entered via fall-through. Don't miss fallthroughs from blocks
terminated by conditional branches. Also, move
isOnlyReachableByFallthrough out of line.

llvm-svn: 68129

Does not work well on 32 bit targets. Bug reported by Albert Graef.

This patch also adds new "-Wllc,option" syntax to pass options to llc.

llvm-svn: 68127

llvm-svn: 68126

possible to run the tests with -asm-verbose defaulting
to false.

llvm-svn: 68124

Patch by Jay Foad.

llvm-svn: 68120

llvm::sys::getOS{Name,Version}.

Right now the implementation just derives from LLVM_HOSTTRIPLE (which
is wrong, but it doesn't look like we have a define for the target
triple). Ideally this routine would actually be able to compute the
triple for targets we care about.

llvm-svn: 68118

llvm-svn: 68113

llvm-svn: 68112

llvm-svn: 68111

llvm-svn: 68109

llvm-svn: 68100

llvm-svn: 68099

Remove the "fast" cases for spill and restore point determination, as these were subtlely wrong in obscure cases.  Patch the testcase
to account for this change.

llvm-svn: 68093

llvm-svn: 68092

llvm-svn: 68087

instructions. This fixes lua.

llvm-svn: 68083

X86 address mode isel tweak. If the base of the address is also used by a CopyToReg (i.e. it's likely live-out), do not fold the sub-expressions into the addressing mode to avoid computing the address twice. The CopyToReg use will be isel'ed to a LEA, re-use it for address instead.

This is not yet enabled.

llvm-svn: 68082

llvm-svn: 68079

only reachable via fall-through edges. This dramatically reduces the
number of labels printed, and thus also the number of labels the
assembler must parse and remember.

llvm-svn: 68073

Loop Index Split can eliminate a loop if it can determin if loop body is executed only once. There was a bug in determining IV based value of the iteration for which the loop body is executed. Fix it.

llvm-svn: 68071

llvm-svn: 68067

When optimzing a mul by immediate into two, the resulting mul's should get a x86 specific node to avoid dag combiner from hacking on them further.

llvm-svn: 68066

Turn a 2-address instruction into a 3-address one when it's profitable even if the two-address operand is killed.
e.g.
%reg1024<def> = MOV r1
%reg1025<def> = ADD %reg1024, %reg1026
r0            = MOV %reg1025

If it's not possible / profitable to commute ADD, then turning ADD into a LEA saves a copy.

llvm-svn: 68065

llvm-svn: 68060

llvm-svn: 68059

llvm-svn: 68057

anything pointer-like, which may or may not actually be a pointer.

llvm-svn: 68056

allow the traits to be specified as well.

llvm-svn: 68055

llvm-svn: 68054

valid argument attributes (zeroext and signext are).

llvm-svn: 68053

docs/TestingGuide.html: correction to prev. text (objdir!=srcdir required for running test-suite), removed refs to llvm-test

llvm-svn: 68051

llvm-svn: 68050

Walking an invalidated iterator is not a good idea.

llvm-svn: 68047

llvm-svn: 68038

llvm-svn: 68037

llvm-svn: 68036

llvm-svn: 68035

llvm-svn: 68034

llvm-svn: 68033

Do not propagate ELF-specific stuff (data.rel) into other targets. This simplifies code and also ensures correctness.

llvm-svn: 68032