Patch created by running:
rg -l parallelForEachN | xargs sed -i '' -c 's/parallelForEachN/parallelFor/'
No behavior change.
Differential Revision: https://reviews.llvm.org/D128140
The error used to look like this:
ld64.lld: error: undefined symbol: _foo
>>> referenced by /path/to/bar.o
Now it displays the name of the function that contains the undefined
reference as well:
ld64.lld: error: undefined symbol: _foo
>>> referenced by /path/to/bar.o:(symbol _baz+0x4)
Differential Revision: https://reviews.llvm.org/D127696
This reverts commit 942f4e3a7c.
The additional change required to avoid the assertion errors seen
previously is:
--- a/lld/MachO/ICF.cpp
+++ b/lld/MachO/ICF.cpp
@@ -443,7 +443,9 @@ void macho::foldIdenticalSections() {
/*relocVA=*/0);
isec->data = copy;
}
- } else {
+ } else if (!isEhFrameSection(isec)) {
+ // EH frames are gathered as hashables from unwindEntry above; give a
+ // unique ID to everything else.
isec->icfEqClass[0] = ++icfUniqueID;
}
}
Differential Revision: https://reviews.llvm.org/D123435
== Background ==
`llvm-mc` generates unwind info in both compact unwind and DWARF
formats. LLD already handles the compact unwind format; this diff gets
us close to handling the DWARF format properly.
== Caveats ==
It's not quite done yet, but I figure it's worth getting this reviewed
and landed first as it's shaping up to be a fairly large code change.
**Known limitations of the current code:**
* Only works for x86_64, for which `llvm-mc` emits "abs-ified"
relocations as described in 618def651b.
`llvm-mc` emits regular relocations for ARM EH frames, which we do not
yet handle correctly.
Since the feature is not ready for real use yet, I've gated it behind a
flag that only gets toggled on during test suite runs. With most of the
new code disabled, we see just a hint of perf regression, so I don't
think it'd be remiss to land this as-is:
base diff difference (95% CI)
sys_time 1.926 ± 0.168 1.979 ± 0.117 [ -1.2% .. +6.6%]
user_time 3.590 ± 0.033 3.606 ± 0.028 [ +0.0% .. +0.9%]
wall_time 7.104 ± 0.184 7.179 ± 0.151 [ -0.2% .. +2.3%]
samples 30 31
== Design ==
Like compact unwind entries, EH frames are also represented as regular
ConcatInputSections that get pointed to via `Defined::unwindEntry`. This
allows them to be handled generically by e.g. the MarkLive and ICF
code. (But note that unlike compact unwind subsections, EH frame
subsections do end up in the final binary.)
In order to make EH frames "look like" a regular ConcatInputSection,
some processing is required. First, we need to split the `__eh_frame`
section along EH frame boundaries rather than along symbol boundaries.
We do this by decoding the length field of each EH frame. Second, the
abs-ified relocations need to be turned into regular Relocs.
== Next Steps ==
In order to support EH frames on ARM targets, we will either have to
teach LLD how to handle EH frames with explicit relocs, or we can try to
make `llvm-mc` emit abs-ified relocs for ARM as well. I'm hoping to do
the latter as I think it will make the LLD implementation both simpler
and faster to execute.
== Misc ==
The `obj-file-with-stabs.s` test had to be updated as the previous
version would trip assertion errors in the code. It appears that in our
attempt to produce a minimal YAML test input, we created a file with
invalid EH frame data. I've fixed this by re-generating the YAML and not
doing any hand-pruning of it.
Reviewed By: #lld-macho, Roger
Differential Revision: https://reviews.llvm.org/D123435
The <internal> symbol was tripping an assertion in getVA() because it
was not marked as used. Per the comment above that symbols creation,
dead stripping has already occurred so marking this symbol as used is
accurate.
Fixes https://github.com/llvm/llvm-project/issues/55565
Differential revision: https://reviews.llvm.org/D126072
Follow-on to {D123276}. Now that we work with an internal
representation of compact unwind entries, we no longer need to template
our UnwindInfoSectionImpl code based on the pointer size of the target
architecture.
I've still kept the split between `UnwindInfoSectionImpl` and
`UnwindInfoSection`. I'd introduced that split in order to do type
erasure, but I think it's still useful to have in order to keep
`UnwindInfoSection`'s definition in the header file clean.
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D123277
{D123302} got me looking deeper at `includeInSymtab`. I thought it was a
little odd that there were excluded (live) symbols for which
`includeInSymtab` was false; we shouldn't have so many different ways to
exclude a symbol. As such, this diff makes the `L`-prefixed-symbol
exclusion code use `includeInSymtab` too. (Note that as part of our
support for `__eh_frame`, we will also be excluding all `__eh_frame`
symbols from the symtab in a future diff.)
Another thing I noticed is that the `emitStabs` code never has to deal
with excluded symbols because `SymtabSection::finalize()` already
filters them out. As such, I've updated the comments and asserts from
{D123302} to reflect this.
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D123433
The previous implementation of UnwindInfoSection materialized
all the compact unwind entries & applied their relocations, then parsed
the resulting data to generate the final unwind info. This design had
some unfortunate conseqeuences: since relocations can only be applied
after their referents have had addresses assigned, operations that need
to happen before address assignment must contort themselves. (See
{D113582} and observe how this diff greatly simplifies it.)
Moreover, it made synthesizing new compact unwind entries awkward.
Handling PR50956 will require us to do this synthesis, and is the main
motivation behind this diff.
Previously, instead of generating a new CompactUnwindEntry directly, we
would have had to generate a ConcatInputSection with a number of
`Reloc`s that would then get "flattened" into a CompactUnwindEntry.
This diff introduces an internal representation of `CompactUnwindEntry`
(the former `CompactUnwindEntry` has been renamed to
`CompactUnwindLayout`). The new CompactUnwindEntry stores references to
its personality symbol and LSDA section directly, without the use of
`Reloc` structs.
In addition to being easier to work with, this diff also allows us to
handle unwind info whose personality symbols are located in sections
placed after the `__unwind_info`.
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D123276
This makes it easier to pinpoint the source of the problem.
TODO: Have more relocation error messages make use of this
functionality.
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D118798
Added some comments (particularly around finalize() and
finalizeContents()) as well as doing some rephrasing / grammar fixes for
existing comments.
Also did some minor style fixups, such as by putting methods together in
a class definition and having fields of similar types next to each
other.
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D118714
In ld.lld, when an ObjFile/BitcodeFile is read in --start-lib state, the file is
given archive semantics. --end-lib closes the previous --start-lib. A build
system can use this feature as an alternative to archives. This patch ports
the feature to lld-macho.
--start-lib and --end-lib are positional, unlike usual ld64 options.
I think the slight drawback does not matter as (a) reusing option names
make build systems convenient (b) `--start-lib a.o b.o --end-lib` conveys more
information than an alternative design: `-objlib a.o -objlib b.o` because
--start-lib makes it clear which objects are in the same conceptual archive.
This provides flexibility (c) `-objlib`/`-filelist` interaction may be weird.
Close https://github.com/llvm/llvm-project/issues/52931
Reviewed By: #lld-macho, Jez Ng, oontvoo
Differential Revision: https://reviews.llvm.org/D116913
D116913 will add LazyObject. Rename LazySymbol to LazyArchive to avoid confusion
and mirror ELF.
Reviewed By: #lld-macho, Jez Ng
Differential Revision: https://reviews.llvm.org/D116914
Follup-up to D107533, where we replaced local syms with non-local.
It doesn't make sense to replace local symbol with lazy.
Differential Revision: https://reviews.llvm.org/D110040
In order to keep signal:noise high for the `__eh_frame` diff, I have teased-out the NFC changes and put them here.
Differential Revision: https://reviews.llvm.org/D114017
Similar to D113702, but for the LSDAs. Clang seems to emit all LSDA
relocs as section relocs, but ld -r can turn those relocs into symbol
ones.
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D113721
Dedup'ing unwind info is tricky because each CUE contains a different
function address, if ICF operated naively and compared the entire
contents of each CUE, entries with identical unwind info but belonging
to different functions would never be considered identical. To work
around this problem, we slice away the function address before
performing ICF. We rely on `relocateCompactUnwind()` to correctly handle
these truncated input sections.
Here are the numbers before and after D109944, D109945, and this diff
were applied, as tested on my 3.2 GHz 16-Core Intel Xeon W:
Without any optimizations:
base diff difference (95% CI)
sys_time 0.849 ± 0.015 0.896 ± 0.012 [ +4.8% .. +6.2%]
user_time 3.357 ± 0.030 3.512 ± 0.023 [ +4.3% .. +5.0%]
wall_time 3.944 ± 0.039 4.032 ± 0.031 [ +1.8% .. +2.6%]
samples 40 38
With `-dead_strip`:
base diff difference (95% CI)
sys_time 0.847 ± 0.010 0.896 ± 0.012 [ +5.2% .. +6.5%]
user_time 3.377 ± 0.014 3.532 ± 0.015 [ +4.4% .. +4.8%]
wall_time 3.962 ± 0.024 4.060 ± 0.030 [ +2.1% .. +2.8%]
samples 47 30
With `-dead_strip` and `--icf=all`:
base diff difference (95% CI)
sys_time 0.935 ± 0.013 0.957 ± 0.018 [ +1.5% .. +3.2%]
user_time 3.472 ± 0.022 6.531 ± 0.046 [ +87.6% .. +88.7%]
wall_time 4.080 ± 0.040 5.329 ± 0.060 [ +30.0% .. +31.2%]
samples 37 30
Unsurprisingly, ICF is now a lot slower, likely due to the much larger
number of input sections it needs to process. But the rest of the
linker only suffers a mild slowdown.
Note that the compact-unwind-bad-reloc.s test was expanded because we
now handle the relocation for CUE's function address in a separate code
path from the rest of the CUE relocations. The extended test covers both
code paths.
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D109946
Previously, our unwind info finalization logic assumed that the LSDA
section referenced by `__compact_unwind` was already finalized before
`__TEXT,__unwind_info` itself. However, that assumption could be broken
by the use of `-rename_section` -- it could be (and is) used to move
`__gcc_except_tab` it into a different segment later in the file.
(__TEXT is always the first non-zerofill segment, so any rename
basically guarantees that the section will be ordered after
`__unwind_info`.)
To handle this case, we compare LSDA relocations instead of their final
values in `UnwindInfoSection::finalize()`, and we actually relocate
those LSDAs in `UnwindInfoSection::writeTo()`. In order to do this, we
need an easy way to track which Symbol a given CUE corresponds to. My
solution was to change our `cuPtrVector` into a vector of indices, with
each index used for both the symbols vector (`symbolsVec`) as well as
the CUE vector (`cuVector`).
This change seems perf neutral. Numbers for linking chromium_framework
on my 16 core Mac Pro:
base diff difference (95% CI)
sys_time 1.248 ± 0.025 1.245 ± 0.026 [ -1.3% .. +0.8%]
user_time 3.588 ± 0.045 3.587 ± 0.037 [ -0.6% .. +0.5%]
wall_time 4.605 ± 0.069 4.595 ± 0.069 [ -1.0% .. +0.5%]
samples 42 26
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D113582
This diff does away with `addEntriesForFunctionsWithoutUnwindInfo()`,
because `addSymbol()` can now determine which functions need those
entries.
While overhauling UnwindInfoSection, I also parallelized the relocation
of the contents of the CUEs. This somewhat offsets the time regression
from creating one InputSection per CUE (which was done in D109944).
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D109945
Compact unwind entries (CUEs) contain pointers to their respective
function symbols. However, during the link process, it's far more useful
to have pointers from the function symbol to the CUE than vice versa.
This diff adds that pointer in the form of `Defined::compactUnwind`.
In particular, when doing dead-stripping, we want to mark CUEs live when
their function symbol is live; and when doing ICF, we want to dedup
sections iff the symbols in that section have identical CUEs. In both
cases, we want to be able to locate the symbols within a given section,
as well as locate the CUEs belonging to those symbols. So this diff also
adds `InputSection::symbols`.
The ultimate goal of this refactor is to have ICF support dedup'ing
functions with unwind info, but that will be handled in subsequent
diffs. This diff focuses on simplifying `-dead_strip` --
`findFunctionsWithUnwindInfo` is no longer necessary, and
`Defined::isLive()` is now a lot simpler. Moreover, UnwindInfoSection no
longer has to check for dead CUEs -- we simply avoid adding them in the
first place.
Additionally, we now support stripping of dead LSDAs, which follows
quite naturally since `markLive()` can now reach them via the CUEs.
Reviewed By: #lld-macho, gkm
Differential Revision: https://reviews.llvm.org/D109944
Sometimes people intentionally re-define a dylib personlity symbol as a local defined symbol as a workaround to a ld -r bug.
As a result, we could see "too many personalities" to encode. This patch tries to handle this case by ignoring the local symbols entirely.
Differential Revision: https://reviews.llvm.org/D107533
This was missed by {D107035}. This fix addresses the following warning:
loop variable 'personality' has type 'const uint32_t &' (aka 'const unsigned int &') but is initialized with type 'const unsigned long long' resulting in a copy [-Wrange-loop-analysis]
In addition to fixing the size, I also removed the const reference,
since there's no performance benefit to avoiding copies of integer-sized
values.
ICF previously operated only within a given OutputSection. We would
merge all CFStrings first, then merge all regular code sections in a
second phase. This worked fine since CFStrings would never reference
regular `__text` sections. However, I would like to expand ICF to merge
functions that reference unwind info. Unwind info references the LSDA
section, which can in turn reference the `__text` section, so we cannot
perform ICF in phases.
In order to have ICF operate on InputSections spanning multiple
OutputSections, we need a way to distinguish InputSections that are
destined for different OutputSections, so that we don't fold across
section boundaries. We achieve this by creating OutputSections early,
and setting `InputSection::parent` to point to them. This is what
LLD-ELF does. (This change should also make it easier to implement the
`section$start$` symbols.)
This diff also folds InputSections w/o checking their flags, which I
think is the right behavior -- if they are destined for the same
OutputSection, they will have the same flags in the output (even if
their input flags differ). I.e. the `parent` pointer check subsumes the
`flags` check. In practice this has nearly no effect (ICF did not become
any more effective on chromium_framework).
I've also updated ICF.cpp's block comment to better reflect its current
status.
Reviewed By: #lld-macho, smeenai
Differential Revision: https://reviews.llvm.org/D105641
Two changess:
- Drop assertions that all symbols are in GOT
- Set allEntriesAreOmitted correctly
Related bug: 50812
Differential Revision: https://reviews.llvm.org/D105364
If the input has compact unwind info but all of it is removed
after dead stripping, we would crash. Now we don't write any
__unwind_info section at all, like ld64.
This is a bit awkward to implement because we only know the final
state of unwind info after UnwindInfoSectionImpl<Ptr>::finalize(),
which is called after sections are added. So add a small amount of
bookkeeping to relocateCompactUnwind() instead (which runs earlier)
so that we can predict what finalize() will do before it runs.
Fixes PR51010.
Differential Revision: https://reviews.llvm.org/D105557
Two bugs:
1. This tries to take the address of the last symbol plus the length
of the last symbol. However, the sorted vector is cuPtrVector,
not cuVector. Also, cuPtrVector has tombstone values removed
and cuVector doesn't. If there was a stripped value at the end,
the "last" element's value was UINT64_MAX, which meant the
sentinel value was one less than the length of that "last"
dead symbol.
2. We have to subtract in.header->addr. For 64-bit binaries that's
(1 << 32) and functionAddress is 32-bit so this is a no-op, but
for 32-bit binaries the sentinel's value was too large.
I believe this has no effect in practice since the first-level
binary search code in libunwind (in UnwindCursor.hpp) does:
uint32_t low = 0;
uint32_t high = sectionHeader.indexCount();
uint32_t last = high - 1;
while (low < high) {
uint32_t mid = (low + high) / 2;
if ((mid == last) ||
(topIndex.functionOffset(mid + 1) > targetFunctionOffset)) {
low = mid;
break;
} else {
low = mid + 1;
}
So the address of the last entry in the first-level table isn't really
checked -- except for the very end, but the check against `last` means
we just run the loop once more than necessary. But it makes `unwinddump` output
look less confusing, and it's what it looks was the intention here.
(No test since I can't think of a way to make FileCheck check that one
number is larger than another.)
Differential Revision: https://reviews.llvm.org/D105404
We have been creating many ConcatInputSections with identical values due
to .subsections_via_symbols. This diff factors out the identical values
into a Shared struct, to reduce memory consumption and make copying
cheaper.
I also changed `callSiteCount` from a uint32_t to a 31-bit field to save an
extra word.
All in all, this takes InputSection from 120 to 72 bytes (and
ConcatInputSection from 160 to 112 bytes), i.e. 30% size reduction in
ConcatInputSection.
Numbers for linking chromium_framework on my 3.2 GHz 16-Core Intel Xeon W:
N Min Max Median Avg Stddev
x 20 4.14 4.24 4.18 4.183 0.027548999
+ 20 4.04 4.11 4.075 4.0775 0.018027756
Difference at 95.0% confidence
-0.1055 +/- 0.0149005
-2.52211% +/- 0.356215%
(Student's t, pooled s = 0.0232803)
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D105305
This is a pretty big refactoring diff, so here are the motivations:
Previously, ICF ran after scanRelocations(), where we emitting
bind/rebase opcodes etc. So we had a bunch of redundant leftovers after
ICF. Having ICF run before Writer seems like a better design, and is
what LLD-ELF does, so this diff refactors it accordingly.
However, ICF had two dependencies on things occurring in Writer: 1) it
needs literals to be deduplicated beforehand and 2) it needs to know
which functions have unwind info, which was being handled by
`UnwindInfoSection::prepareRelocations()`.
In order to do literal deduplication earlier, we need to add literal
input sections to their corresponding output sections. So instead of
putting all input sections into the big `inputSections` vector, and then
filtering them by type later on, I've changed things so that literal
sections get added directly to their output sections during the 'gather'
phase. Likewise for compact unwind sections -- they get added directly
to the UnwindInfoSection now. This latter change is not strictly
necessary, but makes it easier for ICF to determine which functions have
unwind info.
Adding literal sections directly to their output sections means that we
can no longer determine `inputOrder` from iterating over
`inputSections`. Instead, we store that order explicitly on
InputSection. Bloating the size of InputSection for this purpose would
be unfortunate -- but LLD-ELF has already solved this problem: it reuses
`outSecOff` to store this order value.
One downside of this refactor is that we now make an additional pass
over the unwind info relocations to figure out which functions have
unwind info, since want to know that before `processRelocations()`. I've
made sure to run that extra loop only if ICF is enabled, so there should
be no overhead in non-optimizing runs of the linker.
The upside of all this is that the `inputSections` vector now contains
only ConcatInputSections that are destined for ConcatOutputSections, so
we can clean up a bunch of code that just existed to filter out other
elements from that vector.
I will test for the lack of redundant binds/rebases in the upcoming
cfstring deduplication diff. While binds/rebases can also happen in the
regular `.text` section, they're more common in `.data` sections, so it
seems more natural to test it that way.
This change is perf-neutral when linking chromium_framework.
Reviewed By: oontvoo
Differential Revision: https://reviews.llvm.org/D105044
libunwind uses unwind info to find the function address belonging
to the current instruction pointer. libunwind/src/CompactUnwinder.hpp's
step functions read functionStart for UNWIND_X86_64_MODE_STACK_IND
(and for nothing else), so these encodings need a dedicated entry
per function, so that the runtime can get the stacksize off the
`subq` instrunction in the function's prologue.
This matches ld64.
(CompactUnwinder.hpp from https://opensource.apple.com/source/libunwind/
also reads functionStart in a few more cases if `SUPPORT_OLD_BINARIES` is set,
but it defaults to 0, and ld64 seems to not worry about these additional
cases.)
Related upstream bug: https://crbug.com/1220175
Differential Revision: https://reviews.llvm.org/D104978
`icfEqClass` only makes sense on ConcatInputSections since (in contrast
to literal sections) they are deduplicated as an atomic unit.
Similarly, `hasPersonality` and `replacement` don't make sense on
literal sections.
This mirrors LLD-ELF, which stores `icfEqClass` only on non-mergeable
sections.
Reviewed By: #lld-macho, gkm
Differential Revision: https://reviews.llvm.org/D104670
"""Bitcode symbols only exist before LTO runs, and only serve the purpose of
resolving visibility so LTO can better optimize. Running LTO creates ObjFiles
from BitcodeFiles, and those ObjFiles contain regular Defined symbols (with
isec set and all) that will replace the bitcode symbols. So things should
(hopefully) work as-is :)"""
-- https://reviews.llvm.org/rGdbbc8d8333f29cf4ad6f4793da1adf71bbfdac69#inline-6081
Fixes PR50529. With this, lld-linked Chromium base_unittests passes on arm macs.
Surprisingly, no measurable impact on link time.
Differential Revision: https://reviews.llvm.org/D104681
ICF = Identical C(ode|OMDAT) Folding
This is the LLD ELF/COFF algorithm, adapted for MachO. So far, only `-icf all` is supported. In order to support `-icf safe`, we will need to port address-significance tables (`.addrsig` directives) to MachO, which will come in later diffs.
`check-{llvm,clang,lld}` have 0 regressions for `lld -icf all` vs. baseline ld64.
We only run ICF on `__TEXT,__text` for reasons explained in the block comment in `ConcatOutputSection.cpp`.
Here is the perf impact for linking `chromium_framekwork` on a Mac Pro (16-core Xeon W) for the non-ICF case vs. pre-ICF:
```
N Min Max Median Avg Stddev
x 20 4.27 4.44 4.34 4.349 0.043029977
+ 20 4.37 4.46 4.405 4.4115 0.025188761
Difference at 95.0% confidence
0.0625 +/- 0.0225658
1.43711% +/- 0.518873%
(Student's t, pooled s = 0.0352566)
```
Reviewed By: #lld-macho, int3
Differential Revision: https://reviews.llvm.org/D103292
I removed them in rG5de7467e982 but @thakis pointed out that
they were useful to keep, so here they are again. I've also converted
the `!isCoalescedWeak()` asserts into `!shouldOmitFromOutput()` asserts,
since the latter check subsumes the former.
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D104169
`outSecFileOff` and the associated `getFileOffset()` accessors were
unnecessary.
For all the cases we care about, `outSecFileOff` is the same as
`outSecOff`. The only time they deviate is if there are zerofill
sections within a given segment. But since zerofill sections are always
at the end of a segment, the only sections where the two values deviate
are zerofill sections themselves. And we never actually query the
outSecFileOff of zerofill sections.
As for `getFileOffset()`, the only place it was being used was to
calculate the offset of the entry symbol. However, we can compute that
value by just taking the difference between the address of the entry
symbol and the address of the Mach-O header. In fact, this appears to be
what ld64 itself does. This difference is the same as the file offset as
long as there are no intervening zerofill sections, but since `__text`
is the first section in `__TEXT`, this never happens, so our previous
use of `getFileOffset()` was not wrong -- just inefficient.
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D104177
D103977 broke a bunch of stuff as I had only tested the release build
which eliminated asserts.
I've retained the asserts where possible, but I also removed a bunch
instead of adding a whole lot of verbose ConcatInputSection casts.
These fields currently live in the parent InputSection class,
but they should be specific to ConcatInputSection, since the other
InputSection classes (that contain literals) aren't atomically live or
dead -- rather their component string/int literals should have
individual liveness states. (An upcoming diff will add liveness bits for
StringPieces and fixed-sized literals.)
I also factored out some asserts for isCoalescedWeak() in MarkLive.cpp.
We now avoid putting coalesced sections in the `inputSections` vector,
so we don't have to check/assert against it everywhere.
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D103977
Our implementation draws heavily from LLD-ELF's, which in turn delegates
its string deduplication to llvm-mc's StringTableBuilder. The messiness of
this diff is largely due to the fact that we've previously assumed that
all InputSections get concatenated together to form the output. This is
no longer true with CStringInputSections, which split their contents into
StringPieces. StringPieces are much more lightweight than InputSections,
which is important as we create a lot of them. They may also overlap in
the output, which makes it possible for strings to be tail-merged. In
fact, the initial version of this diff implemented tail merging, but
I've dropped it for reasons I'll explain later.
**Alignment Issues**
Mergeable cstring literals are found under the `__TEXT,__cstring`
section. In contrast to ELF, which puts strings that need different
alignments into different sections, clang's Mach-O backend puts them all
in one section. Strings that need to be aligned have the `.p2align`
directive emitted before them, which simply translates into zero padding
in the object file.
I *think* ld64 extracts the desired per-string alignment from this data
by preserving each string's offset from the last section-aligned
address. I'm not entirely certain since it doesn't seem consistent about
doing this; but perhaps this can be chalked up to cases where ld64 has
to deduplicate strings with different offset/alignment combos -- it
seems to pick one of their alignments to preserve. This doesn't seem
correct in general; we can in fact can induce ld64 to produce a crashing
binary just by linking in an additional object file that only contains
cstrings and no code. See PR50563 for details.
Moreover, this scheme seems rather inefficient: since unaligned and
aligned strings are all put in the same section, which has a single
alignment value, it doesn't seem possible to tell whether a given string
doesn't have any alignment requirements. Preserving offset+alignments
for strings that don't need it is wasteful.
In practice, the crashes seen so far seem to stem from x86_64 SIMD
operations on cstrings. X86_64 requires SIMD accesses to be
16-byte-aligned. So for now, I'm thinking of just aligning all strings
to 16 bytes on x86_64. This is indeed wasteful, but implementation-wise
it's simpler than preserving per-string alignment+offsets. It also
avoids the aforementioned crash after deduplication of
differently-aligned strings. Finally, the overhead is not huge: using
16-byte alignment (vs no alignment) is only a 0.5% size overhead when
linking chromium_framework.
With these alignment requirements, it doesn't make sense to attempt tail
merging -- most strings will not be eligible since their overlaps aren't
likely to start at a 16-byte boundary. Tail-merging (with alignment) for
chromium_framework only improves size by 0.3%.
It's worth noting that LLD-ELF only does tail merging at `-O2`. By
default (at `-O1`), it just deduplicates w/o tail merging. @thakis has
also mentioned that they saw it regress compressed size in some cases
and therefore turned it off. `ld64` does not seem to do tail merging at
all.
**Performance Numbers**
CString deduplication reduces chromium_framework from 250MB to 242MB, or
about a 3.2% reduction.
Numbers for linking chromium_framework on my 3.2 GHz 16-Core Intel Xeon W:
N Min Max Median Avg Stddev
x 20 3.91 4.03 3.935 3.95 0.034641016
+ 20 3.99 4.14 4.015 4.0365 0.0492336
Difference at 95.0% confidence
0.0865 +/- 0.027245
2.18987% +/- 0.689746%
(Student's t, pooled s = 0.0425673)
As expected, cstring merging incurs some non-trivial overhead.
When passing `--no-literal-merge`, it seems that performance is the
same, i.e. the refactoring in this diff didn't cost us.
N Min Max Median Avg Stddev
x 20 3.91 4.03 3.935 3.95 0.034641016
+ 20 3.89 4.02 3.935 3.9435 0.043197831
No difference proven at 95.0% confidence
Reviewed By: #lld-macho, gkm
Differential Revision: https://reviews.llvm.org/D102964
Also adds support for live_support sections, no_dead_strip sections,
.no_dead_strip symbols.
Chromium Framework 345MB unstripped -> 250MB stripped
(vs 290MB unstripped -> 236M stripped with ld64).
Doing dead stripping is a bit faster than not, because so much less
data needs to be processed:
% ministat lld_*
x lld_nostrip.txt
+ lld_strip.txt
N Min Max Median Avg Stddev
x 10 3.929414 4.07692 4.0269079 4.0089678 0.044214794
+ 10 3.8129408 3.9025559 3.8670411 3.8642573 0.024779651
Difference at 95.0% confidence
-0.144711 +/- 0.0336749
-3.60967% +/- 0.839989%
(Student's t, pooled s = 0.0358398)
This interacts with many parts of the linker. I tried to add test coverage
for all added `isLive()` checks, so that some test will fail if any of them
is removed. I checked that the test expectations for the most part match
ld64's behavior (except for live-support-iterations.s, see the comment
in the test). Interacts with:
- debug info
- export tries
- import opcodes
- flags like -exported_symbol(s_list)
- -U / dynamic_lookup
- mod_init_funcs, mod_term_funcs
- weak symbol handling
- unwind info
- stubs
- map files
- -sectcreate
- undefined, dylib, common, defined (both absolute and normal) symbols
It's possible it interacts with more features I didn't think of,
of course.
I also did some manual testing:
- check-llvm check-clang check-lld work with lld with this patch
as host linker and -dead_strip enabled
- Chromium still starts
- Chromium's base_unittests still pass, including unwind tests
Implemenation-wise, this is InputSection-based, so it'll work for
object files with .subsections_via_symbols (which includes all
object files generated by clang). I first based this on the COFF
implementation, but later realized that things are more similar to ELF.
I think it'd be good to refactor MarkLive.cpp to look more like the ELF
part at some point, but I'd like to get a working state checked in first.
Mechanical parts:
- Rename canOmitFromOutput to wasCoalesced (no behavior change)
since it really is for weak coalesced symbols
- Add noDeadStrip to Defined, corresponding to N_NO_DEAD_STRIP
(`.no_dead_strip` in asm)
Fixes PR49276.
Differential Revision: https://reviews.llvm.org/D103324
The ELF format has the concept of merge sections (marked by SHF_MERGE),
which contain data that can be safely deduplicated. The Mach-O
equivalents are called literal sections (marked by S_CSTRING_LITERALS or
S_{4,8,16}BYTE_LITERALS). While the Mach-O format doesn't use the word
'merge', to avoid confusion, I've renamed our MergedOutputSection to
ConcatOutputSection. I believe it's a more descriptive name too.
This renaming sets the stage for {D102964}.
Reviewed By: #lld-macho, alexshap
Differential Revision: https://reviews.llvm.org/D102971
Has the effect that `__mh_execute_header` stays in the symbol table of
outputs even after running `strip` on the output. I don't know if that's
important for anything -- my motivation for the patch is just is to make
the output more similar to ld64.
(Corresponds to symbolTableInAndNeverStrip in ld64.)
Differential Revision: https://reviews.llvm.org/D102619
Nico Weber