FreeBSD-2.2 ISA Wavelan driver, 970317 gamma release. ============================================================================== Michael Smith This is a port of Jim Binkley/Portland State University Mobile IP Project's ISA NCR Wavelan drivers to FreeBSD 2.2. At this point, the PCCARD Wavelans are not supported. There are two components to this distribution, the driver and the 'wlconfig' program. Driver ------ Driver sources are in the 'wavelan' directory. Use the supplied KernelDriver script to install these drivers into your kernel source tree : $ KernelDriver -v wavelan [] where is the path to the top of your kernel source tree if not '/sys'. Add a line to your kernel config that reads : device wl0 at isa? port 0x300 net irq ? vector wlintr Change the port value to suit the base address of your card. The IRQ value for the card can be set with the 'wlconfig' program. Other parameters can be set using sysctl, see the wl(4) manpage supplied for more details. wlconfig -------- This program can be used to change the PSA (Parameter Storage Area) values on a Wavelan card. Build wlconfig in-place in the distribution with 'make'; an installation destination is not specified, so it must be installed by hand. Usage is covered in the supplied manpage. The 'vwlconfig' script reqiures Tk installed, and provides a point-n-click GUI interface for configuring Wavelan adapters. PLEASE NOTE: ------------ Some things you should be aware of : - This release is a minor update to the 970305 release incorporating a fixed version of the KernelDriver installation script, and with a corrected directory layout. There are no changes to the driver itself. - My work on this driver was substantially funded by Hilink Internet, who also provided the cards for testing. My thanks to Daniel O'Callaghan for his faith and enthusiasm. - Bug reports are welcome. I may not necessarily have access to the hardware to test them though. - There is strangeness with this driver on my test systems related to output from a fast system being randomly lost, causing throughput hiccups. I haven't been able to track this down yet, but I haven't had it reported by anyone else, so it may be a local quirk.