This combines the two differently named versions of the same function
into a single function utilizing a default argument.
gcc/cp/ChangeLog:
* cp-tree.h (cp_build_qualified_type_real): Rename to ...
(cp_build_qualified_type): ... this. Give its last parameter
a default argument. Remove macro of the same name.
* decl.cc (grokdeclarator): Adjust accordingly.
* pt.cc (tsubst_aggr_type): Likewise.
(rebuild_function_or_method_type): Likewise.
(tsubst): Likewise.
(maybe_dependent_member_ref): Likewise.
(unify): Likewise.
* tree.cc (cp_build_qualified_type_real): Rename to ...
(cp_build_qualified_type): ... this. Adjust accordingly.
A few globs missed in the .c -> .cc transition. Some targets were looking
at both *.c and *.cc, but there are no longer any .c files to scan.
gcc/ChangeLog:
* Makefile.in (TAGS): Look at libcpp/*.cc.
gcc/c/ChangeLog:
* Make-lang.in (c.tags): Look at *.cc.
gcc/cp/ChangeLog:
* Make-lang.in (c++.tags): Just look at *.cc.
gcc/d/ChangeLog:
* Make-lang.in (d.tags): Just look at *.cc.
gcc/fortran/ChangeLog:
* Make-lang.in (fortran.tags): Look at *.cc.
gcc/go/ChangeLog:
* Make-lang.in (go.tags): Look at *.cc.
gcc/objc/ChangeLog:
* Make-lang.in (objc.tags): Look at *.cc.
gcc/objcp/ChangeLog:
* Make-lang.in (obj-c++.tags): Look at *.cc.
As of gdc-12, the lexer expects there 4 bytes of zero padding at the end
of the source buffer to mark the end of input. Sometimes when reading
from stdin, the data at the end of input is garbage rather than zeroes.
Fix that by explicitly calling memset past the end of the buffer.
PR d/105544
gcc/d/ChangeLog:
* d-lang.cc (d_parse_file): Zero padding past the end of the stdin
buffer so the D lexer has a sentinel to stop parsing at.
My r13-680-g0dc8e1e7026d9b commit to add support for Decimal Floating
Point introduced:
case SDmode:
case DDmode:
case TDmode:
which are rejected by gcc-4.8 as build compiler.
This patch replaces them with E_SDmode, E_DDmode and E_TD_mode.
Committed as obvious.
* config/aarch64/aarch64.cc (aarch64_gimplify_va_arg_expr):
Prefix mode names with E_.
In passing, use bool for its return type.
gcc/cp/ChangeLog:
* cp-tree.h (comp_template_args): Change return type to bool.
* pt.cc (comp_template_args): Document default arguments.
Change return type to bool and adjust returns accordingly.
OpenMP 5.2 adds support for firstprivate and allocate clauses on the scope
construct and this patch adds that support to GCC.
5.2 unfortunately (IMNSHO mistakenly) marked scope construct as worksharing,
which implies that it isn't possible to nest inside of it other scope,
worksharing loop, sections, explicit barriers, single etc. which would
make scope far less useful. I'm not implementing that part, keeping the
5.1 behavior here, and will file an issue to revert that for OpenMP 6.0.
But, for firstprivate it keeps the restriction that is now implied from
worksharing construct that listed var can't be private in outer context,
where for reduction 5.1 had similar restriction explicit even for scope
and 5.2 has it implicitly through worksharing construct.
2022-05-31 Jakub Jelinek <jakub@redhat.com>
gcc/
* omp-low.cc (build_outer_var_ref): For code == OMP_CLAUSE_ALLOCATE
allow var to be private in the outer context.
(lower_private_allocate): Pass OMP_CLAUSE_ALLOCATE as last argument
to build_outer_var_ref.
gcc/c/
* c-parser.cc (OMP_SCOPE_CLAUSE_MASK): Add firstprivate and allocate
clauses.
gcc/cp/
* parser.cc (OMP_SCOPE_CLAUSE_MASK): Add firstprivate and allocate
clauses.
gcc/testsuite/
* c-c++-common/gomp/scope-5.c: New test.
* c-c++-common/gomp/scope-6.c: New test.
* g++.dg/gomp/attrs-1.C (bar): Add firstprivate and allocate clauses
to scope construct.
* g++.dg/gomp/attrs-2.C (bar): Likewise.
libgomp/
* testsuite/libgomp.c-c++-common/allocate-1.c (foo): Add testcase for
scope construct with allocate clause.
* testsuite/libgomp.c-c++-common/allocate-3.c (foo): Likewise.
* testsuite/libgomp.c-c++-common/scope-2.c: New test.
As discussed here:
<https://gcc.gnu.org/pipermail/gcc-patches/2021-February/564629.html>,
type_dependent_expression_p should not be called with a type argument.
I promised I'd add an assert so here it is. One place needed adjusting.
PR c++/99080
gcc/cp/ChangeLog:
* pt.cc (type_dependent_expression_p): Assert !TYPE_P.
* semantics.cc (finish_id_expression_1): Handle UNBOUND_CLASS_TEMPLATE
specifically.
This patch is a form of insurance policy in case my patch for PR 7061 runs
into problems on non-x86 targets; the middle-end can add an extra check
that the backend is happy placing SCmode and DImode values in the same
register, before creating a SUBREG. Unfortunately, ix86_modes_tieable_p
currently claims this is not allowed(?), even though the default target
hook for modes_tieable_p is to always return true [i.e. false can be
used to specifically prohibit bad combinations], and the x86_64 ABI
passes SCmode values in DImode registers!. This makes the backend's
modes_tiable_p hook a little more forgiving, and additionally enables
interconversion between SCmode and V2SFmode, and between DCmode and
VD2Fmode, which opens interesting opporutunities in the future.
2022-05-30 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog
* config/i386/i386.cc (ix86_modes_tieable_p): Allow SCmode to be
tieable with DImode on TARGET_64BIT, and SCmode tieable with
V2SFmode, and DCmode with V2DFmode.
ALLOCATE: generate different STAT,ERRMSG results for failures from
allocation of already allocated objects or insufficient virtual memory.
gcc/fortran/ChangeLog:
PR fortran/91300
* libgfortran.h: Define new error code LIBERROR_NO_MEMORY.
* trans-stmt.cc (gfc_trans_allocate): Generate code for setting
ERRMSG depending on result of STAT result of ALLOCATE.
* trans.cc (gfc_allocate_using_malloc): Use STAT value of
LIBERROR_NO_MEMORY in case of failed malloc.
gcc/testsuite/ChangeLog:
PR fortran/91300
* gfortran.dg/allocate_alloc_opt_15.f90: New test.
This patch resolves PR rtl-optimization/101617 where we should generate
the exact same code for (X ? -1 : 1) as we do for ((X ? -1 : 0) | 1).
The cause of the current difference on x86_64 is actually in
ix86_expand_int_movcc that doesn't know that negl;sbbl can be used
to create a -1/0 result depending on whether the input is zero/nonzero.
So for Andrew Pinski's test case:
int f1(int i)
{
return i ? -1 : 1;
}
GCC currently generates:
f1: cmpl $1, %edi
sbbl %eax, %eax // x ? 0 : -1
andl $2, %eax // x ? 0 : 2
subl $1, %eax // x ? -1 : 1
ret
but with the attached patch, now generates:
f1: negl %edi
sbbl %eax, %eax // x ? -1 : 0
orl $1, %eax // x ? -1 : 1
ret
To implement this I needed to add two expanders to i386.md to generate
the required instructions (in both SImode and DImode) matching the
pre-existing define_insns of the same name.
2022-05-30 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog
PR rtl-optimization/101617
* config/i386/i386-expand.cc (ix86_expand_int_movcc): Add a
special case (indicated by negate_cc_compare_p) to generate a
-1/0 mask using neg;sbb.
* config/i386/i386.md (x86_neg<mode>_ccc): New define_expand
to generate an *x86_neg<mode>_ccc instruction.
(x86_mov<mode>cc_0_m1_neg): Likewise, a new define_expand to
generate a *x86_mov<mode>cc_0_m1_neg instruction.
gcc/testsuite/ChangeLog
PR rtl-optimization/101617
* gcc.target/i386/pr101617.c: New test case.
GCC's middle-end provides a default cost model for RTL expressions, for
backends that don't specify their own instruction timings, that can be
summarized as multiplications are COSTS_N_INSNS(4), divisions are
COSTS_N_INSNS(7) and all other operations are COSTS_N_INSNS(1).
This patch tweaks the above definition so that fused-multiply-add
(FMA) and high-part multiplications cost the same as regular multiplications,
or more importantly aren't (by default) considered less expensive. Likewise
the saturating forms of multiplication and division cost the same as the
regular variants. These values can always be changed by the target, but
the goal is to avoid RTL expansion substituting a suitable operation with
its saturating equivalent because it (accidentally) looks much cheaper.
For example, PR 89845 is about implementing division/modulus via highpart
multiply, which may accidentally look extremely cheap.
2022-05-30 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog
* rtlanal.cc (rtx_cost) <MULT>: Treat FMA, SS_MULT, US_MULT,
SMUL_HIGHPART and UMUL_HIGHPART as having the same cost as MULT.
<DIV>: Likewise, SS_DIV and US_DIV have the same default as DIV.
This patch resolves the last piece of PR target/70321 a code quality
(P2 regression) affecting mainline. Currently, for HJ's testcase:
void foo (long long ixi)
{
if (ixi != 14348907)
__builtin_abort ();
}
GCC with -m32 -O2 generates four instructions for the comparison:
movl 16(%esp), %eax
movl 20(%esp), %edx
xorl $14348907, %eax
orl %eax, %edx
but with this patch it now requires only three, making better use of
x86's addressing modes:
movl 16(%esp), %eax
xorl $14348907, %eax
orl 20(%esp), %eax
The solution is to expand "doubleword" equality/inequality expressions
using flag setting COMPARE instructions for the early RTL passes, and
then split them during split1, after STV and before reload.
Hence on x86_64, we now see/allow things like:
(insn 11 8 12 2 (set (reg:CCZ 17 flags)
(compare:CCZ (reg/v:TI 84 [ x ])
(reg:TI 96))) "cmpti.c":2:43 30 {*cmpti_doubleword}
This allows the STV pass to decide whether it's preferrable to perform
this comparison using vector operations, i.e. a pxor/ptest sequence,
or as scalar integer operations, i.e. a xor/xor/or sequence. Alas
this required tweaking of the STV pass to recognize the "new" form of
these comparisons and split out the pxor operation itself. To confirm
this still works as expected I've added a new STV test case:
long long a[1024];
long long b[1024];
int foo()
{
for (int i=0; i<1024; i++)
{
long long t = (a[i]<<8) | (b[i]<<24);
if (t == 0)
return 1;
}
return 0;
}
where with -m32 -O2 -msse4.1 the above comparison with zero should look
like:
punpcklqdq %xmm0, %xmm0
ptest %xmm0, %xmm0
Although this patch includes one or two minor tweaks to provide all the
necessary infrastructure to support conversion of TImode comparisons to
V1TImode (and SImode comparisons to V4SImode), STV doesn't yet implement
these transformations, but this is something that can be considered after
stage 4. Indeed the new convert_compare functionality is split out
into a method to simplify its potential reuse by the timode_scalar_chain
class.
2022-05-30 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog
PR target/70321
* config/i386/i386-expand.cc (ix86_expand_branch): Don't decompose
DI mode equality/inequality using XOR here. Instead generate a
COMPARE for doubleword modes (DImode on !TARGET_64BIT or TImode).
* config/i386/i386-features.cc (gen_gpr_to_xmm_move_src): Use
gen_rtx_SUBREG when NUNITS is 1, i.e. for TImode to V1TImode.
(general_scalar_chain::convert_compare): New function to convert
scalar equality/inequality comparison into vector operations.
(general_scalar_chain::convert_insn) [COMPARE]: Refactor. Call
new convert_compare helper method.
(convertible_comparion_p): Update to match doubleword COMPARE
of two register, memory or integer constant operands.
* config/i386/i386-features.h (general_scalar_chain::convert_compare):
Prototype/declare member function here.
* config/i386/i386.md (cstore<mode>4): Change mode to SDWIM, but
only allow new doubleword modes for EQ and NE operators.
(*cmp<dwi>_doubleword): New define_insn_and_split, to split a
doubleword comparison into a pair of XORs followed by an IOR to
set the (zero) flags register, optimizing the XORs if possible.
* config/i386/sse.md (V_AVX): Include V1TI and V2TI in mode
iterator; V_AVX is (currently) only used by ptest.
(sse4_1 mode attribute): Update to support V1TI and V2TI.
gcc/testsuite/ChangeLog
PR target/70321
* gcc.target/i386/pr70321.c: New test case.
* gcc.target/i386/sse4_1-stv-1.c: New test case.
const_int_operand and other const*_operand predicates do not need
constraints when the constraint is inherited from the range of
constant integer predicate. Remove the constraint in case all
alternatives use the same inherited constraint.
However, when there are operands, commitative with a non-constant
operand, the operand effectively matches e.g.
nonimmediate_operand|const_int_operand rather than just
const_int_operand. We should keep the constraint for
const_int_operand that are in a % pair. See PR 105624.
2022-05-30 Uroš Bizjak <ubizjak@gmail.com>
gcc/ChangeLog:
* config/i386/i386.md: Remove constraints when used with
const_int_operand, const0_operand, const_1_operand, constm1_operand,
const8_operand, const128_operand, const248_operand, const123_operand,
const2367_operand, const1248_operand, const359_operand,
const_4_or_8_to_11_operand, const48_operand, const_0_to_1_operand,
const_0_to_3_operand, const_0_to_4_operand, const_0_to_5_operand,
const_0_to_7_operand, const_0_to_15_operand, const_0_to_31_operand,
const_0_to_63_operand, const_0_to_127_operand, const_0_to_255_operand,
const_0_to_255_mul_8_operand, const_1_to_31_operand,
const_1_to_63_operand, const_2_to_3_operand, const_4_to_5_operand,
const_4_to_7_operand, const_6_to_7_operand, const_8_to_9_operand,
const_8_to_11_operand, const_8_to_15_operand, const_10_to_11_operand,
const_12_to_13_operand, const_12_to_15_operand, const_14_to_15_operand,
const_16_to_19_operand, const_16_to_31_operand, const_20_to_23_operand,
const_24_to_27_operand and const_28_to_31_operand.
* config/i386/mmx.md: Ditto.
* config/i386/sse.md: Ditto.
* config/i386/subst.md: Ditto.
* config/i386/sync.md: Ditto.
gcc/testsuite/ChangeLog:
* gcc.target/i386/pr105624.c: New test.
It's pretty clear that the operand numbers in the MEM_P() checks are
off by one, perhaps due to a copy-and-paste oversight (unlike in most
other places here we're dealing with two outputs).
gcc/
* config/i386/i386.md (bmi2_umul<mode><dwi>3_1): Correct MEM_P()
arguments.
Without this patch, gnat would use `-gnatw?` as the default option for
some of the default warnings.
gcc/ada/
* erroutc.adb (Get_Warning_Option): Don't consider `?` as a
valid option switch.
This enables better integration with tools that handle GNAT's output.
gcc/ada/
* erroutc.ads (Get_Warning_Option): New function returning the
option responsible for a warning if it exists.
* erroutc.adb (Get_Warning_Option): Likewise.
(Get_Warning_Tag): Rely on Get_Warning_Option when possible.
* errout.adb (Output_JSON_Message): Emit option field.
When implementing structural subprogram variants we ignored them in
expansion of the pragma itself, but not in expansion of a recursive
subprogram call. Now fixed.
gcc/ada/
* exp_ch6.adb (Check_Subprogram_Variant): Ignore structural
variants.
Code cleanup; semantics is unaffected.
gcc/ada/
* osint.adb (Locate_File): Change variable to constant and
initialize it by concatenation of directory, file name and NUL.
The GNAT behaviour regarding the Ada RM requirement to support Address
clauses for imported subprograms was documented twice: in section about
packed types (which was a mistake) and in section about address clauses
(where it belongs).
Cleanup related to the use of packed arrays for bitset operations to
detect uses of uninitialized scalars in GNAT.
gcc/ada/
* doc/gnat_rm/implementation_advice.rst (Packed Types): Remove
duplicated and wrongly placed paragraph.
* gnat_rm.texi: Regenerate.
This enables tools that ingest GNAT's output to properly classify these
messages.
gcc/ada/
* inline.adb (Check_Package_Body_For_Inlining): Add insertion
character.
This caused these warnings not to be tagged with the switch causing
them, which is an issue for tools ingesting GNAT's output.
gcc/ada/
* sem_elab.adb (Process_Conditional_ABE_Access_Taken): Add '.f'
insertion characters.
While cleaning up and modifying code for unreferenced warnings we
removed all calls to Defer_Reference, which was the only routine that
populated the Deferred_References table. Consequently, all the code
related to this table became dead.
gcc/ada/
* lib-xref.ads (Deferred_Reference_Entry, Defer_Reference,
Process_Deferred_References, Has_Deferred_Reference): Remove
client API.
* lib-xref.adb (Deferred_References, Defer_Reference,
Has_Deferred_Reference, Process_Deferred_References): Remove
implementation.
* frontend.adb, sem_ch11.adb, sem_ch5.adb, sem_res.adb,
sem_util.adb, sem_warn.adb: Remove uses of Deferred_References.
This patch fixes a bug whereby if a function body has local objects that
are finalizable, and has a return statement that requires generation of
transient finalizable temps, and there are dynamic-sized objects
requiring pushing/popping the (primary) stack at run time, and the
function body has exception handlers, then incorrect code is
generated. In particular, the transient objects are finalized after they
have been deallocated. This can cause seg faults, corrupted heap, and
the like.
Note that if there are no dynamic-sized objects, then the bug does
not occur, because the back end allocates objects of compile-time-known
size based on where they are referenced, rather than which local
block they are declared in.
This patch relies on the fact that an At_End handler and regular
exception handlers CAN coexist in the same handled statement sequence,
which was not true some years ago.
gcc/ada/
* exp_ch7.adb (Wrap_HSS_In_Block): Do not create a new block in
the case of function bodies. We include all subprogram bodies,
because it's harmless for procedures. We cannot easily avoid
creating this block in ALL cases, because some transformations
of (e.g.) task bodies end up moving some code such that the
wrong exception handlers apply to that code.
(Build_Finalizer_Call): Remove code for creating a new block.
This was unreachable code, given that Wrap_HSS_In_Block has
already done that, but with the above change to
Wrap_HSS_In_Block, this code becomes reachable, and triggers
essentially the same bug.
* exp_ch7.ads: Adjust comment.
This patch corrects an error in the compiler whereby spurious
unreferenced warnings are generated on formal parameters of null generic
subprograms.
These changes also fix the parsing of aspects on formal parameters such
that the aspects now get properly set for all formal parameters in a
parameter list.
gcc/ada/
* par-ch6.adb (P_Formal_Part): Set Aspect_Specifications on all
formals instead of just the last in a formal id list.
* sem_ch6.adb (Analyze_Null_Procedure): Mark expanded null
generic procedures as trivial in order to avoid spurious
unreferenced warnings.
Both the `System.Mmap` and `System.Object_Reader` packages are defining
entities named `Offset` and they are both `use`d at the top of
s-dwalin.adb.
Therefore, the references to `Offset` throughout this file are
ambiguous, and GNAT is supposed to complain. Since it does not for the
moment, we fix the ambiguity by declaring a subtype `Offset` at the top
of the file simply renames `System.Object_Reader.Offset`.
gcc/ada/
* libgnat/s-dwalin.adb: Add a subtype declaration to fix the
ambiguity.
The expression given in a Static_Predicate aspect specification is
required to be predicate-static. The implementation of this compile-time
check was incorrect in some cases. There were problems in both
directions: expressions that are not predicate-static were incorrectly
treated as though they were and vice versa. This led to both accepting
invalid code and to rejecting valid code.
gcc/ada/
* sem_ch13.adb (Is_Predicate_Static): Do not generate warnings
about subexpressions of enclosing expressions. Generate warnings
for predicates that are known to be always true or always false,
except in the case where the predicate is expressed as a Boolean
literal. Deal with non-predicate-static expressions that have
been transformed into predicate-static expressions. Add missing
Is_Type_Ref call to N_Membership_Test case.
The presence of the discriminants prevents the values associated with the
components of the parent type from being put into the sub-aggregate built
for the _Parent component.
gcc/ada/
* exp_aggr.adb (Expand_Record_Aggregate.Build_Back_End_Aggregate):
Skip the discriminants at the start of the component list before
looking for the components inherited from the parent in the case
of a tagged extension.
The left-overs of the old freezing code in Make_DT are now redundant since
the semantic analyzer performs the same task for the 'Access attribute.
gcc/ada/
* exp_disp.adb (Make_DT): Remove remaining freezing code.
There is no need to analyze expression functions that are primitives of a
tagged type for its dispatch table because they will be analyzed at the end
of the current scope.
gcc/ada/
* sem_attr.adb (Resolve_Attribute) <Attribute_Access>: Don't analyze
the body of an expression function in the case of a dispatch table.
Changes in GNATprove (translation to Why3 and provers) result in proofs
being much less automatic, and requiring ghost code to detail
intermediate steps. In some cases, it is better to use the new By
mechanism to prove assertions in steps, as this avoids polluting the
proof context for other proofs. For that reason, add units s-spark.ads
and s-spcuop.ads/b to the runtime sources.
gcc/ada/
* Makefile.rtl: Add new units.
* libgnat/s-aridou.adb (Scaled_Divide): Add ghost code for provers.
* libgnat/s-spcuop.adb: New unit for ghost cut operations.
* libgnat/s-spcuop.ads: New unit for ghost cut operations.
* libgnat/s-spark.ads: New unit.
Mention when security hardening features are available in other
languages.
Expand the strub documentation a little, for clarity and completeness.
Add missing 'aliased' and attribute to variables in strub example.
gcc/ada/
* doc/gnat_rm/security_hardening_features.rst: Mention
availability in other languages when applicable.
(Stack Scrubbing): Associate the attribute with types, expand
some comments, fix the example involving access to variables.
* gnat_rm.texi: Regenerate.
The previous change introduced a backward incompatibility in the handling
of a user-defined "=" operator for a class-wide type of a tagged type: it
would previously hide the predefined "=" operator of the tagged type in
the private case, but it no longer does in this case, while it still does
in the nonprivate case.
This hiding is a non-portability, but is fundamentally what the compiler
implements, instead of the RM rule which requires homographs. The reason
lies in the implementation of the "=" operator in GNAT: internally, there
is not a "=" predefined operator for every nonlimited type, but instead
there is a single, universal "=" predefined operator for all the nonlimited
types. The consequence is that the hiding rule implemented in GNAT for "="
is effectively that a user-declared symmetrical "=" operator returning
boolean hides the predefined "=" operator of any type that is covered by
the user-declared operator.
Whether it is desirable to implement the exact RM rule in GNAT is to be
discussed, but existing code relies on the non-portability and would thus
need to be changed.
gcc/ada/
* sem_ch6.adb (New_Overloaded_Entity): Deal specifically with the
overriding of the "=" operator for tagged types.
When static dispatch tables are built for library-level tagged types, the
primitives (the subprogram themselves) are frozen; that's necessary because
their address is taken. However, their profile, i.e. all the types present
therein, is also frozen, which is not necessary after AI05-019 and is also
inconsistent with the handling of attribute references.
The change also removes a couple of pragma Inline on subprograms that are
too large for inlining to bring any benefit.
gcc/ada/
* exp_ch3.adb (Expand_N_Object_Declaration): Adjust call to Make_DT.
* exp_disp.ads (Building_Static_DT): Remove pragma Inline.
(Building_Static_Secondary_DT): Likewise.
(Convert_Tag_To_Interface): Likewise.
(Make_DT): Remove second parameter.
* exp_disp.adb (Make_DT): Likewise.
(Check_Premature_Freezing): Delete.
Pass Do_Freeze_Profile as False in call to Freeze_Entity.
* freeze.ads (Freezing_Library_Level_Tagged_Type): Delete.
* freeze.adb (Freeze_Profile): Remove obsolete code.
(Freeze_Entity): Tweak comment.
After changes in GNATprove, adapt proof. Simply move an assertion up
before it is first needed here.
gcc/ada/
* libgnat/s-arit32.adb (Scaled_Divide32): Move assertion up.
The internal unit System.Generic_Array_Operations defines only generic
subprograms. Thus, pragma Assertion_Policy inside the spec has no
effect, as each instantiation is only subject to the assertion policy at
the program point of the instantiation. Remove this confusing pragma,
and add the pragma inside each generic body making use of additional
assertions or ghost code, so that running time of instantiations is not
impacted by assertions meant for formal verification.
gcc/ada/
PR ada/105303
* libgnat/s-gearop.adb: Add pragma Assertion_Policy in generic
bodies making use of additional assertions or ghost code.
* libgnat/s-gearop.ads: Remove confusing Assertion_Policy.
As a result of other recent changes, the Convert_To_Return_False flag
is never set. The flag can be therefore be deleted.
gcc/ada/
* exp_ch11.adb (Expand_N_Raise_Expression): Remove
Convert_To_Return_False test.
* gen_il-fields.ads: Remove Convert_To_Return_False field.
* gen_il-gen-gen_nodes.adb: Remove use of
Convert_To_Return_False field.
* sinfo.ads: Remove comment describing Convert_To_Return_False
flag.
This patch adds empty constructors to the functional containers so that
we can use them in expression functions.
gcc/ada/
* libgnat/a-cofuma.ads, libgnat/a-cofuma.adb,
libgnat/a-cofuse.ads, libgnat/a-cofuse.adb,
libgnat/a-cofuve.ads, libgnat/a-cofuve.adb: Add empty
constructors.
Removal of the previous kernel registration scheme unearthed mistakes in
the new one, which were:
- The new kernel registration code relied on the binder expansion phase,
which didn't happen because the registration code was already
generated by the binder.
- The kernel handle passed to CUDA_Register_Function was the first eight
bytes of the code of the host-side procedure representing the kernel
rather than its address.
gcc/ada/
* bindgen.adb (Gen_CUDA_Init): Remove code generating CUDA
definitions.
(Gen_CUDA_Defs): New function, generating definitions
initialized by Gen_CUDA_Init.
(Gen_Output_File_Ada): Call Gen_CUDA_Defs instead of
Gen_CUDA_Init.
(Gen_Adainit): Call Gen_CUDA_Init.
