Hello!
> TARGET_BRANCH_PREDICTION_HINTS isn't used at all. This patch removes it.
> Bootstrapped&tested (incl. -m32) on x86_64-unknown-linux-gnu. OK for trunk?
This infrastructure can be used for future targets, so let's leave it as is.
Uros.
Hello,
TARGET_BRANCH_PREDICTION_HINTS isn't used at all. This patch removes it.
Bootstrapped&tested (incl. -m32) on x86_64-unknown-linux-gnu. OK for trunk?
Ciao!
Steven
* final.c (final_forward_branch_p): Remove.
* output.h (final_forward_branch_p): Remove prototype.
* co
Hi David,
I've been auditing rs6000.c after looking at
https://bugzilla.redhat.com/show_bug.cgi?id=837630
We have problems all over the place. :-(
For starters, this code in rs6000_legitimate_offset_address_p.
case TFmode:
...
case TDmode:
case TImode:
if (mode == TFmod
Is this related to the problem described in
http://gcc.gnu.org/ml/gcc-patches/2012-04/msg01511.html ?
David
On Mon, Jun 25, 2012 at 4:43 AM, Dehao Chen wrote:
> During function inlining, a lexical block is added for each cloned
> callee, and source info is attached to this block for addr2line to
On Tue, Jul 10, 2012 at 12:40 AM, Maxim Kuvyrkov wrote:
> On 10/07/2012, at 8:51 AM, Igor Zamyatin wrote:
>
>> Let's try with this patch then. Is it ok for trunk after all
>> appropriate testing is done?
>
> Sorry for the late reply, I've been down with a virus last week.
>
> Uros,
>
> Would you p
Apparently we need to implement DR 1402 in 4.7 in order to fix the
std::pair ABI breakage properly. So here it is: if overload resolution
chooses a non-trivial copy constructor, instead of causing the move
constructor to be deleted, we just don't implicitly declare it.
Tested x86_64-pc-linux-
This looks fine to me, if nobody has any more objections.
Jason
Oleg Endo wrote:
> The attached patch adds a c.torture/compile test case for PR 53886.
> The test case is a slightly modified version of the preprocessed example
> in the PR (added forward decls to remove warnings).
>
> Tested on rev 189361 with
> make check-gcc RUNTESTFLAGS="compile.exp=pr53886.
On 10/07/2012, at 8:51 AM, Igor Zamyatin wrote:
> Let's try with this patch then. Is it ok for trunk after all
> appropriate testing is done?
Sorry for the late reply, I've been down with a virus last week.
Uros,
Would you please approve this patch for i386 provided my review below?
>
> Chang
Oleg Endo wrote:
> This time with the right patch attached. Sorry.
>
> On Mon, 2012-07-09 at 21:09 +0200, Oleg Endo wrote:
>> Hello,
>>
>> The attached patch is a fix for PR 53886, as suggested in the PR.
>> I've also replaced another unchecked usage of 'PATTERN (NEXT_INSN
>> (PREV_INSN (insn))
This time with the right patch attached. Sorry.
On Mon, 2012-07-09 at 21:09 +0200, Oleg Endo wrote:
> Hello,
>
> The attached patch is a fix for PR 53886, as suggested in the PR.
> I've also replaced another unchecked usage of 'PATTERN (NEXT_INSN
> (PREV_INSN (insn)))' in 'find_barrier', just in
Committed as posted.
Thanks.
On Sun, Jul 1, 2012 at 7:33 AM, Gabriel Dos Reis
wrote:
> On Thu, Jun 28, 2012 at 12:50 PM, Sterling Augustine
> wrote:
>> The enclosed patch adds a fix for the pubnames anonymous namespaces contained
>> within named namespaces, and adds an extensive test for the va
Ok. Do you have specific comments on the patch?
thanks,
David
On Sun, Jul 8, 2012 at 4:20 AM, Jason Merrill wrote:
> On 07/07/2012 08:38 PM, Xinliang David Li wrote:
>>>
>>> It seems to me that what you have here are target-specific attributes
>>> that
>>> affect the signature of a function su
http://codereview.appspot.com/5752064/diff/51001/gcc/cgraph.c
File gcc/cgraph.c (right):
http://codereview.appspot.com/5752064/diff/51001/gcc/cgraph.c#newcode1282
gcc/cgraph.c:1282: is needed as the address can be used to do an
indirect call. */
Extend the comment here.
http://codereview.appsp
Let's try with this patch then. Is it ok for trunk after all
appropriate testing is done?
ChangeLog:
2012-07-09 Sergey Melnikov
* config/i386/i386.md (stack_protect_set): Disable the pattern
for Android since Android libc (bionic) does not provide random
value for sta
Tom de Vries wrote:
> 2012-07-06 Tom de Vries
> Richard Guenther
>
> * tree-ssa-ccp.c (optimize_unreachable): New function.
> (execute_fold_all_builtins): Use optimize_unreachable to optimize
> BUILT_IN_UNREACHABLE. Don't optimize after BUILT_IN_UNREACHABLE.
>
>
Since output_addr_const() shows pretty hot in the compiler, I reinstated
the fprint_w() call in place of fprintf().
This patch relies on two things: 2's complement representation for
negative int and that HOST_WIDE_INT is at least as large type as long for
all platforms (will it always be?).
[- gdb, + gdb-patches]
On Mon, Jul 9, 2012 at 11:51 AM, Steven Bosscher wrote:
> On Mon, Jul 9, 2012 at 8:10 PM, Doug Evans wrote:
>> Hi.
>>
>> filename_seen in gdb does a linear search, this patch changes it
>> to use a hash table.
>>
>> Ok to check in?
>
> Why not just use htab_hash_string? Th
Hello,
The attached patch adds a c.torture/compile test case for PR 53886.
The test case is a slightly modified version of the preprocessed example
in the PR (added forward decls to remove warnings).
Tested on rev 189361 with
make check-gcc RUNTESTFLAGS="compile.exp=pr53886.c
--target_board=sh-s
Hello,
The attached patch is a fix for PR 53886, as suggested in the PR.
I've also replaced another unchecked usage of 'PATTERN (NEXT_INSN
(PREV_INSN (insn)))' in 'find_barrier', just in case.
Tested on rev 189361 with
make -k check RUNTESTFLAGS="--target_board=sh-sim
\{-m2/-ml,-m2/-mb,-m2a/-mb,-
On Mon, Jul 9, 2012 at 8:10 PM, Doug Evans wrote:
> Hi.
>
> filename_seen in gdb does a linear search, this patch changes it
> to use a hash table.
>
> Ok to check in?
Why not just use htab_hash_string? The file name canonicalization can
be put in gdb itself.
Ciao!
Steven
Hi.
filename_seen in gdb does a linear search, this patch changes it
to use a hash table.
Ok to check in?
I couldn't think of a good reason to put filename_hash,filename_eq in gdb,
and I like placing them close to where hashtab.c and filename_cmp are defined.
I also couldn't think of a sufficien
Ping ..
On Wed, Jul 4, 2012 at 8:01 AM, Xinliang David Li wrote:
> Comment?
>
> David
>
> On Mon, Jul 2, 2012 at 4:30 PM, Xinliang David Li wrote:
>> I extended the patch a little so that the option can be used to set
>> multiple stack reuse levels: -fstack-reuse=[all|name_vars|none]
>>
>> all:
On 07/02/2012 09:47 PM, Jason Merrill wrote:
FAIL: g++.dg/template/friend13.C -std=c++11 (test for excess errors)
OK, the problem here is that we need to look up Outer::Inner in order to
generate the declaration of foo, and we can't check whether
foo is a friend of Outer until after we've com
On 7/4/12, Jason Merrill wrote:
> On 07/03/2012 04:37 PM, Lawrence Crowl wrote:
>> On 7/1/12, Jason Merrill wrote:
>>> As discussed, I would say that RTTI is currently not permitted
>>> but could be added later.
>>
>> But isn't "could be added later" always true? Other folks have
>> objected to
On 07/06/2012 11:27 AM, Janis Johnson wrote:
> Between June 12 and June 28 I checked in lots of GCC testsuite patches
> to individual tests and to test infrastrure files to eliminate
> non-unique lines in test summaries, as well as a few related cleanups.
> May I backport those to GCC 4.7?
It's be
On Mon, Jul 9, 2012 at 3:23 PM, wrote:
> I also filed a bug here:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53547
Why are you compiling from a different source code base with
profile-use than with profile-generate? I'm surprised this works at
all, I would have expected GCC to just bail out.
We also need to handle NULLPTR_TYPE in type_contains_placeholder. While
I was at it, I added a NULLPTR_TYPE case to type_hash_eq just in case.
Tested x86_64-pc-linux-gnu, applying to trunk and 4.7.
commit 05eb87b7123c9161268a0ef89b160c489a3d4935
Author: Jason Merrill
Date: Mon Jul 9 15:39:06
For aggregate projects and aggregate libraries we may have multiple
references of the same project. Only one is properly initialized
and can be iterated to get the corresponding sources.
Tested on x86_64-pc-linux-gnu, committed on trunk
2012-07-09 Pascal Obry
* prj.adb (For_Every_Proj
On 2012/04/24 22:14:17, asharif wrote:
Ping?
Ping.
I also filed a bug here:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53547
Here is the formatted ChangeLog in case you want a summary of what I
did:
* gcc/ipa-inline.c (edge_badness): Make sure profile is valid before
using it to compute
Prevent mispelling error messages with internal attribute names
CPU, Dispatching_Domain and Interrupt_Priority.
For example:
-- Source --
pragma Ada_2012;
procedure P is
X : Integer;
for X'CPU use 3; -- unrecognized attribute (internal to GNAT)
for X'CPUX us
This is the first step in dealing with the undesirable linkage between
restrictions at compile/bind time (Rident), and restrictions accessible
at run-time (System.Rident). This update stops the compiler from withing
system.rident. No test, since no functional change.
The general problem here is th
This patch reorganizes the Lock_Free restrictions. The compiler issues a
warning whenever a Priority aspect/pragma is given while the lock-free
implementation has been forced by an aspect/pragma.
-- Source --
package T is
protected type P1
with Lock_Free
is
The run time contained code which tried to rearrange the ready queue
in order to achieve Annex D semantics. This code was not very efficient,
and it did not work with multiprocessors. We consider that we should
live with standard Windows semantics.
Tested on x86_64-pc-linux-gnu, committed on trunk
This patch fixes an internal overflow when computing the alignment of a record
component whose size is improperly specified as larger than the largest integer
value.
Compiling the following must yield:
prob.adb:6:38: warning: 4294967168 bits of "Unsigned_Integer_4" unused
prob.adb:13:45: size for
AI05-0283 requires Ada.Streams.Stream_IO to be declared with a pragma
Preelaborate, so that it can be used by other preelaborated units as
well as by Remote_Types and Remote_Call_Interface units. The pragma
has been added to that package along with several other GNAT library
units on which it depe
Required when the unwinder is not the libgcc one (eg: on HP-UX).
No functionnal change.
Tested on x86_64-pc-linux-gnu, committed on trunk
2012-07-09 Tristan Gingold
* a-exexpr-gcc.adb (CleanupUnwind_Handler): Now imported from
raise-gcc.c
* raise-gcc.c (__gnat_cleanupu
This fixes PR53898 by mimicing the other graphite files - leave the
GCC includes included always
to avoid an empty CU on graphite-disabled builds.
Committed as obvious.
Richard.
p
Description: Binary data
On Mon, Jul 9, 2012 at 12:57 PM, Steven Bosscher wrote:
> Hello,
>
> How annoying that it was decided at some point to not always group
> switch case labels in tree-cfgcleanup! This PR53887 and PR53881 before
> it are both cases where the switch lowering to bit tests ICEs because
> the case labels
On Fri, Jun 29, 2012 at 11:34 AM, Jakub Jelinek wrote:
>
> The recent mksysinfo.sh change for glibc 2.16 bits/resource.h unfortunately
> doesn't work, because the sed invocation two lines earlier removes the {}s
> this sed command is looking for.
> The following makes gcc 4.7 as well as trunk buil
Hello,
How annoying that it was decided at some point to not always group
switch case labels in tree-cfgcleanup! This PR53887 and PR53881 before
it are both cases where the switch lowering to bit tests ICEs because
the case labels haven't been grouped. For PR53887 there is a case
label that shares
The declarations of Remote_Types and Remote_Call_Interface packages are
allowed to depend on preelaborated packages, as specified in AI05-206,
as long as the dependence is made using a private with_clause. Before
this fix, the compiler was disallowing such a dependence for RCI packages,
and for Rem
This new routine interfaces through all sources needed to interface
to a library. All the specs are handled and also the bodies when
needed (if there is a pragma inline of the spec is generic for
example).
Tested on x86_64-pc-linux-gnu, committed on trunk
2012-07-09 Pascal Obry
* prj-
The following test illustrates the exemple of an aspect Atomic that applies to
a component of a record type.
-- Source --
procedure Cutdown is
type Rec is record
Comp : aliased Integer;
Acomp : aliased Integer with Atomic => Atom;
end record;
Atomi
Expressions of stream aspects and also aspects Variable_Indexing,
Constant_Indexing, Default_Iterator and Iterator_Element are names that can be
overloaded. This patch checks at the end of the declarations there is an
interpretation of these names that matches the entity obtained at the freeze
poin
This variable is no longer used, so is removed.
Tested on x86_64-pc-linux-gnu, committed on trunk
2012-07-09 Joel Brobecker
* exp_dbug.ads (No_Dollar_In_Label): Delete.
Index: exp_dbug.ads
===
--- exp_dbug.ads(re
As this unit is always compiled for the runtime, remove the #ifndef IN_RTS
parts. Also update comments and clean up code.
No functional change.
Tested on x86_64-pc-linux-gnu, committed on trunk
2012-07-09 Tristan Gingold
* raise-gcc.c (db_indent): Simplify style, improve comments.
>>
>> 1) If -fira_loop_pressure is enabled, it reduces ~24% invariant motions in my
>tests. But it does not help on total code size. Seams there is issue to update
>the
>"regs_needed" after moving an invariant out of the loop (My benchmark logs
>show ~73% cases have more than one invariants moved)
On 9 Jul 2012, at 09:11, Iain Sandoe wrote:
>> crt3.o: $(srcdir)/config/darwin-crt3.c
> regstrapped (all+ada+objc++) on i686-darwin9 with no regressions.
.. but, now I re-check, crt3 is only used on Darwin 8 and earlier;
That will take somewhat longer to check, machine availability is an issue
On Sun, Jul 8, 2012 at 9:44 PM, Steven Bosscher wrote:
> Hello,
>
> A lot of optab_handler checks are against CODE_FOR_nothing, which is
> currently a target dependent number, the last value in "enum
> insn_code". This made target reinitializing slow, so Richard S.
> changed the optab_handler inde
On Sun, Jul 8, 2012 at 9:29 AM, Dehao Chen wrote:
> Hi, Richard,
>
> A test is added. Is it ok?
Ok.
Thanks,
RIchard.
> Thanks,
> Dehao
>
> gcc/testsuite/ChangeLog
>
> 2012-07-08 Dehao Chen
>
> * gcc.dg/debug_info_inline.c: New test.
>
> Index: gcc/testsuite/gcc.dg/debug_info_inline.c
Hi Steven,
On 7 Jul 2012, at 16:34, Steven Bosscher wrote:
> On Sat, Jul 7, 2012 at 5:06 PM, H.J. Lu wrote:
>> Are you sure this the right patch? -fno-tree-dominator-opts is still here.
>
> I am sure it is not the right patch :-)
> Thanks!
>
> Index: libgcc/config/t-darwin
> ==
On Fri, Jul 6, 2012 at 6:36 PM, Tom de Vries wrote:
> On 06/07/12 13:01, Richard Guenther wrote:
>> On Thu, Jul 5, 2012 at 8:45 PM, Tom de Vries wrote:
>>> On 05/07/12 15:30, Michael Matz wrote:
Hi,
On Thu, 5 Jul 2012, Tom de Vries wrote:
> The asserts allow the return res
On 07/09/2012 09:07 AM, Andreas Schwab wrote:
You also get 0 for a partial UCN, which is significant for
forms_identifier_p.
Right. Since 0 is a valid UCN, we can't use the return value both for
the UCN and the predicate; I'd suggest returning the UCN by pointer
parameter.
Jason
Steven Bosscher writes:
> In any case, it's valid for C++11. Jason modified charset.c to
> implement the C++11 change and handle the differences between C99 and
> C++11 (*) but I think he overlooked the two lines at the bottom that
> convert a 0 result to 1.
You also get 0 for a partial UCN, whi
55 matches
Mail list logo
