__attribute__ ((noipa)) void
f_vnx128qi (int8_t a, int8_t b, int8_t c, int8_t d, int8_t e, int8_t f,
int8_t g, int8_t h, int8_t *out)
{
vnx128qi v
= {a, b, c, d, e, f, g, h, a, b, c, d, e, f, g, h,
a, b, c, d, e, f, g, h, a, b, c, d, e, f, g, h,
a, b, c, d, e, f, g, h, a, b, c, d, e, f, g, h,
a, b, c, d, e, f, g, h, a, b, c, d, e, f, g, h,
a, b, c, d, e, f, g, h, a, b, c, d, e, f, g, h,
a, b, c, d, e, f, g, h, a, b, c, d, e, f, g, h,
a, b, c, d, e, f, g, h, a, b, c, d, e, f, g, h,
a, b, c, d, e, f, g, h, a, b, c, d, e, f, g, h};
*(vnx128qi *) out = v;
}
This patch codegen:
f_vnx128qi:
andi a1,a1,0xff
andi a0,a0,0xff
slli a1,a1,8
andi a2,a2,0xff
or a1,a1,a0
slli a2,a2,16
andi a3,a3,0xff
or a2,a2,a1
slli a3,a3,24
andi a4,a4,0xff
or a3,a3,a2
slli a4,a4,32
andi a5,a5,0xff
or a4,a4,a3
slli a5,a5,40
andi a6,a6,0xff
or a5,a5,a4
slli a6,a6,48
or a6,a6,a5
vsetvli a5,zero,e64,m8,ta,ma
ld a5,0(sp)
slli a7,a7,56
or a7,a7,a6
vmv.v.x v8,a7
vs8r.v v8,0(a5)
ret
We support more optimizations cases in the future. But they are not
included in this patch.
* config/riscv/autovec.md (vec_init<mode><vel>): New pattern.
* config/riscv/riscv-protos.h (expand_vec_init): New function.
* config/riscv/riscv-v.cc (class rvv_builder): New class.
(rvv_builder::can_duplicate_repeating_sequence_p): New function.
(rvv_builder::get_merged_repeating_sequence): Ditto.
(expand_vector_init_insert_elems): Ditto.
(expand_vec_init): Ditto.
* config/riscv/vector-iterators.md: New attribute.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/rvv.exp:
* gcc.target/riscv/rvv/autovec/vls-vlmax/insert-1.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/insert-2.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/insert-3.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/insert_run-1.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/insert_run-2.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat-1.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat-2.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat-3.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat-4.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat-5.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat-6.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat_run-1.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat_run-2.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat_run-3.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat_run-4.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat_run-5.c: New test.
* gcc.target/riscv/rvv/autovec/vls-vlmax/repeat_run-6.c: New test.
Haochen Gui [Thu, 11 May 2023 02:10:50 +0000 (10:10 +0800)]
rs6000: Change mode and insn condition for scalar insert exp instruction
gcc/
* config/rs6000/rs6000-builtins.def
(__builtin_vsx_scalar_insert_exp): Replace bif-pattern from xsiexpdp
to xsiexpdp_di.
(__builtin_vsx_scalar_insert_exp_dp): Replace bif-pattern from
xsiexpdpf to xsiexpdpf_di.
* config/rs6000/vsx.md (xsiexpdp): Rename to...
(xsiexpdp_<mode>): ..., set the mode of second operand to GPR and
replace TARGET_64BIT with TARGET_POWERPC64.
(xsiexpdpf): Rename to...
(xsiexpdpf_<mode>): ..., set the mode of second operand to GPR and
replace TARGET_64BIT with TARGET_POWERPC64.
Haochen Gui [Thu, 11 May 2023 02:08:57 +0000 (10:08 +0800)]
rs6000: Change mode and insn condition for scalar extract sig instruction
gcc/
* config/rs6000/rs6000-builtins.def
(__builtin_vsx_scalar_extract_sig): Set return type to const signed
long long.
* config/rs6000/vsx.md (xsxsigdp): Replace TARGET_64BIT with
TARGET_POWERPC64.
Haochen Gui [Thu, 11 May 2023 02:07:01 +0000 (10:07 +0800)]
rs6000: Change mode and insn condition for scalar extract exp instruction
gcc/
* config/rs6000/rs6000-builtins.def
(__builtin_vsx_scalar_extract_exp): Set return type to const signed
int and set its bif-pattern to xsxexpdp_si, move it from power9-64
to power9 catalog.
* config/rs6000/vsx.md (xsxexpdp): Rename to ...
(xsxexpdp_<mode>): ..., set mode of operand 0 to GPR and remove
TARGET_64BIT check.
* doc/extend.texi (scalar_extract_exp): Remove 64-bit environment
requirement when it has a 64-bit argument.
gcc/testsuite/
* gcc.target/powerpc/bfp/scalar-extract-exp-0.c: Remove lp64 check.
* gcc.target/powerpc/bfp/scalar-extract-exp-1.c: Likewise.
* gcc.target/powerpc/bfp/scalar-extract-exp-2.c: Delete as the case
is invalid now.
* gcc.target/powerpc/bfp/scalar-extract-exp-6.c: Remove lp64 check.
Pan Li [Thu, 11 May 2023 06:12:51 +0000 (14:12 +0800)]
Var-Tracking: Typedef pointer_mux<tree_node, rtx_def> as decl_or_value
The decl_or_value is defined as void * before this PATCH. It will take
care of both the tree_node and rtx_def. Unfortunately, given a void
pointer cannot tell the input is tree_node or rtx_def.
Then we have some implicit structure layout requirement similar as
below. Or we will touch unreasonable bits when cast void * to tree_node
or rtx_def.
This behavior blocks the PATCH that extend the rtx_def mode from 8 to
16 bits for running out of machine mode. This PATCH introduced the
pointer_mux to tell the input is tree_node or rtx_def, and decouple
the above implicit dependency.
Signed-off-by: Pan Li <pan2.li@intel.com> Co-Authored-By: Richard Sandiford <richard.sandiford@arm.com> Co-Authored-By: Richard Biener <rguenther@suse.de> Co-Authored-By: Jakub Jelinek <jakub@redhat.com>
gcc/ChangeLog:
* mux-utils.h: Add overload operator == and != for pointer_mux.
* var-tracking.cc: Included mux-utils.h for pointer_tmux.
(decl_or_value): Changed from void * to pointer_mux<tree_node, rtx_def>.
(dv_is_decl_p): Reconciled to the new type, aka pointer_mux.
(dv_as_decl): Ditto.
(dv_as_opaque): Removed due to unnecessary.
(struct variable_hasher): Take decl_or_value as compare_type.
(variable_hasher::equal): Diito.
(dv_from_decl): Reconciled to the new type, aka pointer_mux.
(dv_from_value): Ditto.
(attrs_list_member): Ditto.
(vars_copy): Ditto.
(var_reg_decl_set): Ditto.
(var_reg_delete_and_set): Ditto.
(find_loc_in_1pdv): Ditto.
(canonicalize_values_star): Ditto.
(variable_post_merge_new_vals): Ditto.
(dump_onepart_variable_differences): Ditto.
(variable_different_p): Ditto.
(set_slot_part): Ditto.
(clobber_slot_part): Ditto.
(clobber_variable_part): Ditto.
Gaius Mulley [Thu, 11 May 2023 23:15:28 +0000 (00:15 +0100)]
PR modula2/109810 ICE fix when an array is assigned by a larger string
This patch fixes an ICE when an array variable is assigned with
a string which exceeds the array size. It improves the accuracy
of the virtual token used to indicate the error message.
gcc/m2/ChangeLog:
PR modula2/109810
* gm2-compiler/M2ALU.mod (ConvertConstToType): Use
PrepareCopyString in place of DoCopyString.
* gm2-compiler/M2GenGCC.def (DoCopyString): Rename to ...
(PrepareCopyString): ... this.
* gm2-compiler/M2GenGCC.mod (CodeStatement): Call CodeReturnValue
with a single parameter. Call CodeXIndr with a single parameter.
(CodeReturnValue): Remove parameters and replace with a single
quadno. Reimplement using PrepareCopyString. Issue error
if the string exceeds designator space.
(DoCopyString): Reimplement and rename to ...
(PrepareCopyString): ... this.
(CodeXIndr): Remove parameters and replace with a single
quadno. Reimplement using PrepareCopyString. Issue error
if the string exceeds designator space.
(CodeBecomes): Remove parameters and replace with a single
quadno. Reimplement using PrepareCopyString. Issue error
if the string exceeds designator space.
* gm2-compiler/M2Quads.def (BuildReturn): Rename parameter to
tokreturn.
* gm2-compiler/M2Quads.mod (BuildReturn): Rename parameter to
tokreturn. Rename tokno to tokcombined.
gcc/testsuite/ChangeLog:
PR modula2/109810
* gm2/pim/fail/highice.mod: New test.
Patrick Palka [Thu, 11 May 2023 20:31:33 +0000 (16:31 -0400)]
c++: 'mutable' subobject of constexpr variable [PR109745]
r13-2701-g7107ea6fb933f1 made us correctly accept during constexpr
evaluation 'mutable' member accesses of objects constructed during
that evaluation, while continuing to reject such accesses for constexpr
objects constructed outside of that evaluation, by considering the
CONSTRUCTOR_MUTABLE_POISON flag during cxx_eval_component_reference.
However, this flag is set only for the outermost CONSTRUCTOR of a
constexpr variable initializer, so if we're accessing a 'mutable' member
of a nested CONSTRUCTOR, the flag won't be set and we won't reject the
access. This can lead to us accepting invalid code, as in the first
testcase, or even wrong code generation due to our speculative constexpr
evaluation, as in the second and third testcase.
This patch fixes this by setting CONSTRUCTOR_MUTABLE_POISON recursively
rather than only on the outermost CONSTRUCTOR.
PR c++/109745
gcc/cp/ChangeLog:
* typeck2.cc (poison_mutable_constructors): Define.
(store_init_value): Use it instead of setting
CONSTRUCTOR_MUTABLE_POISON directly.
gcc/testsuite/ChangeLog:
* g++.dg/cpp0x/constexpr-mutable4.C: New test.
* g++.dg/cpp0x/constexpr-mutable5.C: New test.
* g++.dg/cpp1y/constexpr-mutable2.C: New test.
Jonathan Wakely [Thu, 4 May 2023 14:22:07 +0000 (15:22 +0100)]
libstdc++: Use RAII types in strtod-based std::from_chars implementation
This adds auto_locale and auto_ferounding types to use RAII for changing
and restoring the local and floating-point environment when using strtod
to implement std::from_chars.
The destructors for the RAII objects run slightly later than the
previous statements that restored the locale/fenv, but the differences
are just some trivial assignments and an isinf call.
Reviewed-by: Patrick Palka <ppalka@redhat.com>
libstdc++-v3/ChangeLog:
* src/c++17/floating_from_chars.cc [USE_STRTOD_FOR_FROM_CHARS]
(auto_locale, auto_ferounding): New class types.
(from_chars_impl): Use auto_locale and auto_ferounding.
I borked the logic in r13-4526-g5329e1a8e1480d so that the selected
partial specialization of hh_mm_ss::__subseconds might not be able to
represent the correct number of subseconds. This can result in a
truncated value being stored for the subseconds, e.g., 4755859375 gets
truncated to 460892079 because the correct value doesn't fit in
uint_least32_t.
Instead of checking whether the maximum value of the incoming duration
type can be represented, we would need to check whether that maximum value
can be represented after being converted to the correct precision type:
However, this can fail to compile, due to integer overflow in the
constexpr multiplications. Instead, we could limit the check to the case
where the incoming duration has the same period as the precision, where
no conversion is needed and so no overflow can happen. But that seems of
very limited value, as it would only benefit specializations like
hh_mm_ss<duration<int, std::pico>>, which can only represent a
time-of-day between -00:00:00.0215 and +00:00:00.0215 measured in
picoseconds!
Additionally, the hh_mm_ss::__subseconds partial specializations do not
have disjoint constraints, so that some hh_mm_ss specializations result
in ambiguities tying to match a __subseconds partial specialization.
The most practical fix is to just stop using the __fits variable
template in the constraints of the partial specializations. This fixes
the truncated values by not selecting an inappropriate partial
specialization, and fixes the ambiguous match by ensuring the
constraints are disjoint.
Fixing this changes the layout of some specializations, so is an ABI
change. It only affects specializations that have a small (less than
64-bit) representation type and either a very small period (e.g. like
the picosecond specialization above) or a non-power-of-ten period like
ratio<1, 1024>. For example both hh_mm_ss<duration<int, std::pico>> and
hh_mm_ss<duration<int, ratio<1, 1024>> are affected (increasing from 16
bytes to 24 on x86_64), but hh_mm_ss<duration<int, ratio<1, 1000>> and
hh_mm_ss<duration<long, ratio<1, 1024>> are not affected.
libstdc++-v3/ChangeLog:
PR libstdc++/109772
* include/std/chrono (hh_mm_ss::__fits): Remove variable
template.
(hh_mm_ss::__subseconds): Remove __fits from constraints.
* testsuite/std/time/hh_mm_ss/109772.cc: New test.
* testsuite/std/time/hh_mm_ss/1.cc: Adjust expected size for
hh_mm_ss<duration<int, std::pico>>.
Jonathan Wakely [Wed, 10 May 2023 20:30:10 +0000 (21:30 +0100)]
libstdc++: Enforce value_type consistency in strings and streams
P1463R1 made it ill-formed for allocator-aware containers (including
std::basic_string) to use an allocator that has a different value_type
from the container itself. We already enforce that for other containers
(since r8-4828-g866e4d3853ccc0), but not for std::basic_string. We
traditionally accepted it as an extension and rebound the allocator, so
this change only adds the enforcement for C++20 and later.
Similarly, P1148R0 made it ill-formed for strings and streams to use a
traits type that has an incorrect char_type. We already enforce that for
std::basic_string_view, so we just need to add it to std::basic_ios and
std::basic_string.
The assertion for the allocator's value_type caused some testsuite
regressions:
FAIL: 21_strings/basic_string/cons/char/deduction.cc (test for excess errors)
FAIL: 21_strings/basic_string/cons/wchar_t/deduction.cc (test for excess errors)
FAIL: 21_strings/basic_string/requirements/explicit_instantiation/debug.cc (test for excess errors)
FAIL: 21_strings/basic_string/requirements/explicit_instantiation/int.cc (test for excess errors)
The last two are testing the traditional extension that rebinds the
allocator, so need to be disabled for C++20.
The first two are similar to LWG 3076 where an incorrect constructor is
considered for CTAD. In this case, determining that it's not viable
requires instantiating std::basic_string<Iter, char_traits<Iter>, Alloc>
which then fails the new assertion, because Alloc::value_type is not the
same as Iter. This is only a problem because the size_type parameter of
the non-viable constructor is an alias for
_Alloc_traits_impl<A>::size_type which is a nested type, and so the
enclosing basic_string specialization needs to be instantiated. If we
remove the _Alloc_traits_impl wrapper that was added in r12-5413-g2d76292bd6719d, then the definition of size_type no longer
depends on basic_string, and we don't instantiate an invalid
specialization and don't fail the assertion. The work done by
_Alloc_traits_impl::allocate can be done in a _S_allocate function
instead, which is probably more efficient to compile anyway.
libstdc++-v3/ChangeLog:
* config/abi/pre/gnu.ver: Export basic_string::_S_allocate.
* include/bits/basic_ios.h: Add static assertion checking
traits_type::value_type.
* include/bits/basic_string.h: Likewise. Do not rebind
allocator, and add static assertion checking its value_type.
(basic_string::_Alloc_traits_impl): Remove class template.
(basic_string::_S_allocate): New static member function.
(basic_string::assign): Use _S_allocate.
* include/bits/basic_string.tcc (basic_string::_M_create)
(basic_string::reserve, basic_string::_M_replace): Likewise.
* testsuite/21_strings/basic_string/requirements/explicit_instantiation/debug.cc:
Disable for C++20 and later.
* testsuite/21_strings/basic_string/requirements/explicit_instantiation/int.cc:
Likweise.
mtsamis [Mon, 1 Aug 2022 12:11:02 +0000 (14:11 +0200)]
aarch64: convert vector shift + bitwise and + multiply to vector compare
When using SWAR (SIMD in a register) techniques a comparison operation
within such a register can be made by using a combination of shifts,
bitwise and and multiplication. If code using this scheme is
vectorized then there is potential to replace all these operations
with a single vector comparison, by reinterpreting the vector types to
match the width of the SWAR register.
For example, for the test function packed_cmp_16_32, the original
generated code is:
ldr q0, [x0]
add w1, w1, 1
ushr v0.4s, v0.4s, 15
and v0.16b, v0.16b, v2.16b
shl v1.4s, v0.4s, 16
sub v0.4s, v1.4s, v0.4s
str q0, [x0], 16
cmp w2, w1
bhi .L20
with this pattern the above can be optimized to:
ldr q0, [x0]
add w1, w1, 1
cmlt v0.8h, v0.8h, #0
str q0, [x0], 16
cmp w2, w1
bhi .L20
The effect is similar for x86-64.
Factorize vqdmladhq, vqdmladhxq, vqdmlsdhq, vqdmlsdhxq, vqrdmladhq,
vqrdmladhxq, vqrdmlsdhq, vqrdmlsdhxq builtins so that they use the
same parameterized names.
Andrew Pinski [Thu, 11 May 2023 14:59:06 +0000 (07:59 -0700)]
Improve simple_dce for phis that only used in itself
While I was looking at differences before and after r14-569-g21e2ef2dc25de3, I noticed that one phi node was
not being removed.
For an example, while compiling combine.cc, in expand_field_assignment,
we would remove `# pos_51 = PHI <pos_221(31), pos_51(30)>`
but we don't any more since pos_51 has more than zero users
but in this case it is only itself.
This patch improves simple_dce_from_worklist to detect that
case and now we able to remove this phi statement again.
OK? Bootstrapped and tested on x86_64-linux-gnu.
gcc/ChangeLog:
* tree-ssa-dce.cc (simple_dce_from_worklist): For ssa names
defined by a phi node with more than one uses, allow for the
only uses are in that same defining statement.
Robin Dapp [Wed, 26 Apr 2023 18:32:36 +0000 (20:32 +0200)]
RISC-V: Allow vector constants in riscv_const_insns.
This patch adds various vector constants to riscv_const_insns in order
for them to be properly recognized as immediate operands. This then
allows to emit vmv.v.i instructions via autovectorization.
* gcc.target/riscv/rvv/autovec/vmv-imm-rv32.c: New test.
* gcc.target/riscv/rvv/autovec/vmv-imm-rv64.c: New test.
* gcc.target/riscv/rvv/autovec/vmv-imm-template.h: New test.
* gcc.target/riscv/rvv/autovec/vmv-imm-run.c: New test.
Patrick Palka [Thu, 11 May 2023 14:04:25 +0000 (10:04 -0400)]
c++: converted lambda as template argument [PR83258, ...]
r8-1253-g3d2e25a240c711 removed the template argument linkage requirement
in convert_nontype_argument for C++17 (which r9-3836-g4be5c72cf3ea3e later
factored out into invalid_tparm_referent_p), but we need to also remove
the one in convert_nontype_argument_function for benefit of the first and
third testcase which we currently reject even in C++17/20 mode.
And in invalid_tparm_referent_p we're inadvertendly returning false for
the address of a lambda's static op() since it's DECL_ARTIFICIAL, which
currently causes us to reject the second (C++20) testcase. But this
DECL_ARTIFICIAL check seems to be relevant only for VAR_DECL, and in fact
this code path was originally reachable only for VAR_DECL until recently
(r13-6970-gb5e38b1c166357). So this patch restricts the check to VAR_DECL.
Co-authored-by: Jonathan Wakely <jwakely@redhat.com>
PR c++/83258
PR c++/80488
PR c++/97700
gcc/cp/ChangeLog:
* pt.cc (convert_nontype_argument_function): Remove linkage
requirement for C++17 and later.
(invalid_tparm_referent_p) <case ADDR_EXPR>: Restrict
DECL_ARTIFICIAL rejection test to VAR_DECL.
gcc/testsuite/ChangeLog:
* g++.dg/ext/visibility/anon8.C: Don't expect a "no linkage"
error for the template argument &B2:fn in C++17 mode.
* g++.dg/cpp0x/lambda/lambda-conv15.C: New test.
* g++.dg/cpp2a/nontype-class56.C: New test.
* g++.dg/template/function2.C: New test.
Alexandre Oliva [Thu, 11 May 2023 13:06:25 +0000 (10:06 -0300)]
[vxworks] [testsuite] [aarch64] use builtin in pred-not-gen-4.c
On vxworks, isunordered is defined as a macro that ultimately calls a
_Fpcomp function, that GCC doesn't recognize as a builtin, so it
can't optimize accordingly.
Use __builtin_isunordered instead to get the desired code for the
test.
for gcc/testsuite/ChangeLog
* gcc.target/aarch64/sve/pred-not-gen-4.c: Drop math.h
include, call builtin.
The VMSET simplification RVV integer comparision has merged already.
This patch would like to update the comments for the cases that the
define_split will act on.
Signed-off-by: Pan Li <pan2.li@intel.com>
gcc/ChangeLog:
* config/riscv/vector.md: Add comments for simplifying to vmset.
Robin Dapp [Thu, 11 May 2023 08:32:21 +0000 (10:32 +0200)]
RISC-V: Add autovectorization tests for binary integer operations.
This patchs adds scan as well as execution tests for vectorized
binary integer operations. The tests are not comprehensive as
the vector type promotions (vec_unpack, extend etc.) are not implemented
yet. Also, vmulh, vmulhu, and vmulhsu and others are still missing.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/shift-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/shift-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/shift-template.h: New test.
* gcc.target/riscv/rvv/autovec/shift-run.c: New test.
* gcc.target/riscv/rvv/autovec/shift-scalar-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/shift-scalar-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/shift-scalar-template.h: New test.
* gcc.target/riscv/rvv/autovec/shift-scalar-run.c: New test.
* gcc.target/riscv/rvv/autovec/vadd-run-template.h: New test.
* gcc.target/riscv/rvv/autovec/vadd-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vadd-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vadd-template.h: New test.
* gcc.target/riscv/rvv/autovec/vand-run.c: New test.
* gcc.target/riscv/rvv/autovec/vand-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vand-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vand-template.h: New test.
* gcc.target/riscv/rvv/autovec/vdiv-run.c: New test.
* gcc.target/riscv/rvv/autovec/vdiv-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vdiv-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vdiv-template.h: New test.
* gcc.target/riscv/rvv/autovec/vmax-run.c: New test.
* gcc.target/riscv/rvv/autovec/vmax-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vmax-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vmax-template.h: New test.
* gcc.target/riscv/rvv/autovec/vmin-run.c: New test.
* gcc.target/riscv/rvv/autovec/vmin-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vmin-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vmin-template.h: New test.
* gcc.target/riscv/rvv/autovec/vmul-run.c: New test.
* gcc.target/riscv/rvv/autovec/vmul-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vmul-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vmul-template.h: New test.
* gcc.target/riscv/rvv/autovec/vor-run.c: New test.
* gcc.target/riscv/rvv/autovec/vor-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vor-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vor-template.h: New test.
* gcc.target/riscv/rvv/autovec/vrem-run.c: New test.
* gcc.target/riscv/rvv/autovec/vrem-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vrem-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vrem-template.h: New test.
* gcc.target/riscv/rvv/autovec/vsub-run.c: New test.
* gcc.target/riscv/rvv/autovec/vsub-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vsub-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vsub-template.h: New test.
* gcc.target/riscv/rvv/autovec/vxor-run.c: New test.
* gcc.target/riscv/rvv/autovec/vxor-rv32gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vxor-rv64gcv.c: New test.
* gcc.target/riscv/rvv/autovec/vxor-template.h: New test.
Co-authored-by: Michael Collison <collison@rivosinc.com>
Robin Dapp [Wed, 10 May 2023 07:52:43 +0000 (09:52 +0200)]
RISC-V: Split off shift patterns for autovectorization.
This patch splits off the shift patterns of the binop patterns.
This is necessary as the scalar shifts require a Pmode operand
as shift count. To this end, a new iterator any_int_binop_no_shift
is introduced. At a later point when the binops are split up
further in commutative and non-commutative patterns (which both
do not include the shift patterns) we might not need this anymore.
Robin Dapp [Wed, 10 May 2023 07:52:37 +0000 (09:52 +0200)]
RISC-V: Clarify vlmax and length handling.
This patch tries to improve the wrappers that emit either vlmax or
non-vlmax operations. Now, emit_len_op can be used to
emit a regular operation. Depending on whether a length != NULL
is passed either no VLMAX flags are set or we emit a vsetvli and
set VLMAX flags. The patch also adds some comments that describes
some of the rationale of the current handling of vlmax/nonvlmax
operations.
gcc/ChangeLog:
* config/riscv/autovec.md: Use renamed functions.
* config/riscv/riscv-protos.h (emit_vlmax_op): Rename.
(emit_vlmax_reg_op): To this.
(emit_nonvlmax_op): Rename.
(emit_len_op): To this.
(emit_nonvlmax_binop): Rename.
(emit_len_binop): To this.
* config/riscv/riscv-v.cc (emit_pred_op): Add default parameter.
(emit_pred_binop): Remove vlmax_p.
(emit_vlmax_op): Rename.
(emit_vlmax_reg_op): To this.
(emit_nonvlmax_op): Rename.
(emit_len_op): To this.
(emit_nonvlmax_binop): Rename.
(emit_len_binop): To this.
(sew64_scalar_helper): Use renamed functions.
(expand_tuple_move): Use renamed functions.
* config/riscv/riscv.cc (vector_zero_call_used_regs): Use
renamed functions.
* config/riscv/vector.md: Use renamed functions.
Robin Dapp [Mon, 24 Apr 2023 06:27:52 +0000 (08:27 +0200)]
RISC-V: Add vectorized binops and insn_expander helpers.
This patch adds basic binary integer operations support. It is based
on Michael Collison's work and makes use of the existing helpers in
riscv-c.cc. It introduces emit_nonvlmax_binop which, in turn, uses
emit_pred_binop. Setting the destination as well as the mask and the
length are factored out into separate functions.
gcc/ChangeLog:
* config/riscv/autovec.md (<optab><mode>3): Add integer binops.
* config/riscv/riscv-protos.h (emit_nonvlmax_binop): Declare.
* config/riscv/riscv-v.cc (emit_pred_op): New function.
(set_expander_dest_and_mask): New function.
(emit_pred_binop): New function.
(emit_nonvlmax_binop): New function.
Co-authored-by: Michael Collison <collison@rivosinc.com>
Jonathan Wakely [Wed, 10 May 2023 11:20:58 +0000 (12:20 +0100)]
libstdc++: Fix std::abs(__float128) for -NaN and -0.0 [PR109758]
The current implementation of this non-standard overload of std::abs
incorrectly returns a negative value for negative NaNs and negative
zero, because x < 0 is false in both cases.
Use fabsl(long double) or fabsf128(_Float128) if those do the right
thing. Otherwise, use __builtin_signbit(x) instead of x < 0 to detect
negative inputs. This assumes that __builtin_signbit handles __float128
correctly, but that seems to be true for all of GCC, clang and icc.
libstdc++-v3/ChangeLog:
PR libstdc++/109758
* include/bits/std_abs.h (abs(__float128)): Handle negative NaN
and negative zero correctly.
* testsuite/26_numerics/headers/cmath/109758.cc: New test.
Xi Ruoyao [Fri, 24 Jun 2022 12:23:13 +0000 (20:23 +0800)]
fortran: use grep instead of fgrep
fgrep has been deprecated in favor of grep -F for a long time, and the
curren grep release (3.10) prints a warning of fgrep is used. Stop
using fgrep so we won't see the warning.
We can't hard code grep -F here or it may break build on hosts w/o GNU
grep. autoconf documentation contains a warning about this issue and
suggest to use AC_PROG_FGREP and $FGREP, but these are too overkill in
the specific case: there is no way "debian" could be interpreted as an
non-trivial regex, so we can use a plain grep here.
This is one part of the fix for PR109128, along with a corresponding
binutils's linker change. Without this patch, what happens in the
linker, when an unused object in a .a file has offload data, is that
elf_link_is_defined_archive_symbol calls bfd_link_plugin_object_p,
which ends up calling the plugin's claim_file_handler, which then
records the object as one with offload data. That is, the linker never
decides to use the object in the first place, but use of this _p
interface (called as part of trying to decide whether to use the
object) results in the plugin deciding to use its offload data (and a
consequent mismatch in the offload data present at runtime).
The new hook allows the linker plugin to distinguish calls to
claim_file_handler that know the object is being used by the linker
(from ldmain.c:add_archive_element), from calls that don't know it's
being used by the linker (from elf_link_is_defined_archive_symbol); in
the latter case, the plugin should avoid recording the object as one
with offload data.
Thomas Schwinge [Wed, 10 May 2023 07:17:47 +0000 (09:17 +0200)]
Testsuite: Add 'torture-init-done', and use it to conditionalize implicit 'torture-init'
Recent commit d6654a4be3ba44c0d57be7c8a51d76d9721345e1
"Let each 'lto_init' determine the default 'LTO_OPTIONS', and 'torture-init' the 'LTO_TORTURE_OPTIONS'"
made 'torture-init' non-idempotent re 'LTO_TORTURE_OPTIONS', in order to catch
certain classes of errors. Now, most of all '*.exp' files have 'torture-init'
followed by 'set-torture-options' before 'gcc-dg-runtest' etc., and therefore
don't run into the latter's
"Some callers set torture options themselves; don't override those." code.
Some '*.exp' files however do 'torture-init' but not 'set-torture-options', and
therefore we can't any longer conditionalize the implicit 'torture-init' by
'![torture-options-exist]'.
gcc/testsuite/
* lib/torture-options.exp (torture-init-done): Add.
* lib/gcc-dg.exp (gcc-dg-runtest): Use it to conditionalize
implicit 'torture-init'.
* lib/gfortran-dg.exp (gfortran-dg-runtest): Likewise.
* lib/obj-c++-dg.exp (obj-c++-dg-runtest): Likewise.
* lib/objc-dg.exp (objc-dg-runtest): Likewise.
Thomas Schwinge [Tue, 9 May 2023 08:35:27 +0000 (10:35 +0200)]
Testsuite: Add missing 'torture-init'/'torture-finish' around 'LTO_TORTURE_OPTIONS' usage
Recent commit d6654a4be3ba44c0d57be7c8a51d76d9721345e1
"Let each 'lto_init' determine the default 'LTO_OPTIONS', and 'torture-init' the 'LTO_TORTURE_OPTIONS'"
made it a requirement that 'LTO_TORTURE_OPTIONS' usage be within
'torture-init'/'torture-finish', and missed a few cases that didn't have that.
Roger Sayle [Thu, 11 May 2023 07:15:21 +0000 (08:15 +0100)]
match.pd: Simplify popcount(X&Y)+popcount(X|Y) as popcount(X)+popcount(Y)
This patch teaches match.pd to simplify popcount(X&Y)+popcount(X|Y) as
popcount(X)+popcount(Y), and the related simplifications that
popcount(X)+popcount(Y)-popcount(X&Y) is popcount(X|Y). As surprising
as it might seem, this idiom is common in cheminformatics codes
(for Tanimoto coefficient calculations).
2023-05-11 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog
* match.pd <popcount optimizations>: Simplify popcount(X|Y) +
popcount(X&Y) as popcount(X)+popcount(Y). Likewise, simplify
popcount(X)+popcount(Y)-popcount(X&Y) as popcount(X|Y), and
vice versa.
gcc/testsuite/ChangeLog
* gcc.dg/fold-popcount-8.c: New test case.
* gcc.dg/fold-popcount-9.c: Likewise.
* gcc.dg/fold-popcount-10.c: Likewise.
Roger Sayle [Thu, 11 May 2023 07:10:04 +0000 (08:10 +0100)]
match.pd: Simplify popcount/parity of bswap/rotate.
This is the latest iteration of my patch from August 2020
https://gcc.gnu.org/pipermail/gcc-patches/2020-August/552391.html
incorperating feedback and suggestions from reviewers.
This patch to match.pd optimizes away bit permutation operations,
specifically bswap and rotate, in calls to popcount and parity.
2023-05-11 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog
* match.pd <popcount optimizations>: Simplify popcount(bswap(x))
as popcount(x). Simplify popcount(rotate(x,y)) as popcount(x).
<parity optimizations>: Simplify parity(bswap(x)) as parity(x).
Simplify parity(rotate(x,y)) as parity(x).
This is because LoopVectorizer assume target is able to handle
series const vector when we enable binary operations.
Then it will be easily causing ICE like that.
gcc/ChangeLog:
* config/riscv/autovec.md (@vec_series<mode>): New pattern
* config/riscv/riscv-protos.h (expand_vec_series): New function.
* config/riscv/riscv-v.cc (emit_binop): Ditto.
(emit_index_op): Ditto.
(expand_vec_series): Ditto.
(expand_const_vector): Add series vector handling.
* config/riscv/riscv.cc (riscv_const_insns): Enable series vector for testing.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/series-1.c: New test.
* gcc.target/riscv/rvv/autovec/series_run-1.c: New test.
std::add_rvalue_reference is defined as "If T is a function type that
has no cv- or ref- qualifier or an object type, provides a member typedef
type which is T&&, otherwise type is T."
In our case, T is cv-qualified, so the result is T, so we end up with
int () const declval() noexcept;
which is invalid. In other words, this is pretty much like:
using T = int () const;
T fn1(); // bad, fn returning a fn
T& fn2(); // bad, cannot declare reference to qualified function type
T* fn3(); // bad, cannot declare pointer to qualified function type
using U = int ();
U fn4(); // bad, fn returning a fn
U& fn5(); // OK
U* fn6(); // OK
I think is_convertible_helper needs to simulate std::declval better.
To that end, I'm introducing build_trait_object, to be used where
a declval is needed.
PR c++/109680
gcc/cp/ChangeLog:
* method.cc (build_trait_object): New.
(assignable_expr): Use it.
(ref_xes_from_temporary): Likewise.
(is_convertible_helper): Likewise. Check FUNC_OR_METHOD_TYPE_P.
Jason Merrill [Thu, 23 Mar 2023 14:41:29 +0000 (10:41 -0400)]
c++: adjust conversion diagnostics
While looking at PR109247 I made this change to improve diagnostics. I
don't think I'm going ahead with that patch, but this still seems like a
worthy cleanup.
gcc/cp/ChangeLog:
* call.cc (convert_like_internal): Share ck_ref_bind handling
between all bad conversions.
PR fortran/109624
* dump-parse-tree.cc (debug): New function for gfc_namespace.
(gfc_debug_code): Delete forward declaration.
(show_attr): Make sure to print balanced braces.
Andrew Pinski [Wed, 10 May 2023 17:56:15 +0000 (17:56 +0000)]
Add another new testcase
While working on improving min/max detection, this
code (which is reduced from worse_state in ipa-pure-const.cc)
was being miscompiled. Since there was no testcase in the
testsuite yet for this, this patch adds one.
Committed as obvious after testing the testcase via:
make check-gcc RUNTESTFLAGS="execute.exp=20230510-1.c"
François Dumont [Wed, 3 May 2023 04:45:47 +0000 (06:45 +0200)]
libstdc++: [_Hashtable] Implement several small methods implicitly inline
Make implementation of 3 simple _Hashtable methods implicitly inline.
Avoid usage of const_iterator abstraction within _Hashtable implementation.
Replace several usages of __node_type* with expected __node_ptr.
libstdc++-v3/ChangeLog:
* include/bits/hashtable_policy.h
(_NodeBuilder<>::_S_build): Use __node_ptr.
(_ReuseOrAllocNode<>): Use __node_ptr in place of __node_type*.
(_AllocNode<>): Likewise.
(_Equality<>::_M_equal): Remove const_iterator usages. Only preserved
to call std::is_permutation in the non-unique key implementation.
* include/bits/hashtable.h (_Hashtable<>::_M_update_begin()): Capture
_M_begin() once.
(_Hashtable<>::_M_bucket_begin(size_type)): Implement implicitly inline.
(_Hashtable<>::_M_insert_bucket_begin): Likewise.
(_Hashtable<>::_M_remove_bucket_begin): Likewise.
(_Hashtable<>::_M_compute_hash_code): Use __node_ptr rather than
const_iterator.
(_Hashtable<>::find): Likewise.
(_Hashtable<>::_M_emplace): Likewise.
(_Hashtable<>::_M_insert_unique): Likewise.
Jason Merrill [Mon, 6 Feb 2023 23:08:17 +0000 (15:08 -0800)]
c++: be stricter about constinit [CWG2543]
DR 2543 clarifies that constinit variables should follow the language, and
diagnose non-constant initializers (according to [expr.const]) even if they
can actually initialize the variables statically.
Jason Merrill [Tue, 9 May 2023 22:48:11 +0000 (18:48 -0400)]
c++: always check consteval address
The restriction on the "permitted result of a constant expression" to not
refer to an immediate function applies regardless of context. The previous
code tried to only check in cases where we wouldn't get the check in
cp_fold_r, but with the next patch I would need to add another case and it
shouldn't be a problem to always check.
We also shouldn't talk about immediate evaluation when we aren't dealing
with one.
gcc/cp/ChangeLog:
* constexpr.cc (cxx_eval_outermost_constant_expr): Always check
for address of immediate fn.
(maybe_constant_init_1): Evaluate PTRMEM_CST.
Jeff Law [Wed, 10 May 2023 11:25:12 +0000 (05:25 -0600)]
Fix a couple constraints on the H8 in preparation for LRA conversion
So this is the 2nd patch on the way to LRA for the H8.
LRA is more sensitive to getting define_constraint vs define_memory_constraint
vs define_special_memory_constraint correct. than reload.
The H8 port has the "Q" constraint, which is used to indicate memory addresses
that can be used under certain circumstances in various ALU operations. So it
should be a memory constraint. Ideally it'd would be a simple memory
constraint, but it's used in contexts where MEMs are valid only for certain
parts in the H8 family. So it really needs to be a special_memory_constraint.
The "Zz" constraint accepts memory, but the forms are limited and can not be
reloaded into a register. It seems to be working, but I wouldn't be totally
surprised if this got stressed in the right way if it broke.
Anyway, this patch fixes "Q" and "Zz" to be special memory constraints.
Regression tested with gdbsim and pushed to the trunk.
gcc
* config/h8300/constraints.md (Q): Make this a special memory
constraint.
(Zz): Similarly.
Jakub Jelinek [Wed, 10 May 2023 11:20:39 +0000 (13:20 +0200)]
ipa-prop: Fix ipa_get_callee_param_type for calls with argument type mismatches
The PR contains a testcase where the Fortran FE creates FUNCTION_TYPE
which doesn't really match the passed in arguments (FUNCTION_TYPE has
5 arguments, call has 6). Now, I think that is a Fortran FE bug that
should be fixed there, but I think with function pointers one can
create something similar (of course invalid) in C/C++ too,so IMHO IPA
should be also more careful.
The ipa_get_callee_param_type function can return NULL if something goes
wrong and it does e.g. if asked for 7th argument type on a function
with just 5 arguments and similar. But, if a function isn't varargs,
when asked for 6th argument type on a function with just 5 arguments
it actually returns void_type_node because the argument list is in that
case terminated with void_list_node.
The following patch makes sure we don't treat void_list_node as something
holding another argument.
2023-05-10 Jakub Jelinek <jakub@redhat.com>
PR fortran/109788
* ipa-prop.cc (ipa_get_callee_param_type): Don't return TREE_VALUE (t)
if t is void_list_node.
projects / thirdparty / gcc.git / log
