1
0
mirror of https://github.com/RPCS3/llvm.git synced 2024-12-24 04:56:51 +00:00
Commit Graph

29276 Commits

Author SHA1 Message Date
Dan Gohman
099879687f Use getSCEV instead of getUnknown to create a SCEV for a
Constant. This lets ConstantInts be handled as SCEVConstant instead
of SCEVUnknown, as getUnknown no longer has special-case code for
ConstantInt and friends. This usually doesn't affect the final
output, since the constants end up getting folded later, but it
does make intermediate expressions more obvious in many cases.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74459 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 21:31:18 +00:00
Daniel Dunbar
74deeb84bd Don't build LLVMC when configured with --disable-pic (it needs requires shared
module support to build).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74456 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 21:12:26 +00:00
Torok Edwin
6c839926ba Set wasRun to false here on Dan's suggestion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74455 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 21:05:10 +00:00
Devang Patel
1dbc771a83 s/MainCU/ModuleCU/g
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74452 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 20:45:18 +00:00
Devang Patel
70f4426750 Multiple DW_TAG_compile_unit is not used, afaict, on any target.
Update dwarf writer to only emit one DW_TAG_compile_unit per .o file. 


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74449 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 20:38:13 +00:00
Dan Gohman
cffddcd726 Don't cache PHI exit values from exhaustive evaluations, because
an individual exhaustive evaluation reflects only the exit value
implied by an individual exit, which may differ from the actual
exit value of the loop if there are other exits. This fixes PR4477.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74447 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 20:34:13 +00:00
Rafael Espindola
63de5c3b7e FIX PR 4459.
Not sure I understand how the temp register gets used,
but this fixes a bug and introduces no regressions.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74446 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 20:29:59 +00:00
Torok Edwin
48de5d8fd0 Initialize CurSection in constructor.
Not doing so causes some unittests to fail, because CurSection is uninitialized.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74442 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 19:59:10 +00:00
Daniel Dunbar
7908a6a189 Rename MCValue::getCst to getConstant and add MCValue::isConstant.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74440 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 19:51:00 +00:00
Andreas Bolka
2fbb770d40 Relax LDA memory instruction checks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74439 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 18:51:11 +00:00
Torok Edwin
1970a89a49 Call doInitialization(), releaseMemory(), and doFinalization() for on-the-fly passes as well.
Also don't call finalizers for LoopPass if initialization was not called.
Add a unittest that tests that these methods are called, in the proper
order, and the correct number of times.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74438 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 18:49:09 +00:00
Dan Gohman
23737e0423 Simplify this code, and avoid using APInt(). This fixes
(otherwise harmless) uninitialized value warnings that
Duncan found with gcc-4.4.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74437 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 18:25:52 +00:00
Owen Anderson
9917700af4 Add a target-specific DAG combine on X86 to fold the common pattern of
fence-atomic-fence down to just the atomic op.  This is possible thanks to
X86's relatively strong memory model, which guarantees that locked instructions
(which are used to implement atomics) are implicit fences.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74435 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 18:04:45 +00:00
David Greene
3016af5a1d Add processor descriptions for Istanbul and Shanghai.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74429 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 16:54:06 +00:00
David Greene
640a0c1477 Fix a subtarget feature bug.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74428 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 16:51:01 +00:00
David Greene
9b9838dbd4 Add more vector ValueTypes for AVX and other extended vector instruction
sets.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74427 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 16:47:10 +00:00
David Goodwin
c0309b48b5 Rename ARMcmpNZ to ARMcmpZ and use it to represent comparisons that set only the Z flag (i.e. eq and ne). Make ARMcmpZ commutative.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74423 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 15:33:01 +00:00
Duncan Sands
cd1267d2d6 Add triple for OpenBSD.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74422 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 13:36:13 +00:00
Duncan Sands
c91e68a0c4 Include the new file ThumbRegisterInfo.cpp to CMakeLists.txt
to make sure ThumbRegisterInfo.cpp are compiled and linked in.
Patch by Xerxes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74421 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 13:11:32 +00:00
Evan Cheng
055b0310f8 Implement Thumb2 ldr.
After much back and forth, I decided to deviate from ARM design and split LDR into 4 instructions (r + imm12, r + imm8, r + r << imm12, constantpool). The advantage of this is 1) it follows the latest ARM technical manual, and 2) makes it easier to reduce the width of the instruction later. The down side is this creates more inconsistency between the two sub-targets. We should split ARM LDR instruction in a similar fashion later. I've added a README entry for this.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74420 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 07:51:04 +00:00
Mikhail Glushenkov
95c1f5ba64 Make dynamic LLVMC plugins work on Windows (finally!).
Implemented by making lib/CompilerDriver a shared library that holds all the
global static data (CommandLine options, plugin registry) that we unfortunately
have to live with.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74417 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 03:09:15 +00:00
Andreas Bolka
868a302738 Missed one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74416 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 00:53:49 +00:00
Andreas Bolka
acd6f8d5d1 Fix case in LDA util function names.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74415 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-29 00:50:26 +00:00
Andreas Bolka
c6a303052b Print pairwise dependence results, add testcases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74402 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-28 00:35:22 +00:00
Andreas Bolka
f35626d3cd Minimal LDA interface, maximally conservative tester.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74401 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-28 00:21:21 +00:00
Andreas Bolka
707207adae LDA analysis output scaffolding.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74400 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-28 00:16:08 +00:00
Dan Gohman
f7ca1611e2 Don't try to split a loop when the controlling icmp instruction
doesn't have an IV-based operand. This fixes PR4471.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74399 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 22:58:27 +00:00
Dan Gohman
a1baee20c4 Remove the block from the LoopInfo, rather than just the Loop.
LoopInfo will handle removing it from the Loop, as well as updating
its own tables.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74398 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 22:32:36 +00:00
Dan Gohman
4b35f83b91 Teach LoopSimplify how to merge multiple loop exits into a single exit,
when one of them can be converted to a trivial icmp and conditional
branch.

This addresses what is essentially a phase ordering problem.
SimplifyCFG knows how to do this transformation, but it doesn't do so
if the primary block has any instructions in it other than an icmp and
a branch. In the given testcase, the block contains other instructions,
however they are loop-invariant and can be hoisted. SimplifyCFG doesn't
have LoopInfo though, so it can't hoist them. And, it's important that
the blocks be merged before LoopRotation, as it doesn't support
multiple-exit loops.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74396 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 21:30:38 +00:00
Dan Gohman
cfb32203bc More minor code simplifications.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74395 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 21:23:40 +00:00
Dan Gohman
9d59d9f849 Eliminate a layer of indirection in LoopInfo and MachineLoopInfo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74394 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 21:22:48 +00:00
Dan Gohman
1c34375f79 Convert ScalarEvolution to use BumpPtrAllocator and FoldingSet, instead
of a team of individual allocations and a team of std::maps.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74393 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 21:21:31 +00:00
Dan Gohman
267a385342 Change SCEVExpander to use an IRBuilder to emit instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74391 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 21:18:18 +00:00
Anton Korobeynikov
55ad1f22b4 Simplify a bit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74385 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 12:59:03 +00:00
Anton Korobeynikov
a98cbc554c ARM refactoring. Step 2: split RegisterInfo
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74384 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 12:16:40 +00:00
Douglas Gregor
1b573d8b20 Add ThumbInstrInfo.cpp to the CMake makefiles
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74382 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 07:44:59 +00:00
Chris Lattner
f7789c72bd remove a bunch of fixmes (old checking code) and commonize all the
target-specific operand printing functionality.  Yay.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74379 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 05:46:24 +00:00
Chris Lattner
b1acd68fe4 pull @GOT, @GOTOFF, @GOTPCREL handling into isel from the asmprinter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74378 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 05:39:56 +00:00
Chris Lattner
4c1b606ecd simplify some code and eliminate the symbolicAddressesAreRIPRel() predicate.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74377 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 05:24:12 +00:00
Dan Gohman
de53dc03f5 When a value is used multiple times within a single PHI, instructions
inserted to replace that value must dominate all of of the basic
blocks associated with the uses of the value in the PHI, not just
one of them.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74376 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 05:16:57 +00:00
Chris Lattner
553e571b43 fix clang/test/CodeGenObjC/try.m, a basereg doesn't mean no global anymore.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74375 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 04:50:14 +00:00
Chris Lattner
380135cc3e fix a bunch of failures in the X86-64 JIT by tolerating RIP as
a base register.  We just ignore it for now.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74374 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 04:46:33 +00:00
Chris Lattner
74d3f50a80 factor some logic out into a helper function, allow remat of loads from constant
globals.  This implements remat-constant.ll even without aggressive-remat.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74373 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 04:38:55 +00:00
Chris Lattner
18c5987fa3 Reimplement rip-relative addressing in the X86-64 backend. The new
implementation primarily differs from the former in that the asmprinter
doesn't make a zillion decisions about whether or not something will be
RIP relative or not.  Instead, those decisions are made by isel lowering
and propagated through to the asm printer.  To achieve this, we:

1. Represent RIP relative addresses by setting the base of the X86 addr
   mode to X86::RIP.
2. When ISel Lowering decides that it is safe to use RIP, it lowers to
   X86ISD::WrapperRIP.  When it is unsafe to use RIP, it lowers to
   X86ISD::Wrapper as before.
3. This removes isRIPRel from X86ISelAddressMode, representing it with
   a basereg of RIP instead.
4. The addressing mode matching logic in isel is greatly simplified.
5. The asmprinter is greatly simplified, notably the "NotRIPRel" predicate
   passed through various printoperand routines is gone now.
6. The various symbol printing routines in asmprinter now no longer infer
   when to emit (%rip), they just print the symbol.

I think this is a big improvement over the previous situation.  It does have
two small caveats though: 1. I implemented a horrible "no-rip" modifier for
the inline asm "P" constraint modifier.  This is a short term hack, there is
a much better, but more involved, solution.  2. I had to xfail an 
-aggressive-remat testcase because it isn't handling the use of RIP in the
constant-pool reading instruction.  This specific test is easy to fix without
-aggressive-remat, which I intend to do next.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74372 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 04:16:01 +00:00
Chris Lattner
1873d0c225 When doing remat, don't consider uses of non-allocatable physregs. Patch
by Evan.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74370 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 04:06:41 +00:00
Evan Cheng
9cb9e6778c Renaming for consistency.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74368 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 02:26:13 +00:00
Chris Lattner
89da699092 Fix PR4466 by making fastisel set operand flags correctly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74366 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 01:31:51 +00:00
Chris Lattner
5b93ea1ca3 fix a typo that GCC should have caught that causes crashes with -view-*-dags
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74364 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-27 00:57:02 +00:00
David Goodwin
8eba8e0ca3 Remove outdated comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74357 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-26 23:39:02 +00:00
David Goodwin
dcdaebc592 When possible, use "mvn ra, rb" instead of "eor ra, rb, -1" because mvn has a narrow version and eor(i) does not.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74355 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-26 23:13:13 +00:00