Fix PR1739.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43286 91177308-0d34-0410-b5e6-96231b3b80d8

Added VC++ solution file allowing to build clang.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43279 91177308-0d34-0410-b5e6-96231b3b80d8

Silenced a VC++ warning.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43276 91177308-0d34-0410-b5e6-96231b3b80d8

Updated VC++ build system

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43275 91177308-0d34-0410-b5e6-96231b3b80d8

Fix comment and use the "Size" variable that's already provided.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43271 91177308-0d34-0410-b5e6-96231b3b80d8

If there's an unaligned memcpy to/from the stack, don't lower it. Just call the
memcpy library function instead.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43270 91177308-0d34-0410-b5e6-96231b3b80d8

Disable a couple more things for ppcf128.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43267 91177308-0d34-0410-b5e6-96231b3b80d8

Added "ReadEnum" and "WriteEnum" to serialization classes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43265 91177308-0d34-0410-b5e6-96231b3b80d8

This broke lots. Reverting.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43264 91177308-0d34-0410-b5e6-96231b3b80d8

Make DomTreeBase not a FunctionPass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43263 91177308-0d34-0410-b5e6-96231b3b80d8

Lowering a memcpy to the stack is killing PPC. The ARM and X86 backends already
have their own custom memcpy lowering code. This code needs to be factored out
into a target-independent lowering method with hooks to the backend. In the
meantime, just call memcpy if we're trying to copy onto a stack.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43262 91177308-0d34-0410-b5e6-96231b3b80d8

Added preliminary implementation of generic object serialization to bitcode.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43261 91177308-0d34-0410-b5e6-96231b3b80d8

Unbreak the build.  Forgot to commit this file.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43260 91177308-0d34-0410-b5e6-96231b3b80d8

Make DomTree and PostDomTree thin wrappers around DomTreeBase, rather than inheriting from it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43259 91177308-0d34-0410-b5e6-96231b3b80d8

It's possible to commute instrctions with more than 3 operands.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43256 91177308-0d34-0410-b5e6-96231b3b80d8

new testcase

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43252 91177308-0d34-0410-b5e6-96231b3b80d8

llvm/test programs should not be execution programs.  This
will hopefully fix Gabor's Sparc problem.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43251 91177308-0d34-0410-b5e6-96231b3b80d8

isSubRegOf() is a dup of isSubRegister.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43249 91177308-0d34-0410-b5e6-96231b3b80d8

Temporary solution: added a different set of BCTRL_Macho / BCTRL_ELF with right callee-saved defs set for ppc64.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43248 91177308-0d34-0410-b5e6-96231b3b80d8

Fix validation problems, which were somewhat ironically in the "valid html" button code :)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43247 91177308-0d34-0410-b5e6-96231b3b80d8

add a skeleton for part 4

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43246 91177308-0d34-0410-b5e6-96231b3b80d8

complete the codegen chapter

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43245 91177308-0d34-0410-b5e6-96231b3b80d8

Now with valid HTML 4.01!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43244 91177308-0d34-0410-b5e6-96231b3b80d8

Add a link to the "writing an optimization" tutorial.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43243 91177308-0d34-0410-b5e6-96231b3b80d8

Add downloadable code sample for tutorial 2.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43242 91177308-0d34-0410-b5e6-96231b3b80d8

Add the second of the "basic topics" tutorials.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43241 91177308-0d34-0410-b5e6-96231b3b80d8

Reorder the optimization and bitcode tutorials.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43240 91177308-0d34-0410-b5e6-96231b3b80d8

Fix up a broken #include, move code to its own section, add conclusions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43239 91177308-0d34-0410-b5e6-96231b3b80d8

Finish up expr codegen.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43238 91177308-0d34-0410-b5e6-96231b3b80d8

several improvements suggested by Dan, thanks!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43237 91177308-0d34-0410-b5e6-96231b3b80d8

Fix memcpy lowering when addresses are 4-byte aligned but size is not multiple of 4.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43234 91177308-0d34-0410-b5e6-96231b3b80d8

The #include <iterator> isn't needed in this header.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43232 91177308-0d34-0410-b5e6-96231b3b80d8

Strength reduction improvements.

 - Avoid attempting stride-reuse in the case that there are users that
   aren't addresses. In that case, there will be places where the
   multiplications won't be folded away, so it's better to try to
   strength-reduce them.

 - Several SSE intrinsics have operands that strength-reduction can
   treat as addresses. The previous item makes this more visible, as
   any non-address use of an IV can inhibit stride-reuse.

 - Make ValidStride aware of whether there's likely to be a base
   register in the address computation. This prevents it from thinking
   that things like stride 9 are valid on x86 when the base register is
   already occupied.

Also, XFAIL the 2007-08-10-LEA16Use32.ll test; the new logic to avoid
stride-reuse elimintes the LEA in the loop, so the test is no longer
testing what it was intended to test.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43231 91177308-0d34-0410-b5e6-96231b3b80d8

Fix the folding of multiplication into addresses on x86, which was broken
by the recent {U,S}MUL_LOHI changes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43230 91177308-0d34-0410-b5e6-96231b3b80d8

Use ptr type in the immediate field of a BxA instruction so we don't end up selecting 32-bit call instruction for ppc64.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43228 91177308-0d34-0410-b5e6-96231b3b80d8

Add missing paratheses.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43227 91177308-0d34-0410-b5e6-96231b3b80d8

add a 'don't use me' marker.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43226 91177308-0d34-0410-b5e6-96231b3b80d8

Support for expanding extending loads of integers with
funky bit-widths.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43225 91177308-0d34-0410-b5e6-96231b3b80d8

Move the SCEV object factors from being static members of the individual
SCEV subclasses to being non-static member functions of the ScalarEvolution
class.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43224 91177308-0d34-0410-b5e6-96231b3b80d8

Fix up the logic for result expanding the various extension
operations so they work right for integers with funky
bit-widths.  For example, consider extending i48 to i64
on a 32 bit machine.  The i64 result is expanded to 2 x i32.
We know that the i48 operand will be promoted to i64, then
also expanded to 2 x i32.  If we had the expanded promoted
operand to hand, then expanding the result would be trivial.
Unfortunately at this stage we can only get hold of the
promoted operand.  So instead we kind of hand-expand, doing
explicit shifting and truncating to get the top and bottom
halves of the i64 operand into 2 x i32, which are then used
to expand the result.  This is harmless, because when the
promoted operand is finally expanded all this bit fiddling
turns into trivial operations which are eliminated either
by the expansion code itself or the DAG combiner.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43223 91177308-0d34-0410-b5e6-96231b3b80d8

fit in 80 cols :)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43222 91177308-0d34-0410-b5e6-96231b3b80d8

start of chapter 3

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43221 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a few typos I noticed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43220 91177308-0d34-0410-b5e6-96231b3b80d8

Add downloadable code samples.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43219 91177308-0d34-0410-b5e6-96231b3b80d8

Check in part 2: parser and ast.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43218 91177308-0d34-0410-b5e6-96231b3b80d8

Fix some code to make it actually work.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43217 91177308-0d34-0410-b5e6-96231b3b80d8

Put correct link in example

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43216 91177308-0d34-0410-b5e6-96231b3b80d8

add part 1, review appreciated.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43215 91177308-0d34-0410-b5e6-96231b3b80d8

add an outline for "part 2".

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43214 91177308-0d34-0410-b5e6-96231b3b80d8

rename tutorial 1 -> JITTutorial1, make unimplemented links not be dead links.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43213 91177308-0d34-0410-b5e6-96231b3b80d8

Fix an unfolding bug.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43212 91177308-0d34-0410-b5e6-96231b3b80d8

- Only perform the unfolding optimization when the folding in question is modref.
- Remove a bogus assertion.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43211 91177308-0d34-0410-b5e6-96231b3b80d8

add a mechanism for the JIT to invoke a function to lazily create functions as they are referenced.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43210 91177308-0d34-0410-b5e6-96231b3b80d8

llvm-gcc3 is dead, along with it __main.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43209 91177308-0d34-0410-b5e6-96231b3b80d8

Reg2Mem cleanup and optimizations:
 - enable phi instructions demotion to stack
 - create alloca instructions in the entry block

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43208 91177308-0d34-0410-b5e6-96231b3b80d8

LoadLibraryPermanently doesn't throw.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43207 91177308-0d34-0410-b5e6-96231b3b80d8

Add a convenience method for creating EE's.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43206 91177308-0d34-0410-b5e6-96231b3b80d8

Allow for copysign having f80 second argument.
Fixes 5550319.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43205 91177308-0d34-0410-b5e6-96231b3b80d8

Add promote operand support for [su]int_to_fp.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43204 91177308-0d34-0410-b5e6-96231b3b80d8

Some improvements based on feedback from Anders.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43203 91177308-0d34-0410-b5e6-96231b3b80d8

Fix some weird indenting in code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43202 91177308-0d34-0410-b5e6-96231b3b80d8

Use getOrInsertFunction() in tutorial 1.  This makes for shorter, simpler, and better example code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43201 91177308-0d34-0410-b5e6-96231b3b80d8

Add the beginnings of an LLVM tutorial.  If anyone has suggestions, comments,
or would like to contribute, let me know!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43200 91177308-0d34-0410-b5e6-96231b3b80d8

Add result promotion of FP_TO_*INT, fixing CodeGen/X86/trunc-to-bool.ll
with the new legalizer.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43199 91177308-0d34-0410-b5e6-96231b3b80d8

simplify some code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43198 91177308-0d34-0410-b5e6-96231b3b80d8

Implement promote and expand for operands of memcpy and friends.
This fixes CodeGen/X86/mem*.ll.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43197 91177308-0d34-0410-b5e6-96231b3b80d8

Added missing curly braces which renders the if clause useless in debug build.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43196 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a few places vector operations were not getting
the operand's type from the right place.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43195 91177308-0d34-0410-b5e6-96231b3b80d8

Resolve unfold tables ambiguity.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43194 91177308-0d34-0410-b5e6-96231b3b80d8

New test case.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43193 91177308-0d34-0410-b5e6-96231b3b80d8

Local spiller optimization:
Turn a store folding instruction into a load folding instruction. e.g.
     xorl  %edi, %eax
     movl  %eax, -32(%ebp)
     movl  -36(%ebp), %eax
     orl   %eax, -32(%ebp)
=>
     xorl  %edi, %eax
     orl   -36(%ebp), %eax
     mov   %eax, -32(%ebp)
This enables the unfolding optimization for a subsequent instruction which will
also eliminate the newly introduced store instruction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43192 91177308-0d34-0410-b5e6-96231b3b80d8

Don't branch fold inline asm statements.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43191 91177308-0d34-0410-b5e6-96231b3b80d8

Add support for a few more nodes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43190 91177308-0d34-0410-b5e6-96231b3b80d8

Redo "last ppc long double fix" as Chris wants.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43189 91177308-0d34-0410-b5e6-96231b3b80d8

Removed "-arch i386" by popular demand...

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43186 91177308-0d34-0410-b5e6-96231b3b80d8

Now with RUN line!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43185 91177308-0d34-0410-b5e6-96231b3b80d8

Remove the Foundation.h file because not everyone has that.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43184 91177308-0d34-0410-b5e6-96231b3b80d8

This test now passes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43183 91177308-0d34-0410-b5e6-96231b3b80d8

Update this file for 2.0 syntax. Contributed by Jan Rehders

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43182 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a really nasty vector miscompilation bill recently introduced.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43181 91177308-0d34-0410-b5e6-96231b3b80d8

Updated VC++ build system

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43180 91177308-0d34-0410-b5e6-96231b3b80d8

rename ExpandOperation to ExpandOperationResult, as suggested
by Duncan

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43177 91177308-0d34-0410-b5e6-96231b3b80d8

split LowerMEMCPY into LowerMEMCPYCall and LowerMEMCPYInline in the ARM backend.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43176 91177308-0d34-0410-b5e6-96231b3b80d8

Support for expanding ADDE and SUBE.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43175 91177308-0d34-0410-b5e6-96231b3b80d8

If the value types are equal then this routine
asserts in later checks rather than producing
the ordinary load it is supposed to.  Avoid all
such hassles by directly returning an ordinary
load in this case.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43174 91177308-0d34-0410-b5e6-96231b3b80d8

Test byval with a 8 bit aligned struct

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43173 91177308-0d34-0410-b5e6-96231b3b80d8

Add support for byval function whose argument is not 32 bit aligned.
To do this it is necessary to add a "always inline" argument to the
memcpy node. For completeness I have also added this node to memmove
and memset.  I have also added getMem* functions, because the extra
argument makes it cumbersome to use getNode and because I get confused
by it :-)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43172 91177308-0d34-0410-b5e6-96231b3b80d8

Implement a few new operations.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43171 91177308-0d34-0410-b5e6-96231b3b80d8

Implement expansion of SINT_TO_FP and UINT_TO_FP operands.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43170 91177308-0d34-0410-b5e6-96231b3b80d8

implement support for custom expansion of any node type, in one place.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43169 91177308-0d34-0410-b5e6-96231b3b80d8

comment fixes

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43168 91177308-0d34-0410-b5e6-96231b3b80d8

Make use of TLI.ExpandOperation, remove softfloat stuff.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43167 91177308-0d34-0410-b5e6-96231b3b80d8

add expand support for bit_convert result, even allowing custom expansion.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43166 91177308-0d34-0410-b5e6-96231b3b80d8

add a new target hook.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43165 91177308-0d34-0410-b5e6-96231b3b80d8

Add an easy microoptimization I noticed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43164 91177308-0d34-0410-b5e6-96231b3b80d8

New test.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43162 91177308-0d34-0410-b5e6-96231b3b80d8

Negative indices aren't allowed here.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43161 91177308-0d34-0410-b5e6-96231b3b80d8

More ppcf128 issues (maybe the last)?

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43160 91177308-0d34-0410-b5e6-96231b3b80d8

- Added getOpcodeAfterMemoryUnfold(). It doesn't unfold an instruction, but only returns the opcode of the instruction post unfolding.
- Fix some copy+paste bugs.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43153 91177308-0d34-0410-b5e6-96231b3b80d8

Use SmallVectorImpl instead of SmallVector with hardcoded size in MRegister public interface.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43150 91177308-0d34-0410-b5e6-96231b3b80d8