When building gcc with some specific LDFLAGS_FOR_TARGET, e.g.
LDFLAGS_FOR_TARGET=-Wl,-z,relro,-z,now
those flags propagate info linking of target shared libraries,
e.g. lib{ubsan,tsan,stdc++,quadmath,objc,lsan,itm,gphobos,gdruntime,gomp,go,gfortran,atomic,asan}.so.*
but there is one important exception, libgcc_s.so.* linking ignores it.
The following patch fixes that.
Bootstrapped/regtested on x86_64-linux with LDFLAGS_FOR_TARGET=-Wl,-z,relro,-z,now
and verified that libgcc_s.so.* is BIND_NOW when it previously wasn't, and
without any LDFLAGS_FOR_TARGET on x86_64-linux and i686-linux.
There on x86_64-linux I've verified that the libgcc_s.so.1 linking command
line for -m64 is identical except for whitespace to one without the patch,
and for -m32 multilib $(LDFLAGS) actually do supply there an extra -m32
that also repeats later in the @multilib_flags@, which should be harmless.
2021-08-04 Jakub Jelinek <jakub@redhat.com>
* config/t-slibgcc (SHLIB_LINK): Add $(LDFLAGS).
* config/t-slibgcc-darwin (SHLIB_LINK): Likewise.
* config/t-slibgcc-vms (SHLIB_LINK): Likewise.
* config/t-slibgcc-fuchsia (SHLIB_LDFLAGS): Remove $(LDFLAGS).
This patch implements the omp_get_device_num library routine, specified in
OpenMP 5.0.
GOMP_DEVICE_NUM_VAR is a macro symbol which defines name of a "device number"
variable, is defined on the device-side libgomp, has it's address returned to
host-side libgomp during device initialization, and the host libgomp then
sets its value to the designated device number.
libgomp/ChangeLog:
* icv-device.c (omp_get_device_num): New API function, host side.
* fortran.c (omp_get_device_num_): New interface function.
* libgomp-plugin.h (GOMP_DEVICE_NUM_VAR): Define macro symbol.
* libgomp.map (OMP_5.0.2): New version space with omp_get_device_num,
omp_get_device_num_.
* libgomp.texi (omp_get_device_num): Add documentation for new API
function.
* omp.h.in (omp_get_device_num): Add declaration.
* omp_lib.f90.in (omp_get_device_num): Likewise.
* omp_lib.h.in (omp_get_device_num): Likewise.
* target.c (gomp_load_image_to_device): If additional entry for device
number exists at end of returned entries from 'load_image_func' hook,
copy the assigned device number over to the device variable.
* config/gcn/icv-device.c (GOMP_DEVICE_NUM_VAR): Define static global.
(omp_get_device_num): New API function, device side.
* plugin/plugin-gcn.c ("symcat.h"): Add include.
(GOMP_OFFLOAD_load_image): Add addresses of device GOMP_DEVICE_NUM_VAR
at end of returned 'target_table' entries.
* config/nvptx/icv-device.c (GOMP_DEVICE_NUM_VAR): Define static global.
(omp_get_device_num): New API function, device side.
* plugin/plugin-nvptx.c ("symcat.h"): Add include.
(GOMP_OFFLOAD_load_image): Add addresses of device GOMP_DEVICE_NUM_VAR
at end of returned 'target_table' entries.
* testsuite/lib/libgomp.exp
(check_effective_target_offload_target_intelmic): New function for
testing for intelmic offloading.
* testsuite/libgomp.c-c++-common/target-45.c: New test.
* testsuite/libgomp.fortran/target10.f90: New test.
This adds the [[nodiscard]] attribute to all conversion operators,
comparison operators, call operators and non-member functions in
<compare>. Nothing in this header except constructors has side effects.
Signed-off-by: Jonathan Wakely <jwakely@redhat.com>
libstdc++-v3/ChangeLog:
* libsupc++/compare (partial_ordering, weak_ordering)
(strong_ordering, is_eq, is_neq, is_lt, is_lteq, is_gt, is_gteq)
(compare_three_way, strong_order, weak_order, partial_order)
(compare_strong_order_fallback, compare_weak_order_fallback)
(compare_partial_order_fallback, __detail::__synth3way): Add
nodiscard attribute.
* testsuite/18_support/comparisons/categories/zero_neg.cc: Add
-Wno-unused-result to options.
As explained in the PR, the grammar in the Concepts TS means that a [
token following a requires-clause is parsed as part of the
logical-or-expression rather than the start of an attribute. That makes
the following ill-formed when using -fconcepts-ts:
template<typename T> requires foo<T> [[nodiscard]] int f(T);
This change moves all attributes that follow a requires-clause to the
end of the function declarator.
Signed-off-by: Jonathan Wakely <jwakely@redhat.com>
libstdc++-v3/ChangeLog:
PR libstdc++/101782
* include/bits/ranges_base.h (ranges::begin, ranges::end)
(ranges::rbegin, ranges::rend, ranges::size, ranges::ssize)
(ranges::empty, ranges::data): Move attribute to the end of
the declarator.
* include/bits/stl_iterator.h (__gnu_cxx::__normal_iterator)
(common_iterator): Likewise for non-member operator functions.
* include/std/ranges (views::all, views::filter)
(views::transform, views::take, views::take_while, views::drop)
(views::drop_while, views::join, views::lazy_split)
(views::split, views::counted, views::common, views::reverse)
(views::elements): Likewise.
* testsuite/std/ranges/access/101782.cc: New test.
1. Intrinsics in <x86gprintrin.h> only require GPR ISAs. Add
#if defined __MMX__ || defined __SSE__
#pragma GCC push_options
#pragma GCC target("general-regs-only")
#define __DISABLE_GENERAL_REGS_ONLY__
#endif
and
#ifdef __DISABLE_GENERAL_REGS_ONLY__
#undef __DISABLE_GENERAL_REGS_ONLY__
#pragma GCC pop_options
#endif /* __DISABLE_GENERAL_REGS_ONLY__ */
to <x86gprintrin.h> to disable non-GPR ISAs so that they can be used in
functions with __attribute__ ((target("general-regs-only"))).
2. When checking always_inline attribute, if callee only uses GPRs,
ignore MASK_80387 since enable MASK_80387 in caller has no impact on
callee inline.
gcc/
PR target/99744
* config/i386/i386.c (ix86_can_inline_p): Ignore MASK_80387 if
callee only uses GPRs.
* config/i386/ia32intrin.h: Revert commit 5463cee277.
* config/i386/serializeintrin.h: Revert commit 71958f740f.
* config/i386/x86gprintrin.h: Add
#pragma GCC target("general-regs-only") and #pragma GCC pop_options
to disable non-GPR ISAs.
gcc/testsuite/
PR target/99744
* gcc.target/i386/pr99744-3.c: New test.
* gcc.target/i386/pr99744-4.c: Likewise.
* gcc.target/i386/pr99744-5.c: Likewise.
* gcc.target/i386/pr99744-6.c: Likewise.
* gcc.target/i386/pr99744-7.c: Likewise.
* gcc.target/i386/pr99744-8.c: Likewise.
aarch64.c has various routines to test for specific kinds of
vector statement cost. The routines aren't really target-specific,
so following a suggestion from Richi, this patch moves them to a new
section of tree-vectorizer.h.
gcc/
* tree-vectorizer.h (vect_is_store_elt_extraction, vect_is_reduction)
(vect_reduc_type, vect_embedded_comparison_type, vect_comparison_type)
(vect_is_extending_load, vect_is_integer_truncation): New functions,
moved from aarch64.c but given different names.
* config/aarch64/aarch64.c (aarch64_is_store_elt_extraction)
(aarch64_is_reduction, aarch64_reduc_type)
(aarch64_embedded_comparison_type, aarch64_comparison_type)
(aarch64_extending_load_p, aarch64_integer_truncation_p): Delete
in favor of the above. Update callers accordingly.
A change to the way gas interprets the .fpu directive in binutils-2.34
means that issuing .fpu will clear any features set by .arch_extension
that apply to the floating point or simd units. This unfortunately
causes problems for more recent versions of the architecture because
we currently emit .arch, .arch_extension and .fpu directives at
different times and try to suppress redundant changes.
This change addresses this by firstly unifying all the places where we
emit these directives to a single block of code and secondly
(re)emitting all the directives if any changes have been made to the
target options. Whilst this is slightly more than the strict minimum
it should be enough to catch all cases where a change could have
happened. The new code also emits the directives in the order: .arch,
.fpu, .arch_extension. This ensures that the additional architectural
extensions are not removed by a later .fpu directive.
Whilst writing this patch I also noticed that in the corner case where
the last function to be compiled had a non-standard set of
architecture flags, the assembler would add an incorrect set of
derived attributes for the file as a whole. Instead of reflecting the
command-line options it would reflect the flags from the last file in
the function. To address this I've also added a call to re-emit the
flags from the asm_file_end callback so the assembler will be in the
correct state when it finishes processing the intput.
There's some slight churn to the testsuite as a consequence of this,
because previously we had a hack to suppress emitting a .fpu directive
for one specific case, but with the new order this is no-longer
necessary.
gcc/ChangeLog:
PR target/101723
* config/arm/arm-cpus.in (generic-armv7-a): Add quirk to suppress
writing .cpu directive in asm output.
* config/arm/arm.c (arm_identify_fpu_from_isa): New variable.
(arm_last_printed_arch_string): Delete.
(arm_last-printed_fpu_string): Delete.
(arm_configure_build_target): If use of floating-point/SIMD is
disabled, remove all fp/simd related features from the target ISA.
(last_arm_targ_options): New variable.
(arm_print_asm_arch_directives): Add new parameters. Change order
of emitted directives and handle all cases here.
(arm_file_start): Always call arm_print_asm_arch_directives, move
all generation of .arch/.arch_extension here.
(arm_file_end): Call arm_print_asm_arch.
(arm_declare_function_name): Call arm_print_asm_arch_directives
instead of printing .arch/.fpu directives directly.
gcc/testsuite/ChangeLog:
PR target/101723
* gcc.target/arm/cortex-m55-nofp-flag-hard.c: Update expected output.
* gcc.target/arm/cortex-m55-nofp-flag-softfp.c: Likewise.
* gcc.target/arm/cortex-m55-nofp-nomve-flag-softfp.c: Likewise.
* gcc.target/arm/mve/intrinsics/mve_fpu1.c: Convert to dg-do assemble.
Add a non-no-op function body.
* gcc.target/arm/mve/intrinsics/mve_fpu2.c: Likewise.
* gcc.target/arm/pr98636.c (dg-options): Add -mfloat-abi=softfp.
* gcc.target/arm/attr-neon.c: Tighten scan-assembler tests.
* gcc.target/arm/attr-neon2.c: Use -Ofast, convert test to use
check-function-bodies.
* gcc.target/arm/attr-neon3.c: Likewise.
* gcc.target/arm/pr69245.c: Tighten scan-assembler match, but allow
multiple instances.
* gcc.target/arm/pragma_fpu_attribute.c: Likewise.
* gcc.target/arm/pragma_fpu_attribute_2.c: Likewise.
arm_configure_build_target is usually used to reconfigure the
arm_active_target structure, which is then used to reconfigure a
number of other global variables describing the current target.
Occasionally, however, we need to use arm_configure_build_target to
construct a temporary target structure and in that case it is wrong to
try to reconfigure the global variables (although probably harmless,
since arm_option_reconfigure_globals() only looks at
arm_active_target). At the very least, however, this is wasted work,
so it is best not to do it unless needed. What's more, several
callers of arm_configure_build target call
arm_option_reconfigure_globals themselves within a few lines, making
the call from within arm_configure_build_target completely redundant.
So this patch moves the responsibility of calling of
arm_configure_build_target to its callers (only two places needed
updating).
gcc:
* config/arm/arm.c (arm_configure_build_target): Don't call
arm_option_reconfigure_globals.
(arm_option_restore): Call arm_option_reconfigure_globals after
reconfiguring the target.
* config/arm/arm-c.c (arm_pragma_target_parse): Likewise.
This should never happen now if GCC is invoked by the driver, but in
the unusual case of calling cc1 (or its ilk) directly from the command
line the build target's arch_name string can remain NULL. This can
complicate later processing meaning that we need to check for this
case explicitly in some circumstances. Nothing should rely on this
behaviour, so it's simpler to always set the arch_name when
configuring the build target and be done with it.
gcc:
* config/arm/arm.c (arm_configure_build_target): Ensure the target's
arch_name is always set.
The Neon subtract-long/subract-widen instructions can select the top
or bottom half of the operand registers. This selection does not
change the cost of the underlying instruction and this should be
reflected by the RTL cost function.
This patch adds RTL tree traversal in the Neon subtract cost function
to match vec_select high-half of its operands. This traversal
prevents the cost of the vec_select from being added into the cost of
the subtract - meaning that these instructions can now be emitted in
the combine pass as they are no longer deemed prohibitively
expensive.
gcc/ChangeLog:
2021-07-28 Jonathan Wright <jonathan.wright@arm.com>
* config/aarch64/aarch64.c: Traverse RTL tree to prevent cost
of vec_select high-half from being added into Neon subtract
cost.
gcc/testsuite/ChangeLog:
* gcc.target/aarch64/vsubX_high_cost.c: New test.
The Neon add-long/add-widen instructions can select the top or bottom
half of the operand registers. This selection does not change the
cost of the underlying instruction and this should be reflected by
the RTL cost function.
This patch adds RTL tree traversal in the Neon add cost function to
match vec_select high-half of its operands. This traversal prevents
the cost of the vec_select from being added into the cost of the
subtract - meaning that these instructions can now be emitted in the
combine pass as they are no longer deemed prohibitively expensive.
gcc/ChangeLog:
2021-07-28 Jonathan Wright <jonathan.wright@arm.com>
* config/aarch64/aarch64.c: Traverse RTL tree to prevent cost
of vec_select high-half from being added into Neon add cost.
gcc/testsuite/ChangeLog:
* gcc.target/aarch64/vaddX_high_cost.c: New test.
This adjusts the testcase for excess diagnostics emitted by some
targets because of the attribute simd usage like
warning: GCC does not currently support mixed size types for 'simd' functions
on aarch64.
2021-08-05 Richard Biener <rguenther@suse.de>
* gcc.dg/vect/bb-slp-pr101756.c: Add -w.
This patch follows Richi's suggestion to add one optional
argument class loop* root to loops_list's CTOR, it can
provide the ability to construct a visiting list starting
from the given class loop* ROOT rather than the default
tree_root of loops_for_fn (FN), for visiting a subset of
the loop tree.
It unifies all orders of walkings into walk_loop_tree, but
it still uses linear search for LI_ONLY_INNERMOST when
looking at the whole loop tree since it has a more stable
bound.
gcc/ChangeLog:
* cfgloop.h (loops_list::loops_list): Add one optional argument
root and adjust accordingly, update loop tree walking and factor
out to ...
* cfgloop.c (loops_list::walk_loop_tree): ... this. New function.
The scalar storage order does not apply to pointer and vector components.
gcc/
PR tree-optimization/101626
* tree-sra.c (propagate_subaccesses_from_rhs): Do not set the
reverse scalar storage order on a pointer or vector component.
gcc/testsuite/
* gcc.dg/sso-15.c: New test.
In the places where we handle builtin functions, list all
supported ones, and fail if an unexpected one is seen. So if a
new builtin function is added in the future we can detect it,
instead of silently treating it as nonescaping.
Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/339992
gcc/ChangeLog:
* config/i386/sse.md (cond_<code><mode>): New expander.
gcc/testsuite/ChangeLog:
* gcc.target/i386/cond_op_anylogic_d-1.c: New test.
* gcc.target/i386/cond_op_anylogic_d-2.c: New test.
* gcc.target/i386/cond_op_anylogic_q-1.c: New test.
* gcc.target/i386/cond_op_anylogic_q-2.c: New test.
gcc/ChangeLog:
* config/i386/sse.md (cond_<code><mode>): New expander.
gcc/testsuite/ChangeLog:
* gcc.target/i386/cond_op_maxmin_double-1.c: New test.
* gcc.target/i386/cond_op_maxmin_double-2.c: New test.
* gcc.target/i386/cond_op_maxmin_float-1.c: New test.
* gcc.target/i386/cond_op_maxmin_float-2.c: New test.
gcc/ChangeLog:
* config/i386/sse.md (cond_<code><mode>): New expander.
gcc/testsuite/ChangeLog:
* gcc.target/i386/cond_op_maxmin_b-1.c: New test.
* gcc.target/i386/cond_op_maxmin_b-2.c: New test.
* gcc.target/i386/cond_op_maxmin_d-1.c: New test.
* gcc.target/i386/cond_op_maxmin_d-2.c: New test.
* gcc.target/i386/cond_op_maxmin_q-1.c: New test.
* gcc.target/i386/cond_op_maxmin_q-2.c: New test.
* gcc.target/i386/cond_op_maxmin_ub-1.c: New test.
* gcc.target/i386/cond_op_maxmin_ub-2.c: New test.
* gcc.target/i386/cond_op_maxmin_ud-1.c: New test.
* gcc.target/i386/cond_op_maxmin_ud-2.c: New test.
* gcc.target/i386/cond_op_maxmin_uq-1.c: New test.
* gcc.target/i386/cond_op_maxmin_uq-2.c: New test.
* gcc.target/i386/cond_op_maxmin_uw-1.c: New test.
* gcc.target/i386/cond_op_maxmin_uw-2.c: New test.
* gcc.target/i386/cond_op_maxmin_w-1.c: New test.
* gcc.target/i386/cond_op_maxmin_w-2.c: New test.
Update STORE_MAX_PIECES to allow 16/32/64 bytes only if inter-unit move
is enabled since vec_duplicate enabled by inter-unit move is used to
implement store_by_pieces of 16/32/64 bytes.
gcc/
PR target/101742
* config/i386/i386.h (STORE_MAX_PIECES): Allow 16/32/64 bytes
only if TARGET_INTER_UNIT_MOVES_TO_VEC is true.
gcc/testsuite/
PR target/101742
* gcc.target/i386/pr101742a.c: New test.
* gcc.target/i386/pr101742b.c: Likewise.
To avoid stack realignment, call ix86_gen_scratch_sse_rtx to get a
scratch SSE register to copy data with with SSE register from one
memory location to another.
gcc/
PR target/101772
* config/i386/i386-expand.c (ix86_expand_vector_move): Call
ix86_gen_scratch_sse_rtx to get a scratch SSE register to copy
data with SSE register from one memory location to another.
gcc/testsuite/
PR target/101772
* gcc.target/i386/eh_return-2.c: New test.
This patch makes use of the vector permute double immediate
instruction for constant permute vectors.
gcc/ChangeLog:
* config/s390/s390.c (expand_perm_with_vpdi): New function.
(vectorize_vec_perm_const_1): Call expand_perm_with_vpdi.
* config/s390/vector.md (*vpdi1<mode>, @vpdi1<mode>): Enable a
parameterized expander.
(*vpdi4<mode>, @vpdi4<mode>): Likewise.
gcc/testsuite/ChangeLog:
* gcc.target/s390/vector/perm-vpdi.c: New test.
This patch implements the TARGET_VECTORIZE_VEC_PERM_CONST in the IBM Z
backend. The initial implementation only exploits the vector merge
instruction but there is more to come.
gcc/ChangeLog:
* config/s390/s390.c (MAX_VECT_LEN): Define macro.
(struct expand_vec_perm_d): Define struct.
(expand_perm_with_merge): New function.
(vectorize_vec_perm_const_1): New function.
(s390_vectorize_vec_perm_const): New function.
(TARGET_VECTORIZE_VEC_PERM_CONST): Define target macro.
gcc/testsuite/ChangeLog:
* gcc.target/s390/vector/perm-merge.c: New test.
* gcc.target/s390/vector/vec-types.h: New test.
The patch gets rid of the unspec used for the vector permute double
immediate instruction and replaces it with generic rtx.
gcc/ChangeLog:
* config/s390/s390.md (UNSPEC_VEC_PERMI): Remove constant
definition.
* config/s390/vector.md (*vpdi1<mode>, *vpdi4<mode>): New pattern
definitions.
* config/s390/vx-builtins.md (*vec_permi<mode>): Emit generic rtx
instead of an unspec.
gcc/testsuite/ChangeLog:
* gcc.target/s390/zvector/vec-permi.c: Removed.
* gcc.target/s390/zvector/vec_permi.c: New test.
This patch gets rid of the unspecs we were using for the vector merge
instruction and replaces it with generic rtx.
gcc/ChangeLog:
* config/s390/s390-modes.def: Add more vector modes to support
concatenation of two vectors.
* config/s390/s390-protos.h (s390_expand_merge_perm_const): Add
prototype.
(s390_expand_merge): Likewise.
* config/s390/s390.c (s390_expand_merge_perm_const): New function.
(s390_expand_merge): New function.
* config/s390/s390.md (UNSPEC_VEC_MERGEH, UNSPEC_VEC_MERGEL):
Remove constant definitions.
* config/s390/vector.md (V_HW_2): Add mode iterators.
(VI_HW_4, V_HW_4): Rename VI_HW_4 to V_HW_4.
(vec_2x_nelts, vec_2x_wide): New mode attributes.
(*vmrhb, *vmrlb, *vmrhh, *vmrlh, *vmrhf, *vmrlf, *vmrhg, *vmrlg):
New pattern definitions.
(vec_widen_umult_lo_<mode>, vec_widen_umult_hi_<mode>)
(vec_widen_smult_lo_<mode>, vec_widen_smult_hi_<mode>)
(vec_unpacks_lo_v4sf, vec_unpacks_hi_v4sf, vec_unpacks_lo_v2df)
(vec_unpacks_hi_v2df): Adjust expanders to emit non-unspec RTX for
vec merge.
* config/s390/vx-builtins.md (V_HW_4): Remove mode iterator. Now
in vector.md.
(vec_mergeh<mode>, vec_mergel<mode>): Use s390_expand_merge to
emit vec merge pattern.
gcc/testsuite/ChangeLog:
* gcc.target/s390/vector/long-double-asm-in-out-hard-fp-reg.c:
Instead of vpdi with 0 and 5 vmrlg and vmrhg are used now.
* gcc.target/s390/vector/long-double-asm-inout-hard-fp-reg.c: Likewise.
* gcc.target/s390/zvector/vec-types.h: New test.
* gcc.target/s390/zvector/vec_merge.c: New test.
The Neon multiply/multiply-accumulate/multiply-subtract instructions
can select the top or bottom half of the operand registers. This
selection does not change the cost of the underlying instruction and
this should be reflected by the RTL cost function.
This patch adds RTL tree traversal in the Neon multiply cost function
to match vec_select high-half of its operands. This traversal
prevents the cost of the vec_select from being added into the cost of
the multiply - meaning that these instructions can now be emitted in
the combine pass as they are no longer deemed prohibitively
expensive.
gcc/ChangeLog:
2021-07-19 Jonathan Wright <jonathan.wright@arm.com>
* config/aarch64/aarch64.c (aarch64_strip_extend_vec_half):
Define.
(aarch64_rtx_mult_cost): Traverse RTL tree to prevent cost of
vec_select high-half from being added into Neon multiply
cost.
* rtlanal.c (vec_series_highpart_p): Define.
* rtlanal.h (vec_series_highpart_p): Declare.
gcc/testsuite/ChangeLog:
* gcc.target/aarch64/vmul_high_cost.c: New test.
The Neon multiply/multiply-accumulate/multiply-subtract instructions
can take various forms - multiplying full vector registers of values
or multiplying one vector by a single element of another. Regardless
of the form used, these instructions have the same cost, and this
should be reflected by the RTL cost function.
This patch adds RTL tree traversal in the Neon multiply cost function
to match the vec_select used by the lane-referencing forms of the
instructions already mentioned. This traversal prevents the cost of
the vec_select from being added into the cost of the multiply -
meaning that these instructions can now be emitted in the combine
pass as they are no longer deemed prohibitively expensive.
gcc/ChangeLog:
2021-07-19 Jonathan Wright <jonathan.wright@arm.com>
* config/aarch64/aarch64.c (aarch64_strip_duplicate_vec_elt):
Define.
(aarch64_rtx_mult_cost): Traverse RTL tree to prevent
vec_select cost from being added into Neon multiply cost.
gcc/testsuite/ChangeLog:
* gcc.target/aarch64/vmul_element_cost.c: New test.
This patch uses a more accurate scalar iteration estimate when
comparing the epilogue of a constant-iteration loop with a candidate
replacement epilogue.
In the testcase, the patch prevents a 1-to-3-element SVE epilogue
from seeming better than a 64-bit Advanced SIMD epilogue.
gcc/
* tree-vect-loop.c (vect_better_loop_vinfo_p): Detect cases in
which old_loop_vinfo is an epilogue loop that handles a constant
number of iterations.
gcc/testsuite/
* gcc.target/aarch64/sve/cost_model_12.c: New test.
After vect_analyze_loop has successfully analysed a loop for
one base vector mode B1, it considers using following base vector
modes to vectorise an epilogue. However, for VECT_COMPARE_COSTS,
a later mode B2 might turn out to be better than B1 was. Initially
this comparison will be between an epilogue loop (for B2) and a main
loop (for B1). However, in r11-6458 I'd added code to reanalyse the
B2 epilogue loop as a main loop, partly for correctness and partly
for better costing.
This can lead to a situation in which we think that the B2 epilogue
loop was better than the B1 main loop, but that the B2 main loop is
not better than the B1 main loop. There was no dump message to say
that this had happened, which made it look like B2 had still won.
gcc/
* tree-vect-loop.c (vect_analyze_loop): Print a dump message
when a reanalyzed loop fails to be cheaper than the current
main loop.
Ignored functions decls that are compiled at the start of
the assembly have bogus line numbers until the first .file
directive, as reported in PR101575.
The corresponding binutils bug report is
https://sourceware.org/bugzilla/show_bug.cgi?id=28149
The work around for this issue is to emit a dummy .file
directive before the first function is compiled, unless
another .file directive was already emitted previously.
2021-08-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
PR ada/101575
* dwarf2out.c (dwarf2out_assembly_start): Emit a dummy
.file statement when needed.
I believe PR101750 to be a testism. Fix it by giving the class a name.
gcc/testsuite/ChangeLog:
PR tree-optimization/101750
* g++.dg/vect/pr99149.cc: Name class.
This adds a gather vectorization capability to the vectorizer
without target support by decomposing the offset vector, doing
sclar loads and then building a vector from the result. This
is aimed mainly at cases where vectorizing the rest of the loop
offsets the cost of vectorizing the gather.
Note it's difficult to avoid vectorizing the offset load, but in
some cases later passes can turn the vector load + extract into
scalar loads, see the followup patch.
On SPEC CPU 2017 510.parest_r this improves runtime from 250s
to 219s on a Zen2 CPU which has its native gather instructions
disabled (using those the runtime instead increases to 254s)
using -Ofast -march=znver2 [-flto]. It turns out the critical
loops in this benchmark all perform gather operations.
2021-07-30 Richard Biener <rguenther@suse.de>
* tree-vect-data-refs.c (vect_check_gather_scatter):
Include widening conversions only when the result is
still handed by native gather or the current offset
size not already matches the data size.
Also succeed analysis in case there's no native support,
noted by a IFN_LAST ifn and a NULL decl.
(vect_analyze_data_refs): Always consider gathers.
* tree-vect-patterns.c (vect_recog_gather_scatter_pattern):
Test for no IFN gather rather than decl gather.
* tree-vect-stmts.c (vect_model_load_cost): Pass in the
gather-scatter info and cost emulated gathers accordingly.
(vect_truncate_gather_scatter_offset): Properly test for
no IFN gather.
(vect_use_strided_gather_scatters_p): Likewise.
(get_load_store_type): Handle emulated gathers and its
restrictions.
(vectorizable_load): Likewise. Emulate them by extracting
scalar offsets, doing scalar loads and a vector construct.
* gcc.target/i386/vect-gather-1.c: New testcase.
* gfortran.dg/vect/vect-8.f90: Adjust.
Pass MAX_PIECES to op_by_pieces_d::op_by_pieces_d for move, store and
compare.
PR target/101742
* expr.c (op_by_pieces_d::op_by_pieces_d): Add a max_pieces
argument to set m_max_size.
(move_by_pieces_d): Pass MOVE_MAX_PIECES to op_by_pieces_d.
(store_by_pieces_d): Pass STORE_MAX_PIECES to op_by_pieces_d.
(compare_by_pieces_d): Pass COMPARE_MAX_PIECES to op_by_pieces_d.
The easiest way to motivate these additions to match.pd is with the
following example:
unsigned int foo(unsigned char i) {
return i | (i<<8) | (i<<16) | (i<<24);
}
which mainline with -O2 on x86_64 currently generates:
foo: movzbl %dil, %edi
movl %edi, %eax
movl %edi, %edx
sall $8, %eax
sall $16, %edx
orl %edx, %eax
orl %edi, %eax
sall $24, %edi
orl %edi, %eax
ret
but with this patch now becomes:
foo: movzbl %dil, %eax
imull $16843009, %eax, %eax
ret
Interestingly, this transformation is already applied when using
addition, allowing synth_mult to select an optimal sequence, but
not when using the equivalent bit-wise ior or xor operators.
The solution is to use tree_nonzero_bits to check that the
potentially non-zero bits of each operand don't overlap, which
ensures that BIT_IOR_EXPR and BIT_XOR_EXPR produce the same
results as PLUS_EXPR, which effectively generalizes the old
fold_plusminus_mult_expr. Technically, the transformation
is to canonicalize (X*C1)|(X*C2) and (X*C1)^(X*C2) to
X*(C1+C2) where X and X<<C are considered special cases.
2021-08-04 Roger Sayle <roger@nextmovesoftware.com>
Marc Glisse <marc.glisse@inria.fr>
gcc/ChangeLog
* match.pd (bit_ior, bit_xor): Canonicalize (X*C1)|(X*C2) and
(X*C1)^(X*C2) as X*(C1+C2), and related variants, using
tree_nonzero_bits to ensure that operands are bit-wise disjoint.
gcc/testsuite/ChangeLog
* gcc.dg/fold-ior-4.c: New test.
This teaches forwprop to rewrite more vector loads that are only
used in BIT_FIELD_REFs as scalar loads. This provides the
remaining uplift to SPEC CPU 2017 510.parest_r on Zen 2 which
has CPU gathers disabled.
In particular vector load + vec_unpack + bit-field-ref is turned
into (extending) scalar loads which avoids costly XMM/GPR
transitions. To not conflict with vector load + bit-field-ref
+ vector constructor matching to vector load + shuffle the
extended transform is only done after vector lowering.
2021-07-30 Richard Biener <rguenther@suse.de>
* tree-ssa-forwprop.c (pass_forwprop::execute): Split
out code to decompose vector loads ...
(optimize_vector_load): ... here. Generalize it to
handle intermediate widening and TARGET_MEM_REF loads
and apply it to loads with a supported vector mode as well.
The following avoids vectorizing MIN/MAX reductions on bools which,
when ending up as vector(2) <signed-boolean:64> would need to be
adjusted because of the sign change. The fix instead avoids any
reduction vectorization where the result isn't compatible
to the original scalar type since we don't compensate for that
either.
2021-08-04 Richard Biener <rguenther@suse.de>
PR tree-optimization/101756
* tree-vect-slp.c (vectorizable_bb_reduc_epilogue): Make sure
the result of the reduction epilogue is compatible to the original
scalar result.
* gcc.dg/vect/bb-slp-pr101756.c: New testcase.
When parsing default arguments, we need to temporarily clear parser->omp_declare_simd
and parser->oacc_routine, otherwise it can clash with further declarations
inside of e.g. lambdas inside of those default arguments.
2021-08-04 Jakub Jelinek <jakub@redhat.com>
PR c++/101759
* parser.c (cp_parser_default_argument): Temporarily override
parser->omp_declare_simd and parser->oacc_routine to NULL.
* g++.dg/gomp/pr101759.C: New test.
* g++.dg/goacc/pr101759.C: New test.
The file has two identical halves, seems like twice applied patch.
2021-08-04 Jakub Jelinek <jakub@redhat.com>
* gcc.c-torture/execute/ieee/pr29302-1.x: Undo doubly applied patch.
The define_peephole2 which is added by r12-2640-gf7bf03cf69ccb7dc
should only work on general registers, considering that x86 also
supports mov instructions between gpr, sse reg, mask reg, limiting the
peephole2 predicate to general_reg_operand.
gcc/ChangeLog:
PR target/101743
* config/i386/i386.md (peephole2): Refine predicate from
register_operand to general_reg_operand.
The file has two identical halves, seems like twice applied patch.
2021-08-04 Jakub Jelinek <jakub@redhat.com>
* config/t-slibgcc-fuchsia: Undo doubly applied patch.
This makes tail recursion optimization produce a loop structure
manually rather than relying on loop fixup. That also allows the
loop to be marked as finite (it would eventually blow the stack
if it were not).
2021-08-04 Richard Biener <rguenther@suse.de>
PR tree-optimization/101769
* tree-tailcall.c (eliminate_tail_call): Add the created loop
for the first recursion and return it via the new output parameter.
(optimize_tail_call): Pass through new output param.
(tree_optimize_tail_calls_1): After creating all latches,
add the created loop to the loop tree. Do not mark loops for fixup.
* g++.dg/tree-ssa/pr101769.C: New testcase.