[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[bergner at gcc dot gnu dot org]

--- Comment #12 from bergner at gcc dot gnu dot org  2010-06-09 14:15 
---
Subject: Bug 44067

Author: bergner
Date: Wed Jun  9 14:15:11 2010
New Revision: 160479

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=160479
Log:
Backport from mainline:

2010-06-09  Edmar Wienskoski  

PR target/44067
* config/rs6000/rs6000.md (DIFD): Do not split dpfp values for
e500v2 target.

Modified:
branches/ibm/gcc-4_4-branch/gcc/ChangeLog.ibm
branches/ibm/gcc-4_4-branch/gcc/config/rs6000/rs6000.md


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[amodra at gmail dot com]

--- Comment #11 from amodra at gmail dot com  2010-06-09 00:29 ---
Fixed


-- 

amodra at gmail dot com changed:

   What|Removed |Added

 Status|ASSIGNED|RESOLVED
 Resolution||FIXED
   Target Milestone|--- |4.5.1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[amodra at gcc dot gnu dot org]

--- Comment #10 from amodra at gcc dot gnu dot org  2010-06-09 00:20 ---
Subject: Bug 44067

Author: amodra
Date: Wed Jun  9 00:20:27 2010
New Revision: 160454

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=160454
Log:
PR target/44067
* config/rs6000/rs6000.md (DIFD): Do not split dpfp values for
e500v2 target.


Modified:
branches/gcc-4_5-branch/gcc/ChangeLog
branches/gcc-4_5-branch/gcc/config/rs6000/rs6000.md


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[amodra at gcc dot gnu dot org]

--- Comment #9 from amodra at gcc dot gnu dot org  2010-06-09 00:16 ---
Subject: Bug 44067

Author: amodra
Date: Wed Jun  9 00:15:46 2010
New Revision: 160449

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=160449
Log:
PR target/44067
* config/rs6000/rs6000.md (DIFD): Do not split dpfp values for
e500v2 target.


Modified:
trunk/gcc/ChangeLog
trunk/gcc/config/rs6000/rs6000.md


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[amodra at gmail dot com]

--- Comment #8 from amodra at gmail dot com  2010-06-07 23:33 ---
Reassigning since Edmar's identical patch predates mine.


-- 

amodra at gmail dot com changed:

   What|Removed |Added

 AssignedTo|amodra at gmail dot com |edmar at freescale dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[bergner at gcc dot gnu dot org]

--- Comment #7 from bergner at gcc dot gnu dot org  2010-06-07 20:55 ---
I'll note that the thread I pointed to in the previous comment mentioned that I
thought we should be handling DDmode the same as DFmode (normally we do), but
Joseph's post here:

  http://gcc.gnu.org/ml/gcc-patches/2008-06/msg00270.html

shows why we don't want to do that for the e500.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[bergner at gcc dot gnu dot org]

--- Comment #6 from bergner at gcc dot gnu dot org  2010-06-07 20:50 ---
I believe Edmar already bootstrapped and regtested the same patch.

  http://gcc.gnu.org/ml/gcc-patches/2010-06/msg00346.html


-- 

bergner at gcc dot gnu dot org changed:

   What|Removed |Added

 CC||bergner at gcc dot gnu dot
   ||org
URL||http://gcc.gnu.org/ml/gcc-
   ||patches/2010-
   ||06/msg00346.html


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[amodra at gmail dot com]

--- Comment #5 from amodra at gmail dot com  2010-06-07 17:25 ---
Created an attachment (id=20859)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20859&action=view)
fix pr42427 fallout

Would someone with e500 hardware please bootstrap and regression test this
patch?  I'm running powerpc-linux bootstrap to ensure pr42427 doesn't regress.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[amodra at gmail dot com]

--- Comment #4 from amodra at gmail dot com  2010-06-07 06:57 ---
Actually, that's the wrong patch.  The correct one stops
rs6000_split_multireg_move being called in this case, by modifying
define_mode_iterator DIFD in rs6000.md.


-- 

amodra at gmail dot com changed:

   What|Removed |Added

 AssignedTo|unassigned at gcc dot gnu   |amodra at gmail dot com
   |dot org |
 Status|NEW |ASSIGNED
   Last reconfirmed|2010-06-07 02:20:00 |2010-06-07 06:57:49
   date||


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[gcc at breakpoint dot cc]

--- Comment #3 from gcc at breakpoint dot cc  2010-06-07 04:21 ---
Thank you Alan.
Trunk and 4.5 branch is affected, the 4.4 branch just built.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[amodra at gmail dot com]

--- Comment #2 from amodra at gmail dot com  2010-06-07 02:21 ---
*** Bug 44419 has been marked as a duplicate of this bug. ***


-- 

amodra at gmail dot com changed:

   What|Removed |Added

 CC||edmar at freescale dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067

[Bug target/44067] internal compiler error: in rs6000_split_multireg_move, at config/rs6000/rs6000.c:16713

[amodra at gmail dot com]

--- Comment #1 from amodra at gmail dot com  2010-06-07 02:19 ---
First testcase in pr44364 tickles this bug on mainline too.  Looks like we need
the following.

Index: gcc/config/rs6000/rs6000.c
===
--- gcc/config/rs6000/rs6000.c  (revision 160351)
+++ gcc/config/rs6000/rs6000.c  (working copy)
@@ -17013,7 +17013,7 @@ rs6000_split_multireg_move (rtx dst, rtx
((TARGET_HARD_FLOAT && TARGET_DOUBLE_FLOAT) ? DFmode : SFmode);
   else if (ALTIVEC_REGNO_P (reg))
 reg_mode = V16QImode;
-  else if (TARGET_E500_DOUBLE && mode == TFmode)
+  else if (TARGET_E500_DOUBLE && (mode == DFmode || mode == TFmode))
 reg_mode = DFmode;
   else
 reg_mode = word_mode;


-- 

amodra at gmail dot com changed:

   What|Removed |Added

 Status|UNCONFIRMED |NEW
 Ever Confirmed|0   |1
   Last reconfirmed|-00-00 00:00:00 |2010-06-07 02:20:00
   date||


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44067