1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-22 20:43:44 +02:00
llvm-mirror/include/llvm/IR/LLVMContext.h
Michael Kruse a9ec994652 Introduce llvm.loop.parallel_accesses and llvm.access.group metadata.
The current llvm.mem.parallel_loop_access metadata has a problem in that
it uses LoopIDs. LoopID unfortunately is not loop identifier. It is
neither unique (there's even a regression test assigning the some LoopID
to multiple loops; can otherwise happen if passes such as LoopVersioning
make copies of entire loops) nor persistent (every time a property is
removed/added from a LoopID's MDNode, it will also receive a new LoopID;
this happens e.g. when calling Loop::setLoopAlreadyUnrolled()).
Since most loop transformation passes change the loop attributes (even
if it just to mark that a loop should not be processed again as
llvm.loop.isvectorized does, for the versioned and unversioned loop),
the parallel access information is lost for any subsequent pass.

This patch unlinks LoopIDs and parallel accesses.
llvm.mem.parallel_loop_access metadata on instruction is replaced by
llvm.access.group metadata. llvm.access.group points to a distinct
MDNode with no operands (avoiding the problem to ever need to add/remove
operands), called "access group". Alternatively, it can point to a list
of access groups. The LoopID then has an attribute
llvm.loop.parallel_accesses with all the access groups that are parallel
(no dependencies carries by this loop).

This intentionally avoid any kind of "ID". Loops that are clones/have
their attributes modifies retain the llvm.loop.parallel_accesses
attribute. Access instructions that a cloned point to the same access
group. It is not necessary for each access to have it's own "ID" MDNode,
but those memory access instructions with the same behavior can be
grouped together.

The behavior of llvm.mem.parallel_loop_access is not changed by this
patch, but should be considered deprecated.

Differential Revision: https://reviews.llvm.org/D52116

llvm-svn: 349725
2018-12-20 04:58:07 +00:00

358 lines
15 KiB
C++

//===- llvm/LLVMContext.h - Class for managing "global" state ---*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file declares LLVMContext, a container of "global" state in LLVM, such
// as the global type and constant uniquing tables.
//
//===----------------------------------------------------------------------===//
#ifndef LLVM_IR_LLVMCONTEXT_H
#define LLVM_IR_LLVMCONTEXT_H
#include "llvm-c/Types.h"
#include "llvm/IR/DiagnosticHandler.h"
#include "llvm/Support/CBindingWrapping.h"
#include "llvm/Support/Options.h"
#include <cstdint>
#include <memory>
#include <string>
namespace llvm {
class DiagnosticInfo;
enum DiagnosticSeverity : char;
class Function;
class Instruction;
class LLVMContextImpl;
class Module;
class OptPassGate;
template <typename T> class SmallVectorImpl;
class SMDiagnostic;
class StringRef;
class Twine;
namespace yaml {
class Output;
} // end namespace yaml
namespace SyncScope {
typedef uint8_t ID;
/// Known synchronization scope IDs, which always have the same value. All
/// synchronization scope IDs that LLVM has special knowledge of are listed
/// here. Additionally, this scheme allows LLVM to efficiently check for
/// specific synchronization scope ID without comparing strings.
enum {
/// Synchronized with respect to signal handlers executing in the same thread.
SingleThread = 0,
/// Synchronized with respect to all concurrently executing threads.
System = 1
};
} // end namespace SyncScope
/// This is an important class for using LLVM in a threaded context. It
/// (opaquely) owns and manages the core "global" data of LLVM's core
/// infrastructure, including the type and constant uniquing tables.
/// LLVMContext itself provides no locking guarantees, so you should be careful
/// to have one context per thread.
class LLVMContext {
public:
LLVMContextImpl *const pImpl;
LLVMContext();
LLVMContext(LLVMContext &) = delete;
LLVMContext &operator=(const LLVMContext &) = delete;
~LLVMContext();
// Pinned metadata names, which always have the same value. This is a
// compile-time performance optimization, not a correctness optimization.
enum : unsigned {
MD_dbg = 0, // "dbg"
MD_tbaa = 1, // "tbaa"
MD_prof = 2, // "prof"
MD_fpmath = 3, // "fpmath"
MD_range = 4, // "range"
MD_tbaa_struct = 5, // "tbaa.struct"
MD_invariant_load = 6, // "invariant.load"
MD_alias_scope = 7, // "alias.scope"
MD_noalias = 8, // "noalias",
MD_nontemporal = 9, // "nontemporal"
MD_mem_parallel_loop_access = 10, // "llvm.mem.parallel_loop_access"
MD_nonnull = 11, // "nonnull"
MD_dereferenceable = 12, // "dereferenceable"
MD_dereferenceable_or_null = 13, // "dereferenceable_or_null"
MD_make_implicit = 14, // "make.implicit"
MD_unpredictable = 15, // "unpredictable"
MD_invariant_group = 16, // "invariant.group"
MD_align = 17, // "align"
MD_loop = 18, // "llvm.loop"
MD_type = 19, // "type"
MD_section_prefix = 20, // "section_prefix"
MD_absolute_symbol = 21, // "absolute_symbol"
MD_associated = 22, // "associated"
MD_callees = 23, // "callees"
MD_irr_loop = 24, // "irr_loop"
MD_access_group = 25, // "llvm.access.group"
};
/// Known operand bundle tag IDs, which always have the same value. All
/// operand bundle tags that LLVM has special knowledge of are listed here.
/// Additionally, this scheme allows LLVM to efficiently check for specific
/// operand bundle tags without comparing strings.
enum : unsigned {
OB_deopt = 0, // "deopt"
OB_funclet = 1, // "funclet"
OB_gc_transition = 2, // "gc-transition"
};
/// getMDKindID - Return a unique non-zero ID for the specified metadata kind.
/// This ID is uniqued across modules in the current LLVMContext.
unsigned getMDKindID(StringRef Name) const;
/// getMDKindNames - Populate client supplied SmallVector with the name for
/// custom metadata IDs registered in this LLVMContext.
void getMDKindNames(SmallVectorImpl<StringRef> &Result) const;
/// getOperandBundleTags - Populate client supplied SmallVector with the
/// bundle tags registered in this LLVMContext. The bundle tags are ordered
/// by increasing bundle IDs.
/// \see LLVMContext::getOperandBundleTagID
void getOperandBundleTags(SmallVectorImpl<StringRef> &Result) const;
/// getOperandBundleTagID - Maps a bundle tag to an integer ID. Every bundle
/// tag registered with an LLVMContext has an unique ID.
uint32_t getOperandBundleTagID(StringRef Tag) const;
/// getOrInsertSyncScopeID - Maps synchronization scope name to
/// synchronization scope ID. Every synchronization scope registered with
/// LLVMContext has unique ID except pre-defined ones.
SyncScope::ID getOrInsertSyncScopeID(StringRef SSN);
/// getSyncScopeNames - Populates client supplied SmallVector with
/// synchronization scope names registered with LLVMContext. Synchronization
/// scope names are ordered by increasing synchronization scope IDs.
void getSyncScopeNames(SmallVectorImpl<StringRef> &SSNs) const;
/// Define the GC for a function
void setGC(const Function &Fn, std::string GCName);
/// Return the GC for a function
const std::string &getGC(const Function &Fn);
/// Remove the GC for a function
void deleteGC(const Function &Fn);
/// Return true if the Context runtime configuration is set to discard all
/// value names. When true, only GlobalValue names will be available in the
/// IR.
bool shouldDiscardValueNames() const;
/// Set the Context runtime configuration to discard all value name (but
/// GlobalValue). Clients can use this flag to save memory and runtime,
/// especially in release mode.
void setDiscardValueNames(bool Discard);
/// Whether there is a string map for uniquing debug info
/// identifiers across the context. Off by default.
bool isODRUniquingDebugTypes() const;
void enableDebugTypeODRUniquing();
void disableDebugTypeODRUniquing();
using InlineAsmDiagHandlerTy = void (*)(const SMDiagnostic&, void *Context,
unsigned LocCookie);
/// Defines the type of a yield callback.
/// \see LLVMContext::setYieldCallback.
using YieldCallbackTy = void (*)(LLVMContext *Context, void *OpaqueHandle);
/// setInlineAsmDiagnosticHandler - This method sets a handler that is invoked
/// when problems with inline asm are detected by the backend. The first
/// argument is a function pointer and the second is a context pointer that
/// gets passed into the DiagHandler.
///
/// LLVMContext doesn't take ownership or interpret either of these
/// pointers.
void setInlineAsmDiagnosticHandler(InlineAsmDiagHandlerTy DiagHandler,
void *DiagContext = nullptr);
/// getInlineAsmDiagnosticHandler - Return the diagnostic handler set by
/// setInlineAsmDiagnosticHandler.
InlineAsmDiagHandlerTy getInlineAsmDiagnosticHandler() const;
/// getInlineAsmDiagnosticContext - Return the diagnostic context set by
/// setInlineAsmDiagnosticHandler.
void *getInlineAsmDiagnosticContext() const;
/// setDiagnosticHandlerCallBack - This method sets a handler call back
/// that is invoked when the backend needs to report anything to the user.
/// The first argument is a function pointer and the second is a context pointer
/// that gets passed into the DiagHandler. The third argument should be set to
/// true if the handler only expects enabled diagnostics.
///
/// LLVMContext doesn't take ownership or interpret either of these
/// pointers.
void setDiagnosticHandlerCallBack(
DiagnosticHandler::DiagnosticHandlerTy DiagHandler,
void *DiagContext = nullptr, bool RespectFilters = false);
/// setDiagnosticHandler - This method sets unique_ptr to object of DiagnosticHandler
/// to provide custom diagnostic handling. The first argument is unique_ptr of object
/// of type DiagnosticHandler or a derived of that. The third argument should be
/// set to true if the handler only expects enabled diagnostics.
///
/// Ownership of this pointer is moved to LLVMContextImpl.
void setDiagnosticHandler(std::unique_ptr<DiagnosticHandler> &&DH,
bool RespectFilters = false);
/// getDiagnosticHandlerCallBack - Return the diagnostic handler call back set by
/// setDiagnosticHandlerCallBack.
DiagnosticHandler::DiagnosticHandlerTy getDiagnosticHandlerCallBack() const;
/// getDiagnosticContext - Return the diagnostic context set by
/// setDiagnosticContext.
void *getDiagnosticContext() const;
/// getDiagHandlerPtr - Returns const raw pointer of DiagnosticHandler set by
/// setDiagnosticHandler.
const DiagnosticHandler *getDiagHandlerPtr() const;
/// getDiagnosticHandler - transfers owenership of DiagnosticHandler unique_ptr
/// to caller.
std::unique_ptr<DiagnosticHandler> getDiagnosticHandler();
/// Return if a code hotness metric should be included in optimization
/// diagnostics.
bool getDiagnosticsHotnessRequested() const;
/// Set if a code hotness metric should be included in optimization
/// diagnostics.
void setDiagnosticsHotnessRequested(bool Requested);
/// Return the minimum hotness value a diagnostic would need in order
/// to be included in optimization diagnostics. If there is no minimum, this
/// returns None.
uint64_t getDiagnosticsHotnessThreshold() const;
/// Set the minimum hotness value a diagnostic needs in order to be
/// included in optimization diagnostics.
void setDiagnosticsHotnessThreshold(uint64_t Threshold);
/// Return the YAML file used by the backend to save optimization
/// diagnostics. If null, diagnostics are not saved in a file but only
/// emitted via the diagnostic handler.
yaml::Output *getDiagnosticsOutputFile();
/// Set the diagnostics output file used for optimization diagnostics.
///
/// By default or if invoked with null, diagnostics are not saved in a file
/// but only emitted via the diagnostic handler. Even if an output file is
/// set, the handler is invoked for each diagnostic message.
void setDiagnosticsOutputFile(std::unique_ptr<yaml::Output> F);
/// Get the prefix that should be printed in front of a diagnostic of
/// the given \p Severity
static const char *getDiagnosticMessagePrefix(DiagnosticSeverity Severity);
/// Report a message to the currently installed diagnostic handler.
///
/// This function returns, in particular in the case of error reporting
/// (DI.Severity == \a DS_Error), so the caller should leave the compilation
/// process in a self-consistent state, even though the generated code
/// need not be correct.
///
/// The diagnostic message will be implicitly prefixed with a severity keyword
/// according to \p DI.getSeverity(), i.e., "error: " for \a DS_Error,
/// "warning: " for \a DS_Warning, and "note: " for \a DS_Note.
void diagnose(const DiagnosticInfo &DI);
/// Registers a yield callback with the given context.
///
/// The yield callback function may be called by LLVM to transfer control back
/// to the client that invoked the LLVM compilation. This can be used to yield
/// control of the thread, or perform periodic work needed by the client.
/// There is no guaranteed frequency at which callbacks must occur; in fact,
/// the client is not guaranteed to ever receive this callback. It is at the
/// sole discretion of LLVM to do so and only if it can guarantee that
/// suspending the thread won't block any forward progress in other LLVM
/// contexts in the same process.
///
/// At a suspend point, the state of the current LLVM context is intentionally
/// undefined. No assumptions about it can or should be made. Only LLVM
/// context API calls that explicitly state that they can be used during a
/// yield callback are allowed to be used. Any other API calls into the
/// context are not supported until the yield callback function returns
/// control to LLVM. Other LLVM contexts are unaffected by this restriction.
void setYieldCallback(YieldCallbackTy Callback, void *OpaqueHandle);
/// Calls the yield callback (if applicable).
///
/// This transfers control of the current thread back to the client, which may
/// suspend the current thread. Only call this method when LLVM doesn't hold
/// any global mutex or cannot block the execution in another LLVM context.
void yield();
/// emitError - Emit an error message to the currently installed error handler
/// with optional location information. This function returns, so code should
/// be prepared to drop the erroneous construct on the floor and "not crash".
/// The generated code need not be correct. The error message will be
/// implicitly prefixed with "error: " and should not end with a ".".
void emitError(unsigned LocCookie, const Twine &ErrorStr);
void emitError(const Instruction *I, const Twine &ErrorStr);
void emitError(const Twine &ErrorStr);
/// Query for a debug option's value.
///
/// This function returns typed data populated from command line parsing.
template <typename ValT, typename Base, ValT(Base::*Mem)>
ValT getOption() const {
return OptionRegistry::instance().template get<ValT, Base, Mem>();
}
/// Access the object which can disable optional passes and individual
/// optimizations at compile time.
OptPassGate &getOptPassGate() const;
/// Set the object which can disable optional passes and individual
/// optimizations at compile time.
///
/// The lifetime of the object must be guaranteed to extend as long as the
/// LLVMContext is used by compilation.
void setOptPassGate(OptPassGate&);
private:
// Module needs access to the add/removeModule methods.
friend class Module;
/// addModule - Register a module as being instantiated in this context. If
/// the context is deleted, the module will be deleted as well.
void addModule(Module*);
/// removeModule - Unregister a module from this context.
void removeModule(Module*);
};
// Create wrappers for C Binding types (see CBindingWrapping.h).
DEFINE_SIMPLE_CONVERSION_FUNCTIONS(LLVMContext, LLVMContextRef)
/* Specialized opaque context conversions.
*/
inline LLVMContext **unwrap(LLVMContextRef* Tys) {
return reinterpret_cast<LLVMContext**>(Tys);
}
inline LLVMContextRef *wrap(const LLVMContext **Tys) {
return reinterpret_cast<LLVMContextRef*>(const_cast<LLVMContext**>(Tys));
}
} // end namespace llvm
#endif // LLVM_IR_LLVMCONTEXT_H