[Bug tree-optimization/85794] [8/9 Regression][AArch64] ICE in expand_vector_condition in GIMPLE pass: veclower2

2019-01-08 Thread jakub at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85794

--- Comment #6 from Jakub Jelinek  ---
Author: jakub
Date: Tue Jan  8 09:57:10 2019
New Revision: 267687

URL: https://gcc.gnu.org/viewcvs?rev=267687=gcc=rev
Log:
Backported from mainline
2018-11-22  Jakub Jelinek  

PR tree-optimization/85794
* gcc.dg/vect/O3-pr85794.c: New test.

Added:
branches/gcc-8-branch/gcc/testsuite/gcc.dg/vect/O3-pr85794.c
Modified:
branches/gcc-8-branch/gcc/testsuite/ChangeLog

[Bug tree-optimization/85794] [8/9 Regression][AArch64] ICE in expand_vector_condition in GIMPLE pass: veclower2

2018-11-22 Thread jakub at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85794

Jakub Jelinek  changed:

   What|Removed |Added

 Status|NEW |RESOLVED
 CC||jakub at gcc dot gnu.org
 Resolution|--- |FIXED
   Target Milestone|8.3 |8.2

--- Comment #5 from Jakub Jelinek  ---
Fixed by the PR85793 fix, r260317 on the trunk and r260636 backport thereof to
8.x branch.

[Bug tree-optimization/85794] [8/9 Regression][AArch64] ICE in expand_vector_condition in GIMPLE pass: veclower2

2018-11-22 Thread jakub at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85794

--- Comment #4 from Jakub Jelinek  ---
Author: jakub
Date: Thu Nov 22 19:58:50 2018
New Revision: 266387

URL: https://gcc.gnu.org/viewcvs?rev=266387=gcc=rev
Log:
PR tree-optimization/85794
* gcc.dg/vect/O3-pr85794.c: New test.

Added:
trunk/gcc/testsuite/gcc.dg/vect/O3-pr85794.c
Modified:
trunk/gcc/testsuite/ChangeLog

[Bug tree-optimization/85794] [8/9 Regression][AArch64] ICE in expand_vector_condition in GIMPLE pass: veclower2

2018-07-26 Thread jakub at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85794

Jakub Jelinek  changed:

   What|Removed |Added

   Target Milestone|8.2 |8.3

--- Comment #3 from Jakub Jelinek  ---
GCC 8.2 has been released.

[Bug tree-optimization/85794] [8/9 Regression][AArch64] ICE in expand_vector_condition in GIMPLE pass: veclower2

2018-05-16 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85794

--- Comment #2 from Richard Biener  ---
Invalid GIMPLE.  We have

vect_patt_18.16_33 = VEC_COND_EXPR <_24 != vect_cst__23, vect_cst__31,
vect_cst__32>;

with

 
unit-size 
align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0x76a7f3f0 precision:64 min  max >
BLK size  unit-size 
align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0x76a7f5e8 nunits:1>

arg:0  unit-size 
...
arg:1 

the GIMPLE verifier should also catch this.  Testing

Index: gcc/tree-cfg.c
===
--- gcc/tree-cfg.c  (revision 260280)
+++ gcc/tree-cfg.c  (working copy)
@@ -4137,6 +4137,12 @@ verify_gimple_assign_ternary (gassign *s
}
   /* Fallthrough.  */
 case COND_EXPR:
+  if (!is_gimple_val (rhs1)
+ && verify_gimple_comparison (TREE_TYPE (rhs1),
+  TREE_OPERAND (rhs1, 0),
+  TREE_OPERAND (rhs1, 1),
+  TREE_CODE (rhs1)))
+   return true;
   if (!useless_type_conversion_p (lhs_type, rhs2_type)
  || !useless_type_conversion_p (lhs_type, rhs3_type))
{

which will result in

t.c: In function ‘fn1’:
t.c:3:6: error: mismatching comparison operand types
 void fn1() {
  ^~~
long unsigned int
vector(1) long unsigned int
vect_patt_19.16_32 = VEC_COND_EXPR <_9 != vect_cst__28, vect_cst__12,
vect_cst__21>;
t.c:3:6: error: mismatching comparison operand types
long unsigned int
vector(1) long unsigned int
vect_patt_19.16_33 = VEC_COND_EXPR <_64 != vect_cst__28, vect_cst__12,
vect_cst__21>;
during GIMPLE pass: vect
t.c:3:6: internal compiler error: verify_gimple failed

[Bug tree-optimization/85794] [8/9 Regression][AArch64] ICE in expand_vector_condition in GIMPLE pass: veclower2

2018-05-15 Thread ktkachov at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85794

ktkachov at gcc dot gnu.org changed:

   What|Removed |Added

 Status|UNCONFIRMED |NEW
   Last reconfirmed||2018-05-15
 CC||ktkachov at gcc dot gnu.org
   Target Milestone|--- |8.2
 Ever confirmed|0   |1

--- Comment #1 from ktkachov at gcc dot gnu.org ---
Confirmed.