Optimize strchr (s, 0) to s + strlen (s). strchr (s, 0) appears a common
idiom for finding the end of a string, however it is not a very efficient
way of doing so. Strlen is a much simpler operation which is significantly
faster (eg. on x86 strlen is 50% faster for strings of 8 bytes and about
twice as fast as strchr on strings of 1KB).
gcc/
* gimple-fold.c (gimple_fold_builtin_strchr):
New function to optimize strchr (s, 0) to strlen.
(gimple_fold_builtin): Add BUILT_IN_STRCHR case.
testsuite/
* gcc.dg/strlenopt-20.c: Update test.
* gcc.dg/strlenopt-21.c: Likewise.
* gcc.dg/strlenopt-22.c: Likewise.
* gcc.dg/strlenopt-22g.c: Likewise.
* gcc.dg/strlenopt-26.c: Likewise.
* gcc.dg/strlenopt-5.c: Likewise.
* gcc.dg/strlenopt-7.c: Likewise.
* gcc.dg/strlenopt-9.c: Likewise.
From-SVN: r240568
This is a step toward a version of mksysinfo that generates information
for the runtime package. This will be used to generate the
runtime_sysinfo.go file, which is currently directly generated by a
Makefile target.
Reviewed-on: https://go-review.googlesource.com/29683
From-SVN: r240560
During the compiler's order_evaluations() pass, duplicate
call statements were being created and inserted into block
statement lists in situations where there were calls returning
multiple arguments. These duplicates were benign in that they
were ignored during backend code generation, however they
did cause duplicate calls in AST dumps. Tweak the order
evaluation code to handle this case better and avoid inserting
these duplicates.
Fixesgolang/go#17237
Reviewed-on: https://go-review.googlesource.com/29857
From-SVN: r240559
When emitting AST dumps, tag the end of selected statements with with
source file and line number information where available. Example:
tmp.76832448 = 0 // p.go:6
Requires a corresponding change in gcc/go as well as this change to
gofrontend.
Reviewed-on: https://go-review.googlesource.com/29856
* go-linemap.cc (Gcc_linemap::to_string): New method.
From-SVN: r240558
Implement P0018R3, C++17 lambda capture of *this by value as [=,*this]
* parser.c (cp_parser_lambda_introducer): Formatting fix. Pass
true instead of false as by_reference_p to add_capture for 'this'.
Parse '*this' simple-capture.
* lambda.c (build_capture_proxy): Handle '*this' capture by value.
(add_capture): Adjust function comment. For id == this_identifier,
treat by_reference_p as capturing '*this' by reference, i.e. 'this'
by value, and !by_reference_p as capturing '*this' by value.
(add_default_capture): For implicit 'this' capture, always pass
by_reference_p true rather than false.
* g++.dg/cpp1z/lambda-this1.C: New test.
* g++.dg/cpp1z/lambda-this2.C: New test.
From-SVN: r240556
PR c++/77722
* cp-gimplify.c (cp_ubsan_maybe_instrument_return): Instrument also
functions that have just a STATEMENT_LIST instead of BIND_EXPR, or
BIND_EXPR with some statement rather than STATEMENT_LIST as body.
* g++.dg/ubsan/return-4.C: New test.
* g++.dg/ubsan/return-5.C: New test.
* g++.dg/ubsan/return-6.C: New test.
From-SVN: r240555
gcc/testsuite/
* lib/target-supports.exp
(check_effective_target_arm_v8_2a_fp16_scalar_hw): Delete redundant word
in function comment.
From-SVN: r240551
Revert a part of the `d' constraint documentation update made with
commit 111868 ("Use define_constraint for MIPS"),
<https://gcc.gnu.org/ml/gcc-patches/2006-03/msg00460.html>,
<https://gcc.gnu.org/ml/gcc-patches/2006-03/msg00541.html>, which
inadvertently made the registers covered to be referred to as "address"
rather than "general-purpose" registers, and then further clarify the
set of registers allowed with MIPS16 code.
* config/mips/constraints.md (d): Fix documentation.
* doc/md.texi (Machine Constraints): Update accordingly.
From-SVN: r240546
Fix PR gcov-profile/46266
PR gcov-profile/46266
* lib/gcov.exp: Verify that <built-in>.gcov file is not
considered.
* input.h (RESERVED_LOCATION_P): New macro.
* profile.c (branch_prob): Use RESERVED_LOCATION_P and
instread of comparison with UNKNOWN_LOCATION.
From-SVN: r240536
2016-09-27 Richard Biener <rguenther@suse.de>
PR tree-optimization/77745
* tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
When removing redundant stores make sure to check compatibility
of the TBAA state for downstream accesses.
* tree-ssa-sccvn.c (visit_reference_op_store): Likewise for when
value-numbering virtual operands for store matches.
* g++.dg/torture/pr77745.C: New testcase.
From-SVN: r240534
gcc/
PR target/51244
* config/sh/sh.c (sh_rtx_costs): Fix return value of SET of movt and
movrt patterns. Match them before anything else in the SET case.
From-SVN: r240533
PR gcov-profile/7970
PR gcov-profile/16855
PR gcov-profile/44779
* g++.dg/gcov/pr16855.C: New test.
* coverage.c (build_gcov_exit_decl): New function.
(coverage_obj_init): Call the function and generate __gcov_exit
destructor.
* doc/gcov.texi: Document when __gcov_exit function is called.
* libgcov-driver.c (__gcov_init): Do not register a atexit
handler.
(__gcov_exit): Rename from gcov_exit.
* libgcov.h (__gcov_exit): Declare.
From-SVN: r240529
The test has a bunch of hardcoded integer literals that would fit only in a
32 bits+ int, causing overflow warnings for a 16 bit int target like avr.
gcc/testsuite/ChangeLog
2016-09-27 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
* gcc.dg/tree-ssa/builtin-sprintf-warn-1.c: Require int32plus.
From-SVN: r240528
PR fortran/77666
* trans-openmp.c (gfc_omp_private_outer_ref): Return true even for
references to allocatable arrays.
* gfortran.dg/gomp/pr77666.f90: New test.
From-SVN: r240522
LRA likes to use CTR and LR to store CCmode values. Not such a good
idea. All other similar patterns disparage using CTR and LR; do so
here, too.
* config/rs6000/rs6000.md (movcc_internal1): Disparage using CTR or LR.
From-SVN: r240515
PR libstdc++/77727
* include/std/optional (optional(const optional<_Up>&)):
Default-initialize the base and use emplace.
(optional(optional<_Up>&&)): Likewise.
* testsuite/20_util/optional/cons/77727.cc: New.
From-SVN: r240511
2016-09-26 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/77420
* trans-common.c: Handle array elements in equivalence when
the lower and upper bounds of array spec are NULL.
2016-09-26 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/77420
* gfortran.dg/pr77420_1.f90: New test.
* gfortran.dg/pr77420_2.f90: Ditto.
* gfortran.dg/pr77420_3.f90: New test. Requires ...
* gfortran.dg/pr77420_4.f90: this file.
From-SVN: r240506
Fix PR77719
gcc/testsuite/ChangeLog:
2016-09-26 Kugan Vivekanandarajah <kuganv@linaro.org>
PR middle-end/77719
* gfortran.dg/pr77719.f90: New test.
gcc/ChangeLog:
2016-09-26 Kugan Vivekanandarajah <kuganv@linaro.org>
PR middle-end/77719
* tree-ssa-reassoc.c (make_new_ssa_for_def): Use gimple_get_lhs to get lhs
instead of gimple_assign_lhs as stmt can be builtins too.
From-SVN: r240505
2016-09-26 Thomas Preud'homme <thomas.preudhomme@arm.com>
gcc/testsuite/
* gcc.dg/tree-ssa/builtin-sprintf-warn-1.c: Adjust regex to accept
singular form of byte when quantity is unknown.
From-SVN: r240503
gcc/ChangeLog:
* read-rtl.c (read_rtx_code): Rename local "i" to "idx", and use
"c" instead when parsing characters. Move operand parsing into...
(read_rtx_operand): ...this new function, renaming "i" to "idx",
and tightening the scope of various locals.
From-SVN: r240502