Masked loads and store fit naturally with MVE, the instructions being easily
predicated. This adds lowering for the simple cases of masked loads and stores.
It does not yet deal with widening/narrowing or pre/post inc.
The llvm masked load intrinsic will accept a "passthru" value, dictating the
values used for the zero masked lanes. In MVE the instructions write 0 to the
zero predicated lanes, so we need to match a passthru that isn't 0 (or undef)
with a select instruction to pull in the correct data after the load.
We also need to do something with unaligned loads/stores. Currently this uses a
similar method used in big endian, using an VLDRB.8 (and potentially a VREV in
BE). This does mean that the predicate mask is converted from, for example, a
v4i1 to a v16i1. The VLDR instructions are defined as using the first bit of
the relevant mask lane, so this could potentially load different results if the
predicate is little odd. As the input is a v4i1 however, I believe this is OK
and all the bits required should be set in the predicate, making the VLDRB.8
load the same data.
Differential Revision: https://reviews.llvm.org/D66534
llvm-svn: 370329
This adds fp16 VMOVX patterns, using the same patterns as rL362482 with some
adjustments for MVE. It allows us to move fp16 registers without going into and
out of gprs.
VMOVX is able to move the top bits from a fp16 in a fp reg into the bottom bits
of another register, zeroing the rest. This can be used for odd MVE register
lanes. The top bits are not read by fp16 instructions, so no move is required
there if we are dealing with even lanes.
Differential revision: https://reviews.llvm.org/D66793
llvm-svn: 370184
The CodeGen/Thumb2/mve-vaddv.ll test needed to be amended to reflect the
changes from the above patch.
This reverts commit cd53ff6, reapplying 7c6b229.
llvm-svn: 369638
It broke the bots, see e.g. http://lab.llvm.org:8011/builders/clang-cuda-build/builds/36275/
> This patch fixes shifts by a 128/256 bit shift amount. It also fixes
> codegen for shifts of 32 by delegating to LLVM's default optimisation
> instead of emitting a long shift.
>
> Tests that used to generate long shifts of 32 are updated to check for the
> more optimised codegen.
>
> Differential revision: https://reviews.llvm.org/D66519
>
> llvm-svn: 369626
llvm-svn: 369636
This patch fixes shifts by a 128/256 bit shift amount. It also fixes
codegen for shifts of 32 by delegating to LLVM's default optimisation
instead of emitting a long shift.
Tests that used to generate long shifts of 32 are updated to check for the
more optimised codegen.
Differential revision: https://reviews.llvm.org/D66519
llvm-svn: 369626
This patch adds vecreduce_add and the relevant instruction selection for
vaddv.
Differential revision: https://reviews.llvm.org/D66085
llvm-svn: 369245
The widening and narrowing MVE instructions like VLDRH.32 are only permitted to
use low tGPR registers. This means that if they are used for a stack slot,
where the register used is only decided during frame setup, we need to be able
to correctly pick a thumb1 register over a normal GPR.
This attempts to add the required logic into eliminateFrameIndex and
rewriteT2FrameIndex, only picking the FrameReg if it is a valid register for
the operands register class, and picking a valid scratch register for the
register class.
Differential Revision: https://reviews.llvm.org/D66285
llvm-svn: 369108
Two issues:
1. t2CMPri shouldn't use CPSR if it isn't predicated. This doesn't
really have any visible effect at the moment, but it might matter in the
future.
2. The t2CMPri generated for t2WhileLoopStart might need to use a
register that isn't LR.
My team found this because we have a patch to track register liveness
late in the pass pipeline. I'll look into upstreaming it to help catch
issues like this earlier.
Differential Revision: https://reviews.llvm.org/D66243
llvm-svn: 369069
We need to allow any alignment at least 2, not just exactly 2, so that the big
endian loads and stores can be selected successfully. I've also added extra BE
testing for the load and store tests.
Thanks to Oliver for the report.
Differential Revision: https://reviews.llvm.org/D66222
llvm-svn: 368996
Stack loads and stores were already working, but direct stores were not. This
adds the patterns for them, same as predicate loads.
Differential Revision: https://reviews.llvm.org/D66213
llvm-svn: 368988
This adds patterns for selecting trunc instructions from full vectors to i1's
vectors.
Differential Revision: https://reviews.llvm.org/D66201
llvm-svn: 368981
This adds pre- and post- increment and decrements for MVE loads and stores. It
uses the builtin pre and post load/store detection, unlike Neon. Loads are
selected with the code in tryT2IndexedLoad, stores are selected with tablegen
patterns. The immediates have a +/-7bit range, multiplied by the size of the
element.
Differential Revision: https://reviews.llvm.org/D63840
llvm-svn: 368305
This adds some missing patterns for big endian loads/stores, allowing unaligned
loads/stores to also be selected with an extra VREV, which produces better code
than aligning through a stack. Also moves VLDR_P0 to not be LE only, and
adjusts some of the tests to show all that working.
Differential Revision: https://reviews.llvm.org/D65583
llvm-svn: 368304
VLDRH needs to have an alignment of at least 2, including the
widening/narrowing versions. This tightens up the ISel patterns for it and
alters allowsMisalignedMemoryAccesses so that unaligned accesses are expanded
through the stack. It also fixed some incorrect shift amounts, which seemed to
be passing a multiple not a shift.
Differential Revision: https://reviews.llvm.org/D65580
llvm-svn: 368256
This adjusts the load/store tests for better testing of alignments. It also
adds some extra alignment 1 tests, useful for future commits.
llvm-svn: 368255
Currently we check whether LR is stored/loaded to/from inbetween the
loop decrement and loop end pseudo instructions. There's two problems
here:
- It relies on all load/store instructions being labelled as such in
tablegen.
- Actually any use of loop decrement is troublesome because the value
doesn't exist!
So we need to check for any read/write of LR that occurs between the
two instructions and revert if we find anything.
Differential Revision: https://reviews.llvm.org/D65792
llvm-svn: 368130
This adds big endian MVE patterns for bitcasts. They are defined in llvm as
being the same as a store of the existing type and the load into the new. This
means that they have to become a VREV between the two types, working in the
same way that NEON works in big-endian. This also adds some example tests for
bigendian, showing where code is and isn't different.
The main difference, especially from a testing perspective is that vectors are
passed as v2f64, and so are VREV into and out of call arguments, and the
parameters are passed in a v2f64 format. Same happens for inline assembly where
the register class is used, so it is VREV to a v16i8.
So some of this is probably not correct yet, but it is (mostly) self-consistent
and seems to be consistent with how llvm treats vectors. The rest we can
hopefully fix later. More details about big endian neon can be found in
https://llvm.org/docs/BigEndianNEON.html.
Differential Revision: https://reviews.llvm.org/D65581
llvm-svn: 367780
The VREV64 instruction is apparently unpredictable if Qd == Qm, due to the
cross-beat nature of the instruction. This adds an earlyclobber to Qd, which
seems to be the same way we deal with this on other instructions like the
write-back on loads and stores.
Differential Revision: https://reviews.llvm.org/D65502
llvm-svn: 367544
Revert the hardware loop upon finding a LoopEnd that doesn't target
the loop header, instead of asserting a failure.
Differential Revision: https://reviews.llvm.org/D65268
llvm-svn: 367296
This adds the patterns required to transform xor P0, -1 to a VPNOT. The
instruction operands have to change a little for this, adding an in and an out
VCCR reg and using a custom DecodeMVEVPNOT for the decode.
Differential Revision: https://reviews.llvm.org/D65133
llvm-svn: 367192
These are some better patterns for converting between predicates and floating
points. Much like the extends, we select "1"/"-1" or "0" depending on the
predicate value. Or we perform a compare against 0 to convert to a predicate.
Differential Revision: https://reviews.llvm.org/D65103
llvm-svn: 367191
Both WhileLoopStart and LoopEnd may get turned into a cmp and br pair,
so add an implicit def to these pseudo instructions in case that WLS
and LE aren't generated.
Differential Revision: https://reviews.llvm.org/D65275
llvm-svn: 367089
This removes the VCEQ/VCNE/VCGE/VCEQZ/etc nodes, just using two called VCMP and
VCMPZ with an extra operand as the condition code. I believe this will make
some combines simpler, allowing us to just look at these codes and not the
operands. It also helps fill in a missing VCGTUZ MVE selection without adding
extra nodes for it.
Differential Revision: https://reviews.llvm.org/D65072
llvm-svn: 366934
The prevents us from trying to convert an i1 predicate vector to a float, or
vice-versa. Better patterns are possible, which will follow in a subsequent
commit. For now we just expand them.
Differential Revision: https://reviews.llvm.org/D65066
llvm-svn: 366931
MVE VCMP instructions can use a general purpose register as the second operand.
This adds the combines for it, selecting from a compare of a vdup.
Differential Revision: https://reviews.llvm.org/D65061
llvm-svn: 366924
This adds a DeMorgan combine for OR's of compares to turn them into AND's,
helping prevent them from going into and out of gpr registers. It also fills in
the VCLE and VCLT nodes that MVE can select, allowing it to invert more
compares.
Differential Revision: https://reviews.llvm.org/D65059
llvm-svn: 366920
Add a number of folds to convert and(vcmp, vcmp) into a single VPT block, where
the second vcmp becomes predicated on the first.
The VCMP; VPST; VCMP will eventually be converted to VPT; VCMP in the
VPTBlockPass.
Differential Revision: https://reviews.llvm.org/D65058
llvm-svn: 366910
Much like integers, this adds MVE floating point compares and select. It
requires a lot more buildvector/shuffle code because we may need to expand the
compares without mve.fp, and requires support for and/or because of the way we
lower llvm condition codes.
Some original code by David Sherwood
Differential Revision: https://reviews.llvm.org/D65054
llvm-svn: 366909
This adds some basic, "worst case" handling for MVE predicate Or/And/Xor. It
does this by going into and out of GPRs, doing the operation on scalars.
Code by David Sherwood.
Differential Revision: https://reviews.llvm.org/D65053
llvm-svn: 366907
This adds support code for building and shuffling i1 predicate registers. It
generally uses two basic principles, either converting the predicate into an
scalar (through a PREDICATE_CAST) and doing scalar operations on it there, or
by converting the register to an full vector register and back.
Some of the code here is a not super efficient but will hopefully cover most
cases of moving i1 vectors around and can be improved in subsequent patches.
Some code by David Sherwood.
Differential Revision: https://reviews.llvm.org/D65052
llvm-svn: 366890
This adds the very basics for MVE vector predication, adding integer VCMP and
VSEL instruction support. This is done through predicate registers (MVT::v16i1,
MVT::v8i1, MVT::v4i1), but otherwise using same mechanics as NEON to custom
lower setcc's through ARMISD::VCXX nodes (VCEQ, VCGT, VCEQZ, etc).
An extra VCNE was added, as this can be handled sensibly by MVE's expanded
number of VCMP condition codes. (There are also VCLE and VCLT which are added
later).
VPSEL is also added here, simply selecting on the vselect.
Original code by David Sherwood.
Differential Revision: https://reviews.llvm.org/D65051
llvm-svn: 366885
While lowering test.set.loop.iterations, it wasn't checked how the
brcond was using the result and so the wls could branch to the loop
preheader instead of not entering it. The same was true for
loop.decrement.reg.
So brcond and br_cc and now lowered manually when using the hwloop
intrinsics. During this we now check whether the result has been
negated and whether we're using SETEQ or SETNE and 0 or 1. We can
then figure out which basic block the WLS and LE should be targeting.
Differential Revision: https://reviews.llvm.org/D64616
llvm-svn: 366809
ARMLowOverheadLoops would assert a failure if it did not find all the
pseudo instructions that comprise the hardware loop. Instead of doing
this, iterate through all the instructions of the function and revert
any remaining pseudo instructions that haven't been converted.
Differential Revision: https://reviews.llvm.org/D65080
llvm-svn: 366691
We need to ensure that the number of T's is correct when adding multiple
instructions into the same VPT block.
Differential revision: https://reviews.llvm.org/D65049
llvm-svn: 366684
We need to make sure that we are sensibly dealing with vectors of types v2i64
and v2f64, even if most of the time we cannot generate native operations for
them. This mostly adds a lot of testing, plus fixes up a couple of the issues
found. And, or and xor can be legal for v2i64, and shifts combining needs a
slight fixup.
Differential Revision: https://reviews.llvm.org/D64316
llvm-svn: 366106
This adds basic lowering for MVE shifts. There are many shifts in MVE, but the
instructions handled here are:
VSHL (imm)
VSHRu (imm)
VSHRs (imm)
VSHL (vector)
VSHL (register)
MVE, like NEON before it, doesn't have shift right by a vector (or register).
We instead have to negate the amount and shift in the opposite direction. This
means we have to convert any SHR's into a form of SHL (that is still signed or
unsigned) with a negated condition and selecting from there. MVE still does
have shifting by an immediate for SHL, ASR and LSR.
This adds lowering for these and for register forms, which work well for shift
lefts but may require an extra fold of neg(vdup(x)) -> vdup(neg(x)) to potentially
work optimally for right shifts.
Differential Revision: https://reviews.llvm.org/D64212
llvm-svn: 366056
The vmovlb instructions can be uses to sign or zero extend vector registers
between types. This adds some patterns for them and relevant testing. The
VBICIMM generation is also put behind a hasNEON check (as is already done for
VORRIMM).
Code originally by David Sherwood.
Differential Revision: https://reviews.llvm.org/D64069
llvm-svn: 366008
This selects integer VNEG instructions, which can be especially useful with shifts.
Differential Revision: https://reviews.llvm.org/D64204
llvm-svn: 366006
This simply makes the MVE integer min and max instructions legal and adds the
relevant patterns for them.
Differential Revision: https://reviews.llvm.org/D64026
llvm-svn: 366004
This adds support for the floor/ceil/trunc/... series of instructions,
converting to various forms of VRINT. They use the same suffixes as their
floating point counterparts. There is not VTINTR, so nearbyint is expanded.
Also added a copysign test, to show it is expanded.
Differential Revision: https://reviews.llvm.org/D63985
llvm-svn: 366003
This adds the patterns for minnm and maxnm from the fminnum and fmaxnum nodes,
similar to scalar types.
Original patch by Simon Tatham
Differential Revision: https://reviews.llvm.org/D63870
llvm-svn: 366002
This patch addresses a couple of problems:
1) The maximum supported offset of LE is -4094.
2) The offset of WLS also needs to be checked, this uses a
maximum positive offset of 4094.
The use of BasicBlockUtils has been changed because the block offsets
weren't being initialised, but the isBBInRange checks both positive
and negative offsets.
ARMISelLowering has been tweaked because the test case presented
another pattern that we weren't supporting.
llvm-svn: 365749
This adds some handling for VMOVimm, using the same method that NEON uses. We
create VMOVIMM/VMVNIMM/VMOVFPIMM nodes based on the immediate, and select them
using the now renamed ARMvmovImm/etc. There is also an extra 64bit immediate
mode that I have not yet added here.
Code by David Sherwood
Differential Revision: https://reviews.llvm.org/D63884
llvm-svn: 365178
This adds patterns for the simpler VAND, VORR and VEOR bitwise vector
instructions. It also adjusts the top16Zero PatLeaf to not match on vector
instructions, which can otherwise cause problems.
Code written by David Sherwood.
Differential Revision: https://reviews.llvm.org/D63867
llvm-svn: 365113
Passing a vector type over the soft-float ABI involves it being split
into four GPRs, so the first thing that has to happen at the start of
the function is to recombine those into a vector register. The ABI
types all vectors as v2f64, so we need to support BUILD_VECTOR for
that type, which I do in this patch by allowing it to be expanded in
terms of INSERT_VECTOR_ELT, and writing an ISel pattern for that in
turn. Similarly, I provide a rule for EXTRACT_VECTOR_ELT so that a
returned vector can be marshalled back into GPRs.
While I'm here, I've also added ISD::UNDEF to the list of operations
we turn back on in `setAllExpand`, because I noticed that otherwise it
gets expanded into a BUILD_VECTOR with explicit zero inputs, leading
to pointless machine instructions to zero out a vector register that's
about to have every lane overwritten of in any case.
Reviewers: dmgreen, ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63937
llvm-svn: 364910
If you compile with `-mattr=+mve` (enabling integer MVE instructions
but not floating-point ones), then the scalar FP //registers// exist
and it's legal to move things in and out of them, load and store them,
but it's not legal to do arithmetic on them.
In D60708, the calls to `addRegisterClass` in ARMISelLowering that
enable use of the scalar FP registers became conditionalised on
`Subtarget->hasFPRegs()` instead of `Subtarget->hasVFP2Base()`, so
that loads, stores and moves of those registers would work. But I
didn't realise that that would also enable all the operations on those
types by default.
Now, if the target doesn't have basic VFP, we follow up those
`addRegisterClass` calls by turning back off all the nontrivial
operations you can perform on f32 and f64. That causes several
knock-on failures, which are fixed by allowing the `VMOVDcc` and
`VMOVScc` instructions to be selected even if all you have is
`HasFPRegs`, and adjusting several checks for 'is this a double in a
single-precision-only world?' to the more general 'is this any FP type
we can't do arithmetic on?'. Between those, the whole of the
`float-ops.ll` and `fp16-instructions.ll` tests can now run in
MVE-without-FP mode and generate correct-looking code.
One odd side effect is that I had to relax the check lines in that
test so that they permit test functions like `add_f` to be generated
as tailcalls to software FP library functions, instead of ordinary
calls. Doing that is entirely legal, but the mystery is why this is
the first RUN line that's needed the relaxation: on the usual kind of
non-FP target, no tailcalls ever seem to be generated. Going by the
llc messages, I think `SoftenFloatResult` must be perturbing the code
generation in some way, but that's as much as I can guess.
Reviewers: dmgreen, ostannard
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63938
llvm-svn: 364909
Backend changes to enable WLS/LE low-overhead loops for armv8.1-m:
1) Use TTI to communicate to the HardwareLoop pass that we should try
to generate intrinsics that guard the loop entry, as well as setting
the loop trip count.
2) Lower the BRCOND that uses said intrinsic to an Arm specific node:
ARMWLS.
3) ISelDAGToDAG the node to a new pseudo instruction:
t2WhileLoopStart.
4) Add support in ArmLowOverheadLoops to handle the new pseudo
instruction.
Differential Revision: https://reviews.llvm.org/D63816
llvm-svn: 364733
This simply adds integer and floating point VMUL patterns for MVE, same as we
have add and sub.
Differential Revision: https://reviews.llvm.org/D63866
llvm-svn: 364643
This adds handling and tests for a number of floating point math routines,
which have no MVE instructions.
Differential Revision: https://reviews.llvm.org/D63725
llvm-svn: 364641
MVE has instructions to widen as it loads, and narrow as it stores. This adds
the required patterns and legalisation to make them work including specifying
that they are legal, patterns to select them and test changes.
Patch by David Sherwood.
Differential Revision: https://reviews.llvm.org/D63839
llvm-svn: 364636
This fills in the gaps for basic MVE loads and stores, allowing unaligned
access and adding far too many tests. These will become important as
narrowing/expanding and pre/post inc are added. Big endian might still not be
handled very well, because we have not yet added bitcasts (and I'm not sure how
we want it to work yet). I've included the alignment code anyway which maps
with our current patterns. We plan to return to that later.
Code written by Simon Tatham, with additional tests from Me and Mikhail Maltsev.
Differential Revision: https://reviews.llvm.org/D63838
llvm-svn: 364633
We don't have vector operations for these, so they need to be expanded for both
integer and float.
Differential Revision: https://reviews.llvm.org/D63595
llvm-svn: 364631
The same as integer arithmetic, we can add simple floating point MVE addition and
subtraction patterns.
Initial code by David Sherwood
Differential Revision: https://reviews.llvm.org/D63257
llvm-svn: 364629
This adds the first few patterns for MVE code generation, adding simple integer
add and sub patterns.
Initial code by David Sherwood
Differential Revision: https://reviews.llvm.org/D63255
llvm-svn: 364627
This patch adds necessary shuffle vector and buildvector support for ARM MVE.
It essentially adds support for VDUP, VREVs and some VMOVs, which are often
required by other code (like upcoming patches).
This mostly uses the same code from Neon that already generated
NEONvdup/NEONvduplane/NEONvrev's. These have been renamed to ARMvdup/etc and
moved to ARMInstrInfo as they are common to both architectures. Most of the
selection code seems to be applicable to both, but NEON does have some more
instructions making some parts specific.
Most code originally by David Sherwood.
Differential Revision: https://reviews.llvm.org/D63567
llvm-svn: 364626
"To" selects an odd-numbered GPR, and "Te" an even one. There are some
8.1-M instructions that have one too few bits in their register fields
and require registers of particular parity, without necessarily using
a consecutive even/odd pair.
Also, the constraint letter "t" should select an MVE q-register, when
MVE is present. This didn't need any source changes, but some extra
tests have been added.
Reviewers: dmgreen, samparker, SjoerdMeijer
Subscribers: javed.absar, eraman, kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D60709
llvm-svn: 364331
This provides the low-level support to start using MVE vector types in
LLVM IR, loading and storing them, passing them to __asm__ statements
containing hand-written MVE vector instructions, and *if* you have the
hard-float ABI turned on, using them as function parameters.
(In the soft-float ABI, vector types are passed in integer registers,
and combining all those 32-bit integers into a q-reg requires support
for selection DAG nodes like insert_vector_elt and build_vector which
aren't implemented yet for MVE. In fact I've also had to add
`arm_aapcs_vfpcc` to a couple of existing tests to avoid that
problem.)
Specifically, this commit adds support for:
* spills, reloads and register moves for MVE vector registers
* ditto for the VPT predication mask that lives in VPR.P0
* make all the MVE vector types legal in ISel, and provide selection
DAG patterns for BITCAST, LOAD and STORE
* make loads and stores of scalar FP types conditional on
`hasFPRegs()` rather than `hasVFP2Base()`. As a result a few
existing tests needed their llc command lines updating to use
`-mattr=-fpregs` as their method of turning off all hardware FP
support.
Reviewers: dmgreen, samparker, SjoerdMeijer
Subscribers: javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60708
llvm-svn: 364329
Summary:
When identifing instructions that can be folded into a MOVCC instruction,
checking for a predicate operand is not enough, also need to check for
thumb2 function, with restrict-IT, is the machine instruction eligible for
ARMv8 IT or not.
Notes in ARMv8-A Architecture Reference Manual, section "Partial deprecation of IT"
https://usermanual.wiki/Pdf/ARM20Architecture20Reference20ManualARMv8.1667877052.pdf
"ARMv8-A deprecates some uses of the T32 IT instruction. All uses of IT that apply to
instructions other than a single subsequent 16-bit instruction from a restricted set
are deprecated, as are explicit references to the PC within that single 16-bit
instruction. This permits the non-deprecated forms of IT and subsequent instructions
to be treated as a single 32-bit conditional instruction."
Reviewers: efriedma, lebedev.ri, t.p.northover, jmolloy, aemerson, compnerd, stoklund, ostannard
Reviewed By: ostannard
Subscribers: ostannard, javed.absar, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D63474
llvm-svn: 363739
This patch changes MIR stack-id from an integer to an enum,
and adds printing/parsing support for this in MIR files. The default
stack-id '0' is now renamed to 'default'.
This should make MIR tests that have stack objects with different stack-ids
more descriptive. It also clarifies code operating on StackID.
Reviewers: arsenm, thegameg, qcolombet
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D60137
llvm-svn: 363533
Summary:
I found the following case having tail blocks with no successors merging opportunities after block placement.
Before block placement:
bb0:
...
bne a0, 0, bb2:
bb1:
mv a0, 1
ret
bb2:
...
bb3:
mv a0, 1
ret
bb4:
mv a0, -1
ret
The conditional branch bne in bb0 is opposite to beq.
After block placement:
bb0:
...
beq a0, 0, bb1
bb2:
...
bb4:
mv a0, -1
ret
bb1:
mv a0, 1
ret
bb3:
mv a0, 1
ret
After block placement, that appears new tail merging opportunity, bb1 and bb3 can be merged as one block. So the conditional constraint for merging tail blocks with no successors should be removed. In my experiment for RISC-V, it decreases code size.
Author of original patch: Jim Lin
Reviewers: haicheng, aheejin, craig.topper, rnk, RKSimon, Jim, dmgreen
Reviewed By: Jim, dmgreen
Subscribers: xbolva00, dschuff, javed.absar, sbc100, jgravelle-google, aheejin, kito-cheng, dmgreen, PkmX, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D54411
llvm-svn: 363284
Those two subtarget features were awkward because their semantics are
reversed: each one indicates the _lack_ of support for something in
the architecture, rather than the presence. As a consequence, you
don't get the behavior you want if you combine two sets of feature
bits.
Each SubtargetFeature for an FP architecture version now comes in four
versions, one for each combination of those options. So you can still
say (for example) '+vfp2' in a feature string and it will mean what
it's always meant, but there's a new string '+vfp2d16sp' meaning the
version without those extra options.
A lot of this change is just mechanically replacing positive checks
for the old features with negative checks for the new ones. But one
more interesting change is that I've rearranged getFPUFeatures() so
that the main FPU feature is appended to the output list *before*
rather than after the features derived from the Restriction field, so
that -fp64 and -d32 can override defaults added by the main feature.
Reviewers: dmgreen, samparker, SjoerdMeijer
Subscribers: srhines, javed.absar, eraman, kristof.beyls, hiraditya, zzheng, Petar.Avramovic, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D60691
llvm-svn: 361845
The new cortex-m schedule in rL360768 helps performance, but can increase the
amount of high-registers used. This, on average, ends up increasing the
codesize by a fair amount (because less instructions are converted from T2 to
T1). On cortex-m at -Oz, where we are quite size-paranoid, it is better to use
the existing DAG scheduler with the RegPressure scheduling preference (at least
until the issues around T2 vs T1 instructions can be improved).
I have also made sure that the Sched::RegPressure dag scheduler is always
chosen for MinSize.
The test shows one case where we increase the number of registers used.
Differential Revision: https://reviews.llvm.org/D61882
llvm-svn: 360769
This patch adds a simple Cortex-M4 schedule, renaming the existing M3
schedule to M4 and filling in the latencies as-per the Cortex-M4 TRM:
https://developer.arm.com/docs/ddi0439/latest
Most of these are 1, with the important exception being loads taking 2
cycles. A few others are also higher, but I don't believe they make a
large difference. I've repurposed the M3 schedule as the latencies are
mostly the same between the two cores, with the M4 having more FP and
DSP instructions. We also turn on MISched and UseAA for the cores that
now use this.
It also adds some schedule Write's to various instruction to make things
simpler.
Differential Revision: https://reviews.llvm.org/D54142
llvm-svn: 360768
We use both -long-option and --long-option in tests. Switch to --long-option for consistency.
In the "llvm-readelf" mode, -long-option is discouraged as it conflicts with grouped short options and it is not accepted by GNU readelf.
While updating the tests, change llvm-readobj -s to llvm-readobj -S to reduce confusion ("s" is --section-headers in llvm-readobj but --symbols in llvm-readelf).
llvm-svn: 359649
Summary:
Extract the logic for doing reassociations
from DAGCombiner::reassociateOps into a helper
function DAGCombiner::reassociateOpsCommutative,
and use that helper to trigger reassociation
on the original operand order, or the commuted
operand order.
Codegen is not identical since the operand order will
be different when doing the reassociations for the
commuted case. That causes some unfortunate churn in
some test cases. Apart from that this should be NFC.
Reviewers: spatel, craig.topper, tstellar
Reviewed By: spatel
Subscribers: dmgreen, dschuff, jvesely, nhaehnle, javed.absar, sbc100, jgravelle-google, hiraditya, aheejin, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61199
llvm-svn: 359476
The check for creating CBZ in constant island pass recently obtained the
ability to search backwards to find a Cmp instruction. The code in IfCvt should
mirror this to allow more conversions to the smaller form. The common code has
been pulled out into a separate function to be shared between the two places.
Differential Revision: https://reviews.llvm.org/D60090
llvm-svn: 358977
Ifcvt can replicate instructions as it converts them to be predicated. This
stops that from happening on thumb2 targets at minsize where an extra IT
instruction is likely needed.
Differential Revision: https://reviews.llvm.org/D60089
llvm-svn: 358974
This should hopefully lead to minor improvements in code generation, and
more accurate spill/reload comments in assembly.
Also fix isLoadFromStackSlotPostFE/isStoreToStackSlotPostFE so they
don't lead to misleading assembly comments for merged memory operands;
this is technically orthogonal, but in practice the relevant memory
operand lists don't show up without this change.
Differential Revision: https://reviews.llvm.org/D59713
llvm-svn: 356963
The constant island pass currently only looks at the instruction immediately
before a branch for a CMP to fold into a CBZ/CBNZ. This extends it to search
backwards for the instruction that defines CPSR. We need to ensure that the
register is not overridden between the CMP and the branch.
Differential Revision: https://reviews.llvm.org/D59317
llvm-svn: 356336
First step towards PR40800 - I intend to move the float case in a separate future patch.
I had to tweak the (overly reduced) thumb2 test and the x86 widening test change is annoying (no longer rematerializable) but we should address this separately.
Differential Revision: https://reviews.llvm.org/D59244
llvm-svn: 356040
This adds a number of missing Thumb1 opcodes so that the peephole optimiser can
remove redundant CMP instructions.
Reapplying this after the first attempt broke non-thumb1 code as the t2ADDri
instruction can be used with frame indices. In thumb1 we use tADDframe.
Differential Revision: https://reviews.llvm.org/D57833
llvm-svn: 354667
The Arm peephole optimiser code keeps track of both an MI and a SubAdd that can
be used to optimise away a CMP. In the rare case that both are found and not
ruled-out as valid, we could end up setting the flags on the wrong one.
Instead make sure we are using SubAdd if it exists, as it will be closer to the
CMP.
The testcase here is a little theoretical, with a dead def of cpsr. It should
hopefully show the point.
Differential Revision: https://reviews.llvm.org/D58176
llvm-svn: 354018
Currently in Arm code, we allocate LR first, under the assumption that
it needs to be saved anyway. Unfortunately this has the disadvantage
that it will require any instructions using it to be the longer thumb2
instructions, not the shorter thumb1 ones.
This switches the order when we are optimising for minsize, returning to
the default order so that more lower registers can be used. It can end
up requiring more pushed registers, but on average produces smaller code.
Differential Revision: https://reviews.llvm.org/D56008
llvm-svn: 351938
Part of the effort to refactoring frame pointer code generation. We used
to use two function attributes "no-frame-pointer-elim" and
"no-frame-pointer-elim-non-leaf" to represent three kinds of frame
pointer usage: (all) frames use frame pointer, (non-leaf) frames use
frame pointer, (none) frame use frame pointer. This CL makes the idea
explicit by using only one enum function attribute "frame-pointer"
Option "-frame-pointer=" replaces "-disable-fp-elim" for tools such as
llc.
"no-frame-pointer-elim" and "no-frame-pointer-elim-non-leaf" are still
supported for easy migration to "frame-pointer".
tests are mostly updated with
// replace command line args ‘-disable-fp-elim=false’ with ‘-frame-pointer=none’
grep -iIrnl '\-disable-fp-elim=false' * | xargs sed -i '' -e "s/-disable-fp-elim=false/-frame-pointer=none/g"
// replace command line args ‘-disable-fp-elim’ with ‘-frame-pointer=all’
grep -iIrnl '\-disable-fp-elim' * | xargs sed -i '' -e "s/-disable-fp-elim/-frame-pointer=all/g"
Patch by Yuanfang Chen (tabloid.adroit)!
Differential Revision: https://reviews.llvm.org/D56351
llvm-svn: 351049
Add t2TEQrr to the map of instructions with can be reduced down into
a T1 instruction. This is a special case because TEQ just sets the
CPSR and doesn't write to a GPR, which is not the case for EOR. So,
we need to ensure that the EOR can write to the first operand.
Differential Revision: https://reviews.llvm.org/D56255
llvm-svn: 350801
Generalize code in Thumb2InstrInfo::storeRegToStackSlot() and
loadRegToStackSlot() to allow the GPR class or any of its sub-classes
(including hGPR) to be stored/loaded by ARM::t2STRi12/ARM::t2LDRi12.
Differential Revision: https://reviews.llvm.org/D51927
llvm-svn: 346401
The SchedModel allows the addition of ReadAdvances to express that certain
operands of the instructions are needed at a later point than the others.
RegAlloc may add pseudo operands that are not part of the instruction
descriptor, and therefore cannot have any read advance entries. This meant
that in some cases the desired read advance was nullified by such a pseudo
operand, which still had the original latency.
This patch fixes this by making sure that such pseudo operands get a zero
latency during DAG construction.
Review: Matthias Braun, Ulrich Weigand.
https://reviews.llvm.org/D49671
llvm-svn: 345606
The "dead" markings allow existing target-independent optimizations,
like MachineSink, to trigger more frequently. The CPSR defs would have
eventually been marked dead by LiveVariables, so this only affects
optimizations before regalloc.
The ARMBaseInstrInfo.cpp change is fixing a bug which is only visible
with this change: the transform adds a use to an otherwise dead def
of CPSR. This is covered by existing regression tests.
thumb2-tbh.ll breaks for Thumb1 due to MachineLICM changing the
generated code; I'll fix it in D53452.
Differential Revision: https://reviews.llvm.org/D53453
llvm-svn: 345420
Clearing LargeOffsetGEPMap at the end fixes a bug where if a large
offset GEP is in a dead basic block, we fail an assertion when trying
to delete the block due to the asserting VH in LargeOffsetGEPMap.
Differential Revision: https://reviews.llvm.org/D53464
llvm-svn: 345082
A sequence of VMUL and VADD instructions always give the same or better
performance than a fused VMLA instruction on the Cortex-M4 and Cortex-M33.
Executing the VMUL and VADD back-to-back requires the same cycles, but
having separate instructions allows scheduling to avoid the hazard between
these 2 instructions.
Differential Revision: https://reviews.llvm.org/D52289
llvm-svn: 342874
There seems to be a separate command line tokenization issue that
prevents just ':\s*$' from working, since then the pattern argument
isn't quoted, and grep.exe misinterprets the backslash somehow.
llvm-svn: 342430
There is no way in the universe, that doing a full-width division in
software will be faster than doing overflowing multiplication in
software in the first place, especially given that this same full-width
multiplication needs to be done anyway.
This patch replaces the previous implementation with a direct lowering
into an overflowing multiplication algorithm based on half-width
operations.
Correctness of the algorithm was verified by exhaustively checking the
output of this algorithm for overflowing multiplication of 16 bit
integers against an obviously correct widening multiplication. Baring
any oversights introduced by porting the algorithm to DAG, confidence in
correctness of this algorithm is extremely high.
Following table shows the change in both t = runtime and s = space. The
change is expressed as a multiplier of original, so anything under 1 is
“better” and anything above 1 is worse.
+-------+-----------+-----------+-------------+-------------+
| Arch | u64*u64 t | u64*u64 s | u128*u128 t | u128*u128 s |
+-------+-----------+-----------+-------------+-------------+
| X64 | - | - | ~0.5 | ~0.64 |
| i686 | ~0.5 | ~0.6666 | ~0.05 | ~0.9 |
| armv7 | - | ~0.75 | - | ~1.4 |
+-------+-----------+-----------+-------------+-------------+
Performance numbers have been collected by running overflowing
multiplication in a loop under `perf` on two x86_64 (one Intel Haswell,
other AMD Ryzen) based machines. Size numbers have been collected by
looking at the size of function containing an overflowing multiply in
a loop.
All in all, it can be seen that both performance and size has improved
except in the case of armv7 where code size has regressed for 128-bit
multiply. u128*u128 overflowing multiply on 32-bit platforms seem to
benefit from this change a lot, taking only 5% of the time compared to
original algorithm to calculate the same thing.
The final benefit of this change is that LLVM is now capable of lowering
the overflowing unsigned multiply for integers of any bit-width as long
as the target is capable of lowering regular multiplication for the same
bit-width. Previously, 128-bit overflowing multiply was the widest
possible.
Patch by Simonas Kazlauskas!
Differential Revision: https://reviews.llvm.org/D50310
llvm-svn: 339922
The original code attempted to do this, but the std::abs() call didn't
actually do anything due to implicit type conversions. Fix the type
conversions, and perform the correct check for negative immediates.
This probably has very little practical impact, but it's worth fixing
just to avoid confusion in the future, I think.
Differential Revision: https://reviews.llvm.org/D48907
llvm-svn: 336742
Summary:
Two utils methods have essentially the same functionality. This is an attempt to merge them into one.
1. lib/Transforms/Utils/Local.cpp : MergeBasicBlockIntoOnlyPred
2. lib/Transforms/Utils/BasicBlockUtils.cpp : MergeBlockIntoPredecessor
Prior to the patch:
1. MergeBasicBlockIntoOnlyPred
Updates either DomTree or DeferredDominance
Moves all instructions from Pred to BB, deletes Pred
Asserts BB has single predecessor
If address was taken, replace the block address with constant 1 (?)
2. MergeBlockIntoPredecessor
Updates DomTree, LoopInfo and MemoryDependenceResults
Moves all instruction from BB to Pred, deletes BB
Returns if doesn't have a single predecessor
Returns if BB's address was taken
After the patch:
Method 2. MergeBlockIntoPredecessor is attempting to become the new default:
Updates DomTree or DeferredDominance, and LoopInfo and MemoryDependenceResults
Moves all instruction from BB to Pred, deletes BB
Returns if doesn't have a single predecessor
Returns if BB's address was taken
Uses of MergeBasicBlockIntoOnlyPred that need to be replaced:
1. lib/Transforms/Scalar/LoopSimplifyCFG.cpp
Updated in this patch. No challenges.
2. lib/CodeGen/CodeGenPrepare.cpp
Updated in this patch.
i. eliminateFallThrough is straightforward, but I added using a temporary array to avoid the iterator invalidation.
ii. eliminateMostlyEmptyBlock(s) methods also now use a temporary array for blocks
Some interesting aspects:
- Since Pred is not deleted (BB is), the entry block does not need updating.
- The entry block was being updated with the deleted block in eliminateMostlyEmptyBlock. Added assert to make obvious that BB=SinglePred.
- isMergingEmptyBlockProfitable assumes BB is the one to be deleted.
- eliminateMostlyEmptyBlock(BB) does not delete BB on one path, it deletes its unique predecessor instead.
- adding some test owner as subscribers for the interesting tests modified:
test/CodeGen/X86/avx-cmp.ll
test/CodeGen/AMDGPU/nested-loop-conditions.ll
test/CodeGen/AMDGPU/si-annotate-cf.ll
test/CodeGen/X86/hoist-spill.ll
test/CodeGen/X86/2006-11-17-IllegalMove.ll
3. lib/Transforms/Scalar/JumpThreading.cpp
Not covered in this patch. It is the only use case using the DeferredDominance.
I would defer to Brian Rzycki to make this replacement.
Reviewers: chandlerc, spatel, davide, brzycki, bkramer, javed.absar
Subscribers: qcolombet, sanjoy, nemanjai, nhaehnle, jlebar, tpr, kbarton, RKSimon, wmi, arsenm, llvm-commits
Differential Revision: https://reviews.llvm.org/D48202
llvm-svn: 335183
Thumb has more 16-bit encoding space dedicated to ADD than ORR, allowing both a
3-address encoding and a wider range of immediates. So, particularly when
optimizing for code size (but it doesn't make things worse elsewhere) it's
beneficial to select an OR operation to an ADD if we know overflow won't occur.
This is made even better by LLVM's penchant for putting operations in canonical
form by converting the other way.
llvm-svn: 335119
Shows some missed optimizations for the -7929856 and -2166 testcases.
-7929856 is due to a bug in ARMTargetLowering::getARMCmp, I think;
the -2166 case is a missing pattern.
llvm-svn: 335004
We need to clean up the DAG floating-point undef logic.
This process is similar to how we handled integer undef
logic in D43141.
And as we did there, I'm trying to reduce the patch by
changing tests that would probably become meaningless
once we correct FP undef folding.
llvm-svn: 332548
We need to clean up the DAG floating-point undef logic.
This process is similar to how we handled integer undef
logic in D43141.
And as we did there, I'm trying to reduce the patch by
changing tests that would probably become meaningless
once we correct FP undef folding.
llvm-svn: 332547
Summary:
This pass sinks COPY instructions into a successor block, if the COPY is not
used in the current block and the COPY is live-in to a single successor
(i.e., doesn't require the COPY to be duplicated). This avoids executing the
the copy on paths where their results aren't needed. This also exposes
additional opportunites for dead copy elimination and shrink wrapping.
These copies were either not handled by or are inserted after the MachineSink
pass. As an example of the former case, the MachineSink pass cannot sink
COPY instructions with allocatable source registers; for AArch64 these type
of copy instructions are frequently used to move function parameters (PhyReg)
into virtual registers in the entry block..
For the machine IR below, this pass will sink %w19 in the entry into its
successor (%bb.1) because %w19 is only live-in in %bb.1.
```
%bb.0:
%wzr = SUBSWri %w1, 1
%w19 = COPY %w0
Bcc 11, %bb.2
%bb.1:
Live Ins: %w19
BL @fun
%w0 = ADDWrr %w0, %w19
RET %w0
%bb.2:
%w0 = COPY %wzr
RET %w0
```
As we sink %w19 (CSR in AArch64) into %bb.1, the shrink-wrapping pass will be
able to see %bb.0 as a candidate.
With this change I observed 12% more shrink-wrapping candidate and 13% more dead copies deleted in spec2000/2006/2017 on AArch64.
Reviewers: qcolombet, MatzeB, thegameg, mcrosier, gberry, hfinkel, john.brawn, twoh, RKSimon, sebpop, kparzysz
Reviewed By: sebpop
Subscribers: evandro, sebpop, sfertile, aemerson, mgorny, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D41463
llvm-svn: 328237
Enable multiple COPY hints to eliminate more COPYs during register allocation.
Note that this is something all targets should do, see
https://reviews.llvm.org/D38128.
Review: Eli Friedman
llvm-svn: 325327
This patch combines some cases of ARMISD::CMOV for integers that arise in comparisons of the form
a != b ? x : 0
a == b ? 0 : x
and that currently (e.g. in Thumb1) are emitted as branches.
Differential Revision: https://reviews.llvm.org/D34515
llvm-svn: 325323
Discussed here:
http://lists.llvm.org/pipermail/llvm-dev/2018-January/120320.html
In preparation for adding support for named vregs we are changing the sigil for
physical registers in MIR to '$' from '%'. This will prevent name clashes of
named physical register with named vregs.
llvm-svn: 323922
In -debug output we print "pred:" whenever a MachineOperand is a
predicate operand in the instruction descriptor, and "opt:" whenever a
MachineOperand is an optional def in the instruction descriptor.
Differential Revision: https://reviews.llvm.org/D41870
llvm-svn: 322096
As part of the unification of the debug format and the MIR format, print
MBB references as '%bb.5'.
The MIR printer prints the IR name of a MBB only for block definitions.
* find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g'
* find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g'
* find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g'
* grep -nr 'BB#' and fix
Differential Revision: https://reviews.llvm.org/D40422
llvm-svn: 319665
output
As part of the unification of the debug format and the MIR format,
always use `printReg` to print all kinds of registers.
Updated the tests using '_' instead of '%noreg' until we decide which
one we want to be the default one.
Differential Revision: https://reviews.llvm.org/D40421
llvm-svn: 319445
Summary:
This fixes PR35221.
Use pseudo-instructions to let MachineCSE hoist global address computation.
Subscribers: aemerson, javed.absar, kristof.beyls, llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D39871
llvm-svn: 318081
As far as I can tell, this matches gcc: -mfloat-abi determines the
calling convention for all functions except those explicitly defined as
soft-float in the ARM RTABI.
This change only affects cases where the user specifies -mfloat-abi to
override the default calling convention derived from the target triple.
Fixes https://bugs.llvm.org//show_bug.cgi?id=34530.
Differential Revision: https://reviews.llvm.org/D38299
llvm-svn: 316708
Swap the compare operands if the lhs is a shift and the rhs isn't,
as in arm and T2 the shift can be performed by the compare for its
second operand.
Differential Revision: https://reviews.llvm.org/D39004
llvm-svn: 316562
The ARM backend should call setBooleanContents so that it can
use known bits to make some optimizations.
Review: D35821
Patch by Joel Galenson <jgalenson@google.com>
llvm-svn: 311446
Summary:
This patch teaches PostDominatorTree about infinite loops. It is built on top of D29705 by @dberlin which includes a very detailed motivation for this change.
What's new is that the patch also teaches the incremental updater how to deal with reverse-unreachable regions and how to properly maintain and verify tree roots. Before that, the incremental algorithm sometimes ended up preserving reverse-unreachable regions after updates that wouldn't appear in the tree if it was constructed from scratch on the same CFG.
This patch makes the following assumptions:
- A sequence of updates should produce the same tree as a recalculating it.
- Any sequence of the same updates should lead to the same tree.
- Siblings and roots are unordered.
The last two properties are essential to efficiently perform batch updates in the future.
When it comes to the first one, we can decide later that the consistency between freshly built tree and an updated one doesn't matter match, as there are many correct ways to pick roots in infinite loops, and to relax this assumption. That should enable us to recalculate postdominators less frequently.
This patch is pretty conservative when it comes to incremental updates on reverse-unreachable regions and ends up recalculating the whole tree in many cases. It should be possible to improve the performance in many cases, if we decide that it's important enough.
That being said, my experiments showed that reverse-unreachable are very rare in the IR emitted by clang when bootstrapping clang. Here are the statistics I collected by analyzing IR between passes and after each removePredecessor call:
```
# functions: 52283
# samples: 337609
# reverse unreachable BBs: 216022
# BBs: 247840796
Percent reverse-unreachable: 0.08716159869015269 %
Max(PercRevUnreachable) in a function: 87.58620689655172 %
# > 25 % samples: 471 ( 0.1395104988314885 % samples )
... in 145 ( 0.27733680163724345 % functions )
```
Most of the reverse-unreachable regions come from invalid IR where it wouldn't be possible to construct a PostDomTree anyway.
I would like to commit this patch in the next week in order to be able to complete the work that depends on it before the end of my internship, so please don't wait long to voice your concerns :).
Reviewers: dberlin, sanjoy, grosser, brzycki, davide, chandlerc, hfinkel
Reviewed By: dberlin
Subscribers: nhaehnle, javed.absar, kparzysz, uabelho, jlebar, hiraditya, llvm-commits, dberlin, david2050
Differential Revision: https://reviews.llvm.org/D35851
llvm-svn: 310940
`llc -march` is problematic because it only switches the target
architecture, but leaves the operating system unchanged. This
occasionally leads to indeterministic tests because the OS from
LLVM_DEFAULT_TARGET_TRIPLE is used.
However we can simply always use `llc -mtriple` instead. This changes
all the tests to do this to avoid people using -march when they copy and
paste parts of tests.
See also the discussion in https://reviews.llvm.org/D35287
llvm-svn: 309755
When we have a diamond ifcvt the fallthough block will have a branch at the end
of it that disappears when predicated, so discount it from the predication cost.
Differential Revision: https://reviews.llvm.org/D34952
llvm-svn: 307788
The current heuristic in isProfitableToIfCvt assumes we have a branch predictor,
and so gives the wrong answer in some cases when we don't. This patch adds a
subtarget feature to indicate that a subtarget has no branch predictor, and
changes the heuristic in isProfitableToiIfCvt when it's present. This gives a
slight overall improvement in a set of embedded benchmarks on Cortex-M4 and
Cortex-M33.
Differential Revision: https://reviews.llvm.org/D34398
llvm-svn: 306547
The benchmarking summarized in
http://lists.llvm.org/pipermail/llvm-dev/2017-May/113525.html showed
this is beneficial for a wide range of cores.
As is to be expected, quite a few small adaptations are needed to the
regressions tests, as the difference in scheduling results in:
- Quite a few small instruction schedule differences.
- A few changes in register allocation decisions caused by different
instruction schedules.
- A few changes in IfConversion decisions, due to a difference in
instruction schedule and/or the estimated cost of a branch mispredict.
llvm-svn: 306514
This has been deprecated since ARMARM v7-AR, release C.b, published back
in 2012.
This also removes test/CodeGen/Thumb2/ifcvt-neon.ll that originally was
introduced to check that conditionalization of Neon instructions did
happen when generating Thumb2. However, the test had evolved and was no
longer testing that. Rather than trying to adapt that test, this commit
introduces test/CodeGen/Thumb2/ifcvt-neon-deprecated.mir, since we can
now use the MIR framework to write nicer/more maintainable tests.
llvm-svn: 305998
Summary:
In SelectionDAG, when a store is immediately chained to another store
to the same address, elide the first store as it has no observable
effects. This is causes small improvements dealing with intrinsics
lowered to stores.
Test notes:
* Many testcases overwrite store addresses multiple times and needed
minor changes, mainly making stores volatile to prevent the
optimization from optimizing the test away.
* Many X86 test cases optimized out instructions associated with
associated with va_start.
* Note that test_splat in CodeGen/AArch64/misched-stp.ll no longer has
dependencies to check and can probably be removed and potentially
replaced with another test.
Reviewers: rnk, john.brawn
Subscribers: aemerson, rengolin, qcolombet, jyknight, nemanjai, nhaehnle, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D33206
llvm-svn: 303198
Shrink-wrapping uses post-dominators to find a restore point that
post-dominates all the uses of CSR / stack.
The way dominator trees are modeled in LLVM today is that unreachable
blocks are not present in a generic dominator tree, so, an unreachable node is
dominated by anything: include/llvm/Support/GenericDomTree.h:467.
Since for post-dominators, a no-return block is considered
"unreachable", calling findNearestCommonDominator on an unreachable node
A and a non-unreachable node B, will return B, which can be false. If we
find such node, we bail out since there is no good restore point
available.
rdar://problem/30186931
llvm-svn: 303130
During the optimisation of jump tables in the constant island pass,
an extra ADD could be left over, now dead but not removed.
Differential Revision: https://reviews.llvm.org/D31389
llvm-svn: 299634
I previously removed the T2XtPk feature from the ARM backend, but it
looks like I missed some of the tests that were using the feature.
Differential Revision: https://reviews.llvm.org/D30778
llvm-svn: 297386
Removed the HasT2ExtractPack feature and replaced its references
with HasDSP. This then allows the Thumb2 extend instructions to be
selected for ARMv8M +dsp. These instruction descriptions have also
been refactored and more target tests have been added for their isel.
Differential Revision: https://reviews.llvm.org/D29623
llvm-svn: 295452
When generating a floating point comparison we currently unconditionally
generate VCMPE. This has the sideeffect of setting the cumulative Invalid
bit in FPSCR if any of the operands are QNaN.
It is expected that use of a relational predicate on a QNaN value should
raise Invalid. Quoting from the C standard:
The relational and equality operators support the usual mathematical
relationships between numeric values. For any ordered pair of numeric
values exactly one of relationships the less, greater, equal and is true.
Relational operators may raise the floating-point exception when argument
values are NaNs.
The standard doesn't explicitly state the expectation for equality operators,
but the implication and obvious expectation is that equality operators
should not raise Invalid on a QNaN input, as those predicates are wholly
defined on unordered inputs (to return not equal).
Therefore, add a new operand to ARMISD::FPCMP and FPCMPZ indicating if
QNaN should raise Invalid, and pipe that through to TableGen.
llvm-svn: 294945
When choosing the best successor for a block, ordinarily we would have preferred
a block that preserves the CFG unless there is a strong probability the other
direction. For small blocks that can be duplicated we now skip that requirement
as well, subject to some simple frequency calculations.
Differential Revision: https://reviews.llvm.org/D28583
llvm-svn: 293716
The Requires class overrides the target requirements of an instruction,
rather than adding to them, so all ARM instructions need to include the
IsARM predicate when they have overwitten requirements.
This caused the swp and swpb instructions to be allowed in thumb mode
assembly, and the ARM encoding of CDP to be selected in codegen (which
is different for conditional instructions).
Differential Revision: https://reviews.llvm.org/D29283
llvm-svn: 293634
GCC changes the CC between the user-code and the builtins based on the
value of `-target` rather than `-mfloat-abi`. When a HF target is used,
the VFP variant of the AAPCS CC is used. Otherwise, the AAPCS variant
is used. In all cases, the AEABI functions use the AAPCS CC. Adjust
the calling convention based on the target.
Resolves PR30543!
llvm-svn: 291909
This reverts commit ada6595a526d71df04988eb0a4b4fe84df398ded.
This needs a simple probability check because there are some cases where it is
not profitable.
llvm-svn: 291695
When choosing the best successor for a block, ordinarily we would have preferred
a block that preserves the CFG unless there is a strong probability the other
direction. For small blocks that can be duplicated we now skip that requirement
as well.
Differential revision: https://reviews.llvm.org/D27742
llvm-svn: 291609
1.Fix pessimized case in FIXME.
2.Add tests for it.
3.The canonicalisation on shifts results in different sequence for
tests of machine-licm.Correct some check lines.
Differential Revision: https://reviews.llvm.org/D27916
llvm-svn: 290410
This is essentially a recommit of r285893, but with a correctness fix. The
problem of the original commit was that this:
bic r5, r7, #31
cbz r5, .LBB2_10
got rewritten into:
lsrs r5, r7, #5
beq .LBB2_10
The result in destination register r5 is not the same and this is incorrect
when r5 is not dead. So this fix includes checking the uses of the AND
destination register. And also, compared to the original commit, some regression
tests didn't need changing anymore because of this extra check.
For completeness, this was the original commit message:
For the common pattern (CMPZ (AND x, #bitmask), #0), we can do some more
efficient instruction selection if the bitmask is one consecutive sequence of
set bits (32 - clz(bm) - ctz(bm) == popcount(bm)).
1) If the bitmask touches the LSB, then we can remove all the upper bits and
set the flags by doing one LSLS.
2) If the bitmask touches the MSB, then we can remove all the lower bits and
set the flags with one LSRS.
3) If the bitmask has popcount == 1 (only one set bit), we can shift that bit
into the sign bit with one LSLS and change the condition query from NE/EQ to
MI/PL (we could also implement this by shifting into the carry bit and
branching on BCC/BCS).
4) Otherwise, we can emit a sequence of LSLS+LSRS to remove the upper and lower
zero bits of the mask.
1-3 require only one 16-bit instruction and can elide the CMP. 4 requires two
16-bit instructions but can elide the CMP and doesn't require materializing a
complex immediate, so is also a win.
Differential Revision: https://reviews.llvm.org/D27761
llvm-svn: 289794
This recommits r281323, which was backed out for two reasons. One, a selfhost failure, and two, it apparently caused Chromium failures. Actually, the latter was a red herring. The log has expired from the former, but I suspect that was a red herring too (actually caused by another problematic patch of mine). Therefore reapplying, and will watch the bots like a hawk.
For the common pattern (CMPZ (AND x, #bitmask), #0), we can do some more efficient instruction selection if the bitmask is one consecutive sequence of set bits (32 - clz(bm) - ctz(bm) == popcount(bm)).
1) If the bitmask touches the LSB, then we can remove all the upper bits and set the flags by doing one LSLS.
2) If the bitmask touches the MSB, then we can remove all the lower bits and set the flags with one LSRS.
3) If the bitmask has popcount == 1 (only one set bit), we can shift that bit into the sign bit with one LSLS and change the condition query from NE/EQ to MI/PL (we could also implement this by shifting into the carry bit and branching on BCC/BCS).
4) Otherwise, we can emit a sequence of LSLS+LSRS to remove the upper and lower zero bits of the mask.
1-3 require only one 16-bit instruction and can elide the CMP. 4 requires two 16-bit instructions but can elide the CMP and doesn't require materializing a complex immediate, so is also a win.
llvm-svn: 285893
[Reapplying r284580 and r285917 with fix and testing to ensure emitted jump tables for Thumb-1 have 4-byte alignment]
The TBB and TBH instructions in Thumb-2 allow jump tables to be compressed into sequences of bytes or shorts respectively. These instructions do not exist in Thumb-1, however it is possible to synthesize them out of a sequence of other instructions.
It turns out this sequence is so short that it's almost never a lose for performance and is ALWAYS a significant win for code size.
TBB example:
Before: lsls r0, r0, #2 After: add r0, pc
adr r1, .LJTI0_0 ldrb r0, [r0, #6]
ldr r0, [r0, r1] lsls r0, r0, #1
mov pc, r0 add pc, r0
=> No change in prologue code size or dynamic instruction count. Jump table shrunk by a factor of 4.
The only case that can increase dynamic instruction count is the TBH case:
Before: lsls r0, r4, #2 After: lsls r4, r4, #1
adr r1, .LJTI0_0 add r4, pc
ldr r0, [r0, r1] ldrh r4, [r4, #6]
mov pc, r0 lsls r4, r4, #1
add pc, r4
=> 1 more instruction in prologue. Jump table shrunk by a factor of 2.
So there is an argument that this should be disabled when optimizing for performance (and a TBH needs to be generated). I'm not so sure about that in practice, because on small cores with Thumb-1 performance is often tied to code size. But I'm willing to turn it off when optimizing for performance if people want (also note that TBHs are fairly rare in practice!)
llvm-svn: 285690
The TBB and TBH instructions in Thumb-2 allow jump tables to be compressed into sequences of bytes or shorts respectively. These instructions do not exist in Thumb-1, however it is possible to synthesize them out of a sequence of other instructions.
It turns out this sequence is so short that it's almost never a lose for performance and is ALWAYS a significant win for code size.
TBB example:
Before: lsls r0, r0, #2 After: add r0, pc
adr r1, .LJTI0_0 ldrb r0, [r0, #6]
ldr r0, [r0, r1] lsls r0, r0, #1
mov pc, r0 add pc, r0
=> No change in prologue code size or dynamic instruction count. Jump table shrunk by a factor of 4.
The only case that can increase dynamic instruction count is the TBH case:
Before: lsls r0, r4, #2 After: lsls r4, r4, #1
adr r1, .LJTI0_0 add r4, pc
ldr r0, [r0, r1] ldrh r4, [r4, #6]
mov pc, r0 lsls r4, r4, #1
add pc, r4
=> 1 more instruction in prologue. Jump table shrunk by a factor of 2.
So there is an argument that this should be disabled when optimizing for performance (and a TBH needs to be generated). I'm not so sure about that in practice, because on small cores with Thumb-1 performance is often tied to code size. But I'm willing to turn it off when optimizing for performance if people want (also note that TBHs are fairly rare in practice!)
llvm-svn: 284580
Reverts r283938 to reinstate r283867 with a fix.
The original change had an ArrayRef referring to a destroyed temporary
initializer list. Use plain C arrays instead.
llvm-svn: 283942
The high registers are not allocatable in Thumb1 functions, but they
could still be used by inline assembly, so we need to save and restore
the callee-saved high registers (r8-r11) in the prologue and epilogue.
This is complicated by the fact that the Thumb1 push and pop
instructions cannot access these registers. Therefore, we have to move
them down into low registers before pushing, and move them back after
popping into low registers.
In most functions, we will have low registers that are also being
pushed/popped, which we can use as the temporary registers for
saving/restoring the high registers. However, this is not guaranteed, so
we may need to push some extra low registers to ensure that the high
registers can be saved/restored. For correctness, it would be sufficient
to use just one low register, but if we have enough low registers
available then we only need one push/pop instruction, rather than one
per high register.
We can also use the argument/return registers when they are not live,
and the link register when saving (but not restoring), reducing the
number of extra registers we need to push.
There are still a few extreme edge cases where we need two push/pop
instructions, because not enough low registers can be made live in the
prologue or epilogue.
In addition to the regression tests included here, I've also tested this
using a script to generate functions which clobber different
combinations of registers, have different numbers of argument and return
registers (including variadic arguments), allocate different fixed sized
objects on the stack, and do or don't use variable sized allocas and the
__builtin_return_address intrinsic (all of which affect the available
registers in the prologue and epilogue). I ran these functions in a test
harness which verifies that all of the callee-saved registers are
correctly preserved.
Differential Revision: https://reviews.llvm.org/D24228
llvm-svn: 283867
For the common pattern (CMPZ (AND x, #bitmask), #0), we can do some more efficient instruction selection if the bitmask is one consecutive sequence of set bits (32 - clz(bm) - ctz(bm) == popcount(bm)).
1) If the bitmask touches the LSB, then we can remove all the upper bits and set the flags by doing one LSLS.
2) If the bitmask touches the MSB, then we can remove all the lower bits and set the flags with one LSRS.
3) If the bitmask has popcount == 1 (only one set bit), we can shift that bit into the sign bit with one LSLS and change the condition query from NE/EQ to MI/PL (we could also implement this by shifting into the carry bit and branching on BCC/BCS).
4) Otherwise, we can emit a sequence of LSLS+LSRS to remove the upper and lower zero bits of the mask.
1-3 require only one 16-bit instruction and can elide the CMP. 4 requires two 16-bit instructions but can elide the CMP and doesn't require materializing a complex immediate, so is also a win.
llvm-svn: 281323
For the common pattern (CMPZ (AND x, #bitmask), #0), we can do some more efficient instruction selection if the bitmask is one consecutive sequence of set bits (32 - clz(bm) - ctz(bm) == popcount(bm)).
1) If the bitmask touches the LSB, then we can remove all the upper bits and set the flags by doing one LSLS.
2) If the bitmask touches the MSB, then we can remove all the lower bits and set the flags with one LSRS.
3) If the bitmask has popcount == 1 (only one set bit), we can shift that bit into the sign bit with one LSLS and change the condition query from NE/EQ to MI/PL (we could also implement this by shifting into the carry bit and branching on BCC/BCS).
4) Otherwise, we can emit a sequence of LSLS+LSRS to remove the upper and lower zero bits of the mask.
1-3 require only one 16-bit instruction and can elide the CMP. 4 requires two 16-bit instructions but can elide the CMP and doesn't require materializing a complex immediate, so is also a win.
llvm-svn: 281215
The CMPZ #0 disappears during peepholing, leaving just a tADDi3, tADDi8 or t2ADDri. This avoids having to materialize the expensive negative constant in Thumb-1, and allows a shrinking from a 32-bit CMN to a 16-bit ADDS in Thumb-2.
llvm-svn: 281040
The original commit was too aggressive about marking LibCalls as AAPCS. The
libcalls contain libc/libm/libunwind calls which are not AAPCS, but C.
llvm-svn: 280833
All of the builtins are designed to be invoked with ARM AAPCS CC even on ARM
AAPCS VFP CC hosts. Tweak the default initialisation to ARM AAPCS CC rather
than C CC for ARM/thumb targets.
The changes to the tests are necessary to ensure that the calling convention for
the lowered library calls are honoured. Furthermore, these adjustments cause
certain branch invocations to change to branch-and-link since the returned value
needs to be moved across registers (d0 -> r0, r1).
llvm-svn: 280683
The cost of predicating a diamond is only the instructions that are not shared
between the two branches. Additionally If a predicate clobbering instruction
occurs in the shared portion of the branches (e.g. a cond move), it may still
be possible to if convert the sub-cfg. This change handles these two facts by
rescanning the non-shared portion of a diamond sub-cfg to recalculate both the
predication cost and whether both blocks are pred-clobbering.
Fixed 2 bugs before recommitting. Branch instructions must be compared and found
identical before diamond conversion. Also, predicate-clobbering instructions in
the shared prefix disqualifies a potential diamond conversion. Includes tests
for both.
llvm-svn: 279670
There is not an official documented ABI for frame pointers in Thumb2,
but we should try to emit something which is useful.
We use r7 as the frame pointer for Thumb code, which currently means
that if a function needs to save a high register (r8-r11), it will get
pushed to the stack between the frame pointer (r7) and link register
(r14). This means that while a stack unwinder can follow the chain of
frame pointers up the stack, it cannot know the offset to lr, so does
not know which functions correspond to the stack frames.
To fix this, we need to push the callee-saved registers in two batches,
with the first push saving the low registers, fp and lr, and the second
push saving the high registers. This is already implemented, but
previously only used for iOS. This patch turns it on for all Thumb2
targets when frame pointers are required by the ABI, and the frame
pointer is r7 (Windows uses r11, so this isn't a problem there). If
frame pointer elimination is enabled we still emit a single push/pop
even if we need a frame pointer for other reasons, to avoid increasing
code size.
We must also ensure that lr is pushed to the stack when using a frame
pointer, so that we end up with a complete frame record. Situations that
could cause this were rare, because we already push lr in most
situations so that we can return using the pop instruction.
Differential Revision: https://reviews.llvm.org/D23516
llvm-svn: 279506
The following function currently relies on tail-merging for if
conversion to succeed. The common tail of cond_true and cond_false is
extracted, and this then forms a diamond pattern that can be
successfully if converted.
If this block does not get extracted, either because tail-merging is
disabled or the threshold is higher, we should still recognize this
pattern and if-convert it.
Fixed a regression in the original commit. Need to un-reverse branches after
reversing them, or other conversions go awry.
Regression on self-hosting bots with no obvious explanation. Tidied up range
handling to be more obviously correct, but there was no smoking gun.
define i32 @t2(i32 %a, i32 %b) nounwind {
entry:
%tmp1434 = icmp eq i32 %a, %b ; <i1> [#uses=1]
br i1 %tmp1434, label %bb17, label %bb.outer
bb.outer: ; preds = %cond_false, %entry
%b_addr.021.0.ph = phi i32 [ %b, %entry ], [ %tmp10, %cond_false ]
%a_addr.026.0.ph = phi i32 [ %a, %entry ], [ %a_addr.026.0, %cond_false ]
br label %bb
bb: ; preds = %cond_true, %bb.outer
%indvar = phi i32 [ 0, %bb.outer ], [ %indvar.next, %cond_true ]
%tmp. = sub i32 0, %b_addr.021.0.ph
%tmp.40 = mul i32 %indvar, %tmp.
%a_addr.026.0 = add i32 %tmp.40, %a_addr.026.0.ph
%tmp3 = icmp sgt i32 %a_addr.026.0, %b_addr.021.0.ph
br i1 %tmp3, label %cond_true, label %cond_false
cond_true: ; preds = %bb
%tmp7 = sub i32 %a_addr.026.0, %b_addr.021.0.ph
%tmp1437 = icmp eq i32 %tmp7, %b_addr.021.0.ph
%indvar.next = add i32 %indvar, 1
br i1 %tmp1437, label %bb17, label %bb
cond_false: ; preds = %bb
%tmp10 = sub i32 %b_addr.021.0.ph, %a_addr.026.0
%tmp14 = icmp eq i32 %a_addr.026.0, %tmp10
br i1 %tmp14, label %bb17, label %bb.outer
bb17: ; preds = %cond_false, %cond_true, %entry
%a_addr.026.1 = phi i32 [ %a, %entry ], [ %tmp7, %cond_true ], [ %a_addr.026.0, %cond_false ]
ret i32 %a_addr.026.1
}
Without tail-merging or diamond-tail if conversion:
LBB1_1: @ %bb
@ =>This Inner Loop Header: Depth=1
cmp r0, r1
ble LBB1_3
@ BB#2: @ %cond_true
@ in Loop: Header=BB1_1 Depth=1
subs r0, r0, r1
cmp r1, r0
it ne
cmpne r0, r1
bgt LBB1_4
LBB1_3: @ %cond_false
@ in Loop: Header=BB1_1 Depth=1
subs r1, r1, r0
cmp r1, r0
bne LBB1_1
LBB1_4: @ %bb17
bx lr
With diamond-tail if conversion, but without tail-merging:
@ BB#0: @ %entry
cmp r0, r1
it eq
bxeq lr
LBB1_1: @ %bb
@ =>This Inner Loop Header: Depth=1
cmp r0, r1
ite le
suble r1, r1, r0
subgt r0, r0, r1
cmp r1, r0
bne LBB1_1
@ BB#2: @ %bb17
bx lr
llvm-svn: 279168
Created a Thumb2 predicated pattern matcher that uses Thumb2 and
HasT2ExtractPack and used it to redefine the patterns for sxta{b|h}
and uxta{b|h}. Also used the similar patterns to fill in isel pattern
gaps for the corresponding instructions in the ARM backend.
The patch is mainly changes to tests since most of this functionality
appears not to have been tested.
Differential Revision: https://reviews.llvm.org/D23273
llvm-svn: 278207
The important thing I was missing was ensuring newly added constants were kept in topological order. Repositioning the node is correct if the constant is newly added (so it has no topological ordering) but wrong if it already existed - positioning it next in the worklist would break the topological ordering.
Original commit message:
[Thumb] Select a BIC instead of AND if the immediate can be encoded more optimally negated
If an immediate is only used in an AND node, it is possible that the immediate can be more optimally materialized when negated. If this is the case, we can negate the immediate and use a BIC instead;
int i(int a) {
return a & 0xfffffeec;
}
Used to produce:
ldr r1, [CONSTPOOL]
ands r0, r1
CONSTPOOL: 0xfffffeec
And now produces:
movs r1, #255
adds r1, #20 ; Less costly immediate generation
bics r0, r1
llvm-svn: 274543
We were using DAG->getConstant instead of DAG->getTargetConstant. This meant that we could inadvertently increase the use count of a constant if stars aligned, which it did in this testcase. Increasing the use count of the constant could cause ISel to fall over (because DAGToDAG lowering assumed the constant had only one use!)
Original commit message:
[Thumb] Select a BIC instead of AND if the immediate can be encoded more optimally negated
If an immediate is only used in an AND node, it is possible that the immediate can be more optimally materialized when negated. If this is the case, we can negate the immediate and use a BIC instead;
int i(int a) {
return a & 0xfffffeec;
}
Used to produce:
ldr r1, [CONSTPOOL]
ands r0, r1
CONSTPOOL: 0xfffffeec
And now produces:
movs r1, #255
adds r1, #20 ; Less costly immediate generation
bics r0, r1
llvm-svn: 274510
Tail merge was making the assumption that a layout successor or
predecessor was always a cfg successor/predecessor. Remove that
assumption. Changes to tests are necessary because the errant cfg edges
were preventing optimizations.
llvm-svn: 273700
The R_ARM_PLT32 relocation is deprecated and is not produced by MC.
This means that the code being deleted is dead from the .o point of
view and was making the .s more confusing.
llvm-svn: 272909
I'm really not sure why we were in the first place, it's the linker's job to
convert between BL/BLX as necessary. Even worse, using BLX left Thumb calls
that could be locally resolved completely unencodable since all offsets to BLX
are multiples of 4.
rdar://26182344
llvm-svn: 269101
Summary:
While setting kill flags on instructions inside a BUNDLE, we bail out as soon
as we set kill flag on a register. But we are missing a check when all the
registers already have the correct kill flag set. We need to bail out in that
case as well.
This patch refactors the old code and simply makes use of the addRegisterKilled
function in MachineInstr.cpp in order to determine whether to set/remove kill
on an instruction.
Reviewers: apazos, t.p.northover, pete, MatzeB
Subscribers: MatzeB, davide, llvm-commits
Differential Revision: http://reviews.llvm.org/D17356
llvm-svn: 269092
This is better for a few reasons:
+ It matches the other tooling for iOS.
+ It matches EABI in more cases (i.e. Thumb-mode, and in practice we don't
use ARM mode).
+ It leads to infinitesimally smaller code (0.2%, yay!).
rdar://25369506
llvm-svn: 266003
Minimum density for both optsize and non optsize are now options
-sparse-jump-table-density (default 10) for non optsize functions
-dense-jump-table-density (default 40) for optsize functions, which
matches the current default. This improves several benchmarks at google
at the cost of a small codesize increase. For code compiled with -Os,
the old behavior continues
llvm-svn: 264689
Most of the time ARM has the CCR.UNALIGN_TRP bit set to false which
means that unaligned loads/stores do not trap and even extensive testing
will not catch these bugs. However the multi/double variants are not
affected by this bit and will still trap. In effect a more aggressive
load/store optimization will break existing (bad) code.
These bugs do not necessarily manifest in the broken code where the
misaligned pointer is formed but often later in perfectly legal code
where it is accessed. This means recompiling system libraries (which
have no alignment bugs) with a newer compiler will break existing
applications (with alignment bugs) that worked before.
So (under protest) I implemented this safe mode which limits the
formation of multi/double operations to cases that are not affected by
user code (stack operations like spills/reloads) or cases where the
normal operations trap anyway (floating point load/stores). It is
disabled by default.
Differential Revision: http://reviews.llvm.org/D17015
llvm-svn: 262504
Current SCEV expansion will expand SCEV as a sequence of operations
and doesn't utilize the value already existed. This will introduce
redundent computation which may not be cleaned up throughly by
following optimizations.
This patch introduces an ExprValueMap which is a map from SCEV to the
set of equal values with the same SCEV. When a SCEV is expanded, the
set of values is checked and reused whenever possible before generating
a sequence of operations.
The original commit triggered regressions in Polly tests. The regressions
exposed two problems which have been fixed in current version.
1. Polly will generate a new function based on the old one. To generate an
instruction for the new function, it builds SCEV for the old instruction,
applies some tranformation on the SCEV generated, then expands the transformed
SCEV and insert the expanded value into new function. Because SCEV expansion
may reuse value cached in ExprValueMap, the value in old function may be
inserted into new function, which is wrong.
In SCEVExpander::expand, there is a logic to check the cached value to
be used should dominate the insertion point. However, for the above
case, the check always passes. That is because the insertion point is
in a new function, which is unreachable from the old function. However
for unreachable node, DominatorTreeBase::dominates thinks it will be
dominated by any other node.
The fix is to simply add a check that the cached value to be used in
expansion should be in the same function as the insertion point instruction.
2. When the SCEV is of scConstant type, expanding it directly is cheaper than
reusing a normal value cached. Although in the cached value set in ExprValueMap,
there is a Constant type value, but it is not easy to find it out -- the cached
Value set is not sorted according to the potential cost. Existing reuse logic
in SCEVExpander::expand simply chooses the first legal element from the cached
value set.
The fix is that when the SCEV is of scConstant type, don't try the reuse
logic. simply expand it.
Differential Revision: http://reviews.llvm.org/D12090
llvm-svn: 259736
Current SCEV expansion will expand SCEV as a sequence of operations
and doesn't utilize the value already existed. This will introduce
redundent computation which may not be cleaned up throughly by
following optimizations.
This patch introduces an ExprValueMap which is a map from SCEV to the
set of equal values with the same SCEV. When a SCEV is expanded, the
set of values is checked and reused whenever possible before generating
a sequence of operations.
Differential Revision: http://reviews.llvm.org/D12090
llvm-svn: 259662
For historic reasons, the behavior of .align differs between targets.
Fortunately, there are alternatives, .p2align and .balign, which make the
interpretation of the parameter explicit, and which behave consistently across
targets.
This patch teaches MC to use .p2align instead of .align, so that people reading
code for multiple architectures don't have to remember which way each platform
does its .align directive.
Differential Revision: http://reviews.llvm.org/D16549
llvm-svn: 258750
Several (but not all) of the labels that are checked for in this test case
are checked as strings instead of labels. This can cause an apparent test
case failure if they are tested in an appropriately named directory.
For example, one of them that fails:
define zeroext i32 @test2(i32 %A.u, i32 %B.u) {
; A8: test2
; A8: uxtab r0, r0, r1
Output that causes it to fail:
. . .
.file "/home/seurer/llvm/llvm-test2/test/CodeGen/Thumb2/thumb2-uxt_rot.ll"
. . .
.globl test2
.align 1
.type test2,%function
.code 16 @ @test2
.thumb_func
test2:
.fnstart
The "A8: test2" matches on the directory name instead of the label.
llvm-svn: 253702
Note, this was reviewed (and more details are in) http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html
These intrinsics currently have an explicit alignment argument which is
required to be a constant integer. It represents the alignment of the
source and dest, and so must be the minimum of those.
This change allows source and dest to each have their own alignments
by using the alignment attribute on their arguments. The alignment
argument itself is removed.
There are a few places in the code for which the code needs to be
checked by an expert as to whether using only src/dest alignment is
safe. For those places, they currently take the minimum of src/dest
alignments which matches the current behaviour.
For example, code which used to read:
call void @llvm.memcpy.p0i8.p0i8.i32(i8* %dest, i8* %src, i32 500, i32 8, i1 false)
will now read:
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 8 %dest, i8* align 8 %src, i32 500, i1 false)
For out of tree owners, I was able to strip alignment from calls using sed by replacing:
(call.*llvm\.memset.*)i32\ [0-9]*\,\ i1 false\)
with:
$1i1 false)
and similarly for memmove and memcpy.
I then added back in alignment to test cases which needed it.
A similar commit will be made to clang which actually has many differences in alignment as now
IRBuilder can generate different source/dest alignments on calls.
In IRBuilder itself, a new argument was added. Instead of calling:
CreateMemCpy(Dst, Src, getInt64(Size), DstAlign, /* isVolatile */ false)
you now call
CreateMemCpy(Dst, Src, getInt64(Size), DstAlign, SrcAlign, /* isVolatile */ false)
There is a temporary class (IntegerAlignment) which takes the source alignment and rejects
implicit conversion from bool. This is to prevent isVolatile here from passing its default
parameter to the source alignment.
Note, changes in future can now be made to codegen. I didn't change anything here, but this
change should enable better memcpy code sequences.
Reviewed by Hal Finkel.
llvm-svn: 253511
David Green