in cardmgr/probe.c:tcic_probe_at() the function first attempts to determine if any device is in the slot ... and loops checking first 8 words and aborts if ANY are 0xfffff. This is incorrect, since some un-initialized devices may have one or more ports returning 0xffff ... in particular Aironet PC4800 do, and are unusable in slots managed by tcic controllers (like Databook chips). The correct choice would be to check for ALL ports returning 0xffff: /* Anything there? /* for (i = 0; i < 0x10; i+=2) if (tcic_getw(base, i) != 0xffff) break; /* if all registers/ports are floating do an error return - JBass */ if(i == 0x10) return -1;
Current errata use the kernel pcmcia not the add on
This remains broken in both RH8 and RH8 using stock install and no add-on software. In particular Aironet PC4800 cards which worked in previous releases will no longer configure in the reader/adapter cards with DataBook chips, that worked fine in previous releases. I guess I need therefor to re-create the bug report in both RH8 and RH9?????