 |
» |
|
|
|
The
.SUBSPA directive
declares a new subspace or switches back to an old subspace. In
64-bit assembly programs subspaces map directly to the ELF concept
of sections, so the .SUBSPA directive
switches to or creates a new section. For more information, see
“Sections in 64-bit Mode”. Syntax | |
.SUBSPA name [,ACCESS=value][,ALIGN=value][,CODE_ONLY][,COMMON]
[,DUP_COMM][,FIRST][,FROZEN][,LOCK][,QUAD=value]
[,SORT=value][,TSPECIFIC][,UNLOADABLE][,ZERO]...
|
Parameters | |
- name
An identifier that names the current subspace. - ACCESS=value
Specifies the 7-bit value for the access rights
field in the PDIR
(Physical Page Directory for virtual address mapping). Must be 0X2C
for code, and 0X1F
for data subspaces. - ALIGN=value
Specifies a value (which must be a power of 2) on
which the Assembler should align the beginning of the subspace.
The default value is the largest alignment requested within that
subspace, or one if no alignment requests exist. - CODE_ONLY
Specifies that this subspace contains only code. - COMMON
Specifies that this subspace is a common block. - DUP_COMM
Specifies that the initialized data symbols within
this subspace can have duplicate names. When you include this parameter,
multiple occurrences of a universal data symbol can exist and the
linker does not report a "Duplicate Definition"
error. - FIRST
Specifies that the subspace must be allocated exactly
at the beginning of the specified space. - FROZEN
If set, the subspace is locked into physical memory
when the subspace goes into execution. - LOCK
If set, the subspace is locked into physical memory
when the operating system is booted. - QUAD=value
Specifies the quadrant (0 through 3) in which the
Assembler should place this subspace. The default value is zero. - SORT=value
Provides an integer value for the primary sort key.
The linker orders the subspaces in the output object module according
to this primary key. If several subspaces share the same primary
key value, the linker lists these subspaces in the order in which
it processes them. It is suggested that 24 be used for both code
and data subspaces. - TSPECIFIC
Indicates that this space contains thread local
storage data. - UNLOADABLE
Specifies that this subspace is not loadable into
memory. Loadable subspaces must reside in loadable spaces, and unloadable
subspaces must reside in unloadable spaces. - ZERO
Specifies that this subspace contains all zeros
and no data appears in the output file.
Discussion | |
The first time the Assembler encounters a .SUBSPA
directive with a new name, it uses that
name to declare a new subspace. As this is the defining occurrence
of that subspace, optional keywords describe attributes of that
subspace. When the Assembler encounters additional .SUBSPA
directives with that name, it continues
that subspace. In this case, the .SUBSPA
directive can only contain the subspace name; other keywords to
describe the subspace are illegal. Example | |
This example shows some of the standard "subspace"
definitions in a typical assembly language program. .SUBSPA $CODE$, QUAD=0,ALIGN=8,ACCESS=0x2c,SORT=24,CODE_ONLY
.SUBSPA $DATA$, QUAD=1,ALIGN=8,ACCESS=0x1f,SORT=16 .SUBSPA $TBSS$, QUAD=1,ALIGN=8,ACCESS=0x1f,ZERO, TSPECIFIC, SORT=40 |
|
