-
Notifications
You must be signed in to change notification settings - Fork 14.8k
[lldb] Support parsing data symbols from the Wasm name section #153494
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.
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
Conversation
e68b6d3
to
b12ff18
Compare
@llvm/pr-subscribers-lldb Author: Jonas Devlieghere (JDevlieghere) ChangesThis PR adds support for parsing data symbols from the WebAssembly name section. Full diff: https://github.com/llvm/llvm-project/pull/153494.diff 3 Files Affected:
diff --git a/lldb/source/Plugins/ObjectFile/wasm/ObjectFileWasm.cpp b/lldb/source/Plugins/ObjectFile/wasm/ObjectFileWasm.cpp
index 919cc21c32ffd..b3144f28f4913 100644
--- a/lldb/source/Plugins/ObjectFile/wasm/ObjectFileWasm.cpp
+++ b/lldb/source/Plugins/ObjectFile/wasm/ObjectFileWasm.cpp
@@ -251,11 +251,11 @@ bool ObjectFileWasm::ParseHeader() {
static llvm::Expected<std::vector<AddressRange>>
ParseFunctions(SectionSP code_section_sp) {
- DataExtractor code_section_data;
- code_section_sp->GetSectionData(code_section_data);
+ DataExtractor data;
+ code_section_sp->GetSectionData(data);
lldb::offset_t offset = 0;
- const uint64_t function_count = code_section_data.GetULEB128(&offset);
+ const uint64_t function_count = data.GetULEB128(&offset);
if (function_count > std::numeric_limits<uint32_t>::max())
return llvm::createStringError("function count overflows uint32_t");
@@ -263,7 +263,7 @@ ParseFunctions(SectionSP code_section_sp) {
functions.reserve(function_count);
for (uint32_t i = 0; i < function_count; ++i) {
- const uint64_t function_size = code_section_data.GetULEB128(&offset);
+ const uint64_t function_size = data.GetULEB128(&offset);
if (function_size > std::numeric_limits<uint32_t>::max())
return llvm::createStringError("function size overflows uint32_t");
// llvm-objdump considers the ULEB with the function size to be part of the
@@ -281,9 +281,45 @@ ParseFunctions(SectionSP code_section_sp) {
return functions;
}
+static llvm::Expected<std::vector<AddressRange>>
+ParseData(SectionSP data_section_sp) {
+ DataExtractor data;
+ data_section_sp->GetSectionData(data);
+
+ lldb::offset_t offset = 0;
+
+ const uint64_t segment_count = data.GetULEB128(&offset);
+ if (segment_count > std::numeric_limits<uint32_t>::max())
+ return llvm::createStringError("segment count overflows uint32_t");
+
+ std::vector<AddressRange> segments;
+ segments.reserve(segment_count);
+
+ for (uint32_t i = 0; i < segment_count; ++i) {
+ const uint64_t flags = data.GetULEB128(&offset);
+ if (flags > std::numeric_limits<uint32_t>::max())
+ return llvm::createStringError("segment flags overflows uint32_t");
+
+ const uint64_t segment_size = data.GetULEB128(&offset);
+ if (flags > std::numeric_limits<uint32_t>::max())
+ return llvm::createStringError("segment size overflows uint32_t");
+
+ segments.emplace_back(data_section_sp, offset, segment_size);
+
+ std::optional<lldb::offset_t> next_offset =
+ llvm::checkedAddUnsigned(offset, segment_size);
+ if (!next_offset)
+ return llvm::createStringError("segment offset overflows uint64_t");
+ offset = *next_offset;
+ }
+
+ return segments;
+}
+
static llvm::Expected<std::vector<Symbol>>
ParseNames(SectionSP name_section_sp,
- const std::vector<AddressRange> &functions) {
+ const std::vector<AddressRange> &function_ranges,
+ const std::vector<AddressRange> &segment_ranges) {
DataExtractor name_section_data;
name_section_sp->GetSectionData(name_section_data);
@@ -305,17 +341,34 @@ ParseNames(SectionSP name_section_sp,
for (uint64_t i = 0; c && i < count; ++i) {
const uint64_t idx = data.getULEB128(c);
const std::optional<std::string> name = GetWasmString(data, c);
- if (!name || idx >= functions.size())
+ if (!name || idx >= function_ranges.size())
continue;
symbols.emplace_back(
symbols.size(), Mangled(*name), lldb::eSymbolTypeCode,
/*external=*/false, /*is_debug=*/false, /*is_trampoline=*/false,
- /*is_artificial=*/false, functions[idx],
+ /*is_artificial=*/false, function_ranges[idx],
/*size_is_valid=*/true, /*contains_linker_annotations=*/false,
/*flags=*/0);
}
} break;
- case llvm::wasm::WASM_NAMES_DATA_SEGMENT:
+ case llvm::wasm::WASM_NAMES_DATA_SEGMENT: {
+ const uint64_t count = data.getULEB128(c);
+ if (count > std::numeric_limits<uint32_t>::max())
+ return llvm::createStringError("data count overflows uint32_t");
+ for (uint64_t i = 0; c && i < count; ++i) {
+ const uint64_t idx = data.getULEB128(c);
+ const std::optional<std::string> name = GetWasmString(data, c);
+ if (!name || idx >= segment_ranges.size())
+ continue;
+ symbols.emplace_back(
+ symbols.size(), Mangled(*name), lldb::eSymbolTypeData,
+ /*external=*/false, /*is_debug=*/false, /*is_trampoline=*/false,
+ /*is_artificial=*/false, segment_ranges[idx],
+ /*size_is_valid=*/true, /*contains_linker_annotations=*/false,
+ /*flags=*/0);
+ }
+
+ } break;
case llvm::wasm::WASM_NAMES_GLOBAL:
case llvm::wasm::WASM_NAMES_LOCAL:
default:
@@ -336,21 +389,35 @@ void ObjectFileWasm::ParseSymtab(Symtab &symtab) {
assert(m_sections_up && "sections must be parsed");
Log *log = GetLog(LLDBLog::Object);
- // The name section contains names and indexes. First parse the functions from
- // the code section so we can access them by their index.
- SectionSP code_section_sp =
- m_sections_up->FindSectionByType(lldb::eSectionTypeCode, false);
- if (!code_section_sp) {
- LLDB_LOG(log, "Failed to parse Wasm symbol table: no functions section");
- return;
+ // The name section contains names and indexes. First parse the data from the
+ // relevant sections so we can access it by its index.
+ std::vector<AddressRange> function_ranges;
+ std::vector<AddressRange> segment_ranges;
+
+ // Parse the code section.
+ if (SectionSP code_section_sp =
+ m_sections_up->FindSectionByType(lldb::eSectionTypeCode, false)) {
+ llvm::Expected<std::vector<AddressRange>> functions =
+ ParseFunctions(code_section_sp);
+ if (!functions) {
+ LLDB_LOG_ERROR(log, functions.takeError(),
+ "Failed to parse Wasm code section: {0}");
+ return;
+ }
+ function_ranges = *functions;
}
- llvm::Expected<std::vector<AddressRange>> functions =
- ParseFunctions(code_section_sp);
- if (!functions) {
- LLDB_LOG_ERROR(log, functions.takeError(),
- "Failed to parse Wasm functions: {0}");
- return;
+ // Parse the data section.
+ if (SectionSP data_section_sp =
+ m_sections_up->FindSectionByType(lldb::eSectionTypeData, false)) {
+ llvm::Expected<std::vector<AddressRange>> segments =
+ ParseData(data_section_sp);
+ if (!segments) {
+ LLDB_LOG_ERROR(log, segments.takeError(),
+ "Failed to parse Wasm data section: {0}");
+ return;
+ }
+ segment_ranges = *segments;
}
// Parse the name section.
@@ -362,7 +429,7 @@ void ObjectFileWasm::ParseSymtab(Symtab &symtab) {
}
llvm::Expected<std::vector<Symbol>> symbols =
- ParseNames(name_section_sp, *functions);
+ ParseNames(name_section_sp, function_ranges, segment_ranges);
if (!symbols) {
LLDB_LOG_ERROR(log, symbols.takeError(), "Failed to parse Wasm names: {0}");
return;
@@ -408,6 +475,9 @@ void ObjectFileWasm::CreateSections(SectionList &unified_section_list) {
// For this reason Section::GetFileAddress() must return zero for the
// Code section.
vm_addr = 0;
+ } else if (llvm::wasm::WASM_SEC_DATA == sect_info.id) {
+ section_type = eSectionTypeData;
+ section_name = ConstString("data");
} else {
section_type = GetSectionTypeFromName(sect_info.name.GetStringRef());
if (section_type == eSectionTypeOther)
diff --git a/lldb/test/Shell/Symtab/Inputs/simple.wasm.yaml b/lldb/test/Shell/Symtab/Inputs/simple.wasm.yaml
index 165bb53662f40..67b04aa3cf81c 100644
--- a/lldb/test/Shell/Symtab/Inputs/simple.wasm.yaml
+++ b/lldb/test/Shell/Symtab/Inputs/simple.wasm.yaml
@@ -1,3 +1,15 @@
+# clang -target wasm32 -nostdlib -Wl,--no-entry -Wl,--export-all -O0 -g -o simple.wasm simple.c
+# char* str = "data str";
+#
+# int add(int a, int b) {
+# return a + b;
+# }
+#
+# int main() {
+# int i = 1;
+# int j = 2;
+# return add(i, j);
+# }
--- !WASM
FileHeader:
Version: 0x1
@@ -37,13 +49,13 @@ Sections:
Mutable: true
InitExpr:
Opcode: I32_CONST
- Value: 66560
+ Value: 66576
- Index: 1
Type: I32
Mutable: false
InitExpr:
Opcode: I32_CONST
- Value: 1024
+ Value: 1036
- Index: 2
Type: I32
Mutable: false
@@ -55,44 +67,50 @@ Sections:
Mutable: false
InitExpr:
Opcode: I32_CONST
- Value: 1024
+ Value: 1040
- Index: 4
Type: I32
Mutable: false
InitExpr:
Opcode: I32_CONST
- Value: 66560
+ Value: 1040
- Index: 5
Type: I32
Mutable: false
InitExpr:
Opcode: I32_CONST
- Value: 1024
+ Value: 66576
- Index: 6
Type: I32
Mutable: false
InitExpr:
Opcode: I32_CONST
- Value: 66560
+ Value: 1024
- Index: 7
Type: I32
Mutable: false
InitExpr:
Opcode: I32_CONST
- Value: 131072
+ Value: 66576
- Index: 8
Type: I32
Mutable: false
InitExpr:
Opcode: I32_CONST
- Value: 0
+ Value: 131072
- Index: 9
Type: I32
Mutable: false
InitExpr:
Opcode: I32_CONST
- Value: 1
+ Value: 0
- Index: 10
+ Type: I32
+ Mutable: false
+ InitExpr:
+ Opcode: I32_CONST
+ Value: 1
+ - Index: 11
Type: I32
Mutable: false
InitExpr:
@@ -115,6 +133,9 @@ Sections:
- Name: main
Kind: FUNCTION
Index: 3
+ - Name: str
+ Kind: GLOBAL
+ Index: 1
- Name: __main_void
Kind: FUNCTION
Index: 2
@@ -123,34 +144,34 @@ Sections:
Index: 0
- Name: __dso_handle
Kind: GLOBAL
- Index: 1
+ Index: 2
- Name: __data_end
Kind: GLOBAL
- Index: 2
+ Index: 3
- Name: __stack_low
Kind: GLOBAL
- Index: 3
+ Index: 4
- Name: __stack_high
Kind: GLOBAL
- Index: 4
+ Index: 5
- Name: __global_base
Kind: GLOBAL
- Index: 5
+ Index: 6
- Name: __heap_base
Kind: GLOBAL
- Index: 6
+ Index: 7
- Name: __heap_end
Kind: GLOBAL
- Index: 7
+ Index: 8
- Name: __memory_base
Kind: GLOBAL
- Index: 8
+ Index: 9
- Name: __table_base
Kind: GLOBAL
- Index: 9
+ Index: 10
- Name: __wasm_first_page_end
Kind: GLOBAL
- Index: 10
+ Index: 11
- Type: CODE
Functions:
- Index: 0
@@ -169,6 +190,35 @@ Sections:
- Index: 3
Locals: []
Body: 1082808080000F0B
+ - Type: DATA
+ Segments:
+ - SectionOffset: 7
+ InitFlags: 0
+ Offset:
+ Opcode: I32_CONST
+ Value: 1024
+ Content: '646174612073747200'
+ - SectionOffset: 22
+ InitFlags: 0
+ Offset:
+ Opcode: I32_CONST
+ Value: 1036
+ Content: '00040000'
+ - Type: CUSTOM
+ Name: .debug_abbrev
+ Payload: 011101250E1305030E10171B0E11015517000002340049133A0B3B0B02180000030101491300000421004913370B0000052400030E3E0B0B0B0000062400030E0B0B3E0B0000073400030E49133F193A0B3B0B02180000080F0049130000092E01110112064018030E3A0B3B0B271949133F1900000A05000218030E3A0B3B0B491300000B2E01110112064018030E3A0B3B0B49133F1900000C34000218030E3A0B3B0B4913000000
+ - Type: CUSTOM
+ Name: .debug_info
+ Payload: D100000004000000000004017F0000001D005E0000000000000016000000000000000000000002330000000101050300040000033F0000000446000000090005080000000601066B000000080707040000005E000000010105030C040000083F00000009050000002900000004ED00029F5A0000000103CD0000000A02910C690000000103CD0000000A029108670000000103CD000000000B2F0000004C00000004ED00009F0D0000000107CD0000000C029108140000000108CD0000000C029104120000000109CD000000000500000000050400
+ - Type: CUSTOM
+ Name: .debug_ranges
+ Payload: 050000002E0000002F0000007B0000000000000000000000
+ - Type: CUSTOM
+ Name: .debug_str
+ Payload: 696E74007374720063686172006D61696E006A0069002F55736572732F6A6F6E61732F7761736D2D6D6963726F2D72756E74696D652F70726F647563742D6D696E692F706C6174666F726D732F64617277696E2F6275696C64006164640073696D706C652E6300620061005F5F41525241595F53495A455F545950455F5F00636C616E672076657273696F6E2032322E302E306769742028676974406769746875622E636F6D3A6C6C766D2F6C6C766D2D70726F6A6563742E67697420363363633265333930646235376362633430306235313937373162373030356561623166633736612900
+ - Type: CUSTOM
+ Name: .debug_line
+ Payload: 62000000040020000000010101FB0E0D0001010101000000010000010073696D706C652E6300000000000005020500000014050A0A08AD050E0658050C5805032002020001010005022F0000001805070A08BB75050E7505110658050A58050382020F000101
- Type: CUSTOM
Name: name
FunctionNames:
@@ -183,8 +233,17 @@ Sections:
GlobalNames:
- Index: 0
Name: __stack_pointer
+ DataSegmentNames:
+ - Index: 0
+ Name: .rodata
+ - Index: 1
+ Name: .data
- Type: CUSTOM
+ HeaderSecSizeEncodingLen: 2
Name: producers
+ Languages:
+ - Name: C11
+ Version: ''
Tools:
- Name: clang
Version: '22.0.0git'
diff --git a/lldb/test/Shell/Symtab/symtab-wasm.test b/lldb/test/Shell/Symtab/symtab-wasm.test
index fc185cd81a0ec..5374b0c2f2892 100644
--- a/lldb/test/Shell/Symtab/symtab-wasm.test
+++ b/lldb/test/Shell/Symtab/symtab-wasm.test
@@ -1,7 +1,9 @@
# RUN: yaml2obj %S/Inputs/simple.wasm.yaml -o %t.wasm
# RUN: %lldb %t.wasm -o 'image dump symtab'
-# CHECK: Code 0x0000000000000002 {{.*}} __wasm_call_ctors
-# CHECK: Code 0x0000000000000005 {{.*}} add
-# CHECK: Code 0x000000000000002f {{.*}} __original_main
-# CHECK: Code 0x000000000000007c {{.*}} main
+# CHECK: Code 0x0000000000000002 0x0000000000000002 {{.*}} __wasm_call_ctors
+# CHECK: Code 0x0000000000000005 0x0000000000000029 {{.*}} add
+# CHECK: Code 0x000000000000002f 0x000000000000004c {{.*}} __original_main
+# CHECK: Code 0x000000000000007c 0x0000000000000009 {{.*}} main
+# CHECK: Data 0x000000000000022f 0x0000000000000041 {{.*}} .rodata
+# CHECK: Data 0x0000000000000270 0x0000000000000000 {{.*}} .data
|
This PR adds support for parsing data symbols from the WebAssembly name section.
b12ff18
to
2d0a3da
Compare
What I see is parsing the names of data segments, no test checks we can read a symbol that's in those segments. You added a global Or is a "data symbol" a symbol telling us where the data segment is, and that's where the "symbol symbols" will all be and later you'll add code to read those? |
@DavidSpickett Great question and that definitely deserved more explanation from my end. I was a bit rushed in getting the PR up to unlock Adrian who needs the data for the Swift Reflection metadata. In WebAssembly, the data section (
That said, I do think these should also be represented as segments in LLDB, and that's indeed something I have planned to do next. |
So a "data symbol" is nothing to do with symbols the user has defined, these are kind of internal symbols which are section name offsets into this one giant data segment. Add something to the PR description like "data symbols provide the names of the sub-sections of the wasm data segment", whatever the correct form of that is. Just wanted to understand that. Otherwise, I have no other comments. |
This is a continuation of llvm#153494. In a WebAssembly file, the "name" section contains names for the segments in the data section (WASM_NAMES_DATA_SEGMENT). We already parse these as sections, as with this PR, we also create sub-sections for the data segments.
…153494) This PR adds support for parsing the data symbols from the WebAssembly name section, which consists of a name and address range for the segments in the Wasm data section. Unlike other object file formats, Wasm has no symbols for referencing items within those segments (i.e. symbols the user has defined). (cherry picked from commit d0e40ff)
This PR adds support for parsing the data symbols from the WebAssembly name section, which consists of a name and address range for the segments in the Wasm data section. Unlike other object file formats, Wasm has no symbols for referencing items within those segments (i.e. symbols the user has defined).