Hi,

I have made available a new version of the sym53c8xx (896) driver
yesterday that needs some testings I am not able to perform.

Version is 1.2 and the URLs are the following ones:

ftp://ftp.tux.org/pub/roudier/README
ftp://ftp.tux.org/pub/roudier/896/

(Old 810 rev < 16, 815 all rev and 825 rev < 16 are not supported)

The main changes of this version is an heavy rewrite of the interrupt
handling of some SCSI interrupts as Unexpected Disconnect, SCSI Parity
Error and Selection Timeout. These changes are intended to make the driver
more robust when some catastrophic condition happens on the SCSI BUS.
BTW, if your SCSI BUS is quite fine and your SCSI devices are not in the
process to die, the changes will not be triggered by your SCSI subsystem. 

The purpose is not to support not compliant device insertion/removal
neither to encourage any other kind of not compliant things people use to
do with their SCSI BUS(es). However, if it improves recovery in such a
situation, this will not be a bad thing after all. ;-) 

The purpose is:

1 - Not panic or hang the system when a hard disk of a software raid 1/4/5
    array decides to die, allowing user to later stop and fix the
    problem when such an operation will be possible.
2 - Support reasonnably removal/insertion with hardware that are claimed 
    to be compliant but that are not. After all people pay for such craps.
3 - Allows a faulty SCSI BUS to be still usable, allowing time to
    investigate the problem at the own risks of the user, obviously.
4 - Recover from bad luck. :-)

As we all know since years;), anything not SCSI compliant that happens on
a SCSI subsystem may break device's assomptions on the SCSI protocol and
lead to data loss or data corruption. But the only way to test the driver
changes is to play with such bad things. 

So, for once, I will be happy to receive reports about intentionnaly not
SCSI compliant BUS hacking and tampering that makes problems. :-))

Driver 1.2 changes are not really well tested yet but the source looks
quite fine to me. I use this driver version on my personal system. 

Thanks for testing those driver changes and sorry in advance for data
losses caused by the tests.

Regards,
   G�rard.

PS: Many thanks to Russel Johnson from Symbios who responded to my
    questions and gave me advices on the way to handle hot-plugging 
    error conditions with 53c8xx chips. 


-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]

Reply via email to