Commit Graph

5339 Commits

Author SHA1 Message Date
Chris Lattner
e5d8861126 Implement selection of inline asm memory operands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26348 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-24 02:13:12 +00:00
Evan Cheng
d0839f3071 PPC JIT relocation model should be DynamicNoPIC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26338 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-23 22:18:07 +00:00
Evan Cheng
020d2e8e7a - Clean up the lowering and selection code of ConstantPool, GlobalAddress,
and ExternalSymbol.
- Use C++ code (rather than tblgen'd selection code) to match the above
  mentioned leaf nodes. Do not mutate and nodes and do not record the
  selection in CodeGenMap. These nodes should be safe to duplicate. This is
  a performance win.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26335 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-23 20:41:18 +00:00
Chris Lattner
e3f01570c1 Implement the PPC inline asm "L" modifier. This allows us to compile:
long long test(long long X) {
  __asm__("foo %0 %L0 %1 %L1" : "=r"(X): "r"(X));
  return X;
}

to:
        foo r2 r3 r2 r3


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26333 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-23 19:31:10 +00:00
Chris Lattner
e650a6b3f4 "." isn't enough to get a private label on linux, use ".L".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26327 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-23 05:25:02 +00:00
Chris Lattner
205065ae0c add a small and simple case.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26326 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-23 05:17:43 +00:00
Evan Cheng
3032410f9b A couple of new entries.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26325 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-23 02:50:21 +00:00
Evan Cheng
a0ea0539e3 PIC related bug fixes.
1. Various asm printer bug.
2. Lowering bug. Now TargetGlobalAddress is wrapped in X86ISD::TGAWrapper.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26324 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-23 02:43:52 +00:00
Evan Cheng
224ec39cab X86 codegen tweak to use lea in another case:
Suppose base == %eax and it has multiple uses, then instead of
  movl %eax, %ecx
  addl $8, %ecx
use
  leal 8(%eax), %ecx.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26323 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-23 00:13:58 +00:00
Evan Cheng
f1616dadad Missing .globl for weak / link-once .text symbols.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26321 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-22 23:59:57 +00:00
Evan Cheng
4c1aa86657 - Added option -relocation-model to set relocation model. Valid values include static, pic,
dynamic-no-pic, and default.
PPC and x86 default is dynamic-no-pic for Darwin, pic for others.
- Removed options -enable-pic and -ppc-static.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26315 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-22 20:19:42 +00:00
Jim Laskey
0420f2aaf9 Coordinate activities with llvm-gcc4 and dwarf.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26314 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-22 19:02:11 +00:00
Evan Cheng
470a6adc78 Added MMX, SSE1, and SSE2 vector instructions and some simple patterns.
Fixed some existing bugs (wrong predicates, prefixes) at the same time.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26310 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-22 02:26:30 +00:00
Chris Lattner
1efa40f6a4 split register class handling from explicit physreg handling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26308 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-22 00:56:39 +00:00
Chris Lattner
4217ca8dc1 Updates to match change of getRegForInlineAsmConstraint prototype
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26305 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 23:11:00 +00:00
Evan Cheng
4e4c71e423 One more round of reorg so sabre doesn't freak out. :-)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26303 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 20:00:20 +00:00
Evan Cheng
beb07e117d A big more cleaning up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26302 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 19:30:30 +00:00
Evan Cheng
bf156d1ae6 Moving things to their proper places.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26301 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 19:26:52 +00:00
Evan Cheng
ffcb95beab Split instruction info into multiple files, one for each of x87, MMX, and SSE.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26300 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 19:13:53 +00:00
Chris Lattner
a1532bc283 missed optzn
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26299 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 18:29:44 +00:00
Chris Lattner
2deb87f201 The HasNoV9 hack isn't needed here, now that tblgen knows that CustomDAGSchedInserter
instructions are expensive.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26298 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 18:04:32 +00:00
Evan Cheng
747a90d887 Added separate alias instructions for SSE logical ops that operate on non-packed types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26297 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 02:24:38 +00:00
Evan Cheng
7dbc0a3351 Added MMX and XMM packed integer move instructions, movd and movq.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26296 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 01:39:57 +00:00
Evan Cheng
933be3318b Added SSE2 128-bit integer packed types: V16I8, V8I16, V4I32, and V2I64.
Added generic vector types: VR64 and VR128.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26295 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-21 01:38:21 +00:00
Evan Cheng
aea20f50e5 Added x86 integer vector types: 64-bit packed byte integer (v16i8), 64-bit
packed word integer (v8i16), and 64-bit packed doubleword integer (v2i32).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26294 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-20 22:34:53 +00:00
Evan Cheng
755ee8f9aa Some updates
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26292 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-20 19:58:27 +00:00
Evan Cheng
45af8fd8c2 If SSE3 is available, promote FP_TO_UINT i32 to FP_TO_SINT i64 to take
advantage of fisttpll.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26288 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-18 07:26:17 +00:00
Evan Cheng
2b15271571 Added fisttp for fp to int conversion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26283 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-18 02:36:28 +00:00
Evan Cheng
6428302f3d Disable PIC for JIT.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26281 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-18 01:49:25 +00:00
Evan Cheng
5e8b5558f7 Jit does not support PIC yet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26278 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-18 00:57:10 +00:00
Evan Cheng
7ccced634a x86 / Darwin PIC support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26273 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-18 00:15:05 +00:00
Evan Cheng
d2ee218b49 Moved PICEnabled to include/llvm/Target/TargetOptions.h
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26272 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-18 00:08:58 +00:00
Chris Lattner
c2fe97e726 unbreak the build
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26260 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-17 07:09:27 +00:00
Evan Cheng
5298bcc722 Unbreak x86 be
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26259 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-17 07:01:52 +00:00
Nate Begeman
551bf3f800 kill ADD_PARTS & SUB_PARTS and replace them with fancy new ADDC, ADDE, SUBC
and SUBE nodes that actually expose what's going on and allow for
significant simplifications in the targets.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26255 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-17 05:43:56 +00:00
Chris Lattner
a648df2871 add note about div by power of 2
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26253 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-17 04:20:13 +00:00
Nate Begeman
9c3c2e9686 Whoops, didn't mean to check this in yet.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26250 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-17 00:56:19 +00:00
Nate Begeman
c22f357b78 Add a missing and useful pat frag
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26249 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-17 00:51:06 +00:00
Evan Cheng
7634ac4a9b Remind ourselves to revisit the "pxor vs. xorps/xorpd to clear XMM registers"
issue. Need to do more experiments.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26247 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-17 00:04:28 +00:00
Nate Begeman
4c5dcf54ff Kill the x86 pattern isel. boom.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26246 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-17 00:03:04 +00:00
Evan Cheng
dc8acb6420 Remove the entry about using movapd for SSE reg-reg moves.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26245 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-17 00:00:58 +00:00
Evan Cheng
39d1761c70 pxor (for FLD0SS) encoding was missing the OpSize prefix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26244 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-16 23:59:30 +00:00
Chris Lattner
6a6eb7b622 Remove the skeleton target, it doesn't produce useful code and there are
other small targets that do that can be learned from.  They also have
the added advantage of being tested :)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26243 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-16 23:14:50 +00:00
Evan Cheng
fe5cb19405 1. Use pxor instead of xoraps / xorapd to clear FR32 / FR64 registers. This
proves to be worth 20% on Ptrdist/ks. Might be related to dependency
   breaking support.
2. Added FsMOVAPSrr and FsMOVAPDrr as aliases to MOVAPSrr and MOVAPDrr. These
   are used for FR32 / FR64 reg-to-reg copies.
3. Tell reg-allocator to generate MOVSSrm / MOVSDrm and MOVSSmr / MOVSDmr to
   spill / restore FsMOVAPSrr and FsMOVAPDrr.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26241 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-16 22:45:17 +00:00
Evan Cheng
19ade3bf9c Use movaps / movapd to spill / restore V4F4 / V2F8 registers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26240 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-16 21:20:26 +00:00
Nate Begeman
368e18d56a Rework the SelectionDAG-based implementations of SimplifyDemandedBits
and ComputeMaskedBits to match the new improved versions in instcombine.
Tested against all of multisource/benchmarks on ppc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26238 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-16 21:11:51 +00:00
Evan Cheng
77dea9b3c3 MOVAPSrr and MOVAPDrr instruction format should be MRMSrcReg.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26234 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-16 19:34:41 +00:00
Duraid Madina
00d3d44b75 distinguish between objects and register names, now we can have stuff
with names like "f84", "in6" etc etc.

this should fix one or two tests


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26232 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-16 13:12:57 +00:00
Evan Cheng
298ebf2bd8 If the false case is the current basic block, then this is a self loop.
We do not want to emit "Loop: ... brcond Out; br Loop", as it adds an extra
instruction in the loop.  Instead, invert the condition and emit
"Loop: ... br!cond Loop; br Out.

Generalize the fix by moving it from PPCDAGToDAGISel to SelectionDAGLowering.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26231 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-16 08:27:56 +00:00
Evan Cheng
d51425a82d Use movaps / movapd (instead of movss / movsd) to do FR32 / FR64 reg to reg
transfer.

According to the Intel P4 Optimization Manual:

Moves that write a portion of a register can introduce unwanted
dependences. The movsd reg, reg instruction writes only the bottom
64 bits of a register, not to all 128 bits. This introduces a dependence on
the preceding instruction that produces the upper 64 bits (even if those
bits are not longer wanted). The dependence inhibits register renaming,
and thereby reduces parallelism.

Not to mention movaps is shorter than movss.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@26226 91177308-0d34-0410-b5e6-96231b3b80d8
2006-02-16 01:50:02 +00:00