Re: Loading kernel modules permanently

[Rick Stevens <rstevens vitalstream com>]

Les Birman wrote:
> Rick,
> These are the modules needed to be able to run cdrecord.  I've got an IDE
> CDRW drive so the SCSI simulation driver(s) must be loaded.  I've had to
> tell the kernel not to load the the IDE CD-ROM driver when booting by adding
> a line in my lilo.conf file.   Also had to redo the link  in /dev for the
> cdrom to point to the simulated scsi device (cdrom -> scd0).  There is no
> access to the CDRW drive unless I insmod the aforementioned modules.  Also
> can not run cdrecord because it complains about no CDRW drive found (until I
> manually load the modules).  So why do you think these modules should
> autload?

If you use the "hdc=ide-scsi" thing in lilo.conf, the dependencies for
sr_mod, scsi_mod and the rest of the lot _should_ autoload when you use
cdrecord (or gtoaster or x-cdroast or any of the others).

I have three systems with conventional IDE CDROMs as slaves on the first
IDE and IDE CD-Rs as masters on the second IDE and that's how I do it.
One system (home) is a dual-proc 2GHz Athlon box, one is a single proc
P-III/850MHz and the third is a Sony Vaio laptop (1GHz Athlon).  Well,
actually the laptop doesn't have a CD as slave on the first IDE,
it only has the CD-RW/DVD drive as master on the second IDE).

The boot code also creates links from /dev/scd0 (the CD-R due to the
"hdc=ide-scsi" lilo option) to /dev/cdrom1.  So, the regular CD is
/dev/cdrom and the CD-R is /dev/cdrom1.  Simple.

Then again, I'm running kernel 2.4.18.  Under 2.2, your mileage may
vary.
----------------------------------------------------------------------
- Rick Stevens, SSE, VitalStream, Inc.      rstevens vitalstream com -
- 949-743-2010 (Voice)                    http://www.vitalstream.com -
-                                                                    -
-         "If you can't fix it...duct tape it!"  - Tim Allen         -
----------------------------------------------------------------------