MIPS 64-bit ABI does not provide special PC-relative relocation like
R_MIPS_PC32 in 32-bit case. But we can use a "chain of relocation"
defined by N64 ABIs. In that case one relocation record might contain up
to three relocations which applied sequentially. Width of a final relocation
mask applied to the result of relocation depends on the last relocation
in the chain. In case of 64-bit PC-relative relocation we need the following
chain: `R_MIPS_PC32 | R_MIPS_64`. The first relocation calculates an
offset, but does not truncate the result. The second relocation just
apply calculated result as a 64-bit value.
The 64-bit PC-relative relocation might be useful in generation of
`.eh_frame` sections to escape passing `-Wl,-z,notext` flags to linker.
Differential Revision: https://reviews.llvm.org/D80390
Summary:
Support I32/F32 registers in assembler parser and add regression tests of LD/ST
instructions.
Differential Revision: https://reviews.llvm.org/D80777
Summary:
Using a .data() member on a StringRef was discarding the StringRef
size, breaking llvm-exegesis on machines with counter sums (e.g.
Zen2).
Reviewers: oontvoo
Subscribers: mstojanovic, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80982
The OpenMP spec has the task-fulfill event for a call to omp_fulfill_event.
If the task did not yet finish execution, ompt_task_early_fulfill is used,
otherwise ompt_task_late_fulfill.
If a task does not complete, when the execution finishes (i.e., the task goes
in detached mode), ompt_task_detach instead of ompt_task_complete must be
used, when the next task is scheduled.
A test for both cases is included, which only work with clang-11+
Reviewed By: hbae
Differential revision: https://reviews.llvm.org/D80843
This patch adds clang options:
-fbasic-block-sections={all,<filename>,labels,none} and
-funique-basic-block-section-names.
LLVM Support for basic block sections is already enabled.
+ -fbasic-block-sections={all, <file>, labels, none} : Enables/Disables basic
block sections for all or a subset of basic blocks. "labels" only enables
basic block symbols.
+ -funique-basic-block-section-names: Enables unique section names for
basic block sections, disabled by default.
Differential Revision: https://reviews.llvm.org/D68049
Do not spill UNDEF GC values. Instead, replace corresponding
gc.relocate intrinsic with an (arbitrary, but recognizable) constant.
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D80714
Summary:
Combine unmerge(trunc) to enable other merge combines.
Without this combine, the scalar unmerge(trunc(merge))
pattern cannot be combined and easily lead to
hard-to-legalize merge/unmerge artifacts.
Reviewed By: arsenm
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D79567
Keeping in the affine.for to gpu.launch conversions, which should
probably be the affine.parallel to gpu.launch conversion as well.
Differential Revision: https://reviews.llvm.org/D80747
When fixing probability of unreachable edges in
BranchProbabilityInfo::calcMetadataWeights() proportionally distribute
remainder probability over the reachable edges. The old implementation
distributes the remainder probability evenly.
See examples in the fixed tests.
Reviewers: yamauchi, ebrevnov
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80611
These cases all follow the same pattern:
struct A {
friend class X;
//...
class X {};
};
But 'friend class X;' injects 'X' into the surrounding namespace scope,
rather than introducing a class member. So the second 'class X {}' is a
completely different type, which changes the meaning of the earlier name
'X' from '::X' to 'A::X'.
Additionally, the friend declaration is pointless -- members of a class
don't need to be befriended to be able to access private members.
We looked through a truncate to get to the load. So we should be
deleting the truncate first.
There is a check that the node is really unused before deleting
so this didn't cause a functional issue.
Canonicalize on storing FP options in LangOptions instead of
redundantly in CodeGenOptions. Incorporate -ffast-math directly
into the values of those LangOptions rather than considering it
separately when building FPOptions. Build IR attributes from
those options rather than a mix of sources.
We should really simplify the driver/cc1 interaction here and have
the driver pass down options that cc1 directly honors. That can
happen in a follow-up, though.
Patch by Michele Scandale!
https://reviews.llvm.org/D80315
First, do not reserve numSections in the Chunks array. In cases where
there are many non-prevailing sections, this will overallocate memory
which will not be used.
Second, free the memory for sparseChunks after initializeSymbols. After
that, it is never used.
This saves 50MB of 627MB for my use case without affecting performance.
This fixes an unhandled signed integer overflow in AddWithCarry() by
using the llvm::checkedAdd() function. Thats to Vedant Kumar for the
suggestion!
<rdar://problem/60926115>
Differential Revision: https://reviews.llvm.org/D80955
Summary:
Sketch the outline for a new document that explains how to update debug
info in various kinds of code transformations.
Some of the guidelines that belong in HowToUpdateDebugInfo.rst were in
SourceLevelDebugging.rst already under the debugify section. It seems
like the distinction between the two docs ought to be that the former is
more prescriptive, while the latter is more descriptive.
To that end I've consolidated the "how to update debug info" guidelines
which were in SourceLevelDebugging.rst into the new doc, along with the
information about using "debugify" to test transformations. Since we've
added a mir-debugify pass, I've described that as well.
Reviewers: aprantl, jmorse, chrisjackson, dsanders
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80052
The concept of G_GLOBAL_VALUE is nice and simple, but always using it as the
representation for global var addressing until selection time creates some
problems in optimizing accesses in certain code/relocation models.
The problem comes from trying to optimize adrp -> add -> load/store sequences
in the most common "small" code model. These accesses can be optimized into an
adrp -> load with the add offset being folded into the load's immediate field.
If we try to keep all global var references as a single generic instruction
then by the time we get to the complex operand trying to match these, we end up
generating an adrp at the point of use. The real issue here is that we don't
have any form of CSE during selection, so the code size will bloat from many
redundant adrp's.
This patch custom legalizes small code mode non-GOT G_GLOBALs into target ADRP
and a new "target specific generic opcode" G_ADD_LOW. We also teach the
localizer to localize these instructions via the custom hook that was added
recently. Finally, the complex pattern for indexed loads/stores is extended to
try to fold these G_ADD_LOW instructions into the load immediate.
On -O0 CTMark, we see a 0.8% geomean code size improvement. We should also see
some minor performance improvements too.
Differential Revision: https://reviews.llvm.org/D78465
between the two instructions.
If there's a pattern like:
$xA = ADRP foo @PAGE
[some killing use of reg Xb]
$Xb = ADDXri $Xa, 0, @PAGEOFF
CollectLOH would create an AdrpAdd LOH that resulted in the linker optimizing
this sequence into:
$xB = ADR foo
[some killing use of reg $Xb]
... and therefore clobbers the live $Xb register that was used by the
instruction in between.
This was discovered by a GlobalISel patch D78465 which broke up global variable
accesses into two pseudos, which in some cases could be moved apart.
Differential Revision: https://reviews.llvm.org/D80834
In the added testfile, the from argument was recognized as
&Element{SymRegion{reg_$0<long * global_a>},-1 S64b,long}
instead of
reg_$0<long * global_a>.
Like we do for empty std::array, make sure we have assertions in place
for obvious out-of-bounds issues in std::array when the debug mode is
enabled (which isn't by default).
Summary: This changes Clang's generation of OpenMP runtime functions to use the types and functions defined in OpenMPKinds and OpenMPConstants. New OpenMP runtime function information should now be added to OMPKinds.def. This patch also changed the definitions of __kmpc_push_num_teams and __kmpc_copyprivate to match those found in the runtime.
Reviewers: jdoerfert
Reviewed By: jdoerfert
Subscribers: jfb, AndreyChurbanov, openmp-commits, fghanim, hiraditya, sstefan1, cfe-commits, llvm-commits
Tags: #openmp, #clang, #llvm
Differential Revision: https://reviews.llvm.org/D80222
Update for upstream comments. Improve test by writing all the debug
info by hand.
Reviewers: dblaikie, jhenderson
Subscribers: hiraditya, MaskRay, rupprecht, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80168
Summary:
This simplifies the interface by storing the function analysis manager
with the InlineAdvisor, and, thus, not requiring it be passed each time
we inquire for an advice.
Reviewers: davidxl, asbirlea
Subscribers: eraman, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80405
Summary: This is split off from D79192 and exposes APIGenerator (renames to APIIndexer) for use in generating the integrations tests.
Reviewers: sivachandra
Reviewed By: sivachandra
Subscribers: tschuett, ecnelises, libc-commits
Tags: #libc-project
Differential Revision: https://reviews.llvm.org/D80832
The inlinees section contains references to the file checksum table. The
file checksum table in the PDB must have the same layout as the file
checksum table in the object file, so all the existing file id
references should stay valid.
Previously, we would do this:
for all inlined functions:
- lookup filename from checksum and string table
- make that filename absolute
- look up the new file id for that filename up in the new checksum
table
This lead to pdbMakeAbsolute and remove_dots ending up in the hot path.
We should only need to absolutify the source path once, not once every
time we process an inline function from that source file.
This speeds up linking chrome PGO stage 1 net_unittests.exe from 9.203s
to 8.500s (-7.6%). Looking just at time to process symbol records, it
goes from ~2000ms to ~1300ms, which is consistent with the overall
speedup of about 700ms. This will be less noticeable in debug builds,
which have fewer inlined functions records.
This patch implements matrix index expressions
(matrix[RowIdx][ColumnIdx]).
It does so by introducing a new MatrixSubscriptExpr(Base, RowIdx, ColumnIdx).
MatrixSubscriptExprs are built in 2 steps in ActOnMatrixSubscriptExpr. First,
if the base of a subscript is of matrix type, we create a incomplete
MatrixSubscriptExpr(base, idx, nullptr). Second, if the base is an incomplete
MatrixSubscriptExpr, we create a complete
MatrixSubscriptExpr(base->getBase(), base->getRowIdx(), idx)
Similar to vector elements, it is not possible to take the address of
a MatrixSubscriptExpr.
For CodeGen, a new MatrixElt type is added to LValue, which is very
similar to VectorElt. The only difference is that we may need to cast
the type of the base from an array to a vector type when accessing it.
Reviewers: rjmccall, anemet, Bigcheese, rsmith, martong
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D76791
Remove it from target-specific scope which corresponds
to sanitizer_linux.cpp where it lives in the same macro
scope.
Differential Revision: https://reviews.llvm.org/D80864
GNU ld from binutils 2.35 onwards will likely support
--export-dynamic-symbol but with different semantics.
https://sourceware.org/pipermail/binutils/2020-May/111302.html
Differences:
1. -export-dynamic-symbol is not supported
2. --export-dynamic-symbol takes a glob argument
3. --export-dynamic-symbol can suppress binding the references to the definition within the shared object if (-Bsymbolic or -Bsymbolic-functions)
4. --export-dynamic-symbol does not imply -u
I don't think the first three points can affect any user.
For the fourth point, Not implying -u can lead to some archive members unfetched.
Add -u foo to restore the previous behavior.
Exact semantics:
* -no-pie or -pie: matched non-local defined symbols will be added to the dynamic symbol table.
* -shared: matched non-local STV_DEFAULT symbols will not be bound to definitions within the shared object
even if they would otherwise be due to -Bsymbolic, -Bsymbolic-functions, or --dynamic-list.
Reviewed By: psmith
Differential Revision: https://reviews.llvm.org/D80487
SimplifyDemandedVectorElts() bails out on ScalableVectorType
anyway, but we can exit faster with the external check.
Move this to a helper function because there are likely other
vector folds that we can try here.
In this awkward case, we have to emit custom pseudo-constrained FP
wrappers. InstrEmitter concludes that since a mayRaiseFPException
instruction had a chain, it can't add nofpexcept.
Test deferred until mayRaiseFPException is really set on everything.
Summary:
Instead of iterating over all VarLoc IDs in removeEntryValue(), just
iterate over the interval reserved for entry value VarLocs. This changes
the iteration order, hence the test update -- otherwise this is NFC.
This appears to give an ~8.5x wall time speed-up for LiveDebugValues when
compiling sqlite3.c 3.30.1 with a Release clang (on my machine):
```
---User Time--- --System Time-- --User+System-- ---Wall Time--- --- Name ---
Before: 2.5402 ( 18.8%) 0.0050 ( 0.4%) 2.5452 ( 17.3%) 2.5452 ( 17.3%) Live DEBUG_VALUE analysis
After: 0.2364 ( 2.1%) 0.0034 ( 0.3%) 0.2399 ( 2.0%) 0.2398 ( 2.0%) Live DEBUG_VALUE analysis
```
The change in removeEntryValue() is the only one that appears to affect
wall time, but for consistency (and to resolve a pending TODO), I made
the analogous changes for iterating over SpillLocKind VarLocs.
Reviewers: nikic, aprantl, jmorse, djtodoro
Subscribers: hiraditya, dexonsmith, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80684
The AMDGPU non-strict fdiv lowering needs to introduce an FP mode
switch in some cases, and has custom nodes to provide chain/glue for
the intermediate FP operations. We need to propagate nofpexcept here,
but getNode was dropping the flags.
Adding nofpexcept in the AMDGPU custom lowering is left to a future
patch.
Also fix a second case where flags were dropped, but in this case it
seems it just didn't handle this number of operands.
Test will be included in future AMDGPU patch.
Simon Atanasyan