Thanks for the input, the patch seems  to be much more cute now.
Do we still need to play with release/non-release builds, or is it ok
to commit this patch to the trunk as it is?

Changelog:
2012-11-29  Michael Zolotukhin  <michael.v.zolotukhin@intel.com>

        * Makefile.in: Add target mddump, build/genmddump.o.  Extend
        genprogrtl with mddump.
        * genmddump.c: New.

On 27 November 2012 21:34, Richard Henderson <rth@redhat.com> wrote:
> On 11/22/2012 09:48 AM, Kirill Yukhin wrote:
>> +.PHONY: s-mddump
>> +s-mddump: $(BUILD_RTL) $(MD_DEPS) build/genmddump$(build_exeext)
>> +     $(RUN_GEN) build/genmddump$(build_exeext) $(md_file) 2> tmp-mddump.md
>
> I think just
>
> mddump: ...
>         $(RUN_GEN) ... > mddump
>
> will be sufficient.  This is not actually used by the build at all, so we
> don't need to play games with stamp files etc.
>
> There's no need for top-level makefile changes at all.  When you want to
> use this, simply cd into the gcc subdirectory.
>
>> +/* Dump all available rtl queues.  */
>> +void
>> +dump_expanded_md (void)
>
> Why?  Seems to me that you can just have genmddump.c simply use the
> generic read_md_rtx interface, dumping as it goes.  You might also
> consider dumping the pattern_lineno argument as a comment before the
> pattern.  Otherwise it might be tricky to match up the dump pattern
> with the original input file patterns.
>
>
> r~

-- 
---
Best regards,
Michael V. Zolotukhin,
Software Engineer
Intel Corporation.