public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH] vect: Init inside_cost in vect_model_reduction_cost
@ 2021-03-25  6:04 Kewen.Lin
  2021-03-25  6:16 ` Jeff Law
  0 siblings, 1 reply; 2+ messages in thread
From: Kewen.Lin @ 2021-03-25  6:04 UTC (permalink / raw)
  To: GCC Patches

Hi,

I happened to find this possible use of uninitialized inside_cost during
checking the cost for reduction.  Before r11-6587, all the if/else if/else
arms will assign values to inside_cost, so it was fine.  But from r11-6587
the else path won't assign any more, it leads the inside_cost possibly
uninitialized for later use.  Note that it won't have any functionality
issues, just could show one unreasonable cost in dump file.

This is a trivial fix and I guess it can be counted as obvious, but
I'm not sure if it's ok in stage 4.  So is it ok for trunk?


BR,
Kewen
-----
gcc/ChangeLog:

	* tree-vect-loop.c (vect_model_reduction_cost): Init inside_cost.

diff --git a/gcc/tree-vect-loop.c b/gcc/tree-vect-loop.c
index 3e973e774af..38d96fc1634 100644
--- a/gcc/tree-vect-loop.c
+++ b/gcc/tree-vect-loop.c
@@ -4465,7 +4465,7 @@ vect_model_reduction_cost (loop_vec_info loop_vinfo,
                            vect_reduction_type reduction_type,
                            int ncopies, stmt_vector_for_cost *cost_vec)
 {
-  int prologue_cost = 0, epilogue_cost = 0, inside_cost;
+  int prologue_cost = 0, epilogue_cost = 0, inside_cost = 0;
   enum tree_code code;
   optab optab;
   tree vectype;





^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH] vect: Init inside_cost in vect_model_reduction_cost
  2021-03-25  6:04 [PATCH] vect: Init inside_cost in vect_model_reduction_cost Kewen.Lin
@ 2021-03-25  6:16 ` Jeff Law
  0 siblings, 0 replies; 2+ messages in thread
From: Jeff Law @ 2021-03-25  6:16 UTC (permalink / raw)
  To: Kewen.Lin, GCC Patches


On 3/25/2021 12:04 AM, Kewen.Lin via Gcc-patches wrote:
> Hi,
>
> I happened to find this possible use of uninitialized inside_cost during
> checking the cost for reduction.  Before r11-6587, all the if/else if/else
> arms will assign values to inside_cost, so it was fine.  But from r11-6587
> the else path won't assign any more, it leads the inside_cost possibly
> uninitialized for later use.  Note that it won't have any functionality
> issues, just could show one unreasonable cost in dump file.
>
> This is a trivial fix and I guess it can be counted as obvious, but
> I'm not sure if it's ok in stage 4.  So is it ok for trunk?
>
>
> BR,
> Kewen
> -----
> gcc/ChangeLog:
>
> 	* tree-vect-loop.c (vect_model_reduction_cost): Init inside_cost.

OK

jeff


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2021-03-25  6:16 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-03-25  6:04 [PATCH] vect: Init inside_cost in vect_model_reduction_cost Kewen.Lin
2021-03-25  6:16 ` Jeff Law

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).