Hello, Sorry; I did not completed my message; here is the full one:
Thanks a lot for your answer ! Following your answer I did some reading in LDD3 and "Undersatnding the linux kernel" and I want to verify if I may that I understood the interrupts mechanism right regarding my nethooks question: Suppose I will add another hook , of type NF_IP_FORWARD, with a callback method named fw_hook(), to my exsiting NF_IP_PRE_ROUTING hook in my module. As I understand, I can be sure that my_hook() and also fw_hook() will not be called a second time when an interrupt on the NIC arrives (when working on a single CPU) because they both are run under the same context (softirq context probably). So in fact, I do not have to worry about shared data structures which may appear in both hook methods (my_hook() and fw_hook()) Am i right ? John On 10/30/05, Harald Welte <[EMAIL PROTECTED]> wrote: > On Sat, Oct 29, 2005 at 03:39:30PM +0200, John Que wrote: > > Hello, > > I have written a module which uses NF_HOOKS for diverting packets to > > a different destination. > > > > My hook uses NF_IP_PRE_ROUTING. > > > > I call nf_register_hook() where the hook method is called my_hook(). > > > > I see that indeed packets are reaching my hook method , my_hook(). > > > > My question is : > > > > As I understand, the packets are reaching this method every > > interrupt of the Network Card for receiving > > netfilter hooks are not related to interrupts. they're called from > whatever context they're called from, mostly network receive softirq, > sometimes process context (LOCAL_OUT). > > > > Or to be more accurate: > > > > My question is : let's say that a packet sent to my machine reaches > > my_hook() method, > > > > and let's say that another packets are received on the NIC at the > > same time while > > > > the my_hook() method did not finished to process the previous > > packet. (there is a very high > > > > rate of sending packets to my machine). > > > > Is there a possiblility that my_hook() will be called on a second > > and third packets while still > > I suggest you do some basic reading on linux kernel development and the > related contexts. What you're looking for is "what happens when a > hardirq interrupts softirq context". > > Short reply: my_hook() will not be called a second time on the same cpu, > ubt will be called any number of times in parallel on multiple cpus. > > -- > - Harald Welte <[EMAIL PROTECTED]> http://netfilter.org/ > ============================================================================ > "Fragmentation is like classful addressing -- an interesting early > architectural error that shows how much experimentation was going > on while IP was being designed." -- Paul Vixie > > > - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
