http://www.murga-linux.com/puppy/viewto ... 621#219621
I am appending a dot pet with the fixes as well as the corresponding difference files.
The following changes were made:
1. Modemprobe changed to utilize the current /dev/modem target device if the probe for it is successful. Previously, the device found via /dev/modem was ignored, making the probe useless. Also added return of wvdialconf result to pupdial (so code could be added there to correct its report to the user -- when not modem found).
2. To prevent the probe from attributing a successful probe to the wrong device (an old problem), particularly after a modem changeout, most of the PCI-modem scripts delete their device links/nodes if their driver is not loaded, leaving the device link/node for only the active/resident modem(s). In these cases, /dev/modem is also deleted, since it would point to a non-existent link/node. (This fix trusts that Puppy will create only the link/node required by a loaded PCI-modem driver module, as done in alpha5.)
3. Added ttyS... links for the Intel modems, so that modemprobe can find them. But ensured that the actual device name is retained.
4. Restored prioritization links for SmartLink modems, so that USB type checked first, then PCI, then ALSA. The ttySL0 link is not managed, since it does not conflict with the other PCI modems.
The attached dotpet is suitable for testing, since the modem scripts are in both the operational directory and the firmware tarballs. For incorporation into the next alpha, only /usr/sbin/modemprobe and the /lib/module/all-firmware tarballs need be copied into the "master."
BTW, since the "martian" Lucent driver does not work in alpha5, I compiled and ran it, and tested its script successfully in Puppy 4.0. I hope that the loading problems with martian_dev and slamr are related to the SMP kernel, and that they will work in the promised uniprocessor kernel. The slamr problem can be worked around, but the martian problem is a showstopper.
Richard
Unfortunately, I am unable to attach my dotpet file containing the corresponding all-firmware tarballs, receiving the following forum response:
However, apparently the diff listings and the scripts themselves were accepted.Sorry, but the maximum filesize for all Attachments is reached. Please contact the Board Administrator if you have questions.