Commit Graph

339 Commits

Author SHA1 Message Date
Evan Cheng
498c2903e2 CommuteChangesDestination() should check if to-be-commuted instruction defines any register. Also teaches the default commuteInstruction() to commute instruction without definitions (e.g. X86::test / ARM::tsp).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74602 91177308-0d34-0410-b5e6-96231b3b80d8
2009-07-01 08:29:08 +00:00
Evan Cheng
459a7c6b6a Remove special handling of implicit_def. Fix a couple more bugs in liveintervalanalysis and coalescer handling of implicit_def.
Note, isUndef marker must be placed even on implicit_def def operand or else the scavenger will not ignore it. This is necessary because -O0 path does not use liveintervalanalysis, it treats implicit_def just like any other def.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74601 91177308-0d34-0410-b5e6-96231b3b80d8
2009-07-01 08:19:36 +00:00
Evan Cheng
2578ba26e7 Handle IMPLICIT_DEF with isUndef operand marker, part 2. This patch moves the code to annotate machineoperands to LiveIntervalAnalysis. It also add markers for implicit_def that define physical registers. The rest, is just a lot of details.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74580 91177308-0d34-0410-b5e6-96231b3b80d8
2009-07-01 01:59:31 +00:00
Evan Cheng
50564ebc9e Temporarily restore the scavenger implicit_def checking code. MachineOperand isUndef mark is not being put on implicit_def of physical registers (created for parameter passing, etc.).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74519 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-30 09:19:42 +00:00
Evan Cheng
4784f1fc73 Add a bit IsUndef to MachineOperand. This indicates the def / use register operand is defined by an implicit_def. That means it can def / use any register and passes (e.g. register scavenger) can feel free to ignore them.
The register allocator, when it allocates a register to a virtual register defined by an implicit_def, can allocate any physical register without worrying about overlapping live ranges. It should mark all of operands of the said virtual register so later passes will do the right thing.

This is not the best solution. But it should be a lot less fragile to having the scavenger try to track what is defined by implicit_def.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74518 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-30 08:49:04 +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
David Goodwin
caffbd7b01 ORN and BIC tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74289 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-26 16:20:06 +00:00
Evan Cheng
6677f51127 Fix tests: Count -> count.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74282 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-26 07:05:57 +00:00
Evan Cheng
c0ad80fd9e Fix a CodeGenDAGPatterns bug. Check if top level predicates match when it's looking for duplicates.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74276 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-26 05:59:16 +00:00
Evan Cheng
6267422318 Select ADC, SBC, and RSC instead of the ADCS, SBCS, and RSCS when the carry bit def is not used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74228 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-25 23:34:10 +00:00
Evan Cheng
1e249e3705 ISD::ADDE / ISD::SUBE updates the carry bit so they should isle to ADCS and SBCS / RSCS.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74200 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-25 20:59:23 +00:00
Evan Cheng
9f76ed5127 Move thumb and thumb2 tests into separate directories.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@74068 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-24 06:36:07 +00:00
Evan Cheng
a67efd1226 Proper patterns for thumb2 shift and rotate instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73987 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-23 19:39:13 +00:00
Bob Wilson
5bafff36c7 Add support for ARM's Advanced SIMD (NEON) instruction set.
This is still a work in progress but most of the NEON instruction set
is supported.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73919 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-22 23:27:02 +00:00
Evan Cheng
524961e57f It's coalescer, not coaleser.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73902 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-22 21:09:17 +00:00
Bob Wilson
54fc124d72 For Darwin on ARMv6 and newer, make register r9 available for use as a
caller-saved register.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73901 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-22 21:01:46 +00:00
Evan Cheng
81909b7423 Fix another register coalescer crash: forgot to check if the instruction being updated has already been coalesced.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73898 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-22 20:49:32 +00:00
Evan Cheng
8dcbbdd00e hasFP should return true if frame address is taken.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73893 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-22 18:38:48 +00:00
Evan Cheng
694f6c81e8 Fix PR4419: handle defs of partial uses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73816 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-20 04:34:51 +00:00
Evan Cheng
ae69a2a12b Enable arm pre-allocation load / store multiple optimization pass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73791 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-19 23:17:27 +00:00
Eli Friedman
6b7bb42c36 Mark a few Thumb instructions commutable; just happened to spot this
while experimenting.  I'm reasonably sure this is correct, but please 
tell me if these instructions have some strange property which makes this
change unsafe.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73746 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-19 01:43:08 +00:00
Anton Korobeynikov
52237119a9 Initial support for some Thumb2 instructions.
Patch by Viktor Kutuzov and Anton Korzh from Access Softek, Inc.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73622 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-17 18:13:58 +00:00
Anton Korobeynikov
a4e968cb0e Make the test target-neutral
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73547 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-16 20:25:25 +00:00
Anton Korobeynikov
2932795309 GNU as refuses to assemble "pop {}" instruction. Do not emit such
(this is the case when we have thumb vararg function with single
callee-saved register, which is handled separately).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73529 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-16 18:49:08 +00:00
Evan Cheng
67fcf56ac4 If a val# is defined by an implicit_def and it is being removed, all of the copies off the val# were removed. This causes problem later since the scavenger will see uses of registers without defs. The proper solution is to change the copies into implicit_def's instead.
TurnCopyIntoImpDef turns a copy into implicit_def and remove the val# defined by it. This causes an scavenger assertion later if the def reaches other blocks. Disable the transformation if the value live interval extends beyond its def block.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73478 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-16 07:12:58 +00:00
Evan Cheng
2077e18caf ifcvt should ignore cfg where true and false successors are the same.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73423 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-15 21:24:34 +00:00
Evan Cheng
358dec5180 Part 1.
- Change register allocation hint to a pair of unsigned integers. The hint type is zero (which means prefer the register specified as second part of the pair) or entirely target dependent.
- Allow targets to specify alternative register allocation orders based on allocation hint.

Part 2.
- Use the register allocation hint system to implement more aggressive load / store multiple formation.
- Aggressively form LDRD / STRD. These are formed *before* register allocation. It has to be done this way to shorten live interval of base and offset registers. e.g.
v1025 = LDR v1024, 0
v1026 = LDR v1024, 0
=>
v1025,v1026 = LDRD v1024, 0

If this transformation isn't done before allocation, v1024 will overlap v1025 which means it more difficult to allocate a register pair.

- Even with the register allocation hint, it may not be possible to get the desired allocation. In that case, the post-allocation load / store multiple pass must fix the ldrd / strd instructions. They can either become ldm / stm instructions or back to a pair of ldr / str instructions.

This is work in progress, not yet enabled.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73381 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-15 08:28:29 +00:00
Evan Cheng
e7d6df7353 Add a ARM specific pre-allocation pass that re-schedule loads / stores from
consecutive addresses togther. This makes it easier for the post-allocation pass
to form ldm / stm.

This is step 1. We are still missing a lot of ldm / stm opportunities because
of register allocation are not done in the desired order. More enhancements
coming.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73291 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-13 09:12:55 +00:00
Evan Cheng
4a274e573d If killed register is defined by implicit_def, do not clear it since it's live range may overlap another def of same register.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73255 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-12 21:34:26 +00:00
Evan Cheng
cd799b99cb Mark some pattern-less instructions as neverHasSideEffects.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73252 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-12 20:46:18 +00:00
Anton Korobeynikov
f194b0edc9 Add testcase for register scanveger assertion fix in r72755
(double def due to livevars)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73096 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-08 22:54:15 +00:00
Evan Cheng
925492279a Changing allocation ordering from r3 ... r0 back to r0 ... r3. The order change no longer make sense after the coalescing changes we have made since then.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72955 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-05 19:08:58 +00:00
Dan Gohman
ae3a0be92e Split the Add, Sub, and Mul instruction opcodes into separate
integer and floating-point opcodes, introducing
FAdd, FSub, and FMul.

For now, the AsmParser, BitcodeReader, and IRBuilder all preserve
backwards compatability, and the Core LLVM APIs preserve backwards
compatibility for IR producers. Most front-ends won't need to change
immediately.

This implements the first step of the plan outlined here:
http://nondot.org/sabre/LLVMNotes/IntegerOverflow.txt


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72897 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-04 22:49:04 +00:00
Evan Cheng
cd0c4ac7fb A value defined by an implicit_def can be liven to a use BB. This is unfortunate. But register allocator still has to add it to the live-in set of the use BB.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72888 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-04 20:25:48 +00:00
Evan Cheng
1488326156 Re-apply 72756 with fixes. One of those was introduced by we changed MachineInstrBuilder::addReg() interface.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72826 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-04 01:15:28 +00:00
Evan Cheng
70fd60bd57 Temporarily revert 72756 for now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72757 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-03 07:40:47 +00:00
Evan Cheng
9d5fb981b0 Fold preceding / trailing base inc / dec into the single load / store as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72756 91177308-0d34-0410-b5e6-96231b3b80d8
2009-06-03 06:14:58 +00:00
Bob Wilson
04746eae49 Fix pr4058 and pr4059. Do not split i64 or double arguments between r3 and
the stack.  Patch by Sandeep Patel.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72106 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-19 10:02:36 +00:00
Bob Wilson
224c244f56 Fix pr4091: Add support for "m" constraint in ARM inline assembly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72105 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-19 05:53:42 +00:00
Dan Gohman
04623272d2 Add nounwind to a few tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72002 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-18 15:16:49 +00:00
Bob Wilson
74b0ccc577 Fix pr4195: When iterating through predecessor blocks, break out of the loop
after finding the (unique) layout predecessor.  Sometimes a block may be listed
more than once, and processing it more than once in this loop can lead to
inconsistent values for FtTBB/FtFBB, since the AnalyzeBranch method does not
clear these values.  There's no point in continuing the loop regardless.
The testcase for this is reduced from the 2003-05-02-DependentPHI SingleSource
test.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@71536 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-12 03:48:10 +00:00
Bob Wilson
9d928c223f Fix pr4100. Do not remove no-op copies when they are dead. The register
scavenger gets confused about register liveness if it doesn't see them.
I'm not thrilled with this solution, but it only comes up when there are dead
copies in the code, which is something that hopefully doesn't happen much.

Here is what happens in pr4100: As shown in the following excerpt from the
debug output of llc, the source of a move gets reloaded from the stack,
inserting a new load instruction before the move.  Since that source operand
is a kill, the physical register is free to be reused for the destination
of the move.  The move ends up being a no-op, copying R3 to R3, so it is
deleted.  But, it leaves behind the load to reload %reg1028 into R3, and
that load is not updated to show that it's destination operand (R3) is dead.
The scavenger gets confused by that load because it thinks that R3 is live.

Starting RegAlloc of: %reg1025<def,dead> = MOVr %reg1028<kill>, 14, %reg0, %reg0
  Regs have values: 
  Reloading %reg1028 into R3
  Last use of R3[%reg1028], removing it from live set
  Assigning R3 to %reg1025
  Register R3 [%reg1025] is never used, removing it from live set

Alternative solutions might be either marking the load as dead, or zapping
the load along with the no-op copy.  I couldn't see an easy way to do
either of those, though.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@71196 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-07 23:47:03 +00:00
Evan Cheng
caab129cd1 Do not use register as base ptr of pre- and post- inc/dec load / store nodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@71098 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-06 18:25:01 +00:00
Dan Gohman
afc36a9520 Previously, RecursivelyDeleteDeadInstructions provided an option
of returning a list of pointers to Values that are deleted. This was
unsafe, because the pointers in the list are, by nature of what
RecursivelyDeleteDeadInstructions does, always dangling. Replace this
with a simple callback mechanism. This may eventually be removed if
all clients can reasonably be expected to use CallbackVH.

Use this to factor out the dead-phi-cycle-elimination code from LSR
utility function, and generalize it to use the
RecursivelyDeleteTriviallyDeadInstructions utility function.

This makes LSR more aggressive about eliminating dead PHI cycles;
adjust tests to either be less trivial or to simply expect fewer
instructions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@70636 91177308-0d34-0410-b5e6-96231b3b80d8
2009-05-02 18:29:22 +00:00
Bob Wilson
7eb793dd09 Rename file to have the correct suffix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@69380 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-17 20:40:20 +00:00
Bob Wilson
1f595bb429 Use CallConvLower.h and TableGen descriptions of the calling conventions
for ARM.  Patch by Sandeep Patel.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@69371 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-17 19:07:39 +00:00
Dan Gohman
2d1be87ee4 Expand GEPs in ScalarEvolution expressions. SCEV expressions can now
have pointer types, though in contrast to C pointer types, SCEV
addition is never implicitly scaled. This not only eliminates the
need for special code like IndVars' EliminatePointerRecurrence
and LSR's own GEP expansion code, it also does a better job because
it lets the normal optimizations handle pointer expressions just
like integer expressions.

Also, since LLVM IR GEPs can't directly index into multi-dimensional
VLAs, moving the GEP analysis out of client code and into the SCEV
framework makes it easier for clients to handle multi-dimensional
VLAs the same way as other arrays.

Some existing regression tests show improved optimization.
test/CodeGen/ARM/2007-03-13-InstrSched.ll in particular improved to
the point where if-conversion started kicking in; I turned it off
for this test to preserve the intent of the test.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@69258 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-16 03:18:22 +00:00
Dale Johannesen
442b7bfc80 Use the output of the asm so the optimizer won't
delete it.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@69018 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-14 01:51:40 +00:00
Chris Lattner
dbf1e2b08b move a target-specific test into its directory so it isn't run if you
don't configure the ARM target in.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@68843 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-10 23:58:38 +00:00
Bob Wilson
d9df501704 Fix pr3954. The register scavenger asserts for inline assembly with
register destinations that are tied to source operands.  The
TargetInstrDescr::findTiedToSrcOperand method silently fails for inline
assembly.  The existing MachineInstr::isRegReDefinedByTwoAddr was very
close to doing what is needed, so this revision makes a few changes to
that method and also renames it to isRegTiedToUseOperand (for consistency
with the very similar isRegTiedToDefOperand and because it handles both
two-address instructions and inline assembly with tied registers).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@68714 91177308-0d34-0410-b5e6-96231b3b80d8
2009-04-09 17:16:43 +00:00