This patch to the Go frontend uses the backend interface for interface
types. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu.
Committed to mainline.
Ian
diff -r 37dae2a9c21b go/gogo-tree.cc
--- a/go/gogo-tree.cc Wed May 04 08:38:47 2011 -0700
+++ b/go/gogo-tree.cc Wed May 04
Reviewing some old e-mail...
On Feb 21, 2011, David Daney dda...@caviumnetworks.com wrote:
Everything identical to n32, except Pmode == DImode and
POINTERS_EXTEND_UNSIGNED == true.
Here is a patch that allows me to generate plausible looking assembly
for trivial programs.
Neat!
Just one
Hi!
Committed as obvious:
2011-05-06 Jakub Jelinek ja...@redhat.com
PR debug/48902
* var-tracking.c (prepare_call_arguments): Move else before #endif.
--- gcc/var-tracking.c.jj 2011-03-31 08:51:04.0 +0200
+++ gcc/var-tracking.c 2011-05-06 09:18:50.0
On Thu, May 5, 2011 at 12:19, Andrew Pinski pins...@gmail.com wrote:
Is there a reason why this cannot be an option that someone passes on
the command line of GCC instead of a configure option?
I don't think we ever considered that approach.
That's actually a great idea, I think better for our
On Thu, May 5, 2011 at 6:11 PM, Jeff Law l...@redhat.com wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
I should have included these in the last patch of infrastructure changes.
The main change is create_block_for_threading no longer calls
remove_ctrl_stmt_and_useless_edges and
On Thu, May 5, 2011 at 7:02 PM, Xinliang David Li davi...@google.com wrote:
On Thu, May 5, 2011 at 2:16 AM, Richard Guenther
richard.guent...@gmail.com wrote:
On Thu, May 5, 2011 at 12:19 AM, Xinliang David Li davi...@google.com
wrote:
I can think of some more-or-less obvious high-level
On Thu, May 5, 2011 at 5:42 PM, Richard Earnshaw rearn...@arm.com wrote:
On Thu, 2011-05-05 at 14:51 +0800, Guozhi Wei wrote:
Hi
This is the third part of the fixing for
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47855
This patch contains the length computation/refinement for insn
On Thu, 5 May 2011, Hans-Peter Nilsson wrote:
On Thu, 5 May 2011, Richard Guenther wrote:
On Wed, 4 May 2011, Richard Guenther wrote:
On Wed, 4 May 2011, Eric Botcazou wrote:
Hm. I guess people will scream if something breaks (I can't imagine
what though).
AAAaaarghh! Building
On Sun, 2011-05-01 at 10:30 +0300, Ira Rosen wrote:
Ramana Radhakrishnan ramana.radhakrish...@linaro.org wrote on 07/04/2011
03:16:44 PM:
On 07/04/11 08:42, Ira Rosen wrote:
Hi,
This patch makes both outputs of neon_vzip/vuzp/vtrn_internal
explicitly dependent on both
On 04/22/2011 05:21 PM, Chung-Lin Tang wrote:
Also, instead of testing for XEXP(SET_SRC(PATTERN(i3)),1) == const0_rtx
at the top, it now allows CONST_INT_P(XEXP(SET_SRC(PATTERN(i3)),1)),
tries to adjust it by simplify_compare_const() from the last patch, and
then tests if op1 == const0_rtx. This
Hi!
This patch fixes the -fdefault-integer-8 wrappers, tested on x86_64-linux,
committed to trunk and 4.[654] branches.
2011-05-06 Jakub Jelinek ja...@redhat.com
PR fortran/48894
* fortran.c: Include limits.h.
(TO_INT): Define.
(omp_set_dynamic_8_,
On Thu, 2011-04-21 at 12:23 +0100, Andrew Stubbs wrote:
This patch is a repost of the one I previously posted here:
http://gcc.gnu.org/ml/gcc-patches/2010-12/msg00652.html
As requested, I've broken out the other parts of the original patch, and
those have already been reposted
On Thu, 2011-04-21 at 09:02 +0300, Ira Rosen wrote:
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg02172.html
The last version:
ChangeLog:
* doc/invoke.texi (preferred-vector-size): Document.
* params.h (PREFERRED_VECTOR_SIZE): Define.
* config/arm/arm.c
Hi,
when accumulator transformation is performed on a function like
foo(a)
{
if (a 0)
return 1 + foo (a - 1)
return bla();
}
this becomes
foo(a)
{
int tmp = 0;
while (a 0)
tm = 1 + tmp;
return tmp + bla();
}
Before, bla was a tail-call, but after the optimization, it is
On Fri, 2011-05-06 at 11:35 +0100, Richard Sandiford wrote:
Richard Earnshaw rearn...@arm.com writes:
+/* Implements target hook array_mode_supported_p. */
+
+static bool
+arm_array_mode_supported_p (enum machine_mode mode,
+ unsigned HOST_WIDE_INT nelems)
+{
On 2011/5/6 05:57 PM, Paolo Bonzini wrote:
On 04/22/2011 05:21 PM, Chung-Lin Tang wrote:
Also, instead of testing for XEXP(SET_SRC(PATTERN(i3)),1) == const0_rtx
at the top, it now allows CONST_INT_P(XEXP(SET_SRC(PATTERN(i3)),1)),
tries to adjust it by simplify_compare_const() from the last
Hi,
The attached patch is a fix for PR/48807, which is a segfault when compiling
the bfin compiler. The problem appears to be that we're not checking the
return value of cgraph_local_info for null before attempting to dereference it.
This wasn't a problem before, but now cgraph_local_info
Richard Earnshaw rearn...@arm.com writes:
+/* Implements target hook array_mode_supported_p. */
+
+static bool
+arm_array_mode_supported_p (enum machine_mode mode,
+unsigned HOST_WIDE_INT nelems)
+{
+ if (TARGET_NEON
+ (VALID_NEON_DREG_MODE (mode) ||
Jason Merrill ja...@redhat.com writes:
How about type_dependent_expression_p_push instead?
Like this ? Lightly tested. A full bootstrap and regression test is
under way.
Thanks.
gcc/cp/
PR c++/48574 * class.c (fixed_type_or_null): Use
type_dependent_p_push to test if the
Hi,
the attached patch fixes a problem with trampolines in -m31 -mzarch
mode. The elements of the trampoline are pointer size entities so
they have to depend on -m31/-m64 (UNITS_PER_LONG) but not on
-mesa/-mzarch (UNITS_PER_WORD). Fixed with the attached patch.
Committed to mainline and 4.6.
Hello!
We should avoid to allocate r in *_nointeger instructions.
Constraint modifiers apply to a following character, not to the whole
alternative.
2011-05-06 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (*movdf_internal_nointeger): Apply *
constraint modifier to r.
On Fri, May 6, 2011 at 12:44 PM, Zdenek Dvorak rakd...@kam.mff.cuni.cz wrote:
Hi,
when accumulator transformation is performed on a function like
foo(a)
{
if (a 0)
return 1 + foo (a - 1)
return bla();
}
this becomes
foo(a)
{
int tmp = 0;
while (a 0)
tm = 1 + tmp;
The patch below introduces simple accessors for getting at the class or
the type of the `this' parameter. It hides a couple of TYPE_ARG_TYPES
usages and makes the code slightly more obvious, I think.
Tested on x86_64-unknown-linux-gnu. OK to commit?
-Nathan
gcc/cp/
* cp-tree.h
On 05/06/2011 12:56 PM, Chung-Lin Tang wrote:
I'm not sure why it doesn't allow both?
Paolo
Hi Paolo, I'm not sure I understand your meaning of 'both', but before
this patch, it only tested for == const0_rtx, without any attempt of
other cases.
Now it tests
Hi,
This is the first of two patches to add unaligned-access support to the
ARM backend. This is done somewhat differently to Jie Zhang's earlier
patch:
http://gcc.gnu.org/ml/gcc-patches/2010-12/msg01890.html
In that with Jie's patch, *any* pointer dereference would be allowed to
access
On Fri, 6 May 2011, Jan Hubicka wrote:
Hi,
this patch implements thunks as real cgraph nodes instead of alias nodes. I
am not
entirely happy about it, but I can't come with anything better.
The main problem is that thunks can be seen in two ways:
1) As alternative entry points into
On 05/06/2011 07:08 AM, Dodji Seketeli wrote:
Jason Merrillja...@redhat.com writes:
How about type_dependent_expression_p_push instead?
Like this ? Lightly tested. A full bootstrap and regression test is
under way.
OK.
Jason
On Thu, May 5, 2011 at 10:17 PM, Alan Modra amo...@gmail.com wrote:
My fix for PR44266 using the libcall machinery to ensure we had a
proper stack frame allocated for __tls_get_addr calls sloppily used r3
as the arg to the dummy libcall. This made the call seem to depend on
whatever was in r3
Hi,
This is the second of two patches to add unaligned-access support to
the ARM backend. It builds on the first patch to provide support for
unaligned accesses when expanding block moves (i.e. for builtin memcpy
operations). It makes some effort to use load/store multiple
instructions where
On Fri, May 6, 2011 at 04:55, Richard Guenther
richard.guent...@gmail.com wrote:
On Thu, May 5, 2011 at 7:02 PM, Xinliang David Li davi...@google.com wrote:
2) Support of CallInfo for each callsite. This is an annotation, but
more standardized. The callinfo can be used to record information
On 05/06/2011 07:49 AM, Nathan Froyd wrote:
The patch below introduces simple accessors for getting at the class or
the type of the `this' parameter. It hides a couple of TYPE_ARG_TYPES
usages and makes the code slightly more obvious, I think.
Hmm, when I first read the names I expected them
Denis Chertykov schrieb:
2011/5/2 Georg-Johann Lay a...@gjlay.de:
This is a fix for an optimization flaw when a long value is composed
from byte values.
For -fsplit-wide-types (which is still default for avr) the code is
worse than with -fno-split-wide-types. The code for the test case is
Nathan Froyd schrieb:
On Mon, May 02, 2011 at 05:23:48PM +0200, Georg-Johann Lay wrote:
PR45099 is an extension that gives an error when a fixed register is
needed to pass a parameter to a function.
Because the program will show malfunction when such code is generated,
anyway, I think an
Hi,
On Fri, 6 May 2011, Jan Hubicka wrote:
*** dump_cgraph_node (FILE *f, struct cgraph
*** 1874,1880
if (node-only_called_at_exit)
fprintf (f, only_called_at_exit);
! fprintf (f, \n called by: );
for (edge = node-callers; edge; edge =
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 05/06/11 04:44, Zdenek Dvorak wrote:
Hi,
when accumulator transformation is performed on a function like
foo(a)
{
if (a 0)
return 1 + foo (a - 1)
return bla();
}
this becomes
foo(a)
{
int tmp = 0;
while (a 0)
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 05/06/11 03:57, Paolo Bonzini wrote:
On 04/22/2011 05:21 PM, Chung-Lin Tang wrote:
Also, instead of testing for XEXP(SET_SRC(PATTERN(i3)),1) == const0_rtx
at the top, it now allows CONST_INT_P(XEXP(SET_SRC(PATTERN(i3)),1)),
tries to adjust it
On May 6, 2011, at 5:12 AM, Jan Hubicka wrote:
2) As real functions calling the function they are associated with.
Because backend don't handle alternative entry points, we really
implement
thunks as small functions that usually tail call into the associated
functions after
On Wed, May 4, 2011 at 9:45 AM, Richard Guenther
richard.guent...@gmail.com wrote:
On Wed, May 4, 2011 at 3:45 PM, Kai Tietz ktiet...@googlemail.com wrote:
2011/5/4 Richard Guenther richard.guent...@gmail.com:
On Mon, Apr 18, 2011 at 10:17 AM, Andreas Krebbel
kreb...@linux.vnet.ibm.com wrote:
On Thu, Apr 21, 2011 at 03:02:10PM -0400, Michael Meissner wrote:
In looking at some improvements to the powerpc, we wanted to change the
default
for when a table jump is generated vs. a series of if statements. Now, we
could just add a powerpc specific TARGET_CASE_VALUES_THRESHOLD, but I
On Fri, May 6, 2011 at 1:52 AM, Chris Demetriou c...@google.com wrote:
In theory, a more general warning-text-addition mechanism could be useful.
e.g. a flag that said when outputting a warning about flag 'foo',
output this additional text could be useful.
However, we haven't felt the need to
On 05/05/11 21:18, Tobias Burnus wrote:
Changes:
- Remove (not working) critical functions; a normal coarray of LOCK type
should be used instead. (Stub left in until it is removed the the front
end.)
- Added prototypes and stub implementations for
registering/deregistering coarray (currently
On Fri, May 06, 2011 at 12:21:24PM -0400, Michael Meissner wrote:
On Thu, Apr 21, 2011 at 03:02:10PM -0400, Michael Meissner wrote:
In looking at some improvements to the powerpc, we wanted to change the
default
for when a table jump is generated vs. a series of if statements. Now, we
I want propose a more general solution.
1) Generic Annotation Support for gcc IR -- it is used attach to
application/optimization specific annotation to gimple statements and
annotations can be passed around across passes. In gcc, I only see
HISTOGRAM annotation for value profiling, which is
Honza, what do you think of the patch? It actually fixed a regression.
Thanks,
David
On Wed, May 4, 2011 at 4:40 PM, Xinliang David Li davi...@google.com wrote:
Is this patch ok for trunk?
Allowing relative path in -fprofile-generate= is very useful when
running the program remotely -- the
Ping again. Still no review.
On Fri, Apr 15, 2011 at 7:08 AM, Eric Botcazou ebotca...@adacore.com wrote:
Fix annoying gcov filename handling:
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01380.html
(rs6000) Fix thinko in output_profile_hook:
Backported r172788 and r172837 from trunk to google/main.
2011-05-06 Easwaran Raman era...@google.com
Backport r172837:
* cfgexpand.c (stack_var): Remove OFFSET...
(add_stack_var): ...and its reference here...
(expand_stack_vars): ...and here.
On Fri, May 6, 2011 at 14:22, Easwaran Raman era...@google.com wrote:
Backported r172788 and r172837 from trunk to google/main.
Minor nit:
2011-05-06 Easwaran Raman era...@google.com
Backport r172837:
This needs to be indented by 1 tab.
* cfgexpand.c (stack_var): Remove
This small patch to the Go frontend uses the backend representation for
the string type. Bootstrapped and ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to mainline.
Ian
diff -r 434e7ba47e8d go/types.cc
--- a/go/types.cc Thu May 05 23:32:03 2011 -0700
+++ b/go/types.cc Fri May 06
2011-05-06 Jonathan Wakely jwakely@gmail.com
* doc/invoke.texi (-fuse-linker-plugin): Improve grammar.
I was going to commit a smaller version of this patch as obvious (just
the second of the three hunks in the patch) but I spotted a few other
improvements that could be made. I
Thanks. Fixed them (as well as the same issue in some earlier entries).
-Easwaran
On Fri, May 6, 2011 at 11:27 AM, Diego Novillo dnovi...@google.com wrote:
On Fri, May 6, 2011 at 14:22, Easwaran Raman era...@google.com wrote:
Backported r172788 and r172837 from trunk to google/main.
Minor
On Fri, May 6, 2011 at 14:34, Easwaran Raman era...@google.com wrote:
Thanks. Fixed them (as well as the same issue in some earlier entries).
Great, thanks!
On Fri, May 06, 2011 at 06:30:07PM +0200, Jakub Jelinek wrote:
On Fri, May 06, 2011 at 12:21:24PM -0400, Michael Meissner wrote:
On Thu, Apr 21, 2011 at 03:02:10PM -0400, Michael Meissner wrote:
In looking at some improvements to the powerpc, we wanted to change the
default
for when a
This patch to the Go frontend and to libgo adds more uses of the backend
interface for types. There were some changes to libgo because the code
now uses produces a Go type for maps. Previously the map types were
using size_t, but there is no equivalent to size_t in Go. Go instead
has uintptr_t,
OK to check in?
Ok with me. Thanks!
2011-05-06 Philipp Thomas p...@suse.de
* config/mep/mep.c (mep_validate_vliw): Syntax description
should not be translated.
Hi,
while looking at type merging code I noticed that type pairs can be managed
to be ordered by their UIDs. This save some of hashing overhead in one of
most intensively querried hashes.
Also gimple_lookup_type_leader is hot function that is better to be inlined.
I also wonder, why unionfind
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 05/06/11 03:38, Chung-Lin Tang wrote:
Hi Jeff,
I have verified the patch with a native bootstrap + testsuite run on
powerpc-linux (32-bit), results were clean.
Attached is a single patch with the 1+2 combine parts together, with
comments
This patch makes the MIPS options -march=, -mtune= and -mips use the
.opt Enum facility, completing the avoidance of global state in
mips_handle_option.
The list of possible arguments for those options in mips-tables.opt is
generated by awk code in genopt.sh, which reimplements the logic
formerly
The following patch reverted r173158 from google/main -- -fstrict-enums
provides a better
implementation. The test cases are kept with slight modification.
Bootstrap and tested with related test cases.
Ok for google/main?
2011-05-06 David Li davi...@google.com
Revert r173158.
On Fri, May 6, 2011 at 16:53, David Li davi...@google.com wrote:
The following patch reverted r173158 from google/main -- -fstrict-enums
provides a better
implementation. The test cases are kept with slight modification.
Bootstrap and tested with related test cases.
Ok for google/main?
Here, the problem was that in
(*(first + 1) *first) != false
integral promotion changes the expr to have type int. Then
fold_binary_op_with_conditional_arg wants to change this to
(*(first + 1) *first) ? true : false
without changing the type, so the condition of a ?: has the wrong type
I'm going to bootstrap regtest this on x86_64-unknown-linux-gnu
(with again zero testing coverage ...). The patch fixes the
reported ICE with a cross to cris-elf, more testing is appreciated
(though I guess autotesters will pick it up).
Does it look sane?
Yes, I think so, but...
Index:
In 48911, the constexpr expander wasn't properly dealing with
aggregate/string constant array initializers with omitted elements. We
should build up a value-initialization as needed.
Tested x86_64-pc-linux-gnu, applying to trunk and 4.6.
commit b557b9384f1a6509735c25574f1c1d09703e6252
Author:
I noticed a minor tweak I could make to speed this up and figure I might
as well, even though it shouldn't be a significant component of compile
time.
Tested x86_64-pc-linux-gnu, applying to trunk and 4.6.
commit 4c3e6de3e988799dac490b6eb2b762674b5bb9f8
Author: Jason Merrill ja...@redhat.com
On Wed, May 4, 2011 at 11:07 PM, Tobias Burnus bur...@net-b.de wrote:
Before, scalar coarrays were not supported in the coindex intrinsics as they
did not have - on tree level - cobounds attached to them. This patch adds
them.
Additionally, it fixes the algorithm of this_image, which
Am 07.05.2011 00:50, schrieb H.J. Lu:
On Wed, May 4, 2011 at 11:07 PM, Tobias Burnusbur...@net-b.de wrote:
Is the patch OK for the trunk?
This caused: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=4891
That happens if patches do not get approved in the order in which they
were
Hi,
given that the patch has received feedback and I have weekend for fixing the
fallout, I decided to commit the following version today. It contains fix in
visibility handling of thunks that has shown in Mozilla build.
* cgraph.c (cgraph_add_thunk): Create real function node instead
On Thu, May 5, 2011 at 2:20 AM, Jakub Jelinek ja...@redhat.com wrote:
Hi!
My typed DWARF stack changes apparently broke ia64-hpux and H.J.'s out of
tree x32 target. There are several issues:
1) for SUBREG mem_loc_descriptor's 3rd argument was wrong, found by code
inspection
2)
The interface.c patch is to avoid a strange error (actual argument must
be simply contiguous) which is a bit odd if the actual argument is a
scalar. As the dummy was an array, a rank mismatch would have been the
proper error. - The patch simply suppresses the error message such that
the later
68 matches
Mail list logo