2011/5/18 Uros Bizjak ubiz...@gmail.com:
Hello!
2011-05-16 Zuxy Meng zuxy.m...@gmail.com
PR i386/48743
* config/i386/cpuid.h (bit_MMXEXT): New
* config/i386/driver-i386.c (host_detect_local_cpu): Detect Athlon by
the presence of extended MMX instead of extended 3DNow!
PR target/48743
Ontop of the previous LTO type merging cleanup the following patch
should fix PR48849 by making sure that TYPE_CANONICAL computation
of pointer types at LTO streaming time matches that of the middle-end
(and thus later passes that create new pointer types).
LTO bootstrapped and tested on
On Thu, May 19, 2011 at 8:26 PM, Xinliang David Li davi...@google.com wrote:
I have done some SPEC testing evaluating the performance impact of
your patch. They look very positive. LIPO got helped even more than
FDO (I only did SPEC2k LIPO testing).
Did you also check impact on compile-time
On Thu, May 19, 2011 at 7:52 PM, Kai Tietz ktiet...@googlemail.com wrote:
To illustrate in which scenario code in tree-ssa-forwprop doesn't help
is binop-tor4.c
w/o this patch we get
foo (int a, int b, int c)
{
int e;
int d;
int D.2701;
_Bool D.2700;
_Bool D.2699;
_Bool D.2698;
Hi,
On 05/18/2011 11:20 PM, Zdenek Dvorak wrote:
+ sa_cost = (TREE_CODE (expr) != MINUS_EXPR
+ ? shiftadd_cost[speed][mode][m]
+ : (mult == op1
+? shiftsub1_cost[speed][mode][m]
+: shiftsub0_cost[speed][mode][m]));
+ res = new_cost
This patch adds support for 12-bit immediate values for Thumb-2 in ADD and
SUB instructions. We added two new alternatives for *arm_addsi3 which
make use of two new constraints for 12-bit values. Also we modified
costs of PLUS rtx expression.
This patch reduces size of libevas by 1788 bytes
However, did you check that the atomic qualifier is correctly written
re-read in the state (I believe you did, otherwise it probably won't
work). This is needed for plugins using it, or using atomic qualified
fields of existing (or future) structures.
Yes. String options are written as they
I have corrected my patch with your remarks.
Especially about the debug functions, it takes the debug melt flag into
account. Moreover, I have allowed the user to give a message when
calling this function, as usual in MELT debug functions. Giving this
message was a bit difficult (This is the
On 2011/5/20 11:45 AM, Dmitry Plotnikov wrote:
This patch adds support for 12-bit immediate values for Thumb-2 in ADD and
SUB instructions. We added two new alternatives for *arm_addsi3 which
make use of two new constraints for 12-bit values. Also we modified
costs of PLUS rtx expression.
On Fri, 2011-05-20 at 18:06 +0200, Chung-Lin Tang wrote:
On 2011/5/20 11:45 AM, Dmitry Plotnikov wrote:
This patch adds support for 12-bit immediate values for Thumb-2 in ADD and
SUB instructions. We added two new alternatives for *arm_addsi3 which
make use of two new constraints for
This patch replaces a hardcoded . of local label prefix with
LOCAL_LABEL_PREFIX macro in the ARM backend (generating a trampoline in
thumb mode).
The patch also contains a new test to make sure that the local label
.Ltrampoline_start: is generated correctly.
No regressions on arm-none-eabi.
On 2011/5/20 下午 12:10, Richard Earnshaw wrote:
On Fri, 2011-05-20 at 18:06 +0200, Chung-Lin Tang wrote:
On 2011/5/20 11:45 AM, Dmitry Plotnikov wrote:
This patch adds support for 12-bit immediate values for Thumb-2 in ADD and
SUB instructions. We added two new alternatives for *arm_addsi3
On 20/05/11 10:45, Dmitry Plotnikov wrote:
This patch adds support for 12-bit immediate values for Thumb-2 in ADD and
SUB instructions. We added two new alternatives for *arm_addsi3 which
make use of two new constraints for 12-bit values. Also we modified
costs of PLUS rtx expression.
A
On 2011/5/21 12:06 AM, Chung-Lin Tang wrote:
I'll also note here that ADD/SUB are not the only instructions with
12-bit immediate under Thumb-2; so does AND, ORR, etc.
FTR, I was wrong on the above statement. Only add/sub seems to have the
wide constant operands.
CL
On 17/05/11 14:10, Chung-Lin Tang wrote:
On 2011/5/13 04:26 PM, Richard Sandiford wrote:
Richard Sandifordrichard.sandif...@linaro.org writes:
Chung-Lin Tangclt...@codesourcery.com writes:
My fix here simply adds 'reload_completed' as an additional condition
for EPILOGUE_USES to return true
On Fri, May 20, 2011 at 1:31 AM, Richard Guenther rguent...@suse.de wrote:
Ontop of the previous LTO type merging cleanup the following patch
should fix PR48849 by making sure that TYPE_CANONICAL computation
of pointer types at LTO streaming time matches that of the middle-end
(and thus later
On 2011/5/20 下午 07:41, Ramana Radhakrishnan wrote:
On 17/05/11 14:10, Chung-Lin Tang wrote:
On 2011/5/13 04:26 PM, Richard Sandiford wrote:
Richard Sandifordrichard.sandif...@linaro.org writes:
Chung-Lin Tangclt...@codesourcery.com writes:
My fix here simply adds 'reload_completed' as an
There are an increasing number of common object files that get linked
into a range of programs GCC builds and installs for the host,
including but not limited to the core compilers and the drivers.
These have various ad hoc code listing them explicitly for each
program using some of them, with
We should use ptr_mode instead of Pmode for vtable adjustment. This
patch fixed:
FAIL: ext/ext_pointer/1.cc execution test
FAIL: gcc.dg/c99-stdint-1.c (test for excess errors)
FAIL: g++.dg/abi/covariant3.C execution test
FAIL: g++.dg/abi/vcall1.C execution test
FAIL: g++.dg/inherit/covariant17.C
Hi,
for various changes we can run into the situation that frontends fill the
cache with a different number of common nodes than the LTO frontend,
leading to confusion of which numbers mean which tree. This can happen
when some global trees are NULL, or when some global trees are
Hi!
As discussed in the PR, these routines happily look through PHI nodes
into previous loop iteration. f is initialized in the previous
loop iteration with f_10 = d_6 == 3; and in the current iteration
there is f_2 != 0 d_6 == 4 test. As we look through
# f_2 = PHI 0(2), f_10(6) PHI, this is
On Fri, 20 May 2011 11:47:57 +0200
Pierre Vittet pier...@pvittet.com wrote:
I have corrected my patch with your remarks.
Especially about the debug functions, it takes the debug melt flag into
account. Moreover, I have allowed the user to give a message when
calling this function, as
On 05/20/2011 02:21 PM, Joseph S. Myers wrote:
2011-05-20 Joseph Myersjos...@codesourcery.com
* Makefile.in (LIBDEPS): Add libcommon.a.
(LIBS): Likewise.
(GCC_OBJS): Remove diagnostic.o, pretty-print.o and input.o.
(OBJS-common): Remove diagnostic.o, input.o,
Hi,
the attached patch makes libgfortran prefer the x86 asm implementation
of setting FPU traps instead of using glibc feenableexcept(). This
enables to trap on denormals, which glibc doesn't support. Contrary to
the original patch in the PR, this swaps only glibc and x86
priorities, as using the
Ian Lance Taylor i...@google.com writes:
I have committed a patch to update libgo to the current version of the
Go library. This patch includes some minor changes to the Go frontend
and to the testsuite. The patch is too large to include here, but most
of it is simply a copy of the changes
Hi,
On Fri, 20 May 2011, Joseph S. Myers wrote:
(Apart from the arbitrary division between GCC_OBJS and the xgcc link
rule, mentioned above, there are other arbitrary divisions that don't
make sense to me. In particular, the separation between OBJS-common,
OBJS-md and OBJS-archive, all
On Fri, May 20, 2011 at 03:44:07PM +0200, Rainer Orth wrote:
Ian Lance Taylor i...@google.com writes:
I have committed a patch to update libgo to the current version of the
Go library. This patch includes some minor changes to the Go frontend
and to the testsuite. The patch is too large
This patch removes one of the two remaining uses of TYPE_ARG_TYPES in
the ObjC/C++ frontends. (The other one should be addressed in a
different manner.) Given the constraints of the function_args_iterator
interface, I thought rewriting the logic of the loop would make things
slightly clearer.
On 05/20/2011 03:49 PM, Michael Matz wrote:
Hi,
On Fri, 20 May 2011, Joseph S. Myers wrote:
(Apart from the arbitrary division between GCC_OBJS and the xgcc link
rule, mentioned above, there are other arbitrary divisions that don't
make sense to me. In particular, the separation between
As $SUBJECT suggests. It may be worth noting that we now do more work
after this patch (stdarg_p and prototype_p both traverse TYPE_ARG_TYPES
under the hood); one day those will be simple boolean tests.
Tested on x86_64-unknown-linux-gnu with go. OK to commit?
-Nathan
gcc/
* godump.c
Hi!
As discussed in the PR, similarly to dwarf2out.c's reference_to_unused
and late removal of unreferenced symbols, this patch will make sure we don't
reference static symbols that were optimized away.
So far all dbxout was using was give up if DECL_RTL wasn't set, but DECL_RTL
can be set e.g.
There is a crash in vect_is_slp_reduction where use_stmt doesn't get
initialized in the FOR_EACH_IMM_USE_FAST loop.
1718 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, lhs)
(gdb) call debug_tree(lhs)
ssa_name 0x76e3a5a0
type integer_type 0x77f8c7e0 int32_t sizes-gimplified public
On Fri, 20 May 2011, Michael Matz wrote:
Hi,
for various changes we can run into the situation that frontends fill the
cache with a different number of common nodes than the LTO frontend,
leading to confusion of which numbers mean which tree. This can happen
when some global trees are
On Fri, May 20, 2011 at 3:12 PM, Jakub Jelinek ja...@redhat.com wrote:
Hi!
As discussed in the PR, these routines happily look through PHI nodes
into previous loop iteration. f is initialized in the previous
loop iteration with f_10 = d_6 == 3; and in the current iteration
there is f_2 != 0
On Fri, May 20, 2011 at 4:06 PM, Ryan Mansfield rmansfi...@qnx.com wrote:
There is a crash in vect_is_slp_reduction where use_stmt doesn't get
initialized in the FOR_EACH_IMM_USE_FAST loop.
1718 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, lhs)
(gdb) call debug_tree(lhs)
ssa_name
Janne Blomqvist wrote:
the attached patch makes libgfortran prefer the x86 asm implementation
of setting FPU traps instead of using glibc feenableexcept(). This
enables to trap on denormals, which glibc doesn't support.
[...]
Regtested on x86_64-unknown-linux-gnu, Ok for trunk?
OK. Thanks for
Hi,
I checked in this patch to add x32 libstdc++-v3 baseline_symbols.txt.
H.J.
baseline_symbols.txt.diff.bz2
Description: BZip2 compressed data
Hi,
gcc.dg/20020219-1.c is invalid for x32. I checked in this patch to
disable it.
H.J.
commit 5599f023b65b85b5e5320321e64775910a739ff1
Author: H.J. Lu hjl.to...@gmail.com
Date: Fri May 20 06:11:14 2011 -0700
Disable gcc.dg/20020219-1.c for x32.
diff --git
On Fri, May 20, 2011 at 7:21 AM, Joseph S. Myers
jos...@codesourcery.com wrote:
There are an increasing number of common object files that get linked
into a range of programs GCC builds and installs for the host,
including but not limited to the core compilers and the drivers.
the diagnostic
On Wed, May 18, 2011 at 8:36 PM, Uros Bizjak ubiz...@gmail.com wrote:
2011-05-18 Uros Bizjak ubiz...@gmail.com
* config/i386/mingw32.h (OUTPUT_QUOTED_STRING): Fix macro
argument expansion.
Patch was compile tested by crosscompiling to ming32 target.
I went ahead and commit
PING
On Sun, May 15, 2011 at 14:12, Janne Blomqvist
blomqvist.ja...@gmail.com wrote:
Hi,
so, here is take 3 (sigh). Compared to take 2, it no longer uses
stdio, since opening a stdio FILE stream probably malloc()'s a buffer,
which is not async-signal-safe.
Regtested on
This fixes PR49079 - the mem-ref merge broke the trailing
array access detection of get_ref_base_and_extent because of the
embedding of a view-conversion.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to
trunk and branch.
Richard.
2011-05-20 Richard Guenther rguent...@suse.de
Make libstdc++'s abi_check more robust against readelf output format.
libstdc++-abi/abi_check in the libstdc++-v3 testsuite relies on a fixed
number of space separated fields in readelf output. However, the field
count for readelf output can vary where the library contains OS or processor
Ok, for google/integration. Please integrate to google/main and
google/gcc-4_6 as well.
Ollie
On Fri, May 20, 2011 at 10:05 AM, Simon Baldwin sim...@google.com wrote:
Make libstdc++'s abi_check more robust against readelf output format.
libstdc++-abi/abi_check in the libstdc++-v3 testsuite
2011/5/20 Uros Bizjak ubiz...@gmail.com:
On Wed, May 18, 2011 at 8:36 PM, Uros Bizjak ubiz...@gmail.com wrote:
2011-05-18 Uros Bizjak ubiz...@gmail.com
* config/i386/mingw32.h (OUTPUT_QUOTED_STRING): Fix macro
argument expansion.
Patch was compile tested by crosscompiling
On May 20, 2011, at 6:58 AM, Nathan Froyd wrote:
This patch removes one of the two remaining uses of TYPE_ARG_TYPES in
the ObjC/C++ frontends. (The other one should be addressed in a
different manner.) Given the constraints of the function_args_iterator
interface, I thought rewriting the
On Fri, May 20, 2011 at 2:12 AM, Richard Guenther
richard.guent...@gmail.com wrote:
On Thu, May 19, 2011 at 8:26 PM, Xinliang David Li davi...@google.com wrote:
I have done some SPEC testing evaluating the performance impact of
your patch. They look very positive. LIPO got helped even more
Joseph S. Myers jos...@codesourcery.com writes:
I haven't tested building mips-tdump or mips-tfile with this patch,
but don't think it should cause any problems as the static linker
should just ignore the diagnostics objects for those programs. In
principle mips-tfile should use the common
Ping.
On Sat, May 14, 2011 at 8:01 AM, Easwaran Raman era...@google.com wrote:
http://gcc.gnu.org/ml/gcc-patches/2011-05/msg00781.html
Ok to check in this one?
Thanks,
David
On Wed, May 18, 2011 at 12:30 PM, Joseph S. Myers
jos...@codesourcery.com wrote:
On Wed, 18 May 2011, David Li wrote:
+ error (Unrecognized option %s, is_enable ? -fenable :
-fdisable);
+ error (Unknown pass %s specified in %s,
+
gcc-patches-ow...@gcc.gnu.org wrote on 20/05/2011 05:17:47 PM:
On Fri, May 20, 2011 at 4:06 PM, Ryan Mansfield rmansfi...@qnx.com
wrote:
There is a crash in vect_is_slp_reduction where use_stmt doesn't get
initialized in the FOR_EACH_IMM_USE_FAST loop.
1718
On Fri, May 20, 2011 at 8:10 AM, Ollie Wild a...@google.com wrote:
Ok, for google/integration. Please integrate to google/main and
google/gcc-4_6 as well.
Done: r173959, r173960, r173961.
Thanks,
--
Paul Pluzhnikov
This patch fixes -fdump-go-spec to work correctly when the same name is
defined as both a preprocessor macro and an enum constant. This
actually happens with glibc with the name IPPORT_RESERVED. Bootstrapped
and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to
mainline.
Ian
On Fri, 20 May 2011, Sumanth Gundapaneni wrote:
There is no such target as target-mudflap,
The configure file is altered as per your suggestion.
Please remove the remaining toplevel configure changes from the patch.
I moved the code disabling libgcj for particular targets out of the
general
G++ has had a long-standing bug with unqualified name resolution in
templates: if we didn't find any declaration when looking up a name in
the template definition, we would do an additional unqualified lookup at
the point of instantiation. This led to incorrectly finding
namespace-scope
Joseph == Joseph S Myers jos...@codesourcery.com writes:
Joseph 2011-05-20 Joseph Myers jos...@codesourcery.com
Joseph * Make-lang.in ($(XGCJ)$(exeext)): Don't explicitly use intl.o
and
Joseph version.o.
Joseph (JCFDUMP_OBJS): Remove errors.o, version.o and intl.o.
There are more issues on Solaris, though:
/vol/gcc/src/hg/trunk/local/libgo/go/net/cgo_bsd.go:13:25: error: expected
package
/vol/gcc/src/hg/trunk/local/libgo/go/net/cgo_bsd.go:13:30: error: expected ';'
or newline after top level declaration
That patch left a stray reference to $(OBJS-common) in a comment. I've
applied this patch as obvious to fix this.
Index: ChangeLog
===
--- ChangeLog (revision 173962)
+++ ChangeLog (working copy)
@@ -1,3 +1,7 @@
+2011-05-20
On 05/20/2011 06:32 PM, Jason Merrill wrote:
Tested x86_64-pc-linux-gnu. Are the library changes OK for trunk?
The changes look fine to me.
Paolo.
Rainer Orth r...@cebitec.uni-bielefeld.de writes:
IRIX might have similar issues, I'm currently checking.
It didn't, but there are two others:
/vol/gcc/src/hg/trunk/local/libgo/runtime/go-setenv.c: In function 'setenv_c':
/vol/gcc/src/hg/trunk/local/libgo/runtime/go-setenv.c:44:3: error:
On 11-05-20 10:17 AM, Richard Guenther wrote:
On Fri, May 20, 2011 at 4:06 PM, Ryan Mansfieldrmansfi...@qnx.com wrote:
There is a crash in vect_is_slp_reduction where use_stmt doesn't get
initialized in the FOR_EACH_IMM_USE_FAST loop.
1718 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, lhs)
Jakub Jelinek ja...@redhat.com writes:
It broke Linux bootstrap too.
netdb.h includes netinet/in.h which has:
enum
{
IPPORT_ECHO = 7,/* Echo service. */
...
IPPORT_RESERVED = 1024,
...
};
and later on has:
#ifdef __USE_XOPEN2K
/* Highest reserved Internet port
Nathan Froyd froy...@codesourcery.com writes:
gcc/
* godump.c (go_format_type): Don't use TYPE_ARG_TYPES.
case FUNCTION_TYPE:
{
- tree args;
+ tree arg_type;
bool is_varargs;
tree result;
+ function_args_iterator iter;
+ bool seen_arg =
On 05/20/2011 01:01 PM, Gabriel Dos Reis wrote:
Ah, I had always assumed that the previous implementation was exploiting
a license given by the standard which says that both contexts should
yield the same resolution, otherwise the program was ill-formed, no diagnostic
required.
I believe that
Hi all,
I just committed a rather obvious fix for a type extension problem
(after regtesting and approval by Tobias on bugzilla):
http://gcc.gnu.org/viewcvs?root=gccview=revrev=173966
Cheers,
Janus
Some code size and timing number (profile use compile) are collected. Summary:
Compile time for profile-use compilation increase for all cases --
this is probably not a big issue as this is for peak performance.
It is more interesting to look at the size numbers. C++ program size
actually
On Fri, May 20, 2011 at 1:01 PM, Jason Merrill ja...@redhat.com wrote:
On 05/20/2011 01:01 PM, Gabriel Dos Reis wrote:
Ah, I had always assumed that the previous implementation was exploiting
a license given by the standard which says that both contexts should
yield the same resolution,
2011/5/20 Richard Guenther richard.guent...@gmail.com:
On Thu, May 19, 2011 at 7:52 PM, Kai Tietz ktiet...@googlemail.com wrote:
To illustrate in which scenario code in tree-ssa-forwprop doesn't help
is binop-tor4.c
w/o this patch we get
foo (int a, int b, int c)
{
int e;
int d;
int
On Fri, May 20, 2011 at 09:32:16AM -0700, Jason Merrill wrote:
G++ has had a long-standing bug with unqualified name resolution in
templates: if we didn't find any declaration when looking up a name in
the template definition, we would do an additional unqualified lookup at
the point of
On 20 May 2011, at 20:30, Joe Buck wrote:
On Fri, May 20, 2011 at 09:32:16AM -0700, Jason Merrill wrote:
G++ has had a long-standing bug with unqualified name resolution in
templates: if we didn't find any declaration when looking up a name in
the template definition, we would do an
On 05/20/2011 03:45 PM, Christopher Jefferson wrote:
I could see the temptation to introduce this as a mandatory warning for a
while, and only add it under -pedantic. However, it might be easier to just
force people to fix their code.
With the patch I just checked in it's an error by default
While thinking about backporting this patch, I noticed a thinko: I was
applying SCALAR_TYPE_P to the expression itself, not its type. Oops.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 17f3af8efbf2fe2483fd922b42b30585a33095a9
Author: Jason Merrill ja...@redhat.com
Date: Fri May 20
In my earlier change to support noexcept in merge_exception_specifiers,
I failed to consider the case of combining noexcept(false) with
noexcept(true) in the context of implicitly declared member functions;
that combination needs to result in noexcept(false), not noexcept(true).
While I was
Split C++ test x1funcstatic.cc into a C test and a C++ test.
Stream out and in various C++ tree members.
Index: gcc/testsuite/ChangeLog.pph
2011-05-20 Lawrence Crowl cr...@google.com
* x1funcstatic.h: Rename to c1funcstatic.h, as contents are C only.
* x1funcstatic.c: Use
http://codereview.appspot.com/4550064/diff/1/gcc/cp/pph-streamer-in.c
File gcc/cp/pph-streamer-in.c (right):
http://codereview.appspot.com/4550064/diff/1/gcc/cp/pph-streamer-in.c#newcode251
gcc/cp/pph-streamer-in.c:251: {
+static VEC(qualified_typedef_usage_t,gc) *
+pph_stream_read_qual_use_vec
Hi,
more or less straightforward work (*), tuple bits enabled by Jason'
fix for c++/49082 (thanks again!). Tested x86_64-linux, committed.
Thanks,
Paolo.
(*) Modulo the long standing issue we have with pair' move constructor
vs std::map: we can't provide it defaulted per the letter of the
[build] More --enable-threads cleanup
http://gcc.gnu.org/ml/gcc-patches/2011-05/msg00059.html
This might well be obvious.
Also, we usually leave those up to the target maintainers, since
they're target specific. But if it's just a correlation between the
script and a list of
We need to propagate DECL_DELETED_FN in duplicate_decls.
Tested x86_64-pc-linux-gnu, applying to trunk and 4.6.
commit 62d3da092d1252e7e0534fe2220c11b2050013e4
Author: Jason Merrill ja...@redhat.com
Date: Fri May 20 18:05:46 2011 -0400
PR c++/49066
* decl.c (duplicate_decls):
As the bug submitter correctly points out, C++0x scoped enums are not
subject to integral promotion like unscoped enums are. Unfortunately,
this represents an ABI change (at least for varargs), so it's
conditional on -fabi-version.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit
Message
On Thu, Apr 28, 2011 at 01:20, Michael Witten mfwit...@gmail.com wrote:
See the following emails for a few inlined patches
to /trunk/gcc/doc/extend.texi (revision 172911):
[1] Docs: extend.texi: Add missing semicolon for consistency
[2] Docs: extend.texi: Remove trailing blanks
80 matches
Mail list logo