This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] edge profiler considering liveness


> Hi,
> 
> as described in:
> http://gcc.gnu.org/ml/gcc/2004-10/msg00149.html
> 
> I need the edge profiler to take liveness data into account when adding profiling code.
> In order to replace insert_insn_on_edge by safe_insert_insn_on_edge life analysis has
> to be available. Therefore I've added a profiling hook for initializing which performs
> life analysis for the rtl version and is currently empty for the tree version.
> 
> Bootstrapped without regressions on s390 and s390x.
> 
> What do you think about it?

The patch is OK for mainline assuming that will succesfully
profiledbootstrap at least on platform that works currently.

Also please drop explanationary comment to the life_analysis call here:

> + /* Do initialization work for the edge profiler.  */
> + 
> + static void
> + rtl_init_edge_profiler (void)
> + {
> +   life_analysis (NULL, 0);
> + }
> + 

And I blieve that this is not valid coding style, the braces should be
on separate lines.
> + 
> + /* Do initialization work for the edge profiler.  */
> + 
> + static void
> + tree_init_edge_profiler (void) {}
> + 

Thanks,
Honza


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]