Apple Clang in Xcode 14 introduced a new feature for reducing the
overhead of objc_msgSend calls by deduplicating the setup calls for each
individual selector. This works by clang adding undefined symbols for
each selector called in a translation unit, such as `_objc_msgSend$foo`
for calling the `foo` method on any `NSObject`. There are 2
different modes for this behavior, the default directly does the setup
for `_objc_msgSend` and calls it, and the smaller option does the
selector setup, and then calls the standard `_objc_msgSend` stub
function.
The general overview of how this works is:
- Undefined symbols with the given prefix are collected
- The suffix of each matching undefined symbol is added as a string to
`__objc_methname`
- A pointer is added for every method name in the `__objc_selrefs`
section
- A `got` entry is emitted for `_objc_msgSend`
- Stubs are emitting pointing to the synthesized locations
Notes:
- Both `__objc_methname` and `__objc_selrefs` can also exist from object
files, so their contents are merged with our synthesized contents
- The compiler emits method names for defined methods, but not for
undefined symbols you call, but stubs are used for both
- This only implements the default "fast" mode currently just to reduce
the diff, I also doubt many folks will care to swap modes
- This only implements this for arm64 and x86_64, we don't need to
implement this for 32 bit iOS archs, but we should implement it for
watchOS archs in a later diff
Differential Revision: https://reviews.llvm.org/D128108
Some header files used
namespace lld {
namespace macho {
// ...
} // namespace macho
std::string toString(const Type &t);
} // namespace lld
In those files, I didn't use a nested namespace since it's not a big win there.
No behavior change.
Differential Revision: https://reviews.llvm.org/D131354
This method is called on each relocation when parsing input files, so
the overhead of using virtual functions ends up being quite large. We
now have a single non-virtual method, which reads from the appropriate
array of relocation attributes set in the TargetInfo constructor.
This change results in a modest 2.3% reduction in link time for
chromium_framework measured on an x86-64 VPS, and 0.7% on an arm64 Mac.
N Min Max Median Avg Stddev
x 10 11.869417 12.032609 11.935041 11.938268 0.045802324
+ 10 11.581526 11.785265 11.649885 11.659507 0.054634834
Difference at 95.0% confidence
-0.278761 +/- 0.0473673
-2.33502% +/- 0.396768%
(Student's t, pooled s = 0.0504124)
Differential Revision: https://reviews.llvm.org/D130000
This fixes https://github.com/llvm/llvm-project/issues/56238. ld64.lld currently does not generate __dof section in Mach-O, and -no_dtrace_dof option is on by default. However when there are user-defined dtrace symbols, ld64.lld will treat them as undefined symbols, which causes the linking to fail because lld cannot find their definitions. This patch allows ld64.lld to rewrite the instructions calling dtrace symbols to instructions like nop as what ld64 does; therefore, when encountered with user-provided dtrace probes, the linking can still succeed.
I'm not sure whether support for dtrace is expected in lld, so for now I didn't add codes to make lld emit __dof section like ld64, and only made it possible to link with dtrace symbols provided. If this feature is needed, I can add that part in Dtrace.cpp & Dtrace.h.
Reviewed By: int3, #lld-macho
Differential Revision: https://reviews.llvm.org/D129062
Linker optimization hints mark a sequence of instructions used for
synthesizing an address, like ADRP+ADD. If the referenced symbol ends up
close enough, it can be replaced by a faster sequence of instructions
like ADR+NOP.
This commit adds support for 2 of the 7 defined ARM64 optimization
hints:
- LOH_ARM64_ADRP_ADD, which transforms a pair of ADRP+ADD into ADR+NOP
if the referenced address is within +/- 1 MiB
- LOH_ARM64_ADRP_ADRP, which transforms two ADRP instructions into
ADR+NOP if they reference the same page
These two kinds already cover more than 50% of all LOHs in
chromium_framework.
Differential Review: https://reviews.llvm.org/D128093
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
Previously, we only allowed this for DylibSymbols. However, in order to
properly support `-flat_namespace` as well as `-interposable`, we need
to allow this for Defined symbols too. Therefore we hoist the
`lazyBindOffset` and the `stubsHelperIndex` into the parent Symbol
class.
The actual change to support interposition under `-flat_namespace` is in
{D119294}; the NFC changes here have been split out for easier review.
Perf regression isn't stat sig on my 3.2 GHz 16-Core Intel Xeon W linking
chromium_framework:
base diff difference (95% CI)
sys_time 1.227 ± 0.021 1.234 ± 0.031 [ -0.3% .. +1.5%]
user_time 3.665 ± 0.036 3.674 ± 0.035 [ -0.2% .. +0.7%]
wall_time 4.596 ± 0.055 4.609 ± 0.064 [ -0.3% .. +0.9%]
samples 34 47
Max RSS regression is barely stat sig:
base diff difference (95% CI)
time 1003664356.324 ± 15404053.912 1010380403.613 ± 10578309.455 [ +0.0% .. +1.3%]
samples 37 31
Reviewed By: modimo
Differential Revision: https://reviews.llvm.org/D121351
- Don't subtract thunkSize from branchRange. Most places care about
the actual maximal branch range. Subtract thunkSize in the one place
that wants to leave room for a thunk.
- Set it to 0x800_0000 instead of 0xFF_FFFF
- Subtract 4 for the positive branch direction since it's a
two's complement 24bit number sign-extended mutiplied by 4,
so its range is -0x800_0000..+0x7FF_FFFC
- Make boundary checks include the boundary values
This doesn't make a huge difference in practice. It's preparation
for a "real" fix for PR51578 -- but it also lets the repro in comment 0
in that bug place one more thunk before hitting the TODO.
Differential Revision: https://reviews.llvm.org/D108897
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
We had a hardcoded check and a stale TODO, written back when we only had
support for one architecture.
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D102154
@thakis pointed out that `mach_header` and `mach_header_64`
actually have the same set of (used) fields, with the 64-bit version
having extra padding. So we can access the fields we need using the
single `mach_header` type instead of using templates to switch between
the two.
I also spotted a potential issue where hasObjCSection tries to parse a
file w/o checking if it does indeed match the target arch... As such,
I've added a quick magic number check to ensure we don't access invalid
memory during `findCommand()`.
Addresses PR50180.
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D101724
This just parses the `-arch armv7` and emits the right header flags.
The rest will be slowly fleshed out in upcoming diffs.
Reviewed By: #lld-macho, gkm
Differential Revision: https://reviews.llvm.org/D101557
This load command records a range spanning from the end of the load
commands to the end of the `__TEXT` segment. Presumably the kernel will encrypt
all this data.
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D100973
From what I can tell, it's pretty similar to arm64. The two main differences
are:
1. No 64-bit relocations
2. Stub code writes to 32-bit registers instead of 64-bit
Plus of course the various on-disk structures like `segment_command` are using
the 32-bit instead of the 64-bit variants.
Reviewed By: #lld-macho, gkm
Differential Revision: https://reviews.llvm.org/D99822
From what I can tell, it's pretty similar to arm64. The two main differences
are:
1. No 64-bit relocations
2. Stub code writes to 32-bit registers instead of 64-bit
Plus of course the various on-disk structures like `segment_command` are using
the 32-bit instead of the 64-bit variants.
Reviewed By: #lld-macho, gkm
Differential Revision: https://reviews.llvm.org/D99822
The main challenge was handling the different on-disk structures (e.g.
`mach_header` vs `mach_header_64`). I tried to strike a balance between
sprinkling `target->wordSize == 8` checks everywhere (branchy = slow, and ugly)
and templatizing everything (causes code bloat, also ugly). I think I struck a
decent balance by judicious use of type erasure.
Note that LLD-ELF has a similar architecture, though it seems to use more templating.
Linking chromium_framework takes about the same time before and after this
change:
N Min Max Median Avg Stddev
x 20 4.52 4.67 4.595 4.5945 0.044423204
+ 20 4.5 4.71 4.575 4.582 0.056344803
No difference proven at 95.0% confidence
Reviewed By: #lld-macho, oontvoo
Differential Revision: https://reviews.llvm.org/D99633
Move some functions closer to their uses. Move detailed address-assignment logic out of the otherwise abstract `Writer::run()`. This prepares the ground for a diff to implement branch range extension thunks.
* `SyntheticSections.cpp`
** move `needsBinding()` and `prepareBranchTarget()` into `Writer.cpp`
** move `addNonLazyBindingEntries()` adjacent to its use.
* `Writer.cpp`
** move address-assignment logic from `Writer::run()` into new function `Writer::assignAddresses()`
** move `needsBinding()` and `prepareBranchTarget()` from `SyntheticSections.cpp`
* `Target.h`
** remove orphaned decls of `prepareSymbolRelocation()` and `validateRelocationInfo()` which were moved to other files in earlier diffs.
Differential Revision: https://reviews.llvm.org/D98795
This diff required fixing `getEmbeddedAddend` to apply sign
extension to 32-bit values. We were previously passing around wrong
64-bit addend values that became "right" after being truncated back to
32-bit.
I've also made `getEmbeddedAddend` return a signed int, which is similar
to what LLD-ELF does for its `getImplicitAddend`.
`reportRangeError`, `checkUInt`, and `checkInt` are counterparts of similar
functions in LLD-ELF.
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D98387
The previous implementation miscalculated the addend, resulting
in an underflow. This meant that every SIGNED_N section relocation would
be associated with the last subsection (since the addend would now be a
huge number). We were "lucky" that this mistake was typically cancelled
out -- 64-to-32-bit-truncation meant that the final value was correct,
as long as subsections were not rearranged.
Reviewed By: #lld-macho, thakis
Differential Revision: https://reviews.llvm.org/D98385
Also add a few asserts to verify that we are indeed handling an
UNSIGNED relocation as the minued. I haven't made it an actual
user-facing error since I don't think llvm-mc is capable of generating
SUBTRACTOR relocations without an associated UNSIGNED.
Reviewed By: #lld-macho, smeenai
Differential Revision: https://reviews.llvm.org/D97103
I've adjusted the RelocAttrBits to better fit the semantics of
the relocations. In particular:
1. *_UNSIGNED relocations are no longer marked with the `TLV` bit, even
though they can occur within TLV sections. Instead the `TLV` bit is
reserved for relocations that can reference thread-local symbols, and
*_UNSIGNED relocations have their own `UNSIGNED` bit. The previous
implementation caused TLV and regular UNSIGNED semantics to be
conflated, resulting in rebase opcodes being incorrectly emitted for TLV
relocations.
2. I've added a new `POINTER` bit to denote non-relaxable GOT
relocations. This distinction isn't important on x86 -- the GOT
relocations there are either relaxable or non-relaxable loads -- but
arm64 has `GOT_LOAD_PAGE21` which loads the page that the referent
symbol is in (regardless of whether the symbol ends up in the GOT). This
relocation must reference a GOT symbol (so must have the `GOT` bit set)
but isn't itself relaxable (so must not have the `LOAD` bit). The
`POINTER` bit is used for relocations that *must* reference a GOT
slot.
3. A similar situation occurs for TLV relocations.
4. ld64 supports both a pcrel and an absolute version of
ARM64_RELOC_POINTER_TO_GOT. But the semantics of the absolute version
are pretty weird -- it results in the value of the GOT slot being
written, rather than the address. (That means a reference to a
dynamically-bound slot will result in zeroes being written.) The
programs I've tried linking don't use this form of the relocation, so
I've dropped our partial support for it by removing the relevant
RelocAttrBits.
Reviewed By: alexshap
Differential Revision: https://reviews.llvm.org/D97031
This is an initial base commit for ARM64 target arch support. I don't represent that it complete or bug-free, but wish to put it out for review now that some basic things like branch target & load/store address relocs are working.
I can add more tests to this base commit, or add them in follow-up commits.
It is not entirely clear whether I use the "ARM64" (Apple) or "AArch64" (non-Apple) naming convention. Guidance is appreciated.
Differential Revision: https://reviews.llvm.org/D88629
Add per-reloc-type attribute bits and migrate code from per-target file into target independent code, driven by reloc attributes.
Many cleanups
Differential Revision: https://reviews.llvm.org/D95121
This is a refactor to pave the way for supporting paired-ADDEND for ARM64. The only paired reloc type for X86_64 is SUBTRACTOR. In a later diff, I will add SUBTRACTOR for both X86_64 and ARM64.
* s/`getImplicitAddend`/`getAddend`/ because it handles all forms of addend: implicit, explicit, paired.
* add predicate `bool isPairedReloc()`
* check range of `relInfo.r_symbolnum` is internal, unrelated to user-input, so use `assert()`, not `error()`
* minor cleanups & rearrangements in `InputFile::parseRelocations()`
Differential Revision: https://reviews.llvm.org/D90614
Since there is no "weak lazy" lookup, function calls to weak symbols are
always non-lazily bound. We emit both regular non-lazy bindings as well
as weak bindings, in order that the weak bindings may overwrite the
non-lazy bindings if an appropriate symbol is found at runtime. However,
the bound addresses will still be written (non-lazily) into the
LazyPointerSection.
Reviewed By: #lld-macho, smeenai
Differential Revision: https://reviews.llvm.org/D86573
Previously, the BindingEntry struct could only store bindings to offsets
within InputSections. Since the GOTSection and TLVPointerSections are
OutputSections, I handled those in a separate code path. However, this
makes it awkward to support weak bindings properly without code
duplication. This diff allows BindingEntries to point directly to
OutputSections, simplifying the upcoming weak binding implementation.
Along the way, I also converted a bunch of functions taking references
to symbols to take pointers instead. Given how much casting we do for
Symbol (especially in the upcoming weak binding diffs), it's cleaner
this way.
Differential Revision: https://reviews.llvm.org/D86571
Note: What ELF refers to as "TLS", Mach-O seems to refer to as "TLV", i.e.
thread-local variables.
This diff implements support for TLV relocations that reference defined
symbols. On x86_64, TLV relocations are always used with movq opcodes, so for
defined TLVs, we don't need to create a synthetic section to store the
addresses of the symbols -- we can just convert the `movq` to a `leaq`.
One notable quirk of Mach-O's TLVs is that absolute-address relocations
inside TLV-defining sections behave differently -- their addresses are
no longer absolute, but relative to the start of the target section.
(AFAICT, RIP-relative relocations are not allowed in these sections.)
Reviewed By: #lld-macho, compnerd, smeenai
Differential Revision: https://reviews.llvm.org/D85080
Previously, we only supported binding dysyms to the GOT. This
diff adds support for binding them to any arbitrary section. C++
programs appear to use this, I believe for vtables and type_info.
This diff also makes our bind opcode encoding a bit smarter -- we now
encode just the differences between bindings, which will make things
more compact.
I was initially concerned about the performance overhead of iterating
over these relocations, but it turns out that the number of such
relocations is small. A quick analysis of my llvm-project build
directory showed that < 1.3% out of ~7M relocations are RELOC_UNSIGNED
bindings to symbols (including both dynamic and static symbols).
Reviewed By: #lld-macho, smeenai
Differential Revision: https://reviews.llvm.org/D83103
Summary:
Turns out this case is actually really common -- it happens whenever there's
a reference to an `extern` variable that ends up statically linked.
Depends on D80856.
Reviewers: ruiu, pcc, MaskRay, smeenai, alexshap, gkm, Ktwu, christylee
Reviewed By: smeenai
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80857
Summary:
We should be reading / writing our addends / relocated addresses based on
r_length, and not just based on the type of the relocation. But since only
some r_length values are valid for a given reloc type, I've also added some
validation.
ld64 has code to allow for r_length = 0 in X86_64_RELOC_BRANCH relocs, but I'm
not sure how to create such a relocation...
Reviewed By: smeenai
Differential Revision: https://reviews.llvm.org/D80854
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
I considered making a `Target::validate()` method, but I wasn't sure how
I felt about the overhead of doing yet another switch-dispatch on the
relocation type, so I put the validation in `relocateOne` instead...
might be a bit of a micro-optimization, but `relocateOne` does assume
certain things about the relocations it gets, and this error handling
makes that explicit, so it's not a totally unreasonable code
organization.
Reviewed By: smeenai
Differential Revision: https://reviews.llvm.org/D80049
Summary:
This diff implements lazy symbol binding -- very similar to the PLT
mechanism in ELF.
ELF's .plt section is broken up into two sections in Mach-O:
StubsSection and StubHelperSection. Calls to functions in dylibs will
end up calling into StubsSection, which contains indirect jumps to
addresses stored in the LazyPointerSection (the counterpart to ELF's
.plt.got).
Initially, the LazyPointerSection contains addresses that point into one
of the entry points in the middle of the StubHelperSection. The code in
StubHelperSection will push on the stack an offset into the
LazyBindingSection. The push is followed by a jump to the beginning of
the StubHelperSection (similar to PLT0), which then calls into
dyld_stub_binder. dyld_stub_binder is a non-lazily bound symbol, so this
call looks it up in the GOT.
The stub binder will look up the bind opcodes in the LazyBindingSection
at the given offset. The bind opcodes will tell the binder to update the
address in the LazyPointerSection to point to the symbol, so that
subsequent calls don't have to redo the symbol resolution. The binder
will then jump to the resolved symbol.
Depends on D78269.
Reviewers: ruiu, pcc, MaskRay, smeenai, alexshap, gkm, Ktwu, christylee
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78270
This diff implements:
* dylib loading (much of which is being restored from @pcc and @ruiu's
original work)
* The GOT_LOAD relocation, which allows us to load non-lazy dylib
symbols
* Basic bind opcode emission, which tells `dyld` how to populate the GOT
Differential Revision: https://reviews.llvm.org/D76252
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
Keith Smiley