On Thu, Oct 4, 2012 at 5:30 AM, Vladimir Makarov wrote:
> I was going to look at this code too but I was interesting in generation of
> less points and live ranges. It is strange that in my profiles,
> remove_some_program_points_and_update_live_ranges takes 0.6% of compiler
> time on these huge t
On Oct 3, 2012, Jakub Jelinek wrote:
> basically there is a non-addressable parameter in stack slot, and
> vt_canon_true_dep -> canon_true_dependence thinks an argument push
> insn might alias with it, because it doesn't have a MEM_EXPR and
> ao_ref_from_mem fails.
I have a pending (still unrev
On 12-10-03 11:35 AM, Steven Bosscher wrote:
On Wed, Oct 3, 2012 at 4:56 PM, Vladimir Makarov wrote:
On 12-10-03 3:13 AM, Steven Bosscher wrote:
On Tue, Oct 2, 2012 at 3:42 PM, Richard Sandiford
wrote:
+/* Compress pseudo live ranges by removing program points where
+ nothing happens. Com
Hi,
I've committed the patch below to fix sh64-elf build failure.
SHmedia and SHcompact using call cookie require special return insns
and will require an extra work to enable simple_return. I simply
disable it for these targets ATM.
Regards,
kaz
--
2012-10-04 Kaz Kojima
* co
Oleg Endo wrote:
> This adds the two common built-in functions __builtin_thread_pointer and
> __builtin_set_thread_pointer to the SH port.
> I've done it in a way so that hopefully it can be transitioned to target
> independent thread pointer built-ins easily, as suggested by Richard a
> while ago
@@ -1115,7 +1118,8 @@ static const struct attribute_spec rs600
{ NULL,0, 0, false, false, false, NULL, false }
};
-#ifndef MASK_STRICT_ALIGN
+#ifndef OPTION_MASK_STRICT_ALIGN
+#define OPTION_MASK_STRICT_ALIGN 0
#define MASK_STRICT_ALIGN 0
#endif
#ifndef TARGET_PROFILE_KERNEL
Why
The following patch is still awaiting review:
2011-09-19 J"orn Rennecke
* genoutput.c (process_template): Process '*' in '@' alternatives.
* doc/md.texi (node Output Statement): Provide example for the above.
http://gcc.gnu.org/ml/gcc-patches/2012-09/msg01422.html
On Wed, Oct 3, 2012 at 8:29 PM, Mike Stump wrote:
> On Oct 3, 2012, at 4:49 PM, David Edelsohn wrote:
>> Another suggestion from Segher is "_F.", which is both reserve and
>> cannot conflict with C/C++ because identifiers cannot contain ".".
>
> The problem with that are machines that don't like
On Oct 3, 2012, at 4:49 PM, David Edelsohn wrote:
> Another suggestion from Segher is "_F.", which is both reserve and
> cannot conflict with C/C++ because identifiers cannot contain ".".
The problem with that are machines that don't like "." in identifiers (aka
NO_DOT_IN_LABEL)… so one just ha
Another suggestion from Segher is "_F.", which is both reserve and
cannot conflict with C/C++ because identifiers cannot contain ".".
- David
On Tue, 2 Oct 2012, Tom G. Christensen wrote:
> Latest results for 4.5.x
Thanks, Tom!
Gerald
Maybe something like the following:
Index: trans-decl.c
===
--- trans-decl.c(revision 192019)
+++ trans-decl.c(working copy)
@@ -1097,9 +1097,9 @@
/* Also prefix the mangled name. */
if (sym->module)
-
On Tue, 2 Oct 2012, Tom G. Christensen wrote:
> Latest results for 4.6.x
Thanks, applied. A bit unusual to see submissions of older versions
of one branch at this time, and from one and the same submitter...
Gerald
For C and C++, identifiers beginning with underscore and upper case
letter or with two underscores are reserve to the implementation. C++
uses _Z for mangling.
Maybe Fortran could prepend "_F". Something beginning with an
underscore seems like a much better choice, given the rules about
reserved
On Oct 3, 2012, at 1:47 PM, Marc Glisse wrote:
> did you consider making the size of wide_int a template parameter, now that
> we are using C++? All with a convenient typedef or macro so it doesn't show.
> I am asking because in vrp I do some arithmetic that requires 2*N+1 bits
> where N is the
On Wed, Oct 3, 2012 at 6:28 PM, Segher Boessenkool
wrote:
> It was used for old POWER only, which has been removed.
>
> Tested etc.; okay to apply?
>
>
> Segher
>
>
> 2012-10-04 Segher Boessenkool
>
> gcc/
> * config/rs6000/rs6000.c (print_operand) ['A']: Delete.
Okay.
Thanks, David
Sorry, one more time with the right file contents.
On 10/2/12, Richard Guenther wrote:
> You are changing a hashtable used by fold checking, did you test
> with fold checking enabled?
One small thinko fixed. Patch passes tests.
> The cfg.c, dse.c and hash-table.h parts are ok for trunk, I'll
>
On 10/2/12, Richard Guenther wrote:
> On Mon, 1 Oct 2012, Lawrence Crowl wrote:
>
>> Change more non-GTY hash tables to use the new type-safe template hash
>> table.
>> Constify member function parameters that can be const.
>> Correct a couple of expressions in formerly uninstantiated templates.
>
On Wed, 3 Oct 2012, Nick Clifton wrote:
> PS. I have even remembered to include a patch to the html
> documentation as well!
Yes, I was going to mention this appreciatively. ;-)
A minor change I applied on top is the following, adding a closing
and a dash in command-line.
Thanks,
Gerald
Index
It was used for old POWER only, which has been removed.
Tested etc.; okay to apply?
Segher
2012-10-04 Segher Boessenkool
gcc/
* config/rs6000/rs6000.c (print_operand) ['A']: Delete.
---
gcc/config/rs6000/rs6000.c | 11 ---
1 files changed, 0 insertions(+), 11 deletions(
On Oct 3, 2012, at 12:04 PM, Jason Merrill wrote:
> This patch fixes a couple of Darwin issues I noticed with a cross-compiler:
>
> 1) Adds a couple of consts to avoid const-correctness errors.
> 2) Treats visibility "internal" like "hidden" rather than like "default".
> The documentation says
Hello,
Since the -mieee behavior has been fixed, is enabled by default on SH
and the additional flags in libgcc can be removed.
OK?
Cheers,
Oleg
libgcc/ChangeLog:
PR target/33135
* config/sh/t-sh (HOST_LIBGCC2_CFLAGS): Delete.
* config/sh/t-netbsd (HOST_LIBGCC2_CFLAGS):
This patch to libbacktrace fixes the overflow test for the leb128
reading routines. What matters is not the shift after the loop, but the
shift within the loop. This also removes the setting of unit_buf.start
in build_address_map, which was simply wrong and was causing the error
message to print
Hi all,
the attached patch implements an F08 feature, which allows to
distinguish two specific procedures in a generic interface, based on
the POINTER and ALLOCATABLE attribute of their arguments.
In addition to this, the patch fixes a bug in rejecting data actual
arguments passed to procedure fo
i have already converted the vrp code, so i have some guess at where you
are talking about. (of course correct me if i am wrong).
in the code that computes the range when two variables are multiplied
together needs to do a multiplication that produces a result that is
twice as wide as the inp
On Wed, 2012-10-03 at 23:21 +0200, Oleg Endo wrote:
> testsuite/ChangeLog:
>
> PR target/54760
> * gcc.target/sh/pr54706-1.c: New.
> * gcc.target/sh/pr54706-2.c: New.
> * gcc.target/sh/pr54706-3.c: New.
Obviously there is a typo in the file names for the test cases.
Attach
Hello,
This adds the two common built-in functions __builtin_thread_pointer and
__builtin_set_thread_pointer to the SH port.
I've done it in a way so that hopefully it can be transitioned to target
independent thread pointer built-ins easily, as suggested by Richard a
while ago:
http://gcc.gnu.org
On 09/28/2012 11:36 AM, Jeff Law wrote:
On 09/28/2012 09:21 AM, Vladimir Makarov wrote:
On 12-09-28 4:43 AM, Steven Bosscher wrote:
On Fri, Sep 28, 2012 at 12:57 AM, Vladimir Makarov
wrote:
LRA outputs a lot debug information about insns. I found that using
slim
insn/rtl presentation helps a
The following patch takes comments and proposals from Jeff Law's and
Richard Sandiford's reviews of LRA.
The patch was successfully bootstraped on x86/x86-64.
Commited as rev. 192050.
2012-10-03 Vladimir Makarov
* sched-vis.c (print_value_slim): Fix the comment.
* targhooks
On Tue, 2 Oct 2012, Tom G. Christensen wrote:
> Latest results for 4.4.x
Thanks, Tom. Applied.
Gerald
Found using http://scan5.coverity.com/
Build on x86-64-gnu-linux with C/C++/Fortran. I will now do an
all-language build/regtest.
OK when it passes?
(Note to the save_string call: I reduced it by 2: The "+1" in the call
makes it long (out of bounds) and the "+1" in temp_filename_length is
no
David,
David Edelsohn wrote:
I am not sure why you chose a period and how best to correct this.
Well, in principle any name which the user cannot enter would do. (Not
enter: At least not as valid Fortran identifier.)
The reason for choosing "." is that is used elsewhere in
gfortran for su
On Wed, 3 Oct 2012, Kenneth Zadeck wrote:
The patch defines a new datatype, a 'wide_int' (defined in
wide-int.[ch], and this datatype will be used to perform all of the
integer constant math in the compiler. Externally, wide-int is very
similar to double-int except that it does not have the lim
Tobias,
Sorry to return to this patch so long after the discussion, but I have
been investigating testsuite failures on AIX and this patch introduced
a number of Fortran failures.
+ if (sym->module)
+ name = gfc_get_string (".__%s_MOD_%s", sym->module, sym->name);
+ else
+ n
On Wed, Oct 03, 2012 at 01:09:27PM -0700, Dehao Chen wrote:
> Thanks for the comments. The patch was updated as attached.
Ok, thanks.
Jakub
Thanks for the comments. The patch was updated as attached.
Dehao
On Wed, Oct 3, 2012 at 11:46 AM, Jakub Jelinek wrote:
> On Wed, Oct 03, 2012 at 11:26:09AM -0700, Dehao Chen wrote:
>> @@ -6340,6 +6341,20 @@ move_block_to_fn (struct function *dest_cfun, basi
>> SET_USE (use, replace_ss
On Wed, 3 Oct 2012, Gabriel Dos Reis wrote:
On Tue, Oct 2, 2012 at 1:53 PM, Marc Glisse wrote:
* include/c_std/cstdlib (abs(long long)): Define with
__builtin_llabs when we have long long.
(abs(__int128)): Define when we have __int128.
This change is OK
Thanks, I'
The following patch was submitted as approved by Jeff Law and Paolo Bonzini.
The patch was successfully boostraped on x86/x86-64.
Committed as rev. 192048.
2012-10-03 Vladimir Makarov
* reginfo.c (max_regno_since_last_resize): New.
(reg_preferred_class, reg_alternate_class):
Build on x86-64-linux with C/C++/Fortran. I will now do an all-language
build/regtest.
OK when it passes?
Tobias
2012-10-03 Tobias Burnus
* files.c (read_file_guts, _cpp_save_file_entries): Free memory
before returning.
* lex.c (warn_about_normalization): Ditto.
* mkdeps.c (deps_save): D
This patch fixes a couple of Darwin issues I noticed with a cross-compiler:
1) Adds a couple of consts to avoid const-correctness errors.
2) Treats visibility "internal" like "hidden" rather than like
"default". The documentation says that internal is hidden +
processor-specific semantics, so
> thanks for the update!
OK, I will review the patch tomorrow then. It is a good incremental step. I
would certainly like to see the gimple expansion in 4.8 however.
Honza
On Wed, Oct 03, 2012 at 11:26:09AM -0700, Dehao Chen wrote:
> @@ -6340,6 +6341,20 @@ move_block_to_fn (struct function *dest_cfun, basi
> SET_USE (use, replace_ssa_name (op, d->vars_map, dest_cfun->decl));
> }
>
> + for (i = 0; i < EDGE_COUNT (bb->preds); i++)
> + {
> +
Adding myself as write after approval.
Index: MAINTAINERS
===
--- MAINTAINERS (revision 191941)
+++ MAINTAINERS (working copy)
@@ -342,6 +342,7 @@ R. Kelley Cook
kc...@gcc.gnu.org
Christian Cornelssen cc...
Hi,
Attached patch fixes PR54782. phi_arg_location is not correctly
updated in move_block_to_fn. This patch fixes the problem.
Bootstrapped and passed gcc regression tests on x86.
Okay for trunk?
Thanks,
Dehao
gcc/ChangeLog:
2012-10-03 Dehao Chen
PR middle-end/54782
* tree-cfg.c (move_bloc
The ARCompact architecture has some pipelining features that result in
the vanilla branch shortening not always converging.
Moreover, there are some short, complex branch instructions with very short
offsets; replacing them with a multi-insn sequence when the offset doesn't
reach makes the code s
;; Return 1 if op is an operand that can be loaded via the GOT.
-;; or non-special register register field no cr0
(define_predicate "got_operand"
(match_code "symbol_ref,const,label_ref"))
Most likely should be submitted and committed (as obvious) separately.
Yes, will do.
switch
thanks for the update!
David
On Wed, Oct 3, 2012 at 10:37 AM, Steven Bosscher wrote:
> On Wed, Oct 3, 2012 at 6:12 PM, Xinliang David Li wrote:
>> What is the status of switch expansion GIMPLE rewrite? If it is not
>> planned for 4.8, It will be desirable to include this fix into trunk.
>
> I c
On Tue, Oct 2, 2012 at 12:35 PM, H.J. Lu wrote:
> Hi,
>
> This patch checks SSE and YMM state support for -march=native. Tested
> on Linux/x86-64. OK to install?
>
> Thanks.
>
>
> H.J.
> ---
> 2012-10-02 H.J. Lu
>
> PR target/54741
> * config/i386/driver-i386.c (XCR_XFEATURE_
On Wed, Oct 3, 2012 at 6:59 AM, James Lemke wrote:
> Ping..
> @@ -847,7 +1106,6 @@
> && (DEFAULT_ABI != ABI_AIX || SYMBOL_REF_FUNCTION_P
> (op))")))
>
> ;; Return 1 if op is an operand that can be loaded via the GOT.
> -;; or non-special register register field no cr0
>
On Wed, Oct 3, 2012 at 6:12 PM, Xinliang David Li wrote:
> What is the status of switch expansion GIMPLE rewrite? If it is not
> planned for 4.8, It will be desirable to include this fix into trunk.
I could work on it for GCC 4.8 (there's not a lot of work left to be
done for it now) but we haven
On Tue, Oct 2, 2012 at 1:53 PM, Marc Glisse wrote:
>> See what we did in c/cmath and c_global/cmath.
>
>
> Note that llabs is quite different from asin.
Is asin the function you took out of c/cmath?
> __builtin_llabs generates an
> ABS_EXPR, which will later be expanded either to a special ins
On Wed, Oct 03, 2012 at 09:54:45AM -0700, Lawrence Crowl wrote:
> On 10/2/12, Ian Lance Taylor wrote:
> > On Oct 2, 2012 Lawrence Crowl wrote:
> > > On 10/2/12, Richard Guenther wrote:
> > > > You are changing a hashtable used by fold checking, did you
> > > > test with fold checking enabled?
>
On 10/2/12, Ian Lance Taylor wrote:
> On Oct 2, 2012 Lawrence Crowl wrote:
> > On 10/2/12, Richard Guenther wrote:
> > > You are changing a hashtable used by fold checking, did you
> > > test with fold checking enabled?
> >
> > I didn't know I had to do anything beyond the normal make check.
> >
Hi all,
here is a small patch for a wrong-code regression with polymorphic
allocation. The problem is that we falsely detect the allocation
variable to be a polymorphic array (although it is a scalar). For
further details see the PR, in particular comment 4.
Regtested on x86_64-unknown-linux-gnu.
What is the status of switch expansion GIMPLE rewrite? If it is not
planned for 4.8, It will be desirable to include this fix into trunk.
It also helps set up a good base line to test against regression.
thanks,
David
On Tue, Oct 2, 2012 at 6:09 PM, Easwaran Raman wrote:
> Hi,
> This patch pr
Hi Guys,
I am applying the patch below to the RX backend. It adds support for
issuing warning messages when multiple fast interrupt routines are
defined in the same compilation unit. The RX only supports one fast
interrupt so it is probably a mistake to define more than one. If the
pr
On Sat, 15 Sep 2012, Ian Lance Taylor wrote:
>> 2012-09-02 Mark Kettenis
>>
>> * config.gcc (*-*-openbsd4.[3-9]|*-*-openbsd[5-9]*): Set
>> default_use_cxa_atexit to yes.
> This is OK.
I committed this to trunk and plan on doing so for the 4.7 branch
as well (so that OpenBSD can
On Wed, Oct 3, 2012 at 4:56 PM, Vladimir Makarov wrote:
> On 12-10-03 3:13 AM, Steven Bosscher wrote:
>>
>> On Tue, Oct 2, 2012 at 3:42 PM, Richard Sandiford
>> wrote:
+/* Compress pseudo live ranges by removing program points where
+ nothing happens. Complexity of many algorith
Hi Ian,
Can't you just keep a list of the decls for which you have issued the
warning?
Yes - that would work too. In fact I agree that this would be cleaner
solution in my particular case. I'll create a new patch...
How many decls do you expect to be on that list?
Not very many. Mayb
On 12-10-03 3:13 AM, Steven Bosscher wrote:
On Tue, Oct 2, 2012 at 3:42 PM, Richard Sandiford
wrote:
+/* Compress pseudo live ranges by removing program points where
+ nothing happens. Complexity of many algorithms in LRA is linear
+ function of program points number. To speed up the code
On 12-10-03 6:53 AM, Steven Bosscher wrote:
On Mon, Oct 1, 2012 at 1:05 PM, Steven Bosscher wrote:
Hello,
This patch uses the libiberty new-like operators instead of using
xmalloc/xrealloc.
It also adds timevars for the main LRA phases, and it fixes a warning
suggesting a space before a ';' in
On 12-10-03 9:01 AM, Jakub Jelinek wrote:
Hi!
While bootstrapping the PR54519 patches last night, I've noticed a
comparison failure on i686-linux. The problem was -g vs. -g0
swapping two instructions in tree-inline.o, caused by find_modifiable_mems
scanning [head, tail) instead of [head, tail]
On Wed, Oct 3, 2012 at 6:31 AM, Nick Clifton wrote:
>
> Is there a target specific way to add a flag bit to a function decl ?
>
> For the RX port I want to be able to mark a function decl for which I
> have issued a warning message. Unfortunately I could find no easy way
> to do this othe
Hi Guys,
Is there a target specific way to add a flag bit to a function decl ?
For the RX port I want to be able to mark a function decl for which I
have issued a warning message. Unfortunately I could find no easy way
to do this other than stealing one of the generic bits (no_limit_stac
Hi!
While bootstrapping the PR54519 patches last night, I've noticed a
comparison failure on i686-linux. The problem was -g vs. -g0
swapping two instructions in tree-inline.o, caused by find_modifiable_mems
scanning [head, tail) instead of [head, tail] sequence of instructions.
Without -g0 on tre
Hi!
With the PR54519 patch I've just posted, I've noticed, I've noticed on the
same testcase from yesterday's IRC:
static inline void foo (int x, int y) { asm volatile ("nop"); }
static inline void bar (int z) { foo (z, 0); foo (z, 1); }
int main ()
{
bar (0);
bar (1);
return 0;
}
that while
On Tue, Sep 11, 2012 at 03:59:56PM +0200, Jakub Jelinek wrote:
> As discussed in the PR, right now we do a very bad job for debug info
> of partially inlined functions (both when they are kept only partially
> inlined, or when partial inlining is performed, but doesn't seem to be
> useful and foo.p
OK.
Jason
Hi Vlad,
Some comments on lra-spills.c and lra-coalesce.c.
> + The pass creates necessary stack slots and assign spilled pseudos
> + to the stack slots in following way:
s/assign/assigns/
> + (or insn memory constraints) might be not satisfied any more.
s/might be not/might not be/
> +
Il 30/09/2012 11:02, Richard Sandiford ha scritto:
> Uros Bizjak writes:
>> On Thu, Sep 27, 2012 at 8:20 PM, Jakub Jelinek wrote:
>>> On Thu, Sep 27, 2012 at 08:04:58PM +0200, Uros Bizjak wrote:
After some off-line discussion with Richard, attached is v2 of the patch.
2012-09-27 U
On Wed, Oct 03, 2012 at 12:21:02PM +0300, Laurynas Biveinis wrote:
> Hello Basile -
>
> > 2012-10-02 Basile Starynkevitch
> >
> > * gengtype.c (walk_type): Emit mark_hook when inside a
> > struct of a union member.
>
> Can you send me off-list the gengtype output before and a
On Mon, Oct 1, 2012 at 1:05 PM, Steven Bosscher wrote:
> Hello,
>
> This patch uses the libiberty new-like operators instead of using
> xmalloc/xrealloc.
> It also adds timevars for the main LRA phases, and it fixes a warning
> suggesting a space before a ';' in an only-looping for loop.
>
> Boots
Testing was done before posting the patch. It was successful.
Regards
Ganesh
-Original Message-
From: Uros Bizjak [mailto:ubiz...@gmail.com]
Sent: Thursday, September 27, 2012 5:57 PM
To: Gopalasubramanian, Ganesh
Cc: gcc-patches@gcc.gnu.org
Subject: Re: [PATCH, i386]: Fix PR51109, symbo
On Wed, Oct 03, 2012 at 11:14:37AM +0200, Dodji Seketeli wrote:
> Jakub Jelinek a écrit:
>
> > --- gcc/cp/call.c.jj2012-09-27 12:45:49.0 +0200
> > +++ gcc/cp/call.c 2012-10-01 17:53:17.594609236 +0200
> > @@ -557,7 +557,10 @@ null_ptr_cst_p (tree t)
> > {
> >/* Core
Hello Basile -
> 2012-10-02 Basile Starynkevitch
>
> * gengtype.c (walk_type): Emit mark_hook when inside a
> struct of a union member.
Can you send me off-list the gengtype output before and after the fix?
+ const char *structmarkhook = NULL;
struct_mark_hook or just
Jakub Jelinek a écrit:
> --- gcc/cp/call.c.jj 2012-09-27 12:45:49.0 +0200
> +++ gcc/cp/call.c 2012-10-01 17:53:17.594609236 +0200
> @@ -557,7 +557,10 @@ null_ptr_cst_p (tree t)
> {
>/* Core issue 903 says only literal 0 is a null pointer constant. */
>if (cxx_di
This patch has three components. First we set SUPPRESSED mode as the
default for -gnatg mode (including the run-time). This has no effect
right now, but avoids hidden problems that may appear if in future
we make overflow checking the default behavior.
Second, in some obscure cases, the overflow c
When the type of an object is a CPP untagged type, the object is allocated
in memory through the "new" construct, and the object initialization
requires calling its C++ constructor passing defaults, the Ada compiler
did not generate the call to a C++ constructor which has all parameters with
defaul
When the type of an object is a CPP untagged type and the object initialization
requires calling its default C++ constructor, the Ada compiler did not
generate the call to a C++ constructor which has all parameters with
defaults (and hence it covers the default C++ constructor).
// c_class.h
class
This patch adds a clarifying message info when subtype conformance fails, due
to a missing null exclusion indicatar in a formal that must match a controlling
access formal. This Ada 2005 rule was checked partially in the context of
subprogram renamings but not for 'Access attribute references.
Co
This change fixes an oversight in the Connect_Socket variant that includes
a timeout check. The connection is initiated asynchronously, and a select(2)
call is performed to wait for it to complete, with a timeout. When that call
returns however, the connection is not always succsefully established.
On Tue, Oct 2, 2012 at 3:42 PM, Richard Sandiford
wrote:
>
>> +/* Compress pseudo live ranges by removing program points where
>> + nothing happens. Complexity of many algorithms in LRA is linear
>> + function of program points number. To speed up the code we try to
>> + minimize the numbe
82 matches
Mail list logo
