On Wed, Oct 9, 2013 at 8:14 PM, Andrew MacLeod wrote:
> This patch simply moves std_gimplify_va_arg_expr() and the related
> build_va_arg_indirect_ref to gimplify.c where I think it belongs.
> It also moves gimple_fold_indirect_ref out of gimplfy.c and into
> gimple-fold.c.
>
> bootstraps on x86_6
On Wed, Oct 9, 2013 at 11:34 PM, Andrew MacLeod wrote:
> On 10/09/2013 02:18 PM, Andrew MacLeod wrote:
>>
>> On 10/09/2013 02:15 PM, Andrew MacLeod wrote:
>>>
>>>
>>> Fair enough. I'll adjust... the front end files which use that routine
>>> will just have to include gimplify.h
>>>
>> Unless ma
On Thu, Oct 10, 2013 at 2:10 AM, David Malcolm wrote:
> ipa-inline.c:ipa_inline leaks "order" when optimizations are disabled.
> This potentially becomes significant in my JIT branch since the
> compilation code could be invoked many times within one process.
>
> Bootstrapped and regtested against
As a general observation I don't like the
if (pass->has_foo_member_function)
pass->foo_member_function ()
style. It is totally against the idea of having virtual functions. Why
not simply provide stub implementations in the base classes?
Richard.
On Thu, Oct 10, 2013 at 2:41 AM, David Mal
On Thu, Oct 10, 2013 at 5:04 AM, Jeff Law wrote:
> On 08/05/13 02:08, Zhenqiang Chen wrote:
>>
>> Hi
>>
>> The patch reassociates X == CST1 || X == CST2 if popcount (CST2 - CST1) ==
>> 1
>> into ((X - CST1) & ~(CST2 - CST1)) == 0.
>>
>> Bootstrap on x86-64 and ARM chromebook.
>> No make check regr
"Moore, Catherine" writes:
> Hi Richard,
>
> This patch implements a workaround for errors on the PMC-Sierra RM7000
> cpu while executing the dmult or dmultu instruction. The workaround
> is to insert three nops after the dmult/dmultu.
Do you have any more details? E.g. does "dmult $4,$5;addiu
Dear Java maintainers, are you OK with this patch?
- Joey
> -Original Message-
> From: Ian Lance Taylor [mailto:i...@google.com]
> Sent: Thursday, September 12, 2013 3:28
> To: Joey Ye
> Cc: gcc-patches; H.J. Lu; p...@bothner.com; a...@redhat.com; Tom Tromey
> Subject: Re: [PATCH, libgcc]
Richard: Could you review the Gimple part? Thanks!
Joseph: Could you have a look at the C part? Thanks!
Jakub Jelinek wrote:
+ if (loop->latch)
+ FOR_EACH_EDGE (e, ei, loop->latch->succs)
+ {
+ if (e->dest->flags & BB_RTL)
+ break;
I'd prefer Richard to
This patch depends on the middle-end / C FE patch:
http://gcc.gnu.org/ml/gcc-patches/2013-10/msg00655.html
It adds parsing support for #pragma ivdep and annotates the condition of
C-like for loops such that one can later set the vectorization safe
length (loop->safelen) to INT_MAX. I consider
On Thu, Oct 10, 2013 at 04:22:52PM +0800, Joey Ye wrote:
> Dear Java maintainers, are you OK with this patch?
Given the state of gcj that it is now only rarely used and most people
just use OpenJDK instead, wouldn't it be a good idea to just require
that gcj code is linked using gcj driver or, if
> -Original Message-
> From: Jeff Law [mailto:l...@redhat.com]
> Sent: Thursday, October 10, 2013 11:05 AM
> To: Zhenqiang Chen; gcc-patches@gcc.gnu.org
> Subject: Re: [PATCH] Reassociate X == CST1 || X == CST2 if popcount (CST2
-
> CST1) == 1 into ((X - CST1) & ~(CST2 - CST1)) == 0
>
>
Hi,
One of our internal patches is a multilib patch that helps in testing
bare-metal toolchains for v7-a and above.
It's a bit brute force but this is something that we use internally
every night to build a set of base libraries in a hierarchical directory
structure to be used by bare metal
On 09/10/13 21:46, Jeff Law wrote:
> On 08/21/13 03:21, Andreas Krebbel wrote:
>> [RFC] Allow functions calling mcount before prologue to be leaf functions
>> http://gcc.gnu.org/ml/gcc-patches/2013-04/msg00993.html
> I don't think this is necessarily correct for all targets. ISTM the
> ability to
On Thu, Oct 10, 2013 at 05:25:01PM +0800, Zhenqiang Chen wrote:
> > Note I've been suggesting the bits I'm referring to in fold-const.c move
> out
> > into the tree-ssa optimizers. If they fit well into tree-ssa-reassoc.c
> I'd look
> > favorably upon a patch which moved them.
>
> The code is sim
Hi!
The following testcase distilled from Linux kernel is miscompiled,
in *.optimized we end up with asm goto whose only label points to
the fallthru basic block. During expansion, we unfortunately
put further instructions after it into the same basic block
(one in expand_gimple_block:
/* Expan
The frequently called Get_Source_File_Index function is optimized to
be inline and is now a simple array access. This is an internal
optimization only, with no functional effect (except hopefully
a compilation time speed up). No test required
Tested on x86_64-pc-linux-gnu, committed on trunk
2013
Add pragma Preelaborable_Initialization for Root_Storage_Pool type.
Tested on x86_64-pc-linux-gnu, committed on trunk
2013-10-10 Vadim Godunko
* s-stopoo.ads (Root_Storage_Pool): Add pragma
Preelaborable_Initialization.
Index: s-stopoo.ads
This is the front end part of the work for allowing Linker_Section
for types. No test filed yet, waiting for completion of required
back end (gigi) work.
Tested on x86_64-pc-linux-gnu, committed on trunk
2013-10-10 Robert Dewar
* sem_prag.adb (Analyze_Pragma, case Linker_Section): All
On Mon, Aug 05, 2013 at 01:39:50AM -0700, Andrew Pinski wrote:
> Seems like a better place to put this is inside tree-ssa-ifcombine.c
> which handles the case where if(a || b) is split up into if(a) else
> if(b).
> Moving it into tree-ssa-ifcombine.c allows for code to be optimized
> which was writ
Hi!
On Mon, 2013-09-16 10:32:24 -0400, Andrew MacLeod wrote:
> On 09/16/2013 05:05 AM, Richard Biener wrote:
> > On Sat, Sep 14, 2013 at 12:41 PM, Jan-Benedict Glaw
> > wrote:
> > >My Build Robot[1] found this recent commit to break Alpha:
> > >
> > > * tree-flow.h (FREE_SSANAMES): Move
Optimized version of Ada.Containers.Indefinite_Holders is used on
platforms which supports atomic operations.
Tested on x86_64-pc-linux-gnu, committed on trunk
2013-10-10 Vadim Godunko
* a-coinho-shared.ads, a-coinho-shared.adb: New file.
* s-atocou.ads: Add procedure to initi
This patch allows the use of static expressions, including most
importantly literals that accomodate the full range of positive
values for the Address, up to 2**a-1, where a is the size of
address values (most typically 32 or 64).
The following is compiled on a 32-bit machine:
1. with System
The Address attribute definition clause processing warns when overlaying
a small object with a large one, and also avoids giving a warning about
the entity not being referenced. This patch applies the same processing
to corresponding uses of the Address aspect. The following test program
shows the
This patch fixes a visibility hole in the instantiation of child units. Before
this patch, a declaration in the private part of a non-generic common ancestor
was made visible by an instantiation of a child unit in another descendant of
that ancestor.
Compiling p-c.ads must be rejected with:
p
The compiler was expanding choices in variant parts and case
statements and expressions prematurely, so that the ASIS tree
had these expansions, causing malfunction of ASIS tools (notably
gnatpp, the pretty printer) to malfunction. This patch moves
the expansion to the expander where it belongs so
Compiling wrong_aspect.ads must yield:
wrong_aspect.ads:8:62: aspect identifier expected
wrong_aspect.ads:10:50: aspect identifier expected
wrong_aspect.ads:12:50: aspect identifier expected
---
package Wrong_Aspect is
type Bits_1 is mod 2**1 with Size => 1;
type Bits_2 is mod 2
On Mon, Aug 05, 2013 at 04:08:58PM +0800, Zhenqiang Chen wrote:
> ChangeLog
> 2013-08-05 Zhenqiang Chen
>
> * tree-ssa-reassoc.c (optimize_range_tests): Reasociate
> X == CST1 || X == CST2 if popcount (CST2 - CST1) == 1 into
> ((X - CST1) & ~(CST2 - CST1)) == 0.
>
> testsuite
On 10/10/2013 06:44 AM, Jan-Benedict Glaw wrote:
Hi!
On Mon, 2013-09-16 10:32:24 -0400, Andrew MacLeod wrote:
On 09/16/2013 05:05 AM, Richard Biener wrote:
On Sat, Sep 14, 2013 at 12:41 PM, Jan-Benedict Glaw wrote:
My Build Robot[1] found this recent commit to break Alpha:
* tree-
This patch adds the name of entity that is improperly used as the name in a
package or subprogram instantiation.
Compiling generic_test.adb must yield:
generic_test.adb:30:10: "E" is not the name of a generic package
---
procedure Generic_Test is
generic
M : Positive;
package G
For unconstrained types, Ada.Sequential_IO needs to encode length information
prior to actual object representation. This length information is now stored
with the same endianness as the object itself, which means that the same
representation is now generated, independent of platform endianness, if
This is a fairly major reorganization of the way that choices are
handled. The processing of such choices in variants was completely
wrong before, since we attempted to do it when the variant is first
seen, but the statically predicated subtype may not be frozen yet,
so this processing has to be de
When the builder is invoked with -R, shared libraries are not linked
with an Rpath.
Tested on x86_64-pc-linux-gnu, committed on trunk
2013-10-10 Vincent Celier
* mlib-prj.adb (Build_Library): Do not issue link dynamic
libraries with an Rpath, if switch -R was used.
Index: mli
This patch provides the initial implementation of aspect/pragma Refined_Post.
This construct is intended for formal verification proofs.
The syntax of the aspect is as follows:
ASPECT_DEFINITION ::= ( boolean_EXPRESSION )
The syntax of the pragma is as follows:
pragma Refined_Post ( boole
This patch fixes regressions caused by the previous checkin to
handle statically predicated subtypes in variant choices. Error
messages were not being given in full code generation mode.
The following should compile with the indicated errors in both
-gnatc and normal code generation mode.
1.
This patch fixes a name resolution problem in a child instance. involving a
name of the form A.B.C, where A. B and C constitute a generic hierarchy.
Within an instance of C, this name must be rewritten to denote the instances
of A, B and C. Previously the compiler only handled one level of child un
This patch completes the implementation of statically predicated
subtypes as choices in record variants. The following should
compile quietly in both code generation and -gnatct mode:
1. package Predicate_Variant is
2.type Color is
3. (Red, Orange, Yellow, Green, Blue, Indi
Hi,
I think that in order to resolve this very old accepts-invalid, we can
simply adjust the expr to the cv-qualifiers of probe_type. I also double
checked that thing are fine for other combinations of cv-qualified
reference types as parameter and as argument.
Tested x86_64-linux.
Thanks,
P
The two routines Is_Other_Format and Is_Punctuation_Connector
are thus renamed (from Is_Other and Is_Punctuation) in packages
Ada.Wide_Character.Handling and Ada.Wide_Wide_Character.Handling
to reflect the final decision on names.
The following should compile as shown
1. with Ada.Wide_Charac
Previously we gave a warning and ignored all such aspects, including
language-defined ones, which must be diagnosed as illegal. This
patch considers all aspects illegal in this context. Why not? They
are our aspects, and if we are not going to implement them in this
context, they should not be allo
Duplicated index values are now displayed for case statement and
error flags for both duplicated and missing case values are placed
more precisely as shown by the compilation of the following test
compiled with -gnatj60.
1. package dupaggr is
2.type enum is (a, b, c, d, e, f);
3
For reasons explained in source comments, we ignore linker options
that appear in predefined generic units. This patch adds a fatal
error message if this is attempted. This is an internal change
only since it affects only implementation units, so no test needed
Tested on x86_64-pc-linux-gnu, commi
This is part of the previous work on getting variant choices that
involve a statically predicated subtype to work. The one gap was
in the generic case, where we were detecting the error on instances
instead of the template (this also called a failure of ACATS test
B49007A, where we had lots such a
A warning is added for if statements which have no else or elsif,
and a then that is a single null statement, and a condition that
has no obvious side effects. The following is compiled with
-gnatwr.d
1. function WarnOnIf (X : Integer) return Boolean is
2.V : Integer with Volatile;
This patch corrects the string representation of attribute Default_Aspect_Value
when it is applied to scalar types. The string name is for internal debugging
purposes only, no test.
Tested on x86_64-pc-linux-gnu, committed on trunk
2013-10-10 Hristian Kirtchev
* einfo.adb (Write_Field
A subtype of an interface type can appear as a progenitor in a type extension.
The routine that determines whether a given interface is a progenitor of a
type must take this subtype into account.
Compiling and executing main.adb must yield:
The integer: 42
---
with Ultimate_User;
procedure M
This patch adds a number of functions to Ada.Characters.Handling that
were added late on in the design process and missed it into the first
version of this package. The test below tests these missing functions.
It is to be compiled with -gnata and executed:
1. with Ada.Characters.Handling;
This patch provides the initial implementation of aspect/pragma Refined_State.
The construct is intended for formal verification proofs.
The syntax of the aspect/pragma is as follows:
pragma Refined_State (REFINEMENT_LIST);
REFINEMENT_LIST ::=
REFINEMENT_CLAUSE
| (REFINEMENT_CLAU
Jakub> Given the state of gcj that it is now only rarely used and most
Jakub> people just use OpenJDK instead, wouldn't it be a good idea to
Jakub> just require that gcj code is linked using gcj driver or, if
Jakub> linked in any other driver, just using a special non-default
Jakub> option (-flink-
More fun target specific stuff. gimplify.c needs the back end supplied
macros/function for the va-arg padding functions to bootstrap...
confirmed to build stage1 cross compile on rs6000 and mips64.
checked in as revision 203373... hopefully that's the last of them :-P
Andrew
* gimpli
This fixes an unintended side-effect of the PR58464 fix, caching
failed translations has issues with insertion iteration.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2013-10-10 Richard Biener
PR tree-optimization/58656
* tree-ssa-pre.c (phi_transla
Marc> + if (flag_delete_null_pointer_checks
Marc> + && lookup_attribute ("returns_nonnull",
Marc> + TYPE_ATTRIBUTES (TREE_TYPE (fndecl
Marc> + return true;
I wired all this up to gcj, but it tripped over the
flag_delete_null_pointer_checks test, because java/lang.c sets it:
/
This patch addresses an ICE when building qemu for a Mips target in
Yocto. Both gcc-trunk, gcc-4.8 and all of the targets are potentially
affected. The problem occurs because the instruction combine phase uses
two data structures to keep track of registers, reg_stat and
regstat_n_sets_and_refs, but
Hello,
Thanks a lot for all reviews!
On 09 Oct 14:55, Richard Henderson wrote:
> On 10/09/2013 03:28 AM, Kirill Yukhin wrote:
> > +;; CPUID bit AVX512F enables evex encoded scalar and 512-bit fma. It
> > doesn't
> > +;; care about FMA bit, so we enable fma for TARGET_AVX512F even when
> > TARGE
On Thu, Oct 10, 2013 at 4:17 PM, Tom Tromey wrote:
> Marc> + if (flag_delete_null_pointer_checks
> Marc> + && lookup_attribute ("returns_nonnull",
> Marc> + TYPE_ATTRIBUTES (TREE_TYPE (fndecl
> Marc> + return true;
>
> I wired all this up to gcj, but it tripped over the
> flag_d
On Thu, Oct 10, 2013 at 3:10 PM, Andrew MacLeod wrote:
> More fun target specific stuff. gimplify.c needs the back end supplied
> macros/function for the va-arg padding functions to bootstrap...
Err ... don't we have the va_arg gimplify target hooks to hide this
kind of target dependency?
Richa
On Thu, Oct 10, 2013 at 4:30 PM, Richard Biener
wrote:
> On Thu, Oct 10, 2013 at 3:10 PM, Andrew MacLeod wrote:
>> More fun target specific stuff. gimplify.c needs the back end supplied
>> macros/function for the va-arg padding functions to bootstrap...
>
> Err ... don't we have the va_arg gimpl
On Wed, Oct 09, 2013 at 08:17:55PM -0400, Jason Merrill wrote:
> Please add a comment explaining why an invisiref decl is handled specially.
Ok.
> >@@ -1345,6 +1384,30 @@ cplus_decl_attributes (tree *decl, tree
> > || *decl == error_mark_node)
> > return;
> >
> >+ /* Add implicit "omp
Patch updated.
Thanks,
Dehao
On Wed, Oct 9, 2013 at 10:45 PM, Xinliang David Li wrote:
> > /* Program behavior changed, original promoted (and inlined) target is not
> > hot any more. Will avoid promote the original target. */
> > if (total >= info->first * 0.5)
> > return false;
>
> This
Greetings,
For Google b/10860844, I've committed the patch below on google gcc-4_8
(r203381) and integration (r203382) branches.
This forces abi-version to be the latest available.
2013-10-10 Paul Pluizhnikov
* gcc/common.opt (flag_abi_version): Set to 0.
Index: gcc/common.opt
==
Hi,
The libstdc++-v3 testcase atomic/cons/49445.cc fails for a variety of
arm configurations that do not provide atomic builtins because the test
is not gated by dg-require-atomic-builtins
OK ?
/M
2013-10-10 Marcus Shawcroft
* testsuite/29_atomics/atomic/cons/49445.cc
(d
ping^2
Thanks,
Dehao
On Tue, Oct 1, 2013 at 1:28 PM, Dehao Chen wrote:
> Hi,
>
> This patch disables the C++ frontend to add " *INTERNAL* " suffix to
> maybe_in_charge_destructor/constructor. This is needed because these
> functions could be emitted in the debug info, and we would want to
> dema
This is a patch for trunk. It should be backported to 4_8, once it thaws.
OK for trunk?
2013-10-10 Paulo Matos
PR gcov-profile/58682
* ipa-inline.c (inline_small_functions): Update
max_count if new edges are added after inline_call.
Paulo Matos
pr58682.patch
Desc
Hi Paul,
On 10/10/13 17:10, Paulo Matos wrote:
- inline_call (edge, true, &new_indirect_edges, &overall_size, true);
+ if (inline_call (edge, true, &new_indirect_edges, &overall_size,
true))
+{
+ /* Update max_count if new edges were found */
Comments e
On 10/10/2013 03:22 AM, Jakub Jelinek wrote:
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk and 4.8?
>
> 2013-10-10 Jakub Jelinek
>
> PR middle-end/58670
> * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
> if any labels are in FALLTHRU_BB, use
Hi,
lets start with your unification code and then I will update my patch.
>
> The patch is bootstrapped and tested on i386/x86_64 (make check, and stability
> testing on Spec2k, Spec2k6).
Please do not forget to check also -minline-all-stringop.
>
> diff --git a/gcc/config/i386/i386.c b/gcc/con
On 10/10/13 08:17, Tom Tromey wrote:
In
addition, other optimization passes in GCC use this flag to
control global dataflow analyses that eliminate useless checks for
null pointers; these assume that if a pointer is checked after it
has already been dereferenced, i
On Thu, Oct 10, 2013 at 07:26:43AM -0700, Cesar Philippidis wrote:
> This patch addresses an ICE when building qemu for a Mips target in
> Yocto. Both gcc-trunk, gcc-4.8 and all of the targets are potentially
> affected. The problem occurs because the instruction combine phase uses
> two data struc
On 10 October 2013 16:53, Marcus Shawcroft wrote:
> Hi,
>
> The libstdc++-v3 testcase atomic/cons/49445.cc fails for a variety of arm
> configurations that do not provide atomic builtins because the test is not
> gated by dg-require-atomic-builtins
>
> OK ?
OK, thanks.
On 10/10/13 04:00, Andreas Krebbel wrote:
On 09/10/13 21:46, Jeff Law wrote:
On 08/21/13 03:21, Andreas Krebbel wrote:
[RFC] Allow functions calling mcount before prologue to be leaf functions
http://gcc.gnu.org/ml/gcc-patches/2013-04/msg00993.html
I don't think this is necessarily correct for
On 10/10/2013 07:27 AM, Kirill Yukhin wrote:
> Currently, from HW point of view, there're no CPUs which
> feature AVX-512, but not AVX2. So, I believe we may put
> a `TODO` in comment, like this:
>
> +;; CPUID bit AVX512F enables evex encoded scalar and 512-bit fma. It doesn't
> +;; care about FM
On 10/10/2013 10:35 AM, Richard Biener wrote:
On Thu, Oct 10, 2013 at 4:30 PM, Richard Biener
wrote:
On Thu, Oct 10, 2013 at 3:10 PM, Andrew MacLeod wrote:
More fun target specific stuff. gimplify.c needs the back end supplied
macros/function for the va-arg padding functions to bootstrap...
My only comment is that you change the parameters to
c_parser_binary_expression, but don't change the comment above that
function documenting what they are.
--
Joseph S. Myers
jos...@codesourcery.com
"Moore, Catherine" writes:
>> -Original Message-
>> From: Richard Sandiford [mailto:rdsandif...@googlemail.com]
>> Subject: Re: [PATCH] Workaround errata for the PMC-Sierra RM7000 cpu.
>>
>> "Moore, Catherine" writes:
>> > Hi Richard,
>> >
>> > This patch implements a workaround for erro
Hi,
this patch enables X86_TUNE_SSE_TYPELESS_STORES for generic. This is because
it is currently enabled for both Bulldozer and Cores and I think in fact all
modern chips honnors it. I am not sure why it is off for Atom/Slm and Bobcat.
Second change is to enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bul
On Thu, Oct 10, 2013 at 05:14:02PM +, Joseph S. Myers wrote:
> My only comment is that you change the parameters to
> c_parser_binary_expression, but don't change the comment above that
> function documenting what they are.
Thanks, here is the fix for that:
2013-10-10 Jakub Jelinek
From: Trevor Saunders
Hi,
This makes the implementation of stack vectors simpler and easier to use. This
works by
making the size of the on stack
On Wed, 2013-10-09 at 19:56 -0700, Ian Lance Taylor wrote:
> On Wed, Oct 9, 2013 at 5:18 PM, David Malcolm wrote:
> > coverage_init allocates da_file_name using XNEWVEC, but it is never
> > freed. This shows up as a slow memory leak in my JIT code when
> > repeatedly invoking the compiler code wi
On Thu, 2013-10-10 at 09:53 +0200, Richard Biener wrote:
> On Thu, Oct 10, 2013 at 2:10 AM, David Malcolm wrote:
> > ipa-inline.c:ipa_inline leaks "order" when optimizations are disabled.
> > This potentially becomes significant in my JIT branch since the
> > compilation code could be invoked many
On 10/10/2013 08:33 AM, Paolo Carlini wrote:
+ expr_type = TREE_TYPE (expr) = cp_build_qualified_type
+ (TREE_TYPE (expr), cp_type_quals (TREE_TYPE (probe_type)));
Won't that end up being the same as the contents of expr_type before
this statement? Can we just remove
The following patch fixes an inefficiency whereby type unification
was being attempted unnecessarily in LIPO even when there were no
aux modules.
Tested with regression tests and internal LIPO benchmark.
Ok for google/4_8?
Thanks,
Teresa
2013-10-10 Teresa Johnson
* l-ipo.c (cgraph_u
On 10/10/2013 11:01 AM, Jakub Jelinek wrote:
int inhibit_evaluation_warnings;
+ int omp_declare_target_attribute;
Let's keep the comment you had on the variable, just put it on the
member now.
OK with that change.
Jason
Hi,
this patch makes ACCUMULATE_OUTGOING_ARGS to disable itself when function is
cold. I did some extra testing and to my amusement we now seem to output
more compact unwind info when ACCUMULATE_OUTGOING_ARGS is disabled, so this
seems quite consistent code size win.
We actually can do better and
It is better to turn it on when random stress test is also on: param
value of PARAM_LIPO_RANDOM_GROUP_SIZE is not the default value (0).
if (!L_IPO_COMP_MODE || (num_in_fnames == 1 && PARAM_VALUE
(PARAM_LIPO_RANDOM_GROUP_SIZE) == 0))
ok with this change.
David
On Thu, Oct 10, 2013 at 11:27 AM,
On Thu, Oct 10, 2013 at 11:42 AM, Xinliang David Li wrote:
> It is better to turn it on when random stress test is also on: param
> value of PARAM_LIPO_RANDOM_GROUP_SIZE is not the default value (0).
I don't understand why - the random group size is used at dyn-ipa time
to compute module groups,
On 10/09/2013 11:14 AM, Mitchell, Mark wrote:
Unfortunately, I no longer remember this code well. My recollection --
such as it is -- is just that I couldn't find a case where it was
necessary to commit only a single level. You sometimes had to scan an
unbounded number of tokens to complete the
The option/parameter in theory should not have any effect to
profile-use build, but it is usually passed to both profile-gen and
profile-use compilation.
The reason to turn it on is to surface the type of error you have
fixed recently (it is triggered only with trivial module group).
David
On T
On 10/10/13 18:41, Jeff Law wrote:
> On 10/10/13 04:00, Andreas Krebbel wrote:
>> On 09/10/13 21:46, Jeff Law wrote:
>>> On 08/21/13 03:21, Andreas Krebbel wrote:
[RFC] Allow functions calling mcount before prologue to be leaf functions
http://gcc.gnu.org/ml/gcc-patches/2013-04/msg00993.h
Ian?
Ping x4.
On Wed, Sep 18, 2013 at 4:24 PM, Paolo Carlini wrote:
>>> Original message:
>>> http://gcc.gnu.org/ml/gcc-patches/2013-08/msg00394.html
>>
>>
> I'm adding Ian in CC.
--
Paul Pluzhnikov
The patch is ok.
David
On Thu, Oct 10, 2013 at 8:09 AM, Dehao Chen wrote:
> Patch updated.
>
> Thanks,
> Dehao
>
> On Wed, Oct 9, 2013 at 10:45 PM, Xinliang David Li wrote:
>> > /* Program behavior changed, original promoted (and inlined) target is not
>> > hot any more. Will avoid promote th
Hi,
On 10/10/2013 08:26 PM, Jason Merrill wrote:
On 10/10/2013 08:33 AM, Paolo Carlini wrote:
+ expr_type = TREE_TYPE (expr) = cp_build_qualified_type
+(TREE_TYPE (expr), cp_type_quals (TREE_TYPE (probe_type)));
Won't that end up being the same as the contents of expr_type be
I see, we would be using the presence of this parameter to drive lipo
stress-testing in general. I made the change and added a comment to
that effect and will commit it.
Thanks,
Teresa
On Thu, Oct 10, 2013 at 12:03 PM, Xinliang David Li wrote:
> The option/parameter in theory should not have any
On Mon, Oct 7, 2013 at 10:49 PM, Teresa Johnson wrote:
> On Sun, Oct 6, 2013 at 6:51 AM, Jan Hubicka wrote:
>>> The second part ensures that frequencies are correctly updated after
>>> inlining. The problem is that after inlining the frequencies were
>>> being recomputed directly from the corresp
On Wed, Aug 7, 2013 at 11:51 AM, Paul Pluzhnikov wrote:
>
> The following source:
>
> template static void f();
> void g() { f(); }
>
> results in "_Z1fIiEvv" under g++, but in "_ZL1fIiEvv" under clang.
>
> Richard Smith says:
>
> The ABI doesn't cover manglings for local symbols ...
> ...
On 2013-10-09 4:37, Jason Merrill wrote:
On 10/07/2013 05:14 AM, Adam Butcher wrote:
+ /* Forbid ambiguous implicit pack expansions by only allowing
+a single generic type in such a parameter.
+XXX: Maybe allow if explicitly specified temp
On Mon, Oct 7, 2013 at 10:45 PM, Teresa Johnson wrote:
> On Sun, Oct 6, 2013 at 6:43 AM, Jan Hubicka wrote:
>>> 2013-10-03 Teresa Johnson
>>>
>>> * params.def (PARAM_MIN_HOT_RUN_RATIO): New parameter.
>>
>> I would not mention HOT in the parameter name. Blocks are
>> hot/normal/unlik
On Wed, Oct 9, 2013 at 2:51 PM, Cary Coutant wrote:
>> 2013-10-09 Sterling Augustine
>>
>> * dwarf2out.c (include_pubname_in_output): Add conditional on
>> is_declaration_die
>> and debug_generate_pubnames.
>
> OK for google 4.8 branch. Thanks!
>
> -cary
Committed as r/203396.
This patch to libgo lets it report "len out of range" when making a
slice with a very large length. This is not strictly necessary as it
already check for a very large capacity, but since the capacity is often
implicit this patch gives a better error. Bootstrapped and ran Go
testsuite on x86_64-u
On 10/10/2013 04:45 PM, Adam Butcher wrote:
Of course; makes sense. I had confused myself. Not that it's a
function parameter pack, but presumably you would expect
auto f(tuple p)
to be supported?
That makes sense, though it's a (further) extension.
Can we put this in cp_parser, invert
This patch to the Go frontend avoids an extra error for an anonymous
embedded type. The compiler now checks for undefined types elsewhere,
and this error is unnecessary. Bootstrapped and ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to mainline. Will commit to 4.8
branch when it reope
This patch from Chris Manghane changes the Go frontend to use the
backend interface for function declarations and definitions.
Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu.
Committed to mainline.
Ian
2013-10-10 Chris Manghane
* go-gcc.cc (Backend::error_function): Ne
1 - 100 of 106 matches
Mail list logo
