[Bug target/54546] New: SH: Enable -fshrink-wrap

[Bug target/54546] New: SH: Enable -fshrink-wrap

[chrbr at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54546

             Bug #: 54546
           Summary: SH: Enable -fshrink-wrap
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: chrbr@gcc.gnu.org


Created attachment 28169
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28169
simple-return pattern

Implement the simple_return pattern to enable shrink-wrapping, which is
beneficial on SH when the prologue/epilogue is small enough or when not
optimizing for size. Adding the sh_can_use_return_insn_p function so
refinements based on epilogue size can be further added.

However this exposes a -freorder-blocks-and-partition -fprofile-use regression
in the testsuite with that must be investigated.

[Bug target/54546] SH: Enable -fshrink-wrap

[chrbr at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54546

chrbr at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement

[Bug target/54546] SH: Enable -fshrink-wrap

[chrbr at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54546

--- Comment #1 from chrbr at gcc dot gnu.org 2012-09-13 06:40:54 UTC ---
shrink-wrapping exposes a few problems related to partitioning. 

1) Tries to duplicate a Basic block that has only one predecessor coming from a
different partition. Since the edge is not marking with with EDGE_CROSSING
missing across section boundary. Note that even with EDGE_CROSSING information,
the generated code would be quite bad, with extra jumps coming from the
critical path.

The problem is explained with the start of a resolution here
http://gcc.gnu.org/ml/gcc-patches/2012-09/msg00794.html

2) Seem to miss redirecting CROSS_JUMPING jumps. When we have 

jump_insn 205 204 179 4 (set (pc)
        (reg:SI 1 r1 [237])) 316 {indirect_jump_compact}
     (expr_list:REG_CROSSING_JUMP (nil)
        (nil))

we have a  JUMP_LABEL (insn) containing the setting or r1 with the label_rtx,
but 'patch_jump_insn' fails to see it, although it looks like the branch could
be redirected.

[Bug target/54546] SH: Enable -fshrink-wrap

[chrbr at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54546

--- Comment #2 from chrbr at gcc dot gnu.org 2012-09-14 13:22:34 UTC ---
Author: chrbr
Date: Fri Sep 14 13:22:29 2012
New Revision: 191301

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=191301
Log:
       PR target/54546
       * config/sh/sh-protos.h (sh_need_epilogue): Delete.
       (sh_can_use_simple_return_p): Declare.
       * config/sh/sh.c (sh_can_use_simple_return_p): Define.
       (sh_need_epilogue, sh_need_epilogue_known): Delete.
       (sh_output_function_epilogue): Remove sh_need_epilogue_known.
       * config/sh/sh.md (simple_return, return): Define.
       (epilogue): Use inline return rtl.
       (sh_expand_epilogue): Cleanup parameters boolean type.
       * config/sh/iterators.md (any_return): New iterator.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/sh/iterators.md
    trunk/gcc/config/sh/sh-protos.h
    trunk/gcc/config/sh/sh.c
    trunk/gcc/config/sh/sh.md

[Bug target/54546] SH: Enable -fshrink-wrap

[olegendo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54546

--- Comment #3 from Oleg Endo <olegendo at gcc dot gnu.org> 2012-11-01 00:03:11 UTC ---
Even though simple_return is not supported on SHmedia, is it OK to close this
PR?

[Bug target/54546] SH: Enable -fshrink-wrap

[chrbr at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54546

--- Comment #4 from chrbr at gcc dot gnu.org 2012-11-07 13:23:53 UTC ---
There is still the reorder-branch-and-partition issue pending...

(In reply to comment #3)
> Even though simple_return is not supported on SHmedia, is it OK to close this
> PR?

[Bug target/54546] SH: Enable -fshrink-wrap

[chrbr at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54546

chrbr at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
            Version|4.7.0                       |4.8.0
         Resolution|                            |FIXED

--- Comment #5 from chrbr at gcc dot gnu.org 2012-11-09 14:40:15 UTC ---
Remaining partitioning issue fixed with
http://gcc.gnu.org/ml/gcc-patches/2012-11/msg00630.html. Fixed in 4.8