Please read this file completely before posting any questions to the e-mail support! Installation ------------ Copy the driver for your processor to devs:networks/x-surf.device. Don't forget to remove the three digits in the name, the device will not work if the filename has a "000" or a "040" at the end! Make a directory in envarc: named "sana2". This can be done with your favourite directory tool, or by typing in a shell: makedir envarc:sana2 Copy the file x-surf.config to that directory, then restart your computer, so this directory and file is copied automatically to ENV:. If you want to avoid a restart, you can also copy the directory to ENV: manually. Remember that the old device can only be removed from memory if the OpenCount is 0, and you enter "avail flush" in a shell. If this is greek to you, a simple restart will serve your needs. changes from V1.1 ----------------- 1. The new driver now supports more than one X-Surf. Up to eight boards are supported. Use the same device, and the next unit number to access the next board. The order of the unit numbers is taken from the order in the config chain, i.e. the X-Surf in the first slot is Unit 0. 2. The full duplex mode is now supported (20MBit). In order to activate it, the config file "env:sana2/x-surf.config" must contain the keyword "FullDuplex" followed by "1" for the unit you want to activate full duplex for. Example: --- cut --- # **** x-surf config file ***** # Unit0.FullDuplex 0 Unit1.FullDuplex 1 --- cut --- Board number 0 will work with 10Mbit, while board number 1 uses 20MBit. Comments in the file are marked with a # sign. If no configuration file is found, half-duplex mode will be activated. Changes are only stored in the board's EEPROM (non-volatile memory) if the config-file and the contents of the EEPROM are different. 3. Timer.device is now being used for all timing-critical routines. These are: - ISA-PNP - r/w to/from EEPROM This avoids the hickups of the mouse pointer during first open of the device. 4. The device now uses soft-interrupts instead of a task to serve the hardware-IRQs. 5. The size of the driver has been reduced by some optimizing of the code. This has also improved the speed, the performance of this driver is now even better. 6. The Promisc-mode of the driver has been corrected (reception of all packets of the network). Tests ----- The driver is working properly on all test-computers since a couple of days. Some hubs do not display the full-duplex mode, but it can be used (proof: performance stays stable when adding another data stream in the opposite direction). Especially Kingston-Hubs do not seem to display the full duplex mode (LED stays off). We still don't know where the problem with Mac-emulators is located. Please report any changes with this new driver, because we believe that the corrected promisc-mode could be a solution. None of the beta-testers has a Mac-emulator, so we need your help in order to test this. Versions -------- There are versions for 68000, 68020 and 68040 processors. Many customers have asked for specialized versions for 68030 and 68060 processors, and I want to answer these questions here: The different versions are only different regarding the addressing modes of the specific processor. Since 68020 and 68030 are identical in the used addressing modes, it would not make any sense to re-compile for 68030. The same applies to 68060 processors: A re-compilation would result in an identical binary file with a different file name. I hope this answers all questions in this regard, and the e-mail support can answer more important questions. If you have any comments, bugreports or something similar, for example just want to tell us that the driver works as you expected, feel free to write: jens@siliconsonic.de