I am seeing a consistent machine lock-up in RTP when plugging a 5V PCMCIA
card into my laptop running a Win2K/Win98/QNX RTP multi-boot. The laptop is
a Umax ActionBook 850T but I’m pretty sure that the problem is specific to
devp-pccard running with an O2 Micro OZ6832 Cardbus controller.
Insertion and removal of low-voltage cards with RTP is fine, and the 5V
cards work fine under Win2K/Win98. However, as soon as a 5V card is plugged
in running RTP the machine locks (power-off job). If devp-pccard is killed
prior to inserting a 5V card the insertion causes no trouble - no surprise
there. The problem also exists if there is a 5V card inserted when RTP boots
(at the point where devp-pccard loads).
If devp-pccard is loaded just prior to inserting a 5V card using
‘devp-pccard -vvv’ the following manages to get out to the display console
before everything crashes:
Socket 0 returned SCIntMask 0 - Vcontrol 10 - VppLevels 0 - VccLevel 0 -
CtlInd 0 - State 80 - IREQRouting 0 - IFType 1
Card Inserted
5.0V Card
Has anyone seen similar problems with the O2 Micro or any other
PCMCIA/CardBus controller?
I am also having trouble getting a NetGear FA-510 PCMCIA card working with
RTP. This is a low-voltage card so doesn’t cause a lock-up. It is recognised
by nettrap as a 21143 based device (correct) and hence it uses the tulip
driver. However, it seems that interrupts are not being correctly routed
back (although it hooks the correct interrupt line according to the
information as devn-tulip.so loads) to the controller since outgoing packets
work OK but not incoming. Hence a ping of anything except the local host
fails.
These problems are holding up an important and tight-scheduled Neutrino
based development program.
Len Barber.