We verify that the LiveInterval is live at uses and defs, and that all
instructions have a SlotIndex.

Stuff we don't check yet:

- Is the LiveInterval minimal?
- Do all defs correspond to instructions or phis?
- Do all defs dominate all their live ranges?
- Are all live ranges continually reachable from their def?

llvm-svn: 110386

llvm-svn: 110385

llvm-svn: 110384

Split the tag and value members of LVILatticeVal in preparation for expanding the lattice to something that won't fit in two bits.

llvm-svn: 110383

response from getModRefInfo is not useful here. Instead, check for identical
calls only in the NoModRef case.

Reapply r110270, and strengthen it to compensate for the memdep changes.
When both calls are readonly, there is no dependence between them.

llvm-svn: 110382

these, but it's convenient to mangle them when deferring them (in the 99.99%
case where it's not an anonymous union, of course).

llvm-svn: 110381

as a positive consequence the CallSite::getCallee() methods now can be rewritten to be
a bit more efficient

llvm-svn: 110380

because unittest.main() calls sys.exit() before returning.  Fixed by registering
an exit handler for this situation.

llvm-svn: 110379

llvm-svn: 110378

(objc gc and blocks in NeXt runtime).

llvm-svn: 110377

llvm-svn: 110376

do the right thing with mixed-visibility symbols, so disable the visibility
optimization where that's possible, i.e. with template classes (since it's
possible that an arbitrary template might be subject to an explicit
instantiation elsewhere).  447.dealII actually does this.

I've put the code under an option that's currently not hooked up to anything.

llvm-svn: 110374

llvm-svn: 110373

llvm-svn: 110372

llvm-svn: 110371

llvm-svn: 110370

llvm-svn: 110369

register for local access when it's closer to the stack slot being refererenced
than the stack pointer. Make sure to take into account any argument frame
SP adjustments that are in affect at the time.

rdar://8256090

llvm-svn: 110366

llvm-svn: 110363

be killed before being redefined.

These checks are usually disabled, and usually fail when enabled. We de facto
allow live registers to be redefined without a kill, the corresponding
assertions in RegScavenger were removed long ago.

llvm-svn: 110362

Partial fix for PR7792.

llvm-svn: 110361

instructions.

llvm-svn: 110360

llvm-svn: 110359

Partial fix for PR7792.

llvm-svn: 110358

llvm-svn: 110357

we sometimes compare singular iterators, reported by ENABLE_EXPENSIVE_CHECKS.
This fixes PR7825.

llvm-svn: 110355

llvm-svn: 110354

Fixed logic error in UnreachableCodeChecker's marking algorithm that would sometimes allow for multiple sequential statements to be flagged.

llvm-svn: 110353

llvm-svn: 110347

Patch by Francois Pichet!

llvm-svn: 110344

into test/CodeGen/X86, so that they aren't run when the x86 target is
not enabled.

Fix uglygep.ll to not be x86-specific.

llvm-svn: 110343

--- Reverse-merging r110333 into '.':
U    tools/bugpoint/BugDriver.h
U    tools/bugpoint/OptimizerDriver.cpp
U    tools/bugpoint/bugpoint.cpp
U    tools/bugpoint/BugDriver.cpp

llvm-svn: 110341

(objc gc specific).

llvm-svn: 110340

llvm-svn: 110339

llvm-svn: 110338

rely on using a specific x86 triple to test what they want to test.

llvm-svn: 110337

Fixes PR753.

llvm-svn: 110333

llvm-svn: 110330

this breaks something.

I'll wait a few days before cleaning out UnaryOperator::OffsetOf.

llvm-svn: 110328

offsetof expressions.

llvm-svn: 110327