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
Extend the range of calls beyond an architecture's limited branch range by first calling a thunk, which loads the far address into a scratch register (x16 on ARM64) and branches through it.
Other ports (COFF, ELF) use multiple passes with successively-refined guesses regarding the expansion of text-space imposed by thunk-space overhead. This MachO algorithm places thunks during MergedOutputSection::finalize() in a single pass using exact thunk-space overheads. Thunks are kept in a separate vector to avoid the overhead of inserting into the `inputs` vector of `MergedOutputSection`.
FIXME:
* arm64-stubs.s test is broken
* add thunk tests
* Handle thunks to DylibSymbol in MergedOutputSection::finalize()
Differential Revision: https://reviews.llvm.org/D100818
Previously, it was difficult to write code that handled both synthetic
and regular sections generically. We solve this problem by creating a
fake InputSection at the start of every SyntheticSection.
This refactor allows us to handle DSOHandle like a regular Defined
symbol (since Defined symbols must be attached to an InputSection), and
paves the way for supporting `__mh_*header` symbols. Additionally, it
simplifies our binding/rebase code.
I did have to extend Defined a little -- it now has a `linkerInternal`
flag, to indicate that `___dso_handle` should not be in the final symbol
table.
I've also added some additional testing for `___dso_handle`.
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D98545
This makes our error messages more informative. But the bigger motivation is for
LTO symbol resolution, which will be in an upcoming diff. The changes in this
one are largely mechanical.
Reviewed By: #lld-macho, smeenai
Differential Revision: https://reviews.llvm.org/D94316
- most importantly, fix a use-after-free when using thin archives,
by putting the archive unique_ptr to the arena allocator. This
ports D65565 to MachO
- correctly demangle symbol namess from archives in diagnostics
- add a test for thin archives -- it finds this UaF, but only when
running it under asan (it also finds the demangling fix)
- make forceLoadArchive() use addFile() with a bool to have the archive
loading code in fewer places. no behavior change; matches COFF port a
bit better
Differential Revision: https://reviews.llvm.org/D92360
This patch:
- adds an ld64.lld.darwinnew symlink for lld, to go with f2710d4b57,
so that `clang -fuse-ld=lld.darwinnew` can be used to test new
Mach-O lld while it's in bring-up. (The expectation is that we'll
remove this again once new Mach-O lld is the defauld and only Mach-O
lld.)
- lets the clang driver know if the linker is lld (currently
only triggered if `-fuse-ld=lld` or `-fuse-ld=lld.darwinnew` is
passed). Currently only used for the next point, but could be used
to implement other features that need close coordination between
compiler and linker, e.g. having a diag for calling `clang++` instead
of `clang` when link errors are caused by a missing C++ stdlib.
- lets the clang driver pass `-demangle` to Mach-O lld (both old and
new), in addition to ld64
- implements -demangle for new Mach-O lld
- changes demangleItanium() to accept _Z, __Z, ___Z, ____Z prefixes
(and updates one test added in D68014). Mach-O has an extra
underscore for symbols, and the three (or, on Mach-O, four)
underscores are used for block names.
Differential Revision: https://reviews.llvm.org/D91884
They operate like Defined symbols but with no associated InputSection.
Note that `ld64` seems to treat the weak definition flag like a no-op for
absolute symbols, so I have replicated that behavior.
Reviewed By: #lld-macho, smeenai
Differential Revision: https://reviews.llvm.org/D87909
The C++ ABI requires dylibs to pass a pointer to __cxa_atexit which does
e.g. cleanup of static global variables. The C++ spec says that the pointer
can point to any address in one of the dylib's segments, but in practice
ld64 seems to set it to point to the header, so that's what's implemented
here.
Reviewed By: #lld-macho, smeenai
Differential Revision: https://reviews.llvm.org/D83603
That's what ld64 uses for 64-bit targets. I figured it's best to make
this change sooner rather than later since a bunch of our tests are
relying on hardcoded addresses that depend on this value.
Reviewed By: smeenai
Differential Revision: https://reviews.llvm.org/D80177
With this change, basic archive files can be linked together. Input
section discovery has been refactored into a function since archive
files lazily resolve their symbols / the object files containing those
symbols.
Reviewed By: int3, smeenai
Differential Revision: https://reviews.llvm.org/D78342
Summary:
This is the first commit for the new Mach-O backend, designed to roughly
follow the architecture of the existing ELF and COFF backends, and
building off work that @ruiu and @pcc did in a branch a while back. Note
that this is a very stripped-down commit with the bare minimum of
functionality for ease of review. We'll be following up with more diffs
soon.
Currently, we're able to generate a simple "Hello World!" executable
that runs on OS X Catalina (and possibly on earlier OS X versions; I
haven't tested them). (This executable can be obtained by compiling
`test/MachO/relocations.s`.) We're mocking out a few load commands to
achieve this -- for example, we can't load dynamic libraries, but
Catalina requires binaries to be linked against `dyld`, so we hardcode
the emission of a `LC_LOAD_DYLIB` command. Other mocked out load
commands include LC_SYMTAB and LC_DYSYMTAB.
Differential Revision: https://reviews.llvm.org/D75382
Jez Ng