>>>>> "David" == David Edelsohn <dje@watson.ibm.com> writes:
    David> 	The driver also has a potential problem that the descriptors are
    David> defined in a shadow structure and then copied to the real descriptors seen
    David> by the chip, but that only is correct if the copy is atomic -- which it is
    David> not.  The status bits are in the middle of the descriptor structure, so
    David> the "ownership" bit could flip without the descriptor in a completely
    David> self-consitent start.  If some asynchronous event caused the chip to start
    David> before the explicit CSR kick, things could go wrong.

  That's no a problem on 79C792: just put them in SWSTYLE #3, which is
optomized for this kind of copy. The copy itself is unnecessary on PCI anyway.

]         At IETF46 in Washington, DC  --- wireless             |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] mcr@sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another NetBSD/notebook using, kernel hacking, security guy");  [