?? cdbfileformat.txt
字號:
CDB File Format
Lenny Story
SDCC Development Team
Table of Contents
1 Overview
2 Usage
3 Conventions
4 Record Formats
4.1 Basic Record Format
4.2 Module Record
4.3 Symbol Records
4.4 Type Chain Record
4.5 Function Records
4.6 Type Records
4.7 Type Member
4.8 Link Address of Symbol
4.9 Linker Symbol End Address Record
4.10 Linker ASM Line Record
4.11 Linker C-Line Record
5 Source File Example
6 CDB File Example
1 Overview
The CDB File is used to record all of the information
that describes the variables, functions, lines, and
memory items. These records provide the critical
information that allows external utilities to properly
locate and interpret variables, functions, and types.
Development tools such as simulators debuggers and
profilers use these records to analyze and describe the
code in terms of the high level language in which it
was written. The CDB file is therefore the critical
link to bridging the compiled code image to the source
files from which it was built.
2 Usage
CDB files are created when the "--debug" option is
used. Each source file will have its own CDB file
associated with it. When the --debug flag is used
during the link process, the CDB file of the FIRST
source module will contain all the records from all the
source modules linked as well as the linker records.
Code and variables that are removed during the
optimization phase will not be present in the debug
file. Images created for debugging purposes should
always disable as much optimization as possible. The
following options are critical for the completeness of
the resulting debugging file.
--debug Mandatory to produce a debugging file.
--noinduction Induction processing occurs at the very
latest stages of compilation. This results in the
inability to report the location of the resulting
variables. Including this option disables the loop
induction optimization.
--nooverlay Overlay variables will not show up in the
debug file. Including this option will force variables
to the data segment.
3 Conventions
The record examples and grammar shown in this document
are displayed on multiple lines only for the purposes
of readability. The records contained within the CDB
files are always encoded on a single line.
Record grammar is indicated here using a custom format
of the following specifications:
* Record Elements are surrounded using the characters
`<' and '>'.
* Alternation is indicated using the `|' character.
* Non-mandatory items are surrounded using the
characters `{` and `}'.
4 Record Formats
4.1 Basic Record Format
<RecordType><:><RecordSpec>
+---------+------------------------------------+----------------------+
|
Type |
Description |
Form |
+---------+------------------------------------+----------------------+
+---------+------------------------------------+----------------------+
|
M |
Module Record |
Compiler |
+---------+------------------------------------+----------------------+
|
F |
Function Record |
Compiler |
+---------+------------------------------------+----------------------+
|
S |
Symbol Record |
Compiler |
+---------+------------------------------------+----------------------+
|
T |
Structure (Complex Type) Record |
Compiler |
+---------+------------------------------------+----------------------+
|
L |
Linker Record |
Assembler / Linker |
+---------+------------------------------------+----------------------+
All records are ASCII text, with one record per line.
The record type is the first character, followed by a
single colon ":". Sub types are often included as part
of the record specific format.
4.2 Module Record
<M><:><Filename>
+-------------+-------------------------------------------------------------+
|
M |
Module record type indicator |
+-------------+-------------------------------------------------------------+
|
Filename |
The filename of the module that this CDB file represents. |
+-------------+-------------------------------------------------------------+
Purpose
The module record is used to define a source module. It
is usually used at the beginning of each of the module
specific CDB files, and will occur several times in the
final CDB file produced by the linker to represent each
of the source modules. The location of this record
within the file does not necessarily indicate the scope
of the variables that follow. Those variables that have
module specific scope have an indication encoded within
their respective symbol entry.
Examples
M:Timer0
M:_bp
4.3 Symbol Records
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
S |
Symbol record type indicator |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
G |
Scope is global |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
F <Filename> |
Scope is file |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
L <Function> |
Scope is local |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
<Name> |
Symbol name |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
<Level> |
Scope level (see below) |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
<Block> |
Scope block (see below) |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
<TypeChain> |
Type chain record (see type record below) |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
<Address Space> |
Address space code (see table below) |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
<On Stack> |
Is this symbol on the stack? Indicates if the next
parameter is valid. |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
<Stack> |
The stack Offsetorelative to the "bp" variable. (The
libraries may have to be recompiled to include the -
-debug option for this variable to appear in the debug file). |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|
<Reg> |
If the address space is 'R', this field indicates the
register that the symbol is allocated. The register
name is in its native form: R1, AX, etc. |
+-------------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
<S><:>
{ G | F<Filename> | L { <function> | "--null--" }}
<$><Name>
<$><Level>
<$><Block>
<(><TypeRecord><)>
<,><AddressSpace>
<,><OnStack>
<,><Stack>
<,><[><Reg><,>{<Reg><,>}<]>
A symbol record is generated for each named symbol in
the source file; this includes local, global and
parameter symbols.
Blocks and Levels
The level & block are used to further scope local
variables since C allows unique definitions across
different scope blocks. When using the symbol records,
it is always important to include the level and block
information as part of the identification. It is
possible to have two symbols that share the same name,
but have different scope information.
The linker address records contain not only the name of
the symbol, but the Scope information as well, which is
instrumental in determining the correct instantiation
of the symbol.
Linker C line records also contain the Scope
information (see below).
The following code fragment illustrates a simple scope example:
foo()
{
int c; /* block #1 , level #1 */
{
int c; /* block #2, level #2 */
...
}
{
int c; /* block #3 , level #2 */
...
}
}
+------+----------------------------------------------------------+
|
A |
External stack |
+------+----------------------------------------------------------+
|
B |
Internal stack |
+------+----------------------------------------------------------+
|
C |
Code |
+------+----------------------------------------------------------+
|
D |
Code / static segment |
+------+----------------------------------------------------------+
|
E |
Internal ram (lower 128) bytes |
+------+----------------------------------------------------------+
|
F |
External ram |
+------+----------------------------------------------------------+
|
G |
Internal ram |
+------+----------------------------------------------------------+
|
H |
Bit addressable |
+------+----------------------------------------------------------+
|
I |
SFR space |
+------+----------------------------------------------------------+
|
J |
SBIT space |
+------+----------------------------------------------------------+
|
R |
Register space |
+------+----------------------------------------------------------+
|
Z |
Used for function records, or any undefined space code |
+------+----------------------------------------------------------+
4.4 Type Chain Record
<{><Size><}> <DCLType> <,> {<DCLType> <,>} <:> <Sign>
+--------------+------------------------------------------------+
|
<Size> |
The size of the item in decimal. |
+--------------+------------------------------------------------+
|
<DCLType> |
The type encoded using the table below. |
+--------------+------------------------------------------------+
|
<Sign> |
The sign of the item. Encoded as 'U' or 'S'. |
+--------------+------------------------------------------------+
Purpose
The C programming language allows arbitrarily complex
type constructions. Because of this, the CDB file type
designations are organized as a list of basic primitive types.
+--------------+----------------------------+
|
DA <n> |
Array of n elements |
+--------------+----------------------------+
|
DF |
Function |
+--------------+----------------------------+
|
DG |
Generic pointer |
+--------------+----------------------------+
|
DC |
Code pointer |
+--------------+----------------------------+
|
DX |
External ram pointer |
+--------------+----------------------------+
|
DD |
Internal ram pointer |
+--------------+----------------------------+
|
DP |
Paged pointer |
+--------------+----------------------------+
|
DI |
Upper 128 byte pointer |
+--------------+----------------------------+
?? 快捷鍵說明
復制代碼
Ctrl + C
搜索代碼
Ctrl + F
全屏模式
F11
切換主題
Ctrl + Shift + D
顯示快捷鍵
?
增大字號
Ctrl + =
減小字號
Ctrl + -