adaptec 2944 problems on -current alpha

[Bernd Walter]

I have problems to communcate with devices connected to an adaptec 2944
controller with a -current alpha system.
The kernel is compiled with options AHC_ALLOW_MEMIO, because I also
have a 3985 in that machine and SRM did not set IO space on the bridge.
Devices conntect to that controllers A channel work.

ahc1: <Adaptec 2944 SCSI adapter> port 0x1fffd00-0x1fffdff mem 0x7f9c000-0x7f9cfff irq 8 at device 2.0 on pci1
ahc1: interrupting at IRQ 0x0 intA (vec 0x980)
aic7870: Wide Channel A, SCSI Id=15, 16/253 SCBs

(ahc1:A:0:0): refuses synchronous negotiation. Using asynchronous transfers
(probe29:ahc1:0:0:0): Unexpected busfree in Message-in phase
SEQADDR == 0x174
(ahc1:A:2:0): refuses synchronous negotiation. Using asynchronous transfers
(probe31:ahc1:0:2:0): Unexpected busfree in Message-in phase
SEQADDR == 0x174
(ahc1:A:1:0): refuses synchronous negotiation. Using asynchronous transfers
(probe30:ahc1:0:1:0): Unexpected busfree in Message-in phase
SEQADDR == 0x174
(ahc1:A:3:0): refuses synchronous negotiation. Using asynchronous transfers
(probe32:ahc1:0:3:0): Unexpected busfree in Message-in phase
SEQADDR == 0x174
da5: <SEAGATE ST42400N SUN2.1G 0417> Fixed Direct Access SCSI-2 device 
da5: 3.300MB/s transfers, Tagged Queueing Enabled
da5: 2030MB (4159462 512 byte sectors: 255H 63S/T 258C)
da6 at ahc1 bus 0 target 1 lun 0
da6: <SEAGATE ST42400N SUN2.1G 0417> Fixed Direct Access SCSI-2 device 
da6: 3.300MB/s transfers, Tagged Queueing Enabled
da6: 2030MB (4159462 512 byte sectors: 255H 63S/T 258C)
da7 at ahc1 bus 0 target 2 lun 0
da7: <SEAGATE ST42400N SUN2.1G 0417> Fixed Direct Access SCSI-2 device 
da7: 3.300MB/s transfers, Tagged Queueing Enabled
da7: 2030MB (4159462 512 byte sectors: 255H 63S/T 258C)
da8 at ahc1 bus 0 target 3 lun 0
da8: <SEAGATE ST42400N SUN2.1G 0417> Fixed Direct Access SCSI-2 device 
da8: 3.300MB/s transfers, Tagged Queueing Enabled
da8: 2030MB (4159462 512 byte sectors: 255H 63S/T 258C)

I'm not absolutely shure, but I doubt that these drives are not capable
of doing sync transfers.

[70]cicely12# camcontrol inquiry -n da -u 5
Exit 1

[71]cicely12# disklabel da5
disklabel: /dev/da5: Invalid argument
Exit 1
During disklabel I get:
(da5:ahc1:0:0:0): READ CAPACITY. CDB: 25 0 0 0 0 0 0 0 0 0 
(da5:ahc1:0:0:0): CAM Status: SCSI Status Error
(da5:ahc1:0:0:0): SCSI Status: Check Condition
(da5:ahc1:0:0:0): ILLEGAL REQUEST info?:1bc5 asc:20,0
(da5:ahc1:0:0:0): Invalid command operation code: Command byte 0 is invalid
(da5:ahc1:0:0:0): Unretryable error
Opened disk da5 -> 22

[72]cicely12# dd if=/dev/da5 bs=512 count=1 of=/dev/zero
dd: /dev/da5: Invalid argument
Exit 1
And once again I get:
(da5:ahc1:0:0:0): READ CAPACITY. CDB: 25 0 0 0 0 0 0 0 0 0 
(da5:ahc1:0:0:0): CAM Status: SCSI Status Error
(da5:ahc1:0:0:0): SCSI Status: Check Condition
(da5:ahc1:0:0:0): ILLEGAL REQUEST info?:22a6 asc:20,0
(da5:ahc1:0:0:0): Invalid command operation code: Command byte 0 is invalid
(da5:ahc1:0:0:0): Unretryable error
Opened disk da5 -> 22

Interestingly the initial probing got the right capacity somehow.


The terminators on the controller are installed and I don't get bus
error messages.
The drive chain is connected externaly and worked on a differential
controller in a sparc station under NetBSD-1.6.1 before without
problems.

-- 
B.Walter                   BWCT                http://www.bwct.de
ticso at bwct.de                                  info at bwct.de