Hi Richard.
Your last patch for this PR suggested we verify our assumptions wrt a
least common C++ compiler to build trunk with. Since I already had the
offended system at hand (Red Hat Linux 8.0), I decided to investigate a
bit further.
g++ 3.4 builds trunk just fine, but anything prior to
Hi,
The definition of constraint 'X' allows all operands.
`X' - Any operand whatsoever is allowed.
However, invalid memory operands should not be valid input for 'X'.
Please find attached the patch "X_constraint.patch" which ignores
invalid memory operands in constraint 'X'.
Fixes the ICE gcc.d
On Tue, Dec 11, 2012 at 1:37 PM, Jakub Jelinek wrote:
> So, what about this version instead?
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
>
> 2012-12-11 Jakub Jelinek
> Janus Weil
>
> PR fortran/55636
> * gfortran.h (GFC_PREFIX): Define
On Sun, Dec 9, 2012 at 10:49 AM, John David Anglin
wrote:
>
> 2012-12-09 John David Anglin
>
> * mmapio.c: Define MAP_FAILED if not defined.
This is OK.
Thanks.
Ian
On Sun, Dec 9, 2012 at 11:57 AM, John David Anglin
wrote:
>
> #ifndef HAVE_GETEXECNAME
> +#if defined(__hpux) && (defined(__ia64) || defined(_LP64))
> +#include
> +#define getexecname getexecname_hpux
> +
> +static char *
> +getexecname_hpux (void)
> +{
> + struct load_module_desc desc;
> +
> +
Hi Richard,
Thanks for your reply.
The sysroot of /dev/null is basically to force the user or build system to pass a valid --sysroot argument. This helps us to ensure that we only
link against known libraries (that are in a specific location) and there isn't any leakage from the host system.
On Tue, 2012-12-11 at 21:52 +, Richard Sandiford wrote:
> > + mips_dbx_regno[CPRESTORE_SLOT_REGNUM] = IGNORED_DWARF_REGNUM;
> If even fake registers like these are going to be used, then I think
> we should initialise to IGNORED_DWARF_REGNUM rather than INVALID_REGNUM in:
>
> for (i = 0;
Some SPEC2k performance number (with 3 runs on core2):
Push wins over move on 3 benchmarks. Others are noises.
perlbmk : ~+1.9%
gap: ~+1.4%
vortex:~ +0.7%
David
On Tue, Dec 11, 2012 at 2:53 PM, Xinliang David Li wrote:
> The following the O2 size data from SPEC2k. Note that with pus
The following the O2 size data from SPEC2k. Note that with push/pop,
it is a always a net win (negative delta) in terms of total binary or
total loadable section size.
thanks,
David
.text.eh_frame Total_binary
vortex-move 440252 40796 584066
vortex-push 415436 57452 5759
Convert tree-vectorizer.h'_loop_vec_info::peeling_htab from htab_t
to hash_table.
* tree-vectorizer.h
New struct peel_info_hasher.
* tree-vect-loop.c
Update dependent calls and types to match.
* tree-vect-data-refs.c
Fold vect_peeling_hash and vect_peeling_hash_eq into struct peel_info_hasher
Update various htab_t tables to hash_table. Each file is independent.
Update dependent calls and types.
* tree-ssa-strlen.c'decl_to_stridxlist_htab
Fold decl_to_stridxlist_hash into new struct stridxlist_hasher.
* tree-ssa-loop-ivopts.c'ivopts_data::inv_expr_tab
Fold htab_inv_expr_hash and hta
On Wed, Nov 28, 2012 at 5:53 AM, Terry Guo wrote:
> Hello,
>
> Attached patch intends to fix a bug on how to check argument register number
> which should consider the PCS. A test case is also included. Without this
> fix, one of the function argument will be overridden in the case. Tested on
> QE
Convert tree-sra.c'candidates from htab_t to hash_table.
Fold uid_decl_map_hash and uid_decl_map_eq into new struct
uid_decl_hasher. This change moves the definitions from tree-ssa.c
into tree-sra.c and removes the declarations from tree-flow.h
Update dependent calls and types to hash_table.
Te
Enums in anonymous namespace should get anon visibility just like classes.
Tested x86_64-pc-linux-gnu, applying to 4.6, 4.7 and trunk.
commit ebb0d2ecbf7717afcb45f52ab7d7a0d68ae2157b
Author: Jason Merrill
Date: Tue Dec 11 16:34:24 2012 -0500
PR c++/54883
* decl2.c (min_vis_r): Handl
PR libstdc++/55631
* include/ext/alloc_traits.h: Include missing header.
* include/ext/pointer.h: Likewise.
* include/ext/string_conversions.h: Require C++11.
* libsupc++/initializer_list: Reindent.
Tested x86_64-linux, committed to trunk.
commit 4e4d3c624f0
Ping ? Paul, seen that you've contributed fp16.c together with Sandra
Loosemore, perhaps you can review this patch please ?
John.
In message
John Tytgat wrote:
> __gnu_f2h_internal in libgcc converts single-precision floating-point value
> to half-precision value for ARM. I noticed
Hello,
This fixes an old regression about ms-structure-layout in combination
with packed-attribute.
ChangeLog
2012-12-11 Kai Tietz
PR c/52991
* stor-layout.c (start_record_layout): Handle
packed-attribute for ms-structure-layout.
(update_alignment_for_field): Li
On Tue, Dec 11, 2012 at 6:36 AM, Paolo Bonzini wrote:
AM_CXXFLAGS = -Wall -W -Wno-unused-parameter -Wwrite-strings -pedantic
-Wno-long-long -fPIC -fno-builtin -fno-exceptions -fomit-frame-pointer
-funwind-tables -fvisibility=hidden -Wno-variadic-macros
-Wno-c99-extensions
"Steve Ellcey " writes:
> While building libgcc in mips16 mode I found another instance of
> dbx_reg_number aborting that the patch to pr54061 did not fix.
> This code:
>
> extern void __chk_fail (void) __attribute__ ((__noreturn__));
> __strncpy_chk (s1, s2, n, s1len)
> {
> char c;
> char *s
On Thu, Dec 06, 2012 at 01:48:31PM +0100, Steven Bosscher wrote:
> Hello,
>
> This patch has 3 parts:
>
> 1. Benign comment fixes.
>
> 2. Using DF_REF_REG_MEM_P idiom. Also benign.
>
> 3. Real bug fixes for cfglayout mode.
>
> For (3) the fixes are:
> - Pointers to the unlinked parts of the in
Hello,
this patch should help if we ever want to use sunCC to initiate a
bootstrap, though I didn't test with sunCC. Note that using gmp_fprintf
means we have to include stdio.h before gmp.h. I didn't investigate how,
but this seems to already be the case :-) The reallocator cast is just a
ha
On Tue, Dec 11, 2012 at 6:36 AM, Paolo Bonzini wrote:
> Il 11/12/2012 14:47, H.J. Lu ha scritto:
>> On Thu, Dec 6, 2012 at 7:07 AM, H.J. Lu wrote:
>>> On Thu, Nov 29, 2012 at 10:30 AM, H.J. Lu wrote:
Hi,
Since libsanitizer is used for bootstrap and compiled with raw_cxx,
we n
Hi!
As discussed in the PR, this patch forward ports Steven's patch from 4.7
branch to trunk. No need to include pointer-set.h (already included),
I've moved the var definition into the #ifdef so that it isn't an unused
static var and tweaked the comment a little bit.
Bootstrapped/regtested on x
On 12/11/2012 04:09 PM, Dodji Seketeli wrote:
OK. The below should hopefully approach what you have in mind.
Thanks, that's closer to what I was thinking. I'd also like to move the
scan and PACK_EXPANSION_EXTRA_ARGS code back out of the loop.
Jason
Il 11/12/2012 20:27, H.J. Lu ha scritto:
> On Tue, Dec 11, 2012 at 6:36 AM, Paolo Bonzini wrote:
>
>>
>> As a followup please check if AM_MAKEFLAGS is needed at all.
>>
> diff --git a/libsanitizer/Makefile.in b/libsanitizer/Makefile.in
> index 21c2711..53e0be9 100644
> --- a/libsaniti
Ping?
On Thu, Dec 6, 2012 at 1:48 PM, Steven Bosscher wrote:
> Hello,
>
> This patch has 3 parts:
>
> 1. Benign comment fixes.
>
> 2. Using DF_REF_REG_MEM_P idiom. Also benign.
>
> 3. Real bug fixes for cfglayout mode.
>
> For (3) the fixes are:
> - Pointers to the unlinked parts of the insns chai
Jason Merrill writes:
> On 12/11/2012 10:55 AM, Dodji Seketeli wrote:
>> Oops, it seems I was too hasty in trying to do away with the
>> instantiation_yields_no_list_p parameter to
>> gen_elem_of_pack_expansion_instantiation.
>
> I still think that the elem function should just deal with the sing
While building libgcc in mips16 mode I found another instance of
dbx_reg_number aborting that the patch to pr54061 did not fix.
This code:
extern void __chk_fail (void) __attribute__ ((__noreturn__));
__strncpy_chk (s1, s2, n, s1len)
{
char c;
char *s = s1;
if (__builtin_expect (s1len < n, 0
On Tue, Dec 11, 2012 at 3:33 PM, wrote:
> As suggested by Doug, this patch is aimed to google/integration,
> google/main, and google/gcc-4_7.
OK for google branches. No need to add ChangeLog entries, however.
Diego.
As suggested by Doug, this patch is aimed to google/integration,
google/main, and google/gcc-4_7.
https://codereview.appspot.com/6932046/
Vladimir Makarov writes:
> On 12/08/2012 05:30 AM, Richard Sandiford wrote:
>> Vladimir Makarov writes:
>>> Index: lra-constraints.c
>>> ===
>>> --- lra-constraints.c (revision 194307)
>>> +++ lra-constraints.c (working c
Hello,
This patch clusters loop bodies in the CFG graph dumper. Makes for
easier-to-interpret plots and helps debugging.
Bootstrapped&tested on powerpc64-unknown-linux-gnu. OK for trunk?
Ciao!
Steven
graph_dump_loops.diff
Description: Binary data
On Tue, Dec 11, 2012 at 10:12 AM, Jan Hubicka wrote:
>> On 2012.12.11 at 09:13 -0800, Teresa Johnson wrote:
>> > Ping.
>> > Teresa
>> >
>> > On Mon, Nov 19, 2012 at 4:58 PM, Teresa Johnson
>> > wrote:
>> > > This patch was proposed awhile back, and the new working set program
>> > > summary was
On Tue, Dec 11, 2012 at 6:36 AM, Paolo Bonzini wrote:
>
> As a followup please check if AM_MAKEFLAGS is needed at all.
>
diff --git a/libsanitizer/Makefile.in b/libsanitizer/Makefile.in
index 21c2711..53e0be9 100644
--- a/libsanitizer/Makefile.in
+++ b/libsanitizer/Makefile.in
OK.
Jason
OK.
Jason
From: Eric Botcazou
Date: Tue, 11 Dec 2012 19:37:36 +0100
> This is a regression present on mainline and 4.7 branch for the SPARC.
> Reload
> is trying to change the value of a symbol(!) because it is mightily confused
> by an output constraint on a source operand (old pasto in some TLS patte
This is a regression present on mainline and 4.7 branch for the SPARC. Reload
is trying to change the value of a symbol(!) because it is mightily confused
by an output constraint on a source operand (old pasto in some TLS patterns).
Fixed thusly, tested on SPARC64/Linux, applied on all branches
On Tue, Dec 11, 2012 at 10:36 AM, Jakub Jelinek wrote:
>
> 2012-12-11 Jakub Jelinek
>
> PR bootstrap/54926
> * Makefile.am (AM_CFLAGS): Remove -frandom-seed=$@.
> * configure.ac: If --with-target-subdir, add -frandom-seed=$@
> to EXTRA_FLAGS unconditionally, othe
Hi!
Lately this testcase often timesout for me on a busy box,
I don't see a point iterating 2000^3 times, with 400^3 it is much faster
and I still could reproduce the problem before the corresponding fix
and the vectorizer r145171 fix fixed it.
Bootstrapped/regtested on x86_64-linux and i686-linu
On Tue, Dec 11, 2012 at 02:29:18PM +0100, Janus Weil wrote:
> 2012/12/11 Jakub Jelinek :
> > On Tue, Dec 11, 2012 at 12:16:33PM +0100, Janus Weil wrote:
> >> Ok, so here is a new patch, updated according to the suggestions of
> >> David and Jakub. This now only touches the dotted variables, which a
Hi!
Richard noticed that bootstrap with very old GCC as system compiler failed
(as one of the reasons) because -frandom-seed=dwarf.lo etc. was used to
compile the host libbacktrace in stage1. Non-GCC compilers presumably don't
support that option either.
The following patch fixes it by adding it
Hi!
The following testcase on x86_64 (or the f90-intrinsic-bitsize.f on 32-bit
HWI) with -Os shows a bug in discover_iteration_bound_by_body_walk. If some
bound is a -1, -1 HWI, then adding double_int_one to it overflows into 0, 0
HWI and we can return that as maximum number of iterations. We ne
Hi!
When casting ints to other ints, CASE_CONVERT in mark_exp_read handles that
case, but for FLOAT_EXPR it doesn't. I've tried other static casts, but
haven't found anything else that would fail similarly.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk/4.7?
2012-12-11 Jak
In this testcase we get confused by a specialization nested inside a
class template; the members of the specialization end up having
dependent template arguments, and then we get confused when we try to
define one of them in a context that has no template headers.
Fixed by rejecting the defini
> On 2012.12.11 at 09:13 -0800, Teresa Johnson wrote:
> > Ping.
> > Teresa
> >
> > On Mon, Nov 19, 2012 at 4:58 PM, Teresa Johnson
> > wrote:
> > > This patch was proposed awhile back, and the new working set program
> > > summary was split out and committed separately. I would like to
> > > see
On 2012.12.11 at 09:13 -0800, Teresa Johnson wrote:
> Ping.
> Teresa
>
> On Mon, Nov 19, 2012 at 4:58 PM, Teresa Johnson wrote:
> > This patch was proposed awhile back, and the new working set program
> > summary was split out and committed separately. I would like to
> > see if the loop unroller
Hello,
Did you get a chance to look at this patch? I submitted this ~1 month
ago, so thought I would inquire its status.
Thanks,
Balaji V. Iyer.
> -Original Message-
> From: Iyer, Balaji V
> Sent: Monday, November 05, 2012 4:43 PM
> To: Joseph Myers
> Cc: gcc-patches@gcc.gnu.org
On 12/08/2012 05:30 AM, Richard Sandiford wrote:
Vladimir Makarov writes:
Index: lra-constraints.c
===
--- lra-constraints.c (revision 194307)
+++ lra-constraints.c (working copy)
@@ -3329,8 +3329,9 @@ lra_constraints (bool fir
On 12/11/2012 03:21 AM, Jakub Jelinek wrote:
Hi!
As the testcase in the PR (unfortunately only reproduceable after reverting
some inlining patch, so not including that testcase in the patch) shows,
loc_equivalence_change_p isn't good enough for debug_insns, when a REG is
substituted for VOIDmode
On Tue, Dec 11, 2012 at 7:07 AM, Jakub Jelinek wrote:
>
> 2012-12-11 Jakub Jelinek
>
> PR libgcc/55451
> * fixed-bit.c (FIXED_SSADD, FIXED_USADD, FIXED_SSSUB, FIXED_USSUB,
> FIXED_SSNEG): Avoid undefined signed overflows.
This is OK.
Thanks.
Ian
On Tue, Dec 11, 2012 at 1:49 AM, Richard Biener
wrote:
> On Mon, Dec 10, 2012 at 10:07 PM, Mike Stump wrote:
>> On Dec 10, 2012, at 12:42 PM, Xinliang David Li wrote:
>>> I have not measured the CFI size impact -- but conceivably it should
>>> be larger -- which is unfortunate.
>>
>> Code speed
Il 11/12/2012 17:55, H.J. Lu ha scritto:
>> > Does it use any GCC toplevel files except when regenerating the
>> > configury? But I think it's a useful refactoring anyway.
> libsanitizer started as standalone and it didn't work with GCC build.
> We removed all those extra files to get it compile a
On Tue, Dec 11, 2012 at 8:19 AM, Paolo Bonzini wrote:
> Il 11/12/2012 16:03, H.J. Lu ha scritto:
>> diff --git a/libsanitizer/asan/Makefile.am
>> b/libsanitizer/asan/Makefile.am
>> index 3da1db3..45fb3b3 100644
>> --- a/libsanitizer/asan/Makefile.am
>> +++ b/libsanitizer/asan/
On 12/11/2012 11:33 AM, Jakub Jelinek wrote:
I guess it depends on how would make_lvalue_use differ from make_rvalue_use
in the future. To some extent it is lvalue-ish use, but not completely,
while it is taking its address, it is fine if the operand is e.g. const,
which is wrong for lvalues.
On 12/11/2012 10:55 AM, Dodji Seketeli wrote:
Oops, it seems I was too hasty in trying to do away with the
instantiation_yields_no_list_p parameter to
gen_elem_of_pack_expansion_instantiation.
I still think that the elem function should just deal with the single
element case; the cases where w
On Tue, Dec 11, 2012 at 11:19:48AM -0500, Jason Merrill wrote:
> On 12/11/2012 03:16 AM, Jakub Jelinek wrote:
> >+ if (constraint_parsed && !allows_reg && allows_mem)
> >+operand = mark_rvalue_use (TREE_VALUE (t));
>
> This should be mark_lvalue_use, since we're using its address rath
Jakub Jelinek writes:
> Actually, the problem was that libasan was linked in after libpthread.
> Perhaps we need some driver hacks to make sure -lasan comes before -lpthread
> when added automatically for -fsanitize=address (and similarly for -ltsan).
> For now just tweaking dg-options.
> Ok f
On 12/11/2012 03:16 AM, Jakub Jelinek wrote:
+ if (constraint_parsed && !allows_reg && allows_mem)
+ operand = mark_rvalue_use (TREE_VALUE (t));
This should be mark_lvalue_use, since we're using its address rather
than its value. OK with that change.
Jason
Il 11/12/2012 16:03, H.J. Lu ha scritto:
> diff --git a/libsanitizer/asan/Makefile.am b/libsanitizer/asan/Makefile.am
> index 3da1db3..45fb3b3 100644
> --- a/libsanitizer/asan/Makefile.am
> +++ b/libsanitizer/asan/Makefile.am
> @@ -5,6 +5,10 @@ gcc_version := $(shell cat $(top_s
Jason Merrill writes:
> On 12/08/2012 05:12 PM, Dodji Seketeli wrote:
> >>> + else if (arg_from_pack_level_to_prune || has_empty_arg)
> >>> +{
> >>> + /* ... we just return a pack expansion which pattern is PATTERN
> >>> + into which ARGS has been substituted. */
> >>> + *instant
On Tue, Dec 11, 2012 at 06:52:38AM -0800, Ian Lance Taylor wrote:
> I presume that the intent of the obfuscation is to make SSNEG look
> exactly like SSSUB. And SSSUB appears to have the exact problems that
> you are fixing here. So I think you ought to fix both the same way.
Ugh, I was hoping f
On Tue, Dec 11, 2012 at 6:36 AM, Paolo Bonzini wrote:
> As a followup please check if AM_MAKEFLAGS is needed at all.
I think it is just copied from elsewhere. I will take a look after this.
diff --git a/libsanitizer/Makefile.in b/libsanitizer/Makefile.in
index 21c2711..53e0be9 100644
On Tue, Dec 11, 2012 at 4:55 AM, Jakub Jelinek wrote:
>
> This routine, besides aspiring to win obfuscated C contest (not trying to
> address that) contains two undefined signed overflows, which presumably
> show up in arm testing. One overflow is on z = x - y; line,
> where the x looks just like
Il 11/12/2012 14:47, H.J. Lu ha scritto:
> On Thu, Dec 6, 2012 at 7:07 AM, H.J. Lu wrote:
>> On Thu, Nov 29, 2012 at 10:30 AM, H.J. Lu wrote:
>>> Hi,
>>>
>>> Since libsanitizer is used for bootstrap and compiled with raw_cxx,
>>> we need to use explicit -I for libstdc++-v3 header files in
>>> lib
On Tue, 11 Dec 2012, Jakub Jelinek wrote:
> Hi!
>
> As discussed on IRC, fixing this issue is probably 4.9 material, we have
> some plans, but not for 4.8. What GCC generates is actually not incorrect,
> given __builtin_object_size (x, 1) being a high bound, it is just larger than
> strictly req
Hi!
As discussed on IRC, fixing this issue is probably 4.9 material, we have
some plans, but not for 4.8. What GCC generates is actually not incorrect,
given __builtin_object_size (x, 1) being a high bound, it is just larger than
strictly required, but I hope it won't trigger very much on real-wo
gcc/testsuite/ChangeLog
2012-12-11 Kyrylo Tkachov
Missing reference to PR in the testsuite changelog entry.
* gcc.target/arm/pr55642.c: New testcase.
Ok with that change.
Ramana
Hi all,
This patch fixes http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55642
It sets the ce_count attribute for the abssi2 patterns in thumb2.md. If we
don't set it, gcc assumes that the pattern generates only one conditional
instruction and produces
the wrong enclosing IT instruction. The two affec
On Tue, Dec 11, 2012 at 2:58 PM, Steven Bosscher wrote:
> Hello,
>
> With the new GraphViz CFG dumper, the VCG dumper is now redundant. It
> could only dump the CFG just after constructing it, and it doesn't
> work as well as the DOT dumper (especially for multiple functions).
>
> Bootstrapped&tes
Hello,
With the new GraphViz CFG dumper, the VCG dumper is now redundant. It
could only dump the CFG just after constructing it, and it doesn't
work as well as the DOT dumper (especially for multiple functions).
Bootstrapped&tested on powerpc64-unknown-linux-gnu. OK for trunk?
Ciao!
Steven
rem
On Tue, Dec 11, 2012 at 4:14 AM, Richard Biener wrote:
>
> This brings the build-requirements up-to-date with us now requiring
> a C++ host compiler. I optimistically increased the minimum required
> GCC version listed from 2.95 to 3.4 as that is the earliest version
> that could reasonably be ca
On Thu, Dec 6, 2012 at 7:07 AM, H.J. Lu wrote:
> On Thu, Nov 29, 2012 at 10:30 AM, H.J. Lu wrote:
>> Hi,
>>
>> Since libsanitizer is used for bootstrap and compiled with raw_cxx,
>> we need to use explicit -I for libstdc++-v3 header files in
>> libsanitizer. Otherwise, we will get
>>
>> libtool:
2012/12/11 Jakub Jelinek :
> On Tue, Dec 11, 2012 at 12:16:33PM +0100, Janus Weil wrote:
>> Ok, so here is a new patch, updated according to the suggestions of
>> David and Jakub. This now only touches the dotted variables, which are
>> responsible for the AIX trouble. Whether the same prefixing sh
Hi!
This routine, besides aspiring to win obfuscated C contest (not trying to
address that) contains two undefined signed overflows, which presumably
show up in arm testing. One overflow is on z = x - y; line,
where the x looks just like obfuscation (it is always 0), for input equal to
INT_MIN bi
Hi all,
Since the new AArch32 NEON instructions in arm_neon.h are predicated on
__ARM_ARCH 8 the testsuite add_options procedure should also include
-march=armv8-a to make these instructions available.
This makes the new vrnd* tests in gcc.target/arm/neon PASS in situations
where armv8-a is not t
On Tue, Dec 11, 2012 at 12:16:33PM +0100, Janus Weil wrote:
> Ok, so here is a new patch, updated according to the suggestions of
> David and Jakub. This now only touches the dotted variables, which are
> responsible for the AIX trouble. Whether the same prefixing should
> also be applied in other
Jakub Jelinek writes:
> * varasm.c (get_variable_section): Don't return lcomm_section
> for asan_protect_global decls.
> * asan.c (asan_protect_global): Only avoid public common variables.
> Don't call get_variable_section here.
This is OK, thanks.
--
Do
On Dec 11, 2012, at 12:01 AM, Eric Botcazou wrote:
> Even if the Ada language isn't enabled, the Make-lang.in fragment of the Ada
> compiler is processed and invokes gnatls unconditionally, which can result in
> error messages if the command isn't present.
Thank you.
Ok, so here is a new patch, updated according to the suggestions of
David and Jakub. This now only touches the dotted variables, which are
responsible for the AIX trouble. Whether the same prefixing should
also be applied in other cases, we can still decide later.
> The remaining question is if th
On Tue, 2012-12-11 at 11:15 +0100, Andreas Krebbel wrote:
> On 11/12/12 10:16, Kaz Kojima wrote:
> > Andreas Krebbel wrote:
> >> urgs - I'll fix this. Is the patch ok with that change for sh?
> >
> > Yes, the sh portion is OK with that change, though it would be better
> > to fix the users of sh_
This brings the build-requirements up-to-date with us now requiring
a C++ host compiler. I optimistically increased the minimum required
GCC version listed from 2.95 to 3.4 as that is the earliest version
that could reasonably be called a C++98 compatible compiler (yeah,
lawrence will now argue t
On 11/12/12 10:16, Kaz Kojima wrote:
> Andreas Krebbel wrote:
>> urgs - I'll fix this. Is the patch ok with that change for sh?
>
> Yes, the sh portion is OK with that change, though it would be better
> to fix the users of sh_canonicalize_comparison instead of its wrapper
> as pointed out by rth
On Tue, Dec 11, 2012 at 10:12 AM, Eric Botcazou wrote:
>> ... so if would really be a pessimization doing that. Of course handling
>> CONST_DECL in for_each_index is indeed obvious - I just was curios if
>> it was a missed-optimization opportunity as well.
>
> It turns out that, for the same test
On 12/10/2012 06:52 PM, Benjamin De Kosnik wrote:
libstdc++-v3/doc/xsl/customization.xsl.in is marked as
svn:mime-type = application/xml
at least on the 4.7 branch, having some unexpected outcome for svn
diff. If this was unintended, could you change the svn:mime-type
back to text?
This
On Mon, Dec 10, 2012 at 10:07 PM, Mike Stump wrote:
> On Dec 10, 2012, at 12:42 PM, Xinliang David Li wrote:
>> I have not measured the CFI size impact -- but conceivably it should
>> be larger -- which is unfortunate.
>
> Code speed and size are preferable to optimizing dwarf size… :-) I'd let
> ... so if would really be a pessimization doing that. Of course handling
> CONST_DECL in for_each_index is indeed obvious - I just was curios if
> it was a missed-optimization opportunity as well.
It turns out that, for the same testcase, &CONST_DECL is generated on the MIPS
and prepare_decl_r
Andreas Krebbel wrote:
> urgs - I'll fix this. Is the patch ok with that change for sh?
Yes, the sh portion is OK with that change, though it would be better
to fix the users of sh_canonicalize_comparison instead of its wrapper
as pointed out by rth and oleg.
Regards,
kaz
On 11/12/12 00:12, Kaz Kojima wrote:
> Andreas Krebbel wrote:
>> Index: gcc/config/sh/sh.c
>> ===
>> *** gcc/config/sh/sh.c.orig
>> --- gcc/config/sh/sh.c
> [snop]
>> ! static void
>> sh_canonicalize_comparison (enum rtx_code& cmp,
Hi,
on S/390 we have the execute instruction which modifies and executes
an instruction in memory specified by an address operand. On RTL
level (after reload) we embed the target instruction into the execute
pattern. So far we never checked the embedded pattern for validity
what unfortunately ca
Jakub Jelinek writes:
> On Mon, Dec 10, 2012 at 10:44:49PM +0100, Dodji Seketeli wrote:
>> Jakub Jelinek writes:
>>
>> > +++ gcc/asan.c 2012-12-05 15:30:56.069890542 +0100
>> > @@ -1031,7 +1031,7 @@ instrument_builtin_call (gimple_stmt_ite
>> > {
>> >gimple call = gsi_stmt (*iter);
>>
Hi!
As the testcase in the PR (unfortunately only reproduceable after reverting
some inlining patch, so not including that testcase in the patch) shows,
loc_equivalence_change_p isn't good enough for debug_insns, when a REG is
substituted for VOIDmode constant. In that case, the surrounding SUBRE
Hi!
As the testcase shows, for "m" or similar constraints that don't accept
registers, but do accept memory, we want to call *mark_addressable on the
operand, but the call to decay_conversion can change something that was
originally addressable into something non-addressable (be it replacement
of
Even if the Ada language isn't enabled, the Make-lang.in fragment of the Ada
compiler is processed and invokes gnatls unconditionally, which can result in
error messages if the command isn't present.
Tested on x86_64-suse-linux, applied on the mainline.
2012-12-11 Eric Botcazou
* g
94 matches
Mail list logo
