9d0754ada5 dropped MC support required for
optimal macro-fusion alignment in BOLT. Remove the support in BOLT as
performance measurements with large binaries didn't show a significant
improvement.
Test Plan:
macro-fusion alignment was never upstreamed, so no upstream tests are
affected.
Summary:
Use workaround for quadratic behavior inside
AttemptToFoldSymbolOffsetDifference called from BinaryEmitter::emitLSDA.
b06e736982 (commitcomment-142836456)
To avoid accidentally setting the label twice for the same instruction,
which can lead to a "lost" label, introduce getOrSetInstLabel()
function. Rename existing functions to getInstLabel()/setInstLabel() to
make it explicit that they operate on instruction labels. Add an
assertion in setInstLabel() that the instruction did not have a prior
label set.
Make core BOLT functionality more friendly to being used as a
library instead of in our standalone driver llvm-bolt. To
accomplish this, we augment BinaryContext with journaling streams
that are to be used by most BOLT code whenever something needs to
be logged to the screen. Users of the library can decide if logs
should be printed to a file, no file or to the screen, as
before. To illustrate this, this patch adds a new option
`--log-file` that allows the user to redirect BOLT logging to a
file on disk or completely hide it by using
`--log-file=/dev/null`. Future BOLT code should now use
`BinaryContext::outs()` for printing important messages instead of
`llvm::outs()`. A new test log.test enforces this by verifying that
no strings are print to screen once the `--log-file` option is
used.
In previous patches we also added a new BOLTError class to report
common and fatal errors, so code shouldn't call exit(1) now. To
easily handle problems as before (by quitting with exit(1)),
callers can now use
`BinaryContext::logBOLTErrorsAndQuitOnFatal(Error)` whenever code
needs to deal with BOLT errors. To test this, we have fatal.s
that checks we are correctly quitting and printing a fatal error
to the screen.
Because this is a significant change by itself, not all code was
yet ported. Code from Profiler libs (DataAggregator and friends)
still print errors directly to screen.
Co-authored-by: Rafael Auler <rafaelauler@fb.com>
Test Plan: NFC
Fix the bug where merge-fdata unconditionally outputs boltedcollection
line, regardless of whether input files have it set.
Test Plan:
Added bolt/test/X86/merge-fdata-nobat-mode.test which fails without this
fix.
This commit explicitly adds a warm code section, .text.warm, when
-split-functions -split-strategy=cdsplit is used. This replaces the
previous approach of using .text.cold.0 as warm and .text.cold.1 as cold
in 3-way function splitting. NFC.
Use MCAsmBackend::writeNopData() interface to emit NOP instructions on
x86. There are multiple forms of NOP instruction on x86 with different
sizes. Currently, LLVM's assembly/disassembly does not support all forms
correctly which can lead to a breakage of input code semantics, e.g. if
the program relies on NOP instructions for reserving a patch space.
Add "--keep-nops" option to preserve NOP instructions.
We emit a symbol before an instruction for a number of reasons, e.g. for
tracking LocSyms, debug line, or if the instruction has a label
annotation. Currently, we may emit multiple symbols per instruction.
Reuse the same label instead of creating and emitting new ones when
possible. I'm planning to refactor EH labels as well in a separate diff.
Change getLabel() to return a pointer instead of std::optional<> since
an empty label should be treated identically to no label.
Currently minimal alignment of function is hardcoded to 2 bytes.
Add 2 more cases:
1. In case BF is data in code return the alignment of CI as minimal
alignment
2. For aarch64 and riscv platforms return the minimal value of 4 (added
test for aarch64)
Otherwise fallback to returning the 2 as it previously was.
On RISC-V, there are certain relocations that target a specific
instruction instead of a more abstract location like a function or basic
block. Take the following example that loads a value from symbol `foo`:
```
nop
1: auipc t0, %pcrel_hi(foo)
ld t0, %pcrel_lo(1b)(t0)
```
This results in two relocation:
- auipc: `R_RISCV_PCREL_HI20` referencing `foo`;
- ld: `R_RISCV_PCREL_LO12_I` referencing to local label `1` which points
to the auipc instruction.
It is of utmost importance that the `R_RISCV_PCREL_LO12_I` keeps
referring to the auipc instruction; if not, the program will fail to
assemble. However, BOLT currently does not guarantee this.
BOLT currently assumes that all local symbols are jump targets and
always starts a new basic block at symbol locations. The example above
results in a CFG the looks like this:
```
.BB0:
nop
.BB1:
auipc t0, %pcrel_hi(foo)
ld t0, %pcrel_lo(.BB1)(t0)
```
While this currently works (i.e., the `R_RISCV_PCREL_LO12_I` relocation
points to the correct instruction), it has two downsides:
- Too many basic blocks are created (the example above is logically only
one yet two are created);
- If instructions are inserted in `.BB1` (e.g., by instrumentation),
things will break since the label will not point to the auipc anymore.
This patch proposes to fix this issue by teaching BOLT to track labels
that should always point to a specific instruction. This is implemented
as follows:
- Add a new annotation type (`kLabel`) that allows us to annotate
instructions with an `MCSymbol *`;
- Whenever we encounter a relocation type that is used to refer to a
specific instruction (`Relocation::isInstructionReference`), we
register it without a symbol;
- During disassembly, whenever we encounter an instruction with such a
relocation, create a symbol for its target and store it in an offset
to symbol map (to ensure multiple relocations referencing the same
instruction use the same label);
- After disassembly, iterate this map to attach labels to instructions
via the new annotation type;
- During emission, emit these labels right before the instruction.
I believe the use of annotations works quite well for this use case as
it allows us to reliably track instruction labels. If we were to store
them as offsets in basic blocks, it would be error prone to keep them
updated whenever instructions are inserted or removed.
I have chosen to add labels as first-class annotations (as opposed to a
generic one) because the documentation of `MCAnnotation` suggests that
generic annotations are to be used for optional metadata that can be
discarded without affecting correctness. As this is not the case for
labels, a first-class annotation seemed more appropriate.
The trap value used by BOLT was assumed to be single-byte instruction.
It made some functions unaligned on AArch64(e.g exceptions-instrumentation test)
and caused emission failures. Fix that by changing fill value to StringRef.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D158191
BOLT uses MCAsmLayout to calculate the output values of basic blocks.
This means output values are calculated based on a pre-linking state and
any changes to symbol values during linking will cause incorrect values
to be used.
This issue was first addressed in D154604 by adding all basic block
symbols to the symbol table for the linker to resolve them. However, the
runtime overhead of handling this huge symbol table turned out to be
prohibitively large.
This patch solves the issue in a different way. First, a temporary
section containing [input address, output symbol] pairs is emitted to the
intermediary object file. The linker will resolve all these references
so we end up with a section of [input address, output address] pairs.
This section is then parsed and used to:
- Replace BinaryBasicBlock::OffsetTranslationTable
- Replace BinaryFunction::InputOffsetToAddressMap
- Update BinaryBasicBlock::OutputAddressRange
Note that the reason this is more performant than the previous attempt
is that these symbol references do not cause entries to be added to the
symbol table. Instead, section-relative references are used for the
relocations.
Reviewed By: maksfb
Differential Revision: https://reviews.llvm.org/D155604
RuntimeDyld has been deprecated in favor of JITLink. [1] This patch
replaces all uses of RuntimeDyld in BOLT with JITLink.
Care has been taken to minimize the impact on the code structure in
order to ease the inspection of this (rather large) changeset. Since
BOLT relied on the RuntimeDyld API in multiple places, this wasn't
always possible though and I'll explain the changes in code structure
first.
Design note: BOLT uses a JIT linker to perform what essentially is
static linking. No linked code is ever executed; the result of linking
is simply written back to an executable file. For this reason, I
restricted myself to the use of the core JITLink library and avoided ORC
as much as possible.
RuntimeDyld contains methods for loading objects (loadObject) and symbol
lookup (getSymbol). Since JITLink doesn't provide a class with a similar
interface, the BOLTLinker abstract class was added to implement it. It
was added to Core since both the Rewrite and RuntimeLibs libraries make
use of it. Wherever a RuntimeDyld object was used before, it was
replaced with a BOLTLinker object.
There is one major difference between the RuntimeDyld and BOLTLinker
interfaces: in JITLink, section allocation and the application of fixups
(relocation) happens in a single call (jitlink::link). That is, there is
no separate method like finalizeWithMemoryManagerLocking in RuntimeDyld.
BOLT used to remap sections between allocating (loadObject) and linking
them (finalizeWithMemoryManagerLocking). This doesn't work anymore with
JITLink. Instead, BOLTLinker::loadObject accepts a callback that is
called before fixups are applied which is used to remap sections.
The actual implementation of the BOLTLinker interface lives in the
JITLinkLinker class in the Rewrite library. It's the only part of the
BOLT code that should directly interact with the JITLink API.
For loading object, JITLinkLinker first creates a LinkGraph
(jitlink::createLinkGraphFromObject) and then links it (jitlink::link).
For the latter, it uses a custom JITLinkContext with the following
properties:
- Use BOLT's ExecutableFileMemoryManager. This one was updated to
implement the JITLinkMemoryManager interface. Since BOLT never
executes code, its finalization step is a no-op.
- Pass config: don't use the default target passes since they modify
DWARF sections in a way that seems incompatible with BOLT. Also run a
custom pre-prune pass that makes sure sections without symbols are not
pruned by JITLink.
- Implement symbol lookup. This used to be implemented by
BOLTSymbolResolver.
- Call the section mapper callback before the final linking step.
- Copy symbol values when the LinkGraph is resolved. Symbols are stored
inside JITLinkLinker to ensure that later objects (i.e.,
instrumentation libraries) can find them. This functionality used to
be provided by RuntimeDyld but I did not find a way to use JITLink
directly for this.
Some more minor points of interest:
- BinarySection::SectionID: JITLink doesn't have something equivalent to
RuntimeDyld's Section IDs. Instead, sections can only be referred to
by name. Hence, SectionID was updated to a string.
- There seem to be no tests for Mach-O. I've tested a small hello-world
style binary but not more than that.
- On Mach-O, JITLink "normalizes" section names to include the segment
name. I had to parse the section name back from this manually which
feels slightly hacky.
[1] https://reviews.llvm.org/D145686#4222642
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D147544
This patch makes code less readable but it will clean itself after all functions are converted.
Differential Revision: https://reviews.llvm.org/D138665
We should always move jump tables when requested. Previously,
we were not moving jump tables of non-simple functions in relocation
mode. That caused a bug detailed in the attached test case: in PIC
jump tables, we force jump tables to be moved, but if they are not
moved because the function is not simple, we could incorrectly update
original entries in .rodata, corrupting it under special circumstances
(see testcase).
Reviewed By: #bolt, maksfb
Differential Revision: https://reviews.llvm.org/D137357
Simplify the logic of handling sections in BOLT. This change brings more
direct and predictable mapping of BinarySection instances to sections in
the input and output files.
* Only sections from the input binary will have a non-null SectionRef.
When a new section is created as a copy of the input section,
its SectionRef is reset to null.
* RewriteInstance::getOutputSectionName() is removed as the section name
in the output file is now defined by BinarySection::getOutputName().
* Querying BinaryContext for sections by name uses their original name.
E.g., getUniqueSectionByName(".rodata") will return the original
section even if the new .rodata section was created.
* Input file sections (with relocations applied) are emitted via MC with
".bolt.org" prefix. However, their name in the output binary is
unchanged unless a new section with the same name is created.
* New sections are emitted internally with ".bolt.new" prefix if there's
a name conflict with an input file section. Their original name is
preserved in the output file.
* Section header string table is properly populated with section names
that are actually used. Previously we used to include discarded
section names as well.
* Fix the problem when dynamic relocations were propagated to a new
section with a name that matched a section in the input binary.
E.g., the new .rodata with jump tables had dynamic relocations from
the original .rodata.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D135494
This adds a round of checks to memory references, looking for
incorrect references to jump table objects. Fix them by replacing the
jump table reference with another object reference + offset.
This solves bugs related to regular data references in code
accidentally being bound to a jump table, and this reference being
updated to a new (incorrect) location because we moved this jump
table.
Fixes#55004
Reviewed By: #bolt, maksfb
Differential Revision: https://reviews.llvm.org/D134098
To properly set the "_end" symbol, we need to track the last allocatable
address. Simply emitting "_end" at the end of some section is not
sufficient since the order of section allocation is unknown during the
emission step.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D135121
I'm planning to deprecate and eventually remove llvm::empty.
Note that no use of llvm::empty requires the ability of llvm::empty to
determine the emptiness from begin/end only.
In non-relocation mode, every function is emitted in its own section. If
a function is empty, RuntimeDyld will still allocate 1-byte section
for the function and initialize it with zero. As a result, we will
overwrite the first byte of the original function contents with zero.
Such scenario can happen when the input function had only NOP
instructions which BOLT removes by default. Even though such functions
likely cause undefined behavior, it's better to preserve their contents.
Reviewed By: yota9
Differential Revision: https://reviews.llvm.org/D133978
In non-pie binaries BOLT unconditionally converted type encoding
from indirect to absptr, which broke std exceptions since pointers
to their typeinfo were only assigned at runtime in .data section.
In this patch we preserve original encoding so that indirect
remains indirect and can be resolved at runtime, and absolute remains absolute.
Reviewed By: rafauler, maksfb
Differential Revision: https://reviews.llvm.org/D132484
For exception handling, LSDA call sites have to be emitted for each
fragment individually. With this patch, call sites and respective LSDA
symbols are generated and associated with each fragment of their
function, such that they can be used by the emitter.
Reviewed By: maksfb
Differential Revision: https://reviews.llvm.org/D132052
This changes `FunctionFragment` from being used as a temporary proxy
object to access basic block ranges to a heap-allocated object that can
store fragment-specific information.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D132050
This changes `FunctionFragment` from being used as a temporary proxy
object to access basic block ranges to a heap-allocated object that can
store fragment-specific information.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D132050
This patch adds support to generate any number of sections that are
assigned to fragments of functions that are split more than two-way.
With this, a function's *nth* split fragment goes into section
`.text.cold.n`.
This also changes `FunctionLayout::erase` to make sure, that there are
no empty fragments at the end of the function. This sometimes happens
when blocks are erased from the function. To avoid creating symbols
pointing to these fragments, they need to be removed.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D130521
This adds basic fragment awareness in the exception handling passes and
generates the necessary symbols for fragments.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D130520
This changes code emission such that it can emit specific function
fragments instead of scanning all basic blocks of a function and just
emitting those that are hot or cold.
To implement this, `FunctionLayout` explicitly distinguishes the "main"
fragment (i.e. the one that contains the entry block and is associated
with the original symbol) from "split" fragments. Additionally,
`BinaryFunction` receives support for multiple cold symbols - one for
each split fragment.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D130052
If the first block of a fragment is also a landing pad, the landing pad
is not used if an exception is thrown. This is because the landing pad
is at the same start address that the corresponding LSDA describes. In
that case, the offset in the call site records to refer to that landing
pad is zero, and a zero offset is interpreted by the personality
function as "no handler" and ignored.
Reviewed By: Amir
Differential Revision: https://reviews.llvm.org/D132053
Functions that do not contain any code still have to be emitted. This
occurs on AArch64 where functions can consist only of a constant island.
To support fragment semantics in code emission, this commits adds a
guaranteed main fragment to function layout. This fragment might be
empty, but allows us omit checks whether the function is empty in most
places.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D130051
This patch adds a dedicated class to keep track of each function's
layout. It also lays the groundwork for splitting functions into
multiple fragments (as opposed to a strict hot/cold split).
Reviewed By: maksfb
Differential Revision: https://reviews.llvm.org/D129518
Summary:
Introduce NeverAlign fragment type.
The intended usage of this fragment is to insert it before a pair of
macro-op fusion eligible instructions. NeverAlign fragment ensures that
the next fragment (first instruction in the pair) does not end at a
given alignment boundary by emitting a minimal size nop if necessary.
In effect, it ensures that a pair of macro-fusible instructions is not
split by a given alignment boundary, which is a precondition for
macro-op fusion in modern Intel Cores (64B = cache line size, see Intel
Architecture Optimization Reference Manual, 2.3.2.1 Legacy Decode
Pipeline: Macro-Fusion).
This patch introduces functionality used by BOLT when emitting code with
MacroFusion alignment already in place.
The use case is different from BoundaryAlign and instruction bundling:
- BoundaryAlign can be extended to perform the desired alignment for the
first instruction in the macro-op fusion pair (D101817). However, this
approach has higher overhead due to reliance on relaxation as
BoundaryAlign requires in the general case - see
https://reviews.llvm.org/D97982#2710638.
- Instruction bundling: the intent of NeverAlign fragment is to prevent
the first instruction in a pair ending at a given alignment boundary, by
inserting at most one minimum size nop. It's OK if either instruction
crosses the cache line. Padding both instructions using bundles to not
cross the alignment boundary would result in excessive padding. There's
no straightforward way to request instruction bundling to avoid a given
end alignment for the first instruction in the bundle.
LLVM: https://reviews.llvm.org/D97982
Manual rebase conflict history:
https://phabricator.intern.facebook.com/D30142613
Test Plan: sandcastle
Reviewers: #llvm-bolt
Subscribers: phabricatorlinter
Differential Revision: https://phabricator.intern.facebook.com/D31361547
Add functionality to allow splitting code with C++ exceptions in shared
libraries and PIEs. To overcome a limitation in exception ranges format,
for functions with fragments spanning multiple sections, add trampoline
landing pads in the same section as the corresponding throwing range.
Reviewed By: Amir
Differential Revision: https://reviews.llvm.org/D127936
BOLT treats aarch64 objects located in text as empty functions with
contant islands. Emit them with at least 8-byte alignment to the new
text section.
Vladislav Khmelevsky,
Advanced Software Technology Lab, Huawei
Differential Revision: https://reviews.llvm.org/D122097
AArch64 requires CI to be aligned to 8 bytes due to access instructions
restrictions. E.g. the ldr with imm, where imm must be aligned to 8 bytes.
Differential Revision: https://reviews.llvm.org/D122065
The cold text section alignment is set using the maximum alignment value
passed to the emitCodeAlignment. In order to calculate tentetive layout
right we will set the minimum alignment of such sections to the maximum
possible function alignment explicitly.
Differential Revision: https://reviews.llvm.org/D121392
As usual with that header cleanup series, some implicit dependencies now need to
be explicit:
llvm/DebugInfo/DWARF/DWARFContext.h no longer includes:
- "llvm/DebugInfo/DWARF/DWARFAcceleratorTable.h"
- "llvm/DebugInfo/DWARF/DWARFCompileUnit.h"
- "llvm/DebugInfo/DWARF/DWARFDebugAbbrev.h"
- "llvm/DebugInfo/DWARF/DWARFDebugAranges.h"
- "llvm/DebugInfo/DWARF/DWARFDebugFrame.h"
- "llvm/DebugInfo/DWARF/DWARFDebugLoc.h"
- "llvm/DebugInfo/DWARF/DWARFDebugMacro.h"
- "llvm/DebugInfo/DWARF/DWARFGdbIndex.h"
- "llvm/DebugInfo/DWARF/DWARFSection.h"
- "llvm/DebugInfo/DWARF/DWARFTypeUnit.h"
- "llvm/DebugInfo/DWARF/DWARFUnitIndex.h"
Plus llvm/Support/Errc.h not included by a bunch of llvm/DebugInfo/DWARF/DWARF*.h files
Preprocessed lines to build llvm on my setup:
after: 1065629059
before: 1066621848
Which is a great diff!
Discourse thread: https://discourse.llvm.org/t/include-what-you-use-include-cleanup
Differential Revision: https://reviews.llvm.org/D119723
Summary:
Move the annotation to avoid dynamic memory allocations.
Improves the CPU time of instrumenting a large binary by 1% (+-0.8%, p-value 0.01)
Test Plan: NFC
Reviewers: maksfb
FBD30091656
