From owner-freebsd-mobile Wed Dec 6 12: 7:12 2000 From owner-freebsd-mobile@FreeBSD.ORG Wed Dec 6 12:07:04 2000 Return-Path: Delivered-To: freebsd-mobile@freebsd.org Received: by hub.freebsd.org (Postfix, from userid 618) id A2BFF37B401; Wed, 6 Dec 2000 12:07:03 -0800 (PST) Subject: Raylink driver nits (plus patch) To: freebsd-mobile@freebsd.org, dmlb@dmlb.org Date: Wed, 6 Dec 2000 12:07:03 -0800 (PST) X-Mailer: ELM [version 2.4ME+ PL54 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Message-Id: <20001206200703.A2BFF37B401@hub.freebsd.org> From: wpaul@FreeBSD.ORG (Bill Paul) Sender: owner-freebsd-mobile@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I took a quick look at /sys/dev/ray/if_ray.c and noticed a couple things that aren't quite right. I'm including a patch to fix them. The problems are: - In ray_attach(), the code calls ether_ifattach(), then later it calls bpfattach(). This is unnecessary: the whole point of calling ether_ifattach() with the BPF_SUPPORTED flag is that it will do the bpfattach() for you. Calling bpfattach() again is unnecessary, and probably causes bugs, or at the very least a resource leak. - In ray_attach(), the code checks for ifp->if_name being NULL before initializing the ifnet structure info. This isn't necessary anymore. It used to be in FreeBSD 3.x and earlier that there was no if_detach() function, so once you created an ifnet structure you couldn't get rid of it, and you had to be careful not to initialize/attach it twice. Now the test is redundant. - In ray_detach(), the code calls if_detach() when it should be calling ether_ifdetach(). Among other things, this does a bpfdetach(), which isn't being done now and should be. Again, patch is included. The raylink driver maintainer can check this in, or if nobody objects, I can do it. -Bill *** if_ray.c.orig Wed Dec 6 11:54:55 2000 --- if_ray.c Wed Dec 6 11:56:46 2000 *************** *** 491,516 **** /* * Initialise the network interface structure */ ! if (!ifp->if_name) { ! bcopy((char *)&ep->e_station_addr, ! (char *)&sc->arpcom.ac_enaddr, ETHER_ADDR_LEN); ! ifp->if_softc = sc; ! ifp->if_name = "ray"; ! ifp->if_unit = device_get_unit(dev); ! ifp->if_timer = 0; ! ifp->if_flags = (IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST); ! ifp->if_hdrlen = sizeof(struct ieee80211_frame) + ! sizeof(struct ether_header); ! ifp->if_baudrate = 1000000; /* Is this baud or bps ;-) */ ! ifp->if_output = ether_output; ! ifp->if_start = ray_tx; ! ifp->if_ioctl = ray_ioctl; ! ifp->if_watchdog = ray_watchdog; ! ifp->if_init = ray_init_user; ! ifp->if_snd.ifq_maxlen = IFQ_MAXLEN; ! ether_ifattach(ifp, ETHER_BPF_SUPPORTED); ! } /* * Initialise the timers and driver --- 491,514 ---- /* * Initialise the network interface structure */ ! bcopy((char *)&ep->e_station_addr, ! (char *)&sc->arpcom.ac_enaddr, ETHER_ADDR_LEN); ! ifp->if_softc = sc; ! ifp->if_name = "ray"; ! ifp->if_unit = device_get_unit(dev); ! ifp->if_timer = 0; ! ifp->if_flags = (IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST); ! ifp->if_hdrlen = sizeof(struct ieee80211_frame) + ! sizeof(struct ether_header); ! ifp->if_baudrate = 1000000; /* Is this baud or bps ;-) */ ! ifp->if_output = ether_output; ! ifp->if_start = ray_tx; ! ifp->if_ioctl = ray_ioctl; ! ifp->if_watchdog = ray_watchdog; ! ifp->if_init = ray_init_user; ! ifp->if_snd.ifq_maxlen = IFQ_MAXLEN; ! ether_ifattach(ifp, ETHER_BPF_SUPPORTED); /* * Initialise the timers and driver *************** *** 518,524 **** callout_handle_init(&sc->com_timerh); callout_handle_init(&sc->tx_timerh); TAILQ_INIT(&sc->sc_comq); - bpfattach(ifp, DLT_EN10MB, sizeof(struct ether_header)); /* * Print out some useful information --- 516,521 ---- *************** *** 587,593 **** sc->gone = 1; sc->sc_havenet = 0; ifp->if_flags &= ~(IFF_RUNNING | IFF_OACTIVE); ! if_detach(ifp); /* * Stop the runq and wake up anyone sleeping for us. --- 584,590 ---- sc->gone = 1; sc->sc_havenet = 0; ifp->if_flags &= ~(IFF_RUNNING | IFF_OACTIVE); ! ether_ifdetach(ifp, ETHER_BPF_SUPPORTED); /* * Stop the runq and wake up anyone sleeping for us. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-mobile" in the body of the message