Detailed Description

Functions that deal with fixup information.

A loader should setup fixup information using set_fixup().

Classes
struct	fixup_data_t

struct	fixup_handler_t
	Implements the core behavior of a custom fixup. More...

struct	fixup_info_t
	Collect fixup records for the specified range. More...

Functions
THREAD_SAFE bool	is_fixup_custom (fixup_type_t type)
	Is fixup processed by processor module?

idaman bool ida_export	get_fixup (fixup_data_t *fd, ea_t source)
	Get fixup information.

bool	exists_fixup (ea_t source)
	Check that a fixup exists at the given address.

idaman void ida_export	set_fixup (ea_t source, const fixup_data_t &fd)
	Set fixup information. More...

idaman void ida_export	del_fixup (ea_t source)
	Delete fixup information.

idaman const fixup_handler_t *ida_export	get_fixup_handler (fixup_type_t type)
	Get handler of standard or custom fixup.

idaman bool ida_export	apply_fixup (ea_t item_ea, ea_t fixup_ea, int n, bool is_macro)
	Use fixup information for an address. More...

idaman uval_t ida_export	get_fixup_value (ea_t ea, fixup_type_t type)
	Get the operand value. More...

idaman bool ida_export	patch_fixup_value (ea_t ea, const fixup_data_t &fd)
	Patch the fixup bytes. More...

idaman const char *ida_export	get_fixup_desc (qstring *buf, ea_t source, const fixup_data_t &fd)
	Get FIXUP description comment.

idaman int ida_export	calc_fixup_size (fixup_type_t type)
	Calculate size of fixup in bytes (the number of bytes the fixup patches) More...

idaman bool ida_export	get_fixups (fixups_t *out, ea_t ea, asize_t size)

bool	contains_fixups (ea_t ea, asize_t size)
	Does the specified address range contain any fixup information?

idaman void ida_export	gen_fix_fixups (ea_t from, ea_t to, asize_t size)
	Relocate the bytes with fixup information once more (generic function). More...

idaman bool ida_export	handle_fixups_in_macro (refinfo_t *ri, ea_t ea, fixup_type_t other, uint32 macro_reft_and_flags)
	Handle two fixups in a macro. More...

Enumerate addresses with fixup information:
Get the first address with fixup information Returns the first address with fixup information, or BADADDR
idaman ea_t ida_export	get_first_fixup_ea (void)

idaman ea_t ida_export	get_next_fixup_ea (ea_t ea)
	Find next address with fixup information. More...

idaman ea_t ida_export	get_prev_fixup_ea (ea_t ea)
	Find previous address with fixup information. More...

std_get_value()
This internal function takes Tuning options to determine how to get the operand value. It is opposite in meaning to the std_patch_value(). 1) it gets the fixup_handler_t::size bytes at the given address, 2) it shifts the result left by fixup_handler_t::shift bits, 3) it returns the rightmost fixup_handler_t::width bits as a signed value. In details: b) size = 4, width = 28, shift = 2 it gets 4 bytes from the fixup address (the branch insn) then it shifts this dword left by 2 bits (shift=2) then the result is masked with 0xFFFFFFF (width=28) e.g. 0x10000017 the insn 0x4000005C the unshifted value 0x0000005C the masked result
idaman fixup_type_t ida_export	register_custom_fixup (const fixup_handler_t *cfh)
	Register a new custom fixup. More...

idaman bool ida_export	unregister_custom_fixup (fixup_type_t type)
	Unregister a new custom fixup format. More...

idaman fixup_type_t ida_export	find_custom_fixup (const char *name)
	Get id of a custom fixup handler. More...

Macros
#define	FIXUPF_REL 0x0001
	fixup is relative to the linear address `base`. More...

#define	FIXUPF_EXTDEF 0x0002
	target is a location (otherwise - segment). More...

#define	FIXUPF_UNUSED 0x0004
	fixup is ignored by IDA More...

#define	FIXUPF_CREATED 0x0008
	fixup was not present in the input file

#define	FIXUPF_LOADER_MASK 0xF0000000
	additional flags. More...

#define	FHF_VERIFY 0x0001
	verify that the value fits into WIDTH bits. More...

#define	FHF_CODE 0x0002
	verify that ITEM_EA in std_apply() points to an instruction. More...

#define	FHF_FORCE_CODE 0x0004
	if ITEM_EA in std_apply() points to an unknown item, then convert it to code. More...

#define	FHF_ABS_OPVAL 0x0008
	create absolute refinfo in std_apply() because the operand also has the absolute value (usually for o_near operands)

#define	FHF_SIGNED 0x0010
	the operand value is signed. More...

Fixup standard types
#define	FIXUP_OFF8 13
	8-bit offset

#define	FIXUP_OFF16 1
	16-bit offset

#define	FIXUP_SEG16 2
	16-bit base–logical segment base (selector)

#define	FIXUP_PTR16 3
	32-bit long pointer (16-bit base:16-bit offset)

#define	FIXUP_OFF32 4
	32-bit offset

#define	FIXUP_PTR32 5
	48-bit pointer (16-bit base:32-bit offset)

#define	FIXUP_HI8 6
	high 8 bits of 16bit offset

#define	FIXUP_HI16 7
	high 16 bits of 32bit offset

#define	FIXUP_LOW8 8
	low 8 bits of 16bit offset

#define	FIXUP_LOW16 9
	low 16 bits of 32bit offset

#define	V695_FIXUP_VHIGH 10
	obsolete

#define	V695_FIXUP_VLOW 11
	obsolete

#define	FIXUP_OFF64 12
	64-bit offset

#define	FIXUP_OFF8S 14
	8-bit signed offset

#define	FIXUP_OFF16S 15
	16-bit signed offset

#define	FIXUP_OFF32S 16
	32-bit signed offset

#define	FIXUP_CUSTOM 0x8000
	start of the custom types range

Typedefs
typedef uint16	fixup_type_t
	Fixup information structure. More...

typedef qvector< fixup_info_t >	fixups_t

Typedef Documentation

◆ fixup_type_t

typedef uint16 fixup_type_t

Fixup information structure.

see Types of fixups

Function Documentation

◆ set_fixup()

idaman void ida_export set_fixup	(	ea_t	source,
		const fixup_data_t &	fd
	)

Set fixup information.

You should fill fixup_data_t and call this function and the kernel will remember information in the database.

Parameters

source	the fixup source address, i.e. the address modified by the fixup
fd	fixup data

◆ get_next_fixup_ea()

idaman ea_t ida_export get_next_fixup_ea ( ea_t ea )

Find next address with fixup information.

Parameters

ea	current address

Returns: the next address with fixup information, or BADADDR

◆ get_prev_fixup_ea()

idaman ea_t ida_export get_prev_fixup_ea ( ea_t ea )

Find previous address with fixup information.

Parameters

ea	current address

Returns: the previous address with fixup information, or BADADDR

◆ apply_fixup()

idaman bool ida_export apply_fixup	(	ea_t	item_ea,
		ea_t	fixup_ea,
		int	n,
		bool	is_macro
	)

Use fixup information for an address.

This function converts item_ea flags to offsets/segments. For undefined bytes, you may set item_ea == fixup_ea. In this case this function will create an item (byte, word, dword) there.

Parameters

item_ea	start address of item to modify
fixup_ea	address of fixup record
n	0..UA_MAXOP-1 operand number, OPND_ALL one of the operands
is_macro	is the instruction at 'item_ea' a macro? if yes, then partial fixups (HIGH, LOW) won't be applied

Return values

false	no fixup at fixup_ea or it has FIXUPF_UNUSED flag
true	ok, the fixup information was applied

◆ get_fixup_value()

idaman uval_t ida_export get_fixup_value	(	ea_t	ea,
		fixup_type_t	type
	)

Get the operand value.

This function get fixup bytes from data or an instruction at `ea` and convert them to the operand value (maybe partially). It is opposite in meaning to the `patch_fixup_value()`. For example, FIXUP_HI8 read a byte at `ea` and shifts it left by 8 bits, or AArch64's custom fixup BRANCH26 get low 26 bits of the insn at `ea` and shifts it left by 2 bits. This function is mainly used to get a relocation addend.

Parameters

ea	address to get fixup bytes from, the size of the fixup bytes depends on the fixup type.

See also: fixup_handler_t::size

Parameters

type	fixup type

Return values

operand value

◆ patch_fixup_value()

idaman bool ida_export patch_fixup_value	(	ea_t	ea,
		const fixup_data_t &	fd
	)

Patch the fixup bytes.

This function updates data or an instruction at `ea` to the fixup bytes. For example, FIXUP_HI8 updates a byte at `ea` to the high byte of `fd->off`, or AArch64's custom fixup BRANCH26 updates low 26 bits of the insn at `ea` to the value of `fd->off` shifted right by 2.

Parameters

ea	address where data are changed, the size of the changed data depends on the fixup type.

See also: fixup_handler_t::size

Parameters

fd	fixup data

Return values

false the fixup bytes do not fit (e.g. `fd->off` is greater than 0xFFFFFFC for BRANCH26). The database is changed even in this case.

◆ calc_fixup_size()

idaman int ida_export calc_fixup_size ( fixup_type_t type )

Calculate size of fixup in bytes (the number of bytes the fixup patches)

Return values

-1	means error

◆ register_custom_fixup()

idaman fixup_type_t ida_export register_custom_fixup ( const fixup_handler_t * cfh )

Register a new custom fixup.

This function must be called by a processor module or plugin, but not by a file loader. File loaders should use find_custom_fixup() function to find the handler created by the processor module.

Returns: id of the new custom fixup handler with FIXUP_CUSTOM bit set or 0 (e.g. when the custom fixup handler with the same name was already registered).

◆ unregister_custom_fixup()

idaman bool ida_export unregister_custom_fixup ( fixup_type_t type )

Unregister a new custom fixup format.

Should be called by the processor module before the database gets closed.

◆ find_custom_fixup()

idaman fixup_type_t ida_export find_custom_fixup ( const char * name )

Get id of a custom fixup handler.

Parameters

name	name of the custom fixup handler

Returns: id with FIXUP_CUSTOM bit set or 0

◆ gen_fix_fixups()

idaman void ida_export gen_fix_fixups	(	ea_t	from,
		ea_t	to,
		asize_t	size
	)

Relocate the bytes with fixup information once more (generic function).

This function may be called from loader_t::move_segm() if it suits the goal. If loader_t::move_segm is not defined then this function will be called automatically when moving segments or rebasing the entire program. Special parameter values (from = BADADDR, size = 0, to = delta) are used when the function is called from rebase_program(delta).

◆ handle_fixups_in_macro()

idaman bool ida_export handle_fixups_in_macro	(	refinfo_t *	ri,
		ea_t	ea,
		fixup_type_t	other,
		uint32	macro_reft_and_flags
	)

Handle two fixups in a macro.

We often combine two instruction that load parts of a value into one macro instruction. For example:

ARM:   ADRP  X0, #var@PAGE
       ADD   X0, X0, #var@PAGEOFF  --> ADRL X0, var
MIPS:  lui   $v0, %hi(var)
       addiu $v0, $v0, %lo(var)    --> la   $v0, var

When applying the fixups that fall inside such a macro, we should convert them to one refinfo. This function does exactly that. It should be called from the apply() callback of a custom fixup.

Returns: success ('false' means that RI was not changed)

Detailed Description

Classes

Functions

Macros

Typedefs

Typedef Documentation

◆ fixup_type_t

Function Documentation

◆ set_fixup()

◆ get_next_fixup_ea()

◆ get_prev_fixup_ea()

◆ apply_fixup()

◆ get_fixup_value()

◆ patch_fixup_value()

◆ calc_fixup_size()

◆ register_custom_fixup()

◆ unregister_custom_fixup()

◆ find_custom_fixup()

◆ gen_fix_fixups()

◆ handle_fixups_in_macro()