Re: [PATCH] perf/x86/intel: Fix unwind errors from PEBS entries (mk-II)

From: Peter Zijlstra
Date: Tue Aug 07 2018 - 05:06:20 EST

On Mon, Aug 06, 2018 at 04:04:40PM -0700, Fubo Chen wrote:
> On Mon, Aug 6, 2018 at 3:30 PM Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> >
> > On Mon, Aug 06, 2018 at 02:28:18PM -0700, Fubo Chen wrote:
> > > Do you think the patch below is sufficient to suppress the sparse warning?
> >
> > Why would I want to make the code ugly to supress it?
> There are many kernel developers who use sparse to verify the
> correctness of endianness annotations (__be32, __le32, ...). When
> compiling kernel code with sparse every warning that is reported by
> sparse should be analyzed. Most kernel developers consider it annoying
> having to deal with false positive warnings. So I think that is useful
> to suppress false positive sparse warnings if it is possible to
> suppress false positives with a reasonable effort.

Last time I used sparse there were a metric ton of warnings. I really
can't be bothered about one more. Maybe fix sparse if you're bothered?