On 01/30/15 05:29, Martin Liška wrote:
Hi.
Attached patch handles situation where remove_unreachable_nodes is
called before
IPA ICF registers removal hooks. As a result, a deleted callgraph node
is touched.
Tested on x86_64-linux-pc without a new regression seen.
Ready for trunk?
Thanks,
Marti
On 02/05/15 10:13, Sandra Loosemore wrote:
It looks to me like that, when the rewrite of the inline asm sections
was committed last May (r210273), there was general agreement that it
needed additional copy-editing to fix markup problems and the like. But,
nobody ever followed up to do that.
Here
On Thu, Feb 5, 2015 at 3:22 AM, Renlin Li wrote:
> Hi all,
>
> location information is missing when rewrite complex component-wise
> load/store.
>
> This patch add the location info to fix the recently gcc.dg/uninit-13.c and
> gcc.dg/uninit-17.c failures on arm target. The compiler warns uninitial
On February 5, 2015 4:42:50 PM CET, Jakub Jelinek wrote:
>Hi!
>
>On the following testcase we fail -fcompare-debug, because in .LASAN0
>initializer we get a NOP_EXPR optimized away only in one of the cases.
>The reason is that cgraph* is looking at DECL_ABSTRACT_P flag of
>__PRETTY_FUNCTION__ and
On Fri, Feb 06, 2015 at 09:44:40AM +0100, Richard Biener wrote:
> Looks good to me. I wonder if this will also help Aldyh...
>
> I also wonder what it takes to make dwarf2out use DECL_ABSTRACT_P ||
> force_abstract with maintaining that new state and whether that would be
> both cheaper and cleane
Hi,
inline is a patch to avoid using /dev/random on Windows in ssp.c. If it
is getting used there might be a local malicious process supplying fake
random values (e.g. via C:\dev\random) rendering SSP useless.
Comments/review are much appreciated. The patch is against the 4.9 branch:
From 372698
I think it is useful to know that one can build libcc1.so - also as
advertisement. Thus, I propose to include something like the quip in
the attachment.
GDB uses it in version 7.9, which is not yet released but already
branched off.
OK? - Other suggestions or comments?
Tobias
Index: htdocs/gcc
On Fri, Feb 06, 2015 at 09:44:20AM +, Tobias Burnus wrote:
> I think it is useful to know that one can build libcc1.so - also as
> advertisement. Thus, I propose to include something like the quip in
> the attachment.
>
> GDB uses it in version 7.9, which is not yet released but already
> bran
On Thu, 5 Feb 2015, Petar Jovanovic wrote:
> > > This is a follow-up to a change [1] in glibc. It fixes the issue [2]
> > > when jal can not reach a target in different region.
> > Is it not a problem that can be solved with rearranging the order of
> sections in output?
>
> Is not this more cle
On Thu, 5 Feb 2015, Matthew Fortune wrote:
> I'm OK with this change but I'd like Catherine to comment before committing.
> It seems a shame to duplicate the block of code but it is probably just as
> ugly to define a macro for the la/dla instruction.
Native systems have for such ABI dependenci
Maciej W. Rozycki writes:
> On Thu, 5 Feb 2015, Matthew Fortune wrote:
>
> > I'm OK with this change but I'd like Catherine to comment before
> committing.
> > It seems a shame to duplicate the block of code but it is probably just
> as
> > ugly to define a macro for the la/dla instruction.
>
>
On Fri, Feb 06, 2015 at 10:44:20AM +0100, Tobias Burnus wrote:
> I think it is useful to know that one can build libcc1.so - also as
> advertisement. Thus, I propose to include something like the quip in
> the attachment.
>
> GDB uses it in version 7.9, which is not yet released but already
> bran
Hi Paul,
thanks for the review.
Committed as r220474.
Regards,
Andre
On Thu, 5 Feb 2015 15:15:02 +0100
Paul Richard Thomas wrote:
> Dear Andre,
>
> That's fine to commit to trunk.
>
> Thanks for the patch
>
> Paul
>
> On 30 January 2015 at 11:19, Andre Vehreschild wrote:
> > Hi P
Hi!
The
/* Given (xor (ior (xor A B) C) D), where B, C and D are
constants, simplify to (xor (ior A C) (B&~C)^D), canceling
out bits inverted twice and not set by C. Similarly, given
(xor (and (xor A B) C) D), simplify without inverting C in
the xor opera
Ping?
On Fri, 30 Jan 2015, Alexander Monakov wrote:
> Hello,
>
> Recently on gcc-help@ one of the users asked whether they need to set
> -fvar-tracking-assignments on the command line by hand. The documentation may
> be clearer in that this flag has a useful value by default, like
> -fvar-trac
Hi,
While trying to force some particular conditions out of the scheduler
on a private branch, I was toying with some very aggressive
macro-fusion patterns. In doing so, I hit on a latent bug which can
cause miscompilation.
The exact conditions for the bug seem to be:
We go in to macro fusion w
On Fri, 6 Feb 2015, Matthew Fortune wrote:
> > Native systems have for such ABI dependencies, including
> > stuff to set up $gp. Perhaps we could reuse these bits, the licence I
> > think allows us to.
>
> That's a good idea. Perhaps I should take that on as part of some cleanup
> of the MIPS
On 2015.02.05 at 12:57 -0800, H.J. Lu wrote:
>
> We should pass handle, not file, to release_input_file.
> I checked it in as an obvious fix.
This commit causes:
% echo "int main () {}" | gcc -fuse-ld=gold -flto -x c++ -
ld.gold: internal error in remove_writer, at token.h:132
collect2: error:
On Fri, Feb 6, 2015 at 4:56 AM, Markus Trippelsdorf
wrote:
> On 2015.02.05 at 12:57 -0800, H.J. Lu wrote:
>>
>> We should pass handle, not file, to release_input_file.
>> I checked it in as an obvious fix.
>
> This commit causes:
>
> % echo "int main () {}" | gcc -fuse-ld=gold -flto -x c++ -
> ld
On 2015.02.06 at 05:10 -0800, H.J. Lu wrote:
> On Fri, Feb 6, 2015 at 4:56 AM, Markus Trippelsdorf
> wrote:
> > On 2015.02.05 at 12:57 -0800, H.J. Lu wrote:
> >>
> >> We should pass handle, not file, to release_input_file.
> >> I checked it in as an obvious fix.
> >
> > This commit causes:
> >
> >
On Fri, Feb 6, 2015 at 5:18 AM, Markus Trippelsdorf
wrote:
> On 2015.02.06 at 05:10 -0800, H.J. Lu wrote:
>> On Fri, Feb 6, 2015 at 4:56 AM, Markus Trippelsdorf
>> wrote:
>> > On 2015.02.05 at 12:57 -0800, H.J. Lu wrote:
>> >>
>> >> We should pass handle, not file, to release_input_file.
>> >> I
On Fri, Feb 6, 2015 at 5:30 AM, H.J. Lu wrote:
> On Fri, Feb 6, 2015 at 5:18 AM, Markus Trippelsdorf
> wrote:
>> On 2015.02.06 at 05:10 -0800, H.J. Lu wrote:
>>> On Fri, Feb 6, 2015 at 4:56 AM, Markus Trippelsdorf
>>> wrote:
>>> > On 2015.02.05 at 12:57 -0800, H.J. Lu wrote:
>>> >>
>>> >> We sho
On 02/05/2015 08:15 PM, Aldy Hernandez wrote:
Absolutely nothing, basically because #if 0 would have been too obvious. It
was actually there to test if you were on your game. You passed
Yay?
The attached patch still passes the guality tests with no regressions. My
apologies for the lack
On Thursday 2015-02-05 22:49, David Malcolm wrote:
> Tested locally with a suitably hacked up script to work on this
> machine; with this patch the "Show Sources" links work within the
> files installed to /www/gcc/htdocs.
>
> OK for trunk?
>
> maintainer-scripts/ChangeLog:
> * update_web_d
The H8/300 port is setting RTX_FRAME_RELATED_P on random insns outside
the prologue/epilogue. That is causing dwarf2cfi heartburn as seen by
PR64889 (and its dup 64602).
This patch avoids setting RTX_FRAME_RELATED_P on those random insns.
The H8/300 port can now build libssp properly (646
This patch fixes PR target/64205, which was caused due to my previous patch to
add the -mupper-regs support for scalar floating point to occupy the VSX
floating point registers that are overlaid on top of the traditional Altivec
registers, starting with ISA 2.06 (power7).
The problem was, I added
> -Original Message-
> From: Matthew Fortune [mailto:matthew.fort...@imgtec.com]
> Sent: Thursday, February 05, 2015 3:52 PM
> Subject: RE: [PATCH v2][MIPS] fix CRT_CALL_STATIC_FUNCTION macro
> >
> I've put your patch inline below and switched to plain text. I suspect your
> post was bou
> -Original Message-
> From: Matthew Fortune [mailto:matthew.fort...@imgtec.com]
> Sent: Thursday, February 05, 2015 5:24 AM
> To: Moore, Catherine
> Cc: 'gcc-patches@gcc.gnu.org' (gcc-patches@gcc.gnu.org)
> Subject: RE: [PATCH,WWWDOCS] MIPS changes for GCC 5.0
>
> Moore, Catherine writ
On Mon, Nov 3, 2014 at 8:59 AM, Rainer Orth
wrote:
>
> The recent godump changes broke Solaris 11.1+ bootstrap in libgo:
> before, gen-sysinfo.so had
>
> type _zone_net_addr_t struct { zna_family uint16; zna_plen uint16; zna_addru
> struct { znau_addr6 _in6_addr; }; }
>
> which was filtered out
If a hidden weak symbol isn't defined in the TU, we can't assume it will
be defined in another TU at link time. It makes a difference in code
generation when compiling for PIC. If we assume that a hidden weak
undefined symbol is local, the address checking may be optimized out and
leads to the wro
On Thu, Feb 5, 2015 at 2:05 PM, Sriraman Tallam wrote:
> On Thu, Feb 5, 2015 at 11:59 AM, Richard Henderson wrote:
>> On 02/05/2015 11:01 AM, H.J. Lu wrote:
>>> Can you elaborate why it depends on COPY relocation? There
>>> is no COPY relocation on x86-64.
>>
>> Ho hum, we appear to have switche
+&& DECL_CONTEXT (snode->decl)
+&& TREE_CODE (DECL_CONTEXT (snode->decl)) != FUNCTION_DECL)
I think this should be !decl_function_context (snode->decl), in case
there's a class or BLOCK between the symbol and its enclosing function.
Done, also for the iteration through reachable func
Richard Biener writes:
> Looks good to me. I wonder if this will also help Aldyh...
Hmmm, I ran into something very similar. We solved it on the
debug-early branch by avoiding recursively setting DECL_ABSTRACT_P if
the parent was *not* DECL_ABSTRACT_P. Like this:
@@ -18274,7 +18312,8 @@ dwarf
On Feb 6, 2015, at 4:23 AM, Maciej W. Rozycki wrote:
> This consideration made me realise I've had a patch outstanding for some
> 10 years to convert all the `BAL x' instructions there to `BLTZAL $0, x'.
> This has always been a good idea in case implementations recognised the
> special case a
On 02/06/2015 11:42 AM, Aldy Hernandez wrote:
I was actually thinking of using dwarf2out_early_finish() to mop things up as
we generate early (or stream out) other auxiliary tables (pubname_table,
pubtype_table, file_table, etc). More details on that later. If so, can I
leave it as is?
OK.
Mike Stump writes:
> On Feb 6, 2015, at 4:23 AM, Maciej W. Rozycki
> wrote:
> > This consideration made me realise I've had a patch outstanding for
> > some
> > 10 years to convert all the `BAL x' instructions there to `BLTZAL $0,
> x'.
> > This has always been a good idea in case implementations
On Fri, 6 Feb 2015, Mike Stump wrote:
> > This consideration made me realise I've had a patch outstanding for some
> > 10 years to convert all the `BAL x' instructions there to `BLTZAL $0, x'.
> > This has always been a good idea in case implementations recognised the
> > special case and avoi
On Fri, Feb 6, 2015 at 9:41 AM, Michael Meissner
wrote:
> This patch fixes PR target/64205, which was caused due to my previous patch to
> add the -mupper-regs support for scalar floating point to occupy the VSX
> floating point registers that are overlaid on top of the traditional Altivec
> regis
On 02/05/2015 04:18 PM, Alan Modra wrote:
> OK when stage1 opens?
>
> * function.h (struct emit_status): Delete x_first_insn, x_last_insn
> and sequence_stack. Add seq.
> (seq_stack): Delete.
> * function.c (prepare_function_start): Don't access x_last_insn.
> * emit
OK with that change.
Sweet! Thanks for everything.
Though he's been silent, I bet Richi is secretly dancing :-).
Hi all,
This is a backport patch for branch 4.9. You can find the original patch
here: https://gcc.gnu.org/ml/gcc-patches/2014-09/msg00356.html
And it has been commit on the trunk as r215205.
This fixes a few libstdc++-v3 test suite failures.
x86_64 bootstraps Okay, aarch64_be-none-elf libstdc
On Feb 6, 2015, at 9:41 AM, Matthew Fortune wrote:
> Mike Stump writes:
>> On Feb 6, 2015, at 4:23 AM, Maciej W. Rozycki
>> wrote:
>>> This consideration made me realise I've had a patch outstanding for
>>> some
>>> 10 years to convert all the `BAL x' instructions there to `BLTZAL $0,
>> x'.
>>>
Hello!
2015-02-06 Uros Bizjak
* gcc.target/i386/pr64317.c: Compile for 32bit *-*-linux* targets.
(dg-options): Use -fpie instead of -fPIE -pie.
Tested on x86_64-linux-gnu {,m32}, committed to mainline SVN.
Uros.
Index: gcc.target/i386/pr64317.c
===
On Fri, 6 Feb 2015, Mike Stump wrote:
> Personally, the call form of bal in my book should be called call, and
> the non-call form of it should be called bal, but, I realize it is
> likely to late to do much about now. If one went down this path, then
> even changing it away from bal is wrong.
The following patch fixes a typo discovered recently:
https://gcc.gnu.org/ml/gcc/2015-02/msg00040.html
The patch was bootstrapped and tested on x86/x86-64.
Committed as rev.220488.
2015-02-06 Vladimir Makarov
* lra.c (lra_emit_add): Fix a typo in using disp instead of base.
Hi!
As discussed in the PR, for functions that return an aggregate that is not
aggregate_value_p (i.e. returned in registers), using RESULT_DECL is
undesirable, that's not what we normally emit for user code.
So, this patch instead uses a temporary, which is optimized right now as
much as similar
Hi!
As mentioned in the PR, the problem here is that the Ada FE needs to modify
global_options after toplevel.c (process_options), but as for LTO we now use
optimization_{default,current}_node for options of functions without
specifial optimization node, it means the changed options aren't reflect
Hi!
As discussed in PR64340, the clearing of flag_ipa_ra is misplaced too,
because it isn't reflected in optimization_default_node.
Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for
trunk?
2015-02-06 Jakub Jelinek
* toplev.c (process_options): Change flag_ip
Hello,
we currently reject programs of the form
>
> module m
> integer :: s
> end module m
> subroutine s
> use m, only: x => s, x => s
> end subroutine s
with an error stating that S is the name of the current program unit.
Interestingly, the duplicate rename is necessary to trigger it.
> Hi!
>
> As discussed in the PR, for functions that return an aggregate that is not
> aggregate_value_p (i.e. returned in registers), using RESULT_DECL is
> undesirable, that's not what we normally emit for user code.
>
> So, this patch instead uses a temporary, which is optimized right now as
>
On 02/06/15 13:28, Jakub Jelinek wrote:
Hi!
As discussed in PR64340, the clearing of flag_ipa_ra is misplaced too,
because it isn't reflected in optimization_default_node.
Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for
trunk?
2015-02-06 Jakub Jelinek
*
/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/
/sw/src/fink.build/gcc50-5.0.0-1000/gcc-5-20150206/gcc/testsuite/gcc.dg/visibility-22.c
-fno-diagnostics-show-caret -fdiagnostics-color=never -fPIC -lm
-m32 -o ./visibility-22.exe(timeout = 300)
spawn -ignore SIGHUP
/sw/src/fink.build/gcc50
>
> Executing on host:
> /sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/xgcc
> -B/sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/
> /sw/src/fink.build/gcc50-5.0.0-1000/gcc-5-20150206/gcc/testsuite/gcc.dg/visibility-22.c
> -fno-diagnostics-show-caret -fdiagnost
nk.build/gcc50-5.0.0-1000/gcc-5-20150206/gcc/testsuite/gcc.dg/visibility-23.c
-fno-diagnostics-show-caret -fdiagnostics-color=never -fPIC -S -m32 -o
visibility-23.s
contains...
.text
.globl _main
_main:
LFB0:
pushl %ebp
LCFI0:
movl%esp, %ebp
LCFI1:
This bug has gone latent on the trunk; however, the problem still
remains that cse will incorrectly simplify a ZERO/SIGN_EXTRACT in some
cases.
ZERO/SIGN_EXTRACT are somewhat special in that if they are extracting
from a memory operand, that memory operand will always have QImode
regardless of th
The attached patch adds a link to the built HTML docs for libgccjit to
htdocs/onlinedocs/index.html.
Committed to CVS as obvious.
Index: htdocs/onlinedocs/index.html
===
RCS file: /cvs/gcc/wwwdocs/htdocs/onlinedocs/index.html,v
retrie
* config/rs6000/rs6000.c (rs6000_call_aix): Use get_current_sequence.
The rs6000 change is okay.
Thanks, David
On Fri, Feb 6, 2015 at 12:36 PM, Lynn A. Boger
wrote:
>
> I was told by the testers that this patch did not work. The error message
> was:
>
>> panic: interface conversion: interface is nil, not error
>>
>> goroutine 1 [running]:
>> client.$nested1
>> /home/ubuntu/docker.mainline/src/github.c
The jit testsuite attempted to test the gcc_jit_context_compile_to_file
API entrypoint by running "file" on the generated file, and comparing
the result against a regexp.
This approach is unfixable: some hosts won't have "file" installed, and
the output of "file" seems to vary enough from host to
Any idea why HP still sees the tests fail? See comment 8 at
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64467#c8
On 3 February 2015 at 13:49, Renlin Li wrote:
> On 01/02/15 15:08, Jonathan Wakely wrote:
>>
>> I failed to CC gcc-patches on this patch ...
>>
>> On 29/01/15 13:02 +, Jonathan Wakely wrote:
>>
>> diff --git
>> a/libstdc++-v3/testsuite/17_intro/headers/c++200x/all_attributes.cc
>> b/libstdc++-
It's been a long time since I did any significant work on GCC,
and it is unlikely that I'll be doing much for the foreseeable
future.
While I still have some understanding of the modules I used to
maintain, I don't think it is reasonable to have me making
decisions on them. It's been too long and
On Mon, Feb 2, 2015 at 11:37 PM, Jakub Jelinek wrote:
> On Mon, Feb 02, 2015 at 02:51:43PM -0800, Andrew Pinski wrote:
>> While trying to build the GCC 5 with GCC 5, I ran into an ICE when
>> building libcpp at -O0. The problem is the C++ front-end was not
>> folding sizeof(a)/sizeof(a[0]) when p
Hi Sandra,
Index: gcc/ipa-devirt.c
===
polymorphic (indirect) call
- This is callgraph represention of virtual method call. Every
+ This is callgraph representation of virtual method call. Every
polymorphic
; /sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/xgcc
> -B/sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/
> /sw/src/fink.build/gcc50-5.0.0-1000/gcc-5-20150206/gcc/testsuite/gcc.dg/visibility-23.c
> -fno-diagnostics-show-caret -fdiagnostics-color=never -fPIC -S -m32 -o
> vis
5.0.0-1000/darwin_objdir/gcc/xgcc
>> -B/sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/
>> /sw/src/fink.build/gcc50-5.0.0-1000/gcc-5-20150206/gcc/testsuite/gcc.dg/visibility-22.c
>> -fno-diagnostics-show-caret -fdiagnostics-color=never -fPIC -lm
>> -m32 -o ./visibil
On 02/06/2015 06:19 PM, Gerald Pfeifer wrote:
Hi Sandra,
Index: gcc/ipa-devirt.c
===
polymorphic (indirect) call
- This is callgraph represention of virtual method call. Every
+ This is callgraph representation of
>>> FAIL: gcc.dg/visibility-23.c scan-hidden private_extern[ \t_]*_?foo
>>>
>>> with...
>>>
>>> Executing on host:
>>> /sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/xgcc
>>> -B/sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/
>
With the attached patch I create a special version of fbuf_flush that is only
called with list directed I/O. There can be no tabbing back and forth so it is
safe to flush the buffer whenever we want. The bug occurs when the buffer keeps
growing to no end until no more allocations can be made a
t;>> with...
>>>>
>>>> Executing on host:
>>>> /sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/xgcc
>>>> -B/sw/src/fink.build/gcc50-5.0.0-1000/darwin_objdir/gcc/
>>>> /sw/src/fink.build/gcc50-5.0.0-1000/gcc-5-20150206/gcc/testsuite/
After a lot of investigation, I believe that I have learned why stabs
debugging on AIX disabled use of continuations. GDB, IBM DBX and IBM
XLDB are able to work with stab string continuations produced by GCC.
I am enabling it using definitions that match the behavior of IBM XL
compilers to avoid
On 02/06/15 19:32, David Edelsohn wrote:
After a lot of investigation, I believe that I have learned why stabs
debugging on AIX disabled use of continuations. GDB, IBM DBX and IBM
XLDB are able to work with stab string continuations produced by GCC.
I am enabling it using definitions that match
72 matches
Mail list logo
