memory reference information in the backend. Most of this was written by
Florian Brander, cleanup and updating to TOT by me.

llvm-svn: 46556

list just to see if whether the list is empty.

llvm-svn: 46555

llvm-svn: 46554

llvm-svn: 46553

when in a register.

llvm-svn: 46552

llvm-svn: 46551

llvm-svn: 46550

know if all of this is exactly right, but it seems like the right thing to
aim for.  Steve, please review.

llvm-svn: 46549

llvm-svn: 46548

llvm-svn: 46547

llvm-svn: 46546

llvm-svn: 46545

- Expand tabs... (poss 80-col violations, will get them later...)
- Consolidate logic for SelectDFormAddr and SelectDForm2Addr into a single
  function, simplifying maintenance. Also reduced custom instruction
  generation for SPUvecinsert/INSERT_MASK.

llvm-svn: 46544

codegen of calls to functions without a prototype and varargs functions, 
including printf.

llvm-svn: 46543

Make DSE much more aggressive by performing DCE earlier.  Update a testcase to reflect this increased aggressiveness.

llvm-svn: 46542

of one of its internal maps.

llvm-svn: 46541

Many refinements to Sema::MergeVarDecl(). This routine still needs some re-work to fully handle tentative decls.

This includes a fix to bz1908.

llvm-svn: 46540

In practice this can only happen on code with already undefined behavior, 
but this is still a good thing to handle correctly.

llvm-svn: 46539

and StoreSDNode into their common base class LSBaseSDNode. Member
functions getLoadedVT and getStoredVT are replaced with the common
getMemoryVT to simplify code that will handle both loads and stores.

llvm-svn: 46538

auto-propagation of state along a branch.

llvm-svn: 46537

we automatically generate a new successor node along an edge if the checker
did not explicitly do so (i.e., we just propagate the current state).

llvm-svn: 46536

llvm-svn: 46535

llvm-svn: 46534

Rewriting of @synchronized. This has one FIXME in it. But this should allow @sychronized to be rewritten.

llvm-svn: 46533

llvm-svn: 46532

llvm-svn: 46531

have been 'true'.

llvm-svn: 46530

- Some minor optimizations of the encoding format for ObjCAtSynchronizedStmt.
- Reordered methods so that the deserialization methods for 
  ObjCAtSynchronizedStmt appear in alphabetical order with the corresponding
  methods of the other AST nodes.

llvm-svn: 46529

Made some internal API cleanups with creating ConcreteInt values, which lead to
some code reduction in implementing "++", "--".

llvm-svn: 46528

Work in progress. This patch *fixes* x86-64 calls which are modelled as StructRet but really should be return in registers, e.g. _Complex long double, some 128-bit aggregates. This is a short term solution that is necessary only because llvm, for now, cannot model i128 nor call's with multiple results.
Status: This only works for direct calls, and only the caller side is done. Disabled for now.

llvm-svn: 46527

llvm-svn: 46526

type that matters but the operand type.  This fixes
2008-01-08-IllegalCMP.ll which crashed with the new
legalize infrastructure because SETCC with result
type i8 and operand type i64 was being custom expanded
by the X86 backend.  With this fix, the gcc build gets
as far as the first libcall.

llvm-svn: 46525

llvm-svn: 46524

llvm-svn: 46523

Tighten up ASTContext::typesAreCompatible()...it needs to make sure the qualifiers match. The comment and C99 citation for this routine were correct...the code needed to conform to the comment/spec. This fixes the test added below.

Tightening up this routine forced tweaks to Sema::CheckSubtractionOperands() and Sema::CheckCompareOperands(). For example, they both need to operate on the unqualified pointee...

llvm-svn: 46522

behavior is unchanged.

llvm-svn: 46521

llvm-svn: 46520

already used.

llvm-svn: 46519

for analyzing a function.

The initial state for GRConstants now assigns symbolic values to parameters.

llvm-svn: 46517

llvm-svn: 46515