[LLD][COFF] Add Support for auxiliary IAT copy #108610
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Depends on #108608. |
|
@llvm/pr-subscribers-lld @llvm/pr-subscribers-platform-windows Author: Jacek Caban (cjacek) ChangesIn addition to the auxiliary IAT, ARM64EC modules also contain a copy of it. At runtime, the auxiliary IAT is filled with the addresses of actual ARM64EC functions when possible. If patching is detected, the OS may use the IAT copy to revert the auxiliary IAT, ensuring that the call checker is used for calls to imported functions. Full diff: https://github.com/llvm/llvm-project/pull/108610.diff 9 Files Affected:
diff --git a/lld/COFF/Chunks.cpp b/lld/COFF/Chunks.cpp
index ee54fa39fc3d6a..81faa1f3de80fc 100644
--- a/lld/COFF/Chunks.cpp
+++ b/lld/COFF/Chunks.cpp
@@ -422,6 +422,16 @@ void SectionChunk::applyRelocation(uint8_t *off,
const coff_relocation &rel) const {
auto *sym = dyn_cast_or_null<Defined>(file->getSymbol(rel.SymbolTableIndex));
+ // On ARM64EC, the __imp_ symbol references the auxiliary IAT, while the
+ // __imp_aux_ symbol references the regular IAT. However, x86_64 code expects
+ // both to reference the regular IAT, so adjust the symbol if necessary.
+ if (sym && getMachine() == AMD64 && isArm64EC(file->ctx.config.machine)) {
+ if (auto impSym = dyn_cast<DefinedImportData>(sym)) {
+ if (impSym->file->impchkThunk && sym == impSym->file->impECSym)
+ sym = impSym->file->impSym;
+ }
+ }
+
// Get the output section of the symbol for this relocation. The output
// section is needed to compute SECREL and SECTION relocations used in debug
// info.
diff --git a/lld/COFF/DLL.cpp b/lld/COFF/DLL.cpp
index b00a59433319a7..39dcce9fe84837 100644
--- a/lld/COFF/DLL.cpp
+++ b/lld/COFF/DLL.cpp
@@ -730,13 +730,19 @@ void IdataContents::create(COFFLinkerContext &ctx) {
auto chunk = make<AuxImportChunk>(s->file);
auxIat.push_back(chunk);
s->file->impECSym->setLocation(chunk);
+
+ chunk = make<AuxImportChunk>(s->file);
+ auxIatCopy.push_back(chunk);
+ s->file->auxImpCopySym->setLocation(chunk);
}
}
// Terminate with null values.
lookups.push_back(make<NullChunk>(ctx.config.wordsize));
addresses.push_back(make<NullChunk>(ctx.config.wordsize));
- if (ctx.config.machine == ARM64EC)
+ if (ctx.config.machine == ARM64EC) {
auxIat.push_back(make<NullChunk>(ctx.config.wordsize));
+ auxIatCopy.push_back(make<NullChunk>(ctx.config.wordsize));
+ }
for (int i = 0, e = syms.size(); i < e; ++i)
syms[i]->setLocation(addresses[base + i]);
diff --git a/lld/COFF/DLL.h b/lld/COFF/DLL.h
index 48b0f17ca62aeb..afb46f22ec9e14 100644
--- a/lld/COFF/DLL.h
+++ b/lld/COFF/DLL.h
@@ -32,6 +32,7 @@ class IdataContents {
std::vector<Chunk *> hints;
std::vector<Chunk *> dllNames;
std::vector<Chunk *> auxIat;
+ std::vector<Chunk *> auxIatCopy;
};
// Windows-specific.
diff --git a/lld/COFF/Driver.cpp b/lld/COFF/Driver.cpp
index 9994639d5d364a..1b94f10acf80e5 100644
--- a/lld/COFF/Driver.cpp
+++ b/lld/COFF/Driver.cpp
@@ -2448,6 +2448,7 @@ void LinkerDriver::linkerMain(ArrayRef<const char *> argsArr) {
ctx.symtab.addAbsolute("__arm64x_redirection_metadata", 0);
ctx.symtab.addAbsolute("__arm64x_redirection_metadata_count", 0);
ctx.symtab.addAbsolute("__hybrid_auxiliary_iat", 0);
+ ctx.symtab.addAbsolute("__hybrid_auxiliary_iat_copy", 0);
ctx.symtab.addAbsolute("__hybrid_code_map", 0);
ctx.symtab.addAbsolute("__hybrid_code_map_count", 0);
ctx.symtab.addAbsolute("__x64_code_ranges_to_entry_points", 0);
diff --git a/lld/COFF/InputFiles.cpp b/lld/COFF/InputFiles.cpp
index 94ad7f3ceb306a..d9184b04735e83 100644
--- a/lld/COFF/InputFiles.cpp
+++ b/lld/COFF/InputFiles.cpp
@@ -1091,6 +1091,12 @@ void ImportFile::parse() {
}
if (!impECSym)
return;
+
+ StringRef auxImpCopyName = saver().save("__auximpcopy_" + name);
+ auxImpCopySym =
+ ctx.symtab.addImportData(auxImpCopyName, this, auxCopyLocation);
+ if (!auxImpCopySym)
+ return;
}
// If this was a duplicate, we logged an error but may continue;
// in this case, impSym is nullptr.
diff --git a/lld/COFF/InputFiles.h b/lld/COFF/InputFiles.h
index acf221d85ae8f2..5fa93f57ef9e3a 100644
--- a/lld/COFF/InputFiles.h
+++ b/lld/COFF/InputFiles.h
@@ -362,10 +362,12 @@ class ImportFile : public InputFile {
const coff_import_header *hdr;
Chunk *location = nullptr;
- // Auxiliary IAT symbol and chunk on ARM64EC.
+ // Auxiliary IAT symbols and chunks on ARM64EC.
DefinedImportData *impECSym = nullptr;
Chunk *auxLocation = nullptr;
Symbol *auxThunkSym = nullptr;
+ DefinedImportData *auxImpCopySym = nullptr;
+ Chunk *auxCopyLocation = nullptr;
// We want to eliminate dllimported symbols if no one actually refers to them.
// These "Live" bits are used to keep track of which import library members
diff --git a/lld/COFF/Writer.cpp b/lld/COFF/Writer.cpp
index 216db652c10aa8..c2765453aa964e 100644
--- a/lld/COFF/Writer.cpp
+++ b/lld/COFF/Writer.cpp
@@ -916,6 +916,8 @@ void Writer::addSyntheticIdata() {
add(".idata$7", idata.dllNames);
if (!idata.auxIat.empty())
add(".idata$9", idata.auxIat);
+ if (!idata.auxIatCopy.empty())
+ add(".idata$a", idata.auxIatCopy);
}
void Writer::appendECImportTables() {
@@ -2279,6 +2281,11 @@ void Writer::setECSymbols() {
replaceSymbol<DefinedSynthetic>(iatSym, "__hybrid_auxiliary_iat",
idata.auxIat.empty() ? nullptr
: idata.auxIat.front());
+
+ Symbol *iatCopySym = ctx.symtab.findUnderscore("__hybrid_auxiliary_iat_copy");
+ replaceSymbol<DefinedSynthetic>(
+ iatCopySym, "__hybrid_auxiliary_iat_copy",
+ idata.auxIatCopy.empty() ? nullptr : idata.auxIatCopy.front());
}
// Write section contents to a mmap'ed file.
diff --git a/lld/test/COFF/Inputs/loadconfig-arm64ec.s b/lld/test/COFF/Inputs/loadconfig-arm64ec.s
index 8d59d29bbd0212..cb79b5c257e6ec 100644
--- a/lld/test/COFF/Inputs/loadconfig-arm64ec.s
+++ b/lld/test/COFF/Inputs/loadconfig-arm64ec.s
@@ -84,7 +84,7 @@ __chpe_metadata:
.rva __arm64x_extra_rfe_table
.word __arm64x_extra_rfe_table_size
.rva __os_arm64x_dispatch_fptr
- .word 0 // __hybrid_auxiliary_iat_copy
+ .rva __hybrid_auxiliary_iat_copy
.rva __os_arm64x_helper0
.rva __os_arm64x_helper1
.rva __os_arm64x_helper2
diff --git a/lld/test/COFF/arm64ec-import.test b/lld/test/COFF/arm64ec-import.test
index e403daa41f368c..92d7f5517bd426 100644
--- a/lld/test/COFF/arm64ec-import.test
+++ b/lld/test/COFF/arm64ec-import.test
@@ -4,6 +4,7 @@ RUN: split-file %s %t.dir && cd %t.dir
RUN: llvm-mc -filetype=obj -triple=arm64ec-windows test.s -o test.obj
RUN: llvm-mc -filetype=obj -triple=arm64ec-windows icall.s -o icall.obj
RUN: llvm-mc -filetype=obj -triple=arm64ec-windows hybmp.s -o hybmp.obj
+RUN: llvm-mc -filetype=obj -triple=x86_64-windows test.s -o test-x86_64.obj
RUN: llvm-mc -filetype=obj -triple=arm64ec-windows %S/Inputs/loadconfig-arm64ec.s -o loadconfig-arm64ec.obj
RUN: llvm-lib -machine:arm64ec -def:test.def -out:test-arm64ec.lib
RUN: llvm-lib -machine:arm64ec -def:test2.def -out:test2-arm64ec.lib
@@ -17,11 +18,16 @@ Link using x86_64 import library:
RUN: lld-link -machine:arm64ec -dll -noentry -out:out2.dll loadconfig-arm64ec.obj icall.obj hybmp.obj \
RUN: test.obj test-x86_64.lib test2-arm64ec.lib
+Link using x86_64 object file:
+RUN: lld-link -machine:arm64ec -dll -noentry -out:out3.dll loadconfig-arm64ec.obj icall.obj hybmp.obj \
+RUN: test-x86_64.obj test-arm64ec.lib test2-arm64ec.lib
+
RUN: llvm-readobj --coff-imports out.dll | FileCheck --check-prefix=IMPORTS %s
RUN: llvm-readobj --coff-imports out2.dll | FileCheck --check-prefix=IMPORTS %s
+RUN: llvm-readobj --coff-imports out3.dll | FileCheck -check-prefix=IMPORTS %s
IMPORTS: Import {
IMPORTS-NEXT: Name: test.dll
-IMPORTS-NEXT: ImportLookupTableRVA:
+IMPORTS-NEXT: ImportLookupTableRVA: 0x4230
IMPORTS-NEXT: ImportAddressTableRVA: 0x3000
IMPORTS-NEXT: Symbol: data (0)
IMPORTS-NEXT: Symbol: func (0)
@@ -36,6 +42,7 @@ IMPORTS-NEXT: }
RUN: llvm-objdump -d out.dll | FileCheck --check-prefix=DISASM %s
RUN: llvm-objdump -d out2.dll | FileCheck --check-prefix=DISASM %s
+RUN: llvm-objdump -d out3.dll | FileCheck -check-prefix=DISASM %s
DISASM: 180001000: 52800000 mov w0, #0x0 // =0
DISASM-NEXT: 180001004: d65f03c0 ret
@@ -72,23 +79,54 @@ RUN: llvm-readobj --hex-dump=.test out2.dll | FileCheck --check-prefix=TESTSEC %
TESTSEC: 0x180007000 08500000 00300000 10500000 20500000
TESTSEC-NEXT: 0x180007010 08300000 00500000 10300000 20300000
TESTSEC-NEXT: 0x180007020 14100000 28100000 00200000 08100000
-TESTSEC-NEXT: 0x180007030 3c100000
+TESTSEC-NEXT: 0x180007030 3c100000 a0420000
+
+RUN: llvm-readobj --hex-dump=.test out3.dll | FileCheck -check-prefix=TESTSEC-X64 %s
+TESTSEC-X64: 0x180007000 08300000 00300000 10300000 20300000
+TESTSEC-X64-NEXT: 0x180007010 08300000 00500000 10300000 20300000
+TESTSEC-X64-NEXT: 0x180007020 14100000 28100000 00200000 08100000
+TESTSEC-X64-NEXT: 0x180007030 3c100000 a0420000
RUN: llvm-readobj --headers out.dll | FileCheck -check-prefix=HEADERS %s
+RUN: llvm-readobj --headers out2.dll | FileCheck -check-prefix=HEADERS %s
+RUN: llvm-readobj --headers out3.dll | FileCheck -check-prefix=HEADERS %s
HEADERS: LoadConfigTableRVA: 0x4010
HEADERS: IATRVA: 0x3000
HEADERS: IATSize: 0x1000
RUN: llvm-readobj --coff-load-config out.dll | FileCheck -check-prefix=LOADCONFIG %s
+RUN: llvm-readobj --coff-load-config out2.dll | FileCheck -check-prefix=LOADCONFIG %s
+RUN: llvm-readobj --coff-load-config out3.dll | FileCheck -check-prefix=LOADCONFIG %s
LOADCONFIG: AuxiliaryIAT: 0x5000
+LOADCONFIG: AuxiliaryIATCopy: 0x4298
RUN: llvm-readobj --hex-dump=.rdata out.dll | FileCheck -check-prefix=RDATA %s
+RUN: llvm-readobj --hex-dump=.rdata out2.dll | FileCheck -check-prefix=RDATA %s
+RUN: llvm-readobj --hex-dump=.rdata out3.dll | FileCheck -check-prefix=RDATA %s
+RDATA: 0x180004290 2e646c6c 00000000 00000000 00000000
+RDATA-NEXT: 0x1800042a0 14100080 01000000 28100080 01000000
+RDATA-NEXT: 0x1800042b0 00000000 00000000 48100080 01000000
+RDATA-NEXT: 0x1800042c0 00000000 00000000 00000000 00000000
RDATA: 0x180005000 00000000 00000000 14100080 01000000
RDATA-NEXT: 0x180005010 28100080 01000000 00000000 00000000
RDATA-NEXT: 0x180005020 48100080 01000000 00000000 00000000
RUN: llvm-readobj --coff-basereloc out.dll | FileCheck -check-prefix=BASERELOC %s
+RUN: llvm-readobj --coff-basereloc out2.dll | FileCheck -check-prefix=BASERELOC %s
+RUN: llvm-readobj --coff-basereloc out3.dll | FileCheck -check-prefix=BASERELOC %s
BASERELOC: BaseReloc [
+Aux IAT copy:
+BASERELOC: Address: 0x42A0
+BASERELOC-NEXT: }
+BASERELOC-NEXT: Entry {
+BASERELOC-NEXT: Type: DIR64
+BASERELOC-NEXT: Address: 0x42A8
+BASERELOC-NEXT: }
+BASERELOC-NEXT: Entry {
+BASERELOC-NEXT: Type: DIR64
+BASERELOC-NEXT: Address: 0x42B8
+BASERELOC-NEXT: }
+Aux IAT:
BASERELOC-NOT: Address: 0x5000
BASERELOC: Address: 0x5008
BASERELOC-NEXT: }
@@ -119,6 +157,7 @@ arm64ec_data_sym:
.rva func
.rva "#func"
.rva "#t2func"
+ .rva __auximpcopy_func
#--- icall.s
.text
|
mstorsjo
approved these changes
Sep 16, 2024
In addition to the auxiliary IAT, ARM64EC modules also contain a copy of it. At runtime, the auxiliary IAT is filled with the addresses of actual ARM64EC functions when possible. If patching is detected, the OS may use the IAT copy to revert the auxiliary IAT, ensuring that the call checker is used for calls to imported functions.
cjacek
force-pushed
the
arm64ec-iat-copy
branch
from
2185a29 to
2640ed9
Compare
tmsri
pushed a commit
to tmsri/llvm-project
that referenced
this pull request
Sep 19, 2024
In addition to the auxiliary IAT, ARM64EC modules also contain a copy of it. At runtime, the auxiliary IAT is filled with the addresses of actual ARM64EC functions when possible. If patching is detected, the OS may use the IAT copy to revert the auxiliary IAT, ensuring that the call checker is used for calls to imported functions.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
In addition to the auxiliary IAT, ARM64EC modules also contain a copy of it. At runtime, the auxiliary IAT is filled with the addresses of actual ARM64EC functions when possible. If patching is detected, the OS may use the IAT copy to revert the auxiliary IAT, ensuring that the call checker is used for calls to imported functions.