"Maciej W. Rozycki" wrote: > > On Wed, 24 Apr 2002, Michael Pruznick wrote: > > > I'm working on this mips board with a smsc 90e66 south bridge and > > fdc37m812 super io. I'm using the standard pc_keyb.c driver. I only > > see keyboard interrupts and KBD_STAT_OBF set in response to "key up" > > events. I never see them in response to "key down" events. Thus, the > > shell running on the vga console never gets my input (since it is the > > "key down" events that pass the character typed to the shell). > > > > At this point, I'm thinking that the standard driver needs some mods > > to work with the super io's ps2 controller. The smsc doc only covers > > programming the plug and play registers and doesn't give any info about > > programming the ps2 controller. > > An 8042-compatible microcontroller (actually the firmware it runs) may > need to be programmed to a PC/AT-compatible mode. On an i386 it is > typically done by the BIOS. Try dumping configuration data from your chip > and compare it with what is set up in an i386 system. You can dump 32 > bytes of configuration data with the 0x20 command of the 8042 (5 low-order > bits of a command byte specify an address). Writing can be performed > using the 0x60 command (the same semantics). > > Some data is available in the Ralf Brown's interrupt list (look for > "inter60*.zip" files on a SimTel DOS collection's mirror). I have an old > Intel hardcopy document somewhere that describes to some extent the > IBM-defined locations of the configuration data -- I may try to dig it out > and see if I could help you. Anyway, you should probably contact the > chip's manufacturer. Thanks, that seams to be the issue or at least part of it. I dumped offset 0x20-0x3f on several systems. All gave different results. Some helped, some did not. In the case of the ones that helped, all the keys I tried (alpha,num,symbol) worked, until I pressed a shift, control, or alt key, in which case the keyboard was stuck sending the shifted value of all keys. I sent a message to the chip manufacturer, waiting for their response. In all cases, the mouse doesn't work and enabling the mouse via "gpm -t ps2 -m /dev/mouse" or "od -tx1 -w1 /dev/mouse" causes the keyboard to stop sending scancodes (on key up or key down). -- Michael Pruznick, michael_pruznick@mvista.com, www.mvista.com MontaVista Software, 1237 East Arques Ave, Sunnyvale, CA 94085 direct voice/fax:970-266-1108, main office:408-328-9200