8< snip
>
> Single lun blacklisting
> A bug in the single lun blacklisting has been fixed.
>
8< snip
Has it?? Please tell my system its fixed!
I guess you don't have a CD changer to test ;) Well, it is better but ...
The only sure fix is to disable disconnects and remove the device from the
blacklist. That works 100% for concurrent access to two CD's but is likely
to time out for more, and its not recommended if the CD changer's on a bus
with other devices!
So far 2.0.35 and below don't work, 2.0.36-pre with the patch sometimes
works, and sometimes doesn't, depends on which order the luns are started
which also seems to vary depending on which driver is being used, and
2.1.1xx doesn't work - at least 2.1.115-120 don't. 2.1.1xx kernels behave
_exactly_ the same as 2.0.35.
With the patch I have found
dd if=/dev/scd4 of=/dev/null &
dd if=/dev/scd5 of=/dev/null
works - scd5 is a higher lun than scd4 - i.e. works if second command uses a
higher lun than the first.
and
dd if=/dev/scd4 of=/dev/null &
dd if=/dev/scd3 of=/dev/null
doesn't work - scd3 is a lower lun than scd4 - and generates a kernel panic
in allocate_device - i.e. fails if the second command is a lower lun than
the first. This is using the ncr53c8xx driver and NCR MBR-7 (Nakamichi)
Someone else has reported the opposite to this - second command with higher
lun fails with the same kernel panic, second command with lower lun is OK.
This is with the DC390 driver and same type of drive.
I'm looking at it but its a bit "poke and hope" as there's a lot I don't
understand WRT the interface between the drivers and upper layers of code,
much of the code and data structures in scsi.c and ... etc. :) I do
understand SCSI so all is not lost;)
Richard
-- Richard Waltham | Work: richard@digitalinterface.demon.co.uk At home in | Home: dormouse@farsrobt.demon.co.uk Southampton UK | 100421.1276@compuserve.com- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/faq.html