From: Toni I. <ton...@ya...> - 2005-10-15 12:54:49
|
Okay I have compiled version http://195.66.192.167/linux/acx_patches/acx-20051010_txstop_fix.tar.bz2 and it works fine. But why? What is different in this version? What is _txstop_fix? Maybe my kernel is bad and I should recompile it! --- Denis Vlasenko <vd...@il...> wrote: > On Saturday 15 October 2005 15:04, Toni Inger wrote: > > Hello everybody!!! > > I have D-link 520+ and I have Debian 3.1 Sarge > with > > 2.6.12 kernel. I am using firmware from > > > http://195.66.192.167/linux/acx_patches/fw/acx100_1.9.8.b/. > > Version 20051007 worked for me quite good, without > any > > problems. But none from versions: 20051012, > 20051013 > > and 20051014 works for me. Modprobe acx give good > > results. Unfortunatelly, when I run: > > /etc/init.d/networking start, it loads good to the > > moment of authentication. There are many acx: BUG > like > > messages. My Internet provider told me that he is > > using ORINOCO access point. I also attached my > log, > > but when I run modprobe acx debug=0xffff and > > /etc/init.d/networking start, my computer hangs > and > > generates a lot of messages very fast (they are in > > log). > > Greetings, > > Toni Inger > > Oct 15 13:24:22 localhost kernel: 94855772 > ==> acxpci_l_process_rxdesc > Oct 15 13:24:22 localhost kernel: rx: buf 10 full > Oct 15 13:24:22 localhost kernel: rx: buf 11 full > Oct 15 13:24:22 localhost kernel: 94855772 > <== acxpci_l_process_rxdesc > > I do not understand how this can be ever possible. > You should reach the acxlog("rx: tail=%u Ctl_16=%04X > Status=%08X"...) > if you see "rx: buf 11 full" messages. > > More precisely, I fail to understand how in hell > the line "break; /* found it! */" was not > reached: > > static void > log_rxbuffer(const wlandevice_t *priv) > { > const struct rxhostdesc *rxhostdesc; > int i; > > /* no FN_ENTER here, we don't want that */ > > rxhostdesc = priv->rxhostdesc_start; > if (!rxhostdesc) return; > for (i = 0; i < RX_CNT; i++) { > if ((rxhostdesc->Ctl_16 & > cpu_to_le16(DESC_CTL_HOSTOWN)) > && (rxhostdesc->Status & > cpu_to_le32(DESC_STATUS_FULL))) > printk("rx: buf %d full\n", > i); > rxhostdesc++; > } > } > #endif > > static void > acxpci_l_process_rxdesc(wlandevice_t *priv) > { > rxhostdesc_t *hostdesc; > int count,tail; > > FN_ENTER; > > if (unlikely(acx_debug & L_BUFR)) > log_rxbuffer(priv); > > /* First, have a loop to determine the first > descriptor that's > * full, just in case there's a mismatch > between our current > * rx_tail and the full descriptor we're > supposed to handle. */ > count = RX_CNT; > tail = priv->rx_tail; > while (1) { > hostdesc = > &priv->rxhostdesc_start[tail]; > /* advance tail regardless of > outcome of the below test */ > tail = (tail + 1) % RX_CNT; > > if ((hostdesc->Ctl_16 & > cpu_to_le16(DESC_CTL_HOSTOWN)) > && (hostdesc->Status & > cpu_to_le32(DESC_STATUS_FULL))) > break; /* found it! > */ <======================= HERE > > if (--count) /* hmm, no luck: all > descs empty, bail out */ > goto end; > } > > /* now process descriptors, starting with > the first we figured out */ > while (1) { > acxlog(L_BUFR, "rx: tail=%u > Ctl_16=%04X Status=%08X\n", > tail, hostdesc->Ctl_16, > hostdesc->Status); > > ... > -- > vda > > > ------------------------------------------------------- > This SF.Net email is sponsored by: > Power Architecture Resource Center: Free content, > downloads, discussions, > and more. > http://solutions.newsforge.com/ibmarch.tmpl > _______________________________________________ > Acx100-users mailing list > Acx...@li... > https://lists.sourceforge.net/lists/listinfo/acx100-users > __________________________________ Yahoo! Music Unlimited Access over 1 million songs. Try it free. http://music.yahoo.com/unlimited/ |