.\" Copyright (c) 1985 Regents of the University of California.
.\" All rights reserved. The Berkeley software License Agreement
.\" specifies the terms and conditions for redistribution.
.\" @(#)dbx.5.5 6.3 (Berkeley) %G%
dbx \- dbx symbol table information
The compiler symbol information generated
for \fIdbx\fP(1) uses the same structure
as described in \fIstab\fP(5), with additional type and scope information
appended to a symbol's name.
The assembler directive used to describe symbol information
has the following format:
\fBstabs\fR ``\fIstring\fP'',\fIkind\fP,0,\fIsize\fP,\fIvalue\fP
\fIString\fR contains the name, source language type, and scope of the symbol,
\fIkind\fR specifies the memory class
(e.g., external, static, parameter, local, register), and
\fIsize\fR specifies the byte size of the object, if relevant.
The third field (0 above) is unused.
For a global variable or a type, \fIvalue\fR is unused;
for a local variable or parameter, it is the offset from the frame pointer,
for a register variable, it is the associated register number.
The different kinds of stab entries are interpreted by dbx as follows:
The symbol is a global variable (e.g., .comm variable).
The variable's address can be found from the corresponding
ld(1) symbol entry, thus the value field for N_GSYM symbols is ignored.
For example, a global variable ``x'' will have both an N_GSYM entry and
an \fIld\fP(1) entry (e.g., N_BSS + N_EXT).
See \fIa.out\fP(5) for details about these other entries.
The symbol is a procedure or function.
The size field contains the line number of the entry point.
The value field contains the address of the entry point (in the text segment).
The symbol is a statically allocated variable for which
an initial value has been specified.
The value field contains the address of the variable (in the data segment).
The symbol is statically allocated, but not initialized.
The symbol is a register variable whose value is kept in the register
denoted by the value field.
The symbol is a parameter whose value is pushed on the stack before the call.
The value field contains the offset from the argument base pointer
(on the VAX, the ap register).
The symbol is a local variable whose value is stored in the most recently
defined procedure's stack frame.
The value is the (often negative) offset from the frame pointer
(on the VAX, the fp register).
The symbol defines separate compilation information
for pre-linking checking for Berkeley Pascal and DEC Modula-2 programs
For Pascal, the value field contains the line number
that the symbol is defined on.
The value field is not used for Modula-2.
Most of the source level information about a symbol
is stored in the string field of the stab entry.
Since strings are kept in a separate string table in the a.out file,
they can be arbitrarily long.
Thus there are no restrictions on the kind or length of information
in the string field, and it was not necessary
to modify the assembler or loader when extending or modifying
the format of this information.
Below is a grammar describing the syntax of the symbol string.
Except in the case of a constant whose value is a string,
there are no blanks in a symbol string.
NAME: [a-zA-Z_][a-zA-Z_0-9]*
REAL: [+-][0-9]*(\.[0-9][0-9]*|)([eE]([+-]|)[0-9][0-9]*|)
`X' ExportInfo -- export or import information (for N_MOD2 only)
`S' TypeId `,' NumElements `,' NumBits `,' BSTRING
TypeId -- local variable of type TypeId
`r' TypeId -- register variable of type TypeId
`S' TypeId -- module variable of type TypeId (static global in C)
`V' TypeId -- own variable of type TypeId (static local in C)
`G' TypeId -- global variable of type TypeId
Proc -- top level procedure
Proc `,' NAME `,' NAME -- local to first NAME,
-- second NAME is corresponding ld symbol
`Q' -- local procedure (static in C)
`I' -- internal procedure (different calling sequence)
`F' TypeId -- function returning type TypeId
`f' TypeId -- local function
`J' TypeId -- internal function
`p' TypeId -- value parameter of type TypeId
`v' TypeId -- reference parameter of type TypeId
`t' TypeId -- type name for type TypeId
`T' TypeId -- C structure tag name for struct TypeId
INTEGER -- Unique (per compilation) number of type
INTEGER `=' TypeDef -- Definition of type number
INTEGER `=' TypeAttrs TypeDef
-- Type attributes are extra information associated with a type,
-- such as alignment constraints or pointer checking semantics.
-- Dbx interprets some of these, but will ignore rather than complain
-- about any it does not recognize. Therefore this is a way to add
-- extra information for pre-linking checking.
TypeAttrList `,' TypeAttr
`a' INTEGER -- align boundary
`s' INTEGER -- size in bits
`p' INTEGER -- pointer class (e.g., checking)
BSTRING -- something else
`e' EnumList `;' -- enumeration
`*' TypeId -- pointer to TypeId
`S' TypeId -- set of TypeId
`d' TypeId -- file of TypeId
`i' NAME `:' NAME `;' -- imported type ModuleName:Name
`o' NAME `;' -- opaque type
`i' NAME `:' NAME `,' TypeId `;'
`r' TypeId `;' INTEGER `;' INTEGER
`a' TypeId `;' TypeId -- array [TypeId] of TypeId
`A' TypeId -- open array of TypeId
`D' INTEGER `,' TypeId -- N-dim. dynamic array
`E' INTEGER `,' TypeId -- N-dim. subarray
`f' TypeId `;' -- C function type
`f' TypeId `,' NumParams `;' TParamList `;'
`p' NumParams `;' TParamList `;'
`s' ByteSize FieldList `;' -- structure/record
`u' ByteSize FieldList `;' -- C union
NAME `:' TypeId `,' BitOffset `,' BitSize `;'
NAME `:' TypeId `,' PassBy `;'
`f' TypeId `,' NumParams `;' ParamList `;'
`p' NumParams `;' ParamList `;'
A `?' indicates that the symbol information is continued in
This directive can only occur where a `;' would otherwise
separate the fields of a record or constants in an enumeration.
It is useful when the number of elements in one of these lists
dbx(1), stab(5), a.out(5)