DVD players (hardware)

[James Wilkinson]

Daniel Stonier wrote:
> I've been having problems lately with my dvd player - seems to be locking  
> up frequently on dvd's in a movie and repetitiously
> sitting in the one spot trying to read without success. Admittedly some of  
> the problems are due to dvd's with scratches, but it
> seems to do it far more frequently now the player is getting a little aged.
> 
> Anyway, my question is this - how do you kill it? Only way I know of  
> currently is to reboot the computer. Killing the process
> often doesn't work and if it does occasionally, the dvd hardware itself  
> still keeps beating itself senseless.

No-one seems to have answered this one, so I'll see what I can do.

Firstly, you do know about kill -9, don't you? This will kill anything
that can be killed.

If that doesn't work, then the chances are that the process is in the
middle of a system call (and is into kernel mode). When this happens, the
signal is only acted upon when the system call returns. If you have
suitably dodgy hardware (and it sounds like you do), the system call may
not actually return.

Note that this does not mean that the kernel freezes. When a synchronous
call is made that requires a hard disk read, Linux expects to go away for
at least 10 ms (more like 100 ms with a CD or DVD), and wait for the
storage adapter to notify the kernel (with an IRQ and suitable messages:
note that I'm assuming DMA).

That's just what happens here. It's just that the DVD never returns the
data, but doesn't return a read error, so the kernel remembers the
outstanding request indefinitely.

This is probably due to your DVD-ROM drive failing, although it could be
an oddity of the IDE adapter. Do you find the same symptoms with
CD-ROMs?  How about DVD-ROMs? (DVDs are a lot harder for a marginal
DVD-ROM drive to read, but there's not much difference to the IDE side
of things).

HTH,

James.

-- 
E-mail address: james@ | "During the shutdown period I received not one
westexe.demon.co.uk    | single support call, confirming my theory that my
                       | network is indeed perfect, and that all faults are
                       | user-inflicted."