llvm-svn: 28505

vector arguments, not three.

llvm-svn: 28504

llvm-svn: 28503

llvm-svn: 28502

nonccc calls (we were dropping the CC and tail flag).  This broke several
FORTRAN programs.

Testcase here: Regression/Assembler/2006-05-26-VarargsCallEncode.ll

llvm-svn: 28501

llvm-svn: 28500

llvm-svn: 28499

llvm-svn: 28498

llvm-svn: 28497

llvm-svn: 28496

and documentation updates soo.

llvm-svn: 28495

Add lr as a hard coded operand of bx

llvm-svn: 28494

llvm-svn: 28493

llvm-svn: 28492

llvm-svn: 28491

llvm-svn: 28490

No functionality change.

llvm-svn: 28489

llvm-svn: 28488

the program.  This exposes more opportunities for the instcombiner, and implements
vec_shuffle.ll:test6

llvm-svn: 28487

llvm-svn: 28486

extractelement from the SV's source.  This implement vec_shuffle.ll:test[45]

llvm-svn: 28485

llvm-svn: 28484

llvm-svn: 28483

llvm-svn: 28482

llvm-svn: 28481

Patch contributed by Anton Korobeynikov!

llvm-svn: 28480

llvm-svn: 28479

array scopes possibly accessing valid memory in outer subscripts.

llvm-svn: 28478

llvm-svn: 28477

a cycle. This increase the search space and will increase compile time (in
practice it appears to be small, e.g. 176.gcc goes from 62 sec to 65 sec)
that will be addressed later.

llvm-svn: 28476

llvm-svn: 28475

llvm-svn: 28473

llvm-svn: 28471

llvm-svn: 28470

llvm-svn: 28469

llvm-svn: 28468

llvm-svn: 28467

llvm-svn: 28465

llvm-svn: 28464

the copyto/fromregs instead of making the X86ISD::CALL selection code create
them.

llvm-svn: 28463