This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH 2/2] btrace: set/show record btrace cpu
> From: "Metzger, Markus T" <markus.t.metzger@intel.com>
> CC: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>
> Date: Mon, 26 Feb 2018 15:45:36 +0000
> > > +@item set record btrace cpu @var{identifier} Set the processor to be
> > > +used for enabling trace decode errata workarounds.
> >
> > I think we need to say something about just what those "errata workarounds"
> > are, and what are they used for.
>
> I rephrased this to "... for enabling workarounds for processor errata when
> decoding the trace".
It's better, but still not clear enough. What kind of "errata" are we
talking about? The kind described in
https://community.amd.com/thread/186609, for example? And what do the
workarounds do?
If you can explain that to me or give an example, I will try to
propose some text to describe that in the manual.
> > > + add_prefix_cmd ("cpu", class_support, cmd_set_record_btrace_cpu,
> > > + _("\
> > > +Set the cpu to be used for trace decode.\n\n\ The format is
> > > +\"<vendor>: <identifier>\" or \"none\" or \"auto\" (default).
> > ^^
> > So should there be a blank after the colon, or shouldn't there be?
> > The example in the manual says no blank.
>
> White space is ignored. Do we write this explicitly?
Not necessarily. But I'd prefer us to consistently use one of the
forms.
The rest of the patch LGTM, thanks.