This fixes the regression in the frame unwinding support on SPARC/Solaris 11
present on the mainline and 4.6 branch. Tested (by me) on 6 differents OS
versions of Solaris 8, 9 and 10 and on Solaris 11 by Ian and Rainer.
Applied on the mainline and 4.6 branch.
2012-02-15 Eric Botcazou
This is a regression present on the mainline: the compiler aborts on the access
to a component of the value returned by a function call, if the return type is
a volatile record type that isn't controlled or tagged.
Tested on i586-suse-linux, applied on the mainline.
2012-02-15 Eric Botcazou
On Tue, Feb 14, 2012 at 8:52 PM, Richard Henderson r...@redhat.com wrote:
On 02/14/2012 10:26 AM, Uros Bizjak wrote:
#ifdef __x86_64__
+ cfi_def_cfa(%rsi, 0)
movq (%rsi), %rcx
movq 8(%rsi), %rbx
movq 16(%rsi), %rbp
@@ -119,20 +120,21 @@
movq
This is a regression present on the mainline and which stems from enabling
-feliminate-unused-debug-types by default in Ada. We fail to emit the
debugging information for tagged types when only the associated class-wide
type is referenced in the source code.
Tested on i586-suse-linux, applied
This is a regression present on the mainline: the compiler aborts on a deferred
constant initialized with an aggregate and whose type is a discriminated
record type with a variant part and which, ultimately, contains a component
whose type is controlled or tagged.
Tested on i586-suse-linux,
On Wed, 15 Feb 2012, Eric Botcazou wrote:
This fixes PR52178, the failure to bootstrap Ada with LTO (well,
until you hit the next problem). A self-referential DECL_QUALIFIER
makes us think that a QUAL_UNION_TYPE type is of variable-size which
makes us stream that type locally, wrecking
Hello,
This patch fixes a regression when folding a cast cond expression to a
constant in case label.
The problem came from the removal of the lines in
c-common.c:check_case_value
/* ??? Can we ever get nops here for a valid case value? We
shouldn't for C. */
STRIP_TYPE_NOPS (value);
On Tue, Feb 14, 2012 at 6:47 PM, Richard Henderson r...@redhat.com wrote:
On 02/14/2012 08:46 AM, Aldy Hernandez wrote:
The call to ipa_tm_diagnose_tm_safe() does nothing because there are no
longer any calls in the function, since the function call has been inlined:
f ()
{
bb 2:
On 14/02/12 14:10, Jakub Jelinek wrote:
On Tue, Feb 14, 2012 at 02:05:30PM +, Paulo J. Matos wrote:
I think the register allocator will generate good code using % if you
make your predicate nonimmediate_operand in operand 1:
(define_insn iorqi3
[(set (match_operand:QI 0 register_operand
On Wed, Feb 15, 2012 at 4:47 AM, Kenny Simpson
theonetrueke...@yahoo.com wrote:
Are any other recent trunk fixes to be backported - like the wrong-code bugs:
PR tree-optimization/50444, PR tree-optimization/51528, or PR
tree-optimization/46886 ?
I'll backport 46886 but like to defer 50444
On Wed, Feb 15, 2012 at 10:04:15AM +, Paulo J. Matos wrote:
It doesn't look like you can actually do that because, according to
the internals manual:
For a named pattern, the condition (if present) may not depend on
the data in the insn being matched, but only the target-machine-type
On Wed, Feb 15, 2012 at 8:23 AM, Ye Joey joey.ye...@gmail.com wrote:
Ping^2
All hunks apart from the expr.c hunk from 182545 are ok (thus, all hunks
that are explicitly guarded with -fstrict-volatile-bitfields).
Please use -p for diffs, I couldn't spot the place in expr.c you patch,
the
I noticed when working on GAS for darwin that we fail to prepend the
user label prefix for extended identifiers.
fixed as below,
OK for trunk/when?
Iain
gcc:
* config/darwin.h (ASM_OUTPUT_LABELREF): Add user label prefix
for extended identifiers.
diff --git
On 02/15/2012 01:14 PM, Richard Guenther wrote:
This patch by Tobias fixes PR50561.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Thanks Richard.
Tobi
This patch causes all tm tests to fail on arm-none-eabi target that doesn't
support -pthread command line option:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52259
-Original Message-
From: Eric Botcazou [mailto:ebotca...@adacore.com]
Sent: 11 February 2012 19:24
To: Jack Howarth
Cc:
On 02/15/12 03:48, Richard Guenther wrote:
On Tue, Feb 14, 2012 at 6:47 PM, Richard Hendersonr...@redhat.com wrote:
On 02/14/2012 08:46 AM, Aldy Hernandez wrote:
The call to ipa_tm_diagnose_tm_safe() does nothing because there are no longer
any calls in the function, since the function call
This patch to the Go frontend avoids some compiler crashes on erroneous
inputs. It also makes some minor fixes to some comments. Bootstrapped
and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to
mainline.
Ian
diff -r 4f7eddf7a851 go/backend.h
--- a/go/backend.h Tue Feb 14 12:39:26
Georg-Johann Lay wrote:
Jakub Jelinek wrote:
On Wed, Feb 01, 2012 at 02:00:53PM +0100, Georg-Johann Lay wrote:
* gcc.dg/lto/20091013-1_1.c: xfail for avr.
* gcc.dg/lto/20091013-1_2.c: xfail for avr.
If it is just because of the warnings, can't you
add
-Wno-int-to-pointer-cast
On Wed, Feb 15, 2012 at 3:30 PM, Aldy Hernandez al...@redhat.com wrote:
On 02/15/12 03:48, Richard Guenther wrote:
On Tue, Feb 14, 2012 at 6:47 PM, Richard Hendersonr...@redhat.com wrote:
On 02/14/2012 08:46 AM, Aldy Hernandez wrote:
The call to ipa_tm_diagnose_tm_safe() does nothing
Some parts outside the backend use ACCUMULATE_OUTGOING_ARGS without including
tm_p.h which lead to a build warning because avr.h does
#define ACCUMULATE_OUTGOING_ARGS avr_accumulate_outgoing_args()
This patcs moved the prototype from avr-protos.h to avr.h:
* config/avr/avr-protos.h
[rth, please correct me if I have misrepresented the TM semantics and
rules].
On 02/15/12 09:04, Richard Guenther wrote:
Ok, I had a look at the bugreport and am curios why
__attribute__((always_inline))
static void asmfunc(void)
{
__asm__ ();
}
__attribute__((transaction_safe))
static
On Wed, 15 Feb 2012, Christian Bruel wrote:
Removal of this NOP_EXPR if !CAN_HAVE_LOCATION_P fixes the problem.
looks safe from my testing, because the loc is inserted using
'protected_set_expr_location', whereas no loc for a constant case
doesn't seem to introduce new problems with the
On 02/15/2012 06:09 PM, Rainer Orth wrote:
Paolo Bonzini bonz...@gnu.org writes:
This must be a separate macro in acinclude.m4 that is AC_REQUIREd from
gcc_AC_INITFINI_ARRAY.
Here's an updated patch that does so. During testing, I found one
additional complication, though. With the Sun
On Feb 15, 2012, at 2:21 AM, Iain Sandoe wrote:
I noticed when working on GAS for darwin that we fail to prepend the user
label prefix for extended identifiers.
Does this match the native compiler on the platform? This is abi changing, and
we can't do it, unless it mismatches the native
On 02/15/2012 01:48 AM, Richard Guenther wrote:
Hmm. I think you rather want to teach local_pure_const about TM
properties you want to know and have them propagated properly
(of course unless pure/const which is about optimization and has an
easy fallback default yours wouldn't have that -
On Wed, Feb 15, 2012 at 09:29:18AM -0800, Richard Henderson wrote:
On 02/15/2012 01:48 AM, Richard Guenther wrote:
Hmm. I think you rather want to teach local_pure_const about TM
properties you want to know and have them propagated properly
(of course unless pure/const which is about
Prompted by the suggestions in the PR what to do about the new exports
of
_ZNSt19istreambuf_iteratorIcSt11char_traitsIcEEppEv
_ZNSt19istreambuf_iteratorIwSt11char_traitsIwEEppEv
at GLIBCXX_3.4.5 on Solaris, I've followed the preprocess *.ver files
route Jakub suggested to just omit those two on
On 15 Feb 2012, at 17:16, Mike Stump wrote:
On Feb 15, 2012, at 2:21 AM, Iain Sandoe wrote:
I noticed when working on GAS for darwin that we fail to prepend
the user label prefix for extended identifiers.
Does this match the native compiler on the platform? This is abi
changing, and we
On 02/14/2012 12:00 PM, Ulrich Weigand wrote:
Hello,
compiling anything, even just an empty function, with
-fschedule-insns -fsched-pressure -fdump-rtl-sched1
causes the compiler to crash:
Program received signal SIGSEGV, Segmentation fault.
0x0840d9e0 in REG_N_REFS (regno=58) at
Vladimir Makarov wrote:
On 02/14/2012 12:00 PM, Ulrich Weigand wrote:
This is fixed by the following patch. However, I'm not quite sure if this
is
the best way to fix the problem, given that the REG_N_REFS use occurs only
for
debug output purposes -- should we actually do the work to
Like other MIPS ELF targets, the mips-sgi-irix6.5 configuration doesn't
include dbxelf.h, causing objc.dg/stabs-1.m to FAIL. Same cure as in
Richard's patch.
Tested with the appropriate runtest invocations, installed on mainline.
Rainer
2012-02-11 Rainer Orth
The libgo setcontext TLS configure test was failing on Solaris 8 and
9/SPARC, breaking bootstrap due to the unhandled case in
runtime/proc.c. This turned out to be a known issue, documented in the
makecontext(3C) manpage:
NOTES
The semantics of the uc_stack member of the ucontext_t
On Wed, Feb 15, 2012 at 11:59:15AM -0600, Aldy Hernandez wrote:
Hmmm, isn't %K for trees? We're talking gimple, and FUNCTION_DECLs,
which don't have a TREE_BLOCK, here.
Gimple stmts have gimple_block, I guess you'd need to create some tree
and set its TREE_BLOCK and EXPR_LOCATION from
This patch fixes PR 52199, which shows up on power7 systems when the compiler
is vectorizing DImode operations, and the vector duplicate code was not
prepared to accept CONST and SYMBOL_REF's.
I have done a bootstrap and make compare, and there were no regressions. Is
this ok to install? I will
Hello,
While going through some old patches of mine, I noticed this:
config/rl78/rl78.h:#define SMALL_REGISTER_CLASSES 1
config/rx/rx.h:#define SMALL_REGISTER_CLASSES 0
Can't possibly work, that target macro was replaced with a hook.
Attached patch OK for trunk?
Ciao!
On Wed, Feb 15, 2012 at 7:36 PM, Steven Bosscher stevenb@gmail.com wrote:
Hello,
While going through some old patches of mine, I noticed this:
config/rl78/rl78.h:#define SMALL_REGISTER_CLASSES 1
config/rx/rx.h:#define SMALL_REGISTER_CLASSES 0
Can't possibly
I would classify these as obvious but please make sure they go in
before the poison patch ;-)
And this will be needed to make those ports do what was intended:
* config/rl78/rl78.h: Replace SMALL_REGISTER_CLASSES with hook.
* config/rx/rx.h: Remove SMALL_REGISTER_CLASSES.
Alexandre Oliva aol...@redhat.com writes:
On Feb 13, 2012, Richard Sandiford rdsandif...@googlemail.com wrote:
does this avoid the kind of memrefs_conflict_p cycle I was seeing in:
I don't know that it does, I'd missed that bit.
If you still have a preprocessed testcase, I'd be glad to give
On Wed, Feb 15, 2012 at 7:59 PM, DJ Delorie d...@redhat.com wrote:
I would classify these as obvious but please make sure they go in
before the poison patch ;-)
And this will be needed to make those ports do what was intended:
* config/rl78/rl78.h: Replace SMALL_REGISTER_CLASSES with
On Feb 15, 2012, at 9:38 AM, Iain Sandoe wrote:
So, this just brings us into line with the current system compiler.
Ah, the patch is ok. Perfect, thanks.
Here the committed patch approved off-list by Mike Stump which fixed
PR/52042.
Tested on darwin10/11 with x86_64.
PS: this is my first commit so I hope I get it right. Otherwise do not
hesitate to yell at me.
Thanks.
--
Patrick.
2012-02-15 Iain Sandoe ia...@gcc.gnu.org
Patrick
The typo is obvious, but whether or not the wrapped code still works
isn't (to me).
* src/c++98/locale.cc (locale::facet::_S_get_c_locale): Fix
typo.
Please check this in.
-benjamin
This patch introduces a new file builtins.def that is used as central registry
to hold built-ins' information.
The file is used by defining DEF_BUILTIN macre and then including the file as
described in the head comment of builtins.def.
Up to here it's all code clean-up and no functional change.
The typo is obvious, but whether or not the wrapped code still works
isn't (to me).
* src/c++98/locale.cc (locale::facet::_S_get_c_locale): Fix
typo.
Please check this in.
Done. Thanks!
On Feb 15, 2012, at 11:09 AM, Steven Bosscher wrote:
On Wed, Feb 15, 2012 at 7:59 PM, DJ Delorie d...@redhat.com wrote:
I would classify these as obvious but please make sure they go in
before the poison patch ;-)
And this will be needed to make those ports do what was intended:
*
Hello Everyone,
This patch is for the Cilkplus branch affecting the C and C++ compilers.
This patch will fix a problem in metadata dwarf code generation part when we
are trying to catch the register number when a spill occurs.
Thanks,
Balaji V. Iyer.
diff --git a/gcc/ChangeLog.cilk
Rainer Orth r...@cebitec.uni-bielefeld.de writes:
2012-02-13 Rainer Orth r...@cebitec.uni-bielefeld.de
* configure.ac (libgo_cv_lib_makecontext_stack_top): Set on
sparc*-*-solaris2.[89]*.
(MAKECONTEXT_STACK_TOP): Define if appropriate.
On Feb 15, 2012, at 11:18 AM, Patrick Marlier wrote:
PS: this is my first commit so I hope I get it right. Otherwise do not
hesitate to yell at me.
Thanks. I know of two people that will test and yell, if anything goes wrong.
I'll spin up a build...
This patch merges inline functions where a plain declaration is
read from a pph file after an earlier pph file provides the full
definition.
Making this merge happen required refactoring the several pph
routines. Ideally, we should have merge versions of the common tree
streamer. However, we
This patch is another installment in my endless(?) series of copy-edits
to the GCC manual. Here, I've fixed up a number of places to uniformly
use e.g. 32-bit int instead of sometimes that and sometimes 32 bit
and sometimes 32bit. And, as a noun phrase, it's properly 32 bits.
Likewise
Hi,
In PR/52220, eh-1.C fails because the dummy function
__cxa_allocate_exception is used instead of the one from libstdc++.
Indeed, the function is defined in libitm and thus is used.
The fix consists to move the dummy function definitions to darwin-crt-tm.c.
This fix was mainly contributed
On 15 Feb 2012, at 23:39, Patrick Marlier wrote:
PR libitm/52220
* alloc_cpp.cc: No dummy definitions for darwin.
* eh_cpp.cc: Likewise.
libgcc/
2012-02-15 Iain Sandoe ia...@gcc.gnu.org
Patrick Marlier patrick.marl...@gmail.com
PR libitm/52220
On 15 February 2012 17:44, Rainer Orth wrote:
Ok for mainline?
Yes, although I would prefer the config option to be
--enable-libstdcxx-gthreads or --enable-libstdcxx-threads, since I
prefer it to be clear that a config option relates just to libstdc++,
and this only seems to affect libstdc++
On Wed, Feb 15, 2012 at 11:48:29PM +, Iain Sandoe wrote:
On 15 Feb 2012, at 23:39, Patrick Marlier wrote:
PR libitm/52220
* alloc_cpp.cc: No dummy definitions for darwin.
* eh_cpp.cc: Likewise.
libgcc/
2012-02-15 Iain Sandoe ia...@gcc.gnu.org
Patrick Marlier
Hi!
Bootstrapped/regtested on x86_64-linux and i686-linux,
committed to trunk as obvious:
2012-02-16 Jakub Jelinek ja...@redhat.com
PR translation/52264
* cgraphunit.c (verify_cgraph_node): Fix a typo.
--- gcc/cgraphunit.c.jj 2012-02-12 15:46:32.0 +0100
+++
The gccgo frontend was crashing on code like
type T1 struct { F1 *[1]T2 }
type T2 struct { X int }
It needed to know the size of T2 before it was computed. This patch
fixes the problem. I added a testcase as bug417.go in the master Go
testsuite, which I will merged into the gccgo testsuite
Hi!
On this testcase we ICE, because slpeel_tree_peel_loop_to_edge is first
called with a loop that has a virtual PHI and no virtual PHI in the loop
exit bb and this function doesn't update the vop properly after inserting
second loop and adding all the conditional guards, then
Hi!
As mentioned in the PR, if SET_SRC MEM (with REG_EQUIV note) isn't a MEM
that can be used in general operations, but requires some specialized
instruction to load it, decreasing its mem_cost results sometimes in
worse code, the MEM is loaded with a specialized insn, then stored into
a stack
Hi!
On the following testcase we ICE, because now that clone_tree_hash
inserts elements into decl_table hash table, if the hash table is expanded
during that, storing into *slot might either store it at a wrong spot, or
in completely unrelated memory.
Fixed by storing into the hash table
Rainer Orth r...@cebitec.uni-bielefeld.de writes:
The patch allowed the bootstrap to complete successfully, but since
libgo.so now has undefined references to log2 and trunc, all Go
execution still fail.
I went ahead and removed the uses of the libc log2 and trunc functions,
on the basis that
60 matches
Mail list logo
