| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The pass for inlining memcpy in BOLT was currently X86-specific and was
using the instruction `rep movsb`.
This patch implements a static size analysis system for AArch64 memcpy
inlining that extracts copy sizes from preceding instructions to then
use it to generate the optimal width-specific load/store sequences.
|
|
## Change:
* Added `--dump-dot-func` command-line option that allows users to dump
CFGs only for specific functions instead of dumping all functions (the
current only available option being `--dump-dot-all`)
## Usage:
* Users can now specify function names or regex patterns (e.g.,
`--dump-dot-func=main,helper` or `--dump-dot-func="init.*`") to generate
.dot files only for functions of interest
* Aims to save time when analysing specific functions in large binaries
(e.g., only dumping graphs for performance-critical functions identified
through profiling) and we can now avoid reduce output clutter from
generating thousands of unnecessary .dot files when analysing large
binaries
## Testing
The introduced test `dump-dot-func.test` confirms the new option does
the following:
- [x] 1. `dump-dot-func` can correctly filter a specified functions
- [x] 2. Can achieve the above with regexes
- [x] 3. Can do 1. with a list of functions
- [x] No option specified creates no dot files
- [x] Passing in a non-existent function generates no dumping messages
- [x] `dump-dot-all` continues to work as expected
|
|
With --force-patch option, every original function entry point is
overwritten with a trampoline to a new version of the function to
prevent the execution of the original code.
If the function size is too small for the trampoline code, we are forced
to bail out on rewriting the function. That presented a problem on
AArch64 due to LongJmp pass that assumed the presence of the new copy of
the function. If the new copy was not emitted it could have lead to a
relocation overflow.
Run PatchEntries pass before LongJmp and make the latter aware of the
functions that are not going to be emitted. Make --force-patch option
behavior on AArch64 consistent with other architectures.
|
|
Add two additional profile quality stats for CG (call graph) and CFG
(control flow graph) flow conservations besides the CFG discontinuity
stats introduced in #109683. The two new stats quantify how different
"in-flow" is from "out-flow" in the following cases where they should be
equal. The smaller the reported stats, the better the flow conservations
are.
CG flow conservation: for each function that is not a program entry, the
number of times the function is called according to CG ("in-flow")
should be equal to the number of times the transition from an entry
basic block of the function to another basic block within the function
is recorded ("out-flow").
CFG flow conservation: for each basic block that is not a function entry
or exit, the number of times the transition into this basic block from
another basic block within the function is recorded ("in-flow") should
be equal to the number of times the transition from this basic block to
another basic block within the function is recorded ("out-flow").
Use `-v=1` for more detailed bucketed stats, and use `-v=2` to dump
functions / basic blocks with bad flow conservations.
|
|
Identical Code Folding (ICF) folds functions that are identical into one
function, and updates symbol addresses to the new address. This reduces
the size of a binary, but can lead to problems. For example when
function pointers are compared. This can be done either explicitly in
the code or generated IR by optimization passes like Indirect Call
Promotion (ICP). After ICF what used to be two different addresses
become the same address. This can lead to a different code path being
taken.
This is where safe ICF comes in. Linker (LLD) does it using address
significant section generated by clang. If symbol is in it, or an object
doesn't have this section symbols are not folded.
BOLT does not have the information regarding which objects do not have
this section, so can't re-use this mechanism.
This implementation scans code section and conservatively marks
functions symbols as unsafe. It treats symbols as unsafe if they are
used in non-control flow instruction. It also scans through the data
relocation sections and does the same for relocations that reference a
function symbol. The latter handles the case when function pointer is
stored in a local or global variable, etc. If a relocation address
points within a vtable these symbols are skipped.
|
|
Introduce `--print-adr-relaxation` to print after ADR Relaxation pass.
|
|
In a perfect profile, each positive-execution-count block in the
function’s CFG should be reachable from a positive-execution-count
function entry block through a positive-execution-count path. This new
pass checks how well the BOLT input profile satisfies this “CFG
continuity” property.
More specifically, for each of the hottest 1000 functions, the pass
calculates the function’s fraction of basic block execution counts that
is “unreachable”. It then reports the 95th percentile of the
distribution of the 1000 unreachable fractions in a single BOLT-INFO
line. The smaller the reported value is, the better the BOLT profile
satisfies the CFG continuity property.
The default value of 1000 above can be changed via the hidden BOLT
option `-num-functions-for-continuity-check=[N]`. If more detailed stats
are needed, `-v=1` can be added to the BOLT invocation: the hottest N
functions will be grouped into 5 equally-sized buckets, from the hottest
to the coldest; for each bucket, various summary statistics of the
distribution of the fractions and the raw unreachable execution counts
will be reported.
|
|
This reverts commit 750b12f06badc4cdf767139c70090db62358bb44.
The pass should run after splitting phase, but before nop removal
|
|
For non-simple functions we need nop instruction to be presented to
transform ADR to ADRP+ADD sequence, so run this pass before remove nops
pass.
|
|
Add the ability to disable the instruction shortening pass through
--shorten-instructions=false
|
|
InitialDynoStats used to be assigned inside `runAllPasses`, but the
assignment executed before any of the passes. As we've moved
`EstimateEdgeCounts` into a pass out of ProfileReader, it needs to
execute before initial dyno stats are set.
Thus move `InitialDynoStats` into BinaryContext and assignment into
`DynoStatsSetPass`.
|
|
|
|
|
|
Bolt's RepRet pass is x86-specific, no need to add it for non-x86
targets.
|
|
Preserve nop instructions in the Linux kernel since they could be used
for runtime patching.
|
|
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
|
|
(2/2) (#81523)
As part of the effort to refactor old error handling code that
would directly call exit(1), in this patch continue the migration
on libCore, libRewrite and libPasses to use the new BOLTError
class whenever a failure occurs.
Test Plan: NFC
Co-authored-by: Rafael Auler <rafaelauler@fb.com>
|
|
As part of the effort to refactor old error handling code that
would directly call exit(1), in this patch we change the
interface to `BinaryFunctionPass` to return an Error on
`runOnFunctions()`. This gives passes the ability to report a
serious problem to the caller (RewriteInstance class), so the
caller may decide how to best handle the exceptional situation.
Co-authored-by: Rafael Auler <rafaelauler@fb.com>
Test Plan: NFC
|
|
This commit establishes the general structure of the CDSplit strategy in
SplitFunctions without incorporating the exact splitting logic. With
-split-functions -split-strategy=cdsplit, the SplitFunctions pass will
run twice: the first time is before function reordering and functions
are hot-cold split; the second time is after function reordering and
functions are hot-warm-cold split based on the fixed function ordering.
Currently, all functions are hot-warm split after the entry block in the
second splitting pass. Subsequent commits will introduce the precise
splitting logic. NFC.
|
|
Run RemoveNops pass only if --keep-nops is set to false (default).
|
|
Most of the print options are hidden, make hidden them all.
|
|
Just enough features are implemented to process a simple "hello world"
executable and produce something that still runs (including libc calls).
This was mainly a matter of implementing support for various
relocations. Currently, the following are handled:
- R_RISCV_JAL
- R_RISCV_CALL
- R_RISCV_CALL_PLT
- R_RISCV_BRANCH
- R_RISCV_RVC_BRANCH
- R_RISCV_RVC_JUMP
- R_RISCV_GOT_HI20
- R_RISCV_PCREL_HI20
- R_RISCV_PCREL_LO12_I
- R_RISCV_RELAX
- R_RISCV_NONE
Executables linked with linker relaxation will probably fail to be
processed. BOLT relocates .text to a high address while leaving .plt at
its original (low) address. This causes PC-relative PLT calls that were
relaxed to a JAL to not fit their offset in an I-immediate anymore. This
is something that will be addressed in a later patch.
Changes to the BOLT core are relatively minor. Two things were tricky to
implement and needed slightly larger changes. I'll explain those below.
The R_RISCV_CALL(_PLT) relocation is put on the first instruction of a
AUIPC/JALR pair, the second does not get any relocation (unlike other
PCREL pairs). This causes issues with the combinations of the way BOLT
processes binaries and the RISC-V MC-layer handles relocations:
- BOLT reassembles instructions one by one and since the JALR doesn't
have a relocation, it simply gets copied without modification;
- Even though the MC-layer handles R_RISCV_CALL properly (adjusts both
the AUIPC and the JALR), it assumes the immediates of both
instructions are 0 (to be able to or-in a new value). This will most
likely not be the case for the JALR that got copied over.
To handle this difficulty without resorting to RISC-V-specific hacks in
the BOLT core, a new binary pass was added that searches for
AUIPC/JALR pairs and zeroes-out the immediate of the JALR.
A second difficulty was supporting ABS symbols. As far as I can tell,
ABS symbols were not handled at all, causing __global_pointer$ to break.
RewriteInstance::analyzeRelocation was updated to handle these
generically.
Tests are provided for all supported relocations. Note that in order to
test the correct handling of PLT entries, an ELF file produced by GCC
had to be used. While I tried to strip the YAML representation, it's
still quite large. Any suggestions on how to improve this would be
appreciated.
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D145687
|
|
We have mostly harmless data races when running
BinaryContext::calculateEmittedSize() in parallel, while performing
split function pass. However, it is possible to end up in a state
where some MCSymbols are still registered and our clean up
failed. This happens rarely but it does happen, and when it happens,
it is a difficult to diagnose heisenbug. To avoid this, add a new
clean pass to perform a last check on MCSymbols, before they
undergo our final emission pass, to verify that they are in a sane
state. If we fail to do this, we might resolve some symbols to zero
and crash the output binary.
Reviewed By: #bolt, Amir
Differential Revision: https://reviews.llvm.org/D137984
|
|
Linker might relax adrp + ldr got address loading to adrp + add for
local non-preemptible symbols (e.g. hidden/protected symbols in
executable). As usually linker doesn't change relocations properly after
relaxation, so we have to handle such cases by ourselves. To do that
during relocations reading we change LD64 reloc to ADD if instruction
mismatch found and introduce FixRelaxationPass that searches for ADRP+ADD
pairs and after performing some checks we're replacing ADRP target symbol
to already fixed ADDs one.
Vladislav Khmelevsky,
Advanced Software Technology Lab, Huawei
Differential Revision: https://reviews.llvm.org/D138097
|
|
PrintInstructions was unused in BinaryFunction::print() and dump().
Reviewed By: Amir
Differential Revision: https://reviews.llvm.org/D140440
|
|
This patch adds the huge pages support (-hugify) for PIE/no-PIE
binaries. Also returned functionality to support the kernels < 5.10
where there is a problem in a dynamic loader with the alignment of
pages addresses.
Differential Revision: https://reviews.llvm.org/D129107
|
|
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
|
|
The gold linker veneers are written between functions without symbols,
so we to handle it specially in BOLT.
Vladislav Khmelevsky,
Advanced Software Technology Lab, Huawei
Differential Revision: https://reviews.llvm.org/D129260
|
|
Address fuzzer crash
Reviewed By: yota9
Differential Revision: https://reviews.llvm.org/D120696
|
|
This reverts commit 425dda76e9fac93117289fd68a2abdfb1e4a0ba5.
This commit is currently causing BOLT to crash in one of our
binaries and needs a bit more checking to make sure it is safe
to land.
|
|
The gold linker veneers are written between functions without symbols,
so we to handle it specially in BOLT.
Vladislav Khmelevsky,
Advanced Software Technology Lab, Huawei
Differential Revision: https://reviews.llvm.org/D128082
|
|
|
|
Similar to 557efc9a8b68628c2c944678c6471dac30ed9e8e.
This commit handles options where cl::ZeroOrMore is more than one line below
cl::opt.
|
|
A new "cache-aware" strategy for tail duplication.
Differential Revision: https://reviews.llvm.org/D123050
|
|
Convert simple hammocks into cmov based on misprediction rate.
Test Plan:
- Assembly test: `cmov-conversion.s`
- Testing on a binary:
# Bootstrap clang with `-x86-cmov-converter-force-all` and `-Wl,--emit-relocs`
(Release build)
# Collect perf.data:
- `clang++ <opts> bolt/lib/Core/BinaryFunction.cpp -E > bf.cpp`
- `perf record -e cycles:u -j any,u -- clang-15 bf.cpp -O2 -std=c++14 -c -o bf.o`
# Optimize clang-15 with and w/o -cmov-conversion:
- `llvm-bolt clang-15 -p perf.data -o clang-15.bolt`
- `llvm-bolt clang-15 -p perf.data -cmov-conversion -o clang-15.bolt.cmovconv`
# Run perf experiment:
- test: `clang-15.bolt.cmovconv`,
- control: `clang-15.bolt`,
- workload (clang options): `bf.cpp -O2 -std=c++14 -c -o bf.o`
Results:
```
task-clock [delta: -360.21 ± 356.75, delta(%): -1.7760 ± 1.7589, p-value: 0.047951, balance: -6]
instructions [delta: 44061118 ± 13246382, delta(%): 0.0690 ± 0.0207, p-value: 0.000001, balance: 50]
icache-misses [delta: -5534468 ± 2779620, delta(%): -0.4331 ± 0.2175, p-value: 0.028014, balance: -28]
branch-misses [delta: -1624270 ± 1113244, delta(%): -0.3456 ± 0.2368, p-value: 0.030300, balance: -22]
```
Reviewed By: rafauler
Differential Revision: https://reviews.llvm.org/D120177
|
|
Summary:
- variable 'TotalSize' set but not used
- variable 'TotalCallsTopN' set but not used
- use of bitwise '|' with boolean operands
Reviewed By: maksfb
FBD33911129
|
|
Summary:
Reformat code and put options in lexicographical order.
Comparing to clang-format output, manual formatting looks cleaner to me.
(cherry picked from FBD33481692)
|
|
Summary:
Refactor bolt/*/Rewrite to follow the braces rule for if/else/loop from
LLVM Coding Standards.
(cherry picked from FBD33305364)
|
|
Summary: Fix comments at the start of source files.
(cherry picked from FBD33274597)
|
|
Summary:
Since nops are now removed in a separate pass, the profile is consumed
on a CFG with nops. If previously a profile was generated without nops,
the offsets in the profile could be different if branches included nops
either as a source or a destination.
This diff adjust offsets to make the profile reading backwards
compatible.
(cherry picked from FBD33231254)
|
|
Summary:
The patch moves the shortenInstructions and nop remove to separate binary
passes. As a result when llvm-bolt optimizations stage will begin the
instructions of the binary functions will be absolutely the same as it
was in the binary. This is needed for the golang support by llvm-bolt.
Some of the tests must be changed, since bb alignment nops might create
unreachable BBs in original functions.
Vladislav Khmelevsky,
Advanced Software Technology Lab, Huawei
(cherry picked from FBD32896517)
|
|
Summary: Selectively apply clang-format to BOLT code base.
(cherry picked from FBD33119052)
|
|
Summary:
Some optimizations may remove all instructions in a basic block.
The pass will cleanup the CFG afterwards by removing empty basic
blocks and merging duplicate CFG edges.
The normalized CFG is printed under '-print-normalized' option.
(cherry picked from FBD32774360)
|
|
Summary:
Added new functionality of dumping simple functions into assembly.
This includes:
- function control flow (basic blocks, instructions),
- profile information as `FDATA` directives, to be consumed by link_fdata,
- data labels,
- CFI directives,
- symbols for callee functions,
- jump table symbols.
Envisioned usage:
1. Find a function that triggers BOLT crash (e.g. with `bughunter.sh`).
2. Generate reproducer asm source for that function (using `-funcs`).
3. Attach it to an issue.
4. Reduce and include as a test case.
Current limitations:
1. Emitted assembly won't match input file relocations.
2. No DWARF support.
3. Data is not emitted.
(cherry picked from FBD32746857)
|
|
Summary:
Moves source files into separate components, and make explicit
component dependency on each other, so LLVM build system knows how to
build BOLT in BUILD_SHARED_LIBS=ON.
Please use the -c merge.renamelimit=230 git option when rebasing your
work on top of this change.
To achieve this, we create a new library to hold core IR files (most
classes beginning with Binary in their names), a new library to hold
Utils, some command line options shared across both RewriteInstance
and core IR files, a new library called Rewrite to hold most classes
concerned with running top-level functions coordinating the binary
rewriting process, and a new library called Profile to hold classes
dealing with profile reading and writing.
To remove the dependency from BinaryContext into X86-specific classes,
we do some refactoring on the BinaryContext constructor to receive a
reference to the specific backend directly from RewriteInstance. Then,
the dependency on X86 or AArch64-specific classes is transfered to the
Rewrite library. We can't have the Core library depend on targets
because targets depend on Core (which would create a cycle).
Files implementing the entry point of a tool are transferred to the
tools/ folder. All header files are transferred to the include/
folder. The src/ folder was renamed to lib/.
(cherry picked from FBD32746834)
|
