On 21/12/17 05:58, Jeff Law wrote:
On 08/22/2017 11:15 PM, Sebastian Huber wrote:
Hello Jeff,
On 03/08/17 07:11, Sebastian Huber wrote:
On 02/08/17 21:30, Jeff Law wrote:
On 07/24/2017 12:03 AM, Sebastian Huber wrote:
gcc/
PR libgcc/61152
* aarch64/rtems.h: Add GCC Runtime
This patch to the Go frontend by Cherry Zhang improves the escape
analysis diagnostics. This brings the diagnostics closer to those
generated by the gc compiler, which makes porting and debugging the
escape analysis code easier.
- In the gc compiler, the variable expression is represented with
Hi,
On Mon, Dec 18 2017, Richard Biener wrote:
> On Fri, Dec 15, 2017 at 3:19 PM, Martin Jambor wrote:
>>
>> Hello,
>>
>> the patch below prevents creation if fused-multiply-and-add instructions
>> in the widening_mul gimple pass on the Zen-based AMD CPUs and as a
>> result
On Thu, 2017-12-21 at 23:54 +0100, Matthias Klose wrote:
> On 21.12.2017 22:59, Steve Ellcey wrote:
> >
> > > As far as I understand it from Linaro connect and conversations with
> > > Debian/Ubuntu port maintainers, the correct triplet should be:
> > >
> > > aarch64-none-linux-gnu_ilp32
> > >
On 12/21/2017 02:50 PM, Martin Jambor wrote:
> Hi,
>
> the problem in PR 82027 is a general one with thunks that are created
> for IPA-CP clones. At WPA time, IPA-CP creates a clone that is called
> through a thunk. An "artificial thunk" is created in
> duplicate_thunk_for_node() in
On 12/21/2017 10:50 AM, Jakub Jelinek wrote:
> Hi!
>
> The problem here is that the code expects fold_build1 will actually not
> fold, because using gimple_build_assign (..., VIEW_CONVERT_EXPR, temp);
> is valid only if TREE_CODE (temp) == VIEW_CONVERT_EXPR.
> So, either we can replace the
On 12/21/2017 10:54 AM, Jakub Jelinek wrote:
> Hi!
>
> Converting widening mult, or widen_{plus,minus}_expr, or integral mult to
> fma is IMNSHO undesirable with -ftrapv for signed types - the expansion of
> those doesn't detect overflows and by giving up we don't need to worry about
> formerly
On 12/21/2017 01:38 PM, Jakub Jelinek wrote:
> Hi!
>
> When we don't pass an argument at all because it is TYPE_EMPTY_P, we
> shouldn't tweak argument slot alignment based on the alignment of these
> arguments either.
>
> This patch fixes the ICE we issued on pr83487.{c,C}.
> As the compat tests
On 12/21/2017 01:29 PM, Jakub Jelinek wrote:
> Hi!
>
> If copying a substring without %s from the format string into the
> destination does or might overflow, we try to point the caret at the
> character in the format string that will cause the overflow.
> In the first spot in maybe_warn this is
On 21.12.2017 22:59, Steve Ellcey wrote:
> On Thu, 2017-12-21 at 20:55 +, James Greenhalgh wrote:
>> On Thu, Dec 21, 2017 at 06:56:22PM +, Steve Ellcey wrote:
>>>
>>> This one line patch for multi-arch support on Aarch64 and ILP32 was
>>> submitted over a year ago and pinged a number of
OK.
On Thu, Dec 21, 2017 at 12:50 PM, David Malcolm wrote:
> On Wed, 2017-12-20 at 23:56 -0500, Jason Merrill wrote:
>> On Wed, Dec 20, 2017 at 5:14 PM, David Malcolm
>> wrote:
>> > On Mon, 2017-12-11 at 18:39 -0500, Jason Merrill wrote:
>> > > On
On Thu, Dec 21, 2017 at 08:32:28PM -0200, Alexandre Oliva wrote:
> for gcc/c-family/ChangeLog
>
> PR debug/83527
> PR debug/83419
> * c-semantics.c (only_debug_stmts_after_p): New.
> (pop_stmt_list): Clear side effects in debug-only stmt list.
> Check for single
On Tue, Dec 19, 2017 at 4:56 PM, Alexandre Oliva wrote:
> On Dec 14, 2017, Jason Merrill wrote:
>
>> On 12/07/2017 04:04 PM, Alexandre Oliva wrote:
>
>>> For other templates, I ended up writing code to look for
>>> specializations in the hashtables of decl or
On Dec 21, 2017, Jakub Jelinek wrote:
> On Wed, Dec 20, 2017 at 10:02:24PM -0200, Alexandre Oliva wrote:
>> + if (tsi_one_before_end_p (i))
>> +TREE_SIDE_EFFECTS (t) = TREE_SIDE_EFFECTS (tsi_stmt (i));
> So DEBUG_BEGIN_STMTs after the single non-DEBUG_BEGIN_STMT
On 12/21/2017 01:29 PM, Jakub Jelinek wrote:
Hi!
If copying a substring without %s from the format string into the
destination does or might overflow, we try to point the caret at the
character in the format string that will cause the overflow.
In the first spot in maybe_warn this is only done
This is a regression recently introduced on the mainline: the compiler doesn't
use the specified size to access an atomic variable with an address clause.
Tested on x86_64-suse-linux, applied on the mainline.
2017-12-21 Eric Botcazou
* gcc-interface/decl.c
On Thu, 2017-12-21 at 20:55 +, James Greenhalgh wrote:
> On Thu, Dec 21, 2017 at 06:56:22PM +, Steve Ellcey wrote:
> >
> > This one line patch for multi-arch support on Aarch64 and ILP32 was
> > submitted over a year ago and pinged a number of time since then,
> > since no one has
On Thu, Dec 21, 2017 at 01:46:56PM -0800, Cesar Philippidis wrote:
> After thinking about this some more, I decided that it would be better
> expand the offloaded function arguments into individual parameters
> during omp lowering, rather than writing a separate pass later on. I
> don't see too
Hi,
the problem in PR 82027 is a general one with thunks that are created
for IPA-CP clones. At WPA time, IPA-CP creates a clone that is called
through a thunk. An "artificial thunk" is created in
duplicate_thunk_for_node() in cgraphclones.c. This function also
removes now unnecessary
On 12/18/2017 02:58 PM, Cesar Philippidis wrote:
> Jakub,
>
> I'd like your thoughts on the following problem.
>
> One of the offloading bottlenecks with GPU acceleration in OpenACC is
> the nontrivial offloaded function invocation overhead. At present, GCC
> generates code to pass a struct
On 21 December 2017 at 15:24, Tamar Christina wrote:
> The 12/14/2017 20:46, Christophe Lyon wrote:
>> On 14 December 2017 at 11:56, Tamar Christina
>> wrote:
>> > The 12/13/2017 08:49, Christophe Lyon wrote:
>> >> On 12 December 2017 at 18:29,
Hi,
I updated my patch based on all your comments.
the major changes are the following:
1. replace the candidate calls with __builtin_str(n)cmp_eq instead of
__builtin_memcmp_eq;
in builtins.c, when expanding the new __builtin_str(n)cmp_eq call,
expand them first as
On Thu, Dec 21, 2017 at 06:56:22PM +, Steve Ellcey wrote:
> This one line patch for multi-arch support on Aarch64 and ILP32 was
> submitted over a year ago and pinged a number of time since then,
> since no one has objected and since it is only one line I am going
> to check it in as an
Hi!
When we don't pass an argument at all because it is TYPE_EMPTY_P, we
shouldn't tweak argument slot alignment based on the alignment of these
arguments either.
This patch fixes the ICE we issued on pr83487.{c,C}.
As the compat tests show, we are now ABI compatible with clang++ trunk
for
On Thu, Dec 21, 2017 at 06:29:11PM -0200, Alexandre Oliva wrote:
> > PR debug/83419
> > * c-family/c-semantics.c (pop_stmt_list): Propagate side
> > effects from single nondebug stmt to container list.
>
> Oops, this belonged in gcc/c-family/ChangeLog. I'm checking this in to
> fix
Hi!
If copying a substring without %s from the format string into the
destination does or might overflow, we try to point the caret at the
character in the format string that will cause the overflow.
In the first spot in maybe_warn this is only done if avail_range.min ==
avail_range.max and thus
On Dec 20, 2017, Alexandre Oliva wrote:
> for gcc/ChangeLog
> PR debug/83419
> * c-family/c-semantics.c (pop_stmt_list): Propagate side
> effects from single nondebug stmt to container list.
Oops, this belonged in gcc/c-family/ChangeLog. I'm checking this
Hi,
On 21/12/2017 17:04, Jason Merrill wrote:
On Wed, Dec 20, 2017 at 10:37 AM, Paolo Carlini
wrote:
in this error recovery regression, after a sensible error produced by
unqualified_name_lookup_error we ICE much later when gimplify_modify_expr
encounters a
Hello!
Attached patch fixes non-BMI2 shift define-and-split instructions that
remove unnecessary masking of count operand by adding a register
constraints that allows only CX hard register.
2017-12-21 Uros Bizjak
PR target/83467
* config/i386/i386.md (*ashl3_mask):
On Thu, Dec 21, 2017 at 06:16:16PM +, Joseph Myers wrote:
> On Fri, 17 Nov 2017, Michael Meissner wrote:
> The architecture-independent changes are OK. However, I have a comment on
> the target parts:
>
> > +(define_insn "round2"
> > + [(set (match_operand:IEEE128 0
On December 21, 2017 6:54:51 PM GMT+01:00, Jakub Jelinek
wrote:
>Hi!
>
>Converting widening mult, or widen_{plus,minus}_expr, or integral mult
>to
>fma is IMNSHO undesirable with -ftrapv for signed types - the expansion
>of
>those doesn't detect overflows and by giving up we
On December 21, 2017 6:50:45 PM GMT+01:00, Jakub Jelinek
wrote:
>Hi!
>
>The problem here is that the code expects fold_build1 will actually not
>fold, because using gimple_build_assign (..., VIEW_CONVERT_EXPR, temp);
>is valid only if TREE_CODE (temp) == VIEW_CONVERT_EXPR.
>So,
This one line patch for multi-arch support on Aarch64 and ILP32 was
submitted over a year ago and pinged a number of time since then,
since no one has objected and since it is only one line I am going
to check it in as an obvious fix.
Steve Ellcey
sell...@cavium.com
2017-12-21 Andrew Pinski
On December 21, 2017 7:05:10 PM GMT+01:00, Jakub Jelinek
wrote:
>Hi!
>
>The following two testcases show multiple issues in
>reorder_basic_blocks_simple. Both issues only occur if the greedy
>algorithm
>decides to put some basic block before the ENTRY successor. In that
>case
On Wed, Nov 29, 2017 at 4:30 PM, Jason Merrill wrote:
> On Fri, Nov 17, 2017 at 11:04 PM, Nathan Froyd wrote:
>> On Fri, Nov 17, 2017 at 8:50 AM, Jason Merrill wrote:
>>> On Thu, Nov 16, 2017 at 11:21 AM, Nathan Froyd
On Thu, Dec 21, 2017 at 12:33 PM, Nathan Sidwell wrote:
> Jason, this needs a sanity check
>
> 83406 was exposed by Jason's recent patch removing return_type & closure
> from the tree_lambda_expr node:
> Remove unnecessary LAMBDA_EXPR fields.
>
> * cp-tree.h
On Dec 21, 2017, Jeff Law wrote:
> FWIW I wouldn't be surprised if there's other places in reorg that are
> going to need similar fixes.
*nod*. I looked around, and the only place that seemed suspicious was
find_end_label, but I'm not sure we might actually use a debug insn as
On Fri, 17 Nov 2017, Michael Meissner wrote:
> Here is the fixed patch. It fixes the btrunc2 insn to use the correct
> XSRPQI variant for truncf128. I added the float128-hw11.c test as a runtime
> test to make sure round, trunc, ceil, and floor return the correct values.
> The
> machine
On Dec 21, 2017, Jakub Jelinek wrote:
> On Wed, Dec 20, 2017 at 10:02:24PM -0200, Alexandre Oliva wrote:
>> --- a/gcc/c-family/c-semantics.c
>> +++ b/gcc/c-family/c-semantics.c
>> @@ -96,6 +96,15 @@ pop_stmt_list (tree t)
>> + while (!tsi_end_p (i)
>> + &&
Hi!
The following two testcases show multiple issues in
reorder_basic_blocks_simple. Both issues only occur if the greedy algorithm
decides to put some basic block before the ENTRY successor. In that case
reorder_basic_blocks_simple has code to split the ENTRY successor edge
and put the new bb
Hi!
Converting widening mult, or widen_{plus,minus}_expr, or integral mult to
fma is IMNSHO undesirable with -ftrapv for signed types - the expansion of
those doesn't detect overflows and by giving up we don't need to worry about
formerly trapping MULT_EXPR becoming something that isn't trapping
Hi!
The problem here is that the code expects fold_build1 will actually not
fold, because using gimple_build_assign (..., VIEW_CONVERT_EXPR, temp);
is valid only if TREE_CODE (temp) == VIEW_CONVERT_EXPR.
So, either we can replace the fold_build1 with build1, or we should just
use the
On Wed, 2017-12-20 at 23:56 -0500, Jason Merrill wrote:
> On Wed, Dec 20, 2017 at 5:14 PM, David Malcolm
> wrote:
> > On Mon, 2017-12-11 at 18:39 -0500, Jason Merrill wrote:
> > > On 11/10/2017 04:45 PM, David Malcolm wrote:
> > > > Without this, then lvalue_p returns false
Jason, this needs a sanity check
83406 was exposed by Jason's recent patch removing return_type & closure
from the tree_lambda_expr node:
Remove unnecessary LAMBDA_EXPR fields.
* cp-tree.h (LAMBDA_EXPR_CLOSURE): Use TREE_TYPE.
(LAMBDA_EXPR_RETURN_TYPE): Remove.
(struct
On 12/20/2017 03:15 PM, Tom de Vries wrote:
> On 12/20/2017 11:59 PM, Cesar Philippidis wrote:
>> On 12/19/2017 04:39 PM, Tom de Vries wrote:
>>> On 12/20/2017 12:25 AM, Cesar Philippidis wrote:
In CUDA 9, Nvidia removed support for treating the labels of functions
as generic address
On Sun, Dec 17, 2017 at 03:14:08AM +, Segher Boessenkool wrote:
> Hi!
>
> On Mon, Dec 11, 2017 at 02:18:53PM +, James Greenhalgh wrote:
> >
> > In simplify_set we try transforming the paradoxical subreg expression:
> >
> > (set FOO (subreg:M (mem:N BAR) 0))
> >
> > in to:
> >
> > (set
On 12/21/2017 02:13 AM, Martin Liška wrote:
On 12/20/2017 06:45 PM, Jakub Jelinek wrote:
Another thing is that the "/" in there is wrong, so
const char dir_separator_str[] = { DIR_SEPARATOR, '\0' };
char *b = concat (profile_data_prefix, dir_separator_str, pwd, NULL);
needs to be used
On Wed, Dec 20, 2017 at 10:37 AM, Paolo Carlini
wrote:
> in this error recovery regression, after a sensible error produced by
> unqualified_name_lookup_error we ICE much later when gimplify_modify_expr
> encounters a corresponding error_mark_node as second argument of a
OK.
On Thu, Dec 21, 2017 at 10:51 AM, Eric Botcazou wrote:
> Hi,
>
> this is an ICE in the C++ FE on an offsetof construct with __PTRDIFF_MAX__ and
> it happens only with this specific index:
>
> In file included from pr82872.c:4:
> pr82872.c: In function 'size_t foo()':
>
Hi,
this is an ICE in the C++ FE on an offsetof construct with __PTRDIFF_MAX__ and
it happens only with this specific index:
In file included from pr82872.c:4:
pr82872.c: In function 'size_t foo()':
pr82872.c:10:10: internal compiler error: in ignore_overflows, at cp/cvt.c:583
return
Ping?
On 12/15/17 19:59, Bernd Edlinger wrote:
> On 12/15/17 11:51, Jakub Jelinek wrote:
>> On Fri, Dec 15, 2017 at 10:44:54AM +, Bernd Edlinger wrote:
>>> when working on the -Wcast-function-type patch I noticed some rather
>>> ugly and non-portable function type casts that are necessary to
Hi all,
Since support for -mcpu=cortex-a55 and -mcpu=cortex-a75
was added we added support for the +dotprod extension
which these CPUs support.
We already specify as such in the arm-cpus.in entries for
these processors. However the table in driver-arm.c was
not adding +dotproct to the -march
Hi all,
While doing some experimentation with options
in arm-cpus.in I hit an assembler error where
the alphabetical sorting in arm_rewrite_selected_arch
would not sort the string "+simd+fp16" properly
into "+fp16+simd" like gas expects.
The way the comparator function compare_opt_names
is
The 12/14/2017 20:46, Christophe Lyon wrote:
> On 14 December 2017 at 11:56, Tamar Christina wrote:
> > The 12/13/2017 08:49, Christophe Lyon wrote:
> >> On 12 December 2017 at 18:29, Tamar Christina
> >> wrote:
> >> > Hi All,
> >> >
> >> > The
On Wed, Dec 20, 2017 at 10:02:24PM -0200, Alexandre Oliva wrote:
> --- a/gcc/c-family/c-semantics.c
> +++ b/gcc/c-family/c-semantics.c
> @@ -96,6 +96,15 @@ pop_stmt_list (tree t)
> t = l;
> tsi_link_before (, u, TSI_SAME_STMT);
> }
> + while (!tsi_end_p (i)
On 12/20/2017 06:45 PM, Jakub Jelinek wrote:
> Another thing is that the "/" in there is wrong, so
> const char dir_separator_str[] = { DIR_SEPARATOR, '\0' };
> char *b = concat (profile_data_prefix, dir_separator_str, pwd, NULL);
> needs to be used instead.
This looks much nicer, I forgot
On Wed, Dec 20, 2017 at 10:28:53PM -0700, Jeff Law wrote:
> On 12/20/2017 03:18 PM, Jakub Jelinek wrote:
> > In rtl.texi we say:
> > @findex const_vector
> > @item (const_vector:@var{m} [@var{x0} @var{x1} @dots{}])
> > Represents a vector constant. The square brackets stand for the vector
> >
On December 21, 2017 5:40:48 AM GMT+01:00, Jeff Law wrote:
>On 12/15/2017 09:30 AM, Richard Biener wrote:
>> On December 15, 2017 5:27:14 PM GMT+01:00, Jeff Law
>wrote:
>>> On 12/15/2017 01:10 AM, Richard Biener wrote:
On Thu, 14 Dec 2017, Richard Biener
On Wed, Dec 20, 2017 at 11:37 PM, Thomas Koenig wrote:
> Hi Janne,
>
> I think you need a few more test cases, like (compile with -fcheck=all
> and -std=f2018):
>
> module foo
> contains
> subroutine f(n)
> call g(n-1)
> end subroutine f
> subroutine g(n)
> if
On 12/05/2017 10:27 AM, Jakub Jelinek wrote:
> The most important change I've done in the testsuite was pointer-subtract-2.c
> used -fsanitize=address,pointer-subtract, but the function was actually
> doing pointer comparison. Guess that needs to be propagated upstream at
> some point. Another
60 matches
Mail list logo