When I initially implemented the xxperm support in the ISA 3.0 extensions, I
misread the manual. The xxperm (and xxpermr) instruction logically should have
4 arguments, like vperm/vpermr (ouput, 2 inputs that provide the bytes, and the
permute register), but due to encoding issues, it only has 3 V
PING.
On Fri, Jun 24, 2016 at 1:41 PM, Evgeny Stupachenko wrote:
> Hi,
>
> Fix ICE when IPA-cp and target_clones are applied to the same function.
> Is the patch ok for trunk?
>
> Thanks,
> Evgeny
>
> 2016-06-24 Evgeny Stupachenko
>
> gcc/
> * ipa-cp.c (determine_versionability): Do no
Attached patch sets a limit of huge-1 on the character count to avoid the
integer wrap.
Regression tested with adjustment to test case.
OK for trunk?
Regards,
Jerry
2016-07-11 Jerry DeLisle
PR fortran/66310
* simplify.c (gfc_simplify_repeat): Set max repeat to huge - 1 to a
On Mon, Jul 11, 2016 at 4:27 PM, Ian Lance Taylor wrote:
> On Mon, Jul 11, 2016 at 1:05 PM, Brooks Moses wrote:
>> Ping?
>
> This is fine, but please add a comment explaining why the code avoids
> 0-length VLAs. Thanks.
Thanks!
Committed as r238233. This is the comment I added:
+/* Avoid
>
> So something akin to what I initially proposed?
> https://gcc.gnu.org/ml/gcc-patches/2016-07/msg00182.html
I have similar patch to add_expr. I don't think it should have flag_wpa guard:
other FEs are keeping umberged decls, too, just not as often as LTO.
I am not sure about operand_equal_p
On Mon, Jul 11, 2016 at 1:05 PM, Brooks Moses wrote:
> Ping?
This is fine, but please add a comment explaining why the code avoids
0-length VLAs. Thanks.
Ian
> (I suspect I should have added a libiberty maintainer to cc in the first
> place.)
>
> On Mon, Jun 13, 2016 at 9:05 AM, Brooks Moses
On 11 July 2016 at 20:57, Marek Polacek wrote:
>
> 2016-07-11 Marek Polacek
>
> PR c/7652
> * libsupc++/hash_bytes.cc: Use __builtin_fallthrough.
>
> diff --git gcc/libstdc++-v3/libsupc++/hash_bytes.cc
> gcc/libstdc++-v3/libsupc++/hash_bytes.cc
> index 2e5bbfa..818331f 100644
>
> "Marek" == Marek Polacek writes:
Marek> +++ gcc/gcc/java/jcf-dump.c
Marek> @@ -926,6 +926,8 @@ print_constant (FILE *out, JCF *jcf, int index, int
verbosity)
Marek>if (verbosity > 0)
Marek> fprintf (out, "Fieldref: %ld=", (long) JPOOL_USHORT2 (jcf,
index));
Marek>
Sigh, I keep forgetting to attach the patch.
2016-07-11 Michael Meissner
* doc/install.texi (Configuration): Document PowerPC specific
configuration options --enable-lra and --enable-float128.
* configure.ac: Add --enable-lra and --enable-float128 to turn on
-ml
Hi!
As the following testcases show, if device clause argument doesn't have int
type (or some compatible one), then we ICE, because we don't emit valid
GIMPLE.
Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux,
committed to trunk and 6.2.
2016-07-11 Jakub Jelinek
PR
On Mon, Jul 11, 2016 at 04:43:57PM -0400, David Malcolm wrote:
> On Mon, 2016-07-11 at 22:34 +0200, Jakub Jelinek wrote:
> > On Mon, Jul 11, 2016 at 10:23:30PM +0200, Marek Polacek wrote:
> > > On Mon, Jul 11, 2016 at 01:18:02PM -0700, Andi Kleen wrote:
> > > > > I explained why supporting the clas
On Mon, 2016-07-11 at 22:34 +0200, Jakub Jelinek wrote:
> On Mon, Jul 11, 2016 at 10:23:30PM +0200, Marek Polacek wrote:
> > On Mon, Jul 11, 2016 at 01:18:02PM -0700, Andi Kleen wrote:
> > > > I explained why supporting the classic lint style comment
> > > > wouldn't fly.
> > >
> > > Not convincin
Tested on Linux-X64.
The test adjustments are so that the tests are kept valid, which
required adding a bunch of now-required relops to the test types. The new
transparent-but-non-synthesizing aspects of the relops are tested
separately. The constraints are a valid implementation of the
current Re
On Mon, 2016-07-11 at 22:20 +0200, Marek Polacek wrote:
> On Mon, Jul 11, 2016 at 10:11:52PM +0200, Jakub Jelinek wrote:
> > On Mon, Jul 11, 2016 at 10:08:02PM +0200, Eric Botcazou wrote:
> > > > After I'd completed the warning, I kicked off a bootstrap so as
> > > > to add
> > > > various gcc_fall
I'm curious about this. In the process of developing a code analysis
tool, I found some GCC code that was, basically:
switch (v) {
case 0:
if (e) {
do_something();
} else {
case 1:
do_something_else();
}
}
Does this patch mean that the above got fixed? I mean, if you're
going to f
On Mon, Jul 11, 2016 at 10:23:30PM +0200, Marek Polacek wrote:
> On Mon, Jul 11, 2016 at 01:18:02PM -0700, Andi Kleen wrote:
> > > I explained why supporting the classic lint style comment wouldn't fly.
> >
> > Not convincing, it worked fine for 30+ years of lints.
>
> So how can the compiler han
> I think that if *we* refuse to use __builtin_fallthrough, then we can't
> expect users to use it.
No disagreement, but this doesn't change my mind on the thing... IMO there
ought to be a simple & standard way of silencing the warning if it is enabled
by default, otherwise this will very likel
On Mon, Jul 11, 2016 at 3:43 PM, Marek Polacek wrote:
> But then the [[fallthrough]] attribute was
> approved for C++17 [1], and that's what has got me to do all this.
> ...
> I added a new builtin,
> __builtin_fallthrough, that prevents the warning from occurring. It can only
> be used in a swi
On Mon, Jul 11, 2016 at 01:18:02PM -0700, Andi Kleen wrote:
> > I explained why supporting the classic lint style comment wouldn't fly.
>
> Not convincing, it worked fine for 30+ years of lints.
So how can the compiler handle
/* Never ever fall through here */
?
Marek
On Mon, Jul 11, 2016 at 10:11:52PM +0200, Jakub Jelinek wrote:
> On Mon, Jul 11, 2016 at 10:08:02PM +0200, Eric Botcazou wrote:
> > > After I'd completed the warning, I kicked off a bootstrap so as to add
> > > various gcc_fallthrough calls. There were a good amount of them, as
> > > expected. I
> I explained why supporting the classic lint style comment wouldn't fly.
Not convincing, it worked fine for 30+ years of lints.
-Andi
--
a...@linux.intel.com -- Speaking for myself only.
On Thu, Jul 07, 2016 at 03:06:55PM -0400, Jason Merrill wrote:
> On Thu, Jul 7, 2016 at 2:23 PM, Jakub Jelinek wrote:
> > On Thu, Jul 07, 2016 at 12:32:02PM -0400, Jason Merrill wrote:
> >> Hmm, I wonder if walk_tree_1 should walk into DECL_EXPR like it does into
> >> BIND_EXPR_VARS. But your pat
Hi!
add_conv_candidate creates cand->fn which is not a FUNCTION_DECL, but
some type, all spots in convert_like_real assume that if fn is non-NULL, it
is some decl. Just a couple of lines above this hunk we even have a comment
reminding us on cand->fn not having to be a decl:
/* Since cand->
On Mon, Jul 11, 2016 at 10:08:02PM +0200, Eric Botcazou wrote:
> > After I'd completed the warning, I kicked off a bootstrap so as to add
> > various gcc_fallthrough calls. There were a good amount of them, as
> > expected. I also grepped various FALLTHRU/Falls through/...fall
> > thru.../... com
Hi!
REALPART_EXPR and IMAGPART_EXPR are handled like unary expressions, even
though they are references. For !lval that makes no difference, but for
lval it means we can get ADDR_EXPR of INTEGER_CST etc., or trying to store
into an INTEGER_CST.
Fixed by doing roughly what we do for other referen
> After I'd completed the warning, I kicked off a bootstrap so as to add
> various gcc_fallthrough calls. There were a good amount of them, as
> expected. I also grepped various FALLTHRU/Falls through/...fall
> thru.../... comments in config/ and added gcc_fallthroughs to make it
> easier for peo
On July 11, 2016 9:28:43 PM GMT+02:00, Jakub Jelinek wrote:
>Hi!
>
>vect_get_vec_defs handles only one or two arguments, for ternary
>ops like fma vectorizable_operation calls vect_get_vec_defs on the
>first two arguments, and then did
> vec_oprnds2.create (1);
> vec_oprnds2.q
These configuration switches will allow the PowerPC GCC developers to switch
defaults in the compiler to debug the code, before making the decision to flip
the default permanently. In the future, when the defaults have been changed,
these configuration options would allow developers to go back to
Ping?
(I suspect I should have added a libiberty maintainer to cc in the first place.)
On Mon, Jun 13, 2016 at 9:05 AM, Brooks Moses wrote:
> Zero-length variable-length-arrays are not allowed in standard C99,
> and perhaps more importantly, they cause ASAN to complain. (See,
> e.g., https://gc
On Mon, Jul 11, 2016 at 01:01:05PM -0700, Andi Kleen wrote:
> Marek Polacek writes:
> >
> > This warning is enabled by default for C/C++.
>
> That's purely evil. So have to put your non standard builtin all over
> standards compliant code just to shut off a warning that is likely
> common. And yo
On Mon, Jul 11, 2016 at 1:01 PM, Marek Polacek wrote:
> On Mon, Jul 11, 2016 at 12:59:51PM -0700, Andrew Pinski wrote:
>> I don't like it being default turned on or even part of -Wall. As
>> evidence all of the patches that comes after this is the main reason
>> why.
>
> Ok, moving this to -Wextr
On Mon, Jul 11, 2016 at 12:59:51PM -0700, Andrew Pinski wrote:
> I don't like it being default turned on or even part of -Wall. As
> evidence all of the patches that comes after this is the main reason
> why.
Ok, moving this to -Wextra is rather trivial.
Marek
Marek Polacek writes:
>
> This warning is enabled by default for C/C++.
That's purely evil. So have to put your non standard builtin all over
standards compliant code just to shut off a warning that is likely
common. And you're not even supporting the classic lint style comment either.
I think i
2016-07-11 Marek Polacek
PR c/7652
* c-c++-common/Wswitch-unreachable-1.c: Add __builtin_fallthrough.
* c-c++-common/pr44832.c: Likewise.
* g++.dg/lto/20090128_0.C: Likewise.
* g++.dg/opt/pr14029.C: Likewise.
* g++.dg/opt/pr45412.C: Likewise.
On Mon, Jul 11, 2016 at 12:43 PM, Marek Polacek wrote:
> The switch fallthrough has been widely considered a design defect in C, a
> misfeature or, to use Marshall Cline's definition, evil. The overwhelming
> majority of the time you don't want to fall through to the next case, but it
> is
> eas
2016-07-11 Marek Polacek
PR c/7652
* config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass): Likewise.
(rs6000_legitimate_offset_address_p): Likewise.
(rs6000_emit_move): Likewise.
(altivec_expand_ld_builtin): Likewise.
(altivec_expand_st_buil
2016-07-11 Marek Polacek
PR c/7652
* config/alpha/alpha.c (alpha_rtx_costs): Likewise.
(alpha_legitimate_constant_p): Likewise.
(alpha_emit_setcc): Likewise.
(alpha_emit_xfloating_libcall): Likewise.
(alpha_function_value_1): Likewise.
(al
2016-07-11 Marek Polacek
PR c/7652
* objc-encoding.c (encode_type): Add gcc_fallthrough.
diff --git gcc/gcc/objc/objc-encoding.c gcc/gcc/objc/objc-encoding.c
index 41ac6a4..cc64e1b 100644
--- gcc/gcc/objc/objc-encoding.c
+++ gcc/gcc/objc/objc-encoding.c
@@ -622,6 +622,7 @@ enco
2016-07-11 Marek Polacek
PR c/7652
* lto-plugin.c (lto_fallthrough): Define.
(parse_table_entry): Use it.
diff --git gcc/lto-plugin/lto-plugin.c gcc/lto-plugin/lto-plugin.c
index 51afc52..ffdf54a 100644
--- gcc/lto-plugin/lto-plugin.c
+++ gcc/lto-plugin/lto-plugin.c
@@
2016-07-11 Marek Polacek
PR c/7652
* libsupc++/hash_bytes.cc: Use __builtin_fallthrough.
diff --git gcc/libstdc++-v3/libsupc++/hash_bytes.cc
gcc/libstdc++-v3/libsupc++/hash_bytes.cc
index 2e5bbfa..818331f 100644
--- gcc/libstdc++-v3/libsupc++/hash_bytes.cc
+++ gcc/libstdc++-v
2016-07-11 Marek Polacek
PR c/7652
* runtime/heapdump.c (dumpefacetypes): Add break.
diff --git gcc/libgo/runtime/heapdump.c gcc/libgo/runtime/heapdump.c
index d0cfb01..1ce8aa2 100644
--- gcc/libgo/runtime/heapdump.c
+++ gcc/libgo/runtime/heapdump.c
@@ -766,6 +766,7 @@ dumpefa
2016-07-11 Marek Polacek
PR c/7652
* cp-demangle.c (d_print_comp_inner): Use D_FALLTHRU.
(d_print_mod): Likewise.
* cplus-dem.c (demangle_signature): Likewise.
(demangle_fund_type): Likewise.
(do_hpacc_template_const_value): Likewise.
* d-
2016-07-11 Marek Polacek
PR c/7652
* libgomp.h (gomp_fallthrough): Define.
* oacc-init.c (resolve_device): Use gomp_fallthrough.
* testsuite/libgomp.c++/cancel-parallel-2.C: Add __builtin_fallthrough.
* testsuite/libgomp.c/cancel-parallel-2.c: Add __built
2016-07-11 Marek Polacek
PR c/7652
* soft-fp/soft-fp.h (_FP_FALLTHRU): Define.
* soft-fp/op-common.h: Use it.
diff --git gcc/libgcc/soft-fp/op-common.h gcc/libgcc/soft-fp/op-common.h
index 080ef0e..6691f50 100644
--- gcc/libgcc/soft-fp/op-common.h
+++ gcc/libgcc/soft-fp
2016-07-11 Marek Polacek
PR c/7652
* internal.h (CPP_FALLTHRU): Define.
* expr.c (_cpp_parse_expr): Use CPP_FALLTHRU.
* lex.c (search_line_fast): Likewise.
(lex_raw_string): Likewise.
(_cpp_lex_direct): Likewise.
(cpp_token_val_index): Lik
2016-07-11 Marek Polacek
PR c/7652
* libatomic_i.h (libat_fallthrough): New macro.
* gcas.c (libat_compare_exchange): Use libat_fallthrough.
* gexch.c (libat_exchange): Likewise.
* glfree.c (libat_is_lock_free): Likewise.
* gload.c (libat_load): L
2016-07-11 Marek Polacek
PR c/7652
* expr.c (java_truthvalue_conversion): Add gcc_fallthrough.
(type_assertion_hasher::hash): Likewise.
* jcf-dump.c (print_constant): Likewise.
* jcf-io.c (verify_constant_pool): Likewise.
* typeck.c (promote_type)
2016-07-11 Marek Polacek
PR c/7652
* config/i386/driver-i386.c (decode_caches_intel): Likewise.
(detect_caches_cpuid4): Likewise.
* config/i386/i386-c.c (ix86_target_macros_internal): Likewise.
* config/i386/i386.c (function_arg_advance_32): Likewise.
2016-07-11 Marek Polacek
PR c/7652
* go-system.h (go_fallthrough): New macro.
* gofrontend/escape.cc (Escape_analysis_assign::statement): Use it.
(Escape_analysis_assign::assign): Likewise.
* gofrontend/expressions.cc
(Binary_expression::do_get_b
2016-07-11 Marek Polacek
PR c/7652
* config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args): Add
gcc_fallthrough.
* config/aarch64/aarch64-simd.md: Likewise.
* config/aarch64/aarch64.c (aarch64_expand_mov_immediate): Likewise.
(aarch64_print_operan
2016-07-11 Marek Polacek
PR c/7652
* Makefile.in (insn-attrtab.o-warn, insn-dfatab.o-warn,
insn-latencytab.o-warn, insn-output.o-warn, insn-emit.o-warn): Add
-Wno-switch-fallthrough.
* alias.c (find_base_value): Add gcc_fallthrough.
(find_base_ter
2016-07-11 Marek Polacek
PR c/7652
* call.c (add_builtin_candidate): Add gcc_fallthrough.
(add_builtin_candidates): Likewise.
(build_integral_nontype_arg_conv): Likewise.
(build_new_op_1): Likewise.
(convert_like_real): Likewise.
* cfns.h
2016-07-11 Marek Polacek
PR c/7652
* arith.c (eval_intrinsic): Add gcc_fallthrough.
* array.c (gfc_ref_dimen_size): Likewise.
(gfc_array_dimen_size): Likewise.
* cpp.c (gfc_cpp_handle_option): Likewise.
(cb_cpp_error): Likewise.
* decl.c (
2016-07-11 Marek Polacek
PR c/7652
* c-ada-spec.c (print_ada_macros): Add gcc_fallthrough.
(to_ada_name): Likewise.
(dump_generic_ada_node): Likewise.
(dump_nested_type): Likewise.
* c-common.c (warn_if_unused_value): Likewise.
(sizeof_poi
2016-07-11 Marek Polacek
PR c/7652
* c-array-notation.c (expand_array_notations): Add gcc_fallthrough.
* c-decl.c (pop_scope): Likewise.
(grokdeclarator): Likewise.
(get_parm_info): Likewise.
* c-objc-common.c (c_tree_printer): Likewise.
*
2016-07-11 Marek Polacek
PR c/7652
* builtins.c (expand_builtin): Handle BUILT_IN_FALLTHROUGH.
* builtins.def: Add BUILT_IN_FALLTHROUGH.
* common.opt (Wswitch-fallthrough): New option.
* doc/extend.texi: Document __builtin_fallthrough.
* doc/invok
No functional changes.
2016-07-11 Uros Bizjak
* config/i386/predicates.md (x86_64_immediate_operand) :
Hoist common subexpressions.
(x86_64_zext_immediate_operand) : Ditto.
Bootstrapped and regression tested on x86_64-linux-gnu {,-m32}.
Committed to mainline SVN.
Uros.
diff --gi
The switch fallthrough has been widely considered a design defect in C, a
misfeature or, to use Marshall Cline's definition, evil. The overwhelming
majority of the time you don't want to fall through to the next case, but it is
easy to forget to "break" at the end of the case, making this far too
Hi,
It was recently brought to my attention that glibc needs access to complex
multiply and divide for IEEE-128 floating-point in GCC 6.2 in order to move
ahead with the library implementation work. This patch enables this support
using only target-specific changes to avoid any possible effect on
Hi!
Most of the spots in tsubst_baselink that actually access baselink after
it has been assigned lookup_fnfields () test that it is a BASELINK_P, except
one - the BASELINK_OPTYPE update. lookup_fnfields can return
error_mark_node though, perhaps something else too. The patch just follows
what t
Hi!
vect_get_vec_defs handles only one or two arguments, for ternary
ops like fma vectorizable_operation calls vect_get_vec_defs on the
first two arguments, and then did
vec_oprnds2.create (1);
vec_oprnds2.quick_push (vect_get_vec_def_for_operand (op2,
Hi!
For SIZEOF_EXPR, we rely on cp_fold to fold it.
But, for VEC_INIT_EXPR initialization, we actually just genericize it
without ever folding the expressions, so e.g. if the ctor has default args
and some complicated expressions in there, they will never be cp_folded.
This is the only place that
> It is fine to merge decls across static and external flags, but I am not
> sure this is a safe solution to the problem. In C it is perfectly normal to
> have one decl more specified or with different attributes. Like:
>
> extern int a __attribute__ ((warning("bar"));
>
> int a=7 __attribute__
We've decided to apply the following patch to ARM/embedded-6-branch.
Best regards,
Thomas
-- Forwarded Message --
Subject: Re: [PATCH, ARM 3/7, ping1] Fix indentation of FL_FOR_ARCH* definition
after adding support for ARMv8-M
Date: Wednesday 18 May 2016, 14:47:16
From: Kyril
We've decided to apply the following patch to ARM/embedded-6-branch.
Best regards,
Thomas
-- Forwarded Message --
Subject: Re: [PATCH, ARM 4/7, ping1] Factor out MOVW/MOVT availability and
desirability checks
Date: Thursday 07 July 2016, 09:59:53
From: Thomas Preudhomme
To:
We've decided to apply the following patch to ARM/embedded-6-branch.
Best regards,
Thomas
-- Forwarded Message --
Subject: Re: [PATCH, ARM 2/7, ping1] Add support for ARMv8-M
Date: Thursday 07 July 2016, 09:57:08
From: Thomas Preudhomme
To: Kyrill Tkachov
CC: Richard Earnsha
We've decided to apply the following patch to ARM/embedded-6-branch.
Best regards,
Thomas
-- Forwarded Message --
Subject: Re: [PATCH, libgcc/ARM 1/6] Fix Thumb-1 only == ARMv6-M & Thumb-2
only == ARMv7-M assumptions
Date: Wednesday 06 July 2016, 17:20:04
From: Ramana Radhakr
We've decided to apply the following patch to ARM/embedded-6-branch.
Best regards,
Thomas
-- Forwarded Message --
Subject: Re: [PATCH, libgcc/ARM 1a/6] Fix Thumb-1 only == ARMv6-M & Thumb-2
only == ARMv7-M assumptions
Date: Wednesday 06 July 2016, 17:21:24
From: Ramana Radha
On 07/11/2016 09:40 AM, Aldy Hernandez wrote:
On 07/11/2016 11:08 AM, Martin Sebor wrote:
Hey, I'm all for different options. It would definitely be easier for
me :). I wast trying really hard to use -Wvla= and keep the current
-Wvla meaning the same. Though I see your point about using -Wvl
On 07/11/2016 04:52 PM, Kyrill Tkachov wrote:
Based on that, I think that code block is a useful optimisation, we just
need
to take care with immediates.
What do you think?
Yeah, I think the patch is ok.
Bernd
On 06/07/16 11:52, Andre Vieira (lists) wrote:
> On 01/07/16 14:40, Ramana Radhakrishnan wrote:
>>
>>
>> On 13/10/15 18:01, Andre Vieira wrote:
>>> This patch ports the aeabi_idiv routine from Linaro Cortex-Strings
>>> (https://git.linaro.org/toolchain/cortex-strings.git), which was
>>> contribut
On 07/07/16 13:30, mickael guene wrote:
> Hi Andre,
>
> Another feedback on your purecode patch.
> You have to disable casesi pattern since then it will
> generate wrong code with -mpure-code option.
> Indeed it will generate an 'adr rx, .Lx' (aka
> 'subs rx, PC, #offset') which will not work i
This code in store_parm_decls never fired, because we were checking
VOID_TYPE_P on the PARM_DECL itself rather than its type. And we
already check for void parms in grokparms, so there's no point in
doing it here.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 7bc6f4c24e2e853ff18bd61f116e0
On 07/11/2016 11:08 AM, Martin Sebor wrote:
Hey, I'm all for different options. It would definitely be easier for
me :). I wast trying really hard to use -Wvla= and keep the current
-Wvla meaning the same. Though I see your point about using -Wvla* but
using different variables for representi
Hi,
while looking into loop code I noticed that loop_only_exit_p seems overly
simplistic. The purpose of this predicate is to decide whether the loop
must terminate by the exit given (so for example it is known that the number
of iteration test in that particular exit won't overflow because undefin
> On Jul 11, 2016, at 7:44 AM, Jeff Law wrote:
>
> On 07/08/2016 08:26 AM, Martin Liška wrote:
>> Hello
>>
>> Following patch fixes fallout caused by the patch set:
>> https://gcc.gnu.org/ml/gcc-regression/2016-07/msg00097.html
>>
>> Ready after it finished regression tests?
>> Thanks,
>> Mart
Hey, I'm all for different options. It would definitely be easier for
me :). I wast trying really hard to use -Wvla= and keep the current
-Wvla meaning the same. Though I see your point about using -Wvla* but
using different variables for representing -Wvla and -Wvla=blah.
The easiest thing w
Hi,
Sorry for jumping in late, only now I had chance to read through the whole
discussion.
I was looking into similar problem some time ago.
> Index: lto-streamer-out.c
> ===
> --- lto-streamer-out.c(revision 238156)
> +++ lt
On Mon, Jul 11, 2016 at 3:48 PM, H.J. Lu wrote:
> When converting V1TImode register in debug insn, check if it has been
> converted to TImode already.
>
> Tested on x86-64. OK for trunk?
I was looking to use simplify_subreg here, but it is too complicated
for the known purpose.
> H.J.
> ---
> g
On 07/07/2016 09:06 PM, James Bowman wrote:
FT32 makes use of multiple address spaces. When trying to inspect
objects in GDB, GDB was treating them as a straight "const". The cause
seems to be in GCC DWARF2 output.
This output is handled in gcc/gcc/dwarf2out.c, where modified_type_die()
checks t
On 05/07/16 13:57, Kyrill Tkachov wrote:
Hi Bernd,
On 04/07/16 19:02, Bernd Schmidt wrote:
On 07/01/2016 11:18 AM, Kyrill Tkachov wrote:
In this arm wrong-code PR the struct assignment goes wrong when
expanding constructor elements to a register destination
when the constructor elements are s
On 07/08/2016 06:47 AM, Martin Liška wrote:
Hi.
This is my second attempt of the patch where I generate all tests on fly.
Firstly, params-options.h is used to generate a list of options in form of:
"predictable-branch-outcome"=2,0,50
"inline-min-speedup"=10,0,0
"max-inline-insns-single"=400,0,0
On 07/08/2016 08:26 AM, Martin Liška wrote:
Hello
Following patch fixes fallout caused by the patch set:
https://gcc.gnu.org/ml/gcc-regression/2016-07/msg00097.html
Ready after it finished regression tests?
Thanks,
Martin
0001-Fix-Fortran-DO-loop-fallback.patch
From c5dd7ad62f795cce560c7f1b
On 07/08/2016 09:00 AM, David Malcolm wrote:
Ping.
I believe I need review of the selftest.h change; the rest I think I
can self-approve, if need be.
https://gcc.gnu.org/ml/gcc-patches/2016-06/msg01340.html
OK.
jeff
On 07/11/2016 01:55 AM, Martin Liška wrote:
Hello.
During investigation of an IVOPTs issue, I noticed that tree-ssa-loop-ivopts.c
uses a hard-coded
constant AVG_LOOP_NITER. Very similar constant can be seen in cfgloopanal.c.
Thus, I've changed
that to a new param.
Apart from that, I would lik
On 07/10/2016 09:32 AM, Bernd Edlinger wrote:
On 07/10/16 10:23, Eric Botcazou wrote:
I expect it still has an effect due to the same reason as the
DECL_OFFSET_ALIGN thing - memory reference expansion doesn't gather
information from the full reference but tries to re-cover it from
the pieces ret
On 07/10/2016 04:09 PM, Martin Sebor wrote:
On 07/08/2016 05:48 AM, Aldy Hernandez wrote:
I've played with the patch a bit over the weekend and have a few
comments and suggestions (I hope you won't regret encouraging me :)
I like the consistency between -Walloca and -Wvla! (And despite
the volum
On 07/08/2016 05:48 AM, Aldy Hernandez wrote:
[New thread now that I actually have a tested patch :)].
I think detecting potentially problematic uses of alloca would
be useful, especially when done in an intelligent way like in
your patch (as opposed to simply diagnosing every call to
the funct
On 11 July 2016 at 11:10, Aldy Hernandez wrote:
> On 07/10/2016 07:41 PM, Manuel López-Ibáñez wrote:
>>>
>>> +Walloca
>>> +LangEnabledBy(C ObjC C++ ObjC++)
>>> +; in common.opt
>>> +
>>> +Walloca=
>>> +LangEnabledBy(C ObjC C++ ObjC++)
>>> +; in common.opt
>>> +
>>
>>
>> I'm not sure what you think
On 07/08/2016 01:46 PM, David Malcolm wrote:
This patch adds a new phony target to gcc/Makefile.in to make it easy
to run the selftests under valgrind, via "make selftest-valgrind".
This phony target isn't a dependency of anything; it's purely for
convenience (it takes about 4-5 seconds on my box
When converting V1TImode register in debug insn, check if it has been
converted to TImode already.
Tested on x86-64. OK for trunk?
H.J.
---
gcc/
PR target/71801
* config/i386/i386.c (timode_scalar_chain::fix_debug_reg_uses):
Don't convert TImode in debug insn.
gcc/tests
Ping
2016-06-28 15:24 GMT+03:00 Ilya Enkovich :
> On 16 Jun 10:54, Jeff Law wrote:
>> On 05/19/2016 01:44 PM, Ilya Enkovich wrote:
>> >Hi,
>> >
>> >This patch introduces support for loop epilogue combining. This includes
>> >support in cost estimation and all required changes required to mask
>>
Ping
2016-06-23 12:54 GMT+03:00 Ilya Enkovich :
> On 22 Jun 11:42, Jeff Law wrote:
>> On 06/22/2016 10:09 AM, Ilya Enkovich wrote:
>>
>> >>Given the common structure & duplication I can't help but wonder if a
>> >>single
>> >>function should be used for widening/narrowing. Ultimately can't you s
Ping
2016-06-24 10:40 GMT+03:00 Ilya Enkovich :
> On 17 Jun 10:46, Jeff Law wrote:
>> On 06/17/2016 08:33 AM, Ilya Enkovich wrote:
>> >>
>> >>Hmm, there seems to be a level of indirection I'm missing here. We're
>> >>smuggling LOOP_VINFO_ORIG_LOOP_INFO around in loop->aux. Ewww. I thought
>> >>
Ping
2016-06-24 14:46 GMT+03:00 Ilya Enkovich :
> On 16 Jun 18:52, Ilya Enkovich wrote:
>> 2016-06-15 15:00 GMT+03:00 Richard Biener :
>> > On Thu, May 19, 2016 at 9:46 PM, Ilya Enkovich
>> > wrote:
>> >> Hi,
>> >>
>> >> This patch enables vectorization of loop epilogues and low trip count
>> >>
Ping
2016-06-22 17:13 GMT+03:00 Ilya Enkovich :
> On 16 Jun 00:16, Jeff Law wrote:
>> On 05/19/2016 01:40 PM, Ilya Enkovich wrote:
>> >Hi,
>> >
>> >This patch extends vectorizer cost model to include masking cost by
>> >adding new cost model locations and new target hook to compute
>> >masking cos
Ping
2016-06-16 16:42 GMT+03:00 Ilya Enkovich :
> On 20 May 14:40, Ilya Enkovich wrote:
>> > Can you make all these --params then? I think to be useful to users we'd
>> > want
>> > them to be loop pragmas rather than options.
>>
>> OK, I'll change it to params. I didn't think about control via
The following patch fixes PR71816.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2016-07-11 Richard Biener
PR tree-optimization/71816
* tree-ssa-pre.c (compute_avail): Adjust alignment of ref rather
than replacing all of its operands.
> So sth like
>
> Index: gcc/lto-streamer-out.c
> ===
> --- gcc/lto-streamer-out.c (revision 238039)
> +++ gcc/lto-streamer-out.c (working copy)
> @@ -996,7 +996,7 @@ hash_tree (struct streamer_tree_cache_d
>else
>
Alan Modra wrote:
> The reason this fails is that no alternative in altivec_mov
> exactly matches. Ideally reload would choose the Z,v alternative
> (cost 9) and you'd get an address reload for the mem to make it match
> the "Z" constraint. Instead reload chooses the Y,r alternative (cost
> 8) as
1 - 100 of 114 matches
Mail list logo
