current SDep.

llvm-svn: 61064

llvm-svn: 61063

llvm-svn: 61062

llvm-svn: 61061

Make name lookup when we're inside a declarator's scope, such as ClassName::func, work with the new unqualified name lookup code. Test it with default arguments in out-of-line member definitions

llvm-svn: 61060

functions. They work except that name lookup within the default
arguments needs to be deferred until the class definition is complete
(see FIXME in the test).

llvm-svn: 61059

specifiers. Specifically: 
  * Determine when an out-of-line function definition does not match
    any declaration within the class or namespace (including coping
    with overloaded functions).
  * Complain about typedefs and parameters that have scope specifiers.
  * Complain about out-of-line declarations that aren't also
  definitions.
  * Complain about non-static data members being declared out-of-line.
  * Allow cv-qualifiers on out-of-line member function definitions.

llvm-svn: 61058

programs.

llvm-svn: 61056

llvm-svn: 61055

is done. Layout is calculated lazily at code gen type.
This patch changes the name.

llvm-svn: 61054

llvm-svn: 61052

variable sized array allocations.

llvm-svn: 61051

llvm-svn: 61050

just like all other members, and remove the special variables in
CXXRecordDecl to store them. This eliminates a lot of special-case
code for constructors and destructors, including
ActOnConstructor/ActOnDeclarator and special lookup rules in
LookupDecl. The result is far more uniform and manageable.

Diagnose the redeclaration of member functions.

llvm-svn: 61048

llvm-svn: 61047

llvm-svn: 61046

Use GetElementPtrInst::hasAllZeroIndices where possible.

llvm-svn: 61045

llvm-svn: 61044

ivars.

llvm-svn: 61043

Preprocessor: Allocate MacroInfo objects using a BumpPtrAllocator instead using new/delete.  This speeds up -Eonly on Cocoa.h using the regular lexer by 1.8% and the PTHLexer by 3%.

llvm-svn: 61042

llvm-svn: 61041

intrinsics are properly marked nocapture, the fixme should be addressed.

llvm-svn: 61040

llvm-svn: 61039

llvm-svn: 61038

half-solution; the real solution is coming when constructors and
destructors are treated like all other functions by ActOnDeclarator.

llvm-svn: 61037

llvm-svn: 61036

llvm-svn: 61035

together in the same way that we link RecordDecl/CXXRecordDecl nodes. 

Unify ActOnTag and ActOnTagStruct.

Fixes PR clang/2753.

llvm-svn: 61034

llvm-svn: 61033

comments about why we're not getting other cases.

llvm-svn: 61032

llvm-svn: 61031

alignment attribute such that 0 means unaligned.

This will probably require a rebuild of llvm-gcc because of the change to
Attributes.h. If you see many test failures on "make check", please rebuild
your llvm-gcc.

llvm-svn: 61030

and insert vector element.  Modified extract vector element to extend the
result to match the expected promoted type.

llvm-svn: 61029

llvm-svn: 61028

CFG when there is exactly one predecessor where the load is not available.
This is designed to not increase code size but still eliminate partially
redundant loads.  This fires 1765 times on 403.gcc even though it doesn't
do critical edge splitting yet (the most common reason for it to fail).

llvm-svn: 61027

return *just* a clobber of the start block, not other 
random stuff as well.

llvm-svn: 61026

unbreaking it after r61024.

llvm-svn: 61025

llvm-svn: 61024

cleans up the generated code a bit.  This should have the added benefit of
not randomly renaming functions/globals like my previous patch did. :)

llvm-svn: 61023

memdep keeps track of how PHIs affect the pointer in dep queries, which 
allows it to eliminate the load in cases like rle-phi-translate.ll, which
basically end up being:

BB1:
   X = load P
   br BB3
BB2:
   Y = load Q
   br BB3
BB3:
   R = phi [P] [Q]
   load R

turning "load R" into a phi of X/Y.  In addition to additional exposed
opportunities, this makes memdep safe in many cases that it wasn't before
(which is required for load PRE) and also makes it substantially more 
efficient.  For example, consider:


bb1:  // has many predecessors.
   P = some_operator()
   load P

In this example, previously memdep would scan all the predecessors of BB1
to see if they had something that would mustalias P.  In some cases (e.g.
test/Transforms/GVN/rle-must-alias.ll) it would actually find them and end
up eliminating something.  In many other cases though, it would scan and not
find anything useful.  MemDep now stops at a block if the pointer is defined
in that block and cannot be phi translated to predecessors.  This causes it
to miss the (rare) cases like rle-must-alias.ll, but makes it faster by not
scanning tons of stuff that is unlikely to be useful.  For example, this
speeds up GVN as a whole from 3.928s to 2.448s (60%)!.  IMO, scalar GVN 
should be enhanced to simplify the rle-must-alias pointer base anyway, which
would allow the loads to be eliminated.

In the future, this should be enhanced to phi translate through geps and 
bitcasts as well (as indicated by FIXMEs) making memdep even more powerful.

llvm-svn: 61022