During type legalization we often use the SIGN_EXTEND_INREG SDNode.
When this SDNode is legalized during the LegalizeVector phase, it is
scalarized because non-simple types are automatically marked to be expanded.
In this patch we add support for lowering SIGN_EXTEND_INREG manually.
This fixes CodeGen/X86/vec_sext.ll when running with the '-promote-elements'
flag.
llvm-svn: 135144
We have to do this in DAGBuilder instead of DAGCombiner, because the exact bit is lost after building.
struct foo { char x[24]; };
long bar(struct foo *a, struct foo *b) { return a-b; }
is now compiled into
movl 4(%esp), %eax
subl 8(%esp), %eax
sarl $3, %eax
imull $-1431655765, %eax, %eax
instead of
movl 4(%esp), %eax
subl 8(%esp), %eax
movl $715827883, %ecx
imull %ecx
movl %edx, %eax
shrl $31, %eax
sarl $2, %edx
addl %eax, %edx
movl %edx, %eax
llvm-svn: 134695
hasPredecessorHelper function allows predecessors to be cached to speed up
repeated invocations. This fixes PR10186.
X.isPredecessorOf(Y) now just calls Y.hasPredecessor(X)
Y.hasPredecessor(X) calls Y.hasPredecessorHelper(X, Visited, Worklist) with
empty Visited and Worklist sets (i.e. no caching over invocations).
Y.hasPredecessorHelper(X, Visited, Worklist) caches search state in Visited
and Worklist to speed up repeated calls. The Visited set is searched for X
before going to the worklist to further search the DAG if necessary.
llvm-svn: 134592
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
Removed the check that peeks past EXTRA_SUBREG, which I don't think
makes sense any more. Intead treat it as a normal register def. No
significant affect on x86 or ARM benchmarks.
llvm-svn: 133917
Both become <earlyclobber> defs on the INLINEASM MachineInstr, but we
now use two different asm operand kinds.
The new Kind_Clobber is treated identically to the old
Kind_RegDefEarlyClobber for now, but x87 floating point stack inline
assembly does care about the difference.
This will pop a register off the stack:
asm("fstp %st" : : "t"(x) : "st");
While this will pop the input and push an output:
asm("fst %st" : "=&t"(r) : "t"(x));
We need to know if ST0 was a clobber or an output operand, and we can't
depend on <dead> flags for that.
llvm-svn: 133902
1. (((x) & 0xFF00) >> 8) | (((x) & 0x00FF) << 8)
=> (bswap x) >> 16
2. ((x&0xff)<<8)|((x&0xff00)>>8)|((x&0xff000000)>>8)|((x&0x00ff0000)<<8))
=> (rotl (bswap x) 16)
This allows us to eliminate most of the def : Pat patterns for ARM rev16
revsh instructions. It catches many more cases for ARM and x86.
rdar://9609108
llvm-svn: 133503
In Thumb mode we cannot handle GPR virtual registers, even though some
instructions can. When isel is lowering a CopyFromReg, it should limit
itself to subclasses of getRegClassFor(VT).
<rdar://problem/9624323>
llvm-svn: 133210
BranchProbabilityInfo (expect setEdgeWeight which is not available here).
Branch Weights are kept in MachineBasicBlocks. To turn off this analysis
set -use-mbpi=false.
llvm-svn: 133184
This is intended to support using REG_SEQUENCE SDNode's with type MVT::untyped, and is part of the long road to eliminating some of the hacks we currently use to support register pairs and other strange constraints, particularly on ARM NEON.
llvm-svn: 133178
This virtual function will replace allocation_order_begin/end as the one
to override when implementing custom allocation orders. It is simpler to
have one function return an ArrayRef than having two virtual functions
computing different ends of the same array.
Use getRawAllocationOrder() in place of allocation_order_begin() where
it makes sense, but leave some clients that look like they really want
the filtered allocation orders from RegisterClassInfo.
llvm-svn: 133170
GetDemandBits (which must operate on the vector element type).
Fix the a usage of getZeroExtendInReg which must also be done on scalar types.
llvm-svn: 133052
converted to add x,x if x is a undef. add undef, undef does not guarantee
that the resulting low order bit is zero.
Fixes <rdar://problem/9453156> and <rdar://problem/9487392>.
llvm-svn: 133022
types (with power of two types such as 8,16,32 .. 512).
Fix a bug in the integer promotion of bitcast nodes. Enable integer expanding
only if the target of the conversion is an integer (when the type action is
scalarize).
Add handling to the legalization of vector load/store in cases where the saved
vector is integer-promoted.
llvm-svn: 132985
Instead of scalarizing, and doing an element-by-element truncat, use vector
truncate.
Add support for scalarization of vectors: i8 -> <1 x i1> (from Duncan's
testcase).
llvm-svn: 132892
The potential DAGCombine which enforces this more generally messes up some other very fragile patterns, so I'm leaving that alone, at least for now.
llvm-svn: 132809
I've been sitting on this long enough trying to find a test case. I
think the fix should go in now, but I'll keep working on the test case.
llvm-svn: 132701
(only happens when using the -promote-elements option).
The correct legalization order is to first try to promote element. Next, we try
to widen vectors.
llvm-svn: 132648
patch we add a flag to enable a new type legalization decision - to promote
integer elements in vectors. Currently, the rest of the codegen does not support
this kind of legalization. This flag will be removed when the transition is
complete.
llvm-svn: 132394
This patch does not change the behavior of the type legalizer. The codegen
produces the same code.
This infrastructural change is needed in order to enable complex decisions
for vector types (needed by the vector-select patch).
llvm-svn: 132263
The practical effects here are that x86-64 fast-isel can now handle trunc from i8 to i1, and ARM fast-isel can handle many more constructs involving integers narrower than 32 bits (including loads, stores, and many integer casts).
rdar://9437928 .
llvm-svn: 132099
non-zero.
- Teach X86 cmov optimization to eliminate the cmov from ctlz, cttz extension
when the source of X86ISD::BSR / X86ISD::BSF is proven to be non-zero.
rdar://9490949
llvm-svn: 131948
by non-CMP expressions. The executable test case (129821) would test
this as well, if we had an "-O0 -disable-arm-fast-isel" LLVM-GCC
tester. Alas, the ARM assembly would be very difficult to check with
FileCheck.
The thumb2-cbnz.ll test is affected; it generates larger code (tst.w
vs. cmp #0), but I believe the new version is correct.
rdar://problem/9298790
llvm-svn: 131261
If there is a store after the load node, then there is a chain, which means
that there is another user. Thus, asking hasOneUser would fail. Instead we
ask hasNUsesOfValue on the 'data' value.
llvm-svn: 131183
intrinsic call. This prevents it from being reordered so that it appears
*before* the setjmp intrinsic (thus making it completely useless).
<rdar://problem/9409683>
llvm-svn: 131174
more callee-saved registers and introduce copies. Only allows it if scheduling
a node above calls would end up lessen register pressure.
Call operands also has added ABI restrictions for register allocation, so be
extra careful with hoisting them above calls.
rdar://9329627
llvm-svn: 130245
fix bugs exposed by the gcc dejagnu testsuite:
1. The load may actually be used by a dead instruction, which
would cause an assert.
2. The load may not be used by the current chain of instructions,
and we could move it past a side-effecting instruction. Change
how we process uses to define the problem away.
llvm-svn: 130018
On x86 this allows to fold a load into the cmp, greatly reducing register pressure.
movzbl (%rdi), %eax
cmpl $47, %eax
->
cmpb $47, (%rdi)
This shaves 8k off gcc.o on i386. I'll leave applying the patch in README.txt to Chris :)
llvm-svn: 130005
manually and pass all (now) 4 arguments to the mul libcall. Add a new
ExpandLibCall for just this (copied gratuitously from type legalization).
Fixes rdar://9292577
llvm-svn: 129842
2. implement rdar://9289501 - fast isel should fold trivial multiplies to shifts
3. teach tblgen to handle shift immediates that are different sizes than the
shifted operands, eliminating some code from the X86 fast isel backend.
4. Have FastISel::SelectBinaryOp use (the poorly named) FastEmit_ri_ function
instead of FastEmit_ri to simplify code.
llvm-svn: 129666
less trivial things) into a dummy lea. Before we generated:
_test: ## @test
movq _G@GOTPCREL(%rip), %rax
leaq (%rax), %rax
ret
now we produce:
_test: ## @test
movq _G@GOTPCREL(%rip), %rax
ret
This is part of rdar://9289558
llvm-svn: 129662
The basic issue here is that bottom-up isel is matching the branch
and compare, and was failing to fold the load into the branch/compare
combo. Fixing this (by allowing folding into any instruction of a
sequence that is selected) allows us to produce things like:
cmpb $0, 52(%rax)
je LBB4_2
instead of:
movb 52(%rax), %cl
cmpb $0, %cl
je LBB4_2
This makes the generated -O0 code run a bit faster, but also speeds up
compile time by putting less pressure on the register allocator and
generating less code.
This was one of the biggest classes of missing load folding. Implementing
this shrinks 176.gcc's c-decl.s (as a random example) by about 4% in (verbose-asm)
line count.
llvm-svn: 129656
This is done by pushing physical register definitions close to their
use, which happens to handle flag definitions if they're not glued to
the branch. This seems to be generally a good thing though, so I
didn't need to add a target hook yet.
The primary motivation is to generate code closer to what people
expect and rule out missed opportunity from enabling macro-op
fusion. As a side benefit, we get several 2-5% gains on x86
benchmarks. There is one regression:
SingleSource/Benchmarks/Shootout/lists slows down be -10%. But this is
an independent scheduler bug that will be tracked separately.
See rdar://problem/9283108.
Incidentally, pre-RA scheduling is only half the solution. Fixing the
later passes is tracked by:
<rdar://problem/8932804> [pre-RA-sched] on x86, attempt to schedule CMP/TEST adjacent with condition jump
Fixes:
<rdar://problem/9262453> Scheduler unnecessary break of cmp/jump fusion
llvm-svn: 129508
Additional fixes:
Do something reasonable for subtargets with generic
itineraries by handle node latency the same as for an empty
itinerary. Now nodes default to unit latency unless an itinerary
explicitly specifies a zero cycle stage or it is a TokenFactor chain.
Original fixes:
UnitsSharePred was a source of randomness in the scheduler: node
priority depended on the queue data structure. I rewrote the recent
VRegCycle heuristics to completely replace the old heuristic without
any randomness. To make the ndoe latency adjustments work, I also
needed to do something a little more reasonable with TokenFactor. I
gave it zero latency to its consumers and always schedule it as low as
possible.
llvm-svn: 129421
UnitsSharePred was a source of randomness in the scheduler: node
priority depended on the queue data structure. I rewrote the recent
VRegCycle heuristics to completely replace the old heuristic without
any randomness. To make these heuristic adjustments to node latency work,
I also needed to do something a little more reasonable with TokenFactor. I
gave it zero latency to its consumers and always schedule it as low as
possible.
llvm-svn: 129383
induction variable. The preRA scheduler is unaware of induction vars,
so we look for potential "virtual register cycles" instead.
Fixes <rdar://problem/8946719> Bad scheduling prevents coalescing
llvm-svn: 129100
It needed to be moved closer to the setjmp statement, because the code directly
after the setjmp needs to know about values that are on the stack. Also, the
'bitcast' of the function context was causing a dead load. This wouldn't be too
horrible, except that at -O0 it wasn't optimized out, and because it wasn't
using the correct base pointer (if there is a VLA), it would try to access a
value from a garbage address.
<rdar://problem/9130540>
llvm-svn: 128873
transformations in target-specific DAG combines without causing DAGCombiner to
delete the same node twice. If you know of a better way to avoid this (see my
next patch for an example), please let me know.
llvm-svn: 128758
It couldn't be used outside of the file because SDISelAsmOperandInfo
is local to SelectionDAGBuilder.cpp. Making it a static function avoids
a weird linkage dance.
llvm-svn: 128342
so the scheduler can't create new interferences on the copies
themselves. Prior to this fix the scheduler could get stuck in a loop
creating copies.
Fixes PR9509.
llvm-svn: 128164
I'm tired of doing this manually for each checkout.
If anyone knows a better way debug isel for non-trivial tests feel
free to revert and let me know how to do it.
llvm-svn: 128132
not have native support for this operation (such as X86).
The legalized code uses two vector INT_TO_FP operations and is faster
than scalarizing.
llvm-svn: 127951
rather than an int. Thankfully, this only causes LLVM to miss optimizations, not
generate incorrect code.
This just fixes the zext at the return. We still insert an i32 ZextAssert when
reading a function's arguments, but it is followed by a truncate and another i8
ZextAssert so it is not optimized.
llvm-svn: 127766
flexible.
If it returns a register class that's different from the input, then that's the
register class used for cross-register class copies.
If it returns a register class that's the same as the input, then no cross-
register class copies are needed (normal copies would do).
If it returns null, then it's not at all possible to copy registers of the
specified register class.
llvm-svn: 127368
with this before since none of the register tracking or nightly tests
had unschedulable nodes.
This should probably be refixed with a special default Node that just
returns some "don't touch me" values.
Fixes PR9427
llvm-svn: 127263
regs. This is the only change in this checkin that may affects the
default scheduler. With better register tracking and heuristics, it
doesn't make sense to artificially lower the register limit so much.
Added -sched-high-latency-cycles and X86InstrInfo::isHighLatencyDef to
give the scheduler a way to account for div and sqrt on targets that
don't have an itinerary. It is currently defaults to 10 (the actual
number doesn't matter much), but only takes effect on non-default
schedulers: list-hybrid and list-ilp.
Added several heuristics that can be individually disabled for the
non-default sched=list-ilp mode. This helps us determine how much
better we can do on a given benchmark than the default
scheduler. Certain compute intensive loops run much faster in this
mode with the right set of heuristics, and it doesn't seem to have
much negative impact elsewhere. Not all of the heuristics are needed,
but we still need to experiment to decide which should be disabled by
default for sched=list-ilp.
llvm-svn: 127067
type after type legalization has completed. Before then it may simply not be big
enough to hold the shift amount, particularly on x86 which uses a very small type
for shifts (this issue broke stuff in the past which is why LegalizeTypes carefully
uses a large type for shift amounts).
llvm-svn: 127000
Fix the PendingQueue, then disable it because it's not required for
the current schedulers' heuristics.
Fix the logic for the unused list-ilp scheduler.
llvm-svn: 126981
it. It's been assumed up til now that it would be in its immediate
successor. However, this isn't necessarily the case. It could be in one of its
successor's successors.
Modify the code to more thoroughly check for an 'eh.selector' call in
successors. It only looks at a successor if we get there as a result of an
unconditional branch.
Testcase ObjC/exceptions-4.m in r126968.
llvm-svn: 126969
David Greene changed CannotYetSelect() to print the full DAG including multiple
copies of operands reached through different paths in the DAG. Unfortunately
this blows up exponentially in some cases. The depth limit of 100 is way too
high to prevent this -- I'm seeing a message string of 150MB with a depth of
only 40 in one particularly bad case, even though the DAG has less than 200
nodes. Part of the problem is that the printing code is following chain
operands, so if you fail to select an operation with a chain, the printer will
follow all the chained operations back to the entry node.
llvm-svn: 126899
1. Inform users of ADDEs with two 0 operands that it never sets carry
2. Fold other ADDs or ADDCs into the ADDE if possible
It would be neat if we could do the same thing for SETCC+ADD eventually, but we can't do that in target independent code.
llvm-svn: 126557
is possible to do better if the high bit is set in either KnownZero/KnownOne, but
in practice NumSignBits is always 1 when we are zero extending because nothing
is known about that register.
llvm-svn: 126465
Limit the folding of any_ext and sext into the load operation to scalars.
Limit the active-bits trunc optimization to scalars.
Document vector trunc and vector sext in LangRef.
Similar to commit 126080 (for enabling zext).
llvm-svn: 126424
registers at phis. This enables us to eliminate a lot of pointless zexts during
the DAGCombine phase. This fixes <rdar://problem/8760114>.
llvm-svn: 126380
In other words, do not keep track of argument's location. The debugger (gdb) is not prepared to see line table entries for arguments. For the debugger, "second" line table entry marks beginning of function body.
This requires some coordination with debugger to get this working.
- The debugger needs to be aware of prolog_end attribute attached with line table entries.
- The compiler needs to accurately mark prolog_end in line table entries (at -O0 and at -O1+)
llvm-svn: 126155
The DAGCombiner folds the zext into complex load instructions. This patch
prevents this optimization on vectors since none of the supported targets
knows how to perform load+vector_zext in one instruction.
llvm-svn: 126080
the time but presumably my email got lost). Examples where the previous logic
got it wrong: (1) a signed i8 multiply of 64 by 2 overflows, but the high part is
zero; (2) a signed i8 multiple of -128 by 2 overflows, but the high part is all
ones.
llvm-svn: 125748
transformation if we can't legally create a build vector of the correct
type. Check that we can make the transformation first, and add a TODO to
refactor this code with similar cases.
Fixes: PR9223 and rdar://9000350
llvm-svn: 125631
have their low bits set to zero. This allows us to optimize
out explicit stack alignment code like in stack-align.ll:test4 when
it is redundant.
Doing this causes the code generator to start turning FI+cst into
FI|cst all over the place, which is general goodness (that is the
canonical form) except that various pieces of the code generator
don't handle OR aggressively. Fix this by introducing a new
SelectionDAG::isBaseWithConstantOffset predicate, and using it
in places that are looking for ADD(X,CST). The ARM backend in
particular was missing a lot of addressing mode folding opportunities
around OR.
llvm-svn: 125470
generating i8 shift amounts for things like i1024 types. Add
an assert in getNode to prevent this from occuring in the future,
fix the buggy transformation, revert my previous patch, and
document this gotcha in ISDOpcodes.h
llvm-svn: 125465
The DAGCombiner created illegal BUILD_VECTOR operations.
The patch added a check that either illegal operations are
allowed or that the created operation is legal.
llvm-svn: 125435
The bug happens when the DAGCombiner attempts to optimize one of the patterns
of the SUB opcode. It tries to create a zero of type v2i64. This type is legal
on 32bit machines, but the initializer of this vector (i64) is target dependent.
Currently, the initializer attempts to create an i64 zero constant, which fails.
Added a flag to tell the DAGCombiner to create a legal zero, if we require that
the pass would generate legal types.
llvm-svn: 125391
Nadav Rotem