Add myself under Write After Approval.
Commited as r218856.
Thank you,
Prathamesh
Index: MAINTAINERS
===
--- MAINTAINERS (revision 218855)
+++ MAINTAINERS (working copy)
@@ -449,6 +449,7 @@
Michael Koch
Matt Kraai
Jan Kra
This patch is discussed in PR55212
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c77
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c82
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c83
and is to improve code quality with LRA.
SH HI/QImode load/store instructions take r0 as the targ
This patch is discussed in PR55212
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c65
and is to make LRA's register elimination work well on SH.
The problem is SH has very limited add instructions only and expands
rA := rB + N to (set rA (const_int N)) and (set rA (plus rA rB))
instead of (se
This patch is discussed in PR55212
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c40
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c41
and is to fix reload loop issues which happen with movsf_ie/movsi_ie.
It seems that LRA behaves differently from old reload with the choice
of alternati
This patch fixes various issues found when enabling LRA. The essential
works were done by oleg in PR55212.
--
* config/sh/predicates.md (general_movsrc_operand): Allow only
valid plus address expressions.
(general_movdst_operand): Likewise.
(t_reg_operand): Allow (
This patch adds -mlra option to SH. -mno-lra is default ATM.
--
* config/sh/sh.c (sh_lra_p): New function.
(TARGET_LRA_P): Define.
(sh_legitimize_reload_address): Return false if sh_lra_p is true.
* config/sh/sh.opt (mlra): New option.
diff --git a/config/sh/sh.c
This patch adds a target macro discussed in PR55212
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c76
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c78
is to avoid bad codes on SH.
SH has very limited base+displacement addressing and it looks
that the old reload has some mechanism so to
This was discussed in PR55212
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c25
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c58
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c69
and is to fix another ICEs in assign_by_spills.
This patch introduces a new target hook to disable rep
This patch is discussed in PR55212
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c47
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c48
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c49
and is to avoid segfault in remove_pseudos.
In the problematic case in #c47, a call_insn has a no
This was discussed in PR55212
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c52
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c53
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212#c55
and is to fix ICE in assign_by_spills.
In that case, rtlanal.c:decompose_mem_address guesses base and
This patch series is to make SH target use LRA and is discussed
in the PR target/55212:
[SH] Switch to LRA
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55212
I think there are issues which should be solved before defaulting
LRA on SH and the helps from the experts are needed. The series
are cons
On 18 December 2014 at 02:16, Ville Voutilainen
wrote:
> Currently finishing testing on Linux-x64. The only implementation change
> in this patch is the initialization of the iterator variables in
> __minmax_element,
> otherwise it's just adding constexpr.
>
> 2014-12-18 Ville Voutilainen
>
Currently finishing testing on Linux-x64. The only implementation change
in this patch is the initialization of the iterator variables in
__minmax_element,
otherwise it's just adding constexpr.
2014-12-18 Ville Voutilainen
C++14 constexpr min, max, minmax, min_element, max_element
and m
Since we have C++14 sized deallocation it seems we should enable the SD-6
feature test appropriately.
My reading of the announcement and code along with some experimentation
indicates that this feature is allowed in C++11 with -fsized-deallocation.
The feature test macro is thus set whenever fla
Hi,
The test case gcc.target/sh/pr52933-2.c has been failing for a while.
This is because the sh_treg_combine pass optimizes only cbranches but
not cmoves, which on SH are zero-displacement cbranches. The attached
patch makes sh_treg_combine also handle cmoves. This eliminates the
need for some
As testing by Alessandro revealed, vector subscripts weren't properly
handled.
This patch fixes the compiler side (or at least those issues I found).
In particular, for expressions ("get") it wrongly passed a NULL pointer,
additionally, I used the wrong "ar". For it and for assignments/push
(
Hi!
Committed to gomp-4_0-branch in r218849:
commit a6ef9a6d913b6a46872f6f9021d4071e26c6b79c
Author: tschwinge
Date: Wed Dec 17 23:01:26 2014 +
Miscellaneous code cleanup and extension.
gcc/cp/
* parser.c (cp_parser_oacc_clause_vector_length)
(cp_parser_oa
Hi!
Committed to gomp-4_0-branch in r218848:
commit 8fbca6d4d703b08aed3e1c9ae7c227e6d123db24
Author: tschwinge
Date: Wed Dec 17 22:59:07 2014 +
libgomp: Prevent NULL pointer dereference.
libgomp/
* target.c (GOMP_target, GOMP_target_data, GOMP_target_update):
Hi,
When a cbranch immediately follows an SH2A nott insn, the branch
condition can be inverted and the nott insn can be deleted. This is
what the attached patch makes the sh_treg_combine pass do.
Tested with
make -k check RUNTESTFLAGS="--target_board=sh-sim
\{-m2/-ml,-m2/-mb,-m2a/-mb,-m4/-ml,-m4/
Hi!
Committed to gomp-4_0-branch in r218846:
commit cf4f8a78be8da33b991590a1ac24cd001b58bdb7
Author: tschwinge
Date: Wed Dec 17 22:52:20 2014 +
libgomp: Only test with configured (supported) offloading devices.
libgomp/
* testsuite/lib/libgomp.exp: Set offload_tar
Hi!
On Mon, 10 Nov 2014 17:34:30 +0300, Ilya Verbin wrote:
> On 06 Nov 18:55, Jakub Jelinek wrote:
> > Looks mostly good, but:
> >
> > > +# We need more things in site.exp, but automake completely controls the
> > > +# creation of that file; there's no way to append to it without messing
> > >
On Wed, Dec 17, 2014 at 04:14:00PM -0500, Jason Merrill wrote:
> On 12/05/2014 03:28 PM, Jakub Jelinek wrote:
> >3) there is no __has_attribute__ middle-end secondary macro,
> > when it is a built-in macro, it works fine in #ifdef too
>
> I'm not sure what the rationale for the secondary macro w
On Dec 16, 2014, at 4:49 PM, Sergio Durigan Junior wrote:
>
>> Ok.
>
> Can you commit this one for me, please?
Committed revision 218843.
If anyone discovers any nasties, pre-approved to pull the patch out or resolve
the issue in the obvious way. I’d hate to impact the folks with this change
Hi!
Committed to gomp-4_0-branch in r218844:
commit 953f2d61152d9066365c8abb2e4551a17c4d0b83
Author: tschwinge
Date: Wed Dec 17 22:36:40 2014 +
OpenACC testsuite updates.
libgomp/
* testsuite/libgomp.oacc-c++/c++.exp
(check_effective_target_oacc_c): Remove
Hi!
Committed to gomp-4_0-branch in r218842:
commit a192733f55daa0fa24637f89fd3c9a0cd0b8cad3
Author: tschwinge
Date: Wed Dec 17 22:33:35 2014 +
libgomp: OpenACC offloading no longer requires -flto.
libgomp/
* testsuite/libgomp.oacc-c++/c++.exp: Don't add -flto to
Hi!
On Wed, 17 Dec 2014 23:24:17 +0100, I wrote:
> Committed to gomp-4_0-branch in r218839:
>
> commit 1c4f05a68c6d0d5b6137bb6d85a293d16727b389
> Author: tschwinge
> Date: Wed Dec 17 22:23:02 2014 +
>
> libgomp updates.
This has broken libgomp/libgomp_target.h usage from liboffloadmi
Hi!
Committed to gomp-4_0-branch in r218840:
commit febcd8dfdb10fa80edff0880973d1915ca2fef74
Author: tschwinge
Date: Wed Dec 17 22:26:24 2014 +
Use include/gomp-constants.h more actively.
include/
* gomp-constants.h: Update. Change all users.
gcc/
On 12/07/2014 01:52 PM, Sandra Loosemore wrote:
On 11/08/2014 08:57 AM, Sandra Loosemore wrote:
2014-11-08 Sandra Loosemore
Chris Jones
Joshua Conner
gcc/
* config/arm/unknown-elf.h (STARTFILE_SPEC): Add conditional
linking of crtfastmath.o.
* confi
Hi!
On Fri, 12 Dec 2014 21:04:21 +0100, I wrote:
> On Wed, 10 Dec 2014 11:16:08 +0100, Jakub Jelinek wrote:
> > On Wed, Dec 10, 2014 at 11:10:19AM +0100, Thomas Schwinge wrote:
> > > --- /dev/null
> > > +++ gcc/testsuite/c-c++-common/gomp/nesting-1.c
> > > @@ -0,0 +1,77 @@
> > > +void
> > > +f_om
On 11/14/2014 05:46 PM, Sandra Loosemore wrote:
2014-11-14 Sandra Loosemore
Joshua Conner
Chris Jones
gcc/
* doc/invoke.texi (Option Summary): Add -malign-saved-fp-regs.
(ARM options): Document it.
* config/arm/arm.h (arm_stack_offs
Hi!
On Wed, 10 Dec 2014 10:59:34 +0100, Jakub Jelinek wrote:
> On Wed, Dec 10, 2014 at 10:54:13AM +0100, Thomas Schwinge wrote:
> > --- gcc/omp-low.c
> > +++ gcc/omp-low.c
> > @@ -9404,7 +9404,9 @@ build_omp_regions_1 (basic_block bb, struct
> > omp_region *parent,
> >else if (code == GI
I have used the darwin files as templates and made the necessary changes.
It reduces test failures from over 15000 (all run tests) to 27 in 3 files
(patch for PR 48244 applied as well) on NetBSD 7.0 amd64.
--- /dev/null 2014-12-15 07:13:57.0 +
+++ gcc/config/netbsd-protos.h
@@ -0,0
On Dec 17, 2014, at 1:34 AM, Richard Biener wrote:
> Do we even link GCC with libm …?
Yes. :-) If you grew up on C++, you would realize, -lm as you call it, is
just there in C++, always. If you compile with g++ -v, you will even see it.
> Also we don't require a C99 runtime,
With the major
H.J. reported a failure for PR64110 test.
Here is the patch solves it.
The patch was bootstrapped on x86/x86-64/arm.
Committed as rev. 218835.
2014-12-17 Vladimir Makarov
* lra-constraints.c (process_alt_operands): Remove non
allocatable hard regs when considering
i
Hi all,
I have just committed as obvious a one-line fix for an ICE-on-valid
problem with procedure pointer components:
https://gcc.gnu.org/viewcvs/gcc?view=revision&revision=218834
Cheers,
Janus
On 11/27/2014 01:28 AM, tsaund...@mozilla.com wrote:
+ if (warn_override && DECL_VIRTUAL_P (decl) && !DECL_OVERRIDE_P (decl)
Why check DECL_VIRTUAL_P here?
+ warning_at(DECL_SOURCE_LOCATION (decl), OPT_Wsuggest_override,
Space before the (.
Jason
On 11/27/2014 01:28 AM, tsaund...@mozilla.com wrote:
- if (!base_fndecls)
+ if (!base_fndecls.exists ())
This should use is_empty rather than exists. OK with that change.
Jason
On 12/05/2014 03:28 PM, Jakub Jelinek wrote:
3) there is no __has_attribute__ middle-end secondary macro,
when it is a built-in macro, it works fine in #ifdef too
I'm not sure what the rationale for the secondary macro was. Ed?
-cb_has_attribute (cpp_reader *pfile)
+c_common_has_attribute
Now that in C++14 we can have non-constant arguments to constexpr
functions called from another constexpr function, we need to avoid
caching the result, as a later call might produce a different value for
the same non-constant argument, e.g. the address of a local variable.
Tested x86_64-pc-li
I'm uncomfortable with setting TREE_NO_WARNING on a decl just because we
don't want a warning for one particular use of it. How about
suppressing warnings across the call to build_static_cast?
Jason
Hi,
gently pinging this
https://gcc.gnu.org/ml/gcc-patches/2014-12/msg00961.html
On 12/10/2014 08:24 PM, Paolo Carlini wrote:
Hi,
this regression, a spurious warning about taking the address of a
register parameter, happens in C++14 mode due to the use of
force_paren_expr, called by
On 8 September 2014 at 17:25, Alan Lawrence wrote:
> gcc/ChangeLog:
>
> * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): Handle shift
> by 64 by moving const0_rtx.
> (aarch64_ushr_simddi): Delete.
>
> * config/aarch64/aarch64.md (enum unspec): Delete UNSPEC_
Hi for all!
Please, can someone help me, where I found GCC official patchs?
Thanks in advance and best regards,
Ricardo
On 17/12/14 10:28, James Greenhalgh wrote:
>
> Hi,
>
> Reviving this thread:
> https://gcc.gnu.org/ml/gcc-patches/2014-11/msg00143.html
>
> Actually, this is over-engineering. We don't need the generic_sched
> attribute at all. We don't yet have a generic schedule model, and when
> we do get o
On 17 December 2014 at 21:15, Richard Biener wrote:
> On Sun, Dec 14, 2014 at 10:39 PM, Prathamesh Kulkarni
> wrote:
>> Hi,
>> The attached patch flattens cfgloop.h.
>> Created new file cfgloopanal.h which exports function prototypes
>> from cfgloopanal.c. Rest of the patch involves restructuring
Hi Nick,
On Wed, 2014-12-17 17:05:09 +, Nicholas Clifton wrote:
> > 2014-12-17 Jan-Benedict Glaw
> >
> > * config/msp430/msp430.c (msp430_asm_output_addr_const_extra): Fix
> > unused argument warning.
>
> Approved - please apply.
Thanks. Committed as r218828.
MfG, JBG
--
Hi Jan-Benedict,
2014-12-17 Jan-Benedict Glaw
* config/msp430/msp430.c (msp430_asm_output_addr_const_extra): Fix
unused argument warning.
Approved - please apply.
Cheers
Nick
On 17/12/14 16:46, Marcus Shawcroft wrote:
On 17 December 2014 at 15:15, Tejas Belagod wrote:
It isn;t clear to me how far through the various BE patches we need to
get before 59810 is actually resolved?
David's 2 patches
https://gcc.gnu.org/ml/gcc-patches/2014-11/msg01431.html
https://gcc
Hi!
The build robot found this:
g++ -c -g -O2 -DIN_GCC -DCROSS_DIRECTORY_STRUCTURE -fno-exceptions
-fno-rtti -fasynchronous-unwind-tables -W -Wall -Wno-narrowing -Wwrite-strings
-Wcast-qual -Wmissing-format-attribute -Woverloaded-virtual -pedantic
-Wno-long-long -Wno-variadic-macros -Wno-o
On 17 December 2014 at 15:15, Tejas Belagod wrote:
>> It isn;t clear to me how far through the various BE patches we need to
>> get before 59810 is actually resolved?
>
>
> David's 2 patches
>
> https://gcc.gnu.org/ml/gcc-patches/2014-11/msg01431.html
> https://gcc.gnu.org/ml/gcc-patches/2014-10/
Hi,
This change
+2014-12-05 Martin Jambor
+
+ PR ipa/64192
+ * ipa-prop.c (ipa_compute_jump_functions_for_edge): Convert alignment
+ from bits to bytes after checking they are byte-aligned.
+
causes AArch64's vect cost model to vectorize the loops in the
gcc.target/aarch64/vect-movi.c with a
>> gcc/ChangeLog:
>>
>> * config/aarch64/aarch64.md (enum "unspec"): Remove UNSPEC_SSHR64.
>>
>> * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): Change
>> shift
>> amount to 63 if was 64.
>> (aarch64_sshr_simddi): Remove.
OK /Marcus
On 17 December 2014 at 16:19, Alan Lawrence wrote:
> This seems to have slipped under the radar for awhile. Ping.
>
> --Alan
>
>> The sshr_n_64 intrinsics allow performing a signed shift right by 64
>> places. The standard ashrdi3 pattern masks the sign amount with 63, so
>> cannot be used. Howeve
> Well, you'll need FSF approval to relicense - and unless you want to keep
> the same sources used verbatim in both places, the GPL+exception notice is
> the obvious one given such approval. (But in any case, putting in LGPL
> files without a license exception seems a bad idea, because it goes
>
On 12/15/2014 05:24 PM, Jakub Jelinek wrote:
Ok for trunk then.
All the 5 commits are submitted. Thank you very much for your review! :-)
--
Pierre-Marie de Rodat
This seems to have slipped under the radar for awhile. Ping.
--Alan
The sshr_n_64 intrinsics allow performing a signed shift right by 64 places.
The standard ashrdi3 pattern masks the sign amount with 63, so cannot be used.
However, such a shift fills the result by the sign bit, which is iden
On Mon, Dec 15, 2014 at 4:29 PM, Jiong Wang wrote:
>
> On 15/12/14 15:28, Jiong Wang wrote:
>>
>> On 04/12/14 19:32, Jiong Wang wrote:
>>>
>>> On 04/12/14 11:07, Richard Biener wrote:
>>>
On Thu, Dec 4, 2014 at 12:07 PM, Richard Biener
wrote:
>
> On Thu, Dec 4, 2014 at 12:00 PM,
On Sun, Dec 14, 2014 at 10:39 PM, Prathamesh Kulkarni
wrote:
> Hi,
> The attached patch flattens cfgloop.h.
> Created new file cfgloopanal.h which exports function prototypes
> from cfgloopanal.c. Rest of the patch involves restructuring includes.
>
> Boostrapped and tested on x86_64-unknown-linux
On Tue, Dec 16, 2014 at 4:15 PM, Yuri Rumyantsev wrote:
> Hi Richard,
>
> Here is updated patch which includes
> (1) split critical edges for aggressive if conversion.
> (2) delete all stuff related to support of critical edge predication.
> (3) only one function - predicate_scalar_phi performs pr
On Wed, Dec 17, 2014 at 12:17 PM, Martin Liška wrote:
> On 12/11/2014 01:37 PM, Richard Biener wrote:
>>
>> On Wed, Dec 10, 2014 at 1:18 PM, Martin Liška wrote:
>>>
>>> Hello.
>>>
>>> As suggested by Richard, I split compare_operand functions to various
>>> functions
>>> related to a specific com
On 20/11/14 18:19, Marcus Shawcroft wrote:
On 13 November 2014 10:09, David Sherwood wrote:
gcc/:
2014-11-13 David Sherwood
* config/aarch64/aarch64-protos.h (aarch64_simd_attr_length_rglist,
aarch64_reverse_mask): New decls.
* config/aarch64/iterators.
On Wed, Dec 17, 2014 at 10:47 AM, Bin.Cheng wrote:
> On Tue, Dec 16, 2014 at 4:42 PM, Bin.Cheng wrote:
>> On Thu, Dec 11, 2014 at 8:08 PM, Richard Biener
>> wrote:
>>> On Thu, Dec 11, 2014 at 10:56 AM, Bin.Cheng wrote:
On Wed, Dec 10, 2014 at 9:47 PM, Richard Biener
wrote:
> On F
On Wed, Dec 17, 2014 at 03:31:50PM +0100, Uros Bizjak wrote:
> On Wed, Dec 17, 2014 at 2:10 PM, H.J. Lu wrote:
>
> > Add -malign-data={abi|compat,cachineline} to control how GCC aligns
> > variables. "compat" uses increased alignment value compatible with
> > GCC 4.8 and earlier, "abi" uses alig
On Tue, Dec 16, 2014 at 5:05 PM, Martin Liška wrote:
> On 12/11/2014 03:03 PM, Richard Biener wrote:
>>
>> On Thu, Dec 11, 2014 at 2:49 PM, Martin Liška wrote:
>>>
>>> Hello.
>>>
>>> In PR64146, for position independent code IPA ICF should be more careful
>>> about thunk creation.
>>> Patch can b
On Wed, Dec 17, 2014 at 12:07 PM, Martin Liška wrote:
> Hi.
>
> Following patch is taken from
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63573#c18, where
> Honza gave pre-approval for the patch. Jonathan Wakely is waiting for
> approval of the patch
> to commit his new changes.
>
> Tested on x
On Wed, 17 Dec 2014, Jakub Jelinek wrote:
> Hi!
>
> MEM_REF (the only tcc_reference code with 2 operands) has TREE_READONLY set
> whenever all the arguments are TREE_READONLY, which is wrong, if the
> pointer/reference is read-only, it doesn't say anything about whether
> what it points to is als
On Wed, Dec 17, 2014 at 2:10 PM, H.J. Lu wrote:
> Add -malign-data={abi|compat,cachineline} to control how GCC aligns
> variables. "compat" uses increased alignment value compatible with
> GCC 4.8 and earlier, "abi" uses alignment value as specified by the
> psABI, and "cacheline" with increased
On 17/12/14 00:29, Joseph Myers wrote:
On Tue, 16 Dec 2014, Renlin Li wrote:
Hi all,
According to ACLE 2.0, the value of __ARM_SIZEOF_WCHAR_T should be defined in
terms of byte, which means it should be 2 or 4. This patch corrects the error
in arm backend.
What error? builtin_define_type_siz
On Wed, Dec 17, 2014 at 1:54 AM, Uros Bizjak wrote:
> On Tue, Dec 16, 2014 at 8:58 PM, Uros Bizjak wrote:
>
> The (now attached, to avoid mailer mangling) patch also fixes
> mentioned build warnings.
>
> Bootstrapped and regression tested on CentOS 5.11.
>
> OK for mainline?
Thanks. Committed.
On Wed, Dec 17, 2014 at 2:14 PM, Jakub Jelinek wrote:
> On Wed, Dec 17, 2014 at 05:10:36AM -0800, H.J. Lu wrote:
>> PR target/61296
>> * config/i386/i386-opts.h (ix86_align_data): New enum.
>> * config/i386/i386.c (ix86_data_alignment): Return the ABI
>> alignment value for
On Wed, Dec 17, 2014 at 02:14:34PM +0100, Jakub Jelinek wrote:
> On Wed, Dec 17, 2014 at 05:10:36AM -0800, H.J. Lu wrote:
> > PR target/61296
> > * config/i386/i386-opts.h (ix86_align_data): New enum.
> > * config/i386/i386.c (ix86_data_alignment): Return the ABI
> > alignment value
Hi!
MEM_REF (the only tcc_reference code with 2 operands) has TREE_READONLY set
whenever all the arguments are TREE_READONLY, which is wrong, if the
pointer/reference is read-only, it doesn't say anything about whether
what it points to is also read-only.
Fixed thusly, bootstrapped/regtested on x
On Wed, Dec 17, 2014 at 05:10:36AM -0800, H.J. Lu wrote:
> PR target/61296
> * config/i386/i386-opts.h (ix86_align_data): New enum.
> * config/i386/i386.c (ix86_data_alignment): Return the ABI
> alignment value for -malign-data=abi, the cachine line size
> for -malign-
> From: Jakub Jelinek
>
> On Wed, Dec 17, 2014 at 09:33:23AM +0100, Uros Bizjak wrote:
> > Oops, I was under impression that HJ sent the patch on public list.
>
> The ix86_data_alignment change looks too ugly and too complicated to me.
>
> I think it would be much cleaner to add after:
>
> /
On Wed, Dec 17, 2014 at 10:54 AM, Uros Bizjak wrote:
> On Tue, Dec 16, 2014 at 8:58 PM, Uros Bizjak wrote:
>
When building libgo on CentOS 5.11, following warnings appear:
In file included from /usr/include/fcntl.h:38:0,
from sysinfo.c:6:
/home/uros/gcc-b
On December 17, 2014 12:29:44 PM CET, Marek Polacek wrote:
>On Wed, Dec 17, 2014 at 10:12:06AM +0100, Richard Biener wrote:
>> On Wed, 17 Dec 2014, Jakub Jelinek wrote:
>>
>> > On Wed, Dec 17, 2014 at 09:46:44AM +0100, Marek Polacek wrote:
>> > > This adds a transformation of (x & ~m) | (y & m),
On Wed, Dec 17, 2014 at 10:12:06AM +0100, Richard Biener wrote:
> On Wed, 17 Dec 2014, Jakub Jelinek wrote:
>
> > On Wed, Dec 17, 2014 at 09:46:44AM +0100, Marek Polacek wrote:
> > > This adds a transformation of (x & ~m) | (y & m), which (on GIMPLE)
> > > has 4 ops to x ^ ((x ^ y) & m) that has 3
On 12/11/2014 01:37 PM, Richard Biener wrote:
On Wed, Dec 10, 2014 at 1:18 PM, Martin Liška wrote:
Hello.
As suggested by Richard, I split compare_operand functions to various
functions
related to a specific comparison. Apart from that I added fast check for
volatility flag that caused miscomp
On Wed, Dec 17, 2014 at 11:30 AM, Tejas Belagod wrote:
> On 07/12/14 07:38, Jan Hubicka wrote:
>>
>> Hi,
>> this is simplified patch that only adds the equal_address_to predicate
>> (and thus fixes issues
>> with inccorect folding of speculative calls). Hopefully it will mek it
>> easier to handle
Hi.
Following patch is taken from
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63573#c18, where
Honza gave pre-approval for the patch. Jonathan Wakely is waiting for approval
of the patch
to commit his new changes.
Tested on x86_64-linux-pc and no new regression was seen.
Ready for trunk?
Tha
On 07/12/14 07:38, Jan Hubicka wrote:
Hi,
this is simplified patch that only adds the equal_address_to predicate (and
thus fixes issues
with inccorect folding of speculative calls). Hopefully it will mek it easier
to handle
the rest of fold-const incrementally.
Bootstrapped/regtested x86_64-lin
Hi,
Reviving this thread:
https://gcc.gnu.org/ml/gcc-patches/2014-11/msg00143.html
Actually, this is over-engineering. We don't need the generic_sched
attribute at all. We don't yet have a generic schedule model, and when
we do get one, we have other ways of handling it through
aarch64-cores.d
On Tue, Dec 16, 2014 at 8:58 PM, Uros Bizjak wrote:
>>> When building libgo on CentOS 5.11, following warnings appear:
>>>
>>> In file included from /usr/include/fcntl.h:38:0,
>>> from sysinfo.c:6:
>>> /home/uros/gcc-build/gcc/include-fixed/sys/stat.h:317:16: warning:
>>> inline
On Tue, Dec 16, 2014 at 4:42 PM, Bin.Cheng wrote:
> On Thu, Dec 11, 2014 at 8:08 PM, Richard Biener
> wrote:
>> On Thu, Dec 11, 2014 at 10:56 AM, Bin.Cheng wrote:
>>> On Wed, Dec 10, 2014 at 9:47 PM, Richard Biener
>>> wrote:
On Fri, Dec 5, 2014 at 1:15 PM, Bin Cheng wrote:
> Hi,
On Tue, Dec 16, 2014 at 11:53 PM, Jan Hubicka wrote:
>> > gcc/ChangeLog:
>> > 2014-11-18 Dehao Chen
>> >
>> > * auto-profile.c (afdo_annotate_cfg): Invoke update_ssa in the
>> > right
>> > place.
>> > (auto_profile): Recompute inline summary after processing cgraph
>>
On Tue, Dec 16, 2014 at 9:37 PM, Jan Hubicka wrote:
> Hi,
> I have optimized inliner to take advantage that the heap is now in sreal and
> found that using small numbers leads to excessive time (7% of WPA) spent by
> sreal::normalize. This is because the normalization is implemented by loops
> th
On Tue, Dec 16, 2014 at 8:05 PM, Jan Hubicka wrote:
>> On Tue, Dec 16, 2014 at 5:04 PM, Michael Haubenwallner
>> wrote:
>> > Recent commit 218765 adding sreal::to_double() breaks on AIX due to math.h
>> > being included before _LARGE_FILES and __STDC_FORMAT_MACROS being defined
>> > later in conf
On Wed, 17 Dec 2014, Jakub Jelinek wrote:
> On Wed, Dec 17, 2014 at 09:46:44AM +0100, Marek Polacek wrote:
> > This adds a transformation of (x & ~m) | (y & m), which (on GIMPLE)
> > has 4 ops to x ^ ((x ^ y) & m) that has 3 ops on GIMPLE. In fact,
> > the latter is then transformed to (a ^ b) &
On Wed, Dec 17, 2014 at 09:46:44AM +0100, Marek Polacek wrote:
> This adds a transformation of (x & ~m) | (y & m), which (on GIMPLE)
> has 4 ops to x ^ ((x ^ y) & m) that has 3 ops on GIMPLE. In fact,
> the latter is then transformed to (a ^ b) & m ^ a, which also has 3
> ops.
So why don't you tr
> -Original Message-
> From: Jakub Jelinek [mailto:ja...@redhat.com]
> Sent: Wednesday, December 17, 2014 4:20 PM
> To: Terry Guo
> Cc: gcc-patches@gcc.gnu.org
> Subject: Re: [Backport]Is it ok to backport this bug fix to 4.8 branch?
>
> On Wed, Dec 17, 2014 at 04:13:59PM +0800, Terry Guo
This adds a transformation of (x & ~m) | (y & m), which (on GIMPLE)
has 4 ops to x ^ ((x ^ y) & m) that has 3 ops on GIMPLE. In fact,
the latter is then transformed to (a ^ b) & m ^ a, which also has 3
ops.
As discussed in the PR, the original form might also be 3-op on targets
that support someth
On Wed, Dec 17, 2014 at 04:13:59PM +0800, Terry Guo wrote:
> Is it OK to back port revision 209293 to upstream 4.8 branch as it fixed bug
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60663? Thanks.
I don't think it is ok, certainly not for 4.8.4 now. See PR63637.
Jakub
Hi Jakub,
Is it OK to back port revision 209293 to upstream 4.8 branch as it fixed bug
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60663? Thanks.
BR,
Terry
93 matches
Mail list logo
