From mboxrd@z Thu Jan 1 00:00:00 1970 From: Philip Craig Subject: Re: [RFC] [PATCH] clean up nf_log API Date: Thu, 11 May 2006 18:58:42 +1000 Message-ID: <4462FCC2.90205@snapgear.com> References: <20060510191034.GA29531@sunbeam.de.gnumonks.org> <4462E031.4090206@trash.net> <20060511082555.GB10324@sunbeam.de.gnumonks.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Netfilter Development Mailinglist , Patrick McHardy Return-path: To: Harald Welte In-Reply-To: <20060511082555.GB10324@sunbeam.de.gnumonks.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: netfilter-devel-bounces@lists.netfilter.org Errors-To: netfilter-devel-bounces@lists.netfilter.org List-Id: netfilter-devel.vger.kernel.org On 05/11/2006 06:25 PM, Harald Welte wrote: > I agree, I prefer automatic fallback while giving the user the ability > to unload. However, if people automatically unload 'unused' modules, > then there might be a problem (however, for both ULOG and nfnetlink_log, > userspace listeners would increment the refcount, I think). What constitutes in use? If a backend is in use whenever it is the highest priority backend, then blocking unloading would mean nfnetlink_log can never be unloaded. Letting the user choose the highest priority backend would solve that. In which case I think blocking unloading is desirable, we shouldn't unload the backend the user has chosen. And I assume that log rules can also specify a backend? (eg the LOG and ULOG rules always will) That should probably block unloading too. > Also, there is another unresolved problem: What if we actually have > e.g. nfnetlink_log loaded, but it is unable to send a specific packet to > userspace because there is no listener for this Address family or log > group? Should we then also fall back on some other backend? Just doing a printk to notify of the error is enough? We probably should be doing a printk anyway if it is a backend the user has chosen explicitly.