Manuel Bouyer writes:

>I don't remember and I didn't find traces of this in the CVS logs ...

I remember very, _very_ clearly. (I spent a lot of time in the lab
booting test kernels, both with and without a CD in the drive).

Maybe i have the date wrong? Sometime between november and february.

This is at least the third time this particular board/drive has
stopped working with the pciide patches.  I beleive Intel donated
swags of them to various US universities around late 95/early 96:
q(Stanford CS has about 300, CalTech has a lab or two full of them,
etc.)

No blame, but it is a bit frustrating.

I've been round the loop a few times already, and I'd have sent you a
log with the wdc and atapi debug flags turned on, but I hadn't
installed the userland, and trying to recompile locally with
gcc-2.7.2.2 (on a possibly-buggy 1.3H) yielded a kernel that panicked
scanning SCSI children of a bha. Yucko.

Thanks, and I'll try and get the debug output tomorrow.