This patch lowers assignments to vector subscripted designators into the
newly added hlfir.elemental_addr and hlfir.region_assign.
Note that the codegen of these operation to FIR is still TODO and will
still emit a TODO message when trying to compile programs end to end.
Differential Revision: https://reviews.llvm.org/D149962
Make __builtin_c_loc() into an intrinsic function and verify the
special semantic requirements on its actual arguments.
Differential Revision: https://reviews.llvm.org/D149988
Pointer components without default initialization pose some
difficult (or impossible) problems when they appear as right-hand
side targets in pointer assignment statements; they may contain
garbage or stale data that looks enough like a valid descriptor
to cause a crash. Solve the problem by avoiding it -- ensure
that pointers' descriptors are at least minimally established.
Differential Revision: https://reviews.llvm.org/D149979
Currently `CallOpInterface` has a method `getCallableForCallee` to have a consistent way to get the callee from an operation with `CallOpInterface`, but missing a consistent way to set a callee for an operation with `CallOpInterface`.
A set callee method is useful for transformations that operate on `CallOpInterface`, and change the callee, e.g., a pass that specialize function, which clone the callee, and change the `CallOpInterface`'s callee to the cloned version. Without such method, transformation would need to understand the implementation for every operations with `CallOpInterface`, and have a type switch to handle them.
This review adds a method to set callee for operation with `CallOpInterface`.
Reviewed By: gysit, zero9178o
Differential Revision: https://reviews.llvm.org/D149763
Currently complex division is lowered to a fir.divc operation and the
fir.divc is later converted to a sequence of llvm operations to perform
complex division, however this causes issues for extreme values when
the calculations overflow.
This patch changes the lowering of complex division to use the Intrinsic
Call functionality to lower into library calls (for single, double,
extended and quad precisions) or an MLIR complex dialect division operation
(for half and bfloat precisions).
Note 1: If the Complex To Standard conversion of division operation
matures then we can use it for all precisions. Currently it has the
same issues as the conversion of fir.divc.
Note 2: A previous patch (D145808) did the same but during conversion of
the fir.divc operation. But using function calls at that stage leads to
ABI issues since the conversion to LLVM is not aware of the complex target
rewrite.
Note 3: If the patch is accepted, fir.divc can be removed from FIR.
Reviewed By: vzakhari, PeteSteinfeld, DavidTruby
Differential Revision: https://reviews.llvm.org/D149546
I plan to implement lowering from parse tree to HLFIR first for forall
and where to ease testing of the rewrite pass while writing it.
To avoid cryptic errors in ConvertToFir pass about unhandled operations,
this patch already defines the pass that will further lower these
operations and make it throw clear TODO messages.
Differential Revision: https://reviews.llvm.org/D149852
This is the last piece required to lower Forall (except pointer
assignments, where an operation may be needed to deal with bounds
remapping).
Lowering requires symbols to be mapped to memory SSA values produced
by a fir_FortranVariableOpInterface operation. This applies to
forall index-values, that are symbols.
fir.alloca/fir.store/hlfir.declare are not allowed inside the body of
an hlfir.forall that only accept operations with the
hlfir_OrderedAssignmentTreeOpInterface so that the forall structure is
well defined and easy to transform.
Allowing such operations in the forall body would open the doors to
generating ill-formed programs where such operation would be used for
non index-values.
Instead, add an hlfir.forall_index with both required interface to
produce a memory address for a forall index.
As a bonus, since forall index-value are by nature read-only, the
loads of hlfir.forall_index can be canonicalized, which will help
simplifying the hlfir.forall nested code (it is unclear we will be
able to tell MLIR enough about hlfir.forall and hlfir.where structure
so that it could safely do a generic mem-to-reg inside it, and getting
rid of read-effect operations will benefit the forall rewrite pass).
Differential Revision: https://reviews.llvm.org/D149836
Adds a HLFIR operation for the PRODUCT intrinsic according to
the design set out in flang/doc/HighLevelFIR.md
Since the PRODUCT intrinsic is essentially identical to SUM
in terms of its arguments and result characteristics in the
Fortran Standard, the operation definition and subsequent
tests also take the same form.
Differential Revision: https://reviews.llvm.org/D147624
Add hlfir.forall_mask, hlfir.where, and hlfir.elsewhere operations that
are operations that holds (optionally for hlfir.elsewhere) the
evaluation of a logical mask that controls the evaluation of nested
operations.
They allow representing Fortran forall control mask, as well as where
and eslewhere statements/constructs.
They use the OrderedAssignmentTreeOpInterface since they can all be used
inside Forall and their masks should be fully evaluated for all the
index-value set induced by parent Forall before any of the nested
operations in their body is evaluated.
I initially tried making them into a single operation with some attributes
to make a difference, but I felt this made the verifier/parser/printer and
usages messier/tricky compared to making three distinct operations that
represent the three Fortran feature in a vanilla way.
Differential Revision: https://reviews.llvm.org/D149754
This patch adds the hlfir.forall operation and the
OrderAssignmentTreeOpInterface that allows representing Fortran forall.
It uses regions to keep Fortran expression evaluation independent from
each other in the IR. Forall assignments inside hlfir.forall are
represented with hlfir.region_assign which also keeps the IR generated
for each expressions independently.
The goal of this representation is to provide a representation that is
straightforward to generate from Fortran parse tree without any analysis, while
providing enough structure information so that an optimization pass can decide
how to schedule, and save if needed, the evaluations of the Forall and Where
expression and statements. It allows the data dependency analysis to be done at
the HLFIR level.
The OrderAssignmentTreeOpInterface allows ensuring that the Forall/Where
tree structure is kept in the IR. It will allow visiting this tree in
the IR without hard coding the operation structures in the pass.
Differential Revision: https://reviews.llvm.org/D149734
Lower vector subscripted designators as values when they appear outside
of the assignment left-hand side and input IO contexts.
This matches Fortran semantics where vector subscripted designators cannot
be written to outside of the two contexts mentioned above: they are
passed/taken by value where they appear.
This patch uses the added hlfir.element_addr to lower vector designators
in lowering. But when reaching the end of the designator lowering, the
hlfir.element_addr is turned into an hlfir.elemental when lowering is
not asking for the hlfir.elemental_addr.
This approach allows lowering vector subscripted in the same way in
while visiting the designator, and only adapt to the context at the
edge.
The part where lowering uses the hlfir.elemental_addr will be
done in further patch as it requires lowering assignments in the
new hlfir.region_assign op, and there is not codegen yet for these
new operations.
Differential Revision: https://reviews.llvm.org/D149480
hlfir.region_assign is a Region based version of hlfir.assign: the
right-hand side and left-hand-side are evaluated in their own region,
and an optional region can be added to implement user defined
assignment.
This will be used for:
- assignments inside where and forall
- user defined assignments
- assignments to vector subscripted entities.
Rational:
Forall and Where lowering requires solving an expression/assignment
evaluation scheduling problem based on data dependencies between the
variables being assigned and the one used in the expressions.
Keeping left-hand side and right-hand side in their own region will
make it really easy to analyse the dependency and move around the
expression evaluation as a whole. Operation DAGs are hard to scissor out
when the LHS and RHS evaluation are lowered in the same block. The pass
dealing with further forall/where lowering in HLFIR will need to
succeed. It is not acceptable for them to fail splitting the RHS/LHS
evaluation code. Keeping them in independent block is an approach that
cannot fail.
For user defined assignments, having a region allows implementing all
the call details in lowering, and even to allow inlining of the user
assignment, before it is decided if a temporary for the LHS or RHS is
required or not.
The operation description mention "hlfir.elemental_addr" (operation that
will be used for vector subscripted LHS) and "ordered assignment trees"
(concept/inetrface that will be used to represent forall/where structure
in HLFIR). These will be pushed in follow-up patch, but I do not want t
scissor out the descriptions.
Differential Revision: https://reviews.llvm.org/D149442
We have to keep it as a box, since taking box_addr of the optional
box may be invalid.
Reviewed By: jeanPerier
Differential Revision: https://reviews.llvm.org/D149505
Initial support for USE_DEVICE_PTR clause on OMP TARGET DATA directive.
Reviewed By: kiranchandramohan
Differential Revision: https://reviews.llvm.org/D148254
This resolves issues with running out of stack on examples like
https://fortran-lang.discourse.group/t/modern-fortran-sample-code/2019/18
reported by @clementval.
When target rewrite creates alloca(s) around a call, we need to insert
stacksave/stackrestore to free the allocated stack. Better performant
code may be achieved by placing the alloca(s) outside of loops,
but the placement has to behave correctly with regards to OpenMP/OpenACC/etc.
dialect operations that have special representation for "private"
objects. This is a concervative fix for correctness issue.
Differential Revision: https://reviews.llvm.org/D149222
Inlining as a hlfir.elemental will allow the transpose to be inlined
into subsequent operations in some cases. For example,
y = TRANSPOSE(x)
z = y * 2
Will operate in a single loop without creating a temporary for the
TRANSPOSE (unlike the runtime call, which always allocates).
This is in a new SimplifyHLFIRIntriniscs pass. The intention is that some
day that pass might replace the FIR SimplifyIntrinsics pass.
Depends On: D149060
Reviewed By: jeanPerier, vzakhari
Differential Revision: https://reviews.llvm.org/D149067
This patch ports the fopenmp-host-ir-file-path flag from Clang to Flang-new, this flag
is added by the driver to the device pass when doing two phase compilation (device + host).
This flag is then applied to the module when compiling during the OpenMP device phase.
This file can then be utilised during lowering of the OpenMP dialect to LLVM-IR, which
allows the device and host to maintain 1:1 mapping of OpenMP metadata for variables
during lowering via the OpenMPIRBuilders loadOffloadInfoMetadata facilities
(which is used for declare target and I believe target regions as well).
Reviewer: awarzynski
Differential Revision: https://reviews.llvm.org/D148038
Implement semantics for the IGNORE_TKR directive as it is interpreted
by the PGI / NVFORTRAN compiler.
Differential Revision: https://reviews.llvm.org/D148643
The OpenACC dialect data clauses require for variables appearing in data
operations to implement the `PointerLikeType` interface (similarly to
OpenMP dialect). Thus, this interface needs attached to FIR types.
Two of the FIR types, HeapType and LLVMPointerType, did not provide
getElementType - now they do by calling getEleTy. This cleans up the
interface attachment for both OpenACC and OpenMP dialects.
Reviewed By: clementval, jeanPerier
Differential Revision: https://reviews.llvm.org/D148560
The array component indices in the "path" of a fir.slice are zero based
because FIR does not know about the component lower bounds.
When lowering hlfir.designate to FIR for `array%x(i, j)`, convert `i` and
`j` to zero based indices before generating the fir.slice.
Differential Revision: https://reviews.llvm.org/D148627
Introduce conditional code to identify stride of "one element", and simplify the array accesses for that case.
This allows better loop performance in various benchmarks.
Reviewed By: tblah, kiranchandramohan
Differential Revision: https://reviews.llvm.org/D141306
This operation fetches an extent value from a fir.shape. The operation
could just as easily live in the fir namespace, but is only needed for
hlfir lowering so I put it here.
This operation is required to allow one to defer getting the extents of a shape
generated by hlfir.get_shape until after that shape has been resolved
(after bufferization of the hlfir.expr).
This operation will be lowered to FIR as an arith.constant created using
the definition of the fir.shape argument.
Depends on: D146830
Differential Revision: https://reviews.llvm.org/D148220
This is an operation which returns the fir.shape for a hlfir.expr.
A hlfir.expr can be defined by:
- A transformational intrinsic (e.g. hlfir.matmul)
- hlfir.as_expr
- hlfir.elemental
hlfir.elemental is easy because there is a compulsory shape operand.
hlfir.as_expr is defined as operating on a variable (defined using a
hlfir.declare). hlfir.declare has an optional shape argument. The
transformational intrinsics do not have an associated shape.
If all extents are known at compile time, the extents for the shape can
be fetched from the hlfir.expr's type. For example, the result of a
hlfir.matmul with arguments who's extents are known at compile time will
have constant extents which can be queried from the type. In this case
the hlfir.shape_of will be canonicalised to a fir.shape operation using
those extents.
If not all extents are known at compile time, shapes have to be read
from boxes after bufferization. In the case of the transformational
intrinsics, the shape read from the result box can be queried from the
hlfir.declare operation for the buffer allocated to that hlfir.expr (via
the hlfir.as_expr).
Differential Revision: https://reviews.llvm.org/D146830
TYPE(*) arguments fell through in IS_CONTIGUOUS folding
because they are not Expr<SomeType>. Expose entry point for
symbols in IsContiguous and use that.
The added test revealed that IS_CONTIGUOUS was folded to
false for assumed rank arguments. Fix this: the contiguity of
assumed rank without the CONTIGUOUS argument can only be
verified at runtime.
Differential Revision: https://reviews.llvm.org/D148128
/Users/jiefu/llvm-project/flang/runtime/non-tbp-dio.h:41:1: error: 'NonTbpDefinedIoTable' defined as a struct here but previously declared as a class; this is valid, but may result in linker errors under the Microsoft C++ ABI [-Werror,-Wmismatched-tags]
struct NonTbpDefinedIoTable {
^
/Users/jiefu/llvm-project/flang/include/flang/Runtime/io-api.h:26:1: note: did you mean struct here?
class NonTbpDefinedIoTable;
^~~~~
struct
1 error generated.
A fairly recent introduction of runtime I/O APIs called OutputDerivedType()
and InputDerivedType() didn't cover NAMELIST I/O's need to access
non-type-bound generic interfaces for user-defined derived type I/O
when those generic interfaces are defined in some scope other than the
one that defines the derived type.
The patch adds a new data structure shared between lowering
and the runtime that can represent all of the cases that can
arise with non-type-bound defined I/O. It can represent
scopes in which non-type-bound defined I/O generic interfaces
are inaccessible, too, due to IMPORT statements.
The data structure is now an operand to OutputDerivedType() and
InputDerivedType() as well as a data member in the NamelistGroup
structure.
Differential Revision: https://reviews.llvm.org/D148257
After the extraction of the TypeConverter, move the header files
to the include dir so the shared library build is fine.
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D147979
This patch ports OpenMP RTL flags from the shared Clang compiler
options to Flang. As well as adding a limited subset to bbc.
This patch enables the flags below (and any equals or inverse variants)
for Flang that exist in Clang:
-fopenmp-target-debug
-fopenmp-assume-threads-oversubscription
-fopenmp-assume-teams-oversubscription
-fopenmp-assume-no-nested-parallelism
-fopenmp-assume-no-thread-state
For the bbc tool it only utilises the primary variants to minimize
additional complexity in the tool.
The patch also provides FlagAttr generation from these flags. Which
will be lowered to LLVM-IR in a subsequent patch.
Reviewers: kiranchandramohan, awarzynski
Differential Revision: https://reviews.llvm.org/D147324
Implements the PowerPC mtfsf and mtfsfi intrinsics as well as introduces semantic error checking code for PowerPC intrinsics
Reviewed By: klausler
Differential Revision: https://reviews.llvm.org/D144876
Many semantic checks for constraints related to PURE subprograms
can be implemented in terms of Semantics' "definable.h" utilities,
slightly expanded. Replace some particular PURE constraint
checks with calls to WhyNotDefinable(), except for cases that
had better specific error messages, and start checking some
missing constraints with DEALLOCATE statements and local
variable declarations.
Differential Revision: https://reviews.llvm.org/D147389
Scope of changes:
1) Add attribute to OpenMP MLIR dialect which stores target cpu and
target features
2) Store target information in MLIR module
Differential Revision: https://reviews.llvm.org/D146612
Reviewed By: kiranchandramohan
Co-authored-by: Kiran Chandramohan <kiran.chandramohan@arm.com>
Currently, it is OK to have alloca/store/and reboxed to
fir.box<!fir.array<?xnone>> and fir.class<none>, but not simple
fir.box<none>.
This restriction is a legacy from a time where it was thought TYPE(*)
descriptor size would not be statically known, but the way polymorphism
was implemented actually allows knowing its size: a scalar descriptor
with an addendum (in case it is a derived type).
Note that this assumes fir.box<none> are always scalars. There are currently
a few cast from ranked descriptor to !fir.box<None> around runtime calls.
These are simple casts before runtime call, so there are no load/stores
to the resulting fir.box<None> and it is OK.
When assumed rank are supported, some legacy usage of fir.box<none> as the "any"
descriptor in the runtime interface will be replaced to avoid any issues there.
This change will be required to fix an undefined behavior with NULL() that
requires allocation of a fir.box<None>.
Differential Revision: https://reviews.llvm.org/D147237
ASSOCIATED intrinsic TARGET handling is weird for OPTIONAL, because as
opposed to other intrinsic arguments, OPTIONAL allocatable and pointers
may be absent when passed to it, and a diassociated pointer TARGET is not
the same as when TARGET is not provided. Hence, it needs custom
handling in lowering.
The handling was done late (in genIntrinsicCall, without the semantic
context), and assumed it would be possible to retrieve the optionality
aspects, but this is brittle, and hard to share with HLFIR.
Move it in CustomIntrinsicCall that is intended to deal with these
corner case.
Also avoid using fir.box<None> as the related fir.if result, and used
the correct fir.box/fir.class type for the target: using a fir.box<None>
here is risky since fir.box<None> are now meant for scalar TYPE(*), and
the TARGET may be ranked.
Move the introduction of the fir.box<None> around the runtime (when
assumed rank are supported, these will become !fir.box<!fir.array<..xNone>>).
Differential Revision: https://reviews.llvm.org/D147224
Block merging is disabled with FIR: some FIR value should not be promoted
to block arguments, and the region simplification block merging is
promoting all SSA value types to block argument when two blocks are
similar except for the usage these values.
Differential Revision: https://reviews.llvm.org/D147130
While a pass exists to generate basic debug information, currently there is not a corresponding flag to enable it.
This patch adds support for activating this pass at any debug level >= -g1, as well as emiting a warning for higher levels that the functionality is not yet fully implemented.
This patch also adds -g and -gline-tables-only to appear when `flang-new` --help is run
Depends on D142347.
Reviewed By: awarzynski
Differential Revision: https://reviews.llvm.org/D146814
This commit adds the OffloadModuleInterface to the OpenMP dialect,
which will implement future module attribute get/set's for offloading.
Currently it implements set and get's for the omp.is_device attribute,
which is promoted to a real attribute in this commit as well (primarily
to allow switch cases to work nicely with it for future work and to keep
consistency with future module attributes).
This interface is attached to mlir::ModuleOp's on registration of the
OpenMPDialect and should be accessible anywhere the OpenMP
dialect is registered and initialized.
Reviewers: kiranchandramohan, awarzynski
Differential Revision: https://reviews.llvm.org/D146850
Constraint checking for explicit SAVE attributes is more
accurate when done along with other declaration checking, rather
than on the fly during name resolution. This allows us to
catch attempts to attach explicit SAVE attributes to anything
that can't have one (constraints C859, C860).
Also delete `IsSave()`, whose few remaining uses were changed to the
more general `IsSaved()` predicate that seems more correct for
those uses, returning true for both explicit and implied SAVE
attributes.
Differential Revision: https://reviews.llvm.org/D146579
The present I/O infrastructure for user-defined derived type I/O
subroutines works fine for type-bound I/O generic bindings. It also works
for explicit INTERFACE blocks and GENERIC statements that define
UDDIO subroutines in the same scope as the definition of the derived type,
so long as the specific procedures in those bindings are module procedures
or external procedures.
For non-type-bound UDDTIO specific procedures that are dummy procedures,
thunks of inner procedures, or procedure pointers, or that are defined with
interfaces or GENERIC outside the scope of the definition of the derived
type, a new runtime I/O API is needed so that lowering can generate
a call that supplies the appropriate procedure as well as the defined
type instance.
This patch specifies and implements this new runtime API and provides
utility routines for lowering to use to determine whether it should be
called for any particular OutputItem or InputItem in the parse tree.
Differential Revision: https://reviews.llvm.org/D146571
