
segfault it addressed (using an uninitialized test_pmu.alias field).
Signed-off-by: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx>
Thanks,
- Arnaldo
