Merge branch 'topic/vern/script-opt-headers-factoring'

* topic/vern/script-opt-headers-factoring:
  factored CPP source's main header into collection of per-source-file headers
  renamed script optimization Attrs.h header to prepare for factoring large Compile.h
  factored ZAM source's main header into collection of per-source-file headers
This commit is contained in:
Christian Kreibich 2024-10-18 17:51:04 -07:00
commit 2e576b058d
26 changed files with 1372 additions and 1368 deletions

10
CHANGES
View file

@ -1,3 +1,13 @@
7.1.0-dev.408 | 2024-10-18 17:51:04 -0700
* factored CPP source's main header into collection of per-source-file headers (Vern Paxson, Corelight)
* renamed script optimization Attrs.h header to prepare for factoring large Compile.h (Vern Paxson, Corelight)
* factored ZAM source's main header into collection of per-source-file headers (Vern Paxson, Corelight)
* Update doc submodule [nomail] [skip ci] (zeek-bot)
7.1.0-dev.403 | 2024-10-18 09:55:20 -0700 7.1.0-dev.403 | 2024-10-18 09:55:20 -0700
* Fix installation of Python packages in generate docs CI job again (Benjamin Bannier, Corelight) * Fix installation of Python packages in generate docs CI job again (Benjamin Bannier, Corelight)

View file

@ -1 +1 @@
7.1.0-dev.403 7.1.0-dev.408

View file

@ -0,0 +1,17 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Definitions associated with type attributes.
#pragma once
namespace zeek::detail {
enum AttrExprType {
AE_NONE, // attribute doesn't have an expression
AE_CONST, // easy expression - a constant (ConstExpr)
AE_NAME, // easy - a global (NameExpr)
AE_RECORD, // an empty record cast to a given type
AE_CALL, // everything else - requires a lambda, essentially
};
} // namespace zeek::detail

View file

@ -1,17 +1,50 @@
// See the file "COPYING" in the main distribution directory for copyright. // See the file "COPYING" in the main distribution directory for copyright.
// Definitions associated with type attributes. // Methods for tracking attributes associated with Zeek variables/types.
// Attributes arise mainly in the context of constructing types.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
#pragma once public:
// Tracks a use of the given set of attributes, including
// initialization dependencies and the generation of any
// associated expressions.
//
// Returns the initialization info associated with the set of
// attributes.
std::shared_ptr<CPP_InitInfo> RegisterAttributes(const AttributesPtr& attrs);
namespace zeek::detail { // Convenient access to the global offset associated with
// a set of Attributes.
int AttributesOffset(const AttributesPtr& attrs) { return GI_Offset(RegisterAttributes(attrs)); }
enum AttrExprType { // The same, for a single attribute.
AE_NONE, // attribute doesn't have an expression std::shared_ptr<CPP_InitInfo> RegisterAttr(const AttrPtr& attr);
AE_CONST, // easy expression - a constant (ConstExpr)
AE_NAME, // easy - a global (NameExpr)
AE_RECORD, // an empty record cast to a given type
AE_CALL, // everything else - requires a lambda, essentially
};
} // namespace zeek::detail // Returns a mapping of from Attr objects to their associated
// initialization information. The Attr must have previously
// been registered.
auto& ProcessedAttr() const { return processed_attr; }
private:
// Start of methods related to managing script type attributes.
// Attributes arise mainly in the context of constructing types.
// See Attrs.cc for definitions.
//
// Populates the 2nd and 3rd arguments with C++ representations
// of the tags and (optional) values/expressions associated with
// the set of attributes.
void BuildAttrs(const AttributesPtr& attrs, std::string& attr_tags, std::string& attr_vals);
// Returns a string representation of the name associated with
// different attribute tags (e.g., "ATTR_DEFAULT").
static const char* AttrName(AttrTag t);
// Similar for attributes, so we can reconstruct record types.
CPPTracker<Attributes> attributes = {"attrs", false};
// Maps Attributes and Attr's to their global initialization
// information.
std::unordered_map<const Attributes*, std::shared_ptr<CPP_InitInfo>> processed_attrs;
std::unordered_map<const Attr*, std::shared_ptr<CPP_InitInfo>> processed_attr;

File diff suppressed because it is too large Load diff

View file

@ -0,0 +1,51 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods related to generating code for representing script constants
// as run-time values. There's only one nontrivial one of these,
// RegisterConstant() (declared above, as it's public). All the other
// work is done by secondary objects - see InitsInfo.{h,cc} for those.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
public:
// Tracks a Zeek ValPtr used as a constant value. These occur in two
// contexts: directly as constant expressions, and indirectly as elements
// within aggregate constants (such as in vector initializers).
//
// Returns the associated initialization info. In addition, consts_offset
// returns an offset into an initialization-time global that tracks all
// constructed globals, providing general access to them for aggregate
// constants.
std::shared_ptr<CPP_InitInfo> RegisterConstant(const ValPtr& vp, int& consts_offset);
private:
// Maps (non-native) constants to associated C++ globals.
std::unordered_map<const ConstExpr*, std::string> const_exprs;
// Maps the values of (non-native) constants to associated initializer
// information.
std::unordered_map<const Val*, std::shared_ptr<CPP_InitInfo>> const_vals;
// Same, but for the offset into the vector that tracks all constants
// collectively (to support initialization of compound constants).
std::unordered_map<const Val*, int> const_offsets;
// The same as the above pair, but indexed by the string representation
// rather than the Val*. The reason for having both is to enable
// reusing common constants even though their Val*'s differ.
std::unordered_map<std::string, std::shared_ptr<CPP_InitInfo>> constants;
std::unordered_map<std::string, int> constants_offsets;
// Used for memory management associated with const_vals's index.
std::vector<ValPtr> cv_indices;
// For different types of constants (as indicated by TypeTag),
// provides the associated object that manages the initializers
// for those constants.
std::unordered_map<TypeTag, std::shared_ptr<CPP_InitsInfo>> const_info;
// Tracks entries for constructing the vector of all constants
// (regardless of type). Each entry provides a TypeTag, used
// to identify the type-specific vector for a given constant,
// and the offset into that vector.
std::vector<std::pair<TypeTag, int>> consts;

View file

@ -0,0 +1,101 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for generating declarations of functions and lambdas.
// The counterpart to GenFunc.cc.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
// Generates declarations (class, forward reference to C++ function) for the
// given script function.
void DeclareFunc(const FuncInfo& func);
// Similar, but for lambdas.
void DeclareLambda(const LambdaExpr* l, const ProfileFunc* pf);
// Generates code to declare the compiled version of a script function.
// "ft" gives the functions type, "pf" its profile, "fname" its C++ name,
// "body" its AST, "l" if non-nil its corresponding lambda expression, and
// "flavor" whether it's a hook/event/function.
//
// We use two basic approaches. Most functions are represented by a
// "CPPDynStmt" object that's parameterized by a void* pointer to the
// underlying C++ function and an index used to dynamically cast the pointer
// to having the correct type for then calling it. Lambdas, however
// (including "implicit" lambdas used to associate complex expressions with
// &attributes), each have a unique subclass derived from CPPStmt that calls
// the underlying C++ function without requiring a cast, and that holds the
// values of the lambda's captures.
//
// It would be cleanest to use the latter approach for all functions, but
// the hundreds/thousands of additional classes required for doing so
// significantly slows down C++ compilation, so we instead opt for the uglier
// dynamic casting approach, which only requires one additional class.
void CreateFunction(const FuncTypePtr& ft, const ProfileFunc* pf, const std::string& fname, const StmtPtr& body,
int priority, const LambdaExpr* l, FunctionFlavor flavor);
// Used for the case of creating a custom subclass of CPPStmt.
void DeclareSubclass(const FuncTypePtr& ft, const ProfileFunc* pf, const std::string& fname, const std::string& args,
const IDPList* lambda_ids);
// Used for the case of employing an instance of a CPPDynStmt object.
void DeclareDynCPPStmt();
// Generates the declarations (and in-line definitions) associated with
// compiling a lambda.
void BuildLambda(const FuncTypePtr& ft, const ProfileFunc* pf, const std::string& fname, const StmtPtr& body,
const LambdaExpr* l, const IDPList* lambda_ids);
// For a call to the C++ version of a function of type "ft" and with lambda
// captures lambda_ids (nil if not applicable), generates code that binds the
// Interpreter arguments (i.e., Frame offsets) to C++ function arguments, as
// well as passing in the captures.
std::string BindArgs(const FuncTypePtr& ft, const IDPList* lambda_ids);
// Generates the declaration for the parameters for a function with the given
// type, lambda captures (if non-nil), and profile.
std::string ParamDecl(const FuncTypePtr& ft, const IDPList* lambda_ids, const ProfileFunc* pf);
// Returns in p_types the types associated with the parameters for a function
// of the given type, set of lambda captures (if any), and profile.
void GatherParamTypes(std::vector<std::string>& p_types, const FuncTypePtr& ft, const IDPList* lambda_ids,
const ProfileFunc* pf);
// Same, but instead returns the parameter's names.
void GatherParamNames(std::vector<std::string>& p_names, const FuncTypePtr& ft, const IDPList* lambda_ids,
const ProfileFunc* pf);
// Inspects the given profile to find the i'th parameter (starting at 0).
// Returns nil if the profile indicates that the parameter is not used by the
// function.
const ID* FindParam(int i, const ProfileFunc* pf);
// Information associated with a CPPDynStmt dynamic dispatch.
struct DispatchInfo {
std::string cast; // C++ cast to use for function pointer
std::string args; // arguments to pass to the function
bool is_hook; // whether the function is a hook
TypePtr yield; // what type the function returns, if any
};
// An array of cast/invocation pairs used to generate the CPPDynStmt Exec
// method.
std::vector<DispatchInfo> func_casting_glue;
// Maps casting strings to indices into func_casting_glue. The index is
// what's used to dynamically switch to the right dispatch.
std::unordered_map<std::string, int> casting_index;
// Maps functions (using their C++ name) to their casting strings.
std::unordered_map<std::string, std::string> func_index;
// Names for lambda capture ID's. These require a separate space that
// incorporates the lambda's name, to deal with nested lambda's that refer
// to the identifiers with the same name.
std::unordered_map<const ID*, std::string> lambda_names;
// The function's parameters. Tracked so we don't re-declare them.
IDSet params;
// Whether we're compiling a hook.
bool in_hook = false;

View file

@ -0,0 +1,98 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for driving the overall "-O gen-C++" compilation process.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
// Main driver, invoked by constructor.
void Compile(bool report_uncompilable);
// Generate the beginning of the compiled code: run-time functions,
// namespace, auxiliary globals.
void GenProlog();
// The following methods all create objects that track the initializations
// of a given type of value. In each, "tag" is the name used to identify the
// initializer global associated with the given type of value, and "type" is
// its C++ representation. Often "tag" is concatenated with "type" to designate
// a specific C++ type. For example, "tag" might be "Double" and "type" might
// be "ValPtr"; the resulting global's type is "DoubleValPtr".
// Creates an object for tracking values associated with Zeek constants.
// "c_type" is the C++ type used in the initializer for each object; or, if
// empty, it specifies that we represent the value using an index into a
// separate vector that holds the constant.
std::shared_ptr<CPP_InitsInfo> CreateConstInitInfo(const char* tag, const char* type, const char* c_type);
// Creates an object for tracking compound initializers, which are whose
// initialization uses indexes into other vectors.
std::shared_ptr<CPP_InitsInfo> CreateCompoundInitInfo(const char* tag, const char* type);
// Creates an object for tracking initializers that have custom C++ objects
// to hold their initialization information.
std::shared_ptr<CPP_InitsInfo> CreateCustomInitInfo(const char* tag, const char* type);
// Generates the declaration associated with a set of initializations and
// tracks the object to facilitate looping over all so initializations.
// As a convenience, returns the object.
std::shared_ptr<CPP_InitsInfo> RegisterInitInfo(const char* tag, const char* type, std::shared_ptr<CPP_InitsInfo> gi);
// Given the name of a function body that's been compiled, generate code to
// register it at run-time, and track its associated hash so subsequent
// compilations can reuse it.
void RegisterCompiledBody(const std::string& f);
// After compilation, generate the final code. Most of this is in support
// of run-time initialization of various dynamic values.
void GenEpilog();
// Generate the main method of the CPPDynStmt class, doing dynamic dispatch
// for function invocation.
void GenCPPDynStmt();
// Generate a function to load BiFs.
void GenLoadBiFs();
// Generate the main initialization function, which finalizes the run-time
// environment.
void GenFinishInit();
// Generate the function that registers compiled script bodies.
void GenRegisterBodies();
// True if the given function (plus body and profile) is one that should be
// compiled. If non-nil, sets reason to the the reason why, if there's a
// fundamental problem. If however the function should be skipped for other
// reasons, then sets it to nil.
bool IsCompilable(const FuncInfo& func, const char** reason = nullptr);
// The set of functions/bodies we're compiling.
std::vector<FuncInfo>& funcs;
// The global profile of all of the functions.
std::shared_ptr<ProfileFuncs> pfs;
// Script functions that we are able to compile. We compute these ahead
// of time so that when compiling script function A which makes a call to
// script function B, we know whether B will indeed be compiled, or if it'll
// be interpreted due to it including some functionality we don't currently
// support for compilation.
//
// Indexed by the C++ name of the function.
std::unordered_set<std::string> compilable_funcs;
// Tracks which functions/hooks/events have at least one non-compilable body.
// Indexed by the Zeek name of function.
std::unordered_set<std::string> not_fully_compilable;
// Maps functions (not hooks or events) to upstream compiled names.
std::unordered_map<std::string, std::string> hashed_funcs;
// If true, the generated code should run "standalone".
bool standalone = false;
// Hash over the functions in this compilation. This is only needed for
// "seatbelts", to ensure that we can produce a unique hash relating to this
// compilation (*and* its compilation time, which is why these are "seatbelts"
// and likely not important to make distinct).
p_hash_type total_hash = 0;

80
src/script_opt/CPP/Emit.h Normal file
View file

@ -0,0 +1,80 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Low-level methods for emitting code.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
// The following all need to be able to emit code.
friend class CPP_BasicConstInitsInfo;
friend class CPP_CompoundInitsInfo;
friend class IndicesManager;
// Used to create (indented) C++ {...} code blocks. "needs_semi"
// controls whether to terminate the block with a ';' (such as
// for class definitions.
void StartBlock();
void EndBlock(bool needs_semi = false);
void IndentUp() { ++block_level; }
void IndentDown() { --block_level; }
// Various ways of generating code. The multi-argument methods
// assume that the first argument is a printf-style format
// (but one that can only have %s specifiers).
void Emit(const std::string& str) const {
Indent();
fprintf(write_file, "%s", str.c_str());
NL();
}
void Emit(const std::string& fmt, const std::string& arg, bool do_NL = true) const {
Indent();
fprintf(write_file, fmt.c_str(), arg.c_str());
if ( do_NL )
NL();
}
void Emit(const std::string& fmt, const std::string& arg1, const std::string& arg2) const {
Indent();
fprintf(write_file, fmt.c_str(), arg1.c_str(), arg2.c_str());
NL();
}
void Emit(const std::string& fmt, const std::string& arg1, const std::string& arg2, const std::string& arg3) const {
Indent();
fprintf(write_file, fmt.c_str(), arg1.c_str(), arg2.c_str(), arg3.c_str());
NL();
}
void Emit(const std::string& fmt, const std::string& arg1, const std::string& arg2, const std::string& arg3,
const std::string& arg4) const {
Indent();
fprintf(write_file, fmt.c_str(), arg1.c_str(), arg2.c_str(), arg3.c_str(), arg4.c_str());
NL();
}
void Emit(const std::string& fmt, const std::string& arg1, const std::string& arg2, const std::string& arg3,
const std::string& arg4, const std::string& arg5) const {
Indent();
fprintf(write_file, fmt.c_str(), arg1.c_str(), arg2.c_str(), arg3.c_str(), arg4.c_str(), arg5.c_str());
NL();
}
void Emit(const std::string& fmt, const std::string& arg1, const std::string& arg2, const std::string& arg3,
const std::string& arg4, const std::string& arg5, const std::string& arg6) const {
Indent();
fprintf(write_file, fmt.c_str(), arg1.c_str(), arg2.c_str(), arg3.c_str(), arg4.c_str(), arg5.c_str(),
arg6.c_str());
NL();
}
void NL() const { fputc('\n', write_file); }
// Indents to the current indentation level.
void Indent() const;
// File to which we're generating code.
FILE* write_file;
// Indentation level.
int block_level = 0;

147
src/script_opt/CPP/Exprs.h Normal file
View file

@ -0,0 +1,147 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for generating code corresponding with Zeek expression AST nodes
// (Expr objects).
//
// This file is included by Compile.h to insert into the CPPCompiler class.
// These methods are all oriented around returning strings of C++ code;
// they do not directly emit the code, since often the caller will be embedding
// the result in some surrounding context. No effort is made to reduce string
// copying; this isn't worth the hassle, as it takes just a few seconds for
// the compiler to generate 100K+ LOC that clang will then need 10s of seconds
// to compile, so speeding up the compiler has little practical advantage.
// The following enum's represent whether, for expressions yielding native
// values, the end goal is to have the value in (1) native form, (2) instead
// in ValPtr form, or (3) whichever is more convenient to generate (sometimes
// used when the caller knows that the value is non-native).
enum GenType {
GEN_NATIVE,
GEN_VAL_PTR,
GEN_DONT_CARE,
};
// Generate an expression for which we want the result embedded in {}
// initializers (generally to be used in calling a function where we want
// those values to be translated to a vector<ValPtr>).
std::string GenExprs(const Expr* e);
// Generate the value(s) associated with a ListExpr. If true, the "nested"
// parameter indicates that this list is embedded within an outer list, in
// which case it's expanded to include {}'s. It's false if the ListExpr is
// at the top level, such as when expanding the arguments in a CallExpr.
std::string GenListExpr(const Expr* e, GenType gt, bool nested);
// Per-Expr-subclass code generation. The resulting code generally reflects
// the corresponding Eval() or Fold() methods.
std::string GenExpr(const ExprPtr& e, GenType gt, bool top_level = false) { return GenExpr(e.get(), gt, top_level); }
std::string GenExpr(const Expr* e, GenType gt, bool top_level = false);
std::string GenNameExpr(const NameExpr* ne, GenType gt);
std::string GenConstExpr(const ConstExpr* c, GenType gt);
std::string GenAggrAdd(const Expr* e);
std::string GenAggrDel(const Expr* e);
std::string GenIncrExpr(const Expr* e, GenType gt, bool is_incr, bool top_level);
std::string GenCondExpr(const Expr* e, GenType gt);
std::string GenCallExpr(const CallExpr* c, GenType gt, bool top_level);
std::string GenInExpr(const Expr* e, GenType gt);
std::string GenFieldExpr(const FieldExpr* fe, GenType gt);
std::string GenHasFieldExpr(const HasFieldExpr* hfe, GenType gt);
std::string GenIndexExpr(const Expr* e, GenType gt);
std::string GenAssignExpr(const Expr* e, GenType gt, bool top_level);
std::string GenAddToExpr(const Expr* e, GenType gt, bool top_level);
std::string GenRemoveFromExpr(const Expr* e, GenType gt, bool top_level);
std::string GenSizeExpr(const Expr* e, GenType gt);
std::string GenScheduleExpr(const Expr* e);
std::string GenLambdaExpr(const Expr* e);
std::string GenLambdaExpr(const Expr* e, std::string capture_args);
std::string GenIsExpr(const Expr* e, GenType gt);
std::string GenArithCoerceExpr(const Expr* e, GenType gt);
std::string GenRecordCoerceExpr(const Expr* e);
std::string GenTableCoerceExpr(const Expr* e);
std::string GenVectorCoerceExpr(const Expr* e);
std::string GenRecordConstructorExpr(const Expr* e);
std::string GenSetConstructorExpr(const Expr* e);
std::string GenTableConstructorExpr(const Expr* e);
std::string GenVectorConstructorExpr(const Expr* e);
// Generate code for constants that can be expressed directly as C++ constants.
std::string GenVal(const ValPtr& v);
// Helper functions for particular Expr subclasses / flavors.
std::string GenUnary(const Expr* e, GenType gt, const char* op, const char* vec_op = nullptr);
std::string GenBinary(const Expr* e, GenType gt, const char* op, const char* vec_op = nullptr);
std::string GenBinarySet(const Expr* e, GenType gt, const char* op);
std::string GenBinaryString(const Expr* e, GenType gt, const char* op);
std::string GenBinaryPattern(const Expr* e, GenType gt, const char* op);
std::string GenBinaryAddr(const Expr* e, GenType gt, const char* op);
std::string GenBinarySubNet(const Expr* e, GenType gt, const char* op);
std::string GenEQ(const Expr* e, GenType gt, const char* op, const char* vec_op);
std::string GenAssign(const ExprPtr& lhs, const ExprPtr& rhs, const std::string& rhs_native,
const std::string& rhs_val_ptr, GenType gt, bool top_level);
std::string GenDirectAssign(const ExprPtr& lhs, const std::string& rhs_native, const std::string& rhs_val_ptr,
GenType gt, bool top_level);
std::string GenIndexAssign(const ExprPtr& lhs, const ExprPtr& rhs, const std::string& rhs_val_ptr, GenType gt,
bool top_level);
std::string GenFieldAssign(const ExprPtr& lhs, const ExprPtr& rhs, const std::string& rhs_native,
const std::string& rhs_val_ptr, GenType gt, bool top_level);
std::string GenListAssign(const ExprPtr& lhs, const ExprPtr& rhs);
// Support for element-by-element vector operations.
std::string GenVectorOp(const Expr* e, std::string op, const char* vec_op);
std::string GenVectorOp(const Expr* e, std::string op1, std::string op2, const char* vec_op);
// If "all_deep" is true, it means make all of the captures deep copies,
// not just the ones that were explicitly marked as deep copies. That
// functionality is used to support Clone() methods; it's not needed when
// creating a new lambda instance.
std::string GenLambdaClone(const LambdaExpr* l, bool all_deep);
// Returns an initializer list for a vector of integers.
std::string GenIntVector(const std::vector<int>& vec);
// The following are used to generate accesses to elements of extensible
// types. They first check whether the type has been extended (for records,
// beyond the field of interest); if not, then the access is done directly.
// If the access is however to an extended element, then they indirect the
// access through a map that is generated dynamically when the compiled code.
// Doing so allows the compiled code to work in contexts where other extensions
// occur that would otherwise conflict with hardwired offsets/values.
std::string GenField(const ExprPtr& rec, int field);
std::string GenEnum(const TypePtr& et, const ValPtr& ev);
// For record that are extended via redef's, maps fields beyond the original
// definition to locations in the global (in the compiled code) "field_mapping"
// array.
//
// So for each such record, there's a second map of field-in-the-record to
// offset-in-field_mapping.
std::unordered_map<const RecordType*, std::unordered_map<int, int>> record_field_mappings;
// Total number of such mappings (i.e., entries in the inner maps, not the
// outer map).
int num_rf_mappings = 0;
// For each entry in "field_mapping", the record (as a global offset) and
// TypeDecl associated with the mapping.
std::vector<std::pair<int, const TypeDecl*>> field_decls;
// For enums that are extended via redef's, maps each distinct value (that
// the compiled scripts refer to) to locations in the global (in the compiled
// code) "enum_mapping" array.
//
// So for each such enum, there's a second map of value-during-compilation to
// offset-in-enum_mapping.
std::unordered_map<const EnumType*, std::unordered_map<int, int>> enum_val_mappings;
// Total number of such mappings (i.e., entries in the inner maps, not the
// outer map).
int num_ev_mappings = 0;
// For each entry in "enum_mapping", the EnumType (as a global offset) and
// name associated with the mapping.
std::vector<std::pair<int, std::string>> enum_names;

View file

@ -0,0 +1,74 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for generating function/lambda definitions. The counterpart
// to DeclFunc.cc.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
// Driver functions for compiling the body of the given function or lambda.
void CompileFunc(const FuncInfo& func);
void CompileLambda(const LambdaExpr* l, const ProfileFunc* pf);
// Generates the body of the Invoke() method (which supplies the "glue"
// for calling the C++-generated code, for CPPStmt subclasses).
void GenInvokeBody(const std::string& fname, const TypePtr& t, const std::string& args) {
GenInvokeBody(fname + "(" + args + ")", t);
}
void GenInvokeBody(const std::string& call, const TypePtr& t);
// Generates the code for the body of a script function with the given
// type, profile, C++ name, AST, lambda captures (if non-nil), and
// hook/event/function "flavor".
void DefineBody(const FuncTypePtr& ft, const ProfileFunc* pf, const std::string& fname, const StmtPtr& body,
const IDPList* lambda_ids, FunctionFlavor flavor);
// Declare parameters that originate from a type signature of "any" but were
// concretized in this declaration.
void TranslateAnyParams(const FuncTypePtr& ft, const ProfileFunc* pf);
// Generates code to dynamically initialize any events referred to in the
// function.
void InitializeEvents(const ProfileFunc* pf);
// Declare local variables (which are non-globals that aren't parameters or
// lambda captures).
void DeclareLocals(const ProfileFunc* func, const IDPList* lambda_ids);
// Returns the C++ name to use for a given function body.
std::string BodyName(const FuncInfo& func);
// Generate the arguments to be used when calling a C++-generated function.
std::string GenArgs(const RecordTypePtr& params, const Expr* e);
// Functions that we've declared/compiled. Indexed by full C++ name.
std::unordered_set<std::string> compiled_funcs;
// "Simple" functions that we've compiled, i.e., those that have a single
// body and thus can be called directly. Indexed by function name, and
// maps to the C++ name.
std::unordered_map<std::string, std::string> compiled_simple_funcs;
// Maps function bodies to the names we use for them.
std::unordered_map<const Stmt*, std::string> body_names;
// Maps function names to hashes of bodies.
std::unordered_map<std::string, p_hash_type> body_hashes;
// Maps function names to priorities, for hooks & event handlers.
std::unordered_map<std::string, int> body_priorities;
// Maps function names to script locations, for better-than-nothing error
// reporting.
std::unordered_map<std::string, const Location*> body_locs;
// Maps function names to events relevant to them.
std::unordered_map<std::string, std::vector<std::string>> body_events;
// Full type of the function we're currently compiling.
FuncTypePtr func_type;
// Return type of the function we're currently compiling.
TypePtr ret_type;
// Internal name of the function we're currently compiling.
std::string body_name;

127
src/script_opt/CPP/Inits.h Normal file
View file

@ -0,0 +1,127 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for generating run-time initialization of objects relating to
// Zeek values and types.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
public:
// True if the given expression is simple enough that we can generate code
// to evaluate it directly, and don't need to create a separate function per
// RegisterInitExpr() to track it.
static bool IsSimpleInitExpr(const ExprPtr& e);
// Easy access to the global offset and the initialization
// cohort associated with a given type.
int TypeOffset(const TypePtr& t) { return GI_Offset(RegisterType(t)); }
int TypeCohort(const TypePtr& t) { return GI_Cohort(RegisterType(t)); }
int TypeFinalCohort(const TypePtr& t) { return GI_FinalCohort(RegisterType(t)); }
// Tracks expressions used in attributes (such as &default=<expr>).
//
// We need to generate code to evaluate these, via CallExpr's that invoke
// functions that return the value of the expression. However, we can't
// generate that code when first encountering the attribute, because doing
// so will need to refer to the names of types, and initially those are
// unavailable (because the type's representatives, per pfs->RepTypes(), might
// not have yet been tracked). So instead we track the associated
// CallExprInitInfo objects, and after all types have been tracked, then spin
// through them to generate the code.
//
// Returns the associated initialization information.
std::shared_ptr<CPP_InitInfo> RegisterInitExpr(const ExprPtr& e);
// Tracks a C++ string value needed for initialization. Returns
// an offset into the global vector that will hold these.
int TrackString(std::string s) {
auto ts = tracked_strings.find(s);
if ( ts != tracked_strings.end() )
return ts->second;
int offset = ordered_tracked_strings.size();
tracked_strings[s] = offset;
ordered_tracked_strings.emplace_back(s);
return offset;
}
// Tracks a profile hash value needed for initialization. Returns
// an offset into the global vector that will hold these.
int TrackHash(p_hash_type h) {
auto th = tracked_hashes.find(h);
if ( th != tracked_hashes.end() )
return th->second;
int offset = ordered_tracked_hashes.size();
tracked_hashes[h] = offset;
ordered_tracked_hashes.emplace_back(h);
return offset;
}
private:
// Generates code for dynamically generating an expression associated with an
// attribute, via a function call.
void GenInitExpr(std::shared_ptr<CallExprInitInfo> ce_init);
// Returns the name of a function used to evaluate an initialization expression.
std::string InitExprName(const ExprPtr& e);
// Convenience functions for returning the offset or initialization cohort
// associated with an initialization.
int GI_Offset(const std::shared_ptr<CPP_InitInfo>& gi) const { return gi ? gi->Offset() : -1; }
int GI_Cohort(const std::shared_ptr<CPP_InitInfo>& gi) const { return gi ? gi->InitCohort() : 0; }
int GI_FinalCohort(const std::shared_ptr<CPP_InitInfo>& gi) const { return gi ? gi->FinalInitCohort() : 0; }
// Generate code to initialize the mappings for record field offsets for field
// accesses into regions of records that can be extensible (and thus can vary
// at run-time to the offsets encountered during compilation).
void InitializeFieldMappings();
// Same, but for enum types.
void InitializeEnumMappings();
// Generate code to initialize BiFs.
void InitializeBiFs();
// Generate code to initialize strings that we track.
void InitializeStrings();
// Generate code to initialize hashes that we track.
void InitializeHashes();
// Generate code to initialize indirect references to constants.
void InitializeConsts();
// Generate code to initialize globals (using dynamic statements rather than
// constants).
void InitializeGlobals();
// Generate the initialization hook for this set of compiled code.
void GenInitHook();
// Generates code to activate standalone code.
void GenStandaloneActivation();
// Generates code to register the initialization for standalone use, and
// prints to stdout a Zeek script that can load all of what we compiled.
void GenLoad();
// A list of BiFs to look up during initialization. First string is the name
// of the C++ global holding the BiF, the second is its name as known to Zeek.
std::unordered_map<std::string, std::string> BiFs;
// Expressions for which we need to generate initialization-time code.
// Currently, these are only expressions appearing in attributes.
CPPTracker<Expr> init_exprs = {"gen_init_expr", false};
// Maps strings to associated offsets.
std::unordered_map<std::string, int> tracked_strings;
// Tracks strings we've registered in order (corresponding to
// their offsets).
std::vector<std::string> ordered_tracked_strings;
// The same as the previous two, but for profile hashes.
std::vector<p_hash_type> ordered_tracked_hashes;
std::unordered_map<p_hash_type, int> tracked_hashes;

View file

@ -5,7 +5,7 @@
#include "zeek/Desc.h" #include "zeek/Desc.h"
#include "zeek/RE.h" #include "zeek/RE.h"
#include "zeek/ZeekString.h" #include "zeek/ZeekString.h"
#include "zeek/script_opt/CPP/Attrs.h" #include "zeek/script_opt/CPP/AttrExprType.h"
#include "zeek/script_opt/CPP/Compile.h" #include "zeek/script_opt/CPP/Compile.h"
#include "zeek/script_opt/CPP/RuntimeInits.h" #include "zeek/script_opt/CPP/RuntimeInits.h"

View file

@ -5,7 +5,7 @@
#pragma once #pragma once
#include "zeek/Val.h" #include "zeek/Val.h"
#include "zeek/script_opt/CPP/Attrs.h" #include "zeek/script_opt/CPP/AttrExprType.h"
#include "zeek/script_opt/CPP/Func.h" #include "zeek/script_opt/CPP/Func.h"
namespace zeek { namespace zeek {

View file

@ -0,0 +1,34 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for generating code corresponding with Zeek statement AST nodes
// (Stmt objects). For the most part, code generation is straightforward as
// it matches the Exec/DoExec methods of the corresponding Stmt subclasses.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
void GenStmt(const StmtPtr& s) { GenStmt(s.get()); }
void GenStmt(const Stmt* s);
void GenInitStmt(const InitStmt* init);
void GenIfStmt(const IfStmt* i);
void GenWhileStmt(const WhileStmt* w);
void GenReturnStmt(const ReturnStmt* r);
void GenEventStmt(const EventStmt* ev);
void GenSwitchStmt(const SwitchStmt* sw);
void GenTypeSwitchStmt(const Expr* e, const case_list* cases);
void GenTypeSwitchCase(const ID* id, int case_offset, bool is_multi);
void GenValueSwitchStmt(const Expr* e, const case_list* cases);
void GenWhenStmt(const WhenStmt* w);
void GenWhenStmt(const WhenInfo* wi, const std::string& when_lambda, const Location* loc,
std::vector<std::string> local_aggrs);
void GenForStmt(const ForStmt* f);
void GenForOverTable(const ExprPtr& tbl, const IDPtr& value_var, const IDPList* loop_vars);
void GenForOverVector(const ExprPtr& tbl, const IDPtr& value_var, const IDPList* loop_vars);
void GenForOverString(const ExprPtr& str, const IDPList* loop_vars);
void GenAssertStmt(const AssertStmt* a);
// Nested level of loops/switches for which "break"'s should be
// C++ breaks rather than a "hook" break.
int break_level = 0;

View file

@ -0,0 +1,59 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for dealing with Zeek script types.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
public:
// Tracks the given type (with support methods for ones that are complicated),
// recursively including its sub-types, and creating initializations for
// constructing C++ variables representing the types.
//
// Returns the initialization info associated with the type.
std::shared_ptr<CPP_InitInfo> RegisterType(const TypePtr& t);
private:
// "Native" types are those Zeek scripting types that we support using
// low-level C++ types (like "zeek_uint_t" for "count"). Types that we
// instead support using some form of ValPtr representation are "non-native".
bool IsNativeType(const TypePtr& t) const;
// Given an expression corresponding to a native type (and with the given
// script type 't'), converts it to the given GenType.
std::string NativeToGT(const std::string& expr, const TypePtr& t, GenType gt);
// Given an expression with a C++ type of generic "ValPtr", of the given script
// type 't', converts it as needed to the given GenType.
std::string GenericValPtrToGT(const std::string& expr, const TypePtr& t, GenType gt);
// Returns the name of a C++ variable that will hold a TypePtr of the
// appropriate flavor. 't' does not need to be a type representative.
std::string GenTypeName(const Type* t);
std::string GenTypeName(const TypePtr& t) { return GenTypeName(t.get()); }
// Returns the "representative" for a given type, used to ensure that we
// re-use the C++ variable corresponding to a type and don't instantiate
// redundant instances.
const Type* TypeRep(const Type* t) { return pfs->TypeRep(t); }
const Type* TypeRep(const TypePtr& t) { return TypeRep(t.get()); }
// Low-level C++ representations for types, of various flavors.
static const char* TypeTagName(TypeTag tag);
const char* TypeName(const TypePtr& t);
const char* FullTypeName(const TypePtr& t);
const char* TypeType(const TypePtr& t);
// Access to a type's underlying values.
const char* NativeAccessor(const TypePtr& t);
// The name for a type that should be used in declaring an IntrusivePtr to
// such a type.
const char* IntrusiveVal(const TypePtr& t);
// Maps types to indices in the global "CPP__Type__" array.
CPPTracker<Type> types = {"types", true};
// Used to prevent analysis of mutually-referring types from leading to
// infinite recursion. Maps types to their global initialization information
// (or, initially, to nullptr, if they're in the process of being registered).
std::unordered_map<const Type*, std::shared_ptr<CPP_InitInfo>> processed_types;

69
src/script_opt/CPP/Vars.h Normal file
View file

@ -0,0 +1,69 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods related to Zeek script variables and their C++ counterparts.
//
// This file is included by Compile.h to insert into the CPPCompiler class.
public:
// Tracks a global to generate the necessary initialization.
// Returns the associated initialization info.
std::shared_ptr<CPP_InitInfo> RegisterGlobal(const ID* g);
private:
// Generate declarations associated with the given global, and, if it's used
// as a variable (not just as a function being called), track it as such.
void CreateGlobal(const ID* g);
// Register the given identifier as a BiF. If is_var is true then the BiF
// is also used in a non-call context.
void AddBiF(const ID* b, bool is_var);
// Register the given global name. "suffix" distinguishes particular types
// of globals, such as the names of bifs, global (non-function) variables,
// or compiled Zeek functions.
bool AddGlobal(const std::string& g, const char* suffix);
// Tracks that the body we're currently compiling refers to the given event.
void RegisterEvent(std::string ev_name);
// The following match various forms of identifiers to the name used for
// their C++ equivalent.
const char* IDName(const IDPtr& id) { return IDName(id.get()); }
const char* IDName(const ID* id) { return IDNameStr(id).c_str(); }
const std::string& IDNameStr(const ID* id);
// Returns a canonicalized version of a variant of a global made distinct by
// the given suffix.
std::string GlobalName(const std::string& g, const char* suffix) { return Canonicalize(g.c_str()) + "_" + suffix; }
// Returns a canonicalized form of a local identifier's name, expanding its
// module prefix if needed.
std::string LocalName(const ID* l) const;
std::string LocalName(const IDPtr& l) const { return LocalName(l.get()); }
// The same, but for a capture.
std::string CaptureName(const ID* l) const;
std::string CaptureName(const IDPtr& l) const { return CaptureName(l.get()); }
// Returns a canonicalized name, with various non-alphanumeric characters
// stripped or transformed, and guaranteed not to conflict with C++ keywords.
std::string Canonicalize(const char* name) const;
// Returns the name of the global corresponding to an expression (which must
// be a EXPR_NAME).
std::string GlobalName(const ExprPtr& e) { return globals[e->AsNameExpr()->Id()->Name()]; }
// Maps global names (not identifiers) to the names we use for them.
std::unordered_map<std::string, std::string> globals;
// Similar for locals, for the function currently being compiled.
std::unordered_map<const ID*, std::string> locals;
// Retrieves the initialization information associated with the given global.
std::unordered_map<const ID*, std::shared_ptr<CPP_InitInfo>> global_gis;
// Maps event names to the names we use for them.
std::unordered_map<std::string, std::string> events;
// Globals that correspond to variables, not functions.
IDSet global_vars;

101
src/script_opt/ZAM/AM-Opt.h Normal file
View file

@ -0,0 +1,101 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for low-level optimization of the ZAM abstract machine.
//
// This file is included by Compile.h to insert into the ZAMCompiler class.
// Optimizing the low-level compiled instructions.
void OptimizeInsts();
// Tracks which instructions can be branched to via the given
// set of switches.
template<typename T>
void TallySwitchTargets(const CaseMapsI<T>& switches);
// Remove code that can't be reached. True if some removal happened.
bool RemoveDeadCode();
// Collapse chains of gotos. True if some something changed.
bool CollapseGoTos();
// Prune statements that are unnecessary. True if something got
// pruned.
bool PruneUnused();
// For the current state of insts1, compute lifetimes of frame
// denizens (variable(s) using a given frame slot) in terms of
// first-instruction-to-last-instruction during which they're
// relevant, including consideration for loops.
void ComputeFrameLifetimes();
// Given final frame lifetime information, remaps frame members
// with non-overlapping lifetimes to share slots.
void ReMapFrame();
// Given final frame lifetime information, remaps slots in the
// interpreter frame. (No longer strictly necessary.)
void ReMapInterpreterFrame();
// Computes the remapping for a variable currently in the given slot,
// whose scope begins at the given instruction.
void ReMapVar(const ID* id, int slot, zeek_uint_t inst);
// Look to initialize the beginning of local lifetime based on slot
// assignment at instruction inst.
void CheckSlotAssignment(int slot, const ZInstI* inst);
// Track that a local's lifetime begins at the given statement.
void SetLifetimeStart(int slot, const ZInstI* inst);
// Look for extension of local lifetime based on slot usage
// at instruction inst.
void CheckSlotUse(int slot, const ZInstI* inst);
// Extend (or create) the end of a local's lifetime.
void ExtendLifetime(int slot, const ZInstI* inst);
// Returns the (live) instruction at the beginning/end of the loop(s)
// within which the given instruction lies; or that instruction
// itself if it's not inside a loop. The second argument specifies
// the loop depth. For example, a value of '2' means "extend to
// the beginning/end of any loop(s) of depth >= 2".
const ZInstI* BeginningOfLoop(const ZInstI* inst, int depth) const;
const ZInstI* EndOfLoop(const ZInstI* inst, int depth) const;
// True if any statement other than a frame sync uses the given slot.
bool VarIsUsed(int slot) const;
// Find the first non-dead instruction after i (inclusive).
// If follow_gotos is true, then if that instruction is
// an unconditional branch, continues the process until
// a different instruction is found (and report if there
// are infinite loops).
//
// First form returns nil if there's nothing live after i.
// Second form returns insts1.size() in that case.
ZInstI* FirstLiveInst(ZInstI* i, bool follow_gotos = false);
zeek_uint_t FirstLiveInst(zeek_uint_t i, bool follow_gotos = false);
// Same, but not including i.
ZInstI* NextLiveInst(ZInstI* i, bool follow_gotos = false) {
if ( i->inst_num == static_cast<int>(insts1.size()) - 1 )
return nullptr;
return FirstLiveInst(insts1[i->inst_num + 1], follow_gotos);
}
int NextLiveInst(int i, bool follow_gotos = false) { return FirstLiveInst(i + 1, follow_gotos); }
// Mark an instruction as unnecessary and remove its influence on
// other statements. The instruction is indicated as an offset
// into insts1; any labels associated with it are transferred
// to its next live successor, if any.
void KillInst(ZInstI* i) { KillInst(i->inst_num); }
void KillInst(zeek_uint_t i);
// Helper function for propagating control flow (of a given type)
// backwards, when the instruction at the given offset has been killed.
void BackPropagateCFT(int inst_num, ControlFlowType cf_type);
// The same, but kills any successor instructions until finding
// one that's labeled.
void KillInsts(ZInstI* i) { KillInsts(i->inst_num); }
void KillInsts(zeek_uint_t i);

View file

@ -0,0 +1,56 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for managing low-level ZAM control flow, which is implemented
// using go-to branches.
//
// This file is included by Compile.h to insert into the ZAMCompiler class.
void PushNexts() { PushGoTos(nexts); }
void PushBreaks() { PushGoTos(breaks); }
void PushFallThroughs() { PushGoTos(fallthroughs); }
void PushCatchReturns() { PushGoTos(catches); }
void ResolveNexts(const InstLabel l) { ResolveGoTos(nexts, l, CFT_NEXT); }
void ResolveBreaks(const InstLabel l) { ResolveGoTos(breaks, l, CFT_BREAK); }
void ResolveFallThroughs(const InstLabel l) { ResolveGoTos(fallthroughs, l); }
void ResolveCatchReturns(const InstLabel l) { ResolveGoTos(catches, l, CFT_INLINED_RETURN); }
using GoToSet = std::vector<ZAMStmt>;
using GoToSets = std::vector<GoToSet>;
void PushGoTos(GoToSets& gotos);
void ResolveGoTos(GoToSets& gotos, const InstLabel l, ControlFlowType cft = CFT_NONE);
ZAMStmt GenGoTo(GoToSet& v);
ZAMStmt GoToStub();
ZAMStmt GoTo(const InstLabel l);
InstLabel GoToTarget(const ZAMStmt s);
InstLabel GoToTargetBeyond(const ZAMStmt s);
void SetTarget(ZInstI* inst, const InstLabel l, int slot);
// Given a GoTo target, find its live equivalent (first instruction
// at that location or beyond that's live).
ZInstI* FindLiveTarget(ZInstI* goto_target);
// Given an instruction that has a slot associated with the
// given target, updates the slot to correspond with the current
// instruction number of the target.
void ConcretizeBranch(ZInstI* inst, ZInstI* target, int target_slot);
void SetV(ZAMStmt s, const InstLabel l, int v) {
if ( v == 1 )
SetV1(s, l);
else if ( v == 2 )
SetV2(s, l);
else if ( v == 3 )
SetV3(s, l);
else
SetV4(s, l);
}
void SetV1(ZAMStmt s, const InstLabel l);
void SetV2(ZAMStmt s, const InstLabel l);
void SetV3(ZAMStmt s, const InstLabel l);
void SetV4(ZAMStmt s, const InstLabel l);
void SetGoTo(ZAMStmt s, const InstLabel targ) { SetV1(s, targ); }

View file

@ -51,14 +51,22 @@ public:
ZInstAux* aux; ZInstAux* aux;
}; };
// Most of the methods for the compiler are either in separate header source
// files, or in headers generated by auxil/gen-zam. We include these within
// the private part of the compiler class definitions, so a few methods that
// need to be public are specified here directly, rather than via such
// headers.
//
// We declare member variables here, rather than in included headers, since
// many of them are used across different source files, and don't necessarily
// have a natural "home".
class ZAMCompiler { class ZAMCompiler {
public: public:
ZAMCompiler(ScriptFuncPtr f, std::shared_ptr<ProfileFuncs> pfs, std::shared_ptr<ProfileFunc> pf, ScopePtr scope, ZAMCompiler(ScriptFuncPtr f, std::shared_ptr<ProfileFuncs> pfs, std::shared_ptr<ProfileFunc> pf, ScopePtr scope,
StmtPtr body, std::shared_ptr<UseDefs> ud, std::shared_ptr<Reducer> rd); StmtPtr body, std::shared_ptr<UseDefs> ud, std::shared_ptr<Reducer> rd);
~ZAMCompiler(); ~ZAMCompiler();
StmtPtr CompileBody();
const FrameReMap& FrameDenizens() const { return shared_frame_denizens_final; } const FrameReMap& FrameDenizens() const { return shared_frame_denizens_final; }
const std::vector<int>& ManagedSlots() const { return managed_slotsI; } const std::vector<int>& ManagedSlots() const { return managed_slotsI; }
@ -82,6 +90,8 @@ public:
return str_cases; return str_cases;
} }
StmtPtr CompileBody();
void Dump(); void Dump();
private: private:
@ -92,406 +102,27 @@ private:
friend class CatZBI; friend class CatZBI;
friend class MultiZBI; friend class MultiZBI;
void Init(); // The following are used for switch statements, mapping the switch value
void InitGlobals(); // (which can be any atomic type) to a branch target. We have vectors of
void InitArgs(); // them because functions can contain multiple switches.
void InitCaptures(); //
void InitLocals(); // See ZBody.h for their concrete counterparts, which we've already #include'd.
void TrackMemoryManagement();
void ResolveHookBreaks();
void ComputeLoopLevels();
void AdjustBranches();
void RetargetBranches();
void RemapFrameDenizens(const std::vector<int>& inst1_to_inst2);
void CreateSharedFrameDenizens();
void ConcretizeSwitches();
// The following are used for switch statements, mapping the
// switch value (which can be any atomic type) to a branch target.
// We have vectors of them because functions can contain multiple
// switches.
// See ZBody.h for their concrete counterparts, which we've
// already #include'd.
template<typename T> template<typename T>
using CaseMapI = std::map<T, InstLabel>; using CaseMapI = std::map<T, InstLabel>;
template<typename T> template<typename T>
using CaseMapsI = std::vector<CaseMapI<T>>; using CaseMapsI = std::vector<CaseMapI<T>>;
template<typename T> #include "zeek/script_opt/ZAM/AM-Opt.h"
void AdjustSwitchTables(CaseMapsI<T>& abstract_cases); #include "zeek/script_opt/ZAM/Branches.h"
#include "zeek/script_opt/ZAM/Driver.h"
template<typename T> #include "zeek/script_opt/ZAM/Expr.h"
void ConcretizeSwitchTables(const CaseMapsI<T>& abstract_cases, CaseMaps<T>& concrete_cases);
template<typename T>
void DumpCases(const CaseMaps<T>& cases, const char* type_name) const;
void DumpInsts1(const FrameReMap* remappings);
#include "zeek/ZAM-MethodDecls.h"
const ZAMStmt CompileStmt(const StmtPtr& body) { return CompileStmt(body.get()); }
const ZAMStmt CompileStmt(const Stmt* body);
const ZAMStmt CompilePrint(const PrintStmt* ps);
const ZAMStmt CompileExpr(const ExprStmt* es);
const ZAMStmt CompileIf(const IfStmt* is);
const ZAMStmt CompileSwitch(const SwitchStmt* sw);
const ZAMStmt CompileWhile(const WhileStmt* ws);
const ZAMStmt CompileFor(const ForStmt* f);
const ZAMStmt CompileReturn(const ReturnStmt* r);
const ZAMStmt CompileCatchReturn(const CatchReturnStmt* cr);
const ZAMStmt CompileStmts(const StmtList* sl);
const ZAMStmt CompileInit(const InitStmt* is);
const ZAMStmt CompileWhen(const WhenStmt* ws);
const ZAMStmt CompileNext() { return GenGoTo(nexts.back()); }
const ZAMStmt CompileBreak() { return GenGoTo(breaks.back()); }
const ZAMStmt CompileFallThrough() { return GenGoTo(fallthroughs.back()); }
const ZAMStmt CompileCatchReturn() { return GenGoTo(catches.back()); }
const ZAMStmt IfElse(const Expr* e, const Stmt* s1, const Stmt* s2);
const ZAMStmt While(const Stmt* cond_stmt, const Expr* cond, const Stmt* body);
const ZAMStmt InitRecord(IDPtr id, RecordType* rt);
const ZAMStmt InitVector(IDPtr id, VectorType* vt);
const ZAMStmt InitTable(IDPtr id, TableType* tt, Attributes* attrs);
const ZAMStmt ValueSwitch(const SwitchStmt* sw, const NameExpr* v, const ConstExpr* c);
const ZAMStmt TypeSwitch(const SwitchStmt* sw, const NameExpr* v, const ConstExpr* c);
const ZAMStmt GenSwitch(const SwitchStmt* sw, int slot, InternalTypeTag it);
void PushNexts() { PushGoTos(nexts); }
void PushBreaks() { PushGoTos(breaks); }
void PushFallThroughs() { PushGoTos(fallthroughs); }
void PushCatchReturns() { PushGoTos(catches); }
void ResolveNexts(const InstLabel l) { ResolveGoTos(nexts, l, CFT_NEXT); }
void ResolveBreaks(const InstLabel l) { ResolveGoTos(breaks, l, CFT_BREAK); }
void ResolveFallThroughs(const InstLabel l) { ResolveGoTos(fallthroughs, l); }
void ResolveCatchReturns(const InstLabel l) { ResolveGoTos(catches, l, CFT_INLINED_RETURN); }
const ZAMStmt LoopOverTable(const ForStmt* f, const NameExpr* val);
const ZAMStmt LoopOverVector(const ForStmt* f, const NameExpr* val);
const ZAMStmt LoopOverString(const ForStmt* f, const Expr* e);
const ZAMStmt FinishLoop(const ZAMStmt iter_head, ZInstI& iter_stmt, const Stmt* body, int iter_slot,
bool is_table);
const ZAMStmt Loop(const Stmt* body);
const ZAMStmt CompileExpr(const ExprPtr& e) { return CompileExpr(e.get()); }
const ZAMStmt CompileExpr(const Expr* body);
const ZAMStmt CompileIncrExpr(const IncrExpr* e);
const ZAMStmt CompileAppendToExpr(const AppendToExpr* e);
const ZAMStmt CompileAdd(const AggrAddExpr* e);
const ZAMStmt CompileDel(const AggrDelExpr* e);
const ZAMStmt CompileAddToExpr(const AddToExpr* e);
const ZAMStmt CompileRemoveFromExpr(const RemoveFromExpr* e);
const ZAMStmt CompileAssignExpr(const AssignExpr* e);
const ZAMStmt CompileRecFieldUpdates(const RecordFieldUpdatesExpr* e);
const ZAMStmt CompileZAMBuiltin(const NameExpr* lhs, const ScriptOptBuiltinExpr* zbi);
const ZAMStmt CompileAssignToIndex(const NameExpr* lhs, const IndexExpr* rhs);
const ZAMStmt CompileFieldLHSAssignExpr(const FieldLHSAssignExpr* e);
const ZAMStmt CompileScheduleExpr(const ScheduleExpr* e);
const ZAMStmt CompileSchedule(const NameExpr* n, const ConstExpr* c, int is_interval, EventHandler* h,
const ListExpr* l);
const ZAMStmt CompileEvent(EventHandler* h, const ListExpr* l);
const ZAMStmt CompileInExpr(const NameExpr* n1, const NameExpr* n2, const NameExpr* n3) {
return CompileInExpr(n1, n2, nullptr, n3, nullptr);
}
const ZAMStmt CompileInExpr(const NameExpr* n1, const NameExpr* n2, const ConstExpr* c) {
return CompileInExpr(n1, n2, nullptr, nullptr, c);
}
const ZAMStmt CompileInExpr(const NameExpr* n1, const ConstExpr* c, const NameExpr* n3) {
return CompileInExpr(n1, nullptr, c, n3, nullptr);
}
// In the following, one of n2 or c2 (likewise, n3/c3) will be nil.
const ZAMStmt CompileInExpr(const NameExpr* n1, const NameExpr* n2, const ConstExpr* c2, const NameExpr* n3,
const ConstExpr* c3);
const ZAMStmt CompileInExpr(const NameExpr* n1, const ListExpr* l, const NameExpr* n2) {
return CompileInExpr(n1, l, n2, nullptr);
}
const ZAMStmt CompileInExpr(const NameExpr* n, const ListExpr* l, const ConstExpr* c) {
return CompileInExpr(n, l, nullptr, c);
}
const ZAMStmt CompileInExpr(const NameExpr* n1, const ListExpr* l, const NameExpr* n2, const ConstExpr* c);
const ZAMStmt CompileIndex(const NameExpr* n1, const NameExpr* n2, const ListExpr* l, bool in_when);
const ZAMStmt CompileIndex(const NameExpr* n1, const ConstExpr* c, const ListExpr* l, bool in_when);
const ZAMStmt CompileIndex(const NameExpr* n1, int n2_slot, const TypePtr& n2_type, const ListExpr* l,
bool in_when);
const ZAMStmt BuildLambda(const NameExpr* n, ExprPtr le);
const ZAMStmt BuildLambda(int n_slot, ExprPtr le);
// Second argument is which instruction slot holds the branch target.
const ZAMStmt GenCond(const Expr* e, int& branch_v);
const ZAMStmt Call(const ExprStmt* e);
const ZAMStmt AssignToCall(const ExprStmt* e);
const ZAMStmt DoCall(const CallExpr* c, const NameExpr* n);
bool CheckForBuiltIn(const ExprPtr& e, CallExprPtr c);
const ZAMStmt AssignVecElems(const Expr* e);
const ZAMStmt AssignTableElem(const Expr* e);
const ZAMStmt ConstructTable(const NameExpr* n, const Expr* e);
const ZAMStmt ConstructSet(const NameExpr* n, const Expr* e);
const ZAMStmt ConstructRecord(const NameExpr* n, const Expr* e) { return ConstructRecord(n, e, false); }
const ZAMStmt ConstructRecordFromRecord(const NameExpr* n, const Expr* e) { return ConstructRecord(n, e, true); }
const ZAMStmt ConstructRecord(const NameExpr* n, const Expr* e, bool is_from_rec);
const ZAMStmt ConstructVector(const NameExpr* n, const Expr* e);
const ZAMStmt ArithCoerce(const NameExpr* n, const Expr* e);
const ZAMStmt RecordCoerce(const NameExpr* n, const Expr* e);
const ZAMStmt TableCoerce(const NameExpr* n, const Expr* e);
const ZAMStmt VectorCoerce(const NameExpr* n, const Expr* e);
const ZAMStmt Is(const NameExpr* n, const Expr* e);
#include "zeek/script_opt/ZAM/Inst-Gen.h" #include "zeek/script_opt/ZAM/Inst-Gen.h"
#include "zeek/script_opt/ZAM/Low-Level.h"
#include "zeek/script_opt/ZAM/Stmt.h"
#include "zeek/script_opt/ZAM/Vars.h"
int ConvertToInt(const Expr* e) { // Headers auto-generated by gen-zam.
if ( e->Tag() == EXPR_NAME ) #include "zeek/ZAM-MethodDecls.h"
return FrameSlot(e->AsNameExpr()->Id());
else
return e->AsConstExpr()->Value()->AsInt();
}
int ConvertToCount(const Expr* e) {
if ( e->Tag() == EXPR_NAME )
return FrameSlot(e->AsNameExpr()->Id());
else
return e->AsConstExpr()->Value()->AsCount();
}
using GoToSet = std::vector<ZAMStmt>;
using GoToSets = std::vector<GoToSet>;
void PushGoTos(GoToSets& gotos);
void ResolveGoTos(GoToSets& gotos, const InstLabel l, ControlFlowType cft = CFT_NONE);
ZAMStmt GenGoTo(GoToSet& v);
ZAMStmt GoToStub();
ZAMStmt GoTo(const InstLabel l);
InstLabel GoToTarget(const ZAMStmt s);
InstLabel GoToTargetBeyond(const ZAMStmt s);
void SetTarget(ZInstI* inst, const InstLabel l, int slot);
// Given a GoTo target, find its live equivalent (first instruction
// at that location or beyond that's live).
ZInstI* FindLiveTarget(ZInstI* goto_target);
// Given an instruction that has a slot associated with the
// given target, updates the slot to correspond with the current
// instruction number of the target.
void ConcretizeBranch(ZInstI* inst, ZInstI* target, int target_slot);
void SetV(ZAMStmt s, const InstLabel l, int v) {
if ( v == 1 )
SetV1(s, l);
else if ( v == 2 )
SetV2(s, l);
else if ( v == 3 )
SetV3(s, l);
else
SetV4(s, l);
}
void SetV1(ZAMStmt s, const InstLabel l);
void SetV2(ZAMStmt s, const InstLabel l);
void SetV3(ZAMStmt s, const InstLabel l);
void SetV4(ZAMStmt s, const InstLabel l);
void SetGoTo(ZAMStmt s, const InstLabel targ) { SetV1(s, targ); }
const ZAMStmt StartingBlock();
const ZAMStmt FinishBlock(const ZAMStmt start);
bool NullStmtOK() const;
const ZAMStmt EmptyStmt();
const ZAMStmt ErrorStmt();
const ZAMStmt LastInst();
// Adds control flow information to an instruction.
void AddCFT(ZInstI* inst, ControlFlowType cft);
// Returns a handle to state associated with building
// up a list of values.
std::unique_ptr<OpaqueVals> BuildVals(const ListExprPtr&);
// "stride" is how many slots each element of l will consume.
ZInstAux* InternalBuildVals(const ListExpr* l, int stride = 1);
// Returns how many values were added.
int InternalAddVal(ZInstAux* zi, int i, Expr* e);
// Adds the given instruction to the ZAM program. The second
// argument, if true, suppresses generation of any pending
// global/capture store for this instruction.
const ZAMStmt AddInst(const ZInstI& inst, bool suppress_non_local = false);
// Returns the statement just before the given one.
ZAMStmt PrevStmt(const ZAMStmt s);
// Returns the last (interpreter) statement in the body.
const Stmt* LastStmt(const Stmt* s) const;
// Returns the most recent added instruction *other* than those
// added for bookkeeping.
ZInstI* TopMainInst() { return insts1[top_main_inst]; }
bool IsUnused(const IDPtr& id, const Stmt* where) const;
bool IsCapture(const IDPtr& id) const { return IsCapture(id.get()); }
bool IsCapture(const ID* id) const;
int CaptureOffset(const IDPtr& id) const { return IsCapture(id.get()); }
int CaptureOffset(const ID* id) const;
void LoadParam(const ID* id);
const ZAMStmt LoadGlobal(const ID* id);
const ZAMStmt LoadCapture(const ID* id);
int AddToFrame(const ID*);
int FrameSlot(const IDPtr& id) { return FrameSlot(id.get()); }
int FrameSlot(const ID* id);
int FrameSlotIfName(const Expr* e) {
auto n = e->Tag() == EXPR_NAME ? e->AsNameExpr() : nullptr;
return n ? FrameSlot(n->Id()) : -1;
}
int FrameSlot(const NameExpr* id) { return FrameSlot(id->AsNameExpr()->Id()); }
int Frame1Slot(const NameExpr* id, ZOp op) { return Frame1Slot(id->AsNameExpr()->Id(), op); }
int Frame1Slot(const ID* id, ZOp op) { return Frame1Slot(id, op1_flavor[op]); }
int Frame1Slot(const NameExpr* n, ZAMOp1Flavor fl) { return Frame1Slot(n->Id(), fl); }
int Frame1Slot(const ID* id, ZAMOp1Flavor fl);
// The slot without doing any global-related checking.
int RawSlot(const NameExpr* n) { return RawSlot(n->Id()); }
int RawSlot(const ID* id);
bool HasFrameSlot(const ID* id) const;
int NewSlot(const TypePtr& t) { return NewSlot(ZVal::IsManagedType(t)); }
int NewSlot(bool is_managed);
int TempForConst(const ConstExpr* c);
////////////////////////////////////////////////////////////
// The following methods relate to optimizing the low-level
// ZAM function body after it is initially generated. They're
// factored out into ZOpt.cc since they're structurally quite
// different from the methods above that relate to the initial
// compilation.
// Optimizing the low-level compiled instructions.
void OptimizeInsts();
// Tracks which instructions can be branched to via the given
// set of switches.
template<typename T>
void TallySwitchTargets(const CaseMapsI<T>& switches);
// Remove code that can't be reached. True if some removal happened.
bool RemoveDeadCode();
// Collapse chains of gotos. True if some something changed.
bool CollapseGoTos();
// Prune statements that are unnecessary. True if something got
// pruned.
bool PruneUnused();
// For the current state of insts1, compute lifetimes of frame
// denizens (variable(s) using a given frame slot) in terms of
// first-instruction-to-last-instruction during which they're
// relevant, including consideration for loops.
void ComputeFrameLifetimes();
// Given final frame lifetime information, remaps frame members
// with non-overlapping lifetimes to share slots.
void ReMapFrame();
// Given final frame lifetime information, remaps slots in the
// interpreter frame. (No longer strictly necessary.)
void ReMapInterpreterFrame();
// Computes the remapping for a variable currently in the given slot,
// whose scope begins at the given instruction.
void ReMapVar(const ID* id, int slot, zeek_uint_t inst);
// Look to initialize the beginning of local lifetime based on slot
// assignment at instruction inst.
void CheckSlotAssignment(int slot, const ZInstI* inst);
// Track that a local's lifetime begins at the given statement.
void SetLifetimeStart(int slot, const ZInstI* inst);
// Look for extension of local lifetime based on slot usage
// at instruction inst.
void CheckSlotUse(int slot, const ZInstI* inst);
// Extend (or create) the end of a local's lifetime.
void ExtendLifetime(int slot, const ZInstI* inst);
// Returns the (live) instruction at the beginning/end of the loop(s)
// within which the given instruction lies; or that instruction
// itself if it's not inside a loop. The second argument specifies
// the loop depth. For example, a value of '2' means "extend to
// the beginning/end of any loop(s) of depth >= 2".
const ZInstI* BeginningOfLoop(const ZInstI* inst, int depth) const;
const ZInstI* EndOfLoop(const ZInstI* inst, int depth) const;
// True if any statement other than a frame sync uses the given slot.
bool VarIsUsed(int slot) const;
// Find the first non-dead instruction after i (inclusive).
// If follow_gotos is true, then if that instruction is
// an unconditional branch, continues the process until
// a different instruction is found (and report if there
// are infinite loops).
//
// First form returns nil if there's nothing live after i.
// Second form returns insts1.size() in that case.
ZInstI* FirstLiveInst(ZInstI* i, bool follow_gotos = false);
zeek_uint_t FirstLiveInst(zeek_uint_t i, bool follow_gotos = false);
// Same, but not including i.
ZInstI* NextLiveInst(ZInstI* i, bool follow_gotos = false) {
if ( i->inst_num == static_cast<int>(insts1.size()) - 1 )
return nullptr;
return FirstLiveInst(insts1[i->inst_num + 1], follow_gotos);
}
int NextLiveInst(int i, bool follow_gotos = false) { return FirstLiveInst(i + 1, follow_gotos); }
// Mark an instruction as unnecessary and remove its influence on
// other statements. The instruction is indicated as an offset
// into insts1; any labels associated with it are transferred
// to its next live successor, if any.
void KillInst(ZInstI* i) { KillInst(i->inst_num); }
void KillInst(zeek_uint_t i);
// Helper function for propagating control flow (of a given type)
// backwards, when the instruction at the given offset has been killed.
void BackPropagateCFT(int inst_num, ControlFlowType cf_type);
// The same, but kills any successor instructions until finding
// one that's labeled.
void KillInsts(ZInstI* i) { KillInsts(i->inst_num); }
void KillInsts(zeek_uint_t i);
// The first of these is used as we compile down to ZInstI's. // The first of these is used as we compile down to ZInstI's.
// The second is the final intermediary code. They're separate // The second is the final intermediary code. They're separate

View file

@ -0,0 +1,31 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for driving the overall ZAM compilation process.
//
// This file is included by Compile.h to insert into the ZAMCompiler class.
void Init();
void InitGlobals();
void InitArgs();
void InitCaptures();
void InitLocals();
void TrackMemoryManagement();
template<typename T>
void AdjustSwitchTables(CaseMapsI<T>& abstract_cases);
template<typename T>
void ConcretizeSwitchTables(const CaseMapsI<T>& abstract_cases, CaseMaps<T>& concrete_cases);
void ConcretizeSwitches();
void RetargetBranches();
void RemapFrameDenizens(const std::vector<int>& inst1_to_inst2);
void CreateSharedFrameDenizens();
void ResolveHookBreaks();
void ComputeLoopLevels();
void AdjustBranches();
template<typename T>
void DumpCases(const CaseMaps<T>& cases, const char* type_name) const;
void DumpInsts1(const FrameReMap* remappings);

79
src/script_opt/ZAM/Expr.h Normal file
View file

@ -0,0 +1,79 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for ZAM compilation of expression AST nodes (Expr's).
//
// This file is included by Compile.h to insert into the ZAMCompiler class.
const ZAMStmt CompileExpr(const ExprPtr& e) { return CompileExpr(e.get()); }
const ZAMStmt CompileExpr(const Expr* body);
const ZAMStmt CompileIncrExpr(const IncrExpr* e);
const ZAMStmt CompileAppendToExpr(const AppendToExpr* e);
const ZAMStmt CompileAdd(const AggrAddExpr* e);
const ZAMStmt CompileDel(const AggrDelExpr* e);
const ZAMStmt CompileAddToExpr(const AddToExpr* e);
const ZAMStmt CompileRemoveFromExpr(const RemoveFromExpr* e);
const ZAMStmt CompileAssignExpr(const AssignExpr* e);
const ZAMStmt CompileRecFieldUpdates(const RecordFieldUpdatesExpr* e);
const ZAMStmt CompileZAMBuiltin(const NameExpr* lhs, const ScriptOptBuiltinExpr* zbi);
const ZAMStmt CompileAssignToIndex(const NameExpr* lhs, const IndexExpr* rhs);
const ZAMStmt CompileFieldLHSAssignExpr(const FieldLHSAssignExpr* e);
const ZAMStmt CompileScheduleExpr(const ScheduleExpr* e);
const ZAMStmt CompileSchedule(const NameExpr* n, const ConstExpr* c, int is_interval, EventHandler* h,
const ListExpr* l);
const ZAMStmt CompileEvent(EventHandler* h, const ListExpr* l);
const ZAMStmt CompileInExpr(const NameExpr* n1, const NameExpr* n2, const NameExpr* n3) {
return CompileInExpr(n1, n2, nullptr, n3, nullptr);
}
const ZAMStmt CompileInExpr(const NameExpr* n1, const NameExpr* n2, const ConstExpr* c) {
return CompileInExpr(n1, n2, nullptr, nullptr, c);
}
const ZAMStmt CompileInExpr(const NameExpr* n1, const ConstExpr* c, const NameExpr* n3) {
return CompileInExpr(n1, nullptr, c, n3, nullptr);
}
// In the following, one of n2 or c2 (likewise, n3/c3) will be nil.
const ZAMStmt CompileInExpr(const NameExpr* n1, const NameExpr* n2, const ConstExpr* c2, const NameExpr* n3,
const ConstExpr* c3);
const ZAMStmt CompileInExpr(const NameExpr* n1, const ListExpr* l, const NameExpr* n2) {
return CompileInExpr(n1, l, n2, nullptr);
}
const ZAMStmt CompileInExpr(const NameExpr* n, const ListExpr* l, const ConstExpr* c) {
return CompileInExpr(n, l, nullptr, c);
}
const ZAMStmt CompileInExpr(const NameExpr* n1, const ListExpr* l, const NameExpr* n2, const ConstExpr* c);
const ZAMStmt CompileIndex(const NameExpr* n1, const NameExpr* n2, const ListExpr* l, bool in_when);
const ZAMStmt CompileIndex(const NameExpr* n1, const ConstExpr* c, const ListExpr* l, bool in_when);
const ZAMStmt CompileIndex(const NameExpr* n1, int n2_slot, const TypePtr& n2_type, const ListExpr* l, bool in_when);
const ZAMStmt BuildLambda(const NameExpr* n, ExprPtr le);
const ZAMStmt BuildLambda(int n_slot, ExprPtr le);
const ZAMStmt AssignVecElems(const Expr* e);
const ZAMStmt AssignTableElem(const Expr* e);
const ZAMStmt Call(const ExprStmt* e);
const ZAMStmt AssignToCall(const ExprStmt* e);
bool CheckForBuiltIn(const ExprPtr& e, CallExprPtr c);
const ZAMStmt DoCall(const CallExpr* c, const NameExpr* n);
const ZAMStmt ConstructTable(const NameExpr* n, const Expr* e);
const ZAMStmt ConstructSet(const NameExpr* n, const Expr* e);
const ZAMStmt ConstructRecord(const NameExpr* n, const Expr* e) { return ConstructRecord(n, e, false); }
const ZAMStmt ConstructRecordFromRecord(const NameExpr* n, const Expr* e) { return ConstructRecord(n, e, true); }
const ZAMStmt ConstructRecord(const NameExpr* n, const Expr* e, bool is_from_rec);
const ZAMStmt ConstructVector(const NameExpr* n, const Expr* e);
const ZAMStmt ArithCoerce(const NameExpr* n, const Expr* e);
const ZAMStmt RecordCoerce(const NameExpr* n, const Expr* e);
const ZAMStmt TableCoerce(const NameExpr* n, const Expr* e);
const ZAMStmt VectorCoerce(const NameExpr* n, const Expr* e);
const ZAMStmt Is(const NameExpr* n, const Expr* e);

View file

@ -4,8 +4,7 @@
// NameExpr*'s to slots. Some aren't needed, but we provide a complete // NameExpr*'s to slots. Some aren't needed, but we provide a complete
// set mirroring the ZInstI constructors for consistency. // set mirroring the ZInstI constructors for consistency.
// //
// Maintained separately from Compile.h to make it conceptually simple to // This file is included by Compile.h to insert into the ZAMCompiler class.
// add new helpers.
ZInstI GenInst(ZOp op); ZInstI GenInst(ZOp op);
ZInstI GenInst(ZOp op, const NameExpr* v1); ZInstI GenInst(ZOp op, const NameExpr* v1);

View file

@ -0,0 +1,42 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for low-level manipulation of ZAM instructions/statements.
//
// This file is included by Compile.h to insert into the ZAMCompiler class.
const ZAMStmt StartingBlock();
const ZAMStmt FinishBlock(const ZAMStmt start);
bool NullStmtOK() const;
const ZAMStmt EmptyStmt();
const ZAMStmt ErrorStmt();
const ZAMStmt LastInst();
// Adds control flow information to an instruction.
void AddCFT(ZInstI* inst, ControlFlowType cft);
// Returns a handle to state associated with building
// up a list of values.
std::unique_ptr<OpaqueVals> BuildVals(const ListExprPtr&);
// "stride" is how many slots each element of l will consume.
ZInstAux* InternalBuildVals(const ListExpr* l, int stride = 1);
// Returns how many values were added.
int InternalAddVal(ZInstAux* zi, int i, Expr* e);
// Adds the given instruction to the ZAM program. The second
// argument, if true, suppresses generation of any pending
// global/capture store for this instruction.
const ZAMStmt AddInst(const ZInstI& inst, bool suppress_non_local = false);
// Returns the statement just before the given one.
ZAMStmt PrevStmt(const ZAMStmt s);
// Returns the last (interpreter) statement in the body.
const Stmt* LastStmt(const Stmt* s) const;
// Returns the most recent added instruction *other* than those
// added for bookkeeping.
ZInstI* TopMainInst() { return insts1[top_main_inst]; }

48
src/script_opt/ZAM/Stmt.h Normal file
View file

@ -0,0 +1,48 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for ZAM compilation of statement AST nodes (Stmt's).
//
// This file is included by Compile.h to insert into the ZAMCompiler class.
// Note, we first list the AST nodes and then the helper functions, though
// in the definitions source these are intermingled.
const ZAMStmt CompileStmt(const StmtPtr& body) { return CompileStmt(body.get()); }
const ZAMStmt CompileStmt(const Stmt* body);
const ZAMStmt CompilePrint(const PrintStmt* ps);
const ZAMStmt CompileExpr(const ExprStmt* es);
const ZAMStmt CompileIf(const IfStmt* is);
const ZAMStmt CompileSwitch(const SwitchStmt* sw);
const ZAMStmt CompileWhile(const WhileStmt* ws);
const ZAMStmt CompileFor(const ForStmt* f);
const ZAMStmt CompileReturn(const ReturnStmt* r);
const ZAMStmt CompileCatchReturn(const CatchReturnStmt* cr);
const ZAMStmt CompileStmts(const StmtList* sl);
const ZAMStmt CompileInit(const InitStmt* is);
const ZAMStmt CompileWhen(const WhenStmt* ws);
const ZAMStmt CompileNext() { return GenGoTo(nexts.back()); }
const ZAMStmt CompileBreak() { return GenGoTo(breaks.back()); }
const ZAMStmt CompileFallThrough() { return GenGoTo(fallthroughs.back()); }
const ZAMStmt CompileCatchReturn() { return GenGoTo(catches.back()); }
const ZAMStmt IfElse(const Expr* e, const Stmt* s1, const Stmt* s2);
// Second argument is which instruction slot holds the branch target.
const ZAMStmt GenCond(const Expr* e, int& branch_v);
const ZAMStmt While(const Stmt* cond_stmt, const Expr* cond, const Stmt* body);
const ZAMStmt ValueSwitch(const SwitchStmt* sw, const NameExpr* v, const ConstExpr* c);
const ZAMStmt TypeSwitch(const SwitchStmt* sw, const NameExpr* v, const ConstExpr* c);
const ZAMStmt GenSwitch(const SwitchStmt* sw, int slot, InternalTypeTag it);
const ZAMStmt LoopOverTable(const ForStmt* f, const NameExpr* val);
const ZAMStmt LoopOverVector(const ForStmt* f, const NameExpr* val);
const ZAMStmt LoopOverString(const ForStmt* f, const Expr* e);
const ZAMStmt Loop(const Stmt* body);
const ZAMStmt FinishLoop(const ZAMStmt iter_head, ZInstI& iter_stmt, const Stmt* body, int iter_slot, bool is_table);
const ZAMStmt InitRecord(IDPtr id, RecordType* rt);
const ZAMStmt InitVector(IDPtr id, VectorType* vt);
const ZAMStmt InitTable(IDPtr id, TableType* tt, Attributes* attrs);

44
src/script_opt/ZAM/Vars.h Normal file
View file

@ -0,0 +1,44 @@
// See the file "COPYING" in the main distribution directory for copyright.
// Methods for managing Zeek function variables.
//
// This file is included by Compile.h to insert into the ZAMCompiler class.
bool IsUnused(const IDPtr& id, const Stmt* where) const;
bool IsCapture(const IDPtr& id) const { return IsCapture(id.get()); }
bool IsCapture(const ID* id) const;
int CaptureOffset(const IDPtr& id) const { return IsCapture(id.get()); }
int CaptureOffset(const ID* id) const;
void LoadParam(const ID* id);
const ZAMStmt LoadGlobal(const ID* id);
const ZAMStmt LoadCapture(const ID* id);
int AddToFrame(const ID*);
int FrameSlot(const IDPtr& id) { return FrameSlot(id.get()); }
int FrameSlot(const ID* id);
int FrameSlotIfName(const Expr* e) {
auto n = e->Tag() == EXPR_NAME ? e->AsNameExpr() : nullptr;
return n ? FrameSlot(n->Id()) : -1;
}
int FrameSlot(const NameExpr* id) { return FrameSlot(id->AsNameExpr()->Id()); }
int Frame1Slot(const NameExpr* id, ZOp op) { return Frame1Slot(id->AsNameExpr()->Id(), op); }
int Frame1Slot(const ID* id, ZOp op) { return Frame1Slot(id, op1_flavor[op]); }
int Frame1Slot(const NameExpr* n, ZAMOp1Flavor fl) { return Frame1Slot(n->Id(), fl); }
int Frame1Slot(const ID* id, ZAMOp1Flavor fl);
// The slot without doing any global-related checking.
int RawSlot(const NameExpr* n) { return RawSlot(n->Id()); }
int RawSlot(const ID* id);
bool HasFrameSlot(const ID* id) const;
int NewSlot(const TypePtr& t) { return NewSlot(ZVal::IsManagedType(t)); }
int NewSlot(bool is_managed);
int TempForConst(const ConstExpr* c);