Hi Jeff,
On Jun 17, 2014, at 22:42 , Jeff Law wrote:
>> * tree-core.h (tree_block): Add an "end_locus" field, allowing
>> memorization of the end of block source location.
>> * tree.h (BLOCK_SOURCE_END_LOCATION): New accessor.
>> * gimplify.c (gimplify_bind_expr): Propagate t
This makes sure we run cfg-cleanup when we propagate into PHI nodes
or on the FRE/PRE side remove any stmt. Otherwise we can end up
with not removed forwarder blocks.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2014-06-18 Richard Biener
* tree-ssa-propaga
On 18/06/14 06:19, Richard Henderson wrote:
> Trivial fix for missing clobber of the flags over the tlsdesc call.
>
> Ok for all branches?
>
>
OK.
R.
> r~
>
> * config/aarch64/aarch64.md (tlsdesc_small_): Clobber CC_REGNUM.
>
>
> z
>
>
> diff --git a/gcc/config/aarch64/aarch64.md b
On 06/17/2014 10:14 PM, David Malcolm wrote:
On Fri, 2014-06-13 at 12:24 +0200, mliska wrote:
[...snip...]
Statistics about the pass:
Inkscape: 11.95 MB -> 11.44 MB (-4.27%)
Firefox: 70.12 MB -> 70.12 MB (-3.07%)
FWIW, you wrote 70.12 MB here for both before and after for Firefox, but
On Jun 18, 2014, at 09:42 , Olivier Hainque wrote:
>> I assume y'all will add a suitable test to the Ada testsuite and propagate
>> it into the GCC testsuite in due course?
> ISTM that dg-scan-asm for the expected extra .loc's would work, maybe
> restricted to some target we know produces .loc
Hi,
With LTO, -fno-short-enums is ignored, resulting in ABI mis-matching in
linking.
Refer https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61123 for details.
This patch add fshort-enums and fshout-wchar to LTO group.
To check it, a new procedure object-readelf is added in
testsuite/lib/lto.exp a
Hi,
I was quite surprised we still had an assert checking that the target
of a virtual call derived by ipa-cp is among possible ones derived by
ipa-devirt. This is not true for various down-casts and I managed to
trigger it in PR 61540 (where the testcase purposefully invokes
undefined behavior b
This removes the special dce_loop pass in favor of dealing with
scev and niter estimates in dce generally. Likewise it makes
copyprop always cleanup after itself, dealing with scev and niter
estimates. It also makes copyprop not unconditionally schedule
a cfg-cleanup but only do so if copyprop d
On 10/06/14 12:42, Terry Guo wrote:
> Hi There,
>
> The thumb1_reorg function use macro INSN_CODE to find expected instructions.
> But the macro INSN_CODE doesn’t work for label type instruction. The
> INSN_CODE(label_insn) will return the label number. When we have a lot of
> labels and current l
On Tue, 17 Jun 2014, Jeff Law wrote:
> On 06/17/14 07:07, Richard Biener wrote:
> >
> > I felt that -ftree-XXX is bad naming so I went for -fssa-XXX
> > even if that is now inconsistent. Any optinion here? For
> > RTL we simply have unsuffixed names so shall we instead go
> > for -fphiopt? PHI
On 06/17/2014 10:09 PM, Paolo Carlini wrote:
Hi,
On 13/06/14 12:24, mliska wrote:
The optimization is inspired by Microsoft /OPT:ICF optimization
(http://msdn.microsoft.com/en-us/library/bxwfs976.aspx) that merges COMDAT
sections with each function reside in a separate section.
In terms
Hi,
On 18/06/14 10:46, Martin Liška wrote:
As one would expect, there is a function 'size'.
Cool, thanks!
Paolo.
On 06/17/2014 10:50 PM, Rainer Orth wrote:
Jeff Law writes:
On 06/13/14 04:48, mliska wrote:
Hi,
many tests rely on a precise number of scanned functions in a dump file. If
IPA ICF decides to merge some function and(or) read-only variables, counts do
not match.
Martin
Changelog:
201
On Mon, Jun 09, 2014 at 01:43:48PM +0200, Uros Bizjak wrote:
> On Mon, Jun 9, 2014 at 1:34 PM, Kirill Yukhin wrote:
> > Hello Uroš,
> > On 08 Jun 11:26, Uros Bizjak wrote:
> >> On Tue, May 27, 2014 at 12:28 PM, Petr Murzin
> >> wrote:
> >> > Hi,
> >> > I've fixed tests for AVX512, so they could
> -Original Message-
> From: Richard Earnshaw
> Sent: Wednesday, June 18, 2014 4:31 PM
> To: Terry Guo
> Cc: gcc-patches@gcc.gnu.org; Ramana Radhakrishnan
> Subject: Re: [Patch, GCC/Thumb-1]Mishandle the label type insn in function
> thumb1_reorg
>
> On 10/06/14 12:42, Terry Guo wrote:
>
On 18/06/14 10:16, Terry Guo wrote:
>
>
>> -Original Message-
>> From: Richard Earnshaw
>> Sent: Wednesday, June 18, 2014 4:31 PM
>> To: Terry Guo
>> Cc: gcc-patches@gcc.gnu.org; Ramana Radhakrishnan
>> Subject: Re: [Patch, GCC/Thumb-1]Mishandle the label type insn in function
>> thumb1_r
PR61536 is a case where linking fails on arm-linux-gnueabi* and
arm-eabi* systems as the C++ ABI for ARM specifies out of line
comparison operators for typeinfo.
Rev r211355 tightened the symbols exported by libstdc++ a bit too much
which caused some carnage in the test results for arm-linux-g
Hi!
I've committed following change:
--- projects/gomp/index.html25 Oct 2013 07:16:35 - 1.13
+++ projects/gomp/index.html18 Jun 2014 09:48:34 -
@@ -63,9 +63,19 @@ available.
Status
+Jun 18, 2014
+The last major part of Fortran OpenMP v4.0 support has been
+committed into
I just figured that we create &TARGET_MEM_REF [base: a_4, offset: 0]
from within IVOPTs. That pessimizes further passes unnecessarily.
Bootstrap and regtest running on x86_64-unknown-linux-gnu.
Richard.
2014-06-18 Richard Biener
* tree-ssa-address.c (create_mem_ref_raw): Use proper
On Mon, Jun 16, 2014 at 1:53 PM, Venkataramanan Kumar
wrote:
> Hi Maintainers,
>
> This patch fixes the PR 60617 that occurs when we turn on reload pass
> in thumb2 mode.
>
> It occurs for the pattern "*ior_scc_scc" that gets generated for the 3
> argument of the below function call.
>
> JIT:emitS
On Mon, Jun 2, 2014 at 5:47 PM, Charles Baylis
wrote:
> This patch adds support for post-indexed addressing for NEON structure
> memory accesses.
>
> For example VLD1.8 {d0}, [r0], r1
>
>
> Bootstrapped and checked on arm-unknown-gnueabihf using Qemu.
>
> Ok for trunk?
This is OK.
Ramana
>
>
> g
On Wed, Jun 18, 2014 at 11:56:01AM +0200, Richard Biener wrote:
>
> I just figured that we create &TARGET_MEM_REF [base: a_4, offset: 0]
> from within IVOPTs. That pessimizes further passes unnecessarily.
>
> Bootstrap and regtest running on x86_64-unknown-linux-gnu.
Isn't that against the comm
On Tue, Jun 17, 2014 at 4:03 PM, Charles Baylis
wrote:
> On 5 June 2014 07:27, Ramana Radhakrishnan wrote:
>> On Mon, Jun 2, 2014 at 5:47 PM, Charles Baylis
>> wrote:
>>> This patch adds support for post-indexed addressing for NEON structure
>>> memory accesses.
>>>
>>> For example VLD1.8 {d0},
On Wed, Jun 18, 2014 at 3:30 AM, Thomas Preud'homme
wrote:
> Hi everybody,
>
> Thanks to a comment from Richard Biener, the bswap pass take care to not
> perform its optimization is memory is modified between the load of the
> original expression. However, when it replaces these statements by a
On Wed, Jun 18, 2014 at 6:55 AM, Thomas Preud'homme
wrote:
>> From: Richard Biener [mailto:richard.guent...@gmail.com]
>> Sent: Wednesday, June 11, 2014 4:32 PM
>> >
>> >
>> > Is this OK for trunk? Does this bug qualify for a backport patch to
>> > 4.8 and 4.9 branches?
>>
>> This is ok for trunk
On Wed, Jun 18, 2014 at 10:14 AM, Hale Wang wrote:
> Hi,
>
> With LTO, -fno-short-enums is ignored, resulting in ABI mis-matching in
> linking.
>
> Refer https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61123 for details.
>
> This patch add fshort-enums and fshout-wchar to LTO group.
>
> To check it,
On Wed, Jun 18, 2014 at 12:21 PM, Richard Biener
wrote:
> On Wed, Jun 18, 2014 at 10:14 AM, Hale Wang wrote:
>> Hi,
>>
>> With LTO, -fno-short-enums is ignored, resulting in ABI mis-matching in
>> linking.
>>
>> Refer https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61123 for details.
>>
>> This pat
Hello!
Attached patch fixes recently added sibcall insns and their
corresponding peephole2 patterns:
- There is no need for new memory_nox32_operand. A generic
memory_operand can be used, since new insns and peephole2 patterns
should be disabled for TARGET_X32 entirely.
- Adds missing "m" constra
On Wed, 18 Jun 2014, Jakub Jelinek wrote:
> On Wed, Jun 18, 2014 at 11:56:01AM +0200, Richard Biener wrote:
> >
> > I just figured that we create &TARGET_MEM_REF [base: a_4, offset: 0]
> > from within IVOPTs. That pessimizes further passes unnecessarily.
> >
> > Bootstrap and regtest running on
Hi Ramana,
On 18 June 2014 15:29, Ramana Radhakrishnan wrote:
> On Mon, Jun 16, 2014 at 1:53 PM, Venkataramanan Kumar
> wrote:
>> Hi Maintainers,
>>
>> This patch fixes the PR 60617 that occurs when we turn on reload pass
>> in thumb2 mode.
>>
>> It occurs for the pattern "*ior_scc_scc" that get
Jeff Law writes:
> Let's do better this time ;-) Add a testcase for the m68k port which
> verifies we're getting the desired code.
Make sense. Installed with the following test case.
Andreas.
PR rtl-optimization/54555
* gcc.target/m68k/pr54555.c: New test.
diff --git a/gcc/t
The following aims at reducing the number of pointless passes we run
on functions containing no loops. Those are at least two copyprop
and one dce pass (two dce passes when vectorization is enabled,
three dce passes and an additional copyprop pass when any graphite
optimization is enabled).
Simp
Hi All,
Here is a fix for PR 61518 - additional test was added to reject
transformation if reduction variable is not used in reduction
statement only since such reduction will not be vectorized.
Bootstrap and regression testing did not show any new failures.
Is it OK for trunk?
gcc/ChangeLog
2
On Wed, Jun 18, 2014 at 12:42:19PM +0200, Richard Biener wrote:
> Any comments? Does such followup sound realistic or would it be
> better to take the opportunity to move IVOPTs a bit closer to
> RTL expansion and avoid that "pass_tree_no_loop hack"?
I think it is fine to have pass_tree_no_loop p
Hi all,
All other #ifdefs in arm_neon.h that look for fast math use the
__FAST_MATH form rather than __FAST_MATH__. They have the same effect
AFAICS.
This patch fixes the one sticking out.
Committed as obvious with r211779.
Thanks,
Kyrill
2014-06-18 Kyrylo Tkachov
* config/arm/arm_n
On Wed, Jun 18, 2014 at 12:47 PM, Yuri Rumyantsev wrote:
> Hi All,
>
> Here is a fix for PR 61518 - additional test was added to reject
> transformation if reduction variable is not used in reduction
> statement only since such reduction will not be vectorized.
>
> Bootstrap and regression testing
On Wed, 18 Jun 2014, Richard Biener wrote:
>
> This removes the special dce_loop pass in favor of dealing with
> scev and niter estimates in dce generally. Likewise it makes
> copyprop always cleanup after itself, dealing with scev and niter
> estimates. It also makes copyprop not unconditional
2014-06-18 12:24 GMT+02:00 Uros Bizjak :
> Hello!
>
> Attached patch fixes recently added sibcall insns and their
> corresponding peephole2 patterns:
>
> - There is no need for new memory_nox32_operand. A generic
> memory_operand can be used, since new insns and peephole2 patterns
> should be disab
Put a check for capture index.
* genmatch.c (parse_capture): Add condition to check capture index.
(capture_max): New constant.
(stdlib.h): Include.
Thanks and Regards,
Prathamesh
Index: genmatch.c
=
On Wed, Jun 18, 2014 at 2:24 PM, Kai Tietz wrote:
> The following change in predicates.md seems to be a bit premature.
> There is still the point about Darwin's PIC issue for unspec-gotpcrel.
>
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61387
>
> return ANY_QI_REG_P (op);
> })
>
> +;; Retur
> The following change in predicates.md seems to be a bit premature.
> There is still the point about Darwin's PIC issue for unspec-gotpcrel.
The change is indeed incompatible with the patch in pr61387 comment 9.
And without it the failures are back!-(
Kai, what is wrong with Iain's patch in com
On 06/18/14 04:42, Richard Biener wrote:
The following aims at reducing the number of pointless passes we run
on functions containing no loops. Those are at least two copyprop
and one dce pass (two dce passes when vectorization is enabled,
three dce passes and an additional copyprop pass when a
2014-06-18 15:11 GMT+02:00 Dominique Dhumieres :
>
>> The following change in predicates.md seems to be a bit premature.
>> There is still the point about Darwin's PIC issue for unspec-gotpcrel.
>
> The change is indeed incompatible with the patch in pr61387 comment 9.
> And without it the failures
On 06/18/14 01:42, Olivier Hainque wrote:
Hi Jeff,
On Jun 17, 2014, at 22:42 , Jeff Law wrote:
* tree-core.h (tree_block): Add an "end_locus" field, allowing
memorization of the end of block source location.
* tree.h (BLOCK_SOURCE_END_LOCATION): New accessor.
*
On 18 June 2014 11:01, Ramana Radhakrishnan wrote:
> On Mon, Jun 2, 2014 at 5:47 PM, Charles Baylis
> wrote:
>> This patch adds support for post-indexed addressing for NEON structure
>> memory accesses.
>>
>> For example VLD1.8 {d0}, [r0], r1
>>
>>
>> Bootstrapped and checked on arm-unknown-gnuea
On Jun 18, 2014, at 15:48 , Jeff Law wrote:
>> ISTM that dg-scan-asm for the expected extra .loc's would work, maybe
>> restricted to some target we know produces .loc directives.
>>
>> Sounds appropriate ?
> Yea, that should be fine. Most folks test x86-64 linux, so that's going to
> get
On Tue, Jun 17, 2014 at 4:48 PM, Rainer Orth
wrote:
> "William J. Schmidt" writes:
>
>> Index: gcc/testsuite/gfortran.dg/default_format_denormal_2.f90
>> ===
>> --- gcc/testsuite/gfortran.dg/default_format_denormal_2.f90 (revision
On 11/06/14 11:19, Charles Baylis wrote:
> 2014-05-22 Charles Baylis
>
> * config/arm/bpabi.S (__aeabi_uldivmod): Optimise stack pointer
> manipulation.
OK.
R.
> ---
> libgcc/config/arm/bpabi.S | 54
> +--
> 1 file changed, 43 inserti
On 11/06/14 11:19, Charles Baylis wrote:
> 2014-05-22 Charles Baylis
>
> * config/arm/bpabi.S (__aeabi_uldivmod): Perform division using call
> to __udivmoddi4.
OK.
R.
> ---
> libgcc/config/arm/bpabi.S | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a
On 11/06/14 11:19, Charles Baylis wrote:
> 2014-05-22 Charles Baylis
>
> * config/arm/bpabi.S (__aeabi_ldivmod): Optimise stack manipulation.
OK.
R.
> ---
> libgcc/config/arm/bpabi.S | 14 +++---
> 1 file changed, 3 insertions(+), 11 deletions(-)
>
> diff --git a/libgcc/confi
On Tue, Jun 17, 2014 at 6:44 PM, BIll Schmidt
wrote:
> Hi,
>
> As described in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61542, a
> new test case (gcc.dg/vect/vect-nop-move.c) was added in 4.9. This
> exposes a bug on PowerPC little endian for extracting an element from a
> V4SF value that goe
On 11/06/14 11:19, Charles Baylis wrote:
> 2014-05-22 Charles Baylis
>
> * config/arm/bpabi.S (__aeabi_ldivmod): Perform division using
> __udivmoddi4, and fixups for negative operands.
OK.
> ---
> libgcc/config/arm/bpabi.S | 41 -
> 1 fil
On 11/06/14 11:19, Charles Baylis wrote:
> 2014-05-22 Charles Baylis
>
> * config/arm/bpabi.S (__aeabi_ldivmod, __aeabi_uldivmod,
> push_for_divide, pop_for_divide): Use .cfi_* directives for DWARF
> annotations. Fix DWARF information.
OK.
> ---
> libgcc/config/arm/bpabi.S
On 11/06/14 11:19, Charles Baylis wrote:
> 2014-05-22 Charles Baylis
>
> * config/arm/bpabi-v6m.S (__aeabi_uldivmod): Perform division using
> __udivmoddi4.
OK.
R.
> ---
> libgcc/config/arm/bpabi-v6m.S | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/lib
On 11/06/14 11:19, Charles Baylis wrote:
> 2014-05-22 Charles Baylis
>
> * config/arm/bpabi.c (__gnu_uldivmod_helper): Remove.
OK.
R.
> ---
> libgcc/config/arm/bpabi.c | 14 --
> 1 file changed, 14 deletions(-)
>
> diff --git a/libgcc/config/arm/bpabi.c b/libgcc/config/ar
On 17 Jun 21:22, Bernd Schmidt wrote:
> On 06/17/2014 08:20 PM, Ilya Verbin wrote:
> >I don't get this part of the plan. Where a host compiler will look for
> >mkoffloads?
> >
> >E.g., first I configure/make/install the target gcc and corresponding
> >mkoffload with the following options:
> >--e
On Sun, Jun 1, 2014 at 12:27 PM, Tom de Vries wrote:
> Richard,
>
> This patch:
> - adds the for TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS required
> clobbers in CALL_INSN_FUNCTION_USAGE,
> - sets TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS to true, which
> enables
> the fuse-caller-
Hi,
This improves the robustness of the aapcs64 test framework for testing
function return ABI rules. It ensures the test facility functions now
able to see the exact content of return registers right at the moment
when a function returns.
OK for trunk?
Thanks,
Yufeng
gcc/testsuite
A comment in rs6000.h says "cr5 is not supposed to be used". I checked
all ABIs, going as far back as PowerOpen (1994), and found no mention
of this.
Also document cr6 is used by some vector instructions.
Tested on powerpc64-linux, no regressions. Okay to apply?
Segher
2014-06-18 Segher Bo
On 18/06/14 15:16, Yufeng Zhang wrote:
> Hi,
>
> This improves the robustness of the aapcs64 test framework for testing
> function return ABI rules. It ensures the test facility functions now
> able to see the exact content of return registers right at the moment
> when a function returns.
>
On 18 June 2014 11:06, Ramana Radhakrishnan wrote:
>> 2014-06-15 Charles Baylis
>>
>> * config/arm/arm.c (arm_new_rtx_costs): Reduce cost for mem with
>> embedded side effects.
>
> I'm not too thrilled with putting in more special cases that are not
> table driven in there. Can
On 06/18/2014 04:13 PM, Ilya Verbin wrote:
On 17 Jun 21:22, Bernd Schmidt wrote:
On 06/17/2014 08:20 PM, Ilya Verbin wrote:
I don't get this part of the plan. Where a host compiler will look for
mkoffloads?
E.g., first I configure/make/install the target gcc and corresponding mkoffload
with
On 1 June 2014 11:00, Tom de Vries wrote:
> Richard,
>
> This patch:
> - adds the for TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS required
> clobbers in CALL_INSN_FUNCTION_USAGE,
> - sets TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS to true, which
> enables
> the fuse-caller-save optimi
Tested with make info and installed as obvious.
Andreas.
* doc/md.texi (Standard Names): Use @itemx for grouped items.
Remove blank line after @item.
diff --git a/gcc/doc/md.texi b/gcc/doc/md.texi
index e17ffca..1c3a326 100644
--- a/gcc/doc/md.texi
+++ b/gcc/doc/md.texi
@@ -4835,
On Wed, Jun 18, 2014 at 12:55:01PM +0800, Thomas Preud'homme wrote:
> @@ -1646,20 +1648,23 @@ do_shift_rotate (enum tree_code code,
>n->n <<= count;
>break;
> case RSHIFT_EXPR:
> + /* Arithmetic shift of signed type: result is dependent on the value.
> */
> + if (!T
On 11 June 2014 13:55, Richard Earnshaw wrote:
> On 11/06/14 11:19, Charles Baylis wrote:
>> 2014-05-22 Charles Baylis
>>
>> * config/arm/bpabi.S (__aeabi_uldivmod): Fix whitespace.
>> (__aeabi_ldivmod): Fix whitespace.
>
> This is OK, but please wait until the others are ready to g
On 18 June 2014 10:24:16 Martin Jambor wrote:
@@ -3002,10 +3014,8 @@ try_make_edge_direct_virtual_call (struct
cgraph_edge *ie,
if (target)
{
-#ifdef ENABLE_CHECKING
- gcc_assert (possible_polymorphic_call_target_p
-(ie, cgraph_get_node (target)));
-#endif
+ if (!pos
Hi,
This patch improves the code-gen of -marm in the case of two-dimensional
array access.
Given the following code:
typedef struct { int x,y,a,b; } X;
int
f7a(X p[][4], int x, int y)
{
return p[x][y].a;
}
The code-gen on -O2 -marm -mcpu=cortex-a15 is currently
mov r2, r2, as
This time with patch... Apologize.
Yufeng
On 06/18/14 17:31, Yufeng Zhang wrote:
Hi,
This patch improves the code-gen of -marm in the case of two-dimensional
array access.
Given the following code:
typedef struct { int x,y,a,b; } X;
int
f7a(X p[][4], int x, int y)
{
return p[x][y].a;
}
Hi,
On Wed, 18 Jun 2014 09:56:15, David Edelsohn wrote:
>
> On Tue, Jun 17, 2014 at 6:44 PM, BIll Schmidt
> wrote:
>> Hi,
>>
>> As described in https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61542, a
>> new test case (gcc.dg/vect/vect-nop-move.c) was added in 4.9. This
>> exposes a bug on PowerPC l
On Jun 18, 2014, at 3:22 AM, Richard Biener wrote:
>> Space after the *.
>>
>> I think you don't need to copy the LTO harness but you can simply
>> use dg.exp and sth similar to gcc.dg/20081223-1.c (there is an
>> effective target 'lto' to guard for lto support).
>>
>> So simply place the testca
On Wed, Jun 18, 2014 at 05:23:13PM +0200, Marek Polacek wrote:
> On Wed, Jun 18, 2014 at 12:55:01PM +0800, Thomas Preud'homme wrote:
> > @@ -1646,20 +1648,23 @@ do_shift_rotate (enum tree_code code,
> >n->n <<= count;
> >break;
> > case RSHIFT_EXPR:
> > + /* Arithmetic shi
On Jun 18, 2014, at 7:14 AM, Segher Boessenkool
wrote:
> A comment in rs6000.h says "cr5 is not supposed to be used". I checked
> all ABIs,
I usually do a blame and find that change that added it… Ah, there it is,
r341… let’s see, rms did it! :-) Oh well… never mind. Kinda amazing the
b
These patches add ISL AST generation to graphite, which can be chosen
by the fgraphite-code-generator=[isl|cloog] switch. The first patch
makes initial renaming of gloog and gloog_error to
graphite_regenerate_ast_cloog and graphite_regenerate_error,
respectively. The second one adds new files with
On 18/06/2014 21:00, Roman Gareev wrote:
These patches add ISL AST generation to graphite, which can be chosen
by the fgraphite-code-generator=[isl|cloog] switch. The first patch
makes initial renaming of gloog and gloog_error to
graphite_regenerate_ast_cloog and graphite_regenerate_error,
respec
> On 06/13/14 04:24, mliska wrote:
> >
> >You may ask, why the GNU GCC does need such a new optimization. The
> >compiler, having simply better knowledge of a compiled source file,
> >is capable of reaching better results, especially if Link-Time
> >optimization is enabled. Apart from that, GCC imp
On 06/17/14 14:35, Kai Tietz wrote:
I just did retest my testcase with recent source. I can't reproduce
this missed optimization before sched2 pass anymore. I moved second
peephole2 pass just before split_before_sched2 and everything got
caught.
Let's go with this if your idea of using a define
On Wed, Jun 18, 2014 at 10:14 AM, Segher Boessenkool
wrote:
> A comment in rs6000.h says "cr5 is not supposed to be used". I checked
> all ABIs, going as far back as PowerOpen (1994), and found no mention
> of this.
>
> Also document cr6 is used by some vector instructions.
>
> Tested on powerpc6
> On 2014-06-16, 12:12 PM, Robert Suchanek wrote:
> > Pinging for approval.
> >
> > This part of the patch will be needed for MIPS16. The second part to
> enable
> > LRA in MIPS has been already approved.
> >
>
>Sorry, Robert. I thought you are waiting for some Richard's comment
> (actually h
modified_type_die and add_type_attribute take two separate arguments
for whether the type should be const and/or volatile. To help add
more type modifiers pass the requested modifiers as one flag value
to these functions. And introduce helper functions dw_mod_type_flags
and dw_mod_decl_flags to eas
Committed as obvious as Rev. 211806.
Tobias
Index: gcc/fortran/ChangeLog
===
--- gcc/fortran/ChangeLog (Revision 211805)
+++ gcc/fortran/ChangeLog (Arbeitskopie)
@@ -1,3 +1,8 @@
+2014-06-18 Tobias Burnus
+
+ * gfortran.texi (OpenM
> "Joseph" == Joseph S Myers writes:
Tom> This patch adds the plugin to the gcc tree and updates the
Tom> top-level configury.
Following up on your review.
Joseph> I don't see anything obvious that would disable the plugin if
Joseph> plugins are unsupported (e.g. on Windows host) or disable
In this bug we were throwing away the explicit args when doing nested
unification for a function, with the result that we didn't remember them
when proceeding to do unification for the trailing arguments. Fixed by
preserving ARGUMENT_PACK_EXPLICIT_ARGS.
Tested x86_64-pc-linux-gnu, applying to
We were treating a const & member function like a normal const
reference, and binding an rvalue object argument to it. But it doesn't
work that way.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 20a165532a9b0b0dada391716a1fb781af3ec005
Author: Jason Merrill
Date: Wed Jun 18 22:56:25
Hi,
Trunk fails to build on sh4-unknown-linux-gnu with an ICE
during compiling libgomp. See
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61550
for detail. sh.c:prepare_move_operands has the code for
TLS addresses which shouldn't be run if reload in progress
or done.
The attached patch is to fix
On Tue, Jun 17, 2014 at 10:19:06PM -0700, Richard Henderson wrote:
> Trivial fix for missing clobber of the flags over the tlsdesc call.
>
> Ok for all branches?
>
>
> r~
>
> * config/aarch64/aarch64.md (tlsdesc_small_): Clobber CC_REGNUM.
>
pretty sure we need a similar fix for tlsgd_s
On 06/18/2014 03:57 PM, Kyle McMartin wrote:
> pretty sure we need a similar fix for tlsgd_small, since __tls_get_addr
> could clobber CC as well.
As I replied in IRC, no, because tlsgd_small is modeled with an actual
CALL_INSN, and thus call-clobbered registers work as normal.
r~
On Wed, Jun 18, 2014 at 04:04:53PM -0700, Richard Henderson wrote:
> On 06/18/2014 03:57 PM, Kyle McMartin wrote:
> > pretty sure we need a similar fix for tlsgd_small, since __tls_get_addr
> > could clobber CC as well.
>
> As I replied in IRC, no, because tlsgd_small is modeled with an actual
> C
On 06/01/2014 03:00 AM, Tom de Vries wrote:
> +/* Emit call insn with PAT and do aarch64-specific handling. */
> +
> +bool
> +aarch64_emit_call_insn (rtx pat)
> +{
> + rtx insn = emit_call_insn (pat);
> +
> + rtx *fusage = &CALL_INSN_FUNCTION_USAGE (insn);
> + clobber_reg (fusage, gen_rtx_REG (
On 06/01/2014 03:00 AM, Tom de Vries wrote:
> +aarch64_emit_call_insn (rtx pat)
> +{
> + rtx insn = emit_call_insn (pat);
> +
> + rtx *fusage = &CALL_INSN_FUNCTION_USAGE (insn);
> + clobber_reg (fusage, gen_rtx_REG (word_mode, IP0_REGNUM));
> + clobber_reg (fusage, gen_rtx_REG (word_mode, IP1_R
On 06/01/2014 04:27 AM, Tom de Vries wrote:
> + if (TARGET_AAPCS_BASED)
> +{
> + /* For AAPCS, IP and CC can be clobbered by veneers inserted by the
> + linker. We need to add these to allow
> + arm_call_fusage_contains_non_callee_clobbers to return true. */
> + rtx *fusa
On Wed, 18 Jun 2014, Jakub Jelinek wrote:
> I've committed following change:
Cool.
> Status
>
> +Jun 18, 2014
> +The last major part of Fortran OpenMP v4.0 support has been
> +committed into SVN mainline.
> +
> Oct 11, 2013
> The gomp-4_0-branch has been merged into SVN
> -mainline, so GCC 4
On 05/19/2014 07:30 AM, Tom de Vries wrote:
> + for (insn = get_insns (); insn != NULL_RTX; insn = next_insn (insn))
> +{
> + HARD_REG_SET insn_used_regs;
> +
> + if (!NONDEBUG_INSN_P (insn))
> + continue;
> +
> + find_all_hard_reg_sets (insn, &insn_used_regs, false);
> +
>
> From: Jakub Jelinek [mailto:ja...@redhat.com]
> Sent: Thursday, June 19, 2014 1:54 AM
>
> Seems there are actually two spots with this, not just one.
>
> Completely untested fix:
>
> 2014-06-18 Jakub Jelinek
>
> * tree-ssa-math-opts.c (do_shift_rotate, find_bswap_or_nop_1):
> Cast
>
94 matches
Mail list logo
