[llvm-objdump][macho] Add support for ObjC relative method lists #84250
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
alx32
force-pushed
the
06_objdump_rel_sel
branch
from
d33f060 to
3b97012
Compare
|
@llvm/pr-subscribers-llvm-binary-utilities Author: None (alx32) ChangesFor Mach-O, ld64 supports the The code style was matched as close as possible to existing implementation of parsing non-delta method lists. Diff between llvm-objdump and XCode 15.1 otool: Full diff: https://github.com/llvm/llvm-project/pull/84250.diff 4 Files Affected:
diff --git a/llvm/test/tools/llvm-objdump/MachO/AArch64/Inputs/rel-method-lists-arm64.dylib b/llvm/test/tools/llvm-objdump/MachO/AArch64/Inputs/rel-method-lists-arm64.dylib
new file mode 100755
index 00000000000000..9b39fc98fb5d96
Binary files /dev/null and b/llvm/test/tools/llvm-objdump/MachO/AArch64/Inputs/rel-method-lists-arm64.dylib differ
diff --git a/llvm/test/tools/llvm-objdump/MachO/AArch64/Inputs/rel-method-lists-arm64_32.dylib b/llvm/test/tools/llvm-objdump/MachO/AArch64/Inputs/rel-method-lists-arm64_32.dylib
new file mode 100755
index 00000000000000..d3a339057abc34
Binary files /dev/null and b/llvm/test/tools/llvm-objdump/MachO/AArch64/Inputs/rel-method-lists-arm64_32.dylib differ
diff --git a/llvm/test/tools/llvm-objdump/MachO/AArch64/macho-relative-method-lists.test b/llvm/test/tools/llvm-objdump/MachO/AArch64/macho-relative-method-lists.test
new file mode 100644
index 00000000000000..672720a8b2938d
--- /dev/null
+++ b/llvm/test/tools/llvm-objdump/MachO/AArch64/macho-relative-method-lists.test
@@ -0,0 +1,57 @@
+RUN: llvm-objdump --macho --objc-meta-data %p/Inputs/rel-method-lists-arm64_32.dylib | FileCheck %s --check-prefix=CHK32
+RUN: llvm-otool -ov %p/Inputs/rel-method-lists-arm64_32.dylib | FileCheck %s --check-prefix=CHK32
+
+RUN: llvm-objdump --macho --objc-meta-data %p/Inputs/rel-method-lists-arm64.dylib | FileCheck %s --check-prefix=CHK64
+RUN: llvm-otool -ov %p/Inputs/rel-method-lists-arm64.dylib | FileCheck %s --check-prefix=CHK64
+
+CHK32: baseMethods 0x660 (struct method_list_t *)
+CHK32-NEXT: entsize 12 (relative)
+CHK32-NEXT: count 3
+CHK32-NEXT: name 0x144 (0x7ac) instance_method_00
+CHK32-NEXT: types 0x91 (0x6fd) v8@0:4
+CHK32-NEXT: imp 0xffffff18 (0x588) -[MyClass instance_method_00]
+CHK32-NEXT: name 0x13c (0x7b0) instance_method_01
+CHK32-NEXT: types 0x85 (0x6fd) v8@0:4
+CHK32-NEXT: imp 0xffffff28 (0x5a4) -[MyClass instance_method_01]
+CHK32-NEXT: name 0x134 (0x7b4) instance_method_02
+CHK32-NEXT: types 0x79 (0x6fd) v8@0:4
+CHK32-NEXT: imp 0xffffff38 (0x5c0) -[MyClass instance_method_02]
+
+CHK32: baseMethods 0x630 (struct method_list_t *)
+CHK32-NEXT: entsize 12 (relative)
+CHK32-NEXT: count 3
+CHK32-NEXT: name 0x180 (0x7b8) class_method_00
+CHK32-NEXT: types 0xc1 (0x6fd) v8@0:4
+CHK32-NEXT: imp 0xffffff9c (0x5dc) +[MyClass class_method_00]
+CHK32-NEXT: name 0x178 (0x7bc) class_method_01
+CHK32-NEXT: types 0xb5 (0x6fd) v8@0:4
+CHK32-NEXT: imp 0xffffffac (0x5f8) +[MyClass class_method_01]
+CHK32-NEXT: name 0x170 (0x7c0) class_method_02
+CHK32-NEXT: types 0xa9 (0x6fd) v8@0:4
+CHK32-NEXT: imp 0xffffffbc (0x614) +[MyClass class_method_02]
+
+CHK64: baseMethods 0x7d8 (struct method_list_t *)
+CHK64-NEXT: entsize 24
+CHK64-NEXT: count 3
+CHK64-NEXT: name 0x6a4 instance_method_00
+CHK64-NEXT: types 0x6dd v16@0:8
+CHK64-NEXT: imp -[MyClass instance_method_00]
+CHK64-NEXT: name 0x6b7 instance_method_01
+CHK64-NEXT: types 0x6dd v16@0:8
+CHK64-NEXT: imp -[MyClass instance_method_01]
+CHK64-NEXT: name 0x6ca instance_method_02
+CHK64-NEXT: types 0x6dd v16@0:8
+CHK64-NEXT: imp -[MyClass instance_method_02]
+
+CHK64: baseMethods 0x740 (struct method_list_t *)
+CHK64-NEXT: entsize 24
+CHK64-NEXT: count 3
+CHK64-NEXT: name 0x674 class_method_00
+CHK64-NEXT: types 0x6dd v16@0:8
+CHK64-NEXT: imp +[MyClass class_method_00]
+CHK64-NEXT: name 0x684 class_method_01
+CHK64-NEXT: types 0x6dd v16@0:8
+CHK64-NEXT: imp +[MyClass class_method_01]
+CHK64-NEXT: name 0x694 class_method_02
+CHK64-NEXT: types 0x6dd v16@0:8
+CHK64-NEXT: imp +[MyClass class_method_02]
diff --git a/llvm/tools/llvm-objdump/MachODump.cpp b/llvm/tools/llvm-objdump/MachODump.cpp
index 0e6935c0ac5895..21e21e2a1e3ec2 100644
--- a/llvm/tools/llvm-objdump/MachODump.cpp
+++ b/llvm/tools/llvm-objdump/MachODump.cpp
@@ -3519,6 +3519,12 @@ static const char *get_pointer_64(uint64_t Address, uint32_t &offset,
return nullptr;
}
+static const char *get_value_32(uint32_t Address, uint32_t &offset,
+ uint32_t &left, SectionRef &S,
+ DisassembleInfo *info, bool objc_only = false) {
+ return get_pointer_64(Address, offset, left, S, info, objc_only);
+}
+
static const char *get_pointer_32(uint32_t Address, uint32_t &offset,
uint32_t &left, SectionRef &S,
DisassembleInfo *info,
@@ -3661,6 +3667,10 @@ struct class_ro32_t {
#define RO_ROOT (1 << 1)
#define RO_HAS_CXX_STRUCTORS (1 << 2)
+/* Values for method_list{64,32,_delta}_t->entsize */
+#define ML_HAS_DELTAS (1 << 31)
+#define ML_ENTSIZE_MASK 0xFFFF
+
struct method_list64_t {
uint32_t entsize;
uint32_t count;
@@ -3673,6 +3683,12 @@ struct method_list32_t {
/* struct method32_t first; These structures follow inline */
};
+struct method_list_delta_t {
+ uint32_t entsize;
+ uint32_t count;
+ /* struct method_delta_t first; These structures follow inline */
+};
+
struct method64_t {
uint64_t name; /* SEL (64-bit pointer) */
uint64_t types; /* const char * (64-bit pointer) */
@@ -3685,6 +3701,12 @@ struct method32_t {
uint32_t imp; /* IMP (32-bit pointer) */
};
+struct method_delta_t {
+ int32_t name; /* SEL (32-bit delta) */
+ int32_t types; /* const char * (32-bit delta) */
+ int32_t imp; /* IMP (32-bit delta) */
+};
+
struct protocol_list64_t {
uint64_t count; /* uintptr_t (a 64-bit value) */
/* struct protocol64_t * list[0]; These pointers follow inline */
@@ -3974,6 +3996,11 @@ inline void swapStruct(struct method_list32_t &ml) {
sys::swapByteOrder(ml.count);
}
+inline void swapStruct(struct method_list_delta_t &ml) {
+ sys::swapByteOrder(ml.entsize);
+ sys::swapByteOrder(ml.count);
+}
+
inline void swapStruct(struct method64_t &m) {
sys::swapByteOrder(m.name);
sys::swapByteOrder(m.types);
@@ -3986,6 +4013,12 @@ inline void swapStruct(struct method32_t &m) {
sys::swapByteOrder(m.imp);
}
+inline void swapStruct(struct method_delta_t &m) {
+ sys::swapByteOrder(m.name);
+ sys::swapByteOrder(m.types);
+ sys::swapByteOrder(m.imp);
+}
+
inline void swapStruct(struct protocol_list64_t &pl) {
sys::swapByteOrder(pl.count);
}
@@ -4440,8 +4473,103 @@ static void print_layout_map32(uint32_t p, struct DisassembleInfo *info) {
print_layout_map(layout_map, left);
}
+// Return true if this is a delta method list, false otherwise
+static bool print_method_list_delta_t(uint64_t p, struct DisassembleInfo *info,
+ const char *indent,
+ uint32_t pointerBits) {
+ struct method_list_delta_t ml;
+ struct method_delta_t m;
+ const char *r, *name;
+ uint32_t offset, xoffset, left, i;
+ SectionRef S, xS;
+
+ r = get_pointer_32(p, offset, left, S, info);
+ if (r == nullptr)
+ return false;
+ memset(&ml, '\0', sizeof(struct method_list_delta_t));
+ if (left < sizeof(struct method_list_delta_t)) {
+ memcpy(&ml, r, left);
+ outs() << " (method_delta_t entends past the end of the section)\n";
+ } else
+ memcpy(&ml, r, sizeof(struct method_list_delta_t));
+ if (info->O->isLittleEndian() != sys::IsLittleEndianHost)
+ swapStruct(ml);
+ if ((ml.entsize & ML_HAS_DELTAS) == 0)
+ return false;
+
+ outs() << indent << "\t\t entsize " << (ml.entsize & ML_ENTSIZE_MASK)
+ << " (relative) \n";
+ outs() << indent << "\t\t count " << ml.count << "\n";
+
+ p += sizeof(struct method_list_delta_t);
+ offset += sizeof(struct method_delta_t);
+ for (i = 0; i < ml.count; i++) {
+ r = get_value_32(p, offset, left, S, info);
+ if (r == nullptr)
+ return true;
+ memset(&m, '\0', sizeof(struct method_delta_t));
+ if (left < sizeof(struct method_delta_t)) {
+ memcpy(&ml, r, left);
+ outs() << indent << " (method_t entends past the end of the section)\n";
+ } else
+ memcpy(&m, r, sizeof(struct method_delta_t));
+ if (info->O->isLittleEndian() != sys::IsLittleEndianHost)
+ swapStruct(m);
+
+ outs() << indent << "\t\t name " << format("0x%" PRIx32, m.name);
+ uint64_t relNameRefVA = p + offsetof(struct method_delta_t, name);
+ uint64_t absNameRefVA = relNameRefVA + m.name;
+ outs() << " (" << format("0x%" PRIx32, absNameRefVA) << ")";
+
+ // since this is a delta list, absNameRefVA is the address of the
+ // __objc_selrefs entry, so a pointer, not the actual name
+ const char *nameRefPtr =
+ get_pointer_32(absNameRefVA, xoffset, left, xS, info);
+ if (nameRefPtr) {
+ uint32_t pointerSize = pointerBits / CHAR_BIT;
+ if (left < pointerSize)
+ outs() << indent << " (nameRefPtr entends past the end of the section)";
+ else {
+ uint64_t nameVA = 0;
+ memcpy(&nameVA, nameRefPtr, pointerSize);
+ const char *name = get_pointer_32(nameVA, xoffset, left, xS, info);
+ if (name != nullptr)
+ outs() << format(" %.*s", left, name);
+ }
+ }
+ outs() << "\n";
+
+ outs() << indent << "\t\t types " << format("0x%" PRIx32, m.types);
+ uint64_t relTypesVA = p + offsetof(struct method_delta_t, types);
+ uint64_t absTypesVA = relTypesVA + m.types;
+ outs() << " (" << format("0x%" PRIx32, absTypesVA) << ")";
+ name = get_pointer_32(absTypesVA, xoffset, left, xS, info);
+ if (name != nullptr)
+ outs() << format(" %.*s", left, name);
+ outs() << "\n";
+
+ outs() << indent << "\t\t imp " << format("0x%" PRIx32, m.imp);
+ uint64_t relImpVA = p + offsetof(struct method_delta_t, imp);
+ uint64_t absImpVA = relImpVA + m.imp;
+ outs() << " (" << format("0x%" PRIx32, absImpVA) << ")";
+ name = GuessSymbolName(absImpVA, info->AddrMap);
+ if (name != nullptr)
+ outs() << " " << name;
+ outs() << "\n";
+
+ p += sizeof(struct method_delta_t);
+ offset += sizeof(struct method_delta_t);
+ }
+
+ return true;
+}
+
static void print_method_list64_t(uint64_t p, struct DisassembleInfo *info,
const char *indent) {
+ // Try to parse it as a delta list. If successful, just return
+ if (print_method_list_delta_t(p, info, indent, /*pointerBits=*/64))
+ return;
+
struct method_list64_t ml;
struct method64_t m;
const char *r;
@@ -4535,6 +4663,10 @@ static void print_method_list64_t(uint64_t p, struct DisassembleInfo *info,
static void print_method_list32_t(uint64_t p, struct DisassembleInfo *info,
const char *indent) {
+ // Try to parse it as a delta list. If successful, just return
+ if (print_method_list_delta_t(p, info, indent, /*pointerBits=*/32))
+ return;
+
struct method_list32_t ml;
struct method32_t m;
const char *r, *name;
|
alx32
force-pushed
the
06_objdump_rel_sel
branch
from
3b97012 to
54df022
Compare
alx32
force-pushed
the
06_objdump_rel_sel
branch
from
54df022 to
c99ad35
Compare
kyulee-com
reviewed
Mar 13, 2024
llvm/test/tools/llvm-objdump/MachO/AArch64/macho-relative-method-lists.test
Show resolved
Hide resolved
drodriguez
reviewed
Mar 13, 2024
alx32
force-pushed
the
06_objdump_rel_sel
branch
from
a660c89 to
6080be0
Compare
JDevlieghere
approved these changes
Mar 14, 2024
|
LGTM |
kyulee-com
approved these changes
Mar 14, 2024
drodriguez
approved these changes
Mar 14, 2024
|
Hi, this commit probably broke some bots: |
kyulee-com
added a commit
to kyulee-com/llvm-project
that referenced
this pull request
Mar 15, 2024
…sts (llvm#84250)" This reverts llvm#84250, commit aa61006. See build failures: https://lab.llvm.org/buildbot/#/builders/178/builds/7028 https://lab.llvm.org/buildbot/#/builders/182/builds/9282 https://lab.llvm.org/buildbot/#/builders/186/builds/15299 https://lab.llvm.org/buildbot/#/builders/187/builds/14564
kyulee-com
added a commit
that referenced
this pull request
Mar 15, 2024
…sts (#84250)" This reverts #84250, commit aa61006. See build failures: https://lab.llvm.org/buildbot/#/builders/178/builds/7028 https://lab.llvm.org/buildbot/#/builders/182/builds/9282 https://lab.llvm.org/buildbot/#/builders/186/builds/15299 https://lab.llvm.org/buildbot/#/builders/187/builds/14564
|
Follow-up with fix for armv7 host: #85477 |
kyulee-com
pushed a commit
that referenced
this pull request
Mar 16, 2024
) For Mach-O, ld64 supports the -fobjc-relative-method-lists flag which changes the format in which method lists are generated. The format uses delta encoding vs the original direct-pointer encoding. This change adds support to llvm-objdump and llvm-otool for decoding/dumping of method lists in the delta format. Previously, if a binary with this information format was passed to the tooling, it would output invalid information, trying to parse the delta lists as pointer lists. After this change, the tooling will output correct information if a binary in this format is encountered. The output format is closest feasible match to XCode 15.1's otool output. Tests are included for both 32bit and 64bit binaries. The code style was matched as close as possible to existing implementation of parsing non-delta method lists. Diff between llvm-objdump and XCode 15.1 otool:  Note: This is a retry of this PR: #84250 On the original PR, the armv7+armv8 builds were failing due to absolute offsets being different. Co-authored-by: Alex B <[email protected]>
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.
For Mach-O, ld64 supports the
-fobjc-relative-method-listsflag which changes the format in which method lists are generated. The format uses delta encoding vs the original direct-pointer encoding.This change adds support to
llvm-objdumpandllvm-otoolfor decoding/dumping of method lists in the delta format. Previously, if a binary with this information format was passed to the tooling, it would output invalid information, trying to parse the delta lists as pointer lists.After this change, the tooling will output correct information if a binary in this format is encountered.
The output format is closest feasible match to XCode 15.1's otool output. Tests are included for both 32bit and 64bit binaries.
The code style was matched as close as possible to existing implementation of parsing non-delta method lists.
Diff between llvm-objdump and XCode 15.1 otool: