On Wed, 2012-08-29 at 16:50 +0300, Alexey Dobriyan wrote:

On Wed, Aug 29, 2012 at 7:11 AM, Eric Dumazet [off-list ref] wrote:

quoted

I'll polish this patch once LKS/LPC is over...

It should oops in the following way (excuse Gmail please):
PDEO is removed from lists
->pde_users is 0
PDE won't be in purge queue -- no ->release while module is alive

Current code removes PDEO and checks if PDE scheduled for removal atomically.

proc_reg_release                                        remove_proc_entry

                                                        de->proc_fops = NULL;
release = pde_opener_del(inode, file);
rcu_read_lock();
                                                        synchronize_rcu();
fops = rcu_dereference(pde->proc_fops);
if (!fops) {
        rcu_read_unlock();
                         ----------------------------------
                                                        /* NOP */
                                                        while
(atomic_read(&de->pde_users))
                                                                ...
                                                        /* NOP */

pde_openers_purge(de, &purge_queue);
                                                        /* NOP */
                                                        while
(!list_empty(&purge_queue))
                                                                ...
                                                        rmmod

        if (release)
                release(inode, file) /* OOPS */

Fix should be trivial, proper module refcount for example.

As I said, I would do that after LKS/LPC, there is no hurry.