https://gcc.gnu.org/ml/gcc-patches/2017-09/msg00983.html
* cp-demangle.c (is_fnqual_component_type): Reimplement using
FNQUAL_COMPONENT_CASE.
(d_encoding): Hold bare_function_type in local var.
(d_local_name): Build name in both cases and build result once.
Collapse switch-if to single conditional.
* testsuite/demangle-expected: Realign blank lines with tests.
From-SVN: r252802
2017-09-15 Richard Biener <rguenther@suse.de>
PR tree-optimization/82217
* tree-ssa-sccvn.c (visit_phi): Properly handle all VN_TOP
but not undefined case.
* gcc.dg/torture/pr82217.c: New testcase.
From-SVN: r252796
2017-09-15 Paul Thomas <pault@gcc.gnu.org>
PR fortran/82184
trans-decl.c (gfc_trans_deferred_vars): Do not null the 'span'
field if the symbol is either implicitly or explicitly saved.
2017-09-15 Paul Thomas <pault@gcc.gnu.org>
PR fortran/82184
* gfortran.dg/pointer_array_9.f90: New test.
From-SVN: r252781
2017-09-15 Richard Biener <rguenther@suse.de>
PR tree-optimization/68823
* graphite-scop-detection.c (build_alias_set): If we have a
possible dependence check whether we can handle them by just
looking at the DRs DR_ACCESS_FNs.
(build_scops): If build_alias_set fails, fail the SCOP.
From-SVN: r252780
2017-09-14 Tom de Vries <tom@codesourcery.com>
* testsuite/libgomp.c++/cancel-taskgroup-1.C: Remove.
* testsuite/libgomp.c/cancel-taskgroup-1.c: Move to ...
* testsuite/libgomp.c-c++-common/cancel-taskgroup-1.c: ... here.
* testsuite/libgomp.c/c.exp: Include test-cases from
libgomp.c-c++-common.
* testsuite/libgomp.c++/c++.exp: Same. Force c++-mode compilation of .c
files.
From-SVN: r252775
[gcc]
2017-09-14 Michael Meissner <meissner@linux.vnet.ibm.com>
* config/rs6000/rs6000-builtin.def (BU_FLOAT128_1_HW): New macros
to support float128 built-in functions that require the ISA 3.0
hardware.
(BU_FLOAT128_3_HW): Likewise.
(SQRTF128): Add support for the IEEE 128-bit square root and fma
built-in functions.
(FMAF128): Likewise.
(FMAQ): Likewise.
* config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
support for built-in functions that need the ISA 3.0 IEEE 128-bit
floating point instructions.
(rs6000_invalid_builtin): Likewise.
(rs6000_builtin_mask_names): Likewise.
* config/rs6000/rs6000.h (MASK_FLOAT128_HW): Likewise.
(RS6000_BTM_FLOAT128_HW): Likewise.
(RS6000_BTM_COMMON): Likewise.
* config/rs6000/rs6000.md (fma<mode>4_hw): Add a generator
function.
* doc/extend.texi (RS/6000 built-in functions): Document the
IEEE 128-bit floating point square root and fused multiply-add
built-in functions.
[gcc/testsuite]
2017-09-14 Michael Meissner <meissner@linux.vnet.ibm.com>
* gcc.target/powerpc/abs128-1.c: Use __builtin_fabsf128 instead of
__builtin_fabsq.
* gcc.target/powerpc/float128-5.c: Use __builtin_fabsf128 instead
of __builtin_fabsq. Prevent the test from running on 32-bit.
* gcc.target/powerpc/float128-fma1.c: New test.
* gcc.target/powerpc/float128-fma2.c: Likewise.
* gcc.target/powerpc/float128-sqrt1.c: Likewise.
* gcc.target/powerpc/float128-sqrt2.c: Likewise.
From-SVN: r252771
Calls to gcc_jit_context_get_builtin_function that accessed builtins
in sanitizer.def and after (or failed to match any builtin) led to
a crash accessing a NULL builtin name.
The entries with the NULL name came from these lines in sanitizer.def:
/* This has to come before all the sanitizer builtins. */
DEF_BUILTIN_STUB(BEGIN_SANITIZER_BUILTINS, (const char *)0)
[...snip...]
/* This has to come after all the sanitizer builtins. */
DEF_BUILTIN_STUB(END_SANITIZER_BUILTINS, (const char *)0)
This patch updates jit-builtins.c to cope with such entries, fixing the
crash.
gcc/jit/ChangeLog:
PR jit/82174
* jit-builtins.c (matches_builtin): Ignore entries with a NULL
name.
gcc/testsuite/ChangeLog:
PR jit/82174
* jit.dg/test-error-gcc_jit_context_get_builtin_function-unknown-builtin.c:
New test case.
From-SVN: r252769
* config/rs6000/rs6000.c (rs6000_set_up_by_prologue): Make sure the TOC
reg (r2) isn't in the set of registers defined in the prologue.
* gcc.target/powerpc/r2_shrink-wrap.c: New.
From-SVN: r252768
Epilogue vectorisation uses the vectorisation factor of the main loop
as the maximum vectorisation factor allowed for correctness. That makes
sense as a conservatively correct value, since the chosen vectorisation
factor will be strictly less than that anyway. However, once the VF
itself becomes variable, it's easier to carry across the original
maximum VF instead.
2017-09-14 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-vectorizer.h (_loop_vec_info): Add max_vectorization_factor.
(LOOP_VINFO_MAX_VECT_FACTOR): New macro.
(LOOP_VINFO_ORIG_VECT_FACTOR): Replace with...
(LOOP_VINFO_ORIG_MAX_VECT_FACTOR): ...this new macro.
* tree-vect-data-refs.c (vect_analyze_data_ref_dependences): Update
accordingly.
* tree-vect-loop.c (_loop_vec_info::_loop_vec_info): Initialize
max_vectorization_factor.
(vect_analyze_loop_2): Set LOOP_VINFO_MAX_VECT_FACTOR.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252766
The vectoriser sometimes considers lowering "vector" operations into N
scalar word operations. This N needs to be fixed at compile time, so
the condition guarding it needs to change when variable-lengh vectors
are added. This patch puts the condition into a helper routine so that
there's only one place to update.
2017-09-14 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-vectorizer.h (vect_min_worthwhile_factor): Delete.
(vect_worthwhile_without_simd_p): Declare.
* tree-vect-loop.c (vect_worthwhile_without_simd_p): New function.
(vectorizable_reduction): Use it.
* tree-vect-stmts.c (vectorizable_shift): Likewise.
(vectorizable_operation): Likewise.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252765
This patch adds a vectoriser helper routine to calculate how
many copies of a vector statement we need. At present this
is always:
LOOP_VINFO_VECT_FACTOR (loop_vinfo) / TYPE_VECTOR_SUBPARTS (vectype)
but later patches add other cases. Another benefit of using
a helper routine is that it can assert that the division is
exact (which it must be).
2017-09-14 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-vectorizer.h (vect_get_num_copies): New function.
* tree-vect-data-refs.c (vect_get_data_access_cost): Use it.
* tree-vect-loop.c (vectorizable_reduction): Likewise.
(vectorizable_induction): Likewise.
(vectorizable_live_operation): Likewise.
* tree-vect-stmts.c (vectorizable_mask_load_store): Likewise.
(vectorizable_bswap): Likewise.
(vectorizable_call): Likewise.
(vectorizable_conversion): Likewise.
(vectorizable_assignment): Likewise.
(vectorizable_shift): Likewise.
(vectorizable_operation): Likewise.
(vectorizable_store): Likewise.
(vectorizable_load): Likewise.
(vectorizable_condition): Likewise.
(vectorizable_comparison): Likewise.
(vect_analyze_stmt): Pass the slp node to vectorizable_live_operation.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252764
This patch makes the vectoriser use the gimple-fold.h routines
in more cases, instead of vect_init_vector. Later patches want
to use the same interface to handle variable-length vectors.
2017-09-14 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-vect-loop.c (vectorizable_induction): Use gimple_build instead
of vect_init_vector.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252763
This patch adds gimple-fold.h equivalents of build_vector and
build_vector_from_val. Like the other gimple-fold.h routines
they always return a valid gimple value and add any new
statements to a given gimple_seq. In combination with later
patches this reduces the number of force_gimple_operands.
2017-09-14 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* gimple-fold.h (gimple_build_vector_from_val): Declare, and provide
an inline wrapper that provides a location.
(gimple_build_vector): Likewise.
* gimple-fold.c (gimple_build_vector_from_val): New function.
(gimple_build_vector): Likewise.
* tree-vect-loop.c (get_initial_def_for_reduction): Use the new
functions to build the initial value. Always return a gimple value.
(get_initial_defs_for_reduction): Likewise. Only compute
neutral_vec once.
(vect_create_epilog_for_reduction): Don't call force_gimple_operand or
vect_init_vector on the results from get_initial_def(s)_for_reduction.
(vectorizable_induction): Use gimple_build_vector rather than
vect_init_vector.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252762
This patch makes can_vec_perm_p & co. take a vec<>, wrapped in new
typedefs vec_perm_indices and auto_vec_perm_indices. There are two
reasons for doing this for SVE:
(1) it means that the number of elements is bundled with the elements
themselves, and is obviously constant.
(2) it makes it easier to change the "unsigned char" element type to
something wider.
Changing the target hook is left as follow-on work.
2017-09-14 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* target.h (vec_perm_indices): New typedef.
(auto_vec_perm_indices): Likewise.
* optabs-query.h: Include target.h
(can_vec_perm_p): Take a vec_perm_indices *.
* optabs-query.c (can_vec_perm_p): Likewise.
(can_mult_highpart_p): Update accordingly. Use auto_vec_perm_indices.
* tree-ssa-forwprop.c (simplify_vector_constructor): Likewise.
* tree-vect-generic.c (lower_vec_perm): Likewise.
* tree-vect-data-refs.c (vect_grouped_store_supported): Likewise.
(vect_grouped_load_supported): Likewise.
(vect_shift_permute_load_chain): Likewise.
(vect_permute_store_chain): Use auto_vec_perm_indices.
(vect_permute_load_chain): Likewise.
* fold-const.c (fold_vec_perm): Take vec_perm_indices.
(fold_ternary_loc): Update accordingly. Use auto_vec_perm_indices.
Update uses of can_vec_perm_p.
* tree-vect-loop.c (calc_vec_perm_mask_for_shift): Replace the
mode with a number of elements. Take a vec_perm_indices *.
(vect_create_epilog_for_reduction): Update accordingly.
Use auto_vec_perm_indices.
(have_whole_vector_shift): Likewise. Update call to can_vec_perm_p.
* tree-vect-slp.c (vect_build_slp_tree_1): Likewise.
(vect_transform_slp_perm_load): Likewise.
(vect_schedule_slp_instance): Use auto_vec_perm_indices.
* tree-vectorizer.h (vect_gen_perm_mask_any): Take a vec_perm_indices.
(vect_gen_perm_mask_checked): Likewise.
* tree-vect-stmts.c (vect_gen_perm_mask_any): Take a vec_perm_indices.
(vect_gen_perm_mask_checked): Likewise.
(vectorizable_mask_load_store): Use auto_vec_perm_indices.
(vectorizable_store): Likewise.
(vectorizable_load): Likewise.
(perm_mask_for_reverse): Likewise. Update call to can_vec_perm_p.
(vectorizable_bswap): Likewise.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252761
This patch makes build_vector take the elements as a vec<> rather
than a tree *. This is useful for SVE because it bundles the number
of elements with the elements themselves, and enforces the fact that
the number is constant. Also, I think things like the folds can be used
with any generic GNU vector, not just those that match machine vectors,
so the arguments to XALLOCAVEC had no clear limit.
2017-09-14 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree.h (build_vector): Take a vec<tree> instead of a tree *.
* tree.c (build_vector): Likewise.
(build_vector_from_ctor): Update accordingly.
(build_vector_from_val): Likewise.
* gimple-fold.c (gimple_fold_stmt_to_constant_1): Likewise.
* tree-ssa-forwprop.c (simplify_vector_constructor): Likewise.
* tree-vect-generic.c (add_rshift): Likewise.
(expand_vector_divmod): Likewise.
(optimize_vector_constructor): Likewise.
* tree-vect-slp.c (vect_get_constant_vectors): Likewise.
(vect_transform_slp_perm_load): Likewise.
(vect_schedule_slp_instance): Likewise.
* tree-vect-stmts.c (vectorizable_bswap): Likewise.
(vectorizable_call): Likewise.
(vect_gen_perm_mask_any): Likewise. Add elements in order.
* expmed.c (make_tree): Likewise.
* fold-const.c (fold_negate_expr_1): Use auto_vec<tree> when building
a vector passed to build_vector.
(fold_convert_const): Likewise.
(exact_inverse): Likewise.
(fold_ternary_loc): Likewise.
(fold_relational_const): Likewise.
(const_binop): Likewise. Use VECTOR_CST_ELT directly when operating
on VECTOR_CSTs, rather than going through vec_cst_ctor_to_array.
(const_unop): Likewise. Store the reduction accumulator in a
variable rather than an array.
(vec_cst_ctor_to_array): Take the number of elements as a parameter.
(fold_vec_perm): Update calls accordingly. Use auto_vec<tree> for
the new vector, rather than constructing it after the input arrays.
(native_interpret_vector): Use auto_vec<tree> when building
a vector passed to build_vector. Add elements in order.
* tree-vect-loop.c (get_initial_defs_for_reduction): Use
auto_vec<tree> when building a vector passed to build_vector.
(vect_create_epilog_for_reduction): Likewise.
(vectorizable_induction): Likewise.
(get_initial_def_for_reduction): Likewise. Fix indentation of
case statements.
* config/sparc/sparc.c (sparc_handle_vis_mul8x16): Change n_elts
to a vec<tree> *.
(sparc_fold_builtin): Use auto_vec<tree> when building a vector
passed to build_vector.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252760
Previously VECTOR_CST_NELTS (t) read the number of elements from
TYPE_VECTOR_SUBPARTS (TREE_TYPE (t)). There were two ways of handling
this with variable TYPE_VECTOR_SUBPARTS: either forcibly convert the
number to a constant (which is doable) or store the number directly
in the VECTOR_CST. The latter seemed better, since it involves less
pointer chasing and since the tree_node u field is otherwise unused
for VECTOR_CST. It would still be easy to switch to the former in
future if we need to free up the field for someting else.
The patch also changes various bits of VECTOR_CST code to use
VECTOR_CST_NELTS instead of TYPE_VECTOR_SUBPARTS when iterating
over VECTOR_CST_ELTs. Also, when the two are checked for equality,
the patch prefers to read VECTOR_CST_NELTS (which must be constant)
and check against TYPE_VECTOR_SUBPARTS, instead of the other way
around.
2017-09-14 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* tree-core.h (tree_base::u): Add an "nelts" field.
(tree_vector): Use VECTOR_CST_NELTS as the length.
* tree.c (tree_size): Likewise.
(make_vector): Initialize VECTOR_CST_NELTS.
* tree.h (VECTOR_CST_NELTS): Use the u.nelts field.
* cfgexpand.c (expand_debug_expr): Use VECTOR_CST_NELTS instead of
TYPE_VECTOR_SUBPARTS.
* expr.c (const_vector_mask_from_tree): Consistently use "units"
as the number of units, setting it from VECTOR_CST_NELTS.
(const_vector_from_tree): Likewise.
* fold-const.c (negate_expr_p): Use VECTOR_CST_NELTS instead of
TYPE_VECTOR_SUBPARTS for the number of elements in a VECTOR_CST.
(fold_negate_expr_1): Likewise.
(fold_convert_const): Likewise.
(const_binop): Likewise. Differentiate the number of output and
input elements.
(const_unop): Likewise.
(fold_ternary_loc): Use VECTOR_CST_NELTS for the number of elements
in a VECTOR_CST, asserting that it is the same as TYPE_VECTOR_SUBPARTS
in cases that did the opposite.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252758
PR target/81325
* cfgbuild.c (find_bb_boundaries): Ignore debug insns in decisions
if and where to split a bb, except for splitting before debug insn
sequences followed by non-label real insn. Delete debug insns
in between basic blocks.
* g++.dg/cpp0x/pr81325.C: New test.
From-SVN: r252752
In preparation for upgrading libgo to the 1.9 release, this
approximately incorporates https://golang.org/cl/37661 and
https://golang.org/cl/38351.
CL 37661 changed the gc compiler such that the select statement simply
returns an integer which is then used as the argument for a switch.
Since gccgo already worked that way, this just adjusts the switch code
to look like the gc switch code by removing the explicit case index
expression and calculating it from the order of calls to selectsend,
selectrecv, and selectdefault.
CL 38351 simplifies the channel code by not passing the unused channel
type descriptor pointer.
Reviewed-on: https://go-review.googlesource.com/62730
From-SVN: r252749
I encountered this crash while working on upgrading libgo to the 1.9
release. I no longer have the cause of the crash, but it doesn't much
matter, as the policy for crash-on-invalid errors is to fix the crash
but not bother to commit the invalid test case.
Reviewed-on: https://go-review.googlesource.com/62750
From-SVN: r252748
If we have an alias for a struct or array that requires a
type-specific function, don't emit the function with the alias name.
Emit it with the struct/array as usual.
Test case is https://golang.org/cl/62531.
Reviewed-on: https://go-review.googlesource.com/62412
From-SVN: r252747
This adds much of https://golang.org/cl/35731 and
https://golang.org/cl/35732 to the gofrontend code.
This is a step toward updating libgo to the 1.9 release. The
gofrontend already supports type aliases, and this is required for
correct support of type aliases when used as embedded fields.
The change to expressions.cc is to handle the << 1, used for the
newly renamed offsetAnon field, in the constant context used for type
descriptor initialization.
Reviewed-on: https://go-review.googlesource.com/62710
From-SVN: r252746
A normal pointer may not be converted to a notinheap pointer. We were
erroneously permitting a conversion from a normal pointer to a
notinheap unsafe.Pointer, which is useless since unsafe.Pointer is not
marked notinheap. Correct the test to permit a conversion from
unsafe.Pointer to a notinheap pointer, which is the same test that the
gc compiler uses.
The test case for this is in the 1.9 runtime package.
Reviewed-on: https://go-review.googlesource.com/62731
From-SVN: r252745
2017-09-13 Paul Thomas <pault@gcc.gnu.org>
PR fortran/82173
* decl.c (match_char_kind): If the kind expression is
parameterized, save it in saved_kind_expr and set kind = 0.
(gfc_get_pdt_instance): Resolve and simplify before emitting
error on expression kind. Insert a missing simplification after
insertion of kind expressions.
2017-09-13 Paul Thomas <pault@gcc.gnu.org>
PR fortran/82173
* gfortran.dg/pdt_10.f03 : New test.
From-SVN: r252734
2017-09-13 Martin Liska <mliska@suse.cz>
PR middle-end/82154
* stmt.c (expand_sjlj_dispatch_table): Use CASE_LOW when
CASE_HIGH is NULL_TREE.
2017-09-13 Martin Liska <mliska@suse.cz>
PR middle-end/82154
* g++.dg/torture/pr82154.C: New test.
From-SVN: r252728
Since the patch is going through all the definitions anyway, it seemed
like a good opportunity to put the mode argument first, to match the
order for register_move_cost.
2017-09-13 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* target.def (secondary_memory_needed): New hook.
(secondary_reload): Refer to TARGET_SECONDARY_MEMORY_NEEDED
instead of SECONDARY_MEMORY_NEEDED.
(secondary_memory_needed_mode): Likewise.
* hooks.h (hook_bool_mode_reg_class_t_reg_class_t_false): Declare.
* hooks.c (hook_bool_mode_reg_class_t_reg_class_t_false): New function.
* doc/tm.texi.in (SECONDARY_MEMORY_NEEDED): Replace with...
(TARGET_SECONDARY_MEMORY_NEEDED): ...this.
(SECONDARY_MEMORY_NEEDED_RTX): Update reference accordingly.
* doc/tm.texi: Regenerate.
* config/alpha/alpha.h (SECONDARY_MEMORY_NEEDED): Delete.
* config/alpha/alpha.c (alpha_secondary_memory_needed): New function.
(TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
* config/i386/i386.h (SECONDARY_MEMORY_NEEDED): Delete.
* config/i386/i386-protos.h (ix86_secondary_memory_needed): Delete.
* config/i386/i386.c (inline_secondary_memory_needed): Put the
mode argument first and change the reg_class arguments to reg_class_t.
(ix86_secondary_memory_needed): Likewise. Remove the strict parameter.
Make static. Update the call to inline_secondary_memory_needed.
(ix86_register_move_cost): Update the call to
inline_secondary_memory_needed.
(TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
* config/ia64/ia64.h (SECONDARY_MEMORY_NEEDED): Delete commented-out
definition.
* config/ia64/ia64.c (spill_xfmode_rfmode_operand): Refer to
TARGET_SECONDARY_MEMORY_NEEDED rather than SECONDARY_MEMORY_NEEDED
in comment.
* config/mips/mips.h (SECONDARY_MEMORY_NEEDED): Delete.
* config/mips/mips-protos.h (mips_secondary_memory_needed): Delete.
* config/mips/mips.c (mips_secondary_memory_needed): Make static
and match hook interface. Add comment from mips.h.
(TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
* config/mmix/mmix.md (truncdfsf2): Refer to
TARGET_SECONDARY_MEMORY_NEEDED rather than SECONDARY_MEMORY_NEEDED
in comment.
* config/pa/pa-64.h (SECONDARY_MEMORY_NEEDED): Rename to...
(PA_SECONDARY_MEMORY_NEEDED): ...this, and put the mode argument first.
* config/pa/pa.c (TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
(pa_secondary_memory_needed): New function.
* config/pdp11/pdp11.h (SECONDARY_MEMORY_NEEDED): Delete.
* config/pdp11/pdp11-protos.h (pdp11_secondary_memory_needed): Delete.
* config/pdp11/pdp11.c (TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
(pdp11_secondary_memory_needed): Make static and match hook interface.
* config/powerpcspe/powerpcspe.h (SECONDARY_MEMORY_NEEDED): Delete.
* config/powerpcspe/powerpcspe-protos.h
(rs6000_secondary_memory_needed_ptr): Delete.
* config/powerpcspe/powerpcspe.c (rs6000_secondary_memory_needed_ptr):
Delete.
(TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
(rs6000_option_override_internal): Assign to
targetm.secondary_memory_needed rather than
rs6000_secondary_memory_needed_ptr.
(rs6000_secondary_memory_needed): Match hook interface.
(rs6000_debug_secondary_memory_needed): Likewise.
* config/riscv/riscv.h (SECONDARY_MEMORY_NEEDED): Delete.
* config/riscv/riscv.c (riscv_secondary_memory_needed): New function.
(riscv_register_move_cost): Use it instead of SECONDARY_MEMORY_NEEDED.
(TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
* config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED): Delete.
* config/rs6000/rs6000-protos.h (rs6000_secondary_memory_needed_ptr):
Delete.
* config/rs6000/rs6000.c (rs6000_secondary_memory_needed_ptr): Delete.
(TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
(rs6000_option_override_internal): Assign to
targetm.secondary_memory_needed rather than
rs6000_secondary_memory_needed_ptr.
(rs6000_secondary_memory_needed): Match hook interface.
(rs6000_debug_secondary_memory_needed): Likewise.
* config/s390/s390.h (SECONDARY_MEMORY_NEEDED): Delete.
* config/s390/s390.c (s390_secondary_memory_needed): New function.
(TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
* config/sparc/sparc.h (SECONDARY_MEMORY_NEEDED): Delete.
* config/sparc/sparc.c (TARGET_SECONDARY_MEMORY_NEEDED): Redefine.
(sparc_secondary_memory_needed): New function.
* lra-constraints.c (check_and_process_move): Refer to
TARGET_SECONDARY_MEMORY_NEEDED rather than SECONDARY_MEMORY_NEEDED
in comment.
(curr_insn_transform): Likewise.
(process_alt_operands): Use targetm.secondary_memory_needed
instead of TARGET_SECONDARY_MEMORY_NEEDED.
(check_secondary_memory_needed_p): Likewise.
(choose_split_class): Likewise.
* reload.c: Unconditionally include code that was previously
conditional on SECONDARY_MEMORY_NEEDED.
(push_secondary_reload): Use targetm.secondary_memory_needed
instead of TARGET_SECONDARY_MEMORY_NEEDED.
(push_reload): Likewise.
* reload1.c: Unconditionally include code that was previously
conditional on SECONDARY_MEMORY_NEEDED.
(choose_reload_regs): Use targetm.secondary_memory_needed
instead of TARGET_SECONDARY_MEMORY_NEEDED.
(gen_reload): Likewise.
* system.h (SECONDARY_MEMORY_NEEDED): Poison.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252461
This includes a change to LRA. Previously the code was:
if (sclass == NO_REGS && dclass == NO_REGS)
return false;
#ifdef SECONDARY_MEMORY_NEEDED
if (SECONDARY_MEMORY_NEEDED (sclass, dclass, GET_MODE (src))
#ifdef SECONDARY_MEMORY_NEEDED_MODE
&& ((sclass != NO_REGS && dclass != NO_REGS)
|| GET_MODE (src) != SECONDARY_MEMORY_NEEDED_MODE (GET_MODE (src)))
#endif
)
{
*sec_mem_p = true;
return false;
}
#endif
in which the positioning of the second ifdef meant that defining
SECONDARY_MEMORY_NEEDED_MODE to its default value was not a no-op:
without a definition, we would consider using secondary reloads for
mem<-reg and reg<-mem reloads even if the secondary memory has the
same mode as the original mem, while defining it would avoid this.
The latter behaviour seems correct.
The default is different for reload and LRA. For LRA the default is
to use the original mode, while reload promotes smaller-than-word
integral modes to word mode:
if (GET_MODE_BITSIZE (mode) < BITS_PER_WORD && INTEGRAL_MODE_P (mode))
mode = mode_for_size (BITS_PER_WORD,
GET_MODE_CLASS (mode), 0).require ();
Some of the ports that have switched to LRA seemed to have
SECONDARY_MEMORY_NEEDED_MDOEs based on the old reload definition,
and still referred to the reload.c:get_secondary_mem function in
the comments. The patch just keeps them as-is.
2017-09-13 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
gcc/
* target.def (secondary_memory_needed_mode): New hook:
* targhooks.c (default_secondary_memory_needed_mode): Declare.
* targhooks.h (default_secondary_memory_needed_mode): New function.
* doc/tm.texi.in (SECONDARY_MEMORY_NEEDED_MODE): Replace with...
(TARGET_SECONDARY_MEMORY_NEEDED_MODE): ...this.
* doc/tm.texi: Regenerate.
* lra-constraints.c (check_and_process_move): Use
targetm.secondary_memory_needed_mode instead of
TARGET_SECONDARY_MEMORY_NEEDED_MODE.
(curr_insn_transform): Likewise.
* reload.c (get_secondary_mem): Likewise.
* config/alpha/alpha.h (SECONDARY_MEMORY_NEEDED_MODE): Delete.
* config/alpha/alpha.c (alpha_secondary_memory_needed_mode): New
function.
(TARGET_SECONDARY_MEMORY_NEEDED_MODE): Redefine.
* config/i386/i386.h (SECONDARY_MEMORY_NEEDED_MODE): Delete.
* config/i386/i386.c (ix86_secondary_memory_needed_mode): New function.
(TARGET_SECONDARY_MEMORY_NEEDED_MODE): Redefine.
* config/powerpcspe/powerpcspe.h (SECONDARY_MEMORY_NEEDED_MODE):
Delete.
* config/powerpcspe/powerpcspe-protos.h
(rs6000_secondary_memory_needed_mode): Delete.
* config/powerpcspe/powerpcspe.c
(TARGET_SECONDARY_MEMORY_NEEDED_MODE): Redefine.
(rs6000_secondary_memory_needed_mode): Make static.
* config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): Delete.
* config/rs6000/rs6000-protos.h (rs6000_secondary_memory_needed_mode):
Delete.
* config/rs6000/rs6000.c (TARGET_SECONDARY_MEMORY_NEEDED_MODE):
Redefine.
(rs6000_secondary_memory_needed_mode): Make static.
* config/s390/s390.h (SECONDARY_MEMORY_NEEDED_MODE): Delete.
* config/s390/s390.c (s390_secondary_memory_needed_mode): New function.
(TARGET_SECONDARY_MEMORY_NEEDED_MODE): Redefine.
* config/sparc/sparc.h (SECONDARY_MEMORY_NEEDED_MODE): Delete.
* config/sparc/sparc.c (TARGET_SECONDARY_MEMORY_NEEDED_MODE):
Redefine.
(sparc_secondary_memory_needed_mode): New function.
* system.h (TARGET_SECONDARY_MEMORY_NEEDED_MODE): Poison.
Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>
From-SVN: r252455
Committed on behalf of Jackson Woodruff.
---
gcc/
* config/aarch64/constraints.md (Umq): New constraint.
* config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
Change to use Umq.
(mov<mode>): Update condition.
gcc/testsuite/
* gcc.target/aarch64/simd/vect_str_zero.c: Update testcase.
From-SVN: r252387
PR c/82167
* c-typeck.c (c_expr_sizeof_expr): Use the type of expr.value rather
than expr.original_type.
* gcc.dg/pr82167.c: New test.
From-SVN: r252372
This patch changes pr62178.c so that it now scans
for two `ldr`s, one into an `s` register, instead
of a `ld1r` as before. Also add a scan for an mla
instruction.
The `ld1r` was needed when this should have generated
a mla by vector. Now that we can generate an mla by
element instruction and can load directly into the
simd register, it is cheaper to not do the ld1r
which needlessly duplicates the single element used
across the whole vector register.
Committed on behalf of Jackson Woodruff
gcc/testsuite/
* gcc.target/aarch64/pr62178.c: Updated testcase
to scan for two ldrs and an mla.
From-SVN: r252086
As Alexander pointed out in the thread starting at [1] the sort_by_bitpos sorting function
was behaving badly when we had multiple stores at the same position. He fixed that (thanks!)
but we can do better by not returning zero when the bitpositions are equal but by falling back
to comparing the order the stores appear in, which is guaranteed to be unique (barring other
bugs elsewhere).
This patch does that.
Bootstrapped and tested on aarch64-none-linux-gnu.
[1] https://gcc.gnu.org/ml/gcc-patches/2017-07/msg00895.html
* gimple-ssa-store-merging.c (sort_by_bitpos): Compare store order
when bitposition is the same.
From-SVN: r252084
2017-09-13 Nicolas Roche <roche@adacore.com>
* Make-lang.in: In the fallback mechanim, parse the associated .ali
file and try to guess the locations of dependencies.
From-SVN: r252082