Date: Thu, 17 Sep 1998 22:04:56 +1000 (EST) From: Tony Maher <tonym@angis.usyd.edu.au> To: current@FreeBSD.ORG Subject: CAM this and that Message-ID: <199809171204.WAA26692@morgan.angis.su.OZ.AU>
next in thread | raw e-mail | index | archive | help
Hello,
installed CAM last nite (and apart from some operator stupidity concerning
creation of all appropriate devices) and it went well.
A few things:
1. Ignoring all warnings I decided to play with camcontrol.
Tape unit was initially off - switched on, ran camcontrol rescan -
magic... tape drive visible.
Switch tape drive off, ran camcontrol rescan.
Oops cant see tape drive - hmmm lets lock up - time to press the reset
switch (is that why they warn against using camcontrol for novice users
;-)
Ok normally have the tape unit off and only switch on if required.
Used to use scsi -p with `super scsi' device to see it and this worked
fine. Should I be able to do this with camcontrol?
2.
ahc0: <Adaptec 2940 SCSI adapter> rev 0x03 int a irq 11 on pci0.9.0
ahc0: aic7870 Single Channel A, SCSI Id=7, 16/255 SCBs
da0 at ahc0 bus 0 target 0 lun 0
da0: <SEAGATE ST51080N 0913> Fixed Direct Access SCSI2 device
da0: 10.0MB/s transfers (10.0MHz, offset 15), Tagged Queueing Enabled
da0: 1030MB (2109840 512 byte sectors: 255H 63S/T 131C)
da2 at ahc0 bus 0 target 2 lun 0
da2: <SEAGATE ST42100 7614> Fixed Direct Access SCSI2 device
da2: 10.0MB/s transfers (10.0MHz, offset 15), Tagged Queueing Enabled
da2: 1812MB (3711990 512 byte sectors: 255H 63S/T 231C)
da1 at ahc0 bus 0 target 1 lun 0
da1: <SEAGATE ST52160N 0285> Fixed Direct Access SCSI2 device
da1: 10.0MB/s transfers (10.0MHz, offset 15)
da1: 2069MB (4238282 512 byte sectors: 255H 63S/T 263C)
changing root device to da1s1a
So why doesn't my main (and newest disk) do Tagged Queueing?
Reading disk pamphlet and Seagate web site didn't turn up much apart from
using one of the utilities under Netware to send command to turn
taggged queueing on.
Sounds like a a job for camcontrol!
camcontrol modepage -m 0x0a -u 1 -P 3 -e
change
Queue Algorithm Modifier: 0 -> 1
DQue: 1 -> 0
to match the other two disks (and reading /usr/share/misc/scsi_modes
it seemed like the most likely candidates to change).
Unfortunately - this didn't work. Didn't cause any harm either.
Any pointers to documentation for reading?
(BTW man page has modepage -P option taking values 1-4, but actual
program wants 0-3 I think.)
Thats probably enough "testing" of (or should I say fooling around with)
camcontrol for now.
3. Pre CAM I had in kernel config file
options AHC_ALLOW_MEMIO # ok
#options AHC_TAGENABLE # seems ok with AHC_ALLOW_MEMIO,
# actually do get some aborts.
# Maybe old disk?
#options AHC_SCBPAGING_ENABLE # this *breaks* it bad!
I take it AHC_SCBPAGING_ENABLE is now the default (and no longer an
option) since I am seeing 16/255 SCBs rather than the plain 16.
This seems to work under CAM.
Excellent!
4. Soon after booting see
(da2:ahc0:0:2:0): tagged openings now 16
Some limit on this disk set to only 16?
5. This is an old disk and often get (esp. in first 20 mins of
booting) messages like:
(da2:ahc0:0:2:0): SCB 0x9 - timed out in dataout phase, SCSISIGI == 0x4
SEQADDR == 0x10e
SSTAT1 == 0x2
(da2:ahc0:0:2:0): BDR message in message buffer
(da2:ahc0:0:2:0): SCB 0x9 - timed out in dataout phase, SCSISIGI == 0x14
SEQADDR == 0x10e
SSTAT1 == 0x2
(da2:ahc0:0:2:0): no longer in timeout, status = 34b
ahc0: Issued Channel A Bus Reset. 17 SCBs aborted
But system recovers fine!
Currently running make -j4 and all seems ok (it used to break as others
had reported).
Many thanks to all involved.
tonym
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199809171204.WAA26692>