On Tue, Jun 11, 2013 at 09:56:11AM +0930, Alan Modra wrote:
[snip]
It isn't hard to see why we are going wrong. IBM long double is
really a two element array of double, and the rs6000 backend uses
subregs to access the elements. The problem is that lower-subreg
lowers to word_mode, so we
Jan Hubicka hubi...@ucw.cz wrote:
Hi,
this patch implements the type inheritance graph builder. Once the
graph is
built it stays in memory and unchanged thorough the compilation (we do
not
expect to invent new virtual methods during the optimization)
The graph is dumped into new IPA dump file
On 16 August 2013 16:28, Tom Tromey wrote:
Phil == Phil Muldoon pmuld...@redhat.com writes:
Phil Anyway, I have regenerated the patch with the fixes requested.
Thanks.
Phil 2013-08-16 Phil Muldoon pmuld...@redhat.com
Phil PR gcc/53477
I think this should say PR libstdc++/53477
I believe you can get derived objects of derived type not visible in the tu,
so the list of targets does only contain local methods and implicitly all
methods only defined in other tus? That is, you cannot rely no that list to
be complete?
Yes, the list is in majority cases not complete.
On Thu, Aug 08, 2013 at 09:15:54PM +0200, Aldy Hernandez wrote:
On 08/08/13 18:42, Richard Henderson wrote:
On 08/08/2013 02:06 AM, Aldy Hernandez wrote:
The hash is not really mapping the simd DECL to the simduid, since that's
just
a matter of DECL_UID(simduid), but the OMP simd array to
On Aug 17, 2013, at 8:54 AM, Jan Hubicka hubi...@ucw.cz wrote:
Moreover objc apparently never produce any virtual functions/methods.
Objective-C++ might. :-)
Sure, those ought to be regular C++ methods though.
Can someone explain me in greater detail how the objc use works?
Hi,
here is variant of patch that drops the field walking from
gimple_extract_devirt_binfo_from_cst completely. As pointed out
by Jason, it is pointless since all structures have BINFO in C++
and thus get_binfo_at_offset will do the job.
I would like to return the code back eventually to handle
Ping
2013/8/12 Ilya Enkovich enkovich@gmail.com:
2013/8/10 Joseph S. Myers jos...@codesourcery.com:
On Mon, 29 Jul 2013, Ilya Enkovich wrote:
Hi,
Here is updated version of the patch. I removed redundant
mode_for_bound, added comments to BOUND_TYPE and added -mmpx option.
I also fixed
On Thu, Aug 01, 2013 at 09:38:31AM -1000, Richard Henderson wrote:
+ if (simd
+ /*
+ || (fd-sched_kind == OMP_CLAUSE_SCHEDULE_STATIC
+ !fd-have_ordered)*/)
Debugging leftovers or what?
gomp-4_0-branch contains also the || there, but for the trunk merge
I
Jan Hubicka hubi...@ucw.cz wrote:
On Aug 17, 2013, at 8:54 AM, Jan Hubicka hubi...@ucw.cz wrote:
Moreover objc apparently never produce any virtual
functions/methods.
Objective-C++ might. :-)
Sure, those ought to be regular C++ methods though.
Can someone explain me in greater detail
Before this patch, it causes the executor into infinite loop.
Tested under x86_64. I'll test -m32 and check-debug before committing.
Thanks!
--
Tim Shen
changelog
Description: Binary data
emptygroup.patch
Description: Binary data
On Mon, Aug 19, 2013 at 06:07:10PM +0800, Tim Shen wrote:
Before this patch, it causes the executor into infinite loop.
Tested under x86_64. I'll test -m32 and check-debug before committing.
@@ -2371,8 +2372,10 @@
if (__re._M_automaton == nullptr)
return false;
for (auto
Hello Jan,
Just some casual notes.
Jan Hubicka hubi...@ucw.cz a écrit:
[...]
Index: ipa-devirt.c
===
[...]
+/* Brief vocalburary:
[...]
+ OTR = OBJ_TYPE_REF
+ This is Gimple representation of type information of
On 08/17/2013 12:29 AM, Caroline Tice wrote:
OK, I *think* I have done as you requested. I have to try the
environment variable before falling back on stderr (there's a program
we want to use this on that disables the ability to write to stderr).
I have added the secure_getenv stuff as you
All,
The attached patch removes the setting of MULTILIB_DEFAULTS for
arm*-*-linux-gnueabi* targets.
The current setting of MULTILIB_DEFAULTS includes mfloat-abi=hard,
which for arm*-*-linux-gnueabi is not true. This makes generating a
hard-float multilib impossible in this configuration.
In an
Hi,
was having a look to c++/58187 (looks like we are missing a lookup) and
noticed this. It seems obvious to me, I'll commit it later today barring
objections.
Tested x86_64-linux.
Thanks,
Paolo.
2013-08-19 Paolo Carlini paolo.carl...@oracle.com
*
1. In config/gnu-user.h you define GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC to
the main part of GNU_USER_TARGET_LIB_SPEC. GNU_USER_TARGET_LIB_SPEC now
becomes '%{pthread:-lpthread} GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC'.
2. In occurrences of LINUX_OR_ANDROID_LD you continue to use
Hi all,
here is a small patch which does some cleanup to avoid an ICE on
invalid SELECT TYPE statements.
The first three hunks are just cosmetics, and the fourth one also
contains minor refactoring, where I pull some common code out of the
two branches of an if statement. The important part,
Committed as r201835 ...
2013/8/18 Janus Weil ja...@gcc.gnu.org:
Hi all,
here is a pretty-much-trivial patch for a problem with OpenMP and
procedure pointers (proc-ptrs to functions are working, but not
subroutines).
Regtested on x86_64-unknown-linux-gnu. Will commit as obvious tomorrow
On keeps_null_pointer_checks, relative to other targets, we see some NULLs in
sets replaced with/merged into NONLOCAL.
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Ok to apply?
2013-08-14 Joern Rennecke joern.renne...@embecosm.com
*
We need a minimum branch cost of 2 to make the expected optimization happen.
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Ok to apply?
2013-08-16 Joern Rennecke joern.renne...@embecosm.com
* gcc.dg/tree-ssa/vrp87.c [avr-*-*]
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Ok to apply?
2013-08-16 Joern Rennecke joern.renne...@embecosm.com
* gcc.dg/tree-ssa/ssa-dom-thread-4.c [avr-*-*]: Expect 6 times
Threaded.
Index: gcc.dg/tree-ssa/ssa-dom-thread-4.c
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Ok to apply?
2013-08-17 Joern Rennecke joern.renne...@embecosm.com
* gcc.dg/tree-ssa/vrp55.c: Use keeps_null_pointer_checks to determine
correct test response.
Index: gcc.dg/tree-ssa/vrp55.c
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Committed as obvious.
2013-08-17 Joern Rennecke joern.renne...@embecosm.com
* gcc.target/avr/progmem-error-1.cpp: Update linenumber of error.
Index: gcc.target/avr/progmem-error-1.cpp
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Ok to apply?
2013-08-18 Joern Rennecke joern.renne...@embecosm.com
PR testsuite/52641
* gcc.dg/tree-ssa/pr31261.c [int16]: Change expected unsigned type.
* gcc.dg/tree-ssa/ssa-pre-21.c
Can you try and add
pragma Weak_External (Var);
on line 8 and see whether it passes? If so, you can commit the patchlet.
The test still fails with the following change:
--- /opt/gcc/_clean/gcc/testsuite/gnat.dg/specs/linker_alias.ads
2013-08-18 17:39:22.0 +0200
+++
Hello Jan,
Just some casual notes.
Thank you! I will fix the typos shortly!
+static inline bool
+polymorphic_type_binfo_p (tree binfo)
+{
+ /* See if BINFO's type has an virtual table associtated with it. */
+ return BINFO_VTABLE (TYPE_BINFO (BINFO_TYPE (binfo)));
Just for my
This is OK, with function_gnu removed (nothing appears to use it), if no
OS port maintainers object to the changes for their OSes within the next
week.
Hello,
Week is over.
Comitted to MT: http://gcc.gnu.org/ml/gcc-cvs/2013-08/msg00447.html
--
Thanks, K
On Mon, Aug 19, 2013 at 12:26 AM, Alan Modra amo...@gmail.com wrote:
When I made the following change
-#define FIRST_SAVED_GP_REGNO 13
+#define FIRST_SAVED_GP_REGNO (FIXED_R13 ? 14 : 13)
in http://gcc.gnu.org/ml/gcc-patches/2012-04/msg01274.html, I checked
all uses of
On Fri, Aug 16, 2013 at 6:32 PM, Peter Bergner berg...@vnet.ibm.com wrote:
Ok, updated to switch the order of the alternatives. This works...just
like the previous one. I created two versions of the TD test case to
test both alternatives and to make sure we get fmrs on the one and
no fmrs
Hi,
also wondered if we want to do something like this?? And the name of the
helper function?
Thanks,
Paolo.
/
Index: name-lookup.c
===
--- name-lookup.c (revision 201835)
+++ name-lookup.c (working
On Sat, Aug 17, 2013 at 1:44 PM, Jan Hubicka hubi...@ucw.cz wrote:
I added both of these and ran into issues due to profile maintenance.
For example, there were non-zero blocks in the cold section because
pro_and_epilogue split a simple return block that was previously reach
by both hot and
On 08/18/2013 07:42 PM, Iyer, Balaji V wrote:
On 08/16/2013 02:13 PM, Iyer, Balaji V wrote:
+ /* If it is a built-in array notation function, then the return type of
+ the function is the type of the array passed in as array notation. */
Ah, then the comment should say ...is the
That looks good.
Jason
Hi,
this patch makes inheritance graph builder to work on LTO. Nothing but
ODR violation warnings and dump file is produced, yet.
The main change is to make types_same_for_odr and ODR hasher to use
assembler name of virtual tables of the type (if present) to establish type
equality. This is
On Mon, Aug 19, 2013 at 6:17 PM, Jakub Jelinek ja...@redhat.com wrote:
Doesn't everything in between the last added line above and the first added
line below need reindenting by 2 spaces (plus of course transforming any 8
consecutive spaces into tabs)?
Thanks for correcting the indentation :)
-Original Message-
From: Jason Merrill [mailto:ja...@redhat.com]
Sent: Monday, August 19, 2013 9:51 AM
To: Iyer, Balaji V; Rainer Orth
Cc: Jakub Jelinek; gcc-patches@gcc.gnu.org; Marek Polacek
(pola...@redhat.com)
Subject: Re: [PATCH] Fix for PR c/57490
On 08/18/2013 07:42 PM,
On Mon, Aug 19, 2013 at 6:04 AM, Paolo Carlini paolo.carl...@oracle.com wrote:
Hi,
was having a look to c++/58187 (looks like we are missing a lookup) and
noticed this. It seems obvious to me, I'll commit it later today barring
objections.
OK.
Tested x86_64-linux.
Thanks,
Paolo.
On 08/19/2013 10:06 AM, Iyer, Balaji V wrote:
Well, it is described in cilkplus.def. The return type of it changes based on
the array that is passed in. So, it is given a fake type. Thus, we need to fix
it up here.
Right, but it should be fixed up when the CALL_EXPR is created, rather
than
Here's the correct version of changelog.
--
Tim Shen
changelog
Description: Binary data
Hi,
On 08/19/2013 12:07 PM, Tim Shen wrote:
+// 2013-08-08 Tim Shentimshe...@gmail.com
+//
+// Copyright (C) 2013 Free Software Foundation, Inc.
Before committing, please adjust the date.
Thanks,
Paolo.
-Original Message-
From: Jason Merrill [mailto:ja...@redhat.com]
Sent: Monday, August 19, 2013 10:20 AM
To: Iyer, Balaji V; Rainer Orth
Cc: Jakub Jelinek; gcc-patches@gcc.gnu.org; Marek Polacek
(pola...@redhat.com)
Subject: Re: [PATCH] Fix for PR c/57490
On 08/19/2013 10:06 AM,
On 08/19/2013 10:37 AM, Iyer, Balaji V wrote:
I just want to make sure I get what you are saying. Are you suggesting
that I do this in finish_call_expr() instead of cp_build_binary_op() ?
I think build_cxx_call is the right place.
Jason
On 08/19/2013 05:56 AM, Joern Rennecke wrote:
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Ok to apply?
20130819-4
2013-08-17 Joern Renneckejoern.renne...@embecosm.com
* gcc.dg/tree-ssa/vrp55.c: Use keeps_null_pointer_checks to determine
On 08/19/2013 05:53 AM, Joern Rennecke wrote:
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Ok to apply?
20130819-3
2013-08-16 Joern Renneckejoern.renne...@embecosm.com
* gcc.dg/tree-ssa/ssa-dom-thread-4.c [avr-*-*]: Expect 6 times
On 08/19/2013 06:05 AM, Joern Rennecke wrote:
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Ok to apply?
20130819-6
2013-08-18 Joern Renneckejoern.renne...@embecosm.com
PR testsuite/52641
* gcc.dg/tree-ssa/pr31261.c [int16]: Change
Remember it isn't using dominance anymore. The latest patch was
instead ensuring the most frequent path between hot blocks and the
entry/exit are marked hot. That should be better than the dominance
approach used in the earlier version.
Indeed, that looks more resonable approach.
Can you
-- Forwarded message --
From: Tim Shen timshe...@gmail.com
Date: Mon, Aug 19, 2013 at 10:26 PM
Subject: Re: [regex] New enum type syntax_option_type
To: Daniel Krügler daniel.krueg...@gmail.com
Cc: libstdc++ libstd...@gcc.gnu.org
On Sun, Aug 18, 2013 at 8:50 PM, Daniel Krügler
On Sat, Aug 17, 2013 at 1:44 PM, Jan Hubicka hubi...@ucw.cz wrote:
patch for updating counts based on estimated frequencies to address
inlined comdats with 0 profile counts:
013-08-16 Teresa Johnson tejohn...@google.com
* tree-inline.c (copy_bb): Compute count based on frequency.
On 15/08/13 15:10, Charles Baylis wrote:
Hi
The attached patch fixes some tests which fail when testing gcc for a
arm-none-linux-gnueabihf target because they do not expect to be built
with a hard float ABI.
The change in target-supports.exp fixes arm-fp16-ops-5.c and arm-fp16-ops-6.c.
-Original Message-
From: Joern Rennecke [mailto:joern.renne...@embecosm.com]
Sent: Monday, August 19, 2013 5:53 AM
To: gcc-patches@gcc.gnu.org
Cc: Denis Chertykov; Anatoly Sokolov; Weddington, Eric
Subject: RFA: testsuite patches (3/6): [avr]: ssa-dom-thread-4.c:
expect 6 times
-Original Message-
From: Joern Rennecke [mailto:joern.renne...@embecosm.com]
Sent: Monday, August 19, 2013 5:49 AM
To: gcc-patches@gcc.gnu.org
Cc: Denis Chertykov; Anatoly Sokolov; Weddington, Eric
Subject: RFA: testsuite patches (2/6): [avr]: Set required branch cost
for
-Original Message-
From: Joern Rennecke [mailto:joern.renne...@embecosm.com]
Sent: Monday, August 19, 2013 6:02 AM
To: gcc-patches@gcc.gnu.org
Cc: Denis Chertykov; Anatoly Sokolov; Weddington, Eric
Subject: Committed: testsuite patches (5/6): Update error line number
in
This patch emits linkage_name at -gmlt. It also make sure abstract
decls' linkage_names are emitted so that inlined functions can also
find linkage name.
Bootstrapped and passed regression test.
OK for google branches?
OK.
-cary
Hi,
this patchlet adjusts the DECL_NAME of the cloned functions created by
cgraph_create_virtual_clone. It's currently set to a simple copy of the
DECL_ASSEMBLER_NAME, which is mildly annoying because it comprises weird
platform-independent tweaks, which results in a different output in error
- expanded_location s = expand_location (BLOCK_SOURCE_LOCATION (stmt));
+ location_t locus = BLOCK_SOURCE_LOCATION (stmt);
+ expanded_location s = expand_location (locus);
if (dwarf_version = 3 || !dwarf_strict)
{
add_AT_file (die, DW_AT_call_file, lookup_filename
On 08/17/2013 12:29 AM, Caroline Tice wrote:
OK, I *think* I have done as you requested. I have to try the
environment variable before falling back on stderr (there's a program
we want to use this on that disables the ability to write to stderr).
I have added the secure_getenv stuff as you
-Original Message-
From: Jason Merrill [mailto:ja...@redhat.com]
Sent: Monday, August 19, 2013 10:46 AM
To: Iyer, Balaji V; Rainer Orth
Cc: Jakub Jelinek; gcc-patches@gcc.gnu.org; Marek Polacek
(pola...@redhat.com)
Subject: Re: [PATCH] Fix for PR c/57490
On 08/19/2013 10:37 AM,
This patchlet adjusts the DECL_NAME of the cloned functions created by
cgraph_create_virtual_clone. It's currently set to a simple copy of the
DECL_ASSEMBLER_NAME, which is mildly annoying because it comprises weird
platform-independent tweaks, which results in a different output in error
2013/8/19 Joern Rennecke joern.renne...@embecosm.com:
We need a minimum branch cost of 2 to make the expected optimization happen.
Tested for avr with --target_board=atmega128-sim and native on
i686-pc-linuc-gnu.
Ok to apply?
2013-08-16 Joern Rennecke joern.renne...@embecosm.com
On 08/18/13 05:21, Gerald Pfeifer wrote:
On Wed, 26 Jun 2013, Aldy Hernandez wrote:
This is a small cleanup to the Cilk Plus mention in our documentation,
but more importantly, it clarifies that the Cilk Plus implementation
in GCC is only partial.
OK for trunk?
* doc/invoke.texi
gcc.dg/pr27531-1.c was failing with rtl checking enabled for -mips16 -mabi=32
on mips64-linux-gnu. The problem was that mips_adjust_insn_length didn't cope
properly with the new(ish) JUMP_TABLE_DATA rtx, which has no insn code.
Tested on mips64-linux-gnu and applied.
Richard
gcc/
*
mulsize-2.c expected a subtraction, but when multiplying by 9 we should
have a shift and addition instead. While there I made mulsize-1.c and
mulsize-2.c check for a shift, like mulsize-3.c does.
Tested on mips64-linux-gnu and applied.
Richard
gcc/testsuite/
*
On Wed, Aug 14, 2013 at 12:03 AM, Uros Bizjak ubiz...@gmail.com wrote:
On Tue, Aug 13, 2013 at 9:42 PM, Jakub Jelinek ja...@redhat.com wrote:
We right now ICE with -mcmodel=large -fpic on x86_64 on TLS GD and LD
sequences, because obviously we can't call __tls_get_addr@plt there from code
On Mon, Aug 19, 2013 at 8:09 AM, Jan Hubicka hubi...@ucw.cz wrote:
Remember it isn't using dominance anymore. The latest patch was
instead ensuring the most frequent path between hot blocks and the
entry/exit are marked hot. That should be better than the dominance
approach used in the earlier
On Aug 19, 2013, at 7:03 AM, Tim Shen timshe...@gmail.com wrote:
So should I later propose a single patch to make every 8 spaces to a
tab, or just make changes for all my committed files in this patch?
One cannot merely replace 8 spaces with a tab, this doesn't work in general.
If you expand
On 08/16/2013 08:33 AM, David Malcolm wrote:
I also tweaked the traversal hooks for opt_pass to emulate chain_next,
since this is where the really deep callchains could otherwise occur.
See the patch for details (given the subtleties I opted to put big
comments in the relevant routines).
On Mon, 2013-08-19 at 09:17 -0400, David Edelsohn wrote:
The last version is okay.
Jakub ack'd the builtins.{c,def] changes on IRC, so I have committed
this as revision 201849.
And is the *negtd2_fpr hunk ok for the 4.8 branch?
The negtd2_fpr change is okay for the 4.8 branch.
...and
On 08/19/2013 12:40 PM, Iyer, Balaji V wrote:
+ /* If it is a built-in array notation function, then the return type of
+ the function is the element type of the array passed in as array
+ notation (i.e. the first parameter of the function). */
+ if (flag_enable_cilkplus TREE_CODE
Ping.
Thanks,
Teresa
On Mon, Aug 12, 2013 at 6:54 AM, Teresa Johnson tejohn...@google.com wrote:
On Tue, Aug 6, 2013 at 10:23 PM, Teresa Johnson tejohn...@google.com wrote:
On Tue, Aug 6, 2013 at 9:29 AM, Teresa Johnson tejohn...@google.com wrote:
On Tue, Aug 6, 2013 at 9:01 AM, Martin Jambor
Ping.
Thanks,
Teresa
On Sun, Aug 11, 2013 at 9:35 PM, Teresa Johnson tejohn...@google.com wrote:
This patch fixes PR rtl-optimizations/57451 by preventing scopes and
therefore lexical blocks from crossing split section boundaries.
This will prevent debug info generation from using
Patch updated. BTW, do you think this patch should actually go into trunk?
Thanks,
Dehao
Index: include/dwarf2.def
===
--- include/dwarf2.def (revision 201852)
+++ include/dwarf2.def (working copy)
@@ -390,6 +390,8 @@ DW_AT
On 08/19/2013 12:34 PM, Teresa Johnson wrote:
Ping.
Thanks,
Teresa
On Sun, Aug 11, 2013 at 9:35 PM, Teresa Johnson tejohn...@google.com wrote:
This patch fixes PR rtl-optimizations/57451 by preventing scopes and
therefore lexical blocks from crossing split section boundaries.
This will prevent
This patch fixes a bug in assigning discrminator. We should explicitly
call the hash function when finding the next discriminator.
Bootstrapped and passed regression tests.
OK for google branches?
OK. Thanks.
-cary
Patch updated. BTW, do you think this patch should actually go into trunk?
This is OK for google branches. I don't think it can go into trunk
like this, as it relies on get_discriminator_from_locus. In trunk, we
only have the discriminator assigned to a basic block.
-cary
Index:
Patch updated. BTW, do you think this patch should actually go into trunk?
This is OK for google branches. I don't think it can go into trunk
like this, as it relies on get_discriminator_from_locus. In trunk, we
only have the discriminator assigned to a basic block.
Oh, but the dwarf2.def
Dear Teresa,
On 19 August 2013 19:47, Teresa Johnson tejohn...@google.com wrote:
On Mon, Aug 19, 2013 at 8:09 AM, Jan Hubicka hubi...@ucw.cz wrote:
Remember it isn't using dominance anymore. The latest patch was
instead ensuring the most frequent path between hot blocks and the
entry/exit are
On 08/02/2013 07:48 PM, David Malcolm wrote:
GDB 7.0 onwards supports hooks written in Python to improve the
quality-of-life within the debugger. The best known are the
pretty-printing hooks [1], which we already use within libstdc++ for
printing better representations of STL containers.
So as
On 19/08/2013, at 11:09 PM, Pavel Chupin wrote:
...
Thanks. I'm OK with your suggestions.
Please see adjusted patch attached.
All good, thanks!
--
Maxim Kuvyrkov
www.kugelworks.com
On 07/30/2013 09:24 PM, DJ Delorie wrote:
[nickc added for comments about the bits he wrote]
... define these as
(define_predicate msp_general_operand
(match_code mem,reg,subreg,const_int,const,symbol_ref,label_ref
{
int save_volatile_ok = volatile_ok;
volatile_ok = 1;
int ret =
After rerunning test, this will fail one gcc regression test. So I
updated the patch to make sure all test will pass:
Index: gcc/dwarf2out.c
===
--- gcc/dwarf2out.c (revision 201850)
+++ gcc/dwarf2out.c (working copy)
@@ -16545,10
On 08/19/2013 02:32 PM, Jeff Law wrote:
Ok, I think I got it all covered... ended up copying most of the
pushm/popm from rx, added return but not simple_return, and merged
added a peephole for and-bic.
[ ... ignore ... hit send in the wrong window ]
On 08/14/2013 02:55 PM, DJ Delorie wrote:
+#endif /* GCC_MSP430_PROTOS_H */
Index: gcc/config/msp430/predicates.md
===
--- gcc/config/msp430/predicates.md (revision 0)
+++ gcc/config/msp430/predicates.md (revision 0)
+
I'd say it's not as simple as you make it out to be. You can't blindly
combine operations on volatile memory.
I'm not blindly combining them, I'm combining them when I know the
hardware will do the volatile-correct thing.
ie, the programmer may have written them as separate statements and
On 08/19/2013 02:49 PM, DJ Delorie wrote:
I'd say it's not as simple as you make it out to be. You can't blindly
combine operations on volatile memory.
I'm not blindly combining them, I'm combining them when I know the
hardware will do the volatile-correct thing.
You're missing the point. If
On 08/07/2013 01:07 PM, Kirill Yukhin wrote:
+ -1, -1, -1, -1, -1, -1, -1, -1, /* new SSE registers 16-23*/
+ -1, -1, -1, -1, -1, -1, -1, -1, /* new SSE registers 24-31*/
Don't say new, say avx512 -- that comment will be there for 10 years.
@@ -4080,6 +4111,8 @@
On 07/20/2013 03:02 AM, Alexey Makhalov wrote:
Hello!
If delete_unmarked_insns deletes some insn, DF state might be
out of date, and, regs_ever_live might contain unused registers till the end.
Fixed by forcing regs_ever_live update and rerunning df_analyze () at
fini_dce().
I found this bug
On 08/14/2013 12:23 AM, Kirill Yukhin wrote:
+ ;; For AVX512F mask support
+ UNSPEC_KIOR
+ UNSPEC_KXOR
+ UNSPEC_KAND
+ UNSPEC_KANDN
I thought we determined that you didn't need these,
that *Yk as a constraint was sufficient.
+(define_insn kandnmode
+(define_insn kandmode
You're missing the point. If the programmer wrote two statements
which hit volatile memory and you've got some pattern which matches
those two statements, then with your change you'll end up combining
them, that's wrong.
I see nothing in the ISO spec that says you can't combine a volatile
register int a;
extern volatile int b;
a = b;
a |= 0x54;
b = a;
The ISO spec seems to allow gcc to perform those operations in a
single physical insn, as long as the operations on 'b' are all
performed, and in the correct sequence.
And I believe it's also
On Mon, Aug 19, 2013 at 9:56 PM, Jeff Law l...@redhat.com wrote:
On 08/02/2013 07:48 PM, David Malcolm wrote:
GDB 7.0 onwards supports hooks written in Python to improve the
quality-of-life within the debugger. The best known are the
pretty-printing hooks [1], which we already use within
On Mon, 2013-08-19 at 16:49 -0400, DJ Delorie wrote:
I'd say it's not as simple as you make it out to be. You can't blindly
combine operations on volatile memory.
I'm not blindly combining them, I'm combining them when I know the
hardware will do the volatile-correct thing.
ie, the
On 08/19/2013 03:28 PM, Steven Bosscher wrote:
I assume you mean basic blocks and edges? For that, you can already
use the CFG pretty printers with a file attached to a pipe and sent
through XDot (http://code.google.com/p/jrfonseca/wiki/XDot) which is
written in Python and quite easy to
;; All vector modes including V?TImode, used in move patterns.
(define_mode_iterator V16
- [(V32QI TARGET_AVX) V16QI
- (V16HI TARGET_AVX) V8HI
- (V8SI TARGET_AVX) V4SI
- (V4DI TARGET_AVX) V2DI
+ [(V64QI TARGET_AVX512F) (V32QI TARGET_AVX) V16QI
+ (V32HI TARGET_AVX512F) (V16HI
On 08/14/2013 12:26 AM, Kirill Yukhin wrote:
Hello,
Patch was rebased on top of trunk.
It is applicable on top of [2/8] (which was rebased on new trunk today).
Testing:
1. Bootstrap pass.
2. make check shows no regressions.
3. Spec 2000 2006 build show no regressions both with
This patch assigns discriminators for different callsites within the
same BB. This is needed for accurate profile attribution in AutoFDO.
Testing on going.
OK for google branches if test pass?
Thanks,
Dehao
Index: gcc/tree-cfg.c
Hi, Richard,
Could you take a second look at this patch?
Thanks,
Dehao
On Mon, Jun 24, 2013 at 3:45 PM, Dehao Chen de...@google.com wrote:
The original patch has some flaw. The new patch is attached.
Bootstrapped and passed regression tests.
Thanks,
Dehao
On Mon, Jun 24, 2013 at 9:44 AM,
@@ -960,6 +960,7 @@ SCEV_H = tree-scalar-evolution.h $(GGC_H) tree-chrec.h
$(PARAMS_H)
OMEGA_H = omega.h $(PARAMS_H)
TREE_DATA_REF_H = tree-data-ref.h $(OMEGA_H) graphds.h $(SCEV_H)
TREE_INLINE_H = tree-inline.h
+CILK_H = cilk.h
REAL_H = real.h $(MACHMODE_H)
IRA_INT_H = ira.h ira-int.h
Hi All,
I could really use some help here from someone who has a better
understanding of how the config/Makefile system works than I do.
In my libvtv/configure.ac file, I have:
AC_GNU_SOURCE
AC_CHECK_FUNCS([__secure_getenv])
AC_GNU_SOURCE
AC_CHECK_FUNCS([secure_getenv])
This gets translated
On Feb 21, 2013, at 1:35 PM, Jakub Jelinek ja...@redhat.com wrote:
I've committed the following fix for the following testcase.
When scalar_op1 is 0x8000 with 64-bit HWI,
it matches EXACT_POWER_OF_2_OR_ZERO_P, but we should expand it as
negation of the 63 shift
1 - 100 of 112 matches
Mail list logo