On Sun, 10 Aug 2003 16:58:49 +1200 Rik Tindall <[EMAIL PROTECTED]> wrote:
> Hi Chris, & thanks for your detailed script, > > Christopher Sawtell wrote: > > >You can use the -f option to force overwriting a symbolic link > >ln -sf /dev/ttyS15 /dev/modem > > > Will give that a try when all else fails - seems ok (char files, not l > as before = correct?). > Is there s'posed to be only one node pointing to /dev/modem then? Rik, start again. nothing points to /dev/modem, /dev/modem points to the device. Your log file shows the device is /dev/ttyS15, so /dev/modem must point to/dev/ttyS15. To acheive that simply go ln -sf /dev/ttyS15 /dev/modem (target, ie real file, first, symbolic name second) then to check do ls -l /dev/modem, and you should get something like this: lr-xr-xr-x 1 root root 7 Aug 10 17:17 /dev/modem -> /dev/ttyS15 (thats all one line) now if you point kppp at /dev/modem it should work. if it doesn't it may be because the driver isn't working (I just tried to get one of those useless things going the other day and gave up after a while, may try again), or because permissions are wrong on /dev/ttyS15. Please note that permission on /dev/modem are irrelevant. permissions on a link are inherited from the target file (ie /dev/ttyS15) and what ls tells you about /dev/modem's ownership & permissions is misleading. in other words if you want to fiddle with the permissions do it to the target (/dev/ttyS15). I see you are using version 0.9.5 of the driver. I believe there is a version 0.9.6 about somewhere on the net, I found some reference to it when I was trying to get the damn thing going the other day. I nearly wrote "a version" as "aversion" which is what linux (and windows for that matter) users should have to these crap motherboards IMHO! > Is it written dynamically by kppp? I just selected /dev/ttyS0 in kppp > to get the external going - no other amendments. > Do I have to eliminate it from the system completely to work with > /dev/ttyS15? > > >>no deletion done, but: > >> > >>[EMAIL PROTECTED] dev]# ls mode* -la > >>lrwxrwxrwx 1 root root 11 Aug 9 20:45 modem -> > >>/dev/ttyS15 > >> > >> > >Are you sure that that is the correct device node? > >Have you got a multi-port serial card in there? > >If so you will need extra options loaded into the kernel. > > > > > Prev.post of /var/log/messages incl: > > Aug 9 20:50:58 localhost kernel: AC97 modem device found: devnum = > 80000E00, devid = 1039/7013 > Aug 9 20:50:58 localhost kernel: iobase_0=0xd800, > iobase_1=0xd400,irq=5 Aug 9 20:50:58 localhost kernel: PCTel driver > version 0.9.5 [5.05c-4.27.215 (09-14-2001)] (MR) (2002-01-31) with > MANY_PORTS SHARE_IRQ SERIAL_PCI AUDIO_ROUTING[SIS540] enabled. > Aug 9 20:50:58 localhost kernel: PCTel driver built on [Linux > 2.4.20-8 i686 athlon "2.4.20-8 <132116>"] with gcc-gcc (GCC) 3.2.2 > 20030222 (Red Hat Linux 3.2.2-5). > Aug 9 20:50:58 localhost kernel: ttyS15 at 0xd800 (irq = 5) is a > PCTel > * * * * * > Aug 9 20:50:58 localhost kernel: PCI: Found IRQ 5 for device 00:01.6 > Aug 9 20:50:58 localhost kernel: PCI: Sharing IRQ 5 with 00:01.1 > > so yes, it seems. But no "multi-port serial card" - an out-of-the-box > PCChips SiS mobo only. > > >>oh.., so kppp knows its "/dev/modem" option = /dev/ttyS15 already > >then? -> > >> > >We don't know that until you list it:- > >ls -lL /dev/modem > > > crw-rw-rw- 2 root uucp 62, 79 Aug 10 10:44 /dev/modem > & #ls -lL /dev/ttyS15 = > crw-rw-rw- 2 root uucp 62, 79 Aug 10 10:44 /dev/ttyS15 > > so these nodes are linking to themselves, not to each other? c-hars > instead of l-inks? > (same results without the "L") > > >because the link pointing to /dev/ttyS15 may be incorrect. > > > >note the 'L' option flag > >man ls > >explains > > > >Part of your problems might be that you have too many levels of > >symbolic linkage for the application program to plough through. I > >always put the actual device node name for the physical peripheral > >device in the kppp config. > > > Will hunt that out next session. > > >>"Modem ready.. > >> The modem does not respond." > >> > >>when I try that one. > >> > >>(connecting via external on /dev/ttyS0 presently) > >> > >> > >Are you sure of that, I had a system to set up recently in which the > >mapping was the other way around, i.e. > >COM1 -> /dev/ttyS1 > >COM2 -> /dev/ttyS0 > > > >odd, very. I suspect the owner was muddled about the positions of the > >internal connectors to the motherboard. > > > > > That sounds consistent - why sould they numerically reverse? I have no > COM2: Aug 10 15:52:57 localhost pppd[6261]: Connect: ppp0 <--> > /dev/ttyS0 > > >minicom ( and cu in the Taylor uucp package ) are both exceedingly > >useful utilities to fiddle, test and play with serial lines and > >modems > > > >Make sure that the modem has its own serial port and, for best > >performance, its own interrupt. IRQ 5 is normally available, don't > >use IRQ 7, it's for the parallel port. > > > >cat /proc/interrupts > > > CPU0 > 0: 2036979 XT-PIC timer > 1: 11409 XT-PIC keyboard > 2: 0 XT-PIC cascade > 4: 225467 XT-PIC serial * * * * * > 8: 1 XT-PIC rtc > 11: 0 XT-PIC SiS 7018 PCI Audio > 12: 543713 XT-PIC PS/2 Mouse > 14: 58878 XT-PIC ide0 > 15: 92 XT-PIC ide1 > NMI: 0 > ERR: 0 > > > >cat /proc/ioports > > > 03f8-03ff : serial(auto) > 0cf8-0cff : PCI conf1 > a000-afff : PCI Bus #01 > ac00-ac7f : Silicon Integrated Systems [SiS] SiS630 GUI > Accelerator+3D > cc00-ccff : Silicon Integrated Systems [SiS] SiS900 10/100 Ethernet > d000-d0ff : Silicon Integrated Systems [SiS] SiS PCI Audio Accelerator > d000-d0ff : SiS 7018 PCI Audio > d400-d47f : Silicon Integrated Systems [SiS] 56k Winmodem (Smart Link > HAMR5600 compatible) d800-d8ff : Silicon Integrated Systems [SiS] 56k > Winmodem (Smart Link HAMR5600 compatible) > > etc. > > > > >will tell you what's what. > > > >you will probably need to set up the serial port using the setserial > >utility > > > >man setserial > >tells you the options. > >put the needed command in the rc.local file in the unix system > >startup files. > > > Can this actually be used on a /dev/ttyS15, or just on COM ports? > > >If the modem uses an isa socket and purports to be 'plug and play' > >you will almost certainly need to use the isapnp tool kit to set it > >up, and remember to put the appropriate setting into the bios. > > > N/A > > Thanks Chris, I'll work thu these ideas when 'project time' next rolls > round. > Meanwhile, commitment to Linux learning curve is acknowledged. > Thanks to all writers (& readers) for their time. > > Here's a simple Q to end for the weekend: > > Would all this PCTel "onboard modem setup" process be any simpler > under Gentoo? > > Cheers > > ~ rik > >
