On Fri, 16 Dec 2016 02:07:07 +0100, Jonathan Wakely wrote:
> On 15/12/16 22:19 +0100, Jan Kratochvil wrote:
> > Just with the GDB 'compile' project (libcc1) which is planned to be used for
> > all GDB expressions evalation the Xmethods will no longer work.
>
> But then *it can just get compiled, s
On Thu, Dec 15, 2016 at 06:54:43PM -0700, Jeff Law wrote:
>unsigned cnt = 0;
> + bitmap live_bytes = NULL;
> + bitmap orig_live_bytes = NULL;
>
>*use_stmt = NULL;
>
> + /* REF is a memory write. Go ahead and get its base, size, extent
> + information and encode the bytes written
On 12/14/2016 09:19 PM, Jeff Law wrote:
On 12/14/2016 03:56 PM, Martin Sebor wrote:
The -Wnonnull warning improvement (PR c/17308 - nonnull attribute
not as useful as it could be) causes a couple of false positives
in a powerpc64le bootstrap. The attached fix suppresses them.
It passes bootstra
Hi all,
The attached patch regression tested on x86-64-linux.
This adds a static function to allocate and extract the DT format string that
may contain doubled quotes.
OK for trunk, test case also attched.
Jerry
2016-12-16 Jerry DeLisle
PR fortran/78622
* io.c (format_le
This is the final patch in the kit to improve our DSE implementation.
It's based on a observation by Richi. Namely that a read from bytes of
memory that are dead can be ignored. By ignoring such reads we can
sometimes find additional stores that allow us to either eliminate or
trim an earlie
This is the second patch in the kit to improve our DSE implementation.
This patch recognizes when a CONSTRUCTOR assignment could be trimmed at
the head or tail because those bytes are dead.
The first implementation of this turned the CONSTRUCTOR into a memset.
This version actually rewrites t
This is the 3rd patch in the kit to improve our DSE implementation.
This patch supports trimming of the head or tail of a memset, memcpy or
memmove call. It's conceptually similar to trimming CONSTRUCTORS (and
was in fact developed first).
Try as I might, I couldn't find a BZ in our database
This is the first of the 4 part patchkit to address deficiencies in our
DSE implementation.
This patch addresses the P2 regression 33562 which has been a low
priority regression since gcc-4.3. To summarize, DSE no longer has the
ability to detect an aggregate store as dead if subsequent st
This is a 4 part patchkit to address various deficiencies in our DSE
implementation.
BZ33562 was the inspiration for this work. 33562 is a low priority
regression that's been around for a long time. Patch #1 addresses
33562, "aggregate DSE disabled" and also implements trimming of complex
a
On 15/12/16 22:19 +0100, Jan Kratochvil wrote:
On Thu, 15 Dec 2016 15:18:17 +0100, Jonathan Wakely wrote:
I'm going to add Xmethods for all our iterator types so that it will
always be possible to do "print *iter", so if GDB supports Xmethods
then we don't need to register the iterator printers.
This patch by Than McIntosh fixes some comments and field names in the
compiler to match the current libgo sources. Bootstrapped and ran Go
testsuite on x86_64-pc-linux-gnu. Committed to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
On Thu, Dec 15, 2016 at 2:47 PM, Ian Lance Taylor wrote:
> The Go frontend needs to call determine_types even for constant
> expressions, which it was not doing. The problem is that a constant
> expression may include code like unsafe.Sizeof(0). Something needs to
> determine the type of the unt
This patch by Than McIntosh fixes an off-by-one error in the array
length of the value we create for a GC symbol. This was not serious,
it just wasted an element. Bootstrapped and ran Go testsuite on
x86_64-pc-linux-gnu. Committed to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
The Go frontend needs to call determine_types even for constant
expressions, which it was not doing. The problem is that a constant
expression may include code like unsafe.Sizeof(0). Something needs to
determine the type of the untyped 0, and that should be the
determine_types pass.
Implementing
Hello!
Attached patch improves ffs expandsion for TARGET_BMI targets.
Compared to bsf, tzcnt is noticeably faster on AMD processors.
However, since generic target enables TARGET_AVOID_FALSE_DEP_FOR_BMI,
we always expand ffs with bsf, even when using -mbmi.
Attached patch enables TARGET_AVOID_FALS
On Fri, Dec 09, 2016 at 12:52:03PM +, Richard Sandiford wrote:
> This patch renames enum machine_mode to enum machine_mode_enum
> and adds a machine_mode wrapper class.
>
> The previous patch mechanically replaced mode names in case
> statements; this one updates other places that should conti
On Thu, 15 Dec 2016 15:18:17 +0100, Jonathan Wakely wrote:
> I'm going to add Xmethods for all our iterator types so that it will
> always be possible to do "print *iter", so if GDB supports Xmethods
> then we don't need to register the iterator printers.
Just with the GDB 'compile' project (libcc
On 30/11/16 17:39, Wilco Dijkstra wrote:
> The negdi2 patterns for ARM and Thumb-2 are duplicated because Thumb-2
> doesn't support RSC with an immediate. We can however emulate RSC with
> zero using a shifted SBC. If we add this to subsi3_carryin the negdi
> patterns can be merged, simplifying t
2016-12-15 21:41 GMT+01:00 Steve Kargl :
> On Thu, Dec 15, 2016 at 08:38:47PM +0100, Janus Weil wrote:
>> 2016-12-13 19:55 GMT+01:00 Janus Weil :
>> > 2016-12-13 19:19 GMT+01:00 Janne Blomqvist :
>> >> On Tue, Dec 13, 2016 at 8:13 PM, Janus Weil wrote:
>> >>> Hi all,
>> >>>
>> >>> here is a straig
On 30/11/16 17:32, Wilco Dijkstra wrote:
> Thumb uses a special register allocation order to increase the use of low
> registers. Oddly enough, LR appears before R12, which means that LR must
> be saved and restored even if R12 is available. Swapping R12 and LR means
> this simple example now use
On 12/15/2016 05:51 PM, Jakub Jelinek wrote:
This patch fixes what I've found quickly, bootstrapped/regtested on
x86_64-linux and i686-linux, ok for trunk?
Ok.
Bernd
Hi
Here is a fix in the usage of a concept check. There are also many
testsuite failures when using concept checks but this one forbids the
library from being built. I know that concept checks are not really
maintained so maybe the fix is to simply remove those checks.
* include/ext/ran
On Thu, Dec 15, 2016 at 08:38:47PM +0100, Janus Weil wrote:
> 2016-12-13 19:55 GMT+01:00 Janus Weil :
> > 2016-12-13 19:19 GMT+01:00 Janne Blomqvist :
> >> On Tue, Dec 13, 2016 at 8:13 PM, Janus Weil wrote:
> >>> Hi all,
> >>>
> >>> here is a straightforward cleanup patch that makes a few function
Tested on Linux-x64.
2016-12-15 Ville Voutilainen
Implement LWG 2769, Redundant const in the return type of
any_cast(const any&).
* include/std/any (_AnyCast): New.
(any_cast(const any&)): Use it and add an explicit cast for return.
(any_cast(any&)): Likewise.
(any_cast
OK.
On Thu, Dec 15, 2016 at 7:38 AM, Nathan Sidwell wrote:
> 77585 concerns the instantiation of a generic lambda that contains a call to
> a non-dependent non-static member function.
>
> auto lam = [&](auto) { return Share (); };
> r += Eat (lam); // instantation of lambda::operator() here
2016-12-13 19:55 GMT+01:00 Janus Weil :
> 2016-12-13 19:19 GMT+01:00 Janne Blomqvist :
>> On Tue, Dec 13, 2016 at 8:13 PM, Janus Weil wrote:
>>> Hi all,
>>>
>>> here is a straightforward cleanup patch that makes a few functions
>>> return a bool instead of an int. Regtests cleanly on x86_64-linux-
OK.
On Thu, Dec 15, 2016 at 11:44 AM, Jakub Jelinek wrote:
> On Thu, Dec 15, 2016 at 07:40:58AM -0500, Nathan Sidwell wrote:
>> On 12/15/2016 07:26 AM, Jakub Jelinek wrote:
>>
>> > I don't think so. complete_type (error_mark_node) returns error_mark_node,
>> > and COMPLETE_TYPE_P (error_mark_nod
On Thu, 15 Dec 2016, Tamar Christina wrote:
> > Note that on some systems we even disable 64bit floating point support.
> > I suspect this check needs a little re-thinking as I don't think that
> > checking for a specific UNITS_PER_WORD is correct, nor is checking the
> > width of the type. I'm n
2016-12-15 19:51 GMT+03:00 Uros Bizjak :
> On Thu, Dec 15, 2016 at 2:31 PM, Andrew Senkevich
> wrote:
>> 2016-12-14 22:55 GMT+03:00 Uros Bizjak :
>>> On Wed, Dec 14, 2016 at 8:04 PM, Andrew Senkevich
>>> wrote:
>>>
here is the second part of k-mask intrinsics, is it Ok?
>>>
--- a/gcc/co
On 15/12/16 13:11 -0500, David Malcolm wrote:
On Thu, 2016-12-15 at 17:39 +, Jonathan Wakely wrote:
This patch tries to improve the user experience when debugging
container iterators, for cases where some of the typedefs used by the
printers are not in the debuginfo, so gdb.lookup_type() cal
On Thu, Dec 15, 2016 at 05:51:25PM +0100, Jakub Jelinek wrote:
> Hi!
>
> When looking at the noipa attribute, I've been initially changing
> ipa-cp.c, and noticed some bad spellings (various functions called *accross*
> rather than *across*) and tons of bad formatting, sometimes e.g. indentation
>
On Thu, 2016-12-15 at 17:39 +, Jonathan Wakely wrote:
> This patch tries to improve the user experience when debugging
> container iterators, for cases where some of the typedefs used by the
> printers are not in the debuginfo, so gdb.lookup_type() calls fail.
> That happens if the iterator's o
On 15 December 2016 at 19:39, Jonathan Wakely wrote:
> This patch tries to improve the user experience when debugging
> container iterators, for cases where some of the typedefs used by the
> printers are not in the debuginfo, so gdb.lookup_type() calls fail.
> That happens if the iterator's opera
This patch tries to improve the user experience when debugging
container iterators, for cases where some of the typedefs used by the
printers are not in the debuginfo, so gdb.lookup_type() calls fail.
That happens if the iterator's operator*() and operator->() haven't
been instantiated, or if they
Hi!
When looking at the noipa attribute, I've been initially changing
ipa-cp.c, and noticed some bad spellings (various functions called *accross*
rather than *across*) and tons of bad formatting, sometimes e.g. indentation
by 3 or -1 columns etc.
This patch fixes what I've found quickly, bootstr
On Thu, Dec 15, 2016 at 2:31 PM, Andrew Senkevich
wrote:
> 2016-12-14 22:55 GMT+03:00 Uros Bizjak :
>> On Wed, Dec 14, 2016 at 8:04 PM, Andrew Senkevich
>> wrote:
>>
>>> here is the second part of k-mask intrinsics, is it Ok?
>>
>>> --- a/gcc/config/i386/sse.md
>>> +++ b/gcc/config/i386/sse.md
>>
On Thu, Dec 15, 2016 at 07:40:58AM -0500, Nathan Sidwell wrote:
> On 12/15/2016 07:26 AM, Jakub Jelinek wrote:
>
> > I don't think so. complete_type (error_mark_node) returns error_mark_node,
> > and COMPLETE_TYPE_P (error_mark_node) is invalid (should fail TYPE_CHECK in
> > checking compiler).
>
On Wed, Dec 14, 2016 at 01:27:56PM +0100, Richard Biener wrote:
> Ah, ok. But then the size of the memset shouldn't be compared
> against the get_ref_base_and_extend size from src2 but to the
> size of the access of SRC/DEST (clearly looking at the "size" of
> the ADDR_EXPR argument is somewhat bo
On 12/15/2016 03:27 AM, Jakub Jelinek wrote:
On Thu, Dec 15, 2016 at 11:23:14AM +0100, Andreas Schwab wrote:
On Dez 15 2016, Jakub Jelinek wrote:
So, shall we change also the first 3?
Yes, I'd think so.
So here is it in patch form. Is this ok for trunk?
2016-12-15 Jakub Jelinek
On 15/12/16 09:55, Richard Earnshaw (lists) wrote:
On 30/11/16 16:47, Kyrill Tkachov wrote:
Hi all,
In this awkward ICE we have a *load_multiple pattern that is being
transformed in reload from:
(insn 55 67 151 3 (parallel [
(set (reg:SI 0 r0)
(mem/u/c:SI (reg/f:S
This patch adds a cmpstrnsi pattern for rs6000 target to provide
builtin expansion of strncmp(). Perf tests on a power8 system show that
it is 3-10x faster than the glibc strncmp on RHEL7 for lengths under 64
bytes.
Bootstrap/regtest has passed on powerpc64le, in progress for powerpc64,
ok for tru
Ping.
From: Aditya Kumar
Sent: Wednesday, December 7, 2016 11:46 AM
To: libstd...@gcc.gnu.org
Cc: gcc-patches@gcc.gnu.org; hiradi...@msn.com; Aditya Kumar
Subject: [PATCH] improve string find algorithm
Here is an improved version of basic_string::find. The idea is to
split the string fin
Remove the VFP_REGS field by converting its meanings into flag
attributes. The new flag attributes build on each other describing
increasing capabilities. This allows us to do a better job when
inlining functions with differing requiremetns on the fpu environment:
we can now inline A into B if B
Now that everything uses the new ISA features, we can remove the
FEATURES field from the FPU descriptions, along with all the macros
and definitions associated with it.
* arm-fpus.def (ARM_FPU): Remove features field from all definitions.
* arm.h (arm_fpu_feature_set): Delete type
Now we finally have the infrastructure in place we can now derive
details of the FPU from a CPU entry. This patch enables this for the
existing cores that already have an explicit FPU in their product names.
* arm-fpus.def: Add CNAME field to all FPU definitions.
* genopt.sh: Use
It now becomes apparent that it would be better to use the the
cl_target_options as the basis for calling arm_configure_build_target;
it already contains exactly the same fields that we need. I chose not
to rewrite the earlier patches as that would make the progression of
changes seem less logica
Now that we can construct the build target isa from the cl_target_options
data we can use this to determine inlinability. This eliminates the
final remaining use of the FPU features field.
* arm.c (arm_can_inline_p): Use ISA features for determining
inlinability.
---
gcc/config/
Similar to the main ISA, we convert the FPU revision into a set of feature
bits. This permits a more complex set of capability relationships to be
expressed more easily. For now we continue to use the traditional bitmasks.
* arm.h (FPU_FL_VFPv2) New feature bit.
(FPU_FL_VFPv3, F
Now that the isa feature bits are all available in arm_active_target
we can use that for most of the feature tests that are needed.
* arm.h (TARGET_VFPD32): Use arm_active_target.
(TARGET_VFP3): Likewise.
(TARGET_VFP5): Likewise.
(TARGET_VFP_SINGLE): Likewise.
Rather than assuming a specific fpu name has been selected, we work
out the FPU from the ISA properties. This is necessary since once we
have default FPUs selected by the processor, there will be no explicit
entry in the table of fpus to refer to.
This also fixes a bug with the code I added rece
Similar to the new CPU and architecture ISA feature lists, we now add
similar capabilities to each FPU description. We don't use these yet,
that will come in later patches. These follow the same style as the
newly modified flag sets, but use slightly different defaults that
more accurately refle
Now that we can describe the FPU with the standard ISA bits we need to
initialize them. However, the FPU settings can be changed with target build
attributes, so we also need to reset them if things change. This requires
a bit of juggling about with the existing code to ensure that the active
ta
This converts the recently added implicit -mthumb support code to use
the new data structures. Since we have a very simple query and no
initialized copies of the sbitmaps, for now we simply scan the list of
features to look for the one of interest.
* arm-opts.h (struct arm_arch_core_flag
Nothing uses the old feature sets now, so we can delete them entirely.
* arm-cores.def: Remove FLAGS field from all core definitions.
* arm-arches.def: Likewise.
* arm-opts.h (enum processor_type): Remove FLAGS parameter from
ARM_CORES macro.
(arm_arch_core
The arm_fp_model enumeration type has only had one useful value since
the FPA support was removed, and it's no-longer used anywhere. This
patch just cleans that up by removing it.
* arm.h (arm_fp_model): Delete.
---
gcc/config/arm/arm.h | 8
1 file changed, 8 deletions(-)
dif
This patch finishes the job of removing insn_flags and moves the logic
over to using the new data structures. I've added a new boolean
variable to detect when we have ARMv7ve-like capabilities and thus
have 64-bit atomic operations since that would be a complex query and
expensive to do in full.
This patch uses the new ISA data structure to determine which builtins
to add. It entirely eliminates the need for insn_flags to be a global
variable, but we're about to delete that in the following patches, so
for now we leave it as a global.
* arm-builtins.c: Include sbitmap.h.
With the new data structures it is trivial to add a new field and we
aren't (too) limited as to the number we have. This patch adds a new
bit to describe the need for a particular compiler behaviour
modification: in this case a quirk in the cortex-m3.
* arm-isa.h (enum isa_feature): Add
We now start to make more use of the new data structure. This allows
us to eliminate two of the existing static variables,
arm_selected_arch and arm_selected tune.
* arm.c (arm_selected_tune): Delete static variable.
(arm_selected_arch): Likewise.
(arm_configure_build_tar
This patch creates a new data structure for carrying around the data
relating to the current compilation target. The idea behind this is
that this data structure can be updated to reflect the overall
compilation target as new information is gathered (from command line
options) or architectural ex
Make more use of the new data structure for initializing existing
variables.
* arm.c (arm_option_override): Use arm_active_target as source of
information for arm_base_arch and arm_arch_name.
* (arm_file_start): Use arm_active_target for core name.
---
gcc/config/arm/arm.
This patch adds the new ISA data structures. The idea is to use an
sbitmap for carrying these around internally. We don't make much use
of this yet, but will increasingly migrate over to this in the
following patches. All cores and architectures currently have both
old and new encodings for now
We start out by separating the 'tuning flags' in a CPU or architecture
specification into a new field in the data structures. Because there
aren't very many of these (and we'd like to get rid of them entirely,
eventually, moving to entries in the tuning tables), we just use a
simple unsigned word.
As discussed at this year's Cauldron, it has concerned me for a while
now that when a user of the ARM compiler specifies a CPU they also
need to specify which floating-point unit it has (even though the
choice is almost invariably one).
This patch implements the ability to make the selection autom
On Thu, Dec 15, 2016 at 01:57:06PM +0100, Dominik Vogt wrote:
> > > The patch hasn't got a lot of testing yet as I'd like to hear your
> > > opinion on the patch first.
> >
> > I am testing it on powerpc. Please also test on x86?
> >
> > > gcc/ChangeLog-signextend-1
> > >
> > > * combine.c (e
On 14/12/16 22:49 +0100, François Dumont wrote:
On 09/12/2016 16:18, Jonathan Wakely wrote:
But I don't know how to fix this so for the moment I just adapt it
to correctly handle std::__7::string.
But that's not correct. Please try to understand the point I'm making:
The name "std::__7::stri
Hi,
The 32-bit insert.d case in msa-builtins.c is failing with O2 and Os because
the order of the emitted instructions is slightly different compared to the
other optimization levels.
This patch tweaks the regular expression for 32-bit insert.d to accept the
alternate instruction order.
Tested
On 14/12/16 22:49 +0100, François Dumont wrote:
@@ -1321,7 +1328,7 @@ def register_type_printers(obj):
if not _use_type_printing:
return
-for pfx in ('', 'w'):
+for pfx in ('', 'w', vers_nsp, vers_nsp + 'w'):
add_one_type_printer(obj, 'basic_string', pfx + 'string')
This is another partial fix for PR 59170, this time adding checks for
normal mode iterators that are default-constructed, so we don't try to
dereference null pointers.
We still auto-dereference past-the-end iterators, and iterators that
have been invalidated by container mutation. The former can
2016-12-15 14:30 GMT+01:00 Andre Vehreschild :
> Looks good to me.
Thanks, Andre. Committed to trunk as r243691.
Cheers,
Janus
> On Thu, 15 Dec 2016 14:18:28 +0100
> Janus Weil wrote:
>
>> Hi all,
>>
>> the attached patch deals with error recovery only and fixes an
>> ICE-on-invalid problem.
Looks good to me.
- Andre
On Thu, 15 Dec 2016 14:18:28 +0100
Janus Weil wrote:
> Hi all,
>
> the attached patch deals with error recovery only and fixes an
> ICE-on-invalid problem.
>
> Regtests cleanly on x86_64-linux-gnu. Ok for trunk?
>
> Cheers,
> Janus
>
>
> 2016-12-15 Janus Weil
>
As discussed in the PR, when we return a gdb.Value from a printer's
to_string() method GDB converts it to a string using a simplified
format that omits the address part of pointer/reference fields. I
think that's simply wrong, but in order to work with existing versions
of GDB we need to always co
Hi all,
the attached patch deals with error recovery only and fixes an
ICE-on-invalid problem.
Regtests cleanly on x86_64-linux-gnu. Ok for trunk?
Cheers,
Janus
2016-12-15 Janus Weil
PR fortran/78800
* interface.c (compare_allocatable): Avoid additional errors on bad
class decl
On Thu, Dec 15, 2016 at 04:32:34AM -0600, Segher Boessenkool wrote:
> On Wed, Dec 14, 2016 at 01:39:13PM +0100, Dominik Vogt wrote:
> > There may be a slight imprecision in expand_compound_operation.
> > When it encounters a SIGN_EXTEND where it's already known that the
> > sign bit is zero, it may
On Thu, Dec 15, 2016 at 3:23 AM, Jakub Jelinek wrote:
> On Wed, Dec 14, 2016 at 05:10:23PM -0700, Martin Sebor wrote:
>> The regression test is failing on powerpc64le due to the warnings
>> below:
>>
>> FAIL: gcc.dg/torture/pr78515.c -O0 (test for excess errors)
>> Excess errors:
>> /src/gcc/tr
Trivial tweaks.
* python/libstdcxx/v6/printers.py (UniquePointerPrinter.to_string):
Remove redundant parentheses.
(RbtreeIterator, StdRbtreeIteratorPrinter): Add docstrings.
(StdForwardListPrinter.to_string): Remove redundant parentheses.
(StdExpOptionalPri
This makes the Xmethods work for unique_ptr, including
conditionally enabling operator* and operator-> only for non-arrays,
and enabling operator[] only for arrays. And then adds similar
Xmethods for shared_ptr.
* python/libstdcxx/v6/xmethods.py (UniquePtrGetWorker.__init__): Use
On 12/15/2016 07:26 AM, Jakub Jelinek wrote:
I don't think so. complete_type (error_mark_node) returns error_mark_node,
and COMPLETE_TYPE_P (error_mark_node) is invalid (should fail TYPE_CHECK in
checking compiler).
I can write it as
inst = complete_type (inst);
if (inst == error_mark_node
77585 concerns the instantiation of a generic lambda that contains a
call to a non-dependent non-static member function.
auto lam = [&](auto) { return Share (); };
r += Eat (lam); // instantation of lambda::operator() here
During instantiation of the call to Share, maybe_resolve_dummy gets
On Thu, Dec 15, 2016 at 07:14:15AM -0500, Nathan Sidwell wrote:
> On 12/15/2016 03:34 AM, Jakub Jelinek wrote:
> > Hi!
> >
> > I'd like to ping the
> >
> > http://gcc.gnu.org/ml/gcc-patches/2016-12/msg00698.html
> > P0490R0 GB 20: decomposition declaration should commit to tuple
> > interpretati
On 12/15/2016 03:34 AM, Jakub Jelinek wrote:
Hi!
I'd like to ping the
http://gcc.gnu.org/ml/gcc-patches/2016-12/msg00698.html
P0490R0 GB 20: decomposition declaration should commit to tuple interpretation
early
+ if (inst == error_mark_node)
+return NULL_TREE;
This check is unneeded, b
On Thu, Dec 08, 2016 at 09:35:09AM +, Kyrill Tkachov wrote:
> Hi all,
>
> Similar to the previous patch this transforms X-reg UBFIZ instructions into
> W-reg LSL instructions when the UBFIZ operands add up to 32, so we can take
> advantage of the implicit zero-extension to DImode
> when writin
On Thu, Dec 08, 2016 at 09:35:08AM +, Kyrill Tkachov wrote:
> Hi all,
>
> In this patch we split X-register UBFX instructions that extract up to the
> edge of a W-register into a W-register LSR instruction. So for the example in
> the testcase instead of:
> UBFXX0, X0, 24, 8
>
> we'd gen
Hi,
As mentioned in PR77445, the improvements to the jump threading cost model
this year have caused substantial regressions in the amount of jump threading
we do and the performance of workloads which rely on that threading.
This patch represents the low-bar in fixing the performance issues rep
On 14/12/16 18:38 -0300, Felipe Magno de Almeida wrote:
Hello Jonathan,
Sorry for the delay, I was in mid-vacation.
Comments are inline.
On Tue, Dec 6, 2016 at 3:45 PM, Jonathan Wakely wrote:
On 16/09/16 02:53 -0300, Felipe Magno de Almeida wrote:
[snip]
I've tried both approaches. Tem
On 14/12/16 22:49 +0100, François Dumont wrote:
On 09/12/2016 16:18, Jonathan Wakely wrote:
But I don't know how to fix this so for the moment I just adapt it
to correctly handle std::__7::string.
But that's not correct. Please try to understand the point I'm making:
The name "std::__7::stri
On 12/13/2016 03:35 PM, Richard Biener wrote:
> OK from my POV.
>
> Thanks,
> Richard.
Hi.
I prepared backport (and tested) for GCC 5 and 6 branches,
may I install the patch after 6.3 will be released?
Thanks,
Martin
Hi,
Please find attached the patch that optimizes some patterns
in maxmin on same variabes with constants.
Bootstrapped and Regression tested on x86_64 & aarch64-thunder-linux.
Please review the patch and let us know if its okay?
2016-12-15 Andrew Pinski
Naveen H.S
gcc
The patch adds TREE_THIS_VOLATILE check that was removed in r239778.
Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
Ready to be installed?
Martin
>From dc8ec6815fa51b352fe5f1a02d3510022053e0ad Mon Sep 17 00:00:00 2001
From: marxin
Date: Wed, 14 Dec 2016 16:07:56 +0100
On Wed, Dec 14, 2016 at 01:39:13PM +0100, Dominik Vogt wrote:
> There may be a slight imprecision in expand_compound_operation.
> When it encounters a SIGN_EXTEND where it's already known that the
> sign bit is zero, it may replace that with a ZERO_EXTEND (and
> tries to simplify that further). Ho
On Thu, Dec 15, 2016 at 11:23:14AM +0100, Andreas Schwab wrote:
> On Dez 15 2016, Jakub Jelinek wrote:
>
> > So, shall we change also the first 3?
>
> Yes, I'd think so.
So here is it in patch form. Is this ok for trunk?
2016-12-15 Jakub Jelinek
* doc/extend.texi: Clean up @xref{.
Committed as r243687.
Regards,
Toma
> Toma Tabacu writes:
> > microMIPS is not supported on pre-R2 architectures, but the testsuite allows
> > it to be used on pre-R2 architectures, which results in test failures.
> >
> > This patch makes the testsuite upgrade to R2 if -mmicromips is used in a
>
On Dez 15 2016, Jakub Jelinek wrote:
> So, shall we change also the first 3?
Yes, I'd think so.
Andreas.
--
Andreas Schwab, SUSE Labs, sch...@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."
On Thu, Dec 15, 2016 at 11:11:37AM +0100, Andreas Schwab wrote:
> On Dez 15 2016, Jakub Jelinek wrote:
>
> > --- gcc/doc/extend.texi.jj 2016-12-14 20:28:12.0 +0100
> > +++ gcc/doc/extend.texi 2016-12-15 10:56:14.470702563 +0100
> > @@ -1057,7 +1057,7 @@ implements conversions between
> On a high level, presumably there's no real value in keeping the old
> code to "fold" fpclassify. By exposing those operations as integer
> logicals for the fast path, if the FP value becomes a constant during
> the optimization pipeline we'll see the reinterpreted values flowing
> into the new
On Dez 15 2016, Jakub Jelinek wrote:
> --- gcc/doc/extend.texi.jj2016-12-14 20:28:12.0 +0100
> +++ gcc/doc/extend.texi 2016-12-15 10:56:14.470702563 +0100
> @@ -1057,7 +1057,7 @@ implements conversions between @code{__f
> calls.
>
> It is recommended that portable code use t
On Mon, Dec 12, 2016 at 02:19:47PM -0700, Sandra Loosemore wrote:
> One small grammar nit at the end:
>
> > > +It is recommended that code which is intended to be portable use the
> > > +@code{_Float16} type defined by ISO/IEC TS 18661-3:2015
> > > +(@xref{Floating Types}).
>
> Either s/which/tha
On 15/12/16 09:55, Richard Earnshaw (lists) wrote:
> On 30/11/16 16:47, Kyrill Tkachov wrote:
>> Hi all,
>>
>> In this awkward ICE we have a *load_multiple pattern that is being
>> transformed in reload from:
>> (insn 55 67 151 3 (parallel [
>> (set (reg:SI 0 r0)
>> (mem
On 30/11/16 16:47, Kyrill Tkachov wrote:
> Hi all,
>
> In this awkward ICE we have a *load_multiple pattern that is being
> transformed in reload from:
> (insn 55 67 151 3 (parallel [
> (set (reg:SI 0 r0)
> (mem/u/c:SI (reg/f:SI 147) [2 c+0 S4 A32]))
> (set
Ping.
Thanks,
Kyrill
On 08/12/16 11:55, Kyrill Tkachov wrote:
Ping.
https://gcc.gnu.org/ml/gcc-patches/2016-11/msg03078.html
Thanks,
Kyrill
On 30/11/16 16:47, Kyrill Tkachov wrote:
Hi all,
In this awkward ICE we have a *load_multiple pattern that is being transformed
in reload from:
(insn
1 - 100 of 105 matches
Mail list logo
