On Wed, Dec 14, 2016 at 01:27:44PM +0530, Prathamesh Kulkarni wrote:
> 2016-12-14 Jakub Jelinek
> Prathamesh Kulkarni
>
> * tree-ssa-strlen.c (fold_strstr_to_memcmp): New function.
> (strlen_optimize_stmt): Call fold_strstr_to_memcmp.
>
> testsuite/
> * gcc.dg/str
On 11/25/2016 05:18 AM, Tamar Christina wrote:
Hi Joseph,
I have updated the patch with the changes,
w.r.t to the formatting, there are tabs there that seem to be rendered
at 4 spaces wide. In my editor setup at 8 spaces they are correct.
Various random comments, mostly formatting. I do think
On Tue, 13 Dec 2016, Thomas Preudhomme wrote:
> Hi,
>
> Fix in r242869 for PR77673 (bswap loads after end of object) applies cleanly
> on GCC 6 and with trivial fix for GCC 5 (gimple * in GCC 6 -> gimple in GCC
> 5). The backports also bootstrap fine on x86_64-linux-gnu and testsuite shows
> no r
On Tue, 13 Dec 2016, Thomas Preudhomme wrote:
> Hi,
>
> Fix in r242869 for PR77673 was accompanied with r242870 which updated the
> description of the struct symbolic_number modified by the previous patch. It
> did so by rewriting the comment completely but I believe this patch should be
> still
Hi!
GCC before r237607 used to mishandle TLS in -mcmodel=large - forcing
TLS symbols in constant pool, which obviously doesn't work, because they
aren't constant. For all other models aarch64_tls_symbol_p is checked
first.
The following patch backports just the bug fix for this PR part to 6.x,
t
On Tue, Dec 13, 2016 at 10:05:58AM +0100, Richard Biener wrote:
> Hmm, so the "easier" fix would be to always create a debug insn
> right after INSN setting a new debug reg to SRC and then doing the
> replacement with the new debug reg? With the "optimization"
> to not do that for some single use
> I wasn't ever able to get comfortable with the change in behavior that
> Bernd E's patch introduced. Namely that X no longer meant "anything",
> which is how it's been documented for eons.
Understood. I have applied my patchlet then, with the other, more serious LRA
patch I posted yesterday,
On 2016/11/30 10:47 AM, Cesar Philippidis wrote:
> On 11/18/2016 03:24 AM, Jakub Jelinek wrote:
>> On Sat, Nov 12, 2016 at 08:51:00AM -0800, Cesar Philippidis wrote:
>>> On 11/11/2016 02:34 AM, Jakub Jelinek wrote:
On Thu, Nov 10, 2016 at 06:46:46PM +0800, Chung-Lin Tang wrote:
>>>
>>> And her
On 12/08/2016 05:55 AM, Martin Liška wrote:
With the patch applied, one can distinguish between PROB_VERY_UNLIKELY and
real zero probability:
f ()
{
int _1;
[100.00%]:
_1 = __builtin_sprintf (&d2, "%i", 12);
if (_1 != 2)
goto ; [0.04%]
else
goto ; [99.96%]
[0.04%]:
__bu
On Wed, 14 Dec 2016, Jakub Jelinek wrote:
> On Tue, Dec 13, 2016 at 10:05:58AM +0100, Richard Biener wrote:
> > Hmm, so the "easier" fix would be to always create a debug insn
> > right after INSN setting a new debug reg to SRC and then doing the
> > replacement with the new debug reg? With the "
Hi,
It seems I backported the wrong versions of the testcases to
embedded-6-branch. Fixed them with this patch.
Cheers,
Andre
gcc/testsuite/ChangeLog.arm:
2016-12-14 Andre Vieira
* gcc.target/arm/cmse/mainline/hard/cmse-13.c: Fix testsuite.
* gcc.target/arm/cmse/mainline/hard/cmse-7
On Dez 14 2016, Martin Sebor wrote:
> +#define target_dir_max() (target_int_max () + 4932 + 2)
> + if (9864 < prec)
> + prec = 9864;
Should these two magic numbers be macroized?
Andreas.
--
Andreas Schwab, SUSE Labs, sch...@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1
On Wed, Dec 14, 2016 at 09:34:44AM +0100, Jakub Jelinek wrote:
> Here is updated patch that does that. I had to change the combiner
> in a couple of places, because combiner is very unhappy if new instructions
> (with uids above the highest) are inserted into the insn stream,
> e.g. FOR_EACH_LOG_L
On Wed, Dec 14, 2016 at 09:24:18AM +0100, Jakub Jelinek wrote:
> Hi!
>
> GCC before r237607 used to mishandle TLS in -mcmodel=large - forcing
> TLS symbols in constant pool, which obviously doesn't work, because they
> aren't constant. For all other models aarch64_tls_symbol_p is checked
> first.
Hi Thomas,
On 9 December 2016 at 17:56, Thomas Preudhomme
wrote:
> Hi,
>
> This patch moves the definition of arm_arch_core_flags along with the
> declaration of its structure element type to common/config/arm/arm-common.c
> so that it is not *defined* in all object files including tm.h. Otherwis
This is another micro-optimisation in change_zero_ext. If an
(and (lshiftrt ... (N)) (M))
generated by change_zero_ext is equivalent to just
(lshiftrt ... (N))
(because the AND constant selects the N rightmost bits of the
result), strip off the AND.
_But_ I'm still not completely convince
Hi,
On Fri, Dec 09, 2016 at 01:18:25PM +0100, Richard Biener wrote:
>
> The patch looks somewhat backward. It populates the param type from
> the callee but the only thing we really know is the _originating_ type
> from the callers DECL_ARGUMENTS (if the JF is based on a parameter
> which is the
Committed as r243639.
Regards,
Toma
> -Original Message-
> From: Matthew Fortune
> Sent: 13 December 2016 13:46
> To: Toma Tabacu; gcc-patches@gcc.gnu.org
> Cc: catherine_mo...@mentor.com
> Subject: RE: [PATCH, testsuite] MIPS: Skip msa-builtins-err.c if there is no
> assembly output.
>
The following prevents expand_cond_expr_using_cmove from going berzerk
with a large TER tree containing several cond-exprs that eventually
end up failing expansion as cmove. The attempt to re-do TER failed
somehow earlier this year so the following is a simple stop-gap solution.
Bootstrap & regt
On Wed, Dec 14, 2016 at 11:38:29AM +0100, Richard Biener wrote:
>
> The following prevents expand_cond_expr_using_cmove from going berzerk
> with a large TER tree containing several cond-exprs that eventually
> end up failing expansion as cmove. The attempt to re-do TER failed
> somehow earlier t
The following implements Jasons suggestion of using a langhook to
return the size of an aggregate without tail padding that might
be re-used when it is inherited from.
Using this langhook we can fix the size of the representative for the
bitfield properly and thus generate correct code for the t
Hi,
In the case of mips16-attributes.c, even though the (-mips16) option ensures
that -mmicromips will be overriden, there still is a dg-skip-if for -mmicromips.
I think that it is not necessary and it actually decreases test coverage,
because it will cause the testsuite to skip this test instead
This crashes the compiler on ia64.
FAIL: c-c++-common/raw-string-6.c -Wc++-compat (test for excess errors)
Excess errors:
/usr/local/gcc/gcc-20161214/gcc/testsuite/c-c++-common/raw-string-6.c:5:1:
internal compiler error: Segmentation fault
0x4108571f crash_signal
../../gcc
Hi,
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 test.
Tested with mips-mti-elf.
Regards,
Toma
gcc/testsuite/ChangeLog
On 13/12/16 22:51 +0100, François Dumont wrote:
Hi
I have been reported privately by Christophe in copy a regression
resulting from my recent changes to _Rb_tree. I removed a constructor
still necessary in C++03 mode or before. Tests would have shown it if
I had run them in C++03.
For re
On 14/12/16 11:27 +, Jonathan Wakely wrote:
On 13/12/16 22:51 +0100, François Dumont wrote:
Hi
I have been reported privately by Christophe in copy a regression
resulting from my recent changes to _Rb_tree. I removed a
constructor still necessary in C++03 mode or before. Tests would
ha
On 14/12/16 11:27 +, Jonathan Wakely wrote:
On 13/12/16 22:51 +0100, François Dumont wrote:
Hi
I have been reported privately by Christophe in copy a regression
resulting from my recent changes to _Rb_tree. I removed a
constructor still necessary in C++03 mode or before. Tests would
ha
Hi all,
thanks for the input and reviews. I have committed this patch as r243647.
Regards,
Andre
On Sun, 11 Dec 2016 15:49:33 +0100
Paul Richard Thomas wrote:
> Hi Andre,
>
> Thanks for doing this work with the instrumented compiler. It was a
> great help with PR78350.
>
> As for the
On Wed, Dec 14, 2016 at 11:15 AM, Martin Jambor wrote:
> Hi,
>
> On Fri, Dec 09, 2016 at 01:18:25PM +0100, Richard Biener wrote:
>>
>> The patch looks somewhat backward. It populates the param type from
>> the callee but the only thing we really know is the _originating_ type
>> from the callers
* Claudiu Zissulescu [2016-12-05 12:59:12
+0100]:
> Remove old gmonlib from libgcc and reimplemnt profiling using UCB
> counters.
>
> gcc/
> 2016-07-28 Claudiu Zissulescu
>
> * config/arc/arc.h (LINK_SPEC): Tidy up.
> (ENDFILE_SPEC): Likewise.
> (LIB_SPEC): Likewise.
>
* Claudiu Zissulescu [2016-12-05 12:59:13
+0100]:
> Cleanup old patterns.
>
> gcc/
> 2016-10-10 Claudiu Zissulescu
>
> * config/arc/arc.md (call_prof): Remove.
> (call_value_prof): Likewise.
> (sibcall_prof): Likewise.
> (sibcall_value_prof): Likewise.
Looks fine.
On Tue, 13 Dec 2016, Jakub Jelinek wrote:
> Hi!
>
> Sorry for not getting to this earlier.
>
> On Mon, Nov 28, 2016 at 10:50:26AM +0100, Richard Biener wrote:
> > > + else if (gimple_call_builtin_p (defstmt, BUILT_IN_MEMSET)
> > > +&& TREE_CODE (gimple_call_arg (defstmt, 0)) == ADDR_EXPR
>
On Wed, Dec 14, 2016 at 11:54:21AM +0100, Andreas Schwab wrote:
> This crashes the compiler on ia64.
>
> FAIL: c-c++-common/raw-string-6.c -Wc++-compat (test for excess errors)
> Excess errors:
> /usr/local/gcc/gcc-20161214/gcc/testsuite/c-c++-common/raw-string-6.c:5:1:
>
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). However, the pattern is only
replaced if the new set_src_cost() is _lowe
Hi Steve, hi all,
thanks for the review. While going to bed, it came to me that the testcase
checks only half of result of this patch. I therefore took the liberty to
add another testcase, that checks that the other part of the change is also
tested as testsuite/gfortran.dg/coarray_42.f90.
The ch
On 12/13/2016 03:58 PM, Jason Merrill wrote:
I'm still seeing the redundant DECL_LANG_SPECIFIC check.
Sorry, I misunderstood which check you meant was redundant. In the
original patch, the first check wasn't redundant (but the second check
was entirely within the codepath of the first). Bu
Hi Matthias,
> From my point of view this should be backported to the active branches.
> Building GCC 5 and GCC 6 with binutils >=2.26 now results in
>
> $ gcc -c -gz foo.c
> gcc: error: -gz is not supported in this configuration
>
> building these GCC version with binutils 2.25 recognizes this op
Toma Tabacu writes:
> In the case of mips16-attributes.c, even though the (-mips16) option ensures
> that -mmicromips will be overriden, there still is a dg-skip-if for
> -mmicromips.
>
> I think that it is not necessary and it actually decreases test coverage,
> because it will cause the testsu
* Claudiu Zissulescu [2016-11-16 11:18:00
+0100]:
> gcc/
> 2016-07-04 Claudiu Zissulescu
>
> * config/arc/arc.md (mulsidi_600): Changed.
> (umulsidi_600): Likewise.
> (mul64): New pattern.
> (mulu64): Likewise.
> (mulsidi3): Changed.
> (umulsidi3): Likewis
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 test.
>
> Tested with mips-mti-elf.
>
> Regard
Hi!
On Tue, 13 Dec 2016 13:42:23 +0100, Martin Jambor wrote:
> On Tue, Dec 13, 2016 at 12:43:16PM +0100, Jakub Jelinek wrote:
> > On Tue, Dec 13, 2016 at 12:39:01PM +0100, Thomas Schwinge wrote:
> > > On Fri, 9 Dec 2016 14:08:21 +0100, Martin Jambor wrote:
> > > > this is the promised attempt at
On Thu, Dec 08, 2016 at 02:56:56PM -0500, Nathan Sidwell wrote:
> On 12/08/2016 01:05 PM, Jason Merrill wrote:
> > If the problem is the member initializer, we can diagnose that
> > directly rather than wait until we're in a constructor.
>
> What about:
> struct Foo {
> int a;
> char ary[];
> Foo
Hi,
On Wed, Dec 14, 2016 at 02:14:32PM +0100, Thomas Schwinge wrote:
>
> I still couldn't allocate time to review the patch, but at least I now
> have tested it -- no regressions.
Great, thanks!
> As I suppose you want to commit this
> as sooner than later ;-) and you already have approval as I
Status
==
The GCC 6 branch is now frozen for blocking regressions and documentation
fixes only, all changes to the branch require a RM approval now.
Quality Data
Priority # Change from last report
--- ---
P10
P2
On 12/09/2016 08:32 PM, David Malcolm wrote:
Thanks. Unfortunately, applying the "locate_file" patch
https://gcc.gnu.org/ml/gcc-patches/2016-11/msg01186.html
would now introduce a regression in a recently-added test case:
The problem is that this DejaGnu test case uses -fself-test, and
does
On Wed, Dec 14, 2016 at 8:16 AM, Marek Polacek wrote:
> + if (TREE_CODE (type) == ARRAY_TYPE
> + && TYPE_DOMAIN (type) == NULL_TREE
> + && init != NULL_TREE)
> + error_at (DECL_SOURCE_LOCATION (member),
> + "initialization of a flexible array member in a
This patch fixes 78701. The problematic case with a non-type template
arg whose default value is dependent on a previous non-type arg that
failed to be deduced on the first iteration of the loop in
type_unification_real.
Just before the tsubst_template_arg we've already done:
if (T
On Tue, Dec 13, 2016 at 07:02:48PM -0600, Segher Boessenkool wrote:
> On Tue, Dec 13, 2016 at 06:17:17PM -0500, Michael Meissner wrote:
> > > > + else if (mode == V8HImode)
> > > > + {
> > > > + rtx tmp_gpr_si = (GET_CODE (tmp_gpr) == SCRATCH
> > > > +
Committed as r243649.
Regards,
Toma
> Toma Tabacu writes:
> > In the case of mips16-attributes.c, even though the (-mips16) option ensures
> > that -mmicromips will be overriden, there still is a dg-skip-if for
> > -mmicromips.
> >
> > I think that it is not necessary and it actually decreases t
Hi,
Some error tests were forcing assembly output incorrectly, and none of them had
an explanation for why they were using -ffat-lto-objects.
This patch removes dg-skip-if's for -fno-fat-lto-objects and
adds -ffat-lto-objects to the test options instead.
It also adds an explanation for the purpos
On Wed, Dec 14, 2016 at 09:45:12AM -0500, Michael Meissner wrote:
> I looked at the code once again, and in the cases that call the function, they
> always create a GPR temporary register, so I didn't need to reuse the
> destination register. Thanks for the comments. Are these patches ok to check
++-compat (test for excess
> errors)
> Excess errors:
> /usr/local/gcc/gcc-20161214/gcc/testsuite/c-c++-common/raw-string
> -6.c:5:1: internal compiler error: Segmentation fault
> 0x4108571f crash_signal
> ../../gcc/toplev.c:333
> 0x42024dcf _cpp_process_
This is a partial fix for PR 59170, using the debug mode iterator's
members to tell when it's invalid. This doesn't detect past-the-end
iterators, and doesn't help for normal mode where we don't have the
extra info (but we can still detect value-initialized iterators for
normal mode and not ry to
The attached patch takes care of the problem described below.
(I tried to stick the new code into the regular combine-simplify
stuff, but that does not work well. At the time when
combine_simplify_rtx is called, it's not known whether
change_zero_ext will expand the ZERO_EXTRACT to "(and (lshiftr
Ping^2 ?
As a reminder, this patch mimics what aarch64 does wrt to references to weak
symbols such that they are not resolved by the assembler, in case a strong
definition overrides the local one at link time.
Christophe
On 8 December 2016 at 09:17, Christophe Lyon wrote:
> Ping?
>
> On 1 Dece
Apparently, Qt source uses these two builtins. As the removal brings
much pain, I'd like to commit the following patch to restore
__builtin_clzs and __builtin_ctzs on gcc-6 branch.
2016-12-14 Uros Bizjak
PR target/59874
* config/i386/i386.c (enum ix86_builtins): Add IX86_BUILTIN_CLZS
On Wed, Dec 14, 2016 at 03:36:23AM -0600, Segher Boessenkool wrote:
> On Wed, Dec 14, 2016 at 09:34:44AM +0100, Jakub Jelinek wrote:
> > Here is updated patch that does that. I had to change the combiner
> > in a couple of places, because combiner is very unhappy if new instructions
> > (with uids
Ping~
Regards,
Renlin
On 16/06/16 12:04, Renlin Li wrote:
Hi all,
GCC has startfile and endfile spec string built into it.
startfile is used to specify objects files to include at the start of the link
process.
While endfile, on the other hand, is used to specify objects files to include
at
On Wed, Dec 14, 2016 at 04:30:39PM +0100, Uros Bizjak wrote:
> Apparently, Qt source uses these two builtins. As the removal brings
> much pain, I'd like to commit the following patch to restore
> __builtin_clzs and __builtin_ctzs on gcc-6 branch.
>
> 2016-12-14 Uros Bizjak
>
> PR target/5
On Wed, Dec 14, 2016 at 4:33 PM, Jakub Jelinek wrote:
> On Wed, Dec 14, 2016 at 04:30:39PM +0100, Uros Bizjak wrote:
>> Apparently, Qt source uses these two builtins. As the removal brings
>> much pain, I'd like to commit the following patch to restore
>> __builtin_clzs and __builtin_ctzs on gcc-6
On Wed, Dec 14, 2016 at 09:04:01AM -0500, Jason Merrill wrote:
> On Wed, Dec 14, 2016 at 8:16 AM, Marek Polacek wrote:
> > + if (TREE_CODE (type) == ARRAY_TYPE
> > + && TYPE_DOMAIN (type) == NULL_TREE
> > + && init != NULL_TREE)
> > + error_at (DECL_SOURCE_LOCATION (memb
On 12/12/2016 03:21 PM, Bernd Schmidt wrote:
On 12/10/2016 08:58 PM, Segher Boessenkool wrote:
On Thu, Dec 08, 2016 at 01:21:04PM +0100, Bernd Schmidt wrote:
This is another case where an optimization turns a trap_if
unconditional. We have to defer changing the CFG, since the rest of
cprop seem
This fixes the fact that our pretty printers say ungrammatical things
like "std::vector with 1 elements"
* python/libstdcxx/v6/printers.py (num_elements): New function.
(StdMapPrinter.to_string, StdSetPrinter.to_string)
(StdDequePrinter.to_string, Tr1UnorderedSetPrinter.to
On Wed, Dec 14, 2016 at 10:39 AM, Marek Polacek wrote:
> + if (TREE_CODE (type) == ARRAY_TYPE
> + && TYPE_DOMAIN (type) == NULL_TREE
> + && init != NULL_TREE)
> + error_at (DECL_SOURCE_LOCATION (current_function_decl),
> + "member initializer for flexible
Yes, thanks.
On Wed, Dec 14, 2016 at 7:51 AM, Nathan Sidwell wrote:
> On 12/13/2016 03:58 PM, Jason Merrill wrote:
>
>> I'm still seeing the redundant DECL_LANG_SPECIFIC check.
>
>
> Sorry, I misunderstood which check you meant was redundant. In the original
> patch, the first check wasn't redun
OK.
On Wed, Dec 14, 2016 at 9:05 AM, Nathan Sidwell wrote:
> This patch fixes 78701. The problematic case with a non-type template arg
> whose default value is dependent on a previous non-type arg that failed to
> be deduced on the first iteration of the loop in type_unification_real.
>
> Just b
OK, thanks.
On Wed, Dec 14, 2016 at 5:44 AM, Richard Biener wrote:
>
> The following implements Jasons suggestion of using a langhook to
> return the size of an aggregate without tail padding that might
> be re-used when it is inherited from.
>
> Using this langhook we can fix the size of the rep
ping
From: Wilco Dijkstra
Sent: 29 November 2016 11:05
To: GCC Patches
Cc: nd
Subject: [PATCH][ARM] Remove movdi_vfp_cortexa8
Merge the movdi_vfp_cortexa8 pattern into movdi_vfp and remove it to avoid
unnecessary duplication and repeating bugs like PR78439 due to changes being
applied
ping
From: Wilco Dijkstra
Sent: 30 November 2016 17:39
To: GCC Patches
Cc: nd; Bernd Edlinger
Subject: [PATCH][ARM] Merge negdi2 patterns
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
ping
From: Wilco Dijkstra
Sent: 10 November 2016 17:19
To: GCC Patches
Cc: nd
Subject: [PATCH][ARM] Improve max_insns_skipped logic
Improve the logic when setting max_insns_skipped. Limit the maximum size of IT
to MAX_INSN_PER_IT_BLOCK as otherwise multiple IT instructions are needed
ping
From: Wilco Dijkstra
Sent: 30 November 2016 17:32
To: GCC Patches
Cc: nd
Subject: [PATCH][ARM] Improve Thumb allocation order
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 save
ping
From: Wilco Dijkstra
Sent: 05 December 2016 14:52
To: GCC Patches
Cc: nd
Subject: [PATCH][ARM] Remove uses of leaf_function_p
Using leaf_function_p in a backend is dangerous as it incorrectly returns
false if it is called while in a sequence (for example during prolog/epilog
generation
ping
From: Wilco Dijkstra
Sent: 03 November 2016 12:20
To: GCC Patches
Cc: nd
Subject: [PATCH][ARM] Fix ldrd offsets
Fix ldrd offsets of Thumb-2 - for TARGET_LDRD the range is +-1020,
without -255..4091. This reduces the number of addressing instructions
when using DI mode operations
ping
From: Wilco Dijkstra
Sent: 31 October 2016 18:29
To: GCC Patches
Cc: nd
Subject: [RFC][PATCH][AArch64] Cleanup frame pointer usage
This patch cleans up all code related to the frame pointer. On AArch64 we
emit a frame chain even in cases where the frame pointer is not required.
So m
So I've been going back and forth on whether or not to suggest a slight
change in how we're working.
Specifically the practice of lumping multiple fixes into a single patch
-- I realize that it's usually the case that you're finding related
issues so there's a desire to address them as a group.
Hi Wilco,
On 14/12/16 16:37, Wilco Dijkstra wrote:
ping
From: Wilco Dijkstra
Sent: 29 November 2016 11:05
To: GCC Patches
Cc: nd
Subject: [PATCH][ARM] Remove movdi_vfp_cortexa8
Merge the movdi_vfp_cortexa8 pattern into movdi_vfp and remove it to avoid
unnecessary duplication and
ping
From: Wilco Dijkstra
Sent: 02 September 2016 12:31
To: Ramana Radhakrishnan; GCC Patches
Cc: nd
Subject: Re: [PATCH][AArch64 - v3] Simplify eh_return implementation
Ramana Radhakrishnan wrote:
> Can you please file a PR for this and add some testcases ? This sounds like
> a serious
Hi Wilco,
On 14/12/16 16:39, Wilco Dijkstra wrote:
ping
From: Wilco Dijkstra
Sent: 05 December 2016 14:52
To: GCC Patches
Cc: nd
Subject: [PATCH][ARM] Remove uses of leaf_function_p
Using leaf_function_p in a backend is dangerous as it incorrectly returns
false if it is called while in
On Mon, Dec 12, 2016 at 05:03:31PM +0100, Christophe Lyon wrote:
> Hi,
>
> After the recent update from Tamar, I noticed a few discrepancies
> between ARM and AArch64 regarding a few poly64 intrinsics.
>
> This patch:
> - adds vtst_p64 and vtstq_p64 to AArch64's arm_neon.h
> - adds vgetq_lane_p64
On 12/14/2016 02:25 AM, Andreas Schwab wrote:
On Dez 14 2016, Martin Sebor wrote:
+#define target_dir_max() (target_int_max () + 4932 + 2)
+ if (9864 < prec)
+ prec = 9864;
Should these two magic numbers be macroized?
I had considered it and didn't think it was worth it (th
The RTL frontend parses the output of print_rtx_function's "compact"
mode.
Currently non-virtual pseudo regs in such dumps have a '%' prefix,
based on the discussion here (me):
* " Re: [PATCH] Tweaks to print_rtx_function"
* https://gcc.gnu.org/ml/gcc-patches/2016-10/msg01033.html
and here (
On Wed, Dec 14, 2016 at 11:30:17AM -0500, Jason Merrill wrote:
> On Wed, Dec 14, 2016 at 10:39 AM, Marek Polacek wrote:
> > + if (TREE_CODE (type) == ARRAY_TYPE
> > + && TYPE_DOMAIN (type) == NULL_TREE
> > + && init != NULL_TREE)
> > + error_at (DECL_SOURCE_LOCATION (cur
Hi,
During embedded-6-branch testing we discovered that the patch for
PR71607, that I backported to embedded-6-branch, broke cortex-m0 due to
the new implementation of TARGET_USE_BLOCKS_FOR_CONSTANT_P.
What changed is that TARGET_USE_BLOCKS_FOR_CONSTANT_P did not used to be
defined by ARM and thu
On Sat, Nov 26, 2016 at 01:54:34PM -0800, Andrew Pinski wrote:
> On Tue, Nov 1, 2016 at 11:08 AM, Andrew Pinski wrote:
> > On Tue, Nov 17, 2015 at 2:10 PM, Andrew Pinski wrote:
> >> Since ThunderX T88 pass 1 (variant 0) is a ARMv8 part while pass 2
> >> (variant 1)
> >> is an ARMv8.1 part, I nee
On Wed, Dec 14, 2016 at 04:31:19PM +0100, Jakub Jelinek wrote:
> > Could you do a gcc_checking_assert (INSN_UID (INSN) <= max_uid_known)
> > in the LOG_LINKS and INSN_COST defines? It is hard to check if you
> > caught every use.
>
> Like this? Bootstrapped/regtested again on x86_64-linux and i
Kyrill Tkachov wrote:
> On 14/12/16 16:37, Wilco Dijkstra wrote:
>
> > Merge the movdi_vfp_cortexa8 pattern into movdi_vfp and remove it to avoid
> > unnecessary duplication and repeating bugs like PR78439 due to changes being
> > applied only to one of the duplicates.
>
> When this was brought up
On Wed, Dec 14, 2016 at 04:46:09PM +0100, Bernd Schmidt wrote:
> That would be this patch. Tested as before. The two new testcases seem
> to pass with a ppc cross (but I would appreciate if someone were to run
> full tests on ppc).
Thanks, will do.
Segher
2016-12-02 21:31 GMT+03:00 Uros Bizjak :
. . . . .
>>
>> I split this patch after last updates in md files, here is the first
>> part which doesn't change md files.
>> Regtested on x86_64-linux-gnu. Is this part ok?
>
> There is no point to scan for kmovX insn in e.g.:
>
> +/* { dg-final { scan-as
On Wed, Dec 14, 2016 at 11:01:47AM +0100, Dominik Vogt wrote:
> This is another micro-optimisation in change_zero_ext. If an
>
> (and (lshiftrt ... (N)) (M))
>
> generated by change_zero_ext is equivalent to just
>
> (lshiftrt ... (N))
>
> (because the AND constant selects the N rightmost
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
> @@ -1309,12 +1309,30 @@
> ;; Mask variant shift mnemonics
> (define_code_attr mshift [(ashift "shiftl") (lshiftrt "
On Wed, Dec 14, 2016 at 4:39 PM, Uros Bizjak wrote:
> On Wed, Dec 14, 2016 at 4:33 PM, Jakub Jelinek wrote:
>> On Wed, Dec 14, 2016 at 04:30:39PM +0100, Uros Bizjak wrote:
>>> Apparently, Qt source uses these two builtins. As the removal brings
>>> much pain, I'd like to commit the following patc
OK, thanks.
On Wed, Dec 14, 2016 at 12:01 PM, Marek Polacek wrote:
> On Wed, Dec 14, 2016 at 11:30:17AM -0500, Jason Merrill wrote:
>> On Wed, Dec 14, 2016 at 10:39 AM, Marek Polacek wrote:
>> > + if (TREE_CODE (type) == ARRAY_TYPE
>> > + && TYPE_DOMAIN (type) == NULL_TREE
>> > +
On 14/12/2016 12:36, Jonathan Wakely wrote:
On 14/12/16 11:27 +, Jonathan Wakely wrote:
On 13/12/16 22:51 +0100, François Dumont wrote:
I've confirmed that the change fixes these bad test results for C++03
testing:
UNRESOLVED: 12
FAIL: 32
Thanks for confirming, now committed.
Fr
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. Templates were causing problems of not
>> defined instanta
Hi!
As mentioned in the PR, SLOT_TEMP slots may be created during expansion,
and by the time the stv pass is run it is hard to find out what are the
insn ranges where those temporary slots from expansion are live and thus
where we should avoid reusing those even if they have the same mode.
The pa
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::string" does not appear in a symbol name.
Ok, the onl
Hi!
On the following testcase we ICE, because we don't diagnose invalid out of
bounds stores and create garbage CONSTRUCTOR that the middle-end ICEs on.
The following patch enables out of bounds diagnostics even for the case when
lval is true, but because lval = true is used not just for the lhs
o
Hi!
The gimplifier (because of Ada :( ) doesn't recurse to POINTER_TYPE's
type when gimplifying types, so we need a DECL_EXPR to gimplify such
types if they are VLAs. The following patch is an attempt to do that.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2016-12-14 J
On 14 December 2016 at 17:55, James Greenhalgh wrote:
> On Mon, Dec 12, 2016 at 05:03:31PM +0100, Christophe Lyon wrote:
>> Hi,
>>
>> After the recent update from Tamar, I noticed a few discrepancies
>> between ARM and AArch64 regarding a few poly64 intrinsics.
>>
>> This patch:
>> - adds vtst_p64
On Wed, Dec 07, 2016 at 08:24:33PM -0700, Martin Sebor wrote:
> @@ -8622,15 +8635,17 @@ rewrite_call_expr (location_t loc, tree exp, int
> skip, tree fndecl, int n, ...)
> }
>
> /* Validate a single argument ARG against a tree code CODE representing
> - a type. */
> + a type. When NONNUL
1 - 100 of 112 matches
Mail list logo
