On Wednesday 05 December 2001 01:35 am, you wrote:
> > I realize the right (tm) way to fix this is to fix the drivers, but
> > how hard
> > would it be to reload the ide-scsi driver as part of the card recognition
>> ?
>
> You need some configuration framework for this. Not every PCMCIA IDE
> drive is used as ide-scsi; and IDE-pure case obviously works (that is
> why I said that I suspect IDE part is already handled correctly).
True, but the kernel is passed which drives to make ide-scsi, ie
hde=ide-scsi. These parameters seem to be used when when
the module is reloaded manually.
>
> > It's extremely unlikely that a laptop is going to be using a scsi
> > controller,
>
> SCSI controller has nothing to do with ide-scsi.
>
> > or that a desktop machine or server is going to be using a pcmcia
> > controller
> > to run ide-scsi devices.
>
> True.
>
>> Doesn't the ide-scsi driver look for what slots
> > are
> > open when it loads?
>
> Yep. That is why reloading it helps.
>
> > Can the loading / unloading be put in the pcmcia ide startup script?
>
> If you could suggest _clean_ framework to set up when ide-scsi has to be
> loaded ... of course it could be done every time - if it fails it fails,
> if not it does not harm.
>
> > I
> > think I tried to do something along those lines once, but didn't get
> > very far
> > - even though doing this manually worked. (Means I don't program shell
> > scripts worth shit, probably )
>
> Eh? Add modprobe under start) case in /etc/pcmcia/ide script. Or what
> have you done?
>
Have re-added what i thought i did earlier.. probably typed it wrong or
someting. Anyway, after re-adding it, and rebooting, I can see the drive,
and cdrecord can see it as well.
added to /etc/pcmcia/ide two lines:
'start')
[ -n "$VERBOSE" -a -n "$INFO" ] && echo "$INFO"
add_parts "$ADDRESS" "$PARTS" || exit 1
/usr/sbin/updfstab
/sbin/modprobe -r ide-scsi
/sbin/modprobe ide-scsi
;;
