Re:[SLE] wvdial, USB modems and Suse 7.3
I think you're right. I've just installed a 2.4.2 kernel in 7.3 and my usb modem now works. JDL
There seems to be something wrong with the USB in 7.3. I have two USB devices: the modem and an Epson Perfection scanner.
The scanner works with no problems. On the surface all seems well with the modem: it is recognised by Linux and listed in the /proc/bus/usb/devices file, the acm module and the other USB modules are loaded. However, any attempt to read or write to the /dev/ttyACM0 modem device gives an "Invalid argument" error.
I checked the /proc/bus/usb/devices file again. The acm driver is not associated with the modem. OTOH, the scanner driver is associated with the scanner. So, the USB subsystem is successfully loading the acm driver but is not associating it with the modem device. No idea why, and no idea how to fix it :(
I think you're right. I've just installed a 2.4.2 kernel in 7.3 and my usb modem now works.
I finally fixed it by compiling kernel 2.4.13 and ensuring that the usbcore and usb-uhci modules are loaded during boot up in the /etc/rc.config file. I'm also using the latest smppd package. My modem now works :) If Suse continues to rush releases out, it won't do their reputation any good, nor Linux's. Peter
On Sun, Oct 28, 2001 at 08:49:58AM -0000, Peter John Cameron wrote:
I think you're right. I've just installed a 2.4.2 kernel in 7.3 and my usb modem now works.
I finally fixed it by compiling kernel 2.4.13 and ensuring that the usbcore and usb-uhci modules are loaded during boot up in the /etc/rc.config file. I'm also using the latest smppd package. My modem now works :) If Suse continues to rush releases out, it won't do their reputation any good, nor Linux's.
Peter
You make a good point here, but keep two issues seperate in your mind. Firstly a SuSE distribution is of course, a linux kernel modified by SuSE, plus the SuSE specials (yast etc) plus the usual GPL and other open-sourced/free for private use stuff. I have complained before that SuSE bring out new distributions far too quickly, the period between 7.2 and 7.3 was way too short for the kind of testing a distribution needs, however many people you throw at it. I suspect this is bean-counter pressure at work. Secondly there is the Linux kernel, which evolves at a rate of knots, in less than a year 2.4 has gone up to 13 releases, with more on the way before 2.4.6 is started I would guess, I don't think they have even started 2.4.5 yet (the odd-numbered releases are not really for normal human consumption, they are for the developers, so they never get let loose officially). And then there are the Alan Cox releases which contain various patches that are not in the main release. Well, you can read the changelogs for these releases and see that they contain enhancements or fixes, and usually very quickly Mr Mantel's "next" directory contains the Suse modded versions of these. I am very uninterested in new Suse releases, but generally track the kernel releases pretty quickly, but from a reputation point of view it looks pretty bad, and (I always get flamed for quoting this) as someone said "Linux appears to be in a permanent state of Beta test". Then there are people making patches for the kernel not yet accepted by LT. With a Suse kernel it is very difficult, or even hazhardous to try and use these patches since the patch files are usually for the stock kernels, thus the diff's make no sense with a Suse kernel. So Linux is a complex world, if you want to contrast a philosophy you might look at the completely different approach of Free BSD, or even in the Linux world of Debian who the last time I looked are still not "officially" using the 2.4 kernel series. And the 2.2 series still soldiers on with new versions from time to time. That's the way it is. Linux was the idea of one person, and as far as I know he is still in sole charge of when releases of the kernel occur and what is and is not in them. -- Regards Cliff
Cliff, I would have to agree also about Linux. I am pleased SuSE stays so busy as to keep us new and updated releases coming. As a newbie, I have thought at times, the one thing I love about Linux is that the OS & software are continually updated, fixed & made better. On the other hand, the one thing I hate about Linux, is that the OS & software are continually updated, fixed & made better! Such is the world of Open Source though and I think we have to accept that as users. The one thing that still seperates us from the apes though is our ability to reason, think things out. It is our decision how and when to incorporate or add those updates, fixes, etc. and not complain about them, just seek & inquire answers from all the knowledgable people on this list and in the Linux community! ;-) Regards, Lee -------------------------------------
On Sun, Oct 28, 2001 at 08:49:58AM -0000, Peter John Cameron wrote:
I finally fixed it by compiling kernel 2.4.13 and ensuring that the usbcore and usb-uhci modules are loaded during boot up in the /etc/rc.config file. I'm also using the latest smppd package. My modem now works :) If Suse continues to rush releases out, it won't do their reputation any good, nor Linux's.
Peter-------------------------
*>On Tuesday 30 October 2001 05:35 am, Cliff Sarginson, went on about:
Firstly a SuSE distribution is of course, a linux kernel modified by SuSE, plus the SuSE specials (yast etc) plus the usual GPL and other open-sourced/free for private use stuff. I have complained before that SuSE bring out new distributions far too quickly, the period between 7.2 and 7.3 was way too short for the kind of testing a distribution needs,
Secondly there is the Linux kernel, which evolves at a rate of knots, in less than a year 2.4 has gone up to 13 releases, with more on the way before 2.4.6 is started I would guess, I don't think they have even started 2.4.5 yet
I am very uninterested in new Suse releases, but generally track the kernel releases pretty quickly, but from a reputation point of view it looks pretty bad, and (I always get flamed for quoting this) as someone said "Linux appears to be in a permanent state of Beta test".
So Linux is a complex world, if you want to contrast a philosophy you might look at the completely different approach of Free BSD, or even in the Linux world of Debian who the last time I looked are still not "officially" using the 2.4 kernel series. ********************************* -- ---KMail 1.3.1--- SuSE Linux v7.2--- Registered Linux User #225206 /tracerb@sprintmail.com/ *Magic Page Products* *Team Amiga* http://home.sprintmail.com/~tracerb
On Tuesday 30 October 2001 04:35, Cliff Sarginson wrote:
On Sun, Oct 28, 2001 at 08:49:58AM -0000, Peter John Cameron wrote: [snip]
And then there are the Alan Cox releases which contain various patches that are not in the main release.
[snip] How does one distinquish between the kernel released by Linus and the kernel released by Alan. Surely, Alan's kernel qualifies as a 'fork' by any definition of 'fork' that I've read. ??? JLK
On Tue, Oct 30, 2001 at 10:45:17AM -0600, jk05308 wrote:
On Tuesday 30 October 2001 04:35, Cliff Sarginson wrote:
On Sun, Oct 28, 2001 at 08:49:58AM -0000, Peter John Cameron wrote: [snip]
And then there are the Alan Cox releases which contain various patches that are not in the main release.
[snip]
How does one distinquish between the kernel released by Linus and the kernel released by Alan. Surely, Alan's kernel qualifies as a 'fork' by any definition of 'fork' that I've read. ??? JLK
They have the letters "ac" appended to them. -- Regards Cliff
* jk05308
How does one distinquish between the kernel released by Linus and the kernel released by Alan. Surely, Alan's kernel qualifies as a 'fork' by any definition of 'fork' that I've read. ???
The Alan Cox kernels are just a patch against the Linux Kernel, used for testing etc. It does not really qualify as a fork, and when using one of his kernels, the uname -a will report the kernel version with an -acN tag (N denoting what patchlevel it is). -- Mads Martin Joergensen, http://mmj.dk "Why make things difficult, when it is possible to make them cryptic and totally illogic, with just a little bit more effort." -- A. P. J.
I found it a little harder, but have a solution for the Elsa Microlink 56k USB modem: 1. Get 2.4.13 kernel from mirror of suse.com/pub/mantel/next 2. Overwrite acm.c (version 0.21) in 2.4.13/drivers/usb with 0.18 version from 2.4.2 kernel source 3. Copy .config from 2.4.10 source directory to 2.4.13 4. make mrproper; make xconfig (with .config); make dep; make clean; make bzImage; make modules; make modules_install 5. copy bzImage to /boot and rename vmlinuz-2.4.13 6. copy System.map from source to /boot/System.map-2.4.13 7. edit /etc/lilo.conf; mk_initrd -k "vmlinuz-2.4.13" -i "initrd-2.4.13" -s auto; lilo -b /dev/hda (or /dev/fd0 for test) 8. reboot Actually the 2.4.13 kernel works a little better than the 2.4.10 kernel. It connects to my isp, then spews out some rubbish and dies. I don't think we can blame SuSE for this one. It's a kernel problem. And, in fairness, the kernel developers can't be expected to have every USB modem to test. At least with SuSE you have P2P lists, source code to play with, reasonable support and a reasonable chance of getting a solution if you flag it. Thanks for the comments. I wouldn't have thought of the solution without them. JDL :-) Peter John Cameron wrote:
I think you're right. I've just installed a 2.4.2 kernel in 7.3 and my usb modem now works.
I finally fixed it by compiling kernel 2.4.13 and ensuring that the usbcore and usb-uhci modules are loaded during boot up in the /etc/rc.config file. I'm also using the latest smppd package. My modem now works :) If Suse continues to rush releases out, it won't do their reputation any good, nor Linux's.
Peter
participants (6)
-
Cliff Sarginson -
jk05308 -
John Lamb -
Lee O'Malley -
Mads Martin Joergensen -
Peter John Cameron