Skip to content

[lld-macho] Fix branch extension logic compatibility with __objc_stubs #137913

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
May 1, 2025
Merged

[lld-macho] Fix branch extension logic compatibility with __objc_stubs #137913

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
90 changes: 60 additions & 30 deletions lld/MachO/ConcatOutputSection.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -133,8 +133,16 @@ bool TextOutputSection::needsThunks() const {
// we've already processed in this segment needs thunks, so do the
// rest.
bool needsThunks = parent && parent->needsThunks;

// Calculate the total size of all branch target sections
uint64_t branchTargetsSize = in.stubs->getSize();

// Add the size of __objc_stubs section if it exists
if (in.objcStubs && in.objcStubs->isNeeded())
branchTargetsSize += in.objcStubs->getSize();

if (!needsThunks &&
isecAddr - addr + in.stubs->getSize() <=
isecAddr - addr + branchTargetsSize <=
std::min(target->backwardBranchRange, target->forwardBranchRange))
return false;
// Yes, this program is large enough to need thunks.
Expand All @@ -148,11 +156,11 @@ bool TextOutputSection::needsThunks() const {
auto *sym = cast<Symbol *>(r.referent);
// Pre-populate the thunkMap and memoize call site counts for every
// InputSection and ThunkInfo. We do this for the benefit of
// estimateStubsInRangeVA().
// estimateBranchTargetThresholdVA().
ThunkInfo &thunkInfo = thunkMap[sym];
// Knowing ThunkInfo call site count will help us know whether or not we
// might need to create more for this referent at the time we are
// estimating distance to __stubs in estimateStubsInRangeVA().
// estimating distance to __stubs in estimateBranchTargetThresholdVA().
++thunkInfo.callSiteCount;
// We can avoid work on InputSections that have no BRANCH relocs.
isec->hasCallSites = true;
Expand All @@ -161,19 +169,20 @@ bool TextOutputSection::needsThunks() const {
return true;
}

// Since __stubs is placed after __text, we must estimate the address
// beyond which stubs are within range of a simple forward branch.
// This is called exactly once, when the last input section has been finalized.
uint64_t TextOutputSection::estimateStubsInRangeVA(size_t callIdx) const {
// Estimate the address beyond which branch targets (like __stubs and
// __objc_stubs) are within range of a simple forward branch. This is called
// exactly once, when the last input section has been finalized.
uint64_t
TextOutputSection::estimateBranchTargetThresholdVA(size_t callIdx) const {
// Tally the functions which still have call sites remaining to process,
// which yields the maximum number of thunks we might yet place.
size_t maxPotentialThunks = 0;
for (auto &tp : thunkMap) {
ThunkInfo &ti = tp.second;
// This overcounts: Only sections that are in forward jump range from the
// currently-active section get finalized, and all input sections are
// finalized when estimateStubsInRangeVA() is called. So only backward
// jumps will need thunks, but we count all jumps.
// finalized when estimateBranchTargetThresholdVA() is called. So only
// backward jumps will need thunks, but we count all jumps.
if (ti.callSitesUsed < ti.callSiteCount)
maxPotentialThunks += 1;
}
Expand All @@ -200,7 +209,8 @@ uint64_t TextOutputSection::estimateStubsInRangeVA(size_t callIdx) const {
assert(isecEnd - isecVA <= forwardBranchRange &&
"should only finalize sections in jump range");

// Estimate the maximum size of the code, right before the stubs section.
// Estimate the maximum size of the code, right before the branch target
// sections.
uint64_t maxTextSize = 0;
// Add the size of all the inputs, including the unprocessed ones.
maxTextSize += isecEnd;
Expand All @@ -214,21 +224,35 @@ uint64_t TextOutputSection::estimateStubsInRangeVA(size_t callIdx) const {
// 'maxPotentialThunks' overcounts, this is an estimate of the upper limit.
maxTextSize += maxPotentialThunks * target->thunkSize;

// Estimated maximum VA of last stub.
uint64_t maxVAOfLastStub = maxTextSize + in.stubs->getSize();
// Calculate the total size of all late branch target sections
uint64_t branchTargetsSize = 0;

// Add the size of __stubs section
branchTargetsSize += in.stubs->getSize();

// Add the size of __objc_stubs section if it exists
if (in.objcStubs && in.objcStubs->isNeeded())
branchTargetsSize += in.objcStubs->getSize();

// Estimated maximum VA of the last branch target.
uint64_t maxVAOfLastBranchTarget = maxTextSize + branchTargetsSize;

// Estimate the address after which call sites can safely call stubs
// Estimate the address after which call sites can safely call branch targets
// directly rather than through intermediary thunks.
uint64_t stubsInRangeVA = maxVAOfLastStub - forwardBranchRange;
uint64_t branchTargetThresholdVA =
maxVAOfLastBranchTarget - forwardBranchRange;

log("thunks = " + std::to_string(thunkMap.size()) +
", potential = " + std::to_string(maxPotentialThunks) +
", stubs = " + std::to_string(in.stubs->getSize()) + ", isecVA = " +
utohexstr(isecVA) + ", threshold = " + utohexstr(stubsInRangeVA) +
", isecEnd = " + utohexstr(isecEnd) +
", stubs = " + std::to_string(in.stubs->getSize()) +
(in.objcStubs && in.objcStubs->isNeeded()
? ", objc_stubs = " + std::to_string(in.objcStubs->getSize())
: "") +
", isecVA = " + utohexstr(isecVA) + ", threshold = " +
utohexstr(branchTargetThresholdVA) + ", isecEnd = " + utohexstr(isecEnd) +
", tail = " + utohexstr(isecEnd - isecVA) +
", slop = " + utohexstr(forwardBranchRange - (isecEnd - isecVA)));
return stubsInRangeVA;
return branchTargetThresholdVA;
}

void ConcatOutputSection::finalizeOne(ConcatInputSection *isec) {
Expand All @@ -254,7 +278,7 @@ void TextOutputSection::finalize() {

uint64_t forwardBranchRange = target->forwardBranchRange;
uint64_t backwardBranchRange = target->backwardBranchRange;
uint64_t stubsInRangeVA = TargetInfo::outOfRangeVA;
uint64_t branchTargetThresholdVA = TargetInfo::outOfRangeVA;
size_t thunkSize = target->thunkSize;
size_t relocCount = 0;
size_t callSiteCount = 0;
Expand Down Expand Up @@ -297,16 +321,18 @@ void TextOutputSection::finalize() {
if (!isec->hasCallSites)
continue;

if (finalIdx == endIdx && stubsInRangeVA == TargetInfo::outOfRangeVA) {
// When we have finalized all input sections, __stubs (destined
// to follow __text) comes within range of forward branches and
// we can estimate the threshold address after which we can
// reach any stub with a forward branch. Note that although it
// sits in the middle of a loop, this code executes only once.
if (finalIdx == endIdx &&
branchTargetThresholdVA == TargetInfo::outOfRangeVA) {
// When we have finalized all input sections, branch target sections (like
// __stubs and __objc_stubs) (destined to follow __text) come within range
// of forward branches and we can estimate the threshold address after
// which we can reach any branch target with a forward branch. Note that
// although it sits in the middle of a loop, this code executes only once.
// It is in the loop because we need to call it at the proper
// time: the earliest call site from which the end of __text
// (and start of __stubs) comes within range of a forward branch.
stubsInRangeVA = estimateStubsInRangeVA(callIdx);
// (and start of branch target sections) comes within range of a forward
// branch.
branchTargetThresholdVA = estimateBranchTargetThresholdVA(callIdx);
}
// Process relocs by ascending address, i.e., ascending offset within isec
std::vector<Reloc> &relocs = isec->relocs;
Expand All @@ -328,10 +354,14 @@ void TextOutputSection::finalize() {
auto *funcSym = cast<Symbol *>(r.referent);
ThunkInfo &thunkInfo = thunkMap[funcSym];
// The referent is not reachable, so we need to use a thunk ...
if (funcSym->isInStubs() && callVA >= stubsInRangeVA) {
if ((funcSym->isInStubs() ||
(in.objcStubs && in.objcStubs->isNeeded() &&
ObjCStubsSection::isObjCStubSymbol(funcSym))) &&
callVA >= branchTargetThresholdVA) {
assert(callVA != TargetInfo::outOfRangeVA);
// ... Oh, wait! We are close enough to the end that __stubs
// are now within range of a simple forward branch.
// ... Oh, wait! We are close enough to the end that branch target
// sections (__stubs, __objc_stubs) are now within range of a simple
// forward branch.
continue;
}
uint64_t funcVA = funcSym->resolveBranchVA();
Expand Down
2 changes: 1 addition & 1 deletion lld/MachO/ConcatOutputSection.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,7 +80,7 @@ class TextOutputSection : public ConcatOutputSection {
}

private:
uint64_t estimateStubsInRangeVA(size_t callIdx) const;
uint64_t estimateBranchTargetThresholdVA(size_t callIdx) const;

std::vector<ConcatInputSection *> thunks;
};
Expand Down
40 changes: 38 additions & 2 deletions lld/test/MachO/arm64-thunks.s
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,9 +18,28 @@
## with safe_thunks ICF.
# RUN: %lld -arch arm64 -dead_strip -lSystem -U _extern_sym -map %t/thunk.map -o %t/thunk %t/input.o --icf=safe_thunks
# RUN: llvm-objdump --no-print-imm-hex -d --no-show-raw-insn %t/thunk | FileCheck %s
# RUN: llvm-objdump --macho --section-headers %t/thunk > %t/headers.txt
# RUN: llvm-otool -vs __DATA __objc_selrefs %t/thunk >> %t/headers.txt
# RUN: llvm-otool -vs __TEXT __objc_stubs %t/thunk >> %t/headers.txt
# RUN: FileCheck %s --check-prefix=OBJC < %t/headers.txt

# RUN: FileCheck %s --input-file %t/thunk.map --check-prefix=MAP


# OBJC: Sections:
# OBJC: __text
# OBJC-NEXT: __lcxx_override
# OBJC-NEXT: __stubs
# OBJC-NEXT: __stub_helper
# OBJC-NEXT: __objc_stubs

# OBJC: Contents of (__DATA,__objc_selrefs) section
# OBJC-NEXT: {{[0-9a-f]*}} __TEXT:__objc_methname:foo
# OBJC-NEXT: {{[0-9a-f]*}} __TEXT:__objc_methname:bar

# OBJC: Contents of (__TEXT,__objc_stubs) section
# OBJC: _objc_msgSend$bar:
# OBJC: _objc_msgSend$foo:

# MAP: 0x{{[[:xdigit:]]+}} {{.*}} _fold_func_low_addr
# MAP-NEXT: 0x{{[[:xdigit:]]+}} {{.*}} _a
# MAP-NEXT: 0x{{[[:xdigit:]]+}} {{.*}} _b
Expand All @@ -45,7 +64,6 @@
# MAP-NEXT: 0x{{[[:xdigit:]]+}} {{.*}} _e.thunk.1
# MAP-NEXT: 0x{{[[:xdigit:]]+}} {{.*}} _f.thunk.1
# MAP-NEXT: 0x{{[[:xdigit:]]+}} {{.*}} _fold_func_low_addr.thunk.0
# MAP-NEXT: 0x{{[[:xdigit:]]+}} {{.*}} ltmp0.thunk.0
# MAP-NEXT: 0x{{[[:xdigit:]]+}} {{.*}} _z


Expand Down Expand Up @@ -206,6 +224,22 @@

# CHECK: [[#%x, NAN_PAGE + NAN_OFFSET]] <__stubs>:

.section __TEXT,__objc_methname,cstring_literals
lselref1:
.asciz "foo"
lselref2:
.asciz "bar"

.section __DATA,__objc_selrefs,literal_pointers,no_dead_strip
.p2align 3
.quad lselref1
.quad lselref2

.text
.globl _objc_msgSend
_objc_msgSend:
ret

.subsections_via_symbols

.addrsig
Expand Down Expand Up @@ -352,6 +386,8 @@ _main:
bl _fold_func_low_addr
bl _fold_func_high_addr
bl ___nan
bl _objc_msgSend$foo
bl _objc_msgSend$bar
ret

.globl _fold_func_high_addr
Expand Down