With the 2013-09-21 version of trunk, attached test case results in bus
error when simultaneously enabling both -mno-unaligned-access and
-ftree-vectorize.
The error is caused by unaligned vector load via two vldr instructions:
vldr d16, [r3, #-16]
vldr d17, [r3, #-8]
Cause for this error is, e
On Sun, 2 Feb 2014, Joseph S. Myers wrote:
> On Sun, 2 Feb 2014, Richard Sandiford wrote:
>
> > This patch (finally!) moves MIPS over to the soft-fp routines. The main
> > advantage is that we now handle 128-bit long-double exceptions properly
> > on hard-float targets.
>
> Thanks for doing thi
Hi,
this patch fixes the bug in extr_type_from_vtbl_ptr_store that made it to
consider store of construction virtual table or virtual table of virtual base
as store of type's virtual table.
In the testcase we have after early inlining:
virtual C::~C() (struct C * const this)
{
unsigned int i;
The initial commit of the Nios II GCC backend has an ABI conformance bug
in that it initializes the GOT pointer from _GLOBAL_OFFSET_TABLE_ rather
than _gp_got, as specified in the ABI. As a consequence of this,
GCC-generated code can only address 32K of GOT rather than 64K as
documented in the
Le 02/02/2014 21:31, Janus Weil a écrit :
> Hi all,
>
> the attached patch fixes an OOP-related 4.7-only regression. Pretty
> much straightforward, for details see PR.
>
> Regtested on x86_64-unknown-linux-gnu. Ok for 4.7?
>
Yes, thanks.
Mikael
Hi all,
the attached patch fixes an OOP-related 4.7-only regression. Pretty
much straightforward, for details see PR.
Regtested on x86_64-unknown-linux-gnu. Ok for 4.7?
Cheers,
Janus
2014-02-02 Janus Weil
PR fortran/59941
* expr.c (replace_comp): Check for isym to avoid ICE.
2014
Hello!
Attached patch backports 'gcc_update explicit use of "svn"' patch [1]
to 4.7 branch.
Without this patch, contrib/gcc_update command produces corrupted
gcc/REVISION on my F20 system.
2014-02-02 Uros Bizjak
Backport from mainline
2012-05-09 David Edelsohn
* g
>
> basic block 13 only contains
>
> ABNORMAL_DISPATCHER (0);
>
> which is an internal function and will be expanded to empty block:
>
> ;; Generating RTL for gimple basic block 13
>
> ;; ABNORMAL_DISPATCHER (0);
>
> (nil)
>
> compute_bb_predicates goes into an infinite loop due to
OK, h
On Sun, 2 Feb 2014, Marc Glisse wrote:
> An alternative could be to have a helper function that does nothing in C and
> calls default_conversion in C++. Or make the C version of default_conversion
> return its argument unchanged instead of asserting when it sees an unexpected
> tree.
Well, in pri
On Sun, 2 Feb 2014, Richard Sandiford wrote:
> This patch (finally!) moves MIPS over to the soft-fp routines. The main
> advantage is that we now handle 128-bit long-double exceptions properly
> on hard-float targets.
Thanks for doing this. At some point I intend to add soft-fp support for
aft
Hi all,
This is a simple patch which resolves
testsuite/gcc.target/aarch64/aapcs64/ret-func-1.c regression.
Basically, no special operations are needed for vector type i32in128 or
f32in64 in big-endian mode any more due the aarch64 back-end change.
Okay for trunk?
Kind regards,
Renlin Li
d
On Sun, 2 Feb 2014, Joseph S. Myers wrote:
On Sat, 1 Feb 2014, Marc Glisse wrote:
Ping
http://gcc.gnu.org/ml/gcc-patches/2014-01/msg01168.html
As I understand it, this is only relevant to C++ (in C you should have an
INTEGER_CST here, and if you don't then default_conversion won't give you
o
On Sat, 1 Feb 2014, Marc Glisse wrote:
> Ping
> http://gcc.gnu.org/ml/gcc-patches/2014-01/msg01168.html
As I understand it, this is only relevant to C++ (in C you should have an
INTEGER_CST here, and if you don't then default_conversion won't give you
one), so would best be reviewed by Jason no
Hi,
Before r207231, we generate:
;; basic block 12, loop depth 0
;; Invalid sum of incoming frequencies 8210, should be 4809
;;pred: 4
;;13
;;5
;;6
;;7
;;10
;;11
_33 = _setjmp (&float_e
On Sun, Feb 2, 2014 at 5:35 PM, Jakub Jelinek wrote:
> On Sun, Feb 02, 2014 at 04:31:35PM +0100, Uros Bizjak wrote:
>> There is off-by-one error in classify_argument, when processing
>> integer atomic types. When bit_offset is 64, two X86_64_INTEGER CLASS
>> registers should be used.
>>
>> Attache
On Sun, Feb 02, 2014 at 04:31:35PM +0100, Uros Bizjak wrote:
> There is off-by-one error in classify_argument, when processing
> integer atomic types. When bit_offset is 64, two X86_64_INTEGER CLASS
> registers should be used.
>
> Attached patch fixes this off-by-one error, so we won't use size =
Hi,
since we hit can of worms here, I decided to decompose the changes into minimal
patches.
This is first one fixing small bug introduced last July in Martin's change to
add
a flags to passthrough about the type preservation. This does not affect
gcc-4.8
Bootstrapped/regtested x86_64-linux, w
This patch (finally!) moves MIPS over to the soft-fp routines. The main
advantage is that we now handle 128-bit long-double exceptions properly
on hard-float targets.
This also removes the last use of TPBIT in libgcc. It might be worth
removing support for that at some point -- including the ass
This patch implements TARGET_ATOMIC_ASSIGN_EXPAND_FENV for MIPS,
along the lines of the SSE handling.
The #defines aren't very elegant. I might try to do something better
for 4.10, such as add a .def file.
Tested on mips64-linux-gnu and mipsisa64-sde-elf. It fixes the
c11-atomic-exec-5.c failur
This patch adds built-in functions for getting and setting the value of
the FPU's FCSR. Unlike any of the existing built-in functions, these two
can be used in both MIPS16 and non-MIPS16 code.
The functions are needed for the two patches that I'm about to post,
which fix c11-atomic-exec-5.c for m
This patch generalises the code used to handle __mips16_rdhwr stubs so
that it can be reused in the next patch. I also threw in a couple of
#undefs for macros in surrounding code.
Tested on mips64-linux-gnu and mipsisa64-sde-elf. Applied.
Thanks,
Richard
gcc/
* config/mips/mips.c (mip
Hello!
There is off-by-one error in classify_argument, when processing
integer atomic types. When bit_offset is 64, two X86_64_INTEGER CLASS
registers should be used.
Attached patch fixes this off-by-one error, so we won't use size = 0
when bit_offset = 64.
2014-02-02 Uros Bizjak
PR targ
22 matches
Mail list logo
