On Oct 11, 2018, JonY <10wa...@gmail.com> wrote:
> On 10/11/2018 02:57 AM, NightStrike wrote:
>>
>> Except that options typically don't get removed, just deprecated. It
>> seems cleaner to me to drop mingw from the name and make it default to
>> enabled for cygwin.
> It is already enabled for
On Fri, 12 Oct 2018 at 05:37, Jeff Law wrote:
>
> On 10/8/18 11:12 AM, will wray wrote:
> > Hi,
> >
> > A PR to fix Bug 87364 - Pretty print of enumerator never prints the id,
> > always falls back to C-style cast output
> >
> > The fix is tested with the code attached to the bug report and
On 10/10/18, Martin Sebor wrote:
> While writing tests for fixes and enhancements for attribute
> handling I keep finding myself coming up with the same boiler-
> plate code to verify whether an attribute has or has not been
> successfully applied. It's often error-prone because it
> depends on
On 10/11/18, Martin Sebor wrote:
> The manual says that:
>
>The packed attribute specifies that a variable or structure
>field should have the smallest possible alignment--one byte
>for a variable, and one bit for a field...
>
> The variable part doesn't actually reflect reality: GCC
On 8/7/18 2:00 PM, Giuliano Augusto Faulin Belinassi wrote:
> Related with bug 86829, but for hyperbolic trigonometric functions.
> This patch adds substitution rules to both sinh(tanh(x)) -> x / sqrt(1
> - x*x) and cosh(tanh(x)) -> 1 / sqrt(1 - x*x). Notice that the both
> formulas has division
On 10/11/18 1:23 PM, Peter Bergner wrote:
> On 10/11/18 1:18 PM, Peter Bergner wrote:
>> Ok, after working in gdb, I see that the PA-RISC port still uses reload
>> and not LRA, but it too seems to have the same issue of reusing input
>> regs that have REG_DEAD notes, so the question still stands.
On 10/8/18 11:12 AM, will wray wrote:
> Hi,
>
> A PR to fix Bug 87364 - Pretty print of enumerator never prints the id,
> always falls back to C-style cast output
>
> The fix is tested with the code attached to the bug report and additional
> usage over the past few weeks.
>
> Bootstrap and
On 9/27/18 8:13 AM, Tamar Christina wrote:
>>> The error you would get if you do this is very confusing so I thought
>>> since it didn't matter much for the regexp only target triple tests
>>> that just accepting this would be fine.
>> Seems a good thing that that's a noisy failure; the function
On 10/11/18 9:03 PM, Martin Sebor wrote:
> The manual says that:
>
> The packed attribute specifies that a variable or structure
> field should have the smallest possible alignment--one byte
> for a variable, and one bit for a field...
>
> The variable part doesn't actually reflect
The manual says that:
The packed attribute specifies that a variable or structure
field should have the smallest possible alignment--one byte
for a variable, and one bit for a field...
The variable part doesn't actually reflect reality: GCC and all
other compilers I tested ignore the
On 10/11/18 2:58 PM, Martin Sebor wrote:
> While writing tests for attribute flatten I wasn't 100% sure
> from reading the manual if it would respect attribute noinline
> or if it would override it. Turns out it's the former so
> the attached patch mentions it in the manual. Unless there
> are
On 10/11/18 3:09 PM, Paul Koning wrote:
> Updated with an additional item I just debugged.
>
> Since the code that uses the doloop_end pattern does not check the operand
> mode as given in the pattern, the pattern itself may need to do this, and
> that was not documented. In addition, if the
On 10/11/18 5:48 AM, Martin Liška wrote:
> Hi.
>
> This is simplification in ICF where we can use attribute comparison
> function from attrs.c instead of implementing our own.
>
> Patch survives tests on x86_64-linux-gnu.
> Ready for trunk?
> Thanks,
> Martin
>
> gcc/ChangeLog:
>
> 2018-10-10
On 10/11/18 4:12 AM, Paul Hua wrote:
> Hi:
>
> The bug pr87156 make MIPS target bootstrap fail a month ago. The
> attached patch that posted under bugzilla by Jan Hubicka fixed the
> bug.
> Bootstrapped and reg-tested on mips64el-linux-gnu and x86_64-pc-linux-gnu.
> Considering that Jan not very
On 10/11/2018 04:38 PM, Joseph Myers wrote:
On Thu, 11 Oct 2018, Martin Sebor wrote:
Attached is a documentation-only patch to clarify the use case
of multiple distinct format_arg attributes on a single function.
It's far from obvious that the use case makes sense so explicitly
mentioning it
Running make check-c++ RUNTESTFLAGS=dg.exp=decomp31.C will yield
# of unsupported tests 3
because the test (as the only one in cpp1z/) uses
"dg-do compile { target c++17 }" which doesn't work (yet?). This patch
makes it use explicit dg-options as in other tests, so now we get
# of
Recently it came up that no test in the testsuite triggers the RANGE_FOR_STMT
case in potential_constant_expression_1. I came up with this valid test that
tickles that codepath.
I can't use ({ }) instead of a lambda because the constexpr machinery doesn't
handle statement expressions; see
Here potential_constant_expression_1 rejects the testcase because the body of
the for loop calls a non-constexpr function. But the range is empty so the
function would never get called.
The trick with evaluating the for-condition doesn't work here, because we're
dealing with a converted
On Thu, Oct 11, 2018 at 11:35:23AM -0400, Jason Merrill wrote:
> > + /* [dcl.fct.spec]
> > + "the constant-expression, if supplied, shall be a contextually
> > + converted constant expression of type bool." */
> > + expr = build_explicit_specifier (expr,
It's not safe to assume that malloc(n) returns memory aligned to more
than n, so when relying on the guaranteed alignment of malloc ensure
that the number of bytes allocated is at least as large as the
alignment.
PR libstdc++/77691
* include/experimental/memory_resource
On Thu, 2018-10-11 at 10:31 -0400, Jason Merrill wrote:
> On Thu, Oct 11, 2018 at 10:28 AM Jason Merrill
> wrote:
> >
> > On Wed, Oct 10, 2018 at 5:01 PM David Malcolm
> > wrote:
> > > On Tue, 2018-10-09 at 18:38 -0400, Jason Merrill wrote:
> > > > On Tue, Oct 9, 2018 at 1:19 PM David Malcolm
Next patch, next try …
The problem is the following: We have a use-associated symbol (a
function) – which is once referenced (by use association).
We now declare a same-name function in the 'contains' section.
gfortran now parses the function name, fetches the symbol and reports
that one
On Thu, 11 Oct 2018, Martin Sebor wrote:
> Attached is a documentation-only patch to clarify the use case
> of multiple distinct format_arg attributes on a single function.
> It's far from obvious that the use case makes sense so explicitly
> mentioning it should help avoid the mistake of
On 11/10/18 17:37 +0100, Jonathan Wakely wrote:
Avoid a system call when no sleep is required. Sleep in a loop (actually
two loops) to handle interruption by signals.
PR libstdc++/80538
* src/c++11/thread.cc (this_thread::__sleep_for)
[_GLIBCXX_HAVE_SLEEP]: Only call
Attached is a documentation-only patch to clarify the use case
of multiple distinct format_arg attributes on a single function.
It's far from obvious that the use case makes sense so explicitly
mentioning it should help avoid the mistake of assuming that
accepting it is due to the lack of better
On 10/11/18 3:05 PM, Peter Bergner wrote:
> On 10/11/18 2:40 PM, Jeff Law wrote:
>> On 10/11/18 1:23 PM, Peter Bergner wrote:
>>> On 10/11/18 1:18 PM, Peter Bergner wrote:
Ok, after working in gdb, I see that the PA-RISC port still uses reload
and not LRA, but it too seems to have the
Updated with an additional item I just debugged.
Since the code that uses the doloop_end pattern does not check the operand mode
as given in the pattern, the pattern itself may need to do this, and that was
not documented. In addition, if the doloop_end pattern is a define_expand,
there must
On 10/9/18 5:29 PM, Giuliano Augusto Faulin Belinassi wrote:
> Fixed all issues pointed in the previous iteration.
> There is now a significant change regarding how the sin(atan(x))
> constant is calculated, as now it checks for which values such that
> computing 1 + x*x won't overflow. There are
On 10/11/18 2:40 PM, Jeff Law wrote:
> On 10/11/18 1:23 PM, Peter Bergner wrote:
>> On 10/11/18 1:18 PM, Peter Bergner wrote:
>>> Ok, after working in gdb, I see that the PA-RISC port still uses reload
>>> and not LRA, but it too seems to have the same issue of reusing input
>>> regs that have
While writing tests for attribute flatten I wasn't 100% sure
from reading the manual if it would respect attribute noinline
or if it would override it. Turns out it's the former so
the attached patch mentions it in the manual. Unless there
are objections I will commit the change tomorrow.
On 10/11/2018 03:23 PM, Peter Bergner wrote:
On 10/11/18 1:18 PM, Peter Bergner wrote:
Ok, after working in gdb, I see that the PA-RISC port still uses reload
and not LRA, but it too seems to have the same issue of reusing input
regs that have REG_DEAD notes, so the question still stands. It's
On 10/10/18 4:36 PM, Jeff Law wrote:
> On 10/2/18 10:37 AM, Martin Sebor wrote:
>> [3/4] - Change sprintf to use new get_range_strlen overload
>>
>> This change makes use of the new get_range_strlen() overload
>> in gimple-ssa-sprintf.c. This necessitated a few changes to
>> the API but also
This patch makes extensive use of C++11 direct init in
__debug::forward_list.
Doing so I also try to detect useless creation of safe iterators in
debug implementation. In __debug::forward_list there are severals but I
wonder if it is worth fixing those. Most of them are like this:
Hi,
On 11/10/18 20:36, Jason Merrill wrote:
On Wed, Oct 3, 2018 at 8:18 AM Paolo Carlini wrote:
a simple issue, we weren't correctly implementing 7.5.7.4 on the
terminating semicolon. Tested x86_64-linux.
If the missing semicolon is followed by }, let's allow it with a pedwarn.
I see.
On 10/11/18 1:23 PM, Peter Bergner wrote:
> On 10/11/18 1:18 PM, Peter Bergner wrote:
>> Ok, after working in gdb, I see that the PA-RISC port still uses reload
>> and not LRA, but it too seems to have the same issue of reusing input
>> regs that have REG_DEAD notes, so the question still stands.
On 10/11/18 1:18 PM, Peter Bergner wrote:
> Ok, after working in gdb, I see that the PA-RISC port still uses reload
> and not LRA, but it too seems to have the same issue of reusing input
> regs that have REG_DEAD notes, so the question still stands. It's just
> that whatever fix we come up with
In working out how to orchestrate the preprocessor/parser dance that
modules require, I came across the confused logic of
cp_parser_translation_unit. It is only called once, but looks as if it
used to be called multiple times, and retains the scar tissue from that.
It also bails out on the
Here is the same version again with an extra test added:
The popcount expansion uses SIMD instructions acting on 64-bit values.
As a result a popcount of a 32-bit integer requires zero-extension before
moving the zero-extended value into an FP register. This patch adds
support for zero-extended
On Wed, Oct 3, 2018 at 8:18 AM Paolo Carlini wrote:
> a simple issue, we weren't correctly implementing 7.5.7.4 on the
> terminating semicolon. Tested x86_64-linux.
If the missing semicolon is followed by }, let's allow it with a pedwarn.
Jason
OK.
On Thu, Oct 11, 2018 at 1:59 PM Jakub Jelinek wrote:
>
> On Mon, Sep 24, 2018 at 12:48:56PM +0200, Paolo Carlini wrote:
> > as explained in the audit trail, the gcc_assert added by Nathan triggers
> > during error-recovery too, when add_method correctly returns false because
> > it failed to
OK.
On Fri, Sep 21, 2018 at 5:22 PM David Malcolm wrote:
>
> This is v2 of the patch; I managed to bit-rot my own patch due to my
> fix for r264335, which tightened up the "is this meaningful" threshold
> on edit distances when finding spelling correction candidates.
>
> The only change in this
Hi,
On 11/10/18 19:59, Jakub Jelinek wrote:
On Mon, Sep 24, 2018 at 12:48:56PM +0200, Paolo Carlini wrote:
as explained in the audit trail, the gcc_assert added by Nathan triggers
during error-recovery too, when add_method correctly returns false because
it failed to add the method. Thus it
On 10/10/18 7:57 PM, Peter Bergner wrote:
> The problem is, that hard reg %r26 is defined in insn 32, to be used in
> insn 33, so using %r26 as the reload reg is wrong, because it will clobber
> the value we set in insn 32. Looking thru LRA, it looks like LRA assumes
> that for a reload, if one
Hello,
Most VxWorks programs refer to kernel services at some point, and these
often end up too far away for short calls on powerpc.
This patch, originally contributed by Doug, arranges to request linker
relaxation by default for RTPs. This helps many programs transparently,
the feature has been
On Mon, Sep 24, 2018 at 12:48:56PM +0200, Paolo Carlini wrote:
> as explained in the audit trail, the gcc_assert added by Nathan triggers
> during error-recovery too, when add_method correctly returns false because
> it failed to add the method. Thus it seems that we should simply loosen a
> bit
On Thu, 11 Oct 2018, Martin Sebor wrote:
> (Or are there some differences between the underscored forms and
> the one without it)?
They should always behave the same.
> With a few exceptions (like aligned) the built-in returns false
> for attributes that aren't attached to a node. I haven't
On Thu, 11 Oct 2018, Wilco Dijkstra wrote:
> Hi,
>
> > Note that "will ever set errno" includes possibly setting it in the
> > future, since code may be built with one libm version and used with
> > another. So it wouldn't be correct to have a "never sets errno" attribute
> > on glibc logb /
No change in terms of contents.
Committed.
Gerald
Index: steering.html
===
RCS file: /cvs/gcc/wwwdocs/htdocs/steering.html,v
retrieving revision 1.45
diff -u -r1.45 steering.html
--- steering.html 30 Sep 2018 14:38:47 -
On Thu, 11 Oct 2018, Wilco Dijkstra wrote:
> For the new math functions all errno handling is already done in a
> single function so you could do something like: if (want_errno)
> set_errno (error).
want_errno being a global (or TLS) variable gets into the same problems as
_LIB_VERSION; you
Hi!
I've bootstrapped/regtested and committed following 4 backports from trunk
to gcc-8-branch.
Jakub
2018-10-11 Jakub Jelinek
Backported from mainline
2018-09-12 Jakub Jelinek
Andreas Krebbel
PR tree-optimization/86844
*
Ping.
On 10/2/18 9:19 AM, Richard Henderson wrote:
> Changes since v1:
> * Use config/t-slibgcc-libgcc instead of gcc.c changes.
> * Some style fixes.
> * Ifdefs to work with old glibc.
>
> * Force TImode registers into even regnos.
> Required by CASP, allowed by the ABI, and is seen
Wilco Dijkstra writes:
> As mentioned in PR87511, the shift used in aarch64_mask_and_shift_for_ubfiz_p
> should be evaluated as a HOST_WIDE_INT rather than int.
>
> Passes bootstrap, OK for commit and backport?
OK for both, thanks.
Richard
(Reviewing since this is really an rtl handling bug
As mentioned in PR87511, the shift used in aarch64_mask_and_shift_for_ubfiz_p
should be evaluated as a HOST_WIDE_INT rather than int.
Passes bootstrap, OK for commit and backport?
ChangeLog:
2018-10-11 Wilco Dijkstra
gcc/
* config/aarch64/aarch64.c
Avoid a system call when no sleep is required. Sleep in a loop (actually
two loops) to handle interruption by signals.
PR libstdc++/80538
* src/c++11/thread.cc (this_thread::__sleep_for)
[_GLIBCXX_HAVE_SLEEP]: Only call sleep for non-zero values.
Loop while sleep
On 10/11/18 8:20 AM, David Malcolm wrote:
> Mismatched dg-{begin|end}-multiline-output directives are currently
> silently ignored, leading to difficult-to-diagnose test failures
> involving excess output.
>
> This patch makes multiline.exp complain about them.
>
> Successfully regrtested on
.. likewise.
On 04/10/18 12:32, Paolo Carlini wrote:
Hi,
yesterday I didn't notice that we have a separate bug for a similar
issue affecting cp_parser_compound_requirement. Tested x86_64-linux.
Thanks, Paolo.
PS: while working on these issues, I noticed that somewhere else in
the concepts
Hi,
already pinging this because it seems rather straightforward to me...
On 03/10/18 14:18, Paolo Carlini wrote:
Hi,
a simple issue, we weren't correctly implementing 7.5.7.4 on the
terminating semicolon. Tested x86_64-linux.
https://gcc.gnu.org/ml/gcc-patches/2018-10/msg00173.html
On Wed, Oct 3, 2018 at 7:11 PM Marek Polacek wrote:
>
> On Wed, Oct 03, 2018 at 10:24:52AM -0400, Jason Merrill wrote:
> > On Tue, Oct 2, 2018 at 5:25 PM Marek Polacek wrote:
> > >
> > > On Mon, Oct 01, 2018 at 07:47:10PM -0400, Jason Merrill wrote:
> > > > On Mon, Oct 1, 2018 at 6:41 PM Marek
Hi,
> if (math_errhandling & MATH_ERRNO) == 0 a math
> function may still set errno.
>
> it can only set it if there was an error though,
> not arbitrarily clobber it, but this means that
>
> (1) reordering errno access around math calls is
> invalid even with -fno-math-errno.
It's typically
Hi,
I reworked patch use a tuning flag instead of checking explicit for
CPU flavor. I will send soon an update for it, which won't use the
static variable anymore, and uses instead the SCHED-api.
I would like first to get some comments on current version.
Regards,
Kai
Jim Wilson
Since the code that uses the doloop_end pattern does not check the operand mode
as given in the pattern, the pattern itself may need to do this, and that was
not documented. This patch adds that information. It also updates the example
to reflect this.
Ok for trunk?
paul
ChangeLog:
On 10/11/2018 06:04 AM, Joseph Myers wrote:
On Thu, 11 Oct 2018, Martin Sebor wrote:
The attached patch introduces a built-in function called
__builtin_has_attribute that makes some of this possible.
See the documentation and tests for details.
I see nothing in the documentation about
The section from the standard in comment #2 is being misinterpreted.
R609 and R610 concern substrings in the context of section 6.4, which
concerns SCALARS!
Section 6.5 concerns arrays and you will find there:
R618 array-section is data-ref [ ( substring-range ) ] or
complex-part-designator
Hi,
> Note that "will ever set errno" includes possibly setting it in the
> future, since code may be built with one libm version and used with
> another. So it wouldn't be correct to have a "never sets errno" attribute
> on glibc logb / lround / llround / lrint / llrint / fma / remquo
On Thu, Oct 11, 2018 at 10:28 AM Jason Merrill wrote:
>
> On Wed, Oct 10, 2018 at 5:01 PM David Malcolm wrote:
> > On Tue, 2018-10-09 at 18:38 -0400, Jason Merrill wrote:
> > > On Tue, Oct 9, 2018 at 1:19 PM David Malcolm
> > > wrote:
> > > > + /* Emulation of a "move" constructor, but really
On Wed, Oct 10, 2018 at 5:01 PM David Malcolm wrote:
> On Tue, 2018-10-09 at 18:38 -0400, Jason Merrill wrote:
> > On Tue, Oct 9, 2018 at 1:19 PM David Malcolm
> > wrote:
> > > + /* Emulation of a "move" constructor, but really a copy
> > > + constructor. */
> > > +
> > > + name_hint
Another data point: I tried some of the testcases with IBM XL Fortran
compiler and it didn't complain or fail. I have no idea if this
behavior is an IBM extension or how other Fortran compilers behave.
GNU Fortran probably should be compatible with other Fortran
compilers, even if the behavior
On 04/10/18 19:12, Jeff Law wrote:
OK. Thanks for fixing the reader and adding selftests.
Thanks, committed.
Andrew
Apparently this change also breaks CPU SPEC (465.tonto?).
Is this really a correct change? Even if the testcases really are
wrong, it seems that this change is going to break a lot of code in
the wild. This seems like a very bad choice, even if it pedantically
is correct.
Thanks, David
On
Joseph Myers wrote:
> On Mon, 8 Oct 2018, Richard Biener wrote:
>> So I think it would be fine if we'd have -fno-math-errno as documented
>> and then the C library would annotate their math functions according
>> to whether they will ever set errno or not. Once a math function is
>> const or
Since FDPIC requires an architecture >=7, these tests fail because
they enforce and older version. They would pass if the compiler didn't
bail out though.
2018-07-13 Christophe Lyon
* gcc.target/arm/armv6-unaligned-load-ice.c: Add arm_arch
effective-target.
*
uclibc defines bswap_32, so use a different name in this test.
2018-XX-XX Christophe Lyon
gcc/testsuite/
* gcc.target/arm/pr43698.c (bswap_32): Rename as my_bswap_32.
Change-Id: I2591bd911030814331cabf97ee5cf6cf8124b4f3
diff --git a/gcc/testsuite/gcc.target/arm/pr43698.c
Some tests have the "nonpic" guard, but pass on
arm*-*-uclinuxfdpiceabi because it is in PIE mode by default. Rather
than adding this target to all these tests, add the "pie_enabled"
effective target.
2018-XX-XX Christophe Lyon
gcc/testsuite/
* g++.dg/cpp0x/noexcept03.C: Add
Add *-*-uclinux* to tests that work on this target.
2018-XX-XX Christophe Lyon
gcc/testsuite/
* g++.dg/abi/forced.C: Add *-*-uclinux*.
* g++.dg/abi/guard2.C: Likewise.
* g++.dg/ext/cleanup-10.C: Likewise.
* g++.dg/ext/cleanup-11.C: Likewise.
*
Some tests fail on arm*-*-uclinuxfdpiceabi because it generates PIC
code and they don't support it: skip them. They also fail on
arm*-linux* when forcing -fPIC.
2018-XX-XX Christophe Lyon
gcc/testsuite/
* gcc.target/arm/eliminate.c: Accept only nonpic targets.
*
v6-M and v8-M are not supported currently in FDPIC mode, it's better
to skip the tests.
2018-XX-XX Christophe Lyon
Mickaël Guêné
gcc/testsuite/
* gcc.target/arm/atomic-comp-swap-release-acquire-3.c: Skip on
arm*-*-uclinuxfdpiceabi.
*
In FDPIC mode, r9 is saved in addition to other registers, so update
the expected patterns accordingly.
2018-XX-XX Christophe Lyon
Mickaël Guêné
* gcc/testsuite/
* gcc.target/arm/interrupt-1.c: Add scan-assembler pattern for
arm*-*-uclinuxfdpiceabi.
*
Several tests cannot work on ARM-FDPIC for various reasons: skip them,
or skip some directives.
gcc.dg/20020312-2.c: Skip since it forces -fno-pic.
gcc.target/arm/:
* Skip since r9 is clobbered by assembly code:
20051215-1.c
mmx-1.c
pr61948.c
pr77933-1.c
pr77933-2.c
* Skip since the
We call __aeabi_read_tp() to get the thread pointer. Since this is a
function call, we have to restore the FDPIC register afterwards.
2018-XX-XX Christophe Lyon
Mickaël Guêné
gcc/
* config/arm/arm.c (arm_load_tp): Add FDPIC support.
* config/arm/arm.md
Without this, when we are unwinding across a signal frame we can jump
to an even address which leads to an exception.
This is needed in __gnu_persnality_sigframe_fdpic() when restoring the
PC from the signal frame since the PC saved by the kernel has the LSB
bit set to zero.
2018-XX-XX
2018-XX-XX Christophe Lyon
Mickaël Guêné
libgcc/
* unwind-arm-common.inc (ARM_SET_R7_RT_SIGRETURN)
(THUMB2_SET_R7_RT_SIGRETURN, FDPIC_LDR_R12_WITH_FUNCDESC)
(FDPIC_LDR_R9_WITH_GOT, FDPIC_LDR_PC_WITH_RESTORER)
(FDPIC_FUNCDESC_OFFSET,
Support additional relocations: TLS_GD32_FDPIC, TLS_LDM32_FDPIC, and
TLS_IE32_FDPIC.
We do not support the GNU2 TLS dialect.
2018-XX-XX Christophe Lyon
Mickaël Guêné
gcc/
* config/arm/arm.c (tls_reloc): Add TLS_GD32_FDPIC,
TLS_LDM32_FDPIC and TLS_IE32_FDPIC.
In FDPIC mode, the trampoline generated to support pointers to nested
functions looks like:
.wordtrampoline address
.wordtrampoline GOT address
ldr r12, [pc, #8]
ldr r9, [pc, #8]
ldr pc, [pc, #8]
OK.
On Thu, Oct 11, 2018 at 8:52 AM Jakub Jelinek wrote:
>
> Hi!
>
> Except for std::tuple* structured bindings, the VAR_DECLs we create for the
> identifiers aren't actually variables, but placeholders with
> DECL_VALUE_EXPR. If the structured binding is not a reference, it is still
> an
Use local binding rules to decide whether we can use GOTOFFFUNCDESC to
compute the function address.
2018-XX-XX Christophe Lyon
Mickaël Guêné
gcc/
* config/arm/arm.c (arm_local_funcdesc_p): New function.
(legitimize_pic_address): Ensure binding rules on
2018-XX-XX Christophe Lyon
Mickaël Guêné
gcc/
* config/arm/arm.c (arm_compute_save_reg0_reg12_mask): Handle
FDPIC.
(thumb1_compute_save_core_reg_mask): Likewise.
diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c
index 92ae24b..a6dce36 100644
---
The main difference with existing support is that function addresses
are function descriptor addresses instead. This means that all code
dealing with function pointers now has to cope with function
descriptors instead.
For the same reason, Linux kernel helpers can no longer be called by
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.
2018-XX-XX Christophe Lyon
Mickaël Guêné
* libgcc/crtstuff.c: Add support for FDPIC.
Change-Id:
The FDPIC register is hard-coded to r9, as defined in the ABI.
We have to disable tailcall optimizations if we don't know if the
target function is in the same module. If not, we have to set r9 to
the value associated with the target module.
When generating a symbol address, we have to take into
In FDPIC mode, we set -fPIE unless the user provides -fno-PIE, -fpie,
-fPIC or -fpic: indeed FDPIC code is PIC, but we want to generate code
for executables rather than shared libraries by default.
We also make sure to use the --fdpic assembler option, and select the
appropriate linker emulation.
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
like *-*-linux*: in most places, we add *-uclinux* where there is
already *-linux*, or uclinux*
2018-XX-XX Christophe Lyon
Mickaël Guêné
gcc/
* config/arm/arm.opt: Add -mfdpic option.
* doc/invoke.texi: Add documentation for -mfdpic.
diff --git a/gcc/config/arm/arm.opt b/gcc/config/arm/arm.opt
index a1286a4..231c1cb 100644
--- a/gcc/config/arm/arm.opt
Hello,
This patch series implements the GCC contribution of the FDPIC ABI for
ARM targets.
This ABI enables to run Linux on ARM MMU-less cores and supports
shared libraries to reduce the memory footprint.
Without MMU, text and data segments relative distances are different
from one process to
Mismatched dg-{begin|end}-multiline-output directives are currently
silently ignored, leading to difficult-to-diagnose test failures
involving excess output.
This patch makes multiline.exp complain about them.
Successfully regrtested on x86_64-pc-linux-gnu
OK for trunk?
Consider:
demo.c: In function 'test':
demo.c:5:40: error: macro "LOG_2" requires 3 arguments, but only 2 given
5 | LOG_2 ("loading file: %s\n", filename);
|^
This patch adds a note showing the definition of the macro in
question, giving:
demo.c: In
On Thu, 11 Oct 2018, Martin Liška wrote:
> Hi.
>
> As seen in the PR, switch conversion can do better when we return equal
> numbers
> based on index value. I implemented more than that, more precisely I support
> all linear
> transformation based on index value. It's the same what clang is
Hi Jeff,
> So I went back and reviewed all the discussion around this. I'm still
> having trouble getting comfortable with flipping the default -- unless
> we know ahead of time that the target runtime doesn't set errno on any
> of the math routines. That implies a target hook to describe the
>
On Thu, Oct 11, 2018 at 02:56:14PM +0200, Martin Liška wrote:
> As seen in the PR, switch conversion can do better when we return equal
> numbers
> based on index value. I implemented more than that, more precisely I support
> all linear
> transformation based on index value. It's the same what
They are 1:1 the same.
Bootstrapped on x86_64-unknown-linux-gnu, applied.
Richard.
2018-10-11 Richard Biener
* config/i386/x86-tune-costs.h (bdver?_memcpy, bdver?_memset,
bdver?_cost): Unify to ...
(bdver_memcpy, bdver_memset, bdver_cost): ... this.
*
Hi.
As seen in the PR, switch conversion can do better when we return equal numbers
based on index value. I implemented more than that, more precisely I support
all linear
transformation based on index value. It's the same what clang is capable of.
Patch survives testing on x86_64-linux-gnu. I
1 - 100 of 115 matches
Mail list logo
