PR sanitizer/81125
* ubsan.h (enum ubsan_encode_value_phase): New.
(ubsan_encode_value): Change second argument to
enum ubsan_encode_value_phase with default value of
UBSAN_ENCODE_VALUE_GENERIC.
* ubsan.c (ubsan_encode_value): Change second argument to
enum ubsan_encode_value_phase PHASE from bool IN_EXPAND_P,
adjust uses, for UBSAN_ENCODE_VALUE_GENERIC use just
create_tmp_var_raw instead of create_tmp_var and use a
TARGET_EXPR.
(ubsan_expand_bounds_ifn, ubsan_build_overflow_builtin,
instrument_bool_enum_load, ubsan_instrument_float_cast): Adjust
ubsan_encode_value callers.
* g++.dg/ubsan/pr81125.C: New test.
From-SVN: r249376
PR sanitizer/81111
* ubsan.c (ubsan_encode_value): If current_function_decl is NULL,
use create_tmp_var_raw instead of create_tmp_var, mark it addressable
just by setting TREE_ADDRESSABLE on the result and use a TARGET_EXPR.
* g++.dg/ubsan/pr81111.C: New test.
From-SVN: r249375
2017-06-19 Martin Liska <mliska@suse.cz>
PR ipa/80732
* attribs.c (make_dispatcher_decl): Do not append '.ifunc'
to dispatcher function name.
* multiple_target.c (replace_function_decl): New function.
(create_dispatcher_calls): Redirect both edges and references.
2017-06-19 Martin Liska <mliska@suse.cz>
PR ipa/80732
* gcc.target/i386/mvc5.c: Scan indirect_function.
* gcc.target/i386/mvc7.c: Likewise.
* gcc.target/i386/pr80732.c: New test.
From-SVN: r249365
2017-06-19 Richard Biener <rguenther@suse.de>
* tree-ssa-loop-niter.h (estimate_numbers_of_iterations): Take
struct function as arg.
(estimate_numbers_of_iterations): Export overload with loop arg.
(free_numbers_of_iterations_estimates_loop): Use an overload of
free_numbers_of_iterations_estimates instead.
* tree-cfg.c (remove_bb): Adjust.
* tree-cfgcleanup.c (remove_forwarder_block_with_phi): Likewise.
* tree-parloops.c (gen_parallel_loop): Likewise.
* tree-ssa-loop-ivcanon.c (canonicalize_induction_variables):
Likewise.
(tree_unroll_loops_completely): Likewise.
* tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop):
Use an overload instead and export.
(estimated_loop_iterations): Adjust.
(max_loop_iterations): Likewise.
(likely_max_loop_iterations): Likewise.
(estimate_numbers_of_iterations): Take struct function as arg
and adjust.
(loop_exits_before_overflow): Adjust.
(free_numbers_of_iterations_estimates_loop): Use an overload.
* tree-vect-loop.c (vect_analyze_loop_form): Adjust.
* tree-vectorizer.c (vect_free_loop_info_assumptions): Likewise.
From-SVN: r249358
2017-06-16 Carl Love <cel@us.ibm.com>
* gcc/ChangeLog: Update for commit 249311 didn't get committed
* gcc/testsuite/ChangeLog: Update for commit commit 249311 didn't
get committed.
From-SVN: r249336
* pt.c (type_dependent_expression_p): An undeduced auto outside the
template isn't dependent.
* call.c (build_over_call): Instantiate undeduced auto even in a
template.
From-SVN: r249323
The neon-thumb2-move.c test was overriding the options that had been
detected as being necessary to enable Neon. The result was that the
combination of the test's options and those auto-detected were not
compatible with neon leading to a test failure. The correct fix here
is to stick with the options that dg-add-options arm_neon has worked
out.
The thumb2-slow-flash-data tests were relying (incorrectly) on a
particular FPU being enabled by default. These tests are fixed by
adding +fp to the architecture selected.
* gcc.target/arm/neon-thumb2-move.c (dg-options): Don't override
the architecture options added by dg-add-options arm_neon.
* gcc.target/arm/thumb2-slow-flash-data-2.c (dg-opitions): Add +fp
to the architecture.
* gcc.target/arm/thumb3-slow-flash-data-3.c (dg-opitions): Likewise.
* gcc.target/arm/thumb4-slow-flash-data-3.c (dg-opitions): Likewise.
* gcc.target/arm/thumb5-slow-flash-data-3.c (dg-opitions): Likewise.
From-SVN: r249310
-marm and -mthumb are opposites: one cancels out the other. This patch
marks them as such so that the driver will eliminate all but the last
option on the command line. This aids multilib selection which otherwise
can get confused if both are present.
* config/arm/arm.opt (marm): Mark as the negative of of -mthumb.
(mthumb): Mark as the negative of -marm.
From-SVN: r249309
This patch adds the remainder of the main documentation changes. It
adds the changes for -mcpu, -mtune and -mfpu. I've chosen to document
the extension options under -mcpu rather than under -mtune because,
while they are permitted with -mtune, they do not affect the behaviour
of the tuning done by the compiler.
I've also inverted the sense of the table (making the primary index
the extension name and then listing the CPU names to which it applies.
This is because the extensions are much more orthoganal in meaning
here and having a primary entry via the CPU name would lead to
enormous duplication.
Finally, it adds the relevant changes to -mfpu. I haven't stated yet
that any setting of -mfpu other than 'auto' is deprecated, but that is
certainly the long-term goal of this patch series.
* doc/invoke.texi (ARM Options, -mcpu): Document supported
extension options.
(ARM Options, -mtune): Document that this accepts the same
extension options as -mcpu.
(ARM Options, -mfpu): Document addition of -mfpu=auto.
From-SVN: r249308
This adds documentation for the new extension options to -march= on ARM.
I tried a number of different ways of formatting the information, but this
seems the best, given what can be achieved in texinfo format.
* doc/invoke.texi (ARM Options, -march=): Document new syntax and
permitted extensions.
From-SVN: r249307
Reviewing the list of options for the purposes of writing the
documentation revealed that a small number of options were missing.
Mostly these are aliases for existing options, but in a couple of
cases we lacked the ability to disable certain other options.
* config/arm/arm-cpus.in (armv7): Add extension +nofp.
(armv7-r): Add aliases vfpv3xd and vfpv3-d16.
(armv8-m.main): Add option +nodsp.
* config/arm/arm-cpu-cdata.h: Regenerated.
From-SVN: r249306
It looks like the fuchsia port relied on inheriting the multilib rules from
the bare-metal port (the t-arm-elf makefile fragment), but that has now been
rewritten on the assuption that the base architecture is ARMv4t; fuchsia
has a base architecture of ARMv7-a.
To account for this, I've cloned the original t-arm-elf rules into a
new makefile fragment t-fuchsia and arranged for that to be used when
targetting this system.
* config/arm/t-fuchsia: New file.
* config.gcc (arm*-*-fuchsia*): Use it.
From-SVN: r249305
Symbianelf used to build multilib for armv5t with softfp, but that
architecture doesn't really support floating point instructions. This
patch reworks the multilib configuration to use armv5te as the base
when building for floating point.
I'm not sure just how useful the symbian port is these days, so this
has only been very lightly tested (checks that libgcc builds for all
multilib variants). Perhaps we should consider deprecating this
config?
* config/arm/t-symbian: Rewrite for new option infrastructure.
From-SVN: r249304
NB. This configuration does not build in GCC-7 and doesn't build now either.
This patch resets a couple of multlib variables which previously were
not cleared.
It almost certainly needs further work to make it use the new option
framework correctly, but since the library configurations are already
clearly wrong, it's not clear what the changes need to be. In
particular it tries to build a hard-float library for ARM7TDMI in both
ARM and thumb modes, but ARMv4t does not support any floating-point
instructions; furthermore, GCC has never supported a hard-float thumb1
library.
* config/arm/t-phoenix (MULTILIB_REUSE): Clear variable.
(MULTILIB_REQUIRED): Likewise.
From-SVN: r249303
No real change, but for consistency reset all multilib related variables.
* config/arm/t-linux-eabi (MULTILIB_EXCEPTIONS): Set to empty.
(MULTILIB_RESUE): Likewise.
(MULTILIB_MATCHES): Likewise.
(MULTLIB_REQUIRED): Likewise.
From-SVN: r249302
