rewriter in SROA to carry a proper alignment. This involves
interrogating various sources of alignment, etc. This is a more complete
and principled fix to PR13920 as well as related bugs pointed out by Eli
in review and by inspection in the area.

Also by inspection fix the integer and vector promotion paths to create
aligned loads and stores. I still need to work up test cases for
these... Sorry for the delay, they were found purely by inspection.

llvm-svn: 164689

and stores. These will be used in subsequnet patches to SROA to more
systematically manage the alignment on loads and stores.

llvm-svn: 164688

llvm-svn: 164687

llvm-svn: 164686

llvm-svn: 164685

tables in bitmaps when they fit in a target-legal register.

This saves some space, and it also allows for building tables that would
otherwise be deemed too sparse.

One interesting case that this hits is example 7 from
http://blog.regehr.org/archives/320. We currently generate good code
for this when lowering the switch to the selection DAG: we build a
bitmask to decide whether to jump to one block or the other. My patch
will result in the same bitmask, but it removes the need for the jump,
as the return value can just be retrieved from the mask.

llvm-svn: 164684

breaking out the building of lookup tables into a separate class.

llvm-svn: 164682

llvm-svn: 164681

For example, under a Linux chroot, /proc/ might not be mounted.
Therefor, we test if this file exist. If it is the case, use it (the current
behavior). Otherwise, we fall back to the detection used by *BSD.

The issue has been reported initially on the Debian bug tracker:
http://bugs.debian.org/674588

llvm-svn: 164676

llvm-svn: 164675

llvm-svn: 164674

llvm-svn: 164673

llvm-svn: 164671

llvm-svn: 164670

This should really, really fix PR13916. For real this time. The
underlying bug is... a bit more subtle than I had imagined.

The setup is a code pattern that leads to an @llvm.memcpy call with two
equal pointers to an alloca in the source and dest. Now, not any pattern
will do. The alloca needs to be formed just so, and both pointers should
be wrapped in different bitcasts etc. When this precise pattern hits,
a funny sequence of events transpires. First, we correctly detect the
potential for overlap, and correctly optimize the memcpy. The first
time. However, we do simplify the set of users of the alloca, and that
causes us to run the alloca back through the SROA pass in case there are
knock-on simplifications. At this point, a curious thing has happened.
If we happen to have an i8 alloca, we have direct i8 pointer values. So
we don't bother creating a cast, we rewrite the arguments to the memcpy
to dircetly refer to the alloca.

Now, in an unrelated area of the pass, we have clever logic which
ensures that when visiting each User of a particular pointer derived
from an alloca, we only visit that User once, and directly inspect all
of its operands which refer to that particular pointer value. However,
the mechanism used to detect memcpy's with the potential to overlap
relied upon getting visited once per *Use*, not once per *User*. This is
always true *unless* the same exact value is both source and dest. It
turns out that almost nothing actually produces that pattern though.

We can hand craft test cases that more directly test this behavior of
course, and those are included. Also, note that there is a significant
missed optimization here -- we prove in many cases that there is
a non-volatile memcpy call with identical source and dest addresses. We
shouldn't prevent splitting the alloca in that case, and in fact we
should just remove such memcpy calls eagerly. I'll address that in
a subsequent commit.

llvm-svn: 164669

llvm-svn: 164668

Replace calls to getSizeInBits with getExtendedSizeInBits since its already known its an extended type.

llvm-svn: 164667

llvm-svn: 164666

llvm-svn: 164664

llvm-svn: 164663

scalar-to-vector conversion that we cannot handle. For instance, when an invalid
constraint is used in an inline asm statement.
<rdar://problem/12284092>

llvm-svn: 164662

- Instead of embedding 'lock' into each mnemonic of atomic
  instructions except 'xchg', we teach X86 assembly printer to output 'lock'
  prefix similar to or consistent with code emitter.

llvm-svn: 164659

scalar-to-vector conversion that we cannot handle. For instance, when an invalid
constraint is used in an inline asm statement.
<rdar://problem/12284092>

llvm-svn: 164657

Fixed issue with Release build.

llvm-svn: 164654

The Apple buildbots are set up to pass --target to configure for both
cross- and non-cross-compile builds, and the standard autoconf response
to this is to set the program prefix to '<target>-'. Until we can figure
out the proper way to handle this (don't pass --target? pass an explicit
--program-prefix=""? don't auto-populate program_prefix with target_alias?)
it's more important to keep the buildbots running.

This reverts r164633 / ba48ceb1a3802e20e781ef04ea2573ffae2ac414.

llvm-svn: 164651

llvm-svn: 164642

only a missed optimization opportunity if the store is over-aligned, but a
miscompile if the store's new type has a higher natural alignment than the
memcpy did. Fixes PR13920!

llvm-svn: 164641

llvm-svn: 164640

reason we were getting two of the same alloca is because of a memmove/memcpy
which had the same alloca in both the src and dest. Now we detect that case
directly. This has the same testcase as before, but fixes a clang test
CodeGenObjC/exceptions.m which runs clang -O2.

llvm-svn: 164636

Chandler, it's not obvious that it's okay that this alloca gets into the list
twice to begin with. Please review and see whether this is the fix you really
want, but I wanted to get a fix checked in quickly.

llvm-svn: 164634

llvm-svn: 164633

llvm-svn: 164631

llvm-svn: 164630

llvm-svn: 164629

Provide interface in TargetLowering to set or get the minimum number of basic
blocks whereby jump tables are generated for switch statements rather than an
if sequence.

    getMinimumJumpTableEntries() defaults to 4.
    setMinimumJumpTableEntries() allows target configuration.

    This patch changes the default for the Hexagon architecture to 5
    as it improves performance on some benchmarks.

llvm-svn: 164628

llvm-svn: 164627

doesn't transform the trivially unsafe case.

llvm-svn: 164617

- Turn on atomic6432.ll and add specific test case as well

llvm-svn: 164616

When a BL/BLX references a symbol in the same translation unit that is
out of range, use an external relocation. The linker will use this to
generate a branch island rather than a direct reference, allowing the
relocation to resolve correctly.

rdar://12359919

llvm-svn: 164615

llvm-svn: 164614