Commit Graph

49 Commits

Author SHA1 Message Date
Bruno Cardoso Lopes
255fda615d Reapply r105521, this time appending "LLU" to 64 bit
immediates to avoid breaking the build.

llvm-svn: 105652
2010-06-08 22:51:23 +00:00
Chris Lattner
33d0622cdc revert r105521, which is breaking the buildbots with stuff like this:
In file included from X86InstrInfo.cpp:16:
X86GenInstrInfo.inc:2789: error: integer constant is too large for 'long' type
X86GenInstrInfo.inc:2790: error: integer constant is too large for 'long' type
X86GenInstrInfo.inc:2792: error: integer constant is too large for 'long' type
X86GenInstrInfo.inc:2793: error: integer constant is too large for 'long' type
X86GenInstrInfo.inc:2808: error: integer constant is too large for 'long' type
X86GenInstrInfo.inc:2809: error: integer constant is too large for 'long' type
X86GenInstrInfo.inc:2816: error: integer constant is too large for 'long' type
X86GenInstrInfo.inc:2817: error: integer constant is too large for 'long' type

llvm-svn: 105524
2010-06-05 04:17:30 +00:00
Bruno Cardoso Lopes
b05131d907 Initial AVX support for some instructions. No patterns matched
yet, only assembly encoding support.

llvm-svn: 105521
2010-06-05 03:53:24 +00:00
Jakob Stoklund Olesen
83d2cfd6cd Replace the SubRegSet tablegen class with a less error-prone mechanism.
A Register with subregisters must also provide SubRegIndices for adressing the
subregisters. TableGen automatically inherits indices for sub-subregisters to
minimize typing.

CompositeIndices may be specified for the weirder cases such as the XMM sub_sd
index that returns the same register, and ARM NEON Q registers where both D
subregs have ssub_0 and ssub_1 sub-subregs.

It is now required that all subregisters are named by an index, and a future
patch will also require inherited subregisters to be named. This is necessary to
allow composite subregister indices to be reduced to a single index.

llvm-svn: 104704
2010-05-26 17:27:12 +00:00
Jakob Stoklund Olesen
0fefdf4d2a Revert "Replace the SubRegSet tablegen class with a less error-prone mechanism."
This reverts commit 104654.

llvm-svn: 104660
2010-05-26 01:21:14 +00:00
Jakob Stoklund Olesen
a2f0c34e41 Replace the SubRegSet tablegen class with a less error-prone mechanism.
A Register with subregisters must also provide SubRegIndices for adressing the
subregisters. TableGen automatically inherits indices for sub-subregisters to
minimize typing.

CompositeIndices may be specified for the weirder cases such as the XMM sub_sd
index that returns the same register, and ARM NEON Q registers where both D
subregs have ssub_0 and ssub_1 sub-subregs.

It is now required that all subregisters are named by an index, and a future
patch will also require inherited subregisters to be named. This is necessary to
allow composite subregister indices to be reduced to a single index.

llvm-svn: 104654
2010-05-26 00:28:19 +00:00
Jakob Stoklund Olesen
696fbed514 Remove NumberHack entirely.
SubRegIndex instances are now numbered uniquely the same way Register instances
are - in lexicographical order by name.

llvm-svn: 104627
2010-05-25 19:49:33 +00:00
Jakob Stoklund Olesen
d1f4d14609 Switch SubRegSet to using symbolic SubRegIndices
llvm-svn: 104571
2010-05-24 23:03:18 +00:00
Jakob Stoklund Olesen
3a19b732d8 Replace the tablegen RegisterClass field SubRegClassList with an alist-like data
structure that represents a mapping without any dependencies on SubRegIndex
numbering.

This brings us closer to being able to remove the explicit SubRegIndex
numbering, and it is now possible to specify any mapping without inventing
*_INVALID register classes.

llvm-svn: 104563
2010-05-24 21:46:58 +00:00
Jakob Stoklund Olesen
9a54fec092 Add the SubRegIndex TableGen class.
This is the beginning of purely symbolic subregister indices, but we need a bit
of jiggling before the explicit numeric indices can be completely removed.

llvm-svn: 104492
2010-05-24 14:48:12 +00:00
Daniel Dunbar
ee525943d8 tblgen/AsmMatcher: Change AsmOperandClass to allow a list of superclasses instead of just one.
llvm-svn: 104452
2010-05-22 21:02:29 +00:00
Daniel Dunbar
64807873ec tblgen/Target: Add a isAsmParserOnly bit, and teach the disassembler to honor
it.

llvm-svn: 104270
2010-05-20 20:20:32 +00:00
Daniel Dunbar
711d2427dd MC/Matcher: Add support for over-riding the default MatchInstruction function
name (for example, to allow targets to interpose the actual MatchInstruction
function).

llvm-svn: 102987
2010-05-04 00:33:13 +00:00
Evan Cheng
8bd004b033 Add a pseudo instruction REG_SEQUENCE that takes a list of registers and
sub-register indices and outputs a single super register which is formed from
a consecutive sequence of registers.

This is used as register allocation / coalescing aid and it is useful to
represent instructions that output register pairs / quads. For example,
v1024, v1025 = vload <address>
where v1024 and v1025 forms a register pair.

This really should be modelled as
v1024<3>, v1025<4> = vload <address>
but it would violate SSA property before register allocation is done.

Currently we use insert_subreg to form the super register:
v1026 = implicit_def
v1027 - insert_subreg v1026, v1024, 3
v1028 = insert_subreg v1027, v1025, 4
...
      = use v1024
      = use v1028

But this adds pseudo live interval overlap between v1024 and v1025.

We can now modeled it as
v1024, v1025 = vload <address>
v1026 = REG_SEQUENCE v1024, 3, v1025, 4
...
      = use v1024
      = use v1026

After coalescing, it will be
v1026<3>, v1025<4> = vload <address>
...
      = use v1026<3>
      = use v1026

llvm-svn: 102815
2010-05-01 00:28:44 +00:00
Chris Lattner
248e65ec05 trim some spurious references to DwarfWriter. SDIsel really doesn't
need it anymore, so don't addRequire it.

llvm-svn: 100400
2010-04-05 04:09:20 +00:00
Jakob Stoklund Olesen
4c043c50fd Replace TSFlagsFields and TSFlagsShifts with a simpler TSFlags field.
When a target instruction wants to set target-specific flags, it should simply
set bits in the TSFlags bit vector defined in the Instruction TableGen class.

This works well because TableGen resolves member references late:

class I : Instruction {
  AddrMode AM = AddrModeNone;
  let TSFlags{3-0} = AM.Value;
}

let AM = AddrMode4 in
def ADD : I;

TSFlags gets the expected bits from AddrMode4 in this example.

llvm-svn: 100384
2010-04-05 03:10:20 +00:00
Chris Lattner
d678e362d2 use ins/outs
llvm-svn: 98867
2010-03-18 20:55:31 +00:00
Daniel Dunbar
571a3eb0b6 MC/AsmMatcher: Add support for target specific "instruction cleanup" functions,
to allow custom post-processing of matched instructions.

llvm-svn: 98857
2010-03-18 20:05:56 +00:00
Dan Gohman
99c6bcbc13 The mayHaveSideEffects flag is no longer used.
llvm-svn: 97348
2010-02-27 23:47:46 +00:00
Dan Gohman
92b6122204 Fix "the the" and similar typos.
llvm-svn: 95781
2010-02-10 16:03:48 +00:00
Chris Lattner
7acf9be6c4 move target-independent opcodes out of TargetInstrInfo
into TargetOpcodes.h.  #include the new TargetOpcodes.h
into MachineInstr.  Add new inline accessors (like isPHI())
to MachineInstr, and start using them throughout the 
codebase.

llvm-svn: 95687
2010-02-09 19:54:29 +00:00
Dale Johannesen
7a3b1c5ffc DEBUG_VALUE is now variable sized, as it has a
target-dependent memory address representation in it.
Restore X86 printing of DEBUG_VALUE; lowering is
done in X86RegisterInfo using the normal algorithm.

llvm-svn: 93565
2010-01-15 22:22:35 +00:00
Dale Johannesen
c396179c7e Remove DEBUG_DECLARE, looks like we don't need it.
Also, DEBUG_VALUE has side effects.

llvm-svn: 93498
2010-01-15 01:50:44 +00:00
Dale Johannesen
0e21161e32 Add DEBUG_DECLARE. Not used yet.
llvm-svn: 93040
2010-01-09 01:24:25 +00:00
Dale Johannesen
610f0ce9aa Add DEBUG_VALUE. Not used yet.
llvm-svn: 93030
2010-01-08 23:51:25 +00:00
Dan Gohman
b15f87cc4f Mark DBG_LABEL, EH_LABEL, and GC_LABEL as not-duplicable, since
they really are not duplicable.

llvm-svn: 87009
2009-11-12 18:36:19 +00:00
Dan Gohman
3393a4c997 Rename usesCustomDAGSchedInserter to usesCustomInserter, and update a
bunch of associated comments, because it doesn't have anything to do
with DAGs or scheduling. This is another step in decoupling MachineInstr
emitting from scheduling.

llvm-svn: 85517
2009-10-29 18:10:34 +00:00
Evan Cheng
8ea3f4a592 Add instruction flags: hasExtraSrcRegAllocReq and hasExtraDefRegAllocReq. When
set, these flags indicate the instructions source / def operands have special
register allocation requirement that are not captured in their register classes.
Post-allocation passes (e.g. post-alloc scheduler) should not change their
allocations. e.g. ARM::LDRD require the two definitions to be allocated
even / odd register pair.

llvm-svn: 83196
2009-10-01 08:21:18 +00:00
Jakob Stoklund Olesen
31fcbdefbb Introduce the TargetInstrInfo::KILL machine instruction and get rid of the
unused DECLARE instruction.

KILL is not yet used anywhere, it will replace TargetInstrInfo::IMPLICIT_DEF
in the places where IMPLICIT_DEF is just used to alter liveness of physical
registers.

llvm-svn: 83006
2009-09-28 20:32:26 +00:00
Daniel Dunbar
63f93255ae Add 'isCodeGenOnly' bit to Instruction .td records.
- Used to mark fake instructions which don't correspond to an actual machine
   instruction (or are duplicates of a real instruction). This is to be used for
   "special cases" in the .td files, which should be ignored by things like the
   assembler and disassembler. We still need a good solution to handle pervasive
   duplication, like with the Int_ instructions.

 - Set the bit on fake "mov 0" style instructions, which allows turning an
   assembler matcher warning into a hard error.

 - -2 FIXMEs.

llvm-svn: 78731
2009-08-11 22:17:52 +00:00
Daniel Dunbar
186ba504b3 llvm-mc/AsmParser: Allow target to specific a comment delimiter, which will be
used to strip hard coded comments out of .td assembly strings.

llvm-svn: 78716
2009-08-11 20:59:47 +00:00
Daniel Dunbar
15c79b6fd7 llvm-mc/AsmParser: Allow .td users to redefine the names of the methods to call
on target specific operands for testing class membership and converting to
MCInst operands.

llvm-svn: 78597
2009-08-10 21:00:45 +00:00
Daniel Dunbar
749ff1de5a llvm-mc/AsmMatcher: Change assembler parser match classes to their own record
structure.

llvm-svn: 78581
2009-08-10 18:41:10 +00:00
Daniel Dunbar
5a6c69d37d Extend comment on ParserMatchClass .td field, and add some missing
classes for X86.

llvm-svn: 78524
2009-08-09 06:00:04 +00:00
Daniel Dunbar
dff8502076 llvm-mc/AsmParser: Define match classes in the .td file.
-2 FIXMEs.

llvm-svn: 78523
2009-08-09 05:18:30 +00:00
Chris Lattner
58ae68cdc2 fix the column output stuff in the asmwriter from being dynamic and
driven by TAI to being static, driven by tblgen.  This means that a
target doesn't get impacted by this stuff at all if it doesn't opt
into it.

llvm-svn: 78427
2009-08-07 23:13:38 +00:00
Chris Lattner
7205f6c951 fix comment pastos
llvm-svn: 78422
2009-08-07 22:44:56 +00:00
Chris Lattner
6c284cc8cd 1. Introduce a new TargetOperandInfo::getRegClass() helper method
and convert code to using it, instead of having lots of things
   poke the isLookupPtrRegClass() method directly.

2. Make PointerLikeRegClass contain a 'kind' int, and store it in
   the existing regclass field of TargetOperandInfo when the
   isLookupPtrRegClass() predicate is set.  Make getRegClass pass
   this into TargetRegisterInfo::getPointerRegClass(), allowing
   targets to have multiple ptr_rc things.

llvm-svn: 77504
2009-07-29 21:10:12 +00:00
Chris Lattner
adc9ad97f7 make ptr_rc derive from a new PointerLikeRegClass tblgen class.
llvm-svn: 77503
2009-07-29 20:43:05 +00:00
Daniel Dunbar
a364079155 Match X86 register names to number.
llvm-svn: 77404
2009-07-29 00:02:19 +00:00
Evan Cheng
e25ef2e60e 80 col violation.
llvm-svn: 75226
2009-07-10 05:20:19 +00:00
Chris Lattner
7b43ca847d implement support for lowering subregs when preparing to print
LEA64_32r, eliminating a bunch of modifier logic stuff on addr modes.

Implement support for printing mbb labels as operands.

llvm-svn: 73817
2009-06-20 07:03:18 +00:00
Dale Johannesen
8b6ee9e312 Revert 72707 and 72709, for the moment.
llvm-svn: 72712
2009-06-02 03:12:52 +00:00
Dale Johannesen
c08669561e Make the implicit inputs and outputs of target-independent
ADDC/ADDE use MVT::i1 (later, whatever it gets legalized to)
instead of MVT::Flag.  Remove CARRY_FALSE in favor of 0; adjust
all target-independent code to use this format.

Most targets will still produce a Flag-setting target-dependent
version when selection is done.  X86 is converted to use i32
instead, which means TableGen needs to produce different code
in xxxGenDAGISel.inc.  This keys off the new supportsHasI1 bit
in xxxInstrInfo, currently set only for X86; in principle this
is temporary and should go away when all other targets have
been converted.  All relevant X86 instruction patterns are
modified to represent setting and using EFLAGS explicitly.  The
same can be done on other targets.

The immediate behavior change is that an ADC/ADD pair are no
longer tightly coupled in the X86 scheduler; they can be
separated by instructions that don't clobber the flags (MOV).
I will soon add some peephole optimizations based on using
other instructions that set the flags to feed into ADC.

llvm-svn: 72707
2009-06-01 23:27:20 +00:00
Dan Gohman
8393d29bc8 Rename COPY_TO_SUBCLASS to COPY_TO_REGCLASS, and generalize
it accordingly. Thanks to Jakob Stoklund Olesen for pointing
out how this might be useful.

llvm-svn: 68986
2009-04-13 21:06:25 +00:00
Dan Gohman
3873cb7a36 Add a new TargetInstrInfo MachineInstr opcode, COPY_TO_SUBCLASS.
This will be used to replace things like X86's MOV32to32_.

Enhance ScheduleDAGSDNodesEmit to be more flexible and robust
in the presense of subregister superclasses and subclasses. It
can now cope with the definition of a virtual register being in
a subclass of a use.

Re-introduce the code for recording register superreg classes and
subreg classes. This is needed because when subreg extracts and
inserts get coalesced away, the virtual registers are left in
the correct subclass.

llvm-svn: 68961
2009-04-13 15:38:05 +00:00
Dan Gohman
5dad0993a9 Rename isSimpleLoad to canFoldAsLoad, to better reflect its meaning.
llvm-svn: 60487
2008-12-03 18:15:48 +00:00
Dan Gohman
ac6561793c Mark x86's V_SET0 and V_SETALLONES with isSimpleLoad, and teach X86's
foldMemoryOperand how to "fold" them, by converting them into constant-pool
loads. When they aren't folded, they use xorps/cmpeqd, but for example when
register pressure is high, they may now be folded as memory operands, which
reduces register pressure.

Also, mark V_SET0 isAsCheapAsAMove so that two-address-elimination will
remat it instead of copying zeros around (V_SETALLONES was already marked).

llvm-svn: 60461
2008-12-03 05:21:24 +00:00
Evan Cheng
a19ef59d6c Move target independent td files from lib/Target/ to include/llvm/Target so they can be distributed along with the header files.
llvm-svn: 59953
2008-11-24 07:34:46 +00:00