IR: Split Metadata from Value

Split `Metadata` away from the `Value` class hierarchy, as part of
PR21532.  Assembly and bitcode changes are in the wings, but this is the
bulk of the change for the IR C++ API.

I have a follow-up patch prepared for `clang`.  If this breaks other
sub-projects, I apologize in advance :(.  Help me compile it on Darwin
I'll try to fix it.  FWIW, the errors should be easy to fix, so it may
be simpler to just fix it yourself.

This breaks the build for all metadata-related code that's out-of-tree.
Rest assured the transition is mechanical and the compiler should catch
almost all of the problems.

Here's a quick guide for updating your code:

  - `Metadata` is the root of a class hierarchy with three main classes:
    `MDNode`, `MDString`, and `ValueAsMetadata`.  It is distinct from
    the `Value` class hierarchy.  It is typeless -- i.e., instances do
    *not* have a `Type`.

  - `MDNode`'s operands are all `Metadata *` (instead of `Value *`).

  - `TrackingVH<MDNode>` and `WeakVH` referring to metadata can be
    replaced with `TrackingMDNodeRef` and `TrackingMDRef`, respectively.

    If you're referring solely to resolved `MDNode`s -- post graph
    construction -- just use `MDNode*`.

  - `MDNode` (and the rest of `Metadata`) have only limited support for
    `replaceAllUsesWith()`.

    As long as an `MDNode` is pointing at a forward declaration -- the
    result of `MDNode::getTemporary()` -- it maintains a side map of its
    uses and can RAUW itself.  Once the forward declarations are fully
    resolved RAUW support is dropped on the ground.  This means that
    uniquing collisions on changing operands cause nodes to become
    "distinct".  (This already happened fairly commonly, whenever an
    operand went to null.)

    If you're constructing complex (non self-reference) `MDNode` cycles,
    you need to call `MDNode::resolveCycles()` on each node (or on a
    top-level node that somehow references all of the nodes).  Also,
    don't do that.  Metadata cycles (and the RAUW machinery needed to
    construct them) are expensive.

  - An `MDNode` can only refer to a `Constant` through a bridge called
    `ConstantAsMetadata` (one of the subclasses of `ValueAsMetadata`).

    As a side effect, accessing an operand of an `MDNode` that is known
    to be, e.g., `ConstantInt`, takes three steps: first, cast from
    `Metadata` to `ConstantAsMetadata`; second, extract the `Constant`;
    third, cast down to `ConstantInt`.

    The eventual goal is to introduce `MDInt`/`MDFloat`/etc. and have
    metadata schema owners transition away from using `Constant`s when
    the type isn't important (and they don't care about referring to
    `GlobalValue`s).

    In the meantime, I've added transitional API to the `mdconst`
    namespace that matches semantics with the old code, in order to
    avoid adding the error-prone three-step equivalent to every call
    site.  If your old code was:

        MDNode *N = foo();
        bar(isa             <ConstantInt>(N->getOperand(0)));
        baz(cast            <ConstantInt>(N->getOperand(1)));
        bak(cast_or_null    <ConstantInt>(N->getOperand(2)));
        bat(dyn_cast        <ConstantInt>(N->getOperand(3)));
        bay(dyn_cast_or_null<ConstantInt>(N->getOperand(4)));

    you can trivially match its semantics with:

        MDNode *N = foo();
        bar(mdconst::hasa               <ConstantInt>(N->getOperand(0)));
        baz(mdconst::extract            <ConstantInt>(N->getOperand(1)));
        bak(mdconst::extract_or_null    <ConstantInt>(N->getOperand(2)));
        bat(mdconst::dyn_extract        <ConstantInt>(N->getOperand(3)));
        bay(mdconst::dyn_extract_or_null<ConstantInt>(N->getOperand(4)));

    and when you transition your metadata schema to `MDInt`:

        MDNode *N = foo();
        bar(isa             <MDInt>(N->getOperand(0)));
        baz(cast            <MDInt>(N->getOperand(1)));
        bak(cast_or_null    <MDInt>(N->getOperand(2)));
        bat(dyn_cast        <MDInt>(N->getOperand(3)));
        bay(dyn_cast_or_null<MDInt>(N->getOperand(4)));

  - A `CallInst` -- specifically, intrinsic instructions -- can refer to
    metadata through a bridge called `MetadataAsValue`.  This is a
    subclass of `Value` where `getType()->isMetadataTy()`.

    `MetadataAsValue` is the *only* class that can legally refer to a
    `LocalAsMetadata`, which is a bridged form of non-`Constant` values
    like `Argument` and `Instruction`.  It can also refer to any other
    `Metadata` subclass.

(I'll break all your testcases in a follow-up commit, when I propagate
this change to assembly.)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@223802 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Duncan P. N. Exon Smith
2014-12-09 18:38:53 +00:00
parent db7b69e3a6
commit dad20b2ae2
88 changed files with 3370 additions and 1970 deletions

View File

@@ -172,29 +172,29 @@ struct FunctionTypeKeyInfo {
/// the operands.
struct GenericMDNodeInfo {
struct KeyTy {
ArrayRef<Value *> Ops;
ArrayRef<Metadata *> RawOps;
ArrayRef<MDOperand> Ops;
unsigned Hash;
KeyTy(ArrayRef<Value *> Ops)
: Ops(Ops), Hash(hash_combine_range(Ops.begin(), Ops.end())) {}
KeyTy(ArrayRef<Metadata *> Ops)
: RawOps(Ops), Hash(hash_combine_range(Ops.begin(), Ops.end())) {}
KeyTy(GenericMDNode *N, SmallVectorImpl<Value *> &Storage) {
Storage.resize(N->getNumOperands());
for (unsigned I = 0, E = N->getNumOperands(); I != E; ++I)
Storage[I] = N->getOperand(I);
Ops = Storage;
Hash = hash_combine_range(Ops.begin(), Ops.end());
}
KeyTy(GenericMDNode *N)
: Ops(N->op_begin(), N->op_end()), Hash(N->getHash()) {}
bool operator==(const GenericMDNode *RHS) const {
if (RHS == getEmptyKey() || RHS == getTombstoneKey())
return false;
if (Hash != RHS->getHash() || Ops.size() != RHS->getNumOperands())
if (Hash != RHS->getHash())
return false;
for (unsigned I = 0, E = Ops.size(); I != E; ++I)
if (Ops[I] != RHS->getOperand(I))
return false;
return true;
assert((RawOps.empty() || Ops.empty()) && "Two sets of operands?");
return RawOps.empty() ? compareOps(Ops, RHS) : compareOps(RawOps, RHS);
}
template <class T>
static bool compareOps(ArrayRef<T> Ops, const GenericMDNode *RHS) {
if (Ops.size() != RHS->getNumOperands())
return false;
return std::equal(Ops.begin(), Ops.end(), RHS->op_begin());
}
};
static inline GenericMDNode *getEmptyKey() {
@@ -215,29 +215,6 @@ struct GenericMDNodeInfo {
}
};
/// DebugRecVH - This is a CallbackVH used to keep the Scope -> index maps
/// up to date as MDNodes mutate. This class is implemented in DebugLoc.cpp.
class DebugRecVH : public CallbackVH {
/// Ctx - This is the LLVM Context being referenced.
LLVMContextImpl *Ctx;
/// Idx - The index into either ScopeRecordIdx or ScopeInlinedAtRecords that
/// this reference lives in. If this is zero, then it represents a
/// non-canonical entry that has no DenseMap value. This can happen due to
/// RAUW.
int Idx;
public:
DebugRecVH(MDNode *n, LLVMContextImpl *ctx, int idx)
: CallbackVH(n), Ctx(ctx), Idx(idx) {}
MDNode *get() const {
return cast_or_null<MDNode>(getValPtr());
}
void deleted() override;
void allUsesReplacedWith(Value *VNew) override;
};
class LLVMContextImpl {
public:
/// OwnedModules - The set of modules instantiated in this context, and which
@@ -265,6 +242,8 @@ public:
FoldingSet<AttributeSetNode> AttrsSetNodes;
StringMap<MDString> MDStringCache;
DenseMap<Value *, ValueAsMetadata *> ValuesAsMetadata;
DenseMap<Metadata *, MetadataAsValue *> MetadataAsValues;
DenseSet<GenericMDNode *, GenericMDNodeInfo> MDNodeSet;
@@ -301,7 +280,8 @@ public:
ConstantInt *TheFalseVal;
LeakDetectorImpl<Value> LLVMObjects;
LeakDetectorImpl<Metadata> LLVMMDObjects;
// Basic type instances.
Type VoidTy, LabelTy, HalfTy, FloatTy, DoubleTy, MetadataTy;
Type X86_FP80Ty, FP128Ty, PPC_FP128Ty, X86_MMXTy;
@@ -335,32 +315,14 @@ public:
/// CustomMDKindNames - Map to hold the metadata string to ID mapping.
StringMap<unsigned> CustomMDKindNames;
typedef std::pair<unsigned, TrackingVH<MDNode> > MDPairTy;
typedef std::pair<unsigned, TrackingMDNodeRef> MDPairTy;
typedef SmallVector<MDPairTy, 2> MDMapTy;
/// MetadataStore - Collection of per-instruction metadata used in this
/// context.
DenseMap<const Instruction *, MDMapTy> MetadataStore;
/// ScopeRecordIdx - This is the index in ScopeRecords for an MDNode scope
/// entry with no "inlined at" element.
DenseMap<MDNode*, int> ScopeRecordIdx;
/// ScopeRecords - These are the actual mdnodes (in a value handle) for an
/// index. The ValueHandle ensures that ScopeRecordIdx stays up to date if
/// the MDNode is RAUW'd.
std::vector<DebugRecVH> ScopeRecords;
/// ScopeInlinedAtIdx - This is the index in ScopeInlinedAtRecords for an
/// scope/inlined-at pair.
DenseMap<std::pair<MDNode*, MDNode*>, int> ScopeInlinedAtIdx;
/// ScopeInlinedAtRecords - These are the actual mdnodes (in value handles)
/// for an index. The ValueHandle ensures that ScopeINlinedAtIdx stays up
/// to date.
std::vector<std::pair<DebugRecVH, DebugRecVH> > ScopeInlinedAtRecords;
/// DiscriminatorTable - This table maps file:line locations to an
/// integer representing the next DWARF path discriminator to assign to
/// instructions in different blocks at the same location.