These intrinsics are simply expanded to regular icmp/fcmp instructions.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D121594
As a result of making these legal, and tweaking the combine to allow vectors,
we generate vector G_SEXT_INREG during legalization.
The reason we want to make these legal in the first place is to allow for
more combine opportunities. Once those have been done, we can just lower them
back to shifts in the post-legalizer lowering.
This needs to be one commit otherwise we start causing tests to fail due to
incomplete support for selection etc.
Since SROA chooses promotion based on reaching load / stores of allocas, we may run into scenarios in which we alloca a vector, but promote it to an integer. The result of which is the familiar LoadCombine pattern (i.e. ZEXT, SHL, OR). However, instead of coming directly from distinct loads, the elements to be combined are coming from ExtractVectorElements which stem from a shared load.
This patch identifies such a pattern and combines it into a load.
Change-Id: I0bc06588f11e88a0a975cde1fd71e9143e6c42dd
Fix a crash in the FMA combine added by D132837 and amended by D134810.
In cases where the newly created node could be folded, the combiner
would fail this assertion:
llc: DAGCombiner.cpp:268: void (anonymous namespace)::DAGCombiner::AddToWorklist(llvm::SDNode *): Assertion `N->getOpcode() != ISD::DELETED_NODE && "Deleted Node added to Worklist"' failed.
Differential Revision: https://reviews.llvm.org/D135150
Vector support seems to work immediately, as long as we run the combine before
legalization (so the vector SELECTs don't get lowered) and the legalizer rules
are there to enable generation.
Differential Revision: https://reviews.llvm.org/D135047
We have a very common pattern of dispatching between BUILD_VECTOR and SPLAT_VECTOR creation repeated in many cases in code. Common the pattern into a utility function.
This adds a combine that handles
```
(x + y) - y -> x
(x + y) - x -> y
x - (y + x) -> 0 - y
x - (x + z) -> 0 - z
```
On AArch64, we get added benefit for `0 - y` because it can be selected to a
`neg` instruction.
Differential Revision: https://reviews.llvm.org/D135010
The code introduced in https://reviews.llvm.org/D130881 has a bug as it may cause a use-after-free error that can be caught by ASAN.
The bug essentially boils down to iterator invalidation of `DenseMap`. The expression `SDEI[To] = I->second;` may cause `SDEI` to grow if `To` is inserted for the very first time. When that happens, all existing iterators to the map are invalidated as their backing storage has been freed. Accessing `I->second` is then invalid and attempts to access freed memory (as `I` is an iterator of `SDEI`).
This patch fixes that quite simply by first making a copy of `I->second`, and then moving into the possibly newly inserted KV of the ` DenseMap`.
No test attached as I am not sure it is practible to test.
Differential revision: https://reviews.llvm.org/D135019
Includes handling of constants with vector type in isKnownNeverNaN.
For AMDGPU results in not making fcanonicalize during legalization
for vector inputs to fmaxnum_ieee and fminnum_ieee. Does not affect
end result since there is a combine that eliminates fcanonicalize.
Differential Revision: https://reviews.llvm.org/D88573
Before, the isPreLegalize() query in CombinerHelper only checked for the
presence of a LegalizerInfo object. This is problematic when we want to have
a combine actually check for legality in a pre-legalizer combine pass, since
if we pass a LegalizerInfo object to the constructor it causes the combines to
think that we're running *post* legalizer, which isn't true.
This change fixes it to instead check an explicit bool that passes to signal
whether the pass will be run before or after legalization.
Doing so exposed a bug in the extending loads combine, which tried to check for
legality of candidate extending loads if LegalizerInfo was present. Since we
only ran it pre-legalizer and therefore with a null LegalizerInfo, it never
actually ran. Also fixes the legality checks to keep the tests passing.
Differential Revision: https://reviews.llvm.org/D135044
Much like f16 and f32, we shouldn't try to shrink bf16 to smaller fp
constant. The code may not be optimal, but this allows us to legalize
bf16 constants under Arm without errors.
D133866 added the llvm::isNeutralConstant helper to track neutral/passthrough constants
This patch updates foldSelectWithIdentityConstant to use the helper instead of maintaining its own opcode handling
Differential Revision: https://reviews.llvm.org/D134966
Function buildCopyToRegs did not handle properly the case when it should
make wider vector result. It happened, for example, in a function that
returns value of type <2 x f32>, which should be widen to <4 x f32> to
fit XMM register. The function eventually calls
MachineIRBuilder.buildUnmerge, which does not expect that only one
destination register is specified.
Now this case is treated specifically in buildCopyToRegs.
Differential Revision: https://reviews.llvm.org/D128546
Using this helper makes work about neutral elements more easier. Although I only
find one case now, I think it will have more chance to be used since so many
combine works are related to neutral elements.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D133866
In the spirit of D130765 . Get rid of cbranches and/or cmov. Usually shorter, but sometime not, becaus eit's hard to prededict when dependency breaking xor will be introduced.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D134736
The bulk of the implementation is common between 'release' mode (==AOT-ed
model) and 'development' mode (for training), the main difference is
that in development mode, we may also log features (for training logs),
inject scoring information and then produce the log file.
Differential Revision: https://reviews.llvm.org/D133616
In D132837, an existing v_fma combine was extended to regard nested
fma instructions. Originally, the inner FMA was checked for being used
only once. In its current state, this check is missing, which causes
some regressions.
In this patch, this check was added.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D134856
Given something like this:
```
declare signext i16 @signext_callee()
define i32 @caller() {
%res = call i16 @signext_callee()
...
}
```
CallLowering would miss that signext_callee's return value is sign extended,
because it isn't on the call.
Use hasRetAttr on the CallBase to allow us to catch this.
(This now inserts G_ASSERT_SEXT/G_ASSERT_ZEXT like in the original review.)
Differential Revision: https://reviews.llvm.org/D86228
Add a utility function which returns true if the given value is a constant
false value.
This is necessary to port one of the compare simplifications in
TargetLowering::SimplifySetCC.
Differential Revision: https://reviews.llvm.org/D91754
Interestingly, MathExtras.h doesn't use <cmath> declaration, so move it out of
that header and include it when needed.
No functional change intended, but there's no longer a transitive include
fromMathExtras.h to cmath.
This commit adds in two new features to the ML regalloc eviction
analysis that can be used in ML models, a vector of MBB frequencies and
a vector of indicies mapping instructions to their corresponding basic
blocks. This will allow for further experimentation with per-instruction
features and give a lot more flexibility for future experimentation over
how we're extracting MBB frequency data currently.
Reviewed By: mtrofin, jacobhegna
Differential Revision: https://reviews.llvm.org/D134166
D132837 introduced a new DAG combine that used MorphNodeTo to morph an
FMUL into an FMA. It turns out that MorphNodeTo does not properly update
the divergence bit for users of the morphed node, causing an assertion
failure on the new test case:
llc: SelectionDAG.cpp:10486: void llvm::SelectionDAG::VerifyDAGDivergence(): Assertion `calculateDivergence(N) == N->isDivergent() && "Divergence bit inconsistency detected"' failed.
Fixing MorphNodeTo to propagate the divergence bit is tricky because of
the way it is used to select machine instructions, so use getNode and
ReplaceAllUsesOfValueWith instead.
Differential Revision: https://reviews.llvm.org/D134810
On a rv64 without f32 or vector support, this will be passed across
the basic block as an i64. We need use i32 as an intermediate type
with bitcast and anyext/trunc.
Fixes PR58025
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D134758
This removes the ptrtoint from the load's pointer operand, although we
can't entirely eliminate these to get the LSB shift. In a future
patch, this will avoid ptrtoint in the case where the atomic is
overaligned to the word size.
For gathers which load in 8 and 16 bit data then use that data
as an index, the index can be extended to 32 bits instead of
64 bits
Differential Revision: https://reviews.llvm.org/D130692
Previous commit 8b00b24f85 missed to add `int_ceil` anchor for the
llvm.ceil.* section under LangRef.rst
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D134586
This seems to be beneficial overall, except for midpoint-int.ll .
The X86 backend seems to generate zeroing that are not necesary.
Reviewed By: shchenz
Differential Revision: https://reviews.llvm.org/D131260
Add vp.maxnum and vp.minnum which are vector predicted intrinsics of llvm.maxnum
and llvm.minnum.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D134639
This feature implements support for making entries in the exception section
on XCOFF on the direct assembly path using the ".except" pseudo-op. It also
provides functionality to lower entries (comprised of language and reason
codes) into the exception section through the use of annotation metadata
attached to llvm.ppc.trap/trapd/tw/tdw intrinsics. Integrated assembler
support will be provided in another review. https://reviews.llvm.org/D133030
needs to merge first for LIT tests
Reviewed By: shchenz, RKSimon
Differential Revision: https://reviews.llvm.org/D132146
Promote f16 to f32 and use the f32 libcall.
I deleted rv64zfh-half-intrinsics-strict.ll because it only existed due to this issue breaking rv32.
Differential Revision: https://reviews.llvm.org/D134579
This extends the uniform base transform used with scatter/gather to support one-use vector adds-of-splats with a non-zero base. This has the effect of essentially reassociating an add from vector to scalar domain.
The motivation is to improve the lowering of scatter/gather operations fed by complex geps.
Differential Revision: https://reviews.llvm.org/D134472
RISCV doesn't actually support a scaled form of indexed load and store. We previously handled this by forming the scaled SDNode, and then doing custom legalization during lowering. This patch instead adds a callback via TLI to prevent formation entirely.
This has two effects:
* First, the GEP gets expanded (and used). Instead of the shift being created with an SDLoc of the memory operation, it has the SDLoc of the GEP instruction. This avoids the scheduler perturbing IR order when there's no reason to.
* Second, we fix what appears to be a bug in index calculation with RV32. The rules for GEPs require index calculation be done in particular bitwidth, and it appears the custom legalization code got this wrong for the case where index type exceeds pointer width. (Or at least, I trust the generic GEP lowering to be correct a lot more.)
The DAGCombiner change to handle VPScatter/VPGather is technically separate, but is required to prevent a regression on those intrinsics.
Differential Revision: https://reviews.llvm.org/D134382
The transform to fold an add into the base of a scatter/gather was only checking to see if the LHS was a splat. Included test change indicates that splats are not canonicalized to LHS, and that we need to check both sides.
The accessibility level of a typedef or using declaration in a
struct or class was being lost when producing debug information.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D134339
This pass allows a user to dump a MIR function to a dot file
and view it as a graph. It is targeted to provide a similar
functionality as -dot-cfg pass on LLVM-IR. As of now the pass
also support below flags:
-dot-mcfg-only [optional][won't print instructions in the
graph just block name]
-mcfg-dot-filename-prefix [optional][prefix to add to output dot file]
-mcfg-func-name [optional] [specify function name or it's
substring, handy if mir file contains multiple functions and
you need to see graph of just one]
More flags and details can be introduced as per the requirements
in future. This pass is inspired from -dot-cfg IR pass and APIs
are written in almost identical format.
Patch by Yashwant Singh <Yashwant.Singh@amd.com> (yassingh)
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D133709
I don't know what was going on originally with these tests. It seems reasonable
to have the immediate be the same byte alignment unit as the IR, in which case
we need to take the log2 in order to set the right number of low bits.
This fixes a miscompile in chromium.
Differential Revision: https://reviews.llvm.org/D134380
This patch uses the API provided by LiveRangeEdit to handle rematerialization.
It will make future maintenance and improvement more easier.
No functional change.
Differential Revision: https://reviews.llvm.org/D133610
This patch changes a FADD / FMUL => FMA ISel pattern implemented
in D80801 so that it peeks through more than one FMA.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D132837
I'm not sure why the SEXT_INREG was gated on a bitwidth check of the mask
vs element size.
This fixes a miscompile in chromium's skia library.
Differential Revision: https://reviews.llvm.org/D134236
The following changes are necessasy to get the generated tree
matcher to compile:
- In CodeExpansions::declare(), the assert() prevents connecting
two instructions. E.g. the match code
(match (MUL $t, $s1, $s2),
(SUB $d, $t, $s3)),
results in two declarations of $t, one for the def and one for
the use. Removing the assertion allows this construct.
If $t is later used, it is one of the operands, which should be
perfectly fine.
- The code emitted in GIMatchTreeVRegDefPartitioner::generatePartitionSelectorCode()
is not compilable:
- The value of NewInstrID should be emitted, not the name
- Both calls involving getOperand() end with one parenthesis too many
- Swaps generated condition for the partition code in the latter function
It also changes the rules i2p_to_p2i, fabs_fabs_fold, and fneg_fneg_fold
to use the tree matcher for a linear match. These rules are tested by:
CodeGen/AArch64/GlobalISel/combine-fabs.mir
CodeGen/AArch64/GlobalISel/combine-fneg.mir
CodeGen/AArch64/GlobalISel/combine-ptrtoint.mir
CodeGen/AMDGPU/GlobalISel/combine-add-nullptr.mir
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D133257
This patch adds in instruction based features to the regalloc advisor
gated behind a flag so a user can decide at runtime whether or not they
want to enable the feature. The features are only enabled when LLVM is
compiled in MLGO develpment mode (LLVM_HAVE_TF_API) is set to true.
To extract the instruction features, I'm taking a list of segments from
each LiveInterval and noting the start and end SlotIndices. This list is then
sorted based on the start SlotIndex and I iterate through each SlotIndex
to grab instructions, making sure to check for overlaps. This results in
a vector of opcodes and binary mapping matrix that maps live ranges to the
opcodes of the instructions within that LR.
Reviewed By: mtrofin
Differential Revision: https://reviews.llvm.org/D131930
Use salvageDebugInfo for instructions erased as trivially dead in
GlobalISel.
It would be helpful to implement support of G_PTR_ADD and G_FRAME_INDEX
in salvageDebugInfo in future in order to preserve more variable
location.
Reviewed by: arsenm
Differential Revision: https://reviews.llvm.org/D133986
This is a follow-up to D133777, which resolved a use-after-free case but
did not cover all possible memory bugs due to misplacement of loads.
In short, the overall problem was that sinked loads could be moved after
state-modifying instructions leading to memory bugs.
The solution is to restrict load sinking unless it is found to be sound.
i) Within a basic block (to-be-sinked load and select-user are in the same BB),
loads can be sinked only if there is no intervening state-modifying instruction.
This is a conservative approach to avoid resorting to alias analysis to detect
potential memory overlap.
ii) Across basic blocks, sinking of loads is avoided. This is because going over
multiple basic blocks looking for memory conflicts could be computationally
expensive and also unlikely to allow loads to sink. Further, experiments showed
that not sinking these loads has a slight positive performance effect.
Maybe for some of these loads, having some separation allows enough time
for the load to be executed in time for its user. This is not the case for
floating point operations that benefit more from sinking.
The solution in D133777 was essentially undone in this patch,
since the latter is a complete solution to the observed problem.
Overall, the performance impact of this patch is minimal.
Tested on two internal Google workloads with instrPGO.
Search application showed <0.05% perf difference,
while the database one showed a slight improvement,
but not statistically significant.
Reviewed By: davidxl
Differential Revision: https://reviews.llvm.org/D133999
This is a partial port of the code used by the SelectionDAGBuilder to
translate selects.
In particular, see matchSelectPattern in ValueTracking.cpp. This is a
GISel-equivalent of the portion which handles fminnum/fmaxnum/fminimum/fmaximum.
I tried to set it up so it'd be easy to add the non-FP cases. Those are simpler.
On the AArch64-end, it seems like the FP cases are more important for perf
right now, so I bit the bullet and went at the more complicated problem. :)
I elected to do this as a post-legalize combine rather than in the
IRTranslator because
Deciding which fmax/fmin to use can depend on legalization rules
Philosophically-speaking (TM), putting it in a combine just feels cleaner
Being able to enable/disable the combine is handy
Another option would be to use the ValueTracking code in the IRTranslator and
match what SelectionDAGBuilder::visitSelect does. I think that may be somewhat
annoying since we'd need to write lowerings back into the selects in the
legalizer. I'm not strongly opposed to the approach.
We'd also want to be careful with vector selects once that's implemented,
which explicitly check if a vector select is legal on the target. That'd
probably need a hook.
From what I can tell, doing this as a combine is probably a cleaner option
long-term.
Differential Revision: https://reviews.llvm.org/D116702
Currently, it can become extremely costly to compute MayIncreasePressure if the
size of CSUses turns out to be very large. In that case, it's no longer cost
effective to keep computing MayIncreasePressure. Therefore, to limit the amount
of time spent in isProfitableToCSE, we simply conservatively assume
MayIncreasePressure if the size of CSUses is too large. This can reduce overall
compile time by 30% for some benchmarks.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D134003
I want to default all VP operations to Expand. These 2 were blocking
because VE doesn't support them and the tests were expecting them
to fail a specific way. Using Expand caused them to fail differently.
Seemed better to emulate them using operations that are supported.
@simoll mentioned on Discord that VE has some expansion downstream. Not
sure if its done like this or in the VE target.
Reviewed By: frasercrmck, efocht
Differential Revision: https://reviews.llvm.org/D133514
used, and MUL_LOHI is available
Folding into a sra(mul) / srl(mul) into a mulh introduces an extra
multiplication to compute the high half of the multiplication,
while it is more profitable to compute the high and lower halfs with a
single mul_lohi.
Differential Revision: https://reviews.llvm.org/D133768
The IR stack protector pass should insert stack checks before the tail
calls not only the musttail calls. So that the attributes `ssqreq` and
`tail call`, which are emited by llvm-opt, could be both enabled by
llvm-llc.
Reviewed By: compnerd
Differential Revision: https://reviews.llvm.org/D133860
On AArch64, doing the vector truncate separately after the fptoui
conversion can be lowered more efficiently using tbl.4, building on
D133495.
https://alive2.llvm.org/ce/z/T538CC
Depends on D133495
Reviewed By: t.p.northover
Differential Revision: https://reviews.llvm.org/D133496
Similar to using tbl to lower vector ZExts, tbl4 can be used to lower
vector truncates.
The initial version support i32->i8 conversions.
Depends on D120571
Reviewed By: t.p.northover
Differential Revision: https://reviews.llvm.org/D133495
Callee save registers must be preserved, so -fzero-call-used-regs
should not be zeroing them. The previous implementation only did
not zero callee save registers that were saved&restored inside the
function, but we need preserve all of them.
Fixes https://github.com/llvm/llvm-project/issues/57692.
Differential Revision: https://reviews.llvm.org/D133946
The method of counting resource consumption is modified to be based on
"Cycles" value when DFA is not used.
The calculation of ResMII is modified to total "Cycles" and divide it
by the number of units for each resource. Previously, ResMII was
excessive because it was assumed that resources were consumed for
the cycles of "Latency" value.
The method of resource reservation is modified similarly. When a
value of "Cycles" is larger than 1, the resource is considered to be
consumed by 1 for cycles of its length from the scheduled cycle.
To realize this, ResourceManager maintains a resource table for all
slots. Previously, resource consumption was always 1 for 1 cycle
regardless of the value of "Cycles" or "Latency".
In addition, the number of micro operations per cycle is modified to
be constrained by "IssueWidth". To disable the constraint,
--pipeliner-force-issue-width=100 can be used.
For the case of using DFA, the scheduling results are unchanged.
Reviewed By: dpenry
Differential Revision: https://reviews.llvm.org/D133572
This patch contains changes necessary to carry physical condition register (SCC) dependencies through the SDNode scheduler. It adds the edge in the SDNodeScheduler dependency graph instead of inserting the SCC copy between each definition and use. This approach lets the scheduler place instructions in an optimal way placing the copy only when the dependency cannot be resolved.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D133593
This patch extends CodeGenPrepare to lower zext v16i8 -> v16i32 in loops
using a wide shuffle creating a v64i8 vector, selecting groups of 3
zero elements and an element from the input.
This is profitable on AArch64 where such shuffles can be lowered to tbl
instructions, but only in loops, because it requires materializing 4
masks, which can be done in the loop preheader.
This is the only reason the transform is part of CGP. If there's a
better alternative I missed, please let me know. The same goes for the
shouldReplaceZExtWithShuffle hook which guards this. I am not sure if
this transform will be beneficial on other targets, but it seems like
there is no way other convenient way.
This improves the generated code for loops like the one below in
combination with D96522.
int foo(uint8_t *p, int N) {
unsigned long long sum = 0;
for (int i = 0; i < N ; i++, p++) {
unsigned int v = *p;
sum += (v < 127) ? v : 256 - v;
}
return sum;
}
https://clang.godbolt.org/z/Wco866MjY
Reviewed By: t.p.northover
Differential Revision: https://reviews.llvm.org/D120571
All in-tree targets pass pointer-sized ConstantSDNodes to the
method. This overload reduced amount of boilerplate code a bit. This
also makes getCALLSEQ_END consistent with getCALLSEQ_START, which
already takes uint64_ts.
The class priority is expected to be at most 5 bits before it starts
clobbering bits used for other fields. Also clamp the instruction
distance in case we have millions of instructions.
AMDGPU was accidentally overflowing into the global priority bit in
some cases. I think in principal we would have wanted this, but in the
cases I've looked at, it had the counter intuitive effect and
de-prioritized the large register tuple.
Avoid using weird bit hack PPC uses for global priority. The
AllocationPriority field is really 5 bits, and PPC was relying on
overflowing this to 6-bits to forcibly set the global priority
bit. Split this out as a separate flag to avoid having magic behavior
for values above 31.
MachineMemOperand::print can dereference a NULL pointer if TII
is not passed from the printMemOperand. This does not happen while
dumping the DAG/MIR from llc but crashes the debugger if a dump()
method is called from gdb.
Differential Revision: https://reviews.llvm.org/D133903
Get some load-store forwarding cases for big-endian where a larger store covers
a smaller load, and the offset would be 0 and handled on little-endian but on
big-endian the offset is adjusted to be non-zero. The idea is just to shift the
data to make it look like the offset 0 case.
Differential Revision: https://reviews.llvm.org/D130115
https://reviews.llvm.org/D133637 fixes the problem where we should hash raw content of
register mask instead of the pointer to it.
Fix the same issue in `llvm::hash_value()`.
Remove the added API `MachineOperand::getRegMaskSize()` to avoid potential confusion.
Add an assert to emphasize that we probably should hash a machine operand iff it has
associated machine function, but keep the fallback logic in the original change.
Reviewed By: MatzeB
Differential Revision: https://reviews.llvm.org/D133747
There's no real read of the register, so the copy introduced a new
live value. Make sure we introduce a replacement implicit_def instead
of just erasing the copy.
Found from llvm-reduce since it tries to set undef on everything.
When a select is converted to a branch and load instructions are sinked to the true/false blocks,
lifetime intrinsics (if present) could be made unsound if not moved.
This conservatively moves all lifetime intrinsics in a transformed BB to the end block to ensure
preserved lifetime semantics.
Reviewed By: davidxl
Differential Revision: https://reviews.llvm.org/D133777
This is the ultimate fallback code if UADDO isn't supported.
If the target uses 0/1 we used one compare, but if the target doesn't
use 0/1 we emitted two compares. Regardless of boolean constants we
should only need to check that the Result is less than one of the
original operands. So we only need one compare.
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D133708
Some uses of std::make_pair and the std::pair's first/second members
in the ScheduleDAGInstrs.[cpp|h] files were replaced with using of the
vector's emplace_back along with structure bindings from C++17.
This attempts to stop the type promotion pass transforming where it is
not profitable, by not marking PhiNodes as ToPromote and being more
aggressive about pulling extends out of loops.
Differential Revision: https://reviews.llvm.org/D133203
This patch refactors SlotIndex::getInstrDistance to
SlotIndex::getApproxInstrDistance to better describe the actual
functionality of this function. This patch also adds in some additional
comments better documenting the assumptions that this function makes to
increase clarity.
Based on discussion on the LLVM Discourse:
https://discourse.llvm.org/t/odd-behavior-in-slotindex-getinstrdistance/64934/5
Reviewed By: mtrofin, foad
Differential Revision: https://reviews.llvm.org/D133386
The bit masking lowering only works for vectors of scalars, so for pointer
element types we need to add some casting.
Differential Revision: https://reviews.llvm.org/D133672
__declspec(safebuffers) is equivalent to
__attribute__((no_stack_protector)). This information is recorded in
CodeView.
While we are here, add support for strict_gs_check.
I'm planning to deprecate and eventually remove llvm::empty.
I thought about replacing llvm::empty(x) with std::empty(x), but it
turns out that all uses can be converted to x.empty(). That is, no
use requires the ability of std::empty to accept C arrays and
std::initializer_list.
Differential Revision: https://reviews.llvm.org/D133677
MachineOperand::getRegMask() returns a pointer to register mask. We should hash the raw content of register mask instead of its pointer.
Reviewed By: kyulee
Differential Revision: https://reviews.llvm.org/D133637
For remainder:
If (1 << (Bitwidth / 2)) % Divisor == 1, we can add the high and low halves
together and use a (Bitwidth / 2) urem. If (BitWidth /2) is a legal integer
type, this urem will be expand by DAGCombiner using multiply by magic
constant. We do have to take into account that adding high and low
together can produce a carry, making it a (BitWidth / 2)+1 bit number.
So we need to also add back in the carry from the first addition.
For division:
We can use the above trick to compute the remainder, subtract that
remainder from the dividend, then multiply by the multiplicative
inverse of the Divisor modulo (1 << BitWidth).
This is based on the section "Remainder by Summing Digits" in
Hacker's delight.
The remainder trick is similar to a trick you may have learned for
determining if a decimal number is divisible by 3. You can add all the
digits together and see if the sum is divisible by 3. If you're not sure
if the sum is divisible by 3, you can add its digits together. This
can be repeated until you have a single decimal digit. If that digit
is 3, 6, or 9, then the original number is divisible by 3. This works
because 10 % 3 == 1.
gcc already does this same trick. There are additional tricks gcc
does urem as well as srem, udiv, and sdiv that I plan to add in
future patches.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D130862
Also remove new-pass-manager version of ExpandLargeDivRem because there is no way
yet to access TargetLowering in the new pass manager.
Differential Revision: https://reviews.llvm.org/D133691
Simplify extended add/sub (with carry-in and carry-out) to add/sub with
carry (with carry-out only) if carry-in is known to be zero.
Differential Revision: https://reviews.llvm.org/D133702
This was only trying this to relax register class constraints, but
this can also help if there are subranges involved.
This solves a compilation failure for AMDGPU when there is high
pressure created by large register tuples. If one virtual register is
using most of the available budget, we need to be able to evict
subranges.
This solves the immediate failure, but this solution leaves a lot to
be desired. In the relevant testcases, we have 32-element tuples but
most of the uses are operations on 1 element subranges of it. What
we're now getting is a spill and restore of the full 1024 bits and an
extract of the used 32-bits. It would be far better if we introduced a
copy to a new virtual register with a smaller register class and used
narrower spills.
Furthermore, we could probably do a better job if the allocator were
to introduce new subranges where none previously existed in the
highest pressure scenarios. The block and region splits should also
try to split specific subranges out.
The mve-vst3.ll test changes looks like noise to me, but instruction
count increased by one. mve-vst4.ll looks like a solid improvement
with several 16-byte spills eliminated. splitkit-copy-live-lanes.mir
also shows a solid reduction in total spill count.
This could use more tests but it's pretty tiring to come up with cases
that fail on this.
Some uses of std::make_pair and the std::pair's first/second members
in the ScheduleDAGRRList.cpp file were replaced with using of the
vector's emplace_back along with structure bindings from C++17.
Fixes#50098
LLVM uses X19 as the frame base pointer, if it needs to. Meaning you
can get warnings if you clobber that with inline asm.
However, it doesn't explain why. The frame base register is not part
of the ABI so it's pretty confusing why you get that warning out of the blue.
This adds a method to explain a reserved register with X19 as the first one.
The logic is the same as getReservedRegs.
I could have added a return parameter to isASMClobberable and friends
but found that there's a lot of things that call isReservedReg in various
ways.
So while one more method on the pile isn't great design, it is simpler
right now to do it this way and only pay the cost if you are actually using
a reserved register.
Reviewed By: lenary
Differential Revision: https://reviews.llvm.org/D133213
Replacing the following instances of UndefValue with PoisonValue, where the UndefValue is used as an arbitrary value:
- llvm/lib/CodeGen/WinEHPrepare.cpp
`demotePHIsOnFunclets`: RAUW arbitrary value for lingering uses of removed PHI nodes
- llvm/lib/Transforms/Utils/BasicBlockUtils.cpp
`FoldSingleEntryPHINodes`: Removes a self-referential single entry phi node.
- llvm/lib/Transforms/Utils/CallGraphUpdater.cpp
`finalize`: Remove all references to removed functions.
- llvm/lib/Transforms/Utils/ScalarEvolutionExpander.cpp
`cleanup`: the result is not used then the inserted instructions are removed.
- llvm/tools/bugpoint/CrashDebugger.cpp
`TestInts`: the program is cloned and instructions are removed to narrow down source of crash.
Differential Revision: https://reviews.llvm.org/D133640
getNegatedExpression can delete nodes. If the first call to
getNegatedExpression produced a node that the second call also
manages to create, it might get deleted. Use a HandleSDNode to
ensure it has a use to prevent it from being deleted.
Fixes PR57658.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D133602
The LLVM performance tips suggest that allocas should be placed at the
beginning of the entry block. So far, llvm doesn’t provide any helper to
find that position.
Add BasicBlock::getFirstNonPHIOrDbgOrAlloca and IRBuilder::SetInsertPointPastAllocas(Function*)
that get an insert position after the (static) allocas at the start of a
function and use it in ShadowStackGCLowering.
Differential Revision: https://reviews.llvm.org/D132554
LLVM contains a helpful function for getting the size of a C-style
array: `llvm::array_lengthof`. This is useful prior to C++17, but not as
helpful for C++17 or later: `std::size` already has support for C-style
arrays.
Change call sites to use `std::size` instead.
Differential Revision: https://reviews.llvm.org/D133429
This patch introduces the priority analysis and the priority advisor,
the default implementation, and the scaffolding for introducing the
other implementations of the advisor.
Reviewed By: mtrofin
Differential Revision: https://reviews.llvm.org/D132835
Interpret MD_pcsections in AsmPrinter emitting the requested metadata to
the associated sections. Functions and normal instructions are handled.
Differential Revision: https://reviews.llvm.org/D130879
Propagate (most) PC sections metadata to MachineInstr when GlobalISel is
doing instruction selection.
This change results in support for architectures using GlobalISel (such
as -O0 with AArch64). Not all instructions may be supported yet, and
requires further target-specific handling (such as done for AArch64
pseudo-atomics). Expanding supported instructions is planned on a
case-by-case basis and new use cases for PC sections metadata.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D130886
When expanding IR atomics to target-specific atomics, copy all
!pcsections Metadata to expanded atomics automatically.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D130885
Propagate PC sections metadata to MachineInstr when FastISel is doing
instruction selection.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D130884
Add a new entry to SDNodeExtraInfo to propagate PCSections through
SelectionDAG.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D130882
Details:
Currently CodeGenPrepare is very time consuming in handling big functions.
Old Algorithm :
It iterate each BB in function, and go on handle very instructions in BB.
Due to some instruction optimizations may affect the BBs' dominate tree.
The old logic will re-iterate and try optimize for each BB.
Suppose we have a big function with 20000 BBs, If we handled the last BB
with fine tuning the dominate tree. We need totally re-iterate and try optimize
the 20000 BBs from the beginning.
The Complex is near N!
And we really encounter somes big tests (> 20000 BBs) that cost more than 30
mins in this pass. (Debug version compiler will cost 2 hours here)
What this patch do for huge function ?
It mainly changes the iteration way for optimization.
1 We do optimizeBlock for each BB (that is same with old way).
And, in the meaning time, If BB is changed/updated in the optimization, it will
be put into FreshBBs (try do optimizeBlock again).
The new created BB at previous iteration will also put into FreshBBs.
2 For the BBs which not updated at previous iteration, we directly skip it.
Strictly speaking, here may miss some opportunity, but the probability is very
small.
3 For Instructions in single BB, we do optimizeInst for each instruction.
If optimizeInst change the instruction dominator in this BB, rather than break
and go back to optimize the first BB (the old way), we directly iterate
instructions (to do optimizeInst) in this updated BB again (the new way).
What this patch do for small/normal (not huge) function ?
It is same with the Old Algorithm. (NFC)
Reviewed By: LuoYuanke
Differential Revision: https://reviews.llvm.org/D129352
This patch is essentially an alternative to https://reviews.llvm.org/D75836 and was mentioned by @lhames in a comment.
The gist of the issue is that Mach-O has restrictions on which kind of sections are allowed after debug info has been emitted, which is also properly asserted within LLVM. Problem is that stack maps are currently emitted as one of the last sections in each target-specific AsmPrinter so far, which would cause the assertion to trigger. The current approach of special casing for the `__LLVM_STACKMAPS` section is not viable either, as downstream users can overwrite the stackmap format using plugins, which may want to use different sections.
This patch fixes the issue by emitting the stack map earlier, right before debug info is emitted. The way this is implemented is by taking the choice when to emit the StackMap away from the target AsmPrinter and doing so in the base class. The only disadvantage of this approach is that the `StackMaps` member is now part of the base class, even for targets that do not support them. This is functionaly not a problem however, as emitting an empty `StackMaps` is a no-op.
Differential Revision: https://reviews.llvm.org/D132708
During SelectionDAG legalization SDNodes with associated extra info may
be replaced with a new SDNode. Preserve associated extra info on
ReplaceAllUsesWith and remove entries in DeallocateNode.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D130881
For information infrequently attached to SDNodes, it is useful to
provide a way to add this information out-of-line. This is already done
for call-site specific information.
Rename CallSiteDbgInfo to NodeExtraInfo in preparation of adding
additional information not necessarily related to call sites only.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D130880
This adds the ExpandLargeDivRem to the default pass pipeline.
The limit at which it expands div/rem instructions is configured
via a new TargetTransformInfo hook (default: no expansion)
X86, Arm and AArch64 backends implement this hook to expand div/rem
instructions with more than 128 bits.
Differential Revision: https://reviews.llvm.org/D130076
Provide MachineInstr::setPCSection(), to propagate relevant metadata
through the backend. Use ExtraInfo to store the metadata.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D130876
The main difference is that this preserves intermediate rounding steps,
which the other route doesn't. This aligns bfloat16 more with half
floats, which use this path on most targets.
I didn't understand what the difference was between these softening
approaches when I first added bfloat lowerings, would be nice if we only
had one of them.
Based on @pengfei 's D131502
Differential Revision: https://reviews.llvm.org/D133207
The issue was with processing two subregs of the same reg are used in the same
instruction (e.g. inline asm): "def early-clobber" and other just "def".
Register coalescer ran in bad recursion if the early clobbered subreg is second
in the following sequence of COPYs.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D127136
MachineVerifier tried to checkLivenessAtDef() ignoring it is actually a subreg.
The issue was with processing two subregs of the same reg are used in the same
instruction (e.g. inline asm): "def early-clobber" and other just "def".
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D126661
For some indices we can simply extract the fixed-length subvector from the
low half of the scalable vector, for example when the index is less than the
minimum number of elements in the low half. For all other cases we can
expand the operation through the stack by storing out the vector and
reloading the fixed-length part we need.
Fixes https://github.com/llvm/llvm-project/issues/55412
Tests added here:
CodeGen/AArch64/sve-extract-fixed-from-scalable-vector.ll
Differential Revision: https://reviews.llvm.org/D117499
Similar to #57402 - we were calling isGuaranteedNotToBeUndefOrPoison on the freeze operand (with Depth = 0), but wasn't accounting for the fact that a later isGuaranteedNotToBeUndefOrPoison assertion will call from the new node (with Depth = 0 as well) - which will then recursively call isGuaranteedNotToBeUndefOrPoison for its operands with Depth = 1
Fixes#57554
The way ComputeNumSignBits was being used was only correct if
OuterBitSize is exactly 2x InnerBitSize. Which is always true,
but not obviously so. Comparing ComputeMaxSignificantBits to
InnerBitSize feels more correct.
CGP uses a raw `getInstructionCost(I, TargetTransformInfo::TCK_SizeAndLatency) >= TCC_Expensive` check to see if its better to move an expensive instruction used in a select behind a branch instead.
This is causing issues with upcoming improvements to TCK_SizeAndLatency costs on X86 as we need to use TCK_SizeAndLatency as an uop count (so its compatible with various target-specific buffer sizes - see D132288), but we can have instructions that have a low TCK_SizeAndLatency value but should still be treated as 'expensive' (FDIV for example) - by adding a isExpensiveToSpeculativelyExecute wrapper we can keep the current behaviour but still add an x86 override in a future patch when the cost tables are updated to compensate.
[MachineFunctionPass] Support -filter-passes for -print-changed
-filter-passes specifies a `PassID` (a lower-case dashed-separated pass name,
also used by -print-after, -stop-after, etc) instead of a CamelCasePass.
`-filter-passes=CamelCaseNewPMPass` seems like a workaround for new PM passes before
we can use lower-case dashed-separated pass names (as used by `-passes=`).
Example:
```
# getPassName() is "IRTranslator". PassID is "irtranslator"
llc -mtriple=aarch64 -print-changed -filter-passes=irtranslator < print-changed-machine.ll
```
Close https://github.com/llvm/llvm-project/issues/57453
Reviewed By: aeubanks
Differential Revision: https://reviews.llvm.org/D133055
DwarfEhPrepare inserts calls to _Unwind_Resume into landing pads.
If _Unwind_Resume happens to be defined in the same module and
debug info is used, then this leads to a verifier error:
inlinable function call in a function with debug info must
have a !dbg location
call void @_Unwind_Resume(ptr %exn.obj) #0
Fix this by assigning a dummy location to the call. (As this
happens in the backend, inlining is not actually relevant here.)
Fixes https://github.com/llvm/llvm-project/issues/57469.
Differential Revision: https://reviews.llvm.org/D133095
Some targets like RISC-V require operands to be inspected to
determine if an instruction is similar to a move.
Spotted while investigating code differences between using an ADDI
vs an ADDIW. RISC-V has the isAsCheapAsAMove flag for ADDI, but
the TII hook checks the immediate is 0 or the register is X0. ADDIW
is never generated with X0 or with an immediate of 0 so it doesn't
have the isAsCheapAsAMove flag.
I don't know enough about the PRE code to write a test for this yet.
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D132981
Warn if `.size` is specified for a function symbol. The size of a
function symbol is determined solely by its content.
I noticed this simplification was possible while debugging #57427, but
this change doesn't fix that specific issue.
Differential Revision: https://reviews.llvm.org/D132929
We feed the result from the first extractShiftForRotate call into the second, and that result might no longer be a shift op (usually due to constant folding).
NOTE: We REALLY need to stop creating nodes on the fly inside extractShiftForRotate!
Fixes Issue #57474
We were calling isGuaranteedNotToBeUndefOrPoison on operands (with Depth = 0), but wasn't accounting for the fact that a later isGuaranteedNotToBeUndefOrPoison assertion will call from the new node (with Depth = 0 as well) - which will then recursively call isGuaranteedNotToBeUndefOrPoison for its operands with Depth = 1
Fixes#57402
IIUC, the conversion part is not part of atomic operations and fences should be put around converted atomic operations.
This also fixes atomic load of floating point values which requires fence on PowerPC.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D127609
The provided testcase would previously fail with an assertion due to later down below trying to allocate registers for `token` return types and arguments. This is especially problematic as the process would then exit instead of falling back to using FastIsel.
This patch fixes that by simply explicitly failing translation if either of these intrinsics are encountered.
Fixes https://github.com/llvm/llvm-project/issues/57349
Differential Revision: https://reviews.llvm.org/D132974
widenScalarDst updates the insert point to after MI, so
widenScalarSrc must be called before widenScalarDst. Otherwise
The updated Src values will appear after MI and break SSA. e.g.:
%14:_(s64), %15:_(s1) = G_UADDE %9:_, %11:_, %13:_
becomes
%14:_(s64), %16:_(s32) = G_UADDE %9:_, %11:_, %17:_
%15:_(s1) = G_TRUNC %16:_(s32)
%17:_(s32) = G_ZEXT %13:_(s1)
Differential Revision: https://reviews.llvm.org/D132547
Change-Id: Ie3458747a6879433f4d5ab9939d2bd102dd0f2db
Mostly just modeled after vp.fneg except there is a
"functional instruction" for fneg while fabs is always an
intrinsic.
Reviewed By: fakepaper56
Differential Revision: https://reviews.llvm.org/D132793
This patch replaces calls to greatestCommonDivisor with std::gcd where
two arguments are of the same type. This means that
std::common_type_t of the argument type is the same as the argument
type.
We could drop calls to std::abs in some cases, but that's left for
another patch.
This patch replaces calls to greatestCommonDivisor with std::gcd where
both arguments are known to be of unsigned. This means that
std::common_type_t of the two argument types should just be the wider
one of the two.
This patch replaces getLCMSize with std::lcm, a C++17 feature.
Note that all the arguments are of unsigned with no implicit type
conversion as they are passed to getLCMSize.
Adds a pass ExpandLargeDivRem to expand div/rem instructions
with more than 128 bits into a loop computing that value.
As discussed on https://reviews.llvm.org/D120327, this approach has the advantage
that it is independent of the runtime library. This also helps the clang driver,
which otherwise would need to understand enough about the runtime library
to know whether to allow _BitInts with more than 128 bits.
Targets are still free to disable this pass and instead provide a faster
implementation in a runtime library.
Fixes https://github.com/llvm/llvm-project/issues/44994
Differential Revision: https://reviews.llvm.org/D126644
This patch follows the InstCombine approach of stripping poison generating flags (nsw/nuw from add/sub etc.) to allow us to push a freeze() through the op. Unlike InstCombine it doesn't retain any flags, but we have plenty of DAG folds that do the same thing already. We assert that the newly generated op isGuaranteedNotToBeUndefOrPoison.
Similar to the ValueTracking approach, isGuaranteedNotToBeUndefOrPoison has been updated to confirm that if an op can't create undef/poison and its operands are guaranteed not to be undef/poison - then its not undef/poison. This is just for the generic opcodes - target specific opcodes will need to do this manually just in case they have some special cases.
Differential Revision: https://reviews.llvm.org/D132333
While this does not matter for most targets, when building for Arm Morello,
we have to mark the symbol as a function and add size information, so that
LLD can correctly evaluate relocations against the local symbol.
Since Morello is an out-of-tree target, I tried to reproduce this with
in-tree backends and with the previous reviews applied this results in
a noticeable difference when targeting Thumb.
Background: Morello uses a method similar Thumb where the encoding mode is
specified in the LSB of the symbol. If we don't mark the target as a
function, the relocation will not have the LSB set and calls will end up
using the wrong encoding mode (which will almost certainly crash).
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D131429
D132080 introduced a bug leading to `RegisterClassInfo` caches not
getting invalidated when there was exactly one more CSR register added.
Differential Revision: https://reviews.llvm.org/D132606
The KCFI sanitizer, enabled with `-fsanitize=kcfi`, implements a
forward-edge control flow integrity scheme for indirect calls. It
uses a !kcfi_type metadata node to attach a type identifier for each
function and injects verification code before indirect calls.
Unlike the current CFI schemes implemented in LLVM, KCFI does not
require LTO, does not alter function references to point to a jump
table, and never breaks function address equality. KCFI is intended
to be used in low-level code, such as operating system kernels,
where the existing schemes can cause undue complications because
of the aforementioned properties. However, unlike the existing
schemes, KCFI is limited to validating only function pointers and is
not compatible with executable-only memory.
KCFI does not provide runtime support, but always traps when a
type mismatch is encountered. Users of the scheme are expected
to handle the trap. With `-fsanitize=kcfi`, Clang emits a `kcfi`
operand bundle to indirect calls, and LLVM lowers this to a
known architecture-specific sequence of instructions for each
callsite to make runtime patching easier for users who require this
functionality.
A KCFI type identifier is a 32-bit constant produced by taking the
lower half of xxHash64 from a C++ mangled typename. If a program
contains indirect calls to assembly functions, they must be
manually annotated with the expected type identifiers to prevent
errors. To make this easier, Clang generates a weak SHN_ABS
`__kcfi_typeid_<function>` symbol for each address-taken function
declaration, which can be used to annotate functions in assembly
as long as at least one C translation unit linked into the program
takes the function address. For example on AArch64, we might have
the following code:
```
.c:
int f(void);
int (*p)(void) = f;
p();
.s:
.4byte __kcfi_typeid_f
.global f
f:
...
```
Note that X86 uses a different preamble format for compatibility
with Linux kernel tooling. See the comments in
`X86AsmPrinter::emitKCFITypeId` for details.
As users of KCFI may need to locate trap locations for binary
validation and error handling, LLVM can additionally emit the
locations of traps to a `.kcfi_traps` section.
Similarly to other sanitizers, KCFI checking can be disabled for a
function with a `no_sanitize("kcfi")` function attribute.
Relands 67504c9549 with a fix for
32-bit builds.
Reviewed By: nickdesaulniers, kees, joaomoreira, MaskRay
Differential Revision: https://reviews.llvm.org/D119296
The KCFI sanitizer, enabled with `-fsanitize=kcfi`, implements a
forward-edge control flow integrity scheme for indirect calls. It
uses a !kcfi_type metadata node to attach a type identifier for each
function and injects verification code before indirect calls.
Unlike the current CFI schemes implemented in LLVM, KCFI does not
require LTO, does not alter function references to point to a jump
table, and never breaks function address equality. KCFI is intended
to be used in low-level code, such as operating system kernels,
where the existing schemes can cause undue complications because
of the aforementioned properties. However, unlike the existing
schemes, KCFI is limited to validating only function pointers and is
not compatible with executable-only memory.
KCFI does not provide runtime support, but always traps when a
type mismatch is encountered. Users of the scheme are expected
to handle the trap. With `-fsanitize=kcfi`, Clang emits a `kcfi`
operand bundle to indirect calls, and LLVM lowers this to a
known architecture-specific sequence of instructions for each
callsite to make runtime patching easier for users who require this
functionality.
A KCFI type identifier is a 32-bit constant produced by taking the
lower half of xxHash64 from a C++ mangled typename. If a program
contains indirect calls to assembly functions, they must be
manually annotated with the expected type identifiers to prevent
errors. To make this easier, Clang generates a weak SHN_ABS
`__kcfi_typeid_<function>` symbol for each address-taken function
declaration, which can be used to annotate functions in assembly
as long as at least one C translation unit linked into the program
takes the function address. For example on AArch64, we might have
the following code:
```
.c:
int f(void);
int (*p)(void) = f;
p();
.s:
.4byte __kcfi_typeid_f
.global f
f:
...
```
Note that X86 uses a different preamble format for compatibility
with Linux kernel tooling. See the comments in
`X86AsmPrinter::emitKCFITypeId` for details.
As users of KCFI may need to locate trap locations for binary
validation and error handling, LLVM can additionally emit the
locations of traps to a `.kcfi_traps` section.
Similarly to other sanitizers, KCFI checking can be disabled for a
function with a `no_sanitize("kcfi")` function attribute.
Reviewed By: nickdesaulniers, kees, joaomoreira, MaskRay
Differential Revision: https://reviews.llvm.org/D119296
The diff modifies ext-tsp code layout algorithm in the following ways:
(i) fixes merging of cold block chains (this is a port of D129397);
(ii) adjusts the cost model utilized for optimization;
(iii) adjusts some APIs so that the implementation can be used in BOLT; this is
a prerequisite for D129895.
The only non-trivial change is (ii). Here we introduce different weights for
conditional and unconditional branches in the cost model. Based on the new model
it is slightly more important to increase the number of "fall-through
unconditional" jumps, which makes sense, as placing two blocks with an
unconditional jump next to each other reduces the number of jump instructions in
the generated code. Experimentally, this makes a mild impact on the performance;
I've seen up to 0.2%-0.3% perf win on some benchmarks.
Reviewed By: hoy
Differential Revision: https://reviews.llvm.org/D129893
This patch adds a Type operand to the TLI isCheapToSpeculateCttz/isCheapToSpeculateCtlz callbacks, allowing targets to decide whether branches should occur on a type-by-type/legality basis.
For X86, this patch proposes to allow CTTZ speculation for i8/i16 types that will lower to promoted i32 BSF instructions by masking the operand above the msb (we already do something similar for i8/i16 TZCNT). This required a minor tweak to CTTZ lowering - if the src operand is known never zero (i.e. due to the promotion masking) we can remove the CMOV zero src handling.
Although BSF isn't very fast, most CPUs from the last 20 years don't do that bad a job with it, although there are some annoying passthrough EFLAGS dependencies. Additionally, now that we emit 'REP BSF' in most cases, we are tending towards assuming this will most likely be executed as a TZCNT instruction on any semi-modern CPU.
Differential Revision: https://reviews.llvm.org/D132520
Based off Issue #57283 - we need to try harder to ensure we're not creating nodes on-the-fly - so make sure we're just using SelectionDAG for analysis where possible
extractShiftForRotate may fail to return canonicalized shifts due to constant folding or other simplification that can occur in getNode()
Fixes Issue #57283
(ctpop x) == 1 --> (x != 0) && ((x & x-1) == 0)
Adjust the legality check to avoid the poor codegen on AArch64.
We probably only want to use popcount on this pattern when it
is a single instruction.
fixes#57225
Differential Revision: https://reviews.llvm.org/D132237
This patch builds on prior support patches to enable support for
variadic debug values in InstrRefLDV, allowing DBG_VALUE_LISTs to
have their ranges extended.
Differential Revision: https://reviews.llvm.org/D128212
musttail should be honored even in the presence of attributes like "disable-tail-calls". SelectionDAG properly handles this.
Update LangRef to explicitly mention that this is the semantics of musttail.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D132193
This patch adds the last of the changes required to enable
DBG_VALUE_LIST handling in InstrRefLDV, handling variadic debug values
during the transfer tracking step. Most of the changes are fairly
straightforward, and based around tracking multiple locations per
variable in TransferTracker::VLocTracker.
Differential Revision: https://reviews.llvm.org/D128211
In preparation for supporting DBG_VALUE_LIST in InstrRefLDV, this patch
adds the logic for emitting DBG_VALUE_LIST instructions from
InstrRefLDV. The logical changes here are fairly simple, with the main
change being that instead of directly prepending offsets to the DIExpr,
we use appendOpsToArg to modify the expression for individual debug
operands in the expression. The function emitLoc is also changed to take
a list of debug ops, with an empty list meaning an undef value.
Differential Revision: https://reviews.llvm.org/D128209
CodeGenPrepare pass can sink pointer comparison across statepoint
to the point of use (see comment in IR/SafepointIRVerifier.cpp)
Due to specifics of statepoints, it is still legal to have tied
def and use rewritten to the same register in TwoAddress pass.
However, properly updating LiveIntervals and LiveVariables becomes
complicated. For simplicity, let's fall back to generic handling of
tied registers when we detect such case.
TODO: This fixes functional (assertion) failure. Ideally we should
try to recompute new live range/liveness in place.
Reviewed By: skatkov
Differential Revision: https://reviews.llvm.org/D132255
In preparation for adding support for DBG_VALUE_LIST instructions in
InstrRefLDV, this patch updates the logic for joining variables at block
joins to support joining variables that use multiple debug operands.
This is one of the more meaty "logical" changes, although the line count
isn't too high - this changes pickVPHILoc to find a valid joined
location for every operand, with part of the function being split off
into pickValuePHILoc which finds a location for a single operand.
Differential Revision: https://reviews.llvm.org/D128180
This interface allows a target to reject a proposed
SMS schedule. For Hexagon/PowerPC, all schedules
are accepted, leaving behavior unchanged. For ARM,
schedules which exceed register pressure limits are
rejected.
Also, two RegisterPressureTracker methods now need to be public so
that register pressure can be computed by more callers.
Reapplication of D128941/(reversion:D132037) with small fix.
Differential Revision: https://reviews.llvm.org/D132170
This completes the client side transition to the OperandValueInfo version of this routine. Backend TTI implementations still use the prior versions for now.
This patch fixes:
llvm/lib/CodeGen/LiveDebugValues/InstrRefBasedImpl.h:330:5: error:
anonymous types declared in an anonymous union are an extension
[-Werror,-Wnested-anon-types]
In preparation for allowing InstrRefBasedLDV to handle DBG_VALUE_LIST,
this patch updates the internal representation that it uses to represent
debug values to store a list of values. This is one of the more
significant changes in terms of line count, but is fairly simple and
should not affect the output of this pass.
Differential Revision: https://reviews.llvm.org/D128177
`RegisterClassInfo` caches information like allocation orders and reuses
it for multiple machine functions where possible. However the `MCPhysReg
*CalleeSavedRegs` field used to test whether the set of callee saved
registers changed did not work: After D28566
`MachineRegisterInfo::getCalleeSavedRegs()` can return dynamically
computed CSR sets that are only valid while the `MachineRegisterInfo`
object of the current function exists.
This changes the code to make a copy of the CSR list instead of keeping
a possibly invalid pointer around.
Differential Revision: https://reviews.llvm.org/D132080
Currently, InstrRefLDV only handles DBG_VALUE instructions, not
DBG_VALUE_LIST, and as a result of this it handles these instructions
using functions that only work for that type of debug value, i.e. using
getOperand(0) to get the debug operand. This patch changes this to use
the generic debug value functions, such as getDebugOperand and
isDebugOffsetImm, as well as adding an IsVariadic field to the
DbgValueProperties class and a few other minor changes to acknowledge
DBG_VALUE_LISTs. Note that this patch does not add support for
DBG_VALUE_LIST here, but is a precursor to other patches that do add
that support.
Differential Revision: https://reviews.llvm.org/D128174
In the InstrRefBasedImpl for LiveDebugValues, we attempt to propagate
debug values through basic blocks in part by checking to see whether all
a variable's incoming debug values to a BB "agree", i.e. whether their
properties match and they refer to the same underlying value.
Prior to this patch, the check for agreement between incoming values
relied on exact equality, which meant that a VPHI and a Def DbgValue
that referred to the same underlying value would be seen as disagreeing.
This patch changes this behaviour to treat them as referring to the same
value, allowing the shared value to propagate into the BB.
Differential Revision: https://reviews.llvm.org/D125953
Following the comment's thread of D117235, I added checks for the widening + splitting case, which also causes a split with one of the resulting vectors to be empty. Due to the same issues described in that same thread, the `fixed-vectors-strided-store.ll` test is missing the widening + splitting case, while the same case in the `strided-vpload.ll` test requires to manually split the loaded vector.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D121784
Test for a case we observed after the initial implementation of D129997
landed, in which case we observed a crash while building the ppc64le
Linux kernel. In that case, we had one block with two exits, both to the
same successor. Removing one of the exits corrupted the
successor/predecessor lists.
So when we have an INLINEASM_BR, check a few things for each indirect
target:
1. that it exists.
2. that it is listed in our successors.
3. that its predecessor list contains the parent MBB of INLINEASM_BR.
This would have caught the regression discovered after D129997 landed,
after the pass that was problematic (early-tailduplication) rather than
getting a stack trace in a later pass (regalloc) that doesn't understand
the anomaly and crashes.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D130290
Registers used for arguments are listed as "live-ins" into the starting
basic block. This means we don't have to go through a potentially
expensive search through all possible argument registers when we only
care about used argument registers.
Differential Revision: https://reviews.llvm.org/D132181
This patch introduces the priority analysis and the priority advisor,
the default implementation, and the scaffolding for introducing the
other implementations of the advisor.
Reviewed By: mtrofin
Differential Revision: https://reviews.llvm.org/D131220
Extends findMoreOptimalIndexType to allow ISD::BUILD_VECTOR based
indices to be truncated when such truncation is lossless. This can
enable the use of 32bit gather/scatter indices thus making it less
likely to have to split a gather/scatter in two.
Depends on D125194
Differential Revision: https://reviews.llvm.org/D130533
* Replace getUserCost with getInstructionCost, covering all cost kinds.
* Remove getInstructionLatency, it's not implemented by any backends, and we should fold the functionality into getUserCost (now getInstructionCost) to make it easier for targets to handle the cost kinds with their existing cost callbacks.
Original Patch by @samparker (Sam Parker)
Differential Revision: https://reviews.llvm.org/D79483
TragetLowering had two last InstructionCost related `getTypeLegalizationCost()`
and `getScalingFactorCost()` members, but all other costs are processed in TTI.
E.g. it is not comfortable to use other TTI members in these two functions
overrided in a target.
Minor refactoring: `getTypeLegalizationCost()` now doesn't need DataLayout
parameter - it was always passed from TTI.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D117723
This is a potentially better alternative to D131452 that also
should avoid the infinite loop bug from:
issue #56403
This is again a minimal fix to reduce merging pain for the
release. But if this makes sense, then we might want to guard
all of the RTLIB generation (and other libcalls?) with a
similar name check.
Differential Revision: https://reviews.llvm.org/D131521
Improve copy statistics:
- Count copies from or to physical registers: They are used to model function parameters and calling conventions and the register allocator optimizes for them.
- Check physical registers assigned to virtual registers and stop counting "identity" `COPY`s where source and destination is the same physical registers; they will be removed in the `virtregmap` pass anyway.
Differential Revision: https://reviews.llvm.org/D131932
The current machine function splitter is reliant on profile data to do profile summary analysis to split blocks into cold section. This may sometimes limit the usage of machine function splitter especially in cases where we could do some form of static analysis to split out cold blocks if profile data is absent or profile data which may be faulty (Consider Sample PGO).
Of all code that could statically be marked cold Exception handling blocks are one of them (In fact BFI framework also tends to mark them as cold), and the most in size contribution. In my experiments I found out Exception handling pads and all code reachable from there account for up to 6-8% of the .text section on modern production binaries. This patch introduces a flag to split out all Exception handling blocks and blocks only reachable from Exceptional Handling pad to cold section. This flag has shown to give a performance win of up to 0.1% in terms of average cycles and instructions executed on internal facebook search service.
Reviewed By: snehasish
Differential Revision: https://reviews.llvm.org/D131824
This reverts commit 8c4aea438c.
Needed because buildbot failures (warnings) gave a clue that there was
a functional bug in the ARM rejection logic.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D132037
This interface allows a target to reject a proposed
SMS schedule. For Hexagon/PowerPC, all schedules
are accepted, leaving behavior unchanged. For ARM,
schedules which exceed register pressure limits are
rejected.
Also, two RegisterPressureTracker methods now need to be public so
that register pressure can be computed by more callers.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D128941
There are two different senses in which a block can be "address-taken".
There can be a BlockAddress involved, which means we need to map the
IR-level value to some specific block of machine code. Or there can be
constructs inside a function which involve using the address of a basic
block to implement certain kinds of control flow.
Mixing these together causes a problem: if target-specific passes are
marking random blocks "address-taken", if we have a BlockAddress, we
can't actually tell which MachineBasicBlock corresponds to the
BlockAddress.
So split this into two separate bits: one for BlockAddress, and one for
the machine-specific bits.
Discovered while trying to sort out related stuff on D102817.
Differential Revision: https://reviews.llvm.org/D124697
This patch fixes an issue where an instruction reading a whole register would be moved during register allocation into a spot where one of the subregisters was dead.
The code to check whether an instruction can be rematerialized at a given point or not was already checking for subranges to ensure that subregisters are live, but only when the instruction being moved was using a subregister, this patch changes that so the subranges are checked even when the moved instruction uses the full register.
This patch also adds a case to the original test for the subrange checking that trigger the issue described above.
The original subrange checking code was introduced in this revision: https://reviews.llvm.org/D115278
And I've encountered this issue on AMDGPUs while working with DPC++: https://github.com/intel/llvm/issues/6209
Essentially the greedy register allocator attempts to move the following instruction:
```
%3961:vreg_64 = V_LSHLREV_B64_e64 3, %3078:vreg_64, implicit $exec
```
From `@3440` into the body of a loop `@16312`, but `%3078` has the following live ranges:
```
%3078 [2224r,2240r:0)[2240r,3488B:1)[16192B,38336B:1) 0@2224r 1@2240r L0000000000000003 [2224r,3440r:0) 0@2224r L000000000000000C [2240r,3488B:0)[16192B,38336B:0) 0@2240r
```
So `@16312e` `%3078.sub1` is alive but `%3078.sub0` is dead, so this instruction being moved there leads to invalid memory accesses as `3078.sub0` ends up being trashed and the result of this instruction is used as part of an address calculation for a load.
On the original ticket this issue showed up on gfx906 and gfx90a but not on gfx908, this turned out to be because on gfx908 instead of moving the shift instruction into the loop, its value is spilled into an ACC register, gfx906 doesn't have ACC registers and for gfx90a ACC registers are used like regular vector registers and so aren't used for spilling.
With this patch the original application from the DPC++ ticket works properly on gfx906, and the result of the shift instruction is correctly spilled instead of moving the instruction in the loop.
Original Author: npmiller
Reviewed by: rampitec
Submitted by: rampitec
Differential Revision: https://reviews.llvm.org/D131884
Currently we treat initializers with init_seg(compiler/lib) as similar
to any other init_seg, they simply have a global variable in the proper
section (".CRT$XCC" for compiler/".CRT$XCL" for lib) and are added to
llvm.used. However, this doesn't match with how LLVM sees normal (or
init_seg(user)) initializers via llvm.global_ctors. This
causes issues like incorrect init_seg(compiler) vs init_seg(user)
ordering due to GlobalOpt evaluating constructors, and the
ability to remove init_seg(compiler/lib) initializers at all.
Currently we use 'A' for priorities less than 200. Use 200 for
init_seg(compiler) (".CRT$XCC") and 400 for init_seg(lib) (".CRT$XCL"),
which do not append the priority to the section name. Priorities
between 200 and 400 use ".CRT$XCC${Priority}". This allows for
some wiggle room for people/future extensions that want to add
initializers between compiler and lib.
Fixes#56922
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D131910
Static variables declared within a routine or lexical block should
be emitted with a non-qualified name. This allows the variables to
be visible to the Visual Studio watch window.
Differential Revision: https://reviews.llvm.org/D131400
When no landing pads exist for a function, `@LPStart` is undefined and must be omitted.
EH table is generally not emitted for functions without landing pads, except when the personality function is uknown (`!isNoOpWithoutInvoke(classifyEHPersonality(Per))`). In that case, we must omit `@LPStart` even when machine function splitting is enabled.
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D131626
Currently there is no way to add in development features to the ML
regalloc evict advisor which is useful to have when working on feature
engineering/improving the current model. This patch adds in the ability
to add in development features to the ML regalloc evict advisor which
are gated by a runtime flag and not added in at all if not compiled in
LLVM development mode. This sets the stage for future work where we are
planning on upstreaming some of the newer features that we are currently
experimenting with.
Reviewed By: mtrofin
Differential Revision: https://reviews.llvm.org/D131209
This is a followup to D131350, which caused another problem for i64
types being split into i32 on i32 targets. This patch tries to make sure
that either Illegal types are OK, or that the element types of a
buildvector are legal and bigger than or equal to the size of the
original elements.
Differential Revision: https://reviews.llvm.org/D131883
This patch really just extends D39946 towards stores as well as loads.
While the patch is in SelectionDAGBuilder, it only applies to AVR (the
only target that supports unaligned atomic operations).
Differential Revision: https://reviews.llvm.org/D128483
The outer signext_inreg is redundant in the following:
Fold (signext_inreg (extract_subvector (zext|anyext|sext iN_value to _) _) from iN)
-> (extract_subvector (signext iN_value to iM))
Tests are precommitted and clone those by analogy from the AND case in
the same file. Add a negative test to check extension width is handled
correctly.
This patch supersedes D130700.
Differential Revision: https://reviews.llvm.org/D131503
These are guaranteed not to create undef/poison (although they may pass through) - the associated ISD::VALUETYPE node is also guaranteed never to generate poison
Reland commit 719658d078
The base RA support infrastructure that only allow a specific register
class be allocated in RA pss. Since greedy RA, basic RA derived from
base RA, they all allow allocating specific register class. Fast RA
doesn't support allocating register for specific register class. This
patch is to enable ShouldAllocateClass in fast RA, so that it can
support allocating register for specific register class.
Differential Revision: https://reviews.llvm.org/D131825
SimplifyMultipleUseDemandedBits shouldn't be creating general nodes like this - although we allow bitcasts, even general constant folding is avoided.
Removing it causes a number of regressions that need addressing first, but I've added a TODO for now.
TLS debug on AIX is not ready for now.
The location generated in no-integrated-as mode is wrong and
in integrated-as mode causes AIX linker error.
Reviewed By: Esme
Differential Revision: https://reviews.llvm.org/D130245
Expand TypePromotion pass to try to promote PHI-nodes in loops that are the
operand of a ZExt, using the ZExt's result type to determine the Promote Width.
Differential Revision: https://reviews.llvm.org/D111237
This patch makes the variants of `mm*_cast*` intel intrinsics that use `shufflevector(freeze(poison), ..)` emit efficient assembly.
(These intrinsics are planned to use `shufflevector(freeze(poison), ..)` after shufflevector's semantics update; relevant thread: D103874)
To do so, this patch
1. Updates `LowerAVXCONCAT_VECTORS` in X86ISelLowering.cpp to recognize `FREEZE(UNDEF)` operand of `CONCAT_VECTOR` in addition to `UNDEF`
2. Updates X86InstrVecCompiler.td to recognize `insert_subvector` of `FREEZE(UNDEF)` vector as its first operand.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D130339
canCreateUndefOrPoison currently only handles unary ops, but we intend to change that soon - this more closely matches the pushFreezeToPreventPoisonFromPropagating behaviour where the freeze is pushed up to a single operand value, as long as all others are guaranteed not to be poison/undef.
However, pushFreezeToPreventPoisonFromPropagating would freeze all uses of the value - whilst this variant requires the frozen value to be only used in the op - we can look at generalize multiple uses later if the need arises.
Currently fcopysign for VLS vectors lowers through NEON even when the
vector width is wider than a NEON vector, causing bad codegen as the
vectors are split. This patch causes SVE to be used for these vectors
instead, giving much better codegen on wide VLS vectors.
Reviewed By: paulwalker-arm
Differential Revision: https://reviews.llvm.org/D128642
This is a follow-up patch to D130999. In the test, the MIR contains an
unreachable MBB but the code attempts to look it up in MLocs. This
patch fixes this issue by checking for the default-constructed value.
rdar://97226240
Differential Revision: https://reviews.llvm.org/D131453
The register operand of DBG_VALUE is not selected to a proper register
bank in both AArch64 and X86. This would cause getRegClass crash after
global ISel. After discussion, we think the MIR should assume all
vritual register should be set proper register class after global ISel,
so this patch is to fix the gap of DBG_VALUE for AArch64 and X86.
Differential Revision: https://reviews.llvm.org/D129037
The previous code overwrites VRMap for prologue stages during Phi
generation if a register spans many stages.
As a result, the wrong register is used as the one coming from
the prologue in Phis at later stages. (A process exists to correct
this, but it does not work in all cases.)
In addition, VRMap for prologue must be preserved until addBranches().
This patch fixes them by separating the map for Phis into a different
variable (VRMapPhi).
Reviewed By: bcahoon
Differential Revision: https://reviews.llvm.org/D127840
The build breakages should be addressed by d4abdd2e3d:
[CMake] Check CMAKE_CXX_STANDARD and error if it's to old
Thanks to Tobias and Roy for addressing these issues.
This patch adds basic support for a DAG variant of the canCreateUndefOrPoison call and updates DAGCombiner::visitFREEZE to use it, further Opcodes (including target specific Opcodes) can be handled when we have test coverage.
So far, I've left visitFREEZE to just use this for unary nodes (which currently means the existing BITCAST/FREEZE cases) - later patches will add other unary opcodes (with test coverage) and we can also refactor visitFREEZE to support a general number of operands like we do in InstCombinerImpl::pushFreezeToPreventPoisonFromPropagating.
I'm not aware of any vector test freeze coverage so the DemandedElts (and the Depth) args are not being used yet - but they are in place. Similarly we will be able to handle poison generating SDNodeFlags as and when it becomes an issue.
Part of the work for D106675 / PR50468
Differential Revision: https://reviews.llvm.org/D130646
This patch emits table lookup in expandCTTZ.
Context -
https://reviews.llvm.org/D113291 transforms set of IR instructions to
cttz intrinsic but there are some targets which does not support CTTZ or
CTLZ. Hence, I generate a table lookup in TargetLowering::expandCTTZ().
Differential Revision: https://reviews.llvm.org/D128911
FoldConstantArithmetic can fold constant vectors hidden behind bitcasts (e.g. vXi64 -> v2Xi32 on 32-bit platforms), but currently bails if either vector contains undef elements. These undefs can often occur due to SimplifyDemandedBits/VectorElts calls recognising that the upper bits are often unnecessary (e.g. funnel-shift/rotate implicit-modulo and AND masks).
This patch adds a basic 'FoldValueWithUndef' handler that will attempt to constant fold if one or both of the ops are undef - so far this just handles the AND and MUL cases where we always fold to zero.
The RISCV codegen increase is interesting - it looks like the BUILD_VECTOR lowering was loading a constant pool entry but now (with all elements defined constant) it can materialize the constant instead?
Differential Revision: https://reviews.llvm.org/D130839
D129150 added a combine from shuffles to And that creates a BUILD_VECTOR
of constant elements. We need to ensure that the elements are of a legal
type, to prevent asserts during lowering.
Fixes#56970.
Differential Revision: https://reviews.llvm.org/D131350
A const reference is preferred over a non-null const pointer.
`Type *` is kept as is to match the other overload.
Reviewed By: davidxl
Differential Revision: https://reviews.llvm.org/D131197
1) Overloaded (instruction-based) method is a wrapper around the current (opcode-based) method.
2) This patch also changes a few callsites (VectorCombine.cpp,
SLPVectorizer.cpp, CodeGenPrepare.cpp) to call the overloaded method.
3) This is a split of D128302.
Differential Revision: https://reviews.llvm.org/D131114
Fraser Cormack