Re: frequent slab corruption (since a long time)

[Dave Jones <davej@redhat.com>]

On Thu, Aug 03, 2006 at 06:40:42PM +0100, Alan Cox wrote:
 > Ar Mer, 2006-08-02 am 15:49 -0700, ysgrifennodd David Miller:
 > > > 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?
 > > 
 > > It is in the primary code paths at least, all callers of init_dev()
 > > (which increments tty->count) grab the mutex and also release_dev()
 > > grabs the mutex around tty->count manipulations.
 > 
 > I've been auditing tty code and its joyously bad but only in harmless
 > places so far except for one.
 > 
 > init_dev (and caller) relies on tty_mutex to ensure that the
 > driver->ttys list is protected from things going away.
 > 
 > release_mem() removes stuff from the said list and frees memory. It is
 > not always called under tty_mutex and that appears very dubious to me at
 > the moment although tty->closing and the BKL *might* be sufficient.

Against my better judgment I was poring over that code until the wee
hours last night, and one thing crossed my mind re: the assumptions made
about the BKL in that subsystem.  Now that the BKL is preemtible, do
any of those assumptions break ?

		Dave


-- 
http://www.codemonkey.org.uk