Good morning.

James Morris wrote:
> > - Updating the sel_netif_hash table is protected by spinlock_irqsave().
> > This method is same as my RCU-patch.
>
> Why use irqsave? This code is never contended from irq context.

Hmm, I worried about using spinlock() without _irqsave(), since I faced to the another problem as you know. Your notice is right.

> > # And, I wondered if 'sel_netif_total' is really necessary.
>
> This prevents the cache from growing without bounds.

But a result of sel_netif_insert() was not checked in the original code. It's assumed the number of 'netif' is less than SEL_NETIF_HASH_MAX. And, Its upper limitation is the number of the real network interfaces. Why do you think the cache may grow without bounds?

Thanks.

--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.