Bug#793670: mount: bad optical disk can place mount command into
From: Ben Hutchings <ben@decadent.org.uk>
> On Sun, 26 Jul 2015 04:32:23 -0700 "Dallas E. Legan" <aw585@lafn.org> wrote=
> :
> > Package: mount
> > Version: 2.26.2-6
> > Severity: normal
> > Tags: upstream
> >=20
> > Dear Maintainer,
> > I've found trying to mount an optical disk that turns out to be bad,
> > (this is the case I've encountered, may possibly happen with other media)
> > can cause the mount command to go into a noninteruptable sleep state,
> > waiting for response from the hardware that never arrives.
>
> Block device I/O is normally uninterruptible; that's standard Unix
> behaviour that is unlikely to be changed.
OK.
It's only reading the data necessary to allow it to mount the
read only disk. If that I/O is interrupted, then there
won't be data needed for that mounting, so the mount command
must fail. It can't use guesses or typical values.
But for a standard 'ROM' like optical disk,
the data can't be corrupted any more than it already is by
the mount failing.
Anyway, that's my perspective as an ignoramous.
>
> However there should be a timeout in the driver. How long did you wait?
I can't recall any specifics, but a reasonable amount of time for the
disk to get mounted.
Once stuck in the uninterruptible sleep mode, the mount command
would seem to go on forever.
At some point, I have to remove the disk to return it to the library. :-)
>
> > This latter can interfere in placing the computer in hibernate or
> > performing a clean shutdown.
> > My understanding is that the command could alternatively be implimented t=
> o
> > use a "killable" state, similar to the uninteruptable sleep,
> > except that the process can be killed.
> > If there are reason's the noninteruptable sleep must be used
> > most of the time, perhaps providing a switch to enforce
> > "killable" when the media is of unproven quality would be possible.
>
> Well, it's not as simple as that. The driver would have to allow the
> I/O to be cancelled.
Guess I'll just have to accept it.
The drift I'm getting is that this is a problem with the driver,
not the mount command. ??
>
> > Thanks for any consideration.
>
> Ben.
> Hoare's Law of Large Problems:
> Inside every large problem is a small problem struggling to get out=
> ...
Thanks for your analysis.
Regards,
Dallas E. Legan II
legan@acm.org / aw585@lafn.org /
http://www.lafn.org/~aw585/index.html
---------------------------------------------
This message was sent using Endymion MailMan.
http://www.endymion.com/products/mailman/
Reply to: