Re: Re: cdrom ioctls

"Also sprach Steven Hartley:"
> I came across the same problem with the same IOCTL about 1-2 months ago
> and decided to just hard code that that ioctl returned always (drive
> ready), when I go back to the enbd project, I try our your code.

Yes - I am kinda working on the code (SPONSOR me somebody .. I need an
ia64 arch machine for experiments and I can get cpu and mobo n ebay for
about E150).  I should be doing some trials in a little while.

There is this "special" mechanism built into the enbd ioctl architecture
that can take care of the return value by transporting it back in an
extra buffer.  I'm just seeing how it all fits, since I've never used
this "special" mechanism before - it should allow one to customize
transfers on a per-ioctl basis, by using ones own methods, but I've
never used it before ...  it's just there.

The reason it's hard to move the return value across any other way is
that the return value ends up in a kernel request's "errors" field,
where anything apart from zero means baaaaad.  That's why I carefully
zeroed all positive ioctl returns before. 

Peter