Jakub Jelinek [Mon, 26 Feb 2018 19:46:34 +0000 (20:46 +0100)]
re PR debug/83917 (with -mcall-ms2sysv-xlogues, stepping into x86 tail-call restore stub gives bad backtrace)
PR debug/83917
* config/i386/i386-asm.h (PACKAGE_VERSION, PACKAGE_NAME,
PACKAGE_STRING, PACKAGE_TARNAME, PACKAGE_URL): Undefine between
inclusion of auto-target.h and auto-host.h.
(USE_GAS_CFI_DIRECTIVES): Define if not defined already based on
__GCC_HAVE_DWARF2_CFI_ASM.
(cfi_startproc, cfi_endproc, cfi_adjust_cfa_offset,
cfi_def_cfa_register, cfi_def_cfa, cfi_register, cfi_offset, cfi_push,
cfi_pop): Define.
* config/i386/cygwin.S: Don't include auto-host.h here, just
define USE_GAS_CFI_DIRECTIVES to 1 or 0 and include i386-asm.h.
(cfi_startproc, cfi_endproc, cfi_adjust_cfa_offset,
cfi_def_cfa_register, cfi_register, cfi_push, cfi_pop): Remove.
* config/i386/resms64fx.h: Add cfi_* directives.
* config/i386/resms64x.h: Likewise.
x86-64 is supposed to have asynchronous unwind tables by default, but
there is nothing that reflects the change in the (relative) frame
address after .LIND0. That region really has to be moved outside of
the .cfi_startproc/.cfi_endproc bracket.
This patch adds TARGET_INDIRECT_BRANCH_REGISTER to force indirect
branch via register whenever -mindirect-branch= is used. Now,
-mindirect-branch=thunk-extern -O2 on x86-64 GNU/Linux generates:
Eric Botcazou [Mon, 26 Feb 2018 16:29:30 +0000 (16:29 +0000)]
re PR rtl-optimization/83496 (wrong code generated with -Os -mbranch-cost=1)
PR rtl-optimization/83496
* reorg.c (steal_delay_list_from_target): Change REDUNDANT array from
booleans to RTXes. Call fix_reg_dead_note on every non-null element.
(steal_delay_list_from_fallthrough): Call fix_reg_dead_note on a
redundant insn, if any.
(relax_delay_slots): Likewise.
(update_reg_unused_notes): Rename REDUNDANT_INSN to OTHER_INSN.
This patch prevents pattern-matching of fold-left SLP reduction chains,
which the previous patch for 83965 didn't handle properly. It only
stops the last statement in the group from being matched, but that's
enough to cause the group to be dissolved later.
A better fix would be to put all the information about the reduction
on the the first statement in the reduction chain, so that every
statement in the group can tell what the group is doing. That doesn't
seem like stage 4 material though.
2018-02-26 Richard Sandiford <richard.sandiford@linaro.org>
gcc/
PR tree-optimization/83965
* tree-vect-patterns.c (vect_reassociating_reduction_p): Assume
that grouped statements are part of a reduction chain. Return
true if the statement is not marked as a reduction itself but
is part of a group.
(vect_recog_dot_prod_pattern): Don't check whether the statement
is part of a group here.
(vect_recog_sad_pattern): Likewise.
(vect_recog_widen_sum_pattern): Likewise.
gcc/testsuite/
PR tree-optimization/83965
* gcc.dg/vect/pr83965-2.c: New test.
H.J. Lu [Mon, 26 Feb 2018 15:29:30 +0000 (15:29 +0000)]
i386: Update -mfunction-return= for return with pop
When -mfunction-return= is used, simple_return_pop_internal should pop
return address into ECX register, adjust stack by bytes to pop from stack
and jump to the return thunk via ECX register.
Tested on i686 and x86-64.
PR target/84530
* config/i386/i386-protos.h (ix86_output_indirect_jmp): Remove
the bool argument.
(ix86_output_indirect_function_return): New prototype.
(ix86_split_simple_return_pop_internal): Likewise.
* config/i386/i386.c (indirect_return_via_cx): New.
(indirect_return_via_cx_bnd): Likewise.
(indirect_thunk_name): Handle return va CX_REG.
(output_indirect_thunk_function): Create alias for
__x86_return_thunk_[re]cx and __x86_return_thunk_[re]cx_bnd.
(ix86_output_indirect_jmp): Remove the bool argument.
(ix86_output_indirect_function_return): New function.
(ix86_split_simple_return_pop_internal): Likewise.
* config/i386/i386.md (*indirect_jump): Don't pass false
to ix86_output_indirect_jmp.
(*tablejump_1): Likewise.
(simple_return_pop_internal): Change it to define_insn_and_split.
Call ix86_split_simple_return_pop_internal to split it for
-mfunction-return=.
(simple_return_indirect_internal): Call
ix86_output_indirect_function_return instead of
ix86_output_indirect_jmp.
[Patch AArch64] Turn on frame pointer / partial fix for PR84521
This fixes a GCC-8 regression that we accidentally switched off frame
pointers in the AArch64 backend when changing the defaults in the common
parts of the code. This breaks an ABI decision that was made in GCC at
the dawn of the port with respect to having a frame pointer at all
times. If we really want to turn this off lets have a discussion around
that separately.
For now turn this back on and I believe this will leave PR84521 latent
again with -fomit-frame-pointer and (hopefully) make the ruby issue go
away. I'm asking Sudi to pick that up.
Bootstrapped and regression tested on AArch64-none-linux-gnu but I see
one regression in gcc.c-torture/execute/960419-2.c which needs to be
looked at next (PR84528, thanks Kyrill).
Some command-line options have been deprecated for a long time. This
patch adds a warning for them, so that we can remove them in GCC 9
without surprising any users.
* config/rs6000/rs6000.opt (mvrsave=no, mvrsave=yes, isel=no,
isel=yes): Warn for these deprecated options.
Thomas Koenig [Sun, 25 Feb 2018 09:02:32 +0000 (09:02 +0000)]
re PR fortran/78238 ([OOP] ICE: verify_gimple failed, with -fdefault-integer-8)
2018-02-25 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/78238
* gfortran.h (gfc_integer_4_kind): Define.
* resolve.c (resolve_select_type): Make sure that the
kind of c->high is gfc_integer_4_kind.
2018-02-25 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/78238
* gfortran.dg/select_type_40.f90: New test.
Marek Polacek [Sat, 24 Feb 2018 13:10:44 +0000 (13:10 +0000)]
re PR c++/83692 (Rejects valid constexpr with unrelated code fixing problem)
PR c++/83692
* constexpr.c (maybe_constant_init_1): New function.
(maybe_constant_init): Make it a wrapper around maybe_constant_init_1.
(cxx_constant_init): New function.
* cp-tree.h (cxx_constant_init): Declare.
* typeck2.c (store_init_value): Call cxx_constant_init instead of
cxx_constant_value. Move the maybe_constant_init call under an 'else'.
Carl Love [Fri, 23 Feb 2018 17:22:10 +0000 (17:22 +0000)]
rs6000-builtin.def: Change VSIGNED2 and VUNSIGNED2 macro expansions from BU_VSX_2 to BU_P8V_VSX_2...
gcc/ChangeLog:
2018-02-23 Carl Love <cel@us.ibm.com>
* config/rs6000/rs6000-builtin.def: Change VSIGNED2 and VUNSIGNED2
macro expansions from BU_VSX_2 to BU_P8V_VSX_2 and BU_VSX_OVERLOAD_2 to
BU_P8V_OVERLOAD_2.
* config/rs6000/rs6000-c.c: Change VSX_BUILTIN_VEC_VSIGNED2 to
P8V_BUILTIN_VEC_VSIGNED2. Change VSX_BUILTIN_VEC_VUNSIGNED2 to
P8V_BUILTIN_VEC_VUNSIGNED2.
gcc/testsuite/ChangeLog:
2018-02-23 Carl Love <cel@us.ibm.com>
* gcc.target/powerpc/builtins-3-runnable.c: Move tests for vec_float2,
vec_signed2 and vec_unsigned2 to new Power 8 test file.
* gcc.target/powerpc/builtins-3-runnable-p8.c: New test file for
Power 8 tests.
Paul Thomas [Fri, 23 Feb 2018 12:35:30 +0000 (12:35 +0000)]
re PR fortran/83148 (ICE: crash_signal from toplev.c:325)
2018-02-23 Paul Thomas <pault@gcc.gnu.org>
PR fortran/83148
* trans-const.c : Clean up some whitespace issues.
* trans-expr.c (gfc_conv_initializer): If an iso_c_binding
derived type has a kind value of zero, set it to the default
integer kind.
2018-02-23 Paul Thomas <pault@gcc.gnu.org>
PR fortran/83148
* gfortran.dg/class_68.f90: New test.
Janne Blomqvist [Fri, 23 Feb 2018 09:07:24 +0000 (11:07 +0200)]
PR 84519 Handle optional QUIET specifier for STOP and ERROR STOP
Fortran 2018 adds a new QUIET specifier for the STOP and ERROR STOP
statements, in order to suppress the printing of signaling FP
exceptions and the stop code. This patch adds the necessary library
changes, but for now the new specifier is not parsed and the frontend
unconditionally adds a false value for the new argument.
Regtested on x86_64-pc-linux-gnu.
gcc/fortran/ChangeLog:
2018-02-23 Janne Blomqvist <jb@gcc.gnu.org>
PR fortran/84519
* trans-decl.c (gfc_build_builtin_function_decls): Add bool
argument to stop and error stop decls.
* trans-stmt.c (gfc_trans_stop): Add false value to argument
lists.
Vladimir Makarov [Thu, 22 Feb 2018 21:17:51 +0000 (21:17 +0000)]
re PR target/81572 (gcc-7 regression: unnecessary vector regmove on compare)
2018-02-22 Vladimir Makarov <vmakarov@redhat.com>
PR target/81572
* lra-int.h (LRA_UNKNOWN_ALT, LRA_NON_CLOBBERED_ALT): New macros.
* lra.c (lra_set_insn_recog_data, lra_update_insn_recog_data): Use
LRA_UNKNOWN_ALT.
* lra-constraints.c (curr_insn_transform): Set up
LRA_NON_CLOBBERED_ALT for moves processed on the fast path. Use
LRA_UNKNOWN_ALT.
(remove_inheritance_pseudos): Use LRA_UNKNOWN_ALT.
* lra-eliminations.c (spill_pseudos): Ditto.
(process_insn_for_elimination): Ditto.
* lra-lives.c (reg_early_clobber_p): Use the new macros.
* lra-spills.c (spill_pseudos): Use LRA_UNKNOWN_ALT and
LRA_NON_CLOBBERED_ALT.
H.J. Lu [Thu, 22 Feb 2018 17:09:06 +0000 (17:09 +0000)]
i386: Add __x86_indirect_thunk_nt_reg for -fcf-protection -mcet
nocf_check attribute can be used with -fcf-protection -mcet to disable
control-flow check by adding NOTRACK prefix before indirect branch.
When -mindirect-branch=thunk-extern -mindirect-branch-register is added,
indirect branch via register, "notrack call/jmp reg", is converted to
call/jmp __x86_indirect_thunk_nt_reg
When running on machines with CET enabled, __x86_indirect_thunk_nt_reg
can be updated to
notrack jmp reg
at run-time to restore NOTRACK prefix in the original indirect branch.
Since we don't support -mindirect-branch=thunk-extern, CET and MPX at
the same time, -mindirect-branch=thunk-extern is disallowed with
-fcf-protection=branch and -fcheck-pointer-bounds.
Tested on i686 and x86-64.
gcc/
PR target/84176
* config/i386/i386.c (ix86_set_indirect_branch_type): Issue an
error when -mindirect-branch=thunk-extern, -fcf-protection=branch
and -fcheck-pointer-bounds are used together.
(indirect_thunk_prefix): New enum.
(indirect_thunk_need_prefix): New function.
(indirect_thunk_name): Replace need_bnd_p with need_prefix. Use
"_nt" instead of "_bnd" for NOTRACK prefix.
(output_indirect_thunk): Replace need_bnd_p with need_prefix.
(output_indirect_thunk_function): Likewise.
(): Likewise.
(ix86_code_end): Update output_indirect_thunk_function calls.
(ix86_output_indirect_branch_via_reg): Replace
ix86_bnd_prefixed_insn_p with indirect_thunk_need_prefix.
(ix86_output_indirect_branch_via_push): Likewise.
(ix86_output_function_return): Likewise.
* doc/invoke.texi: Document -mindirect-branch=thunk-extern is
incompatible with -fcf-protection=branch and
-fcheck-pointer-bounds.
Due to there being no multilib mapping for Armv8-R, default multilib
builts for -march=armv4t with softfloat floating-point arithmetic is
being used. This patch maps it instead to the existing Armv7 multilibs.
Note that mapping for single-precision Armv8-R has been left out due to
there being no Arm implementation of that architecture variant.
2018-02-22 Thomas Preud'homme <thomas.preudhomme@arm.com>
gcc/
* config/arm/t-multilib: Map Armv8-R to Armv7 multilibs.
gcc/testsuite/
* gcc.target/arm/multilib.exp: Add tests for Armv8-R multilib mappings.
Janne Blomqvist [Thu, 22 Feb 2018 16:14:21 +0000 (18:14 +0200)]
PR 78534, 84509 Fix libgfortran API for PAUSE statement
This patch changes the libgfortran API for the PAUSE statement. By
passing a GFC_INTEGER_8 it handles -fdefault-integer-8, and for the
character version passing the length as a size_t.
Regtested on x86_64-pc-linux-gnu, committed as obvious.
gcc/fortran/ChangeLog:
2018-02-22 Janne Blomqvist <jb@gcc.gnu.org>
PR 78534
PR 84509
* trans-decl.c (gfc_build_builtin_function_decls): Pass
gfc_int8_type node to pause_numeric, size_type_node to
pause_string.
* trans-stmt.c (gfc_trans_pause): Likewise.
libgfortran/ChangeLog:
2018-02-22 Janne Blomqvist <jb@gcc.gnu.org>
PR 78534
PR 84509
* runtime/pause.c (pause_numeric): Modify to take GFC_INTEGER_8
argument.
(pause_string): Modify to take size_t character length argument.
Janne Blomqvist [Thu, 22 Feb 2018 09:44:31 +0000 (11:44 +0200)]
Character length cleanup for Coarray Fortran library
Following the change to use size_t for Fortran character lengths (PR
78534), this patch modifies the Coarray ABI in a similar way. The
single-image implementation that is included in libgfortran is
updated, but this needs corresponding work in the OpenCoarray library
as well for multi-image support. I also fixed the types for the STOP
and ERROR STOP implementation in libgfortran, as the calling of them
is somewhat intertwined with the calling of the corresponding CAF
functions. I'll send the OpenCoarray changes as a separate pull
request to the OpenCoarrays repository.
Regtested on x86_64-pc-linux-gnu.
gcc/fortran/ChangeLog:
2018-02-22 Janne Blomqvist <jb@gcc.gnu.org>
* gfortran.texi: Update Coarray API description.
* trans-decl.c (gfc_build_builtin_function_decls): Use size_t for
character lengths, int for exit codes.
(generate_coarray_sym_init): Use size_t for character length.
* trans-intrinsic.c (conv_co_collective): Likewise.
* trans-stmt.c (gfc_trans_lock_unlock): Likewise.
(gfc_trans_event_post_wait): Likewise.
(gfc_trans_sync): Likewise.
(gfc_trans_stop): Use size_t for character lengths, int for exit
codes.
libgfortran/ChangeLog:
2018-02-22 Janne Blomqvist <jb@gcc.gnu.org>
* libgfortran.h (stop_string): Use size_t for character length.
* runtime/stop.c (stop_string): Likewise.
(error_stop_string): Likewise.
(stop_numeric): Use int for exit code.
(error_stop_numeric): Likewise.
* caf/libcaf.h: Remove stdint.h include.
(_gfortran_caf_register): Use size_t for character length.
(_gfortran_caf_deregister): Likewise.
(_gfortran_caf_sync_all): Likewise.
(_gfortran_caf_sync_memory): Likewise.
(_gfortran_caf_sync_images): Likewise.
(_gfortran_caf_stop_numeric): Use int for exit code.
(_gfortran_caf_stop_str): Use size_t for character length.
(_gfortran_caf_error_stop_str): Likewise.
(_gfortran_caf_error_stop): Use int for exit code.
(_gfortran_caf_co_broadcast): Use size_t for character length.
(_gfortran_caf_co_sum): Likewise.
(_gfortran_caf_co_min): Likewise.
(_gfortran_caf_co_max): Likewise.
(_gfortran_caf_co_reduce): Likewise.
(_gfortran_caf_lock): Likewise.
(_gfortran_caf_unlock): Likewise.
(_gfortran_caf_event_post): Likewise.
(_gfortran_caf_event_wait): Likewise.
* caf/mpi.c (_gfortran_caf_register): Update implementation to
match prototype.
(_gfortran_caf_deregister): Likewise.
(_gfortran_caf_sync_all): Likewise.
(_gfortran_caf_sync_images): Likewise.
(_gfortran_caf_error_stop_str): Likewise.
(_gfortran_caf_error_stop): Likewise.
* caf/single.c (caf_internal_error): Likewise.
(_gfortran_caf_register): Likewise.
(_gfortran_caf_deregister): Likewise.
(_gfortran_caf_sync_all): Likewise.
(_gfortran_caf_sync_memory): Likewise.
(_gfortran_caf_sync_images): Likewise.
(_gfortran_caf_stop_numeric): Likewise.
(_gfortran_caf_stop_str): Likewise.
(_gfortran_caf_error_stop_str): Likewise.
(_gfortran_caf_error_stop): Likewise.
(_gfortran_caf_co_broadcast): Likewise.
(_gfortran_caf_co_sum): Likewise.
(_gfortran_caf_co_min): Likewise.
(_gfortran_caf_co_max): Likewise.
(_gfortran_caf_co_reduce): Likewise.
(_gfortran_caf_event_post): Likewise.
(_gfortran_caf_event_wait): Likewise.
(_gfortran_caf_lock): Likewise.
(_gfortran_caf_unlock): Likewise.
Jakub Jelinek [Thu, 22 Feb 2018 08:28:42 +0000 (09:28 +0100)]
re PR tree-optimization/84503 (store-merging miscompilation on powerpc64 with -O3 since r241789)
PR tree-optimization/84503
* gimple-ssa-store-merging.c (merged_store_group::merge_into): Compute
width as info->bitpos + info->bitsize - start.
(merged_store_group::merge_overlapping): Simplify width computation.
(check_no_overlap): New function.
(imm_store_chain_info::try_coalesce_bswap): Compute expected
start + width and last_order of the group, fail if check_no_overlap
fails.
(imm_store_chain_info::coalesce_immediate_stores): Don't merge info
to group if check_no_overlap fails.
* gcc.dg/pr84503-1.c: New test.
* gcc.dg/pr84503-2.c: New test.
rs6000: Delete meaningless arguments to define_{expand,split,peephole2}
This removes the (usually empty) constraint arguments to define_expand,
define_split, and define_peephole2: this argument is meaningless and
just noise.
* config/rs6000/altivec.md: Delete contraint arguments to
define_expand, define_split, and define_peephole2, and in
define_insn_and_split if always unused.
* config/rs6000/darwin.md: Ditto.
* config/rs6000/dfp.md: Ditto.
* config/rs6000/rs6000.md: Ditto.
* config/rs6000/sync.md: Ditto.
* config/rs6000/vector.md: Ditto.
* config/rs6000/vsx.md: Ditto.
This patch changes the remaining cases in our machine description files
to use brace blocks instead of double-quoted strings as the output
control string. This increases readability by making the blocks look
more like normal C code, mostly because backslash quoting is no longer
needed. It also removes such quoting where it was still there (usually
harmless but always confusing). and it writes "\n\t" as "\;" in one
place where we didn't already.
* config/rs6000/altivec.md: Write output control strings as braced
blocks instead of double-quoted strings.
* config/rs6000/darwin.md: Ditto.
* config/rs6000/rs6000.md: Ditto.
* config/rs6000/vector.md: Ditto.
* config/rs6000/vsx.md: Ditto.
Martin Liska [Wed, 21 Feb 2018 14:05:45 +0000 (15:05 +0100)]
Add "native" as a valid option value for -march= on aarch64 (PR driver/83193).
2018-02-21 Martin Liska <mliska@suse.cz>
PR driver/83193
* config/aarch64/aarch64.c (aarch64_print_hint_for_core_or_arch):
Add "native" as a possible value.
* config/aarch64/aarch64.h (HAVE_LOCAL_CPU_DETECT): Define
the macro when native cpu detection is available.
Martin Liska [Wed, 21 Feb 2018 14:05:17 +0000 (15:05 +0100)]
Add "native" as a valid option value for -march= on arm (PR driver/83193).
2018-02-21 Martin Liska <mliska@suse.cz>
PR driver/83193
* common/config/arm/arm-common.c (arm_print_hint_for_arch_option):
Add "native" as a possible value.
* config/arm/arm.h (HAVE_LOCAL_CPU_DETECT): Define the macro
when native cpu detection is available.
Jakub Jelinek [Wed, 21 Feb 2018 10:35:46 +0000 (11:35 +0100)]
re PR tree-optimization/84478 (pdftex miscompilation on i386)
PR tree-optimization/84478
* gimple-fold.h (get_range_strlen): Add a bool argument defaulted to
false.
* gimple-fold.c (get_range_strlen): Make minlen const and assume it
can't be NULL. Change FUZZY from bool to int, for 1 add PHI/COND_EXPR
support which is conservatively correct, for 2 only stay conservative
for maxlen. Formatting and comment capitalization fixes. Add STRICT
argument to the 2 argument get_range_strlen, adjust 6 arg
get_range_strlen caller and clear minmaxlen[0] and [1] if it returned
false.
(get_maxval_strlen): Adjust 6 arg get_range_strlen caller.
(gimple_fold_builtin_strlen): Pass true as last argument to
get_range_strlen.
* gcc.c-torture/execute/pr84478.c: New test.
Co-Authored-By: Martin Sebor <msebor@redhat.com>
From-SVN: r257870
Martin Sebor [Tue, 20 Feb 2018 20:22:01 +0000 (20:22 +0000)]
PR middle-end/84095 - false-positive -Wrestrict warnings for memcpy within array
gcc/ChangeLog:
PR middle-end/84095
* gimple-ssa-warn-restrict.c (builtin_memref::extend_offset_range): New.
(builtin_memref::set_base_and_offset): Same. Handle inner references.
(builtin_memref::builtin_memref): Factor out parts into
set_base_and_offset and call it.
gcc/testsuite/ChangeLog:
PR middle-end/84095
* c-c++-common/Warray-bounds-3.c: Adjust text of expected warnings.
* c-c++-common/Wrestrict.c: Same.
* gcc.dg/Wrestrict-6.c: Same.
* gcc.dg/Warray-bounds-27.c: New test.
* gcc.dg/Wrestrict-8.c: New test.
* gcc.dg/Wrestrict-9.c: New test.
* gcc.dg/pr84095.c: New test.