Indeed -- I don't have such a card to test. (I didn't read your earlier message closely enough and missed the fact that this is merely a card similar to the axiom card)
I think I understand why the original configuration outb()s are written in that order: In the new order, the outputs are enabled for a short time before they are driven with the 0xff value. Perhaps a better sequence would be outb(0xff, board->base_addr+0); outb(0xff, board->base_addr+1); outb(0xff, board->base_addr+2); outb(board->port1config, board->base_addr+3); + outb(0xff, board->base_addr+0); + outb(0xff, board->base_addr+1); + outb(0xff, board->base_addr+2); so that 8255-clones which accept output values while configured as inputs do not drive "unknown" values after the switch, but boards which don't accept output values while in input mode are driven with the proper initial output value as early as possible. Does this order work on your card? As for setting the 0x80 bit of the configuration register, if it turns out the Access and Axiom boards can't be satisfied with a single setting then I suppose a module parameter could be used for that purpose. Jeff ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users