RA(4) RA(4)
NAME
ra - DEC MSCP disks (RA60, RA80, RA81)
DESCRIPTION
Ra* refer to disk drives conforming to DEC's Mass Storage
Control Protocol standard: drives such as the RA81 connected
via controllers such as the UDA50. Files with minor device
numbers 0 through 7 refer to various portions of drive 0,
minor devices 8 through 16 refer to drive 1, and so on up to
63. These files have a blocksize of 1K bytes; if 64 is
added to the minor device number, the resulting file has 4K
byte blocks, and if used as a filesystem, a slightly differ-
ent structure.
The range and size of the pseudo-drives for each drive are
as follows. Sizes are measured in hardware sectors (512
bytes).
disk start length
0 0 10240
1 10240 20480
2 30720 249848
3 280568 249848
4 530416 249848
5 780264 110808
6 30720 749544
7 0 arbitrarily large
Partition 7 covers the entire disk: 891072 sectors on an
RA81.
The ra files discussed above access the disk via the
system's normal buffering mechanism and may be read and
written without regard to physical disk records. There is
also a `raw' interface which provides for direct transmis-
sion between the disk and the user's read or write buffer.
A single read or write call results in exactly one I/O oper-
ation and therefore raw I/O is considerably more efficient
when many words are transmitted. The names of the raw RA
files begin with rra and end with a number which selects the
same disk as the corresponding ra file.
In raw I/O the buffer must begin on a word boundary, and
counts should be a multiple of 512 bytes (a disk block).
Likewise lseek(2) calls should specify a multiple of 512
bytes.
Several ioctl(2) calls apply to the raw devices. They are
probably useful only for bad block handling.
RA(4) RA(4)
UIOCHAR The third argument to ioctl is a pointer to a
struct ud_unit:
struct ud_unit {
daddr_t radsize;/* size of the disk */
daddr_t rctsize;/* size of the RCT, including pad */
long medium;
short tracksz;
short groupsz;
short cylsz;
char rbns;
char copies;
};
which the system fills with geometry parameters
for the drive.
UIORRCT The third argument points to a struct ud_rctbuf:
struct ud_rctbuf {
caddr_t buf;
int lbn;
};
Buf points to a 512-byte buffer, into which block
lbn of the replacement and caching table (RCT) is
read. As many copies of the RCT as necessary are
examined to find a readable copy of the block.
UIOWRCT The third argument points to a struct ud_rctbuf;
block lbn of the RCT is written in all copies.
UIOREPL The third argument points to a struct ud_repl :
struct ud_repl {
daddr_t replbn;/* good block */
daddr_t lbn; /* bad block */
short prim; /* nonzero if primary replacement */
};
A REPLACE command is sent to the controller,
requesting that attempts to access logical block
lbn henceforth be revectored to replacement block
replbn. Prim should be set nonzero if and only if
replbn is the primary replacement block for lbn.
FILES
/dev/ra*, /dev/rra*
SEE ALSO
rarepl(8)
MSCP Basic Disk Functions Manual
DEC Standard Disk Format Specification
BUGS
In raw I/O read and write(2) truncate file offsets to 512-
RA(4) RA(4)
byte block boundaries, and write scribbles on the tail of
incomplete blocks. Thus, in programs that are likely to
access raw devices, read, write and lseek(2) should always
deal in 512-byte multiples.
UIORRCT and UIOWRCT will misbehave if invoked on a partition
that doesn't start at the beginning of the disk. Partition
7 (the whole disk) is the best choice.
This driver should be split into a generic disk driver and a
port driver, which communicate by exchanging MSCP packets.