Module `ida_segment`

Functions that deal with segments.

IDA requires that all program addresses belong to segments (each address must belong to exactly one segment). The situation when an address doesn't belong to any segment is allowed as a temporary situation only when the user changes program segmentation. Bytes outside a segment can't be converted to instructions, have names, comments, etc. Each segment has its start address, ending address and represents a contiguous range of addresses. There might be unused holes between segments.

Each segment has its unique segment selector. This selector is used to distinguish the segment from other segments. For 16-bit programs the selector is equal to the segment base paragraph. For 32-bit programs there is special array to translate the selectors to the segment base paragraphs. A selector is a 32/64 bit value.

The segment base paragraph determines the offsets in the segment. If the start address of the segment == (base << 4) then the first offset in the segment will be 0. The start address should be higher or equal to (base << 4). We will call the offsets in the segment 'virtual addresses'. So, the virtual address of the first byte of the segment is

(start address of segment - segment base linear address)

For IBM PC, the virtual address corresponds to the offset part of the address. For other processors (Z80, for example), virtual addresses correspond to Z80 addresses and linear addresses are used only internally. For MS Windows programs the segment base paragraph is 0 and therefore the segment virtual addresses are equal to linear addresses.

Global variables

var ADDSEG_FILLGAP

fill gap between new segment and previous one. i.e. if such a gap exists, and this gap is less than 64K, then fill the gap by extending the previous segment and adding .align directive to it. This way we avoid gaps between segments. too many gaps lead to a virtual array failure. it cannot hold more than ~1000 gaps.

var ADDSEG_IDBENC

'name' and 'sclass' are given in the IDB encoding; non-ASCII bytes will be decoded accordingly

var ADDSEG_NOAA

do not mark new segment for auto-analysis

var ADDSEG_NOSREG

set all default segment register values to BADSEL (undefine all default segment registers)

var ADDSEG_NOTRUNC

don't truncate the new segment at the beginning of the next segment if they overlap. destroy/truncate old segments instead.

var ADDSEG_OR_DIE

qexit() if can't add a segment

var ADDSEG_QUIET

silent mode, no "Adding segment..." in the messages window

var ADDSEG_SPARSE

use sparse storage method for the new ranges of the created segment. please note that the ranges that were already enabled before creating the segment will not change their storage type.

var CSS_BREAK

memory reading process stopped by user

var CSS_NODBG

debugger is not running

var CSS_NOMEM

not enough memory (might be because the segment is too big)

var CSS_NORANGE

could not find corresponding memory range

var CSS_OK

var MAX_GROUPS

max number of segment groups

var MAX_SEGM_TRANSLATIONS

max number of segment translations

var MOVE_SEGM_CHUNK

Too many chunks are defined, can't move.

var MOVE_SEGM_DEBUG

Debugger segments cannot be moved.

var MOVE_SEGM_IDP

IDP module forbids moving the segment.

var MOVE_SEGM_INVAL

Invalid argument (delta/target does not fit the address space)

var MOVE_SEGM_LOADER

The segment has been moved but the loader complained.

var MOVE_SEGM_MAPPING

Memory mapping ranges of addresses hinder segment movement.

var MOVE_SEGM_ODD

Cannot move segments by an odd number of bytes.

var MOVE_SEGM_OK

all ok

var MOVE_SEGM_ORPHAN

Orphan bytes hinder segment movement.

var MOVE_SEGM_PARAM

The specified segment does not exist.

var MOVE_SEGM_ROOM

Not enough free room at the target address.

var MOVE_SEGM_SOURCEFILES

Source files ranges of addresses hinder segment movement.

var MSF_FIXONCE

call loader only once with the special calling method. valid for rebase_program(). see loader_t::move_segm.

var MSF_LDKEEP

keep the loader in the memory (optimization)

var MSF_NETNODES

move netnodes instead of changing inf.netdelta (this is slower); valid for

rebase_program()

var MSF_NOFIX

don't call the loader to fix relocations

var MSF_PRIORITY

loader segments will overwrite any existing debugger segments when moved. valid for move_segm()

var MSF_SILENT

don't display a "please wait" box on the screen

var SEGMOD_KEEP

keep information (code & data, etc)

var SEGMOD_KEEP0

flag for internal use, don't set

var SEGMOD_KEEPSEL

do not try to delete unused selector

var SEGMOD_KILL

disable addresses if segment gets shrinked or deleted

var SEGMOD_NOMOVE

don't move info from the start of segment to the new start address (for set_segm_start())

var SEGMOD_SILENT

be silent

var SEGMOD_SPARSE

use sparse storage if extending the segment (for set_segm_start(), set_segm_end())

var SEGPERM_EXEC

Execute.

var SEGPERM_MAXVAL

SEGPERM_MAXVAL = 7

var SEGPERM_READ

Read.

var SEGPERM_WRITE

Write.

var SEG_ABSSYM

segment with definitions of absolute symbols

var SEG_BSS

uninitialized segment

var SEG_CODE

code segment

var SEG_COMM

segment with communal definitions

var SEG_DATA

data segment

var SEG_GRP

group of segments

var SEG_IMEM

internal processor memory & sfr (8051)

var SEG_IMP

SEG_IMP = 4

var SEG_MAX_BITNESS_CODE

SEG_MAX_BITNESS_CODE = 2

var SEG_MAX_SEGTYPE_CODE

SEG_MAX_SEGTYPE_CODE = 12

var SEG_NORM

unknown type, no assumptions

var SEG_NULL

zero-length segment

var SEG_UNDF

undefined segment type (not used)

var SEG_XTRN

segment with 'extern' definitions. no instructions are allowed

var SFL_COMORG

IDP dependent field (IBM PC: if set, ORG directive is not commented out)

var SFL_DEBUG

Is the segment created for the debugger?. Such segments are temporary and do not have permanent flags.

var SFL_HEADER

Header segment (do not create offsets to it in the disassembly)

var SFL_HIDDEN

Is the segment hidden?

var SFL_HIDETYPE

Hide segment type (do not print it in the listing)

var SFL_LOADER

Is the segment created by the loader?

var SFL_OBOK

Orgbase is present? (IDP dependent field)

var SNAP_ALL_SEG

Take a snapshot of all segments.

var SNAP_CUR_SEG

Take a snapshot of current segment.

var SNAP_LOAD_SEG

Take a snapshot of loader segments.

var SREG_NUM

Maximum number of segment registers is 16 (see segregs.hpp)

var saAbs

Absolute segment.

var saGroup

Segment group.

var saRel1024Bytes

1024 bytes

var saRel128Bytes

128 bytes

var saRel2048Bytes

2048 bytes

var saRel32Bytes

32 bytes

var saRel4K

This value is used by the PharLap OMF for page (4K) alignment. It is not supported by LINK.

var saRel512Bytes

512 bytes

var saRel64Bytes

64 bytes

var saRelByte

Relocatable, byte aligned.

var saRelDble

Relocatable, aligned on a double word (4-byte) boundary.

var saRelPage

Relocatable, aligned on 256-byte boundary.

var saRelPara

Relocatable, paragraph (16-byte) aligned.

var saRelQword

8 bytes

var saRelWord

Relocatable, word (2-byte) aligned.

var saRel_MAX_ALIGN_CODE

saRel_MAX_ALIGN_CODE = 14

var scCommon

Common. Combine by overlay using maximum size.

var scGroup

Segment group.

var scPriv

Private. Do not combine with any other program segment.

var scPub

Public. Combine by appending at an offset that meets the alignment requirement.

var scPub2

As defined by Microsoft, same as C=2 (public).

var scPub3

As defined by Microsoft, same as C=2 (public).

var scStack

Stack. Combine as for C=2. This combine type forces byte alignment.

var sc_MAX_COMB_CODE

sc_MAX_COMB_CODE = 7

Functions

def add_segm(*args) ‑> bool

add_segm(para, start, end, name, sclass, flags=0) -> bool

Add a new segment, second form. Segment alignment is set to saRelByte. Segment combination is "public" or "stack" (if segment class is "STACK"). Addressing mode of segment is taken as default (16bit or 32bit). Default segment registers are set to BADSEL. If a segment already exists at the specified range of addresses, this segment will be truncated. Instructions and data in the old segment will be deleted if the new segment has another addressing mode or another segment base address.

para: (C++: ea_t) segment base paragraph. if paragraph can't fit in 16bit, then a new

selector is allocated and mapped to the paragraph.

start: (C++: ea_t) start address of the segment. if start==BADADDR then start <-

to_ea(para,0).

end: (C++: ea_t) end address of the segment. end address should be higher than start

address. For emulate empty segments, use SEG_NULL segment type. If the end address is lower than start address, then fail. If end==BADADDR, then a segment up to the next segment will be created (if the next segment doesn't exist, then 1 byte segment will be created). If 'end' is too high and the new segment would overlap the next segment, 'end' is adjusted properly.

name: (C++: const char *) name of new segment. may be nullptr

sclass: (C++: const char *) class of the segment. may be nullptr. type of the new segment is

modified if class is one of predefined names:

"CODE" -> SEG_CODE
"DATA" -> SEG_DATA
"CONST" -> SEG_DATA
"STACK" -> SEG_BSS
"BSS" -> SEG_BSS
"XTRN" -> SEG_XTRN
"COMM" -> SEG_COMM
"ABS" -> SEG_ABSSYM

flags: (C++: int) Add segment flags

retval 1: ok

retval 0: failed, a warning message is displayed

def add_segm_ex(*args) ‑> bool

add_segm_ex(NONNULL_s, name, sclass, flags) -> bool

Add a new segment. If a segment already exists at the specified range of addresses, this segment will be truncated. Instructions and data in the old segment will be deleted if the new segment has another addressing mode or another segment base address.

NONNULL_s: (C++: segment_t *)

name: (C++: const char *) name of new segment. may be nullptr. if specified, the segment is

immediately renamed

sclass: (C++: const char *) class of the segment. may be nullptr. if specified, the segment

class is immediately changed

flags: (C++: int) Add segment flags

retval 1: ok

retval 0: failed, a warning message is displayed

def add_segment_translation(*args) ‑> bool

add_segment_translation(segstart, mappedseg) -> bool

Add segment translation.

segstart: (C++: ea_t) start address of the segment to add translation to

mappedseg: (C++: ea_t) start address of the overlayed segment

retval 1: ok

retval 0: too many translations or bad segstart

def allocate_selector(*args) ‑> sel_t

allocate_selector(segbase) -> sel_t

Allocate a selector for a segment unconditionally. You must call this function before calling add_segm_ex(). add_segm() calls this function itself, so you don't need to allocate a selector. This function will allocate a new free selector and setup its mapping using find_free_selector() and set_selector() functions.

segbase: (C++: ea_t) a new segment base paragraph

return: the allocated selector number

def change_segment_status(*args) ‑> int

change_segment_status(s, is_deb_segm) -> int

Convert a debugger segment to a regular segment and vice versa. When converting debug->regular, the memory contents will be copied to the database.

s: (C++: segment_t *) segment to modify

is_deb_segm: (C++: bool) new status of the segment

return: Change segment status result codes

def del_segm(*args) ‑> bool

del_segm(ea, flags) -> bool

Delete a segment.

ea: (C++: ea_t) any address belonging to the segment

flags: (C++: int) Segment modification flags

retval 1: ok

retval 0: failed, no segment at 'ea'.

def del_segment_translations(*args) ‑> void

del_segment_translations(segstart)

Delete the translation list

segstart: (C++: ea_t) start address of the segment to delete translation list

def del_selector(*args) ‑> void

del_selector(selector)

Delete mapping of a selector. Be wary of deleting selectors that are being used in the program, this can make a mess in the segments.

selector: (C++: sel_t) number of selector to remove from the translation table

def find_free_selector(*args) ‑> sel_t

find_free_selector() -> sel_t

Find first unused selector.

return: a number >= 1

def find_selector(*args) ‑> sel_t

find_selector(base) -> sel_t

Find a selector that has mapping to the specified paragraph.

base: (C++: ea_t) paragraph to search in the translation table

return: selector value or base

def get_defsr(*args) ‑> sel_t

get_defsr(s, reg) -> sel_t

Deprecated, use instead:

value = s.defsr[reg]

s: segment_t *

reg: int

def get_first_seg(*args) ‑> segment_t *

get_first_seg() -> segment_t

Get pointer to the first segment.

def get_group_selector(*args) ‑> sel_t

get_group_selector(grpsel) -> sel_t

Get common selector for a group of segments.

grpsel: (C++: sel_t) selector of group segment

return: common selector of the group or 'grpsel' if no such group is found

def get_last_seg(*args) ‑> segment_t *

get_last_seg() -> segment_t

Get pointer to the last segment.

def get_next_seg(*args) ‑> segment_t *

get_next_seg(ea) -> segment_t

Get pointer to the next segment.

ea: (C++: ea_t)

def get_prev_seg(*args) ‑> segment_t *

get_prev_seg(ea) -> segment_t

Get pointer to the previous segment.

ea: (C++: ea_t)

def get_segm_base(*args) ‑> ea_t

get_segm_base(s) -> ea_t

Get segment base linear address. Segment base linear address is used to calculate virtual addresses. The virtual address of the first byte of the segment will be (start address of segment - segment base linear address)

s: (C++: const segment_t *) pointer to segment

return: 0 if s == nullptr, otherwise segment base linear address

def get_segm_by_name(*args) ‑> segment_t *

get_segm_by_name(name) -> segment_t

Get pointer to segment by its name. If there are several segments with the same name, returns the first of them.

name: (C++: const char *) segment name. may be nullptr.

return: nullptr or pointer to segment structure

def get_segm_by_sel(*args) ‑> segment_t *

get_segm_by_sel(selector) -> segment_t

Get pointer to segment structure. This function finds a segment by its selector. If there are several segments with the same selectors, the last one will be returned.

selector: (C++: sel_t) a segment with the specified selector will be returned

return: pointer to segment or nullptr

def get_segm_class(*args) ‑> qstring *

get_segm_class(s) -> str

Get segment class. Segment class is arbitrary text (max 8 characters).

s: (C++: const segment_t *) pointer to segment

return: size of segment class (-1 if s==nullptr or bufsize<=0)

def get_segm_name(*args) ‑> qstring *

get_segm_name(s, flags=0) -> str

Get true segment name by pointer to segment.

s: (C++: const segment_t *) pointer to segment

flags: (C++: int) 0-return name as is; 1-substitute bad symbols with _ 1 corresponds

to GN_VISIBLE

return: size of segment name (-1 if s==nullptr)

def get_segm_num(*args) ‑> int

get_segm_num(ea) -> int

Get number of segment by address.

ea: (C++: ea_t) linear address belonging to the segment

return: -1 if no segment occupies the specified address. otherwise returns

number of the specified segment (0..get_segm_qty()-1)

def get_segm_para(*args) ‑> ea_t

get_segm_para(s) -> ea_t

Get segment base paragraph. Segment base paragraph may be converted to segment base linear address using to_ea() function. In fact, to_ea(get_segm_para(s), 0) == get_segm_base(s).

s: (C++: const segment_t *) pointer to segment

return: 0 if s == nullptr, the segment base paragraph

def get_segm_qty(*args) ‑> int

get_segm_qty() -> int

Get number of segments.

def get_segment_alignment(*args) ‑> char const *

get_segment_alignment(align) -> char const *

Get text representation of segment alignment code.

align: (C++: uchar)

return: text digestable by IBM PC assembler.

def get_segment_cmt(*args) ‑> qstring *

get_segment_cmt(s, repeatable) -> str

Get segment comment.

s: (C++: const segment_t *) pointer to segment structure

repeatable: (C++: bool) 0: get regular comment. 1: get repeatable comment.

return: size of comment or -1

def get_segment_combination(*args) ‑> char const *

get_segment_combination(comb) -> char const *

Get text representation of segment combination code.

comb: (C++: uchar)

return: text digestable by IBM PC assembler.

def get_segment_translations(*args) ‑> ssize_t

get_segment_translations(transmap, segstart) -> ssize_t

Get segment translation list.

transmap: (C++: eavec_t *) vector of segment start addresses for the translation list

segstart: (C++: ea_t) start address of the segment to get information about

return: -1 if no translation list or bad segstart. otherwise returns size of

translation list.

def get_selector_qty(*args) ‑> size_t

get_selector_qty() -> size_t

Get number of defined selectors.

def get_visible_segm_name(*args) ‑> qstring *

get_visible_segm_name(s) -> str

Get segment name by pointer to segment.

s: (C++: const segment_t *) pointer to segment

return: size of segment name (-1 if s==nullptr)

def getn_selector(*args) ‑> sel_t *, ea_t *

getn_selector(n) -> bool

Get description of selector (0..get_selector_qty()-1)

n: (C++: int)

def getnseg(*args) ‑> segment_t *

getnseg(n) -> segment_t

Get pointer to segment by its number.

warning: Obsoleted because it can slow down the debugger (it has to refresh the

whole memory segmentation to calculate the correct answer)

n: (C++: int) segment number in the range (0..get_segm_qty()-1)

return: nullptr or pointer to segment structure

def getseg(*args) ‑> segment_t *

getseg(ea) -> segment_t

Get pointer to segment by linear address.

ea: (C++: ea_t) linear address belonging to the segment

return: nullptr or pointer to segment structure

def is_finally_visible_segm(*args) ‑> bool

is_finally_visible_segm(s) -> bool

See SFL_HIDDEN, SCF_SHHID_SEGM.

s: (C++: segment_t *)

def is_miniidb(*args) ‑> bool

is_miniidb() -> bool

Is the database a miniidb created by the debugger?.

return: true if the database contains no segments or only debugger segments

def is_segm_locked(*args) ‑> bool

is_segm_locked(segm) -> bool

Is a segment pointer locked?

segm: (C++: const segment_t *) segment_t const *

def is_spec_ea(*args) ‑> bool

is_spec_ea(ea) -> bool

Does the address belong to a segment with a special type?. (SEG_XTRN, SEG_GRP, SEG_ABSSYM, SEG_COMM)

ea: (C++: ea_t) linear address

def is_spec_segm(*args) ‑> bool

is_spec_segm(seg_type) -> bool

Has segment a special type?. (SEG_XTRN, SEG_GRP, SEG_ABSSYM, SEG_COMM)

seg_type: (C++: uchar)

def is_visible_segm(*args) ‑> bool

is_visible_segm(s) -> bool

See SFL_HIDDEN.

s: (C++: segment_t *)

def lock_segm(*args) ‑> void

lock_segm(segm, lock)

Lock segment pointer Locked pointers are guaranteed to remain valid until they are unlocked. Ranges with locked pointers cannot be deleted or moved.

segm: (C++: const segment_t *) segment_t const *

lock: (C++: bool)

def move_segm(*args) ‑> move_segm_code_t

move_segm(s, to, flags=0) -> move_segm_code_t

This function moves all information to the new address. It fixes up address sensitive information in the kernel. The total effect is equal to reloading the segment to the target address. For the file format dependent address sensitive information, loader_t::move_segm is called. Also IDB notification event idb_event::segm_moved is called.

s: (C++: segment_t *) segment to move

to: (C++: ea_t) new segment start address

flags: (C++: int) Move segment flags

return: Move segment result codes

def move_segm_start(*args) ‑> bool

move_segm_start(ea, newstart, mode) -> bool

Move segment start. The main difference between this function and set_segm_start() is that this function may expand the previous segment while set_segm_start() never does it. So, this function allows to change bounds of two segments simultaneously. If the previous segment and the specified segment have the same addressing mode and segment base, then instructions and data are not destroyed - they simply move from one segment to another. Otherwise all instructions/data which migrate from one segment to another are destroyed.

note: this function never disables addresses.

ea: (C++: ea_t) any address belonging to the segment

newstart: (C++: ea_t) new start address of the segment note that segment start

address should be higher than segment base linear address.

mode: (C++: int) policy for destroying defined items

0: if it is necessary to destroy defined items, display a dialog box and ask confirmation
1: if it is necessary to destroy defined items, just destroy them without asking the user
-1: if it is necessary to destroy defined items, don't destroy them (i.e. function will fail)
-2: don't destroy defined items (function will succeed)

retval 1: ok

retval 0: failed, a warning message is displayed

def move_segm_strerror(*args) ‑> char const *

move_segm_strerror(code) -> char const *

Return string describing error MOVE_SEGM_... code.

code: (C++: move_segm_code_t) enum move_segm_code_t

def rebase_program(*args) ‑> int

rebase_program(delta, flags) -> int

Rebase the whole program by 'delta' bytes.

delta: (C++: adiff_t) number of bytes to move the program

flags: (C++: int) Move segment flags it is recommended to use MSF_FIXONCE so that

the loader takes care of global variables it stored in the database

return: Move segment result codes

def segm_adjust_diff(*args) ‑> adiff_t

segm_adjust_diff(s, delta) -> adiff_t

Truncate and sign extend a delta depending on the segment.

s: (C++: const segment_t *) segment_t const *

delta: (C++: adiff_t)

def segm_adjust_ea(*args) ‑> ea_t

segm_adjust_ea(s, ea) -> ea_t

Truncate an address depending on the segment.

s: (C++: const segment_t *) segment_t const *

ea: (C++: ea_t)

def segtype(*args) ‑> uchar

segtype(ea) -> uchar

Get segment type.

ea: (C++: ea_t) any linear address within the segment

return: Segment types, SEG_UNDF if no segment found at 'ea'

def sel2ea(*args) ‑> ea_t

sel2ea(selector) -> ea_t

Get mapping of a selector as a linear address.

selector: (C++: sel_t) number of selector to translate to linear address

return: linear address the specified selector is mapped to. if there is no

mapping, returns to_ea(selector,0);

def sel2para(*args) ‑> ea_t

sel2para(selector) -> ea_t

Get mapping of a selector.

selector: (C++: sel_t) number of selector to translate

return: paragraph the specified selector is mapped to. if there is no mapping,

returns 'selector'.

def set_defsr(*args) ‑> void

set_defsr(s, reg, value)

Deprecated, use instead:

s.defsr[reg] = value

s: segment_t *

reg: int

value: sel_t

def set_group_selector(*args) ‑> int

set_group_selector(grp, sel) -> int

Create a new group of segments (used OMF files).

grp: (C++: sel_t) selector of group segment (segment type is SEG_GRP) You should

create an 'empty' (1 byte) group segment It won't contain anything and will be used to redirect references to the group of segments to the common selector.

sel: (C++: sel_t) common selector of all segments belonging to the segment You should

create all segments within the group with the same selector value.

return: 1 ok 0 too many groups (see MAX_GROUPS)

def set_segm_addressing(*args) ‑> bool

set_segm_addressing(s, bitness) -> bool

Change segment addressing mode (16, 32, 64 bits). You must use this function to change segment addressing, never change the 'bitness' field directly. This function will delete all instructions, comments and names in the segment

s: (C++: segment_t *) pointer to segment

bitness: (C++: size_t) new addressing mode of segment

2: 64bit segment
1: 32bit segment
0: 16bit segment

return: success

def set_segm_base(*args) ‑> bool

set_segm_base(s, newbase) -> bool

Internal function.

s: (C++: segment_t *)

newbase: (C++: ea_t)

def set_segm_class(*args) ‑> int

set_segm_class(s, sclass, flags=0) -> int

Set segment class.

s: (C++: segment_t *) pointer to segment (may be nullptr)

sclass: (C++: const char *) segment class (may be nullptr). If segment type is SEG_NORM and

segment class is one of predefined names, then segment type is changed to:

"CODE" -> SEG_CODE
"DATA" -> SEG_DATA
"STACK" -> SEG_BSS
"BSS" -> SEG_BSS
if "UNK" then segment type is reset to SEG_NORM.

flags: (C++: int) Add segment flags

retval 1: ok, name is good and segment is renamed

retval 0: failure, name is nullptr or bad or segment is nullptr

def set_segm_end(*args) ‑> bool

set_segm_end(ea, newend, flags) -> bool

Set segment end address. The next segment is shrinked to allow expansion of the specified segment. The kernel might even delete the next segment if necessary. The kernel will ask the user for a permission to destroy instructions or data going out of segment scope if such instructions exist.

ea: (C++: ea_t) any address belonging to the segment

newend: (C++: ea_t) new end address of the segment

flags: (C++: int) Segment modification flags

retval 1: ok

retval 0: failed, a warning message is displayed

def set_segm_name(*args) ‑> int

set_segm_name(s, name, flags=0) -> int

Rename segment. The new name is validated (see validate_name). A segment always has a name. If you hadn't specified a name, the kernel will assign it "seg###" name where ### is segment number.

s: (C++: segment_t *) pointer to segment (may be nullptr)

name: (C++: const char *) new segment name

flags: (C++: int) ADDSEG_IDBENC or 0

retval 1: ok, name is good and segment is renamed

retval 0: failure, name is bad or segment is nullptr

def set_segm_start(*args) ‑> bool

set_segm_start(ea, newstart, flags) -> bool

Set segment start address. The previous segment is trimmed to allow expansion of the specified segment. The kernel might even delete the previous segment if necessary. The kernel will ask the user for a permission to destroy instructions or data going out of segment scope if such instructions exist.

ea: (C++: ea_t) any address belonging to the segment

newstart: (C++: ea_t) new start address of the segment note that segment start

address should be higher than segment base linear address.

flags: (C++: int) Segment modification flags

retval 1: ok

retval 0: failed, a warning message is displayed

def set_segment_cmt(*args) ‑> void

set_segment_cmt(s, cmt, repeatable)

Set segment comment.

s: (C++: const segment_t *) pointer to segment structure

cmt: (C++: const char *) comment string, may be multiline (with ' '). maximal size is 4096 bytes. Use empty str ("") to delete comment

repeatable: (C++: bool) 0: set regular comment. 1: set repeatable comment.

def set_segment_translations(*args) ‑> bool

set_segment_translations(segstart, transmap) -> bool

Set new translation list.

segstart: (C++: ea_t) start address of the segment to add translation to

transmap: (C++: const eavec_t &) vector of segment start addresses for the translation list. If

transmap is empty, the translation list is deleted.

retval 1: ok

retval 0: too many translations or bad segstart

def set_selector(*args) ‑> int

set_selector(selector, paragraph) -> int

Set mapping of selector to a paragraph. You should call this function _before_ creating a segment which uses the selector, otherwise the creation of the segment will fail.

selector: (C++: sel_t) number of selector to map

if selector == BADSEL, then return 0 (fail)
if the selector has had a mapping, old mapping is destroyed
if the selector number is equal to paragraph value, then the mapping is destroyed because we don't need to keep trivial mappings.

paragraph: (C++: ea_t) paragraph to map selector

retval 1: ok

retval 0: failure (bad selector or too many mappings)

def set_visible_segm(*args) ‑> void

set_visible_segm(s, visible)

See SFL_HIDDEN.

s: (C++: segment_t *)

visible: (C++: bool)

def setup_selector(*args) ‑> sel_t

setup_selector(segbase) -> sel_t

Allocate a selector for a segment if necessary. You must call this function before calling add_segm_ex(). add_segm() calls this function itself, so you don't need to allocate a selector. This function will allocate a selector if 'segbase' requires more than 16 bits and the current processor is IBM PC. Otherwise it will return the segbase value.

segbase: (C++: ea_t) a new segment base paragraph

return: the allocated selector number

def std_out_segm_footer(*args) ‑> void

std_out_segm_footer(ctx, seg)

Generate segment footer line as a comment line. This function may be used in IDP modules to generate segment footer if the target assembler doesn't have 'ends' directive.

ctx: (C++: struct outctx_t &) outctx_t &

seg: (C++: segment_t *)

def take_memory_snapshot(*args) ‑> bool

take_memory_snapshot(type) -> bool

Take a memory snapshot of the running process.

type: (C++: int) specifies which snapshot we want (see SNAP_ Snapshot types)

return: success

def update_segm(*args) ‑> bool

update_segm(s) -> bool

s: segment_t *

Classes

class lock_segment (*args)

Proxy of C++ lock_segment class.

__init__(self, _segm) -> lock_segment

_segm: segment_t const *

class segment_defsr_array (*args)

Proxy of C++ wrapped_array_t< sel_t,SREG_NUM > class.

__init__(self, data) -> segment_defsr_array

data: unsigned long long (&)[SREG_NUM]

Instance variables

var bytes : bytevec_t: _get_bytes(self) -> bytevec_t
var data: data

class segment_t (*args)

Proxy of C++ segment_t class.

__init__(self) -> segment_t

Ancestors

range_t

Instance variables

var align

Segment alignment codes

var bitness

Number of bits in the segment addressing

0: 16 bits
1: 32 bits
2: 64 bits

var color

the segment color

var comb

Segment combination codes

var defsr : wrapped_array_t< sel_t,SREG_NUM >

default segment register values. first element of this array keeps information about value of processor_t::reg_first_sreg

var end_ea

end_ea

var flags

Segment flags

var name

use get/set_segm_name() functions

var orgbase

this field is IDP dependent. you may keep your information about the segment here

var perm

Segment permissions (0 means no information)

var sclass

use get/set_segm_class() functions

var sel

segment selector - should be unique. You can't change this field after creating the segment. Exception: 16bit OMF files may have several segments with the same selector, but this is not good (no way to denote a segment exactly) so it should be fixed in the future.

var start_ea

start_ea

var type

segment type (see Segment types). The kernel treats different segment types differently. Segments marked with '*' contain no instructions or data and are not declared as 'segments' in the disassembly.

Methods

def abits(self, *args) ‑> int: abits(self) -> int
Get number of address bits.
def abytes(self, *args) ‑> int: abytes(self) -> int
Get number of address bytes.
def clr_comorg(self, *args) ‑> void: clr_comorg(self)
def clr_ob_ok(self, *args) ‑> void: clr_ob_ok(self)
def comorg(self, *args) ‑> bool: comorg(self) -> bool
def is_16bit(self, *args) ‑> bool: is_16bit(self) -> bool
Is a 16-bit segment?
def is_32bit(self, *args) ‑> bool: is_32bit(self) -> bool
Is a 32-bit segment?
def is_64bit(self, *args) ‑> bool: is_64bit(self) -> bool
Is a 64-bit segment?
def is_header_segm(self, *args) ‑> bool: is_header_segm(self) -> bool
def is_hidden_segtype(self, *args) ‑> bool: is_hidden_segtype(self) -> bool
def is_loader_segm(self, *args) ‑> bool: is_loader_segm(self) -> bool
def is_visible_segm(self, *args) ‑> bool: is_visible_segm(self) -> bool
def ob_ok(self, *args) ‑> bool: ob_ok(self) -> bool
def set_comorg(self, *args) ‑> void: set_comorg(self)
def set_debugger_segm(self, *args) ‑> void: set_debugger_segm(self, debseg)
debseg: bool
def set_header_segm(self, *args) ‑> void: set_header_segm(self, on)
on: bool
def set_hidden_segtype(self, *args) ‑> void: set_hidden_segtype(self, hide)
hide: bool
def set_loader_segm(self, *args) ‑> void: set_loader_segm(self, ldrseg)
ldrseg: bool
def set_ob_ok(self, *args) ‑> void: set_ob_ok(self)
def set_visible_segm(self, *args) ‑> void: set_visible_segm(self, visible)
visible: bool
def update(self, *args) ‑> bool: update(self) -> bool
Update segment information. You must call this function after modification of segment characteristics. Note that not all fields of segment structure may be modified directly, there are special functions to modify some fields.

return: success
def use64(self, *args) ‑> bool: is_64bit(self) -> bool
Is a 64-bit segment?

Inherited members

range_t:
- clear
- compare
- contains
- empty
- extend
- intersect
- overlaps
- size