Re: CD drives on vax

[Johnny Billquist <bqt%update.uu.se@localhost>]

On 2012-06-21 13:10, Anders Magnusson wrote:
> On 06/21/2012 01:02 PM, David Brownlee wrote:
> > On 21 June 2012 11:59, Martin Husemann <martin%duskware.de@localhost> wrote:
> > > On Thu, Jun 21, 2012 at 11:49:32AM +0100, David Brownlee wrote:
> > >
> > > > Ideally I'd like sysinst to be able to default to the "right" cd
> > > > device when possible, by detecting if a scsi cd* or MSCP ra* CD drive
> > > > is attached.
> > > Sysctl hw.disknames lists them, we only need to check for ra* on vax.
> > >
> > > Is there a way to test wether a ra* cd drive has a medium inserted?
> > > Sysinst uses TEST UNIT READY on scsi/atapi devices via a special ioctl,
> > > we could add support for that to the ra driver.
> > I thought of replacing CD_NAME in sysinst with an array of potential
> > drives, and sysinst picking the first one found in hw.disknames, but
> > that is complicated by the fact ra* is used for MSCP attached hard
> > drives as well, though they are differentiated in dmesg:
> > ra0 at mscpbus1 drive 0: RRD40 <-- MSCP CD drive
> > ra1 at mscpbus1 drive 1: RA82   <-- RA82 MSCP disk
> >
> > So maybe an md routine which returns the default CD partition. On most
> > ports it could return just "cd0a" while on vax it could pick the first
> > RRD40 marked ra* in dmesg, or fallback to cd0?
> It would be simple to set a separate device name for mscp cd:s, floppies
> are handled separately.
> See src/sys/dev/mscp/mscp_disk.c:ramatch() and rxmatch().  The same
> could be done for RRD40.

I'm not sure it will be that easy. The ra have a major number of 9, 
while rx have a major number of 12. However, cd have another major 
number, and that will be associated with the mi cd driver. Can we 
somehow get this to play along with a md cd driver?

        Johnny