Commit Graph

102 Commits

Author SHA1 Message Date
Quentin Colombet 0a905042cd [ARM] Do not generate invalid encoding for stack adjust, even if this is just
temporary.

Because of that:
1. The machine verifier was complaining on such code.
2. The generate code worked just because the thumb reduction size pass fixed the
opcode.

rdar://problem/20749824

llvm-svn: 236247
2015-04-30 18:52:49 +00:00
Eric Christopher 34085832f8 Remove the need to cache the subtarget in the ARM TargetRegisterInfo
classes. Replace the frame pointer initialization with a static function
that'll look it up via the subtarget on the MachineFunction.

llvm-svn: 232010
2015-03-12 05:12:31 +00:00
Bob Wilson 8e29dec986 Fix handling of negative offsets for AddrModeT2_i8s4 in rewriteT2FrameIndex.
This is a follow up to r230233 to fix something that I noticed by
inspection. The AddrModeT2_i8s4 addressing mode does not support
negative offsets. I spent a good chunk of the day trying to come up with
a testcase for this but was not successful. This addressing mode is used
to spill and restore GPRPair registers in Thumb2 code and that does not
happen often. We also make very limited used of negative offsets when
lowering frame indexes. I am going ahead with the change anyway, because
I am pretty confident that it is correct. I also added a missing assertion
to check that the low bits of the scaled offset are zero.

llvm-svn: 230297
2015-02-24 01:37:31 +00:00
Bob Wilson 89e94fc3ad Fix incorrect immediate size for AddrModeT2_i8s4 in rewriteT2FrameIndex.
The natural way to handle this addressing mode would be to say that it has
8 bits and gets scaled by 4, but since the MC layer is expecting the scaling
to be already reflected in the immediate value, we have been setting the
Scale to 1. That's fine, but then NumBits needs to be adjusted to reflect
the effective increase in the range of the immediate. That adjustment was
missing.

The consequence is that the register scavenger can fail.
The estimateRSStackSizeLimit() function in ARMFrameLowering.cpp correctly
assumes that the AddrModeT2_i8s4 address mode can handle scaled offsets up to
1020. Under just the right circumstances, we fail to reserve space for the
scavenger because it thinks that nothing will be needed. However, the overly
pessimistic behavior in rewriteT2FrameIndex causes some frame indexes to be
out of range and require scavenged registers, and so the scavenger asserts.

Unfortunately I have not been able to come up with a testcase for this. I
can only reproduce it on an internal branch where the frame layout and
register allocation is slightly different than trunk. We really need a
way to serialize MachineInstr-level IR to write reasonable tests for things
like this.

rdar://problem/19909005

llvm-svn: 230233
2015-02-23 16:57:19 +00:00
Akira Hatanaka dc08c30df9 [ARM] In dynamic-no-pic mode, ARM's post-RA pseudo expansion was incorrectly
expanding pseudo LOAD_STATCK_GUARD using instructions that are normally used
in pic mode. This patch fixes the bug.

<rdar://problem/17886592>

llvm-svn: 214614
2014-08-02 05:40:40 +00:00
Akira Hatanaka e5b6e0d231 [stack protector] Fix a potential security bug in stack protector where the
address of the stack guard was being spilled to the stack.

Previously the address of the stack guard would get spilled to the stack if it
was impossible to keep it in a register. This patch introduces a new target
independent node and pseudo instruction which gets expanded post-RA to a
sequence of instructions that load the stack guard value. Register allocator
can now just remat the value when it can't keep it in a register. 

<rdar://problem/12475629>

llvm-svn: 213967
2014-07-25 19:31:34 +00:00
Craig Topper a9253267a9 Prune includes in ARM target.
llvm-svn: 204548
2014-03-22 23:51:00 +00:00
Tim Northover c9432eb9e5 ARM: remove unnecessary state-tracking during frame lowering.
ResolveFrameIndex had what appeared to be a very nasty hack for when the
frame-index referred to a callee-saved register. In this case it "adjusted" the
offset so that the address was correct if (and only if) the MachineInstr
immediately followed the respective push.

This "worked" for all forms of GPR & DPR but was only ever used to set the
frame pointer itself, and once this was put in a more sensible location the
entire state-tracking machinery it relied on became redundant. So I stripped
it.

The only wrinkle is that "add r7, sp, #0" might theoretically be slower (need
an actual ALU slot) compared to "mov r7, sp" so I added a micro-optimisation
that also makes emitARMRegUpdate and emitT2RegUpdate also work when NumBytes ==
0.

No test changes since there shouldn't be any functionality change.

llvm-svn: 194025
2013-11-04 23:04:15 +00:00
Richard Barton 87dacc38b8 Add hint disassembly syntax for 16-bit Thumb hint instructions.
Patch by Artyom Skrobov

llvm-svn: 192972
2013-10-18 14:09:49 +00:00
Weiming Zhao 286304a317 Fix PR 17372: Emitting PLD for stack address for ARM Thumb2
t2PLDi12, t2PLDi8, t2PLDs was omitted in Thumb2InstrInfo.
This patch fixes it.

llvm-svn: 191441
2013-09-26 17:25:10 +00:00
Tilmann Scheller 841a9ccfed Reverting 190043 for now.
Solution is not sufficient to prevent 'mov pc, lr' being emitted for jump table code.
Test case doesn't trigger the added functionality.

llvm-svn: 190047
2013-09-05 11:59:43 +00:00
Tilmann Scheller a1787a5835 ARM: Add GPR register class excluding LR for use with the ADR instruction.
This improves code generation for jump tables by avoiding the emission of "mov pc, lr" which could fool the processor into believing this is a return from a function causing mispredicts. The code generation logic for jump tables uses ADR to materialize the address of the jump target.

Patch by Daniel Stewart!
   

llvm-svn: 190043
2013-09-05 11:10:31 +00:00
Bill Wendling f95178e679 Don't cache the instruction and register info from the TargetMachine, because
the internals of TargetMachine could change.

llvm-svn: 183488
2013-06-07 05:54:19 +00:00
Michael J. Spencer df1ecbd734 Replace Count{Leading,Trailing}Zeros_{32,64} with count{Leading,Trailing}Zeros.
llvm-svn: 182680
2013-05-24 22:23:49 +00:00
Tim Northover 798697d662 ARM: Use ldrd/strd to spill 64-bit pairs when available.
This allows common sp-offsets to be part of the instruction and is
probably faster on modern CPUs too.

llvm-svn: 179977
2013-04-21 11:57:07 +00:00
Jakob Stoklund Olesen b159b5ff0d Remove the explicit MachineInstrBuilder(MI) constructor.
Use the version that also takes an MF reference instead.

It would technically be possible to extract an MF reference from the MI
as MI->getParent()->getParent(), but that would not work for MIs that
are not inserted into any basic block.

Given the reasonably small number of places this constructor was used at
all, I preferred the compile time check to a run time assertion.

llvm-svn: 170588
2012-12-19 21:31:56 +00:00
Jakob Stoklund Olesen 9de596e650 Remove all references to TargetInstrInfoImpl.
This class has been merged into its super-class TargetInstrInfo.

llvm-svn: 168760
2012-11-28 02:35:17 +00:00
Jakob Stoklund Olesen 702bcc3bcf Remove the TII::scheduleTwoAddrSource() hook.
It never does anything when running 'make check', and it get's in the
way of updating live intervals in 2-addr.

The hook was originally added to help form IT blocks in Thumb2 code
before register allocation, but the pass ordering has changed since
then, and we run if-conversion after register allocation now.

When the MI scheduler is enabled, there will be no less than two
schedulers between 2-addr and Thumb2ITBlockPass, so this hook is
unlikely to help anything.

llvm-svn: 161794
2012-08-13 21:52:57 +00:00
Craig Topper c7242e054d Convert more uses of XXXRegisterClass to &XXXRegClass. No functional change since they are equivalent.
llvm-svn: 155188
2012-04-20 07:30:17 +00:00
Craig Topper 1fcf5bcae1 Prune some includes
llvm-svn: 153502
2012-03-27 07:54:11 +00:00
Craig Topper f6e7e12f75 Remove unnecessary llvm:: qualifications
llvm-svn: 153500
2012-03-27 07:21:54 +00:00
Craig Topper 188ed9d56e Reorder includes to match coding standards. Fix an issue or two exposed by that.
llvm-svn: 152978
2012-03-17 07:33:42 +00:00
Jim Grosbach 617f84ddbd ARM implement TargetInstrInfo::getNoopForMachoTarget()
Without this hook, functions w/ a completely empty body (including no
epilogue) will cause an MCEmitter assertion failure.

For example,
define internal fastcc void @empty_function() {
  unreachable
}

rdar://10947471

llvm-svn: 151673
2012-02-28 23:53:30 +00:00
Jia Liu b22310fda6 Emacs-tag and some comment fix for all ARM, CellSPU, Hexagon, MBlaze, MSP430, PPC, PTX, Sparc, X86, XCore.
llvm-svn: 150878
2012-02-18 12:03:15 +00:00
Jakob Stoklund Olesen 4fad5b2b9e Handle regmask operands in ARMInstrInfo.
llvm-svn: 150833
2012-02-17 19:23:15 +00:00
Jay Foad 465101bb0e Make use of MachinePointerInfo::getFixedStack. This removes all mention
of PseudoSourceValue from lib/Target/.

llvm-svn: 144632
2011-11-15 07:34:52 +00:00
Jim Grosbach 1b8457a84c Thumb1 ADD/SUB SP instructions are predicable in Thumb2 mode.
Add the predicate operand to the instructions. Update the back end
accordingly where the instructions are used. Restrict the SP operands
to actually only be SP, as otherwise these break assembly parsing for the
normal instruction variants.

llvm-svn: 138445
2011-08-24 17:46:13 +00:00
Owen Anderson 12d13efa21 Handle new register classes in Thumb2 mode. Should fix the ARM buildbots.
llvm-svn: 137364
2011-08-11 21:52:38 +00:00
Evan Cheng a20cde31e7 Sink ARMMCExpr and ARMAddressingModes into MC layer. First step to separate ARM MC code from target.
llvm-svn: 135636
2011-07-20 23:34:39 +00:00
Jim Grosbach e9cc901814 Refact ARM Thumb1 tMOVr instruction family.
Merge the tMOVr, tMOVgpr2tgpr, tMOVtgpr2gpr, and tMOVgpr2gpr instructions
into tMOVr. There's no need to keep them separate. Giving the tMOVr
instruction the proper GPR register class for its operands is sufficient
to give the register allocator enough information to do the right thing
directly.

llvm-svn: 134204
2011-06-30 23:38:17 +00:00
Jim Grosbach b98ab91e39 Thumb1 register to register MOV instruction is predicable.
Fix a FIXME and allow predication (in Thumb2) for the T1 register to
register MOV instructions. This allows some better codegen with
if-conversion (as seen in the test updates), plus it lays the groundwork
for pseudo-izing the tMOVCC instructions.

llvm-svn: 134197
2011-06-30 22:10:46 +00:00
Jim Grosbach cfe3b14d77 Kill dead code.
llvm-svn: 134131
2011-06-30 02:23:05 +00:00
Jim Grosbach a8a8067dec Remove redundant Thumb2 ADD/SUB SP instruction definitions.
Unlike Thumb1, Thumb2 does not have dedicated encodings for adjusting the
stack pointer. It can just use the normal add-register-immediate encoding
since it can use all registers as a source, not just R0-R7. The extra
instruction definitions are just duplicates of the normal instructions with
the (not well enforced) constraint that the source register was SP.

llvm-svn: 134114
2011-06-29 23:25:04 +00:00
Evan Cheng 1e210d08d8 Merge XXXGenRegisterNames.inc into XXXGenRegisterInfo.inc
llvm-svn: 134024
2011-06-28 20:07:07 +00:00
Evan Cheng 6cc775f905 - Rename TargetInstrDesc, TargetOperandInfo to MCInstrDesc and MCOperandInfo and
sink them into MC layer.
- Added MCInstrInfo, which captures the tablegen generated static data. Chang
TargetInstrInfo so it's based off MCInstrInfo.

llvm-svn: 134021
2011-06-28 19:10:37 +00:00
Anton Korobeynikov e7410dd0d5 Preliminary support for ARM frame save directives emission via MI flags.
This is just very first approximation how the stuff should be done
(e.g. ARM-only for now). More to follow.

llvm-svn: 127101
2011-03-05 18:43:32 +00:00
Evan Cheng 666cf56668 Guard against de-referencing MBB.end().
llvm-svn: 126192
2011-02-22 07:07:59 +00:00
Evan Cheng 87a9f19f9c Skipping over debugvalue instructions to determine whether the split spot is in a IT block. rdar://9030770
llvm-svn: 126159
2011-02-21 23:40:47 +00:00
Evan Cheng 62c7b5bf76 Making use of VFP / NEON floating point multiply-accumulate / subtraction is
difficult on current ARM implementations for a few reasons.
1. Even though a single vmla has latency that is one cycle shorter than a pair
   of vmul + vadd, a RAW hazard during the first (4? on Cortex-a8) can cause
   additional pipeline stall. So it's frequently better to single codegen
   vmul + vadd.
2. A vmla folowed by a vmul, vmadd, or vsub causes the second fp instruction to
   stall for 4 cycles. We need to schedule them apart.
3. A vmla followed vmla is a special case. Obvious issuing back to back RAW
   vmla + vmla is very bad. But this isn't ideal either:
     vmul
     vadd
     vmla
   Instead, we want to expand the second vmla:
     vmla
     vmul
     vadd
   Even with the 4 cycle vmul stall, the second sequence is still 2 cycles
   faster.

Up to now, isel simply avoid codegen'ing fp vmla / vmls. This works well enough
but it isn't the optimial solution. This patch attempts to make it possible to
use vmla / vmls in cases where it is profitable.

A. Add missing isel predicates which cause vmla to be codegen'ed.
B. Make sure the fmul in (fadd (fmul)) has a single use. We don't want to
   compute a fmul and a fmla.
C. Add additional isel checks for vmla, avoid cases where vmla is feeding into
   fp instructions (except for the #3 exceptional case).
D. Add ARM hazard recognizer to model the vmla / vmls hazards.
E. Add a special pre-regalloc case to expand vmla / vmls when it's likely the
   vmla / vmls will trigger one of the special hazards.

Work in progress, only A+B are enabled.

llvm-svn: 120960
2010-12-05 22:04:16 +00:00
Evan Cheng debf9c502a Two sets of changes. Sorry they are intermingled.
1. Fix pre-ra scheduler so it doesn't try to push instructions above calls to
   "optimize for latency". Call instructions don't have the right latency and
   this is more likely to use introduce spills.
2. Fix if-converter cost function. For ARM, it should use instruction latencies,
   not # of micro-ops since multi-latency instructions is completely executed
   even when the predicate is false. Also, some instruction will be "slower"
   when they are predicated due to the register def becoming implicit input.
   rdar://8598427

llvm-svn: 118135
2010-11-03 00:45:17 +00:00
Owen Anderson f31f33ea89 Thread the determination of branch prediction hit rates back through the if-conversion heuristic APIs. For now,
stick with a constant estimate of 90% (branch predictors are good!), but we might find that we want to provide
more nuanced estimates in the future.

llvm-svn: 115364
2010-10-01 22:45:50 +00:00
Owen Anderson 671d57865e Provide an option to restore old-style if-conversion heuristics for Thumb2.
llvm-svn: 115339
2010-10-01 20:28:06 +00:00
Owen Anderson 88af7d00fc Part one of switching to using a more sane heuristic for determining if-conversion profitability.
Rather than having arbitrary cutoffs, actually try to cost model the conversion.

For now, the constants are tuned to more or less match our existing behavior, but these will be
changed to reflect realistic values as this work proceeds.

llvm-svn: 114973
2010-09-28 18:32:13 +00:00
Chris Lattner e3d864b857 convert targets to the new MF.getMachineMemOperand interface.
llvm-svn: 114391
2010-09-21 04:39:43 +00:00
Evan Cheng bf4070756f Teach if-converter to be more careful with predicating instructions that would
take multiple cycles to decode.
For the current if-converter clients (actually only ARM), the instructions that
are predicated on false are not nops. They would still take machine cycles to
decode. Micro-coded instructions such as LDM / STM can potentially take multiple
cycles to decode. If-converter should take treat them as non-micro-coded
simple instructions.

llvm-svn: 113570
2010-09-10 01:29:16 +00:00
Jim Grosbach d343166a0b Many Thumb2 instructions can reference the full ARM register set (i.e.,
have 4 bits per register in the operand encoding), but have undefined
behavior when the operand value is 13 or 15 (SP and PC, respectively).
The trivial coalescer in linear scan sometimes will merge a copy from
SP into a subsequent instruction which uses the copy, and if that
instruction cannot legally reference SP, we get bad code such as:
  mls r0,r9,r0,sp
instead of:
  mov r2, sp
  mls r0, r9, r0, r2

This patch adds a new register class for use by Thumb2 that excludes
the problematic registers (SP and PC) and is used instead of GPR
for those operands which cannot legally reference PC or SP. The
trivial coalescer explicitly requires that the register class
of the destination for the COPY instruction contain the source
register for the COPY to be considered for coalescing. This prevents
errant instructions like that above.

PR7499

llvm-svn: 109842
2010-07-30 02:41:01 +00:00
Jakob Stoklund Olesen d7b33002dd Replace copyRegToReg with copyPhysReg for ARM.
llvm-svn: 108078
2010-07-11 06:33:54 +00:00
Bob Wilson 83b993a977 The t2MOVi16 and t2MOVTi16 instructions do not set CPSR. Trying to add
a CPSR operand to them causes an assertion failure, so apparently these
instructions haven't been getting a lot of use.

llvm-svn: 107147
2010-06-29 16:25:11 +00:00
Evan Cheng 0c30739cbb Change if-cvt options to something that actually as useable.
llvm-svn: 107121
2010-06-29 05:37:59 +00:00
Evan Cheng 02b184de5b Change if-conversion block size limit checks to add some flexibility.
llvm-svn: 106901
2010-06-25 22:42:03 +00:00