2007-12-30
 

target_remove_breakpoint

target_insert_breakpoint (bp_tgt)

target_remove_hw_breakpoint (bp_tgt)

target_insert_hw_breakpoint (bp_tgt)

TARGET_REGION_OK_FOR_HW_WATCHPOINT (addr, len)

Return non-zero if hardware watchpoints can be used to watch a region whose address is addr and whose length

in bytes is len.

target_insert_watchpoint (addr, len, type)

target_remove_watchpoint (addr, len, type)

Insert or remove a hardware watchpoint starting at addr, for len bytes. type is the watchpoint type, one of the

possible values of the enumerated data type target_hw_bp_type, defined by `breakpoint.h' as follows:

 enum target_hw_bp_type
   {
     hw_write   = 0, /* Common (write) HW watchpoint */
     hw_read    = 1, /* Read    HW watchpoint */
     hw_access  = 2, /* Access (read or write) HW watchpoint */
     hw_execute = 3  /* Execute HW breakpoint */
   };

These two macros should return 0 for success, non-zero for failure.

target_stopped_data_address (addr_p)

If the inferior has some watchpoint that triggered, place the address associated with the watchpoint at the location pointed to by addr_p and return non-zero. Otherwise, return zero. Note that this primitive is used by GDB only ontargets that support data-read or data-access type watchpoints, so targets that have support only for data-write watchpoints need not implement these primitives.

HAVE_STEPPABLE_WATCHPOINT

If defined to a non-zero value, it is not necessary to disable a watchpoint to step over it.

int gdbarch_have_nonsteppable_watchpoint (gdbarch)

If it returns a non-zero value, GDB should disable a watchpoint to step the inferior over it.

HAVE_CONTINUABLE_WATCHPOINT

If defined to a non-zero value, it is possible to continue the inferior after a watchpoint has been hit.

CANNOT_STEP_HW_WATCHPOINTS

If this is defined to a non-zero value, GDB will remove all watchpoints before stepping the inferior.

STOPPED_BY_WATCHPOINT (wait_status)

Return non-zero if stopped by a watchpoint. wait_status is of the type struct target_waitstatus, defined by `target.h'. Normally, this macro is defined to invoke the function pointed to by the to_stopped_by_watchpoint member of the structure(of the type target_ops, defined on `target.h') that describes the target-specific operations; to_stopped_by_watchpoint ignores the wait_status argument.

GDB does not require the non-zero value returned by STOPPED_BY_WATCHPOINT to be 100% correct, so if a target cannot  determine for sure whether the inferior stopped due to a watchpoint, it could return non-zero "just in case".

x86 Watchpoints :请参考 GDBINT英文原版.

Observing changes in GDB internals (眼拙,未能明白讲的是啥)

Symbol Handling

当需要读取符号表的时候,symbol_file_add 会调用此函数,参数是新分配的一个fym_fns,其bfd field 是新符号表对应的BFD.

xyz_new_init()

放弃当前的symbols时,symbol_file_add 调用此函数.

xyz_symfile_read(struct sym_fns *sf, CORE_ADDR addr, int mainline)

symbol_file_add 调用此函数获取具体的符号表:psymtabs or symtabs. sf 是调用初始化函数时的那个sym_fns.

xyz_psymtab_to_symtab (struct partial_symtab *pst)

Partial Symbol Tables

• Full symbol tables (symtabs) :包含关于符号和地址的主信息.

• Partial symbol tables (psymtabs):包含足够的信息去读取full symbol table.

• Minimal symbol tables (msymtabs):非调试用symbols.

GDB使用的内部类型.

Type Codes (e.g., TYPE_CODE_PTR, TYPE_CODE_ARRAY).

属于基本类型或者派生类型. 典型情况下几个基本类型 FT_* 映射到一种TYPE_CODE_* , 通过其bit长度,是否是signed的等熟悉进行区分.

历史原因造成的,对应于基本类型.(GDB的维护人员其实打算把这些internal type给搞掉的: builtin_type_int (gdbtypes.c)基本上和 a TYPE_CODE_INT (c-lang.c)是一样的.(对应于FT_INTEGER).区别在于builtin_type 不和任何objfile有关联,而`c-lang.c' 则搞了很多 TYPE_CODE_INT, 每个都和特定的objfile相关.

COFF format :  System V Release 3 (SVR3) Unix, 符号表有缺陷(比如include的头文件不能解析),对应文件coffread.c

ECOFF         :  COFF 扩展版本,Mips and Alpha workstations, mipsread.c

XCOFF         :  IBM RS/6000 running AIX

PE              : Windows 95 and NT use, 基本上是COFF.

ELF             : System V Release 4 (SVR4) Unix. ELF 类似COFF但是解决了COFF的许多不足, elfread.c

SOM            : HP(not to be confused with IBM's SOM, which is a cross-language ABI),`somread.c'.

Debugging File Formats

独立于obj文件的调试信息.

Stabs

stabs原本是a.out中的信息,但是COFF,ELF和其他obj文件也含有有这个信息.dbxread.c:基本的stabs处理和封装,stabsread.c:干活的地方.

COFF :coff文件也含有私有的debugging信息,不太常用,扩展性不好.

Mips debug (Third Eye): ECOFF 含有的特殊调试信息, mdebugread.c

DWARF 2 : DWARF 1的下一版,但和第一版不兼容,dwarf2read.c

SOM: 和COFF类似.

Adding a New Symbol Reader to GDB

如果使用现有的obj文件,就简单的多.否则,你需要先将新的 obj文件支持加到BFD. GDB 通过一组swaping 函数(request comment),使用具体的BFD接口, 对于特殊的target(如COFF),可能还需要一层封装,因为不同的platform可能不一样,这些接口应该在bfd/libxyz.h中进行描述.

Memory Management for Symbol Files

一个symbol file的符号信息,存储在objfile_obstack里(request comment), unload 一个objfile的时候内存自动释放. 所以也不要在一个obj文件中引用另一个obj的符号. 和用户相关的一些数(request comment)据和type也是存在于这个obstack里的,但是objfile unload的时候会copy到global的内存里,所以不会丢失.

Language Support

gdb/config/arch/xyz.mh'

包含host和native的配置.host configuration 现在由Autoconf处理,HOST信息包含一些定义:XM_FILE=xm-xyz.h,还可能有CC, SYSV_DEFINE, XM_CFLAGS, XM_ADD_FILES, XM_CLIBS, XM_CDEPS, 请参考"Makefile.in".

gdb/config/arch/xm-xyz.h'

这个文件以前包含在xyz机器上运行gdb需要的一些定义和信息,现在通过Autoconf来实现.新的host和native配置不需要这个文件了.

完成GDB的配置后,需要很多的宏需要定义.这里列出了一些:

GDBINIT_FILENAME  : GDB初始化文件名,一般是.gdbinit

NO_STD_REGS : This macro is deprecated.

SIGWINCH_HANDLER 

If your host defines SIGWINCH, you can define this to be the name of a function to be called if SIGWINCH is received.

SIGWINCH_HANDLER_BODY

Define this to expand into code that will define the function named by the expansion of SIGWINCH_HANDLER.

ALIGN_STACK_ON_STARTUP

Define this if your system is of a sort that will crash in tgetent if the stack happens not to be longword-aligned when main is called. This is a rare situation, but is known to occur on several different types of systems.

CRLF_SOURCE_FILES

Define this if host files use \r\n rather than \n as a line terminator. This will cause source file listings to omit \r characters when printing and it will allow \r\n line endings of files which are "sourced" by gdb. It must be possible to open files in binary mode using O_BINARY or, for fopen, "rb".

DEFAULT_PROMPT

The default value of the prompt string (normally "(gdb) ").

DEV_TTY

The name of the generic TTY device, defaults to "/dev/tty".

FOPEN_RB

Define this if binary files are opened the same way as text files.

HAVE_MMAP

In some cases, use the system call mmap for reading symbol tables. For some machines this allows for sharing and quick updates.

HAVE_TERMIO

Define this if the host system has termio.h.

INT_MAX INT_MIN LONG_MAX UINT_MAX ULONG_MAX

Values for host-side constants.

ISATTY

Substitute for isatty, if not available.

LONGEST

This is the longest integer type available on the host. If not defined, it will default to long long or long, depending on CC_HAS_LONG_LONG.

CC_HAS_LONG_LONG

Define this if the host C compiler supports long long. This is set by the configure script.

PRINTF_HAS_LONG_LONG

Define this if the host can handle printing of long long integers via the printf format conversion specifier ll. This is set by the configure script.

HAVE_LONG_DOUBLE

Define this if the host C compiler supports long double. This is set by the configure script.

PRINTF_HAS_LONG_DOUBLE

Define this if the host can handle printing of long double float-point numbers via the printf format conversion specifier Lg. This is set by the configure script.

SCANF_HAS_LONG_DOUBLE

Define this if the host can handle the parsing of long double float-point numbers via the scanf format conversion specifier Lg. This is set by the configure script.

LSEEK_NOT_LINEAR

Define this if lseek (n) does not necessarily move to byte number n in the file. This is only used when reading source files. It is normally faster to define CRLF_SOURCE_FILES when possible.

L_SET

This macro is used as the argument to lseek (or, most commonly, bfd_seek). FIXME, should be replaced by SEEK_SET instead, which is the POSIX equivalent.

NORETURN

If defined, this should be one or more tokens, such as volatile, that can be used in both the declaration and definition of functions to indicate that they never return. The default is already set correctly if compiling with GCC. This will almost never need to be defined.

ATTR_NORETURN

If defined, this should be one or more tokens, such as __attribute__ ((noreturn)), that can be used in the declarations of functions to indicate that they never return. The default is already set correctly if compiling with GCC. This will almost never need to be defined.

SEEK_CUR

SEEK_SET

Define these to appropriate value for the system lseek, if not already defined.

STOP_SIGNAL

This is the signal for stopping GDB. Defaults to SIGTSTP. (Only redefined for the Convex.)

USG

Means that System V (prior to SVR4) include files are in use. (FIXME: This symbol is abused in `infrun.c', `regex.c', and `utils.c' for other things, at the moment.)

lint

Define this to help placate lint in some situations.

volatile

Define this to override the defaults of __volatile__ or /**/.

GDB_OSABI_UNINITIALIZED : Used for struct gdbarch_info if ABI is still uninitialized.

GDB_OSABI_UNKNOWN: The ABI of the inferior is unknown. The default gdbarch settings for the architecture will be used.

GDB_OSABI_SVR4:UNIX System V Release 4.

GDB_OSABI_LINUX  GDB_OSABI_WINCE  GDB_OSABI_GO32(DJGPP) 

OS ABI framework 接口函数:
const char *gdbarch_osabi_name (enum gdb_osabi osabi)

void gdbarch_register_osabi (...)

void gdbarch_register_osabi_sniffer (.....)

enum gdb_osabi gdbarch_lookup_osabi (bfd *abfd)

void generic_elf_osabi_sniff_abi_tag_sections (bfd *abfd, asection *sect, void *obj)

   每个gdbarch对应一个BFD architecture,并有一个常量bfd_arch_xxx. 通过函数register_gdbarch_init 注册gdbarch ,通常在 _initialize_filename 中调用注册函数,这样就加入了GDB启动流程. (细节请参考GDBINT)

GDB 采用的target machine模型是十分简单的:GDB 假定一种机器包含一组寄存器和一块内存,每个寄存器有一定的大小.GDB把寄存器的一切封装到像 gdbarch_register_name 这样的一系列函数中,继而做到正确处理.GDB 支持各种endian的机器:big-endian, little-endian, bi-endian(request commnet).

Pointers Are Not Always Addresses

有些特殊的平台,同一个word,其用于code pointer和datapointe的时候有不同的解释,比如D10V,同样是0xC020,当用于数据地址的时候代表地址0xC020,但是用于代码地址的时候,需要转换一下,代表0x30080(详 见D10V或者GDBINT). 所以在GDB中,对于address和pointer是区分对待的.address代表字节地址,而pointer指向特定的数据类型:data or code or other...., pointer需要转换以下才是address,大部分平台是相等的值,部分平台需要移位等操作. 但是大部分平台上,这两者是等价的,以下函数用于转换pointer和address: (详见GDBINT)

CORE_ADDR extract_typed_address (void *buf, struct type *type)

CORE_ADDR store_typed_address (void *buf, struct type *type, CORE_ADDR addr)

CORE_ADDR value_as_address (struct value *val)

CORE_ADDR value_from_pointer (struct type *type, CORE_ADDR addr)

CORE_ADDR gdbarch_pointer_to_address (struct gdbarch *current_gdbarch, struct type *type, char *buf)

void gdbarch_address_to_pointer (struct gdbarch *current_gdbarch, struct type *type, char *buf, CORE_ADDR addr)

Address Classes

当不同种类的address可以而从调试信息中(如DWARF2)获取时, 应该定义以下宏来支持GDB认知这些区别,或者给GDB user提供类型信息.

int gdbarch_address_class_type_flags (struct gdbarch *current_gdbarch, int byte_size, int dwarf2_addr_class)

Raw and Virtual Register Representations

注:这里描述的内容已经过时,请参考下一节.

一些体系结构在寄存器和内存中对一个值的表达方式是不同的.用GDB的术语来定义,在寄存器中的表达方式叫raw representation,在内存中的表达方式叫做virtual representation.GDB用struct value objects来表示这种差异.

在几乎所有的体系上,数据类型的virtual和raw representations 是一致的.但是有例外,如:

  x86 提示支持一种80-bit long double type,但是存储到内存时,占用12个byte:浮点数占用头是个byte,后两个byte没有用(对齐要求),因此,x86 80-bit floating-point type 就是raw representation,twelve-byte loosely-packed arrangement 是virtual representation.

  另外,一些64-bit MIPS,使用64-bit的寄存器当作32-bit registers使用,高位是垃圾数据.所以,64-bit form, 带有垃圾数据的这种表达形式是raw representation, 32-bit 形式是virtual representation.

  GDB's register file, registers, 采用raw format, GDB remote protocol transmits register values也是raw format.如要如下的函数支持raw和virtual模式的转换.

int REGISTER_CONVERTIBLE (int reg)   /*是否需要转换*/

int DEPRECATED_REGISTER_RAW_SIZE (int reg) /*返回raw格式占用的字节数*/

int DEPRECATED_REGISTER_VIRTUAL_SIZE (int reg)/*vitural 格式占用的字节数*/

struct type *DEPRECATED_REGISTER_VIRTUAL_TYPE (int reg) /*返回vitual representation*/

void REGISTER_CONVERT_TO_VIRTUAL (int reg, struct type *type, char *from, char *to)

void REGISTER_CONVERT_TO_RAW (struct type *type, int reg, char *from, char *to)

Maintainer's note: GDB操作寄存器的方式正在变革中,详见 A.R. Index and Bug Tracking Database 2002.

这里解决的问题和上一节一样:x86 80-bit 浮点寄存器,还有Alpha 可以把32 bit integer values 存到floating-point registers.

这种问题也可以定义如下的宏来解决:
int gdbarch_convert_register_p (struct gdbarch *gdbarch, int reg) /*是否必要转换*/

void gdbarch_register_to_value (struct gdbarch *gdbarch, int reg, struct type *type, char *from, char *to)

void gdbarch_value_to_register (struct gdbarch *gdbarch, struct type *type, int reg, char *from, char *to)/*转换到raw format*/

void REGISTER_CONVERT_TO_TYPE (int regnum, struct type *type, char *buf) /*refer to `mips-tdep.c',事情很复杂(request comments)

Target Conditionals

这里讨论的宏和函数用于定义新的target.

Register and Bit structure:

BITS_BIG_ENDIAN :如果target的bis endin和byte endianness不同,则定义此值.

TARGET_CHAR_BIT /* defaults to 8.*/


int gdbarch_char_signed (gdbarch) /*non-zero, 如果char通常是带符号的整数的话*//*unsigned on IBM S/390, RS6000, and PowerPC*/


int gdbarch_double_bit (gdbarch) /*默认是 8 * TARGET_CHAR_BIT.*/


int gdbarch_float_bit (gdbarch) /* defaults to 4 * TARGET_CHAR_BIT*/


int gdbarch_int_bit (gdbarch) /*默认是 4 * TARGET_CHAR_BIT.*/


int gdbarch_long_bit (gdbarch)/*defaults to 4 * TARGET_CHAR_BIT.*/


int gdbarch_long_double_bit (gdbarch)/*defaults to 2 * gdbarch_double_bit (gdbarch).*/


int gdbarch_long_long_bit (gdbarch)/*defaults to 2 * gdbarch_long_bit (gdbarch).*/


int gdbarch_ptr_bit (gdbarch)/* defaults to gdbarch_int_bit (gdbarch).*/


int gdbarch_short_bit (gdbarch)/* defaults to 2 * TARGET_CHAR_BIT.*/


int gdbarch_cannot_store_register (gdbarch, regnum)/*比如PC,状态寄存器等*/


int gdbarch_convert_register_p (gdbarch, regnum, struct type *type)/*need raw <-> vitral ?*/

void gdbarch_value_to_register (gdbarch, frame, type, buf) /* convert value to raw (register)*/

int gdbarch_sp_regnum (gdbarch) /*返回堆栈指针寄存器的编号*/
register_reggroup_p (gdbarch, regnum, reggroup)/*float_reggroup,vector_reggroup...restore_reggroup..*/
DEPRECATED_REGISTER_VIRTUAL_SIZE (reg)/*raw and virtal for reg and memory*/
DEPRECATED_REGISTER_VIRTUAL_TYPE (reg)
struct type *register_type (gdbarch, reg)/*refer to raw virtual register representations*/
REGISTER_CONVERT_TO_VIRTUAL(reg, type, from, to)

int gdbarch_stab_reg_to_regnum (gdbarch, stab_regnr)/*optinal,stab register stab_regnr 到GDB regnum转换*/
void gdbarch_store_return_value (gdbarch, type, regcache, valbuf)/*been deprecated by gdbarch_return_value*/

gdbarch_ps_regnum (gdbarch  /* 返回处理器状态寄存器编号用于解析 "$ps"*/

int gdbarch_sdb_reg_to_regnum (gdbarch, sdb_regnr)/*sdb_regnr到GDB regnum的转换,optional*/

Stack Frame:


SOFTWARE_SINGLE_STEP_P() /*如果无hardware的单步机制*/


SOFTWARE_SINGLE_STEP(signal, insert_breakpoints_p)/*refer to `sparc-tdep.c' and `rs6000-tdep.c'*/


DEPRECATED_FRAME_SAVED_PC(frame) /* 返回frame中的PC地址*//*deprecated. See gdbarch_unwind_pc.*/


CORE_ADDR gdbarch_unwind_pc (next_frame)/*Return the retrun address*/


CORE_ADDR gdbarch_unwind_sp (gdbarch, next_frame) /* Return the frame's inner most stack address*/


CORE_ADDR frame_align (gdbarch, address)


int gdbarch_frame_red_zone_size (gdbarch) /*ABI reserverd stack address,refer AMD64 (nee x86-64) ABI */ 


DEPRECATED_FRAME_CHAIN(frame) /* retern calling frame*/


DEPRECATED_FRAME_CHAIN_VALID(chain, thisframe)/* 0:最外层frame,此时PC一般位于ctr0.o*/


DEPRECATED_FRAME_INIT_SAVED_REGS(frame)/* See `frame.h'*/


int gdbarch_frame_num_args (gdbarch, frame)/*返回frame的参数个数(request comments)*/

void gdbarch_extract_return_value (gdbarch, type, regbuf, valbuf) /*deprecated by gdbarch_return_value */

DEPRECATED_FP_REGNUM /* need to be defined if DEPRECATED_TARGET_READ_FP is not defined.*/

DEPRECATED_FRAMELESS_FUNCTION_INVOCATION(fi)/*如果函数调用fi无stack frame,返回0*/

FUNCTION_EPILOGUE_SIZE /*比如COFF targets,无epilogue信息*/

DEPRECATED_FUNCTION_START_OFFSET /*一般是0,从函数地址到其第一条指令的字节长度,如VAX*/
int gdbarch_inner_than (gdbarch, lhs, rhs) /* is lhs inner than rhs?*/

gdbarch_get_longjmp_target /*longjmp要跳到的PC地址*/

Address pointer:

int address_class_name_to_type_flags (gdbarch, name, type_flags_ptr) /*如合法,设置class flag*/
int address_class_name_to_type_flags_p (gdbarch) /* address_class_name_to_type_flags是否被定义*/


int gdbarch_address_class_type_flags (gdbarch, byte_size, dwarf2_addr_class)/*从指针的bytesize和调试信息,给出address class type*/

int gdbarch_address_class_type_flags_p (gdbarch)/* gdbarch_address_class_type_flags_p 是否被定义*/

const char *gdbarch_address_class_type_flags_to_name (gdbarch, type_flags) /*address class to name*/

int gdbarch_address_class_type_flags_to_name_p (gdbarch) /*gdbarch_address_class_type_flags_to_name 是否定义*/

void gdbarch_address_to_pointer (gdbarch, type, buf, addr) /*将类型为type的 addr转换为指针存入buf*/

CORE_ADDR gdbarch_integer_to_address (gdbarch, type, buf) /*如果需要从int到address的转换,定义此函数*/

CORE_ADDR gdbarch_pointer_to_address (gdbarch, type, buf)

DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS(regbuf)/*When defined,extract raw regbuf to CORE_ADDR */

DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS_P() /* Predicate for DEPRECATED_EXTRACT_STRUCT_VALUE_ADDRESS.*/

Compiler and BFD:

int gdbarch_believe_pcc_promotion (gdbarch)/*编译其是否把char 或者short提升为int?*/


int gdbarch_dwarf2_reg_to_regnum (gdbarch, dwarf2_regnr)/*for dwarf2*/


int gdbarch_ecoff_reg_to_regnum (gdbarch, ecoff_regnr)

int gdbarch_dwarf_reg_to_regnum (gdbarch, dwarf_regnr)/*dwarf_regnr 到GDB regnum转换,should optional*/

GCC_COMPILED_FLAG_SYMBOL/*(Currently only defined for the Delta 68.)*/

GCC2_COMPILED_FLAG_SYMBOL/*(Currently only defined for the Delta 68.)*/

SOFUN_ADDRESS_MAYBE_MISSING /*表示stabs-format debugging information中N_FUN(函数起始和结束地址)和N_SO(compilation unit 的起始和结束地址)是0, 需要以其他方式确定*/  

PROCESS_LINENUMBER_HOOK /*XCOFF reading hook*/

VARIABLES_INSIDE_BLOCK (desc, gcc_p) /*For dbx-style debugging information, if the compiler puts variable declarations inside LBRAC/RBRAC blocks, this should be defined to be nonzero. refer to GDBINT*/

BreakPoint wathpoint

BREAKPOINT :定义breakpoint使用的指令,不能长于最短指令.如果采用trap则不必定义.如果才用非法指令或者其他值则必须定义TARGET_HAS_HARDWARE_WATCHPOINTS

BIG_BREAKPOINT LITTLE_BREAKPOINT :类似BREAKPOINT, 但是用于 bi-endian targets.

int gdbarch_memory_insert_breakpoint (gdbarch, bp_tgt)

gdbarch_memory_remove_breakpoint (gdbarch, bp_tgt)  /*有默认函数,除非有特殊处理,否则无需这两个函数.*/

Motorola M68K:

BPT_VECTOR : Define this to be the 4-bit location of the breakpoint trap vector. If not defined, it will default to 0xf.
REMOTE_BPT_VECTOR: Defaults to 1.

const char *gdbarch_name_of_malloc (gdbarch) : A string containing the name of the function to call in order to allocate some memory in the inferior. The default value is "malloc".

Adding a New Target

为了建立一个target,需要添加以下文件:

gdb/config/arch/ttt.mt

      makefile,定义 `TDEPFILES=...' and `TDEPLIBS=...'. 指定`TM_FILE= tm-ttt.h':描述tty的头文件. 

gdb/ttt-tdep.c 

或许根本就不需要这个文件.如果 `tm-ttt.h' 太复杂,可以把函数放到这里

gdb/arch-tdep.c

gdb/arch-tdep.h

registers, stack ...的描述, it is included by `ttt-tdep.h'. /*为cpu相同的target所共享*/

gdb/config/arch/tm-ttt.h

`tm.h'是这个文件的一个link,configure自动创建,registers, stack frame format instructions描述.新的target不应该使用这个文件了.

`gdb/config/arch/tm-arch.h'

新target无需此文件了. 同cpu的target可以共享.

如果是为一个新的os添加支持(cpu已存在),则只需要添加 `config/tm-os.h' ,描述此os的特殊facilities即可,比如extra symbol table info; the breakpoint instruction needed; 等等),然后写一个 `arch/tm-os.h',就include`tm-arch.h'和config/tm-os.h即可.

Target Descriptions

Target Descriptions Implementation

  GDB 连接到一个new target之前,恢复成一个默认的target,连接完成后,通过 target_find_description 找到新的target. 新的target可以由用户指定的XML文件来获取,或者远程的qXfer:features:read (also XML), 或者在target vector里定制的to_read_description函数来获取target.比如远程的target支持判断一个MIPS target是32-bit还是64-bit(根据 `g' packet的大小).

  发现一个target description后,GDB 通过gdbarch_update_p创建新的gdbarch. 优先根据 `<architecture>'来确定调用那个 gdbarch initialization 函数.然后,调用初始化函数,初始化函数可以调整这个architecture:比如通过to_read_description确定属性集.

Adding Target Described Register Support

Target descriptions通过体系特定的支持函数,可以报告更多的寄存器到target.

一个target description或者没有寄存器,或者有完整的寄存器集.

如果 tdesc_has_registers 返回1, 代表这个description包含寄存器,architecture's gdbarch_init函数应该完成如下功能:

• 通过tdesc_data_alloc 分配内存(在搜索匹配的gdbarch后者分配新的之前进行)

• 通过tdesc_find_feature,根据name确定标准属性集.

• 通过tdesc_numbered_register 和 tdesc_numbered_register_choices 确定standard features需要的registers.

• 如果需要的feature缺失,返回NULL, 如standard feature缺失必要寄存器,返回NULL.

• 释放分配的数据,除非调用了tdesc_use_registers.
  

• Call set_gdbarch_num_regs as usual, with a number higher than any fixed number passed to tdesc_numbered_register.

• 返回新创建的gdbarch之前,调用 tdesc_use_registers.
  

调用tdesc_use_registers后, 这个architecture的register_name, register_type, 和 register_reggroup_p routines将不会被调用;这些信息将从target description获取. num_regs 或许增加,因为可能有新的寄存器加进来.

Pseudo-registers require some extra care:

• Using tdesc_numbered_register allows the architecture to give constant register numbers to standard architectural registers, e.g. as an enum in `arch-tdep.h'. But because pseudo-registers are always numbered above num_regs, which may be increased by the description, constant numbers can not be used for pseudos. They must be numbered relative to num_regs instead.

• The description will not describe pseudo-registers, so the architecture must call set_tdesc_pseudo_register_name, set_tdesc_pseudo_register_type, and set_tdesc_pseudo_register_reggroup_p to supply routines describing pseudo registers. These routines will be passed internal register numbers, so the same routines used for the gdbarch equivalents are usually suitable.

http://sourceware.org/ml/gdb-patches/2000-07/msg00127.html

Target Vector Definition

  target vector 定义GD对target system的抽象接口.GDB 包含30-40这样的接口,但每个configuration只有几个.

Managing Execution State

  target vector 状态:

completely inactive :not pushed on the target stack

active but not running:pushed, but not connected to a fully manifested inferior

completely active (pushed, with an accessible inferior).

persistent connections to a target even when the target has exited or not yet started.如:

使用target sim连接到simulator是,并不创建进程,寄存器和内存在run前都不可访问.类似的, kill之后,程序也不能继续运行.但是这两种情况下GDB仍然连接到simulator.

如果target仅仅支持complete activation,则在其to_open函数中,应该使用push_target将自己压入堆栈(rfc), 在其to_mourn_inferior函数中通过unpush_target将自己出栈.

如果target支持partial和complete activation,就不应该在to_mourn_inferior中出栈(但是仍然需要在to_open中入栈),而应调用 target_mark_running或者target_mark_exited. 只要任何时候inferior fully active就应该调用target_mark_running(如在to_create_inferior和to_attach中), 而inferior inactive(to_mourn_inferior)时就掉用 target_mark_exited.在to_kill中,应该调用target_mourn_inferior确保target进入inactive状态.

Existing Targets

File Targets

executables 和 core files 都有target vectors.

Standard Protocol and Remote Stubs

GDB 文件remote.c 通过串口和target system通讯.GDB 提供一系列的stubs: stub集成到target系统中作为GDB的通讯断点.其名称一般是*-stub.c. 

  GDB user's manual 描述如何将stub嵌入target中:下面的例子讨论集成SPARC stub 到操作系统:

stub中的trap处理代码假设trap_low的接口如下:

1. %l1 和 %l2 分别包含pc 和 npc;

2. traps 以禁止;

3. 已经在正确的trap window(rfc).

满足以上条件时,可以从hw trap vector直接跳到stub.一般情况下stub使用的trap,操作系统是不使用的(一般是不可恢复性错误),也没有共享问题(有也可以支持). 最重要的trap之一ta 1(rfc):用于单步执行或者breakpoint.

Native Debugging

如下文件控制native support的配置:

`gdb/config/arch/xyz.mh'

makefile. 通过NATDEPFILES 列出所有体系相关的.o文件 . 通过NAT_FILE= nm-xyz.h指定native support的头文件.其他可选项:`NAT_CFLAGS', `NAT_ADD_FILES', `NAT_CLIBS', `NAT_CDEPS'参考`Makefile.in'.

`gdb/config/arch/nm-xyz.h'

`nm.h' 是到此文件的一个link(configure 自动创建). 包含C的宏定义,描述native system环境,如child process control和core file.

`gdb/xyz-nat.c'

native support的C代码.(有些系统都没有这个).

许多通用函数可以为很多系统共享,如果你的系统可以使用这些函数,使用NATDEPFILES定义所需要的.o即可.否则,你需要重新实现这些函数,并放到xyz-nat.c中.

`inftarg.c'

包含 target_ops vector ,支持通过ptrache和wait的Unix child processes.

`procfs.c'

target_ops vector 支持通过/proc控制Unix child processes.

`fork-child.c'

unix形式的fork和exec创建子进程.

`infptrace.c'

采用unix ptrace接口的被调试程序控制.

Native core file Support

`core-aout.c::fetch_core_registers()'

读取corefile的registers. 此函数调用register_addr(). GDB使用BFD读取core file, 事实上,所有machines都应该使用 core-aout.c, 仅提供fetch_core_registers 即可(xyz-nat.c,or REGISTER_U_ADDR in nm-xyz.h).

`core-aout.c::register_addr()'

如果nm-xyz.h file中定义了REGISTER_U_ADDR(addr, blockend, regno), 其功能应该如下:将addr设置成'user' struct中GDB寄存器regno的偏移. blockend 是'upgae'(u.u_ar0)内的偏移.如果定义了 REGISTER_U_ADDR , `core-aout.c' 会定义register_addr() 函数,并使用这个宏.如果没有定义此宏,但是又要使用fetch_core_registers(), 则必须自己定义 register_addr(), 并置于xyz-nat.c  如果自己定义了fetch_core_registers(), 就不用把register_addr()单独拿出来定义了.

当在一个新的OS上运行GDB native时,为了确保能够调试core file,就要写代码支持解析这个OS的core file,或者修改`bfd/trad-core.c'. 首先,定义一个寄存器结构来描述OS Core file中的寄存器(比如core file 的u-area),将定义core file header的头文件包含进来(u-area or a struct core). 然后,修改trad_unix_core_file_p 使用这些信息建立core file总的data segment, stack segment和其他段的信息(shared library ,control information), 以及"registers" segment(如果有两组寄存器,比如integer和float),即 "reg2" segment. 这些section 的信息使BFD知道如何从corefile读取这些信息. 然后,修改GDB:定义fetch_core_registers,使用通用的core-aout.c,或者自己的`xyz-nat.c'. GDB使用这个函数来将寄存器的值转移到GDB的"registers" array中.

   如果你的系统使用`/proc'控制进程,并使用ELF format的core files,则可以使用相同的函数来读取进程和core file的寄存器.

Native Conditionals

`nm-system.h'中可能存在的configure:

CHILD_PREPARE_TO_STORE

If the machine stores all registers at once in the child process, then define this to ensure that all values are correct. This usually entails a read from the child.

[Note that this is incorrectly defined in `xm-system.h' files currently.]

FETCH_INFERIOR_REGISTERS

如果自己定义fetch_inferior_registers 和 store_inferior_registers ,定义此宏.否则`infptrace.c'将编译进来.

int gdbarch_fp0_regnum (gdbarch)

返回首个浮点寄存器的编号.

int gdbarch_get_longjmp_target (gdbarch)

在ECstation和Iris中,这个是一个native-dependent的参数,大多数系统是target dependnet的.此函数确定longjmp将要跳转到的PC地址,(假定我们正好停在一个longjmp brakpoint中).(see GDBINT)

I386_USE_GENERIC_WATCHPOINTS

An x86-based machine can define this to use the generic x86 watchpoint support; see I386_USE_GENERIC_WATCHPOINTS.

ONE_PROCESS_WRITETEXT

定义此宏后,如果breakpoint插入失败,警告用户另一个进程可能使用同一个executable.

PROC_NAME_FMT

定义一个/proc 设备的name格式.定义在`nm.h',覆盖`procfs.c'中的默认定义.

SHELL_COMMAND_CONCAT

启动inferior时,将此字符串作为shell 命令的前缀.

SHELL_FILE

使用这个shell去启动inferior,默认是"/bin/sh".

SOLIB_ADD (filename, from_tty, targ, readsyms)

用于取filename中的symb到GDB的 symbol table.如果 readsyms是0,symbols将不读入,但是其他处理照常进行.

SOLIB_CREATE_INFERIOR_HOOK (rfc)

Define this to expand into any shared-library-relocation code that you want to be run just after the child process has been forked.

START_INFERIOR_TRAPS_EXPECTED

一般GDB启动一个inferior时, 又两次trap:因此shell运行时,一次program 自己运行时.如果实际上不是2次,定义此宏为正确的值.

CLEAR_SOLIB

See `objfiles.c'.

Support Libraries

BFD

identifying executable and core files

access to sections of files

BFD 解析文件头,确定像.text .data这些section 的名字,虚拟地址,大小等.

specialized core file support

BFD 提供函数用于确定core file的failing signal,以及此core file是否匹配一个特定的可执行文件.

locating the symbol information

opcodes

广泛用于binutils.提供反汇编支持.

readline

libiberty

此库提供一系列的函数用于补足,替换和扩展操作系统提供的功能.GDB使用很多这个库提供的特性:如 C++ demangler,IEEE floating format support, 输入选项解析`getopt', `obstack'扩展等.

obstacks in GDB

 obstack 是一个内存分配和释放机制.每个 obstack 是一个内存池,操作方式类似堆栈. 内存以LIFO方式在obstck上分配和释放.

主要用于object files管理. 每个obj文件有一个对应的obstack,许多对象在obstack上分配:....(略)

 在不同时间分配的对象,一次性的释放.

gnu-regex

Regex conditionals. 

Array Containers

see `vec.h' (略)

例子:

DEF_VEC_P(tree);   // non-managed tree vector.

struct my_struct {
  VEC(tree) *v;      // A (pointer to) a vector of tree pointers.
};

struct my_struct *s;

if (VEC_length(tree, s->v)) { we have some contents }
VEC_safe_push(tree, s->v, decl); // append some decl onto the end
for (ix = 0; VEC_iterate(tree, s->v, ix, elt); ix++)
  { do something with elt }

VEC_length  VEC_empty VEC_last  ....


Coding

Cleanups

Cleanups 是GDB使用的一个通用机制.cleanup 会在以后的合适时间执行:命令完成,error处理等.

使用方式:

struct cleanup *old_chain;

     Declare a variable which will hold a cleanup chain handle.

old_chain = make_cleanup (function, arg);

Make a cleanup which will cause function to be called with arg (a char *) later. The result, old_chain, is a handle that can later be passed to do_cleanups or discard_cleanups. Unless you are going to call do_cleanups or discard_cleanups, you can ignore the result from make_cleanup.

do_cleanups (old_chain);

Do all cleanups added to the chain since the corresponding make_cleanup call was made.

discard_cleanups (old_chain);

Same as do_cleanups except that it just removes the cleanups from the chain and does not call the specified functions.

Per-architecture module data

向gdbarch 添加module specific per-architecture data-pointers的机制:(see GDBINT)

Function: struct gdbarch_data *gdbarch_data_register_pre_init (gdbarch_data_pre_init_ftype *pre_init)

pre_init is used to, on-demand, allocate an initial value for a per-architecture data-pointer using the architecture's obstack (passed in as a parameter). Since pre_init can be called during architecture creation, it is not parameterized with the architecture. and must not call modules that use per-architecture data.

Function: struct gdbarch_data *gdbarch_data_register_post_init (gdbarch_data_post_init_ftype *post_init)

post_init is used to obtain an initial value for a per-architecture data-pointer after. Since post_init is always called after architecture creation, it both receives the fully initialized architecture and is free to call modules that use per-architecture data (care needs to be taken to ensure that those other modules do not try to call back to this module as that will create in cycles in the initialization call graph).

These functions return a struct gdbarch_data that is used to identify the per-architecture data-pointer added for that module.

The per-architecture data-pointer is accessed using the function:

Function: void *gdbarch_data (struct gdbarch *gdbarch, struct gdbarch_data *data_handle)

Given the architecture arch and module data handle data_handle (returned by gdbarch_data_register_pre_init or gdbarch_data_register_post_init), this function returns the current value of the per-architecture data-pointer. If the data pointer is NULL, it is first initialized by calling the corresponding pre_init or post_init method.

Wrapping Output Lines

输出换行的标准. 细节参考GDBINT.

GDB Coding Standards

ISO C

GDB assumes an ISO/IEC 9899:1990 (a.k.a. ISO C90) compliant compiler.

GDB does not assume an ISO C or POSIX compliant C library.

Memory Management

GDB does not use the functions malloc, realloc, calloc, free and asprintf.

GDB uses the functions xmalloc, xrealloc and xcalloc....

Compiler Warnings

Formatting

Comments

C Usage

Function Prototypes

Internal Error Recovery

File Names

Include Files

Clean Design and Portable Implementation

 GDB多年的经验告诉我们许多可移植性方面的问题:不能假定任何target的byte order(values, object files, and instructions).必须使用SWAP_TARGET_AND_HOST宏. 

所有操作target的函数必须使用 target_ops vector. 不能假定host和target是同一台机器(除了native),尤其不能假定target机器的头文件可以在host上使用.

在新的machine上编译GDB需要很多配置工作,需要修改许多头文件和配置脚本.假定新的host是xyz (e.g., `sun4'), 其配置名是 arch-xvend-xos (e.g., `sparc-sun-sunos4') :

• 在根目录,编辑 `config.sub' 添加 arch, xvend, and xos 到architecture列表.添加xyz作为arch-xvend-xos的别名.运行以下命令测试修改:
  ./config.sub xyz
  

  ./config.sub arch-xvend-xos
  (或许需要port BFD)
  

• 配置GDB:编辑`gdb/configure.host'时期认识你的系统,设置 gdb_host 为 xyz, 编辑`gdb/configure.tgt',设置gdb_target为合适值,如xyz.
• 最后,开始写对应的host-, native-,  target-dependent的 `.h' 和 `.c' 文件.......(不少啊).

Other

16. Versions and Branches

17. Start of New Year Procedure

18. Releasing GDB

19. Testsuite

20. Hints

A. GDB Currently available observers

B. GNU Free Documentation License