Hi!
I've discovered a pasto, which breaks the following testcase.
Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux,
committed to trunk, will backport to 9.2 too.
2019-07-13 Jakub Jelinek
PR c/91149
* c-omp.c (c_omp_split_clauses): Fix a pasto in
OMP
Hi!
The constructs with order(concurrent) clause have some extra restrictions,
this patch diagnoses those that are easy to diagnose.
Bootstrapped/regtested on x86_64-linux and i686-linux, committed to trunk.
2019-07-13 Jakub Jelinek
* gimplify.c (struct gimplify_omp_ctx): Add order_c
a result,
>it falls back on using the version.in file verbatim, meaning that
>the "DATE" placeholder doesn't get replaced with the snapshot date,
>and the "-git" suffix doesn't get stripped. More precisely, we get
>snapshots called "gdb-8.3.50-DATE-g
On Fri, Jul 12, 2019 at 4:59 PM Ren Kimura wrote:
>
> This patch fixes a Bug 90924.
> simple_object_elf functions don't load section table 0 of ELF file, which is
> not a useful.
> However If e_shstrndx in ELF header points to a section table 0 (i.e.
> e_shstrndx == 0), a calculation of offset t
This patch fixes a Bug 90924.
simple_object_elf functions don't load section table 0 of ELF file, which is
not a useful.
However If e_shstrndx in ELF header points to a section table 0 (i.e.
e_shstrndx == 0), a calculation of offset to string section table causes
integer overflow at every line "
Oh. I missed a ChangeLog entry.
Hold on... I'll send v3 patch.
Thanks
2019年7月12日(金) 19:37 Ren Kimura :
>
> This patch fixes a Bug 90924.
> simple_object_elf functions don't load section table 0 of ELF file, which is
> not a useful.
> However If e_shstrndx in ELF header points to a section table
This patch fixes a Bug 90924.
simple_object_elf functions don't load section table 0 of ELF file, which is
not a useful.
However If e_shstrndx in ELF header points to a section table 0 (i.e.
e_shstrndx == 0), a calculation of offset to string section table causes
integer overflow at every line "
"DATE" placeholder doesn't get replaced with the snapshot date,
and the "-git" suffix doesn't get stripped. More precisely, we get
snapshots called "gdb-8.3.50-DATE-git.tar" instead of (e.g.)
"gdb-8.3-20190712.tar".
For those who have been getting snaps
On Fri, Jul 12, 2019 at 3:04 PM Ren Kimura wrote:
>
> This patch fixes a Bug 90924.
> simple_object_elf functions don't load section table 0 of ELF file, which is
> not a useful.
> However If e_shstrndx in ELF header points to a section table 0 (i.e.
> e_shstrndx == 0), a calculation of offset t
On Fri, Jul 12, 2019 at 02:38:59PM -0400, Marek Polacek wrote:
> On Fri, Jul 12, 2019 at 02:34:37PM -0400, Matthew Beliveau wrote:
> > @@ -1628,6 +1629,16 @@ c_do_switch_warnings (splay_tree cases, location_t
> > switch_location,
> >if (cond && tree_int_cst_compare (cond, value))
> > c
This patch fixes a Bug 90924.
simple_object_elf functions don't load section table 0 of ELF file, which is
not a useful.
However If e_shstrndx in ELF header points to a section table 0 (i.e.
e_shstrndx == 0), a calculation of offset to string section table causes
integer overflow at every line "
On 7/12/19 2:36 PM, Giuliano Belinassi wrote:
> Hi,
>
> It seems that this patch wasn't merged to trunk yet. Is there any
> problem to it?
Nope. Just assumed you had access and would commit it yourself. I'll
go ahead and commit it for you shortly.
jeff
Hi,
It seems that this patch wasn't merged to trunk yet. Is there any
problem to it?
On 07/02, Jeff Law wrote:
> On 7/2/19 6:36 AM, Giuliano Belinassi wrote:
> > This patch makes lto-dump the symtab callgraph in graphviz DOT format.
> >
> > Previously, the suggested patch had some minor and chec
In order to fix 89906 in GCC 8, we need to backport 86098.
I think the patch is safe to be backported.
Tested x86_64-linux, ok for 8?
2018-06-12 Jason Merrill
PR c++/86098 - ICE with template placeholder for TTP.
* typeck.c (structural_comptypes) [TEMPLATE_TYPE_PARM]: Check
During GCC-9, the codegen for unreachable switch case statements changed
such that the (undefined) behaviour of reaching such statements is directed
to one of the existing switch cases. This means that the testcase which
deals with the old behaviour can no longer work (and there is nothing to test
On Fri, Jul 12, 2019 at 02:34:37PM -0400, Matthew Beliveau wrote:
> @@ -1628,6 +1629,16 @@ c_do_switch_warnings (splay_tree cases, location_t
> switch_location,
>if (cond && tree_int_cst_compare (cond, value))
> continue;
>
> + /* If the enumerator is defined in a system heade
Hi,
This should fix the concerns you had! Now the function only gets the
location from the type, and then only look for the name if the type's
location is in the header.
On Tue, Jul 9, 2019 at 5:21 PM Jason Merrill wrote:
>
> On 7/9/19 11:18 AM, Matthew Beliveau wrote:
> > This patch suppresses t
Segher Boessenkool writes:
> On Thu, Jul 11, 2019 at 06:09:10PM +0100, Gaius Mulley wrote:
>> Segher Boessenkool writes:
>>
>> > On Thu, Jul 11, 2019 at 12:49:44PM +0100, Gaius Mulley wrote:
>> >> Matthias Klose writes:
>> >> > powerpc64le-linux-gnu fails to build (search for "unfinished" in t
Hi,
I've now backported this to all active branches so the issue can no
longer occur:
Turn off ipa-ra in builtins test (PR91059)
The gcc.c-torture/execute/builtins/lib directory contains a reimplementation
of many C library string functions, which causes non-trivial register allocation
bugs with
On 7/12/19 3:35 AM, Richard Biener wrote:
On Thu, Jul 11, 2019 at 7:43 PM Martin Sebor wrote:
Attached is a patch that adds a new parameter to limit the number
of SSA_NAME assignments for GCC to follow in iterative or recursive
algorithms. Purely as a proof of concept the patch introduces
the
On Thu, Jul 11, 2019 at 03:44:02PM -0400, Michael Meissner wrote:
> --- gcc/config/rs6000/rs6000-protos.h (revision 273371)
> +++ gcc/config/rs6000/rs6000-protos.h (working copy)
> @@ -154,6 +154,18 @@ extern align_flags rs6000_loop_align (rt
> extern void rs6000_split_logical (rtx [], enum rtx_co
On 7/12/19 9:36 AM, Jason Merrill wrote:
On Fri, Jul 12, 2019 at 11:14 AM Martin Sebor wrote:
On 7/12/19 5:42 AM, Jonathan Wakely wrote:
On 12/07/19 10:24 +0200, Jakub Jelinek wrote:
On Mon, Jul 08, 2019 at 03:56:51PM -0600, Martin Sebor wrote:
A couple of GCC's Coding Conventions call to
On 7/8/19 1:39 AM, JiangNing OS wrote:
> Hi Jeff and Richard B.,
>
> Following your tips, I've found a much simpler solution in tree-ssa-phiopt.c.
> Attached is the new patch. Review again, please!
>
> Thanks a lot!
> -Jiangning
>
>> -Original Message-
>> From: Jeff Law
>> Sent: Saturd
On 7/12/19 9:26 AM, Jonathan Wakely wrote:
On 12/07/19 09:14 -0600, Martin Sebor wrote:
On 7/12/19 5:42 AM, Jonathan Wakely wrote:
On 12/07/19 10:24 +0200, Jakub Jelinek wrote:
On Mon, Jul 08, 2019 at 03:56:51PM -0600, Martin Sebor wrote:
A couple of GCC's Coding Conventions call to
1) Use
On 7/11/19 10:08 PM, Andrew Pinski wrote:
> On Mon, Jul 8, 2019 at 12:39 AM JiangNing OS
> wrote:
>>
>> Hi Jeff and Richard B.,
>>
>> Following your tips, I've found a much simpler solution in
>> tree-ssa-phiopt.c. Attached is the new patch. Review again, please!
>
> /* Prove that we can move
On Wed, 19 Jun 2019 at 20:07, Jonathan Wakely wrote:
>
> On Wed, 19 Jun 2019 at 20:05, Joel Sherrill wrote:
> >
> > Hi
> >
> > I was double checking the C++17 support in GCC for someone and the text at
> > this URL states
> > the support is experimental and gives the impression that the support i
Committed to CVS.
Index: htdocs/gcc-9/changes.html
===
RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-9/changes.html,v
retrieving revision 1.71
diff -u -r1.71 changes.html
--- htdocs/gcc-9/changes.html 10 Jul 2019 13:03:55 - 1.71
+++ htd
On 11/07/19 20:45 +0100, Jonathan Wakely wrote:
+ // Repeat for volatile std::atomic
+ if constexpr (std::atomic::is_always_lock_free)
Thanks to Uros for pointing out this typo. Fixed by the attached
patch, committed to trunk.
commit 991bdaf97870e2775a15314823759aec7fd79599
Author: redi
Da
My tester reports ~22000 unexpected failures for the c6x port. The vast
majority are due to a linker warning when the type of the .far section
changes.
It turns out we can emit two declarations for the .far section. One
with @nobits, the other without. And since the .far section is a bss
secti
On Thu, Jul 11, 2019 at 06:09:10PM +0100, Gaius Mulley wrote:
> Segher Boessenkool writes:
>
> > On Thu, Jul 11, 2019 at 12:49:44PM +0100, Gaius Mulley wrote:
> >> Matthias Klose writes:
> >> > powerpc64le-linux-gnu fails to build (search for "unfinished" in the
> >> > build log)
> >> >
> >> >
On Fri, Jul 12, 2019 at 11:14 AM Martin Sebor wrote:
>
> On 7/12/19 5:42 AM, Jonathan Wakely wrote:
> > On 12/07/19 10:24 +0200, Jakub Jelinek wrote:
> >> On Mon, Jul 08, 2019 at 03:56:51PM -0600, Martin Sebor wrote:
> >>> A couple of GCC's Coding Conventions call to
> >>>
> >>> 1) Use the struc
On 12/07/19 09:14 -0600, Martin Sebor wrote:
On 7/12/19 5:42 AM, Jonathan Wakely wrote:
On 12/07/19 10:24 +0200, Jakub Jelinek wrote:
On Mon, Jul 08, 2019 at 03:56:51PM -0600, Martin Sebor wrote:
A couple of GCC's Coding Conventions call to
1) Use the struct keyword for plain old data (POD)
On 7/12/19 5:42 AM, Jonathan Wakely wrote:
On 12/07/19 10:24 +0200, Jakub Jelinek wrote:
On Mon, Jul 08, 2019 at 03:56:51PM -0600, Martin Sebor wrote:
A couple of GCC's Coding Conventions call to
1) Use the struct keyword for plain old data (POD) types.
https://www.gnu.org/software/gcc/
On Fri, Jul 12, 2019 at 7:42 AM Jonathan Wakely wrote:
>
> On 12/07/19 10:24 +0200, Jakub Jelinek wrote:
> >On Mon, Jul 08, 2019 at 03:56:51PM -0600, Martin Sebor wrote:
> >> A couple of GCC's Coding Conventions call to
> >>
> >> 1) Use the struct keyword for plain old data (POD) types.
> >>
On 7/12/19 8:44 AM, Martin Liška wrote:
On 7/11/19 7:25 PM, Jason Merrill wrote:
On 7/11/19 3:02 AM, Jakub Jelinek wrote:
On Thu, Jul 11, 2019 at 08:48:52AM +0200, Martin Liška wrote:
--- a/gcc/c-family/c-opts.c
+++ b/gcc/c-family/c-opts.c
@@ -501,6 +501,8 @@ c_common_handle_option (size_t sco
On Fri, 12 Jul 2019 at 08:06, Richard Sandiford
wrote:
>
> Christophe Lyon writes:
> > In FDPIC, we need to make sure __do_global_dtors_aux and frame_dummy
> > are referenced by their address, not by pointers to the function
> > descriptors.
> >
> > 2019-XX-XX Christophe Lyon
> > Mickaël
On Fri, Jul 12, 2019 at 02:11:16PM +0200, Richard Biener wrote:
> Am I correct that doloop HW implementations are constrainted
> by a decrement of one?
GCC's doloop patterns are. Not all hardware is.
Segher
Christophe Lyon writes:
> On Fri, 12 Jul 2019 at 08:49, Richard Sandiford
> wrote:
>>
>> Christophe Lyon writes:
>> > The new arm-uclinuxfdpiceabi target behaves pretty much like
>> > arm-linux-gnueabi. In order the enable the same set of features, we
>> > have to update several configure script
On Fri, 12 Jul 2019 at 08:49, Richard Sandiford
wrote:
>
> Christophe Lyon writes:
> > The new arm-uclinuxfdpiceabi target behaves pretty much like
> > arm-linux-gnueabi. In order the enable the same set of features, we
> > have to update several configure scripts that generally match targets
> >
On 7/11/19 7:25 PM, Jason Merrill wrote:
> On 7/11/19 3:02 AM, Jakub Jelinek wrote:
>> On Thu, Jul 11, 2019 at 08:48:52AM +0200, Martin Liška wrote:
>>> --- a/gcc/c-family/c-opts.c
>>> +++ b/gcc/c-family/c-opts.c
>>> @@ -501,6 +501,8 @@ c_common_handle_option (size_t scode, const char *arg,
>>> HO
The following fixes the PR.
Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
Richard.
2019-07-12 Richard Biener
PR tree-optimization/91145
* tree-vect-slp.c (vect_build_slp_tree_2): Fix reduction
chain check.
* gcc.dg/torture/pr91145.c: New te
On Wed, 10 Jul 2019, Kewen.Lin wrote:
> Hi all,
>
> I'd like to gentle ping the below patch:
> https://gcc.gnu.org/ml/gcc-patches/2019-06/msg01225.html
>
> The previous version for more context/background:
> https://gcc.gnu.org/ml/gcc-patches/2019-06/msg01126.html
>
> Thanks a lot in advance!
On 12/07/19 12:20 +0100, Jonathan Wakely wrote:
On 11/07/19 20:45 +0100, Jonathan Wakely wrote:
This adds the new atomic types from C++2a, as proposed by P0019 and
P0020. To reduce duplication the calls to the compiler's atomic
built-ins are wrapped in new functions in the __atomic_impl namespac
On 11/07/19 20:45 +0100, Jonathan Wakely wrote:
This adds the new atomic types from C++2a, as proposed by P0019 and
P0020. To reduce duplication the calls to the compiler's atomic
built-ins are wrapped in new functions in the __atomic_impl namespace.
These functions are currently only used by std
This defines the equivalent of C++2a's std::type_identity_t alias but
for use in C++11 and later. This can be used to replace __detail::__idt
in the string_view headers, which previously used common_type_t because
the one argument specialization of common_type_t was simply the identity
transform (
On 12/07/19 10:24 +0200, Jakub Jelinek wrote:
On Mon, Jul 08, 2019 at 03:56:51PM -0600, Martin Sebor wrote:
A couple of GCC's Coding Conventions call to
1) Use the struct keyword for plain old data (POD) types.
https://www.gnu.org/software/gcc/codingrationale.html#struct
and
2) Use t
On Fri, Jul 12, 2019 at 12:36:13PM +0100, Kwok Cheung Yeung wrote:
> --- a/gcc/omp-general.c
> +++ b/gcc/omp-general.c
> @@ -48,6 +48,20 @@ omp_find_clause (tree clauses, enum omp_clause_code kind)
>return NULL_TREE;
> }
>
> +/* Return true if DECL is a Fortran optional argument. */
> +
> +b
This adds testcases exercising the use of optional arguments in the
various OpenACC directives. Where applicable, both the present and
non-present cases are tested, with an integer, array of integers and
allocatable array of integers as the argument.
libgomp/
* testsuite/libgom
This patch fixes a similar situation that occurs with the use_device
clause, where the lowering would result in a null dereference if applied
to a non-present optional argument. This patch builds a conditional
check that skips the dereference if the argument is non-present, and
ensures that opt
This patch allows allocatable arrays passed as Fortran optional
arguments to be used in OpenACC. The GIMPLE code generated by the
current lowering unconditionally attempts to access fields within the
structure representing the array, resulting in a null dereference if the
array is non-present.
Reference types used by Fortran often need to be treated specially in
the OACC lowering to deal with the referenced object as well as the
reference itself. However, as optional arguments can be null, they are
pointer types rather than reference types, so the code to detect these
situations need
Fortran pass-by-reference optional arguments behave much like normal
Fortran arguments when lowered to GENERIC/GIMPLE, except they can be
null (representing a non-present argument).
Some parts of libgomp (those dealing with updating mappings) currently
do not expect to take a null address and
This patchset allows the use of Fortran optional arguments in OpenACC
programs in accordance with section 2.17 of the OpenACC 2.6 specification.
These patches were originally posted at
https://gcc.gnu.org/ml/gcc-patches/2019-01/msg01750.html for the OG8
branch. This version is targeted at trun
Current IPA-cp only generates cost-evaluating predicate for conditional
statement like
"if (param cmp const_val)", it is too simple and conservative. This patch
generalizes the
process to handle the form as T(param), a mathematical transformation on the
function
parameter, in which the parameter
On 11/07/19 20:45 +0100, Jonathan Wakely wrote:
This adds the new atomic types from C++2a, as proposed by P0019 and
P0020. To reduce duplication the calls to the compiler's atomic
built-ins are wrapped in new functions in the __atomic_impl namespace.
These functions are currently only used by std
Richard Biener writes:
> On Mon, Jul 8, 2019 at 4:41 PM Richard Sandiford
> wrote:
>>
>> Richard Biener writes:
>> > On Sun, Jul 7, 2019 at 9:07 PM Jeff Law wrote:
>> >>
>> >> On 7/7/19 3:45 AM, Richard Sandiford wrote:
>> >> > DCE tries to delete dead stores to local data and also tries to ins
On Thu, Jul 11, 2019 at 12:44 PM Alexandre Oliva wrote:
>
> ... and here's a patch that uses a try/finally/else gimplefe test to
> demonstrate the GIMPLE_EH_ELSE lowering problem (might_throw3 is tagged
> as [LP 1] rather than [LP 2]), and fixes it.
>
> Regstrapped on x86_64-linux-gnu. Ok to inst
On Thu, Jul 11, 2019 at 12:41 PM Alexandre Oliva wrote:
>
> On Jul 4, 2019, Richard Biener wrote:
>
> > Yeah. For other stuff we're simply looking at CPP_NAME and
> > string-matching, see c_parser_gimple_compound_statement
> > where you'd probably hook this into.
>
> Here's a working patch that
Richard Biener writes:
> On Fri, Jul 12, 2019 at 12:28 PM Richard Sandiford
> wrote:
>>
>> Richard Biener writes:
>> > On Fri, Jul 12, 2019 at 10:00 AM Richard Sandiford
>> > wrote:
>> >>
>> >> This change is needed to avoid a regression in gcc.dg/ifcvt-3.c
>> >> for a later patch. Without it,
Eric Botcazou writes:
>> ...trapping on signalling NaNs is enough for the field to be true.
>
> So what's the point in encoding this? The main distinction, and the only one
> relevant for the RTL middle-end, is whether the operator traps on quiet NaNs
> since this can change the comparison inst
On Fri, Jul 12, 2019 at 12:28 PM Richard Sandiford
wrote:
>
> Richard Biener writes:
> > On Fri, Jul 12, 2019 at 10:00 AM Richard Sandiford
> > wrote:
> >>
> >> This change is needed to avoid a regression in gcc.dg/ifcvt-3.c
> >> for a later patch. Without it, we enter CSE with a dead compariso
Richard Biener writes:
> On Fri, Jul 12, 2019 at 10:00 AM Richard Sandiford
> wrote:
>>
>> This change is needed to avoid a regression in gcc.dg/ifcvt-3.c
>> for a later patch. Without it, we enter CSE with a dead comparison left
>> by if-conversion and then eliminate the second (live) compariso
> ...trapping on signalling NaNs is enough for the field to be true.
So what's the point in encoding this? The main distinction, and the only one
relevant for the RTL middle-end, is whether the operator traps on quiet NaNs
since this can change the comparison instruction emitted by the back-end
On Thu, 11 Jul 2019, Andre Vieira (lists) wrote:
> Hi Richard(s),
>
> I am trying to tackle PR88915 and get GCC to further vectorize the "fallback"
> loop when doing loop versioning as it does when loop versioning is not
> required.
>
> I have a prototype patch that needs further testing, but at
This is another installment in the famous series "how can this have worked for
so long?". The delayed branches scheduling pass, aka dbr in reorg.c, uses its
own dataflow algorithm implemented in resource.c to do liveness analysis after
starting from seeds obtained from DF. It's a forward scan
On Thu, 11 Jul 2019, Jan Hubicka wrote:
> > > On Thu, 11 Jul 2019, Jan Hubicka wrote:
> > >
> > > > Hi,
> > > > this patch makes nonoverlapping_component_refs_since_match_p to accept
> > > > paths with non-trivial MEM_REFs and TMRs assuming that they have same
> > > > semantics.
> > >
> > > Hmm.
On Fri, Jul 12, 2019 at 9:49 AM Richard Sandiford
wrote:
>
> I'd made it a precondition of vector_builder::elt that the encoding
> must have been fully populated and that all implicit elements are
> therefore defined. But for one of the AArch64 patches I'm working on,
> it'd be convenient to be a
On Thu, 11 Jul 2019, Kewen.Lin wrote:
> Hi Richard,
>
> on 2019/7/10 下午7:50, Richard Biener wrote:
> > On Mon, 8 Jul 2019, Kewen.Lin wrote:
> >
> >
> > + tree rhs = gimple_assign_rhs1 (oe1def);
> > + tree vec = TREE_OPERAND (rhs, 0);
> > + tree vec_type = TREE_TYPE (vec);
> > +
>
Eric Botcazou writes:
>> +/* Invoke T (CODE, ORDER, SIGNEDNESS, CAN_TRAP) for each comparison, where:
>> +
>> + - CODE is the rtl comparison code
>> + - ORDER is the OR of the conditions under which CODE returns true
>> + - SIGNEDNESS is the signedness of COND, or 0 if it is sign-agnostic
>>
On Fri, Jul 12, 2019 at 10:00 AM Richard Sandiford
wrote:
>
> This change is needed to avoid a regression in gcc.dg/ifcvt-3.c
> for a later patch. Without it, we enter CSE with a dead comparison left
> by if-conversion and then eliminate the second (live) comparison in
> favour of the dead one.
On Thu, Jul 11, 2019 at 7:43 PM Martin Sebor wrote:
>
> Attached is a patch that adds a new parameter to limit the number
> of SSA_NAME assignments for GCC to follow in iterative or recursive
> algorithms. Purely as a proof of concept the patch introduces
> the parameter into -Warray-bounds where
On Thu, Jul 11, 2019 at 7:04 PM Eric Botcazou wrote:
>
> > After your patch does behavior change when trying to break on a line
> > with a return stmt inside a debugger?
>
> Note that every patch in the series was tested against GDB too, so hopefully
> this would have been caught... But the answe
> +/* Invoke T (CODE, ORDER, SIGNEDNESS, CAN_TRAP) for each comparison, where:
> +
> + - CODE is the rtl comparison code
> + - ORDER is the OR of the conditions under which CODE returns true
> + - SIGNEDNESS is the signedness of COND, or 0 if it is sign-agnostic
> + - CAN_TRAP is true if CO
On 11/07/19 20:45 +0100, Jonathan Wakely wrote:
This adds the new atomic types from C++2a, as proposed by P0019 and
P0020. To reduce duplication the calls to the compiler's atomic
built-ins are wrapped in new functions in the __atomic_impl namespace.
These functions are currently only used by std
This patch aims to fix PR69678 caused by PGO indirect call profiling
performance issues.
Currently the default profile can only use the single indirect target
that called more than 50%. This patch leverages multiple indirect
targets use in LTO-WPA and LTO-LTRANS stage, as a result, function
specia
IPA does not construct executability predicate for default case of switch
statement.
So execution cost of default case is not properly evaluated in IPA-cp, this
might
prevent function clone for function containing switch statement, if certain
non-default
case is proved to be executed after const
On Mon, Jul 08, 2019 at 03:56:51PM -0600, Martin Sebor wrote:
> A couple of GCC's Coding Conventions call to
>
> 1) Use the struct keyword for plain old data (POD) types.
> https://www.gnu.org/software/gcc/codingrationale.html#struct
>
> and
>
> 2) Use the class keyword for non-POD type
Ping. I think this is the only one of the series left.
(I should probably have treated the patch as obvious, but it always
seems wrong to do that retrospectively.)
Thanks,
Richard
Richard Sandiford writes:
> This patch is part of a series that fixes ambiguous attribute
> uses in .md files, i.e
This patch adds rtx folds for:
(and (cmp1 x y) (cmp2 x y)) -> (cmp3 x y)
(ior (cmp1 x y) (cmp2 x y)) -> (cmp3 x y)
(if_then_else (cmp1 x y) (cmp2 x y) (const_int 0))
-> (and (cmp1 x y) (cmp2 x y))
As always, there are plenty of other things that could be added.
This just happened to be
This patch rewrites some jump.c routines to operate on flags that
describe comparisons rather than handling each comparison code
separately. This in turn makes it easier to add some new routines
in the next patch.
Maybe the change isn't worth it for swap_condition, unsigned_condition
and signed_c
This change is needed to avoid a regression in gcc.dg/ifcvt-3.c
for a later patch. Without it, we enter CSE with a dead comparison left
by if-conversion and then eliminate the second (live) comparison in
favour of the dead one. That's functionally correct in itself, but it
meant that we'd combine
Hi!
This adds parsing support for a new clause, but currently just ignores it
besides checking some restrictions (will add further ones later).
For worksharing loop the spec says:
"If an order(concurrent) clause is present, then after assigning the iterations
of the associated
loops to their resp
This patch extends the support for "@..." pattern names so that
the patterns can have different numbers of operands. This allows
things like binary and ternary operations to be handled in a
consistent way, a bit like optabs. The generators assert that
the number of operands passed is correct for
I'd made it a precondition of vector_builder::elt that the encoding
must have been fully populated and that all implicit elements are
therefore defined. But for one of the AArch64 patches I'm working on,
it'd be convenient to be able to look back at previous elements while
building up the encoding
Richard Sandiford writes:
> This patch rewrites the way simplify_subreg handles constants.
> It uses similar native_encode/native_decode routines to the
> tree-level handling of VIEW_CONVERT_EXPR, meaning that we can
> move between rtx constants and the target memory image of them.
>
> The main po
85 matches
Mail list logo
