Commit 229752afe315 ("VEC_COND_EXPR optimizations") has improved code
generation: we no longer need "vx x,x,-1", which turned out to be
superfluous. Instead, we simply swap 0 and -1 arguments of the
preceding "vsel".
gcc/testsuite/ChangeLog:
2020-11-23 Ilya Leoshkevich <iii@linux.ibm.com>
* gcc.target/s390/zvector/autovec-double-quiet-uneq.c: Expect
that "vx" is not emitted.
* gcc.target/s390/zvector/autovec-float-quiet-uneq.c: Likewise.
This patch implements the following set of changes:
1. If a component flag of -ffast-math (or -funsafe-math-optimizations)
is explicitly set (or reset) on the command line, this should override
any implicit change due to -f(no-)fast-math, no matter in which order
the flags come on the command line. This change affects all flags.
2. Any component flag modified from its default by -ffast-math should
be reset to the default by -fno-fast-math. This was previously
not done for the following flags:
-fcx-limited-range
-fexcess-precision=
3. Once -ffinite-math-only is true, the -f(no-)signaling-nans flag has
no meaning (if we have no NaNs at all, it does not matter whether
there is a difference between quiet and signaling NaNs). Therefore,
it does not make sense for -ffast-math to imply -fno-signaling-nans.
(This is also a documentation change.)
4. -ffast-math is documented to imply -fno-rounding-math, however the
latter setting is the default anyway; therefore it does not make
sense to try to modify it from its default setting.
5. The __FAST_MATH__ preprocessor macro should be defined if and only
if all the component flags of -ffast-math are set to the value that
is documented as the effect of -ffast-math. The following flags
were currently *not* so tested:
-fcx-limited-range
-fassociative-math
-freciprocal-math
-frounding-math
(Note that we should still *test* for -fno-rounding-math here even
though it is not set as to 4. -ffast-math -frounding-math should
not set the __FAST_MATH__ macro.)
This is also a documentation change.
2020-11-24 Ulrich Weigand <uweigand@de.ibm.com>
gcc/
* doc/invoke.texi (-ffast-math): Remove mention of -fno-signaling-nans.
Clarify conditions when __FAST_MATH__ preprocessor macro is defined.
* opts.c (common_handle_option): Pass OPTS_SET to set_fast_math_flags
and set_unsafe_math_optimizations_flags.
(set_fast_math_flags): Add OPTS_SET argument, and use it to avoid
setting flags already explicitly set on the command line. In the !set
case, also reset x_flag_cx_limited_range and x_flag_excess_precision.
Never reset x_flag_signaling_nans or x_flag_rounding_math.
(set_unsafe_math_optimizations_flags): Add OPTS_SET argument, and use
it to avoid setting flags already explicitly set on the command line.
(fast_math_flags_set_p): Also test x_flag_cx_limited_range,
x_flag_associative_math, x_flag_reciprocal_math, and
x_flag_rounding_math.
gcc/ada/
* exp_attr.adb (Expand_N_Attribute_Reference): Replace calls to
Sloc with a local constant Loc; remove call to
Analyze_And_Resolve and return, which is exactly what happens
anyway (and other branches in the Constrained declare block
appear to rely on analysis, resolution and returning happening
in all cases).
* sem_util.adb: Remove useless parens.
gcc/ada/
* sem_type.adb (Add_One_Interp.Is_Universal_Operation): Account
for universal_access = operator.
(Disambiguate): Take into account preference on universal_access
= operator when relevant.
(Disambiguate.Is_User_Defined_Anonymous_Access_Equality): New.
gcc/ada/
* exp_util.adb (Is_Finalizable_Transient): Take into account return
statements containing N_Expression_With_Actions. Also clean up a
condition to make it more readable.
* exp_ch6.adb: Fix typo.
gcc/ada/
* sem_ch13.adb (Validate_Literal_Aspect): Add support for named
numbers and in particular overload of the Real_Literal function.
* sem_res.adb (Resolve): Add support for named numbers in
Real_Literal and Integer_Literal resolution.
* einfo.adb, einfo.ads (Related_Expression,
Set_Related_Expression): Allow E_Function.
* uintp.ads (UI_Image_Max): Bump size of buffer to avoid loosing
precision.
* sem_eval.adb: Fix typo in comment.
* libgnat/a-nbnbin.adb, libgnat/a-nbnbin.ads (From_String):
Return a Valid_Big_Integer.
* libgnat/a-nbnbre.adb, libgnat/a-nbnbre.ads (From_String): New
variant taking two strings. Return a Valid_Big_Real.
gcc/ada/
* sem_ch12.adb (Analyze_Associations) <Explicit_Freeze_Check>: Test
that the instance is in a statement sequence instead of local scope.
(Freeze_Subprogram_Body): Use the special delayed placement with
regard to the parent instance only if its Sloc is strictly greater.
(Install_Body): Likewise.
gcc/ada/
* sem_ch13.adb (Validate_Literal_Aspect): Call to Base_Type
needed in order to correctly check result type of String_Literal
function when the first named subtype differs from the base
type (e.g.:
type T is range 1 .. 10 with String_Literal => ... ;
).
gcc/ada/
* sem_prag.adb (Analyze_Global_Item): Handle specially the
current instance of a PO.
* sem_util.ads (Is_Effectively_Volatile,
Is_Effectively_Volatile_For_Reading): Add parameter
Ignore_Protected.
* sem_util.adb (Is_Effectively_Volatile,
Is_Effectively_Volatile_For_Reading): Add parameter
Ignore_Protected to compute the query results ignoring protected
objects/types.
(Is_Effectively_Volatile_Object,
Is_Effectively_Volatile_Object_For_Reading): Adapt to new
signature.
gcc/ada/
* exp_ch6.adb (Add_Cond_Expression_Extra_Actual): Simplify
handling of function calls and remove bug in handling of
transient objects. Minor reformatting along the way.
gcc/ada/
* sem_aggr.adb (Resolve_Delta_Array_Aggregate): If the choice is
a subtype_indication then call
Resolve_Discrete_Subtype_Indication; both for choices
immediately inside array delta aggregates and inside
iterated_component_association within array delta aggregates.
gcc/ada/
* exp_spark.adb (Expand_SPARK_Array_Aggregate,
Expand_SPARK_N_Aggregate): Remove, no longer needed.
* sem_aggr.adb (Resolve_Iterated_Component_Association): Only
remove references in the analyzed expression when generating
code and the expression needs to be analyzed anew after being
rewritten into a loop.
As the following testcase shows, unlike char, int or long long sized
__builtin_*_overflow{,_p}, for short sized one in most cases the ce1 pass
doesn't optimize the jo/jno or jc/jnc jumps with setting of a pseudo to 0/1
into seto/setc. The reason is missing *setcc_hi_1* pattern. The following
patch implements it using mode iterators so that on i486 and pentium?
one can get the zero extension through and instead of movzbw.
2020-11-24 Jakub Jelinek <jakub@redhat.com>
PR target/97950
* config/i386/i386.md (*setcc_si_1_and): Macroize into...
(*setcc_<mode>_1_and): New define_insn_and_split with SWI24 iterator.
(*setcc_si_1_movzbl): Macroize into...
(*setcc_<mode>_1_movzbl): New define_insn_and_split with SWI24
iterator.
* gcc.target/i386/pr97950.c: New test.
Currently the __builtin_clear_padding expansion code emits no code for
full words that don't have any padding bits, and most of the time if
the only padding bytes are from the start of the word it attempts to merge
them with previous padding store (via {}) or if the only padding bytes are
from the end of the word, it attempts to merge it with following padding
bytes. For everything else it was using a RMW, except when it found
an aligned char/short/int covering all the padding bytes and all those
padding bytes were all ones in that store.
The following patch changes it, such that we only use RMW if the padding has
any bytes which have some padding and some non-padding bits (i.e. bitfields
are involved), often it is the same amount of instructions in the end and
avoids being thread-unsafe unless necessary (and avoids having to wait for
the reads to make it into the CPU). So, if there are no bitfields,
the function will just store some zero bytes, shorts, ints, long longs etc.
where needed.
2020-11-24 Jakub Jelinek <jakub@redhat.com>
* gimple-fold.c (clear_padding_flush): If a word contains only 0
or 0xff bytes of padding other than all set, all clear, all set
followed by all clear or all clear followed by all set, don't emit
a RMW operation on the whole word or parts of it, but instead
clear the individual bytes of padding. For paddings of one byte
size, don't use char[1] and {}, but instead just char and 0.
This testcase started failing with r8-2090 and works again starting
with r11-4755.
2020-11-24 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/97964
* gcc.dg/tree-ssa/pr97964.c: New test.
In particular, more precisely highlight what applies generally vs. the special
handling for the current 'parloops'-based OpenACC 'kernels' implementation.
gcc/
* omp-expand.c (expand_oacc_for): More explicit checking of which
OMP constructs we're expecting.
The diagnostics produced by 'dg-optimized', 'dg-missed' aren't error
diagnostics (fatal, meaning: causes compilation to fail) but rather warning
diagnostics (non-fatal, doesn't cause compilation to fail). Thus, same as
'dg-message', these should use 'saved-dg-warning' instead of 'saved-dg-error',
which then prints: "(test for *warnings*, line [...]) instead of currently:
"(test for *errors*, line [...])".
This is a small bug-fix for commit ed2d9d3720adef3a260b8a55e17e744352a901fc
"dumpfile.c: use prefixes other than 'note: ' for
MSG_{OPTIMIZED_LOCATIONS|MISSED_OPTIMIZATION}", which added 'dg-optimized',
'dg-missed'.
gcc/testsuite/
* lib/gcc-dg.exp (dg-optimized, dg-missed): Use 'saved-dg-warning'
instead of 'saved-dg-error'.
'process-message' would like the 'msgprefix' argument without trailing space.
This is a small bug-fix for commit ed2d9d3720adef3a260b8a55e17e744352a901fc
"dumpfile.c: use prefixes other than 'note: ' for
MSG_{OPTIMIZED_LOCATIONS|MISSED_OPTIMIZATION}", which added 'dg-optimized',
'dg-missed'.
gcc/testsuite/
* lib/gcc-dg.exp (dg-optimized, dg-missed): Fix 'process-message'
call.
* gcc.dg/vect/nodump-vect-opt-info-1.c: Demonstrate.
* gcc.dg/vect/nodump-vect-opt-info-2.c: Likewise.
c_parser_binary_expression was using build2 to create a temporary holder
for binary expression that c_parser_atomic and c_finish_omp_atomic can then
handle. The latter performs then all the needed checking.
Unfortunately, build2 performs some checking too, e.g. PLUS_EXPR vs.
POINTER_PLUS_EXPR or matching types of the arguments, nothing we can guarantee
at the parsing time. So we need something like C++ build_min_nt*. This
patch implements that inline.
2020-11-24 Jakub Jelinek <jakub@redhat.com>
PR c/97958
* c-parser.c (c_parser_binary_expression): For omp atomic binary
expressions, use make_node instead of build2 to avoid checking build2
performs.
* c-c++-common/gomp/pr97958.c: New test.
The PR38359 change made the -1 >> x to -1 optimization less useful by
requiring that the x must be non-negative.
Shifts by negative amount are UB, but we for historic reasons had in some
(but not all) places some hack to treat shifts by negative value as the
other direction shifts by the negated amount.
The following patch just removes that special handling, instead we punt on
optimizing those (and ideally path isolation should catch that up and turn
those into __builtin_unreachable, perhaps with __builtin_warning next to
it). Folding the shifts in some places as if they were rotates and in other
as if they were saturating just leads to inconsistencies.
For C++ constexpr diagnostics and -fpermissive, I've added code to pretend
fold-const.c has not changed, without -fpermissive it will be an error
anyway and I think it is better not to change all the diagnostics.
During x86_64-linux and i686-linux bootstrap/regtest, my statistics
gathering patch noted 185 unique -m32/-m64 x TU x function_name x shift_kind
x fold-const/tree-ssa-ccp cases. I have investigated the
64 ../../gcc/config/i386/i386.c x86_output_aligned_bss LSHIFT_EXPR wide_int_bitop
64 ../../gcc/config/i386/i386-expand.c emit_memmov LSHIFT_EXPR wide_int_bitop
64 ../../gcc/config/i386/i386-expand.c ix86_expand_carry_flag_compare LSHIFT_EXPR wide_int_bitop
64 ../../gcc/expmed.c expand_divmod LSHIFT_EXPR wide_int_bitop
64 ../../gcc/lra-lives.c process_bb_lives LSHIFT_EXPR wide_int_bitop
64 ../../gcc/rtlanal.c nonzero_bits1 LSHIFT_EXPR wide_int_bitop
64 ../../gcc/varasm.c optimize_constant_pool.isra LSHIFT_EXPR wide_int_bitop
cases and all of them are either during jump threading (dom) or during PRE.
For jump threading, the most common case is 1 << floor_log2 (whatever) where
floor_log2 is return HOST_BITS_PER_WIDE_INT - 1 - clz_hwi (x);
and clz_hwi is if (x == 0) return HOST_BITS_PER_WIDE_INT; return __builtin_clz* (x);
and so has range [-1, 63] and a comparison against == 0 which makes the
threader think it might be nice to jump thread the case leading to 1 << -1.
I think it is better to keep the 1 << -1 s in the IL for this and let path
isolation turn that into __builtin_unreachable () if the user wishes so.
2020-11-24 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/96929
* fold-const.c (wide_int_binop) <case LSHIFT_EXPR, case RSHIFT_EXPR>:
Return false on negative second argument rather than trying to handle
it as shift in the other direction.
* tree-ssa-ccp.c (bit_value_binop) <case LSHIFT_EXPR,
case RSHIFT_EXPR>: Punt on negative shift count rather than trying
to handle it as shift in the other direction.
* match.pd (-1 >> x to -1): Remove tree_expr_nonnegative_p check.
* constexpr.c (cxx_eval_binary_expression): For shifts by constant
with MSB set, emulate older wide_int_binop behavior to preserve
diagnostics and -fpermissive behavior.
* gcc.dg/tree-ssa/pr96929.c: New test.
Commit r11-3393 improved the epilogue loop handling of partial
vectors and we won't use partial vectors to vectorize a single
iteration scalar loop any more.
The affected test cases have only one single iteration in their
epilogues, so we shouldn't expect the vectorization with
partial vector there.
Tested with explicit --param=vect-partial-vector-usage=1 and
default enablement.
gcc/testsuite/ChangeLog:
* gcc.dg/vect/slp-perm-1.c: Adjust for partial vectors.
* gcc.dg/vect/slp-perm-5.c: Likewise.
* gcc.dg/vect/slp-perm-6.c: Likewise.
* gcc.dg/vect/slp-perm-7.c: Likewise.
The issue here is that rpo vn may eliminate target ssa_name referred to in
redundant_ssa_names, and thus ifcvt_local_dce may replace candidate
ssa_name with invalid ssa_name resulting in incorrect IR. The patch simply
does ssa_name replacement before calling do_rpo_vn, which fixes the issue.
gcc/
2020-11-24 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
PR tree-optimization/97849
* tree-if-conv.c (tree_if_conversion): Move ssa_name
replacement code from ifcvt_local_dce to this function
before calling do_rpo_vn.
gcc/testsuite/
2020-11-24 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
PR tree-optimization/97849
* gcc.dg/tree-ssa/pr97849.c: New test.
gcc/ChangeLog:
* tree-cfg.c (dump_function_to_file): Print type attributes
and return type.
gcc/testsuite/ChangeLog:
* gcc.dg/attr-access-5.c: New test.
As noted in bug 95630, C11 removed a restriction in C99 on comparing
pointers to compatible complete and incomplete types (this was one of
the changes in N1439, which was largely a terminological change to
make incomplete types a subset of object types rather than a different
kind of type). Implement that change by using pedwarn_c99 with
OPT_Wpedantic for this diagnostic.
Bootstrapped with no regressions for x86_64-pc-linux-gnu.
gcc/c/
2020-11-23 Joseph Myers <joseph@codesourcery.com>
PR c/95630
* c-typeck.c (build_binary_op): Use pedwarn_c99 with OPT_Wpedantic
for comparisons of complete and incomplete pointers.
gcc/testsuite/
2020-11-23 Joseph Myers <joseph@codesourcery.com>
PR c/95630
* gcc.dg/c11-compare-incomplete-1.c,
gcc.dg/c11-compare-incomplete-2.c,
gcc.dg/c99-compare-incomplete-1.c,
gcc.dg/c99-compare-incomplete-2.c: New tests.
when Fortran functions pass array descriptors they receive as a
parameter to another function, they actually rebuild it. Thanks to
work done mainly by Feng, IPA-CP can already handle the cases when
they pass directly the values loaded from the original descriptor.
Unfortunately, perhaps the most important one, stride, is first
checked against zero and is replaced with one in that case:
_12 = *a_11(D).dim[0].stride;
if (_12 != 0)
goto <bb 4>; [50.00%]
else
goto <bb 3>; [50.00%]
<bb 3>
// empty BB
<bb 4>
# iftmp.22_9 = PHI <_12(2), 1(3)>
...
parm.6.dim[0].stride = iftmp.22_9;
...
__x_MOD_foo (&parm.6, b_31(D));
in the most important and hopefully common cases, the incoming value
is already 1 and we fail to propagate it.
I would therefore like to propose the following way of encoding this
situation in pass-through jump functions using using ASSERTT_EXPR
operation code meaning that if the incoming value is the same as the
"operand" in the jump function, it is passed on, otherwise the result
is unknown. This of course captures only the single (but most
important) case but is an improvement and does not need enlarging the
jump function structure and is simple to pattern match. Encoding that
zero needs to be changed to one would need another field and matching
it would be slightly more complicated too.
gcc/
2020-06-12 Martin Jambor <mjambor@suse.cz>
* ipa-prop.h (ipa_pass_through_data): Expand comment describing
operation.
* ipa-prop.c (analyze_agg_content_value): Detect new special case and
encode it as ASSERT_EXPR.
* ipa-cp.c (values_equal_for_ipcp_p): Move before
ipa_get_jf_arith_result.
(ipa_get_jf_arith_result): Special case ASSERT_EXPR.
gcc/testsuite/
2020-06-12 Martin Jambor <mjambor@suse.cz>
* gfortran.dg/ipcp-array-2.f90: New test.
So the primary purpose of this patch is to make it easier to write tests for
removal of useless test/compare insns on the H8.
In simplest terms the costing model in the H8 port tends to encourage changing
something like:
x = y + 4;
if (x == 0)
into:
x = y + 4;
if (y == -4)
This is a marginal de-optimization on the H8. So fixing it makes the code
ever-so-slightly better in isolation. Fixing this also improves redundant
test/compare elimination and makes writing tests for redundant test/compare
elimination far easier.
gcc/
* config/h8300/h8300.c (h8300_rtx_costs): Handle the various
comparison rtx codes too.
