Commit Graph

20563 Commits

Author SHA1 Message Date
Evan Cheng
4102eb57bb 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
2007-10-22 22:11:27 +00:00
Dan Gohman
ac72058dd6 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
2007-10-22 20:44:10 +00:00
Dan Gohman
02e4fa7d5f 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
2007-10-22 20:40:42 +00:00
Dan Gohman
8368805e25 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
2007-10-22 20:22:24 +00:00
Evan Cheng
3311876c3c 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
2007-10-22 19:46:19 +00:00
Evan Cheng
76500d52be Add missing paratheses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43227 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-22 19:42:28 +00:00
Duncan Sands
fec3ad3e5f 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
2007-10-22 19:00:05 +00:00
Dan Gohman
246b2564d3 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
2007-10-22 18:31:58 +00:00
Duncan Sands
245741d2a1 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
2007-10-22 18:26:21 +00:00
Evan Cheng
42b08be973 Fix an unfolding bug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43212 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-22 03:03:20 +00:00
Evan Cheng
cbfb9b27ef - 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
2007-10-22 03:01:44 +00:00
Chris Lattner
d958a5a9fe 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
2007-10-22 02:50:12 +00:00
Chris Lattner
ec2fcafbea 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
2007-10-22 02:39:47 +00:00
Anton Korobeynikov
a024e8ceda 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
2007-10-21 23:05:16 +00:00
Chris Lattner
c6185038b8 LoadLibraryPermanently doesn't throw.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43207 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-21 22:58:11 +00:00
Chris Lattner
8b5295b7bb 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
2007-10-21 22:57:11 +00:00
Dale Johannesen
61c7ef34e3 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
2007-10-21 01:07:44 +00:00
Chris Lattner
d0e908e830 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
2007-10-20 22:57:56 +00:00
Chris Lattner
be5a0a4668 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
2007-10-20 04:32:38 +00:00
Chris Lattner
aaeb0c8666 simplify some code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43198 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-20 04:09:48 +00:00
Chris Lattner
d20154c762 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
2007-10-20 04:07:07 +00:00
Evan Cheng
b9d5e7cdc9 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
2007-10-20 04:01:47 +00:00
Dale Johannesen
e526962fc5 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
2007-10-20 00:07:52 +00:00
Evan Cheng
f7c9695a38 Resolve unfold tables ambiguity.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43194 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 23:50:58 +00:00
Evan Cheng
66f7163545 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
2007-10-19 21:23:22 +00:00
Bill Wendling
80629c85f1 Don't branch fold inline asm statements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43191 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 21:09:55 +00:00
Duncan Sands
7f5f6851c3 Add support for a few more nodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43190 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 20:29:48 +00:00
Dale Johannesen
b6210fc92b 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
2007-10-19 20:29:00 +00:00
Chris Lattner
f185e67844 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
2007-10-19 16:47:35 +00:00
Chris Lattner
1c4d492b94 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
2007-10-19 15:28:47 +00:00
Rafael Espindola
7b73a5d6de 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
2007-10-19 14:35:17 +00:00
Duncan Sands
c8ca2365e5 Support for expanding ADDE and SUBE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43175 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 13:06:17 +00:00
Duncan Sands
5d868b1ffe 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
2007-10-19 13:05:40 +00:00
Rafael Espindola
5c0d6ed325 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
2007-10-19 10:41:11 +00:00
Chris Lattner
1f39dc42cd Implement a few new operations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43171 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 04:46:45 +00:00
Chris Lattner
50f4395d86 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
2007-10-19 04:32:47 +00:00
Chris Lattner
c99df1b848 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
2007-10-19 04:14:36 +00:00
Chris Lattner
26cb2862e9 comment fixes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43168 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 04:08:28 +00:00
Chris Lattner
597ab48ca0 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
2007-10-19 03:58:25 +00:00
Chris Lattner
c224a53d7a 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
2007-10-19 03:33:14 +00:00
Chris Lattner
f58dac3169 add a new target hook.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43165 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 03:31:45 +00:00
Chris Lattner
3a7c33a853 Add an easy microoptimization I noticed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43164 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 03:29:26 +00:00
Bill Wendling
1851898e75 Negative indices aren't allowed here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43161 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 01:10:49 +00:00
Dale Johannesen
fabd32deb0 More ppcf128 issues (maybe the last)?
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43160 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-19 00:59:18 +00:00
Evan Cheng
f0a0cddbcd - 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
2007-10-18 22:40:57 +00:00
Evan Cheng
58184e6878 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
2007-10-18 21:29:24 +00:00
Devang Patel
99db6add3d Try again.
Instead of loading small global string from memory, use
integer constant.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43148 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-18 19:52:32 +00:00
Owen Anderson
b388ca9544 Allow GVN to eliminate redundant calls to functions without side effects.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43147 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-18 19:39:33 +00:00
Christopher Lamb
a4c791072f Fix a misnamed parameter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43145 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-18 19:29:45 +00:00
Christopher Lamb
91ee18c1f7 Fix a typo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@43144 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-18 19:28:55 +00:00