NS3000/iX Error Messages Reference Manual (36923-90043)

420 Chapter18
PCI LAPBMUX Link Error Messages
SDI Driver Specific Status Values
2440 CLAS0001 MESSAGE: Cannot send a port message.
CAUSE: While shutting down the link, the driver attempted to free ADA
memory previously obtained from the MPE I/O system, but
encountered an error trying to send a memory request message to the
I/O Memory Port (“Cause” = 32-bit status from the call to send_msg).
ACTION: This is a warning that some memory resources may have been
lost. The driver will attempt to continue. Unless this happens
frequently, the system should continue to run.
If the problem happens every time, enable link tracing via
:LINKCONTROL then reproduce the link shutdown problem (the trace file
will close automatically). Save the resulting trace file
(NMTC####.PUB.SYS) on tape for Hewlett-Packard analysis, and see
Appendix A , “Submitting a CR.
2480 CLAS0001 MESSAGE: Cannot send a port message.
CAUSE: While starting up the link, the driver attempted to request ADA
memory from the MPE I/O system, but encountered an error trying to
send a memory request message to the I/O Memory Port
(“Cause” = 32-bit status from the call to send_msg).
ACTION: This is a fatal software error. The driver cannot operate the
adapter card without ADA memory. After reporting this error, the
driver will return a configuration error, then die and wait to be shut
down; the configuration routine should shut it down automatically.
The problem can be caused either by lack of memory or by a bug. If you
suspect high memory usage and have the GLANCE tool, run it and
check system memory usage. If usage is high, try closing some
applications or subsystems, then retry link startup. If the problem
persists, shut down the system to clear memory, then reboot but try
starting the link earlier, before too many other applications get going.
If still low on memory, you may need to add more memory hardware.
First, however, Hewlett-Packard can try to determine if more memory
would help, or if instead there is a bug. Use NMMGR to make sure all
logging classes are enabled for the link type, for disk logging. Reproduce
the problem, then collect all NM log files (NMLG####.PUB.SYS) and see
Appendix A , “Submitting a CR.
2560 CLAS0001 MESSAGE: Bad message for current state.
CAUSE: The driver received a new bind (rendezvous) from an upper
layer protocol, at a time when the driver was not able to process these
requests, such as before receiving its initial configuration data, during
shutdown, or after a fatal software error (“Cause” = 16-bit encoded
value, giving driver input event code and current state: HP use only).
ACTION: This problem most likely occurred while starting a new
network subsystem. Attempt to determine which subsystem was being
started when the error occurred and, if possible, which upper layer