On Tue, Jan 9, 2018 at 10:41 AM, Jakub Jelinek wrote:
> On Tue, Jan 09, 2018 at 10:30:31AM -0500, Jason Merrill wrote:
>> On Mon, Jan 8, 2018 at 6:55 PM, Jakub Jelinek wrote:
>> > The assert there assumes we never evaluate a statement list to
>> > DEBUG_BEGIN_STMT, but it breaks appart when a BIN
Committed as obvious after regression testing on x86_64-*-freebsd.
2018-01-09 Steven G. Kargl
PR fortran/83742
* expr.c (gfc_is_simply_contiguous): Check for NULL pointer.
2018-01-09 Steven G. Kargl
PR fortran/83742
* gfortran.dg/contiguous_6.f90: New test.
Hi,
I'd like to gently point to two pending patches of mine:
The updated fix for PR 81055 ("[6/7/8 Regression] ICE with invalid
initializer for array new")
https://gcc.gnu.org/ml/gcc-patches/2017-12/msg01428.html
and also
PR 78344 ("ICE on invalid c++ code (tree check: expected
Hi Ian,
> This patch updates libgo to the Go1.10beta1 release. The final Go
> 1.10 release is expected around February 1, so it's not clear how the
> release timing is going to work with GCC 8. In any case this updates
> GCC to something pretty close to the final Go 1.10 release.
>
> A few chang
This patch by Cherry Zhang adds a new -fgo-debug-escape-hash option
to the Go frontend. This option takes a string that can be used to
select specific functions on which to run escape analysis. This makes
it easier to find escape analysis bugs by doing a binary search over
all the functions, enab
Whilst developing the location-wrapper patch kit, I accumulated various
source files for which my work-in-progress patched-cc1plus had different
behavior to an unpatched cc1plus: some of these were crashes, others
were erroneous diagnostics.
All of these are now fixed, but it seems appropriate to
To answer some of my own questions about how this works (or doesn't work):
There are two different types in libgcc called TFtype, one in
quad-float128.h and one in libgcc2.h. The one in quad-float128.h is
mapped to KFmode in the case where TFmode is IBM long double, so functions
such as __mulk
We should also adjust stack_realign_offset for the largest alignment of
stack slot actually used when stack realignment isn't needed. This is
required to keep stack frame properly aligned to satisfy the largest
alignment of stack slots.
Tested on Linux/i686 and Linux/x86-64.
OK for trunk?
Thank
In the Go frontend, if a local variable's address is taken and passed
out of its lexical scope, the GCC middle-end may reuse the stack slot
for the variable, not knowing it is still live through a pointer.
This patch by Cherry Zhang creates a top-level temporary variable and
let the user-defined va
I found a few problems in the previous revision:
1) It didn't handle the simple case of member arrays of const
struct objects (only member arrays of const arrays of structs
were handled).
2) The array_ctor_elt() function returned a narrow empty string
for an uninitialized CONSTRUCTOR el
On Tue, Jan 9, 2018 at 12:26 PM, Rainer Orth
wrote:
> Hi Ian,
>
>> This patch updates libgo to the Go1.10beta1 release. The final Go
>> 1.10 release is expected around February 1, so it's not clear how the
>> release timing is going to work with GCC 8. In any case this updates
>> GCC to somethin
This fixes unique_ptr so it doesn't print as unique_ptr, and
fixes a problem in the type-printers that can result in a gdb.error
exception (see https://sourceware.org/PR22686).
PR libstdc++/80276
* python/libstdcxx/v6/printers.py (SharedPointerPrinter)
(UniquePointerPrinte
After cunrolling the inner loop, the remaining loop in the testcase
has a single 32-bit access and a group of 64-bit accesses. We first
try to vectorise at 128 bits (VF 4), but decide not to for cost reasons.
We then try with 64 bits (VF 2) instead. This means that the group
of 64-bit accesses us
Hi,
I’d like to ping the following patch the 2nd time:
https://gcc.gnu.org/ml/gcc-patches/2017-12/msg01454.html
I already addressed all the issues raised for the first version of the patch,
and retested on both X86 and Aarch64, without any issue.
I’d really like to make it into GCC 8.
thank
The following patch changes the x86 intrinsic compat headers to use #error
instead of #warning. We do this for two reasons. Firstly, we want the user
to really be sure they want/need to use the x86 intrinsic compat support
before doing so, because a warning is too easy to ignore. Secondly, we do
On 09/01/18 18:50 +, Jonathan Wakely wrote:
On 09/01/18 15:02 +, Jonathan Wakely wrote:
On 09/01/18 14:59 +, Jonathan Wakely wrote:
On 04/01/18 11:22 +0100, Juraj Oršulić wrote:
Hi Jonathan (and the libstdc++ list). Can we revive this? I sent the
patches for improving the smart poi
On 01/09/18 18:50, Richard Earnshaw (lists) wrote:
> On 09/01/18 17:36, Bernd Edlinger wrote:
>> Richard Earnshaw wrote:
>> > Let me give an example, we use the generic code expansion if we
>> > encounter the builtin when generating code for Thumb on pre-ARMv7
>> > devices. We don't have ins
I saw one regression with this patch:
PASS-FAIL: gcc.dg/torture/vshuf-v8qi.c -O2 execution test
Could you confirm this? Note: I'm testing with your other (pre-2018 at
least) patches applied at the same time (smax etc, anddi, bswaphi) so it
might be an interaction, but the code looks like a m
On 01/05/2018 02:01 AM, Richard Biener wrote:
> On Tue, 28 Nov 2017, Richard Biener wrote:
>
>>
>> The following adds a new target hook, targetm.vectorize.split_reduction,
>> which allows the target to specify a preferred mode to perform the
>> final reducion on using either vector shifts or scala
PR target/78585 has been fixed for GCC 7 by
commit 7ed04d053eead43d87dff40fb4e2904219afc4d5
Author: jakub
Date: Wed Nov 30 13:02:07 2016 +
* config/i386/i386.c (dimode_scalar_chain::convert_op): Avoid
sharing the SUBREG rtx between move and following insn.
git-svn-id: svn+
This patch to the Go frontend by Cherry Zhang delays escaping sliced
arrays. Arrays that are sliced are set to escape in type checking,
very early in compilation. The escape analysis runs later but cannot
undo it. This patch changes it to not escape in the early stage.
Later the escape analysis
This patch to the Go frontend by Cherry Zhang stack allocates a
non-escaping make of a slice type with a constant length and capacity.
Bootstrapped on x86_64-pc-linux-gnu. Committed to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
===
This patch to the Go frontend by Cherry Zhang adds support for the
go:noescape compilation pragma across packages.
https://golang.org/cl/83876 added support for the go:noescape pragma,
but only for functions called from the same package. The pragma did
not take effect for exported functions that a
This patch to the Go frontend by Cherry Zhang moves some checks of
escape state earlier, from get_backend to Mark_address_taken. This
permits us to reclaim escape analysis Nodes before kicking off the
backend (although that is not done in this patch). Also it makes it
easier to check variables a
This patch to the Go frontend by Cherry Zhang makes non-escaping
Bound_method_expression not heap allocate. It stack allocates the
closure when it does not escape. Bootstrapped on x86_64-pc-linux-gnu.
Committed to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
Hi Steve,
Here are the results of compiling with the OpenCoarrays “caf” compiler wrapper,
which uses -fcoarray=lib -lcaf_mpi:
1. random_init(repeatable=.true., image_distinct=.true.) gives repeatable
sequences that are distinct on each image.
2. random_init(.true., .false.) gives repeatable seq
This patch to the Go frontend by Cherry Zhang stack allocates defer
thunk when possible, which is to say when the defer statement is not
in a loop. Bootstrapped on x86_64-pc-linux-gnu. Committed to
mainline.
Ian
Index: gcc/go/gofrontend/MERGE
=
On 01/05/2018 12:25 PM, Sudakshina Das wrote:
> Hi Jeff
>
> On 05/01/18 18:44, Jeff Law wrote:
>> On 01/04/2018 08:35 AM, Sudakshina Das wrote:
>>> Hi
>>>
>>> The bug reported a particular test di-longlong64-sync-1.c failing when
>>> run on arm-linux-gnueabi with options -mthumb -march=armv5t -O[g
This patch to the Go frontend by Cherry Zhang sets captured variable
addresses to nonescape until the flood phase. The escape analysis
pass models closures by flowing captured variable address to the
closure node. However, the escape state for the address expressions
remained unset as ESCAPE_UNKN
On 01/07/2018 03:26 AM, Richard Sandiford wrote:
> r254296 added support for (const ...) wrappers around vectors,
> but in the end the agreement was to use a variable-length
> encoding of CONST_VECTOR (and VECTOR_CST) instead. This patch
> therefore reverts the bits that are no longer needed.
>
>
This patch to the Go frontend by Cherry Zhang uses a temporary
variable for stack allocation when possible. Currently, an allocation
expression that can be allocated on stack is implemented with
__builtin_alloca, which turns into __morestack_allocate_stack_space,
which may call the C malloc functi
On 01/08/2018 08:05 PM, Martin Sebor wrote:
> On 01/08/2018 04:28 AM, Richard Biener wrote:
>> On Sat, Jan 6, 2018 at 11:04 PM, Martin Sebor wrote:
>>> Bug 83671 - Fix for false positive reported by -Wstringop-overflow
>>> does not work at -O1, points out that the string length range
>>> optimizat
When parsing code and implicit typing is used, the
type of an entity used as a character length is not
known until after resolution. The attach patch
checks the type of length and response accordingly.
Regression tested on x86_64-*-freebsd. Ok to commit?
2018-01-09 Steven G. Kargl
P
This patch by Tony Reix adds some aix build tags to new libgo files.
Bootstrapped on x86_64-pc-linux-gnu. Committed to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
===
--- gcc/go/gofrontend/MERGE (revision 256412)
+++ gcc/go/gofr
On Tue, Jan 9, 2018 at 6:54 AM, Segher Boessenkool
wrote:
> On Tue, Jan 09, 2018 at 12:23:42PM +, Wilco Dijkstra wrote:
>> Segher Boessenkool wrote:
>> > On Mon, Jan 08, 2018 at 0:25:47PM +, Wilco Dijkstra wrote:
>> >> > Always pairing two registers together *also* degrades code quality.
>
This patch by Tony Reix adds AIX support to cmd/go in libgo. For
gccgo code avoid --whole-archive and -(. Use -blibpath instead of
-rpath. Bootstrapped and ran Go testsuite on x86_64-pc-linux-gnu.
Committed to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
This libgo patch by Cherry Zhang adds some noescape cases to the
runtime package. This is in preparation of turning on escape analysis
for the
runtime.
- In gccgo, systemstack is implemented with mcall, which is not
go:noescape. Wrap the closure in noescape so the escape analysis
does not th
This libgo patch by Cherry Zhang works around some escaping code in
runtime/export_test.go. When compiling the runtime, local variables
and closures may not be heap allocated. In one test, there is a go
statement with a closure. In the gc compiler, it distinguishes
capturing variable by value vs
On Wed, Jan 10, 2018 at 3:18 AM, Steve Kargl
wrote:
> When parsing code and implicit typing is used, the
> type of an entity used as a character length is not
> known until after resolution. The attach patch
> checks the type of length and response accordingly.
> Regression tested on x86_64-*-fre
On Tue, Jan 9, 2018 at 10:29 PM, H.J. Lu wrote:
> We should also adjust stack_realign_offset for the largest alignment of
> stack slot actually used when stack realignment isn't needed. This is
> required to keep stack frame properly aligned to satisfy the largest
> alignment of stack slots.
>
>
> This patch updates libgo to the Go1.10beta1 release. The final Go
> 1.10 release is expected around February 1, so it's not clear how the
> release timing is going to work with GCC 8. In any case this updates
> GCC to something pretty close to the final Go 1.10 release.
This patch introduces a
101 - 141 of 141 matches
Mail list logo
