Alex Coplan [Mon, 7 Sep 2020 14:23:44 +0000 (15:23 +0100)]
aarch64: Remove redundant mult patterns
Following on from the previous commit to fix up the syntax for
add/sub/adds/subs and friends with a sign/zero-extended operand, this
patch removes the "mult" variants of these patterns which are all
redundant.
This patch removes the following patterns from the AArch64 backend:
but this add instruction uses architecturally-invalid syntax: the width
of the third operand conflicts with the width of the extension
specifier. The third operand is only permitted to be an x register when
the extension specifier is (u|s)xtx.
This instruction, and analogous insns for adds, sub, subs, and cmp, are
rejected by clang, but accepted by binutils. Assembling and
disassembling such an insn with binutils gives the architecturally-valid
version in the disassembly:
This patch fixes several patterns in the AArch64 backend to use the
standard syntax as specified in the Arm ARM such that GCC's output can
be assembled by assemblers other than GAS.
Richard Biener [Mon, 7 Sep 2020 12:26:46 +0000 (14:26 +0200)]
improve SLP vect dumping
This adds additional dumping helping in particular basic-block
vectorization SLP dump reading plus showing what we actually
generate code from.
2020-09-07 Richard Biener <rguenther@suse.de>
* tree-vect-slp.c (vect_analyze_slp_instance): Dump
stmts we start SLP analysis from, failure and splitting.
(vect_schedule_slp): Dump SLP graph entry and root stmt
we are about to emit code for.
Martin Storsjö [Mon, 7 Sep 2020 11:18:42 +0000 (13:18 +0200)]
gcc: Make strchr return value pointers const
This fixes compilation of codepaths for dos-like filesystems
with Clang. When built with clang, it treats C input files as C++
when the compiler driver is invoked in C++ mode, triggering errors
when the return value of strchr() on a pointer to const is assigned
to a pointer to non-const variable.
This matches similar variables outside of the ifdefs for dos-like
path handling.
2020-09-07 Martin Storsjö <martin@martin.st>
gcc/
* dwarf2out.c (file_name_acquire): Make a strchr return value
pointer to const.
libcpp/
* files.c (remap_filename): Make a strchr return value pointer
to const.
Fortran: Fixes for pointer function call as variable (PR96896)
gcc/fortran/ChangeLog:
PR fortran/96896
* resolve.c (get_temp_from_expr): Also reset proc_pointer +
use_assoc attribute.
(resolve_ptr_fcn_assign): Use information from the LHS.
gcc/testsuite/ChangeLog:
PR fortran/96896
* gfortran.dg/ptr_func_assign_4.f08: Update dg-error.
* gfortran.dg/ptr-func-3.f90: New test.
Tom de Vries [Mon, 7 Sep 2020 09:54:27 +0000 (11:54 +0200)]
[libatomic, testsuite] Add missing include in atomic-generic.c
When compiling atomic-generic.c from the libatomic testsuite, we run into:
...
$ gcc src/libatomic/testsuite/libatomic.c/atomic-generic.c -latomic
src/libatomic/testsuite/libatomic.c/atomic-generic.c: In function ‘main’:
src/libatomic/testsuite/libatomic.c/atomic-generic.c:31:7: warning: \
implicit declaration of function ‘memcmp’ [-Wimplicit-function-declaration]
if (memcmp (&a, &zero, size))
^~~~~~
...
Fix this by adding the missing string.h include.
Tested on x86_64.
libatomic/ChangeLog:
* testsuite/libatomic.c/atomic-generic.c: Include string.h.
Jakub Jelinek [Mon, 7 Sep 2020 07:54:38 +0000 (09:54 +0200)]
lto: Stream edge goto_locus [PR94235]
The following patch adds streaming of edge goto_locus (both LOCATION_LOCUS
and LOCATION_BLOCK from it), the PR shows a testcase (inappropriate for
gcc testsuite) where the lack of streaming of goto_locus results in worse
debug info.
Earlier version of the patch (without the output_function changes) failed
miserably, because on the order mismatch - input_function would
first input_cfg, then input_eh_regions and then input_bb (all of which now
have locations), while output_function used output_eh_regions, then output_bb
and then output_cfg. *_cfg went to a separate stream...
Now, is there a reason why the order is different?
If the intent is that the cfg could be read separately from the rest of
function or vice versa, alternatively we'd need to clear_line_info ();
before output_eh_regions and before/after output_cfg to make them
independent.
2020-09-07 Jakub Jelinek <jakub@redhat.com>
PR debug/94235
* lto-streamer-out.c (output_cfg): Also stream goto_locus for edges.
Use bp_pack_var_len_unsigned instead of streamer_write_uhwi to stream
e->dest->index and e->flags.
(output_function): Call output_cfg before output_ssa_name, rather than
after streaming all bbs.
* lto-streamer-in.c (input_cfg): Stream in goto_locus for edges.
Use bp_unpack_var_len_unsigned instead of streamer_read_uhwi to stream
in dest_index and edge_flags.
Richard Biener [Fri, 4 Sep 2020 13:33:19 +0000 (15:33 +0200)]
code generate live lanes in basic-block vectorization
The following adds the capability to code-generate live lanes in
basic-block vectorization using lane extracts from vector stmts
rather than keeping the original scalar code around for those.
This eventually makes previously not profitable vectorizations
profitable (the live scalar code was appropriately costed so
are the lane extracts now), without considering the cost model
this patch doesn't add or remove any basic-block vectorization
capabilities.
The patch re/ab-uses STMT_VINFO_LIVE_P in basic-block vectorization
mode to tell whether a live lane is vectorized or whether it is
provided by means of keeping the scalar code live.
The patch is a first step towards vectorizing sequences of
stmts that do not end up in stores or vector constructors though.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
2020-09-04 Richard Biener <rguenther@suse.de>
* tree-vectorizer.h (vectorizable_live_operation): Adjust.
* tree-vect-loop.c (vectorizable_live_operation): Vectorize
live lanes out of basic-block vectorization nodes.
* tree-vect-slp.c (vect_bb_slp_mark_live_stmts): New function.
(vect_slp_analyze_operations): Analyze live lanes and their
vectorization possibility after the whole SLP graph is final.
(vect_bb_slp_scalar_cost): Adjust for vectorized live lanes.
* tree-vect-stmts.c (can_vectorize_live_stmts): Adjust.
(vect_transform_stmt): Call can_vectorize_live_stmts also for
basic-block vectorization.
* gcc.dg/vect/bb-slp-46.c: New testcase.
* gcc.dg/vect/bb-slp-47.c: Likewise.
* gcc.dg/vect/bb-slp-32.c: Adjust.
d: Fix ICE in create_tmp_var, at gimple-expr.c:482
Array concatenate expressions were creating more SAVE_EXPRs than what
was necessary. The internal error itself was the result of a forced
temporary being made on a TREE_ADDRESSABLE type.
Richard Biener [Fri, 4 Sep 2020 10:18:38 +0000 (12:18 +0200)]
tree-optimization/96931 - clear ctrl-altering flag more aggressively
The testcase shows that we fail to clear gimple_call_ctrl_altering_p
when the last abnormal edge goes away, causing an edge insert to
a loop header edge when we have preheaders to split the edge
unnecessarily.
The following addresses this by more aggressively clearing the
flag in cleanup_call_ctrl_altering_flag.
2020-09-04 Richard Biener <rguenther@suse.de>
PR tree-optimization/96931
* tree-cfgcleanup.c (cleanup_call_ctrl_altering_flag): If
there's a fallthru edge and no abnormal edge the call is
no longer control-altering.
(cleanup_control_flow_bb): Pass down the BB to
cleanup_call_ctrl_altering_flag.
Jakub Jelinek [Fri, 4 Sep 2020 09:55:13 +0000 (11:55 +0200)]
lto: Remove stream_input_location_now
As discussed yesterday, stream_input_location_now has been used in 3
remaining places. For ERT_MUST_NOT_THROW, I believe the failure_loc
location is stable at least until the apply_cache after the bbs are all
read, and the locations do not include BLOCK, so we can use normal
stream_input_location, and the two input_struct_function_base also
shouldn't include BLOCK and are stable at least until that same apply_cache
after reading all bbs, so again we can use the location cache.
2020-09-04 Jakub Jelinek <jakub@redhat.com>
* lto-streamer.h (stream_input_location_now): Remove declaration.
* lto-streamer-in.c (stream_input_location_now): Remove.
(input_eh_region, input_struct_function_base): Use
stream_input_location instead of stream_input_location_now.
Jakub Jelinek [Fri, 4 Sep 2020 09:53:28 +0000 (11:53 +0200)]
lto: Ensure we force a change for file/line/column after clear_line_info
As discussed yesterday:
On the streamer out side, we call clear_line_info
in multiple spots which resets the current_* values to something, but on the
reader side, we don't have corresponding resets in the same location, just have
the stream_* static variables that keep the current values through the
entire stream in (so across all the clear_line_info spots in a single LTO
object but also across jumping from one LTO object to another one).
Now, in an earlier version of my patch it actually broke LTO bootstrap
(and a lot of LTO testcases), so for the BLOCK case I've solved it by
clear_line_info setting current_block to something that should never appear,
which means that in the LTO stream after the clear_line_info spots including
the start of the LTO stream we force the block change bit to be set and thus
BLOCK to be streamed and therefore stream_block from earlier to be
ignored. But for the rest I think that is not the case, so I wonder if we
don't sometimes end up with wrong line/column info because of that, or
please tell me what prevents that.
clear_line_info does:
ob->current_file = NULL;
ob->current_line = 0;
ob->current_col = 0;
ob->current_sysp = false;
while I think NULL current_file is something that should likely be different
from expanded_location (...).file (UNKNOWN_LOCATION/BUILTINS_LOCATION are
handled separately and not go through the caching), I think line number 0
can sometimes occur and especially column 0 occurs frequently if we ran out
of location_t with columns info. But then we do:
bp_pack_value (bp, ob->current_file != xloc.file, 1);
bp_pack_value (bp, ob->current_line != xloc.line, 1);
bp_pack_value (bp, ob->current_col != xloc.column, 1);
and stream the details only if the != is true. If that happens immediately
after clear_line_info and e.g. xloc.column is 0, we would stream 0 bit and
not stream the actual value, so on read-in it would reuse whatever
stream_col etc. were before. Shouldn't we set some ob->current_* new bit
that would signal we are immediately past clear_line_info which would force
all these != checks to non-zero? Either by oring something into those
tests, or perhaps:
if (ob->current_reset)
{
if (xloc.file == NULL)
ob->current_file = "";
if (xloc.line == 0)
ob->current_line = 1;
if (xloc.column == 0)
ob->current_column = 1;
ob->current_reset = false;
}
before doing those bp_pack_value calls with a comment, effectively forcing
all 6 != comparisons to be true?
2020-09-04 Jakub Jelinek <jakub@redhat.com>
* lto-streamer.h (struct output_block): Add reset_locus member.
* lto-streamer-out.c (clear_line_info): Set reset_locus to true.
(lto_output_location_1): If reset_locus, clear it and ensure
current_{file,line,col} is different from xloc members.
Martin Jambor [Thu, 3 Sep 2020 20:43:49 +0000 (22:43 +0200)]
sra: Avoid SRAing if there is an aout-of-bounds access (PR 96820)
The testcase causes and ICE in the SRA verifier on x86_64 when
compiling with -m32 because build_user_friendly_ref_for_offset looks
at an out-of-bounds array_ref within an array_ref which accesses an
offset which does not fit into a signed 32bit integer and turns it
into an array-ref with a negative index.
The best thing is probably to bail out early when encountering an out
of bounds access to a local stack-allocated aggregate (and let the DSE
just delete such statements) which is what the patch does.
I also glanced over to the initial candidate vetting routine to make
sure the size would fit into HWI and noticed that it uses unsigned
variants whereas the rest of SRA operates on signed offsets and
sizes (because get_ref_and_extent does) and so changed that for the
sake of consistency. These ancient checks operate on sizes of types
as opposed to DECLs but I hope that any issues potentially arising
from that are basically hypothetical.
gcc/ChangeLog:
2020-08-28 Martin Jambor <mjambor@suse.cz>
PR tree-optimization/96820
* tree-sra.c (create_access): Disqualify candidates with accesses
beyond the end of the original aggregate.
(maybe_add_sra_candidate): Check that candidate type size fits
signed uhwi for the sake of consistency.
gcc/testsuite/ChangeLog:
2020-08-28 Martin Jambor <mjambor@suse.cz>
PR tree-optimization/96820
* gcc.dg/tree-ssa/pr96820.c: New test.
Will Schmidt [Mon, 20 Jul 2020 15:51:37 +0000 (10:51 -0500)]
[PATCH, rs6000] Fix vector long long subtype (PR96139)
Hi,
This corrects an issue with the powerpc vector long long subtypes.
As reported by SjMunroe, when building some code with -Wall, and
attempting to print an element of a "long long vector" with a
long long printf format string, we will report an error because
the vector sub-type was improperly defined as int.
When defining a V2DI_type_node we use a TARGET_POWERPC64 ternary to
define the V2DI_type_node with "vector long" or "vector long long".
We also need to specify the proper sub-type when we define the type.
PR target/96139
2020-09-03 Will Schmidt <will_schmidt@vnet.ibm.com>
gcc/ChangeLog:
* config/rs6000/rs6000-call.c (rs6000_init_builtin): Update V2DI_type_node
and unsigned_V2DI_type_node definitions.
gcc/testsuite/ChangeLog:
* gcc.target/powerpc/pr96139-a.c: New test.
* gcc.target/powerpc/pr96139-b.c: New test.
* gcc.target/powerpc/pr96139-c.c: New test.
Jakub Jelinek [Thu, 3 Sep 2020 19:53:40 +0000 (21:53 +0200)]
c++: Fix another PCH hash_map issue [PR96901]
The recent libstdc++ changes caused lots of libstdc++-v3 tests FAILs
on i686-linux, all of them in the same spot during constexpr evaluation
of a recursive _S_gcd call.
The problem is yet another hash_map that used the default hasing of
tree keys through pointer hashing which is preserved across PCH write/read.
During PCH handling, the addresses of GC objects are changed, which means
that the hash values of the keys in such hash tables change without those
hash tables being rehashed. Which in the fundef_copies_table case usually
means we just don't find a copy of a FUNCTION_DECL body for recursive uses
and start from scratch. But when the hash table keeps growing, the "dead"
elements in the hash table can sometimes reappear and break things.
In particular what I saw under the debugger is when the fundef_copies_table
hash map has been used on the outer _S_gcd call, it didn't find an entry for
it, so returned a slot with *slot == NULL, which is treated as that the
function itself is used directly (i.e. no recursion), but that addition of
a hash table slot caused the recursive _S_gcd call to actually find
something in the hash table, unfortunately not the new *slot == NULL spot,
but a different one from the pre-PCH streaming which contained the returned
toplevel (non-recursive) call entry for it, which means that for the
recursive _S_gcd call we actually used the same trees as for the outer ones
rather than a copy of those, which breaks constexpr evaluation.
2020-09-03 Jakub Jelinek <jakub@redhat.com>
PR c++/96901
* tree.h (struct decl_tree_traits): New type.
(decl_tree_map): New typedef.
* constexpr.c (fundef_copies_table): Change type from
hash_map<tree, tree> * to decl_tree_map *.
Harald Anlauf [Thu, 3 Sep 2020 18:33:14 +0000 (20:33 +0200)]
PR fortran/96890 - Wrong answer with intrinsic IALL
The IALL intrinsic would always return 0 when the DIM and MASK arguments
were present since the initial value of repeated BIT-AND operations was
set to 0 instead of -1.
libgfortran/ChangeLog:
* m4/iall.m4: Initial value for result should be -1.
* generated/iall_i1.c (miall_i1): Generated.
* generated/iall_i16.c (miall_i16): Likewise.
* generated/iall_i2.c (miall_i2): Likewise.
* generated/iall_i4.c (miall_i4): Likewise.
* generated/iall_i8.c (miall_i8): Likewise.
Marek Polacek [Wed, 26 Aug 2020 12:27:33 +0000 (08:27 -0400)]
c++: Fix P0960 in member init list and array [PR92812]
This patch nails down the remaining P0960 case in PR92812:
struct A {
int ar[2];
A(): ar(1, 2) {} // doesn't work without this patch
};
Note that when the target object is not of array type, this already
works:
struct S { int x, y; };
struct A {
S s;
A(): s(1, 2) { } // OK in C++20
};
because build_new_method_call_1 takes care of the P0960 magic.
It proved to be quite hairy. When the ()-list has more than one
element, we can always create a CONSTRUCTOR, because the code was
previously invalid. But when the ()-list has just one element, it
gets all kinds of difficult. As usual, we have to handle a("foo")
so as not to wrap the STRING_CST in a CONSTRUCTOR. Always turning
x(e) into x{e} would run into trouble as in c++/93790. Another
issue was what to do about x({e}): previously, this would trigger
"list-initializer for non-class type must not be parenthesized".
I figured I'd make this work in C++20, so that given
struct S { int x, y; };
you can do
S a[2];
[...]
A(): a({1, 2}) // initialize a[0] with {1, 2} and a[1] with {}
It also turned out that, as an extension, we support compound literals:
F (): m((S[1]) { 1, 2 })
so this has to keep working as before. Moreover, make sure not to trigger
in compiler-generated code, like =default, where array assignment is allowed.
I've factored out a function that turns a TREE_LIST into a CONSTRUCTOR
to simplify handling of P0960.
paren-init35.C also tests this with vector types.
gcc/cp/ChangeLog:
PR c++/92812
* cp-tree.h (do_aggregate_paren_init): Declare.
* decl.c (do_aggregate_paren_init): New.
(grok_reference_init): Use it.
(check_initializer): Likewise.
* init.c (perform_member_init): Handle initializing an array from
a ()-list. Use do_aggregate_paren_init.
Jakub Jelinek [Thu, 3 Sep 2020 18:11:43 +0000 (20:11 +0200)]
c++: Disable -frounding-math during manifestly constant evaluation [PR96862]
As discussed in the PR, fold-const.c punts on floating point constant
evaluation if the result is inexact and -frounding-math is turned on.
/* Don't constant fold this floating point operation if the
result may dependent upon the run-time rounding mode and
flag_rounding_math is set, or if GCC's software emulation
is unable to accurately represent the result. */
if ((flag_rounding_math
|| (MODE_COMPOSITE_P (mode) && !flag_unsafe_math_optimizations))
&& (inexact || !real_identical (&result, &value)))
return NULL_TREE;
Jonathan said that we should be evaluating them anyway, e.g. conceptually
as if they are done with the default rounding mode before user had a chance
to change that, and e.g. in C in initializers it is also ignored.
In fact, fold-const.c for C initializers turns off various other options:
/* Perform constant folding and related simplification of initializer
expression EXPR. These behave identically to "fold_buildN" but ignore
potential run-time traps and exceptions that fold must preserve. */
int saved_signaling_nans = flag_signaling_nans;\
int saved_trapping_math = flag_trapping_math;\
int saved_rounding_math = flag_rounding_math;\
int saved_trapv = flag_trapv;\
int saved_folding_initializer = folding_initializer;\
flag_signaling_nans = 0;\
flag_trapping_math = 0;\
flag_rounding_math = 0;\
flag_trapv = 0;\
folding_initializer = 1;
So, shall cxx_eval_outermost_constant_expr instead turn off all those
options (then warning_sentinel wouldn't be the right thing to use, but given
the 8 or how many return stmts in cxx_eval_outermost_constant_expr, we'd
need a RAII class for this. Not sure about the folding_initializer, that
one is affecting complex multiplication and division constant evaluation
somehow.
Jonathan Wakely [Thu, 3 Sep 2020 15:26:16 +0000 (16:26 +0100)]
libstdc++: Add workaround for weird std::tuple error [PR 96592]
This "fix" makes no sense, but it avoids an error from G++ about
std::is_constructible being incomplete. The real problem is elsewhere,
but this "fixes" the regression for now.
libstdc++-v3/ChangeLog:
PR libstdc++/96592
* include/std/tuple (_TupleConstraints<true, T...>): Use
alternative is_constructible instead of std::is_constructible.
* testsuite/20_util/tuple/cons/96592.cc: New test.
Jonathan Wakely [Thu, 3 Sep 2020 11:38:50 +0000 (12:38 +0100)]
libstdc++: Optimise GCD algorithms
The current std::gcd and std::chrono::duration::_S_gcd algorithms are
both recursive. This is potentially expensive to evaluate in constant
expressions, because each level of recursion makes a new copy of the
function to evaluate. The maximum number of steps is bounded
(proportional to the number of decimal digits in the smaller value) and
so unlikely to exceed the limit for constexpr nesting, but the memory
usage is still suboptimal. By using an iterative algorithm we avoid
that compile-time cost. Because looping in constexpr functions is not
allowed until C++14, we need to keep the recursive implementation in
duration::_S_gcd for C++11 mode.
For std::gcd we can also optimise runtime performance by using the
binary GCD algorithm.
libstdc++-v3/ChangeLog:
* include/std/chrono (duration::_S_gcd): Use iterative algorithm
for C++14 and later.
* include/std/numeric (__detail::__gcd): Replace recursive
Euclidean algorithm with iterative version of binary GCD algorithm.
* testsuite/26_numerics/gcd/1.cc: Test additional inputs.
* testsuite/26_numerics/gcd/gcd_neg.cc: Adjust dg-error lines.
* testsuite/26_numerics/lcm/lcm_neg.cc: Likewise.
* testsuite/experimental/numeric/gcd.cc: Test additional inputs.
* testsuite/26_numerics/gcd/2.cc: New test.
Jakub Jelinek [Thu, 3 Sep 2020 10:51:01 +0000 (12:51 +0200)]
lto: Cache location_ts including BLOCKs in GIMPLE streaming [PR94311]
As mentioned in the PR, when compiling valgrind even on fairly small
testcase where in one larger function the location keeps oscillating
between a small line number and 8000-ish line number in the same file
we very quickly run out of all possible location_t numbers and because of
that emit non-sensical line numbers in .debug_line.
There are ways how to decrease speed of depleting location_t numbers
in libcpp, but the main reason of this is that we use
stream_input_location_now for streaming in location_t for gimple_location
and phi arg locations. libcpp strongly prefers that the locations
it is given are sorted by the different files and by line numbers in
ascending order, otherwise it depletes quickly no matter what and is much
more costly (many extra file changes etc.).
The reason for not caching those were the BLOCKs that were streamed
immediately after the location and encoded into the locations (and for PHIs
we failed to stream the BLOCKs altogether).
This patch enhances the location cache to handle also BLOCKs (but not for
everything, only for the spots we care about the BLOCKs) and also optimizes
the size of the LTO stream by emitting a single bit into a pack whether the
BLOCK changed from last case and only streaming the BLOCK tree if it
changed.
2020-09-03 Jakub Jelinek <jakub@redhat.com>
PR lto/94311
* gimple.h (gimple_location_ptr, gimple_phi_arg_location_ptr): New
functions.
* streamer-hooks.h (struct streamer_hooks): Add
output_location_and_block callback. Fix up formatting for
output_location.
(stream_output_location_and_block): Define.
* lto-streamer.h (class lto_location_cache): Fix comment typo. Add
current_block member.
(lto_location_cache::input_location_and_block): New method.
(lto_location_cache::lto_location_cache): Initialize current_block.
(lto_location_cache::cached_location): Add block member.
(struct output_block): Add current_block member.
(lto_output_location): Formatting fix.
(lto_output_location_and_block): Declare.
* lto-streamer.c (lto_streamer_hooks_init): Initialize
streamer_hooks.output_location_and_block.
* lto-streamer-in.c (lto_location_cache::cmp_loc): Also compare
block members.
(lto_location_cache::apply_location_cache): Handle blocks.
(lto_location_cache::accept_location_cache,
lto_location_cache::revert_location_cache): Fix up function comments.
(lto_location_cache::input_location_and_block): New method.
(lto_location_cache::input_location): Implement using
input_location_and_block.
(input_function): Invoke apply_location_cache after streaming in all
bbs.
* lto-streamer-out.c (clear_line_info): Set current_block.
(lto_output_location_1): New function, moved from lto_output_location,
added block handling.
(lto_output_location): Implement using lto_output_location_1.
(lto_output_location_and_block): New function.
* gimple-streamer-in.c (input_phi): Use input_location_and_block
to input and cache both location and block.
(input_gimple_stmt): Likewise.
* gimple-streamer-out.c (output_phi): Use
stream_output_location_and_block.
(output_gimple_stmt): Likewise.
Look at fullest view when checking for static types in unnesting
When seeing if any bound involved in a type is an uplevel reference,
we must look at the fullest view of a type, since that's what the
backends will do. Similarly for private types. We introduce
Get_Fullest_View for that purpose.
* sem_util.ads, sem_util.adb (Get_Fullest_View): New procedure.
* exp_unst.adb (Check Static_Type): Do all processing on fullest
view of specified type.
Optimize memory broadcast for constant vector under AVX512.
For constant vector having one duplicated value, there's no need to put
whole vector in the constant pool, using embedded broadcast instead.
2020-07-09 Hongtao Liu <hongtao.liu@intel.com>
gcc/ChangeLog:
PR target/87767
* config/i386/i386-features.c
(replace_constant_pool_with_broadcast): New function.
(constant_pool_broadcast): Ditto.
(class pass_constant_pool_broadcast): New pass.
(make_pass_constant_pool_broadcast): Ditto.
(remove_partial_avx_dependency): Call
replace_constant_pool_with_broadcast under TARGET_AVX512F, it
would save compile time when both pass rpad and cpb are
available.
(remove_partial_avx_dependency_gate): New function.
(class pass_remove_partial_avx_dependency::gate): Call
remove_partial_avx_dependency_gate.
* config/i386/i386-passes.def: Insert new pass after combine.
* config/i386/i386-protos.h
(make_pass_constant_pool_broadcast): Declare.
* config/i386/sse.md (*avx512dq_mul<mode>3<mask_name>_bcst):
New define_insn.
(*avx512f_mul<mode>3<mask_name>_bcst): Ditto.
* config/i386/avx512fintrin.h (_mm512_set1_ps,
_mm512_set1_pd,_mm512_set1_epi32, _mm512_set1_epi64): Adjusted.
gcc/testsuite/ChangeLog:
PR target/87767
* gcc.target/i386/avx2-broadcast-pr87767-1.c: New test.
* gcc.target/i386/avx512f-broadcast-pr87767-1.c: New test.
* gcc.target/i386/avx512f-broadcast-pr87767-2.c: New test.
* gcc.target/i386/avx512f-broadcast-pr87767-3.c: New test.
* gcc.target/i386/avx512f-broadcast-pr87767-4.c: New test.
* gcc.target/i386/avx512f-broadcast-pr87767-5.c: New test.
* gcc.target/i386/avx512f-broadcast-pr87767-6.c: New test.
* gcc.target/i386/avx512f-broadcast-pr87767-7.c: New test.
* gcc.target/i386/avx512vl-broadcast-pr87767-1.c: New test.
* gcc.target/i386/avx512vl-broadcast-pr87767-1.c: New test.
* gcc.target/i386/avx512vl-broadcast-pr87767-2.c: New test.
* gcc.target/i386/avx512vl-broadcast-pr87767-3.c: New test.
* gcc.target/i386/avx512vl-broadcast-pr87767-4.c: New test.
* gcc.target/i386/avx512vl-broadcast-pr87767-5.c: New test.
* gcc.target/i386/avx512vl-broadcast-pr87767-6.c: New test.
liuhongt [Mon, 31 Aug 2020 02:54:13 +0000 (10:54 +0800)]
Adjust testcase.
gcc/testsuite/ChangeLog:
PR target/96246
PR target/96855
PR target/96856
PR target/96857
* g++.target/i386/avx512bw-pr96246-2.C: Add runtime check for
AVX512BW.
* g++.target/i386/avx512vl-pr96246-2.C: Add runtime check for
AVX512BW and AVX512VL
* g++.target/i386/avx512f-helper.h: New header.
* gcc.target/i386/pr92658-avx512f.c: Add
-mprefer-vector-width=512 to avoid impact of different default
mtune which gcc is built with.
* gcc.target/i386/avx512bw-pr95488-1.c: Ditto.
* gcc.target/i386/pr92645-4.c: Add -mno-avx512f to avoid
impact of different default march which gcc is built with.
Iain Buclaw [Mon, 31 Aug 2020 20:42:10 +0000 (22:42 +0200)]
d: __vectors unsupported in hardware should be rejected at compile-time.
gcc/d/ChangeLog:
PR d/96869
* d-builtins.cc (build_frontend_type): Don't expose intrinsics that
use unsupported vector types.
* d-target.cc (Target::isVectorTypeSupported): Restrict to supporting
only if TARGET_VECTOR_MODE_SUPPORTED_P is true. Don't allow complex
or boolean vector types.
gcc/testsuite/ChangeLog:
PR d/96869
* gdc.dg/simd.d: Removed.
* gdc.dg/cast1.d: New test.
* gdc.dg/gdc213.d: Compile with target vect_sizes_16B_8B.
* gdc.dg/gdc284.d: Likewise.
* gdc.dg/gdc67.d: Likewise.
* gdc.dg/pr96869.d: New test.
* gdc.dg/simd1.d: New test.
* gdc.dg/simd10447.d: New test.
* gdc.dg/simd12776.d: New test.
* gdc.dg/simd13841.d: New test.
* gdc.dg/simd13927.d: New test.
* gdc.dg/simd15123.d: New test.
* gdc.dg/simd15144.d: New test.
* gdc.dg/simd16087.d: New test.
* gdc.dg/simd16697.d: New test.
* gdc.dg/simd17237.d: New test.
* gdc.dg/simd17695.d: New test.
* gdc.dg/simd17720a.d: New test.
* gdc.dg/simd17720b.d: New test.
* gdc.dg/simd19224.d: New test.
* gdc.dg/simd19627.d: New test.
* gdc.dg/simd19628.d: New test.
* gdc.dg/simd19629.d: New test.
* gdc.dg/simd19630.d: New test.
* gdc.dg/simd2a.d: New test.
* gdc.dg/simd2b.d: New test.
* gdc.dg/simd2c.d: New test.
* gdc.dg/simd2d.d: New test.
* gdc.dg/simd2e.d: New test.
* gdc.dg/simd2f.d: New test.
* gdc.dg/simd2g.d: New test.
* gdc.dg/simd2h.d: New test.
* gdc.dg/simd2i.d: New test.
* gdc.dg/simd2j.d: New test.
* gdc.dg/simd7951.d: New test.
* gdc.dg/torture/array2.d: New test.
* gdc.dg/torture/array3.d: New test.
* gdc.dg/torture/simd16488a.d: New test.
* gdc.dg/torture/simd16488b.d: New test.
* gdc.dg/torture/simd16703.d: New test.
* gdc.dg/torture/simd19223.d: New test.
* gdc.dg/torture/simd19607.d: New test.
* gdc.dg/torture/simd3.d: New test.
* gdc.dg/torture/simd4.d: New test.
* gdc.dg/torture/simd7411.d: New test.
* gdc.dg/torture/simd7413a.d: New test.
* gdc.dg/torture/simd7413b.d: New test.
* gdc.dg/torture/simd7414.d: New test.
* gdc.dg/torture/simd9200.d: New test.
* gdc.dg/torture/simd9304.d: New test.
* gdc.dg/torture/simd9449.d: New test.
* gdc.dg/torture/simd9910.d: New test.
Iain Buclaw [Mon, 31 Aug 2020 17:27:15 +0000 (19:27 +0200)]
d: Only test with default permutation flags for runnable tests.
Unless the test explicitly requests, all compilable tests as well as
fail_compilation tests will be ran without any extra flags.
The C++ tests now are checked against shared D runtime library.
gcc/testsuite/ChangeLog:
* lib/gdc-utils.exp (gdc-convert-test): Handle LINK directive.
Set PERMUTE_ARGS as DEFAULT_DFLAGS only for runnable tests.
(gdc-do-test): Set default action of compilable tests to compile.
Test SHARED_OPTION on runnable_cxx tests.
Jonathan Wakely [Wed, 2 Sep 2020 17:51:28 +0000 (18:51 +0100)]
c++: Stop defining true, false and bool as macros in <stdbool.h>
Since r216679 these macros have only been defined in C++98 mode, rather
than all modes. That is permitted as a GNU extension because that header
doesn't exist in the C++ standard until C++11, so we can make it do
whatever we want for C++98. But as discussed in the PR c++/60304
comments, these macros shouldn't ever be defined for C++.
This patch removes the macro definitions for C++98 too.
The new test already passed for C++98 (and the conversion is ill-formed
in C++11 and later) so this new test is arguably unnecessary.
gcc/ChangeLog:
PR c++/60304
* ginclude/stdbool.h (bool, false, true): Never define for C++.
gcc/testsuite/ChangeLog:
PR c++/60304
* g++.dg/warn/Wconversion-null-5.C: New test.
Jonathan Wakely [Wed, 2 Sep 2020 14:17:24 +0000 (15:17 +0100)]
libstdc++: Fix three-way comparison for std::array [PR 96851]
The spaceship operator for std::array uses memcmp when the
__is_byte<value_type> trait is true, but memcmp isn't usable in
constexpr contexts. Also, memcmp should only be used for unsigned byte
types, because it gives the wrong answer for signed chars with negative
values.
We can simply check std::is_constant_evaluated() so that we don't use
memcmp during constant evaluation.
To fix the problem of using memcmp for inappropriate types, this patch
adds new __is_memcmp_ordered and __is_memcmp_ordered_with traits. These
say whether using memcmp will give the right answer for ordering
operations such as lexicographical_compare and three-way comparisons.
The new traits can be used in several places, and can also be used to
implement my suggestion in PR 93059 comment 37 to use memcmp for
unsigned integers larger than one byte on big endian targets.
libstdc++-v3/ChangeLog:
PR libstdc++/96851
* include/bits/cpp_type_traits.h (__is_memcmp_ordered):
New trait that says if memcmp can be used for ordering.
(__is_memcmp_ordered_with): Likewise, for two types.
* include/bits/deque.tcc (__lex_cmp_dit): Use new traits
instead of __is_byte and __numeric_traits.
(__lexicographical_compare_aux1): Likewise.
* include/bits/ranges_algo.h (__lexicographical_compare_fn):
Likewise.
* include/bits/stl_algobase.h (__lexicographical_compare_aux1)
(__is_byte_iter): Likewise.
* include/std/array (operator<=>): Likewise. Only use memcmp
when std::is_constant_evaluated() is false.
* testsuite/23_containers/array/comparison_operators/96851.cc:
New test.
* testsuite/23_containers/array/tuple_interface/get_neg.cc:
Adjust dg-error line numbers.
MSP430: Skip gcc.dg/pr55940.c in the small memory model
In the MSP430 small memory model, there is a 16-bit address space and
pointer arithmetic wraps around the address space, so any calculated
address is always within this range.
In this test, pointer arithmetic wraps when 0x1000 is added to the
address of a variable, causing the resulting address to be unexpectedly
less than 0x2000, which breaks the test.
gcc/testsuite/ChangeLog:
* gcc.dg/pr55940.c: Skip for msp430 unless -mlarge is specified.
Jonathan Wakely [Wed, 2 Sep 2020 12:27:57 +0000 (13:27 +0100)]
libstdc++: Break header cycle between <new> and <exception>
The <new> and <exception> headers each include each other, which makes
building them as header-units "exciting". The <new> header only needs
the definition of std::exception (in order to derive from it) which is
already in its own header, so just include that.
libstdc++-v3/ChangeLog:
* include/bits/stl_iterator.h: Include <bits/exception_defines.h>
for definitions of __try, __catch and __throw_exception_again.
(counted_iterator::operator++(int)): Use __throw_exception_again
instead of throw.
* libsupc++/new: Include <bits/exception.h> not <exception>.
* libsupc++/new_opvnt.cc: Include <bits/exception_defines.h>.
* testsuite/18_support/destroying_delete.cc: Include
<type_traits> for std::is_same_v definition.
* testsuite/20_util/variant/index_type.cc: Qualify size_t.
Jakub Jelinek [Wed, 2 Sep 2020 10:18:46 +0000 (12:18 +0200)]
fortran: Fix o'...' boz to integer/real conversions [PR96859]
The standard says that excess digits from boz are truncated.
For hexadecimal or binary, the routines copy just the number of digits
that will be needed, but for octal we copy number of digits that
contain one extra bit (for 8-bit, 32-bit or 128-bit, i.e. kind 1, 4 and 16)
or two extra bits (for 16-bit or 64-bit, i.e. kind 2 and 8).
The clearing of the first bit is done correctly by changing the first digit
if it is 4-7 to one smaller by 4 (i.e. modulo 4).
The clearing of the first two bits is done by changing 4 or 6 to 0
and 5 or 7 to 1, which is incorrect, because we really want to change the
first digit to 0 if it was even, or to 1 if it was odd, so digits
2 and 3 are mishandled by keeping them as is, rather than changing 2 to 0
and 3 to 1.
2020-09-02 Jakub Jelinek <jakub@redhat.com>
PR fortran/96859
* check.c (gfc_boz2real, gfc_boz2int): When clearing first two bits,
change also '2' to '0' and '3' to '1' rather than just handling '4'
through '7'.
BPF is an ELF-based target, so it definitely benefits from using
elfos.h. This patch makes the target to use it, and removes
superfluous definitions from bpf.h which are better defined in
elfos.h.
Note that BPF, despite being an ELF target, doesn't use DWARF. At
some point it will generate DWARF when generating xBPF (-mxbpf) and
BTF when generating plain eBPF, but for the time being it just
generates stabs.
Marek Polacek [Tue, 1 Sep 2020 15:44:16 +0000 (11:44 -0400)]
c++: Allow new char[4]{"foo"} [PR77841]
Currently, we allow new char[]{"foo"}, but not new char[4]{"foo"}.
We should accept the latter too: [dcl.init.list]p3.3 says to treat
this as [dcl.init.string].
We were rejecting this code because we never called reshape_init before
the digest_init in build_new_1. reshape_init handles [dcl.init.string]
by unwrapping the STRING_CST from its enclosing { }, and digest_init
assumes that reshape_init has been called for aggregates anyway, and an
array is an aggregate.
This change evaluates __glibcxx_assert checks unconditionally when a
function is being constant evaluated (when std::is_constant_evaluated()
is true). If the check fails, compilation will fail with an error.
If the function isn't being constant evaluated, the normal runtime check
will be done if enabled by _GLIBCXX_ASSERTIONS or _GLIBCXX_DEBUG, the
same as before.
Tangentially, the __glibcxx_assert and _GLIBCXX_PARALLEL_ASSERT macros
are changed to expand to 'do { } while (false)' when assertions are
disabled, instead of expanding to nothing. This avoids -Wempty-body
warnings when a disabled assertion is used in an 'if' or 'else'
statement e.g.
if constexpr (/* precondition is testable */)
__glibcxx_assert(precondition);
a.C:9:27: warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body]
9 | __glibcxx_assert(precondition);
| ^
libstdc++-v3/ChangeLog:
PR libstdc++/71960
* include/bits/c++config (__glibcxx_assert_impl): Remove
do-while so that uses of the macro need to add it.
(__glibcxx_assert): Rename macro for runtime assertions
to __glibcxx_assert_2.
(__glibcxx_assert_1): Define macro for constexpr assertions.
(__glibcxx_assert): Define macro for constexpr and runtime
assertions.
* include/bits/range_access.h (ranges::advance): Remove
redundant precondition checks during constant evaluation.
* include/parallel/base.h (_GLIBCXX_PARALLEL_ASSERT): Always
use do-while in macro expansion.
* include/std/ranges (iota_view::iota_view(W, B)): Remove
redundant braces.
Peter Bergner [Tue, 1 Sep 2020 18:47:44 +0000 (13:47 -0500)]
rs6000: MMA built-in dies with incorrect sharing of tree nodes error
When we expand our MMA built-ins into gimple, we erroneously reused the
accumulator memory reference for both the source input value as well as
the destination output value. This led to a tree sharing error.
The solution is to create separate memory references for the input
and output values.
2020-09-01 Peter Bergner <bergner@linux.ibm.com>
gcc/
PR target/96808
* config/rs6000/rs6000-call.c (rs6000_gimple_fold_mma_builtin): Do not
reuse accumulator memory reference for source and destination accesses.
gcc/testsuite/
PR target/96808
* gcc.target/powerpc/pr96808.c: New test.
The chrono::duration constructor that converts from another duration
type is meant to be constrained so that it doesn't participate in
overload resolution if the ratio of the periods cannot be represented as
a std::ratio.
Because our std::ratio_divide is not SFINAE-friendly the evaluation of
__is_harmonic results in an error outside the immediate context when an
overflow occurs. I intend to make ratio_divide (and ratio_multiply)
SFINAE-friendly in a future patch, but for now this patch just
introduces a new SFINAE-friendly alias template for the division.
The standard doesn't require it, but it also seems right to constrain
the constructor with std::is_convertible_v<_Rep2, rep>.
libstdc++-v3/ChangeLog:
* include/std/chrono (duration::_S_gcd(intmax_t, intmax_t)):
New helper function for finding GCD of two positive intmax_t
values.
(duration::__divide): New helper alias for dividing one period
by another.
(duration::__is_harmonic): Use __divide not ratio_divide.
(duration(const duration<R2, P2>&)): Require the duration rep
types to be convertible.
* testsuite/20_util/duration/cons/dr2094.cc: New test.
* testsuite/20_util/duration/requirements/reduced_period.cc:
Fix definition of unused member functions in test type.
* testsuite/20_util/duration/requirements/typedefs_neg2.cc:
Adjust expected errors.
Roger Sayle [Tue, 1 Sep 2020 11:03:21 +0000 (12:03 +0100)]
PR middle-end/90597: gcc_assert ICE in layout_type
This patch fixes the default implementation of TARGET_VECTOR_ALIGNMENT,
known as default_vector_alignment, using the same logic as my earlier
nvptx patch, as the ICE caused by TYPE_SIZE(type) being zero during
error handling in gcc.dg/attr-vector_size.c is common among backends,
and is known in bugzilla as PR middle-end/90597, apparently a recent
regression.
2020-09-01 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog:
PR middle-end/90597
* targhooks.c (default_vector_alignment): Return at least the
GET_MODE_ALIGNMENT for the type's mode.
Mark Eggleston [Mon, 1 Jun 2020 07:15:31 +0000 (08:15 +0100)]
Fortran : ICE on invalid code PR95398
The CLASS_DATA macro is used to shorten the code accessing the derived
components of an expressions type specification. If the type is not
BT_CLASS the derived pointer is NULL resulting in an ICE. To avoid
dereferencing a NULL pointer the type should be BT_CLASS.
2020-09-01 Steven G. Kargl <kargl@gcc.gnu.org>
gcc/fortran
PR fortran/95398
* resolve.c (resolve_select_type): Add check for BT_CLASS
type before using the CLASS_DATA macro which will have a
NULL pointer to derive components if it isn't BT_CLASS.
2020-09-01 Mark Eggleston <markeggleston@gcc.gnu.org>
gcc/testsuite
PR fortran/95398
* gfortran.dg/pr95398.f90: New test.
Richard Biener [Thu, 27 Aug 2020 12:21:33 +0000 (14:21 +0200)]
rtl-optimization/96812 - remap dependence info on RTL loop unrolling
This carries over the PR87609 fix also to RTL loop unrolling. The
gcc.dg/torture/pr90328.c testcase otherwise is miscompiled with
the tree-ssa-address.c hunk (or alternatively with -fno-ivopts
on master). I've tried to find the correct abstraction and
adjusted two other duplicate_insn_chain users for which I do not
have testcases. There may be other insn-chain copying routines
that could be affected but hopefully most appropriately go through
CFG hooks.
2020-08-27 Richard Biener <rguenther@suse.de>
PR rtl-optimization/96812
* tree-ssa-address.c (copy_ref_info): Also copy dependence info.
* cfgrtl.h (duplicate_insn_chain): Adjust prototype.
* cfgrtl.c (duplicate_insn_chain): Remap dependence info
if requested.
(cfg_layout_duplicate_bb): Make sure we remap dependence info.
* modulo-sched.c (duplicate_insns_of_cycles): Remap dependence
info.
(generate_prolog_epilog): Adjust.
* config/c6x/c6x.c (hwloop_optimize): Remap dependence info.
Jakub Jelinek [Tue, 1 Sep 2020 07:17:58 +0000 (09:17 +0200)]
openmp: Check for PARM_DECL before using C_ARRAY_PARAMETER or DECL_ARRAY_PARAMETER_P [PR96867]
The C++ macro performs a PARM_DECL_CHECK, so will ICE if not tested on a PARM_DECL,
C_ARRAY_PARAMETER doesn't, but probably should, otherwise it is testing e.g.
C_DECL_VARIABLE_SIZE on VAR_DECLs.
2020-09-01 Jakub Jelinek <jakub@redhat.com>
PR c++/96867
* c-typeck.c (handle_omp_array_sections_1): Test C_ARRAY_PARAMETER
only on PARM_DECLs.
* semantics.c (handle_omp_array_sections_1): Test
DECL_ARRAY_PARAMETER_P only on PARM_DECLs.
Kewen Lin [Tue, 1 Sep 2020 02:37:41 +0000 (02:37 +0000)]
test/rs6000: Add Power9 and up as vect_len target
Power9 supports vector with length in bytes load/store, this patch
is to teach check_effective_target_vect_len_load_store to take it
and its laters as effective vector with length targets.
Also supplement the documents for has_arch_pwr*.
Bootstrapped/regtested on powerpc64le-linux-gnu P8, also on
powerpc64le-linux-gnu P9 with explicit usage setting.
David Malcolm [Mon, 31 Aug 2020 20:20:55 +0000 (16:20 -0400)]
analyzer: gather builtin/internal fn handling into switch statements
Clean up this code in preparation for fixing PR analyzer/96798.
gcc/analyzer/ChangeLog:
* region-model.cc (region_model::on_call_pre): Gather handling of
builtins and of internal fns into switch statements. Handle
"alloca" and BUILT_IN_ALLOCA_WITH_ALIGN.
David Malcolm [Mon, 31 Aug 2020 13:00:23 +0000 (09:00 -0400)]
analyzer: fix ICE on unknown index in CONSTRUCTOR [PR96860]
PR analyzer/96860 reports an ICE inside CONSTRUCTOR-handling with
--param analyzer-max-svalue-depth=0 when attempting to build a
binding_map for the CONSTRUCTOR's values.
The issue is that when handling (index, value) pairs for initializing
an array, the index values for the elements exceeds the svalue
complexity limit, and the index is thus treated as unknown, leading to
a symbolic rather than concrete offset for each array element.
This patch updates the CONSTRUCTOR-handling code so that it can
fail, returning an unknown value for the overall value of the
constructor for this case, fixing the ICE.
gcc/analyzer/ChangeLog:
PR analyzer/96860
* region.cc (decl_region::get_svalue_for_constructor): Support
apply_ctor_to_region failing.
* store.cc (binding_map::apply_ctor_to_region): Add failure
handling.
(binding_map::apply_ctor_val_to_range): Likewise.
(binding_map::apply_ctor_pair_to_child_region): Likewise. Replace
assertion that child_base_offset is not symbolic with error
handling.
* store.h (binding_map::apply_ctor_to_region): Convert return type
from void to bool.
(binding_map::apply_ctor_val_to_range): Likewise.
(binding_map::apply_ctor_pair_to_child_region): Likewise.
gcc/testsuite/ChangeLog:
PR analyzer/96860
* gcc.dg/analyzer/pr96860-1.c: New test.
* gcc.dg/analyzer/pr96860-2.c: New test.
Marek Polacek [Thu, 13 Aug 2020 18:56:13 +0000 (14:56 -0400)]
c++: Implement P1009: Array size deduction in new-expressions.
This patch implements C++20 P1009, allowing code like
new double[]{1,2,3}; // array bound will be deduced
Since this proposal makes the initialization rules more consistent, it is
applied to all previous versions of C++ (thus, effectively, all the way back
to C++11).
My patch is based on Jason's patch that handled the basic case. I've
extended it to work with ()-init and also the string literal case.
Further testing revealed that to handle stuff like
new int[]{t...};
in a template, we have to consider such a NEW_EXPR type-dependent.
Obviously, we first have to expand the pack to be able to deduce the
number of elements in the array.
Curiously, while implementing this proposal, I noticed that we fail
to accept
new char[4]{"abc"};
so I've assigned 77841 to self. I think the fix will depend on the
build_new_1 hunk in this patch.
The new tree.c function build_constructor_from_vec helps us morph
a vector into a CONSTRUCTOR more efficiently.
gcc/cp/ChangeLog:
PR c++/93529
* call.c (build_new_method_call_1): Use build_constructor_from_vec
instead of build_tree_list_vec + build_constructor_from_list.
* init.c (build_new_1): Handle new char[]{"foo"}. Use
build_constructor_from_vec instead of build_tree_list_vec +
build_constructor_from_list.
(build_new): Deduce the array size in new-expression if not
present. Handle ()-init. Handle initializing an array from
a string literal.
* parser.c (cp_parser_new_type_id): Leave [] alone.
(cp_parser_direct_new_declarator): Allow [].
* pt.c (type_dependent_expression_p): In a NEW_EXPR, consider
array types whose dimension has to be deduced type-dependent.
PR c++/93529
* g++.dg/cpp0x/sfinae4.C: Adjust expected result after P1009.
* g++.dg/cpp2a/new-array1.C: New test.
* g++.dg/cpp2a/new-array2.C: New test.
* g++.dg/cpp2a/new-array3.C: New test.
* g++.dg/cpp2a/new-array4.C: New test.
David Malcolm [Mon, 24 Aug 2020 13:33:42 +0000 (09:33 -0400)]
analyzer: fix ICE on RANGE_EXPR in CONSTRUCTORs [PR96763]
gcc/analyzer/ChangeLog:
PR analyzer/96763
* store.cc (binding_map::apply_ctor_to_region): Handle RANGE_EXPR
by calling a new binding_map::apply_ctor_val_to_range subroutine.
Split out the existing non-CONSTRUCTOR-handling code to a new
apply_ctor_pair_to_child_region subroutine.
(binding_map::apply_ctor_val_to_range): New.
(binding_map::apply_ctor_pair_to_child_region): New, split out
from binding_map::apply_ctor_to_region as noted above.
* store.h (binding_map::apply_ctor_val_to_range): New decl.
(binding_map::apply_ctor_pair_to_child_region): New decl.
gcc/testsuite/ChangeLog:
PR analyzer/96763
* g++.dg/analyzer/pr96763.C: New test.
Iain Buclaw [Mon, 31 Aug 2020 11:31:04 +0000 (13:31 +0200)]
d: Fix ICEs in the front-end when pointer size is 16-bit.
In the lowering of `bt*' intrinsics, some integer constants had
mismatched types, and bitsize was set to the wrong value.
In base_vtable_offset, the base offset value was calculated incorrectly.
The TypeInfo_Class object is comprised of 18 pointers and 1 uint field,
so now the internal classinfo type size is used instead.
gcc/d/ChangeLog:
* d-target.cc (Target::_init): Don't set classinfosize.
* d-tree.h (base_vtable_offset): Move under typeinfo.cc section.
* decl.cc (base_vtable_offset): Move to...
* typeinfo.cc (base_vtable_offset): ...here. Get base offset from
internal TypeInfo_Class type.
* intrinsics.cc (expand_intrinsic_bt): Use pointer TYPE_SIZE for
setting bitsize value. Build integer constants of correct type.
gcc/
PR target/96551
* config/i386/sse.md (vec_unpacku_float_hi_v16si): For vector
compare to integer mask, don't use gen_rtx_LT, use
ix86_expand_mask_vec_cmp instead.
(vec_unpacku_float_hi_v16si): Ditto.
gcc/testsuite
* gcc.target/i386/avx512f-pr96551-1.c: New test.
* gcc.target/i386/avx512f-pr96551-2.c: New test.
Jakub Jelinek [Mon, 31 Aug 2020 08:51:02 +0000 (10:51 +0200)]
tree-cfg: Improve gimple switch verification
When looking at the verification, I have noticed a bug in it.
The verification that CASE_HIGH (if present) has the same type as CASE_LOW
is only performed for the case label 2 and higher, case label 1 (the first
one after the default label) isn't checked.
The following patch fixes that, it will uselessly also compare
TREE_TYPE (CASE_LOW (elt)) != elt_type for the case label 1, but I think
that isn't that expensive and helps readability of the code.
2020-08-31 Jakub Jelinek <jakub@redhat.com>
* tree-cfg.c (verify_gimple_switch): If the first non-default case
label has CASE_HIGH, verify it has the same type as CASE_LOW.
testcase merges all the 4 scalar constant pool entries into the CONST_VECTOR
one.
I'm punting for section anchors and not doing it in the per-function (i.e.
non-shared) constant pools simply because I don't know them well enough,
don't know whether backends use the offsets for something etc.
For section anchors, I guess it would need to be done before (re)computing the
offsets and arrange for the desc->mark < 0 entries not to be considered as
objects in the object block, for non-shared pools, perhaps it would be
enough to call the new function from output_constant_pool before calling
recompute_pool_offsets and adjust recompute_pool_offsets to ignore
desc->mark < 0.
Here is an adjusted patch that ought to merge even the same sized different
mode vectors with the same byte representation, etc.
It won't really help with avoiding the multiple reads of the constant in the
same function, but as you found, your patch doesn't help with that either.
Your patch isn't really incompatible with what the patch below does, though
I wonder whether a) it wouldn't be better to always canonicalize to an
integral mode with as few elts as possible even e.g. for floats b) whether
asserting that it simplify_rtx succeeds is safe, whether it shouldn't just
canonicalize if the canonicalization works and just do what it previously
did otherwise.
The following patch puts all pool entries which can be natively encoded
into a vector, sorts it by decreasing size, determines minimum size
of a pool entry and adds hash elts for each (aligned) min_size or wider
power of two-ish portion of the pool constant in addition to the whole pool
constant byte representation.
This is the version that passed bootstrap/regtest on both x86_64-linux and
i686-linux. In both bootstraps/regtests together, it saved (from the
statistics I've gathered) 63104 .rodata bytes (before constant merging),
in 6814 hits of the data->desc->mark = ~(*slot)->desc->labelno;.
2020-08-31 Jakub Jelinek <jakub@redhat.com>
PR middle-end/54201
* varasm.c: Include alloc-pool.h.
(output_constant_pool_contents): Emit desc->mark < 0 entries as
aliases.
(struct constant_descriptor_rtx_data): New type.
(constant_descriptor_rtx_data_cmp): New function.
(struct const_rtx_data_hasher): New type.
(const_rtx_data_hasher::hash, const_rtx_data_hasher::equal): New
methods.
(optimize_constant_pool): New function.
(output_shared_constant_pool): Call it if TARGET_SUPPORTS_ALIASES.
Kewen Lin [Mon, 31 Aug 2020 00:52:03 +0000 (19:52 -0500)]
testsuite: Update some vect cases for partial vectors
This patch is to adjust some existing vectorization test cases
to work well with the newly introduced partial vector usages.
Bootstrapped/regtested on aarch64-linux-gnu and powerpc64le-linux-gnu
P9 (with explicit param vect-partial-vector-usage=1 and enablement on
check_effective_target_vect_partial_vectors_usage_1 check).
2020-8-20 José Rui Faustino de Sousa <jrfsousa@gmail.com>
gcc/fortran/ChangeLog:
PR fortran/96728
* module.c (module_peek_char): Peek ahead function.
(parse_integer): Add code for parsing signed integers.
(parse_atom): Add code to handle signed integers.
(peek_atom): Add code to handle signed integers.
gcc/testsuite/ChangeLog:
PR fortran/96728
* gfortran.dg/PR96728.f90: New test.
2020-8-20 José Rui Faustino de Sousa <jrfsousa@gmail.com>
gcc/fortran/ChangeLog:
PR fortran/94110
* interface.c (gfc_compare_actual_formal): Add code to also raise
the actual argument cannot be an assumed-size array error when the
dummy arguments are deferred-shape or assumed-rank pointer.
gcc/testsuite/ChangeLog:
PR fortran/94110
* gfortran.dg/PR94110.f90: New test.