llvm-svn: 127913

The llvm.dbg.value intrinsic refers to SSA values, not virtual registers, so we
should be able to extend the range of a value by tracking that value through
register copies. This greatly improves the debug value tracking for function
arguments that for some reason are copied to a second virtual register at the
end of the entry block.

We only extend the debug value range where its register is killed. All original
llvm.dbg.value locations are still respected.

Copies from physical registers are ignored. That should not be a problem since
the entry block already adds DBG_VALUE instructions for the virtual registers
holding the function arguments.

llvm-svn: 127912

llvm-svn: 127909

'ar'. Have to figure out how to make libLTO even lazier.

llvm-svn: 127901

llvm-svn: 127900

llvm-svn: 127899

- Emit mad instead of mad.rn for shader model 1.0
- Emit explicit mov.u32 instructions for reading global variables
- (most PTX instructions cannot take global variable immediates)

llvm-svn: 127895

Factor out the 64-bit specific bits into a helper function and add an
equivalent that loads the 32-bit sections. This allows using llvm-rtdyld on ARM.

llvm-svn: 127892

llvm-svn: 127891

llvm-svn: 127888

llvm-svn: 127887

llvm-svn: 127886

Add a bone-simple utility to load a MachO object into memory, look for
a function (main) in it, and run that function directly. This will be used
as a test and development platform for MC-JIT work regarding symbol resolution,
dynamic lookup, etc..

Code by Daniel Dunbar.

llvm-svn: 127885

For example, on 32-bit architecture, don't promote all uses of the IV
to 64-bits just because one use is a 64-bit cast.
Alternate implementation of the patch by Arnaud de Grandmaison.

llvm-svn: 127884

llvm-svn: 127883

For now, only the default segments are supported.

llvm-svn: 127875

llvm-svn: 127874

llvm-svn: 127873

On MSVCRT and compatible, output of %e is incompatible to Posix by default. Number of exponent digits should be at least 2. "%+03d"

FIXME: Implement our formatter in future!
llvm-svn: 127872

makes valgrind stop complaining about uninitialized variables being read when it
accesses a bitfield (category) that shares its bits with these variables.

llvm-svn: 127871

functions and initializers, just report the declarations present in
the module.

The motivation is to open the way for using the lazy module parsing,
which should speed up clients that just want a symbol list (nm, ar).

This is slightly less precise, but since both -strip-dead-prototypes
and -globaldce are part of the standard pipeline, this shouldn't
change the result for clang/dragonegg produced binaries.

Any decl in an IL file was also put there because a FE expected it
to be necessary, so this should not be a problem for "-O0 -emit-llvm".

As a sanity check, I have bootstrapped clang on linux and built
firefox on both linux and darwin. A clang bootstrap on darwin
with LTO fails with or without this patch because, ironically,
the linker doesn't like the combination of dead_strip and LTO
when building libLTO.so :-)

llvm-svn: 127870

Stack slot real estate is virtually free compared to registers, so it is
advantageous to spill earlier even though the same value is now kept in both a
register and a stack slot.

Also eliminate redundant spills by extending the stack slot live range
underneath reloaded registers.

This can trigger a dead code elimination, removing copies and even reloads that
were only feeding spills.

llvm-svn: 127868

llvm-svn: 127867

Tweak CrashRecoveryContextCleanup::createCleanup() to use the 'delete' cleanup as opposed to the 'destructor' cleanup (reclaims more memory).

llvm-svn: 127865

llvm-svn: 127859

llvm-svn: 127858

This is not supposed to happen, but I have seen the x86 rematter getting
confused when rematerializing partial redefs.

llvm-svn: 127857

and early clobbers.

Assert when trying to find an undefined value.

llvm-svn: 127856

Add new CrashRecoveryContextCleanup subclass: CrashRecoveryContextDeleteCleanup.  This deletes the object, not just calls its destructor.

llvm-svn: 127855

llvm-svn: 127853

comparisons on x86.  Essentially, the way this works is that SUB+SBB sets
the relevant flags the same way a double-width CMP would.

This is a substantial improvement over the generic lowering in LLVM. The output
is also shorter than the gcc-generated output; I haven't done any detailed
benchmarking, though.

llvm-svn: 127852

Augment CrashRecoveryContext to have registered "cleanup" objects that can be used to release resources during a crash.

llvm-svn: 127849

llvm-svn: 127845

Remove the offending logic and update the test cases.

llvm-svn: 127843

llvm-svn: 127842

llvm-svn: 127840

SCEV may generate expressions composed of multiple pointers, which can
lead to invalid GEP expansion. Until we can teach SCEV to follow strict
pointer rules, make sure no bad GEPs creep into IR.
Fixes rdar://problem/9038671.

llvm-svn: 127839

llvm-svn: 127837

instead of copying.

llvm-svn: 127835

This is done by lowering dbg.declare intrinsic into dbg.value intrinsic.
Radar 9143931.

llvm-svn: 127834