On Fri, May 10, 2013 at 02:55:07AM -0400, Shakthi Kannan wrote:
> Hi,
>
> The attached patch adds LAST_SOURCE_COLUMN to pp_verbatim
> function in the while loop present in
> diagnostic_report_current_module(). This makes the output
> consistent for any error parsing program as stated in the bug.
>
On 05/10/2013 03:55 AM, Alan Modra wrote:
On Thu, May 09, 2013 at 08:47:19AM -0700, Richard Henderson wrote:
On 02/22/2013 02:51 AM, Alan Modra wrote:
Richard, can I bother you with a review request?
http://gcc.gnu.org/ml/gcc-patches/2013-02/msg00544.html
Sorry for the truly excessive delay h
gcc/testsuite/ChangeLog
2013-05-10 Sebastian Huber
PR target/55033
* gcc.target/powerpc/pr55033.c: New.
---
gcc/testsuite/gcc.target/powerpc/pr55033.c | 21 +
1 files changed, 21 insertions(+), 0 deletions(-)
create mode 100644 gcc/testsuite/gcc.target/po
On Thu, 9 May 2013, Jakub Jelinek wrote:
> Hi!
>
> The bit_rotate: code in fold-const.c handles only a subset of rotates,
> and e.g. with C++ FE when sizeof isn't folded early fold_binary often
> even isn't called with something that could recognize the rotate pattern,
> even if bit_rotate: has b
On Tue, 7 May 2013, Richard Biener wrote:
>
> The following patch is a first step towards being able to enable
> vectorizing of a subset of all vectorizable functions at -O2 by
> default. Analysis of Polyhedron (loop heavy code) shows that
> the cost of doing vectorizer analysis is in the noise
This fixes PR57214 - the simple constant propagation pass we perform
over unrolled loop bodies does not properly avoid propagating across
abnormal edges.
Bootstrap & regtest running on x86_64-unknown-linux-gnu, I'll install
this on trunk and the 4.8 branch where the bug is latent.
Richard.
2013
I run into the following issue on the Fortran-Dev branch. I get an ICE due to a
tree-checking assert for fold-const.c's fold_indirect_ref_1, which has
tree min_val = size_zero_node;
if (type_domain && TYPE_MIN_VALUE (type_domain))
min_val = TYPE_MIN_VAL
On Thu, May 9, 2013 at 10:24 PM, Jakub Jelinek wrote:
> On Thu, May 09, 2013 at 09:47:49PM +0200, Uros Bizjak wrote:
>> On Thu, May 9, 2013 at 8:45 PM, Jakub Jelinek wrote:
>> > 2013-05-09 Jakub Jelinek
>> >
>> > * config/i386/i386.md (rotateinv): New code attr.
>> > (*3_1, *si
Hi,
- Original Message -
| From: "Marek Polacek"
|
| Maybe
| PR preprocessor/42014
| * diagnostic.c: Print LAST_SOURCE_COLUMN as well.
| ?
\--
Sure!
SK
--
Shakthi Kannan
skannan at redhat dot com
On Thu, May 9, 2013 at 12:55 AM, Marc Glisse wrote:
> Hello,
>
> here are a few more changes to fold-const.c so vectors can use the existing
> optimizations. Note that I made fold_truth_not_expr safe for use with vector
> BIT_NOT_EXPR.
>
> Passes bootstrap+testsuite on x86_64-linux-gnu.
>
> 2013-0
On Tue, May 7, 2013 at 11:26 PM, Marc Glisse wrote:
> Hello,
>
> this patch is about the use of TYPE_PRECISION for non-scalar types. For
> complex types, it is unused (always 0) and for vectors, the field is used to
> store the log of the number of elements, so in both cases we shouldn't use
> the
On Fri, 10 May 2013, Richard Biener wrote:
On Thu, May 9, 2013 at 12:55 AM, Marc Glisse wrote:
Hello,
here are a few more changes to fold-const.c so vectors can use the existing
optimizations. Note that I made fold_truth_not_expr safe for use with vector
BIT_NOT_EXPR.
Passes bootstrap+testsu
OK for 4.7?
Ok - I did say ok if no fallout in my original review :)
regards
Ramana
Thanks,
Matt
2013-01-04 Matthew Gretton-Dann
Backport from mainline.
2012-11-29 Matthew Gretton-Dann
PR target/54974
* config/arm/arm.md (thumb2_pool_range, pool_r
GCC 4.7 added some additional checks for type-bound procedure
overriding. However, doing so it weakened the check whether the nonpass
argument has the same type.
While for normal arguments, passing the parent type to an extended type
is fine, for overriding the type (of nonpass arguments) must
Chung-Lin Tang writes:
> +case UNSPEC:
> + /* Reach for a contained symbol. */
> + return nonzero_address_p (XVECEXP (x, 0, 0));
I don't think this is safe. UNSPECs really are unspecified :-),
so we can't assume that (unspec X) is nonzero simply because X is.
Richard
This patch makes the MIPS16 length checks use the style of predicates
and constraints that were added for microMIPS. Some constraints and
predicates are shared between the two ISA encodings, while others are
only useful for one or the other.
This is supposed to be a 1:1 change. I haven't tried t
Hi,
The following patch fixes an issue with the inline hardware multiply used in
the RL78 G13 target.
The G13 target does not have a multiply instruction, but achieves this using
set of data and control registers.
This patch adds an additional 'nop' after control and data registers are loaded
in t
> On Thu, May 09, 2013 at 03:52:20PM +0200, Martin Jambor wrote:
> > On Tue, May 07, 2013 at 11:43:33PM +0200, Jan Hubicka wrote:
> > > > Hi,
> > > >
> > > > the problem in PR 57084 is that late PRE devirtualization creates a
> > > > direct call to a decl fro which we only have an inlined call gra
> On 05/07/13 23:13, Teresa Johnson wrote:
> >--
> >Revised patch that fixes failures encountered when enabling
> >-freorder-blocks-and-partition, including the failure reported in PR 53743.
> >
> >This includes new verification code to ensure no cold blocks dominate hot
> >bloc
On Fri, May 10, 2013 at 12:57 AM, Xinliang David Li wrote:
> On Thu, May 9, 2013 at 3:40 PM, Steven Bosscher wrote:
>> This patch mixes up things badly from the point of
>> what-depends-on-what, the whole approach looks wrong to me.
>
>
> Do you mean the 'source file dependency' or 'logical depend
On Fri, May 10, 2013 at 10:44 AM, Tobias Burnus
wrote:
> I run into the following issue on the Fortran-Dev branch. I get an ICE due to
> a tree-checking assert for fold-const.c's fold_indirect_ref_1, which has
>
> tree min_val = size_zero_node;
> if (type_domain && TYP
On Fri, May 10, 2013 at 12:03 PM, Marc Glisse wrote:
> On Fri, 10 May 2013, Richard Biener wrote:
>
>> On Thu, May 9, 2013 at 12:55 AM, Marc Glisse wrote:
>>>
>>> Hello,
>>>
>>> here are a few more changes to fold-const.c so vectors can use the
>>> existing
>>> optimizations. Note that I made fol
On Thu, 9 May 2013, Sriraman Tallam wrote:
> Then, I plan to add the following hooks to libgcc (in a different patch) :
>
> int set_mock_cpu_is (const char *cpu);
> int set_mock_cpu_supports (const char *isa);
> int init_mock_cpu (); // Clear the values of the mock cpu.
Those names are in the us
> Hi!
>
> This is something I've noticed while working on the rotate recognizer
> patch I've just posted. We emit say
> roll %eax
> instead of
> roll $1, %eax
> because the former is shorter, but emit
> roll $31, %eax
> instead of the equivalent, but shorter
> rorl %eax
> The following pa
Hi!
Note, the patch has been already committed.
On Fri, May 10, 2013 at 03:48:50PM +0200, Jan Hubicka wrote:
> Going this route you will also need to update
>
> [(set_attr "type" "rotate1")
>(set (attr "length_immediate")
> (if_then_else
>(and (match_operand 1 "const1_operand"
2013/5/10 Sebastian Huber :
> gcc/testsuite/ChangeLog
> 2013-05-10 Sebastian Huber
>
> PR target/55033
> * gcc.target/powerpc/pr55033.c: New.
> ---
> gcc/testsuite/gcc.target/powerpc/pr55033.c | 21 +
> 1 files changed, 21 insertions(+), 0 deletions(-)
> c
Hi,
this is the issue about the signatures of the erase member functions of
the sequence containers. Mostly rather straightfoward stuff within the
limits of the current infrastructure: the various _M_const_case are
normally simple enough, I only mention the rather ugly std::vector one,
requir
This bug talks about ICEs trying to capture a VLA by value. We don't
support that, but we can give a more helpful error. In the comments
Paolo also noticed that with the new support for capture by reference,
we wrongly warn about returning a reference to the captured variable.
Tested x86_64-
Oddly, after you assign to a variable, you can't use its old value
anymore...
Tested x86_64-pc-linux-gnu, applying to trunk, 4.8, 4.7.
commit 16ee0ee2c495bc475e0ba015c6b9fd4707692196
Author: jason
Date: Fri May 10 14:17:37 2013 +
PR c++/57047
* semantics.c (cxx_fold_indirect_re
v2: Format changes
gcc/testsuite/ChangeLog
2013-05-10 Sebastian Huber
PR target/55033
* gcc.target/powerpc/pr55033.c: New.
---
gcc/testsuite/gcc.target/powerpc/pr55033.c | 23 +++
1 files changed, 23 insertions(+), 0 deletions(-)
create mode 100644 gcc/t
On Thu, May 9, 2013 at 2:42 PM, Diego Novillo wrote:
> On 2013-05-08 01:13 , Teresa Johnson wrote:
>>
>> Somehow Rietveld didn't upload the patch properly. I've attached the
>> patch to this email instead. Here is the description:
>
>
> Rietveld has turned out to be far less useful that I had hope
When I added the check for instantiation-dependent expressions, I
treated statement-expressions and some qualified-ids as
instantiation-dependent even when they don't involve template
parameters. In this PR, that causes problems because array
initialization expands into a statement-expression
Hi!
This patch allows us to vectorize attached testcases again, rot1.c
and rot3.c e.g. with -O3 -mavx, rot2.c only with -O3 -mavx2.
The problem with rot2.c is that the rotation count is there
vect_external_def, but when we insert some stmts into the pattern
def seq (the (-j) & 31), we turn it int
Hi!
I've apparently missed one spot, where store of zero value was assumed,
but not actually verified. Fixed thusly, bootstrapped/regtested on
x86_64-linux and i686-linux, ok for trunk/4.8/4.7?
2013-05-10 Jakub Jelinek
PR tree-optimization/57230
* tree-ssa-strlen.c (handle_ch
Hi!
Related to the previously posted patch, I've noticed that
on the attached testcase we don't optimize the strlen into
constant, even when we easily could.
Implemented thusly, bootstrapped/regtested on x86_64-linux and i686-linux,
ok for trunk?
2013-05-10 Jakub Jelinek
PR tree-opti
On Fri, May 10, 2013 at 4:52 AM, Jan Hubicka wrote:
>> On 05/07/13 23:13, Teresa Johnson wrote:
>> >--
>> >Revised patch that fixes failures encountered when enabling
>> >-freorder-blocks-and-partition, including the failure reported in PR 53743.
>> >
>> >This includes new veri
Hi!
Our rotate expansion if rotate optab isn't present for the respective
mode looks unsafe for rotations by variable count if that count could
be 0, because then it invokes right or left shift by bitsize.
While on most targets the hw will probably do the right thing
(it is fine if x << 32 will ei
Committed to 4.7, 4.8 and head.
On 5/8/2013 5:06 AM, Sebastian Huber wrote:
This patch should be applied to all open GCC branches.
This patch removes the mthumb/march=armv7 since it is virtually equal to
the mthumb/march=armv7-m multilib. Add variants for ARMv7-A and
ARMv7-R.
I tried to use t
> -Original Message-
> From: Ramana Radhakrishnan
> Sent: Friday, May 10, 2013 18:13
> To: Matthew Gretton-Dann
> Cc: gcc-patches@gcc.gnu.org; Richard Earnshaw; d...@canonical.com; Patch
> Tracking; Richard Biener; Joey Ye
> Subject: Re: [RFA/ARM/4.7] Fix PR54974: Thumb literal pools don't
Analyzing code generated for PPC with LRA, I found that sometimes
call-saved registers are overused. It creates long prologs/epilogs.
The following patch fixes it.
The problem was in that BB frequency might be zero.
Committed to trunk as rev. 198792.
The same patch was committed into lra-mi
On Thu, May 9, 2013 at 3:40 PM, Steven Bosscher wrote:
> On Thu, May 9, 2013 at 11:42 PM, Diego Novillo wrote:
>> On 2013-05-08 01:13 , Teresa Johnson wrote:
>>> -static void
>>> +void
>>> emit_barrier_after_bb (basic_block bb)
>>> {
>>>rtx barrier = emit_barrier_after (BB_END (bb));
>>> -
Hi Tobias,
> GCC 4.7 added some additional checks for type-bound procedure overriding.
> However, doing so it weakened the check whether the nonpass argument has the
> same type.
>
> While for normal arguments, passing the parent type to an extended type is
> fine, for overriding the type (of nonp
> "Shakthi" == Shakthi Kannan writes:
Shakthi> 2013-05-10 Shakthi Kannan
Shakthi> PR preprocessor/42014
Shakthi> * gcc/diagnostic.c: Added LAST_SOURCE_COLUMN in while loop.
You should mention the function name in there.
See the GNU Coding Standards for the format.
Shakthi> -
Hi Janus,
Janus Weil wrote:
in principle I think your idea to tighten up the type check by
symmetrizing it is ok.
I actually kind of copied it from the proc-pointer assignment check,
which does the same.
However, I'm not sure if the place where you do it is the most
preferable: I think it
Hi,
I thought I was hitting the ipa-inline-transform.c:263 assert with
this patch when LTO-building Mozilla Firefox but it turned out it was
caused by a different patch so I'm handling that there.
On Thu, May 02, 2013 at 02:37:26PM +0200, Jan Hubicka wrote:
> > 2013-03-22 Martin Jambor
> >
>
.. looks like the prettyprinting code needs again adjustments:
FAIL: libstdc++-prettyprinters/cxx11.cc print uom
FAIL: libstdc++-prettyprinters/cxx11.cc print uomm
FAIL: libstdc++-prettyprinters/cxx11.cc print uos
FAIL: libstdc++-prettyprinters/cxx11.cc print uoms
Thanks,
Paolo.
Hi,
I have noticed that whereas all of the dumps in ipa-inline*, ipa-cp.c,
ipa-prop.c and elsewhere dump node UIDs, function dump_cgraph_node,
which is the core of call graph dumping, dumps node->symbol.uid. This
unnecessarily makes it more difficult to see what has happened to a
node in the grap
Hi,
>> However, I'm not sure if the place where you do it is the most
>> preferable: I think it should rather go inside of
>> check_dummy_characteristics (since any check for the dummy
>> characteristics should include the 'strict' type check, and not only a
>> check of type compatibility, cf. F08
Hi,
PR 57084 has alerted me that IN LTO we do not devirtualize calls
during ipa-cp and ipa-inline that we do in non-LTO compilations. The
reason is that the type comparison in get_binfo_at_offset is failing,
because the two types do not have the same TYPE_MAIN_VARIANT but
TYPE_CANONICAL should be
> Hi!
>
> Note, the patch has been already committed.
>
> On Fri, May 10, 2013 at 03:48:50PM +0200, Jan Hubicka wrote:
> > Going this route you will also need to update
> >
> > [(set_attr "type" "rotate1")
> >(set (attr "length_immediate")
> > (if_then_else
> >(and (match_oper
> OK, the updated patch is below (the only change is the extra asserts),
> re-bootstrapped and re-tested on x86_64-linux. I will commit it on
> Monday if there are no objections.
No objections from my side. Path looks OK.
Thanks,
Honza
> 2013-05-10 Martin Jambor
>
> * tree.c (get_binfo_at_offset): Use types_compatible_p to compare
> types.
>
> Index: src/gcc/tree.c
> ===
> --- src.orig/gcc/tree.c
> +++ src/gcc/tree.c
> @@ -11483,7 +11483,7 @@ get_bin
Hi,
as we discover targets of previously indirect calls in ipa-inline and
ipa-cp, we sometimes figure out that the targets are not a function.
One typical example is when NULL is passed as a function pointer
parameter, another is when C++ member-pointer points to a virtual
function and the overloa
Now we don't store the module info if the module is not exported or
has any aux module (to compress the profile data size). Thus it's
normal that a primary module entry cannot be found. This patch
suppresses the messages printed when the primary module is not found.
Bootstrapped and passed regress
Hi,
On Fri, May 10, 2013 at 07:24:06PM +0200, Jan Hubicka wrote:
> > 2013-05-10 Martin Jambor
> >
> > * tree.c (get_binfo_at_offset): Use types_compatible_p to compare
> > types.
> >
> > Index: src/gcc/tree.c
> > ===
> >
On 05/10/2013 11:27 AM, Martin Jambor wrote:
Hi,
as we discover targets of previously indirect calls in ipa-inline and
ipa-cp, we sometimes figure out that the targets are not a function.
One typical example is when NULL is passed as a function pointer
parameter, another is when C++ member-point
Is it only auto fdo that doesn't store the module info if the module
is not exported or has aux modules? Note that this will prevent usage
of my script that enables annotation of gcov dump info with function
names, which relies on accessing the primary module info to obtain the
module name (which i
On Fri, May 10, 2013 at 6:34 AM, Joseph S. Myers
wrote:
> On Thu, 9 May 2013, Sriraman Tallam wrote:
>
>> Then, I plan to add the following hooks to libgcc (in a different patch) :
>>
>> int set_mock_cpu_is (const char *cpu);
>> int set_mock_cpu_supports (const char *isa);
>> int init_mock_cpu ();
On Fri, May 10, 2013 at 10:47 AM, Teresa Johnson wrote:
> Is it only auto fdo that doesn't store the module info if the module
> is not exported or has aux modules? Note that this will prevent usage
> of my script that enables annotation of gcov dump info with function
> names, which relies on acc
ok. Would be nicer if there is a way to tell this from other error cases though.
David
On Fri, May 10, 2013 at 11:00 AM, Dehao Chen wrote:
> On Fri, May 10, 2013 at 10:47 AM, Teresa Johnson wrote:
>> Is it only auto fdo that doesn't store the module info if the module
>> is not exported or has
Hi
In function dw2_output_indirect_constant_1 a new var decl is created. Only
When the variable is not PUBLIC it is allocated static storage. Does anybody
know why the variable is not allocated static storage by marking TREE_STATIC
when it is PUBLIC?
The following patch marks the STATIC flag in a
On Fri, May 10, 2013 at 11:00 AM, Dehao Chen wrote:
> On Fri, May 10, 2013 at 10:47 AM, Teresa Johnson wrote:
>> Is it only auto fdo that doesn't store the module info if the module
>> is not exported or has aux modules? Note that this will prevent usage
>> of my script that enables annotation of
Martin Jambor wrote:
>Hi,
>
>On Fri, May 10, 2013 at 07:24:06PM +0200, Jan Hubicka wrote:
>> > 2013-05-10 Martin Jambor
>> >
>> >* tree.c (get_binfo_at_offset): Use types_compatible_p to compare
>> >types.
>> >
>> > Index: src/gcc/tree.c
>> > =
Right now, the error was prevented from the profile_creation tool
side. In GCC, we only assume that the all exported modules will have
primary module entry stored in profile. We can also check it when
reading in profile, but that might be too costly because every
compilation (even for those files t
On Fri, May 10, 2013 at 11:32 AM, Teresa Johnson wrote:
> On Fri, May 10, 2013 at 11:00 AM, Dehao Chen wrote:
>> On Fri, May 10, 2013 at 10:47 AM, Teresa Johnson
>> wrote:
>>> Is it only auto fdo that doesn't store the module info if the module
>>> is not exported or has aux modules? Note that
Hello!
2013-05-10 Uros Bizjak
* config/i386/i386.md (memory): Handle sseishft1.
* config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
(*vec_extractv2di_1): Ditto.
Tested on x86_64-pc-linux-gnu {,-m32}, committed to mainline SVN.
Uros.
Index: i386.md
==
> Martin Jambor wrote:
>
> >Hi,
> >
> >On Fri, May 10, 2013 at 07:24:06PM +0200, Jan Hubicka wrote:
> >> > 2013-05-10 Martin Jambor
> >> >
> >> > * tree.c (get_binfo_at_offset): Use types_compatible_p to compare
> >> > types.
> >> >
> >> > Index: src/gcc/tree.c
> >> > =
On May 10, 2013, at 5:12 AM, Richard Biener wrote:
> Somewhere we talked about naming new C++ files wide-int.cc
> (or was it .C or .cxx?)
for the C++ library:
$ find . -name \*.cc | wc -l
5605
$ find . -name \*.C | wc -l
1
:-)
Go also prefers .cc.
> movwax, %h2
> movw0x2, ax ; MDAH
> nop ; mdc += mdal * mdah
> + nop ; Additional nop for MAC
>
> mov a, #0x40
> mov !0xf00e8, a ; MDUC
Shouldn't the MOV after the nop add the extra cycle here?
> movwax, %H2
>
Jan Hubicka wrote:
>> Martin Jambor wrote:
>>
>> >Hi,
>> >
>> >On Fri, May 10, 2013 at 07:24:06PM +0200, Jan Hubicka wrote:
>> >> > 2013-05-10 Martin Jambor
>> >> >
>> >> > * tree.c (get_binfo_at_offset): Use types_compatible_p to
>compare
>> >> > types.
>> >> >
>> >> > Ind
Hi
sparc64*-*-rtems* ends up with __svr4__ defined. The attached
patch corrects that.
OK to apply?
013-05-10 Joel Sherrill
* config.gcc (sparc64*-*-rtems*): Use sp64-rtemself.h.
RTEMS target should not have -D__svr4__ in CPP_SUBTARGET_SPEC.
* config/sparc/sp64-rtemse
On Fri, May 10, 2013 at 5:54 PM, Teresa Johnson wrote:
> The main issue I had here, and why I made this change, is that we go
> in and out of cfglayout mode several times after bb partitioning and
> then out_of_cfglayout. The problem was that when we subsequently went
> in and out of cfglayout mode
On Fri, 10 May 2013, Sriraman Tallam wrote:
> On Fri, May 10, 2013 at 6:34 AM, Joseph S. Myers
> wrote:
> > On Thu, 9 May 2013, Sriraman Tallam wrote:
> >
> >> Then, I plan to add the following hooks to libgcc (in a different patch) :
> >>
> >> int set_mock_cpu_is (const char *cpu);
> >> int set_
> There shouldn't be any forwarder blocks in cfg layout mode. What did
> you need this for?
>
> BTW2: We badly need to figure out a way to create test cases for FDO... :-(
We have gcc.dg/tree-prof and friends. What do you need to add?
Honza
On Fri, May 10, 2013 at 11:10 PM, Jan Hubicka wrote:
>> There shouldn't be any forwarder blocks in cfg layout mode. What did
>> you need this for?
>>
>> BTW2: We badly need to figure out a way to create test cases for FDO... :-(
>
> We have gcc.dg/tree-prof and friends. What do you need to add?
I have a instruction pattern for my port that reads in part:
[(parallel [(set (subreg:DI (match_operand:TI 0 "register_operand" "=r") 0)
(mem:DI (plus:DI (match_operand:DI 1 "register_operand" "r")
(match_operand:DI 2 "const_int_operand"
""
On Fri, May 10, 2013 at 11:51 PM, Mike Stump wrote:
> - if (DF_REF_LOC (*ref) == recog_data.operand_loc[op])
> - break;
> + {
> + if (DF_REF_LOC (*ref) == recog_data.operand_loc[op])
> + break;
> + /* DF_REF_LOC can be (subreg:DI (reg:TI 5) 8) and
> +
On Fri, May 10, 2013 at 11:51 PM, Mike Stump wrote:
> I have a instruction pattern for my port that reads in part:
>
> [(parallel [(set (subreg:DI (match_operand:TI 0 "register_operand" "=r") 0)
>(mem:DI (plus:DI (match_operand:DI 1 "register_operand"
> "r")
>
On May 10, 2013, at 3:29 PM, Steven Bosscher wrote:
> Your web.c patch looks correct to me, but I can't approve it.
Thanks. Now that you point out the DF accessor, it all makes perfect sense.
:-) I've fixed one other instance that has to be as wrong for all the same
reasons. Though my port
On 05/10/2013 08:49 AM, Jakub Jelinek wrote:
Hi!
Related to the previously posted patch, I've noticed that
on the attached testcase we don't optimize the strlen into
constant, even when we easily could.
Implemented thusly, bootstrapped/regtested on x86_64-linux and i686-linux,
ok for trunk?
20
On 05/10/2013 08:53 AM, Jakub Jelinek wrote:
Hi!
Our rotate expansion if rotate optab isn't present for the respective
mode looks unsafe for rotations by variable count if that count could
be 0, because then it invokes right or left shift by bitsize.
While on most targets the hw will probably do
Hello,
This unch_memory in struct resources is a left-over from
RTX_UNCHANGING_P, but it looks like the change-over to MEM_READONLY_P
was done incorrectly: The resource_conflicts_p code now reports
conflicts for insns reading readonly memory and insns reading "normal"
memory or no memory at all.
> > Note that I had to make a few changes (fixes?) in the MI portions of
> > gcc to avoid problems I encountered, I don't know if these changes are
> > "correct" or if there are better ways to avoid those cases. Those
>
> In any case, they should best be posted in separate messages, each one
>
> > Note that I had to make a few changes (fixes?) in the MI portions of
> > gcc to avoid problems I encountered, I don't know if these changes are
> > "correct" or if there are better ways to avoid those cases. Those
>
> In any case, they should best be posted in separate messages, each one
>
>
> The existing check should work ok with lto. If not then we should figure out
> why we do not merge the main variants properly.
Hmm, adding:
Index: tree.c
===
--- tree.c (revision 198796)
+++ tree.c (working copy)
@@ -11
> seems to bring a lot of positives and by quick inspection most of them seem
> like same classes. I will try to find some time to debug this more. But
> glancing over the the dumps, I see many of them just have different name
> spaces. Do we even attempt to merge namespace_decl? How types from s
Assuming the patch has been tested on a public port, it is ok for commit.
kenny
On 05/10/2013 06:52 PM, Mike Stump wrote:
On May 10, 2013, at 3:29 PM, Steven Bosscher wrote:
Your web.c patch looks correct to me, but I can't approve it.
Thanks. Now that you point out the DF accessor, it all m
On Fri, May 10, 2013 at 5:12 AM, Richard Biener wrote:
>
> Somewhere we talked about naming new C++ files wide-int.cc
> (or was it .C or .cxx?)
Definitely not .C.
I prefer .cc.
Ian
On Fri, May 10, 2013 at 2:00 PM, Steven Bosscher wrote:
> On Fri, May 10, 2013 at 5:54 PM, Teresa Johnson wrote:
>> The main issue I had here, and why I made this change, is that we go
>> in and out of cfglayout mode several times after bb partitioning and
>> then out_of_cfglayout. The problem was
2013/5/10 Sebastian Huber :
> v2: Format changes
>
> gcc/testsuite/ChangeLog
> 2013-05-10 Sebastian Huber
>
> PR target/55033
> * gcc.target/powerpc/pr55033.c: New.
> ---
> gcc/testsuite/gcc.target/powerpc/pr55033.c | 23 +++
> 1 files changed, 23 insertion
90 matches
Mail list logo
