is - IPI channel driver for Sun IPI string controllers

---

CONFIG -- SUN-4, SPARCsystem 600MP SERIES

controller	isc0	at vme32d32 ? csr 0x01080000 priority 2 vector isintr 0x4c
channel		is0	at isc0 ipi_addr 0x00000	# channel 0 slave 0
controller	isc1	at vme32d32 ? csr 0x01080400 priority 2 vector isintr 0x4d
channel		is1	at isc1 ipi_addr 0x00100	# channel 0 slave 1
controller	isc2	at vme32d32 ? csr 0x01080800 priority 2 vector isintr 0x4e
channel		is2	at isc2 ipi_addr 0x00200	# channel 0 slave 2
controller	isc3	at vme32d32 ? csr 0x01080c00 priority 2 vector isintr 0x4f
channel		is3	at isc3 ipi_addr 0x00300	# channel 0 slave 3
controller	isc4	at vme32d32 ? csr 0x01081000 priority 2 vector isintr 0x50
channel		is4	at isc4 ipi_addr 0x00400	# channel 0 slave 4

DESCRIPTION

The first four controller and channel lines given in the synopsis section above correspond to the first, second, third, and fourth ISP-80 IPI disk controllers in a Sun system; the fifth controller and channel pair are currently available on SPARCsystem 600MP series machines only. These controllers are treated as integrated channel/disk-controller cards, in that there are two drivers that manage them. This driver, the is driver, presents a generic IPI-3 interface to the id driver, which handles the controller as if it were a channel-attached IPI-3 controller. See id.4s

The csr value gives the VME address of the controller. The ipi_addr field gives the 3-byte channel, slave, and facility address supported by the controller; the facility address portion is ignored and should be specified as zero.

SEE ALSO

DIAGNOSTICS

iscn: channel reset timed out

The probe routine, called during system initialization, determined that the controller reset, initiated by the boot, had not completed in the allotted time. The controller cannot be used.

iscn: bad IPI address addr

The specified address contained a slave address larger than 7.

iscn: interrupt vector not specified in config

The configuration did not specify an interrupt vector for the controller.

isn: refnum lookup on success failed. refnum refnum

A completion interrupt occurred for a command which was unknown to the controller. The interrupt is ignored.

isn: response too short. len l min m response r

The response packet presented in an interrupt was too short to contain a valid header. The response is printed, but otherwise the interrupt is ignored.

isn: response too long. max m len l truncating

The response packet is longer than the maximum, m, for this controller,

isn: is_reset_slave: resetting slave

The driver is resetting the controller under direction from the id disk driver. This occurs during recovery from timed out requests, also known as missing interrupts.

isn: is_reset_slave: slave reset not enabled

The id driver attempted to reset the controller to recovery from an error, but the driver has been set to disable such resets.

isn: ctlr reset timed out

The reset of the controller took longer than expected and has presumably failed.

iscn: response reg r not same as packet refnum ref

After reading a response packet, the response register was read and did not match the command reference number in the response packet. This could indicate a problem with the driver or the controller.

iscn: error status s response r

The controller presented error status code s and a response register value of r. These will be interpreted by messages that follow.

iscn: ctlr fault f - bus error on cmd refnum r

The controller reported a bus error while fetching the command.

iscn: ctlr fault f - timeout on cmd refnum r

The controller reported a VME timeout while fetching the command.

iscn: ctlr fault f - invalid command reg write on cmd refnum r

The controller indicates that the command register was written after the controller has panicked.

iscn: ctlr fault f - unknown fault code on cmd refnum r

The controller has presented an unknown fault code in its status.

Created by unroff & hp-tools. © somebody (See intro for details). All Rights Reserved. Last modified 11/5/97