aboutsummaryrefslogtreecommitdiff
path: root/lib/CodeGen/CGObjCMac.cpp
AgeCommit message (Collapse)Author
2009-05-04Don't allow clients to traverse into superclass synthesized propertiesDaniel Dunbar
via CollectObjCIvars. - In places where we need them, we should have the implementation and access the properties through it. This is a fairly substantial functionality change: 1. @encode no longer encodes synthesized ivars, ever. 2. The ivar layout bitmap no longer encodes information for synthesized ivars in superclasses. Well, actually I had already broken that, but it is intentional now. We are now differing substantially from llvm-gcc and gcc here. However, in my opinion this fundamentally *must* work if non-fragile classes are to work. Without this change, the result of @encode and the ivar layout depend on the order that the implementation is seen in a file (if it is in the same file with its superclass). Since both scenarios should work the same, our behavior is now consistent with gcc behavior as if an implementation is never seen following an implementation of its superclass. Note that #2 is only a functionality change when (A) an implementation appears in the same translation unit with the implementation of its superclass, and (B) the superclass has synthesized ivars. My belief is that this situation does not occur in practice. I am not yet sure of the role/semantics of @encode when synthesized ivars are present... it's use is fairly unsound in a non-fragile world. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70822 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Inline GetFieldBaseOffset into sole callsite.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70813 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Avoid recomputing field offsets.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70812 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Normalize formattingDaniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70810 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Use the implementation decl for looking up offset while building theDaniel Dunbar
ivar layout. - The layout needs access to synthesized ivars. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70798 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03It turns out BuildAggrIvarLayout wasn't even using the shadow struct,Daniel Dunbar
just computing it! git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70779 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Lift common subexpression, remove dead "base" variable.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70778 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Factor out BuildAggrIvarRecordLayout routine.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70777 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Lift out GetGCAttrTypeForType routine.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70776 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Add constructors for GC_IVAR and SKIP_SCAN, tighten up uses.Daniel Dunbar
Lift up a size calculation and note some asymmetries. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70775 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Normalize style, remove a dead assert.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70771 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Use ASTRecordLayout for computing ivar offsets instead of shadowDaniel Dunbar
struct. - We still need to do more lookup than necessary because ivars don't live in a reasonable DeclContext. - The only remaining client of the interface shadow struct is the ivar layout bitmap. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70756 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Add a ComputeIvarBaseOffset overload taking an implementationDaniel Dunbar
decl. Only this routine will be suitable for computing the offset of a synthesized ivar. - No functionality change. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70696 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Compute Objective-C metadata size information from the record layout,Daniel Dunbar
not the shadow structure. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70691 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Remove unused argument.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70684 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Coalesce the ivar offset calculation further.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70683 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-03Use type from ivar instead of from shadow struct field.Daniel Dunbar
- No functionality change. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70674 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-30Remove a warning when this file is compiled optimized.Fariborz Jahanian
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70518 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-30API for message dispatch of methods returning floatsFariborz Jahanian
to match gcc's closely. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70493 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-29Undid setting of the flag for msg_Send for 32bit code gen.Fariborz Jahanian
It seems to effect code gen. Add a FIXME instead. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70423 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-29Export lazy references of .objc_class_name of class namesFariborz Jahanian
referenced in a category implementation meta-data (Next objc 32bit abi). git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70407 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-29Type of msgSend message dispatch API is a vararg.Fariborz Jahanian
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70404 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-26Remove getIntegerConstantExprValue in favor of using EvaluateAsInt.Eli Friedman
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70145 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-26split ObjC and C++ Statements out into their own headers.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70105 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-24Minor refactoring. No intended change in behavior.Fariborz Jahanian
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69988 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-24Minor refactoring. No change in functionality.Fariborz Jahanian
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69979 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-24Some code clean up of objc2's bitmap layout.Fariborz Jahanian
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69970 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-23Removed bunch of FIXMEs no longer needed.Fariborz Jahanian
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69896 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-23Mark IMAGE_INFO as constant on x86_64-darwin.Daniel Dunbar
- This shouldn't change anything, we never actually access it, but this is consistent with llvm-gcc (and 32-bit) git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69880 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-23Use std::sort instead of qsort.Daniel Dunbar
- Notably, there was a memory error here, SkipIvars does not have to be the same size as IvarsInfo. - Fariborz, please check. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69850 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-23Eliminate the three SmallVectors in ObjCImplDecl (for instanceDouglas Gregor
methods, class methods, and property implementations) and instead place all of these entities into the DeclContext. This eliminates more linear walks when looking for class or instance methods and should make PCH (de-)serialization of ObjCDecls trivial (and lazy). git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69849 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22ivar layout bitmap is alive!Fariborz Jahanian
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69838 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22Reapply r69771, with updates & fixes:Daniel Dunbar
Rework the shadow struct that is layed out for Objective-C classes. - Superclasses are now always laid out in their shadow structure at the first field. - Prior to this, the entire class heirarchy was flattened into a single structure which meant that alignment, padding, and bitfields were incorrect (the ASTRecordLayout was correct however, which meant our debug info didn't coincide with ivar offsets, for example). - This is still very suboptimal (for example, ivar are looked up recursively, but I believe the ivar layout itself is now at least close to correct. - <rdar://problem/6773388> error: objc[29823]: layout bitmap sliding backwards git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69811 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22Remove lookupFieldDeclFromIvar from ObjCIvarDecl interface.Daniel Dunbar
- This is only used by CGObjCRuntime now. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69800 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22Don't convert interface types (to structs) as part of CodeGenTypes.Daniel Dunbar
- This has pros and cons, but for now the pros seem to significantly outway the con. The con is that we will always need to cast in the runtime implementation to a struct type, if we wish to access an interface directly. The pros are: - Avoid the cost of generating types which are used. Most manipulation of Objective-C objects is done through messages, and only the implementation of a class will directly access memory. Previously, we would convert the type even if it only appear as a function parameter, for example. - We don't need to worry about incomplete types, and UpdateCompletedType for interfaces is gone. - It becomes easier to narrow the interface to the shadow struct for Objective-C interfaces (so it can be eliminated). Currently the runtimes still use the CodeGenTypes machinery to generate the LLVM structure they need via ConvertTagDecl, but this can eventually be replaced. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69797 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22Add CGObjCRuntime::GetConcreteClassStruct to encapsulate access to theDaniel Dunbar
underlying llvm::StructType for an interface. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69796 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22Emit meta data using the Ivar, not a looked up FieldDecl.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69790 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22Merge ivar access amongst the three runtimes.Daniel Dunbar
- For now, this means we are always doing the address computations by hand instead of constructing a proper GEP. Right now, however, this is less important than having fewer entry points to dealing with Objective-C interface layout. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69787 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22Make ObjCInterfaceDecl's const in some more places.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69775 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22Revert r69771, I missed some (obvious) details. :/Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69773 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22Rework the shadow struct that is layed out for Objective-C classes.Daniel Dunbar
- Superclasses are now always laid out their shadow structure at the first field. - Prior to this, the entire class heirarchy was flattened into a single structure which meant that alignment, padding, and bitfields weren't packed correctly (the ASTRecordLayout was correct however, which meant our debug info didn't coincide with ivar offsets, for example). - This is still very suboptimal, but I believe the ivar layout itself is now at least close to correct. - <rdar://problem/6773388> error: objc[29823]: layout bitmap sliding backwards git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69771 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22move 64-bit abi functions to lazy model, everything is lazy now, yay.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69767 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22remove the last of the non-lazy objc runtime functions for the 32-bit ABI,Chris Lattner
7 left for 64-bit ABI. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69766 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22number of non-lazy runtime functions from 9 -> 4.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69765 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22make message send functions lazy, we're down from 14 non-lazy functions to 9.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69764 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22move more EH stuff to being lazily created. An empty .m file now Chris Lattner
produces just 14 dead "declares" in llvm ir instead of 19. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69763 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-22make try/catch objc runtime functions be lazily generated.Chris Lattner
rdar://6809612 git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69762 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-21Make sure to mark the interface as completed when we see anDaniel Dunbar
@implementation that closes a @class delcaration. - I don't know how to make a test case for this, but this strengthens the invariants that hold internally. The functionality change here is the edit to SemaDeclObjC.cpp. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69728 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-21More objc2's ivar layout bitmap. No change in functionality.Fariborz Jahanian
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69695 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-21Kill ASTContext::[gs]etFieldForDecl, instead we just lookup thingsDaniel Dunbar
when we need them -- which is exactly what some code was already doing! - No intended functionality change. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@69648 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-07-27 23:49:14 +0000