--- Additional Comments From reichelt at gcc dot gnu dot org 2005-01-30
23:02 ---
Just for reference, this is now tracked in
http://bugs.kde.org/show_bug.cgi?id=97042
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18089
--- Additional Comments From jakub at gcc dot gnu dot org 2005-01-14 18:35
---
Are you sure this is not just a bug in valgrind?
I have verified that with current CVS I get
ERROR SUMMARY: 37 errors from 21 contexts (suppressed: 12 from 1)
while if I rebuild stage1's combine.o with
--- Additional Comments From jakub at gcc dot gnu dot org 2005-01-14 20:20
---
BTW, if I replace all
subl$-2147483648, %eax
instructions in real.s by:
xorl$-2147483648, %eax
(note that all such instructions are followed either by shrl, or addl, so
any differences
--- Additional Comments From jakub at gcc dot gnu dot org 2005-01-14 20:36
---
To prove this is a valgrind bug and not GCC bug, I wrote a small self-contained
testcase on which valgrind complains:
#define EXP_BITS (32 - 5)
struct real_value
{
unsigned int cl : 2;
unsigned int sign :
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-01-14
21:45 ---
So closing as invalid, please report this to valgrind then.
--
What|Removed |Added
--- Additional Comments From reichelt at gcc dot gnu dot org 2005-01-10
09:33 ---
Just another data point:
Reverting the patch on current mainline makes the problem disappear.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18089
--- Additional Comments From reichelt at gcc dot gnu dot org 2005-01-08
21:06 ---
Roger, the problem appears with your patch for PR17151:
http://gcc.gnu.org/ml/gcc-cvs/2004-09/msg01231.html
It looks like this causes a miscompilation of the stage2 compiler.
--
What
--- Additional Comments From reichelt at gcc dot gnu dot org 2005-01-07
15:07 ---
I just stumbled over this one, too.
Even shorter testcase: valgrind --tool=memcheck cc1 -quiet vg.c
==
int foo() { return 1.0 + 1.1; }
==
