aboutsummaryrefslogtreecommitdiff
path: root/lib/CodeGen/CGStmt.cpp
AgeCommit message (Collapse)Author
2010-05-23PR5863: Don't erase unreachable BBs which have an associated cleanup size.Benjamin Kramer
This works around a crash where malloc reused the memory of an erased BB for a new BB leaving old cleanup information pointing at the new block. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@104472 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-21Don't remove the break/continue scope of a for loop until after we'veDouglas Gregor
emitted the increment expression. Fixes PR7189. If someone knows how to write a useful test for this, I'd be grateful. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@104335 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-17Ensure that destructors are called for NRVO'd objects when theDouglas Gregor
function does not return. Thanks to Eli for pointing out this corner case. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@103941 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-15Implement a simple form of the C++ named return value optimization forDouglas Gregor
return statements. We perform NRVO only when all of the return statements in the function return the same variable. Fixes some link failures in Boost.Interprocess (which is relying on NRVO), and probably improves performance for some C++ applications. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@103867 91177308-0d34-0410-b5e6-96231b3b80d8
2010-05-12Use end location of DeclStmt to mark stop point. Devang Patel
This is meaningful for blocks. This patch fixes bunch of test failures in gdb testsuite. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@103533 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-24Fix a place in inline asm lowering which was creating a TruncInst with aDan Gohman
pointer operand. This fixes an abort on MultiSource/Applications/ClamAV/libclamav_mbox.c. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@102238 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-23Implement PR6845. We allow matching constraints to have differentChris Lattner
input and output types when the smaller value isn't mentioned in the asm string. Extend this support from integers to also allowing fp values to be mismatched (if not mentioned in the asm string). git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@102188 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-21I failed to notice that my last patch wasn't doing as much as it couldJohn McCall
because EmitBranch actually clears the insert point. This version actually accomplishes what I initially wanted. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@101998 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-21Teach EmitBlock to put the target block immediately after the current blockJohn McCall
(if there's a current block). The chief advantage of doing this is that it lets us pick blocks (e.g. EH blocks) to push to the end of the function so that fallthrough happens consistently --- i.e. it gives us the flexibility of ordering blocks as we please without having to change the order in which we generate code. There are standard (?) optimization passes which can do some of that for us, but better to generate reasonable code to begin with. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@101997 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-13Fix debug info for cleanup block.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@101100 91177308-0d34-0410-b5e6-96231b3b80d8
2010-04-07a ridiculous amount of propagation through the backend later, Chris Lattner
have the code generate slap a srcloc metadata on inline asm nodes. This allows us to diagnose invalid inline asms with such nice diagnostics as: <inline asm>:1:2: error: unrecognized instruction abc incl %eax ^ asm.c:2:12: note: generated from here __asm__ ("abc incl %0" : "+r" (X)); ^ 2 diagnostics generated. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@100608 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-24When returning from a function that has a reference return type, useDouglas Gregor
EmitReferenceBindingToExpr() rather than assuming we have an lvalue. This is just the lowest hanging fruit for PR6024, which still requires a bit of work. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@99447 91177308-0d34-0410-b5e6-96231b3b80d8
2010-03-03fix PR6475, we were doing side-effecting stuff in an assert.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@97669 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-08Reapply r95393, without the change to CGExpr. I was wrong in assuming that theDaniel Dunbar
element type always matched the converted LLVM type for ExprType. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@95596 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-05Revert r95393, which broke Clang's self-host.Douglas Gregor
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@95430 91177308-0d34-0410-b5e6-96231b3b80d8
2010-02-05IRgen: Use hasAggregateLLVMType instead of isSingleValueType() for cases thatDaniel Dunbar
need to deal with aggregates specially; this is consistent with the rest of IRgen. Also, simplify EmitParmDecl and don't worry about using Decl::getNameAsString. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@95393 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-30More asm cleanup.Anders Carlsson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@94920 91177308-0d34-0410-b5e6-96231b3b80d8
2010-01-30Yay for more StringRefs.Anders Carlsson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@94917 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-25Implement proper cleanup semantics for condition variables in forDouglas Gregor
statements, e.g., for(; X x = X(); ) { ... } Daniel or Anders, please review! git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@89832 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-25Eliminate CXXConditionDeclExpr with extreme prejudice.Douglas Gregor
All statements that involve conditions can now hold on to a separate condition declaration (a VarDecl), and will use a DeclRefExpr referring to that VarDecl for the condition expression. ForStmts now have such a VarDecl (I'd missed those in previous commits). Also, since this change reworks the Action interface for if/while/switch/for, use FullExprArg for the full expressions in those expressions, to ensure that we're emitting Note that we are (still) not generating the right cleanups for condition variables in for statements. That will be a follow-on commit. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@89817 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-24Clean up the AST for while loops and fix several problems withDouglas Gregor
cleanups for while loops: 1) Make sure that we destroy the condition variable of a while statement each time through the loop for, e.g., while (shared_ptr<WorkInt> p = getWorkItem()) { // ... } 2) Make sure that we always enter a new cleanup scope for the body of the while loop, even when there is no compound expression, e.g., while (blah) RAIIObject raii(blah+1); git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@89800 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-24Explicitly store the condition variable within switch statements, andDouglas Gregor
make sure that this variable is destroyed when we exit the switch statement. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@89776 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-24Introduce cleanup scopes for "if" statements in two places:Douglas Gregor
- Outside the "if", to ensure that we destroy the condition variable at the end of the "if" statement rather than at the end of the block containing the "if" statement. - Inside the "then" and "else" branches, so that we emit then- or else-local cleanups at the end of the corresponding block when the block is not a compound statement. To make adding these new cleanup scopes easier (and since switch/do/while will all need the same treatment), added the CleanupScope RAII object to introduce a new cleanup scope and make sure it gets cleaned up. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@89773 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-23Explicitly track the condition variable within an "if" statement,Douglas Gregor
rather than burying it in a CXXConditionDeclExpr (that occassionally hides behind implicit conversions). Similar changes for switch, while, and do-while will follow, then the removal of CXXConditionDeclExpr. This commit is the canary. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@89717 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-13Wherein the TargetInfo argument to Preprocessor is made 'const' and propogated.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@87087 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-12 "Attach debug info with llvm instructions" mode was enabled a month ago. ↵Devang Patel
Now make it permanent and remove old way of inserting intrinsics to encode debug info for locations and types. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@87007 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-06update comment, pointed out by GaborChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@86263 91177308-0d34-0410-b5e6-96231b3b80d8
2009-11-06indirectbr seems to work! Rip out the old code.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@86256 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-28Implement clang support for indirect branch and address of labelChris Lattner
using the new LLVM support for this. This is temporarily hiding behind horrible and ugly #ifdefs until the time when the optimizer is stable (hopefully a week or so). Until then, lets make it "opt in" :) git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@85446 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-13reimplement codegen for indirect goto with the following advantages:Chris Lattner
1. CGF now has fewer bytes of state (one pointer instead of a vector). 2. The generated code is determinstic, instead of getting labels in 'map order' based on pointer addresses. 3. Clang now emits one 'indirect goto switch' for each function, instead of one for each indirect goto. This fixes an M*N = N^2 IR size issue when there are lots of address-taken labels and lots of indirect gotos. 4. This also makes the default cause do something useful, reducing the size of the jump table needed (by one). git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@83952 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-06Add support to attach debug info to an instruction.Devang Patel
This is not yet enabled. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@83399 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-27CodeGen for try statements. (We just emit the body for now).Anders Carlsson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@82910 91177308-0d34-0410-b5e6-96231b3b80d8
2009-09-09Remove tabs, and whitespace cleanups.Mike Stump
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@81346 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-13Update for LLVM API change.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@78946 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-05Update for LLVM API change.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@78259 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-04Revert the fix for PR3800, it broke things.Anders Carlsson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@78084 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-02Remove old code that was causing the new test to fail.Anders Carlsson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@77862 91177308-0d34-0410-b5e6-96231b3b80d8
2009-08-02Don't evaluate inout constraints twice. Fixes PR3800.Anders Carlsson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@77854 91177308-0d34-0410-b5e6-96231b3b80d8
2009-07-30Canonicalize else spacing.Mike Stump
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@77629 91177308-0d34-0410-b5e6-96231b3b80d8
2009-07-24Update for LLVM API change.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@77012 91177308-0d34-0410-b5e6-96231b3b80d8
2009-07-19Detect when the current generation point is unreachable after emittingDaniel Dunbar
expressions. - This generally catches the important case of noreturn functions. - With the last two changes, we are down to 152 unreachable blocks emitted on 403.gcc, vs the 1805 we started with. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@76364 91177308-0d34-0410-b5e6-96231b3b80d8
2009-07-19Fix thinko.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@76362 91177308-0d34-0410-b5e6-96231b3b80d8
2009-07-19Avoid generation of dead code in a few more situations.Daniel Dunbar
- Emit variable declarations as "simple", we want to avoid forcing the creation of a dummy basic block, but still need to make the variable available for later use. - With that, we can now skip IRgen for other unreachable statements (which don't define a label). - Anders, I added two fixmes on calls to EmitVLASize, can you check them? git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@76361 91177308-0d34-0410-b5e6-96231b3b80d8
2009-07-16Update for LLVM API change.Owen Anderson
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@76090 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-27Add IRGen support for return statements in functions with reference Eli Friedman
type. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@72459 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-26Fixup codegen for volatile structs in the trivial cases (a a=a and a=a=a).Mike Stump
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@72439 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-21Use v.data() instead of &v[0] when SmallVector v might be empty.Jay Foad
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@72210 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-04"Fix" a problem with debug info in the presence of always_inlineChris Lattner
function calls. For a program like this: #include <stdio.h> static __inline__ __attribute__((always_inline)) int bar(int x) { return 4; } int main() { int X = bar(4); printf("%d\n", X); } clang was not outputing any debug info for the body of main(). This is because the backend is getting confused by the region_start/end that clang is emitting for block scopes. For now, just disable these (matching llvm-gcc), this stuff is in progress of rework anyway. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70889 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-04Remove unnecessary copy of constraint info.Daniel Dunbar
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70835 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-04Remove unnecessary push_back (at least, I think it's unnecessary); Eli Friedman
hopefully, this fixes PR4144 without any regressions. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@70823 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-05-06 08:29:02 +0000