Ping
> -Original Message-
> From: Joey Ye
> Sent: Thursday, December 20, 2012 17:53
> To: gcc-patches@gcc.gnu.org; Ramana Radhakrishnan
> Cc: Joey Ye
> Subject: [PATCH][ARM][thumb1] Reduce lr save for leaf function with non-
> far jump
>
> Current GCC thumb1 has an annoying problem that a
Need to document the parameter in doc/invoke.texi.
Ok with that change.
David
On Fri, Jan 4, 2013 at 5:28 PM, Rong Xu wrote:
> Hi,
>
> This patch adjusts the single target indirect call promotion heuristics.
> (1) it uses bb counts (bb_all), instead of count "all" which only counts
> the target
On Fri, Jan 4, 2013 at 2:32 PM, Xinliang David Li wrote:
> Looks good -- but better with followup :
> 1) give a warning when the parameter to the option is not allowed;
> 2) add test cases if possible.
Made all the changes. Modified the test case to check if the segment
splitting API is invoked.
Hi,
This patch adjusts the single target indirect call promotion heuristics.
(1) it uses bb counts (bb_all), instead of count "all" which only counts
the targets with the same modules as the caller.
(becaue __gcov_indirect_call_callee is file scope statics).
Using overall bb counts makes more sen
ok.
thanks,
David
On Fri, Jan 4, 2013 at 3:38 PM, Paul Pluzhnikov wrote:
> On Fri, Jan 4, 2013 at 3:27 PM, Xinliang David Li wrote:
>> I saw only one test case fix patch from Paolo in trunk. Does this
>> patch include more local fixes?
>
> There was an earlier patch which touched g++.dg/parse/
On Fri, Jan 4, 2013 at 3:27 PM, Xinliang David Li wrote:
> I saw only one test case fix patch from Paolo in trunk. Does this
> patch include more local fixes?
There was an earlier patch which touched g++.dg/parse/error1{1,2}.C:
r191712 | paolo | 2012-09-25 07:44:52 -0700 (Tue, 25 Sep 2012) | 15
This patch "removes" -fno-whole-file. (Actually, it turns it into "Ignore".)
Reasoning:
* -fwhole-file/-fno-whole-file was added in 4.5 to make the transition
easier; -fwhole-file is the default since 4.6.
* There are many wrong-code issues and probably also some ICEs with
-fno-whole file.
I saw only one test case fix patch from Paolo in trunk. Does this
patch include more local fixes?
David
On Fri, Jan 4, 2013 at 3:03 PM, Paul Pluzhnikov wrote:
> On Fri, Jan 4, 2013 at 9:41 AM, Xinliang David Li wrote:
>> ok.
>
> The patch as sent caused some breakage, I had to adjust test cases
On Fri, Jan 4, 2013 at 9:41 AM, Xinliang David Li wrote:
> ok.
The patch as sent caused some breakage, I had to adjust test cases a bit.
Submitting attached patch.
Thanks,
--
Paul Pluzhnikov
Index: gcc/testsuite/g++.old-deja/g++.other/crash28.C
=
Janne Blomqvist writes:
> diff --git a/libgfortran/io/file_pos.c b/libgfortran/io/file_pos.c
> index c8ecc3a..bf2250a 100644
> --- a/libgfortran/io/file_pos.c
> +++ b/libgfortran/io/file_pos.c
> @@ -140,15 +140,21 @@ unformatted_backspace (st_parameter_filepos *fpp,
> gfc_unit *u)
> }
>
On Fri, Jan 4, 2013 at 2:22 PM, Leif Ekblad wrote:
> Change log:
> * config/gthr.m4: Added rdos thread header.
> * gcc/config/i386/i386.c: Provided a way to define a default setting for
> medium memory model and PIC. Provided a way to define a default value for
> large-data-threshold.
> * gcc/conf
Looks good -- but better with followup :
1) give a warning when the parameter to the option is not allowed;
2) add test cases if possible.
David
On Fri, Jan 4, 2013 at 2:19 PM, Sriraman Tallam wrote:
> Attached new patch.
>
> Thanks,
> -Sri.
>
> On Fri, Jan 4, 2013 at 9:12 AM, Rong Xu wrote:
>
OK. I know I can do that, but I would need to do it for every syscall since
every syscall can potentially clobber rbx.
Also, it is very strange that it is only one of the inlines the compiler
complains about. I have another inline (which uses rbx as input), but that
doesn't generate any error:
On Fri, Jan 4, 2013 at 1:34 AM, Richard Earnshaw wrote:
> On 03/01/13 22:39, Andrew Pinski wrote:
>>
>> Hi,
>>For aarch64, we don't CSE some cmp away. This patch fixes the case
>> where we are CSE across some basic-blocks like:
>> int f(int a, int b)
>> {
>>if(a> return 1;
>>if(a
Change log:
* config/gthr.m4: Added rdos thread header.
* gcc/config/i386/i386.c: Provided a way to define a default setting for
medium memory model and PIC. Provided a way to define a default value for
large-data-threshold.
* gcc/config/i386/i386.h: Defined default value for medium memory model
Attached new patch.
Thanks,
-Sri.
On Fri, Jan 4, 2013 at 9:12 AM, Rong Xu wrote:
> The code looks fine to me. Please consider David's comments about the
> option name.
>
> -Rong
>
> On Thu, Jan 3, 2013 at 9:14 PM, Xinliang David Li wrote:
>> Is it better to change the option to something like:
Hi,
currently byte swapped unformatted IO can be quite slow compared to
the same code with no byte swapping. There are two major reasons for
this:
1) The byte swapping code path resorts to transferring data element by
element, leading to a lot of overhead in the IO library.
2) The function used
On Fri, Jan 04, 2013 at 10:39:05PM +0100, Leif Ekblad wrote:
> I just tried the patch, but it seems to produce some problems for
> me. The other patch which used a 64-bit specific register (r15)
> instead of rbx was easier to adapt to. The problem for me is that
> syscalls might clobber higher-half
I just tried the patch, but it seems to produce some problems for me. The
other patch which used a 64-bit specific register (r15) instead of rbx was
easier to adapt to. The problem for me is that syscalls might clobber
higher-half of all 32-bit registers, and that I cannot use the stack to
pres
Dear Tobias,
s/Argument is neither a pointer/allocatble/Argument is neither a
pointer/allocatable/
Why have you all the asserts for se.pre and se.post? Did you have
trouble with this or is there some hidden nasty that I am unaware of?
Should you not either return rapidly for an intrinsic type or
Dear Tobias,
This one is also OK for trunk. It does cross my mind, however, that
we should offer STRIDE as a gcc extension, in anticipation of F201x.
Cheers
Paul
On 31 December 2012 13:17, Tobias Burnus wrote:
> The attached patch adds a new - internal only - intrinsic, STRIDE, which
> return
Dear Tobias,
This one is 'obvious' - OK for trunk.
Thanks
Paul
On 31 December 2012 13:16, Tobias Burnus wrote:
> This simple patch fixes a wrong indent and adds a _final component to the
> virtual tables generated for intrinsic types.
>
> This patch not only prepares the trunk for finalization
Dear Tobias,
Committed as revision 194916.
I am leaving the PR open to deal with 4.7 just as soon as I have
downloaded it!
Thanks again for the careful review.
Cheers
Paul
On 2 January 2013 23:48, Tobias Burnus wrote:
> Dear Paul,
>
> First, the new patch is fine from my side. (Although,
I filed http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55743 (my first
upstream gcc bug so be gentle :-)
Details are there but the short version is that the limits.h inclusion
in libgcc2.c is now a relic because the constants that it brings
in are no longer used (since
http://repo.or.cz/w/official-gc
PR 54800 points out a minor optimization uncovered by cppcheck. This
optimization is not important, but we might as well fix it in case
cppcheck comes up with something useful.
Tested by Iain Sandoe (thanks!). Committed to mainline.
Ian
2013-01-04 Ian Lance Taylor
PR other/54800
> Hmm, how would anyone else get at the "padding" info dealing with the
> parallel? This looks broken if we need access to the type :/
Yes, you need to access the type for return values, but that's not really new.
> Eric, any comments?
Richard's patch seems fine to me, modulo the name of the ro
Hello!
As suggested by rth at [1].
2013-01-04 Uros Bizjak
* config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
unconditionally.
(ix86_expand_move): Ditto.
(ix86_zero_extend_to_Pmode): Ditto.
(ix86_expand_call): Ditto.
(ix86_expan
I've merged trunk revision 194911 to the gccgo branch.
Ian
Dear Tobias,
Thanks for the review. The patched trunk is just now bootstrapping
and regtesting prior to commitment.
I have added a check for co-indexed selectors in resolve.c and have
added tests in select_type_31.f03 for co-indexing and vector indices.
I am now turning my attention to FINAL :-
ok.
thanks,
David
On Fri, Jan 4, 2013 at 9:32 AM, Paul Pluzhnikov wrote:
> Back-port revision 194909 to google/gcc-4_7 branch:
> http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=194909
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54526
>
> Google ref: b/7427993
>
> Index: gcc/testsuite/g++.old
Back-port revision 194909 to google/gcc-4_7 branch:
http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=194909
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54526
Google ref: b/7427993
Index: gcc/testsuite/g++.old-deja/g++.other/crash28.C
===
The code looks fine to me. Please consider David's comments about the
option name.
-Rong
On Thu, Jan 3, 2013 at 9:14 PM, Xinliang David Li wrote:
> Is it better to change the option to something like:
>
> split_segment|nosplit-segment
> or split_segment=yes|no
>
>
> David
>
> On Thu, Jan 3, 2013
On Fri, Jan 04, 2013 at 08:44:13AM -0800, Andrew Pinski wrote:
> On Fri, Jan 4, 2013 at 4:54 AM, Jakub Jelinek wrote:
> > I've run a script to notice gcc maintained files with FSF copyright that
> > have been modified in 2011 and/or 2012 (according to svn log, ignoring
> > r168438 and r184997 comm
When an anonymous class gets a linkage name from a typedef, we
recalculate its visibility; anonymous types have internal linkage, but
types with linkage names have external linkage and therefore can have
visibility. But we were forgetting to also recalculate the visibility
of any nested types,
On Fri, Jan 4, 2013 at 4:54 AM, Jakub Jelinek wrote:
> Hi!
>
> I've run a script to notice gcc maintained files with FSF copyright that
> have been modified in 2011 and/or 2012 (according to svn log, ignoring
> r168438 and r184997 commits), but didn't have years 2011 and/or 2012
> included in Copy
OK.
Jason
On Fri, Jan 4, 2013 at 2:06 AM, Andreas Schwab wrote:
> "H.J. Lu" writes:
>
>> On Thu, Jan 3, 2013 at 10:09 AM, Andreas Schwab
>> wrote:
>>> "H.J. Lu" writes:
>>>
diff --git a/libjava/Makefile.am b/libjava/Makefile.am
index c6c84e4..dd08a4f 100644
--- a/libjava/Makefile.am
The following reverts a change I did during the various PRE cleanups.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2013-01-04 Richard Biener
PR tree-optimization/55862
* tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
translating
On Fri, 4 Jan 2013, Jakub Jelinek wrote:
> Hi!
>
> As discussed in the PR, as all INTEGRAL_TYPE_P types with the same
> sign/precision are usually considered compatible
> (useless_type_conversion_p), during gimplify_one_sizepos the gimplifier
> can change e.g. an integer expression into e.g. VAR_
Assembler could be fixed easily (I'll contribute the patch to MacPorts
cctools port) and the library is loading and working now.
So I suggest disabling the use of 64 bit instructions and registers
when targetting PowerPC Darwin 32 bit ABI can be considered working
and a patch to gcc shouldn't
Hi!
As discussed in the PR, as all INTEGRAL_TYPE_P types with the same
sign/precision are usually considered compatible
(useless_type_conversion_p), during gimplify_one_sizepos the gimplifier
can change e.g. an integer expression into e.g. VAR_DECL with ENUMERAL_TYPE,
and when the gimplifier later
Hi,
the patch below fixes PR 55579 by not disqualifying a candidate
variable when it does not have any real replacements but has
ones for debug statements only.
analyze_access_subtree has to return the same value regardless of
MAY_HAVE_DEBUG_STMTS or we get debug comparison errors. I also have a
Hi!
I've run a script to notice gcc maintained files with FSF copyright that
have been modified in 2011 and/or 2012 (according to svn log, ignoring
r168438 and r184997 commits), but didn't have years 2011 and/or 2012
included in Copyright lines. I've kept the preexisting style, so
where year rang
Fortran 2008 allows:
integer :: pointer => init_data
and
type t
integer :: pointer => init_data
end type t
The current check in gfc_check_assign_symbol was only called for former
and for constructors, but not for the type definition. Additionally,
BT_CLASS wasn't handled. I also imp
Hello,
Le 04/01/2013 00:23, Tobias Burnus a écrit :
NULL with MOLD should be rejected as (default) initialization
expression. From F2008:
R506 null-init is function-reference
C512 (R506) The function-reference shall be a reference to the intrinsic
function NULL with no arguments.
"null-init" o
On Fri, Jan 04, 2013 at 12:05:49PM +0100, Marek Polacek wrote:
> On Fri, Jan 04, 2013 at 11:42:24AM +0100, Jakub Jelinek wrote:
> > I wonder what is the point of using %qs with a fixed string, and ' should be
> > probably replaced by %< and %>, so perhaps
> > "argument to %<-O%> should
On Fri, Jan 04, 2013 at 11:42:24AM +0100, Jakub Jelinek wrote:
> I wonder what is the point of using %qs with a fixed string, and ' should be
> probably replaced by %< and %>, so perhaps
> "argument to %<-O%> should be a non-negative integer, "
> "%, % or %");
So, like
On Thu, Jan 3, 2013 at 7:42 PM, Richard Sandiford
wrote:
> Richard Biener writes:
>> On Wed, Jan 2, 2013 at 5:36 PM, Richard Sandiford
>> wrote:
>>> Richard Biener writes:
On Sun, Dec 23, 2012 at 10:43 AM, Richard Sandiford
wrote:
> Some of the maths builtins can expand to a call
On Thu, 3 Jan 2013, Martin Jambor wrote:
> Hi,
>
> the patch below fixes PR 55755 which was in the compiler for years.
> The problem is that a replacement of a bit-field can have a larger
> TYPE_SIZE than the type of the field and so creating a V_C_E from it
> to the field type may result in inva
On Fri, Jan 04, 2013 at 02:59:46AM +0100, Marek Polacek wrote:
> This patch clarifies the error message when using e.g. -Obar option,
> except non-negative numbers we accept some other levels too.
> Bootstrapped on x86_64-linux. Ok for trunk?
>
> 2013-01-04 Marek Polacek
>
> PR middle-e
This fixes PR55863, a folding missed optimization caused by
folding -X - 1 to ~X. This makes data dependence analysis fail
for offsetted-by-1 accesses (in some cases).
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Richard.
2013-01-04 Richard Biener
PR middl
Hi Guys,
I am checking in the patch below to fix a small problem with libgcc
for the V850. The CALLT support functions were being compiled even if
the particular multilib concerned did not support the CALLT insn.
Normally this would not matter, but the linker will now report an
unknown
This is an internal cleanup, no functional effect, no test
Tested on x86_64-pc-linux-gnu, committed on trunk
2013-01-04 Robert Dewar
* sinfo.ads: Clean up order of N_xxx subtypes
Index: sinfo.ads
===
--- sinfo.ads (rev
On 29/11/12 14:42, Matthew Gretton-Dann wrote:
On 24 November 2012 00:27, Ramana Radhakrishnan
wrote:
On Wed, Nov 21, 2012 at 7:59 PM, Matthew Gretton-Dann
wrote:
[snip]
The fix is to decrease the pool_range of all insns by 2 when generating
Thumb code. There is no need to change neg_pool_r
"H.J. Lu" writes:
> On Thu, Jan 3, 2013 at 10:09 AM, Andreas Schwab wrote:
>> "H.J. Lu" writes:
>>
>>> diff --git a/libjava/Makefile.am b/libjava/Makefile.am
>>> index c6c84e4..dd08a4f 100644
>>> --- a/libjava/Makefile.am
>>> +++ b/libjava/Makefile.am
>>> @@ -594,7 +594,7 @@ lib_gnu_awt_xlib_la
Hi,
On 01/03/2013 10:56 PM, Jason Merrill wrote:
On 01/03/2013 05:44 AM, Paolo Carlini wrote:
+ /* C++11 - 2.5 p3, bullet 2. */
Please flesh out this comment some more.
Ok, I extended it like this.
Thanks,
Paolo.
/
Index: gcc/cp/parser.c
===
On 03/01/13 22:39, Andrew Pinski wrote:
Hi,
For aarch64, we don't CSE some cmp away. This patch fixes the case
where we are CSE across some basic-blocks like:
int f(int a, int b)
{
if(ab)
return -1;
return 0;
}
--- CUT ---
To fix this, I implemented the target hook
TARGET_FIXED_CON
This patch corrects an obvious latent bug in Insert_Actions, namely
that in the case of an expression with actions node, the actions
were inserted in reverse order from the calls. As far as we know,
this bug is only latent (found from code reading), so no test is
required for this fix.
Tested on x
Correct some missing cases of setting this switch, document it in
usage. Primarily documentation, so no new test needed. Also makes
clear that this switch is not available in VMS. No test needed.
Tested on x86_64-pc-linux-gnu, committed on trunk
2013-01-04 Robert Dewar
* gnat_ugn.texi
This patch improves the error message when compiling a unit with an aspect
specification in an older Ada mode, and the first aspect specification is for
a boolean aspect whose value is defaulted.
Compiling the following:
gcc -c r1.ads
must yield:
r1.ads:2:16: aspect specification is an Ad
This patch provides the initial implementation of aspect Global. This construct
is intended for formal verification proofs.
The syntax of the aspect is as follows:
global_specification ::= (moded_global_list {, moded_global_list})
| global_list
This patch delays the generation of index checks for the following cases of
Loop_Entry attribute references:
Prefix'Loop_Entry (Expr)
Prefix'Loop_Entry (Expr1, Expr2, ... ExprN)
Even though these constructs appear initially as attribute references, analysis
converts them into indexed compon
62 matches
Mail list logo