Messages in this thread | | | Date | Tue, 7 Mar 2000 19:18:34 +0100 | From | Jamie Lokier <> | Subject | Re: Changes in the sockets: SIGIO handling |
| |
kuznet@ms2.inr.ac.ru wrote: > But do not forget that even generation of such signals is not a zero > overhead, which can be saved as rule. F.e. both old and rt signals > are not generated at all when someone _sleeps_ in read()/write(). > >From the viewpoint of previous mails, it is bug. Look, another > signal may wake us, then read() returns partial read. Oops, event is lost.
If you've got a partial read's worth of data pending, you shouldn't be sleeping. If you receive a signal during sleep (because the buffer was empty when you went to sleep), you should wake up and return -EINTR, even if the buffer is partially full now.
Well you don't have to do that, but that allows you to avoid this particular "event lost" bug.
-- Jamie
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |