On Thu, Feb 1, 2018 at 2:44 AM, Jeff Law wrote:
>
> This is V2 of the patch to fix 84128.
>
> What's change since V1. The -fstack-check path for non-linux systems
> has been restored to its previous (correct) behavior. Only the behavior
> for -fstack-check on linux systems and -fstack-clash-prot
On Tue, Jan 30, 2018 at 12:53 PM, Koval, Julia wrote:
> Thank you for your comments, fixed them and rebased Ice Lake patch on top of
> it. Ok for trunk?
>
> Bitmask patch changelog:
>
> gcc/c-family/
> * c-common.h (omp_clause_mask): Move to wide_int_bitmask.h.
>
> gcc/
> * config
On Thu, Feb 1, 2018 at 1:16 AM, H.J. Lu wrote:
>
> We currently read and write beyond the builtin jmpbuf on ILP32 targets
> where Pmode == DImode and ptr_mode == SImode. Since the builtin jmpbuf
> is an array of 5 pointers, ptr_mode should be used to save and restore
> frame and program pointers.
On 01/31/2018 09:34 PM, Paul Richard Thomas wrote:
This fixes the reduced testcase provided by Tom de Vries in comment #7
of the PR. Committed as 'obvious' as r257262. Will await a report from
Tom as to whether or not this fixes the original problem
Hi Paul,
it does.
I've tested r257266,
Hi Richard,
On 31 January 2018 at 21:39, Richard Biener wrote:
> On Wed, Jan 31, 2018 at 11:28 AM, Kugan Vivekanandarajah
> wrote:
>> Hi Richard,
>>
>> Thanks for the review.
>> On 25 January 2018 at 20:04, Richard Biener
>> wrote:
>>> On Wed, Jan 24, 2018 at 10:56 PM, Kugan Vivekanandarajah
>
This is V2 of the patch to fix 84128.
What's change since V1. The -fstack-check path for non-linux systems
has been restored to its previous (correct) behavior. Only the behavior
for -fstack-check on linux systems and -fstack-clash-protection has been
changed.
release_scratch_regsiter_on_entry
This patch to the Go frontend changes value methods to always check
that the pointer they are passed is not nil. We already dereference
the pointer to copy the value, but if the method does not use the
value then the pointer dereference may be optimized away. Do an
explicit nil check so that we g
On 01/31/2018 03:41 AM, Uros Bizjak wrote:
>>
>> The one concern I have is do we need to tell the CFI machinery that
>> %eax's value was restored to its entry value?
>
> Can you or someone that knows CFI stuff please investigate this a bit?
> I'm not expert in this area, and I don't feel comfortab
On 01/31/2018 10:36 AM, Renlin Li wrote:
Hi there,
I have a patch to fix to regressions we observed in armhf native
environment.
To effectively check out of range for format string, a target type
should be
used. And according to the standard, int type is used for "width" and
"precision"
field o
We currently read and write beyond the builtin jmpbuf on ILP32 targets
where Pmode == DImode and ptr_mode == SImode. Since the builtin jmpbuf
is an array of 5 pointers, ptr_mode should be used to save and restore
frame and program pointers. Since x86 only saves stack pointer in
stack save area,
This patch to libbacktrace closes any debuginfo files that we opened.
Bootstrapped and ran libbacktrace tests on x86_64-pc-linux-gnu. This
should fix https://golang.org/issue/23626. Committed to mainline.
Ian
2018-01-31 Ian Lance Taylor
* elf.c (elf_add): Close descriptor if we use a debug
This patch teaches install_parm_decl (part of the PTX .param
optimization during omp-lowering) not to extract the identifier string
of artificial decls. Apparently, GCC's OpenACC testsuite did not have
any tests that contained local arrays which used alloca for storage
allocation, or else this prob
On Wed, Jan 31, 2018 at 10:18:46PM +0200, Janne Blomqvist wrote:
> This patch fixes the regression by increasing the limit where we fall
> back to runtime to 2**28 elements, which is the same limit where
> previous releases failed. The are still bugs in the runtime
> evaluation, so in many cases lo
On 01/31/2018 01:28 AM, Eric Botcazou wrote:
>> -fstack-check and -fstack-clash both potentially create a loop for stack
>> probing. In that case they both need a scratch register to hold the
>> loop upper bound.
>>
>> The code to allocate a scratch register first starts with the
>> caller-saved r
On 01/31/2018 10:33 PM, Segher Boessenkool wrote:
Hi!
On Wed, Jan 31, 2018 at 08:47:08PM +0100, Uros Bizjak wrote:
We should skip RTXes that will result in certain ICE here, and this is
what the attached patch does.
2018-01-31 Uros Bizjak
PR target/84123
* combine.c (change_zero_
While refining the COMDAT group on Solaris 10 patch
https://gcc.gnu.org/ml/gcc-patches/2018-01/msg02257.html
I noticed that the comdat_group effective-target keyword didn't work
when using Solaris as. No wonder, actually, given the different
syntaxes used:
* with sparc as
.grou
On Wed, 31 Jan 2018, Marek Polacek wrote:
> This fixes one particular instance of the "tokens coming from system headers
> may
> suppress a warning" problem. While we need a more general solution, for GCC 8
> we can at least fix this spot.
>
> Bootstrapped/regtested on x86_64-linux, ok for trun
Hi!
On Wed, Jan 31, 2018 at 08:47:08PM +0100, Uros Bizjak wrote:
> We should skip RTXes that will result in certain ICE here, and this is
> what the attached patch does.
>
> 2018-01-31 Uros Bizjak
>
> PR target/84123
> * combine.c (change_zero_ext): Check if hard register satisfies
>
This fixes one particular instance of the "tokens coming from system headers may
suppress a warning" problem. While we need a more general solution, for GCC 8
we can at least fix this spot.
Bootstrapped/regtested on x86_64-linux, ok for trunk?
2018-01-31 Marek Polacek
PR c/81779
Hi!
We ICE on the following test because rest_of_insert_endbranch
separates a setjmp call from the following NOTE_INSN_CALL_ARG_LOCATION
that must always immediately follow the call.
No other note or debug insn (which aren't around after var-tracking anyway)
needs to follow the call, so the loop i
Hi!
Some spots in the vectorizer create generic COND_EXPRs that in one of the
branches compute some +/-/* arithmetics. When -ftrapv, the gimplifier
ICEs on this as it may trap, we can't emit code with multiple basic blocks
from the APIs and don't want to evaluate it if the guarding condition is
f
Hi!
If the gfc_match_omp_variable_list calls fail, they don't modify head,
so *head means referencing NULL. In that case we don't really need to
free anything, nothing has been allocated, just need to emit a syntax error.
Bootstrapped/regtested on x86_64-linux and i686-linux, committed to trunk.
This fixes the reduced testcase provided by Tom de Vries in comment #7
of the PR. Committed as 'obvious' as r257262. Will await a report from
Tom as to whether or not this fixes the original problem
Paul
2018-01-31 Paul Thomas
PR fortran/84088
* trans-expr.c (gfc_conv_procedure_ca
Hello world,
I have just committed the attached patch as obvious and simple.
Before, the C binding label was ignored for generating
function prototypes with the -fc-prototypes option.
2018-01-31 Thomas Koenig
* dump-parse-tree.c (write_proc): Use sym_name (which may
be sym->b
This patch fixes the regression by increasing the limit where we fall
back to runtime to 2**28 elements, which is the same limit where
previous releases failed. The are still bugs in the runtime
evaluation, so in many cases longer characters will still fail, so
print a warning message.
Regtested o
Hi again,
On 24/01/2018 16:58, Paolo Carlini wrote:
Hi,
I'm looking into this rather mild regression, which should be
relatively easy to fix. In short, Jason's fix for c++/54325 moved an
abstract_virtuals_error_sfinae check from build_aggr_init_expr to
build_cplus_new therefore the testcase
Hi,
The vec-cmpne-long.c tests targets -mcpu=power8, but always dg-do-run, even
on P7 (or earlier) systems. Add a dg-do run requirement for p8vector_hw to
prevent illegal instruction errors.
(noticed during review of test results from a P7).
Sniff tests look good.
OK for trunk?
Th
Hello!
As shown in the PR, alpha specific testcase hits the above ICE when
combine pass is trying to simplify:
(insn 13 12 16 2 (set (reg:SI 141 [ ID ])
(zero_extend:SI (subreg:QI (reg:DI 48 $f16 [ ID ]) 0))) 48
{zero_extendqisi2}
via change_zero_ext (combine.c):
11486 else if (
Hi,
I just committed the patch below as obvious.
Index: ChangeLog
===
--- ChangeLog (revision 257256)
+++ ChangeLog (working copy)
@@ -1,5 +1,10 @@
2018-01-31 Janne Blomqvist
+ * trans-const.c (gfc_conv_string_init): U
Hi,
Noticed during a review of test results on an AIX platform, a few
of the (p9 target specific) tests are expecting to see Power9 codegen, but
are failing to build due to the environment missing P9 assembler support. Thus,
we need to add the power9_vector_ok test requirement to these tests.
On Wed, Jan 31, 2018 at 11:11 AM, augustine.sterl...@gmail.com
wrote:
> On Tue, Jan 30, 2018 at 8:02 PM, Max Filippov wrote:
>>
>> libgcc/
>> 2018-01-31 Max Filippov
>>
>> * config/xtensa/ieee754-df.S (__adddf3_aux): Add
>> .literal_position directive.
>> * config/xtens
On Tue, Jan 30, 2018 at 8:02 PM, Max Filippov wrote:
>
> libgcc/
> 2018-01-31 Max Filippov
>
> * config/xtensa/ieee754-df.S (__adddf3_aux): Add
> .literal_position directive.
> * config/xtensa/ieee754-sf.S (__addsf3_aux): Likewise.
This is fine, but when did it stop wo
Segher Boessenkool writes:
> Hi!
>
> On Fri, Jan 26, 2018 at 01:25:51PM +, Richard Sandiford wrote:
>>if (SCALAR_INT_MODE_P (inmode))
>> new_out_reg = gen_lowpart_SUBREG (outmode, reg);
>>else
>> -new_out_reg = gen_rtx_SUBREG (outmode, reg, 0);
>> +{
>>
Eric Botcazou writes:
>> Tested on SPARC64/Linux and ARM/EABI, applied on mainline and 7 branch.
>
> As discussed in the audit trail, this beefs up the internal documentation
> about WORD_REGISTER_OPERATIONS.
>
> Tested with 'make doc', applied on mainline and 7 branch.
>
>
> 2018-01-31 Eric Bot
The following patch fixes ICE for
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82444
The patch was successfully bootstrapped on x86_64.
Committed as rev. 257254.
Index: ChangeLog
===
--- ChangeLog (revision 257252)
+++ Chang
On Tue, Jan 30, 2018 at 04:43:20PM -0600, Will Schmidt wrote:
> Hi,
> Some VSX function has previosly crept into the altivec-13 testcase. In
> particular,
> anything 'vector long long' and 'vector double', causing issues on platforms
> without
> VSX support. So, break that content out into i
This patch to the Go frontend, based on one by Cherry Zhang, lowers
expression types in the lowering pass. This ensures that array types
with complicated length expressions are handled correctly.
This required some adjustment of constant expression types to not
report too many errors for circular
The big GO symbol rename missed a case in libgo's C code, in the
fieldtrack support. This patch fixes it. Bootstrapped and ran Go
testsuite on x86_64-pc-linux-gnu. Committed to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
===
--- g
This libgo patch changes the gotest script to accept symbols with a
leading dot, as found on AIX. While we're here stop doing fgrep -v
'$', symbol names no longer contain '$' anyhow. Bootstrapped and ran
Go testsuite on x86_64-pc-linux-gnu. Committed to mainline.
Ian
Index: gcc/go/gofrontend/ME
On 1/31/18 11:39 AM, Segher Boessenkool wrote:
> On Mon, Jan 29, 2018 at 08:55:35PM -0600, Peter Bergner wrote:
>>
>> Either that, or I could still call candidates_list_and_hint() and just
>> throw the hint away, since it's meaningless.
>
> It's less code, simpler code, so that would be best I thi
OK.
On Wed, Jan 31, 2018 at 12:45 PM, Jakub Jelinek wrote:
> On Wed, Jan 31, 2018 at 11:02:36AM -0500, Jason Merrill wrote:
>> > In this case I was just extending existing warning and wanted
>> > consistency with that. Both can be changed in a GCC9 follow-up,
>> > or if Jason/Nathan want it now,
On Tue, Jan 30, 2018 at 03:45:17PM +, Renlin Li wrote:
> Hi Richard,
>
> Thanks for the review!
>
> On 29/01/18 20:23, Richard Sandiford wrote:
> >
> > The patch looks good to me FWIW. How about adding something like
> > the following testcase?
> >
> >
On Mon, Jan 29, 2018 at 09:41:12PM -0700, Jeff Law wrote:
> On 01/07/2018 10:09 AM, Segher Boessenkool wrote:
> >> --- a/gcc/testsuite/gcc.target/powerpc/pr56605.c
> >> +++ b/gcc/testsuite/gcc.target/powerpc/pr56605.c
> >> @@ -1,7 +1,7 @@
> >> /* PR rtl-optimization/56605 */
> >> -/* { dg-do compi
On Wed, Jan 31, 2018 at 11:02:36AM -0500, Jason Merrill wrote:
> > In this case I was just extending existing warning and wanted
> > consistency with that. Both can be changed in a GCC9 follow-up,
> > or if Jason/Nathan want it now, even for GCC8, sure.
So, do you want %qE as in the patch or %E?
On Mon, Jan 29, 2018 at 08:55:35PM -0600, Peter Bergner wrote:
>
> Either that, or I could still call candidates_list_and_hint() and just
> throw the hint away, since it's meaningless.
It's less code, simpler code, so that would be best I think yes.
Segher
Hi there,
I have a patch to fix to regressions we observed in armhf native environment.
To effectively check out of range for format string, a target type should be
used. And according to the standard, int type is used for "width" and
"precision"
field of format string.
Here target_strtol10 is
On Mon, Jan 29, 2018 at 08:37:09PM -0600, Peter Bergner wrote:
> On 1/29/18 6:30 PM, Segher Boessenkool wrote:
> > On Mon, Jan 29, 2018 at 02:54:15PM -0600, Peter Bergner wrote:
> > Why don't you want that? It let's the compiler say "hey silly human who
> > can hardly type his own name correctly(*
On 01/31/2018 08:39 AM, David Malcolm wrote:
PR 84136 reports an ICE within sccvn_dom_walker when handling a
C/C++ source file that overuses the labels-as-values extension.
The code in question stores a jump label into a global, and then
jumps to it from another function, which ICEs after inlinin
On Wed, Jan 31, 2018 at 11:21 AM, Jakub Jelinek wrote:
> On Wed, Jan 31, 2018 at 11:05:17AM -0500, Jason Merrill wrote:
>> > 2018-01-24 Jakub Jelinek
>> >
>> > PR c++/83993
>> > * constexpr.c (cxx_eval_outermost_constant_expr): Don't clear
>> > TREE_CONSTANT on ADDR_EXPR
On Mon, Jan 29, 2018 at 11:07:12PM -0700, Jeff Law wrote:
> OK. Makes me wonder how many big endian LRA targets are getting
> significant use.
powerpc and powerpc64 are BE, but we don't often have QImode and HImode
(all instructions work on 32-bit or 64-bit words).
Segher
Hi!
On Fri, Jan 26, 2018 at 01:25:51PM +, Richard Sandiford wrote:
> if (SCALAR_INT_MODE_P (inmode))
> new_out_reg = gen_lowpart_SUBREG (outmode, reg);
> else
> - new_out_reg = gen_rtx_SUBREG (outmode, reg, 0);
> + {
> + poly_uint64 offset =
On Wed, Jan 31, 2018 at 11:05:17AM -0500, Jason Merrill wrote:
> > 2018-01-24 Jakub Jelinek
> >
> > PR c++/83993
> > * constexpr.c (cxx_eval_outermost_constant_expr): Don't clear
> > TREE_CONSTANT on ADDR_EXPRs.
> >
> > * g++.dg/init/pr83993-2.C: New test.
> >
> >
On Wed, Jan 24, 2018 at 6:22 PM, Jakub Jelinek wrote:
> Hi!
>
> cxx_eval_outermost_constant_expr clears TREE_CONSTANT on ADDR_EXPRs that
> aren't considered by C++ constant expressions, but that breaks middle-end
> which relies on TREE_CONSTANT being set on ADDR_EXPR where the address
> is constan
On Thu, Jan 25, 2018 at 5:53 PM, Jakub Jelinek wrote:
> On Thu, Jan 25, 2018 at 01:13:56PM -0700, Martin Sebor wrote:
>> On 01/24/2018 04:19 PM, Jakub Jelinek wrote:
>> > Hi!
>> >
>> > In constexpr evaluation of array references for arrays with unknown bounds,
>> > we need to diagnose out of bound
On 01/31/2018 01:28 AM, Eric Botcazou wrote:
>> -fstack-check and -fstack-clash both potentially create a loop for stack
>> probing. In that case they both need a scratch register to hold the
>> loop upper bound.
>>
>> The code to allocate a scratch register first starts with the
>> caller-saved r
OK.
On Mon, Jan 29, 2018 at 5:00 PM, Paolo Carlini wrote:
> Hi,
>
> On 29/01/2018 21:41, Jason Merrill wrote:
>>
>> On Mon, Jan 29, 2018 at 10:45 AM, Paolo Carlini
>> wrote:
>>>
>>> the fix for c++/81236 removed some special code for dependent_p from
>>> finish_id_expression, and now finish_qual
Hi!
I'd like to ping following patches:
http://gcc.gnu.org/ml/gcc-patches/2018-01/msg02066.html
- PR83993 - fix constexpr handling of arrays with unknown bound
http://gcc.gnu.org/ml/gcc-patches/2018-01/msg02067.html
- PR83993 - don't clear TREE_CONSTANT on ADDR_EXPRs in constexpr.c
Thanks
PR 84136 reports an ICE within sccvn_dom_walker when handling a
C/C++ source file that overuses the labels-as-values extension.
The code in question stores a jump label into a global, and then
jumps to it from another function, which ICEs after inlining:
void* a;
void foo() {
if ((a = &&l))
OK.
On Wed, Jan 31, 2018 at 6:28 AM, Marek Polacek wrote:
> This fixes an ICE-on-invalid, where we're passing error_mark_node to
> useless_type_conversion_p which can't handle that.
>
> Bootstrapped/regtested on x86_64-linux, ok for trunk?
>
> 2018-01-31 Marek Polacek
>
> PR c++/84138
I've applied this patch to openacc-gcc-7-branch which teaches the
gimplifier how to pass certain OpenACC reduction variables as
firstprivate, and not with an implicit copy directive. This is matches
the default behavior for the implicit data mappings of scalar variables
inside OpenACC parallel regi
This patch implements the original suggestion for fixing PR 81635:
use range info in split_constant_offset to see whether a conversion
of a wrapping type can be split. The range info problem described in:
https://gcc.gnu.org/ml/gcc-patches/2017-08/msg01002.html
seems to have been fixed.
The p
> Tested on SPARC64/Linux and ARM/EABI, applied on mainline and 7 branch.
As discussed in the audit trail, this beefs up the internal documentation
about WORD_REGISTER_OPERATIONS.
Tested with 'make doc', applied on mainline and 7 branch.
2018-01-31 Eric Botcazou
PR rtl-optimization
On Wed, 31 Jan 2018, Christophe Lyon wrote:
> On 30 January 2018 at 11:47, Jakub Jelinek wrote:
> > On Tue, Jan 30, 2018 at 11:07:50AM +0100, Richard Biener wrote:
> >>
> >> I have been asked to push this change, fixing (somewhat) the impreciseness
> >> of costing constant/invariant vector uses i
This patch to libgo renames the variable TestAddr6. The script
libgo/testsuite/gotest picks up names that start with Test to find the
test functions to run. On most systems it only looks for functions,
but on PPC64, because of the function descriptors used for ELF ABI v1,
it also looks for data s
Maxim Kuvyrkov writes:
>> On Jan 31, 2018, at 4:33 PM, Wilco Dijkstra wrote:
>>
>> Richard Sandiford wrote:
>>
>>> This was the original intent, but was changed in r213708. TBH I'm not
>>> sure what the second hunk in that revision fixed, since model_index is
>>> supposed to return an index gr
On 30 January 2018 at 11:47, Jakub Jelinek wrote:
> On Tue, Jan 30, 2018 at 11:07:50AM +0100, Richard Biener wrote:
>>
>> I have been asked to push this change, fixing (somewhat) the impreciseness
>> of costing constant/invariant vector uses in SLP stmts. The previous
>> code always just consider
> On Jan 31, 2018, at 4:33 PM, Wilco Dijkstra wrote:
>
> Richard Sandiford wrote:
>
>> This was the original intent, but was changed in r213708. TBH I'm not
>> sure what the second hunk in that revision fixed, since model_index is
>> supposed to return an index greater than all valid indices wh
Richard Sandiford wrote:
> This was the original intent, but was changed in r213708. TBH I'm not
> sure what the second hunk in that revision fixed, since model_index is
> supposed to return an index greater than all valid indices when passed
> an instruction outside the current block. Maxim, do
On Tue, 30 Jan 2018, Richard Biener wrote:
> On Tue, 30 Jan 2018, Richard Biener wrote:
>
> >
> > This patch tries to deal with the "easy" part of a function ABI,
> > the return value location, in vectorization costing. The testcase
> > shows that if we vectorize the returned value but the func
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2018-01-31 Richard Biener
PR tree-optimization/84132
* tree-data-ref.c (analyze_miv_subscript): Properly
check whether evolution_function_is_affine_multivariate_p
before calling gcd_of_step
Wilco Dijkstra writes:
> The comparison order for SCHED_PRESSURE_MODEL is incorrect. If either
> instruction is not in target_bb, the ordering is not well defined. To fix
> this, give all instructions in target_bb the highest priority and sort all
> other instructions behind it. This way instru
On Wed, 31 Jan 2018, Wilco Dijkstra wrote:
> The comparison order for SCHED_PRESSURE_MODEL is incorrect. If either
> instruction is not in target_bb, the ordering is not well defined. To fix
> this, give all instructions in target_bb the highest priority and sort all
> other instructions behind
The comparison order for SCHED_PRESSURE_MODEL is incorrect. If either
instruction is not in target_bb, the ordering is not well defined. To fix
this, give all instructions in target_bb the highest priority and sort all
other instructions behind it. This way instructions in target_bb will be
sort
> OK, so you think the test above should check
> && (!TYPE_NONALIASED_COMPONENT (t1) || TYPE_STRUCTURAL_EQUALITY_P (t1))
> && (!TYPE_NONALIASED_COMPONENT (t2) || TYPE_STRUCTURAL_EQUALITY_P (t2))
I'm not sure about TYPE_STRUCTURAL_EQUALITY_P (does it matter in LTO mode?)
but, yes, TYPE_NONALIASE
This fixes an ICE-on-invalid, where we're passing error_mark_node to
useless_type_conversion_p which can't handle that.
Bootstrapped/regtested on x86_64-linux, ok for trunk?
2018-01-31 Marek Polacek
PR c++/84138
* cp-gimplify.c (cp_fold): Check if X is an error node before
> > 2018-01-30 Jan Hubicka
> >
> >gcc:
> >PR lto/83954
> >* lto-symtab.c (warn_type_compatibility_p): Silence false positive
> >for type match warning on arrays of pointers.
> >
> >gcc/testsuite:
> >PR lto/83954
> >* gcc.dg/lto/pr83954.h:
On Wed, Jan 31, 2018 at 12:02 PM, Eric Botcazou wrote:
>> 2018-01-30 Jan Hubicka
>>
>>gcc:
>>PR lto/83954
>>* lto-symtab.c (warn_type_compatibility_p): Silence false positive
>>for type match warning on arrays of pointers.
>>
>>gcc/testsuite:
>>PR
On Tue, Jan 30, 2018 at 11:11 PM, Aldy Hernandez wrote:
> Hi!
>
> [Note: Jakub has mentioned that missing -Warray-bounds regressions should be
> punted to GCC 9. I think this particular one is easy pickings, but if this
> and/or the rest of the -Warray-bounds regressions should be marked as GCC 9
> 2018-01-30 Jan Hubicka
>
>gcc:
>PR lto/83954
>* lto-symtab.c (warn_type_compatibility_p): Silence false positive
>for type match warning on arrays of pointers.
>
>gcc/testsuite:
>PR lto/83954
>* gcc.dg/lto/pr83954.h: New testcase.
>
On Tue, Dec 19, 2017 at 4:36 PM, Bin Cheng wrote:
> HI,
> This patch backports r254778 and test case in r244815 to GCC6. Bootstrap and
> test on x86_64. Is it OK?
Ok.
Richard.
> Thanks,
> bin
>
> 2017-12-18 Bin Cheng
>
> Backport from mainline
> 2017-11-15 Bin Cheng
>
>
On Wed, Jan 31, 2018 at 6:35 AM, Jeff Law wrote:
> Whee, fun, this appears to have been broken for a very long time,
> possibly since the introduction of -fstack-check in 2010. Thankfully it
> only affects 32 bit and only in relatively constrained circumstances.
>
> -fstack-check and -fstack-clas
On Wed, Jan 31, 2018 at 11:28 AM, Kugan Vivekanandarajah
wrote:
> Hi Richard,
>
> Thanks for the review.
> On 25 January 2018 at 20:04, Richard Biener
> wrote:
>> On Wed, Jan 24, 2018 at 10:56 PM, Kugan Vivekanandarajah
>> wrote:
>>> Hi All,
>>>
>>> Here is a patch for popcount builtin detectio
On Tue, Jan 30, 2018 at 7:29 PM, Aldy Hernandez wrote:
> Hi!
>
> As discussed in the PR, the ICE here happens in dump_generic_node:
>
> case FUNCTION_TYPE:
> case METHOD_TYPE:
> ...
> if (TYPE_NAME (node) && DECL_NAME (TYPE_NAME (node)))
> dump_decl_name (pp, TYPE_NAME (node)
On Wed, Jan 31, 2018 at 11:28:26AM +0100, Jan Hubicka wrote:
> > On Wed, Jan 31, 2018 at 10:16:10AM +0100, Jan Hubicka wrote:
> > > > On Wed, Jan 31, 2018 at 09:51:31AM +0100, Christophe Lyon wrote:
> > > > > Even with Rainer's fix, I've noticed that the new test fails on
> > > > > arm/aarch64:
>
Hi Richard,
Thanks for the review.
On 25 January 2018 at 20:04, Richard Biener wrote:
> On Wed, Jan 24, 2018 at 10:56 PM, Kugan Vivekanandarajah
> wrote:
>> Hi All,
>>
>> Here is a patch for popcount builtin detection similar to LLVM. I
>> would like to queue this for review for next stage 1.
>>
> On Wed, Jan 31, 2018 at 10:16:10AM +0100, Jan Hubicka wrote:
> > > On Wed, Jan 31, 2018 at 09:51:31AM +0100, Christophe Lyon wrote:
> > > > Even with Rainer's fix, I've noticed that the new test fails on
> > > > arm/aarch64:
> > > > FAIL: g++.dg/torture/pr81360.C -O2 -flto -fuse-linker-plugin
> I think this caused:
>
> .././src/gcc/config/sparc/sparc.md: In function ‘rtx_def*
> gen_vxworks_load_got()’: .././src/gcc/config/sparc/sparc.md:1811:17: error:
> ‘global_offset_table_rtx’ was not declared in this scope operands[0] =
> global_offset_table_rtx;
>
> on sparc-wrs-vxworks. (Might
WORD_REGISTER_OPERATIONS target.
2018-01-31 Eric Botcazou
* gcc.c-torture/execute/20180131-1.c: New test.
--
Eric BotcazouIndex: combine.c
===
--- combine.c (revision 257139)
+++ combine.c (working copy)
@@ -13245,18
Eric Botcazou writes:
> This fixes a somewhat obscure interaction between alloca and setjmp/longjmp
> in
> PIC mode on the SPARC architecture. The problem is that the canonical PIC
> register on SPARC (%l7) is call-saved like on other architectures but, unlike
> on other architectures, not (a
Hi,
This patch fixes invalid profile count information in vectorization peeling.
Current implementation is a bit confusing for me since it tries to compute
an overall probability based on scaling probability and change of estimated
niters. This patch does it in two steps. Firstly it does the scal
On Wed, Jan 31, 2018 at 09:46:32AM +, Kyrill Tkachov wrote:
> Hi all,
>
> This test has been failing since forever, it has never passed AFAIK.
> The PR details the vectoriser deficiency.
> I propose we xfail this with a reference to the PR.
>
> Ok for trunk?
Yes please. We're long overdue on
Hi all,
This test has been failing since forever, it has never passed AFAIK.
The PR details the vectoriser deficiency.
I propose we xfail this with a reference to the PR.
Ok for trunk?
Thanks,
Kyrill
2018-01-31 Kyrylo Tkachov
PR tree-optimization/64946
* gcc.target/aarch64/vect-abs
When compiling MariaDB 10.2 (which uses cfi directives in inline asm,
but gets the conditionals wrong) on Solaris 11.4 with an
x86_64-pc-solaris2.11 gcc using gas, I noticed that this one incorrectly
doesn't use cfi directives although it could/should.
Looking closer, I noticed that the logic in g
As described in the PR, cc1 SEGVs compiling gcc.dg/rtl/x86_64/final.c on
targets that default to -fno-dwarf2-cfi-asm. However, the test is
compile-only, so doesn't depend on the toolchain's support for cfi
directives. Besides, it scans the output for those, so depends on
-fdwarf2-cfi-asm anyway,
On Wed, Jan 31, 2018 at 10:16:10AM +0100, Jan Hubicka wrote:
> > On Wed, Jan 31, 2018 at 09:51:31AM +0100, Christophe Lyon wrote:
> > > Even with Rainer's fix, I've noticed that the new test fails on
> > > arm/aarch64:
> > > FAIL: g++.dg/torture/pr81360.C -O2 -flto -fuse-linker-plugin
> > > -fno
> On Wed, Jan 31, 2018 at 09:51:31AM +0100, Christophe Lyon wrote:
> > Even with Rainer's fix, I've noticed that the new test fails on arm/aarch64:
> > FAIL: g++.dg/torture/pr81360.C -O2 -flto -fuse-linker-plugin
> > -fno-fat-lto-objects scan-ipa-dump icf "Equal symbols: 0"
>
> With -flto -fno
On Wed, Jan 31, 2018 at 09:51:31AM +0100, Christophe Lyon wrote:
> Even with Rainer's fix, I've noticed that the new test fails on arm/aarch64:
> FAIL: g++.dg/torture/pr81360.C -O2 -flto -fuse-linker-plugin
> -fno-fat-lto-objects scan-ipa-dump icf "Equal symbols: 0"
With -flto -fno-fat-lto-obj
On 30 January 2018 at 23:42, Jakub Jelinek wrote:
> On Tue, Jan 30, 2018 at 09:05:31PM +0100, Rainer Orth wrote:
>> > this patch fixed ICE that was introduced by Richard Standiford's change to
>> > reorder
>> > can and want_inline predicates to reduce amount of work done to verify
>> > inlining
> -fstack-check and -fstack-clash both potentially create a loop for stack
> probing. In that case they both need a scratch register to hold the
> loop upper bound.
>
> The code to allocate a scratch register first starts with the
> caller-saved registers as they're zero cost. Then it'll use any
> The original warning was
>
> /vol/gcc/src/hg/trunk/local/gcc/testsuite/gnat.dg/lto20_pkg.ads:7:13:
> warning: type of 'lto20_pkg__proc' does not match original declaration
> [-Wlto-type-mismatch]
>
> so just removing the dg-excess-errors seems the right thing to do.
> Tested with the appropriat
100 matches
Mail list logo
