The bug has been lying dormant, but apparently was never exposed, until
after rL324941 because we didn't return the correct result
for shifts with undef operands.
llvm-svn: 325010
Here are the number of additional debug values salvaged in a stage2
build of clang:
63 SALVAGE: MUL
1250 SALVAGE: SDIV
(No values were salvaged from `srem` instructions in this experiment,
but it's a simple case to handle so we might as well.)
llvm-svn: 324976
Here are the number of additional debug values salvaged in a stage2
build of clang:
1912 SALVAGE: ASHR
405 SALVAGE: LSHR
249 SALVAGE: SHL
llvm-svn: 324975
Also make a drive-by-fix of a bug in the subregister scan code that
only triggers with an incomplete or otherwise very irregular machine
description.
rdar://problem/37404493
llvm-svn: 324972
Summary:
This change is part of step five in the series of changes to remove alignment argument from
memcpy/memmove/memset in favour of alignment attributes. In particular, this changes the
creation of memcpys in the SafeStack pass to set the alignment of the destination object to
its stack alignment while separately setting the source byval arguments alignment to its
alignment.
Steps:
Step 1) Remove alignment parameter and create alignment parameter attributes for
memcpy/memmove/memset. ( rL322965, rC322964, rL322963 )
Step 2) Expand the IRBuilder API to allow creation of memcpy/memmove with differing
source and dest alignments. ( rL323597 )
Step 3) Update Clang to use the new IRBuilder API. ( rC323617 )
Step 4) Update Polly to use the new IRBuilder API. ( rL323618 )
Step 5) Update LLVM passes that create memcpy/memmove calls to use the new IRBuilder API,
and those that use use MemIntrinsicInst::[get|set]Alignment() to use [get|set]DestAlignment()
and [get|set]SourceAlignment() instead. (rL323886, rL323891, rL324148, rL324273, rL324278,
rL324384, rL324395, rL324402, rL324626, rL324642, rL324653, rL324654, rL324773, rL324774,
rL324781, rL324784 )
Step 6) Remove the single-alignment IRBuilder API for memcpy/memmove, and the
MemIntrinsicInst::[get|set]Alignment() methods.
Reference
http://lists.llvm.org/pipermail/llvm-dev/2015-August/089384.htmlhttp://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html
Reviewers: eugenis, bollu
Reviewed By: eugenis
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D42710
llvm-svn: 324955
This started by noticing that scalar and vector types were producing different results with div ops in PR36305:
https://bugs.llvm.org/show_bug.cgi?id=36305
...but the problem is bigger. I couldn't keep it straight without a table, so I'm attaching that as a PDF to
the review. The x86 tests in undef-ops.ll correspond to that table.
Green means that instsimplify and the DAG agree on the result for all types.
Red means the DAG was returning undef when IR was not.
Yellow means the DAG was returning a non-undef result when IR returned undef.
This patch assumes that we're currently doing the right thing in IR.
Note: I couldn't find any problems with lowering vector constants as the code comments were warning,
but those comments were written long ago in rL36413 .
Differential Revision: https://reviews.llvm.org/D43141
llvm-svn: 324941
If merging them, the dllexport attribute needs to be brought along
to the new GlobalAlias.
Differential Revision: https://reviews.llvm.org/D43192
llvm-svn: 324937
Rather than encode the absence of a checksum with a Kind variant, instead put
both the kind and value in a struct and wrap it in an Optional.
Differential Revision: http://reviews.llvm.org/D43043
llvm-svn: 324928
Armv8.1-A added an atomic load-clear instruction (which performs bitwise
and with the complement of it's operand), but not a load-and
instruction. Our current code-generation for atomic load-and always
inserts an MVN instruction to invert its argument, even if it could be
folded into a constant or another instruction.
This adds lowering early in selection DAG to convert a load-and
operation into an xor with -1 and a load-clear, allowing the normal DAG
optimisations to work on it.
To do this, I've had to add a new ISD opcode, ATOMIC_LOAD_CLR. I don't
see any easy way to do this with an AArch64-specific ISD node, because
the code-generation for atomic operations assumes the SDNodes are of
type AtomicSDNode.
I've left the old tablegen patterns in because they are still needed for
global isel.
Differential revision: https://reviews.llvm.org/D42478
llvm-svn: 324908
Add a common -trap-unreachable option, similar to the target
specific hexagon equivalent, which has been replaced. This
turns unreachable instructions into traps, which is useful for
debugging.
Differential Revision: https://reviews.llvm.org/D42965
llvm-svn: 324880
Instead of reserving 0xF00 bytes for the fixed length portion of the CodeView
symbol name, calculate the actual length of the fixed length portion.
Differential Revision: https://reviews.llvm.org/D42125
llvm-svn: 324850
This reverses instcombine's demanded bits' transform which always tries to clear bits in constants.
As noted in PR35792 and shown in the test diffs:
https://bugs.llvm.org/show_bug.cgi?id=35792
...we can do better in codegen by trying to form -1. The x86 sub test shows a missed opportunity.
I did investigate changing instcombine's behavior, but it would be more work to change
canonicalization in IR. Clearing bits / shrinking constants can allow killing instructions,
so we'd have to figure out how to not regress those cases.
Differential Revision: https://reviews.llvm.org/D42986
llvm-svn: 324839
This allows us to recognise more saturation patterns and also simplify some MINMAX codegen that was failing to combine CMPGE comparisons to a legal CMPGT.
Differential Revision: https://reviews.llvm.org/D43014
llvm-svn: 324837
SelectionDAG::getBoolConstant was recently introduced. At the time I didn't know getConstTrueVal existed, but I think getBoolConstant is better as it will use the source VT to make sure it can properly detect floating point if it is configured differently.
llvm-svn: 324832
Extend salvageDebugInfo to preserve the debug info from a dead 'or'
with a constant.
Patch by Ismail Badawi!
Differential Revision: https://reviews.llvm.org/D43129
llvm-svn: 324764
* Use uleb128 for code offsets in the LSDA call site table.
* Omit the TTBase offset if the type table is empty.
This change can reduce the size of the DWARF/Itanium LSDA by about half.
Patch by Ryan Prichard!
llvm-svn: 324750
Rely on the assembler to finalize the layout of the DWARF/Itanium
exception-handling LSDA. Rather than calculate the exact size of each
thing in the LSDA, use assembler directives:
To emit the offset to the TTBase label:
.uleb128 .Lttbase0-.Lttbaseref0
.Lttbaseref0:
To emit the size of the call site table:
.uleb128 .Lcst_end0-.Lcst_begin0
.Lcst_begin0:
... call site table entries ...
.Lcst_end0:
To align the type info table:
... action table ...
.balign 4
.long _ZTIi
.long _ZTIl
.Lttbase0:
Using assembler directives simplifies the compiler and allows switching
the encoding of offsets in the call site table from udata4 to uleb128 for
a large code size savings. (This commit does not change the encoding.)
The combination of the uleb128 followed by a balign creates an unfortunate
dependency cycle that the assembler must sometimes resolve either by
padding an LEB or by inserting zero padding before the type table. See
PR35809 or GNU as bug 4029.
Patch by Ryan Prichard!
llvm-svn: 324749
r314974 introduced insertion of DEBUG_VALUEs after
each redefinition of debug value register in the slot index range.
In case the instruction redefining the debug value register
was a terminator, machine verifier would complain since it
enforces the rule of no non-terminator instructions
following the first terminator.
Differential Revision: https://reviews.llvm.org/D42801
llvm-svn: 324734
When adding operands to machine instructions in case of
RegisterSDNodes, generate a COPY node in case the register class
does not match the one in the instruction definition.
Differental Revision: https://reviews.llvm.org/D35561
llvm-svn: 324733
Summary:
The class contained arrays of two structures (DataArray and HashData).
These structures were in 1:1 correspondence, and one of them contained
pointers to the other (and *both* contained a "Name" field). By merging
these two structures into one, we can save a bit of space without
negatively impacting much of anything.
Reviewers: JDevlieghere, aprantl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43073
llvm-svn: 324724
Add verification for copies involving generic registers if they are
compatible - ie if it is a generic copy, then the types are the
same, and if a COPY b/w generic and target virtual register, then
the sizes should be the same. Only checks if there are no sub registers
involved for now.
https://reviews.llvm.org/D37775
llvm-svn: 324696
This addresses review feedback for D42940. The topological sort is
slightly more expensive but it can now also detect cycles in the
dependencies and actually works correctly.
rdar://problem/37217988
Differential Review: https://reviews.llvm.org/D43036
llvm-svn: 324677
Many in SimplifySetCC and FoldSetCC try to create true or false constants. Some of them query getBooleanContents to figure out whether to use all ones or just 1 for true. But many places do not check and just use 1 without ensuring the VT has an i1 scalar type. Note sure if those places only trigger before type legalization so they only see an i1
type?
To cleanup the inconsistency and reduce some duplicated code, this patch adds a getBoolConstant method to SelectionDAG that takes are of querying getBooleanContents and doing the right thing.
Differential Revision: https://reviews.llvm.org/D43037
llvm-svn: 324634
We're passing the binary op that uses the load instead of the load.
Noticed by inspection. Not sure how to test this because this just prevents the introduction of an extend that will later be truncated and will probably be combined out.
llvm-svn: 324568
The truncate is being used to replace other users of of the load, but we checked that the load only has one use so there are no other uses to replace.
llvm-svn: 324567
Instead of:
%bb.1: derived from LLVM BB %for.body
print:
bb.1.for.body:
Also use MIR syntax for MBB attributes like "align", "landing-pad", etc.
llvm-svn: 324563
The truncate is only needed if the load has additional users. It used to get passed to extendSetCCUses so was created early, but that's no longer the case.
llvm-svn: 324562
Travel all chains paths to first non-tokenfactor node can be
exponential work. Add simple redundency check to avoid this.
Fixes PR36264.
llvm-svn: 324491
This patch is the LLVM part of fixing the issues, described in
https://bugs.llvm.org/show_bug.cgi?id=36168
* The representation of enumerator values in the debug info metadata now
contains a boolean flag isUnsigned, which determines how the bits of
the value are interpreted.
* The DW_TAG_enumeration type DIE now always (for DWARF version >= 3)
includes a DW_AT_type attribute, which refers to the underlying
integer type, as suggested in DWARFv4 (5.7 Enumeration Type Entries).
* The debug info metadata for enumeration type contains (in flags)
indication whether this is a C++11 "fixed enum".
* For C++11 enumeration with a fixed underlying type, the DIE also
includes the DW_AT_enum_class attribute (for DWARF version >= 4).
* Encoding of enumerator constants uses DW_FORM_sdata for signed values
and DW_FORM_udata for unsigned values, as suggested by DWARFv4 (7.5.4
Attribute Encodings).
The changes should be backwards compatible:
* the isUnsigned attribute is optional and defaults to false.
* if the underlying type for the enumeration is not available, the
enumerator values are considered signed.
* the FixedEnum flag defaults to clear.
* the bitcode format for DIEnumerator stores the unsigned flag bit #1 of
the first record element, so the format does not change and the zero
previously stored there is consistent with the false default for
IsUnsigned.
Differential Revision: https://reviews.llvm.org/D42734
llvm-svn: 324489
With fixes from rL324341.
Original commit message:
[MergeICmps] Enable the MergeICmps Pass by default.
Summary: Now that PR33325 is fixed, this should always improve the generated code.
Reviewers: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D42793
llvm-svn: 324465
X86 currently has a late DAG combine after cttz/ctlz are turned into BSR+BSF+CMOV to detect this and remove the CMOV. But we should be able to do this much earlier and avoid creating the cmov all together.
For the changed AMDGPU test case it appears that previously the i8 cttz was type legalized to i16 which introduced an OR with 256 in order to limit the result to 8 on the widened type. At this point the result is known to never be zero, but nothing checked that. Then operation legalization is told to promote all i16 cttz to i32. This introduces an extend and a truncate and another OR with 65536 to limit the result to 16. With the DAG combiner change we are able to prevent the creation of the second OR since the opcode will have been changed to cttz_zero_undef after the first OR. I the lack of the OR caused the instruction to change to v_ffbl_b32_sdwa
Differential Revision: https://reviews.llvm.org/D42985
llvm-svn: 324427
n Rust, an enum that carries data in the variants is, essentially, a
discriminated union. Furthermore, the Rust compiler will perform
space optimizations on such enums in some situations. Previously,
DWARF for these constructs was emitted using a hack (a magic field
name); but this approach stopped working when more space optimizations
were added in https://github.com/rust-lang/rust/pull/45225.
This patch changes LLVM to allow discriminated unions to be
represented in DWARF. It adds createDiscriminatedUnionType and
createDiscriminatedMemberType to DIBuilder and then arranges for this
to be emitted using DWARF's DW_TAG_variant_part and DW_TAG_variant.
Note that DWARF requires that a discriminated union be represented as
a structure with a variant part. However, as Rust only needs to emit
pure discriminated unions, this is what I chose to expose on
DIBuilder.
Patch by Tom Tromey!
Differential Revision: https://reviews.llvm.org/D42082
llvm-svn: 324426
See D42509 for the original version of this.
Basically, there are two significant changes to behavior here:
- addLiveOuts always adds all pristine registers (even if a block has
no successors).
- addLiveOuts and addLiveOutsNoPristines always add all callee-saved
registers for return blocks (including conditional return blocks).
I cleaned up the functions a bit to make it clear these properties hold.
Differential Revision: https://reviews.llvm.org/D42655
llvm-svn: 324422
VLAs may refer to a previous DIE to express the DW_AT_count of their
type. Clang generates an artificial "vla_expr" variable for this. If
this DIE hasn't been created yet LLVM asserts. This patch fixes this
by sorting the local variables so that dependencies come before they
are needed. It also replaces the linear scan in DWARFFile with a
std::map, which can be faster.
Differential Revision: https://reviews.llvm.org/D42940
llvm-svn: 324412
Instruction Selection
Cleanup cycle/validity checks in ISel (IsLegalToFold,
HandleMergeInputChains) and X86 (isFusableLoadOpStore). Now do a full
search for cycles / dependencies pruning the search when topological
property of NodeId allows.
As part of this propogate the NodeId-based cutoffs to narrow
hasPreprocessorHelper searches.
Reviewers: craig.topper, bogner
Subscribers: llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D41293
llvm-svn: 324359
Summary: Now that PR33325 is fixed, this should always improve the generated code.
Reviewers: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D42793
llvm-svn: 324317
Summary:
This method is trying to use the truncate node to find which SETCC operand should be replaced directly with the extended load.
This used to work correctly because all uses of the original load were replaced by the truncate before this function was called. So this was used to effectively bypass the truncate and find the load under it.
All but one of the callers now call this before the truncate has replaced the laod so the setcc doesn't yet use the truncate. To account for this we should pass the original load instead.
I changed the order of that one caller to make this work there too.
I don't have a test case because this is probably hidden by later DAG combines causing the extend and truncate to cancel out. I assume this way is a little more efficient and matches what was originally intended.
Reviewers: RKSimon, spatel, niravd
Reviewed By: niravd
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D42878
llvm-svn: 324311
Summary:
If the load is already an extended load we should be using the memory VT for the legality check, not just the VT of the current extension.
I don't have a test case, just noticed it while investigating some load extension improvements.
Reviewers: RKSimon, spatel, niravd
Reviewed By: niravd
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D42783
llvm-svn: 324181
Better to assume that any value type may be commuted, not just MVTs.
No test case right now, but discovered while investigating possible shuffle combines.
llvm-svn: 324179
When handling vectors with non byte-sized elements, reverse the order of the
elements in the built integer if the target is Big-Endian.
SystemZ tests updated.
Review: Eli Friedman, Ulrich Weigand.
https://reviews.llvm.org/D42786
llvm-svn: 324063
When getNode() is called to create an EXTRACT_VECTOR_ELT, assert that
the result VT is at least as wide as the vector element type.
Review: Eli Friedman
llvm-svn: 324061
Example situation:
```
BB0:
%0 = ...
use %0
; ...
condjump BB1
jmp BB2
BB1:
%0 = ... ; rematerialized def from above (from earlier split step)
jmp BB2
BB2:
; ...
use %0
```
%0 will have a live interval with 3 value numbers (for the BB0, BB1 and
BB2 parts). Now SplitKit tries and succeeds in rematerializing the value
number in BB2 (This only works because it is a secondary split so
SplitKit is can trace this back to a single original def).
We need to recompute all live ranges affected by a value number that we
rematerialize. The case that we missed before is that when the value
that is rematerialized is at a join (Phi VNI) then we also have to
recompute liveness for the predecessor VNIs.
rdar://35699130
Differential Revision: https://reviews.llvm.org/D42667
llvm-svn: 324039
We were only checking the element count, but not the total width. This could cause illegal bitcasts to be created if for example the output was 512-bits, but N1 is 256 bits, and the extraction size was 128-bits.
Fixes PR36199
Differential Revision: https://reviews.llvm.org/D42809
llvm-svn: 324002
Until we support extending loads properly we're going to fall back for these.
We already handle stores in the same way, so this is just being consistent.
llvm-svn: 324001
Increment the field list member count for base classes and virtual base
classes.
Differential Revision: https://reviews.llvm.org/D41874
llvm-svn: 324000
Summary:
This change extends MachineCopyPropagation to do COPY source forwarding
and adds an additional run of the pass to the default pass pipeline just
after register allocation.
This version of this patch uses the newly added
MachineOperand::isRenamable bit to avoid forwarding registers is such a
way as to violate constraints that aren't captured in the
Machine IR (e.g. ABI or ISA constraints).
This change is a continuation of the work started in D30751.
Reviewers: qcolombet, javed.absar, MatzeB, jonpa, tstellar
Subscribers: tpr, mgorny, mcrosier, nhaehnle, nemanjai, jyknight, hfinkel, arsenm, inouehrs, eraman, sdardis, guyblank, fedor.sergeev, aheejin, dschuff, jfb, myatsina, llvm-commits
Differential Revision: https://reviews.llvm.org/D41835
llvm-svn: 323991
As shown in the example in PR34994:
https://bugs.llvm.org/show_bug.cgi?id=34994
...we can return a very wrong answer (inf instead of 0.0) for square root when
using a reciprocal square root estimate instruction.
Here, I've conditionalized the filtering out of denorms based on the function
having "denormal-fp-math"="ieee" in its attributes. The other options for this
attribute are 'preserve-sign' and 'positive-zero'.
So we don't generate this extra code by default with just '-ffast-math' (because
then there's no denormal attribute string at all), but it works if you specify
'-ffast-math -fdenormal-fp-math=ieee' from clang.
As noted in the review, there may be other problems in clang that affect the
results depending on platform (Linux x86 at least), but this should allow
creating the desired codegen.
Differential Revision: https://reviews.llvm.org/D42323
llvm-svn: 323981
Summary:
In Instruction Selection UpdateChains replaces all matched Nodes'
chain references including interior token factors and deletes them.
This may allow nodes which depend on these interior nodes but are not
part of the set of matched nodes to be left with a dangling dependence.
Avoid this by doing the replacement for matched non-TokenFactor nodes.
Fixes PR36164.
Reviewers: jonpa, RKSimon, bogner
Subscribers: llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D42754
llvm-svn: 323977
Summary:
This change expands the amount of registers stashed by the entry and
`__xray_CustomEvent` trampolines.
We've found that since the `__xray_CustomEvent` trampoline calls can show up in
situations where the scratch registers are being used, and since we don't
typically want to affect the code-gen around the disabled
`__xray_customevent(...)` intrinsic calls, that we need to save and restore the
state of even the scratch registers in the handling of these custom events.
Reviewers: pcc, pelikan, dblaikie, eizan, kpw, echristo, chandlerc
Reviewed By: echristo
Subscribers: chandlerc, echristo, hiraditya, davide, dblaikie, llvm-commits
Differential Revision: https://reviews.llvm.org/D40894
llvm-svn: 323940
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
Summary:
Call MRI.freezeReservedRegs() on functions created during outlining so
that calls to isReserved() by the verifier called after this pass won't
assert.
Reviewers: MatzeB, qcolombet, paquette
Subscribers: mcrosier, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D42749
llvm-svn: 323905
Summary:
This change is part of step five in the series of changes to remove alignment argument from
memcpy/memmove/memset in favour of alignment attributes. In particular, this changes the
CodeGenPrepare pass to be more aggressive in improving the source and destination alignments
of memcpy/memmove/memset by exploiting our new ability to record independent alignments
for each argument.
Steps:
Step 1) Remove alignment parameter and create alignment parameter attributes for
memcpy/memmove/memset. ( rL322965, rC322964, rL322963 )
Step 2) Expand the IRBuilder API to allow creation of memcpy/memmove with differing
source and dest alignments. ( rL323597 )
Step 3) Update Clang to use the new IRBuilder API. ( rC323617 )
Step 4) Update Polly to use the new IRBuilder API. ( rL323618 )
Step 5) Update LLVM passes that create memcpy/memmove calls to use the new IRBuilder API,
and those that use use MemIntrinsicInst::[get|set]Alignment() to use [get|set]DestAlignment()
and [get|set]SourceAlignment() instead. ( rL323886 )
Step 6) Remove the single-alignment IRBuilder API for memcpy/memmove, and the
MemIntrinsicInst::[get|set]Alignment() methods.
Reference
http://lists.llvm.org/pipermail/llvm-dev/2015-August/089384.htmlhttp://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html
llvm-svn: 323891
This commit came as a result for revert of patch r317579 (originally
committed as r317100). The patch made CFI instructions duplicable, because
their existence in the epilogue block was affecting the Tail duplication
pass. However, duplicating blocks with CFI instructions was an issue for
compact unwind info on Darwin, which is why the patch was reverted.
This patch allows duplicating tails with CFI instructions, though they are
not duplicable, by copying them 'manually'.
Patch by Djordje Kovacevic.
Differential Revision: https://reviews.llvm.org/D40979
llvm-svn: 323883
Summary:
Instruction Selection preserves relative orders of all nodes save
TokenFactors which we treat specially. As a result Node Ids for
TokenFactors may violate the topological ordering and should not be
considered as valid pruning candidates in predecessor search.
Fixes PR35316.
Reviewers: RKSimon, hfinkel
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D42701
llvm-svn: 323880
In D41587, @mssimpso discovered that the order of some patterns for
AArch64 was sub-optimal. I thought a bit about how we could avoid that
case in the future. I do not think there is a need for evaluating all
patterns for now. But this patch adds an extra (expensive) check, that
evaluates the latencies of all patterns, and ensures that the latency
saved decreases for subsequent patterns.
This catches the sub-optimal order fixed in D41587, but I am not
entirely happy with the check, as it only applies to sub-optimal
patterns seen while building with EXPENSIVE_CHECKS on. It did not
discover any other sub-optimal pattern ordering.
Reviewers: Gerolf, spatel, mssimpso
Reviewed By: Gerolf, mssimpso
Differential Revision: https://reviews.llvm.org/D41766
llvm-svn: 323873
When selecting a split candidate for region splitting, the register allocator tries to predict which candidate will have the cheapest spill cost.
Global splitting may cause the creation of local intervals, and they might spill.
This patch makes RA take into account the spill cost of local split intervals in use blocks (we already take into account the spill cost in through blocks).
A flag ("-condsider-local-interval-cost") controls weather we do this advanced cost calculation (it's on by default for X86 target, off for the rest).
Differential Revision: https://reviews.llvm.org/D41585
Change-Id: Icccb8ad2dbf13124f5d97a18c67d95aa6be0d14d
llvm-svn: 323870
In Thumb 1, with the new ADDCARRY / SUBCARRY the scheduler may need to do
copies CPSR ↔ GPR but not all Thumb1 targets implement them.
The schedule can attempt, before attempting a copy, to clone the instructions
but it does not currently do that for nodes with input glue. In this patch we
introduce a target-hook to let the hook decide if a glued machinenode is still
eligible for copying. In this case these are ARM::tADCS and ARM::tSBCS .
As a follow-up of this change we should actually implement the copies for the
Thumb1 targets that do implement them and restrict the hook to the targets that
can't really do such copy as these clones are not ideal.
This change fixes PR35836.
Differential Revision: https://reviews.llvm.org/D42051
llvm-svn: 323857
Sometimes users do not specify data layout in LLVM assembly and let llc set the
data layout by target triple after loading the LLVM assembly.
Currently the parser checks alloca address space no matter whether the LLVM
assembly contains data layout definition, which causes false alarm since the
default data layout does not contain the correct alloca address space.
The parser also calls verifier to check debug info and updating invalid debug
info. Currently there is no way to let the verifier to check debug info only.
If the verifier finds non-debug-info issues the parser will fail.
For llc, the fix is to remove the check of alloca addr space in the parser and
disable updating debug info, and defer the updating of debug info and
verification to be after setting data layout of the IR by target.
For other llvm tools, since they do not override data layout by target but
instead can override data layout by a command line option, an argument for
overriding data layout is added to the parser. In cases where data layout
overriding is necessary for the parser, the data layout can be provided by
command line.
Differential Revision: https://reviews.llvm.org/D41832
llvm-svn: 323826
Introduce an extension to support passing linker options to the linker.
These would be ignored by older linkers, but newer linkers which support
this feature would be able to process the linker.
Emit a special discarded section `.linker-option`. The content of this
section is a pair of strings (key, value). The key is a type identifier for
the parameter. This allows for an argument free parameter that will be
processed by the linker with the value being the parameter. As an example,
`lib` identifies a library to be linked against, traditionally the `-l`
argument for Unix-based linkers with the parameter being the library name.
Thanks to James Henderson, Cary Coutant, Rafael Espinolda, Sean Silva
for the valuable discussion on the design of this feature.
llvm-svn: 323783
PR36061 showed that during the expansion of ISD::FPOWI, that there
was an incorrect zero extension of the integer argument which for
MIPS64 would then give incorrect results. Address this with the
existing mechanism for correcting sign extensions.
This resolves PR36061.
Thanks to James Cowgill for reporting the issue!
Reviewers: atanasyan, hfinkel
Differential Revision: https://reviews.llvm.org/D42537
llvm-svn: 323781
Legal if we have hardware support for floating point, libcalls
otherwise.
Also add the necessary support for libcalls in the legalizer helper.
llvm-svn: 323726
When a function return value can't be directly lowered, such as
returning an i128 on WebAssembly, as indicated by the CanLowerReturn
target hook, SelectionDAGBuilder can translate it to return the
value through a hidden sret-like argument.
If such a function has an argument with the "returned" attribute,
the attribute can't be automatically lowered, because the function
no longer has a normal return value. For now, just discard the
"returned" attribute.
This fixes PR36128.
llvm-svn: 323715
When RAFast sees liveins in on a basic block, it uses that information
to initialize the availability of the registers. The called
method uses an instruction as one of its argument and in the liveins
case, RAFast was dereferencing MBB::begin which can be MBB::end for
empty basic block.
Change the API of definePhysReg to use MachineBasicBlock::iterator
instead of MachineInstr so that we don't dereference an
invalid iterator while making the call.
rdar://problem/36952401
llvm-svn: 323710
Summary:
Apparently, we missed on constraining register classes of VReg-operands of all the instructions
built from a destination pattern but the root (top-level) one. The issue exposed itself
while selecting G_FPTOSI for armv7: the corresponding pattern generates VTOSIZS wrapped
into COPY_TO_REGCLASS, so top-level COPY_TO_REGCLASS gets properly constrained,
while nested VTOSIZS (or rather its destination virtual register to be exact) does not.
Fixing this by issuing GIR_ConstrainSelectedInstOperands for every nested GIR_BuildMI.
https://bugs.llvm.org/show_bug.cgi?id=35965
rdar://problem/36886530
Patch by Roman Tereshin
Reviewers: dsanders, qcolombet, rovka, bogner, aditya_nandakumar, volkan
Reviewed By: dsanders, qcolombet, rovka
Subscribers: aemerson, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D42565
llvm-svn: 323692
Prior to committing r323681, we decided to change pick() to identity() since
it wasn't clear from the name what pick() did. However, identity() isn't a very
good name either since it implies that no changes are made. For some reason,
naming it changeTo() didn't occur to me until just after the commit. This
should resolve the lack of clarity that pick() had while still implying that
it changes the MIR.
llvm-svn: 323689
Rafael pointed out that `hasInternalLinkage() || hasPrivateLinkage()` is
equivalent to `hasLocalLinkage()` in post-commit review.
I'm intentionally not updating the comment, partly because I like it
being explicit, and partly because "global symbols with local linkage"
sounds like an oxymoron.
llvm-svn: 323688
Summary:
As discussed in D42244, we have difficulty describing the legality of some
operations. We're not able to specify relationships between types.
For example, declaring the following
setAction({..., 0, s32}, Legal)
setAction({..., 0, s64}, Legal)
setAction({..., 1, s32}, Legal)
setAction({..., 1, s64}, Legal)
currently declares these type combinations as legal:
{s32, s32}
{s64, s32}
{s32, s64}
{s64, s64}
but we currently have no means to say that, for example, {s64, s32} is
not legal. Some operations such as G_INSERT/G_EXTRACT/G_MERGE_VALUES/
G_UNMERGE_VALUES have relationships between the types that are currently
described incorrectly.
Additionally, G_LOAD/G_STORE currently have no means to legalize non-atomics
differently to atomics. The necessary information is in the MMO but we have no
way to use this in the legalizer. Similarly, there is currently no way for the
register type and the memory type to differ so there is no way to cleanly
represent extending-load/truncating-store in a way that can't be broken by
optimizers (resulting in illegal MIR).
It's also difficult to control the legalization strategy. We've added support
for legalizing non-power of 2 types but there's still some hardcoded assumptions
about the strategy. The main one I've noticed is that type0 is always legalized
before type1 which is not a good strategy for `type0 = G_EXTRACT type1, ...` if
you need to widen the container. It will converge on the same result eventually
but it will take a much longer route when legalizing type0 than if you legalize
type1 first.
Lastly, the definition of legality and the legalization strategy is kept
separate which is not ideal. It's helpful to be able to look at a one piece of
code and see both what is legal and the method the legalizer will use to make
illegal MIR more legal.
This patch adds a layer onto the LegalizerInfo (to be removed when all targets
have been migrated) which resolves all these issues.
Here are the rules for shift and division:
for (unsigned BinOp : {G_LSHR, G_ASHR, G_SDIV, G_UDIV})
getActionDefinitions(BinOp)
.legalFor({s32, s64}) // If type0 is s32/s64 then it's Legal
.clampScalar(0, s32, s64) // If type0 is <s32 then WidenScalar to s32
// If type0 is >s64 then NarrowScalar to s64
.widenScalarToPow2(0) // Round type0 scalars up to powers of 2
.unsupported(); // Otherwise, it's unsupported
This describes everything needed to both define legality and describe how to
make illegal things legal.
Here's an example of a complex rule:
getActionDefinitions(G_INSERT)
.unsupportedIf([=](const LegalityQuery &Query) {
// If type0 is smaller than type1 then it's unsupported
return Query.Types[0].getSizeInBits() <= Query.Types[1].getSizeInBits();
})
.legalIf([=](const LegalityQuery &Query) {
// If type0 is s32/s64/p0 and type1 is a power of 2 other than 2 or 4 then it's legal
// We don't need to worry about large type1's because unsupportedIf caught that.
const LLT &Ty0 = Query.Types[0];
const LLT &Ty1 = Query.Types[1];
if (Ty0 != s32 && Ty0 != s64 && Ty0 != p0)
return false;
return isPowerOf2_32(Ty1.getSizeInBits()) &&
(Ty1.getSizeInBits() == 1 || Ty1.getSizeInBits() >= 8);
})
.clampScalar(0, s32, s64)
.widenScalarToPow2(0)
.maxScalarIf(typeInSet(0, {s32}), 1, s16) // If type0 is s32 and type1 is bigger than s16 then NarrowScalar type1 to s16
.maxScalarIf(typeInSet(0, {s64}), 1, s32) // If type0 is s64 and type1 is bigger than s32 then NarrowScalar type1 to s32
.widenScalarToPow2(1) // Round type1 scalars up to powers of 2
.unsupported();
This uses a lambda to say that G_INSERT is unsupported when type0 is bigger than
type1 (in practice, this would be a default rule for G_INSERT). It also uses one
to describe the legal cases. This particular predicate is equivalent to:
.legalFor({{s32, s1}, {s32, s8}, {s32, s16}, {s64, s1}, {s64, s8}, {s64, s16}, {s64, s32}})
In terms of performance, I saw a slight (~6%) performance improvement when
AArch64 was around 30% ported but it's pretty much break even right now.
I'm going to take a look at constexpr as a means to reduce the initialization
cost.
Future work:
* Make it possible for opcodes to share rulesets. There's no need for
G_LSHR/G_ASHR/G_SDIV/G_UDIV to have separate rule and ruleset objects. There's
no technical barrier to this, it just hasn't been done yet.
* Replace the type-index numbers with an enum to get .clampScalar(Type0, s32, s64)
* Better names for things like .maxScalarIf() (clampMaxScalar?) and the vector rules.
* Improve initialization cost using constexpr
Possible future work:
* It's possible to make these rulesets change the MIR directly instead of
returning a description of how to change the MIR. This should remove a little
overhead caused by parsing the description and routing to the right code, but
the real motivation is that it removes the need for LegalizeAction::Custom.
With Custom removed, there's no longer a requirement that Custom legalization
change the opcode to something that's considered legal.
Reviewers: ab, t.p.northover, qcolombet, rovka, aditya_nandakumar, volkan, reames, bogner
Reviewed By: bogner
Subscribers: hintonda, bogner, aemerson, mgorny, javed.absar, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D42251
llvm-svn: 323681
Summary:
Fix a few places that were modifying code after register
allocation to set the renamable bit correctly to avoid failing the
validation added in D42449.
llvm-svn: 323675
Summary:
The improvements to the LegalizerInfo discussed in D42244 require that
LegalizerInfo::LegalizeAction be available for use in other classes. As such,
it needs to be moved out of LegalizerInfo. This has been done separately to the
next patch to minimize the noise in that patch.
llvm-svn: 323669
Microsoft Visual Studio rejects the static constexpr static list of
atoms even though it's valid C++. This provides a workaround to unbreak
the bots.
llvm-svn: 323667
MSVC complains that the constexpr "expression did not evaluate to a
constant". Trying to make it happy by adding a `const` specifier as
suggested in https://stackoverflow.com/questions/37574343.
llvm-svn: 323659
This patch adds support for generating accelerator tables in dsymutil.
This feature was already present in our internal repository but not yet
upstreamed because it requires changes to the Apple accelerator table
implementation.
Differential revision: https://reviews.llvm.org/D42501
llvm-svn: 323655
This patch renames DwarfAccelTable.{h,cpp} to AccelTable.{h,cpp} and
moves the header to the include dir so it is accessible by the
dsymutil implementation.
Differential revision: https://reviews.llvm.org/D42529
llvm-svn: 323654
This patch refactors the way data is stored in the accelerator table and
makes them truly generic. There have been several attempts to do this in
the past:
- D8215 & D8216: Using a union and partial hardcoding.
- D11805: Using inheritance.
- D42246: Using a callback.
In the end I didn't like either of them, because for some reason or
another parts of it felt hacky or decreased runtime performance. I
didn't want to completely rewrite them as I was hoping that we could
reuse parts for the successor in the DWARF standard. However, it seems
less and less likely that there will be a lot of opportunities for
sharing code and/or an interface.
Originally I choose to template the whole class, because it introduces
no performance overhead compared to the original implementation.
We ended up settling on a hybrid between a templated method and a
virtual call to emit the data. The motivation is that we don't want to
increase code size for a feature that should soon be superseded by the
DWARFv5 accelerator tables. While the code will continue to be used for
compatibility, it won't be on the hot path. Furthermore this does not
regress performance compared to Apple's internal implementation that
already uses virtual calls for this.
A quick summary for why these changes are necessary: dsymutil likes to
reuse the current implementation of the Apple accelerator tables.
However, LLDB expects a slightly different interface than what is
currently emitted. Additionally, in dsymutil we only have offsets and no
actual DIEs.
Although the patch suggests a lot of code has changed, this change is
pretty straightforward:
- We created an abstract class `AppleAccelTableData` to serve as an
interface for the different data classes.
- We created two implementations of this class, one for type tables and
one for everything else. There will be a third one for dsymutil that
takes just the offset.
- We use the supplied class to deduct the atoms for the header which
makes the structure of the table fully self contained, although not
enforced by the interface as was the case for the fully templated
approach.
- We renamed the prefix from DWARF- to Apple- to make space for the
future implementation of .debug_names.
This change is NFC and relies on the existing tests.
Differential revision: https://reviews.llvm.org/D42334
llvm-svn: 323653
Summary:
When emitting the location for a global variable with fragmented debug
expressions, make sure that the offset pieces, which represent
optimized-out parts of the variable, are emitted before their succeeding
fragments' expressions. Previously, if the succeeding fragment's
location was a symbol, the offset piece was emitted after, rather than
before, that symbol's expression. This effectively meant that the symbols
were associated with the wrong parts of the variable.
This fixes PR36085.
Patch by: David Stenberg
Reviewers: aprantl, probinson, dblaikie
Reviewed By: aprantl
Subscribers: JDevlieghere, llvm-commits
Tags: #debug-info
Differential Revision: https://reviews.llvm.org/D42527
llvm-svn: 323644
Summary:
There's a check in the code to only check getSetCCResultType after LegalOperations or if the type is MVT::i1. But the i1 check is only allowing scalar types through. I think it should check that the scalar type is MVT::i1 so that it will work for vectors.
The changed test already does this combine with AVX512VL where getSetCCResultType returns vXi1. But with avx512f and no VLX getSetCCResultType returns a type matching the width of the input type.
Reviewers: spatel, RKSimon
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D42619
llvm-svn: 323631
This patch moves the DJB hash to support. This is consistent with other
hashing algorithms living there. The hash is used by the DWARF
accelerator tables. We're doing this now because the hashing function is
needed by dsymutil and we don't want to link against libBinaryFormat.
Differential revision: https://reviews.llvm.org/D42594
llvm-svn: 323616
The code was using getValueSizeInBits and combining with the result of a call to DAG.ComputeNumSignBits. But for vector types getValueSizeInBits returns the width of the full vector while ComputeNumSignBits is going to give a number no larger than the width of a single element. So we should be using getScalarValueSizeInBits to get the element width.
llvm-svn: 323583
We weren't converting the immediate ConstantFP during legalization, which caused
the wrong bit patterns to be emitted for half type FP constants.
Fixes PR36106.
llvm-svn: 323582
One common source of blocks with no successors is calls to noreturn
functions; we want to preserve pristine registers in case they throw an
exception.
The whole pristine register thing is messy (we should really prefer to
explicitly model registers), but this fills a hole in the model for now.
Fixes https://bugs.llvm.org/show_bug.cgi?id=36073.
Differential Revision: https://reviews.llvm.org/D42509
llvm-svn: 323559
Previously some targets printed their own message at the start of Select to indicate what they were selecting. For the targets that didn't, it means there was no print of the root node before any custom handling in the target executed. So if the target did something custom and never called SelectNodeCommon, no print would be made. For the targets that did print a message in Select, if they didn't custom handle a node SelectNodeCommon would reprint the root node before walking the isel table.
It seems better to just print the message before the call to Select so all targets behave the same. And then remove the root node printing from SelectNodeCommon and just leave a message that says we're starting the table search.
There were also some oddities in blank line behavior. Usually due to a \n after a call to SelectionDAGNode::dump which already inserted a new line.
llvm-svn: 323551
Summary: This is the producer side for DWARF v5 string offsets tables. The reader/consumer
side was committed with r321295. All compile and type units in a module share a
contribution to the string offsets table. Indirect strings use the strx{1,2,3,4} index forms.
Reviewers: dblaikie, aprantl, JDevliegehere
Differential Revision: https://reviews.llvm.org/D42021
llvm-svn: 323546
Add support for printing / parsing the addrspace of a MachineMemOperand.
Fixes PR35970.
Differential Revision: https://reviews.llvm.org/D42502
llvm-svn: 323521
https://reviews.llvm.org/D41373
The various components are
GICombinerHelper contains transformations that are common to all
targets. Targets can pick and choose which transformations (at
function/opcode granularity) each pass uses via configuring a
GICombinerInfo.
GICombiner contains some common code and it does the traversal,
driving of combines, worklist management and iterating until
convergence.
GICombinerInfo is an interface with a virtual method called combine.
The combiner info will allow targets to pick and choose (or
implement their own specific combines). CombineInfos can make
use of available combines in GICombineHelper to configure the
transformations for a particular pass. Currently this approach allows
cherry picking transformations from helpers (at function/opcode
granularity) and also allows early returning on specific
transformations. Targets also get to prioritize whether target specific
combines run before/after the opt-in generic combines. Ideally we would
like this part to be configured by both C++ and Tablegen. The
CombinerInfo also has a field which indicates how to deal with
IllegalOps (ie - should we allow to create them/or legalize them?).
A CombinerPass would configure a CombinerInfo, create the GICombiner
with the Info, and call
GICombiner::combineMachineInstrs(MachineFunction&).
This organization is very similar to the GISelLegalizer.
llvm-svn: 323392
Apparently checking the pass structure isn't enough to ensure that we don't fall
back to FastISel, as it's set up as part of the SelectionDAGISel.
llvm-svn: 323369
Summary:
`getAction(const InstrAspect &) const` breaks encapsulation by exposing
the smaller components that are used to decide how to legalize an
instruction.
This is a problem because we need to change the implementation of
LegalizerInfo so that it's able to describe particular type combinations
rather than just cartesian products of types.
For example, declaring the following
setAction({..., 0, s32}, Legal)
setAction({..., 0, s64}, Legal)
setAction({..., 1, s32}, Legal)
setAction({..., 1, s64}, Legal)
currently declares these type combinations as legal:
{s32, s32}
{s64, s32}
{s32, s64}
{s64, s64}
but we currently have no means to say that, for example, {s64, s32} is
not legal. Some operations such as G_INSERT/G_EXTRACT/G_MERGE_VALUES/
G_UNMERGE_VALUES has relationships between the types that are currently
described incorrectly.
Additionally, G_LOAD/G_STORE currently have no means to legalize non-atomics
differently to atomics. The necessary information is in the MMO but we have no
way to use this in the legalizer. Similarly, there is currently no way for the
register type and the memory type to differ so there is no way to cleanly
represent extending-load/truncating-store in a way that can't be broken by
optimizers (resulting in illegal MIR).
This patch introduces LegalityQuery which provides all the information
needed by the legalizer to make a decision on whether something is legal
and how to legalize it.
Reviewers: ab, t.p.northover, qcolombet, rovka, aditya_nandakumar, volkan, reames, bogner
Reviewed By: bogner
Subscribers: bogner, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D42244
llvm-svn: 323342
Summary:
This patch implements the codegen of DWARF debug info for non-constant
'count' fields for DISubrange.
This is patch [2/3] in a series to extend LLVM's DISubrange Metadata
node to support debugging of C99 variable length arrays and vectors with
runtime length like the Scalable Vector Extension for AArch64. It is
also a first step towards representing more complex cases like arrays
in Fortran.
Reviewers: echristo, pcc, aprantl, dexonsmith, clayborg, kristof.beyls, dblaikie
Reviewed By: aprantl
Subscribers: fhahn, aemerson, rengolin, JDevlieghere, llvm-commits
Differential Revision: https://reviews.llvm.org/D41696
llvm-svn: 323323
Summary:
This patch extends the DISubrange 'count' field to take either a
(signed) constant integer value or a reference to a DILocalVariable
or DIGlobalVariable.
This is patch [1/3] in a series to extend LLVM's DISubrange Metadata
node to support debugging of C99 variable length arrays and vectors with
runtime length like the Scalable Vector Extension for AArch64. It is
also a first step towards representing more complex cases like arrays
in Fortran.
Reviewers: echristo, pcc, aprantl, dexonsmith, clayborg, kristof.beyls, dblaikie
Reviewed By: aprantl
Subscribers: rnk, probinson, fhahn, aemerson, rengolin, JDevlieghere, llvm-commits
Differential Revision: https://reviews.llvm.org/D41695
llvm-svn: 323313
For the included test case, the DAG transformation
concat_vectors(scalar, undef) -> scalar_to_vector(sclr)
would attempt to create a v2i32 vector for a v9i8
concat_vector. Bail out to avoid creating a bitcast with
mismatching sizes later on.
Differential Revision: https://reviews.llvm.org/D42379
llvm-svn: 323312
Merging such globals loses the dllexport attribute. Add a test
to check that normal globals still are merged.
Differential Revision: https://reviews.llvm.org/D42127
llvm-svn: 323307
https://reviews.llvm.org/D42402
A lot of these copies are useless (copies b/w VRegs having the same
regclass) and should be cleaned up.
llvm-svn: 323291
Summary:
This adds an -mllvm flag that forces the use of a runtime function call to
get the unsafe stack pointer, the same that is currently used on non-x86, non-aarch64 android.
The call may be inlined.
Reviewers: pcc
Subscribers: aemerson, kristof.beyls, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D37405
llvm-svn: 323259
Fix a bug in ScheduleDAGMILive::scheduleMI which causes BotRPTracker not tracking CurrentBottom in some rare cases involving llvm.dbg.value.
This issues causes amdgcn target to assert when compiling some user codes with -g.
Differential Revision: https://reviews.llvm.org/D42394
llvm-svn: 323214
If in complex addressing mode the difference is in GV then
base reg should not be installed because we plan to use
base reg as a merge point of different GVs.
This is a fix for PR35980.
Reviewers: reames, john.brawn, santosh
Reviewed By: john.brawn
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D42230
llvm-svn: 323192
Summary:
First, we need to explain the core of the vulnerability. Note that this
is a very incomplete description, please see the Project Zero blog post
for details:
https://googleprojectzero.blogspot.com/2018/01/reading-privileged-memory-with-side.html
The basis for branch target injection is to direct speculative execution
of the processor to some "gadget" of executable code by poisoning the
prediction of indirect branches with the address of that gadget. The
gadget in turn contains an operation that provides a side channel for
reading data. Most commonly, this will look like a load of secret data
followed by a branch on the loaded value and then a load of some
predictable cache line. The attacker then uses timing of the processors
cache to determine which direction the branch took *in the speculative
execution*, and in turn what one bit of the loaded value was. Due to the
nature of these timing side channels and the branch predictor on Intel
processors, this allows an attacker to leak data only accessible to
a privileged domain (like the kernel) back into an unprivileged domain.
The goal is simple: avoid generating code which contains an indirect
branch that could have its prediction poisoned by an attacker. In many
cases, the compiler can simply use directed conditional branches and
a small search tree. LLVM already has support for lowering switches in
this way and the first step of this patch is to disable jump-table
lowering of switches and introduce a pass to rewrite explicit indirectbr
sequences into a switch over integers.
However, there is no fully general alternative to indirect calls. We
introduce a new construct we call a "retpoline" to implement indirect
calls in a non-speculatable way. It can be thought of loosely as
a trampoline for indirect calls which uses the RET instruction on x86.
Further, we arrange for a specific call->ret sequence which ensures the
processor predicts the return to go to a controlled, known location. The
retpoline then "smashes" the return address pushed onto the stack by the
call with the desired target of the original indirect call. The result
is a predicted return to the next instruction after a call (which can be
used to trap speculative execution within an infinite loop) and an
actual indirect branch to an arbitrary address.
On 64-bit x86 ABIs, this is especially easily done in the compiler by
using a guaranteed scratch register to pass the target into this device.
For 32-bit ABIs there isn't a guaranteed scratch register and so several
different retpoline variants are introduced to use a scratch register if
one is available in the calling convention and to otherwise use direct
stack push/pop sequences to pass the target address.
This "retpoline" mitigation is fully described in the following blog
post: https://support.google.com/faqs/answer/7625886
We also support a target feature that disables emission of the retpoline
thunk by the compiler to allow for custom thunks if users want them.
These are particularly useful in environments like kernels that
routinely do hot-patching on boot and want to hot-patch their thunk to
different code sequences. They can write this custom thunk and use
`-mretpoline-external-thunk` *in addition* to `-mretpoline`. In this
case, on x86-64 thu thunk names must be:
```
__llvm_external_retpoline_r11
```
or on 32-bit:
```
__llvm_external_retpoline_eax
__llvm_external_retpoline_ecx
__llvm_external_retpoline_edx
__llvm_external_retpoline_push
```
And the target of the retpoline is passed in the named register, or in
the case of the `push` suffix on the top of the stack via a `pushl`
instruction.
There is one other important source of indirect branches in x86 ELF
binaries: the PLT. These patches also include support for LLD to
generate PLT entries that perform a retpoline-style indirection.
The only other indirect branches remaining that we are aware of are from
precompiled runtimes (such as crt0.o and similar). The ones we have
found are not really attackable, and so we have not focused on them
here, but eventually these runtimes should also be replicated for
retpoline-ed configurations for completeness.
For kernels or other freestanding or fully static executables, the
compiler switch `-mretpoline` is sufficient to fully mitigate this
particular attack. For dynamic executables, you must compile *all*
libraries with `-mretpoline` and additionally link the dynamic
executable and all shared libraries with LLD and pass `-z retpolineplt`
(or use similar functionality from some other linker). We strongly
recommend also using `-z now` as non-lazy binding allows the
retpoline-mitigated PLT to be substantially smaller.
When manually apply similar transformations to `-mretpoline` to the
Linux kernel we observed very small performance hits to applications
running typical workloads, and relatively minor hits (approximately 2%)
even for extremely syscall-heavy applications. This is largely due to
the small number of indirect branches that occur in performance
sensitive paths of the kernel.
When using these patches on statically linked applications, especially
C++ applications, you should expect to see a much more dramatic
performance hit. For microbenchmarks that are switch, indirect-, or
virtual-call heavy we have seen overheads ranging from 10% to 50%.
However, real-world workloads exhibit substantially lower performance
impact. Notably, techniques such as PGO and ThinLTO dramatically reduce
the impact of hot indirect calls (by speculatively promoting them to
direct calls) and allow optimized search trees to be used to lower
switches. If you need to deploy these techniques in C++ applications, we
*strongly* recommend that you ensure all hot call targets are statically
linked (avoiding PLT indirection) and use both PGO and ThinLTO. Well
tuned servers using all of these techniques saw 5% - 10% overhead from
the use of retpoline.
We will add detailed documentation covering these components in
subsequent patches, but wanted to make the core functionality available
as soon as possible. Happy for more code review, but we'd really like to
get these patches landed and backported ASAP for obvious reasons. We're
planning to backport this to both 6.0 and 5.0 release streams and get
a 5.0 release with just this cherry picked ASAP for distros and vendors.
This patch is the work of a number of people over the past month: Eric, Reid,
Rui, and myself. I'm mailing it out as a single commit due to the time
sensitive nature of landing this and the need to backport it. Huge thanks to
everyone who helped out here, and everyone at Intel who helped out in
discussions about how to craft this. Also, credit goes to Paul Turner (at
Google, but not an LLVM contributor) for much of the underlying retpoline
design.
Reviewers: echristo, rnk, ruiu, craig.topper, DavidKreitzer
Subscribers: sanjoy, emaste, mcrosier, mgorny, mehdi_amini, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D41723
llvm-svn: 323155
1. ReachingDefsAnalysis - Allows to identify for each instruction what is the “closest” reaching def of a certain register. Used by BreakFalseDeps (for clearance calculation) and ExecutionDomainFix (for arbitrating conflicting domains).
2. ExecutionDomainFix - Changes the variant of the instructions in order to minimize domain crossings.
3. BreakFalseDeps - Breaks false dependencies.
4. LoopTraversal - Creatws a traversal order of the basic blocks that is optimal for loops (introduced in revision L293571). Both ExecutionDomainFix and ReachingDefsAnalysis use this to determine the order they will traverse the basic blocks.
This also included the following changes to ExcecutionDepsFix original logic:
1. BreakFalseDeps and ReachingDefsAnalysis logic no longer restricted by a register class.
2. ReachingDefsAnalysis tracks liveness of reg units instead of reg indices into a given reg class.
Additional changes in affected files:
1. X86 and ARM targets now inherit from ExecutionDomainFix instead of ExecutionDepsFix. BreakFalseDeps also was added to the passes they activate.
2. Comments and references to ExecutionDepsFix replaced with ExecutionDomainFix and BreakFalseDeps, as appropriate.
Additional refactoring changes will follow.
This commit is (almost) NFC.
The only functional change is that now BreakFalseDeps will break dependency for all register classes.
Since no additional instructions were added to the list of instructions that have false dependencies, there is no actual change yet.
In a future commit several instructions (and tests) will be added.
This is the first of multiple patches that fix bugzilla https://bugs.llvm.org/show_bug.cgi?id=33869
Most of the patches are intended at refactoring the existent code.
Additional relevant reviews:
https://reviews.llvm.org/D40331https://reviews.llvm.org/D40332https://reviews.llvm.org/D40333https://reviews.llvm.org/D40334
Differential Revision: https://reviews.llvm.org/D40330
Change-Id: Icaeb75e014eff96a8f721377783f9a3e6c679275
llvm-svn: 323087
This was completely broken, but hopefully fixed by this patch.
In cases where it is needed, a vector with non byte-sized elements is stored
by extracting, zero-extending, shift:ing and or:ing the elements into an
integer of the same width as the vector, which is then stored.
Review: Eli Friedman, Ulrich Weigand
https://reviews.llvm.org/D42100#inline-369520https://bugs.llvm.org/show_bug.cgi?id=35520
llvm-svn: 323042
`llvm.used` contains a list of pointers to named values which the
compiler, assembler, and linker are required to treat as if there is a
reference that they cannot see. Ensure that the symbols are preserved
by adding an explicit `-include` reference to the linker command.
llvm-svn: 323017
Previously, the DIBuilder didn't expose functionality to set its compile unit
in any other way than calling createCompileUnit. This meant that the outliner,
which creates new functions, had to create a new compile unit for its debug
info.
This commit adds an optional parameter in the DIBuilder's constructor which
lets you set its CU at construction.
It also changes the MachineOutliner so that it keeps track of the DISubprograms
for each outlined sequence. If debugging information is requested, then it
uses one of the outlined sequence's DISubprograms to grab a CU. It then uses
that CU to construct the DISubprogram for the new outlined function.
The test has also been updated to reflect this change.
See https://reviews.llvm.org/D42254 for more information. Also see the e-mail
discussion on D42254 in llvm-commits for more context.
llvm-svn: 322992
Sanjay Patel