Re: atapi cdrom on laptop

[Marco S Hyman <marc@snafu.org>]

Steve Slater writes:
 > The GENERIC kernel had all of the specific devices set as wildcards
 > and I had both controllers enabled. I tried the changes used here
 > and even a few more. wdc1 is not even recognized at boot so
 > I'm assuming the toshiba 550 only has a single controller.

Does it work on a hard reset, e.g. a power cycle?  I ask because
this is what I found on an older Toshiba CDT 400

    wdc probe tests for presence of a controller by writing two values,
    one to a write only register and one to a read-write register.  The
    two registers are then read.  If the write only register returns the
    value written or if the read-write register does not return the
    value written the device is assumed not to exist.

    The problem is that the second controller on my Toshiba portable
    fails this test unless a power cycle (or hard reset) is performed.

I've got a local patch that forces a controller reset on wdc1 before
the probe and it has solved my prob.  However, that is not a general
solution.  Toby suggested:

    One thing you might try, is to do a controller reset inside a shutdown
    hook (shutdownhook_establish()), that might do the trick on soft reboots.

One of these days I'll code that up.  If it works it would probably be
ok to commit as the reset would only be performed if the device is known
to exist.

// marc