[PATCH] PR50325 store_bit_field: Fix for big endian targets

Hi, starting with r177691 store_bit_field is used for source operands which cannot be covered exactly with word mode chunks. So far store_bit_field does not seem to handle that correctly. With the attached patch extract_bit_field is used for the remaining bits of the source on big endian targets

[PATCH] Remove myself from sparc port MAINTAINERS

Hi! I haven't done any SPARC related work in the last 5 years and did a very bad job as SPARC maintainer, I don't even have any bootable SPARC boxes anymore, fortunately we have lots of other SPARC maintainers, Eric does a good job at it and DaveM works on new SPARC stuff. I should have done this

[google] Backport r178995 from trunk to google/gcc-4_6 (issue5075046)

Hi This patch fixed PR 49452, thus decreased 30+ dejaGNU failures for target arm. Tested with buildit and arm regression testing on qemu. OK for google/gcc-4_6? thanks Carrot ChangeLog 2011-09-21 Guozhi Wei Backport r178995 from trunk. 2011-09-20 Wei Guozhi

Re: Define FLAGS_TO_PASS for libquadmath

On 09/20/2011 09:51 PM, Joseph S. Myers wrote: Some target libraries have # Subdir rules rely on $(FLAGS_TO_PASS) FLAGS_TO_PASS = $(AM_MAKEFLAGS) in their Makefile.am. This was introduced to libstdc++-v3 by to fix a problem described in

Trivial spelling fixes ("informations")

Hi, The Lintian tool used in Debian includes a number of checks for spelling mistakes. It currently complains about "informations" which appears in graphite output as well as code comments. The attached patches fix the code (output) and the comments, separately. I believe they're trivial enough n

Re: [C++ Patch] PR 50454 [Take 2]

OK. Jason

Re: [SH] AND/OR/XOR costs calculation

Oleg Endo wrote: > the attached patch corrects the costs calculation for AND / OR / XOR > operations that could be done with an 8 bit immediate value and R0. > The change doesn't seem to have an impact on the generated code but will > be (probably) required to fix PR 49263. A few minor nits: >

Re: [Mingw-users] [patch] --enable-dynamic-string default for mingw-w64

On 9/20/2011 12:29 PM, Charles Wilson wrote: On 9/20/2011 10:12 AM, Kai Tietz wrote: 2011/9/20 Charles Wilson: So, this would be a change in current mingw.org behavior. I *was* under the impression that this workaround for the old "can't pass empty strings across DLL boundary" problem[1] was n

[Patch][Cilkplus] Fix inling in function calls that spawn

Hello Everyone,    This patch is for the Cilkplus branch. In C and C++ compilers, at optimization -O and -O2, when there are inlined functions inside a function that calls a spawn it was giving an error. This patch should fix that. Thanks, Balaji V. Iyer. diff --git a/gcc/ChangeLog.

Re: [C++ Patch] PR 50454 [Take 2]

Hi, On 09/21/2011 03:25 AM, Jason Merrill wrote: On 09/20/2011 08:38 PM, Paolo Carlini wrote: + if (int128_integer_type_node == NULL_TREE) +error ("%<__int128%> is not supported by this target"); In this case we should unset explicit_int128. Also remove this code: else if (

Re: [PATCH] C++11, implement delegating constructors

Looks good, thanks. Jason

Re: [C++ Patch] PR 50454 [Take 2]

On 09/20/2011 08:38 PM, Paolo Carlini wrote: + if (int128_integer_type_node == NULL_TREE) + error ("%<__int128%> is not supported by this target"); In this case we should unset explicit_int128. Also remove this code: else if (explicit_int128 && TREE_CODE (type) != INTEGER_T

Re: [PATCH 3/7] Emit macro expansion related diagnostics

On 09/20/2011 10:05 AM, Dodji Seketeli wrote: Jason Merrill writes: In linemap_macro_map_loc_to_def_point you get the token number and then use that to index into MACRO_MAP_LOCATIONS. Can't you use the same token number to index into macro->exp.tokens instead? No, because a macro argument

[SH] AND/OR/XOR costs calculation

Hello, the attached patch corrects the costs calculation for AND / OR / XOR operations that could be done with an 8 bit immediate value and R0. The change doesn't seem to have an impact on the generated code but will be (probably) required to fix PR 49263. Tested with make -k check RUNTESTFLAGS="

Go patch committed: Fix channels with element types of size zero

This trivial patch to the Go library fixes using channels with element types of size zero. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian diff -r f87ff36963ee libgo/runtime/go-new-channel.c --- a/libgo/runtime/go-new-channel.c Tue Sep 20 17:10:50 2011

[C++ Patch] PR 50454 [Take 2]

Hi, per the preceding discussion. Tested x86_64-linux. Ok for mainline? Thanks, Paolo. /// /cp 2011-09-21 Paolo Carlini PR c++/50454 * decl.c (grokdeclarator): Consistently handle both __int128 and unsigned __int128 with -pedantic; suppress diagn

Re: Go patch committed: Better error for unknown name in package

Eric Botcazou writes: >> This patch improves the error message that the Go frontend issues when >> an unknown name in a package is used as type. It also fixes some >> followon errors. > > s/identifer/identifier/ in the error message. Fixed with the obvious patch. Committed to mainline. Thanks

Re: [PATCH] C++11, implement delegating constructors

On 21 September 2011 02:18, Ville Voutilainen wrote: > Ok, a new patch: I just noticed that I have tabs disabled. You may want to tabify the patch.

Re: [PATCH] C++11, implement delegating constructors

At Tue, 20 Sep 2011 19:05:32 -0400, Jason Merrill wrote: > We don't need a new variable, just use basetype. TYPE_MAIN_VARIANT > shouldn't affect the comparison to current_class_type. > > > + error ("mem-initializer for %qD follows constructor delegation; " > > +"previous target c

Re: Ping^3: PR 50113/50061: Fix ABI breakage from emit_library_call_value_1 patch

On 09/16/11 17:14, Richard Sandiford wrote: > Ping for this patch to emit_library_call_value_1: > > http://gcc.gnu.org/ml/gcc-patches/2011-08/msg00735.html > > which fixes a bootstrap failure on MIPS since: > > http://gcc.gnu.org/ml/gcc-patches/2011-06/msg02341.html > > Tested on mips64

Re: [PATCH] C++11, implement delegating constructors

On 09/20/2011 06:17 PM, Ville Voutilainen wrote: I find it hard to see what exactly changes there. Even if I restore the indentation level, git still shows a diff. Perhaps there were tabs in it... Yes, the change is probably tabs vs. spaces. else if (TYPE_P (name)) { basetyp

Re: [Fortran,committed] Remove redundant word.

On Tue, Sep 20, 2011 at 3:55 PM, H.J. Lu wrote: > On Tue, Sep 20, 2011 at 9:29 AM, Steve Kargl > wrote: >> I've committed the following patch under the obviously >> correct rule. >> >> 2011-09-20  Steven G. Kargl   >> >>        * check.c (gfc_check_c_sizeof): Remove redundant word. >> >> > > You

Re: [Fortran,committed] Remove redundant word.

On Tue, Sep 20, 2011 at 9:29 AM, Steve Kargl wrote: > I've committed the following patch under the obviously > correct rule. > > 2011-09-20  Steven G. Kargl   > >        * check.c (gfc_check_c_sizeof): Remove redundant word. > > You may need to update gfortran.dg/storage_size_2.f08: http://gcc.

Re: Go patch committed: Better error for unknown name in package

Eric Botcazou writes: >> This patch improves the error message that the Go frontend issues when >> an unknown name in a package is used as type. It also fixes some >> followon errors. > > s/identifer/identifier/ in the error message. Argh, thanks. Will fix. Ian

Re: Go patch committed: Better error for unknown name in package

> This patch improves the error message that the Go frontend issues when > an unknown name in a package is used as type. It also fixes some > followon errors. s/identifer/identifier/ in the error message. -- Eric Botcazou

Go patch committed: Better error for unknown name in package

This patch improves the error message that the Go frontend issues when an unknown name in a package is used as type. It also fixes some followon errors. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian diff -r 4e6476a9ca36 go/expressions.cc --- a/go/exp

Re: [PATCH] C++11, implement delegating constructors

At Tue, 20 Sep 2011 17:05:55 -0400, Jason Merrill wrote: > Try to avoid reformatting lines that you don't actually change. I find it hard to see what exactly changes there. Even if I restore the indentation level, git still shows a diff. Perhaps there were tabs in it... > > + if (init == void_ty

Go patch committed: Fix make of slice of zero-size type

This trivial patch to the Go library fixes calling make on a slice of elements whose size is zero. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian diff -r ce7423ad4e50 libgo/runtime/go-make-slice.c --- a/libgo/runtime/go-make-slice.c Tue Sep 20 13:57:32

Re: [PATCH] Fix AVR ICE

On 09/20/11 19:09, Tom de Vries wrote: > > * final.c (final): Handle if JUMP_LABEL is not LABEL_P. This is ok. Bernd

Re: [PATCH] C++11, implement delegating constructors

On 09/20/2011 01:29 PM, Ville Voutilainen wrote: 2011-09-20 Ville Voutilainen Implement delegating constructors. Based on an original patch by Pedro Lamarao. Looks good, just a few minor comments: - build2 (EQ_EXPR, boolean_type_node, -

Go patch committed: Implement goto restrictions

The Go language has clarified the use of goto. It is now invalid to use goto to jump into an inner block, and to use goto to jump across a variable declaration. This patch to the Go frontend implements these restrictions, and adjust one library and a few test cases accordingly. Bootstrapped and r

Re: Ping: C-family stack check for threads

Please post your pings under a more meaningful subject line that indicates that this is an ARM back-end and middle-end patch. It's not a C-family patch, C-family maintainers can't give it any useful review. -- Joseph S. Myers jos...@codesourcery.com

Re: [C++-11] User defined literals

On 09/19/2011 10:07 PM, Ed Smith-Rowland wrote: On 09/19/2011 05:11 PM, Jason Merrill wrote: Can't you store the string and the suffix, and then interpret the number later if you end up calling an operator that takes the value? I could and I thought abut it but from the wording (2.14.8 p3 and

Define FLAGS_TO_PASS for libquadmath

Some target libraries have # Subdir rules rely on $(FLAGS_TO_PASS) FLAGS_TO_PASS = $(AM_MAKEFLAGS) in their Makefile.am. This was introduced to libstdc++-v3 by to fix a problem described in

Re: inline-analysis improvement

> the problem here is that we consider ¶m.foo to be a memory reference, > while it is not. The following patch should fix it, but because it changes > the behaviour of inline heuristics, I will run bechmarks tonight before > committing it. Thanks in advance. Can you add PR tree-optimization/50433

C++ PATCH to when constexpr violations are diagnosed

For a normal function, we want a constexpr violation to be diagnosed immediately so that the programmer can fix it. But for a template, we don't want to complain about a particular instantiation if other instantiations could be usable as constexpr functions. I've gone through a couple of diff

C++ PATCH to avoid double copying of default arguments

While working on something else I noticed that we were copying the trees in a default argument twice: once in break_out_target_exprs and then again with an explicit unshare_expr. So I removed the redundant unshare_expr. This didn't cause any regressions by itself, but just to be safe I also c

Re: [C++ Patch] PR 50454

Hi, > On 09/20/2011 10:19 AM, Paolo Carlini wrote: >> + if (explicit_int128 && pedantic && ! in_system_header) >> +pedwarn (input_location, OPT_pedantic, >> + "ISO C++ does not support %<__int128%> for %qs", name); > > Yep, like that. But we also want the check for null integer128_t

Re: [C++ Patch] PR 50454

On 09/20/2011 10:19 AM, Paolo Carlini wrote: + if (explicit_int128 && pedantic && ! in_system_header) +pedwarn (input_location, OPT_pedantic, +"ISO C++ does not support %<__int128%> for %qs", name); Yep, like that. But we also want the check for null integer128_type_node. Jas

Re: devirtualize final virtual function calls

On 09/20/2011 12:43 PM, Roberto Agostino Vitillo wrote: - if (DECL_VINDEX (fn)&& ! (flags& LOOKUP_NONVIRTUAL) - && resolves_to_fixed_type_p (instance, 0)) + if (DECL_VINDEX (fn)&& ((! (flags& LOOKUP_NONVIRTUAL) + && resolves_to_fixed_type_

RE: [Patch,AVR]: Fix PR50449

> -Original Message- > From: Georg-Johann Lay [mailto:a...@gjlay.de] > Sent: Tuesday, September 20, 2011 12:28 PM > To: gcc-patches@gcc.gnu.org > Cc: Weddington, Eric; Anatoly Sokolov; Denis Chertykov > Subject: [Patch,AVR]: Fix PR50449 > > This is fix for a minor performance regression

[Patch,AVR]: Fix PR50449

This is fix for a minor performance regression introduced by my changes in trunk r175956: To load a 32-bit constant like 1 into R2, 4.6 uses CLR R2 CLR R3 MOVW R4,R2 INC R2 whereas trunk prints the longer CLR R2 INC R2 CLR R3 CLR R4 CLR R5 Th

[RFA] Add libiberty/argv.c:countargv

Hi. Part of the abstraction of an argv is a count of the number elements. This patch adds countargv which I have use for in gdb. Ok to check in? 2011-09-20 Doug Evans include/ * libiberty.h (countargv): Declare. libiberty/ * argv.c (countargv): New function.

Re: [PATCH] C++11, implement delegating constructors

At Tue, 20 Sep 2011 20:29:27 +0300, Ville Voutilainen wrote: > --- a/gcc/testsuite/g++.dg/template/meminit1.C > +++ b/gcc/testsuite/g++.dg/template/meminit1.C > @@ -2,7 +2,7 @@ > template > struct S > { > - S() : S() {} // { dg-error "base" } > + S() : S() {} > }; > > -S s; // { dg-message

[PATCH] alignaddr and edge VIS intrinsic refinements

We provided the big-endian 'alignaddr' but not the little-endian version 'alignaddrl', add it. The edge operations are meant to operate on addresses and return a very small integer bitmask so use Pmode and return SImode. Update documentation and intrinsic header, as needed. Committed to trunk.

[trans-mem] Post-merge regressions, part 1

Handle the "new" MEM_REF in a couple of places that were missed. The memopt-10.c test for some reason expected the restoration via logging functions. But we're saving a simple "int", so the fact that the compiler chooses a local temp variable seems a completely reasonable choice. The memopt-14.c

[PATCH] C++11, implement delegating constructors

Tested on Linux/X86-32. 2011-09-20 Ville Voutilainen Implement delegating constructors. Based on an original patch by Pedro Lamarao. * cp-tree.h (enum cpp0x_warn_str): Add CPP0X_DELEGATING_CTORS. * error.c (maybe_warn_cpp0x): Adjust. * parser

Re: inline-analysis improvement

Hi Honza, On 15 Sep 2011, at 11:27, Jan Hubicka wrote: Bootstrapped/regtested x86_64-linux, will commit it later today. Honza * ipa-inline-analysis.c (add_condition): Add conditions parameter; simplify obviously true clauses. (and_predicates, or_predicates): Add conditi

[PATCH] Fix AVR ICE

Hi Bernd, this patch fixes an ICE in final.c:final() when building for AVR with --enable-checking=yes,rtl. The ICE happens when the jump label is a return, but expected to be a label. Tested on X86_64 and AVR (courtesy of Johann). ok for trunk? Thanks, - Tom 2011-09-20 Tom de Vries

Go patch committed: Check for duplicate parameter/result names

This patch to the Go frontend checks for duplicate parameter/result names in function declarations and interface method signatures. Previously I only checked for duplicates in function definitions. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian diff -r 3

devirtualize final virtual function calls

When calling a final virtual function or a virtual function that belongs to a final class there is no need for a virtual lookup. E.g.: struct A final { virtual void foo () { } }; struct B { virtual void foo () final { } }; void fun(A* a, B* b) { a->foo(); b->foo(); } with this

[Fortran,committed] Remove redundant word.

I've committed the following patch under the obviously correct rule. 2011-09-20 Steven G. Kargl * check.c (gfc_check_c_sizeof): Remove redundant word. Index: check.c === --- check.c (revision 179008) +++ check.c

Ping: C-family stack check for threads

ping references http://gcc.gnu.org/ml/gcc-patches/2011-09/msg00310.html http://gcc.gnu.org/ml/gcc-patches/2011-08/msg00216.html http://gcc.gnu.org/ml/gcc-patches/2011-07/msg00281.html http://gcc.gnu.org/ml/gcc-patches/2011-07/msg00149.html http://gcc.gnu.org/ml/gcc-patches/2011-06/msg01872.html h

Re: [PATCH] derive alias information from named address spaces.

On 09/20/2011 01:46 AM, Bingfeng Mei wrote: >if (MEM_P (rtlx) && MEM_P (rtly) >&& MEM_ADDR_SPACE (rtlx) != MEM_ADDR_SPACE (rtly)) > -return 0; > +{ > + if (!targetm.addr_space.subset_p (MEM_ADDR_SPACE (rtlx), > +MEM_ADDR_SPACE (rtly))

Go patch committed: Emit errors for unused values

The other Go compiler issues errors for unused values, including unused calls to builtin functions that return no value. gccgo already had a warning (under -Wunused-value) but not an error. This patch makes gccgo compatible by issuing an error, and also issues errors for builtin functions. Boots

Re: [patch] --enable-dynamic-string default for mingw-w64

On 9/20/2011 10:12 AM, Kai Tietz wrote: > 2011/9/20 Charles Wilson : http://cygwin.com/acronyms/index.html#PCYMTNQREAIYR >> So, this would be a change in current mingw.org behavior. I *was* >> under the impression that this workaround for the old "can't pass >> empty strings across DLL boundary"

Re: [patch] --enable-dynamic-string default for mingw-w64

On Tuesday 20 September 2011 15:12:30, Kai Tietz wrote: > Yes. If you read last comment of the thread you are citing, then you > would notice that for static-libstdc++ version the issue is still > present. So to allow users to use also static-libstdc++ variant, this > option is still necessary.

Re: [patch] --enable-dynamic-string default for mingw-w64

On 9/20/2011 22:12, Kai Tietz wrote: >> I'm not really opposed to making this change for i*86-pc-mingw -- and >> now's the time to do it, as the recently released 4.6.1 mingw.org gcc >> broke the C++ abi anyway, thanks to thiscall. > > Here I am a bit curious? How is 4.6.1 affected by new thiscal

Re: [C++ Patch] PR 50454

On 09/20/2011 03:45 PM, Paolo Carlini wrote: On 09/20/2011 03:12 PM, Jason Merrill wrote: No, I think we want to complain about __int128 when -pedantic unless we're in a system header, as C does. The right fix is to move the __int128 code out of the if (unsigned_p || ...) block, and to suppre

Re: [patch] --enable-dynamic-string default for mingw-w64

2011/9/20 Charles Wilson : > On 9/20/2011 9:20 AM, JonY wrote: >> On 9/20/2011 13:59, Kai Tietz wrote: >>> 2011/9/20 JonY: Its been used in the automated toolchain builds for sometime, seems like a good idea to enable it by default. It can be easily changed to match for all mingw as

Re: v{extract,insert,broadcast,perm2}{i,f}128

On Tue, Sep 20, 2011 at 4:07 PM, Uros Bizjak wrote: >>> No, you define mode attribute like >>> >>> (define_mode_attr somesuffix [(V2DI "%i") (V2DF "f") ...]) >>> >>> where integer modes map to %i and FP modes to "f". >>> >>> Then, just return "vextract128\t{$0x1, %1, %0|%0, %1, 0x1}" >> >> I can

Re: [patch] --enable-dynamic-string default for mingw-w64

于 2011/9/20 21:52, Charles Wilson 写道: On 9/20/2011 9:20 AM, JonY wrote: On 9/20/2011 13:59, Kai Tietz wrote: 2011/9/20 JonY: Its been used in the automated toolchain builds for sometime, seems like a good idea to enable it by default. It can be easily changed to match for all mingw as well if

Re: v{extract,insert,broadcast,perm2}{i,f}128

On Tue, Sep 20, 2011 at 4:00 PM, Jakub Jelinek wrote: > On Tue, Sep 20, 2011 at 03:47:56PM +0200, Uros Bizjak wrote: >> No, you define mode attribute like >> >> (define_mode_attr somesuffix [(V2DI "%i") (V2DF "f") ...]) >> >> where integer modes map to %i and FP modes to "f". >> >> Then, just retu

Re: [PATCH 3/7] Emit macro expansion related diagnostics

Jason Merrill writes: > On 09/20/2011 03:23 AM, Dodji Seketeli wrote: >> Jason Merrill writes: > >>> Certainly all the calls to tokens_buff_add_token pass src->src_loc for >>> the second. So why don't we look up the second location in the macro >>> definition when we need it rather than store a

Re: [PATCH 3/7] Emit macro expansion related diagnostics

On 09/20/2011 03:23 AM, Dodji Seketeli wrote: Jason Merrill writes: Certainly all the calls to tokens_buff_add_token pass src->src_loc for the second. So why don't we look up the second location in the macro definition when we need it rather than store a copy in the map? Because when you h

Re: v{extract,insert,broadcast,perm2}{i,f}128

On Tue, Sep 20, 2011 at 03:47:56PM +0200, Uros Bizjak wrote: > No, you define mode attribute like > > (define_mode_attr somesuffix [(V2DI "%i") (V2DF "f") ...]) > > where integer modes map to %i and FP modes to "f". > > Then, just return "vextract128\t{$0x1, %1, %0|%0, %1, 0x1}" I can surely do

Re: [PATCH 5/7] Add line map statistics to -fmem-report output

OK. Jason

Re: [patch] --enable-dynamic-string default for mingw-w64

On 9/20/2011 9:20 AM, JonY wrote: > On 9/20/2011 13:59, Kai Tietz wrote: >> 2011/9/20 JonY: >>> Its been used in the automated toolchain builds for sometime, >>> seems like a good idea to enable it by default. It can be >>> easily changed to match for all mingw as well if needed. >> >> This patch

Re: v{extract,insert,broadcast,perm2}{i,f}128

On Tue, Sep 20, 2011 at 3:38 PM, Jakub Jelinek wrote: > On Tue, Sep 20, 2011 at 07:15:42AM +0200, Uros Bizjak wrote: >> Perhaps use mode attribute that defines "f" for FP modes and "%i" for >> integer modes. "%i" is further processed in ix86_output_operand  (or >> perhaps ASM_OUTPUT_OPCODE, simila

Re: [C++ Patch] PR 50454

On 09/20/2011 03:12 PM, Jason Merrill wrote: No, I think we want to complain about __int128 when -pedantic unless we're in a system header, as C does. The right fix is to move the __int128 code out of the if (unsigned_p || ...) block, and to suppress the pedwarn when in_system_header. Ok, I'l

Re: v{extract,insert,broadcast,perm2}{i,f}128

On Tue, Sep 20, 2011 at 07:15:42AM +0200, Uros Bizjak wrote: > Perhaps use mode attribute that defines "f" for FP modes and "%i" for > integer modes. "%i" is further processed in ix86_output_operand (or > perhaps ASM_OUTPUT_OPCODE, similar to %v ?) to output "i" for > TARGET_AVX2 or "f" otherwise.

Re: [PATCH] Fix PR48571, remove (most) array-ref re-construction

On Tue, Aug 30, 2011 at 4:59 AM, Richard Guenther wrote: > > I've run into PR48571 again, and after having fun with out data-dep > analysis code I indeed believe that we should not try to re-construct > ARRAY_REFs from code involving pointer arithmetic.  ARRAY_REFs > have constraints imposed on th

Re: [patch] --enable-dynamic-string default for mingw-w64

On 9/20/2011 13:59, Kai Tietz wrote: > 2011/9/20 JonY: >> Hi, >> >> Its been used in the automated toolchain builds for sometime, seems like >> a good idea to enable it by default. It can be easily changed to match >> for all mingw as well if needed. >> >> OK for trunk? >> >> Index: libstdc++-v3/co

Re: [C++ Patch] PR 50454

On 09/20/2011 08:09 AM, Paolo Carlini wrote: as I tried to analyze/explain in the audit trail, apparently we are doing something bogus in grokdeclarator wrt unsigned __int128, which leads to the latter being wrongly rejected with -pedantic-errors, at variance with plain __int128 or __int128_t / _

[C++ Patch] PR 50454

Hi, as I tried to analyze/explain in the audit trail, apparently we are doing something bogus in grokdeclarator wrt unsigned __int128, which leads to the latter being wrongly rejected with -pedantic-errors, at variance with plain __int128 or __int128_t / __uint128_t for that matter. Thus I pr

Re: [VTA, PR49310] O(n+m)-ish emit_notes

On Tue, Sep 20, 2011 at 08:19:02AM -0300, Alexandre Oliva wrote: > > Those need to be emitted uncached > > Would you please remind me why that is? The reason totally escapes me. For NOTE_INSN_VAR_LOCATION, we are caching because we prefer to create shorter .debug_loc location lists over longer o

Re: [VTA, PR49310] O(n+m)-ish emit_notes

On Sep 20, 2011, Jakub Jelinek wrote: > What worries me is that we suddenly emit far less than half of the > previous call_site_value parameters: Thanks for the measurements, I'll look into that. > Those need to be emitted uncached Would you please remind me why that is? The reason totally es

[PATCH, PR43814] Assume function arguments of pointer type are aligned.

Hi Richard, I have a patch for PR43814. It introduces an option that assumes that function arguments of pointer type are aligned, and uses that information in tree-ssa-ccp. This enables the memcpy in pr43814-2.c to be inlined. I tested the patch successfully on-by-default on x86_64 and i686 (both

Re: [PATCH 5/7] Add line map statistics to -fmem-report output

Jason Merrill writes: > It's hard to compare macro_maps_allocated_size with > macro_maps_used_size since you have to do subtraction to discover both > values from what is printed. OK, I have changed the statistics components accordingly. > And do the addition in dump_line_table_statistics rath

Re: [PATCH][ARM] -m{cpu,tune,arch}=native

On 09/09/11 12:55, Richard Earnshaw wrote: The part number field is meaningless outside of the context of a a specific vendor -- only taken as a pair can they refer to a specific part. So why is the vendor field hard-coded rather than factored into the table of parts. Maybe it would be better t

[arm-embedded] Backport mainline r177357

Backport mainline r177357 to ARM/embedded-4_6-branch. Committed. 2011-09-20 Terry Guo Backport r177357 from mainline 2011-08-04 Ian Bolton * gcc.target/arm/vfp-1.c: no large negative offsets on Thumb2.

Re: [VTA, PR49310] O(n+m)-ish emit_notes

Hi! First of all, I'm worried about generated debug info quality. obj81{4,5,6,7}/gcc/cc1plus below are x86_64/i686 unpatched trunk stage3 compilers (with var-tracking.c patched afterwards and make cc1plus in the stage3 gcc directory, so it is the same code, just built with different compiler) and

Re: [Trunk/GCC 4.6] Re: [google] Omit date from Fortran .mod files for reproducible builds

On Fri, Sep 16, 2011 at 21:06, Diego Novillo wrote: > On Fri, Jan 28, 2011 at 13:00, Diego Novillo wrote: >> On Fri, Jan 28, 2011 at 06:19, Tobias Burnus wrote: >>> We (Janne and I) think this patch can also be applied to the GCC 4.6 trunk; >>> as the date is never read there is also no .mod ABI

Re: inline-analysis improvement

> > * ipa-inline-analysis.c (add_condition): Add conditions parameter; > > simplify obviously true clauses. > > (and_predicates, or_predicates): Add conditions parameter. > > (inline_duplication_hoook): Update. > > (mark_modified): New function. > > (unmodified_parm): New fu

[arm-embedded] Backport mainline

Backport mainline r174578 & r174577 to ARM/embedded-4_6-branch. Committed. 2011-09-20 Jiangning Liu Backport r174578 from mainline 2011-06-02 Julian Brown * config/arm/arm-cores.def (arm1156t2-s, arm1156t2f-s): Use v6t2 tuning. * config/arm/arm-prot

[arm-embedded] Backport mainline r174341 r174310

Backport mainline r174341 & r174310 to ARM/embedded-4_6-branch. Committed. 2011-09-20 Bin Cheng Backport r174341 from mainline 2011-05-27 Janis Johnson * g++.dg/tree-ssa-pr43411.C: Rename function to be inlined and replace dump search. 2011-09-20 Bin Chen

Ping: [ARM] Loosen MODES_TIEABLE_P

Ping for: http://gcc.gnu.org/ml/gcc-patches/2011-09/msg00626.html which allows taking the vector subreg of a structure mode to be treated as "cheap" (because direct register references are allowed). At the moment we treat them as expensive. Richard

Ping: PR middle-end/48660: Assigning to BLKmode RESULT_DECL

Ping for: http://gcc.gnu.org/ml/gcc-patches/2011-09/msg00558.html which fixes an expand-time bug in thunks that return BLKmode structures in registers. Tested on x86_64-linux-gnu and arm-linux-gnueabi. Richard

RE: [Patch, testsuite, arm] Skip the arch conflict to enable case neon-thumb2-move pass on more targets.

Hello, > > > > I suppose you want a torture that excercises different -march/-mtune > > combinations then. > > > > But can't you do the pruning somewhere in an .exp file then instead > > of sprinkling it all over the tests itself? > > > > It seems not. At present the multilib options are all add

[arm-embedded] Backport mainline r175246 r175009

Backport mainline r175246 r175009 to ARM/embedded-4_6-branch. Committed. 2011-09-20 Jiangning Liu Backport r175246 from mainline 2011-06-21 Ira Rosen PR testsuite/49443 * gcc.dg/vect/vect-peel-3.c: Expect to fail on vect_no_align targets. *

[arm-embedded] Backport mainline r174549 r178906

Backport mainline r178906, r174549 to ARM/embedded-4_6-branch. Committed. 2011-09-20 Terry Guo Backport r178906 from mainline 2011-09-16 Terry Guo * gcc.dg/tree-ssa/foldconst-3.c: Don't use short enums. 2011-09-20 Terry Guo Backport r174549 from mainli

[PATCH] derive alias information from named address spaces.

Hello, The following patch is to derive more alias information from named address spaces using existing target hook. It was discussed in http://gcc.gnu.org/ml/gcc/2011-09/msg00180.html Bootstrapped and tested on x86-64. OK for trunk? Thanks, Bingfeng 2011-09-20 Bingfeng Mei * alias

[ARM] Fix a performance regression from the fix for PR49030

My fix for PR49030 had the unintended side-effect of forcing libcalls to be used for some DImode comparisons. The problem (which I should have noticed at the time, sorry) is that arm_comparison_operator is used for both expanders and insns. The patch fixed the definition for insns, but made it to

[arm-embedded] Backport mainline r176339

Backport mainline 176339 to ARM/embedded-4_6-branch. Committed. 2011-09-20 Bin Cheng Backport r176339 from mainline 2011-07-15 Yufeng Zhang * config/locale/newlib/ctype_members.cc: New file. * acinclude.m4 (GLIBCXX_ENABLE_CLOCALE): Add a new C locale

[arm-embedded] Backport mainline fixed-point support

Backport following mainline revs to ARM/embedded-4_6-branch: r174187, r176036, r177015, r177017, r177018, r177019, 177022, 177025 Committed. 2011-09-20 Jiangning Liu Backport r177025 from mainline 2011-08-01 Julian Brown * configure.ac (fixed-point): Add ARM suppo

Re: [PATCH 3/7] Emit macro expansion related diagnostics

Jason Merrill writes: > It is sounding to me like the first location (xI) gets you the next > virtual location in the unwinding process, whereas the second location > (yI) gets you the spelling location of the token in the definition of > a macro. Right. > Certainly all the calls to tokens_buff