Currently lowering sets the extents of assumed-size array to "undef"
which was OK as long as the value was not expected to be read.
But when interfacing with the runtime and when passing assumed-size to
assumed-rank, this last extent may be read and must be -1 as specified
in the BIND(C) case in 18.5.3 point 5.
Set this value to -1, and update all the lowering code that was looking
for an undef defining op to identify assumed-size: much safer to
propagate and use semantic info here, the previous check actually did
not work if the array was used in an internal procedure (defining op not
visible anymore).
@clementval and @agozillon, I left assumed-size extent to zero in the
acc/omp bounds op as it was, please double check that is what you want
(I can imagine -1 may create troubles here, and 0 makes some sense as it
would lead to no data transfer).
This also allows removing special cases in UBOUND/LBOUND lowering.
Also disable allocation of cray pointee. This was never intended and
would now lead to crashes with the -1 value for assumed-size cray
pointee.
acc.loop was redesigned in https://reviews.llvm.org/D159229. This patch
updates the lowering to match the new op.
DO CONCURRENT construct will be added in a follow up patch.
Note that the pre-commit ci will fail until D159229 is merged.
Depends on #67355
This brings `createBodyOfOp` to its final intended form. First, input
privatization is performed, then the recursive lowering takes place, and
finally the output privatization (lastprivate) is done.
This enables fixing a known issue with infinite loops inside of an
OpenMP region, and the fix is included in this patch.
Fixes https://github.com/llvm/llvm-project/issues/74348.
Recursive lowering [5/5]
---------
Co-authored-by: Kiran Chandramohan <kiran.chandramohan@arm.com>
For simplicity, lowering relies on semantics expansion of parent
components in designators.
This was not done in `call x%p()` where `p` is a procedure component
pointer of a parent component of `x`.
Do it and turn lowering TODO into a new lowering TODO for `call bar(x%type_bound_procedure)` (passing a tybe bound procedure is allowed as an extension, but lowering does not handle this extension yet. This is a lowering issue, will do in different patch).
This is a slightly more slimmed down and up-to-date version of the older
PR from here: https://reviews.llvm.org/D144203, written by @jsjodin,
which has already under gone some review.
This PR places allocas in the alloca address space specified by the
provided data layout (default is 0 for all address spaces, unless
explicitly specified by the layout), and then will cast these alloca's
to the program address space if this address space is different from the
allocation address space. For most architectures data layouts, this will
be a no-op, as they have a flat address space. But in the case of AMDGPU
it will result in allocas being placed in the correct address space (5,
private), and then casted into the correct program address space (0,
generic). This results in correct (partially, a follow up PR will be
forthcoming soon) generation of allocations inside of device code.
This PR is in addition to the work by @skatrak in this PR:
https://github.com/llvm/llvm-project/pull/69599 and adds seperate and
neccesary functionality of casting alloca's from their address space to
the program address space, both are independent PRs, although there is
some minor overlap e.g. this PR incorporates some of the useful helper
functions from 69599, so whichever lands first will need a minor rebase.
Co-author: jsjodin
In
CALL FOO
PRINT *, ABS(FOO)
we currently resolve the first FOO to a global external subprogram, but
then the second FOO is treated as an implicitly typed local variable.
This happens because the name FOO is not present in the local scope.
Fix by adding FOO to the local scope using a place-holding
HostAssocDetails symbol whose existence prevents the creation of another
FOO in the local scope. The symbol stored in the parser::Name parse tree
nodes or used in typed expressions will all continue to point to the
global external subprogram.
Resolves llvm-test-suite/Fortran/gfortran/regression/pr71859.f90.
Recognize Cray pointees as such when they are declared as assumed size
arrays, and don't emit a bogus error message about implied shape arrays.
Fixes https://github.com/llvm/llvm-project/issues/77330.
The checking of calls to the intrinsic subroutine MOVE_ALLOC is not
insisting that its first two arguments be whole allocatable variables or
components. Fix, move the code into check-calls.cpp (a better home for
such things), and clean up the tests.
Fixes https://github.com/llvm/llvm-project/issues/77230.
The Fortran standard defines real MOD and MODULO with expressions like
MOD(a,p) = a - AINT(a/p)*p. Unfortunately, these definitions have poor
accuracy when a is much larger in magnitude than p, and every Fortran
compiler uses better algorithms instead.
Fixes llvm-test-suite/Fortran/gfortran/regression/mod_large_1.f90.
There are some very odd (even for Fortran) rules in F'2023 subclause
19.4 (paras 6 & 8) pertaining to the index variables of FORALL and DO
CONCURRENT constructs/statements, and they are not currently implemented
correctly.
Although these index variables are construct entities, they have
restrictions in the standard that would essentially allow them to also
be variables in their enclosing scopes. If their names are present in
the enclosing scope, and the construct does not have an explicit type
specification for its indices, then the names in the enclosing scope
must either be scalar variables or COMMON blocks, and their type must be
integer.
Reimplement these restrictions largely with portability warnings rather
than hard errors. Retain the semantic interpretation that the type of an
untyped index variable be taken from the type of a variable of the same
name in the enclosing scope, if it exists, although that bit of the
standard could be interpreted otherwise.
Fixes https://github.com/llvm/llvm-project/issues/76978.
Support \uNNNN and \uNNNNNNNN escape sequences for CHARACTER(KIND=2) and
CHARACTER(KIND=4) literal constants for better GNU Fortran
compatibility.
Fixes llvm-test-suite/Fortran/gfortran/regression/achar_6.F90 and
.../widechar_1.f90.
Introduce `genNestedEvaluations` that will lower all evaluations nested
in the given, accouting for a potential COLLAPSE directive.
Recursive lowering [2/5]
When a CharBlock starts with an expanded macro but does not end in this
macro expansion, GetProvenanceRange fails to return a ProvenanceRange
which may cause error message to be emitted without location or lowering
to emit code without source location (which is problematic if this code
contains calls to procedures defined in the same file since LLVM will
later crash with the error:
"inlinable function call in a function with a DISubprogram location must
have a debug location"
Fix this situation by returning the ProvenanceRange starting at the
replaced macro reference.
reference to gfortran fdate
https://gcc.gnu.org/onlinedocs/gfortran/FDATE.html
usage:
```fortran
CHARACTER(32) :: time
CALL fdate(time)
WRITE(*,*) time
```
fdate is used in the ECP proxy application
https://proxyapps.exascaleproject.org/app/minismac2d/f904467142/ref/smac2d.f (L1570)
`fdate` now produce the same result on flang, compare to gfortran, where
If the length is too short to fit completely, blank return.
```fortran
character(20) :: string
call fdate(string)
write(*, *) string, "X"
```
```bash
$ ../build-release/bin/flang-new test.f90
$ ./a.out
X
```
If length if larger than it requires(24), fill the rest of buffer space.
```fortran
character(30) :: string
call fdate(string)
write(*, *) string, "X"
```
```bash
$ ../build-release/bin/flang-new test.f90
$ ./a.out
Wed Nov 15 16:59:13 2023 X
```
The length value is hardcoded, because:
```c++
// Day Mon dd hh:mm:ss yyyy\n\0 is 26 characters, e.g.
// Tue May 26 21:51:03 2015\n\0
```
---------
Co-authored-by: Yi Wu <yiwu02@wdev-yiwu02.arm.com>
This patch add support of intrinsics Fortran 2008 EXECUTE_COMMAND_LINE.
The patch contains both the lowering and the runtime code and works on
both Windows and Linux. The patch contains a list of commits, to convey
the authorship and the history of changes. Some implementation specifics
or status has been added to `flang/docs/Intrinsics.md`.
I have provided a summary of the usage and the options required for the
`EXECUTE_COMMAND_LINE intrinsic`. The intrinsic supports both a
synchronous
(by default) and an asynchronous option.
| System | Mode | Implemention |
|---------|-------|---------------------------|
| Linux | Sync | std::system() |
| Windows | Sync | std::system() |
| Linux | Async | fork() |
| Windows | Async | CreateProcess |
Support for the SYSTEM GNU extension will be added in a separate PR.
Co-authored with @jeffhammond
---------
Signed-off-by: Jeff Hammond <jeff.science@gmail.com>
Co-authored-by: Jeff Hammond <jeff.science@gmail.com>
Co-authored-by: Yi Wu <yiwu02@wdev-yiwu02.arm.com>
If -nogpulib option is passed by the user, then the OpenMP device
runtime is not used and we should not emit globals to configure
debugging at compile-time for the device runtime.
Link to -nogpulib flag implementation for Clang:
https://reviews.llvm.org/D125314
fir::isPolymorphic was returning false for TYPE(*) assumed-size arrays
causing bad fir.rebox to be created when passing a polymorphic actual
argument to such TYPE(*) dummy.
Fix fir::isAssumedSize to return true for fir.ref<fir.array<none>> and
fir.ref<none>.
@cabreraam, I found this bug when testing your patch, although it is not
caused by it, so you may hit it when passing TYPE(*) deferred shape of
to assumed size TYPE(*) with a different rank.
Add `notify-type` to `iso_fortran_env` module. Add `notify-wait-stmt` to
the parser and add checks for constraints on the statement, `C1177` and
`C1178`, from the Fortran 2023 standard. Add three semantics tests for
`notify-wait-stmt`.
The I/O runtime's API allows -1 to be passed for a unit number in a
READ, WRITE, or PRINT statement, where it gets replaced by 5 or 6 as
appropriate. This turns out to have been a bad idea, as it prevents the
I/O runtime from detecting and reporting a program's invalid attempt to
use -1 as an I/O unit number. So just pass 5 or 6 as appropriate.
The compiler requires that a Cray pointee have a SEQUENCE type, but a
recent bug report points out that a BIND(C) type should also be
accepted.
Fixes https://github.com/llvm/llvm-project/issues/76529.
Shift the data from lower to higher order bits when memcpy the value in
the namelist in the big endian environment
---------
Co-authored-by: Mark Danial <mark.danial@ibm.com>
Co-authored-by: Kelvin Li <kli@ca.ibm.com>
A zero 'd' digit count in EX output editing has a meaning that's
distinct from other numeric output editing descriptors, and I missed
this in my initial implementation of the feature. d==0 means that the
runtime should emit hexadecimal digits after the (hexa)decimal point
until all of the rest of them would be zero.
C_FUNLOC was not handling procedure pointer argument correctly, the
issue lied in `hlfir::convertToBox` that did not handle procedure
pointers.
I modified the interface of `hlfir::convertToXXX` to take values on the
way because hlfir::Entity are fundamentally an mlir::Value with type
guarantees, so they should be dealt with by value as mlir::Value are
(they are very small).
This is a lot less complex than the data case where the shape has to be
accounted for, so the implementation is done inline.
One corner case will not be supported correctly for now: the case where
POINTER and TARGET points to the same internal procedure may return
false because lowering is creating fir.embox_proc each time the address
of an internal procedure is taken, so different thunk for the same
internal procedure/host link may be created and compare to false. This
will be fixed in a later patch that moves creating of internal procedure
fir.embox_proc in the host so that the addresses are the same when the
host link is the same. This change is required to properly support the
required lifetime of internal procedure addresses anyway (should be the
always be the lifetime of the host, even when the address is taken in an
internal procedure).
This patch enables more numeric (mod, sum, matmul, etc.) APIs,
and some others.
I added new macros to disable warnings about using C++ STD methods
like operators of std::complex, which do not have __device__ attribute.
This may probably result in unresolved references, if the header files
implementation relies on libstdc++. I will need to follow up on this.
Lower procedure pointer components, except in the context of structure
constructor (left TODO).
Procedure pointer components lowering share most of the lowering logic
of procedure poionters with the following particularities:
- They are components, so an hlfir.designate must be generated to
retrieve the procedure pointer address from its derived type base.
- They may have a PASS argument. While there is no dispatching as with
type bound procedure, special care must be taken to retrieve the derived
type component base in this case since semantics placed it in the
argument list and not in the evaluate::ProcedureDesignator.
These components also bring a new level of recursive MLIR types since a
fir.type may now contain a component with an MLIR function type where
one of the argument is the fir.type itself. This required moving the
"derived type in construction" stackto the converter so that the object
and function type lowering utilities share the same state (currently the
function type utilty would end-up creating a new stack when lowering its
arguments, leading to infinite loops). The BoxedProcedurePass also
needed an update to deal with this recursive aspect.
The function `instantiateVariable` in Bridge.cpp has the following code:
```
if (var.getSymbol().test(
Fortran::semantics::Symbol::Flag::OmpThreadprivate))
Fortran::lower::genThreadprivateOp(*this, var);
if (var.getSymbol().test(
Fortran::semantics::Symbol::Flag::OmpDeclareTarget))
Fortran::lower::genDeclareTargetIntGlobal(*this, var);
```
Implement `handleOpenMPSymbolProperties` in OpenMP.cpp, move the above
code there, and have `instantiateVariable` call this function instead.
This would further separate OpenMP-related details into OpenMP.cpp.
There was some discussion on discourse[1] about allowing call to FIR
generation functions from other part of lowering belonging to OpenMP.
This solution exposes a simple `genEval` member function on the
`AbstractConverter` so that IR generation for PFT Evaluation objects can
be called from lowering outside of the FirConverter but not exposing it.
[1] https://discourse.llvm.org/t/openmp-lowering-from-pft-to-fir/75263
This patch replaces uses of StringRef::{starts,ends}with with
StringRef::{starts,ends}_with for consistency with
std::{string,string_view}::{starts,ends}_with in C++20.
I'm planning to deprecate and eventually remove
StringRef::{starts,ends}with.
The adds a hlfir minloc intrinsic, similar to the minval intrinsic
already added, to help in the lowering of minloc. The idea is to later
add maxloc too, and from there add a simplification for producing minloc
with inlined elemental and hopefully less temporaries.
In the context of C/Fortran interoperability (BIND(C)), it is possible
to give the VALUE attribute to a BIND(C) derived type dummy, which
according to Fortran 2018 18.3.6 - 2. (4) implies that it must be passed
like the equivalent C structure value. The way C structure value are
passed is ABI dependent.
LLVM does not implement the C struct ABI passing for LLVM aggregate type
arguments. It is up to the front-end, like clang is doing, to split the
struct into registers or pass the struct on the stack (llvm "byval") as
required by the target ABI.
So the logic for C struct passing sits in clang. Using it from flang
requires setting up a lot of clang context and to bridge FIR/MLIR
representation to clang AST representation for function signatures (in
both directions). It is a non trivial task.
See
https://stackoverflow.com/questions/39438033/passing-structs-by-value-in-llvm-ir/75002581#75002581.
Since BIND(C) struct are rather limited as opposed to generic C struct
(e.g. no bit fields). It is easier to provide a limited implementation
of it for the case that matter to Fortran.
This patch:
- Updates the generic target rewrite pass to keep track of both the new
argument type and attributes. The motivation for this is to be able to
tell if a previously marshalled argument is passed in memory (it is a C
pointer), or if it is being passed on the stack (has the byval llvm
attributes).
- Adds an entry point in the target specific codegen to marshal struct
arguments, and use it in the generic target rewrite pass.
- Implements limited support for the X86-64 case. So far, the support
allows telling if a struct must be passed in register or on the stack,
and to deal with the stack case. The register case is left TODO in this
patch.
The X86-64 ABI implemented is the System V ABI for AMD64 version 1.0
Lowering doesn't use it, and won't need it in the future; it emits
in-line code for conversions and for assignments to scalars, and uses
the general Assign() routine for arrays.
Defined unformatted I/O is not allowed except from/to an external unit.
This restriction prohibits an INQUIRE(IOLENGTH=n) statement from using
derived types with defined unformatted output in its I/O list. The
runtime currently detects this case and crashes with an internal error;
this patch defines a new I/O error enum and causes the program to crash
with a more useful message.
The function `genCommonBlockMember` is not specific to OpenMP, and it
could very well be a common utility. Move it to ConvertVariable.cpp
where it logically belongs.
Preliminary patch to change lowering/code generation to use
llvm::DataLayout information instead of generating "sizeof" GEP (see
https://github.com/llvm/llvm-project/issues/71507).
Fortran Semantic analysis needs to know about the target type size and
alignment to deal with common blocks, and intrinsics like
C_SIZEOF/TRANSFER. This information should be obtained from the
llvm::DataLayout so that it is consistent during the whole compilation
flow.
This change is changing flang-new and bbc drivers to:
1. Create the llvm::TargetMachine so that the data layout of the target
can be obtained before semantics.
2. Sharing bbc/flang-new set-up of the
SemanticConstext.targetCharateristics from the llvm::TargetMachine. For
now, the actual part that set-up the Fortran type size and alignment
from the llvm::DataLayout is left TODO so that this change is mostly an
NFC impacting the drivers.
3. Let the lowering bridge set-up the mlir::Module datalayout attributes
since it is doing it for the target attribute, and that allows the llvm
data layout information to be available during lowering.
For flang-new, the changes are code shuffling: the `llvm::TargetMachine`
instance is moved to `CompilerInvocation` class so that it can be used
to set-up the semantic contexts. `setMLIRDataLayout` is moved to
`flang/Optimizer/Support/DataLayout.h` (it will need to be used from
codegen pass for fir-opt target independent testing.)), and the code
setting-up semantics targetCharacteristics is moved to
`Tools/TargetSetup.h` so that it can be shared with bbc.
As a consequence, LLVM targets must be registered when running
semantics, and it is not possible to run semantics for a target that is
not registered with the -triple option (hence the power pc specific
modules can only be built if the PowerPC target is available.
`llvm.fcmp` does support fast math attributes therefore so should
`arith.cmpf`.
The heavy churn in flang tests are because flang sets
`fastmath<contract>` by default on all operations that support the fast
math interface. Downstream users of MLIR should not be so effected.
This was requested in https://github.com/llvm/llvm-project/issues/74263
