Some of the maths builtins can expand to a call followed by a bit
of postprocessing. With 4.8's PARALLEL return optimisations, these
embedded calls might return a PARALLEL of pseudos, but the postprocessing
isn't prepared to deal with that. This leads to an ICE in builtins-53.c
on n32 and n64 mip
Fixes a warning seen with -mabi=64 on mips64-linux-gnu. The test was
for an rtl optimisation, so the choice between "int" and "long" doesn't
matter for the original ILP32 target.
Tested on mips64-linux-gnu and applied.
Richard
gcc/testsuite/
* gcc.target/mips/pr55315.c: Cast to long ra
The loop in this test is supposed to trigger a likely (annulled)
back-branch. However, "n" is naturally in $4 on entry to the loop
and naturally in $2 on the back edge. Usually the move from $4
to $2 was outside the loop, but for -funroll-loops it was at the
end of the loop body, and could be saf
Please ignore that patch. I found a much simpler way, which is a tiny
patch and actually also works for the commented part of the test case of
the previous patch (i.e. the missed-optimization bit).
Namely, I now simply only tag all the symbols which are written to the
.mod file. That way, ther
Le 23/12/2012 15:29, Tobias Burnus a écrit :
Please ignore that patch. I found a much simpler way, which is a tiny
patch and actually also works for the commented part of the test case of
the previous patch (i.e. the missed-optimization bit).
Namely, I now simply only tag all the symbols which a
The minimum uint128_t value is an all-zeros double_int and the maximum
value is an all-ones double_int. Truncating these values to int and
then sign-extending them gives the same all-zeros and all-ones values,
so tree-vrp.c:range_fits_type_p concludes that uint128_t fits in an int.
This showed up
Some of the maths builtins try to expand via an optab and fall back
to an expand_call. The optabs path tends to clobber "target",
so the original target is lost by the time we call expand_call. E.g.:
/* Compute into TARGET.
Set TARGET to wherever the result comes back. */
target = expa
Mikael Morin wrote:
Le 23/12/2012 15:29, Tobias Burnus a écrit :
Namely, I now simply only tag all the symbols which are written to the
.mod file. That way, there should be never a symbol which outside used
(no wrong-code issues).
Yeah, I thought about that too. But then, I was worried to hav
URL: http://gcc.gnu.org/PR32193
Signed-off-by: Mike Frysinger
2012-12-23 Mike Frysinger
PR libgomp/32193
* configure.ac: Call AC_ARG_ENABLE(werror).
(XCFLAGS): Add -Werror when enable_werror is not no.
* configure: Regenerated.
---
libgomp/configure| 16 ++
URL: http://gcc.gnu.org/PR53679
Signed-off-by: Mike Frysinger
---
libgo/configure| 15 ---
libgo/configure.ac | 7 +--
2 files changed, 17 insertions(+), 5 deletions(-)
diff --git a/libgo/configure b/libgo/configure
index 04fa89d..49cc4a9 100755
--- a/libgo/configure
+++ b/l
We use this in the sim tree currently. Rather than require people to
have pkg-config installed, include it in the config/ dir.
Signed-off-by: Mike Frysinger
2012-12-23 Mike Frysinger
* pkg.m4: New file from pkg-config-0.27.
---
config/pkg.m4 | 199 ++
On 12/22/2012 06:02 PM, Paolo Carlini wrote:
Well, we still reject it after the patch
My 4.7 and trunk compilers both accept it (with -std=c++11, of course).
Jason
On 12/21/2012 07:35 AM, Dodji Seketeli wrote:
else if (TREE_TYPE (t)
&& INTEGRAL_OR_ENUMERATION_TYPE_P (TREE_TYPE (t))
- && !TREE_CONSTANT (t))
+ && !TREE_CONSTANT (t)
+ /* Class template and alias template arguments should be OK. */
+ && !DECL_
On Sun, Dec 23, 2012 at 11:04 PM, Jason Merrill wrote:
> On 12/21/2012 07:35 AM, Dodji Seketeli wrote:
>>
>> else if (TREE_TYPE (t)
>>&& INTEGRAL_OR_ENUMERATION_TYPE_P (TREE_TYPE (t))
>> - && !TREE_CONSTANT (t))
>> + && !TREE_CONSTANT (t)
>> + /* Class te
On 12/22/2012 11:03 AM, Dodji Seketeli wrote:
[1]: The relacement of the VAR_DECL by its initializer is done by
decay_conversion by callig decl_constant_value_safe. That replacement
doesn't happen if processing_template_decl is not set. That's why it
doesn't happen for the class template instan
15 matches
Mail list logo
