2010-06-09 00:52:24 +08:00
|
|
|
//===-- Disassembler.cpp ----------------------------------------*- C++ -*-===//
|
|
|
|
//
|
|
|
|
// The LLVM Compiler Infrastructure
|
|
|
|
//
|
|
|
|
// This file is distributed under the University of Illinois Open Source
|
|
|
|
// License. See LICENSE.TXT for details.
|
|
|
|
//
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
|
|
|
|
#include "lldb/Core/Disassembler.h"
|
|
|
|
|
|
|
|
// C Includes
|
|
|
|
// C++ Includes
|
2016-03-03 08:51:40 +08:00
|
|
|
#include <cstdio>
|
|
|
|
#include <cstring>
|
|
|
|
#include <limits>
|
|
|
|
|
2010-06-09 00:52:24 +08:00
|
|
|
// Other libraries and framework includes
|
|
|
|
// Project includes
|
|
|
|
#include "lldb/lldb-private.h"
|
|
|
|
#include "lldb/Core/Error.h"
|
|
|
|
#include "lldb/Core/DataBufferHeap.h"
|
|
|
|
#include "lldb/Core/DataExtractor.h"
|
|
|
|
#include "lldb/Core/Debugger.h"
|
2011-04-06 02:46:00 +08:00
|
|
|
#include "lldb/Core/EmulateInstruction.h"
|
2010-06-09 00:52:24 +08:00
|
|
|
#include "lldb/Core/Module.h"
|
|
|
|
#include "lldb/Core/PluginManager.h"
|
2011-04-22 13:08:45 +08:00
|
|
|
#include "lldb/Core/RegularExpression.h"
|
2010-06-09 00:52:24 +08:00
|
|
|
#include "lldb/Core/Timer.h"
|
2012-08-23 01:17:09 +08:00
|
|
|
#include "lldb/Interpreter/OptionValue.h"
|
|
|
|
#include "lldb/Interpreter/OptionValueArray.h"
|
|
|
|
#include "lldb/Interpreter/OptionValueDictionary.h"
|
|
|
|
#include "lldb/Interpreter/OptionValueString.h"
|
|
|
|
#include "lldb/Interpreter/OptionValueUInt64.h"
|
<rdar://problem/11757916>
Make breakpoint setting by file and line much more efficient by only looking for inlined breakpoint locations if we are setting a breakpoint in anything but a source implementation file. Implementing this complex for a many reasons. Turns out that parsing compile units lazily had some issues with respect to how we need to do things with DWARF in .o files. So the fixes in the checkin for this makes these changes:
- Add a new setting called "target.inline-breakpoint-strategy" which can be set to "never", "always", or "headers". "never" will never try and set any inlined breakpoints (fastest). "always" always looks for inlined breakpoint locations (slowest, but most accurate). "headers", which is the default setting, will only look for inlined breakpoint locations if the breakpoint is set in what are consudered to be header files, which is realy defined as "not in an implementation source file".
- modify the breakpoint setting by file and line to check the current "target.inline-breakpoint-strategy" setting and act accordingly
- Modify compile units to be able to get their language and other info lazily. This allows us to create compile units from the debug map and not have to fill all of the details in, and then lazily discover this information as we go on debuggging. This is needed to avoid parsing all .o files when setting breakpoints in implementation only files (no inlines). Otherwise we would need to parse the .o file, the object file (mach-o in our case) and the symbol file (DWARF in the object file) just to see what the compile unit was.
- modify the "SymbolFileDWARFDebugMap" to subclass lldb_private::Module so that the virtual "GetObjectFile()" and "GetSymbolVendor()" functions can be intercepted when the .o file contenst are later lazilly needed. Prior to this fix, when we first instantiated the "SymbolFileDWARFDebugMap" class, we would also make modules, object files and symbol files for every .o file in the debug map because we needed to fix up the sections in the .o files with information that is in the executable debug map. Now we lazily do this in the DebugMapModule::GetObjectFile()
Cleaned up header includes a bit as well.
llvm-svn: 162860
2012-08-30 05:13:06 +08:00
|
|
|
#include "lldb/Symbol/Function.h"
|
2010-06-09 00:52:24 +08:00
|
|
|
#include "lldb/Symbol/ObjectFile.h"
|
|
|
|
#include "lldb/Target/ExecutionContext.h"
|
|
|
|
#include "lldb/Target/Process.h"
|
2013-12-06 09:12:00 +08:00
|
|
|
#include "lldb/Target/SectionLoadList.h"
|
2013-11-04 17:33:30 +08:00
|
|
|
#include "lldb/Target/StackFrame.h"
|
2010-06-09 00:52:24 +08:00
|
|
|
#include "lldb/Target/Target.h"
|
|
|
|
|
|
|
|
#define DEFAULT_DISASM_BYTE_SIZE 32
|
|
|
|
|
|
|
|
using namespace lldb;
|
|
|
|
using namespace lldb_private;
|
|
|
|
|
2012-08-02 02:50:59 +08:00
|
|
|
DisassemblerSP
|
2013-03-02 08:26:47 +08:00
|
|
|
Disassembler::FindPlugin (const ArchSpec &arch, const char *flavor, const char *plugin_name)
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
|
|
|
Timer scoped_timer (__PRETTY_FUNCTION__,
|
2011-03-26 02:03:16 +08:00
|
|
|
"Disassembler::FindPlugin (arch = %s, plugin_name = %s)",
|
|
|
|
arch.GetArchitectureName(),
|
|
|
|
plugin_name);
|
2010-06-09 00:52:24 +08:00
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
DisassemblerCreateInstance create_callback = nullptr;
|
2011-03-26 02:03:16 +08:00
|
|
|
|
|
|
|
if (plugin_name)
|
|
|
|
{
|
2013-05-11 05:47:16 +08:00
|
|
|
ConstString const_plugin_name (plugin_name);
|
|
|
|
create_callback = PluginManager::GetDisassemblerCreateCallbackForPluginName (const_plugin_name);
|
2011-03-26 02:03:16 +08:00
|
|
|
if (create_callback)
|
|
|
|
{
|
2013-03-02 08:26:47 +08:00
|
|
|
DisassemblerSP disassembler_sp(create_callback(arch, flavor));
|
2011-03-26 02:03:16 +08:00
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
if (disassembler_sp)
|
2012-08-02 02:50:59 +08:00
|
|
|
return disassembler_sp;
|
2011-03-26 02:03:16 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
else
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2016-03-03 08:51:40 +08:00
|
|
|
for (uint32_t idx = 0; (create_callback = PluginManager::GetDisassemblerCreateCallbackAtIndex(idx)) != nullptr; ++idx)
|
2011-03-26 02:03:16 +08:00
|
|
|
{
|
2013-03-02 08:26:47 +08:00
|
|
|
DisassemblerSP disassembler_sp(create_callback(arch, flavor));
|
2010-06-09 00:52:24 +08:00
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
if (disassembler_sp)
|
2012-08-02 02:50:59 +08:00
|
|
|
return disassembler_sp;
|
2011-03-26 02:03:16 +08:00
|
|
|
}
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
2012-08-02 02:50:59 +08:00
|
|
|
return DisassemblerSP();
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
|
|
|
|
2013-03-02 08:26:47 +08:00
|
|
|
DisassemblerSP
|
|
|
|
Disassembler::FindPluginForTarget(const TargetSP target_sp, const ArchSpec &arch, const char *flavor, const char *plugin_name)
|
|
|
|
{
|
2016-03-03 08:51:40 +08:00
|
|
|
if (target_sp && flavor == nullptr)
|
2013-03-02 08:26:47 +08:00
|
|
|
{
|
|
|
|
// FIXME - we don't have the mechanism in place to do per-architecture settings. But since we know that for now
|
|
|
|
// we only support flavors on x86 & x86_64,
|
|
|
|
if (arch.GetTriple().getArch() == llvm::Triple::x86
|
|
|
|
|| arch.GetTriple().getArch() == llvm::Triple::x86_64)
|
|
|
|
flavor = target_sp->GetDisassemblyFlavor();
|
|
|
|
}
|
|
|
|
return FindPlugin(arch, flavor, plugin_name);
|
|
|
|
}
|
|
|
|
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
static void
|
|
|
|
ResolveAddress (const ExecutionContext &exe_ctx,
|
|
|
|
const Address &addr,
|
|
|
|
Address &resolved_addr)
|
|
|
|
{
|
|
|
|
if (!addr.IsSectionOffset())
|
|
|
|
{
|
|
|
|
// If we weren't passed in a section offset address range,
|
|
|
|
// try and resolve it to something
|
2011-09-22 12:58:26 +08:00
|
|
|
Target *target = exe_ctx.GetTargetPtr();
|
|
|
|
if (target)
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
{
|
2011-09-22 12:58:26 +08:00
|
|
|
if (target->GetSectionLoadList().IsEmpty())
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
{
|
2011-09-22 12:58:26 +08:00
|
|
|
target->GetImages().ResolveFileAddress (addr.GetOffset(), resolved_addr);
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2011-09-22 12:58:26 +08:00
|
|
|
target->GetSectionLoadList().ResolveLoadAddress (addr.GetOffset(), resolved_addr);
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
}
|
|
|
|
// We weren't able to resolve the address, just treat it as a
|
|
|
|
// raw address
|
|
|
|
if (resolved_addr.IsValid())
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
resolved_addr = addr;
|
|
|
|
}
|
2010-07-01 07:03:03 +08:00
|
|
|
|
|
|
|
size_t
|
2016-03-03 08:51:40 +08:00
|
|
|
Disassembler::Disassemble(Debugger &debugger,
|
|
|
|
const ArchSpec &arch,
|
|
|
|
const char *plugin_name,
|
|
|
|
const char *flavor,
|
|
|
|
const ExecutionContext &exe_ctx,
|
|
|
|
SymbolContextList &sc_list,
|
|
|
|
uint32_t num_instructions,
|
|
|
|
uint32_t num_mixed_context_lines,
|
|
|
|
uint32_t options,
|
|
|
|
Stream &strm)
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2010-07-01 07:03:03 +08:00
|
|
|
size_t success_count = 0;
|
|
|
|
const size_t count = sc_list.GetSize();
|
|
|
|
SymbolContext sc;
|
|
|
|
AddressRange range;
|
2011-04-23 10:04:55 +08:00
|
|
|
const uint32_t scope = eSymbolContextBlock | eSymbolContextFunction | eSymbolContextSymbol;
|
|
|
|
const bool use_inline_block_range = true;
|
2016-03-03 08:51:40 +08:00
|
|
|
for (size_t i = 0; i < count; ++i)
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2016-03-03 08:51:40 +08:00
|
|
|
if (!sc_list.GetContextAtIndex(i, sc))
|
2010-07-01 07:03:03 +08:00
|
|
|
break;
|
2011-04-23 10:04:55 +08:00
|
|
|
for (uint32_t range_idx = 0; sc.GetAddressRange(scope, range_idx, use_inline_block_range, range); ++range_idx)
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2011-03-26 02:03:16 +08:00
|
|
|
if (Disassemble (debugger,
|
|
|
|
arch,
|
|
|
|
plugin_name,
|
2013-03-02 08:26:47 +08:00
|
|
|
flavor,
|
2011-03-26 02:03:16 +08:00
|
|
|
exe_ctx,
|
|
|
|
range,
|
|
|
|
num_instructions,
|
|
|
|
num_mixed_context_lines,
|
2011-06-22 09:39:49 +08:00
|
|
|
options,
|
2011-03-26 02:03:16 +08:00
|
|
|
strm))
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2010-07-01 07:03:03 +08:00
|
|
|
++success_count;
|
|
|
|
strm.EOL();
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
2010-07-01 07:03:03 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return success_count;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool
|
2016-03-03 08:51:40 +08:00
|
|
|
Disassembler::Disassemble(Debugger &debugger,
|
|
|
|
const ArchSpec &arch,
|
|
|
|
const char *plugin_name,
|
|
|
|
const char *flavor,
|
|
|
|
const ExecutionContext &exe_ctx,
|
|
|
|
const ConstString &name,
|
|
|
|
Module *module,
|
|
|
|
uint32_t num_instructions,
|
|
|
|
uint32_t num_mixed_context_lines,
|
|
|
|
uint32_t options,
|
|
|
|
Stream &strm)
|
2010-07-01 07:03:03 +08:00
|
|
|
{
|
|
|
|
SymbolContextList sc_list;
|
2011-01-27 14:44:37 +08:00
|
|
|
if (name)
|
2010-07-01 07:03:03 +08:00
|
|
|
{
|
2011-01-27 14:44:37 +08:00
|
|
|
const bool include_symbols = true;
|
2012-02-11 06:52:19 +08:00
|
|
|
const bool include_inlines = true;
|
2011-01-27 14:44:37 +08:00
|
|
|
if (module)
|
|
|
|
{
|
2016-03-03 08:51:40 +08:00
|
|
|
module->FindFunctions(name,
|
|
|
|
nullptr,
|
|
|
|
eFunctionNameTypeAuto,
|
|
|
|
include_symbols,
|
|
|
|
include_inlines,
|
|
|
|
true,
|
|
|
|
sc_list);
|
2011-01-27 14:44:37 +08:00
|
|
|
}
|
2011-09-22 12:58:26 +08:00
|
|
|
else if (exe_ctx.GetTargetPtr())
|
2011-01-27 14:44:37 +08:00
|
|
|
{
|
2011-09-22 12:58:26 +08:00
|
|
|
exe_ctx.GetTargetPtr()->GetImages().FindFunctions (name,
|
2013-05-18 08:11:21 +08:00
|
|
|
eFunctionNameTypeAuto,
|
2012-02-11 06:52:19 +08:00
|
|
|
include_symbols,
|
|
|
|
include_inlines,
|
2011-09-22 12:58:26 +08:00
|
|
|
false,
|
|
|
|
sc_list);
|
2010-07-01 07:03:03 +08:00
|
|
|
}
|
|
|
|
}
|
2011-01-27 14:44:37 +08:00
|
|
|
|
|
|
|
if (sc_list.GetSize ())
|
|
|
|
{
|
|
|
|
return Disassemble (debugger,
|
|
|
|
arch,
|
2011-03-26 02:03:16 +08:00
|
|
|
plugin_name,
|
2013-03-02 08:26:47 +08:00
|
|
|
flavor,
|
2011-01-27 14:44:37 +08:00
|
|
|
exe_ctx,
|
2011-03-22 09:48:42 +08:00
|
|
|
sc_list,
|
|
|
|
num_instructions,
|
2011-01-27 14:44:37 +08:00
|
|
|
num_mixed_context_lines,
|
2011-06-22 09:39:49 +08:00
|
|
|
options,
|
2011-01-27 14:44:37 +08:00
|
|
|
strm);
|
|
|
|
}
|
2010-07-01 07:03:03 +08:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2010-10-06 11:09:58 +08:00
|
|
|
lldb::DisassemblerSP
|
2016-03-03 08:51:40 +08:00
|
|
|
Disassembler::DisassembleRange(const ArchSpec &arch,
|
|
|
|
const char *plugin_name,
|
|
|
|
const char *flavor,
|
|
|
|
const ExecutionContext &exe_ctx,
|
|
|
|
const AddressRange &range,
|
|
|
|
bool prefer_file_cache)
|
2010-10-06 11:09:58 +08:00
|
|
|
{
|
|
|
|
lldb::DisassemblerSP disasm_sp;
|
|
|
|
if (range.GetByteSize() > 0 && range.GetBaseAddress().IsValid())
|
|
|
|
{
|
2013-03-02 08:26:47 +08:00
|
|
|
disasm_sp = Disassembler::FindPluginForTarget(exe_ctx.GetTargetSP(), arch, flavor, plugin_name);
|
2010-10-06 11:09:58 +08:00
|
|
|
|
|
|
|
if (disasm_sp)
|
|
|
|
{
|
2016-03-03 08:51:40 +08:00
|
|
|
size_t bytes_disassembled = disasm_sp->ParseInstructions(&exe_ctx, range, nullptr, prefer_file_cache);
|
2010-10-06 11:09:58 +08:00
|
|
|
if (bytes_disassembled == 0)
|
|
|
|
disasm_sp.reset();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return disasm_sp;
|
|
|
|
}
|
|
|
|
|
2011-12-15 07:49:37 +08:00
|
|
|
lldb::DisassemblerSP
|
2013-03-29 07:42:53 +08:00
|
|
|
Disassembler::DisassembleBytes (const ArchSpec &arch,
|
|
|
|
const char *plugin_name,
|
|
|
|
const char *flavor,
|
|
|
|
const Address &start,
|
|
|
|
const void *src,
|
|
|
|
size_t src_len,
|
|
|
|
uint32_t num_instructions,
|
|
|
|
bool data_from_file)
|
2011-12-15 07:49:37 +08:00
|
|
|
{
|
|
|
|
lldb::DisassemblerSP disasm_sp;
|
|
|
|
|
2013-03-29 07:42:53 +08:00
|
|
|
if (src)
|
2011-12-15 07:49:37 +08:00
|
|
|
{
|
2013-03-02 08:26:47 +08:00
|
|
|
disasm_sp = Disassembler::FindPlugin(arch, flavor, plugin_name);
|
2011-12-15 07:49:37 +08:00
|
|
|
|
|
|
|
if (disasm_sp)
|
|
|
|
{
|
2013-03-29 07:42:53 +08:00
|
|
|
DataExtractor data(src, src_len, arch.GetByteOrder(), arch.GetAddressByteSize());
|
2011-12-15 07:49:37 +08:00
|
|
|
|
|
|
|
(void)disasm_sp->DecodeInstructions (start,
|
|
|
|
data,
|
|
|
|
0,
|
2012-03-07 06:24:44 +08:00
|
|
|
num_instructions,
|
2013-03-29 07:42:53 +08:00
|
|
|
false,
|
|
|
|
data_from_file);
|
2011-12-15 07:49:37 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return disasm_sp;
|
|
|
|
}
|
|
|
|
|
2010-07-01 07:03:03 +08:00
|
|
|
bool
|
2016-03-03 08:51:40 +08:00
|
|
|
Disassembler::Disassemble(Debugger &debugger,
|
|
|
|
const ArchSpec &arch,
|
|
|
|
const char *plugin_name,
|
|
|
|
const char *flavor,
|
|
|
|
const ExecutionContext &exe_ctx,
|
|
|
|
const AddressRange &disasm_range,
|
|
|
|
uint32_t num_instructions,
|
|
|
|
uint32_t num_mixed_context_lines,
|
|
|
|
uint32_t options,
|
|
|
|
Stream &strm)
|
2010-07-01 07:03:03 +08:00
|
|
|
{
|
|
|
|
if (disasm_range.GetByteSize())
|
|
|
|
{
|
2013-03-02 08:26:47 +08:00
|
|
|
lldb::DisassemblerSP disasm_sp (Disassembler::FindPluginForTarget(exe_ctx.GetTargetSP(), arch, flavor, plugin_name));
|
2010-07-01 07:03:03 +08:00
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
if (disasm_sp)
|
2010-07-01 07:03:03 +08:00
|
|
|
{
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
AddressRange range;
|
|
|
|
ResolveAddress (exe_ctx, disasm_range.GetBaseAddress(), range.GetBaseAddress());
|
|
|
|
range.SetByteSize (disasm_range.GetByteSize());
|
2013-03-29 07:42:53 +08:00
|
|
|
const bool prefer_file_cache = false;
|
|
|
|
size_t bytes_disassembled = disasm_sp->ParseInstructions (&exe_ctx, range, &strm, prefer_file_cache);
|
2010-06-09 00:52:24 +08:00
|
|
|
if (bytes_disassembled == 0)
|
|
|
|
return false;
|
2011-03-26 02:03:16 +08:00
|
|
|
|
2013-07-31 10:19:15 +08:00
|
|
|
bool result = PrintInstructions (disasm_sp.get(),
|
|
|
|
debugger,
|
|
|
|
arch,
|
|
|
|
exe_ctx,
|
|
|
|
num_instructions,
|
|
|
|
num_mixed_context_lines,
|
|
|
|
options,
|
|
|
|
strm);
|
|
|
|
|
|
|
|
// FIXME: The DisassemblerLLVMC has a reference cycle and won't go away if it has any active instructions.
|
|
|
|
// I'll fix that but for now, just clear the list and it will go away nicely.
|
|
|
|
disasm_sp->GetInstructionList().Clear();
|
|
|
|
return result;
|
2011-03-22 09:48:42 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool
|
2016-03-03 08:51:40 +08:00
|
|
|
Disassembler::Disassemble(Debugger &debugger,
|
|
|
|
const ArchSpec &arch,
|
|
|
|
const char *plugin_name,
|
|
|
|
const char *flavor,
|
|
|
|
const ExecutionContext &exe_ctx,
|
|
|
|
const Address &start_address,
|
|
|
|
uint32_t num_instructions,
|
|
|
|
uint32_t num_mixed_context_lines,
|
|
|
|
uint32_t options,
|
|
|
|
Stream &strm)
|
2011-03-22 09:48:42 +08:00
|
|
|
{
|
|
|
|
if (num_instructions > 0)
|
|
|
|
{
|
2013-03-29 07:42:53 +08:00
|
|
|
lldb::DisassemblerSP disasm_sp (Disassembler::FindPluginForTarget(exe_ctx.GetTargetSP(),
|
|
|
|
arch,
|
|
|
|
flavor,
|
|
|
|
plugin_name));
|
2016-03-03 08:51:40 +08:00
|
|
|
if (disasm_sp)
|
2011-03-22 09:48:42 +08:00
|
|
|
{
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
Address addr;
|
|
|
|
ResolveAddress (exe_ctx, start_address, addr);
|
2013-03-29 07:42:53 +08:00
|
|
|
const bool prefer_file_cache = false;
|
|
|
|
size_t bytes_disassembled = disasm_sp->ParseInstructions (&exe_ctx,
|
|
|
|
addr,
|
|
|
|
num_instructions,
|
|
|
|
prefer_file_cache);
|
2011-03-22 09:48:42 +08:00
|
|
|
if (bytes_disassembled == 0)
|
|
|
|
return false;
|
2013-07-31 10:19:15 +08:00
|
|
|
bool result = PrintInstructions (disasm_sp.get(),
|
|
|
|
debugger,
|
|
|
|
arch,
|
|
|
|
exe_ctx,
|
|
|
|
num_instructions,
|
|
|
|
num_mixed_context_lines,
|
|
|
|
options,
|
|
|
|
strm);
|
|
|
|
|
|
|
|
// FIXME: The DisassemblerLLVMC has a reference cycle and won't go away if it has any active instructions.
|
|
|
|
// I'll fix that but for now, just clear the list and it will go away nicely.
|
|
|
|
disasm_sp->GetInstructionList().Clear();
|
|
|
|
return result;
|
2011-03-22 09:48:42 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool
|
2016-03-03 08:51:40 +08:00
|
|
|
Disassembler::PrintInstructions(Disassembler *disasm_ptr,
|
|
|
|
Debugger &debugger,
|
|
|
|
const ArchSpec &arch,
|
|
|
|
const ExecutionContext &exe_ctx,
|
|
|
|
uint32_t num_instructions,
|
|
|
|
uint32_t num_mixed_context_lines,
|
|
|
|
uint32_t options,
|
|
|
|
Stream &strm)
|
2011-03-22 09:48:42 +08:00
|
|
|
{
|
|
|
|
// We got some things disassembled...
|
|
|
|
size_t num_instructions_found = disasm_ptr->GetInstructionList().GetSize();
|
2010-07-01 07:03:03 +08:00
|
|
|
|
2011-03-22 09:48:42 +08:00
|
|
|
if (num_instructions > 0 && num_instructions < num_instructions_found)
|
|
|
|
num_instructions_found = num_instructions;
|
|
|
|
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
const uint32_t max_opcode_byte_size = disasm_ptr->GetInstructionList().GetMaxOpcocdeByteSize ();
|
2011-03-22 09:48:42 +08:00
|
|
|
uint32_t offset = 0;
|
|
|
|
SymbolContext sc;
|
|
|
|
SymbolContext prev_sc;
|
|
|
|
AddressRange sc_range;
|
2016-03-03 08:51:40 +08:00
|
|
|
const Address *pc_addr_ptr = nullptr;
|
2013-11-04 17:33:30 +08:00
|
|
|
StackFrame *frame = exe_ctx.GetFramePtr();
|
2011-09-22 12:58:26 +08:00
|
|
|
|
2013-07-09 01:56:02 +08:00
|
|
|
TargetSP target_sp (exe_ctx.GetTargetSP());
|
|
|
|
SourceManager &source_manager = target_sp ? target_sp->GetSourceManager() : debugger.GetSourceManager();
|
|
|
|
|
2011-09-22 12:58:26 +08:00
|
|
|
if (frame)
|
2014-10-11 07:07:36 +08:00
|
|
|
{
|
2011-09-22 12:58:26 +08:00
|
|
|
pc_addr_ptr = &frame->GetFrameCodeAddress();
|
2014-10-11 07:07:36 +08:00
|
|
|
}
|
2011-04-23 10:04:55 +08:00
|
|
|
const uint32_t scope = eSymbolContextLineEntry | eSymbolContextFunction | eSymbolContextSymbol;
|
|
|
|
const bool use_inline_block_range = false;
|
2015-02-14 07:24:21 +08:00
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
const FormatEntity::Entry *disassembly_format = nullptr;
|
2015-02-14 07:24:21 +08:00
|
|
|
FormatEntity::Entry format;
|
|
|
|
if (exe_ctx.HasTargetScope())
|
|
|
|
{
|
|
|
|
disassembly_format = exe_ctx.GetTargetRef().GetDebugger().GetDisassemblyFormat ();
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
FormatEntity::Parse("${addr}: ", format);
|
|
|
|
disassembly_format = &format;
|
|
|
|
}
|
|
|
|
|
|
|
|
// First pass: step through the list of instructions,
|
|
|
|
// find how long the initial addresses strings are, insert padding
|
|
|
|
// in the second pass so the opcodes all line up nicely.
|
|
|
|
size_t address_text_size = 0;
|
|
|
|
for (size_t i = 0; i < num_instructions_found; ++i)
|
|
|
|
{
|
|
|
|
Instruction *inst = disasm_ptr->GetInstructionList().GetInstructionAtIndex (i).get();
|
|
|
|
if (inst)
|
|
|
|
{
|
|
|
|
const Address &addr = inst->GetAddress();
|
|
|
|
ModuleSP module_sp (addr.GetModule());
|
|
|
|
if (module_sp)
|
|
|
|
{
|
|
|
|
const uint32_t resolve_mask = eSymbolContextFunction | eSymbolContextSymbol;
|
|
|
|
uint32_t resolved_mask = module_sp->ResolveSymbolContextForAddress(addr, resolve_mask, sc);
|
|
|
|
if (resolved_mask)
|
|
|
|
{
|
|
|
|
StreamString strmstr;
|
2016-03-03 08:51:40 +08:00
|
|
|
Debugger::FormatDisassemblerAddress(disassembly_format, &sc, nullptr, &exe_ctx, &addr, strmstr);
|
2015-02-14 07:24:21 +08:00
|
|
|
size_t cur_line = strmstr.GetSizeOfLastLine();
|
|
|
|
if (cur_line > address_text_size)
|
|
|
|
address_text_size = cur_line;
|
|
|
|
}
|
|
|
|
sc.Clear(false);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-10-11 07:07:36 +08:00
|
|
|
for (size_t i = 0; i < num_instructions_found; ++i)
|
2011-03-22 09:48:42 +08:00
|
|
|
{
|
|
|
|
Instruction *inst = disasm_ptr->GetInstructionList().GetInstructionAtIndex (i).get();
|
|
|
|
if (inst)
|
|
|
|
{
|
Many improvements to the Platform base class and subclasses. The base Platform
class now implements the Host functionality for a lot of things that make
sense by default so that subclasses can check:
int
PlatformSubclass::Foo ()
{
if (IsHost())
return Platform::Foo (); // Let the platform base class do the host specific stuff
// Platform subclass specific code...
int result = ...
return result;
}
Added new functions to the platform:
virtual const char *Platform::GetUserName (uint32_t uid);
virtual const char *Platform::GetGroupName (uint32_t gid);
The user and group names are cached locally so that remote platforms can avoid
sending packets multiple times to resolve this information.
Added the parent process ID to the ProcessInfo class.
Added a new ProcessInfoMatch class which helps us to match processes up
and changed the Host layer over to using this new class. The new class allows
us to search for processs:
1 - by name (equal to, starts with, ends with, contains, and regex)
2 - by pid
3 - And further check for parent pid == value, uid == value, gid == value,
euid == value, egid == value, arch == value, parent == value.
This is all hookup up to the "platform process list" command which required
adding dumping routines to dump process information. If the Host class
implements the process lookup routines, you can now lists processes on
your local machine:
machine1.foo.com % lldb
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode
92742 92710 username usergroup username usergroup i386-apple-darwin debugserver
This of course also works remotely with the lldb-platform:
machine1.foo.com % lldb-platform --listen 1234
machine2.foo.com % lldb
(lldb) platform create remote-macosx
Platform: remote-macosx
Connected: no
(lldb) platform connect connect://localhost:1444
Platform: remote-macosx
Triple: x86_64-apple-darwin
OS Version: 10.6.7 (10J869)
Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386
Hostname: machine1.foo.com
Connected: yes
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation
99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
The lldb-platform implements everything with the Host:: layer, so this should
"just work" for linux. I will probably be adding more stuff to the Host layer
for launching processes and attaching to processes so that this support should
eventually just work as well.
Modified the target to be able to be created with an architecture that differs
from the main executable. This is needed for iOS debugging since we can have
an "armv6" binary which can run on an "armv7" machine, so we want to be able
to do:
% lldb
(lldb) platform create remote-ios
(lldb) file --arch armv7 a.out
Where "a.out" is an armv6 executable. The platform then can correctly decide
to open all "armv7" images for all dependent shared libraries.
Modified the disassembly to show the current PC value. Example output:
(lldb) disassemble --frame
a.out`main:
0x1eb7: pushl %ebp
0x1eb8: movl %esp, %ebp
0x1eba: pushl %ebx
0x1ebb: subl $20, %esp
0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18
0x1ec3: popl %ebx
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
0x1edb: leal 213(%ebx), %eax
0x1ee1: movl %eax, (%esp)
0x1ee4: calll 0x1f1e ; puts
0x1ee9: calll 0x1f0c ; getchar
0x1eee: movl $20, (%esp)
0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6
0x1efa: movl $12, %eax
0x1eff: addl $20, %esp
0x1f02: popl %ebx
0x1f03: leave
0x1f04: ret
This can be handy when dealing with the new --line options that was recently
added:
(lldb) disassemble --line
a.out`main + 13 at test.c:19
18 {
-> 19 printf("Process: %i\n\n", getpid());
20 puts("Press any key to continue..."); getchar();
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
Modified the ModuleList to have a lookup based solely on a UUID. Since the
UUID is typically the MD5 checksum of a binary image, there is no need
to give the path and architecture when searching for a pre-existing
image in an image list.
Now that we support remote debugging a bit better, our lldb_private::Module
needs to be able to track what the original path for file was as the platform
knows it, as well as where the file is locally. The module has the two
following functions to retrieve both paths:
const FileSpec &Module::GetFileSpec () const;
const FileSpec &Module::GetPlatformFileSpec () const;
llvm-svn: 128563
2011-03-31 02:16:51 +08:00
|
|
|
const Address &addr = inst->GetAddress();
|
|
|
|
const bool inst_is_at_pc = pc_addr_ptr && addr == *pc_addr_ptr;
|
2011-03-22 09:48:42 +08:00
|
|
|
|
|
|
|
prev_sc = sc;
|
|
|
|
|
2012-02-24 09:59:29 +08:00
|
|
|
ModuleSP module_sp (addr.GetModule());
|
|
|
|
if (module_sp)
|
2011-03-22 09:48:42 +08:00
|
|
|
{
|
2012-02-24 09:59:29 +08:00
|
|
|
uint32_t resolved_mask = module_sp->ResolveSymbolContextForAddress(addr, eSymbolContextEverything, sc);
|
2011-03-22 09:48:42 +08:00
|
|
|
if (resolved_mask)
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
Many improvements to the Platform base class and subclasses. The base Platform
class now implements the Host functionality for a lot of things that make
sense by default so that subclasses can check:
int
PlatformSubclass::Foo ()
{
if (IsHost())
return Platform::Foo (); // Let the platform base class do the host specific stuff
// Platform subclass specific code...
int result = ...
return result;
}
Added new functions to the platform:
virtual const char *Platform::GetUserName (uint32_t uid);
virtual const char *Platform::GetGroupName (uint32_t gid);
The user and group names are cached locally so that remote platforms can avoid
sending packets multiple times to resolve this information.
Added the parent process ID to the ProcessInfo class.
Added a new ProcessInfoMatch class which helps us to match processes up
and changed the Host layer over to using this new class. The new class allows
us to search for processs:
1 - by name (equal to, starts with, ends with, contains, and regex)
2 - by pid
3 - And further check for parent pid == value, uid == value, gid == value,
euid == value, egid == value, arch == value, parent == value.
This is all hookup up to the "platform process list" command which required
adding dumping routines to dump process information. If the Host class
implements the process lookup routines, you can now lists processes on
your local machine:
machine1.foo.com % lldb
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode
92742 92710 username usergroup username usergroup i386-apple-darwin debugserver
This of course also works remotely with the lldb-platform:
machine1.foo.com % lldb-platform --listen 1234
machine2.foo.com % lldb
(lldb) platform create remote-macosx
Platform: remote-macosx
Connected: no
(lldb) platform connect connect://localhost:1444
Platform: remote-macosx
Triple: x86_64-apple-darwin
OS Version: 10.6.7 (10J869)
Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386
Hostname: machine1.foo.com
Connected: yes
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation
99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
The lldb-platform implements everything with the Host:: layer, so this should
"just work" for linux. I will probably be adding more stuff to the Host layer
for launching processes and attaching to processes so that this support should
eventually just work as well.
Modified the target to be able to be created with an architecture that differs
from the main executable. This is needed for iOS debugging since we can have
an "armv6" binary which can run on an "armv7" machine, so we want to be able
to do:
% lldb
(lldb) platform create remote-ios
(lldb) file --arch armv7 a.out
Where "a.out" is an armv6 executable. The platform then can correctly decide
to open all "armv7" images for all dependent shared libraries.
Modified the disassembly to show the current PC value. Example output:
(lldb) disassemble --frame
a.out`main:
0x1eb7: pushl %ebp
0x1eb8: movl %esp, %ebp
0x1eba: pushl %ebx
0x1ebb: subl $20, %esp
0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18
0x1ec3: popl %ebx
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
0x1edb: leal 213(%ebx), %eax
0x1ee1: movl %eax, (%esp)
0x1ee4: calll 0x1f1e ; puts
0x1ee9: calll 0x1f0c ; getchar
0x1eee: movl $20, (%esp)
0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6
0x1efa: movl $12, %eax
0x1eff: addl $20, %esp
0x1f02: popl %ebx
0x1f03: leave
0x1f04: ret
This can be handy when dealing with the new --line options that was recently
added:
(lldb) disassemble --line
a.out`main + 13 at test.c:19
18 {
-> 19 printf("Process: %i\n\n", getpid());
20 puts("Press any key to continue..."); getchar();
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
Modified the ModuleList to have a lookup based solely on a UUID. Since the
UUID is typically the MD5 checksum of a binary image, there is no need
to give the path and architecture when searching for a pre-existing
image in an image list.
Now that we support remote debugging a bit better, our lldb_private::Module
needs to be able to track what the original path for file was as the platform
knows it, as well as where the file is locally. The module has the two
following functions to retrieve both paths:
const FileSpec &Module::GetFileSpec () const;
const FileSpec &Module::GetPlatformFileSpec () const;
llvm-svn: 128563
2011-03-31 02:16:51 +08:00
|
|
|
if (num_mixed_context_lines)
|
|
|
|
{
|
|
|
|
if (!sc_range.ContainsFileAddress (addr))
|
|
|
|
{
|
2011-04-23 10:04:55 +08:00
|
|
|
sc.GetAddressRange (scope, 0, use_inline_block_range, sc_range);
|
Many improvements to the Platform base class and subclasses. The base Platform
class now implements the Host functionality for a lot of things that make
sense by default so that subclasses can check:
int
PlatformSubclass::Foo ()
{
if (IsHost())
return Platform::Foo (); // Let the platform base class do the host specific stuff
// Platform subclass specific code...
int result = ...
return result;
}
Added new functions to the platform:
virtual const char *Platform::GetUserName (uint32_t uid);
virtual const char *Platform::GetGroupName (uint32_t gid);
The user and group names are cached locally so that remote platforms can avoid
sending packets multiple times to resolve this information.
Added the parent process ID to the ProcessInfo class.
Added a new ProcessInfoMatch class which helps us to match processes up
and changed the Host layer over to using this new class. The new class allows
us to search for processs:
1 - by name (equal to, starts with, ends with, contains, and regex)
2 - by pid
3 - And further check for parent pid == value, uid == value, gid == value,
euid == value, egid == value, arch == value, parent == value.
This is all hookup up to the "platform process list" command which required
adding dumping routines to dump process information. If the Host class
implements the process lookup routines, you can now lists processes on
your local machine:
machine1.foo.com % lldb
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode
92742 92710 username usergroup username usergroup i386-apple-darwin debugserver
This of course also works remotely with the lldb-platform:
machine1.foo.com % lldb-platform --listen 1234
machine2.foo.com % lldb
(lldb) platform create remote-macosx
Platform: remote-macosx
Connected: no
(lldb) platform connect connect://localhost:1444
Platform: remote-macosx
Triple: x86_64-apple-darwin
OS Version: 10.6.7 (10J869)
Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386
Hostname: machine1.foo.com
Connected: yes
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation
99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
The lldb-platform implements everything with the Host:: layer, so this should
"just work" for linux. I will probably be adding more stuff to the Host layer
for launching processes and attaching to processes so that this support should
eventually just work as well.
Modified the target to be able to be created with an architecture that differs
from the main executable. This is needed for iOS debugging since we can have
an "armv6" binary which can run on an "armv7" machine, so we want to be able
to do:
% lldb
(lldb) platform create remote-ios
(lldb) file --arch armv7 a.out
Where "a.out" is an armv6 executable. The platform then can correctly decide
to open all "armv7" images for all dependent shared libraries.
Modified the disassembly to show the current PC value. Example output:
(lldb) disassemble --frame
a.out`main:
0x1eb7: pushl %ebp
0x1eb8: movl %esp, %ebp
0x1eba: pushl %ebx
0x1ebb: subl $20, %esp
0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18
0x1ec3: popl %ebx
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
0x1edb: leal 213(%ebx), %eax
0x1ee1: movl %eax, (%esp)
0x1ee4: calll 0x1f1e ; puts
0x1ee9: calll 0x1f0c ; getchar
0x1eee: movl $20, (%esp)
0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6
0x1efa: movl $12, %eax
0x1eff: addl $20, %esp
0x1f02: popl %ebx
0x1f03: leave
0x1f04: ret
This can be handy when dealing with the new --line options that was recently
added:
(lldb) disassemble --line
a.out`main + 13 at test.c:19
18 {
-> 19 printf("Process: %i\n\n", getpid());
20 puts("Press any key to continue..."); getchar();
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
Modified the ModuleList to have a lookup based solely on a UUID. Since the
UUID is typically the MD5 checksum of a binary image, there is no need
to give the path and architecture when searching for a pre-existing
image in an image list.
Now that we support remote debugging a bit better, our lldb_private::Module
needs to be able to track what the original path for file was as the platform
knows it, as well as where the file is locally. The module has the two
following functions to retrieve both paths:
const FileSpec &Module::GetFileSpec () const;
const FileSpec &Module::GetPlatformFileSpec () const;
llvm-svn: 128563
2011-03-31 02:16:51 +08:00
|
|
|
|
|
|
|
if (sc != prev_sc)
|
|
|
|
{
|
|
|
|
if (offset != 0)
|
|
|
|
strm.EOL();
|
|
|
|
|
2015-02-14 07:24:21 +08:00
|
|
|
sc.DumpStopContext(&strm, exe_ctx.GetProcessPtr(), addr, false, true, false, false, true);
|
Many improvements to the Platform base class and subclasses. The base Platform
class now implements the Host functionality for a lot of things that make
sense by default so that subclasses can check:
int
PlatformSubclass::Foo ()
{
if (IsHost())
return Platform::Foo (); // Let the platform base class do the host specific stuff
// Platform subclass specific code...
int result = ...
return result;
}
Added new functions to the platform:
virtual const char *Platform::GetUserName (uint32_t uid);
virtual const char *Platform::GetGroupName (uint32_t gid);
The user and group names are cached locally so that remote platforms can avoid
sending packets multiple times to resolve this information.
Added the parent process ID to the ProcessInfo class.
Added a new ProcessInfoMatch class which helps us to match processes up
and changed the Host layer over to using this new class. The new class allows
us to search for processs:
1 - by name (equal to, starts with, ends with, contains, and regex)
2 - by pid
3 - And further check for parent pid == value, uid == value, gid == value,
euid == value, egid == value, arch == value, parent == value.
This is all hookup up to the "platform process list" command which required
adding dumping routines to dump process information. If the Host class
implements the process lookup routines, you can now lists processes on
your local machine:
machine1.foo.com % lldb
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode
92742 92710 username usergroup username usergroup i386-apple-darwin debugserver
This of course also works remotely with the lldb-platform:
machine1.foo.com % lldb-platform --listen 1234
machine2.foo.com % lldb
(lldb) platform create remote-macosx
Platform: remote-macosx
Connected: no
(lldb) platform connect connect://localhost:1444
Platform: remote-macosx
Triple: x86_64-apple-darwin
OS Version: 10.6.7 (10J869)
Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386
Hostname: machine1.foo.com
Connected: yes
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation
99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
The lldb-platform implements everything with the Host:: layer, so this should
"just work" for linux. I will probably be adding more stuff to the Host layer
for launching processes and attaching to processes so that this support should
eventually just work as well.
Modified the target to be able to be created with an architecture that differs
from the main executable. This is needed for iOS debugging since we can have
an "armv6" binary which can run on an "armv7" machine, so we want to be able
to do:
% lldb
(lldb) platform create remote-ios
(lldb) file --arch armv7 a.out
Where "a.out" is an armv6 executable. The platform then can correctly decide
to open all "armv7" images for all dependent shared libraries.
Modified the disassembly to show the current PC value. Example output:
(lldb) disassemble --frame
a.out`main:
0x1eb7: pushl %ebp
0x1eb8: movl %esp, %ebp
0x1eba: pushl %ebx
0x1ebb: subl $20, %esp
0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18
0x1ec3: popl %ebx
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
0x1edb: leal 213(%ebx), %eax
0x1ee1: movl %eax, (%esp)
0x1ee4: calll 0x1f1e ; puts
0x1ee9: calll 0x1f0c ; getchar
0x1eee: movl $20, (%esp)
0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6
0x1efa: movl $12, %eax
0x1eff: addl $20, %esp
0x1f02: popl %ebx
0x1f03: leave
0x1f04: ret
This can be handy when dealing with the new --line options that was recently
added:
(lldb) disassemble --line
a.out`main + 13 at test.c:19
18 {
-> 19 printf("Process: %i\n\n", getpid());
20 puts("Press any key to continue..."); getchar();
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
Modified the ModuleList to have a lookup based solely on a UUID. Since the
UUID is typically the MD5 checksum of a binary image, there is no need
to give the path and architecture when searching for a pre-existing
image in an image list.
Now that we support remote debugging a bit better, our lldb_private::Module
needs to be able to track what the original path for file was as the platform
knows it, as well as where the file is locally. The module has the two
following functions to retrieve both paths:
const FileSpec &Module::GetFileSpec () const;
const FileSpec &Module::GetPlatformFileSpec () const;
llvm-svn: 128563
2011-03-31 02:16:51 +08:00
|
|
|
strm.EOL();
|
|
|
|
|
|
|
|
if (sc.comp_unit && sc.line_entry.IsValid())
|
|
|
|
{
|
2013-07-09 01:56:02 +08:00
|
|
|
source_manager.DisplaySourceLinesWithLineNumbers (sc.line_entry.file,
|
|
|
|
sc.line_entry.line,
|
|
|
|
num_mixed_context_lines,
|
|
|
|
num_mixed_context_lines,
|
|
|
|
((inst_is_at_pc && (options & eOptionMarkPCSourceLine)) ? "->" : ""),
|
|
|
|
&strm);
|
Many improvements to the Platform base class and subclasses. The base Platform
class now implements the Host functionality for a lot of things that make
sense by default so that subclasses can check:
int
PlatformSubclass::Foo ()
{
if (IsHost())
return Platform::Foo (); // Let the platform base class do the host specific stuff
// Platform subclass specific code...
int result = ...
return result;
}
Added new functions to the platform:
virtual const char *Platform::GetUserName (uint32_t uid);
virtual const char *Platform::GetGroupName (uint32_t gid);
The user and group names are cached locally so that remote platforms can avoid
sending packets multiple times to resolve this information.
Added the parent process ID to the ProcessInfo class.
Added a new ProcessInfoMatch class which helps us to match processes up
and changed the Host layer over to using this new class. The new class allows
us to search for processs:
1 - by name (equal to, starts with, ends with, contains, and regex)
2 - by pid
3 - And further check for parent pid == value, uid == value, gid == value,
euid == value, egid == value, arch == value, parent == value.
This is all hookup up to the "platform process list" command which required
adding dumping routines to dump process information. If the Host class
implements the process lookup routines, you can now lists processes on
your local machine:
machine1.foo.com % lldb
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode
92742 92710 username usergroup username usergroup i386-apple-darwin debugserver
This of course also works remotely with the lldb-platform:
machine1.foo.com % lldb-platform --listen 1234
machine2.foo.com % lldb
(lldb) platform create remote-macosx
Platform: remote-macosx
Connected: no
(lldb) platform connect connect://localhost:1444
Platform: remote-macosx
Triple: x86_64-apple-darwin
OS Version: 10.6.7 (10J869)
Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386
Hostname: machine1.foo.com
Connected: yes
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation
99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
The lldb-platform implements everything with the Host:: layer, so this should
"just work" for linux. I will probably be adding more stuff to the Host layer
for launching processes and attaching to processes so that this support should
eventually just work as well.
Modified the target to be able to be created with an architecture that differs
from the main executable. This is needed for iOS debugging since we can have
an "armv6" binary which can run on an "armv7" machine, so we want to be able
to do:
% lldb
(lldb) platform create remote-ios
(lldb) file --arch armv7 a.out
Where "a.out" is an armv6 executable. The platform then can correctly decide
to open all "armv7" images for all dependent shared libraries.
Modified the disassembly to show the current PC value. Example output:
(lldb) disassemble --frame
a.out`main:
0x1eb7: pushl %ebp
0x1eb8: movl %esp, %ebp
0x1eba: pushl %ebx
0x1ebb: subl $20, %esp
0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18
0x1ec3: popl %ebx
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
0x1edb: leal 213(%ebx), %eax
0x1ee1: movl %eax, (%esp)
0x1ee4: calll 0x1f1e ; puts
0x1ee9: calll 0x1f0c ; getchar
0x1eee: movl $20, (%esp)
0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6
0x1efa: movl $12, %eax
0x1eff: addl $20, %esp
0x1f02: popl %ebx
0x1f03: leave
0x1f04: ret
This can be handy when dealing with the new --line options that was recently
added:
(lldb) disassemble --line
a.out`main + 13 at test.c:19
18 {
-> 19 printf("Process: %i\n\n", getpid());
20 puts("Press any key to continue..."); getchar();
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
Modified the ModuleList to have a lookup based solely on a UUID. Since the
UUID is typically the MD5 checksum of a binary image, there is no need
to give the path and architecture when searching for a pre-existing
image in an image list.
Now that we support remote debugging a bit better, our lldb_private::Module
needs to be able to track what the original path for file was as the platform
knows it, as well as where the file is locally. The module has the two
following functions to retrieve both paths:
const FileSpec &Module::GetFileSpec () const;
const FileSpec &Module::GetPlatformFileSpec () const;
llvm-svn: 128563
2011-03-31 02:16:51 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
2011-03-22 09:48:42 +08:00
|
|
|
else
|
|
|
|
{
|
2013-02-23 12:12:47 +08:00
|
|
|
sc.Clear(true);
|
2011-03-22 09:48:42 +08:00
|
|
|
}
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
2011-03-22 09:48:42 +08:00
|
|
|
|
2014-10-11 07:07:36 +08:00
|
|
|
const bool show_bytes = (options & eOptionShowBytes) != 0;
|
2016-03-03 08:51:40 +08:00
|
|
|
inst->Dump(&strm, max_opcode_byte_size, true, show_bytes, &exe_ctx, &sc, &prev_sc, nullptr, address_text_size);
|
Many improvements to the Platform base class and subclasses. The base Platform
class now implements the Host functionality for a lot of things that make
sense by default so that subclasses can check:
int
PlatformSubclass::Foo ()
{
if (IsHost())
return Platform::Foo (); // Let the platform base class do the host specific stuff
// Platform subclass specific code...
int result = ...
return result;
}
Added new functions to the platform:
virtual const char *Platform::GetUserName (uint32_t uid);
virtual const char *Platform::GetGroupName (uint32_t gid);
The user and group names are cached locally so that remote platforms can avoid
sending packets multiple times to resolve this information.
Added the parent process ID to the ProcessInfo class.
Added a new ProcessInfoMatch class which helps us to match processes up
and changed the Host layer over to using this new class. The new class allows
us to search for processs:
1 - by name (equal to, starts with, ends with, contains, and regex)
2 - by pid
3 - And further check for parent pid == value, uid == value, gid == value,
euid == value, egid == value, arch == value, parent == value.
This is all hookup up to the "platform process list" command which required
adding dumping routines to dump process information. If the Host class
implements the process lookup routines, you can now lists processes on
your local machine:
machine1.foo.com % lldb
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
94727 244 username usergroup username usergroup x86_64-apple-darwin Xcode
92742 92710 username usergroup username usergroup i386-apple-darwin debugserver
This of course also works remotely with the lldb-platform:
machine1.foo.com % lldb-platform --listen 1234
machine2.foo.com % lldb
(lldb) platform create remote-macosx
Platform: remote-macosx
Connected: no
(lldb) platform connect connect://localhost:1444
Platform: remote-macosx
Triple: x86_64-apple-darwin
OS Version: 10.6.7 (10J869)
Kernel: Darwin Kernel Version 10.7.0: Sat Jan 29 15:17:16 PST 2011; root:xnu-1504.9.37~1/RELEASE_I386
Hostname: machine1.foo.com
Connected: yes
(lldb) platform process list
PID PARENT USER GROUP EFF USER EFF GROUP TRIPLE NAME
====== ====== ========== ========== ========== ========== ======================== ============================
99556 244 username usergroup username usergroup x86_64-apple-darwin trustevaluation
99548 65539 username usergroup username usergroup x86_64-apple-darwin lldb
99538 1 username usergroup username usergroup x86_64-apple-darwin FileMerge
94943 1 username usergroup username usergroup x86_64-apple-darwin mdworker
94852 244 username usergroup username usergroup x86_64-apple-darwin Safari
The lldb-platform implements everything with the Host:: layer, so this should
"just work" for linux. I will probably be adding more stuff to the Host layer
for launching processes and attaching to processes so that this support should
eventually just work as well.
Modified the target to be able to be created with an architecture that differs
from the main executable. This is needed for iOS debugging since we can have
an "armv6" binary which can run on an "armv7" machine, so we want to be able
to do:
% lldb
(lldb) platform create remote-ios
(lldb) file --arch armv7 a.out
Where "a.out" is an armv6 executable. The platform then can correctly decide
to open all "armv7" images for all dependent shared libraries.
Modified the disassembly to show the current PC value. Example output:
(lldb) disassemble --frame
a.out`main:
0x1eb7: pushl %ebp
0x1eb8: movl %esp, %ebp
0x1eba: pushl %ebx
0x1ebb: subl $20, %esp
0x1ebe: calll 0x1ec3 ; main + 12 at test.c:18
0x1ec3: popl %ebx
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
0x1edb: leal 213(%ebx), %eax
0x1ee1: movl %eax, (%esp)
0x1ee4: calll 0x1f1e ; puts
0x1ee9: calll 0x1f0c ; getchar
0x1eee: movl $20, (%esp)
0x1ef5: calll 0x1e6a ; sleep_loop at test.c:6
0x1efa: movl $12, %eax
0x1eff: addl $20, %esp
0x1f02: popl %ebx
0x1f03: leave
0x1f04: ret
This can be handy when dealing with the new --line options that was recently
added:
(lldb) disassemble --line
a.out`main + 13 at test.c:19
18 {
-> 19 printf("Process: %i\n\n", getpid());
20 puts("Press any key to continue..."); getchar();
-> 0x1ec4: calll 0x1f12 ; getpid
0x1ec9: movl %eax, 4(%esp)
0x1ecd: leal 199(%ebx), %eax
0x1ed3: movl %eax, (%esp)
0x1ed6: calll 0x1f18 ; printf
Modified the ModuleList to have a lookup based solely on a UUID. Since the
UUID is typically the MD5 checksum of a binary image, there is no need
to give the path and architecture when searching for a pre-existing
image in an image list.
Now that we support remote debugging a bit better, our lldb_private::Module
needs to be able to track what the original path for file was as the platform
knows it, as well as where the file is locally. The module has the two
following functions to retrieve both paths:
const FileSpec &Module::GetFileSpec () const;
const FileSpec &Module::GetPlatformFileSpec () const;
llvm-svn: 128563
2011-03-31 02:16:51 +08:00
|
|
|
strm.EOL();
|
2011-03-22 09:48:42 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
break;
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
|
|
|
}
|
2011-03-22 09:48:42 +08:00
|
|
|
|
|
|
|
return true;
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
|
|
|
|
2010-07-01 07:03:03 +08:00
|
|
|
bool
|
2016-03-03 08:51:40 +08:00
|
|
|
Disassembler::Disassemble(Debugger &debugger,
|
|
|
|
const ArchSpec &arch,
|
|
|
|
const char *plugin_name,
|
|
|
|
const char *flavor,
|
|
|
|
const ExecutionContext &exe_ctx,
|
|
|
|
uint32_t num_instructions,
|
|
|
|
uint32_t num_mixed_context_lines,
|
|
|
|
uint32_t options,
|
|
|
|
Stream &strm)
|
2010-07-01 07:03:03 +08:00
|
|
|
{
|
|
|
|
AddressRange range;
|
2013-11-04 17:33:30 +08:00
|
|
|
StackFrame *frame = exe_ctx.GetFramePtr();
|
2011-09-22 12:58:26 +08:00
|
|
|
if (frame)
|
2010-07-01 07:03:03 +08:00
|
|
|
{
|
2011-09-22 12:58:26 +08:00
|
|
|
SymbolContext sc(frame->GetSymbolContext(eSymbolContextFunction | eSymbolContextSymbol));
|
2010-07-01 07:03:03 +08:00
|
|
|
if (sc.function)
|
|
|
|
{
|
|
|
|
range = sc.function->GetAddressRange();
|
|
|
|
}
|
2012-03-08 05:03:09 +08:00
|
|
|
else if (sc.symbol && sc.symbol->ValueIsAddress())
|
2010-07-01 07:03:03 +08:00
|
|
|
{
|
2015-06-26 05:46:34 +08:00
|
|
|
range.GetBaseAddress() = sc.symbol->GetAddressRef();
|
2012-03-08 05:03:09 +08:00
|
|
|
range.SetByteSize (sc.symbol->GetByteSize());
|
2010-07-01 07:03:03 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2011-09-22 12:58:26 +08:00
|
|
|
range.GetBaseAddress() = frame->GetFrameCodeAddress();
|
2010-07-01 07:03:03 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (range.GetBaseAddress().IsValid() && range.GetByteSize() == 0)
|
|
|
|
range.SetByteSize (DEFAULT_DISASM_BYTE_SIZE);
|
|
|
|
}
|
|
|
|
|
2011-03-26 02:03:16 +08:00
|
|
|
return Disassemble (debugger,
|
|
|
|
arch,
|
|
|
|
plugin_name,
|
2013-03-02 08:26:47 +08:00
|
|
|
flavor,
|
2011-03-26 02:03:16 +08:00
|
|
|
exe_ctx,
|
|
|
|
range,
|
|
|
|
num_instructions,
|
|
|
|
num_mixed_context_lines,
|
2011-06-22 09:39:49 +08:00
|
|
|
options,
|
2011-03-26 02:03:16 +08:00
|
|
|
strm);
|
2010-07-01 07:03:03 +08:00
|
|
|
}
|
|
|
|
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
Instruction::Instruction(const Address &address, AddressClass addr_class) :
|
2011-03-26 02:03:16 +08:00
|
|
|
m_address (address),
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
m_address_class (addr_class),
|
2012-02-14 08:22:51 +08:00
|
|
|
m_opcode(),
|
|
|
|
m_calculated_strings(false)
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
|
|
|
}
|
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
Instruction::~Instruction() = default;
|
2010-06-09 00:52:24 +08:00
|
|
|
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
AddressClass
|
|
|
|
Instruction::GetAddressClass ()
|
|
|
|
{
|
|
|
|
if (m_address_class == eAddressClassInvalid)
|
|
|
|
m_address_class = m_address.GetAddressClass();
|
|
|
|
return m_address_class;
|
|
|
|
}
|
2010-06-09 00:52:24 +08:00
|
|
|
|
2012-05-10 10:52:23 +08:00
|
|
|
void
|
|
|
|
Instruction::Dump (lldb_private::Stream *s,
|
|
|
|
uint32_t max_opcode_byte_size,
|
|
|
|
bool show_address,
|
|
|
|
bool show_bytes,
|
2014-10-11 07:07:36 +08:00
|
|
|
const ExecutionContext* exe_ctx,
|
|
|
|
const SymbolContext *sym_ctx,
|
|
|
|
const SymbolContext *prev_sym_ctx,
|
2015-02-14 07:24:21 +08:00
|
|
|
const FormatEntity::Entry *disassembly_addr_format,
|
|
|
|
size_t max_address_text_size)
|
2012-05-10 10:52:23 +08:00
|
|
|
{
|
2013-01-05 07:52:35 +08:00
|
|
|
size_t opcode_column_width = 7;
|
2012-05-10 10:52:23 +08:00
|
|
|
const size_t operand_column_width = 25;
|
|
|
|
|
|
|
|
CalculateMnemonicOperandsAndCommentIfNeeded (exe_ctx);
|
|
|
|
|
|
|
|
StreamString ss;
|
|
|
|
|
|
|
|
if (show_address)
|
|
|
|
{
|
2015-02-05 06:00:53 +08:00
|
|
|
Debugger::FormatDisassemblerAddress (disassembly_addr_format, sym_ctx, prev_sym_ctx, exe_ctx, &m_address, ss);
|
2015-02-14 07:24:21 +08:00
|
|
|
ss.FillLastLineToColumn (max_address_text_size, ' ');
|
2012-05-10 10:52:23 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
if (show_bytes)
|
|
|
|
{
|
|
|
|
if (m_opcode.GetType() == Opcode::eTypeBytes)
|
|
|
|
{
|
|
|
|
// x86_64 and i386 are the only ones that use bytes right now so
|
|
|
|
// pad out the byte dump to be able to always show 15 bytes (3 chars each)
|
|
|
|
// plus a space
|
|
|
|
if (max_opcode_byte_size > 0)
|
|
|
|
m_opcode.Dump (&ss, max_opcode_byte_size * 3 + 1);
|
|
|
|
else
|
|
|
|
m_opcode.Dump (&ss, 15 * 3 + 1);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2013-10-11 03:17:07 +08:00
|
|
|
// Else, we have ARM or MIPS which can show up to a uint32_t
|
|
|
|
// 0x00000000 (10 spaces) plus two for padding...
|
2012-05-10 10:52:23 +08:00
|
|
|
if (max_opcode_byte_size > 0)
|
|
|
|
m_opcode.Dump (&ss, max_opcode_byte_size * 3 + 1);
|
|
|
|
else
|
|
|
|
m_opcode.Dump (&ss, 12);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-10-11 07:07:36 +08:00
|
|
|
const size_t opcode_pos = ss.GetSizeOfLastLine();
|
2012-05-10 10:52:23 +08:00
|
|
|
|
2013-01-05 07:52:35 +08:00
|
|
|
// The default opcode size of 7 characters is plenty for most architectures
|
|
|
|
// but some like arm can pull out the occasional vqrshrun.s16. We won't get
|
|
|
|
// consistent column spacing in these cases, unfortunately.
|
|
|
|
if (m_opcode_name.length() >= opcode_column_width)
|
|
|
|
{
|
|
|
|
opcode_column_width = m_opcode_name.length() + 1;
|
|
|
|
}
|
|
|
|
|
2012-05-10 10:52:23 +08:00
|
|
|
ss.PutCString (m_opcode_name.c_str());
|
|
|
|
ss.FillLastLineToColumn (opcode_pos + opcode_column_width, ' ');
|
2013-03-02 08:26:47 +08:00
|
|
|
ss.PutCString (m_mnemonics.c_str());
|
2012-05-10 10:52:23 +08:00
|
|
|
|
|
|
|
if (!m_comment.empty())
|
|
|
|
{
|
|
|
|
ss.FillLastLineToColumn (opcode_pos + opcode_column_width + operand_column_width, ' ');
|
|
|
|
ss.PutCString (" ; ");
|
|
|
|
ss.PutCString (m_comment.c_str());
|
|
|
|
}
|
|
|
|
s->Write (ss.GetData(), ss.GetSize());
|
|
|
|
}
|
|
|
|
|
2011-04-06 07:22:54 +08:00
|
|
|
bool
|
|
|
|
Instruction::DumpEmulation (const ArchSpec &arch)
|
|
|
|
{
|
2016-03-03 08:51:40 +08:00
|
|
|
std::unique_ptr<EmulateInstruction> insn_emulator_ap(EmulateInstruction::FindPlugin(arch, eInstructionTypeAny, nullptr));
|
|
|
|
if (insn_emulator_ap)
|
|
|
|
{
|
|
|
|
insn_emulator_ap->SetInstruction(GetOpcode(), GetAddress(), nullptr);
|
2011-04-26 12:39:08 +08:00
|
|
|
return insn_emulator_ap->EvaluateInstruction (0);
|
2016-03-03 08:51:40 +08:00
|
|
|
}
|
2011-04-06 07:22:54 +08:00
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2015-08-26 14:04:54 +08:00
|
|
|
bool
|
|
|
|
Instruction::HasDelaySlot ()
|
|
|
|
{
|
|
|
|
// Default is false.
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2011-04-22 13:08:45 +08:00
|
|
|
OptionValueSP
|
|
|
|
Instruction::ReadArray (FILE *in_file, Stream *out_stream, OptionValue::Type data_type)
|
|
|
|
{
|
|
|
|
bool done = false;
|
|
|
|
char buffer[1024];
|
|
|
|
|
|
|
|
OptionValueSP option_value_sp (new OptionValueArray (1u << data_type));
|
|
|
|
|
|
|
|
int idx = 0;
|
|
|
|
while (!done)
|
|
|
|
{
|
|
|
|
if (!fgets (buffer, 1023, in_file))
|
|
|
|
{
|
2011-09-19 02:59:15 +08:00
|
|
|
out_stream->Printf ("Instruction::ReadArray: Error reading file (fgets).\n");
|
2011-04-22 13:08:45 +08:00
|
|
|
option_value_sp.reset ();
|
|
|
|
return option_value_sp;
|
|
|
|
}
|
|
|
|
|
|
|
|
std::string line (buffer);
|
|
|
|
|
2013-01-26 02:06:21 +08:00
|
|
|
size_t len = line.size();
|
2011-04-22 13:08:45 +08:00
|
|
|
if (line[len-1] == '\n')
|
|
|
|
{
|
|
|
|
line[len-1] = '\0';
|
|
|
|
line.resize (len-1);
|
|
|
|
}
|
|
|
|
|
|
|
|
if ((line.size() == 1) && line[0] == ']')
|
|
|
|
{
|
|
|
|
done = true;
|
|
|
|
line.clear();
|
|
|
|
}
|
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
if (!line.empty())
|
2011-04-22 13:08:45 +08:00
|
|
|
{
|
|
|
|
std::string value;
|
2013-04-04 05:37:16 +08:00
|
|
|
static RegularExpression g_reg_exp ("^[ \t]*([^ \t]+)[ \t]*$");
|
|
|
|
RegularExpression::Match regex_match(1);
|
|
|
|
bool reg_exp_success = g_reg_exp.Execute (line.c_str(), ®ex_match);
|
2011-04-22 13:08:45 +08:00
|
|
|
if (reg_exp_success)
|
2013-04-04 05:37:16 +08:00
|
|
|
regex_match.GetMatchAtIndex (line.c_str(), 1, value);
|
2011-04-22 13:08:45 +08:00
|
|
|
else
|
|
|
|
value = line;
|
|
|
|
|
|
|
|
OptionValueSP data_value_sp;
|
|
|
|
switch (data_type)
|
|
|
|
{
|
|
|
|
case OptionValue::eTypeUInt64:
|
|
|
|
data_value_sp.reset (new OptionValueUInt64 (0, 0));
|
2015-02-20 19:14:59 +08:00
|
|
|
data_value_sp->SetValueFromString (value);
|
2011-04-22 13:08:45 +08:00
|
|
|
break;
|
|
|
|
// Other types can be added later as needed.
|
|
|
|
default:
|
|
|
|
data_value_sp.reset (new OptionValueString (value.c_str(), ""));
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2011-04-28 06:04:39 +08:00
|
|
|
option_value_sp->GetAsArray()->InsertValue (idx, data_value_sp);
|
2011-04-22 13:08:45 +08:00
|
|
|
++idx;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return option_value_sp;
|
|
|
|
}
|
|
|
|
|
|
|
|
OptionValueSP
|
|
|
|
Instruction::ReadDictionary (FILE *in_file, Stream *out_stream)
|
|
|
|
{
|
|
|
|
bool done = false;
|
|
|
|
char buffer[1024];
|
|
|
|
|
|
|
|
OptionValueSP option_value_sp (new OptionValueDictionary());
|
|
|
|
static ConstString encoding_key ("data_encoding");
|
|
|
|
OptionValue::Type data_type = OptionValue::eTypeInvalid;
|
|
|
|
|
|
|
|
|
|
|
|
while (!done)
|
|
|
|
{
|
|
|
|
// Read the next line in the file
|
|
|
|
if (!fgets (buffer, 1023, in_file))
|
|
|
|
{
|
|
|
|
out_stream->Printf ("Instruction::ReadDictionary: Error reading file (fgets).\n");
|
|
|
|
option_value_sp.reset ();
|
|
|
|
return option_value_sp;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Check to see if the line contains the end-of-dictionary marker ("}")
|
|
|
|
std::string line (buffer);
|
|
|
|
|
2013-01-26 02:06:21 +08:00
|
|
|
size_t len = line.size();
|
2011-04-22 13:08:45 +08:00
|
|
|
if (line[len-1] == '\n')
|
|
|
|
{
|
|
|
|
line[len-1] = '\0';
|
|
|
|
line.resize (len-1);
|
|
|
|
}
|
|
|
|
|
|
|
|
if ((line.size() == 1) && (line[0] == '}'))
|
|
|
|
{
|
|
|
|
done = true;
|
|
|
|
line.clear();
|
|
|
|
}
|
|
|
|
|
|
|
|
// Try to find a key-value pair in the current line and add it to the dictionary.
|
2016-03-03 08:51:40 +08:00
|
|
|
if (!line.empty())
|
2011-04-22 13:08:45 +08:00
|
|
|
{
|
2013-04-04 05:37:16 +08:00
|
|
|
static RegularExpression g_reg_exp ("^[ \t]*([a-zA-Z_][a-zA-Z0-9_]*)[ \t]*=[ \t]*(.*)[ \t]*$");
|
|
|
|
RegularExpression::Match regex_match(2);
|
|
|
|
|
|
|
|
bool reg_exp_success = g_reg_exp.Execute (line.c_str(), ®ex_match);
|
2011-04-22 13:08:45 +08:00
|
|
|
std::string key;
|
|
|
|
std::string value;
|
|
|
|
if (reg_exp_success)
|
|
|
|
{
|
2013-04-04 05:37:16 +08:00
|
|
|
regex_match.GetMatchAtIndex (line.c_str(), 1, key);
|
|
|
|
regex_match.GetMatchAtIndex (line.c_str(), 2, value);
|
2011-04-22 13:08:45 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
out_stream->Printf ("Instruction::ReadDictionary: Failure executing regular expression.\n");
|
|
|
|
option_value_sp.reset();
|
|
|
|
return option_value_sp;
|
|
|
|
}
|
|
|
|
|
|
|
|
ConstString const_key (key.c_str());
|
|
|
|
// Check value to see if it's the start of an array or dictionary.
|
|
|
|
|
|
|
|
lldb::OptionValueSP value_sp;
|
|
|
|
assert (value.empty() == false);
|
|
|
|
assert (key.empty() == false);
|
|
|
|
|
|
|
|
if (value[0] == '{')
|
|
|
|
{
|
|
|
|
assert (value.size() == 1);
|
|
|
|
// value is a dictionary
|
|
|
|
value_sp = ReadDictionary (in_file, out_stream);
|
2016-03-03 08:51:40 +08:00
|
|
|
if (!value_sp)
|
2011-04-22 13:08:45 +08:00
|
|
|
{
|
|
|
|
option_value_sp.reset ();
|
|
|
|
return option_value_sp;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else if (value[0] == '[')
|
|
|
|
{
|
|
|
|
assert (value.size() == 1);
|
|
|
|
// value is an array
|
|
|
|
value_sp = ReadArray (in_file, out_stream, data_type);
|
2016-03-03 08:51:40 +08:00
|
|
|
if (!value_sp)
|
2011-04-22 13:08:45 +08:00
|
|
|
{
|
|
|
|
option_value_sp.reset ();
|
|
|
|
return option_value_sp;
|
|
|
|
}
|
|
|
|
// We've used the data_type to read an array; re-set the type to Invalid
|
|
|
|
data_type = OptionValue::eTypeInvalid;
|
|
|
|
}
|
|
|
|
else if ((value[0] == '0') && (value[1] == 'x'))
|
|
|
|
{
|
|
|
|
value_sp.reset (new OptionValueUInt64 (0, 0));
|
2015-02-20 19:14:59 +08:00
|
|
|
value_sp->SetValueFromString (value);
|
2011-04-22 13:08:45 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2013-01-26 02:06:21 +08:00
|
|
|
size_t len = value.size();
|
2011-04-22 13:08:45 +08:00
|
|
|
if ((value[0] == '"') && (value[len-1] == '"'))
|
|
|
|
value = value.substr (1, len-2);
|
|
|
|
value_sp.reset (new OptionValueString (value.c_str(), ""));
|
|
|
|
}
|
|
|
|
|
|
|
|
if (const_key == encoding_key)
|
|
|
|
{
|
|
|
|
// A 'data_encoding=..." is NOT a normal key-value pair; it is meta-data indicating the
|
|
|
|
// data type of an upcoming array (usually the next bit of data to be read in).
|
|
|
|
if (strcmp (value.c_str(), "uint32_t") == 0)
|
|
|
|
data_type = OptionValue::eTypeUInt64;
|
|
|
|
}
|
|
|
|
else
|
2011-04-28 06:04:39 +08:00
|
|
|
option_value_sp->GetAsDictionary()->SetValueForKey (const_key, value_sp, false);
|
2011-04-22 13:08:45 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return option_value_sp;
|
|
|
|
}
|
|
|
|
|
2011-04-20 07:30:03 +08:00
|
|
|
bool
|
|
|
|
Instruction::TestEmulation (Stream *out_stream, const char *file_name)
|
|
|
|
{
|
|
|
|
if (!out_stream)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
if (!file_name)
|
|
|
|
{
|
2011-04-22 04:27:45 +08:00
|
|
|
out_stream->Printf ("Instruction::TestEmulation: Missing file_name.");
|
2011-04-20 07:30:03 +08:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
FILE *test_file = fopen (file_name, "r");
|
|
|
|
if (!test_file)
|
|
|
|
{
|
2011-04-22 04:27:45 +08:00
|
|
|
out_stream->Printf ("Instruction::TestEmulation: Attempt to open test file failed.");
|
2011-04-20 07:30:03 +08:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
char buffer[256];
|
2011-04-22 13:08:45 +08:00
|
|
|
if (!fgets (buffer, 255, test_file))
|
2011-04-20 07:30:03 +08:00
|
|
|
{
|
2011-04-22 13:08:45 +08:00
|
|
|
out_stream->Printf ("Instruction::TestEmulation: Error reading first line of test file.\n");
|
2011-04-20 07:30:03 +08:00
|
|
|
fclose (test_file);
|
|
|
|
return false;
|
|
|
|
}
|
2011-04-22 13:08:45 +08:00
|
|
|
|
|
|
|
if (strncmp (buffer, "InstructionEmulationState={", 27) != 0)
|
|
|
|
{
|
|
|
|
out_stream->Printf ("Instructin::TestEmulation: Test file does not contain emulation state dictionary\n");
|
|
|
|
fclose (test_file);
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Read all the test information from the test file into an OptionValueDictionary.
|
|
|
|
|
|
|
|
OptionValueSP data_dictionary_sp (ReadDictionary (test_file, out_stream));
|
2016-03-03 08:51:40 +08:00
|
|
|
if (!data_dictionary_sp)
|
2011-04-20 07:30:03 +08:00
|
|
|
{
|
2011-04-22 13:08:45 +08:00
|
|
|
out_stream->Printf ("Instruction::TestEmulation: Error reading Dictionary Object.\n");
|
2011-04-20 07:30:03 +08:00
|
|
|
fclose (test_file);
|
|
|
|
return false;
|
|
|
|
}
|
2011-04-22 13:08:45 +08:00
|
|
|
|
|
|
|
fclose (test_file);
|
|
|
|
|
2011-04-28 06:04:39 +08:00
|
|
|
OptionValueDictionary *data_dictionary = data_dictionary_sp->GetAsDictionary();
|
2011-04-22 13:08:45 +08:00
|
|
|
static ConstString description_key ("assembly_string");
|
|
|
|
static ConstString triple_key ("triple");
|
|
|
|
|
|
|
|
OptionValueSP value_sp = data_dictionary->GetValueForKey (description_key);
|
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
if (!value_sp)
|
2011-04-22 13:08:45 +08:00
|
|
|
{
|
|
|
|
out_stream->Printf ("Instruction::TestEmulation: Test file does not contain description string.\n");
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
SetDescription (value_sp->GetStringValue());
|
2016-03-03 08:51:40 +08:00
|
|
|
|
2011-04-22 13:08:45 +08:00
|
|
|
value_sp = data_dictionary->GetValueForKey (triple_key);
|
2016-03-03 08:51:40 +08:00
|
|
|
if (!value_sp)
|
2011-04-22 13:08:45 +08:00
|
|
|
{
|
|
|
|
out_stream->Printf ("Instruction::TestEmulation: Test file does not contain triple.\n");
|
|
|
|
return false;
|
|
|
|
}
|
2011-04-20 07:30:03 +08:00
|
|
|
|
2011-04-22 13:08:45 +08:00
|
|
|
ArchSpec arch;
|
|
|
|
arch.SetTriple (llvm::Triple (value_sp->GetStringValue()));
|
2011-04-20 07:30:03 +08:00
|
|
|
|
|
|
|
bool success = false;
|
2016-03-03 08:51:40 +08:00
|
|
|
std::unique_ptr<EmulateInstruction> insn_emulator_ap(EmulateInstruction::FindPlugin(arch, eInstructionTypeAny, nullptr));
|
|
|
|
if (insn_emulator_ap)
|
2011-04-22 13:08:45 +08:00
|
|
|
success = insn_emulator_ap->TestEmulation (out_stream, arch, data_dictionary);
|
2011-04-20 07:30:03 +08:00
|
|
|
|
|
|
|
if (success)
|
2011-04-22 04:27:45 +08:00
|
|
|
out_stream->Printf ("Emulation test succeeded.");
|
2011-04-20 07:30:03 +08:00
|
|
|
else
|
2011-04-22 04:27:45 +08:00
|
|
|
out_stream->Printf ("Emulation test failed.");
|
2011-04-20 07:30:03 +08:00
|
|
|
|
|
|
|
return success;
|
|
|
|
}
|
|
|
|
|
2011-04-06 07:22:54 +08:00
|
|
|
bool
|
|
|
|
Instruction::Emulate (const ArchSpec &arch,
|
2011-04-26 12:39:08 +08:00
|
|
|
uint32_t evaluate_options,
|
2011-04-06 07:22:54 +08:00
|
|
|
void *baton,
|
2011-05-10 04:18:18 +08:00
|
|
|
EmulateInstruction::ReadMemoryCallback read_mem_callback,
|
|
|
|
EmulateInstruction::WriteMemoryCallback write_mem_callback,
|
|
|
|
EmulateInstruction::ReadRegisterCallback read_reg_callback,
|
|
|
|
EmulateInstruction::WriteRegisterCallback write_reg_callback)
|
2011-04-06 07:22:54 +08:00
|
|
|
{
|
2016-03-03 08:51:40 +08:00
|
|
|
std::unique_ptr<EmulateInstruction> insn_emulator_ap(EmulateInstruction::FindPlugin(arch, eInstructionTypeAny, nullptr));
|
|
|
|
if (insn_emulator_ap)
|
|
|
|
{
|
|
|
|
insn_emulator_ap->SetBaton(baton);
|
|
|
|
insn_emulator_ap->SetCallbacks(read_mem_callback, write_mem_callback, read_reg_callback, write_reg_callback);
|
|
|
|
insn_emulator_ap->SetInstruction(GetOpcode(), GetAddress(), nullptr);
|
|
|
|
return insn_emulator_ap->EvaluateInstruction(evaluate_options);
|
|
|
|
}
|
2011-04-06 07:22:54 +08:00
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2012-05-10 10:52:23 +08:00
|
|
|
uint32_t
|
|
|
|
Instruction::GetData (DataExtractor &data)
|
|
|
|
{
|
2012-08-07 09:44:58 +08:00
|
|
|
return m_opcode.GetData(data);
|
2012-05-10 10:52:23 +08:00
|
|
|
}
|
|
|
|
|
2010-10-06 11:09:58 +08:00
|
|
|
InstructionList::InstructionList() :
|
2010-06-09 00:52:24 +08:00
|
|
|
m_instructions()
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
InstructionList::~InstructionList() = default;
|
2010-06-09 00:52:24 +08:00
|
|
|
|
|
|
|
size_t
|
2010-10-06 11:09:58 +08:00
|
|
|
InstructionList::GetSize() const
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
|
|
|
return m_instructions.size();
|
|
|
|
}
|
|
|
|
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
uint32_t
|
|
|
|
InstructionList::GetMaxOpcocdeByteSize () const
|
|
|
|
{
|
|
|
|
uint32_t max_inst_size = 0;
|
|
|
|
collection::const_iterator pos, end;
|
|
|
|
for (pos = m_instructions.begin(), end = m_instructions.end();
|
|
|
|
pos != end;
|
|
|
|
++pos)
|
|
|
|
{
|
|
|
|
uint32_t inst_size = (*pos)->GetOpcode().GetByteSize();
|
|
|
|
if (max_inst_size < inst_size)
|
|
|
|
max_inst_size = inst_size;
|
|
|
|
}
|
|
|
|
return max_inst_size;
|
|
|
|
}
|
|
|
|
|
2010-10-06 11:09:58 +08:00
|
|
|
InstructionSP
|
2013-01-26 02:06:21 +08:00
|
|
|
InstructionList::GetInstructionAtIndex (size_t idx) const
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2010-10-06 11:09:58 +08:00
|
|
|
InstructionSP inst_sp;
|
2010-06-09 00:52:24 +08:00
|
|
|
if (idx < m_instructions.size())
|
2010-10-06 11:09:58 +08:00
|
|
|
inst_sp = m_instructions[idx];
|
|
|
|
return inst_sp;
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
|
|
|
|
Added support for the new ".apple_objc" accelerator tables. These tables are
in the same hashed format as the ".apple_names", but they map objective C
class names to all of the methods and class functions. We need to do this
because in the DWARF the methods for Objective C are never contained in the
class definition, they are scattered about at the translation unit level and
they don't even have attributes that say the are contained within the class
itself.
Added 3 new formats which can be used to display data:
eFormatAddressInfo
eFormatHexFloat
eFormatInstruction
eFormatAddressInfo describes an address such as function+offset and file+line,
or symbol + offset, or constant data (c string, 2, 4, 8, or 16 byte constants).
The format character for this is "A", the long format is "address".
eFormatHexFloat will print out the hex float format that compilers tend to use.
The format character for this is "X", the long format is "hex float".
eFormatInstruction will print out disassembly with bytes and it will use the
current target's architecture. The format character for this is "i" (which
used to be being used for the integer format, but the integer format also has
"d", so we gave the "i" format to disassembly), the long format is
"instruction".
Mate the lldb::FormatterChoiceCriterion enumeration private as it should have
been from the start. It is very specialized and doesn't belong in the public
API.
llvm-svn: 143114
2011-10-28 01:55:14 +08:00
|
|
|
void
|
|
|
|
InstructionList::Dump (Stream *s,
|
|
|
|
bool show_address,
|
|
|
|
bool show_bytes,
|
|
|
|
const ExecutionContext* exe_ctx)
|
|
|
|
{
|
|
|
|
const uint32_t max_opcode_byte_size = GetMaxOpcocdeByteSize();
|
|
|
|
collection::const_iterator pos, begin, end;
|
2015-02-05 06:00:53 +08:00
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
const FormatEntity::Entry *disassembly_format = nullptr;
|
2015-02-05 06:00:53 +08:00
|
|
|
FormatEntity::Entry format;
|
|
|
|
if (exe_ctx && exe_ctx->HasTargetScope())
|
2014-10-11 07:07:36 +08:00
|
|
|
{
|
2015-02-05 06:00:53 +08:00
|
|
|
disassembly_format = exe_ctx->GetTargetRef().GetDebugger().GetDisassemblyFormat ();
|
2014-10-11 07:07:36 +08:00
|
|
|
}
|
2015-02-05 06:00:53 +08:00
|
|
|
else
|
|
|
|
{
|
|
|
|
FormatEntity::Parse("${addr}: ", format);
|
|
|
|
disassembly_format = &format;
|
|
|
|
}
|
|
|
|
|
Added support for the new ".apple_objc" accelerator tables. These tables are
in the same hashed format as the ".apple_names", but they map objective C
class names to all of the methods and class functions. We need to do this
because in the DWARF the methods for Objective C are never contained in the
class definition, they are scattered about at the translation unit level and
they don't even have attributes that say the are contained within the class
itself.
Added 3 new formats which can be used to display data:
eFormatAddressInfo
eFormatHexFloat
eFormatInstruction
eFormatAddressInfo describes an address such as function+offset and file+line,
or symbol + offset, or constant data (c string, 2, 4, 8, or 16 byte constants).
The format character for this is "A", the long format is "address".
eFormatHexFloat will print out the hex float format that compilers tend to use.
The format character for this is "X", the long format is "hex float".
eFormatInstruction will print out disassembly with bytes and it will use the
current target's architecture. The format character for this is "i" (which
used to be being used for the integer format, but the integer format also has
"d", so we gave the "i" format to disassembly), the long format is
"instruction".
Mate the lldb::FormatterChoiceCriterion enumeration private as it should have
been from the start. It is very specialized and doesn't belong in the public
API.
llvm-svn: 143114
2011-10-28 01:55:14 +08:00
|
|
|
for (begin = m_instructions.begin(), end = m_instructions.end(), pos = begin;
|
|
|
|
pos != end;
|
|
|
|
++pos)
|
|
|
|
{
|
|
|
|
if (pos != begin)
|
|
|
|
s->EOL();
|
2016-03-03 08:51:40 +08:00
|
|
|
(*pos)->Dump(s, max_opcode_byte_size, show_address, show_bytes, exe_ctx, nullptr, nullptr, disassembly_format, 0);
|
Added support for the new ".apple_objc" accelerator tables. These tables are
in the same hashed format as the ".apple_names", but they map objective C
class names to all of the methods and class functions. We need to do this
because in the DWARF the methods for Objective C are never contained in the
class definition, they are scattered about at the translation unit level and
they don't even have attributes that say the are contained within the class
itself.
Added 3 new formats which can be used to display data:
eFormatAddressInfo
eFormatHexFloat
eFormatInstruction
eFormatAddressInfo describes an address such as function+offset and file+line,
or symbol + offset, or constant data (c string, 2, 4, 8, or 16 byte constants).
The format character for this is "A", the long format is "address".
eFormatHexFloat will print out the hex float format that compilers tend to use.
The format character for this is "X", the long format is "hex float".
eFormatInstruction will print out disassembly with bytes and it will use the
current target's architecture. The format character for this is "i" (which
used to be being used for the integer format, but the integer format also has
"d", so we gave the "i" format to disassembly), the long format is
"instruction".
Mate the lldb::FormatterChoiceCriterion enumeration private as it should have
been from the start. It is very specialized and doesn't belong in the public
API.
llvm-svn: 143114
2011-10-28 01:55:14 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2010-06-09 00:52:24 +08:00
|
|
|
void
|
2010-10-06 11:09:58 +08:00
|
|
|
InstructionList::Clear()
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2016-03-03 08:51:40 +08:00
|
|
|
m_instructions.clear();
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
void
|
2010-10-06 11:09:58 +08:00
|
|
|
InstructionList::Append (lldb::InstructionSP &inst_sp)
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
|
|
|
if (inst_sp)
|
|
|
|
m_instructions.push_back(inst_sp);
|
|
|
|
}
|
|
|
|
|
2012-03-09 12:10:47 +08:00
|
|
|
uint32_t
|
2015-05-12 05:12:33 +08:00
|
|
|
InstructionList::GetIndexOfNextBranchInstruction(uint32_t start, Target &target) const
|
2012-03-09 12:10:47 +08:00
|
|
|
{
|
|
|
|
size_t num_instructions = m_instructions.size();
|
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
uint32_t next_branch = std::numeric_limits<uint32_t>::max();
|
2015-05-12 05:12:33 +08:00
|
|
|
size_t i;
|
|
|
|
for (i = start; i < num_instructions; i++)
|
2012-03-09 12:10:47 +08:00
|
|
|
{
|
|
|
|
if (m_instructions[i]->DoesBranch())
|
|
|
|
{
|
|
|
|
next_branch = i;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2015-05-12 05:12:33 +08:00
|
|
|
|
|
|
|
// Hexagon needs the first instruction of the packet with the branch.
|
|
|
|
// Go backwards until we find an instruction marked end-of-packet, or
|
|
|
|
// until we hit start.
|
|
|
|
if (target.GetArchitecture().GetTriple().getArch() == llvm::Triple::hexagon)
|
|
|
|
{
|
|
|
|
// If we didn't find a branch, find the last packet start.
|
2016-03-03 08:51:40 +08:00
|
|
|
if (next_branch == std::numeric_limits<uint32_t>::max())
|
2015-05-12 05:12:33 +08:00
|
|
|
{
|
|
|
|
i = num_instructions - 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
while (i > start)
|
|
|
|
{
|
|
|
|
--i;
|
|
|
|
|
|
|
|
Error error;
|
|
|
|
uint32_t inst_bytes;
|
|
|
|
bool prefer_file_cache = false; // Read from process if process is running
|
|
|
|
lldb::addr_t load_addr = LLDB_INVALID_ADDRESS;
|
|
|
|
target.ReadMemory(m_instructions[i]->GetAddress(),
|
|
|
|
prefer_file_cache,
|
|
|
|
&inst_bytes,
|
|
|
|
sizeof(inst_bytes),
|
|
|
|
error,
|
|
|
|
&load_addr);
|
|
|
|
// If we have an error reading memory, return start
|
|
|
|
if (!error.Success())
|
|
|
|
return start;
|
|
|
|
// check if this is the last instruction in a packet
|
|
|
|
// bits 15:14 will be 11b or 00b for a duplex
|
|
|
|
if (((inst_bytes & 0xC000) == 0xC000) ||
|
|
|
|
((inst_bytes & 0xC000) == 0x0000))
|
|
|
|
{
|
|
|
|
// instruction after this should be the start of next packet
|
|
|
|
next_branch = i + 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
if (next_branch == std::numeric_limits<uint32_t>::max())
|
2015-05-12 05:12:33 +08:00
|
|
|
{
|
|
|
|
// We couldn't find the previous packet, so return start
|
|
|
|
next_branch = start;
|
|
|
|
}
|
|
|
|
}
|
2012-03-09 12:10:47 +08:00
|
|
|
return next_branch;
|
|
|
|
}
|
|
|
|
|
|
|
|
uint32_t
|
2014-01-28 07:43:24 +08:00
|
|
|
InstructionList::GetIndexOfInstructionAtAddress (const Address &address)
|
2012-03-09 12:10:47 +08:00
|
|
|
{
|
2013-01-26 02:06:21 +08:00
|
|
|
size_t num_instructions = m_instructions.size();
|
2016-03-03 08:51:40 +08:00
|
|
|
uint32_t index = std::numeric_limits<uint32_t>::max();
|
2013-01-26 02:06:21 +08:00
|
|
|
for (size_t i = 0; i < num_instructions; i++)
|
2012-03-09 12:10:47 +08:00
|
|
|
{
|
|
|
|
if (m_instructions[i]->GetAddress() == address)
|
|
|
|
{
|
|
|
|
index = i;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return index;
|
|
|
|
}
|
2010-06-09 00:52:24 +08:00
|
|
|
|
2014-01-28 07:43:24 +08:00
|
|
|
uint32_t
|
|
|
|
InstructionList::GetIndexOfInstructionAtLoadAddress (lldb::addr_t load_addr, Target &target)
|
|
|
|
{
|
|
|
|
Address address;
|
|
|
|
address.SetLoadAddress(load_addr, &target);
|
|
|
|
return GetIndexOfInstructionAtAddress(address);
|
|
|
|
}
|
|
|
|
|
2010-06-09 00:52:24 +08:00
|
|
|
size_t
|
2013-03-29 07:42:53 +08:00
|
|
|
Disassembler::ParseInstructions (const ExecutionContext *exe_ctx,
|
|
|
|
const AddressRange &range,
|
|
|
|
Stream *error_strm_ptr,
|
|
|
|
bool prefer_file_cache)
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2011-09-22 12:58:26 +08:00
|
|
|
if (exe_ctx)
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2011-09-22 12:58:26 +08:00
|
|
|
Target *target = exe_ctx->GetTargetPtr();
|
|
|
|
const addr_t byte_size = range.GetByteSize();
|
2016-03-03 08:51:40 +08:00
|
|
|
if (target == nullptr || byte_size == 0 || !range.GetBaseAddress().IsValid())
|
2011-09-22 12:58:26 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
DataBufferHeap *heap_buffer = new DataBufferHeap (byte_size, '\0');
|
|
|
|
DataBufferSP data_sp(heap_buffer);
|
|
|
|
|
|
|
|
Error error;
|
2013-03-29 07:42:53 +08:00
|
|
|
lldb::addr_t load_addr = LLDB_INVALID_ADDRESS;
|
|
|
|
const size_t bytes_read = target->ReadMemory (range.GetBaseAddress(),
|
2011-09-22 12:58:26 +08:00
|
|
|
prefer_file_cache,
|
|
|
|
heap_buffer->GetBytes(),
|
|
|
|
heap_buffer->GetByteSize(),
|
2013-03-29 07:42:53 +08:00
|
|
|
error,
|
|
|
|
&load_addr);
|
2011-09-22 12:58:26 +08:00
|
|
|
|
|
|
|
if (bytes_read > 0)
|
|
|
|
{
|
|
|
|
if (bytes_read != heap_buffer->GetByteSize())
|
|
|
|
heap_buffer->SetByteSize (bytes_read);
|
|
|
|
DataExtractor data (data_sp,
|
|
|
|
m_arch.GetByteOrder(),
|
|
|
|
m_arch.GetAddressByteSize());
|
2013-03-29 07:42:53 +08:00
|
|
|
const bool data_from_file = load_addr == LLDB_INVALID_ADDRESS;
|
2016-03-03 08:51:40 +08:00
|
|
|
return DecodeInstructions(range.GetBaseAddress(), data, 0, std::numeric_limits<uint32_t>::max(), false,
|
|
|
|
data_from_file);
|
2011-09-22 12:58:26 +08:00
|
|
|
}
|
2012-05-26 01:05:55 +08:00
|
|
|
else if (error_strm_ptr)
|
|
|
|
{
|
|
|
|
const char *error_cstr = error.AsCString();
|
|
|
|
if (error_cstr)
|
|
|
|
{
|
|
|
|
error_strm_ptr->Printf("error: %s\n", error_cstr);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else if (error_strm_ptr)
|
|
|
|
{
|
|
|
|
error_strm_ptr->PutCString("error: invalid execution context\n");
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-03-22 09:48:42 +08:00
|
|
|
size_t
|
2013-03-29 07:42:53 +08:00
|
|
|
Disassembler::ParseInstructions (const ExecutionContext *exe_ctx,
|
|
|
|
const Address &start,
|
|
|
|
uint32_t num_instructions,
|
|
|
|
bool prefer_file_cache)
|
2011-03-22 09:48:42 +08:00
|
|
|
{
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
m_instruction_list.Clear();
|
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
if (exe_ctx == nullptr || num_instructions == 0 || !start.IsValid())
|
2011-03-22 09:48:42 +08:00
|
|
|
return 0;
|
|
|
|
|
2011-09-22 12:58:26 +08:00
|
|
|
Target *target = exe_ctx->GetTargetPtr();
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
// Calculate the max buffer size we will need in order to disassemble
|
|
|
|
const addr_t byte_size = num_instructions * m_arch.GetMaximumOpcodeByteSize();
|
2011-03-22 09:48:42 +08:00
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
if (target == nullptr || byte_size == 0)
|
2011-03-22 09:48:42 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
DataBufferHeap *heap_buffer = new DataBufferHeap (byte_size, '\0');
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
DataBufferSP data_sp (heap_buffer);
|
2011-03-22 09:48:42 +08:00
|
|
|
|
|
|
|
Error error;
|
2013-03-29 07:42:53 +08:00
|
|
|
lldb::addr_t load_addr = LLDB_INVALID_ADDRESS;
|
|
|
|
const size_t bytes_read = target->ReadMemory (start,
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
prefer_file_cache,
|
|
|
|
heap_buffer->GetBytes(),
|
|
|
|
byte_size,
|
2013-03-29 07:42:53 +08:00
|
|
|
error,
|
|
|
|
&load_addr);
|
|
|
|
|
|
|
|
const bool data_from_file = load_addr == LLDB_INVALID_ADDRESS;
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
|
|
|
|
if (bytes_read == 0)
|
|
|
|
return 0;
|
|
|
|
DataExtractor data (data_sp,
|
|
|
|
m_arch.GetByteOrder(),
|
|
|
|
m_arch.GetAddressByteSize());
|
|
|
|
|
|
|
|
const bool append_instructions = true;
|
|
|
|
DecodeInstructions (start,
|
|
|
|
data,
|
|
|
|
0,
|
|
|
|
num_instructions,
|
2013-03-29 07:42:53 +08:00
|
|
|
append_instructions,
|
|
|
|
data_from_file);
|
Added the ability to get the min and max instruction byte size for
an architecture into ArchSpec:
uint32_t
ArchSpec::GetMinimumOpcodeByteSize() const;
uint32_t
ArchSpec::GetMaximumOpcodeByteSize() const;
Added an AddressClass to the Instruction class in Disassembler.h.
This allows decoded instructions to know know if they are code,
code with alternate ISA (thumb), or even data which can be mixed
into code. The instruction does have an address, but it is a good
idea to cache this value so we don't have to look it up more than
once.
Fixed an issue in Opcode::SetOpcodeBytes() where the length wasn't
getting set.
Changed:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc);
To:
bool
SymbolContextList::AppendIfUnique (const SymbolContext& sc,
bool merge_symbol_into_function);
This function was typically being used when looking up functions
and symbols. Now if you lookup a function, then find the symbol,
they can be merged into the same symbol context and not cause
multiple symbol contexts to appear in a symbol context list that
describes the same function.
Fixed the SymbolContext not equal operator which was causing mixed
mode disassembly to not work ("disassembler --mixed --name main").
Modified the disassembler classes to know about the fact we know,
for a given architecture, what the min and max opcode byte sizes
are. The InstructionList class was modified to return the max
opcode byte size for all of the instructions in its list.
These two fixes means when disassemble a list of instructions and dump
them and show the opcode bytes, we can format the output more
intelligently when showing opcode bytes. This affects any architectures
that have varying opcode byte sizes (x86_64 and i386). Knowing the max
opcode byte size also helps us to be able to disassemble N instructions
without having to re-read data if we didn't read enough bytes.
Added the ability to set the architecture for the disassemble command.
This means you can easily cross disassemble data for any supported
architecture. I also added the ability to specify "thumb" as an
architecture so that we can force disassembly into thumb mode when
needed. In GDB this was done using a hack of specifying an odd
address when disassembling. I don't want to repeat this hack in LLDB,
so the auto detection between ARM and thumb is failing, just specify
thumb when disassembling:
(lldb) disassemble --arch thumb --name main
You can also have data in say an x86_64 file executable and disassemble
data as any other supported architecture:
% lldb a.out
Current executable set to 'a.out' (x86_64).
(lldb) b main
(lldb) run
(lldb) disassemble --arch thumb --count 2 --start-address 0x0000000100001080 --bytes
0x100001080: 0xb580 push {r7, lr}
0x100001082: 0xaf00 add r7, sp, #0
Fixed Target::ReadMemory(...) to be able to deal with Address argument object
that isn't section offset. When an address object was supplied that was
out on the heap or stack, target read memory would fail. Disassembly uses
Target::ReadMemory(...), and the example above where we disassembler thumb
opcodes in an x86 binary was failing do to this bug.
llvm-svn: 128347
2011-03-27 03:14:58 +08:00
|
|
|
|
2011-03-22 09:48:42 +08:00
|
|
|
return m_instruction_list.GetSize();
|
|
|
|
}
|
|
|
|
|
2010-06-09 00:52:24 +08:00
|
|
|
//----------------------------------------------------------------------
|
|
|
|
// Disassembler copy constructor
|
|
|
|
//----------------------------------------------------------------------
|
2013-03-02 08:26:47 +08:00
|
|
|
Disassembler::Disassembler(const ArchSpec& arch, const char *flavor) :
|
2010-06-09 00:52:24 +08:00
|
|
|
m_arch (arch),
|
|
|
|
m_instruction_list(),
|
2013-03-02 08:26:47 +08:00
|
|
|
m_base_addr(LLDB_INVALID_ADDRESS),
|
|
|
|
m_flavor ()
|
2010-06-09 00:52:24 +08:00
|
|
|
{
|
2016-03-03 08:51:40 +08:00
|
|
|
if (flavor == nullptr)
|
2013-03-02 08:26:47 +08:00
|
|
|
m_flavor.assign("default");
|
|
|
|
else
|
|
|
|
m_flavor.assign(flavor);
|
2015-02-07 14:03:49 +08:00
|
|
|
|
|
|
|
// If this is an arm variant that can only include thumb (T16, T32)
|
|
|
|
// instructions, force the arch triple to be "thumbv.." instead of
|
|
|
|
// "armv..."
|
2015-08-21 08:13:37 +08:00
|
|
|
if ((arch.GetTriple().getArch() == llvm::Triple::arm || arch.GetTriple().getArch() == llvm::Triple::thumb)
|
2015-02-07 14:03:49 +08:00
|
|
|
&& (arch.GetCore() == ArchSpec::Core::eCore_arm_armv7m
|
|
|
|
|| arch.GetCore() == ArchSpec::Core::eCore_arm_armv7em
|
|
|
|
|| arch.GetCore() == ArchSpec::Core::eCore_arm_armv6m))
|
|
|
|
{
|
|
|
|
std::string thumb_arch_name (arch.GetTriple().getArchName().str());
|
|
|
|
// Replace "arm" with "thumb" so we get all thumb variants correct
|
|
|
|
if (thumb_arch_name.size() > 3)
|
|
|
|
{
|
|
|
|
thumb_arch_name.erase(0, 3);
|
|
|
|
thumb_arch_name.insert(0, "thumb");
|
|
|
|
}
|
|
|
|
m_arch.SetTriple (thumb_arch_name.c_str());
|
|
|
|
}
|
2010-06-09 00:52:24 +08:00
|
|
|
}
|
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
Disassembler::~Disassembler() = default;
|
2010-06-09 00:52:24 +08:00
|
|
|
|
2010-10-06 11:09:58 +08:00
|
|
|
InstructionList &
|
2010-06-09 00:52:24 +08:00
|
|
|
Disassembler::GetInstructionList ()
|
|
|
|
{
|
|
|
|
return m_instruction_list;
|
|
|
|
}
|
|
|
|
|
2010-10-06 11:09:58 +08:00
|
|
|
const InstructionList &
|
2010-06-09 00:52:24 +08:00
|
|
|
Disassembler::GetInstructionList () const
|
|
|
|
{
|
|
|
|
return m_instruction_list;
|
|
|
|
}
|
2011-04-20 07:30:03 +08:00
|
|
|
|
|
|
|
//----------------------------------------------------------------------
|
|
|
|
// Class PseudoInstruction
|
|
|
|
//----------------------------------------------------------------------
|
2016-03-03 08:51:40 +08:00
|
|
|
|
2011-04-20 07:30:03 +08:00
|
|
|
PseudoInstruction::PseudoInstruction () :
|
|
|
|
Instruction (Address(), eAddressClassUnknown),
|
|
|
|
m_description ()
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
2016-03-03 08:51:40 +08:00
|
|
|
PseudoInstruction::~PseudoInstruction() = default;
|
2011-04-20 07:30:03 +08:00
|
|
|
|
|
|
|
bool
|
2013-03-13 09:55:16 +08:00
|
|
|
PseudoInstruction::DoesBranch ()
|
2011-04-20 07:30:03 +08:00
|
|
|
{
|
|
|
|
// This is NOT a valid question for a pseudo instruction.
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2015-08-26 14:04:54 +08:00
|
|
|
bool
|
|
|
|
PseudoInstruction::HasDelaySlot ()
|
|
|
|
{
|
|
|
|
// This is NOT a valid question for a pseudo instruction.
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2011-04-20 07:30:03 +08:00
|
|
|
size_t
|
|
|
|
PseudoInstruction::Decode (const lldb_private::Disassembler &disassembler,
|
|
|
|
const lldb_private::DataExtractor &data,
|
2013-01-26 02:06:21 +08:00
|
|
|
lldb::offset_t data_offset)
|
2011-04-20 07:30:03 +08:00
|
|
|
{
|
|
|
|
return m_opcode.GetByteSize();
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
|
|
|
PseudoInstruction::SetOpcode (size_t opcode_size, void *opcode_data)
|
|
|
|
{
|
|
|
|
if (!opcode_data)
|
|
|
|
return;
|
|
|
|
|
|
|
|
switch (opcode_size)
|
|
|
|
{
|
|
|
|
case 8:
|
|
|
|
{
|
|
|
|
uint8_t value8 = *((uint8_t *) opcode_data);
|
2013-12-10 03:45:33 +08:00
|
|
|
m_opcode.SetOpcode8 (value8, eByteOrderInvalid);
|
2011-04-20 07:30:03 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case 16:
|
|
|
|
{
|
|
|
|
uint16_t value16 = *((uint16_t *) opcode_data);
|
2013-12-10 03:45:33 +08:00
|
|
|
m_opcode.SetOpcode16 (value16, eByteOrderInvalid);
|
2011-04-20 07:30:03 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case 32:
|
|
|
|
{
|
|
|
|
uint32_t value32 = *((uint32_t *) opcode_data);
|
2013-12-10 03:45:33 +08:00
|
|
|
m_opcode.SetOpcode32 (value32, eByteOrderInvalid);
|
2011-04-20 07:30:03 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case 64:
|
|
|
|
{
|
|
|
|
uint64_t value64 = *((uint64_t *) opcode_data);
|
2013-12-10 03:45:33 +08:00
|
|
|
m_opcode.SetOpcode64 (value64, eByteOrderInvalid);
|
2011-04-20 07:30:03 +08:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
|
|
|
PseudoInstruction::SetDescription (const char *description)
|
|
|
|
{
|
|
|
|
if (description && strlen (description) > 0)
|
|
|
|
m_description = description;
|
|
|
|
}
|