FREQUENTLY ASKED QUESTIONS AND SOME COMMON TROUBLES =================================================== Intended audience: ECI ADSL USB modem driver (up to version 0.12) users. Before reading the following: the following Q/A list is *not* exhaustive. It is also possible that you won't find any Q/A that matches your problem. Please read them anyway, they may provide you help in fixing your problem anyway. And even if your problem is listed below, it is possible that the answers don't help you anymore to solve it. Then, please contact us, we will try to help you finding a solution. 1 - Driver compilation, installation, and configuration 2 - Modem initialization 3 - Modem synchronization 4 - Connection problems 5 - Various topics 1.0 ================================================================================ Q: I don't know if my modem is supported. I don't see my modem in the list (using eciadsl-config-tk or eciadsl-config-text). -------------------------------------------------------------------------------- A: Check the pre-requisites. Check at http://eciadsl.flashtux.org/modems.php?lang=en Install the latest driver version, then launch eciadsl-start or eciadsl-doctor, if they report errors, try to launch eciadsl-probe-device to check your modem compatibility. If still uncertain, contact us! 1.1 ================================================================================ Q: `./configure` fails and returns some errors :-\ -------------------------------------------------------------------------------- A: Please report us the output of `./configure`. Check which options (see `./configure --help`) could help you. For instance, you may use --disable-gcc-check if `./configure` fails in finding gcc or checking its version number whereas gcc is installed and is in your $PATH. This error may happen on some systems, which include a modified gcc (Debian for instance). 1.2 ================================================================================ Q: I get the following error when I run `make` or `make install`: Makefile:1: Makefile.config: No such file or directory make: *** No rule to make target `Makefile.config'. Stop. -------------------------------------------------------------------------------- A: Did you run `./configure`? You should have! (>= 0.6-pre4 only) 1.3 ================================================================================ Q: `eciadsl-config-tk` immediately fails, saying: wish: command not found -------------------------------------------------------------------------------- A: Install tcl/tk, or run `eciadsl-config-tk`. Even if tcl/tk is not mandatory, check out the pre-requisites. 1.4 ================================================================================ Q: eciadsl-probe-device shows me VID1/PID1 that are the same as VID2/VID2, is this normal? -------------------------------------------------------------------------------- A: You must reset your modem (unplug/wait/replug) before launching eciadsl-probe-device. Please read the WARNING displayed by this script when it starts. This may be due to more general problems problems in your system's USB configuration. Either hotplug interfers with the driver, or kernel's USB modules have problems with your hardware (might be a kernel bug too). Also check if dabusb module is showm in /var/log/messages (see dabusb-related question above). 1.5 ================================================================================ Q: eciadsl-start, eciadsl-config-tk or eciadsl-doctor report me an unsupported modem! -------------------------------------------------------------------------------- A: See 1.4. If your plugged modem isn't showed in /proc/bus/usb/devices, look at /var/log/messages. If you see such lines appearing during the last boot: kernel: usb_control/bulk_msg: timeout kernel: usb.c: USB device not accepting new address=2 (error=-110) there's a problem between you hardware/BIOS/kernel config, it's an interrupts issue. It may worth disabling APIC support in kernel or at boot time (use noapic when booting the kernel, but it may come from other IRQ conflicts between hardware devices, or some other kernel configuration issue, or also to disable/enable USB legacy support in BIOS. Such problems are common issues with some USB chipsets like old VIA and SiS 700x. If you don't know which chipset you have or if you want to get confirmation that your problem is a hardware/kernel driver issue, please contact us. If confirmed, a possible solution would be to buy a USB HUB to plug into a free PCI slot (don't try to plug an external USB HUB in a USB port from your embedded USB HUB, this will not work, of course since the problem comes from the on-board chipset). Don't buy such PCI card if you cannot test it before! Also check the question that deals with the dabusb module, if dabusb is loaded, you won't be able to use your modem with this driver until you disable the loading of dabusb. 1.6 ================================================================================ Q: I don't know my provider's DNS IP. -------------------------------------------------------------------------------- A: Most people need them (except those who use ISP DHCP for instance), if you can't find them at your provider's website or in the configuration CD or the paper they've send to you, or even under MS Windows, there's a way to guess them. Here is the way to do this under Linux: > host -t ns This returns you a bunch of DNS server names. You can ping some of them in order to get the IP. For instance: > host -t ns clix.com clix.com. name server ns4.dnsmanaged.com. clix.com. name server ns1.dnsmanaged.com. clix.com. name server ns7.dnsmanaged.com. clix.com. name server ns6.dnsmanaged.com. > ping -c 1 -q ns4.dnsmanaged.com PING ns4.dnsmanaged.com (192.31.80.34) from 10.0.0.99 : 56(84) bytes of data. --- ns4.dnsmanaged.com ping statistics --- 1 packets transmitted, 1 received, 0% loss, time 0ms rtt min/avg/max/mdev = 169.466/169.466/169.466/0.000 ms The IP of this ns4.dnsmanaged.com is 192.31.80.34. Repeat the ping command with another DNS server name, then you'll get at least 2 DNS IP to use :-). If the config tool from the ECIADSL driver don't allow you to select your provider's DNS, select 'Other' and type in at least one DNS IP to use. 2.0 ================================================================================ Q: eciadsl-start, eciadsl-config-tk or eciadsl-doctor's output shows: /proc/bus/usb: No such file or directory -------------------------------------------------------------------------------- A: Check your system's USB configuration. This may be due to a lack of support for usbdevfs in your kernel configuration. If kernel has support for it, try to mount it using: > mount -t usbdevfs none /proc/bus/usb This can be set in your /etc/fstab. To get his automounted at boot time, add this line: usbdevfs /proc/bus/usb usbdevfs defaults 0 0 Nota: there will be nothing in /proc/bus/usb until a host controller module is also loaded. 2.1 ================================================================================ Q: ecidoctor reports me that the DABUSB module is loaded, or eciadsl-start says dabusb has been found, but couldn't remove it! -------------------------------------------------------------------------------- A: Hotplug is probably enabled, and it wrongly detects your modems as an audio device and loads dabusb module in order to add support for this audio device. If /etc/hotplug/blacklist exists, edit it and add a line containing the word 'dabusb' (without the quotes) to it. Restart Linux. If you cannot find such file whereas hotplug is installed and enabled, there must be another way to configure it, but you can also apply the following method (a bit rough): Boot your Linux machine with your modem *UNPLUGGED*, then: You can remove the dabusb module from your system using eciadsl-config-tk or eciadsl-config-text. You can also directly call eciadsl-remove-dabusb (in /usr/local/bin by default). Or, manually type the following command: > modprobe -r dabusb && rm -f $(modprobe -l | grep dabusb) && depmod -a If this kernel has been compiled by hand, don't forget to remove dabusb support from the kernel configuration too. 2.2 ================================================================================ Q: No way to remove dabusb! At each boot, it is loaded again! -------------------------------------------------------------------------------- A: See 2.1, but use the manual method. 2.3 ================================================================================ Q: I try to remove the dabusb module, but always get: dabusb: Device or resource busy -------------------------------------------------------------------------------- A: See 2.2. 2.4 ================================================================================ Q: My modem seems initialized at boot, the LEDs are flashing or fixed. Does this mean that my system recognized my modem and that I don't have to install the driver? -------------------------------------------------------------------------------- A: No! This is not friendly at all. That's a dabusb module issue. See 2.3. 2.5 ================================================================================ Q: `eciadsl-doctor` reports me: HDLC support is buggy, you should apply the HDLC patch to your kernel source. -------------------------------------------------------------------------------- A: Follow the pre-requisites, read the documentation about the N_HDLC problem. If you want to use the 'persist' option of pppd in order to reconnect automatically when disconnected from the provider, then you will have to patch your kernel or use a kernel >= 2.4.18-pre3. You can find the patch n_hdlc.c.diff in the drivers' archive of the Speedtouch modem: http://speedtouch.sourceforge.net/ These are instructions about how to use it: > cd /usr/src/linux > patch -p1 --dry-run < /path/to/n_hdlc.c.diff If no error message is returned by the previous command, type the following line to make the source's patch: > patch -p1 < /path/to/n_hdlc.c.diff then jump to the kernel configuration question (5.3). 2.6 ================================================================================ Q: eciadsl-start says: eciadsl-firmware: timeout -------------------------------------------------------------------------------- A: Your modem is probably not supported, start to read this FAQ from the beginning. 2.7 ================================================================================ Q: eciadsl-start/eciadsl-probe-device/eciadsl-doctor cannot find my modem. -------------------------------------------------------------------------------- A: Your modem is probably not supported or there's a problem with your USB HUB hardware, it might also be a bug in the kernel modules. 3.0 ================================================================================ Q: eciadsl-start says: eciadsl-synch: timeout -------------------------------------------------------------------------------- A: See 3.1. 3.1 ================================================================================ Q: eciadsl-start says: eciadsl-synch: failed -------------------------------------------------------------------------------- A: See 3.2. Known problem: under GNOME or KDE, eciadsl-synch may not get the synch. Try the latest version, or contact us. This may also happen if you run eciadsl-start while your CPU is under heavy load. 3.2 ================================================================================ Q: eciadsl-start is stalled at block xxx. -------------------------------------------------------------------------------- A: Follow the pre-requisites. Ensure that the phone line is properly plugged at both ends (modem and wall jack), and that DSL microfilters are placed between the wall adapters and any device (telephone, fax, answering devices), check if your microfilters are OK. Upgrade your eciadsl driver version (0.5 owners: get version 0.6 or the latest CVS). Try the other synch .bin's (available online). 4.0 ================================================================================ Q: /var/log/messages shows CHAP or PAP failures. -------------------------------------------------------------------------------- A: Check your /etc/ppp/pap-secrets or /etc/ppp/chap-secrets files. Entries must be like this: "username" * "userpassword" * You can configure this using eciadsl-config-tk. Following the pppd version, the syntax of pap-secrets and chap-secrets may differ. If so, try to modify them by hand or contact us. 4.1 ================================================================================ Q: eciadsl-start went OK, but I still cannot use Internet access? -------------------------------------------------------------------------------- A: See 4.0. If you can ping an IP but cannot ping a hostname, check at your /etc/resolv.conf file, it must include your provider's DNS IP. For instance (French Wanadoo ones): nameserver 193.252.19.3 nameserver 193.252.19.4 You can set this using eciadsl-config-tk. If you still cannot access a hostname using nslookup, try the following commands and report us the output: > route -n (or netstat -rn) Kernel IP routing table Destination Gateway Genmask Flags MSS Window irtt Iface * 80.14.50.1 0.0.0.0 255.255.255.255 UH 40 0 0 ppp0 * 10.0.0.0 0.0.0.0 255.255.255.0 U 40 0 0 eth0 * 127.0.0.0 0.0.0.0 255.0.0.0 U 40 0 0 lo 0.0.0.0 80.14.50.1 0.0.0.0 UG 40 0 0 ppp0 The lines with a leading * are optional, your eth local network should differ from 10.0.0.0 (if you have one). The UG line must appear, this indicates the default route to the gateway. > ifconfig ppp0 ppp0 Link encap:Point-to-Point Protocol inet addr:80.14.50.227 P-t-P:80.14.50.1 Mask:255.255.255.255 UP POINTOPOINT RUNNING NOARP MULTICAST MTU:1500 Metric:1 RX packets:80951 errors:0 dropped:0 overruns:0 frame:0 TX packets:85395 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:3 The IP and other values will be different from the above ones. If ppp0 interface is not up, check for error lines in /var/log/messages or /var/log/ppp after having launched eciadsl-start. If the default route (UG) is set to eth0, remove it: > route del default dev eth0 or remove the "gateway" line from your /etc/sysconfig/network, or disable your LAN network before launching eciadsl-start. If the default route is NOT set to ppp0 even if ppp0 is up: > route add default dev ppp0 This may be a firewall issue. Try to disable your firewall, if this works now, you have to adjust your firewall's settings. Mandrake 9 users: if you see lines like the following in your /var/log/messages: Shorewall:INPUT:REJECT:IN=ppp0 .... then you can be sure that it's a firewall issue. 4.1.1 ================================================================================ Q: ppp0 seems up (in /var/log/message and eciadsl-start's output) but eciadsl-start still says "no ppp0 device found" after a while. The problem could be encountered as well if your configuration requires tap0 or tun0 instead of ppp0. -------------------------------------------------------------------------------- A: This may happen is net-tools package is not installed or not properly configured for the user that runs eciadsl-start (usually root). eciadsl-start needs `route` and `ifconfig` from this package and also /sbin to be somewhere in the user's PATH. Following the system, you may need /sbin or /usr/sbin or /usr/local/sbin or both, please check. 4.2 ================================================================================ Q: I got disconnected after a while, pppd's persist options seems to work but I cannot access to anything on Internet. -------------------------------------------------------------------------------- A: Do you know that pppd calls /etc/ppp/ip-down when it got desconnected and ip-up once reconnected? Maybe your distro's default ip-down restores basic network route table or removes the DNS in /etc/resolv.conf? 4.3 ================================================================================ Q: After a few minutes of Internet access, nothing seems to work anymore. No IP ping, neither http access nor nslookup. -------------------------------------------------------------------------------- A: Often seen on some recent distros (MDK9, Slackware 8.1 and RH8). Check the USB driver you're using with lsmod command. If you're using uhci, unload it (modprobe -r uhci) and load usb-uhci (modprobe usb-uhci). Did you see a disconnection (and a reconnection maybe)? In /var/log/messages or /var/log/ppp? If so, jump to the previous question. Else it might be network/firewall configuration related, but not sure. Try `network down` is you have but don't need local network. 4.4 ================================================================================ Q: I get LCP timeout lines in /var/log/messages then a modem hangup (and eventually another problem more between both) like this: sent [LCP ConfReq id=0x1 ] .. LCP: timeout sending Config-Requests Connection terminated. -------------------------------------------------------------------------------- A: If your kernel really OK? N_HDLC support is OK? Check the pre-requisites first (see the README file). This may be a synch issue. Try the other synch .bin available at the official download page. See 4.5 also, because CHAP/PAP authentication problems are not always verbosely reported by pppd. Questions 4.6 to 4.9 may help too. See also below questions related to international users (5.4). 4.4.1 ================================================================================ Q: I get something like this: pppd[1954]: pppd 2.4.1 started by root, uid 0 pppd[1954]: Using interface ppp0 pppd[1954]: Connect: ppp0 <--> /dev/pts/1 endless.. or, after a while: pppd[1954]: Child process /usr/local/bin/eciadsl-pppoeci -vpi 8 -vci 35 -vendor 0x0915 -product 0x8000 -mode VCM_RFC2364 (pid 1955) terminated with signal 2 pppd[1954]: Modem hangup pppd[1954]: Connection terminated. pppd[1954]: Exit. -------------------------------------------------------------------------------- A: Check your PAP and/or CHAP files (/etc/ppp/pap-secrets and /etc/ppp/chap-secrets). This may happen with broken authentication even if pppd is not verbose regarding to this in /var/log/message. See also Q 4.4. 4.5 ================================================================================ Q: Sometimes I see some LCP timeouts in /var/log/messages, but my connection is still up or it doesn't disconnect every time. -------------------------------------------------------------------------------- A: Try to uncomment the two lines that deal with LCP in /etc/ppp/peers/adsl, and maybe to slightly increase the values. This may be due to your provider too, wait for a few minutes/hours if you still can't connect. If these LCP timeouts don't affect your PPP connection, simply forget them or try another synch .bin. 4.6 ================================================================================ Q: I get a modem hangup or I'm not able to connect at all, and I see these lines in /var/log/messages: ioctl(PPP..): Inappropriate ioctl for device Connexion failed -------------------------------------------------------------------------------- A: Check your kernel config (below). Upgrade your eciadsl driver version. Try another synch .bin. This can happen when the invocation of eciadsl-pppoeci in the /etc/ppp/peers/adsl file is corrupted or wrong (probably because it has been edited by hand?). For instance, the use of "-vendor 0915" is wrong. The exact syntax is "-vendor 0x0915". Use eciadsl-config-tk to properly configure the driver! More generally, this may also come from a bad configuration, for instance if the PPP mode is not the one you need. Enable verbose logging to eciadsl-pppoeci (-v 2), and if you see in eciadsl-pppoeci's log file these lines: hi! I'm the parent process, I handle the endpoint 0x07 file descriptors: fdin=3, fdout=4 error loading N_HDLC you must ensure that your kernel has n_hdlc module (pre-requisite) see the relevant question (5.3)! 4.7 ================================================================================ Q: I get some "USB timeout" lines in /var/log/messages. -------------------------------------------------------------------------------- A: Try to uncomment the MTU line in /etc/ppp/peers/adsl. Eventually modify this MTU value (see `man pppd`). This may also an kernel/hardware issue (see the following questions). 4.8 ================================================================================ Q: PPP line sometimes goes down (modem hangup). -------------------------------------------------------------------------------- A: This may be due to high traffic on the line or to any provider problem/daily disconnection (non-exhaustive list). Enable the 'persist' option for pppd in /etc/ppp/peers/adsl (this requires N_HDLC support in the kernel, see the kernel configuration related question below). Verify that this persist option is set (by default it is). Decreasing the MTU in /etc/ppp/peers/adsl to 1000 or even to 512 may help. You may also try using a third-party autoconnect script, and of course check if your configuration has been properly set. 4.9 ================================================================================ Q: I get modem hangups but my problem doesn't match the previous questions or my modem goes OFF! -------------------------------------------------------------------------------- A: Is your USB chipset an old VIA one or a SiS 700x? Some are known to be buggy and to have problems regarding to the USB power supply or a buggy kernel module. If you modem goes OFF (no power), this might be your case (VIA). On some systems (bugged USB chipset or old processors), this may happen when you use multiple USB device that sucks a lot of power (video devices, hdd). This may also happen if you CPU is heavily loaded (burning CD or so) or when the USB bus is heavily loaded too (webcam use, etc.). 4.10 ================================================================================ Q: I see strange characters on the terminal/console after eciadsl-start completes, and I get no PPP connection: Connect Modem ... ~ÿ}#À!}!}!} }4}"}&} } } } }%}&øïpÆ} -------------------------------------------------------------------------------- A: You are not drunk, pppd cannot communicate with eciadsl-pppoeci (part of the driver) and it must come from a bad pppd version, from a bad ppp configuration inside the kernel, or because you're using the driver on a non supported system. Also check your /etc/ppp/peers/adsl file. The eciadsl-pppoeci invocation might be corrupted also, or the file might not exist at all (this is a known bug). 4.11 ================================================================================ Q: I get kernel panic/oops. -------------------------------------------------------------------------------- A: This may come from a kernel module bug (OHCI?), badly supported hardware, heavily loaded CPU or USB bus, or any kind of global system problem (even a bug in the driver of course). 4.12 ================================================================================ Q: I get this message in /var/log/messages: kernel: usb-uhci.c: ENXIO 80000xxx, flags 0, urb c5205d60, burb c2ad1120 -------------------------------------------------------------------------------- A: Try to disable all other USB devices to get sure that this problem is related to the use of the modem. 4.13 ================================================================================ Q: eciadsl-doctor says: Modem hangup Connection terminated. ... usb_control/bulk_msg: timeout -------------------------------------------------------------------------------- A: Too many possible reasons! Please contact us. See 1.5 too. 4.14 ================================================================================ Q: My transfert rate is very slow (for instance 2-5KB/sec instead of the 40KB/sec expected) -------------------------------------------------------------------------------- A: Try with another synch .bin. Recently, some users from Belgium have had such problems, since their provider now deliver 3.3Mbit/sec bandwidth. You can also play with the MTU value. 4.15 ================================================================================ Q: I use ISP DHCP, `eciadsl-start` goes OK but I still can't access internet. -------------------------------------------------------------------------------- A: Check your routing table: typing `route -n` as root. If no UG line is found (no default route), it might be a DHCP client issue on your side. Most common DHCP clients are pump, dhclient and dhcpcd. Check if your DHCP client is installed, else install another or a more recent version Note: this problem has been reported by users in Finland, Sweden or some countries in Asia. 4.16 ================================================================================ Q: eciadsl-start says: eciadsl-synch: failed to create shared semaphore: No space left on device -------------------------------------------------------------------------------- A: That's a bug introduced in version 0.6. It should be fixed in version higher than 0.7. Please upgrade with cvs or any version > 0.7. 5.0 ================================================================================ Q: Is it possible to launch eciadsl-start at boot time? -------------------------------------------------------------------------------- A: Yes it is, using init.d for instance, or /etc/ppp/ppp_on_boot for Debian users. Making it possible using init.d: Pre-requisites: - ECIADSL driver installed and configured - the ECIADSL sources package, available at: http://eciadsl.flashtux.org/download.php?lang=en Copy the rc.adsl file to /etc/init.d: > cp rc.adsl /etc/init.d > chmod +x /etc/init.d/rc.adsl (the rc.adsl file is provided with version >= 0.7 (or CVS)) For Debian, just type as root: update-rc.d rc.adsl defaults 15 For other distributions, see below : In /etc/rc.d, each rc*.d directory corresponds to an init's runlevel. For instance, runlevel 5 is the one that leads to a graphical login (this is default on most Linux systems, but sometimes it's 3), so /etc/rc.d/rc5.d belongs to that level. You will find more information about init levels: > man inittab or look at your /etc/inittab file. To know which default init level is reached at boot, find a line like the following in /etc/inittab file: id:5:initdefault: Here, default level is 5. For the operations below, let's say we use init level 5 at boot. You have to know that when your system enters an init level, it calls all the K* files from the last level, then call all the S* files from the new level. All these K* and S* files are in /etc/rc.d/rc.d they belong to. In fact, these S* and K* files are symlink, and point to scripts located (most of the times) in /etc/init.d. Eash S* file is like this: Sxxyyyyy, where xx is a 2-digit number, and yyyyy a service name (the name is human friendly, but you can set it as anything you want). All Sxxyyyyy files are called following the xx number (increasingly), then by the service name (yyyyy, alphabetical increasing order). For instance: > ls /etc/rc.d/rc5.d shows S01foo S15bar S99dummy K10makemyday K80whatthefuck (take care, it's only an example). So when init level 5 in entered, it will call S01foo, S15bar then S99dummy. When init level 5 is left, this is the same for K* files. But following the system, they are called in the same order (by xx increasingly), or in a reverse order (by xx decreasingly). Now you know how all this stuff works, you can make the relevant links to /etc/init.d in /etc/rc.d/rc5.d: > ln -s /etc/init.d/rc.adsl /etc/rc.d/rc5.d/S90adsl > ln -s /etc/init.d/rc.adsl /etc/rc.d/rc5.d/K90adsl Before you create these links, ensure that the S90 is relevant, this means that S90* will be called after all other network services. Of course, verify that K90 is relevant too. According to some systems (reverse kill order) this would be better to use K00, but simply think of the best order to kill the connection. Once all the symlinks are done, you can test it by rebooting your system. You can also test it without rebooting: close your X11 session, open a root session in console mode, then: > init 3 then > init 5 Understood what it does? It means leaving init level 5 to init level 3, then coming back to init level 5. The relevant S* and K* links will be called! 5.1 ================================================================================ Q: I cannot connect in console mode whereas it works fine under X11. -------------------------------------------------------------------------------- A: This may be a framebuffer issue. Try to boot your Linux without framebuffer console support. 5.2 ================================================================================ Q: I cannot connect under X11 whereas it works fine in console mode. -------------------------------------------------------------------------------- A: See 5.1. 5.3 ================================================================================ Q: How to properly configure kernel from the sources to get USB/PPP/N_HDLC support? -------------------------------------------------------------------------------- A: Type the following commands: > cd /usr/src/linux > make -s menuconfig --- General setup [*] System V IPC .. USB support ---> Support for USB [ ] USB verbose debug messages --- Miscellaneous USB options [*] Preliminary USB device filesystem [ ] Enforce USB bandwidth allocation (EXPERIMENTAL) [ ] Long timeout for slow-responding devices (some MGE Ellipse UPSes) --- USB Host Controller Drivers < > EHCI HCD (USB 2.0) support (EXPERIMENTAL) UHCI (Intel PIIX4, VIA, ...) support UHCI Alternate Driver (JE) support OHCI (Compaq, iMacs, OPTi, SiS, ALi, ...) support .. --- USB Multimedia devices .. < > DABUSB driver .. Character devices ---> .. [*] Non-standard serial port support HDLC line discipline support .. Network device support ---> .. PPP (point-to-point protocol) support [ ] PPP multilink support (EXPERIMENTAL) [ ] PPP filtering PPP support for async serial ports PPP support for sync tty ports PPP Deflate compression PPP BSD-Compress compression < > PPP over Ethernet (EXPERIMENTAL) < > PPP over ATM (EXPERIMENTAL) .. Following the kind of PPP encapulsation your provider uses, you may need extra settings in your kernel configuration. PPPoE or IPoATM users, please refer to 5.3.1. Once the kernel configuration file stored, type: > make -s dep modules modules_install && depmod -a or whatever command you use to compile and install the kernal modules. Of course your whole kernel config must be properly set, and the loadable module support turned ON. Also `make -s bzImage` if PPP or USB support was previously in kernel instead of as modules! You can enable all USB Host Controller Drivers but must use the relevant one only! DABUSB must be disabled. Upgrade your modutils package. Then you must see these available modules when you type `modprobe -l`: usbcore and usb-uhci or ush-ohci or uhci and ppp_generic ppp_async ppp_synctty bsd_comp ppp_deflate and also n_hdlc Add these following lines in your /etc/modules.conf if they are missing: alias char-major-108 ppp_generic alias /dev/ppp ppp_generic alias tty-ldisc-3 ppp_async alias tty-ldisc-13 n_hdlc alias tty-ldisc-14 ppp_synctty alias ppp-compress-21 bsd_comp alias ppp-compress-24 ppp_deflate alias ppp-compress-26 ppp_deflate alias char-major-180 usbcore and alias usb-hostadapter usb-uhci or alias usb-hostadapter usb-ohci or alias usb-hostadapter uhci then > touch /etc/modules.conf /lib/modules//modules.dep 5.3.1 ================================================================================ Q: How to properly configure kernel from the sources if I use PPPoE, Bridged Ethernet (RFC1483B) or IPoATM (RFC1483R)? -------------------------------------------------------------------------------- A: The kernel configuration and compilation principles are still the same (see 5.3), but you need this: Network device support ---> .. [*] Network device support .. Universal TUN/TAP device driver support .. PPP (point-to-point protocol) support .. PPP over ATM (EXPERIMENTAL) then type: > mkdir /dev/net > mknod /dev/net/tun0 c 10 200 > ln -s /dev/net/tun0 /dev/net/tun PPPoE users: read the INSTALL file, there's some software you need to install and configure. 5.4 ================================================================================ Q: I'm not from France, will the driver work with my modem and provider? -------------------------------------------------------------------------------- A: Verify that your modem is supported. If your modem isn't in the list of the supported modem and is NOT in the list of UN-supported modem, ask us for further help. Check which PPP protocol/encapsulation is used under MS Windows by your provider. Also check if DHCP is used by your provider or if you use a static IP. Use eciadsl-config-tk or eciadsl-config-text to properly use it. If it's not supported by the driver or if you are uncertain, get the latest version of the driver or contact us. FYI: RFC1483 VC-MUX (or null) PPPoA == RFC2364 VC-MUX aka RFC2364 PPPoATM NULL Encapsulation RFC1483 LLC PPPoA == RFC2364 LLC Routed RFC1483 BRIDGED ETH with NO FCS (ethernet over ADSL, ETHoA or ETHoATM) is a mode typically used to encapsulate ethernet frames, and is often used in conjunction to PPPoE, ethernet static IP, or DHCP over ADSL. This mode is very flexible, either your provider gives you a user/password to connect with PPPoE, or a static IP and gateway, or you use a DHCP client to connect. VCM RFC1483 BRIDGED ETH is a similar mode to RFC1483 BRIDGED ETH with NO FCS but not so widely used. LLC RFC1483 ROUTED IP is used for direct transmition of IP datagrams over ADSL (IPoATM or IPoA) typically in the case of static IP, this protocol requires a static IP and a gateway given by your provider. VCM RFC1483 ROUTED IP is a similar mode to LLC RFC1483 ROUTED IP but not so widely used. SNAP is an encapsulation layer more, between ethernet and LLC or between IP and LLC. For more information, please refer to the control application of your modem under MS Windows, it should list the mode used, or to your provider. You can also read RFC1483 and RFC2364 which describe all theses modes in detail. 5.6 ================================================================================ Q: Can I use my USB modem with a USB2.0 HUB? -------------------------------------------------------------------------------- A: We encountered problems when the modem or another device is plugged to a USB 2.0 HUB, which is currently not supported. In most cases, USB2.0 HUB module has to be unloaded. That's why eciadsl-start now unload ehci-hcd module unconditionally when it founds it loaded. If the USB2.0 ports are managed by usb-uhci or usb-ohci, this can work, we have experienced such working configurations. 5.7 ================================================================================ Q: I have several hardware USB HUBs or several USB devices. Is this a problem? -------------------------------------------------------------------------------- A: Might be. If you don't succeed in initializing your modem, try to disable all additional hardware USB HUBs. We only have partial experience about this. Many users have webcams, mice or other USB devices added to their USB modem, both working at the same time. This doesn't exclude problems with some USB devices anyway. Sometimes the order of the devices in the USB ports may be a problem too (to be verified). Also try your modem with all other USB devices unplugged. If you have USB 1.1 and 2.0 HUBs together in your machine, see 5.6. 5.8 ================================================================================ Q: Switching from Linux to MS Windows, I cannot use my modem and have to re-install the MS Windows driver to make it working again. -------------------------------------------------------------------------------- A: Unload all USB modules related to you modem when leaving Linux. This could be done automatically, through the USB manager of your Linux distribution maybe, or using init.d, (think of cascading modules unloading using the post-remove or pre-remove commands in /etc/modules.conf file). If it still doesn't work, you have to manually unplug you modem, wait for a few seconds so that it's resetted, and replug it. Then, you can use it under MS Windows. If the problem still persist, contact us. 5.8.1 ================================================================================ Q: Switching from MS Windows to Linux, eciadsl-start says that the firmware is already loaded and fails. -------------------------------------------------------------------------------- A: You have to manually unplug you modem, wait for a few seconds so that it's resetted, and replug it. Then you can retry eciadsl-start. 5.9 ================================================================================ Q: Can I use this driver under *BSD? -------------------------------------------------------------------------------- A: A BSD port of the driver is under development. Still not officially supported! 5.10 ================================================================================ Q: Can I use this driver under GNU/Hurd, Darwin, QNX or BeOS or other systems? -------------------------------------------------------------------------------- A: The driver is known to be running on many GNU/Linux systems, but also on *BSD (work under progress) other non-Linux kernel based systems are not supported. 5.11 ================================================================================ Q: How can I stop the PPP connection? -------------------------------------------------------------------------------- A: `eciadsl-stop` If you really want to reset your modem, unload the module which manages your USB HUB (`modprobe -r usb-uhci` for instance). But this may lead to problems because some other USB devices may belong to this module (!), and you may encounter problems re-starting eciadsl-start (need to run it twice or maybe to unplug/replug your modem then launch `eciadsl-start`). You can also automatically stop the running PPP connection when rebooting or shutting down your system. The proper way to configure this is describbed in Q 5.0. 5.12 ================================================================================ Q: Is it possible to run eciadsl-start as any user? -------------------------------------------------------------------------------- A: Simply ensure that sudo is installed on your system. Edit your /etc/sudoers file as root and add a line: username ALL=NOPASSWD:/usr/local/bin/eciadsl-start where username is the name of the user you want to allow to run eciadsl-start. You can do this for many users or different scripts, adding the same line with a relevant username and script path/name. Granted users can now run eciadsl-start, typing: > sudo eciadsl-start See 5.13 now. 5.13 ================================================================================ Q: When I run eciadsl-start or sudo eciadsl-start, I get this error: nice: pppd: no such file or directory -------------------------------------------------------------------------------- A: Check if pppd is installed on your system and try "su -" to get root privileges (instead of "su"). For sudo: Otherwise, in the console or terminal as root, type: > PATH="/sbin:/usr/sbin:/usr/local/sbin:$PATH" sudo eciadsl-start If this works, add the following line to your normal user's ~/.bashrc or ~/.profile (*): > export PATH="/sbin:/usr/sbin:/usr/local/sbin:$PATH" Next time you will open a terminal (if you've modified ~/.bashrc) or next time you login in console or graphical environment (if you've modified ~/.profile), running sudo eciadsl-start should work. You can also create a small script containing the first command, then you will only have to run this script. Don't forget to chmod 777 this script so that it becomes executable. (*) caution: following the Linux distro you use, these files may not exist or some other ones be used instead. 5.14 ================================================================================ Q: My provider use PPPoE (RFC1483, RFC2516), how can I manage to configure this? -------------------------------------------------------------------------------- A: First, don't use kernel's PPPoE support. It's experimental in 2.4 kernels and used by default in Debian kernel. Use rp-pppoe instead as a userland tool to manage the PPP connection. See 5.3.1 and 5.4. Of course, the PPP mode defined in the ECIADSL config must be one of the supported PPPoE modes (see `eciadsl-pppoeci --list`). You can find rp-pppoe here: http://www.roaringpenguin.com/pppoe Install and configure it so that it uses tap0 as ETH interface. Run eciadsl-start, then use rp-pppoe to enable the PPP line (see rp-pppoe documentation). 5.15 ================================================================================ Q: My provider use a PPP mode I cannot find in the PPP modes list. What can I do? -------------------------------------------------------------------------------- A: Please contact us. 5.16 ================================================================================ Q: I don't know which PPP mode is used by my provider. What can I do? -------------------------------------------------------------------------------- A: See 5.15. 5.17 ================================================================================ Q: When EciAdsl is running, some apps like KDE run very slowly. What can I do? -------------------------------------------------------------------------------- A: Look at ifconfig command output if "lo" (loopback) interface is here. Add it if necessary (for example: ifconfig lo up).