I selected "no ALPHA test drivers", but still got
PCI bridge optimization (experimental)
Kernel/User network link driver
offered.
in the sources scsi/eata_{dma,pio}.c the date of the new versions is
older than the versions of 1.3.97. I guess the author forgot to
change the year to 96!
I bug reported for 1.3.85 is still there: When I insert a bad CD into
my Toshiba SCSI CD drive, I get a completely locked machine
(interrupts off) after the following messages:
# dd if=/dev/cdrom of=/dev/null count=1
scsi0 channel 0 : resetting for second half of retries
SCSI bus is being reset for host 0 channel 0
SCSI error: host 0 id2 lun 0 return code = 18000002
Sense class 7, sense error 0, extended sense 3
sr_photocd: ioctl error(TOSHIBA#1): 0x18000002
Comments: I have an Adaptec 2940 with one disk and one CD-ROM. The CD
inserted was _not_ a photo cd as the message might indicate. In ront
of the message with "Sense class" there is one graphical character (a
tab?). The inability of doing anything with keyboard and mouse
suggests that the interrupts were off when the machine was dead (Yes,
I've waited long enough (several minutes)).
I would expect that the driver tries resetting the bus or the device
several times, but not that the system dies. This is very bad. Maybe
it's a problem in the aic7xxx driver. Can anybody with a really
damaged CD reproduce this effect?
BTW: Someone forgot to change the driver version for aic7xxx in
1.3.98. Or is this an inofficial pre-patch?
Ulrich Windl