Re: [Patch,optimization]: Optimized changes in the estimate register pressure cost.

On Wed, Sep 30, 2015 at 12:00 AM, Pat Haugen wrote: > On 09/25/2015 11:51 PM, Ajit Kumar Agarwal wrote: >> >> I have made the following changes in the estimate_reg_pressure_cost >> function used >> by the loop invariant and IVOPTS. >> >> Earlier the estimate_reg_pressure cost uses the cost of n_ne

Re: PR66311: Fix extension in mpz->wide_int conversions

On Aug 5, 2015, at 7:14 AM, Richard Sandiford wrote: > wi::from_mpz reads the absolute value of the mpz and then negates the > result if the mpz is negative. When the top bit of the most-significant > HOST_WIDE_INT in the absolute value is set, we implicitly sign- > rather than zero-extend it to

[PATCH] Properly handle -miamcu and -march=lakemont

-miamcu specifies how parameters are passed to functions and how value is returned from a function. Inside function, we can use instructions supported by -march=XXX. -miamcu -march=haswell can generate x87, SSE and AVX instructions as long as the IA MCU psABI is followed. But since Lakemont pro

Re: [PATCH, sh][v3] musl support for sh

On Fri, 2015-10-16 at 17:06 +0100, Szabolcs Nagy wrote: > Revision of > https://gcc.gnu.org/ml/gcc-patches/2015-04/msg01636.html > > The musl dynamic linker name is /lib/ld-musl-sh{-nofpu}{-fdpic}.so.1 > > New in this revision: > > Add -fdpic to the name, will be useful with the pending sh2 FDPI

Re: [PATCH] 2015-10-15 Benedikt Huber Philipp Tomsich

On Fri, 2015-10-16 at 15:47 +0100, Marcus Shawcroft wrote: > On 16 October 2015 at 15:31, Benedikt Huber > wrote: > > I introduced this in revision 7 due to a request from James Greenhalgh. > > https://gcc.gnu.org/ml/gcc-patches/2015-10/msg00963.html > > > >> Given that this is all so mechanical,

[PATCH] change a function argument from rtx to rtx_insn *

From: Trevor Saunders Hi, $subject, just a small improvement to rtx_insn type safety I happened to notice. bootstrapped + regtested x86_64-linux-gnu, committing to trunk. Trev gcc/ChangeLog: 2015-10-16 Trevor Saunders * lra-constraints.c (add_next_usage_insn): Change argument t

Re: [PATCH, wwwdocs] Add -march=skylake-avx512 to gcc-6/changes.html.

On Fri, 16 Oct 2015, Kirill Yukhin wrote: > Is it ok to install? Yes, just add a "the" before "following". > This switch was backported to gcc-5. > Is it ok to create a new section `GCC 5.3' and put it there > or I need to wait for actual release? How about if you add it now, and add "(pending)"

Re: [PATCH] PR66870 PowerPC64 Enable gold linker with split stack

On Sun, Oct 11, 2015 at 6:07 AM, Alan Modra wrote: > > gcc/ > * config/rs6000/sysv4.h (TARGET_CAN_SPLIT_STACK_64BIT): Don't define. > * config/rs6000/linux64.h (TARGET_CAN_SPLIT_STACK): Define. > (TARGET_CAN_SPLIT_STACK_64BIT): Define. > gcc/go/ > * gospec.c (saw_op

Re: [c++-delayed-folding] First stab at convert_to_integer

On 10/16/2015 07:35 AM, Marek Polacek wrote: This code path seems to be for pushing a conversion down into a binary expression. We shouldn't do this at all when we aren't folding. I tend to agree, but this case is tricky. What's this code about is e.g. for int fn (long p, long o) { return

Re: [PATCH] PR fortran/67987 -- character lengths cannot be negative

On Fri, Oct 16, 2015 at 02:55:27PM -0700, Steve Kargl wrote: > On Fri, Oct 16, 2015 at 11:39:51PM +0200, FX wrote: > > > - why remove the -Wsurprising warning? it seems a good case > > for -Wsurprising: legal code, but dubious anyway > > > > OK after you ponder that second point. > > > > F90

[PATCH] PR target/67995: __attribute__ ((target("arch=XXX"))) enables unsupported ISA

When processing __attribute__ ((target("arch=XXX"))), we should clear the ISA bits in x_ix86_isa_flags first to avoid leaking ISA from command line. Tested on x86-64. OK for trunk? Thanks. H.J. --- gcc/ PR target/67995 * config/i386/i386.c (ix86_valid_target_attribute_tree): If

Re: [PATCH] PR fortran/67987 -- character lengths cannot be negative

On Fri, Oct 16, 2015 at 11:39:51PM +0200, FX wrote: > > 2015-10-16 Steven G. Kargl > > > > PR fortran/67987 > > * decl.c (char_len_param_value): Unwrap unlong line. If LEN < 0, > > then force it to zero pre Fortran Standards. > > * resolve.c (gfc_resolve_substring_charlen): Un

Re: [PATCH: RL78] libgcc fixes for divmodsi, divmodhi and divmodqi

> This is regression tested for RL78 -msim. Please let me know if it is > OK to commit. I've committed this patch for you. Thanks! > Best Regards, > Kaushik > > Changelog: > 2015-08-21 Kaushik Phatak > > * config/rl78/divmodqi.S: Return 0x00 by default for div by 0. > * config/rl78/divmods

Re: Add VIEW_CONVERT_EXPR to operand_equal_p

On October 16, 2015 5:55:08 PM GMT+02:00, Eric Botcazou wrote: >> I wasn't aware that x86/IA-64 is still broken. I am flying to NY >tomorrow >> but will try to take a look. The ICEs are not caused by >operand_equal_p >> changes, but the change to useless_type_conversion to ignore mode on >> aggr

Re: [PATCH] PR fortran/67987 -- character lengths cannot be negative

> 2015-10-16 Steven G. Kargl > > PR fortran/67987 > * decl.c (char_len_param_value): Unwrap unlong line. If LEN < 0, > then force it to zero pre Fortran Standards. > * resolve.c (gfc_resolve_substring_charlen): Unwrap unlong line. > If 'start' is larger than 'end

Re: [PATCH] PR fortran/67987 -- character lengths cannot be negative

On Fri, Oct 16, 2015 at 10:17:34PM +0200, FX wrote: > > The attach patch enforces the Fortran Standard's requirement > > that character length must be great than or equal to zero. > > 4.4.3.2. "If the character length parameter value evaluates to > a negative value, the length of character entitie

[committed, PATCH] Disable X86_TUNE_ALWAYS_FANCY_MATH_387 for Lakemont

Since Lakemont processor doesn't have 387, we should disable X86_TUNE_ALWAYS_FANCY_MATH_387 for Lakemont. * i386/x86-tune.def (X86_TUNE_ALWAYS_FANCY_MATH_387): Disable for Lakemont. --- gcc/config/i386/x86-tune.def | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git

[Patch, MIPS] Frame header optimization for MIPS (part 2)

Here is the second part of the MIPS frame header optimization patch. The first part avoided allocating a frame header if it knew that none of the functions that it called would need it. This part uses a frame header to save callee saved registers if doing so will allow the function to avoid havi

Re: [PATCH] PR fortran/67987 -- character lengths cannot be negative

> The attach patch enforces the Fortran Standard's requirement > that character length must be great than or equal to zero. We've got to be careful about this. The standard (F2008) has this to say about character lengths: 4.4.3.1. "The number of characters in the string is called the length of t

Re: config header file reduction patch checked in.

On 10/16/2015 03:49 PM, Andrew MacLeod wrote: On 10/12/2015 04:04 AM, Jeff Law wrote: On 10/08/2015 07:37 AM, Andrew MacLeod wrote: On 10/07/2015 06:02 PM, Jeff Law wrote: I'm slightly concerned about the darwin, windows and solaris bits. The former primarily because Darwin has been a gener

Re: [Patch, fortran] COMMON block error recovery: PR 67758 (second pass)

On Tue, Oct 06, 2015 at 07:52:16PM +0200, Mikael Morin wrote: > > Dominique noticed that the test coming with the preceding PR67758 patch > [1] was failing if compiled as free form. > [1] https://gcc.gnu.org/ml/gcc-patches/2015-10/msg00301.html > > The problem is again an inconsistent state, but

config header file reduction patch checked in.

On 10/12/2015 04:04 AM, Jeff Law wrote: On 10/08/2015 07:37 AM, Andrew MacLeod wrote: On 10/07/2015 06:02 PM, Jeff Law wrote: I'm slightly concerned about the darwin, windows and solaris bits. The former primarily because Darwin has been a general source of pain, and in the others because I

Re: [PATCH] c/67882 - improve -Warray-bounds for invalid offsetof

On 10/16/2015 06:27 AM, Bernd Schmidt wrote: On 10/09/2015 04:55 AM, Martin Sebor wrote: Gcc attempts to diagnose invalid offsetof expressions whose member designator is an array element with an out-of-bounds index. The logic in the function that does this detection is incomplete, leading to fal

[committed, PATCH] Replace printf with __builtin_printf

Index: ChangeLog === --- ChangeLog (revision 228921) +++ ChangeLog (working copy) @@ -1,5 +1,10 @@ 2015-10-16 H.J. Lu + * gcc.target/i386/iamcu/test_basic_64bit_returning.c (main): + Replace printf with __builtin_

[committed, PATCH] Properly initialize u5

Index: ChangeLog === --- ChangeLog (revision 228920) +++ ChangeLog (working copy) @@ -1,3 +1,8 @@ +2015-10-16 H.J. Lu + + * gcc.target/i386/iamcu/test_passing_unions.c (main): Properly + initialize u5. + 2015-10-16

[PATCH] PR fortran/67987 -- character lengths cannot be negative

The attach patch enforces the Fortran Standard's requirement that character length must be great than or equal to zero. The fix submitted here supercedes the fix for PR fortran/31250, which silently converted a negative string length to zero. In removing the fix for 31250, a regression occurred, be

Re: [PATCH, rs6000][v3] powerpc musl libc support

On 16/10/15 17:35, Segher Boessenkool wrote: Hi! On Fri, Oct 16, 2015 at 04:58:06PM +0100, Szabolcs Nagy wrote: #if DEFAULT_LIBC == LIBC_UCLIBC -#define CHOOSE_DYNAMIC_LINKER(G, U) "%{mglibc:" G ";:" U "}" +#define CHOOSE_DYNAMIC_LINKER(G, U, M) \ + "%{mglibc:" G ";:%{mmusl:" M ";:" U "}}"

[gomp4] small oacc cleanup

A small patch committed to gomp4. 1) extract_oacc_routine_gwv does more work than necessary, we just nee to check if there are oacc function attributes. 2) We still want to run the oacc_device_lower pass, even if errors were detected. (a) we still need to process the loop markers etc (b

[PATCH] Use GET_MODE_BITSIZE to get vector natural alignment

Since GET_MODE_ALIGNMENT is defined by psABI and the biggest alignment is 4 byte for IA MCU psABI, we should use GET_MODE_BITSIZE to get vector natural alignment to check misaligned vector move. OK for trunk? Thanks. H.J. --- * config/i386/i386.c (ix86_expand_vector_move): Use GE

[c++-delayed-folding] First stab at convert_to_integer

I felt like it'd be good to resolve some unfinished business in convert_to_integer before starting messing with other convert_to_* functions. This is a response to . > > + if (!dofold) > > +{ > > + expr = build1 (CONVE

[hsa] Allow gridification of loop pre_bodies

Hi, the patch below allows gridification and thus fast execution on HSA GPUs of loops even when they have some statements in their pre-bodies. It also moves the bulk of target construct preparation for gridification to even before omp scanning, which should considerably ease transition to OpenMP

[PATCH] Fix def_test_returning_type in iamcu/test_basic_returning.c

Use union to check float return bits to avoid converting from integer to float when comparing float return value. I will check it in after regression test. H.J. --- * gcc.target/i386/iamcu/test_basic_returning.c (def_test_returning_type): Use union to check float return bits. ---

Re: [PATCH] c/67882 - improve -Warray-bounds for invalid offsetof

On Fri, 16 Oct 2015, Bernd Schmidt wrote: > > +// The following expression is silently accepted as an extension > > +// because it simply forms the equivalent of a just-past-the-end > > +// address. > > +__builtin_offsetof (A, a1_1 [0][1]),// extension > > Hmm, do we really wa

Re: [PATCH] Eliminate libgo.so dependency for go and gofmt tools in gotools/

On Fri, Oct 16, 2015 at 9:28 AM, Lynn A. Boger wrote: > To make it easier to use gotools go and gofmt, add -static-libgo when > linking them. This makes it nicer > in all cases since it eliminates the dependency on libgo.so for these tools, > but it is mostly helpful for > the cross compiled case

[PATCH] 2015-10-15 Benedikt Huber Philipp Tomsich

* config/aarch64/aarch64-builtins.c: Builtins for rsqrt and rsqrtf. * config/aarch64/aarch64-protos.h: Declare. * config/aarch64/aarch64-simd.md: Matching expressions for frsqrte and frsqrts. * config/aarch64/aarch64-tuning-flags.def: Added recip_sqrt.

[PATCH v9][aarch64] Implemented reciprocal square root (rsqrt) estimation in -ffast-math

This ninth revision of the patch: * Removes unnecessary typedef. Ok for check in. Benedikt Huber (1): 2015-10-15 Benedikt Huber Philipp Tomsich gcc/ChangeLog | 20 gcc/config/aarch64/aarch64-builtins.c | 115 ++

Re: [PATCH, rs6000][v3] powerpc musl libc support

Hi! On Fri, Oct 16, 2015 at 04:58:06PM +0100, Szabolcs Nagy wrote: > #if DEFAULT_LIBC == LIBC_UCLIBC > -#define CHOOSE_DYNAMIC_LINKER(G, U) "%{mglibc:" G ";:" U "}" > +#define CHOOSE_DYNAMIC_LINKER(G, U, M) \ > + "%{mglibc:" G ";:%{mmusl:" M ";:" U "}}" > #elif DEFAULT_LIBC == LIBC_GLIBC > -#de

Re: Do not describe -std=c11 etc. as experimental in c.opt help text

The changes to C++ option descriptions are OK in the absence of C++ front-end maintainer objections within 96 hours. -- Joseph S. Myers jos...@codesourcery.com

[PATCH] Eliminate libgo.so dependency for go and gofmt tools in gotools/

To make it easier to use gotools go and gofmt, add -static-libgo when linking them. This makes it nicer in all cases since it eliminates the dependency on libgo.so for these tools, but it is mostly helpful for the cross compiled case. When building for use in the cross compiled case, this allo

Re: RFC: always default to -mno-unaligned-access for bare-metal ARM

On 16/10/15 16:28, Sandra Loosemore wrote: > Recently I tracked down a target crash problem in an ARM EABI configuration > running on a Cortex-A9 board to an unaligned access fault. The startup code > provided by the customer for this board doesn't enable the MMU, and unaligned > access suppo

[gomp4.5] Allow arbitrary low-bound on C/C++ array sections in reduction clauses

Hi! The public draft required low bound in array sections to be zero (but it was unclear if at compile time or at runtime only). After discussions, the whole restriction got removed, so now low-bound can be arbitrary. The following patch implements that. Committed to gomp-4_5-branch after retest

[PATCH, sh][v3] musl support for sh

Revision of https://gcc.gnu.org/ml/gcc-patches/2015-04/msg01636.html The musl dynamic linker name is /lib/ld-musl-sh{-nofpu}{-fdpic}.so.1 New in this revision: Add -fdpic to the name, will be useful with the pending sh2 FDPIC support. 2015-10-16 Gregor Richards Szabolcs Nagy

[PATCH, rs6000][v3] powerpc musl libc support

Revision of https://gcc.gnu.org/ml/gcc-patches/2015-04/msg01640.html This adds musl libc support to powerpc: - The dynamic-linker name is /lib/ld-musl-powerpc{64}{le}{-sf}.so.1 - The include order is changed for musl as in linux.h. New in this revision: Soft float abi detection through "{msoft

Re: Add VIEW_CONVERT_EXPR to operand_equal_p

> I wasn't aware that x86/IA-64 is still broken. I am flying to NY tomorrow > but will try to take a look. The ICEs are not caused by operand_equal_p > changes, but the change to useless_type_conversion to ignore mode on > aggregate types. Sure, but I'd like to avoid hiding new problems against p

Re: Benchmarks of v2 (was Re: [PATCH 0/5] RFC: Overhaul of diagnostics (v2))

On Wed, 2015-10-14 at 11:00 +0200, Richard Biener wrote: > On Tue, Oct 13, 2015 at 5:32 PM, David Malcolm wrote: > > On Thu, 2015-09-24 at 10:15 +0200, Richard Biener wrote: > >> On Thu, Sep 24, 2015 at 2:25 AM, David Malcolm wrote: > >> > On Wed, 2015-09-23 at 15:36 +0200, Richard Biener wrote:

[PATCH, rs6000] Enable secureplt by default on musl

The musl dynamic loader can only deal with secure-plt, make it the default. Split out from https://gcc.gnu.org/ml/gcc-patches/2015-04/msg01640.html for easier review (independent of the rest of the patch). gcc/ChangeLog: 2015-10-16 Gregor Richards Szabolcs Nagy * confi

[gomp4,committed] Handle bind clause in dump_omp_clause

Hi, this patch handles the oacc bind clause in dump_omp_clause. Committed to gomp-4_0-branch. Thanks, - Tom Handle bind clause in dump_omp_clause 2015-10-16 Tom de Vries * tree-pretty-print.c (dump_omp_clause): Handle bind clause. --- gcc/tree-pretty-print.c | 6 ++ 1 file changed, 6

RFC: always default to -mno-unaligned-access for bare-metal ARM

Recently I tracked down a target crash problem in an ARM EABI configuration running on a Cortex-A9 board to an unaligned access fault. The startup code provided by the customer for this board doesn't enable the MMU, and unaligned access support requires the MMU to be enabled per http://infoc

Re: [PATCH 5/9] i386: Add address spaces for fs/gs segments

On 08/10/2015 06:59, Richard Henderson wrote: > +/* Address space support. > + > + This is not "far pointers" in the 16-bit sense, but an easy way > + to use %fs and %gs segment prefixes. Therefore: > + > +(a) All address spaces have the same modes, > +(b) All address spaces have the

[PATCH] tree-scalar-evolution.c: Handle LSHIFT by constant

This lets the vectorizer handle some simple strides expressed using left-shift rather than mul, e.g. a[i << 1] (whereas previously only a[i * 2] would have been handled). This patch does *not* handle the general case of shifts - neither a[i << j] nor a[1 << i] will be handled; that would be a sign

[gomp4, committed] Add nohost clause support in dump_omp_clause

Hi, this patch adds nohost clause support in dump_omp_clause. Furthermore, it fixes an infinite recursion bug when handling unrecognized clauses. Committed to gomp-4_0-branch. Thanks, - Tom Add nohost clause support in dump_omp_clause 2015-10-16 Tom de Vries * tree-pretty-print.c (dump_

Re: Do not use TYPE_CANONICAL in useless_type_conversion

Jan Hubicka writes: >> Jan Hubicka writes: >> >> > Does the patch in https://gcc.gnu.org/ml/gcc-patches/2015-10/msg00902.html >> > help? >> >> No, it doesn't. >> > Andreas, > I am sorry for getting late to this. I hoped that the alternative patch by > Alexandre would fix this. > I still don

Re: [PATCH] 2015-10-15 Benedikt Huber Philipp Tomsich

On 16 October 2015 at 15:31, Benedikt Huber wrote: > I introduced this in revision 7 due to a request from James Greenhalgh. > https://gcc.gnu.org/ml/gcc-patches/2015-10/msg00963.html > >> Given that this is all so mechanical, I'd have a preference towards >> refactoring this to loop over some str

[PATCH, wwwdocs] Add -march=skylake-avx512 to gcc-6/changes.html.

Hello, Patch in the bottom adds mentioning of new `march=skylake-avx512' to gcc-6/changes.html. Is it ok to install? This switch was backported to gcc-5. Is it ok to create a new section `GCC 5.3' and put it there or I need to wait for actual release? -- Thanks, K Index: htdocs/gcc-6/changes.ht

[AArch64] Update comments on the usage of X30 in FIXED_REGISTERS and CALL_USED_REGISTERS

The patch https://gcc.gnu.org/ml/gcc-patches/2014-09/msg02654.html from last year changed the definition of LR in CALL_USED_REGISTERS, but didn't update the comment above the #define to reflect the new usage. This patch bring the comment inline with the implementation. OK for trunk? Thanks. 20

Re: [PATCH] 2015-10-15 Benedikt Huber Philipp Tomsich

I introduced this in revision 7 due to a request from James Greenhalgh. https://gcc.gnu.org/ml/gcc-patches/2015-10/msg00963.html > Given that this is all so mechanical, I'd have a preference towards > refactoring this to loop over some structured data. Do you mean, that I should get rid of the ty

Re: [PATCH] 2015-10-15 Benedikt Huber Philipp Tomsich

On 16 October 2015 at 14:59, Benedikt Huber wrote: > + typedef struct > + { > +tree type_node; > +const char *builtin_name; > +int function_code; > + } builtin_decls_data; Please address Oleg's comment. Cheers /Marcus

Re: [PATCH 1/7] Libsanitizer merge from upstream r249633.

On 16 October 2015 at 14:59, Maxim Ostapenko wrote: > Yeah, thanks. Just wondering if I should step back until they are resolved > upstream or we can have another merge in the future (stage3 is coming ...)? Well, right now, the support is patchy, experimental, but it's reasonably stable. From the

[PATCH] 2015-10-15 Benedikt Huber Philipp Tomsich

* config/aarch64/aarch64-builtins.c: Builtins for rsqrt and rsqrtf. * config/aarch64/aarch64-protos.h: Declare. * config/aarch64/aarch64-simd.md: Matching expressions for frsqrte and frsqrts. * config/aarch64/aarch64-tuning-flags.def: Added recip_sqrt.

Re: [PATCH 1/7] Libsanitizer merge from upstream r249633.

On 16/10/15 16:48, Renato Golin wrote: On 14 October 2015 at 19:38, Renato Golin wrote: On 14 October 2015 at 19:21, Evgenii Stepanov wrote: Wait. As Jakub correctly pointed out in the other thread, there is no obvious reason why there could not be a single shadow offset value that would work

[PATCH v8][aarch64] Implemented reciprocal square root (rsqrt) estimation in -ffast-math

This eighth revision of the patch: * Style improvements. Ok for check in. Benedikt Huber (1): 2015-10-15 Benedikt Huber Philipp Tomsich gcc/ChangeLog | 20 gcc/config/aarch64/aarch64-builtins.c | 115 ++

[patch] Document options for building and linking to libstdc++fs.a

This documents how to use the Filsystem TS library. Committed to trunk. commit 2d8dfef4311b51a2743f5ab722d467792c7c32dd Author: Jonathan Wakely Date: Fri Oct 16 14:53:46 2015 +0100 Document options for Filesystem TS library * doc/xml/manual/configure.xml: Document --enabl

Re: [PR67383][ARM][4.9]Backport of "Allow any register for DImode values in Thumb2"

Hi Ramana, On 16/10/15 11:52, Ramana Radhakrishnan wrote: On Thu, Oct 15, 2015 at 03:01:24PM +0100, Renlin Li wrote: Hi all, This is a backport patch to loosen restrictions on core registers for DImode values in Thumb2. It fixes PR67383. In this particular case, reload tries to spill a hard r

[Ada] Pragma Constant_After_Elaboration

This patch implements the legality rules of pragma Constant_After_Elaboration: The Boolean aspect Constant_After_Elaboration may be specified as part of the declaration of a library level variable. The semantic checks of this annotation cannot be performed by the compiler as this requires f

[Ada] Spurious error on SPARK_Mode in generic package instantiation

This patch modifies the generic instantiation to ensure that a context with a missing SPARK_Mode annotation is treated as having SPARK_Mode set to Off. This ensures that the following SPARK UG rule 9.4.1 Code where SPARK_Mode is Off shall not enclose code where Spark_Mode is On. However, if

Re: [PATCH 1/7] Libsanitizer merge from upstream r249633.

On 14 October 2015 at 19:38, Renato Golin wrote: > On 14 October 2015 at 19:21, Evgenii Stepanov > wrote: >> Wait. As Jakub correctly pointed out in the other thread, there is no >> obvious reason why there could not be a single shadow offset value >> that would work for all 3 possible VMA setti

[PATCH] Correctly fill up cgraph_node::local.versionable flag.

Hello. I've been working on HSA branch, where we have a cloning pass running with all optimization levels. The patch makes computation of cgraph_node::local.versionability independent on IPA CP and uses the flag to verify that a function can be cloned. The patch can bootstrap on x86_64-linux-pc

Re: [PATCH] i386: Use the STC bb-reorder algorithm at -Os (PR67864)

On Fri, Oct 16, 2015 at 02:55:54PM +0200, Bernd Schmidt wrote: > On 10/16/2015 02:53 PM, Segher Boessenkool wrote: > >For x86, STC still gives better results for optimise-for-size than > >"simple" does. So use STC at -Os as well. > > For how many targets is this true, and for the others, what is

[Ada] Check suppression in Ada.Containers

This patch implements two new check names (Container_Checks and Tampering_Check) that may be used with pragma Suppress. Suppressing Tampering_Check suppresses checks for "tampering with cursors" and "tampering with elements". If pragma Suppress(Tampering_Check) is in force at the point of instantia

Re: [PATCH][haifa-sched] model load/store multiples properly in autoprefetcher scheduling

On 16/10/15 04:55, Vladimir Makarov wrote: On 10/15/2015 11:27 AM, Kyrill Tkachov wrote: On 15/10/15 11:16, Bernd Schmidt wrote: On 10/15/2015 11:40 AM, Kyrill Tkachov wrote: The code that analyzes the offsets of the loads/stores doesn't try to handle load/store-multiple insns. These appear

[Ada] Minor cleanup in finalization support of the runtime

This removes a couple of redundant/unused things. No functional changes. Tested on x86_64-pc-linux-gnu, committed on trunk 2015-10-16 Eric Botcazou * a-tags.ads (Parent_Size): Remove obsolete pragma Export. * s-finmas.ads (Header_Offset): Delete. * s-finmas.adb (Heade

[HSA] HSA back-end improvement

Hello. Attached patch set applies a bunch of small changes to HSA back-end. Patches have been installed to hsa branch. Martin >From 10cf42ce8c0199471271edea80bb0cd717b6f0d1 Mon Sep 17 00:00:00 2001 From: marxin Date: Fri, 9 Oct 2015 14:36:31 +0200 Subject: [PATCH 1/8] HSA: fix types in switch to

[PATCH][AArch64] Add support for 64-bit vector-mode ldp/stp

Hi all, We already support load/store-pair operations on the D-registers when they contain an FP value, but the peepholes/sched-fusion machinery that do all the hard work currently ignore 64-bit vector modes. This patch adds support for fusing loads/stores of 64-bit vector operands into ldp an

Re: [PATCH] i386: Use the STC bb-reorder algorithm at -Os (PR67864)

On 10/16/2015 02:53 PM, Segher Boessenkool wrote: For x86, STC still gives better results for optimise-for-size than "simple" does. So use STC at -Os as well. For how many targets is this true, and for the others, what is the biggest win from "simple"? If the list of targets which get patches

[PATCH] mn10300: Use the STC bb-reorder algorithm at -Os

For mn10300, STC still gives better results for optimise-for-size than "simple" does. So use STC at -Os as well. Is this okay for trunk? Segher 2015-10-16 Segher Boessenkool * common/config/mn10300/mn10300-common.c (mn10300_option_optimization_table) : Use REORDER_

[PATCH] i386: Use the STC bb-reorder algorithm at -Os (PR67864)

For x86, STC still gives better results for optimise-for-size than "simple" does. So use STC at -Os as well. Is this okay for trunk? Segher 2015-10-16 Segher Boessenkool PR rtl-optimization/67864 * common/config/i386/i386-common.c (ix86_option_optimization_table) :

[Ada] Cleanups in inter-unit inlining engine

This removes a component in the record attached to every subprogram considered for inter-unit inlining, which doesn't serve any useful purpose. In addition, this fixes a small inconsistency in the code driving inter-unit inlining from the front-end. The code was at the same time discarding the su

Re: [PATCH] 2015-10-15 Benedikt Huber Philipp Tomsich

On Thu, 2015-10-15 at 22:03 +, Benedikt Huber wrote: > > +/* Add builtins for reciprocal square root. */ > + > +void > +aarch64_init_builtin_rsqrt (void) > +{ > + tree fndecl = NULL; > + tree ftype = NULL; > + > + tree V2SF_type_node = build_vector_type (float_type_node, 2); > + tree V2D

[Ada] Optization of predicate checks

This patch marks the generated predicate functions as Pure, so that the back-end can optimize redundant calls to these functions when inlining and high level of optimization are requested. This is a performance enhancement, no change in behavior. Tested on x86_64-pc-linux-gnu, committed on trunk

[Ada] Crash on illegal program with -gnatf.

This patch fixes a crash in the compiler when reporting an error on an illegal prefixed call whose prefix is overloaded, one of its interpretations has an untagged type, and All_Errors_Mode is set. No short example available. Tested on x86_64-pc-linux-gnu, committed on trunk 2015-10-16 Ed Schon

[Ada] Premature finalization leads to wrong short circuit result

This patch modifies the expansion of expression_with_actions nodes to force the evaluation of the expression when its type is Boolean. This prevents "leaks" of dependencies on transient controlled objects which lead to incorrect results in short circuit operators. -- Source -- ---

Re: [PATCH 1/2] s/390: Implement "target" attribute.

On Fri, Sep 25, 2015 at 02:59:41PM +0100, Dominik Vogt wrote: > The following set of two patches implements the function > __attribute__ ((target("..."))) and the corresponding #pragma GCC > target("...") on S/390. It comes with certain limitations: > > * It is not possible to change any options

[PATCH] Fix PR67975, teach SCCVN basic control equivalency for PHI value-numbering

The following patch teaches SCCVN to value-number two PHI nodes the same even when they are in a different basic-block. To do that we have to prove equivalency of the edge predicates into the PHI (and of course equivalence of the PHI arguments). The patch handles the simple case of PHI nodes wit

Re: [PATCH] c/67882 - improve -Warray-bounds for invalid offsetof

On 10/09/2015 04:55 AM, Martin Sebor wrote: Gcc attempts to diagnose invalid offsetof expressions whose member designator is an array element with an out-of-bounds index. The logic in the function that does this detection is incomplete, leading to false negatives. Since the result of the expressi

[Ada] Improper initialization of elementary parameters in entry calls

The compiler was incorrectly default initializing entry call parameters in cases where the parameter is of an access type or of a scalar type whose Default_Value aspect is set. The access type case caused a problem for CodePeer in the case where the formal parameter is a null-excluding access type,

[gomp4,committed] Handle device-resident and link map kinds in dump_omp_clause

Hi, this patch fixes an ICE when compiling c-c++-common/goacc/declare-1.c with -fdump-tree-omplower. Committed to gomp-4_0-branch. Thanks, - Tom Handle device-resident and link map kinds in dump_omp_clause 2015-10-16 Tom de Vries * tree-pretty-print.c (dump_omp_clause): Handle device-re

Re: [gomp4.1] depend nowait support for target {update,{enter,exit} data}

Hi, On Thu, Oct 15, 2015 at 04:01:56PM +0200, Jakub Jelinek wrote: > Hi! > > CCing various people, because I'd like to have something that won't work on > XeonPhi only. thanks. However, I have not paid too much attention to OMP tasks yet. Nevertheless, let me try to answer some of the question

Re: [PATCH 7/7] Libsanitizer merge from upstream r249633.

On 14/10/15 15:12, Jakub Jelinek wrote: On Wed, Oct 14, 2015 at 03:02:22PM +0300, Maxim Ostapenko wrote: On 14/10/15 14:06, Jakub Jelinek wrote: On Wed, Oct 14, 2015 at 01:51:44PM +0300, Maxim Ostapenko wrote: Ok, got it. The first solution would require changes in libsanitizer because heurist

[PATCH][Testsuite] Turn on 64-bit-vector tests for AArch64

This enables tests bb-slp-11.c and bb-slp-26.c for AArch64. Both of these are currently passing on little- and big-endian. (Tested on aarch64-none-linux-gnu and aarch64_be-none-elf). OK for trunk? gcc/testsuite/ChangeLog: * lib/target-supports.exp (check_effective_target_vect64): Add AA

Re: [PATCH] Random shuffle moveable: container size

Committed to trunk, thanks for the patch.

[Patch AArch64 63304] Fix issue with global state.

Hi, Jiong pointed out privately that there was a thinko in the way in which the global state was being set and reset. I don't like adding such global state but Tested on aarch64-none-elf with no regressions Bootstrapped and regression tested on aarch64-none-linux-gnu Ok to apply ? regards

Re: Do not describe -std=c11 etc. as experimental in c.opt help text

Ping^2. On Fri, Oct 09, 2015 at 03:50:02PM +0200, Marek Polacek wrote: > Jason: ping. > > On Fri, Oct 02, 2015 at 05:35:39PM +0200, Marek Polacek wrote: > > On Thu, Oct 01, 2015 at 05:01:26PM +, Joseph Myers wrote: > > > I noticed that c.opt still described -std=c11 and related options as > >

Re: Move some bit and binary optimizations in simplify and match

+(match (logical_inverted_value @0) + (truth_not @0)) That's good. +/* Simplify ~X & X as zero. */ +(simplify + (bit_and:c (convert? truth_valued_p@0) (convert? (logical_inverted_value @0))) + { build_zero_cst (type); }) That's not what Richard meant. We already have: /* X & !X -> 0. */ (s

[Ada] Minimize the save/restore of Ghost_Mode

This patch minimizes the stack-like handling of global variable Ghost_Mode when processing Ghost code. The patch addresses references to Ghost entities within the expanded code for pragma Contract_Cases. Tested on x86_64-pc-linux-gnu, committed on trunk 2015-10-16 Hristian Kirtchev *

Re: refactoring TARGET_PTRMEMFUNC_VBIT_LOCATION checks

On 10/16/2015 12:48 PM, Christian Bruel wrote: I'm not sure. at each point of the macro, we have the current alignment == FUNCTION_BOUNDARY, because we are just returning from the sequence build_lang_decl/make_node so it looks like DECL_ALIGN (fn) = MAX (MINIMUM_METHOD_BOUNDARY, DECL_ALIGN (

Re: [patch] Minor adjustment to gimplify_addr_expr

> Sure, if that works it's pre-approved. Your original patch is also ok > (though I still > think it's incomplete - but we'll wait until a testcase comes up with > the assert). It passed a bootstrap/regtest cycle on x86-64/Linux so I have installed it. 2015-10-16 Eric Botcazou * tre

Re: Drop CONSTRUCTOR comparsion from ipa-icf-gimple

> On Fri, Oct 16, 2015 at 1:46 AM, Richard Biener > wrote: > > On Fri, Oct 16, 2015 at 5:12 AM, Jan Hubicka wrote: > >> Hi, > >> as Richard noticed in my port of the code to operand_equal_p, the checking > >> of > >> CONSTURCTOR in ipa-icf-gimple is incomplete missing the index checks. > >> It i

Re: [PR67383][ARM][4.9]Backport of "Allow any register for DImode values in Thumb2"

On Thu, Oct 15, 2015 at 03:01:24PM +0100, Renlin Li wrote: > Hi all, > > This is a backport patch to loosen restrictions on core registers > for DImode values in Thumb2. > > It fixes PR67383. In this particular case, reload tries to spill a > hard register, and use next register together as a pai

Re: Drop CONSTRUCTOR comparsion from ipa-icf-gimple

On Fri, Oct 16, 2015 at 1:46 AM, Richard Biener wrote: > On Fri, Oct 16, 2015 at 5:12 AM, Jan Hubicka wrote: >> Hi, >> as Richard noticed in my port of the code to operand_equal_p, the checking of >> CONSTURCTOR in ipa-icf-gimple is incomplete missing the index checks. >> It is also unnecesary si

update in Ada maintainers

Robert Dewar is no longer with us unfortunately (http://www.adacore.com/press/adacore-president-robert-dewar-1945-2015/), and Geert is no longer working for AdaCore, nor working on GNAT. I've also updated my email. Committed on trunk. * MAINTAINERS: Update list of Ada maintainers and ema

  1   2   >