On Wed, Oct 31, 2012 at 11:34:10AM -0700, Wei Mi wrote:
+static tree
+get_init_decl (void)
+{
+ tree typ;
+ static tree decl;
+
+ if (decl != NULL)
+return decl;
+ typ = build_function_type_list (void_type_node, NULL_TREE);
+ decl = build_func_decl (typ, "__tsan_init");
+ return decl;
+}
gen_doloop_end_split creates a pattern that sets pc, hence emit_jump_insn
has to be used instead of emit_insn.
Committed as obvious.
2012-11-01 Joern Rennecke
PR target/55160
* config/sh/sh.md (doloop_end): Use emit_jump_insn.
Index: config/sh/sh.md
==
On Wed, Oct 31, 2012 at 11:00:17PM -0700, Xinliang David Li wrote:
> > + /* Below are things we do not instrument
> > + (no possibility of races or not implemented yet). */
> > + if (/* Compiler-emitted artificial variables. */
> > + (DECL_P (expr) && DECL_ARTIFICIAL (expr))
> > +
On Wed, Oct 31, 2012 at 4:10 PM, Jakub Jelinek wrote:
> Hi!
>
> Just a couple of random comments:
>
> On Wed, Oct 31, 2012 at 11:34:10AM -0700, Wei Mi wrote:
>> gcc/ChangeLog:
>> 2012-10-31 Wei Mi
>
> If Dmitry wrote parts of the patch, it would be nice to mention
> him in the ChangeLog too.
>
> --fno-default-inline @gol
> +-fmove-loop-invariants -fmudflap -fmudflapir -fmudflapth
> -fno-branch-count-reg @gol
> +-ftsan -ftsan-ignore -fno-default-inline @gol
Change -ftsan to -fthread-sanitizer
> -fno-defer-pop -fno-function-cse -fno-guess-branch-probability @gol
> -fno-inline -fno-mat
Quoting Richard Sandiford :
But what I'm trying to get at is: why can't the backend tell
shorten_branches about the amount of alignment/misalignment
that the target wants, and where? Via an attribute or a hook,
I don't mind which. But it should be declarative, rather than
a part of the shorten
Hi,
When debugging optimized code, it is always confusing when gdb jumped
to a place that has never been executed. This is because compiler
performs some aggressive code motion that moves an instruction outside
of its original residing basic block.
This patch tries to fix this problem by resettin
On Wed, 2012-10-31 at 09:02 -0500, Peter Bergner wrote:
> On Wed, 2012-10-31 at 14:55 +0100, Jakub Jelinek wrote:
> > On Wed, Oct 31, 2012 at 08:53:31AM -0500, Peter Bergner wrote:
> > > Great. Jakub, were you going to commit your change or did you want me
> > > to do that?
> >
> > I haven't test
On Wed, Oct 31, 2012 at 6:57 PM, H.J. Lu wrote:
> On Wed, Oct 31, 2012 at 3:53 PM, H.J. Lu wrote:
>> On Wed, Oct 31, 2012 at 7:29 AM, Eric Botcazou wrote:
>>>
>>> > It failed with revision 188008.
>>>
>>> OK, thanks. So the testcase never compiled on the trunk (except for about
>>> 24
>>> hours
On Wed, Oct 31, 2012 at 7:14 PM, H.J. Lu wrote:
> On Wed, Oct 31, 2012 at 6:57 PM, H.J. Lu wrote:
>> On Wed, Oct 31, 2012 at 3:53 PM, H.J. Lu wrote:
>>> On Wed, Oct 31, 2012 at 7:29 AM, Eric Botcazou
>>> wrote:
> It failed with revision 188008.
OK, thanks. So the testcase
On Wed, Oct 31, 2012 at 6:57 PM, H.J. Lu wrote:
> On Wed, Oct 31, 2012 at 3:53 PM, H.J. Lu wrote:
>> On Wed, Oct 31, 2012 at 7:29 AM, Eric Botcazou wrote:
>>>
>>> > It failed with revision 188008.
>>>
>>> OK, thanks. So the testcase never compiled on the trunk (except for about
>>> 24
>>> hours
On Wed, Oct 31, 2012 at 3:53 PM, H.J. Lu wrote:
> On Wed, Oct 31, 2012 at 7:29 AM, Eric Botcazou wrote:
>>
>> > It failed with revision 188008.
>>
>> OK, thanks. So the testcase never compiled on the trunk (except for about
>> 24
>> hours between 188009 & 188118) or did it compile before 188008
This changes the forward_list::assign() members to assign to existing
elements instead of destroying them and reallocating new ones, as
allowed by the Sequence Container requirements. The copy assignment
operator already did that, so now it uses assign(). For QoI we still
support non-assignable t
From: Andi Kleen
This adds a new C/C++ option to force
__attribute__((no_instrument_function)) on every function compiled.
This is useful together with LTO. You may want to have the whole
program compiled with -pg and have to specify that in the LTO
link, but want to disable it for some specific
Hi,
When -fcompare_debug is used, what we really want to do is to compare
instructions between the -g version and -gtoggle version. However,
current dump file still contains the source line in its rtl dump. This
patch changes to only dump rtl without dumping its source info.
Bootstrapped and pass
(CC list trimmed.)
On Wed, 31 Oct 2012, Kirill Yukhin wrote:
> Hi,
> This patch introduces a new RTL expression called define_subst and
> required by it define_subst_attr.
>
> The new feature allows to make MD-files more compact - it defines a
> rule by which a parser could generate modified versi
Hi,
the testcase bellow started to fail because we now inline the call of _Exit to
call to bar. Fixed thus.
Honza
Index: ChangeLog
===
--- ChangeLog (revision 193049)
+++ ChangeLog (working copy)
@@ -1,3 +1,7 @@
+2012-10-31 Jan
Hi!
Just a couple of random comments:
On Wed, Oct 31, 2012 at 11:34:10AM -0700, Wei Mi wrote:
> gcc/ChangeLog:
> 2012-10-31 Wei Mi
If Dmitry wrote parts of the patch, it would be nice to mention
him in the ChangeLog too.
> * Makefile.in (tsan.o): New
> * passes.c (init_optimi
Hi Jason,
Just wanted to be sure you saw this. I'm hoping to get it in for stage 1.
http://gcc.gnu.org/ml/gcc-patches/2012-10/msg02684.html
Sterling
> On Fri, Oct 26, 2012 at 8:05 AM, Jan Hubicka wrote:
> >> Hi,
> >>
> >> On Tue, Oct 23, 2012 at 3:03 AM, Jan Hubicka wrote:
> >> >> Ping.
> >> >>
> >> >>
> >> >> On Wed, Oct 17, 2012 at 1:48 PM, Easwaran Raman
> >> >> wrote:
> >> >> > Hi,
> >> >> > This patch fixes bugs introduced by my previ
On Tue, Oct 30, 2012 at 10:43 PM, Teresa Johnson wrote:
> Sure, I will give this a try after your verification patch tests
> complete. Does this mean that the patch you posted above to
> force_nonfallthru_and_redirect is no longer needed either? I'll see if
> I can avoid the need for some of my fix
On 31 October 2012 22:46, Marc Glisse wrote:
> On Wed, 31 Oct 2012, Jonathan Wakely wrote:
>
>> On 31 October 2012 22:14, François Dumont wrote:
>>>
>>> Here is the patch I came to. I use the 'universal reference' like you
>>> propose but some tests started to fail because I think gcc called it
On Wed, Oct 31, 2012 at 7:29 AM, Eric Botcazou wrote:
>
> > It failed with revision 188008.
>
> OK, thanks. So the testcase never compiled on the trunk (except for about
> 24
> hours between 188009 & 188118) or did it compile before 188008 at some
> point?
>
> --
> Eric Botcazou
It was OK until
This patch removes the unused ebitmap, and then removes some sbitmap functions
only used by ebitmap. The functions removed are:
SET_BIT_WITH_POPCOUNT
RESET_BIT_WITH_POPCOUNT
bitmap_copy_n
bitmap_range_empty_p
sbitmap_popcount
In addition, two functions have been made private to the implementatio
On Wed, 31 Oct 2012, Jonathan Wakely wrote:
On 31 October 2012 22:14, François Dumont wrote:
Here is the patch I came to. I use the 'universal reference' like you
propose but some tests started to fail because I think gcc called it instead
of the move constructor.
Ah of course. The defau
On 31 October 2012 22:14, François Dumont wrote:
> On 10/30/2012 10:26 PM, Jonathan Wakely wrote:
>
> Are you sure all GDB 7.x should work ? I have gdb 7.1 and when running
> pretty printers tests I have:
>
> Spawning: gdb -nw -nx -quiet -batch -ex "python print
> gdb.lookup_global_symbol"
> ES
On 10/30/2012 10:26 PM, Jonathan Wakely wrote:
I try to adapt pretty printer code but haven't been able to test it as I
don't have the necessary gdb version and don't have time to update it at the
moment. If you prefer I can leave it untouched.
Please try to install a newer GDB, building it
Dear Tobias,
>
> Looks obvious if you ask me ...
..and to me too. OK for trunk.
Thanks
Paul
This patch to libgo changes the goc2c program to change the Go type
"int" to the C type "intgo". This is in preparation for changing the Go
type "int" to be 64 bits on x86_64. The goc2c program used to be used
by the gc compiler, but it no longer is, so while I was there I removed
the gc support.
On 2012-11-01 07:31, Aldy Hernandez wrote:
> + // Hmmm, the front-end should have caught outer aborts without
> + // an outer transaction. Bail and hope for the best.
> + tree attrs = get_attrs_for (current_function_decl);
> + if (!attrs || !lookup_attribute
+ // The front-end should have caught outer aborts without
+ // an outer transaction.
+ gcc_unreachable ();
Err, no the final check there would be if the current function includes
attribute may_throw_abort_outer (sp). Just return here I think for now.
Lik
On Wed, Oct 31, 2012 at 12:58 PM, Christophe Lyon
wrote:
> On 30.10.2012 17:59, Teresa Johnson wrote:
>>
>> On Tue, Oct 30, 2012 at 9:26 AM, Steven Bosscher
>> wrote:
>>>
>>> Hello,
>>>
>>> Hot/cold partitioning is apparently a hot topic all of a sudden, which
>>> is a good thing of course, becau
On 30.10.2012 17:59, Teresa Johnson wrote:
On Tue, Oct 30, 2012 at 9:26 AM, Steven Bosscher wrote:
Hello,
Hot/cold partitioning is apparently a hot topic all of a sudden, which
is a good thing of course, because it's in need of some TLC.
The attached patch adds another check the RTL cfg check
On Oct 31, 2012, at 7:05 AM, Richard Biener wrote:
> You have an artificial limit on what 'len' can be.
No. There is no limit, and nothing artificial. We take the maximum of the
needs of the target, the maximum of the front-ends and the maximum of the
mid-end and the back-end. We can drop a
On 2012-10-31 19:09, Eric Botcazou wrote:
> Original message at:
> http://gcc.gnu.org/ml/gcc-patches/2012-10/msg00013.html
>
> Thanks in advance.
>
Ok.
r~
On Oct 31, 2012, at 6:54 AM, Richard Biener wrote:
> I propose that no wide-int member function
> may _change_ it's len (to something larger).
We never do that, so, we already do as you wish. We construct wide ints, and
we have member functions to construct values. We need to construct values
On Wed, Oct 31, 2012 at 4:36 AM, Richard Biener
wrote:
> On Wed, Oct 24, 2012 at 4:02 AM, Easwaran Raman wrote:
>> On Tue, Oct 23, 2012 at 2:52 AM, Richard Biener
>> wrote:
>>> On Mon, Oct 22, 2012 at 8:31 PM, Easwaran Raman wrote:
On Mon, Oct 22, 2012 at 12:59 AM, Richard Biener
wro
On Oct 31, 2012, at 5:44 AM, Richard Biener wrote:
> the
> fact that len ends up being mutable is another thing I dislike about
> wide-int.
We expose len for construction only, it is non-mutating. During construction,
there is no previous value.
> If wide-ints are cheap then all ops should be
Dehao's patch will make the debugging of the following code (-g -O2)
less jumpy. After the testing of x > 0, it should go to line 'a++'.
Without the fix, when stepping through 'abc', the lines covered are
6, 4, 11, 13. With the fix it should be 6, 9, 11, 13 -- much better.
David
1. int x;
On Wed, 31 Oct 2012, Kenneth Zadeck wrote:
Richi,
Let me explain to you what a broken api is. I have spent the last week
screwing around with tree-vpn and as of last night i finally got it to work.
In tree-vpn, it is clear that double-int is the precise definition of a
broken api.
The tr
OK.
Jason
The following patch fixes
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55150
The patch was successfully bootstrapped on x86/x86-64.
Committed as rev. 193042.
2012-10-31 Vladimir Makarov
PR middle-end/55150
* lra-constraints.c (lra_constraints): Update debug insn info
Hi,
The patch is about ThreadSanitizer. ThreadSanitizer is a data race
detector for C/C++ programs. It contains two parts: instrumentation
and runtime library. This patch is the first part, and runtime will be
included in the second part. Dmitry(dvyu...@google.com) is the author
of this part, and
Jakub,
it is hard from all of the threads to actually distill what the real
issues are here. So let me start from a clean slate and state them simply.
Richi has three primary objections:
1) that we can do all of this with a templated version of double-int.
2) that we should not be passing in
On 10/31/2012 09:49 AM, Richard Biener wrote:
> On Tue, Oct 30, 2012 at 10:05 PM, Kenneth Zadeck
> wrote:
>> jakub,
>>
>> i am hoping to get the rest of my wide integer conversion posted by nov 5.
>> I am under some adverse conditions here: hurricane sandy hit her pretty
>> badly. my house is hoo
On Fri, Oct 26, 2012 at 8:05 AM, Jan Hubicka wrote:
>> Hi,
>>
>> On Tue, Oct 23, 2012 at 3:03 AM, Jan Hubicka wrote:
>> >> Ping.
>> >>
>> >>
>> >> On Wed, Oct 17, 2012 at 1:48 PM, Easwaran Raman wrote:
>> >> > Hi,
>> >> > This patch fixes bugs introduced by my previous patch to propagate
>> >>
On Oct 31, 2012, at 2:45 AM, Richard Biener wrote:.
> My comment was for isolated code parts that are being rewritten
> (I think it was the wide-int class). Consistency comes first.
In the case of wide int, we only use references in one very narrow way. We use
const T& as parameters instead of
This patch renames sbitmap iterators to unify them with the bitmap iterators.
Remove the unused EXECUTE_IF_SET_IN_SBITMAP_REV, which has an unconventional
interface.
Rename the sbitmap_iter_* functions to match bitmap's bmp_iter_* functions.
Add an additional parameter to the initialization and n
On 10/31/2012 06:13 PM, Florian Weimer wrote:
+ if (outer_nelts_check != NULL && inner_size.is_one())
Uhm, I will add the missing space before commit. Sorry.
--
Florian Weimer / Red Hat Product Security Team
On 10/31/2012 10:05 AM, Jakub Jelinek wrote:
I don't see any == double_int_one (or zero) comparisons in grep,
so I'd say inner_size.is_one () should be used instead (which is used
pretty frequently). Ditto in the second spot.
Otherwise the patch looks good to me, but I'd like Jason to chime in
On Wed, 31 Oct 2012, Tobias Burnus wrote:
> Jakub Jelinek:
> > I think it would be nice if you also posted the changes you did to
> > test-ldouble.c and libm-test.inc, so that next time we could more easily
> > test it again.
>
> See attachment. (I didn't do it properly at first, thus, I had to p
> With "ASSOCIATE (A => array)", one generates internally "A" as AS_DEFERRED
> array. However, it is neither a pointer nor allocatable, unless "array" is.
>
> When passing "A" as actual argument to a non-descriptor dummy, trans-array.c
> assumed that the actual argument had no descriptor, which lea
On Wed, 31 Oct 2012, Tobias Burnus wrote:
> Tobias Burnus wrote:
> > libquadmath's math functions are based on (but not identical to) GLIBC's
> > sysdeps/ieee754/ldbl-128 functions. In the attached patch, I have ported the
> > bug fixes from GLIBC over to libquadmath. Hopefully, the port is comple
On 12-10-31 12:33 PM, Gary Funck wrote:
On 10/28/12 20:43:05, Vladimir Makarov wrote:
The following patch fixes PR55106. A value in GENERAL_REGS is
inherited into a move with destination pseudo of SSE_REGS. It
results into secondary move for which inheritance is tried again an
again. It means
With "ASSOCIATE (A => array)", one generates internally "A" as
AS_DEFERRED array. However, it is neither a pointer nor allocatable,
unless "array" is.
When passing "A" as actual argument to a non-descriptor dummy,
trans-array.c assumed that the actual argument had no descriptor, which
lead to
On 10/28/12 20:43:05, Vladimir Makarov wrote:
> The following patch fixes PR55106. A value in GENERAL_REGS is
> inherited into a move with destination pseudo of SSE_REGS. It
> results into secondary move for which inheritance is tried again an
> again. It means cycling LRA passes.
>
> The pat
Tobias Burnus writes:
> diff --git a/math/libm-test.inc b/math/libm-test.inc
> index 1e067fe..81b40b6 100644
> --- a/math/libm-test.inc
> +++ b/math/libm-test.inc
> @@ -226,5 +226,5 @@ static FLOAT max_error, real_max_error, imag_max_error;
>
> #define BUILD_COMPLEX(real, imag) \
> - ({ __com
Jakub Jelinek:
I think it would be nice if you also posted the changes you did to
test-ldouble.c and libm-test.inc, so that next time we could more
easily test it again.
See attachment. (I didn't do it properly at first, thus, I had to
propagate the changes to the right files …)
Tobias
/* C
On Wed, Oct 31, 2012 at 04:52:19PM +0100, Tobias Burnus wrote:
> Tobias Burnus wrote:
> >libquadmath's math functions are based on (but not identical to)
> >GLIBC's sysdeps/ieee754/ldbl-128 functions. In the attached patch,
> >I have ported the bug fixes from GLIBC over to libquadmath.
> >Hopefully
Tobias Burnus wrote:
libquadmath's math functions are based on (but not identical to)
GLIBC's sysdeps/ieee754/ldbl-128 functions. In the attached patch, I
have ported the bug fixes from GLIBC over to libquadmath. Hopefully,
the port is complete and correct.
Slightly updated version, committed
On 10/31/2012 09:30 AM, Richard Sandiford wrote:
Richard Biener writes:
But that means that wide_int has to model a P-bit operation as a
"normal" len*HOST_WIDE_INT operation and then fix up the result
after the fact, which seems unnecessarily convoluted.
It does that right now. The operation
Joern Rennecke writes:
> Quoting Richard Sandiford :
>>> The length variation for the ARC are not alike: there are branches that
>>> are subject to branch shortening in the usual way, but they might
>>> shrink when other things shrink. When we are iterating starting at
>>> minimal length and incr
I was not planning to do that mangling for 4.8.My primary
justification for getting it in publicly now is that there are a large
number of places where the current compiler (both at the tree and rtl
levels) do not do optimization of the value is larger than a single
hwi.My code generali
> Yeah. But please also check gdb testsuite for this kind of patches.
This patch also passed gdb testsuite.
Thanks,
Dehao
>
> Jakub
On Wed, Oct 31, 2012 at 2:34 AM, Richard Biener
wrote:
> On Wed, Oct 31, 2012 at 12:57 AM, Xinliang David Li
> wrote:
>> It will make the location info for the newly synthesized stmt more
>> deterministic, I think.
>
> Maybe, but it will increase the jumpiness in the debugger without actually
>
Committed as obvious.
2012-10-31 Joern Rennecke
* expr.c (can_move_by_pieces): Apply ATTRIBUTE_UNUSED to len.
Index: expr.c
===
--- expr.c (revision 193034)
+++ expr.c (working copy)
@@ -841,7 +841,7 @@ widest_i
> It failed with revision 188008.
OK, thanks. So the testcase never compiled on the trunk (except for about 24
hours between 188009 & 188118) or did it compile before 188008 at some point?
--
Eric Botcazou
On 10/31/2012 08:44 AM, Richard Biener wrote:
On Wed, Oct 31, 2012 at 1:22 PM, Richard Sandiford
wrote:
Richard Biener writes:
On Wed, Oct 31, 2012 at 1:05 PM, Richard Sandiford
wrote:
Richard Biener writes:
On Wed, Oct 31, 2012 at 11:43 AM, Richard Sandiford
wrote:
Richard Biener wri
Hi,
On 10/31/2012 02:50 PM, Jason Merrill wrote:
On 10/30/2012 07:44 PM, Paolo Carlini wrote:
straightforward fix. The below, which regtests fine, simply acts on any
MULT_EXPR as TYPE_SIZE, which I think should be fine, but, in case the
idea is basically Ok, we could also consider the more comp
On Wed, Oct 31, 2012 at 10:04:58AM -0400, Kenneth Zadeck wrote:
> if one looks at where intel is going, they are doing exactly the
> same thing.The difference is that they like to add the
> operations one at a time rather than just do a clean implementation
> like we did. Soon they will get t
On 10/31/2012 10:24 AM, Richard Biener wrote:
On Wed, Oct 31, 2012 at 3:18 PM, Kenneth Zadeck
wrote:
On 10/31/2012 10:05 AM, Richard Biener wrote:
On Wed, Oct 31, 2012 at 2:54 PM, Kenneth Zadeck
wrote:
On 10/31/2012 08:11 AM, Richard Biener wrote:
On Wed, Oct 31, 2012 at 1:05 PM, Richard S
On 10/30/2012 05:49 PM, Sriraman Tallam wrote:
AFAIU, this should not be a problem. For duplicate declarations,
duplicate_decls should merge them and they should never be seen here.
Did I miss something?
With extern "C" functions you can have multiple declarations of the same
function in diffe
On 10/31/2012 09:54 AM, Richard Biener wrote:
On Wed, Oct 31, 2012 at 2:30 PM, Richard Sandiford
wrote:
Richard Biener writes:
But that means that wide_int has to model a P-bit operation as a
"normal" len*HOST_WIDE_INT operation and then fix up the result
after the fact, which seems unnecess
On Wed, Oct 31, 2012 at 3:18 PM, Kenneth Zadeck
wrote:
>
> On 10/31/2012 10:05 AM, Richard Biener wrote:
>>
>> On Wed, Oct 31, 2012 at 2:54 PM, Kenneth Zadeck
>> wrote:
>>>
>>> On 10/31/2012 08:11 AM, Richard Biener wrote:
On Wed, Oct 31, 2012 at 1:05 PM, Richard Sandiford
wrote:
On 10/31/2012 10:05 AM, Richard Biener wrote:
On Wed, Oct 31, 2012 at 2:54 PM, Kenneth Zadeck
wrote:
On 10/31/2012 08:11 AM, Richard Biener wrote:
On Wed, Oct 31, 2012 at 1:05 PM, Richard Sandiford
wrote:
Richard Biener writes:
On Wed, Oct 31, 2012 at 11:43 AM, Richard Sandiford
wrote:
On Wed, 31 Oct 2012, Jan Hubicka wrote:
> > On Wed, Oct 31, 2012 at 01:30:02PM +0100, Richard Biener wrote:
> > > On Wed, 31 Oct 2012, Jan Hubicka wrote:
> > > > > unroll you mean. Because unrolling mutates the CFG too much.
> > > > > Well - it was just a starting point, populating -Og with as li
On Wed, 2012-10-31 at 14:55 +0100, Jakub Jelinek wrote:
> On Wed, Oct 31, 2012 at 08:53:31AM -0500, Peter Bergner wrote:
> > Great. Jakub, were you going to commit your change or did you want me
> > to do that?
>
> I haven't tested it, you did, so please do that yourself. Thanks.
I tested it on
On Wed, 31 Oct 2012, Jan Hubicka wrote:
> > > - FOR_EACH_VEC_ELT (edge, exits, i, ex)
> > > + if (loop->any_upper_bound)
> > > {
> > > - if (!just_once_each_iteration_p (loop, ex->src))
> > > - continue;
> > > + if (dump_file && (dump_flags & TDF_DETAILS))
> > > + fprintf (dump_fi
On Wed, Oct 31, 2012 at 2:54 PM, Kenneth Zadeck
wrote:
>
> On 10/31/2012 08:11 AM, Richard Biener wrote:
>>
>> On Wed, Oct 31, 2012 at 1:05 PM, Richard Sandiford
>> wrote:
>>>
>>> Richard Biener writes:
On Wed, Oct 31, 2012 at 11:43 AM, Richard Sandiford
wrote:
>
> Richar
> On Wed, Oct 31, 2012 at 01:30:02PM +0100, Richard Biener wrote:
> > On Wed, 31 Oct 2012, Jan Hubicka wrote:
> > > > unroll you mean. Because unrolling mutates the CFG too much.
> > > > Well - it was just a starting point, populating -Og with as little
> > > > as possible and 100% profitable tran
jakub
my port has 256 bit integers. They are done by strapping together all
of the elements of a vector unit.
if one looks at where intel is going, they are doing exactly the same
thing.The difference is that they like to add the operations one at
a time rather than just do a clean imple
> > - FOR_EACH_VEC_ELT (edge, exits, i, ex)
> > + if (loop->any_upper_bound)
> > {
> > - if (!just_once_each_iteration_p (loop, ex->src))
> > - continue;
> > + if (dump_file && (dump_flags & TDF_DETAILS))
> > + fprintf (dump_file, "Found loop %i to be finite: upper bound is
>
Hello Alan,
maybe it is better to use a require effective target instead of the
{ target powerpc*-*-eabi* powerpc*-*-elf* powerpc*-*-linux* }
patterns scattered around in the testsuite? One problem with this is that test
cases for one of these will likely also work with powerpc*-*-rtems*. I
On Wed, Oct 31, 2012 at 08:53:31AM -0500, Peter Bergner wrote:
> Great. Jakub, were you going to commit your change or did you want me
> to do that?
I haven't tested it, you did, so please do that yourself. Thanks.
Jakub
On Wed, Oct 31, 2012 at 09:44:50AM -0400, Kenneth Zadeck wrote:
> The tree-vpn uses an infinite-precision view of arithmetic. However,
> that infinite precision is implemented on top of a finite, CARVED IN
> STONE, base that is and will always be without a patch like this,
> 128 bits on an x86-64.
Quoting Richard Biener :
On Wed, Oct 31, 2012 at 2:06 PM, Joern Rennecke
wrote:
Quoting Richard Biener :
Maybe we should split the thing then into a adjust_insn_length attribute
without the iteration parameter
Attributes don't get any parameter but the instruction, and don't apply
to dela
On Wed, Oct 31, 2012 at 2:30 PM, Richard Sandiford
wrote:
> Richard Biener writes:
>>> But that means that wide_int has to model a P-bit operation as a
>>> "normal" len*HOST_WIDE_INT operation and then fix up the result
>>> after the fact, which seems unnecessarily convoluted.
>>
>> It does that
On 10/31/2012 08:11 AM, Richard Biener wrote:
On Wed, Oct 31, 2012 at 1:05 PM, Richard Sandiford
wrote:
Richard Biener writes:
On Wed, Oct 31, 2012 at 11:43 AM, Richard Sandiford
wrote:
Richard Biener writes:
On Thu, Oct 25, 2012 at 12:55 PM, Kenneth Zadeck
wrote:
On 10/25/2012 06:42 A
On Wed, 2012-10-31 at 10:12 +0100, Richard Biener wrote:
> On Tue, Oct 30, 2012 at 9:23 PM, Peter Bergner wrote:
> > On Tue, 2012-10-30 at 20:37 +0100, Jakub Jelinek wrote:
> >> On Tue, Oct 30, 2012 at 02:03:44PM -0500, Peter Bergner wrote:
> >> > Ok, then I'll bootstrap and regtest your suggested
On 10/30/2012 07:44 PM, Paolo Carlini wrote:
straightforward fix. The below, which regtests fine, simply acts on any
MULT_EXPR as TYPE_SIZE, which I think should be fine, but, in case the
idea is basically Ok, we could also consider the more complex but more
sophisticated variably_modified_type_p
Richi,
Let me explain to you what a broken api is. I have spent the last week
screwing around with tree-vpn and as of last night i finally got it to
work. In tree-vpn, it is clear that double-int is the precise
definition of a broken api.
The tree-vpn uses an infinite-precision view of a
Richard Biener writes:
>> But that means that wide_int has to model a P-bit operation as a
>> "normal" len*HOST_WIDE_INT operation and then fix up the result
>> after the fact, which seems unnecessarily convoluted.
>
> It does that right now. The operations are carried out in a loop
> over len HO
On Tue, Oct 30, 2012 at 6:50 PM, Eric Botcazou wrote:
> This will hopefully fix the build failure reported by Diego. Apart from
> adding missing dependencies, this also removes redundant command lines.
>
> Tested on x86_64-suse-linux, applied on the mainline and 4.7 branch.
>
>
> 2012-10-30 Eric
On Wed, Oct 31, 2012 at 2:06 PM, Joern Rennecke
wrote:
> Quoting Richard Biener :
>
>> Maybe we should split the thing then into a adjust_insn_length attribute
>> without the iteration parameter
>
>
> Attributes don't get any parameter but the instruction, and don't apply
> to delay slot SEQUENCEs
Quoting Richard Biener :
Maybe we should split the thing then into a adjust_insn_length attribute
without the iteration parameter
Attributes don't get any parameter but the instruction, and don't apply
to delay slot SEQUENCEs.
On Wed, 31 Oct 2012, Jakub Jelinek wrote:
On Tue, Oct 30, 2012 at 01:14:40PM +0100, Marc Glisse wrote:
On Tue, 30 Oct 2012, Marek Polacek wrote:
On Tue, Oct 30, 2012 at 08:05:13AM +0100, Marc Glisse wrote:
Hello,
one more optimization that needed help for vectors, it crashed on
(x
On 10/31/2012 20:01, Jonathan Wakely wrote:
> On 31 October 2012 11:23, JonY wrote:
>> On 10/31/2012 19:12, Jonathan Wakely wrote:
>>>
>>> It looks like the workaround is in mingw not in GCC, so is it a
>>> problem that it won't be possible to use GCC 4.8 with existing mingw
>>> versions, or are us
On Wed, Oct 31, 2012 at 5:42 AM, Eric Botcazou wrote:
>> Your change caused:
>>
>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55142
>
> Please check whether it worked before Richard's fix (r188009).
>
It failed with revision 188008.
--
H.J.
On Fri, 2012-10-26 at 17:05 +0200, Jan Hubicka wrote:
> > Hi,
> >
> > On Tue, Oct 23, 2012 at 3:03 AM, Jan Hubicka wrote:
> > >> Ping.
> > >>
> > >>
> > >> On Wed, Oct 17, 2012 at 1:48 PM, Easwaran Raman
> > >> wrote:
> > >> > Hi,
> > >> > This patch fixes bugs introduced by my previous patch
On Wed, Oct 31, 2012 at 1:33 PM, Joern Rennecke
wrote:
> Quoting Richard Sandiford :
>
>>> It's about describing complex interactions of length adjustments that
>>> derive from branch shortening and length added for (un)alignment for
>>> scheduling purposes. Expressed naively, these can lead to c
1 - 100 of 160 matches
Mail list logo
