If the instruction window is < NumRegs/2, pressure tracking is not
likely to be effective. The scheduler has to process a very large
number of tiny blocks. We want this to be fast.
llvm-svn: 189991
Register pressure tracking is half the complexity of the
scheduler. It's useful to be able to turn it off for compile time and
performance comparisons.
llvm-svn: 189987
This reverts commit r189913.
Talked with Eric on IRC. I am going to XFAIL the failing test since it
is using what Eric described as "the member hack" which was needed on
that old GDB.
Sorry for the noise!
llvm-svn: 189914
This won't affect the kinds of hashes we test for as we actually
do hashing based on form and attribute. Change the fission-hash
testcase one last time to handle DW_AT_comp_dir.
llvm-svn: 189840
There was one case that we could hit a DebugValue where I didn't think
to check. DebugValues are evil. No checkinable test case, sorry. It's
an obvious fix.
llvm-svn: 189717
Created SUPressureDiffs array to hold the per node PDiff computed during DAG building.
Added a getUpwardPressureDelta API that will soon replace the old
one. Compute PressureDelta here from the precomputed PressureDiffs.
Updating for liveness will come next.
llvm-svn: 189640
Revert unintentional commit (of an unreviewed change).
Original commit message:
Add getUnrollingPreferences to TTI
Allow targets to customize the default behavior of the generic loop unrolling
transformation. This will be used by the PowerPC backend when targeting the A2
core (which is in-order with a deep pipeline), and using more aggressive
defaults is important.
llvm-svn: 189566
Allow targets to customize the default behavior of the generic loop unrolling
transformation. This will be used by the PowerPC backend when targeting the A2
core (which is in-order with a deep pipeline), and using more aggressive
defaults is important.
llvm-svn: 189565
This uses the TargetSubtargetInfo::useAA() function to control the defaults of
the -combiner-alias-analysis and -combiner-global-alias-analysis options.
llvm-svn: 189564
There are several optional (off-by-default) features in CodeGen that can make
use of alias analysis. These features are important for generating code for
some kinds of cores (for example the (in-order) PPC A2 core). This adds a
useAA() function to TargetSubtargetInfo to allow these features to be enabled
by default on a per-subtarget basis.
Here is the first use of this function: To control the default of the
-enable-aa-sched-mi feature.
llvm-svn: 189563
when we can. Migrate from using blocks when we're adding just a
single attribute and floating point values are an unsigned, not signed,
bag of bits.
Update all test cases accordingly.
llvm-svn: 189419
We want to convert code like (or (srl N, 8), (shl N, 8)) into (srl (bswap N),
const), but this is only valid if the bits above 16 on the source pattern are
0, the checks we were doing on this were slightly wrong before.
llvm-svn: 189348
is constructing from as an input and keep the same unique identifier.
We can use this to connect items which must stay in the .o file
(e.g. pubnames and pubtypes) to the skeleton cu rather than having
duplicate unique numbers for the sections and needing to do lookups
based on MDNode.
llvm-svn: 189293
If we have a binary operation like ISD:ADD, we can set the result type
equal to the result type of one of its operands rather than using
TargetLowering::getPointerTy().
Also, any use of DAG.getIntPtrConstant(C) as an operand for a binary
operation can be replaced with:
DAG.getConstant(C, OtherOperand.getValueType());
llvm-svn: 189227
This adds minimal support to the SelectionDAG for handling address spaces
with different pointer sizes. The SelectionDAG should now correctly
lower pointer function arguments to the correct size as well as generate
the correct code when lowering getelementptr.
This patch also updates the R600 DataLayout to use 32-bit pointers for
the local address space.
v2:
- Add more helper functions to TargetLoweringBase
- Use CHECK-LABEL for tests
llvm-svn: 189221
We currently emit labels with the prefix Lllvm$workaround$fake$stub$ if
the target's MCAsmInfo has getLinkOnceDirective() mapped to something
interesting. This was apparently a work around introduced in r31033 for
binutils that we don't need anymore.
llvm-svn: 189187
Estimate the cyclic critical path within a single block loop. If the
acyclic critical path is longer, then the loop will exhaust OOO
resources after some number of iterations. If lag between the acyclic
critical path and cyclic critical path is longer the the time it takes
to issue those loop iterations, then aggressively schedule for
latency.
llvm-svn: 189120
This will be used to compute the cyclic critical path and to
update precomputed per-node pressure differences.
In the longer term, it could also be used to speed up LiveInterval
update by avoiding visiting all global vreg users.
llvm-svn: 189118
...so that it can be used for z too. Most of the code is the same.
The only real change is to use TargetTransformInfo to test when a sqrt
instruction is available.
The pass is opt-in because at the moment it only handles sqrt.
llvm-svn: 189097
When truncated vector stores were being custom lowered in
VectorLegalizer::LegalizeOp(), the old (illegal) and new (legal) node pair
was not being added to LegalizedNodes list. Instead of the legalized
result being passed to VectorLegalizer::TranslateLegalizeResult(),
the result was being passed back into VectorLegalizer::LegalizeOp(),
which ended up adding a (new, new) pair to the list instead.
This was causing an assertion failure when a custom lowered truncated
vector store was the last instruction a basic block and the VectorLegalizer
was unable to find it in the LegalizedNodes list when updating the
DAG root.
llvm-svn: 188953
The small utility function that pattern matches Base + Index +
Offset patterns for loads and stores fails to recognize the base
pointer for loads/stores from/into an array at offset 0 inside a
loop. As a result DAGCombiner::MergeConsecutiveStores was not able
to merge all stores.
This commit fixes the issue by adding an additional pattern match
and also a test case.
Reviewer: Nadav
llvm-svn: 188936
Summary:
LLVM would generate DWARF with version 3 in the .debug_pubname and
.debug_pubtypes version fields. This would lead SGI dwarfdump to fail
parsing the DWARF with (in the instance of .debug_pubnames) would exit
with:
dwarfdump ERROR: dwarf_get_globals: DW_DLE_PUBNAMES_VERSION_ERROR (123)
This fixes PR16950.
Reviewers: echristo, dblaikie
Reviewed By: echristo
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D1454
llvm-svn: 188869
SystemZTargetLowering::emitStringWrapper() previously loaded the character
into R0 before the loop and made R0 live on entry. I'd forgotten that
allocatable registers weren't allowed to be live across blocks at this stage,
and it confused LiveVariables enough to cause a miscompilation of f3 in
memchr-02.ll.
This patch instead loads R0 in the loop and leaves LICM to hoist it
after RA. This is actually what I'd tried originally, but I went for
the manual optimisation after noticing that R0 often wasn't being hoisted.
This bug forced me to go back and look at why, now fixed as r188774.
We should also try to optimize null checks so that they test the CC result
of the SRST directly. The select between null and the SRST GPR result could
then usually be deleted as dead.
llvm-svn: 188779
Post-RA LICM keeps three sets of registers: PhysRegDefs, PhysRegClobbers
and TermRegs. When it sees a definition of R it adds all aliases of R
to the corresponding set, so that when it needs to test for membership
it only needs to test a single register, rather than worrying about
aliases there too. E.g. the final candidate loop just has:
unsigned Def = Candidates[i].Def;
if (!PhysRegClobbers.test(Def) && ...) {
to test whether register Def is multiply defined.
However, there was also a shortcut in ProcessMI to make sure we didn't
add candidates if we already knew that they would fail the final test.
This shortcut was more pessimistic than the final one because it
checked whether _any alias_ of the defined register was multiply defined.
This is too conservative for targets that define register pairs.
E.g. on z, R0 and R1 are sometimes used as a pair, so there is a
128-bit register that aliases both R0 and R1. If a loop used
R0 and R1 independently, and the definition of R0 came first,
we would be able to hoist the R0 assignment (because that used
the final test quoted above) but not the R1 assignment (because
that meant we had two definitions of the paired R0/R1 register
and would fail the shortcut in ProcessMI).
This patch just uses the same check for the ProcessMI shortcut as
we use in the final candidate loop.
llvm-svn: 188774
Previously, generation of stack protectors was done exclusively in the
pre-SelectionDAG Codegen LLVM IR Pass "Stack Protector". This necessitated
splitting basic blocks at the IR level to create the success/failure basic
blocks in the tail of the basic block in question. As a result of this,
calls that would have qualified for the sibling call optimization were no
longer eligible for optimization since said calls were no longer right in
the "tail position" (i.e. the immediate predecessor of a ReturnInst
instruction).
Then it was noticed that since the sibling call optimization causes the
callee to reuse the caller's stack, if we could delay the generation of
the stack protector check until later in CodeGen after the sibling call
decision was made, we get both the tail call optimization and the stack
protector check!
A few goals in solving this problem were:
1. Preserve the architecture independence of stack protector generation.
2. Preserve the normal IR level stack protector check for platforms like
OpenBSD for which we support platform specific stack protector
generation.
The main problem that guided the present solution is that one can not
solve this problem in an architecture independent manner at the IR level
only. This is because:
1. The decision on whether or not to perform a sibling call on certain
platforms (for instance i386) requires lower level information
related to available registers that can not be known at the IR level.
2. Even if the previous point were not true, the decision on whether to
perform a tail call is done in LowerCallTo in SelectionDAG which
occurs after the Stack Protector Pass. As a result, one would need to
put the relevant callinst into the stack protector check success
basic block (where the return inst is placed) and then move it back
later at SelectionDAG/MI time before the stack protector check if the
tail call optimization failed. The MI level option was nixed
immediately since it would require platform specific pattern
matching. The SelectionDAG level option was nixed because
SelectionDAG only processes one IR level basic block at a time
implying one could not create a DAG Combine to move the callinst.
To get around this problem a few things were realized:
1. While one can not handle multiple IR level basic blocks at the
SelectionDAG Level, one can generate multiple machine basic blocks
for one IR level basic block. This is how we handle bit tests and
switches.
2. At the MI level, tail calls are represented via a special return
MIInst called "tcreturn". Thus if we know the basic block in which we
wish to insert the stack protector check, we get the correct behavior
by always inserting the stack protector check right before the return
statement. This is a "magical transformation" since no matter where
the stack protector check intrinsic is, we always insert the stack
protector check code at the end of the BB.
Given the aforementioned constraints, the following solution was devised:
1. On platforms that do not support SelectionDAG stack protector check
generation, allow for the normal IR level stack protector check
generation to continue.
2. On platforms that do support SelectionDAG stack protector check
generation:
a. Use the IR level stack protector pass to decide if a stack
protector is required/which BB we insert the stack protector check
in by reusing the logic already therein. If we wish to generate a
stack protector check in a basic block, we place a special IR
intrinsic called llvm.stackprotectorcheck right before the BB's
returninst or if there is a callinst that could potentially be
sibling call optimized, before the call inst.
b. Then when a BB with said intrinsic is processed, we codegen the BB
normally via SelectBasicBlock. In said process, when we visit the
stack protector check, we do not actually emit anything into the
BB. Instead, we just initialize the stack protector descriptor
class (which involves stashing information/creating the success
mbbb and the failure mbb if we have not created one for this
function yet) and export the guard variable that we are going to
compare.
c. After we finish selecting the basic block, in FinishBasicBlock if
the StackProtectorDescriptor attached to the SelectionDAGBuilder is
initialized, we first find a splice point in the parent basic block
before the terminator and then splice the terminator of said basic
block into the success basic block. Then we code-gen a new tail for
the parent basic block consisting of the two loads, the comparison,
and finally two branches to the success/failure basic blocks. We
conclude by code-gening the failure basic block if we have not
code-gened it already (all stack protector checks we generate in
the same function, use the same failure basic block).
llvm-svn: 188755
This adds a llvm.copysign intrinsic; We already have Libfunc recognition for
copysign (which is turned into the FCOPYSIGN SDAG node). In order to
autovectorize calls to copysign in the loop vectorizer, we need a corresponding
intrinsic as well.
In addition to the expected changes to the language reference, the loop
vectorizer, BasicTTI, and the SDAG builder (the intrinsic is transformed into
an FCOPYSIGN node, just like the function call), this also adds FCOPYSIGN to a
few lists in LegalizeVector{Ops,Types} so that vector copysigns can be
expanded.
In TargetLoweringBase::initActions, I've made the default action for FCOPYSIGN
be Expand for vector types. This seems correct for all in-tree targets, and I
think is the right thing to do because, previously, there was no way to generate
vector-values FCOPYSIGN nodes (and most targets don't specify an action for
vector-typed FCOPYSIGN).
llvm-svn: 188728
Until gdb supports the new accelerator tables we should add the
pubnames section so that gdb_index can be generated from gold
at link time. On darwin we already emit the accelerator tables
and so don't need to worry about pubnames.
llvm-svn: 188708
- split WidenVecRes_Binary into WidenVecRes_Binary and WidenVecRes_BinaryCanTrap
- WidenVecRes_BinaryCanTrap preserves the original behaviour for operations
that can trap
- WidenVecRes_Binary simply widens the operation and improves codegen for
3-element vectors by allowing widening and promotion on x86 (matches the
behaviour of unary and ternary operation widening)
- use WidenVecRes_Binary for operations on integers.
Reviewed by: nrotem
llvm-svn: 188699
We had previously been asserting when faced with a FCOPYSIGN f64, ppcf128 node
because there was no way to expand the FCOPYSIGN node. Because ppcf128 is the
sum of two doubles, and the first double must have the larger magnitude, we
can take the sign from the first double. As a result, in addition to fixing the
crash, this is also an optimization.
llvm-svn: 188655
We check this in many/all other cases, just missed this one it seems.
Perhaps it'd be worth unifying this so we never emit zero-length
DW_AT_names.
llvm-svn: 188649
Teach the generic instruction selection helper functions to constrain
the register classes of their input operands. For non-physical register
references, the generic code needs to be careful not to mess that up
when replacing references to result registers. As the comment indicates
for MachineRegisterInfo::replaceRegWith(), it's important to call
constrainRegClass() first.
rdar://12594152
llvm-svn: 188593
Generalize r188163 to cope with return types other than MVT::i32, just
as the existing visitMemCmpCall code did. I've split this out into a
subroutine so that it can be used for other upcoming patches.
I also noticed that I'd used the wrong API to record the out chain.
It's a load that uses DAG.getRoot() rather than getRoot(), so the out
chain should go on PendingLoads. I don't have a testcase for that because
we don't do any interesting scheduling on z yet.
llvm-svn: 188540
When new virtual registers are created during splitting/spilling, defer
creation of the live interval until we need to use the live interval.
Along with the recent commits to notify LiveRangeEdit when new virtual
registers are created, this makes it possible for functions like
TargetInstrInfo::loadRegFromStackSlot() and
TargetInstrInfo::storeRegToStackSlot() to create multiple virtual
registers as part of the process of generating loads/stores for
different register classes, and then have the live intervals for those
new registers computed when they are needed.
llvm-svn: 188437
Add a delegate class to MachineRegisterInfo with a single virtual
function, MRI_NoteNewVirtualRegister(). Update LiveRangeEdit to inherit
from this delegate class and override the definition of the callback
with an implementation that tracks the newly created virtual registers.
llvm-svn: 188435
Track new virtual registers by register number, rather than by the live
interval created for them. This is the first step in separating the
creation of new virtual registers and new live intervals. Eventually
live intervals will be created and populated on demand after the virtual
registers have been created and used in instructions.
llvm-svn: 188434
A common idiom is to use zero and all-ones as sentinal values and to
check for both in a single conditional ("x != 0 && x != (unsigned)-1").
That generates code, for i32, like:
testl %edi, %edi
setne %al
cmpl $-1, %edi
setne %cl
andb %al, %cl
With this transform, we generate the simpler:
incl %edi
cmpl $1, %edi
seta %al
Similar improvements for other integer sizes and on other platforms. In
general, combining the two setcc instructions into one is better.
rdar://14689217
llvm-svn: 188315
LowerCallTo returns a pair with the return value of the call as the first
element and the chain associated with the return value as the second element. If
we lower a call that has a void return value, LowerCallTo returns an SDValue
with a NULL SDNode and the chain for the call. Thus makeLibCall by just
returning the first value makes it impossible for you to set up the chain so
that the call is not eliminated as dead code.
I also updated all references to makeLibCall to reflect the new return type.
llvm-svn: 188300
Summary:
We need to do two things:
- Initialize BSSSection in MCObjectFileInfo::InitCOFFMCObjectFileInfo
- Teach TargetLoweringObjectFileCOFF::SelectSectionForGlobal what to do
with it
This fixes PR16861.
Reviewers: rnk
Reviewed By: rnk
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D1361
llvm-svn: 188244
CUs.
Currently only hashes the name of CUs and the names of any children,
but it's an obvious first step to show the framework. The testcase
should continue to be correct, however, as it's an empty TU.
llvm-svn: 188243
If the tail-callee and caller give the same bits via the same signext/zeroext
attribute then a tail-call should be allowed, since the extension has already
been done by the callee.
llvm-svn: 188159
This patch decouples the stack protector pass so that we can support stack
protector implementations that do not use the IR level generated stack protector
fail basic block.
No codesize increase is caused by this change since the MI level tail merge pass
properly merges together the fail condition blocks (see the updated test).
llvm-svn: 188105
Previously the asserts were only checking that RHS and LHS were the same type and had the same element type as the result. All downstream code for ISD::VECTOR_SHUFFLE requires the types to be the same.
Also removed one unnecessary check of matched element counts that was present in the code.
llvm-svn: 188051
For most libm ISD nodes, TargetLoweringBase::initActions sets the default
scalar-type action to Expand, and leaves the vector-type action default as
Legal. This is not appropriate for the new ISD::FROUND node (which no backend
but PowerPC handles explicitly).
Fixes PR16842.
llvm-svn: 188048
the type exists.
Fix up cases where we weren't checking for optional types and add
an assert to addType to make sure we catch this in the future.
Fix up a testcase that was using the tag for DW_TAG_array_type
when it meant DW_TAG_enumeration_type.
llvm-svn: 187963
This reverts commit r77814.
We were sticking global constants in the .data section instead of in the
.rdata section when emitting for COFF.
This fixes PR16831.
llvm-svn: 187956
Original commit message:
Stop emitting weak symbols into the "coal" sections.
The Mach-O linker has been able to support the weak-def bit on any symbol for
quite a while now. The compiler however continued to place these symbols into a
"coal" section, which required the linker to map them back to the base section
name.
Replace the sections like this:
__TEXT/__textcoal_nt instead use __TEXT/__text
__TEXT/__const_coal instead use __TEXT/__const
__DATA/__datacoal_nt instead use __DATA/__data
<rdar://problem/14265330>
llvm-svn: 187939
All libm floating-point rounding functions, except for round(), had their own
ISD nodes. Recent PowerPC cores have an instruction for round(), and so here I'm
adding ISD::FROUND so that round() can be custom lowered as well.
For the most part, this is straightforward. I've added an intrinsic
and a matching ISD node just like those for nearbyint() and friends. The
SelectionDAG pattern I've named frnd (because ISD::FP_ROUND has already claimed
fround).
This will be used by the PowerPC backend in a follow-up commit.
llvm-svn: 187926
This change came about primarily because of two issues in the existing code.
Niether of:
define i64 @test1(i64 %val) {
%in = trunc i64 %val to i32
tail call i32 @ret32(i32 returned %in)
ret i64 %val
}
define i64 @test2(i64 %val) {
tail call i32 @ret32(i32 returned undef)
ret i32 42
}
should be tail calls, and the function sameNoopInput is responsible. The main
problem is that it is completely symmetric in the "tail call" and "ret" value,
but in reality different things are allowed on each side.
For these cases:
1. Any truncation should lead to a larger value being generated by "tail call"
than needed by "ret".
2. Undef should only be allowed as a source for ret, not as a result of the
call.
Along the way I noticed that a mismatch between what this function treats as a
valid truncation and what the backends see can lead to invalid calls as well
(see x86-32 test case).
This patch refactors the code so that instead of being based primarily on
values which it recurses into when necessary, it starts by inspecting the type
and considers each fundamental slot that the backend will see in turn. For
example, given a pathological function that returned {{}, {{}, i32, {}}, i32}
we would consider each "real" i32 in turn, and ask if it passes through
unchanged. This is much closer to what the backend sees as a result of
ComputeValueVTs.
Aside from the bug fixes, this eliminates the recursion that's going on and, I
believe, makes the bulk of the code significantly easier to understand. The
trade-off is the nasty iterators needed to find the real types inside a
returned value.
llvm-svn: 187787
This virtual function can be implemented by targets to specify the type
to use for the index operand of INSERT_VECTOR_ELT, EXTRACT_VECTOR_ELT,
INSERT_SUBVECTOR, EXTRACT_SUBVECTOR. The default implementation returns
the result from TargetLowering::getPointerTy()
The previous code was using TargetLowering::getPointerTy() for vector
indices, because this is guaranteed to be legal on all targets. However,
using TargetLowering::getPointerTy() can be a problem for targets with
pointer sizes that differ across address spaces. On such targets,
when vectors need to be loaded or stored to an address space other than the
default 'zero' address space (which is the address space assumed by
TargetLowering::getPointerTy()), having an index that
is a different size than the pointer can lead to inefficient
pointer calculations, (e.g. 64-bit adds for a 32-bit address space).
There is no intended functionality change with this patch.
llvm-svn: 187748
Function attributes are the future! So just query whether we want to realign the
stack directly from the function instead of through a random target options
structure.
llvm-svn: 187618
For a testcase like the following:
typedef unsigned long uint64_t;
typedef struct {
uint64_t lo;
uint64_t hi;
} blob128_t;
void add_128_to_128(const blob128_t *in, blob128_t *res) {
asm ("PAND %1, %0" : "+Q"(*res) : "Q"(*in));
}
where we'll fail to allocate the register for the output constraint,
our matching input constraint will not find a register to match,
and could try to search past the end of the current operands array.
On the idea that we'd like to attempt to keep compilation going
to find more errors in the module, change the error cases when
we're visiting inline asm IR to return immediately and avoid
trying to create a node in the DAG. This leaves us with only
a single error message per inline asm instruction, but allows us
to safely keep going in the general case.
llvm-svn: 187470
When registers must be live throughout the scheduling region, increase
the limit for the register class. Once we exceed the original limit,
they will be spilled, and there's no point further reducing pressure.
This isn't a perfect heuristics but avoids a situation where the
scheduler could become trapped by trying to achieve the impossible.
llvm-svn: 187436
This patch prevents the following combine when the input vector is used more
than once.
insert_vector_elt (build_vector elt0, ..., eltN), NewEltIdx, idx
=>
build_vector elt0, ..., NewEltIdx, ..., eltN
The reasons are:
- Building a vector may be expensive, so try to reuse the existing part of a
vector instead of creating a new one (think big vectors).
- elt0 to eltN now have two users instead of one. This may prevent some other
optimizations.
llvm-svn: 187396
update testcase to make sure we generate debug info for walrus
by adding a non-trivial constructor and verify that we don't
emit an ODR signature for the type.
llvm-svn: 187393
32-bit symbols have "_" as global prefix, but when forming the name of
COMDAT sections this prefix is ignored. The current behavior assumes that
this prefix is always present which is not the case for 64-bit and names
are truncated.
llvm-svn: 187356
There doesn't appear to be any reason to put this variable on the heap.
I'm suspicious of the LexicalScope above that we stuff in a map and then
delete afterward, but I'm just trying to get the valgrind bot clean.
llvm-svn: 187301
Adds unit tests for it too.
Split BasicBlockUtils into an analysis-half and a transforms-half, and put the
analysis bits into a new Analysis/CFG.{h,cpp}. Promote isPotentiallyReachable
into llvm::isPotentiallyReachable and move it into Analysis/CFG.
llvm-svn: 187283
Merge consecutive if-regions if they contain identical statements.
Both transformations reduce number of branches. The transformation
is guarded by a target-hook, and is currently enabled only for +R600,
but the correctness has been tested on X86 target using a variety of
CPU benchmarks.
Patch by: Mei Ye
llvm-svn: 187278
type units.
Initially this support is used in the computation of an ODR checker
for C++. For now we're attaching it to the DIE, but in the future
it will be attached to the type unit.
This also starts breaking out types into the separation for type
units, but without actually splitting the DIEs.
In preparation for hashing the DIEs this adds a DIEString type
that contains a StringRef with the string contained at the label.
llvm-svn: 187213
CustomLowerNode was not being called during SplitVectorOperand,
meaning custom legalization could not be used by targets.
This also adds a test case for NVPTX that depends on this custom
legalization.
Differential Revision: http://llvm-reviews.chandlerc.com/D1195
Attempt to fix the buildbots by making the X86 test I just added platform independent
llvm-svn: 187202
This reverts commit 187198. It broke the bots.
The soft float test probably needs a -triple because of name differences.
On the hard float test I am getting a "roundss $1, %xmm0, %xmm0", instead of
"vroundss $1, %xmm0, %xmm0, %xmm0".
llvm-svn: 187201
CustomLowerNode was not being called during SplitVectorOperand,
meaning custom legalization could not be used by targets.
This also adds a test case for NVPTX that depends on this custom
legalization.
Differential Revision: http://llvm-reviews.chandlerc.com/D1195
llvm-svn: 187198
The previous change to local live range allocation also suppressed
eviction of local ranges. In rare cases, this could result in more
expensive register choices. This commit actually revives a feature
that I added long ago: check if live ranges can be reassigned before
eviction. But now it only happens in rare cases of evicting a local
live range because another local live range wants a cheaper register.
The benefit is improved code size for some benchmarks on x86 and armv7.
I measured no significant compile time increase and performance
changes are noise.
llvm-svn: 187140
Also avoid locals evicting locals just because they want a cheaper register.
Problem: MI Sched knows exactly how many registers we have and assumes
they can be colored. In cases where we have large blocks, usually from
unrolled loops, greedy coloring fails. This is a source of
"regressions" from the MI Scheduler on x86. I noticed this issue on
x86 where we have long chains of two-address defs in the same live
range. It's easy to see this in matrix multiplication benchmarks like
IRSmk and even the unit test misched-matmul.ll.
A fundamental difference between the LLVM register allocator and
conventional graph coloring is that in our model a live range can't
discover its neighbors, it can only verify its neighbors. That's why
we initially went for greedy coloring and added eviction to deal with
the hard cases. However, for singly defined and two-address live
ranges, we can optimally color without visiting neighbors simply by
processing the live ranges in instruction order.
Other beneficial side effects:
It is much easier to understand and debug regalloc for large blocks
when the live ranges are allocated in order. Yes, global allocation is
still very confusing, but it's nice to be able to comprehend what
happened locally.
Heuristics could be added to bias register assignment based on
instruction locality (think late register pairing, banks...).
Intuituvely this will make some test cases that are on the threshold
of register pressure more stable.
llvm-svn: 187139
There's no need to specify a flag to omit frame pointer elimination on non-leaf
nodes...(Honestly, I can't parse that option out.) Use the function attribute
stuff instead.
llvm-svn: 187093
Prior to this patch, IfConverter may widen the cases where a sequence of
instructions were executed because of the way it uses nested predicates. This
result in incorrect execution.
For instance, Let A be a basic block that flows conditionally into B and B be a
predicated block.
B can be predicated with A.BrToBPredicate into A iff B.Predicate is less
"permissive" than A.BrToBPredicate, i.e., iff A.BrToBPredicate subsumes
B.Predicate.
The IfConverter was checking the opposite: B.Predicate subsumes
A.BrToBPredicate.
<rdar://problem/14379453>
llvm-svn: 187071
Use the function attributes to pass along the stack protector buffer size.
Now that we have robust function attributes, don't use a command line option to
specify the stack protecto buffer size.
llvm-svn: 186863
These floats all represented block frequencies anyway, so just use the
BlockFrequency class directly.
Some floating point computations remain in tryLocalSplit(). They are
estimating spill weights which are still floats.
llvm-svn: 186435
Original commit message:
Remove floating point computations from SpillPlacement.cpp.
Patch by Benjamin Kramer!
Use the BlockFrequency class instead of floats in the Hopfield network
computations. This rescales the node Bias field from a [-2;2] float
range to two block frequencies BiasN and BiasP pulling in opposite
directions. This construct has a more predictable behavior when block
frequencies saturate.
The per-node scaling factors are no longer necessary, assuming the block
frequencies around a bundle are consistent.
This patch can cause the register allocator to make different spilling
decisions. The differences should be small.
llvm-svn: 186434
We can have a FrameSetup in one basic block and the matching FrameDestroy
in a different basic block when we have struct byval. In that case, SPAdj
is not zero at beginning of the basic block.
Modify PEI to correctly set SPAdj at beginning of each basic block using
DFS traversal. We used to assume SPAdj is 0 at beginning of each basic block.
PEI had an assert SPAdjCount || SPAdj == 0.
If we have a Destroy <n> followed by a Setup <m>, PEI will assert failure.
We can add an extra condition to make sure the pairs are matched:
The pairs start with a FrameSetup.
But since we are doing a much better job in the verifier, this patch removes
the check in PEI.
PR16393
llvm-svn: 186364
1> on every path through the CFG, a FrameSetup <n> is always followed by a
FrameDestroy <n> and a FrameDestroy is always followed by a FrameSetup.
2> stack adjustments are identical on all CFG edges to a merge point.
3> frame is destroyed at end of a return block.
PR16393
llvm-svn: 186350
There is a comment at the top of DAGTypeLegalizer::PerformExpensiveChecks
which, in part, says:
// Note that these invariants may not hold momentarily when processing a node:
// the node being processed may be put in a map before being marked Processed.
Unfortunately, this assert would be valid only if the above-mentioned invariant
held unconditionally. This was causing llc to assert when, in fact,
everything was fine.
Thanks to Richard Sandiford for investigating this issue!
Fixes PR16562.
llvm-svn: 186338
Address calculation for gather/scather in vectorized code can incur a
significant cost making vectorization unbeneficial. Add infrastructure to add
cost.
Tests and cost model for targets will be in follow-up commits.
radar://14351991
llvm-svn: 186187
MF is normally initialized in AsmPrinter::SetupMachineFunction, but if the file
contains only globals (no functions), then we need this to be initialized
because, when encountering an error, lowerConstant() references it.
This should fix the non-deterministic failures of
test/CodeGen/X86/nonconst-static-iv.ll, etc.
llvm-svn: 186068
When computing currently-live registers, the register scavenger excludes undef
uses. As a result, undef uses are ignored when computing the restore points of
registers spilled into the emergency slots. While the register scavenger
normally excludes from consideration, when scavenging, registers used by the
current instruction, we need to not exclude undef uses. Otherwise, we might end
up requiring more emergency spill slots than we have (in the case where the
undef use *is* the currently-spilled register).
Another bug found by llvm-stress.
llvm-svn: 186067
Change the informal convention of DBG_VALUE machine instructions so that
we can express a register-indirect address with an offset of 0.
The old convention was that a DBG_VALUE is a register-indirect value if
the offset (operand 1) is nonzero. The new convention is that a DBG_VALUE
is register-indirect if the first operand is a register and the second
operand is an immediate. For plain register values the combination reg,
reg is used. MachineInstrBuilder::BuildMI knows how to build the new
DBG_VALUES.
rdar://problem/13658587
llvm-svn: 185966
Because integer BUILD_VECTOR operands may have a larger type than the result's
vector element type, and all operands must have the same type, when widening a
BUILD_VECTOR node by adding UNDEFs, we cannot use the vector element type, but
rather must use the type of the existing operands.
Another bug found by llvm-stress.
llvm-svn: 185960
in-tree implementations of TargetLoweringBase::isFMAFasterThanMulAndAdd in
order to resolve the following issues with fmuladd (i.e. optional FMA)
intrinsics:
1. On X86(-64) targets, ISD::FMA nodes are formed when lowering fmuladd
intrinsics even if the subtarget does not support FMA instructions, leading
to laughably bad code generation in some situations.
2. On AArch64 targets, ISD::FMA nodes are formed for operations on fp128,
resulting in a call to a software fp128 FMA implementation.
3. On PowerPC targets, FMAs are not generated from fmuladd intrinsics on types
like v2f32, v8f32, v4f64, etc., even though they promote, split, scalarize,
etc. to types that support hardware FMAs.
The function has also been slightly renamed for consistency and to force a
merge/build conflict for any out-of-tree target implementing it. To resolve,
see comments and fixed in-tree examples.
llvm-svn: 185956
When folding sub x, x (and other similar constructs), where x is a vector, the
result is a vector of zeros. After type legalization, make sure that the input
zero elements have a legal type. This type may be larger than the result's
vector element type.
This was another bug found by llvm-stress.
llvm-svn: 185949
This patch broke `make check-asan` on Mac, causing ld warnings like the following one:
ld: warning: direct access in __GLOBAL__I_a to global weak symbol
___asan_mapping_scale means the weak symbol cannot be overridden at
runtime. This was likely caused by different translation units being
compiled with different visibility settings.
The resulting test binaries crashed with incorrect ASan warnings.
llvm-svn: 185923
The Mach-O linker has been able to support the weak-def bit on any symbol for
quite a while now. The compiler however continued to place these symbols into a
"coal" section, which required the linker to map them back to the base section
name.
Replace the sections like this:
__TEXT/__textcoal_nt instead use __TEXT/__text
__TEXT/__const_coal instead use __TEXT/__const
__DATA/__datacoal_nt instead use __DATA/__data
<rdar://problem/14265330>
llvm-svn: 185872
Since the pool indexes are necessarily sequential and contiguous, just
insert things in the right place rather than having to sort the sequence
after the fact.
No functionality change.
llvm-svn: 185842
This fixes a bug (found by llvm-stress) in
DAGTypeLegalizer::PromoteIntRes_BUILD_VECTOR where it assumed that the result
type would always be larger than the original operands. This is not always
true, however, with boolean vectors. For example, promoting a node of type v8i1
(where the operands will be of type i32, the type to which i1 is promoted) will
yield a node with a result vector element type of i16 (and operands of type
i32). As a result, we cannot blindly assume that we can ANY_EXTEND the operands
to the result type.
llvm-svn: 185794
This fixes an oversight that Intrinsic::nearbyint was not being mapped to
ISD::FNEARBYINT (thus fixing the over-optimistic cost we were assigning to
nearbyint calls for some targets).
llvm-svn: 185783
Obviously the personality function should be emitted as language handler
instead of the hard coded _GCC_specific_handler. The language specific
data must be placed after the unwind information therefore it must not
be emitted into a separate section.
Reviewed by Charles Davis and Nico Rieck.
llvm-svn: 185761
ReduceLoadWidth unconditionally drops extensions from loads. Limit it to the
case when all of the bits the extension would otherwise produce are dropped by
the shrink. It would be possible to shrink the load in more cases by merging
the extensions, but this isn't trivial and a very rare case. I left a TODO for
that case.
Fixes PR16551.
llvm-svn: 185755
This prevents the emission of DAG-generated vreg definitions after a
tail call be dropping them entirely (on the grounds that nothing could
use them anyway, and they interfere with O0 CodeGen).
llvm-svn: 185754
The stack coloring pass has code to delete stores and loads that become
trivially dead after coloring. Extend it to cope with single instructions
that copy from one frame index to another.
The testcase happens to show an example of this kicking in at the moment.
It did occur in Real Code too though.
llvm-svn: 185705
The stack coloring pass renumbered frame indexes with a loop of the form:
for each frame index FI
for each instruction I that uses FI
for each use of FI in I
rename FI to FI'
This caused problems if an instruction used two frame indexes F0 and F1
and if F0 was renamed to F1 and F1 to F2. The first time we visited the
instruction we changed F0 to F1, then we changed both F1s to F2.
In other words, the problem was that SSRefs recorded which instructions
used an FI, but not which MachineOperands and MachineMemOperands within
that instruction used it.
This is easily fixed for MachineOperands by walking the instructions
once and processing each operand in turn. There's already a loop to
do that for dead store elimination, so it seemed more efficient to
fuse the two at the block level.
MachineMemOperands are more tricky because they can be shared between
instructions. The patch handles them by making SSRefs an array of
MachineMemOperands rather than an array of MachineInstrs. We might end
up processing the same MachineMemOperand twice, but that's OK because
we always know from the SSRefs index what the original frame index was.
llvm-svn: 185703
Eric Christopher