[Bug tree-optimization/32544] gcc 4.2.1 miscompiles Mesa's r300 DRI driver with -ftree-vrp

[pinskia at gcc dot gnu dot org]

--- Comment #9 from pinskia at gcc dot gnu dot org  2007-08-08 09:07 ---
This should work on the trunk also via the pointer plus branch merge where we
no longer get a multiplication in a pointer type.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32544

[Bug tree-optimization/32544] gcc 4.2.1 miscompiles Mesa's r300 DRI driver with -ftree-vrp

[pixel at mandriva dot com]

--- Comment #8 from pixel at mandriva dot com  2007-08-08 08:45 ---
Known to work: 4.1.2 4.3.0
Known to fail: 4.2.1


-- 

pixel at mandriva dot com changed:

   What|Removed |Added

 CC||pixel at mandriva dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32544

[Bug tree-optimization/32544] gcc 4.2.1 miscompiles Mesa's r300 DRI driver with -ftree-vrp

[pixel at mandriva dot com]

--- Comment #7 from pixel at mandriva dot com  2007-08-07 22:45 ---
the following patch fixes the testcase ivopts-issue.c, but breaks everything
else... someone else will find a correct patch :)

--- gcc-4.2.1/gcc/tree-ssa-loop-ivopts.c.pix2006-10-06 21:32:04.0
+0200
+++ gcc-4.2.1/gcc/tree-ssa-loop-ivopts.c2007-08-08 00:23:49.0
+0200
@@ -3061,7 +3061,9 @@ get_computation_aff (struct loop *loop,
   expr = fold_convert (ctype, expr);
 }

-  if (TYPE_UNSIGNED (utype))
+  if (POINTER_TYPE_P (utype))
+uutype = ctype;
+  else if (TYPE_UNSIGNED (utype))
 uutype = utype;
   else
 {


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32544

[Bug tree-optimization/32544] gcc 4.2.1 miscompiles Mesa's r300 DRI driver with -ftree-vrp

[pixel at mandriva dot com]

--- Comment #6 from pixel at mandriva dot com  2007-08-07 22:41 ---
more info: the tree-vrp issue is raised by ivopts pass:

without ivopts:

  D.1541_13 = 3 - i_31;
  D.1542_14 = (unsigned int) D.1541_13;
  D.1543_15 = D.1542_14 * 4;
  D.1544_16 = (int *) D.1543_15;

with ivopts:

  D.1607_62 = (int *) i_31;
  D.1608_63 = D.1607_62 * -4B;
  D.1544_64 = D.1608_63 + 12B;
  D.1544_16 = D.1544_64;

  vrp finds that D.1608_63 is 0, and so replaces D.1544_64 with 12

  D.1607_62 = (int *) i_31;
  D.1608_63 = D.1607_62 * -4B;
  D.1544_64 = D.1608_63 + 12B;
  D.1544_16 = D.1544_64;


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32544

[Bug tree-optimization/32544] gcc 4.2.1 miscompiles Mesa's r300 DRI driver with -ftree-vrp

[pixel at mandriva dot com]

--- Comment #5 from pixel at mandriva dot com  2007-08-07 22:33 ---
Created an attachment (id=14038)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14038&action=view)
miscompiles with gcc 4.2 -O2

here is a simple test case reproducing the issue.

eg:

% gcc -O2 ivopts-issue.c && ./a.out && gcc -O2 -fno-ivopts ivopts-issue.c &&
./a.out 
0 0 1 2 
2 2 1 1 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32544

[Bug tree-optimization/32544] gcc 4.2.1 miscompiles Mesa's r300 DRI driver with -ftree-vrp

[rguenth at gcc dot gnu dot org]

--- Comment #4 from rguenth at gcc dot gnu dot org  2007-07-04 09:49 ---
Sorry, I can't see what is wrong.  There is no effective difference in assembly
with/without -ftree-vrp.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32544

[Bug tree-optimization/32544] gcc 4.2.1 miscompiles Mesa's r300 DRI driver with -ftree-vrp

[bero at arklinux dot org]

--- Comment #3 from bero at arklinux dot org  2007-06-29 14:17 ---
r300SetupPixelShader()


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32544

[Bug tree-optimization/32544] gcc 4.2.1 miscompiles Mesa's r300 DRI driver with -ftree-vrp

[rguenth at gcc dot gnu dot org]

--- Comment #2 from rguenth at gcc dot gnu dot org  2007-06-29 08:46 ---
Can you narrow it down to a specific function?


-- 

rguenth at gcc dot gnu dot org changed:

   What|Removed |Added

 CC||rguenth at gcc dot gnu dot
   ||org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32544

[Bug tree-optimization/32544] gcc 4.2.1 miscompiles Mesa's r300 DRI driver with -ftree-vrp

[bero at arklinux dot org]

--- Comment #1 from bero at arklinux dot org  2007-06-29 08:16 ---
Created an attachment (id=13804)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=13804&action=view)
Preprocessed source, bzip2-ed

Attaching preprocessed source of r300_state.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32544