Add new target builtins __builtin_cpu_is_intel_corei7 and
__builtin_cpu_is_amdfam10.
* config/i386/i386-cpuinfo.c (__processor_model): Add new members
__cpu_is_intel_corei7 and __cpu_is_amdfam10.
(get_amd_cpu): Set __cpu_is_amdfam10.
(get_intel_cpu): Set __cpu_is_in
When there are no arguments for a parameter pack, an expansion between
parens in a variable initializer reduces to value-initialization. This
isn't valid for auto. Fixed by checking for incomplete type.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 380724b25cd317ea8b6434cdd1215e831ac2
Richard Henderson wrote:
> Ok except,
>
>> +static inline void
>> +atomic_read_barrier (void)
>> +{
>> + __sync_synchronize ();
>> +}
>> +
>> +static inline void
>> +atomic_write_barrier (void)
>> +{
>> + __sync_synchronize ();
>> +}
>
> We no longer use these since the library has been convert
On Fri, Dec 16, 2011 at 2:49 PM, Joseph S. Myers
wrote:
> On Wed, 14 Dec 2011, H.J. Lu wrote:
>
>> Hi,
>>
>> This patch adds -mandroid support on Linux/i386, similar to Linux/ARM.
>> OK for trunk in stage 1?
>
> This does not appear to be a patch against current trunk, but against a
> much older v
On Fri, Dec 16, 2011 at 3:27 PM, Joseph S. Myers
wrote:
> On Thu, 15 Dec 2011, H.J. Lu wrote:
>
>> Hi,
>>
>> Linux/x86 target is configured as i686-android-linux. Changing it to
>> i686-linuxandroid requires changes in many packages. This patch adds
>> DEFAULT_LIBC=LIBC_BIONIC and ANDROID_DEFAULT
Hi,
when I was fixing PR 50622, somehow I lost a hunk in the patch along
the way. I punished myself by making up an ICEing testcase which is
included alongside the fix below. The fix is needed for trunk and
also in the 4.6 branch.
I have bootstrapped and tested the patch on x86_64-linux on top
This follows the other logical patterns, andsi3 and iorsi3, which
also accept reg_or_int_operand and split post-reload.
For the atomic optabs, I have two choices: either rely on this
post-reload splitting, or call arm_split_constant myself. Making
all of the logicals work similarly seemed cleane
I noticed this while testing my atomic optabs patch with -mthumb.
This is identical to what we actually output in the current sync code:
> arm_output_asm_insn (emit, 0, operands, "cmp\t%%0, %%1");
> if (is_di)
> {
> arm_output_it (emit, "", "eq");
> arm_outp
There's no "good" place for this. The description in Makefile.def that libitm
uses c++ is only used insofar as the dependencies for libitm -- it's built
after libstdc++ as the language support library. If we put this into the
toplevel configure.ac directly, we have to write a bunch of shell co
From: Eric Botcazou
Date: Sat, 17 Dec 2011 00:37:43 +0100
> 2011-12-16 Eric Botcazou
>
> * config/sparc/sparc.md (UNSPEC_FRAME_BLOCKAGE): New constant.
> (frame_blockage): New expander.
> (frame_blockage): New instruction.
> * config/sparc/sparc.c (sparc_expand_prologu
On 12/16/2011 03:05 PM, Kaz Kojima wrote:
> * configure.tgt: Add sh* case.
> * config/sh/target.h: New file.
> * config/sh/sjlj.S: New file.
> * config/linux/sh/futex_bits.h: New file.
Ok except,
> +static inline void
> +atomic_read_barrier (void)
> +{
> + __sync_synchron
In regular mode (with register windows), the compiler can make use of three
sequences to establish the frame:
1. a single 'save' instruction taking an immediate as decrement,
2. a 'save' instruction taking an immediate followed by an 'add' instruction
taking another immediate as decrement,
3.
On Thu, 15 Dec 2011, H.J. Lu wrote:
> Hi,
>
> Linux/x86 target is configured as i686-android-linux. Changing it to
> i686-linuxandroid requires changes in many packages. This patch adds
> DEFAULT_LIBC=LIBC_BIONIC and ANDROID_DEFAULT=1 to *-android-*. OK for
> trunk in stage 1?
I think using th
On Thu, 15 Dec 2011, Georg-Johann Lay wrote:
> Index: gcc/config/avr/genmultilib.awk
> ===
> --- gcc/config/avr/genmultilib.awk(revision 0)
> +++ gcc/config/avr/genmultilib.awk(revision 0)
> @@ -0,0 +1,255 @@
This new file ne
Lance Taylor
PR middle-end/51592
* gcc.dg/20111216-1.c: New test.
Index: gcc.dg/20111216-1.c
===
--- gcc.dg/20111216-1.c (revision 0)
+++ gcc.dg/20111216-1.c (revision 0)
@@ -0,0 +1,17 @@
+/* { dg-do compile
Ok, Cary's explanation makes sense. Please update the comments to make
it clearer.
http://codereview.appspot.com/5416043/diff/1/gcc/config/i386/i386.c
File gcc/config/i386/i386.c (right):
http://codereview.appspot.com/5416043/diff/1/gcc/config/i386/i386.c#newcode10927
gcc/config/i386/i386.c:109
Hi,
The attached patch is to port libitm to SH. It's a updated
one of the patch in
http://gcc.gnu.org/ml/gcc-patches/2011-11/msg02382.html
The patch is tested with "make -k check" and the result is
=== libitm Summary ===
# of expected passes23
# of expected failures
On Wed, 14 Dec 2011, H.J. Lu wrote:
> Hi,
>
> This patch adds -mandroid support on Linux/i386, similar to Linux/ARM.
> OK for trunk in stage 1?
This does not appear to be a patch against current trunk, but against a
much older version.
> diff --git a/gcc/config/i386/linux.h b/gcc/config/i386/l
Ok for google branches -- not applicable to trunk.
David
On Fri, Dec 16, 2011 at 2:37 PM, Rong Xu wrote:
> Hi,
>
> This patch fixes the insane values in runs and run_max field of the
> program summary. This can happen when the gcda files manually removed
> from the output directory after the mer
On 12/15/2011 01:23 AM, Tom de Vries wrote:
> PR tree-optimization/51491
> * tree-ssa-ccp.c (insert_clobber_before_stack_restore)
> (gsi_prev_dom_bb_nondebug, insert_clobbers_for_var): New function.
> (ccp_fold_stmt): Use insert_clobbers_for_var after a successful
> fo
Hi,
This patch fixes the insane values in runs and run_max field of the
program summary. This can happen when the gcda files manually removed
from the output directory after the merge happens.
This is for google branch only.
Tested with (make; make check).
-Rong
2011-12-16 Rong Xu
literal_type_p doesn't like to see incomplete types, so let's check for
that case first.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 0418652f7ecc0286565abc805d4728872eac32b9
Author: Jason Merrill
Date: Fri Dec 16 14:04:01 2011 -0500
PR c++/51461
* decl.c (check_static_var
For PR 48035, Jakub factored build_zero_init_1 out of build_zero_init to
avoid clobbering the area occupied by virtual bases in the complete
type, but not part of a base subobject. For 50618, I changed
build_aggr_init_1 to use it. This testcase shows another case of the
same problem: in this
http://codereview.appspot.com/5416043/diff/12001/gcc/config/i386/i386.c
File gcc/config/i386/i386.c (left):
http://codereview.appspot.com/5416043/diff/12001/gcc/config/i386/i386.c#oldcode10928
gcc/config/i386/i386.c:10928: if (current_function_decl != NULL_TREE &&
If I understand correctly, this
On Mon, 12 Dec 2011, Andrew Pinski wrote:
> * config/mips/mips-cpus.def: Add Octeon2.
You need to update the documentation for MIPS -march= in invoke.texi,
which lists the supported CPUs.
--
Joseph S. Myers
jos...@codesourcery.com
On 12/16/2011 03:16 AM, Chung-Lin Tang wrote:
> Hi, under powerpc targets, using -mrelocatable under some cases triggers
> an error during final assembly generation: rs6000_assemble_integer()
> calls varasm.c:unlikely_text_section_p(), and the call chain eventually
> gets to where the section htab
On 12/16/2011 09:44 AM, Michael Zolotukhin wrote:
>> Why? SSSE3 *really can* do arbitrary permutation. If you say that
>> isn't enough, then there's something wrong in the back end, and all
>> you're doing is papering over a bug.
> Yes, sure, SSSE3 is enough for that. I checked for AVX in vect_pe
On Fri, 2011-12-16 at 11:05 -0800, Xinliang David Li wrote:
> On Fri, Dec 16, 2011 at 3:35 AM, Torvald Riegel wrote:
> > For Transactional Memory support, we also create versions of functions
> > (see code in trunk, e.g., in trans-mem.c). Right now, that's a single
> > instrumented version of the
I have uploaded a new patch set with all the mentioned changes made. If
a function has the target attribute it will not be touched by the
autoclone pass.
Also fixed some test cases which were broken because the clone names
used '_' instead of '.' for suffixing.
On 2011/12/16 19:39:47, davidxl wr
On Sat, 10 Dec 2011, Richard Henderson wrote:
> +/* Begin by defaulting to whatever options were given to the compiler. */
> +int GTM_hwcap HIDDEN = 0
> +#ifdef __VFP_FP__
> + | HWCAP_ARM_VFP
> +#endif
> +#ifdef __IWMMXT__
> + | HWCAP_ARM_IWMMXT
> +#endif
> + ;
It doesn't matter much since th
http://codereview.appspot.com/5416043/diff/12001/gcc/config/i386/i386.c
File gcc/config/i386/i386.c (left):
http://codereview.appspot.com/5416043/diff/12001/gcc/config/i386/i386.c#oldcode10928
gcc/config/i386/i386.c:10928: if (current_function_decl != NULL_TREE &&
I am not sure how the hack you
On 11-12-16 12:35 , dvyu...@google.com wrote:
On 2011/12/16 16:00:06, Diego Novillo wrote:
On 11-12-16 09:54 , Dmitriy Vyukov wrote:
> This is for google-main branch.
> Ignore thunks in ThreadSanitizer pass.
>
> Index: gcc/tree-tsan.c
> ===
Dear all,
this patch fixes on of the FIXMEs in coarray/poly_run_1.f90 [and extends
the test case] by doing in check.c the same for coarrays as already done
for arrays.
Additionally, I encountered bugs in the constraint checking for
polymorphic variables, thus, I fixed those and added a test
http://codereview.appspot.com/5490054/diff/1011/config/i386/i386.c
File config/i386/i386.c (right):
http://codereview.appspot.com/5490054/diff/1011/config/i386/i386.c#newcode26569
config/i386/i386.c:26569: +mversion_for_core2 (tree *optimization_node,
-> mversionable_for_core2_p ?
http://codere
On 12/03/2011 01:18 PM, Dodji Seketeli wrote:
What I am proposing in the patch below is to wait until the closing
'}' of a given scope, walk the statements of that scope, and fixup the
locations of CLEANUP_STMTs by setting them the current location, i.e,
the location of the closing '}'.
I think
Hi!
This testcase apparently fails on 32-bit hppa (both hpux and linux),
because the struct there is passed by callee copied reference
(unlike most other targets or even 64-bit hppa).
So "Will be eliminated" is printed 3 times instead of 4 times.
This patch just XFAILs it.
Tested with x86_64-lin
OK.
Jason
Gabriel Dos Reis writes:
> On Fri, Dec 16, 2011 at 10:40 AM, Dodji Seketeli wrote:
>> Hello,
>
>> So I am thinking that maybe letting expand_default_init represent the
>> invalid initialization would prevent us from getting into this case.
>> For that to work I had to make
>> build_constexpr_con
Jason Merrill writes:
> On 12/16/2011 11:40 AM, Dodji Seketeli wrote:
>>/* It's OK to skip a member with a trivial constexpr ctor.
>> A constexpr ctor that isn't trivial should have been
>> added in by now. */
>>gcc_checking_assert (!TYPE_HAS_C
On Fri, 9 Dec 2011, Richard Earnshaw wrote:
> I think we've reached the point where the following target
> configurations should be End-of-Life'd. As such, I'd like to mark them
> as deprecated in gcc-4.7, prior to removal after the branch.
I agree it's time to remove these old-ABI ports - and o
On 12/16/2011 11:40 AM, Dodji Seketeli wrote:
/* It's OK to skip a member with a trivial constexpr ctor.
A constexpr ctor that isn't trivial should have been
added in by now. */
gcc_checking_assert (!TYPE_HAS_COMPLEX_DFLT (ftype));
I
On 16/12/2011 09:01, Kai Tietz wrote:
> 2011/12/15 Dave Korn:
>> { dg-options "-mno-align-double" { target i?86-*-cygwin* i?86-*-mingw* } }
>> { dg-additional-options "-mno-ms-bitfields" { target i?86-*-mingw* } }
>>
>> ... so that MinGW gets both and Cygwin only the one it wants? (Actually the
>
> Why? SSSE3 *really can* do arbitrary permutation. If you say that
> isn't enough, then there's something wrong in the back end, and all
> you're doing is papering over a bug.
Yes, sure, SSSE3 is enough for that. I checked for AVX in vect_perm
just because there is no check for SSSE3 and no test
On Fri, Dec 16, 2011 at 10:40 AM, Dodji Seketeli wrote:
> Hello,
> So I am thinking that maybe letting expand_default_init represent the
> invalid initialization would prevent us from getting into this case.
> For that to work I had to make
> build_constexpr_constructor_member_initializers try a
This is for google-main branch.
Ignore thunks in ThreadSanitizer pass.
Index: gcc/tree-tsan.c
===
--- gcc/tree-tsan.c (revision 182401)
+++ gcc/tree-tsan.c (working copy)
@@ -426,6 +426,10 @@
ignore_init = 1;
}
+
On 12/15/2011 11:21 PM, Michael Zolotukhin wrote:
> E.g., in vect-114.c we need permutation only for reversing vector
> elements, even ssse3 isn't required for it ...
Sure.
> while in slp-perm-9 such permutation isn't enough.
Why? SSSE3 *really can* do arbitrary permutation. If you say that
is
Hello,
In the example of the patch below, the invalid initialization for the
'a' member in the context of the constructor of B (in
expand_default_init) yields an error_mark_node, and rightfully so.
But then expand_default_init doesn't generate any (expression)
statement to represent that invalid
Hi!
This PR got also fixed by my PR51576 patch, as that one was a bootstrap
issue with non-default language, this one adds a testcase for it into
the testsuite.
Regtested on x86_64-linux, committed to trunk.
2011-12-16 Jakub Jelinek
PR tree-optimization/51575
* g++.dg/opt/pr5
Alpha/VMS claims to be ELF like, but it is not enough to support T M clone
registry.
This patch disables it as it broke compilation. The right way would be not to
define OBJECT_FORMAT_ELF.
Committed on trunk.
Tristan.
2011-12-16 Tristan Gingold
* config/alpha/vms.h (USE_TM_CLONE_
On 11-12-16 09:54 , Dmitriy Vyukov wrote:
This is for google-main branch.
Ignore thunks in ThreadSanitizer pass.
Index: gcc/tree-tsan.c
===
--- gcc/tree-tsan.c (revision 182401)
+++ gcc/tree-tsan.c (working copy)
@@ -426,6 +
Now that transparent aliases are heavily used by VMS to translate names, we
shouldn't directly emit symbol name but use assemble_name.
This patch fixes some abuse.
Committed on trunk.
Tristan.
2011-12-16 Tristan Gingold
* config/alpha/alpha.c (alpha_use_linkage): Follow transparent
The default object suffix for VMS is .obj, so we need to force .o.
Ironically the issue was in vms/t-vms, but wasn't caught until a native
compiler was built.
Committed on trunk.
Tristan.
2011-12-16 Tristan Gingold
* config/vms/t-vms (vms.o, vms-c.o): Force output filename.
--- a/g
Hi!
Bootstrap failed with go enabled at least on i686-linux for the last day
or so, some changes in libgo crypto apparently started triggering a
recently introduced problem in replace_uses_by. fold_stmt isn't the
only way a stmt can be changed by replace_uses_by, in the sha1.go
case replace_exp c
Hi,
libiberty/pex-unix.c uses the global variable 'environ', which is available on
VMS but with a DEC-C name.
This patch adds a translation for it in the VMS table.
Committed on trunk.
Tristan.
2011-12-16 Tristan Gingold
* config/vms/vms.c (VMS_CRTL_GLOBAL): Define.
(vms_pa
This is for google-main branch.
Ignore thunks in ThreadSanitizer pass.
Index: gcc/tree-tsan.c
===
--- gcc/tree-tsan.c (revision 182401)
+++ gcc/tree-tsan.c (working copy)
@@ -426,6 +426,11 @@
ignore_init = 1;
}
+
Rainer Orth writes:
> The following patch is necessary to restore Solaris Go bootstrap after
> your patch went in, to fix
>
> /vol/gcc/src/hg/trunk/local/libgo/go/os/stat.go:22:29: error: argument 1 has
> incompatible type (cannot use type syscall.Timestruc as type syscall.Timespec)
> /vol/gcc/s
This fixes PR51572 - as with LTO we do not call debug_hooks->type_decl
for each global TYPE_DECL we run into the issue that if those are
emitted lazily they always end up in the limbo DIE list but are
later not allowed there (we ICE processing the limbo list). The
fix is to not put those on the l
PING 7: For reviewing the modified CR16 port.
Hello,
Can some one please review the updated patch and let me know if any more
changes are required to be done in it?
Rainer had suggested few important changes last time. After making those
changes, the modified patch was posted at following URL:
For Transactional Memory support, we also create versions of functions
(see code in trunk, e.g., in trans-mem.c). Right now, that's a single
instrumented version of the original code but having different
transactional instrumentations available might be worthwhile in the
future.
Is there a chance
Hi, under powerpc targets, using -mrelocatable under some cases triggers
an error during final assembly generation: rs6000_assemble_integer()
calls varasm.c:unlikely_text_section_p(), and the call chain eventually
gets to where the section htab is queried for ".text.unlikely", and
fails because of
Ian Lance Taylor writes:
> I have committed a patch to update libgo to the weekly.2011-12-02
> release of the master library. In this message I have only included the
> diffs to files that are specific to the gccgo copy of the library.
> Bootstrapped and ran Go testsuite on x86_64-unknown-linux-
> 2011-12-15 Enkovich Ilya
>
> PR target/50038
> * implicit-zee.c: Delete.
> * ree.c: New file.
> * Makefile.in: Replace implicit-zee.c with ree.c.
> * config/i386/i386.c (ix86_option_override_internal): Set
> flag_ree for 32 bit platform.
> * common.opt
On 16.12.2011 11:53, Jakub Jelinek wrote:
On Fri, Dec 16, 2011 at 09:47:52AM +0400, Andrey Belevantsev wrote:
Thus, if an insn has any note that prevents it from being cloned, we
need to fix the above place. (If an insn should not even be moved
from its block, then CANT_MOVE should be set to 1
2011/12/15 Dave Korn :
> On 15/12/2011 10:33, Kai Tietz wrote:
>
>> -// { dg-options "-mno-align-double" { target i?86-*-cygwin* i?86-*-mingw* }
>> }
>> +// As for mingw target the the ms-bitfield switch is activated by default,
>> +// make sure for this test that it is disabled.
>> +// { dg-optio
On Thu, 15 Dec 2011, Patrick Marlier wrote:
> In PR51280, LTO does ICE because the object file uses TM builtin but the TM is
> not enabled.
> In the patch, it displays a error message if the builtin is not defined and
> due to TM.
> I moved is_tm_builtin() from calls.c to trans-mem.c. I splitted i
> REG_DEAD/REG_UNUSED notes are recomputed by df, I don't think
> anything recomputes REG_EQUAL/REG_EQUIV.
Correct, passes need to adjust the REG_EQUAL/REG_EQUIV notes if need be.
--
Eric Botcazou
On Thu, 15 Dec 2011, Andrew Pinski wrote:
> On Thu, Dec 15, 2011 at 4:17 AM, Richard Guenther wrote:
> > + fprintf (mstream, "\t@-touch -r %s %s.tem && mv %s.tem
> > %s\n",
> > + input_name, input_name, input_name, input_name);
>
> Is touch -r portable?
Sho
67 matches
Mail list logo
