objc: Add support for per-location warning groups.

gcc/objc/ChangeLog:

* objc-act.c (objc_maybe_build_modify_expr): Replace direct uses
of TREE_NO_WARNING with warning_suppressed_p, and suppress_warning.
(objc_build_incr_expr_for_property_ref): Same.
(objc_build_struct): Same.
(synth_module_prologue): Same.
* objc-gnu-runtime-abi-01.c (gnu_runtime_01_initialize): Same.
* objc-next-runtime-abi-01.c (next_runtime_01_initialize): Same.
* objc-next-runtime-abi-02.c (next_runtime_02_initialize): Same.

libcc1: Add support for per-location warning groups.

libcc1/ChangeLog:

* libcp1plugin.cc (record_decl_address): Replace a direct use
of TREE_NO_WARNING with suppress_warning.

fortran: Add support for per-location warning groups.

gcc/fortran/ChangeLog:

* trans-array.c (trans_array_constructor): Replace direct uses
of TREE_NO_WARNING with warning_suppressed_p, and suppress_warning.
* trans-decl.c (gfc_build_qualified_array): Same.
(gfc_build_dummy_array_decl): Same.
(generate_local_decl): Same.
(gfc_generate_function_code): Same.
* trans-openmp.c (gfc_omp_clause_default_ctor): Same.
(gfc_omp_clause_copy_ctor): Same.
* trans-types.c (get_dtype_type_node): Same.
(gfc_get_desc_dim_type): Same.
(gfc_get_array_descriptor_base): Same.
(gfc_get_caf_vector_type): Same.
(gfc_get_caf_reference_type): Same.
* trans.c (gfc_create_var_np): Same.

ada: Add support for per-location warning groups.

gcc/ada/ChangeLog:

* gcc-interface/trans.c (Handled_Sequence_Of_Statements_to_gnu):
Replace TREE_NO_WARNING with suppress_warning.
(gnat_gimplify_expr): Same.
* gcc-interface/utils.c (gnat_pushdecl): Same.

libbacktrace: improve XCOFF support

libbacktrace/ChangeLog:
2021-06-28  Clément Chigot  <clement.chigot@atos.net>

* xcoff.c (SSUBTYP_DWRNGES): New define.
(xcoff_add): Use correct XCOFF DWARF section subtype
for DEBUG_RANGES. Remove lineoff workaround.
Adjust base_address.
(xcoff_initialize_syminfo): Adapt to new base_address.
(xcoff_lookup_pc): Likewise.
(xcoff_initialize_fileline): Likewise.

libiberty: copy over .BTF section when using LTO

libiberty/ChangeLog:

* simple-object.c (handle_lto_debug_sections): Copy over .BTF section.

Enable BTF generation in the BPF backend

This patch changes the BPF GCC backend in order to use the DWARF debug
hooks and therefore enables the user to generate BTF debugging
information with -gbtf.  Generating BTF is crucial when compiling BPF
programs, since the CO-RE (compile-once, run-everwhere) mechanism
used by the kernel BPF loader relies on it.

Note that since in eBPF it is not possible to unwind frames due to the
restrictive nature of the target architecture, we are disabling the
generation of CFA in this target.

2021-06-28  David Faust <david.faust@oracle.com>

* config/bpf/bpf.c (bpf_expand_prologue): Do not mark insns as
frame related.
(bpf_expand_epilogue): Likewise.
* config/bpf/bpf.h (DWARF2_FRAME_INFO): Define to 0.
Do not define DBX_DEBUGGING_INFO.

CTF/BTF documentation

This commit documents the new command line options introduced by the
CTF and BTF debug formats.

2021-06-28  Indu Bhagat  <indu.bhagat@oracle.com>

* doc/invoke.texi: Document the CTF and BTF debug info options.

CTF/BTF testsuites

This commit adds a new testsuite for the CTF debug format.

2021-06-28  Indu Bhagat  <indu.bhagat@oracle.com>
    David Faust  <david.faust@oracle.com>

gcc/testsuite/

* lib/gcc-dg.exp (gcc-dg-frontend-supports-ctf): New procedure.
(gcc-dg-debug-runtest): Add -gctf support.
* gcc.dg/debug/btf/btf-1.c: New test.
* gcc.dg/debug/btf/btf-2.c: Likewise.
* gcc.dg/debug/btf/btf-anonymous-struct-1.c: Likewise.
* gcc.dg/debug/btf/btf-anonymous-union-1.c: Likewise.
* gcc.dg/debug/btf/btf-array-1.c: Likewise.
* gcc.dg/debug/btf/btf-bitfields-1.c: Likewise.
* gcc.dg/debug/btf/btf-bitfields-2.c: Likewise.
* gcc.dg/debug/btf/btf-bitfields-3.c: Likewise.
* gcc.dg/debug/btf/btf-cvr-quals-1.c: Likewise.
* gcc.dg/debug/btf/btf-enum-1.c: Likewise.
* gcc.dg/debug/btf/btf-forward-1.c: Likewise.
* gcc.dg/debug/btf/btf-function-1.c: Likewise.
* gcc.dg/debug/btf/btf-function-2.c: Likewise.
* gcc.dg/debug/btf/btf-int-1.c: Likewise.
* gcc.dg/debug/btf/btf-pointers-1.c: Likewise.
* gcc.dg/debug/btf/btf-struct-1.c: Likewise.
* gcc.dg/debug/btf/btf-typedef-1.c: Likewise.
* gcc.dg/debug/btf/btf-union-1.c: Likewise.
* gcc.dg/debug/btf/btf-variables-1.c: Likewise.
* gcc.dg/debug/btf/btf.exp: Likewise.
* gcc.dg/debug/ctf/ctf-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-anonymous-struct-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-anonymous-union-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-array-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-array-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-array-3.c: Likewise.
* gcc.dg/debug/ctf/ctf-array-4.c: Likewise.
* gcc.dg/debug/ctf/ctf-attr-mode-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-attr-used-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-bitfields-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-bitfields-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-bitfields-3.c: Likewise.
* gcc.dg/debug/ctf/ctf-bitfields-4.c: Likewise.
* gcc.dg/debug/ctf/ctf-complex-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-cvr-quals-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-cvr-quals-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-cvr-quals-3.c: Likewise.
* gcc.dg/debug/ctf/ctf-cvr-quals-4.c: Likewise.
* gcc.dg/debug/ctf/ctf-enum-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-enum-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-file-scope-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-float-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-forward-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-forward-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-func-index-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-function-pointers-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-function-pointers-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-function-pointers-3.c: Likewise.
* gcc.dg/debug/ctf/ctf-functions-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-int-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-objt-index-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-pointers-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-pointers-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-preamble-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-skip-types-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-skip-types-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-skip-types-3.c: Likewise.
* gcc.dg/debug/ctf/ctf-skip-types-4.c: Likewise.
* gcc.dg/debug/ctf/ctf-skip-types-5.c: Likewise.
* gcc.dg/debug/ctf/ctf-skip-types-6.c: Likewise.
* gcc.dg/debug/ctf/ctf-str-table-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-struct-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-struct-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-struct-array-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-struct-pointer-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-struct-pointer-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-typedef-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-typedef-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-typedef-3.c: Likewise.
* gcc.dg/debug/ctf/ctf-typedef-struct-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-typedef-struct-2.c: Likewise.
* gcc.dg/debug/ctf/ctf-typedef-struct-3.c: Likewise.
* gcc.dg/debug/ctf/ctf-union-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-variables-1.c: Likewise.
* gcc.dg/debug/ctf/ctf-variables-2.c: Likewise.
* gcc.dg/debug/ctf/ctf.exp: Likewise.

CTF/BTF debug formats

This commit introduces support for generating CTF debugging
information and BTF debugging information from GCC.

2021-06-28  Indu Bhagat  <indu.bhagat@oracle.com>
    David Faust  <david.faust@oracle.com>
    Jose E. Marchesi  <jose.marchesi@oracle.com>
    Weimin Pan  <weimin.pan@oracle.com>

gcc/

* Makefile.in: Add ctfc.*, ctfout.c and btfout.c files to
GTFILES.  Add new object files.
* common.opt: Add CTF and BTF debug info options.
* btfout.c: New file.
* ctfc.c: Likewise.
* ctfc.h: Likewise.
* ctfout.c: Likewise.
* dwarf2ctf.c: Likewise.
* dwarf2ctf.h: Likewise.
* dwarf2cfi.c (dwarf2out_do_frame): Acknowledge CTF_DEBUG and
BTF_DEBUG.
* dwarf2out.c (dwarf2out_source_line): Likewise.
(dwarf2out_finish): Skip emitting DWARF if CTF or BTF are to
be generated.
(debug_format_do_cu): New function.
(dwarf2out_early_finish): Traverse DIEs and emit CTF/BTF for
them if requested.
Include dwarf2ctf.c.
* final.c (dwarf2_debug_info_emitted_p): Acknowledge DWARF-based debug
formats.
* flag-types.h (enum debug_info_type): Add CTF_DEBUG and BTF_DEBUG.
(CTF_DEBUG): New bitmask.
(BTF_DEBUG): Likewise.
(enum ctf_debug_info_levels): New enum.
* gengtype.c (open_base_files): Handle ctfc.h.
(main): Handle uint32_t type.
* flags.h (btf_debuginfo_p): New definition.
(dwarf_based_debuginfo_p): Likewise.
* opts.c (debug_type_names): Add entries for CTF and BTF.
(btf_debuginfo_p): New function.
(dwarf_based_debuginfo_p): Likewise.
(common_handle_option): Handle -gctfN and -gbtf options.
(set_debug_level): Set CTF_DEBUG, BTF_DEBUG whenever appropriate.
* toplev.c (process_options): Inform the user and ignore -gctfLEVEL if
frontend is not C.

include/

* ctf.h: New file.
* btf.h: Likewise.

libiberty/

* simple-object.c (handle_lto_debug_sections): Copy over .ctf
sections.

dejagnu: modularize gcc-dg-debug-runtest a bit

Move some functionality into a procedure of its own. This is only so that when
the patch for ctf comes along, the gcc-dg-debug-runtest procedure looks bit
more uniform.

gcc/testsuite/ChangeLog:

* lib/gcc-dg.exp (gcc-dg-target-supports-debug-format): New procedure.

dwarf: externalize some DWARF internals for needs of CTF/BTF

This patch externalizes some internal DIE structures and their attributes
for the use of DWARF-based debug formats like CTF and BTF.

The following functions which were previously defined as static in
dwarf2out.c are now non-static, and extern prototypes for them have
been added to dwarf2out.h:

- get_AT
- AT_int
- AT_class
- AT_loc
- get_AT_ref
- get_AT_string
- get_AT_class
- AT_unsigned
- get_AT_unsigned
- get_AT_flag
- add_name_attribute
- new_die_raw
- base_type_die
- lookup_decl_die
- get_AT_file

Note how this patch doens't change the names of these functions to
avoid a massive renaming in dwarf2out.c, but in the future we probably
want these functions to sport a dw_* prefix.

Also, some type definitions have been moved from dwarf2out.c to
dwarf2out.h:

- dw_attr_node
- struct dwarf_file_data

Finally, three new accessor functions have been added to dwarf2out.c
with prototypes in dwarf2out.h:

- dw_get_die_child
- dw_get_die_sib
- dw_get_die_tag

2021-06-28  Jose E. Marchesi  <jose.marchesi@oracle.com>

* dwarf2out.c (AT_class): Function is no longer static.
(AT_int): Likewise.
(AT_unsigned): Likewise.
(AT_loc): Likewise.
(get_AT): Likewise.
(get_AT_string): Likewise.
(get_AT_flag): Likewise.
(get_AT_unsigned): Likewise.
(get_AT_ref): Likewise.
(new_die_raw): Likewise.
(lookup_decl_die): Likewise.
(base_type_die): Likewise.
(add_name_attribute): Likewise.
(add_AT_int): Likewise.
(add_AT_unsigned): Likewise.
(add_AT_loc): Likewise.
(dw_get_die_tag): New function.
(dw_get_die_child): Likewise.
(dw_get_die_sib): Likewise.
(struct dwarf_file_data): Move from here to dwarf2out.h
(struct dw_attr_struct): Likewise.
* dwarf2out.h: Analogous changes.

ipa-sra: Introduce a mini-DCE to tree-inline.c (PR 93385)

I was asked by Richi to split my fix for PR 93385 for easier review
into IPA-SRA materialization refactoring and the actual DCE addition.
This is the second part that actually contains the DCE of statements
that IPA-SRA should not leave behind because they can have problematic
side effects, even if they are useless, so that we do not depend on
tree-dce to remove them for correctness.

The patch fixes the problem by doing a def-use walk when materializing
clones, marking which statements should not be copied and which
SSA_NAMEs do not need to be computed because eventually they would be
DCEd.  We do this on the original function body and tree-inline simply
does not copy statements which are "dead."

The only complication is removing dead argument calls because that
needs to be communicated to callee redirection code using the
infrastructure introduced by the previous patch.

I added all testcases of the original patch to this one, although some
probably test behavior introduced in the previous patch.

gcc/ChangeLog:

2021-05-12  Martin Jambor  <mjambor@suse.cz>

PR ipa/93385
* ipa-param-manipulation.h (class ipa_param_body_adjustments): New
members m_dead_stmts and m_dead_ssas.
* ipa-param-manipulation.c
(ipa_param_body_adjustments::mark_dead_statements): New function.
(ipa_param_body_adjustments::common_initialization): Call it on
all removed but not split parameters.
(ipa_param_body_adjustments::ipa_param_body_adjustments): Initialize
new mwmbers.
(ipa_param_body_adjustments::modify_call_stmt): Remove arguments that
are dead.
* tree-inline.c (remap_gimple_stmt): Do not copy dead statements, reset
dead debug statements.
(copy_phis_for_bb): Do not copy dead PHI nodes.

gcc/testsuite/ChangeLog:

2021-03-22  Martin Jambor  <mjambor@suse.cz>

PR ipa/93385
* gcc.dg/ipa/pr93385.c: New test.
* gcc.dg/ipa/ipa-sra-23.c: Likewise.
* gcc.dg/ipa/ipa-sra-24.c: Likewise.
* g++.dg/ipa/ipa-sra-4.C: Likewise.

ipa-sra: Restructure how cloning and call redirection communicate (PR 93385)

I was asked by Richi to split my fix for PR 93385 for easier review
into IPA-SRA materialization refactoring and the actual DCE addition.
Fortunately it was mostly natural except for a temporary weird
condition in ipa_param_body_adjustments::modify_call_stmt.
Additionally.  In addition to the patch I posted previously, this one
also deallocated the newly added summary in toplev::finalize and fixes
a mistakenly uninitialized field.

This is the first part which basically replaces performed_splits in
clone_info and the code which generates it, keeps it up-to-date and
consumes it with new edge summaries which are much nicer.  It simply
contains 1) a mapping from the original argument indices to the actual
indices in the call statement as it is now, 2) information needed to
identify arguments representing pass-through IPA-SRA splits with which
have been added to the call arguments in place of an original
argument/reference and 3) a delta to the index where va_args may start
- so basically directly all the information that the consumer of
performed_splits had to compute and we also do not need the weird
dummy declarations.

The main disadvantage is that the information has to be created (and
kept up-to-date) for all call graph edges associated with the given
statement from all clones (including inline clones) of the clone where
splitting or removal happened first.  But all of this happens during
clone materialization so the only effect on WPA memory consumption is
the removal of a pointer from clone_info.

The statement modification code also has to know the statement from
the original function in order to be able to locate the edge summaries
which at this point are still keyed to these.  However, the code is
already quite heavily dependant on how things are structured in
tree-inline.c and in order to fix bugs like these it probably has to
be.

The subsequent patch needs this new information to be able to remove
arguments from calls during materialization and communicate this
information to the call redirection.

gcc/ChangeLog:

2021-05-17  Martin Jambor  <mjambor@suse.cz>

PR ipa/93385
* symtab-clones.h (clone_info): Removed member param_adjustments.
* ipa-param-manipulation.h: Adjust initial comment to reflect how we
deal with pass-through splits now.
(ipa_param_performed_split): Removed.
(ipa_param_adjustments::modify_call): Adjusted parameters.
(class ipa_param_body_adjustments): Adjusted parameters of
register_replacement, modify_gimple_stmt and modify_call_stmt.
(ipa_verify_edge_has_no_modifications): Declare.
(ipa_edge_modifications_finalize): Declare.
* cgraph.c (cgraph_edge::redirect_call_stmt_to_callee): Remove
performed_splits processing, pas only edge to padjs->modify_call,
check that call arguments were not modified if they should not have
been.
* cgraphclones.c (cgraph_node::create_clone): Do not copy performed
splits.
* ipa-param-manipulation.c (struct pass_through_split_map): New type.
(ipa_edge_modification_info): Likewise.
(ipa_edge_modification_sum): Likewise.
(ipa_edge_modifications): New edge summary.
(ipa_verify_edge_has_no_modifications): New function.
(transitive_split_p): Removed.
(transitive_split_map): Likewise.
(init_transitive_splits): Likewise.
(ipa_param_adjustments::modify_call): Adjusted to use the new edge
summary instead of performed_splits.
(ipa_param_body_adjustments::register_replacement): Drop dummy
parameter, set base_index of the created ipa_param_body_replacement.
(phi_arg_will_live_p): New function.
(ipa_param_body_adjustments::common_initialization): Do not create
IPA_SRA dummy decls.
(simple_tree_swap_info): Removed.
(remap_split_decl_to_dummy): Likewise.
(record_argument_state_1): New function.
(record_argument_state): Likewise.
(ipa_param_body_adjustments::modify_call_stmt): New parameter
orig_stmt.  Do not work with dummy decls, save necessary info about
changes to ipa_edge_modifications.
(ipa_param_body_adjustments::modify_gimple_stmt): New parameter
orig_stmt, pass it to modify_call_stmt.
(ipa_param_body_adjustments::modify_cfun_body): Adjust call to
modify_gimple_stmt.
(ipa_edge_modifications_finalize): New function.
* tree-inline.c (remap_gimple_stmt): Pass original statement to
modify_gimple_stmt.
(copy_phis_for_bb): Do not copy dead PHI nodes.
(expand_call_inline): Do not remap performed_splits.
(update_clone_info): Likewise.
* toplev.c: Include ipa-param-manipulation.h.
(toplev::finalize): Call ipa_edge_modifications_finalize.

libstdc++: Remove redundant explicit instantiations

These function templates are explicitly specialized for char and wchar_t
streambufs, so the explicit instantiations do nothing. Remove them, to
avoid confusion.

libstdc++-v3/ChangeLog:

* include/bits/streambuf.tcc (__copy_streambufs_eof): Remove
explicit instantiation declarations.
* src/c++11/streambuf-inst.cc (__copy_streambufs_eof): Remove
explicit instantiation definitions.

libstdc++: Fix backwards logic in 17_intro/names.cc test [PR 97088]

I meant to undef the names that clash with newlib headers for newlib,
but I only undef'd them for non-newlib targets. This means they still
cause errors for newlib, and aren't tested for other targets.

This fixes the test to check those names for non-newlib targets, and to
undef them to avoid errors for newlib.

libstdc++-v3/ChangeLog:

PR libstdc++/97088
* testsuite/17_intro/names.cc: Fix #if condition for names used
by newlib headers.

libstdc++: Allow unique_ptr<Incomplete[]>::operator[] [PR 101236]

PR libstdc++/101236 shows that LLVM depends on being able to use
unique_ptr<T[]>::operator[] when T is incomplete. This is undefined, but
previously worked with libstdc++. When I added the conditional noexcept
to that operator we started to diagnose the incomplete type.

This change restores support for that case, by making the noexcept
condition check that the type is complete before checking whether
indexing on the pointer can throw.  A workaround for PR c++/101239 is
needed to avoid a bogus error where G++ fails to do SFINAE on the
ill-formed p[n] expression and gets an ICE. Instead of checking that the
p[n] expression is valid in the trailing-return-type, we only check that
the element_type is complete.

libstdc++-v3/ChangeLog:

PR libstdc++/101236
* include/bits/unique_ptr.h (unique_ptr<T[], D>::operator[]):
Fail gracefully if element_type is incomplete.
* testsuite/20_util/unique_ptr/cons/incomplete.cc: Clarify that
the standard doesn't require this test to work for array types.
* testsuite/20_util/unique_ptr/lwg2762.cc: Check that incomplete
types can be used with array specialization.
* testsuite/20_util/unique_ptr/101236.cc: New test.

libstdc++: Implement LWG 415 for std::ws

For C++11 std::ws changed to be an unformatted input function, meaning
it constructs a sentry and sets badbit on exceptions.

libstdc++-v3/ChangeLog:

* doc/xml/manual/intro.xml: Document LWG 415 change.
* doc/html/manual/bugs.html: Regenerate.
* include/bits/istream.tcc (ws): Create sentry and catch
exceptions.
* testsuite/27_io/basic_istream/ws/char/lwg415.cc: New test.
* testsuite/27_io/basic_istream/ws/wchar_t/lwg415.cc: New test.

mklog: Handle correctly long lines.

contrib/ChangeLog:

* mklog.py: Handle correctly long lines.
* test_mklog.py: Test it.

Duplicate the range information of the phi onto the new ssa_name

Since match_simplify_replacement uses gimple_simplify, there is a new
ssa name created sometimes and then we go and replace the phi edge with
this new ssa name, the range information on the phi is lost.
Placing this in replace_phi_edge_with_variable is the best option instead
of doing it in each time replace_phi_edge_with_variable is called which is
what is done today.

OK? Bootstrapped and tested on x86_64-linux-gnu with no regressions.

gcc/ChangeLog:

* tree-ssa-phiopt.c (replace_phi_edge_with_variable): Duplicate range
info if we're the only things setting the target PHI.
(value_replacement): Don't duplicate range here.
(minmax_replacement): Likewise.

tree-optimization/101229 - fix vectorizer SLP hybrid detection with PHIs

This fixes the missing handling of PHIs in gimple_walk_op which causes
the new vectorizer SLP hybrid detection scheme to fail.

2021-06-28  Richard Biener  <rguenther@suse.de>

PR tree-optimization/101229
* gimple-walk.c (gimple_walk_op): Handle PHIs.

* gcc.dg/torture/pr101229.c: New testcase.

v850: silent 2 warnings

Silents:

/home/marxin/Programming/gcc/gcc/config/v850/v850.c: In function ‘char* construct_dispose_instruction(rtx)’:
/home/marxin/Programming/gcc/gcc/config/v850/v850.c:2690:22: warning: ‘%s’ directive writing up to 99 bytes into a region of size between 79 and 89 [-Wformat-overflow=]
 2690 |       sprintf (buff, "dispose %d {%s}, r31", stack_bytes / 4, regs);
      |                      ^~~~~~~~~~~~~~~~~~~~~~                   ~~~~
/home/marxin/Programming/gcc/gcc/config/v850/v850.c:2690:15: note: ‘sprintf’ output between 18 and 127 bytes into a destination of size 100
 2690 |       sprintf (buff, "dispose %d {%s}, r31", stack_bytes / 4, regs);
      |       ~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/marxin/Programming/gcc/gcc/config/v850/v850.c: In function ‘char* construct_prepare_instruction(rtx)’:
/home/marxin/Programming/gcc/gcc/config/v850/v850.c:2814:22: warning: ‘%s’ directive writing up to 99 bytes into a region of size 91 [-Wformat-overflow=]
 2814 |       sprintf (buff, "prepare {%s}, %d", regs, (- stack_bytes) / 4);
      |                      ^~~~~~~~~~~~~~~~~~  ~~~~
/home/marxin/Programming/gcc/gcc/config/v850/v850.c:2814:15: note: ‘sprintf’ output between 14 and 123 bytes into a destination of size 100
 2814 |       sprintf (buff, "prepare {%s}, %d", regs, (- stack_bytes) / 4);
      |       ~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

gcc/ChangeLog:

* config/v850/v850.c (construct_dispose_instruction): Allocate
a bigger buffer.
(construct_prepare_instruction): Likewise.

v850: add v850_can_inline_p target hook

gcc/ChangeLog:

* config/v850/v850.c (v850_option_override): Build default
target node.
(v850_can_inline_p): New.  Allow MASK_PROLOG_FUNCTION to be
ignored for inlining.
(TARGET_CAN_INLINE_P): New.

tree-optimization/101207 - fix BB reduc permute elide with life stmts

This fixes breakage of live lane extracts from permuted loads we elide
from BB reduction vectorization by handling the un-permuting the same
as in the regular eliding code - apply the reverse permute to
both the scalar stmts and the load permutation.

2021-06-28  Richard Biener  <rguenther@suse.de>

PR tree-optimization/101207
* tree-vect-slp.c (vect_optimize_slp): Do BB reduction
permute eliding for load permutations properly.

* gcc.dg/vect/bb-slp-pr101207.c: New testcase.

tree-optimization/101173 - fix interchange dependence checking

This adjusts the loop interchange dependence checking to disallow
an outer loop dependence distance of zero.

2021-06-23  Richard Biener  <rguenther@suse.de>

PR tree-optimization/101173
* gimple-loop-interchange.cc
(tree_loop_interchange::valid_data_dependences): Disallow outer
loop dependence distance of zero.

* gcc.dg/torture/pr101173.c: New testcase.

For 128/256-bit vec_cond_expr, When mask operands is lt reg const0_rtx, blendv can be used instead of avx512 mask.

gcc/ChangeLog:

PR target/100648
* config/i386/sse.md (*avx_cmp<mode>3_lt): New
define_insn_and_split.
(*avx_cmp<mode>3_ltint): Ditto.
(*avx2_pcmp<mode>3_3): Ditto.
(*avx2_pcmp<mode>3_4): Ditto.
(*avx2_pcmp<mode>3_5): Ditto.

gcc/testsuite/ChangeLog:

PR target/100648
* g++.target/i386/avx2-pr54700-2.C: Adjust testcase.
* g++.target/i386/avx512vl-pr54700-1a.C: New test.
* g++.target/i386/avx512vl-pr54700-1b.C: New test.
* g++.target/i386/avx512vl-pr54700-2a.C: New test.
* g++.target/i386/avx512vl-pr54700-2b.C: New test.
* gcc.target/i386/avx512vl-pr100648.c: New test.
* gcc.target/i386/avx512vl-blendv-1.c: New test.
* gcc.target/i386/avx512vl-blendv-2.c: New test.

Fold blendv builtins into gimple.

Fold __builtin_ia32_pblendvb128 (a, b, c) as VEC_COND_EXPR (c < 0, b,
a), similar for float version but with mask operand VIEW_CONVERT_EXPR
to same sized integer vectype.

gcc/ChangeLog:

* config/i386/i386-builtin.def (IX86_BUILTIN_BLENDVPD256,
IX86_BUILTIN_BLENDVPS256, IX86_BUILTIN_PBLENDVB256,
IX86_BUILTIN_BLENDVPD, IX86_BUILTIN_BLENDVPS,
IX86_BUILTIN_PBLENDVB128): Replace icode with
CODE_FOR_nothing.
* config/i386/i386.c (ix86_gimple_fold_builtin): Fold blendv
builtins.
* config/i386/sse.md (*<sse4_1_avx2>_pblendvb_lt_subreg_not):
New pre_reload splitter.

gcc/testsuite/ChangeLog:

* gcc.target/i386/funcspec-8.c: Replace
__builtin_ia32_blendvpd with  __builtin_ia32_roundps_az.
* gcc.target/i386/blendv-1.c: New test.
* gcc.target/i386/blendv-2.c: New test.

Daily bump.

Fix PR 101230: ICE in fold_cond_expr_with_comparison

This fixes PR 101230 where I had messed up and forgot that
invert_tree_comparison can return ERROR_MARK if the comparsion
is not invertable (floating point types).

Committed as obvious after a bootstrap/test on x86_64-linux-gnu-gnu

gcc/ChangeLog:

PR middle-end/101230
* fold-const.c (fold_ternary_loc): Check
the return value of invert_tree_comparison.

gcc/testsuite/ChangeLog:

* gcc.dg/torture/pr101230-1.c: New test.

aix: Add AIX 7.3 configuration and SPDX License Identifiers.

The anticipated release of AIX 7.3 has been announced.  This
patch adds the configuration bits based on AIX 7.2 configuration.

gcc/ChangeLog:

* config.gcc: Add SPDX License Identifier.
(powerpc-ibm-aix789): Default to aix73.h.
(powerpc-ibm-aix7.2.*.*): New stanza.
* config/rs6000/aix72.h: Add SPDX License Identifier.
* config/rs6000/aix73.h: New file.

Daily bump.

c++: access scope during partial spec matching [PR96204]

Here, when determining whether the partial specialization matches
has_type_member<Child>, we do so from the scope of where the template-id
appears rather than from the scope of the specialization, and this
causes us to select the partial specialization (since Child::type is
accessible from Parent).  When we later instantiate this partial
specialization, we've entered the scope of the specialization and so
substitution into e.g. the DECL_CONTEXT of has_type_member::value fails
with access errors since the friend declaration that we relied on to
choose the partial specialization no longer applies.

It seems the appropriate access scope from which to perform partial
specialization matching is the specialization itself (similar to how
we check access of base-clauses), which is what this patch implements.

PR c++/96204

gcc/cp/ChangeLog:

* pt.c (instantiate_class_template_1): Enter the scope of the
type when calling most_specialized_partial_spec.

gcc/testsuite/ChangeLog:

* g++.dg/template/access40.C: New test.
* g++.dg/template/access40a.C: New test.

except: remove dwarf2out.h dependency

When thinking about the CTF debug patchset dwarf2out.h split, I noticed that
except.c only needs macros from dwarf2.h, nothing from dwarf2out.h.

gcc/ChangeLog:

* except.c: #include "dwarf2.h" instead of "dwarf2out.h".

c++: constexpr aggr init of empty class [PR101040]

This is basically the aggregate initializer version of PR97566; as in that
bug, we are trying to initialize empty field 'obj' in 'single' when there's
no CONSTRUCTOR entry for the 'single' base class subobject of 'derived'.  As
with that bug, the fix is to stop trying to add entries for empty fields,
this time in cxx_eval_bare_aggregate.

The change to the other function isn't necessary for this version of
the patch, but seems worthwhile for robustness anyway.

PR c++/101040
PR c++/97566

gcc/cp/ChangeLog:

* class.c (is_empty_field): Handle null argument.
* constexpr.c (cxx_eval_bare_aggregate): Discard initializer
for empty field.

gcc/testsuite/ChangeLog:

* g++.dg/cpp2a/no_unique_address13.C: New test.

Lower for loops before lowering cond in genmatch

While converting some fold_cond_expr_with_comparison
to match, I found that I wanted to use "for cnd (cond vec_cond)"
but that was not causing the lowering of cond to happen.
What was happening was the lowering of the for loop
was happening after the lowering of the cond. So
swapping was the correct thing to do but it also
means we need to copy for_subst_vec in lower_cond.

OK?  Bootstrapped and tested on x86_64-linux-gnu with no regressions.

gcc/ChangeLog:

* genmatch.c (lower_cond): Copy for_subst_vec
for the simplify also.
(lower): Swap the order for lower_for and lower_cond.

Reset the range info on the moved instruction in PHIOPT

I had missed this when wrote the patch which allowed the
gimple to be moved from inside the conditional as it.  It
was also missed in the review.  Anyways the range information
needs to be reset for the moved gimple as it was under a
conditional and the flow has changed to be unconditional.
I have not seen any testcase in the wild that produces wrong code
yet which is why there is no testcase but this is similar to what
the other code in phiopt does so after moving those to match, there
might be some.

OK? Bootstrapped and tested on x86_64-linux-gnu with no regressions.

gcc/ChangeLog:

* tree-ssa-phiopt.c (match_simplify_replacement): Reset
flow senatitive info on the moved ssa set.

Expand the comparison argument of fold_cond_expr_with_comparison

To make things slightly easiler to convert fold_cond_expr_with_comparison
over to match.pd, expanding the arg0 argument into 3 different arguments
is done. Also this was simple because we don't use arg0 after grabbing
the code and the two operands.
Also since we do this, we don't need to fold the comparison to
get the inverse but just use invert_tree_comparison directly.

OK? Bootstrapped and tested on x86_64-linux-gnu with no regressions.

gcc/ChangeLog:

* fold-const.c (fold_cond_expr_with_comparison):
Exand arg0 into comp_code, arg00, and arg01.
(fold_ternary_loc): Use invert_tree_comparison
instead of fold_invert_truthvalue for the case
where we have A CMP B ? C : A.

Daily bump.

c++: Failure to delay noexcept parsing with ptr-operator [PR100752]

We weren't passing 'flags' to the recursive call to cp_parser_declarator
in the ptr-operator case and as an effect, delayed parsing of noexcept
didn't work as advertised.  The following change passes more than just
CP_PARSER_FLAGS_DELAY_NOEXCEPT but that doesn't seem to break anything.

I'm now also passing member_p and static_p, as a consequence, two tests
needed small tweaks.

PR c++/100752

gcc/cp/ChangeLog:

* parser.c (cp_parser_declarator): Pass flags down to
cp_parser_declarator.  Also pass static_p/member_p.

gcc/testsuite/ChangeLog:

* g++.dg/cpp0x/noexcept69.C: New test.
* g++.dg/parse/saved1.C: Adjust dg-error.
* g++.dg/template/crash50.C: Likewise.

jit: fix test-vector-* failures

Fix failures seen on i686 due to relying on exact floating-point
equality when testing results of vector division.

gcc/testsuite/ChangeLog:
* jit.dg/test-vector-rvalues.cc (check_div): Add specialization
for v4f, to avoid relying on exact floating-point equality.
* jit.dg/test-vector-types.cc (check_div): Likewise.

Signed-off-by: David Malcolm <dmalcolm@redhat.com>

jit: fix test-asm failures on i?86

On i686, test_i386_basic_asm_4 has:
  error: inconsistent operand constraints in an 'asm'
and test_i386_basic_asm_5 has:
  /tmp/libgccjit-9FsLie/fake.s:9: Error: bad register name `%rdi'
  /tmp/libgccjit-9FsLie/fake.s:10: Error: bad register name `%rsi'

This is only intended as a smoketest of asm support, so only run
it on x86_64.

gcc/testsuite/ChangeLog:
* jit.dg/test-asm.c: Remove i?86-*-* from target specifier.
* jit.dg/test-asm.cc: Likewise.

Signed-off-by: David Malcolm <dmalcolm@redhat.com>

PR middle-end/101216 - spurious notes for function calls

PR middle-end/101216

gcc/ChangeLog:

* calls.c (maybe_warn_rdwr_sizes): Use the no_warning constant.

gcc/testsuite/ChangeLog:

* gcc.dg/Wnonnull-7.c: New test.

libstdc++: Avoid intercepting exception in ostream::write

Currently if ostream::write fails and sets badbit and that causes an
exception, we will catch the exception, set badbit again, and rethrow
the exception.

This change delays setting badbit until after the try-catch block, so
that if it causes an exception we don't need to catch and rethrow it.

This removes the last remaining use of _M_write, so it can be made
private (or removed entirely for versioned namespace builds, where ABI
compatibility is not required). All other uses of _M_write were replaced
by calls to __ostream_insert, so make _M_write use that too.

libstdc++-v3/ChangeLog:

* include/bits/ostream.tcc (basic_ostream::write): Call sputn
directly instead of using _M_write. Do setstate(__err) all
outside the try-catch block.
* include/std/ostream (basic_ostream::_M_write): Declare
private. Use __ostream_insert. Do not define for the versioned
namespace.

libstdc++: Implement LWG 581 for std:ostream::flush()

LWG 581 changed ostream::flush() to an unformatted output function for
C++11, but it was never implemented in libstdc++.

libstdc++-v3/ChangeLog:

* doc/xml/manual/intro.xml: Document LWG 581 change.
* doc/html/manual/bugs.html: Regenerate.
* include/bits/basic_ios.tcc: Whitespace.
* include/bits/ostream.tcc (basic_ostream::flush()): Construct
sentry.
* testsuite/27_io/basic_ostream/flush/char/2.cc: Check
additional cases.
* testsuite/27_io/basic_ostream/flush/char/exceptions_badbit_throw.cc:
Likewise.
* testsuite/27_io/basic_ostream/flush/wchar_t/2.cc: Likewise.
* testsuite/27_io/basic_ostream/flush/wchar_t/exceptions_badbit_throw.cc:
Likewise.

libstdc++: Fix exception handling in std::ostream seek functions

N3168 added the requirement that the [ostream.seeks] functions create a
sentry object. Nothing in the requirements of those functions says
anything about catching exceptions and setting badbit.

As well as not catching exceptions, this change results in another
observable behaviour change. Previously seeking on a stream with eofbit
set would work (as long as badbit and failbit weren't set). The
construction of a sentry causes failbit to be set when eofbit is set,
which causes the seek to fail. It is necessary to clear the eofbit
before seeking now.

libstdc++-v3/ChangeLog:

* include/bits/ostream.tcc (sentry): Only set failbit if badbit
is set, not if eofbit is set.
(tellp, seekp, seekp): Create sentry object. Do not set badbit
on exceptions.
* testsuite/27_io/basic_ostream/seekp/char/exceptions_badbit_throw.cc:
Adjust expected behaviour.
* testsuite/27_io/basic_ostream/seekp/wchar_t/exceptions_badbit_throw.cc:
Likewise.
* testsuite/27_io/basic_ostream/tellp/char/exceptions_badbit_throw.cc:
Likewise.
* testsuite/27_io/basic_ostream/tellp/wchar_t/exceptions_badbit_throw.cc:
Likewise.
* testsuite/27_io/basic_ostream/seekp/char/n3168.cc: New test.
* testsuite/27_io/basic_ostream/seekp/wchar_t/n3168.cc: New test.
* testsuite/27_io/basic_ostream/tellp/char/n3168.cc: New test.
* testsuite/27_io/basic_ostream/tellp/wchar_t/n3168.cc: New test.

libstdc++: Remove noexcept from syncbuf::swap (LWG 3498)

The proposed resolution for the inconsistent noexcept-specifiers in the
spec is to remove it from bto hthe assignment operator and swap.

libstdc++-v3/ChangeLog:

* include/std/syncstream (basic_syncbuf::swap()): Remove
noexcept, as per LWG 3498.

libstdc++: More workarounds in 17_intro/names.cc test [PR 97088]

Conditionally #undef some more names that are used in system headers.

libstdc++-v3/ChangeLog:

PR libstdc++/97088
* testsuite/17_intro/names.cc: Undef more names for newlib and
also for arm-none-linux-gnueabi.
* testsuite/experimental/names.cc: Disable PCH.

testsuite/101114: Adjust libgomp.c-c++-common/struct-elem-5.c testcase

The dg-shouldfail testcase libgomp.c-c++-common/struct-elem-5.c does not
properly fail for non-shared address space offloading. Adjust testcase
to limit testing only for "target offload_device_nonshared_as".

libgomp/ChangeLog:

PR testsuite/101114
* testsuite/libgomp.c-c++-common/struct-elem-5.c:
Add "target offload_device_nonshared_as" condition for enabling test.

libstdc++: Make use of __builtin_bit_cast for simd

The __bit_cast function was a hack to achieve what __builtin_bit_cast
can do, therefore use __builtin_bit_cast if possible. However,
__builtin_bit_cast cannot be used to cast from/to fixed_size_simd, since
it isn't trivially copyable (in the language sense — in principle it
is). Therefore add __proposed::simd_bit_cast to enable the use case
required in the test framework.

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd.h (__bit_cast): Implement via
__builtin_bit_cast #if available.
(__proposed::simd_bit_cast): Add overloads for simd and
simd_mask, which use __builtin_bit_cast (or __bit_cast #if not
available), which return an object of the requested type with
the same bits as the argument.
* include/experimental/bits/simd_math.h: Use simd_bit_cast
instead of __bit_cast to allow casts to fixed_size_simd.
(copysign): Remove branch that was only required if __bit_cast
cannot be constexpr.
* testsuite/experimental/simd/tests/bits/test_values.h: Switch
from __bit_cast to __proposed::simd_bit_cast since the former
will not cast fixed_size objects anymore.

MAINTAINERS: Add myself for write after approval and DCO

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
ChangeLog:

* MAINTAINERS: Add myself for write after approval and DCO

Use right shifts to eliminate redundant test/compare insns on the H8

gcc/
* config/h8300/h8300.c (select_cc_mode): Handle ASHIFTRT and LSHIFTRT.

tree-optimization/101202 - fix ICE with failed backedge SLP nodes

This fixes an ICE with failed backedge SLP nodes still in the graph
while doing permute optimization by explicitely handling those.

2021-06-25  Richard Biener  <rguenther@suse.de>

PR tree-optimization/101202
* tree-vect-slp.c (vect_optimize_slp): Explicitely handle
failed nodes.

* gcc.dg/torture/pr101202.c: New testcase.

Fixup reduction info on addsub SLP pattern

gcc.dg/vect/pr96854.c shows we need to copy over reduction info
to the SLP pattern as already done for the complex patterns.

2021-06-25  Richard Biener  <rguenther@suse.de>

* tree-vect-slp-patterns.c (addsub_pattern::build): Copy
STMT_VINFO_REDUC_DEF from the original representative.

add -ltrans-objects lto-plugin debug option

This adds a -ltrans-objects option to lto-plugin that by-passes
lto-wrapper invocation and instead feeds LD the final LTRANS objects
directly from the response file given as argument to the option.
This allows LD issues involving the linker-plugin path to be
debugged in an easier way with just the IR objects (their symtab)
and the LTRANS objects as testcase.

I've tested the path re-building stage2 build/genmatch from an
LTO bootstrap and got a bit-identical executable by adding
-plugin-opt=-ltrans-objects=y to the original collect2 invocation,
seeding y with the final objects as printed by building genmatch
with -save-temps -v.

2021-06-22  Richard Biener  <rguenther@suse.de>

lto-plugin/
* lto-plugin.c (ltrans_objects): New global.
(all_symbols_read_handler): If -ltrans-objects was specified,
add the output files from the specified file directly.
(process_option): Handle -ltrans-objects.

testsuite: avoid no-stack-protector-attr-3 fail on mips*-*-*

On MIPS a call to __stack_chk_fail needs an additional .reloc pseudo-op,
so "stack_chk_fail" will appear two times.

gcc/testsuite/

* g++.dg/no-stack-protector-attr-3.C (dg-final): Adjust for MIPS.

middle-end: add support for per-location warning groups.

gcc/ChangeLog:

* builtins.c (warn_string_no_nul): Replace uses of TREE_NO_WARNING,
gimple_no_warning_p and gimple_set_no_warning with
warning_suppressed_p, and suppress_warning.
(c_strlen): Same.
(maybe_warn_for_bound): Same.
(warn_for_access): Same.
(check_access): Same.
(expand_builtin_strncmp): Same.
(fold_builtin_varargs): Same.
* calls.c (maybe_warn_nonstring_arg): Same.
(maybe_warn_rdwr_sizes): Same.
* cfgexpand.c (expand_call_stmt): Same.
* cgraphunit.c (check_global_declaration): Same.
* fold-const.c (fold_undefer_overflow_warnings): Same.
(fold_truth_not_expr): Same.
(fold_unary_loc): Same.
(fold_checksum_tree): Same.
* gimple-array-bounds.cc (array_bounds_checker::check_array_ref): Same.
(array_bounds_checker::check_mem_ref): Same.
(array_bounds_checker::check_addr_expr): Same.
(array_bounds_checker::check_array_bounds): Same.
* gimple-expr.c (copy_var_decl): Same.
* gimple-fold.c (gimple_fold_builtin_strcpy): Same.
(gimple_fold_builtin_strncat): Same.
(gimple_fold_builtin_stxcpy_chk): Same.
(gimple_fold_builtin_stpcpy): Same.
(gimple_fold_builtin_sprintf): Same.
(fold_stmt_1): Same.
* gimple-ssa-isolate-paths.c (diag_returned_locals): Same.
* gimple-ssa-nonnull-compare.c (do_warn_nonnull_compare): Same.
* gimple-ssa-sprintf.c (handle_printf_call): Same.
* gimple-ssa-store-merging.c (imm_store_chain_info::output_merged_store): Same.
* gimple-ssa-warn-restrict.c (maybe_diag_overlap): Same.
* gimple-ssa-warn-restrict.h: Adjust declarations.
(maybe_diag_access_bounds): Replace uses of TREE_NO_WARNING,
gimple_no_warning_p and gimple_set_no_warning with
warning_suppressed_p, and suppress_warning.
(check_call): Same.
(check_bounds_or_overlap): Same.
* gimple.c (gimple_build_call_from_tree): Same.
* gimplify.c (gimplify_return_expr): Same.
(gimplify_cond_expr): Same.
(gimplify_modify_expr_complex_part): Same.
(gimplify_modify_expr): Same.
(gimple_push_cleanup): Same.
(gimplify_expr): Same.
* omp-expand.c (expand_omp_for_generic): Same.
(expand_omp_taskloop_for_outer): Same.
* omp-low.c (lower_rec_input_clauses): Same.
(lower_lastprivate_clauses): Same.
(lower_send_clauses): Same.
(lower_omp_target): Same.
* tree-cfg.c (pass_warn_function_return::execute): Same.
* tree-complex.c (create_one_component_var): Same.
* tree-inline.c (remap_gimple_op_r): Same.
(copy_tree_body_r): Same.
(declare_return_variable): Same.
(expand_call_inline): Same.
* tree-nested.c (lookup_field_for_decl): Same.
* tree-sra.c (create_access_replacement): Same.
(generate_subtree_copies): Same.
* tree-ssa-ccp.c (pass_post_ipa_warn::execute): Same.
* tree-ssa-forwprop.c (combine_cond_expr_cond): Same.
* tree-ssa-loop-ch.c (ch_base::copy_headers): Same.
* tree-ssa-loop-im.c (execute_sm): Same.
* tree-ssa-phiopt.c (cond_store_replacement): Same.
* tree-ssa-strlen.c (maybe_warn_overflow): Same.
(handle_builtin_strcpy): Same.
(maybe_diag_stxncpy_trunc): Same.
(handle_builtin_stxncpy_strncat): Same.
(handle_builtin_strcat): Same.
* tree-ssa-uninit.c (get_no_uninit_warning): Same.
(set_no_uninit_warning): Same.
(uninit_undefined_value_p): Same.
(warn_uninit): Same.
(maybe_warn_operand): Same.
* tree-vrp.c (compare_values_warnv): Same.
* vr-values.c (vr_values::extract_range_for_var_from_comparison_expr): Same.
(test_for_singularity): Same.

* gimple.h (warning_suppressed_p): New function.
(suppress_warning): Same.
(copy_no_warning): Same.
(gimple_set_block): Call gimple_set_location.
(gimple_set_location): Call copy_warning.

cp: add support for per-location warning groups.

gcc/cp/ChangeLog:

* call.c (build_over_call): Replace direct uses of TREE_NO_WARNING
with warning_suppressed_p, suppress_warning, and copy_no_warning, or
nothing if not necessary.
(set_up_extended_ref_temp): Same.
* class.c (layout_class_type): Same.
* constraint.cc (constraint_satisfaction_value): Same.
* coroutines.cc (finish_co_await_expr): Same.
(finish_co_yield_expr): Same.
(finish_co_return_stmt): Same.
(build_actor_fn): Same.
(coro_rewrite_function_body): Same.
(morph_fn_to_coro): Same.
* cp-gimplify.c (genericize_eh_spec_block): Same.
(gimplify_expr_stmt): Same.
(cp_genericize_r): Same.
(cp_fold): Same.
* cp-ubsan.c (cp_ubsan_instrument_vptr): Same.
* cvt.c (cp_fold_convert): Same.
(convert_to_void): Same.
* decl.c (wrapup_namespace_globals): Same.
(grokdeclarator): Same.
(finish_function): Same.
(require_deduced_type): Same.
* decl2.c (no_linkage_error): Same.
(c_parse_final_cleanups): Same.
* except.c (expand_end_catch_block): Same.
* init.c (build_new_1): Same.
(build_new): Same.
(build_vec_delete_1): Same.
(build_vec_init): Same.
(build_delete): Same.
* method.c (defaultable_fn_check): Same.
* parser.c (cp_parser_fold_expression): Same.
(cp_parser_primary_expression): Same.
* pt.c (push_tinst_level_loc): Same.
(tsubst_copy): Same.
(tsubst_omp_udr): Same.
(tsubst_copy_and_build): Same.
* rtti.c (build_if_nonnull): Same.
* semantics.c (maybe_convert_cond): Same.
(finish_return_stmt): Same.
(finish_parenthesized_expr): Same.
(cp_check_omp_declare_reduction): Same.
* tree.c (build_cplus_array_type): Same.
* typeck.c (build_ptrmemfunc_access_expr): Same.
(cp_build_indirect_ref_1): Same.
(cp_build_function_call_vec): Same.
(warn_for_null_address): Same.
(cp_build_binary_op): Same.
(unary_complex_lvalue): Same.
(cp_build_modify_expr): Same.
(build_x_modify_expr): Same.
(convert_for_assignment): Same.

c-family: add support for per-location warning groups.

gcc/c-family/ChangeLog:

* c-common.c (c_wrap_maybe_const): Remove TREE_NO_WARNING.
(c_common_truthvalue_conversion): Replace direct uses of
TREE_NO_WARNING with warning_suppressed_p, suppress_warning, and
copy_no_warning.
(check_function_arguments_recurse): Same.
* c-gimplify.c (c_gimplify_expr): Same.
* c-warn.c (overflow_warning): Same.
(warn_logical_operator): Same.
(warn_if_unused_value): Same.
(do_warn_unused_parameter): Same.

c: add support for per-location warning groups.

gcc/ChangeLog:
* tree.h (warning_suppressed_at, copy_warning,
warning_suppressed_p, suppress_warning): New functions.

gcc/c/ChangeLog:

* c-decl.c (pop_scope): Replace direct uses of TREE_NO_WARNING with
warning_suppressed_p, suppress_warning, and copy_no_warning.
(diagnose_mismatched_decls): Same.
(duplicate_decls): Same.
(grokdeclarator): Same.
(finish_function): Same.
(c_write_global_declarations_1): Same.
* c-fold.c (c_fully_fold_internal): Same.
* c-parser.c (c_parser_expr_no_commas): Same.
(c_parser_postfix_expression): Same.
* c-typeck.c (array_to_pointer_conversion): Same.
(function_to_pointer_conversion): Same.
(default_function_array_conversion): Same.
(convert_lvalue_to_rvalue): Same.
(default_conversion): Same.
(build_indirect_ref): Same.
(build_function_call_vec): Same.
(build_atomic_assign): Same.
(build_unary_op): Same.
(c_finish_return): Same.
(emit_side_effect_warnings): Same.
(c_finish_stmt_expr): Same.
(c_omp_clause_copy_ctor): Same.

Add support for per-location warning groups.

gcc/ChangeLog:

* Makefile.in (OBJS-libcommon): Add diagnostic-spec.o.
* gengtype.c (open_base_files): Add diagnostic-spec.h.
* diagnostic-spec.c: New file.
* diagnostic-spec.h: New file.
* tree.h (no_warning, all_warnings, suppress_warning_at): New
declarations.
* warning-control.cc: New file.

Revert x86_order_regs_for_local_alloc changes in r12-1669.

Still put general regs as first alloca order.

gcc/ChangeLog:

PR target/101185
* config/i386/i386.c (x86_order_regs_for_local_alloc):
Revert r12-1669.

gcc/testsuite/ChangeLog

PR target/101185
* gcc.target/i386/bitwise_mask_op-3.c: Add xfail to
temporarily avoid regression, eventually xfail should be
removed.

Daily bump.

Add a testcase to confirm the equivalence's are being checked by EVRP.

* gcc.dg/tree-ssa/evrp30.c: New.

Only register relations on live edges

Register a relation on a conditional edge only if the LHS supports
this edge being taken.

gcc/
PR tree-optimization/101189
* gimple-range-fold.cc (fold_using_range::range_of_range_op): Pass
LHS range of condition to postfold routine.
(fold_using_range::postfold_gcond_edges): Only process the TRUE or
FALSE edge if the LHS range supports it being taken.
* gimple-range-fold.h (postfold_gcond_edges): Add range parameter.

gcc/testsuite/
* gcc.dg/tree-ssa/pr101189.c: New.

Fix relation query of equivalences.

When looking for relations between equivalencies, a typo was causing
the wrong bitmap to be checked. Effect was is missed them.
Plus don't dump blocks which don't exist.

* value-relation.cc (equiv_oracle::dump): Do not dump NULL blocks.
(relation_oracle::find_relation_block): Check correct bitmap.
(relation_oracle::dump): Do not dump NULL blocks.

Correctly unify recomputation with existing range.

When propagating the on-entry cache, new block ranges are calculated
by combining all the incoming edges and comparing to the old value.
When a recomputation was performed on an edge, it didn't take into account
that the value in the block may already be better than a potential
recompuation... Thus a worse values was sometimes propagated.
Fixed by simply calling the now correct range_on_edge the cache provides.

* gimple-range-cache.cc (ranger_cache::propagate_cache): Call
range_on_edge instead of manually calculating.

Fix comment typo.

* range-op.cc: Fix comment.

c++: alias CTAD and aggregate deduction cand [PR98832]

During alias CTAD, we're accidentally ignoring the aggregate deduction
candidate for the underlying template because this guide is added
separately via maybe_aggr_guide (which doesn't yet handle alias
templates) instead of via deduction_guides_for (which does).  This patch
makes maybe_aggr_guide handle alias templates in a manner similar to
deduction_guides_for.

PR c++/98832

gcc/cp/ChangeLog:

* pt.c (maybe_aggr_guide): Handle alias templates appropriately.

gcc/testsuite/ChangeLog:

* g++.dg/cpp2a/class-deduction-alias9.C: New test.

c++: requires-expression folding [PR101182]

Here we're crashing because cp_fold_function walks into the (templated)
requirements of a requires-expression outside a template, but the
folding routines aren't prepared to handle templated trees.  This patch
fixes this by making cp_fold use evaluate_requires_expr to fold a
requires-expression as a whole, which also means we no longer need to
explicitly do so during gimplification.  (Note that we delay folding
of such requires-expressions for sake of better diagnostics when one is
used as the condition of a failed static_assert.)

PR c++/101182

gcc/cp/ChangeLog:

* constraint.cc (evaluate_requires_expr): Adjust function comment.
* cp-gimplify.c (cp_genericize_r) <case REQUIRES_EXPR>: Move to ...
(cp_fold) <case REQUIRES_EXPR>: ... here.

gcc/testsuite/ChangeLog:

* g++.dg/cpp2a/concepts-requires25.C: New test.

c: Fix up c_parser_has_attribute_expression [PR101176]

This function keeps src_range member of the result uninitialized, which at
least under valgrind can show up later when those uninitialized location_t's
can make it into the IL or location_t hash tables.

2021-06-24  Jakub Jelinek  <jakub@redhat.com>

PR c/101176
* c-parser.c (c_parser_has_attribute_expression): Set source range for
the result.

c: Fix C cast error-recovery [PR101171]

The following testcase ICEs during error-recovery, as build_c_cast calls
note_integer_operands on error_mark_node and that wraps it into
C_MAYBE_CONST_EXPR which is unexpected and causes ICE later on.
Seems most other callers of note_integer_operands check early if something
is error_mark_node and return before calling note_integer_operands on it.

The following patch fixes it by not calling on error_mark_node, another
possibility would be to handle error_mark_node in note_integer_operands and
just return it.

2021-06-24  Jakub Jelinek  <jakub@redhat.com>

PR c/101171
* c-typeck.c (build_c_cast): Don't call note_integer_operands on
error_mark_node.

* gcc.dg/pr101171.c: New test.

i386: Add pack/unpack patterns for 64bit vectors [PR89021]

2021-06-24  Uroš Bizjak  <ubizjak@gmail.com>

gcc/
PR target/89021
* config/i386/i386-expand.c (ix86_expand_sse_unpack):
Handle V8QI and V4HI modes.
* config/i386/mmx.md (sse4_1_<any_extend:code>v4qiv4hi2):
New insn pattern.
(sse4_1_<any_extend:code>v4qiv4hi2): Ditto.
(mmxpackmode): New mode attribute.
(vec_pack_trunc_<mmxpackmode:mode>): New expander.
(mmxunpackmode): New mode attribute.
(vec_unpacks_lo_<mmxunpackmode:mode>): New expander.
(vec_unpacks_hi_<mmxunpackmode:mode>): Ditto.
(vec_unpacku_lo_<mmxunpackmode:mode>): Ditto.
(vec_unpacku_hi_<mmxunpackmode:mode>): Ditto.
* config/i386/i386.md (extsuffix): Move from ...
* config/i386/sse.md: ... here.

gcc/testsuite/

PR target/89021
* gcc.dg/vect/vect-nb-iter-ub-3.c (dg-additional-options):
Add --param vect-epilogues-nomask=0.
* gcc.target/i386/pr97249-1.c (foo): Add #pragma GCC unroll
to avoid loop vectorization.
(foo1): Ditto.
(foo2): Ditto.

libstdc++: Fix internal names: add missing underscores

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd_math.h
(_GLIBCXX_SIMD_MATH_CALL2_): Rename arg2_ to __arg2.
(_GLIBCXX_SIMD_MATH_CALL3_): Rename arg2_ to __arg2 and arg3_ to
__arg3.

libstdc++: Ensure unrolled loops inline the lambda

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd.h (__execute_on_index_sequence)
(__execute_on_index_sequence_with_return)
(__call_with_n_evaluations, __call_with_subscripts): Add flatten
attribute.

libstdc++: Avoid raising fp exceptions in trunc, floor, and ceil

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd_x86.h (_S_trunc, _S_floor)
(_S_ceil): Set bit 8 (_MM_FROUND_NO_EXC) on AVX and SSE4.1
roundp[sd] calls.

libstdc++: Fix condition when AVX512F ldexp implementation is used

This improves codegen of ldexp if AVX512VL is available.

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd_x86.h (_S_ldexp): The AVX512F
implementation doesn't require a _VecBltnBtmsk ABI tag, it
requires either a 64-Byte input (in which case AVX512F must be
available) or AVX512VL.

libstdc++: Minor simd_math cleanups

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd_math.h: Undefine internal
macros after use.
(frexp): Move #if to a more sensible position and reformat
preceding code.
(logb): Call _SimdImpl::_S_logb for fixed_size instead of
duplicating the code here.
(modf): Simplify condition.

libstdc++: Remove incorrect fabs(simd) overload

fabs(int) returns double, this one didn't. This overload is not
specified in the Parallelism TS 2. Also remove the comment about labs
and llabs: it doesn't belong here.

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd_math.h (fabs): Remove
fabs(simd<integral>) overload.

libstdc++: Improve simd fixed_size codegen

Sometimes fixed_size objects will get unnecessarily copied on the stack.
The simd implementation should never pass _SimdTuple by value to avoid
requiring the optimizer to see through these copies.

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd_converter.h
(_SimdConverter::operator()): Pass _SimdTuple by const-ref.
* include/experimental/bits/simd_fixed_size.h
(_GLIBCXX_SIMD_FIXED_OP): Pass binary operator _SimdTuple
arguments by const-ref.
(_S_masked_unary): Pass _SimdTuple by const-ref.

libstdc++: Remove dead code in simd

This helper type became unused at some point.

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd_fixed_size.h
(_AbisInSimdTuple): Removed.

libstdc++: Improve copysign(simd) codegen

This also resolves a test failure on aarch64 with -ffast-math and
fixed_size<N> with large N.

Signed-off-by: Matthias Kretz <m.kretz@gsi.de>
libstdc++-v3/ChangeLog:

* include/experimental/bits/simd.h: Add missing operator~
overload for simd<floating-point> to __float_bitwise_operators.
* include/experimental/bits/simd_builtin.h
(_SimdImplBuiltin::_S_complement): Bitcast to int (and back) to
implement complement for floating-point vectors.
* include/experimental/bits/simd_fixed_size.h
(_SimdImplFixedSize::_S_copysign): New function, forwarding to
copysign implementation of _SimdTuple members.
* include/experimental/bits/simd_math.h (copysign): Call
_SimdImpl::_S_copysign for fixed_size arguments. Simplify
generic copysign implementation using the new ~ operator.

libstdc++: Fix typos and markdown errors in new simd/README.md

Signed-off-by: Jonathan Wakely <jwakely@redhat.com>
libstdc++-v3/ChangeLog:

* testsuite/experimental/simd/README.md: Fix typos.

libstdc++: Implement LWG 2762 for std::unique_ptr::operator*

The LWG issue proposes to add a conditional noexcept-specifier to
std::unique_ptr's dereference operator. The issue is currently in
Tentatively Ready status, but even if it isn't voted into the draft, we
can do it as a conforming extensions. This commit also adds a similar
noexcept-specifier to operator[] for the unique_ptr<T[], D> partial
specialization.

Also ensure that all dereference operators for shared_ptr are noexcept,
and adds tests for the std::optional accessors modified by the issue,
which were already noexcept in our implementation.

Signed-off-by: Jonathan Wakely <jwakely@redhat.com>
libstdc++-v3/ChangeLog:

* include/bits/shared_ptr_base.h (__shared_ptr_access::operator[]):
Add noexcept.
* include/bits/unique_ptr.h (unique_ptr::operator*): Add
conditional noexcept as per LWG 2762.
* testsuite/20_util/shared_ptr/observers/array.cc: Check that
dereferencing cannot throw.
* testsuite/20_util/shared_ptr/observers/get.cc: Likewise.
* testsuite/20_util/optional/observers/lwg2762.cc: New test.
* testsuite/20_util/unique_ptr/lwg2762.cc: New test.

Emit .file 0 directive earlier in DWARF 5

When the assembler supports it, the compiler automatically passes --gdwarf-5
to it, which has an interesting side effect: any assembly instruction prior
to the first .file directive defines a new line associated with .file 0 in
the .debug_line section and of course the numbering of these implicit lines
has nothing to do with that of the source code.  This can be problematic in
Ada when we do not generate .file/.loc directives for compiled-generated
functions to avoid too jumpy a debugging experience.

gcc/
* dwarf2out.c (dwarf2out_assembly_start): Emit .file 0 marker here..
(dwarf2out_finish): ...instead of here.

Fix --gdwarf-5 configure tests for Windows

The issues are that 1) they use readelf instead of objdump and 2) they use
ELF syntax in the assembly code.

gcc/
* configure.ac (--gdwarf-5 option): Use objdump instead of readelf.
(working --gdwarf-4/--gdwarf-5 for all sources): Likewise.
(--gdwarf-4 not refusing generated .debug_line): Adjust for Windows.
* configure: Regenerate.

Add cscope.out to git ignore.

ChangeLog:

* .gitignore: Add entry for cscope.out.

Merge vec_addsub patterns

This merges the vec_addsub<mode>3 patterns using a mode attribute
for the vec_merge merge operand.

2021-06-18  Richard Biener  <rguenther@suse.de>

* config/i386/sse.md (vec_addsubv4df3, vec_addsubv2df3,
vec_addsubv8sf3, vec_addsubv4sf3): Merge into ...
(vec_addsub<mode>3): ... using a new addsub_cst mode attribute.

Add x86 addsub SLP pattern

This addds SLP pattern recognition for the SSE3/AVX [v]addsubp{ds} v0, v1
instructions which compute { v0[0] - v1[0], v0[1], + v1[1], ... }
thus subtract, add alternating on lanes, starting with subtract.

It adds a corresponding optab and direct internal function,
vec_addsub$a3 and renames the existing i386 backend patterns to
the new canonical name.

The SLP pattern matches the exact alternating lane sequence rather
than trying to be clever and anticipating incoming permutes - we
could permute the two input vectors to the needed lane alternation,
do the addsub and then permute the result vector back but that's
only profitable in case the two input or the output permute will
vanish - something Tamars refactoring of SLP pattern recog should
make possible.

2021-06-17  Richard Biener  <rguenther@suse.de>

* config/i386/sse.md (avx_addsubv4df3): Rename to
vec_addsubv4df3.
(avx_addsubv8sf3): Rename to vec_addsubv8sf3.
(sse3_addsubv2df3): Rename to vec_addsubv2df3.
(sse3_addsubv4sf3): Rename to vec_addsubv4sf3.
* config/i386/i386-builtin.def: Adjust.
* internal-fn.def (VEC_ADDSUB): New internal optab fn.
* optabs.def (vec_addsub_optab): New optab.
* tree-vect-slp-patterns.c (class addsub_pattern): New.
(slp_patterns): Add addsub_pattern.
* tree-vect-slp.c (vect_optimize_slp): Disable propagation
across CFN_VEC_ADDSUB.
* tree-vectorizer.h (vect_pattern::vect_pattern): Make
m_ops optional.
* doc/md.texi (vec_addsub<mode>3): Document.

* gcc.target/i386/vect-addsubv2df.c: New testcase.
* gcc.target/i386/vect-addsubv4sf.c: Likewise.
* gcc.target/i386/vect-addsubv4df.c: Likewise.
* gcc.target/i386/vect-addsubv8sf.c: Likewise.
* gcc.target/i386/vect-addsub-2.c: Likewise.
* gcc.target/i386/vect-addsub-3.c: Likewise.

df: Fix up handling of paradoxical subregs in debug insns [PR101170]

The recent addition of gcc_assert (regno < endregno); triggers during
glibc build on m68k.
The problem is that RA decisions shouldn't depend on expressions in
DEBUG_INSNs and those expressions can contain paradoxical subregs of certain
pseudos.  If RA then decides to allocate the pseudo to a register
with very small hard register REGNO, we can trigger the new assert,
as (int) subreg_regno_offset may be negative on big endian and the small
REGNO + the negative offset can wrap around.

The following patch in that case records the range from the REGNO 0 to
endregno, before the addition of the assert as both regno and endregno are
unsigned it wouldn't record anything at all silently.

2021-06-24  Jakub Jelinek  <jakub@redhat.com>

PR middle-end/101170
* df-scan.c (df_ref_record): For paradoxical big-endian SUBREGs
where regno + subreg_regno_offset wraps around use 0 as starting
regno.

* gcc.dg/pr101170.c: New test.

stor-layout: Avoid DECL_BIT_FIELD_REPRESENTATIVE with NULL TREE_TYPE [PR101172]

finish_bitfield_representative has an early out if the field after a
bitfield has error_mark_node type, but that early out leads to TREE_TYPE
of the DECL_BIT_FIELD_REPRESENTATIVE being NULL, which breaks assumptions
on code that uses the DECL_BIT_FIELD_REPRESENTATIVE during error-recovery.

The following patch instead sets TREE_TYPE of the representative to
error_mark_node, something the users can deal with better.  At this point
the representative can be set as DECL_BIT_FIELD_REPRESENTATIVE for multiple
bitfields, so making sure that we clear the DECL_BIT_FIELD_REPRESENTATIVE
instead would be harder (but doable, e.g. with the error_mark_node TREE_TYPE
set by this patch set some flag in the caller and if the flag is there, walk
all the fields once again and clear all DECL_BIT_FIELD_REPRESENTATIVE that
have error_mark_node TREE_TYPE).

2021-06-24  Jakub Jelinek  <jakub@redhat.com>

PR middle-end/101172
* stor-layout.c (finish_bitfield_representative): If nextf has
error_mark_node type, set repr type to error_mark_node too.

* gcc.dg/pr101172.c: New test.

IBM Z: Define NO_PROFILE_COUNTERS

s390 glibc does not need counters in the .data section, since it stores
edge hits in its own data structure.  Therefore counters only waste
space and confuse diffing tools (e.g. kpatch), so don't generate them.

gcc/ChangeLog:

* config/s390/s390.c (s390_function_profiler): Ignore labelno
parameter.
* config/s390/s390.h (NO_PROFILE_COUNTERS): Define.

gcc/testsuite/ChangeLog:

* gcc.target/s390/mnop-mcount-m31-mzarch.c: Adapt to the new
prologue size.
* gcc.target/s390/mnop-mcount-m64.c: Likewise.

Fix SLP permute propagation error

This fixes SLP permute propagation to not propagate across operations
that have different semantics on different lanes like for example
the recently added COMPLEX_ADD_ROT90.

2021-06-24  Richard Biener  <rguenther@suse.de>

* tree-vect-slp.c (vect_optimize_slp): Do not propagate
across operations that have different semantics on different
lanes.

openmp: in_reduction clause support on target construct

This patch adds support for in_reduction clause on target construct, though
for now only for synchronous targets (without nowait clause).
The encountering thread in that case runs the target task and blocks until
the target region ends, so it is implemented by remapping it before entering
the target, initializing the private copy if not yet initialized for the
current thread and then using the remapped addresses for the mapping
addresses.
For nowait combined with in_reduction the patch contains a hack where the
nowait clause is ignored.  To implement it correctly, I think we would need
to create a new private variable for the in_reduction and initialize it before
doing the async target and adjust the map addresses to that private variable
and then pass a function pointer to the library routine with code where the callback
would remap the address to the current threads private variable and use in_reduction
combiner to combine the private variable we've created into the thread's copy.
The library would then need to make sure that the routine is called in some thread
participating in the parallel (and not in an unshackeled thread).

2021-06-24  Jakub Jelinek  <jakub@redhat.com>

gcc/
* tree.h (OMP_CLAUSE_MAP_IN_REDUCTION): Document meaning for OpenMP.
* gimplify.c (gimplify_scan_omp_clauses): For OpenMP map clauses
with OMP_CLAUSE_MAP_IN_REDUCTION flag partially defer gimplification
of non-decl OMP_CLAUSE_DECL.  For OMP_CLAUSE_IN_REDUCTION on
OMP_TARGET user outer_ctx instead of ctx for placeholders and
initializer/combiner gimplification.
* omp-low.c (scan_sharing_clauses): Handle OMP_CLAUSE_MAP_IN_REDUCTION
on target constructs.
(lower_rec_input_clauses): Likewise.
(lower_omp_target): Likewise.
* omp-expand.c (expand_omp_target): Temporarily ignore nowait clause
on target if in_reduction is present.
gcc/c-family/
* c-common.h (enum c_omp_region_type): Add C_ORT_TARGET and
C_ORT_OMP_TARGET.
* c-omp.c (c_omp_split_clauses): For OMP_CLAUSE_IN_REDUCTION on
combined target constructs also add map (always, tofrom:) clause.
gcc/c/
* c-parser.c (omp_split_clauses): Pass C_ORT_OMP_TARGET instead of
C_ORT_OMP for clauses on target construct.
(OMP_TARGET_CLAUSE_MASK): Add in_reduction clause.
(c_parser_omp_target): For non-combined target add
map (always, tofrom:) clauses for OMP_CLAUSE_IN_REDUCTION.  Pass
C_ORT_OMP_TARGET to c_finish_omp_clauses.
* c-typeck.c (handle_omp_array_sections): Adjust ort handling
for addition of C_ORT_OMP_TARGET and simplify, mapping clauses are
never present on C_ORT_*DECLARE_SIMD.
(c_finish_omp_clauses): Likewise.  Handle OMP_CLAUSE_IN_REDUCTION
on C_ORT_OMP_TARGET, set OMP_CLAUSE_MAP_IN_REDUCTION on
corresponding map clauses.
gcc/cp/
* parser.c (cp_omp_split_clauses): Pass C_ORT_OMP_TARGET instead of
C_ORT_OMP for clauses on target construct.
(OMP_TARGET_CLAUSE_MASK): Add in_reduction clause.
(cp_parser_omp_target): For non-combined target add
map (always, tofrom:) clauses for OMP_CLAUSE_IN_REDUCTION.  Pass
C_ORT_OMP_TARGET to finish_omp_clauses.
* semantics.c (handle_omp_array_sections_1): Adjust ort handling
for addition of C_ORT_OMP_TARGET and simplify, mapping clauses are
never present on C_ORT_*DECLARE_SIMD.
(handle_omp_array_sections): Likewise.
(finish_omp_clauses): Likewise.  Handle OMP_CLAUSE_IN_REDUCTION
on C_ORT_OMP_TARGET, set OMP_CLAUSE_MAP_IN_REDUCTION on
corresponding map clauses.
* pt.c (tsubst_expr): Pass C_ORT_OMP_TARGET instead of C_ORT_OMP for
clauses on target construct.
gcc/testsuite/
* c-c++-common/gomp/target-in-reduction-1.c: New test.
* c-c++-common/gomp/clauses-1.c: Add in_reduction clauses on
target or combined target constructs.
libgomp/
* testsuite/libgomp.c-c++-common/target-in-reduction-1.c: New test.
* testsuite/libgomp.c-c++-common/target-in-reduction-2.c: New test.
* testsuite/libgomp.c++/target-in-reduction-1.C: New test.
* testsuite/libgomp.c++/target-in-reduction-2.C: New test.

predcom: Refactor more by encapsulating global states

This patch is to encapsulate global states into a class and
making their accessors as member functions, remove some
consequent useless clean up code, and do some clean up with
RAII.

Bootstrapped/regtested on powerpc64le-linux-gnu P9,
x86_64-redhat-linux and aarch64-linux-gnu, also
bootstrapped on ppc64le P9 with bootstrap-O3 config.

gcc/ChangeLog:

* tree-predcom.c (class pcom_worker): New class.
(release_chain): Renamed to...
(pcom_worker::release_chain): ...this.
(release_chains): Renamed to...
(pcom_worker::release_chains): ...this.
(aff_combination_dr_offset): Renamed to...
(pcom_worker::aff_combination_dr_offset): ...this.
(determine_offset): Renamed to...
(pcom_worker::determine_offset): ...this.
(class comp_ptrs): New class.
(split_data_refs_to_components): Renamed to...
(pcom_worker::split_data_refs_to_components): ...this,
and update with class comp_ptrs.
(suitable_component_p): Renamed to...
(pcom_worker::suitable_component_p): ...this.
(filter_suitable_components): Renamed to...
(pcom_worker::filter_suitable_components): ...this.
(valid_initializer_p): Renamed to...
(pcom_worker::valid_initializer_p): ...this.
(find_looparound_phi): Renamed to...
(pcom_worker::find_looparound_phi): ...this.
(add_looparound_copies): Renamed to...
(pcom_worker::add_looparound_copies): ...this.
(determine_roots_comp): Renamed to...
(pcom_worker::determine_roots_comp): ...this.
(determine_roots): Renamed to...
(pcom_worker::determine_roots): ...this.
(single_nonlooparound_use): Renamed to...
(pcom_worker::single_nonlooparound_use): ...this.
(remove_stmt): Renamed to...
(pcom_worker::remove_stmt): ...this.
(execute_pred_commoning_chain): Renamed to...
(pcom_worker::execute_pred_commoning_chain): ...this.
(execute_pred_commoning): Renamed to...
(pcom_worker::execute_pred_commoning): ...this.
(struct epcc_data): New member worker.
(execute_pred_commoning_cbck): Call execute_pred_commoning
with pcom_worker pointer.
(find_use_stmt): Renamed to...
(pcom_worker::find_use_stmt): ...this.
(find_associative_operation_root): Renamed to...
(pcom_worker::find_associative_operation_root): ...this.
(find_common_use_stmt): Renamed to...
(pcom_worker::find_common_use_stmt): ...this.
(combinable_refs_p): Renamed to...
(pcom_worker::combinable_refs_p): ...this.
(reassociate_to_the_same_stmt): Renamed to...
(pcom_worker::reassociate_to_the_same_stmt): ...this.
(stmt_combining_refs): Renamed to...
(pcom_worker::stmt_combining_refs): ...this.
(combine_chains): Renamed to...
(pcom_worker::combine_chains): ...this.
(try_combine_chains): Renamed to...
(pcom_worker::try_combine_chains): ...this.
(prepare_initializers_chain): Renamed to...
(pcom_worker::prepare_initializers_chain): ...this.
(prepare_initializers): Renamed to...
(pcom_worker::prepare_initializers): ...this.
(prepare_finalizers_chain): Renamed to...
(pcom_worker::prepare_finalizers_chain): ...this.
(prepare_finalizers): Renamed to...
(pcom_worker::prepare_finalizers): ...this.
(tree_predictive_commoning_loop): Renamed to...
(pcom_worker::tree_predictive_commoning_loop): ...this, adjust
some calls and remove some cleanup code.
(tree_predictive_commoning): Adjusted to use pcom_worker instance.
(static variable looparound_phis): Remove.
(static variable name_expansions): Remove.

refactor SLP permute propagation

This refactors SLP permute propagation to record the outgoing permute
separately from the incoming/materialized one.  Instead of separate
arrays/bitmaps I've now created a struct to represent the state.

2021-06-23  Richard Biener  <rguenther@suse.de>

* tree-vect-slp.c (slpg_vertex): New struct.
(vect_slp_build_vertices): Adjust.
(vect_optimize_slp): Likewise.  Maintain an outgoing permute
and a materialized one.

tree-optimization/101105 - fix runtime alias test optimization

We were ignoring DR_STEP for VF == 1 which is OK only in case
the scalar order is preserved or both DR steps are the same.

2021-06-23  Richard Biener  <rguenther@suse.de>

PR tree-optimization/101105
* tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
Only ignore steps when they are equal or scalar order is preserved.

* gcc.dg/torture/pr101105.c: New testcase.

i386: Add vashlm3/vashrm3/vlshrm3 to enable vectorization of vector shift vector. [PR98434]

Add expanders for vashl<VI12_AVX512BW>, vlshr<VI12_AVX512BW>,
vashr<VI1_AVX512BW> and vashr<v32hi,v16hi,v4di,v8di>.

Besides there's some assumption in expand_mult_const that mul and
add must be available at the same time, but for i386, addv8qi is
restricted under TARGET_64BIT, but mulv8qi not, that could cause ICE.
So restrict mulv8qi and shiftv8qi under TARGET_64BIT.

gcc/ChangeLog:

PR target/98434
* config/i386/i386-expand.c (ix86_expand_vec_interleave):
Adjust comments for ix86_expand_vecop_qihi2.
(ix86_expand_vecmul_qihi): Renamed to ..
(ix86_expand_vecop_qihi2): Adjust function prototype to
support shift operation, add static to definition.
(ix86_expand_vec_shift_qihi_constant): Add static to definition.
(ix86_expand_vecop_qihi): Call ix86_expand_vecop_qihi2 and
ix86_expand_vec_shift_qihi_constant.
* config/i386/i386-protos.h (ix86_expand_vecmul_qihi): Deleted.
(ix86_expand_vec_shift_qihi_constant): Deleted.
* config/i386/sse.md (VI12_256_512_AVX512VL): New mode
iterator.
(mulv8qi3): Call ix86_expand_vecop_qihi directly, add
condition TARGET_64BIT.
(mul<mode>3): Ditto.
(<insn><mode>3): Ditto.
(vlshr<mode>3): Extend to support avx512 vlshr.
(v<insn><mode>3): New expander for
vashr/vlshr/vashl.
(v<insn>v8qi3): Ditto.
(vashrv8hi3<mask_name>): Renamed to ..
(vashr<mode>3): And extend to support V16QImode for avx512.
(vashrv16qi3): Deleted.
(vashrv2di3<mask_name>): Extend expander to support avx512
instruction.

gcc/testsuite/ChangeLog:

PR target/98434
* gcc.target/i386/pr98434-1.c: New test.
* gcc.target/i386/pr98434-2.c: New test.
* gcc.target/i386/avx512vl-pr95488-1.c: Adjust testcase.

Daily bump.

c++: excessive instantiation during CTAD [PR101174]

We set DECL_CONTEXT on implicitly generated deduction guides so that
their access is consistent with that of the constructor.  But this
apparently leads to excessive instantiation in some cases, ultimately
because instantiation of a deduction guide should be independent of
instantiation of the resulting class specialization, but setting the
DECL_CONTEXT of the former to the latter breaks this independence.

To fix this, this patch makes push_access_scope handle artificial
deduction guides specifically rather than setting their DECL_CONTEXT
in build_deduction_guide.  We could alternatively make the class
befriend the guide via DECL_BEFRIENDING_CLASSES, but that wouldn't
be a complete fix and would break class-deduction-access3.C below
since friendship isn't transitive.

PR c++/101174

gcc/cp/ChangeLog:

* pt.c (push_access_scope): For artificial deduction guides,
set the access scope to that of the constructor.
(pop_access_scope): Likewise.
(build_deduction_guide): Don't set DECL_CONTEXT on the guide.

libstdc++-v3/ChangeLog:

* testsuite/23_containers/multiset/cons/deduction.cc:
Uncomment CTAD example that was rejected by this bug.
* testsuite/23_containers/set/cons/deduction.cc: Likewise.

gcc/testsuite/ChangeLog:

* g++.dg/cpp1z/class-deduction-access3.C: New test.
* g++.dg/cpp1z/class-deduction91.C: New test.

doc/lto.texi: List slim object format as the default

Slim LTO object files have been the default for quite a while, since:
  commit e9f67e625c2a4225a7169d7220dcb85b6fdd7ca9
  Author:     Jan Hubicka <hubicka@gcc.gnu.org>
  common.opt (ffat-lto-objects): Disable by default.

That commit did not update lto.texi, so do it now.

gcc/ChangeLog:

* doc/lto.texi (Design Overview): Update that slim objects are
the default.

Signed-off-by: Dimitar Dimitrov <dimitar@dinux.eu>