aboutsummaryrefslogtreecommitdiff
path: root/lib/CodeGen
AgeCommit message (Collapse)Author
2011-07-13Convert the standard default-construction loops to use phis andJohn McCall
partial destruction. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@135033 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-13per john's advice, speculatively lower uses of forward-declared enums toChris Lattner
i32. They almost always end up this way in the end anyway, and if we get lucky, this avoids generating some bitcasts. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@135032 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-13PR10337 reminds me that calls return values, lets handle them justChris Lattner
like arguments. Thanks PR10337! :) git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@135030 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-13Aggressive dead code elimination.John McCall
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@135029 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-13Generalize the routine for destroying an object with staticJohn McCall
storage duration, then explicitly exempt ownership-qualified types from it. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@135028 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-13Switch delete[] IR-generation over to the destroy framework,John McCall
which implicitly makes it EH-safe as well. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@135025 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-13When compiling ::delete for a class with a virtual destructor, callDouglas Gregor
the complete destructor and then invoke the global delete operator. Previously, we would invoke the deleting destructor, which calls the wrong delete operator. Fixes PR10341. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@135021 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Revert r134946Bruno Cardoso Lopes
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@135004 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Generalize Cleanup::Emit's "isForEH" parameter into a setJohn McCall
of flags. No functionality change. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134997 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Add more compiler workarounds. Should fix the build with old GCCs and MSVC.Benjamin Kramer
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134995 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Ugh, use this compiler workaround again.John McCall
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134989 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Switch field destruction over to use the new destroyer-based APIJohn McCall
and kill a lot of redundant code. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134988 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Second attempt at de-constifying LLVM Types in FunctionType::get(),Jay Foad
StructType::get() and TargetData::getIntPtrType(). git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134982 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Work around a problem with a static helper's formulation in releaseChandler Carruth
builds introduced in r134972: lib/CodeGen/CGExpr.cpp:1294:7: error: no matching function for call to 'EmitBitCastOfLValueToProperType' lib/CodeGen/CGExpr.cpp:1278:1: note: candidate function not viable: no known conversion from 'CGBuilderTy' (aka 'IRBuilder<false>') to 'llvm::IRBuilder<> &' for 1st argument This fixes the issue by passing CodeGenFunction on down, and using its builder directly rather than passing just the builder down. This may not be the best / cleanest fix, Chris please review. It at least fixes builds. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134977 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Fix a problem Eli ran into where we now reject incomplete arrays of Chris Lattner
uncompleted struct types. We now do what llvm-gcc does and compile them into [i8 x 0]. If the type is later completed, we make sure that it is appropriately cast. We compile the terrible example to something like this now: %struct.A = type { i32, i32, i32 } @g = external global [0 x i8] define void @_Z1fv() nounwind { entry: call void @_Z3fooP1A(%struct.A* bitcast ([0 x i8]* @g to %struct.A*)) ret void } declare void @_Z3fooP1A(%struct.A*) define %struct.A* @_Z2f2v() nounwind { entry: ret %struct.A* getelementptr inbounds ([0 x %struct.A]* bitcast ([0 x i8]* @g to [0 x %struct.A]*), i32 0, i64 1) } git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134972 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12insert a bitcast in the 'expand' case of argument passing when needed. ThisChris Lattner
fixes the -m32 build of oggenc. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134971 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12fix an unintended behavior change in the type system rewrite, which caused ↵Chris Lattner
us to compile stuff like this: typedef struct { int x, y, z; } foo_t; foo_t g; into: %"struct.<anonymous>" = type { i32, i32, i32 } we now get: %struct.foo_t = type { i32, i32, i32 } This doesn't change the behavior of the compiler, but makes the IR much easier to read. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134969 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12fix PR10335 by watching out for IR type compatibility in call argument lists.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134966 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12add some aggressive assertions that call arguments match up to calleeChris Lattner
expectations in EmitCall. Next patch will take this further. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134962 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Fix a typo!Bruno Cardoso Lopes
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134952 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Do the same as r134946 for arrays. Add more testcases for avx x86_64 argBruno Cardoso Lopes
passing. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134951 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Revert r134888 (and related patches in other trees). It was causingBill Wendling
an assert on Darwin llvm-gcc builds. Assertion failed: (castIsValid(op, S, Ty) && "Invalid cast!"), function Create, file /Users/buildslave/zorg/buildbot/smooshlab/slave-0.8/build.llvm-gcc-i386-darwin9-RA/llvm.src/lib/VMCore/Instructions.cpp, line 2067. etc. http://smooshlab.apple.com:8013/builders/llvm-gcc-i386-darwin9-RA/builds/2354 --- Reverse-merging r134888 into '.': U lib/CodeGen/CodeGenModule.cpp git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134950 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Fix one x86_64 abi issue and the test to actually look for the right thing,Bruno Cardoso Lopes
which is: { <4 x float>, <4 x float> } should continue to go through memory. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134946 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-12Do full-expression cleanups in a much more sensible way that still letsJohn McCall
people write useful cleanup classes. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134942 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-11Reapply r134754, which turns out to be working correctly and alsoBruno Cardoso Lopes
add one more testcase. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134934 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-11Emit partial destruction of structs with initializer lists.John McCall
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134913 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-11De-constify Types in StructType::get() and TargetData::getIntPtrType().Jay Foad
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134893 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-11Fix a lot of problems with the partial destruction of arrays:John McCall
- an off-by-one error in emission of irregular array limits for InitListExprs - use an EH partial-destruction cleanup within the normal array-destruction cleanup - get the branch destinations right for the empty check Also some refactoring which unfortunately obscures these changes. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134890 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-11De-constify Types in FunctionType::get().Jay Foad
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134888 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10Use cached types; no functionality change.John McCall
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134880 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10implement a nice new optimization: CodeGenTypes::UpdateCompletedTypeChris Lattner
is called whenever a tag type is completed. We previously used that as the sign to layout the codegen representation for the tag type, which worked but meant that we laid out *every* completed type, whether it was used or not. Now we just lay out the type if we've already seen it somehow else. This means that we lay out types we've used but haven't seen a body for, but we don't lay out tons of stuff that noone cares about. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134866 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10enhance EmitLValueForFieldInitialization to do the proper pointer ↵Chris Lattner
adjustment, allowing us to revert the other half of r134860. Now things are back to a relatively tidy state. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134865 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10revert part of r134860, which is empirically unnecessary after the proper fixChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134864 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10keep track of whether being in a RS_StructPointer stateChris Lattner
caused us to skip layout out a function accurately. If so, flush the type cache for both the function and struct case to ensure that any pointers to the functions get recomputed. This is overconservative, but with this patch clang can build itself again. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134863 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10change EmitLValueForField to cast the returned lvalue to the rightChris Lattner
type, even when in the struct case. This was one root issue that was causing type mismatches throughout the compiler. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134862 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10Fix the clang bootstrap and Jay's testcase from llvm-dev by being completelyChris Lattner
conservative when converting a functiontype to IR when in a "pointer within a struct" context. This has the unfortunate sideeffect of compiling all function pointers inside of structs into "{}*" which, though correct, is ugly. This has the positive side effect of being correct, and it is pretty straight-forward to improve on this. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134861 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10when emitting pointer load from an lvalue or storing to an lvalue,Chris Lattner
do an explicit bitcast to whatever ConvertType produces. This will go with the next patch. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134860 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10how about we initialize RecursionState.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134855 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-10Rename CGT::VerifyFuncTypeComplete to isFuncTypeConvertible sinceChris Lattner
it is a predicate, not an action. Change the return type to be a bool, not the incomplete member. Enhace it to detect the recursive compilation case, allowing us to compile Eli's testcase on llvmdev: struct T { struct T (*p)(void); } t; into: %struct.T = type { {}* } @t = common global %struct.T zeroinitializer, align 8 git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134853 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-09when an enum type is completed, only flush the type cache whenChris Lattner
the enum has already been converted. If not, there cannot be any types built on top of it, so there is no need to flush the cache. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134841 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-09clang side to match the LLVM IR type system rewrite patch.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134831 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-09More compiler workarounds. I have to admit that I was notJohn McCall
expecting so much concentrated oddity on what seemed like a trivial feature. Thanks to François Pichet for doing the MSVC legwork here. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134813 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-09GCC 4.2 compatibility hack.John McCall
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134785 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-09A number of array-related IR-gen cleanups.John McCall
- Emit default-initialization of arrays that were partially initialized with initializer lists with a loop, rather than emitting the default initializer N times; - support destroying VLAs of non-trivial type, although this is not yet exposed to users; and - support the partial destruction of arrays initialized with initializer lists when an initializer throws an exception. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134784 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-08Change -mno-mmx to be more compatible with gcc. Specifically, -mno-mmx ↵Eli Friedman
should not imply -mno-sse. Note that because we don't usually touch the MMX registers anyway, all -mno-mmx needs to do is tweak the x86-32 calling convention a little for vectors that look like MMX vectors, and prevent the definition of __MMX__. clang doesn't actually stop the user from using MMX inline asm operands or MMX builtins in -mno-mmx mode; as a QOI issue, it would be nice to diagnose, but I doubt it really matters much. <rdar://problem/9694837> git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134770 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-08Revert x86_64 ABI changes until I have time to check the items raised by Eli.Bruno Cardoso Lopes
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134765 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-08Introduce __builtin_expect() intrinsic support.Jakub Staszak
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134761 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-08Add support for AVX 256-bit in the x86_64 ABI (as in the 0.99.5 draft)Bruno Cardoso Lopes
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134754 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-08Add codegen support for the fma/fmal/fmaf builtins.Cameron Zwarich
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134743 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-07Layout the code for trapping arithmetic so that the overflow case comes afterBill Wendling
the normal case. Before, for this: $ cat t.c int test(int x) { return x * 2; } We would get this: addl %edi, %edi jno LBB0_2 ## BB#1: ## %overflow ud2 LBB0_2: ## %nooverflow movl %edi, %eax popq %rbp ret Now we get this: addl %edi, %edi jo LBB0_2 ## BB#1: ## %nooverflow movl %edi, %eax popq %rbp ret LBB0_2: ## %overflow ud2 <rdar://problem/8283919> git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@134642 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-07-19 12:10:58 +0000