Re: frequent slab corruption (since a long time)

[Alan Cox <alan@lxorguk.ukuu.org.uk>]

Ar Mer, 2006-08-02 am 18:23 -0400, ysgrifennodd Dave Jones:
> None of the code manipulating tty->count seems to be under
> the tty_mutex.  Should it be ?
> Or is this protected through some other means?

Its old lock_kernel() code so its a prime suspect for most offences.

Given the age of the reports it appears that the tty buffering changes
are too new. The ldisc locking changes are a candidate but shouldn't be
doing anything that breaks the kernel lock stuff. Will look tomorrow see
if anything strikes me about that lot.

The tty_mutex primarily protected current->signal.tty (except if you are
using SELinux which has some extremely dubious looking code for tty
handling). Someone ought to review flush_unauthorized_files() although
it doesn't fit this problem.

Alan