Module ida_offset

"Being an offset" is a characteristic of an operand. This means that operand or its part represent offset from some address in the program. This linear address is called "offset base". Some operands may have 2 offsets simultaneously. Generally, IDA doesn't handle this except for Motorola outer offsets. Thus there may be two offset values in an operand: simple offset and outer offset.

Outer offsets are handled by specifying special operand number: it should be ORed with OPND_OUTER value.

See bytes.hpp for further explanation of operand numbers.

Functions

add_refinfo_dref(insn: insn_t const &, _from: ea_t, ri: refinfo_t, opval: adiff_t, type: dref_t, opoff: int) ‑> ea_t

add_refinfo_dref(insn, _from, ri, opval, type, opoff) -> ea_t Add xrefs for a reference from the given instruction ( insn_t::ea). This function creates a cross references to the target and the base. insn_t::add_off_drefs() calls this function to create xrefs for 'offset' operand.

insn: (C++: const insn_t &) the referencing instruction
from: (C++: ea_t) the referencing instruction/data address
ri: (C++: const refinfo_t &) reference info block from the database
opval: (C++: adiff_t) operand value (usually op_t::value or op_t::addr)
type: (C++: dref_t) type of xref
opoff: (C++: int) offset of the operand from the start of instruction
return: the target address of the reference

calc_basevalue(target: ea_t, base: ea_t) ‑> ea_t

calc_basevalue(target, base) -> ea_t Calculate the value of the reference base.

target: (C++: ea_t)
base: (C++: ea_t)

calc_offset_base(ea: ea_t, n: int) ‑> ea_t

calc_offset_base(ea, n) -> ea_t Try to calculate the offset base This function takes into account the fixup information, current ds and cs values.

ea: (C++: ea_t) the referencing instruction/data address
n: (C++: int) operand number

• 0: first operand

• 1: second operand

• ...

• 7: eighth operand return: output base address or BADADDR

calc_probable_base_by_value(ea: ea_t, off: uval_t) ‑> ea_t

calc_probable_base_by_value(ea, off) -> ea_t Try to calculate the offset base. 2 bases are checked: current ds and cs. If fails, return BADADDR

ea: (C++: ea_t)
off: (C++: uval_t)

calc_reference_data(target: ea_t *, base: ea_t *, _from: ea_t, ri: refinfo_t, opval: adiff_t) ‑> bool

calc_reference_data(target, base, _from, ri, opval) -> bool Calculate the target and base addresses of an offset expression. The calculated target and base addresses are returned in the locations pointed by 'base' and 'target'. In case 'ri.base' is BADADDR, the function calculates the offset base address from the referencing instruction/data address. The target address is copied from ri.target. If ri.target is BADADDR then the target is calculated using the base address and 'opval'. This function also checks if 'opval' matches the full value of the reference and takes in account the memory-mapping.

target: (C++: ea_t *) output target address
base: (C++: ea_t *) output base address
from: (C++: ea_t) the referencing instruction/data address
ri: (C++: const refinfo_t &) reference info block from the database
opval: (C++: adiff_t) operand value (usually op_t::value or op_t::addr)
return: success

calc_target(*args) ‑> ea_t

calc_target(_from, opval, ri) -> ea_t Retrieve refinfo_t structure and calculate the target.

from: (C++: ea_t)
opval: (C++: adiff_t)
ri: refinfo_t const &

calc_target(_from, ea, n, opval) -> ea_t

from: ea_t
ea: ea_t
n: int
opval: adiff_t

can_be_off32(ea: ea_t) ‑> ea_t

can_be_off32(ea) -> ea_t Does the specified address contain a valid OFF32 value?. For symbols in special segments the displacement is not taken into account. If yes, then the target address of OFF32 will be returned. If not, then BADADDR is returned.

ea: (C++: ea_t)

get_default_reftype(ea: ea_t) ‑> reftype_t

get_default_reftype(ea) -> reftype_t Get default reference type depending on the segment.

ea: (C++: ea_t)
return: one of REF_OFF8, REF_OFF16, REF_OFF32, REF_OFF64

get_offbase(ea: ea_t, n: int) ‑> ea_t

get_offbase(ea, n) -> ea_t Get offset base value

ea: (C++: ea_t) linear address
n: (C++: int) 0..UA_MAXOP-1 operand number
return: offset base or BADADDR

get_offset_expr(ea: ea_t, n: int, ri: refinfo_t, _from: ea_t, offset: adiff_t, getn_flags: int = 0) ‑> qstring *

get_offset_expr(ea, n, ri, _from, offset, getn_flags=0) -> int See get_offset_expression()

ea: (C++: ea_t)
n: (C++: int)
ri: (C++: const refinfo_t &) refinfo_t const &
from: (C++: ea_t)
offset: (C++: adiff_t)
getn_flags: (C++: int)

get_offset_expression(ea: ea_t, n: int, _from: ea_t, offset: adiff_t, getn_flags: int = 0) ‑> qstring *

get_offset_expression(ea, n, _from, offset, getn_flags=0) -> int Get offset expression (in the form "offset name+displ"). This function uses offset translation function ( processor_t::translate) if your IDP module has such a function. Translation function is used to map linear addresses in the program (only for offsets).

Example: suppose we have instruction at linear address 0x00011000: mov ax, [bx+7422h] and at ds:7422h: array dw ... We want to represent the second operand with an offset expression, so then we call: get_offset_expresion(0x001100, 1, 0x001102, 0x7422, buf); | | | | | | | | | +output buffer | | | +value of offset expression | | +address offset value in the instruction | +the second operand +address of instruction and the function will return a colored string: offset array

ea: (C++: ea_t) start of instruction or data with the offset expression
n: (C++: int) operand number (may be ORed with OPND_OUTER)

• 0: first operand

• 1: second operand

• ...

• 7: eighth operand from: (C++: ea_t) linear address of instruction operand or data referring to the name. This address will be used to get fixup information, so it should point to exact position of operand in the instruction. offset: (C++: adiff_t) value of operand or its part. The function will return text representation of this value as offset expression. getn_flags: (C++: int) combination of:

• GETN_APPZERO: meaningful only if the name refers to a structure. appends the struct field name if the field offset is zero

• GETN_NODUMMY: do not generate dummy names for the expression but pretend they already exist (useful to verify that the offset expression can be represented) retval 0: can't convert to offset expression retval 1: ok, a simple offset expression retval 2: ok, a complex offset expression

op_offset(*args) ‑> bool

op_offset(ea, n, type_and_flags, target=BADADDR, base=0, tdelta=0) -> bool See op_offset_ex()

ea: (C++: ea_t)
n: (C++: int)
type_and_flags: (C++: uint32)
target: (C++: ea_t)
base: (C++: ea_t)
tdelta: (C++: adiff_t)

op_offset_ex(ea: ea_t, n: int, ri: refinfo_t) ‑> bool

op_offset_ex(ea, n, ri) -> bool Convert operand to a reference. To delete an offset, use clr_op_type() function.

ea: (C++: ea_t) linear address. if 'ea' has unexplored bytes, try to convert them to

• no segment: fail

• 16bit segment: to 16bit word data

• 32bit segment: to dword n: (C++: int) operand number (may be ORed with OPND_OUTER)

• 0: first

• 1: second

• ...

• 7: eighth operand

• OPND_MASK: all operands ri: (C++: const refinfo_t *) reference information return: success

op_plain_offset(ea: ea_t, n: int, base: ea_t) ‑> bool

op_plain_offset(ea, n, base) -> bool Convert operand to a reference with the default reference type.

ea: (C++: ea_t)
n: (C++: int)
base: (C++: ea_t)