This appendix lists error messages and their explanations and solutions. The error messages fall into two groups:
messages printed by IRIX command programs
messages printed during boot or found in SYSLOG
Listed here are messages associated with programs that use or operate the MCO. Typically, only setmon and SSE diagnostics cause these errors.
ERROR: couldn't get number of boards
Explanation: The operating system prevented the software from manipulating the RealityEngine graphics board.
Solution: Check software revisions.
ERROR: couldn't open /dev/graphics
Explanation: The operating system prevented the software from manipulating the RealityEngine graphics board.
Solution: Check software revisions.
setmon: can't open vof file, filename
Explanation: The specified MCO format file is corrupt or is from an older revision of system software. For the 5.0.1 software release only, a CD containing the MCO video output formats (VOFs) was provided. As of release 5.1, the VOFs are incorporated into the standard system software subsystem package eoe1.sw.unix. This error message appears if the older 5.0.1 VOFs are loaded on top of the 5.1 VOFs.
Solution: Identify and repair the file. If necessary, reinstall the eoe1.sw.unix software.
setmon: cannot detect MCO in appropriate address. Is software installed properly (EEPROM written)?
Explanation: Some hardware responded at the specified VME address, but the EEPROM at that address did not contain the MCO identification sequence.
Solution: Ensure that the MCO has its EEPROM correctly loaded by running the /usr/gfx/ucode/RE/vs2/vs2prom program (or /usr/gfx/ucode/REC/vs2/vs2prom or /usr/gfx/ucode/REV/vs2/vs2prom, depending on whether you are running RealityEngine, Crimson, or Onyx).
setmon: internal error in filename
Explanation: The specified MCO format file was corrupt.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
setmon: no MCO board is present -- rerun setmon without -S
Explanation: Nothing responded in the address specified in the DG2 EEPROM (where the address of the MCO board is specified).
Solution: Check the address of the MCO as set on MCO board DIP switches for correctness. Use the /usr/gfx/ucode/RE/vs2/vs2addrset program (or /usr/gfx/ucode/REC/vs2/vs2addrset or /usr/gfx/ucode/REV/vs2/vs2addrset, depending on whether you are running RealityEngine, Crimson, or Onyx) to set and validate the address of the MCO in the DG2 EEPROM.
setmon: unexpected EOF in filename
Explanation: The specified MCO format file was not readable.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
The Multi-Channel Option hardware is present, but the EEPROM seems to be outdated or corrupt. Run setmon -S to repair it.
Explanation: Software has detected one of the following problems with the Multi-Channel Option EEPROM:
the stored revision number does not match that which the software expects
the size of the stored data does not match that which the software expects
the checksum stored in the EEPROM is not internally consistent
Solution: Run setmon -S to rewrite the EEPROM.
VOF is not for VS2 (magic_number)
Explanation: The specified MCO format file did not contain the proper magic number. The magic number found in the file is shown.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
Listed here are messages that can be printed during boot or that appear in the file /usr/adm/SYSLOG. In the latter case, each message has the gfxinit: prefix.
Couldn't load DG2's special MCO VOF (filename): reason
Explanation: The special DG2 format file used when running MCO could not be loaded. The name of the file is printed, as is the reason the file could not be loaded.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
Could not open MCO VDRC file filename: reason
Explanation: The initialization software could not load a needed MCO file. The reason is specified.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
Error opening /dev/graphics in routine routine-name - reason
Explanation: The operating system prevented initialization from determining information about or initializing the graphics system.
Solution: Ensure that graphics have stopped before you attempt whatever caused this error. Check software revisions.
Failure in GFX_ATTACH_BOARD ioctl in routine-name - reason
Explanation: The operating system prevented initialization from determining information about or initializing the graphics system.
Solution: Ensure that graphics have stopped before you attempt whatever caused this error. Check software revisions.
Failure in GFX_MAPALL ioctl in routine-name
Explanation: The operating system prevented initialization from determining information about or initializing the graphics system.
Solution: Ensure that graphics have stopped before you attempt whatever caused this error. Check software revisions.
Failure in VENICE_VS2_SET_SHADOW ioctl in routine-name - reason
Explanation: An internal software error has caused a problem in the specified routine.
Solution: Report this problem to technical support.
graphics getboardinfo() failed during MCO initialization: reason
Explanation: The operating system prevented initialization from determining information about the graphics system.
Solution: Ensure that graphics have stopped before you attempt whatever caused this error. Check software revisions.
Invalid pipe number during MCO initialization
Explanation: A pipe number was specified that was out of range of available pipes.
Solution: Report this error to technical support.
MCO: bad base address address, using default
Explanation: The DG2 EEPROM contains an address for MCO that is invalid because of range or alignment.
Solution: Rewrite the address in the DG2 EEPROM with the /usr/gfx/ucode/RE/vs2/vs2prom program (or /usr/gfx/ucode/REC/vs2/vs2prom or /usr/gfx/ucode/REV/vs2/vs2prom, depending on whether you are running RealityEngine, Crimson, or Onyx).
MCO: board found at address
Explanation: The MCO board was successfully located and identified.
Solution: No resolution is necessary. Consider this a happy message.
MCO: board on pipe <pipe-number> found at <address>, VME adapter <Adapter>, but not properly connected to graphics via DVI cable
Explanation: An MCO was detected at the specified location. However, a check of the DVI cable between MCO and DG2 detected a problem.
Solution: Check connection of DVI cable between MCO and DG2. If connection appears correct, you can use the MCO Data Path diagnostic and tehn process of elimination to determine if the problem lies in the DG2 board, MCO board, paddleboard, or cabling.
MCO: can't map VME adapter number <address> (10 slot <slot> adapter <adapter>
Explanation: The RealityEngine device driver was unable to map the specified adapter number on which it believed the MCO was located. This indicates an error in system software.
Solution: Contact technical support.
MCO hardware is present, but unusable because the number of RM boards has changed from oldcount to newcount. Rerun setmon.
Explanation: Each MCO format configuration is unique to the graphics hardware configuration. When setmon was run previously, the MCO's EEPROM was loaded with a configuration specified for the number of RM (Raster Memory) boards in place at that time. This message is printed when initialization detects that the number of RM boards has changed. Therefore, the format configuration in the MCO EEPROM is out of date.
Solution: Run setmon again, loading the MCO EEPROM properly.
MCO: unable to read signature from <address> (bad read from MCO eeprom)
Explanation: The address specified in the message is a valid address for a Multi-Channel Option, so the RealityEngine device driver attempted to read the MCO signature from that location. The MCO signature was not detected.
Solution: You will see this message whenever a device other than MCO is found in a potential address used by MCO; if another device does exist in that location, you may ignore this message.
If an MCO exists at that location, the on-board EEPROM is not responding as an MCO. Initialize the MCO board's EEPROM with the vs2prom program.
MCO: Using MCO on pipe <pipe-number> - address <address>, VME adapter <adapter>
Explanation: The RealityEngine device driver has successfully detected an MCO in the specified location.
Solution: None necessary. This behavior is correct.
Performing MCO initialization on hardware not a RealityEngine
Explanation: Initialization is being performed on a system with incorrect graphics installed; the MCO does not work.
Solution: Install RealityEngine graphics.
RE pipe <pipe-number>: can't read dg2 eeprom
Explanation: The RealityEngine device driver was unable to read the DG2 EEPROM to determine whether MCO should be enabled for output. Because the status of MCO could not be determined, output will be directed to DG2 until setmon is run.
Solution: Run setmon to reset the DG2 EEPROM.
The address of the MCO responds, but not as a MCO. Check EEPROM validity.
Explanation: A check was made to determine whether the MCO was present. The MCO identification sequence should be stored in the MCO's EEPROM. This MCO's EEPROM did not contain the identification sequence.
Solution: Initialize the MCO board's EEPROM with the vs2prom program. Check that the address of the MCO as specified in the DG2 EEPROM is valid; correct it, if necessary, with the /usr/gfx/ucode/RE/vs2/vs2addrset program (or /usr/gfx/ucode/REC/vs2/vs2addrset or /usr/gfx/ucode/REV/vs2/vs2addrset, depending on whether you are running RealityEngine, Crimson, or Onyx).
The MCO hardware is present, but unusable because DG2 hardware is down-rev
Explanation: The MCO requires a DG2 of at least a certain revision level. The installed DG2 board revision is not sufficiently recent.
Solution: Install a DG2 board of recent revision.
The Multi-Channel Option hardware is present, but the EEPROM seems to be outdated or corrupt. Run setmon -S to repair it.
Explanation: Software has detected one of the following problems with the Multi-Channel Option EEPROM:
the stored revision number does not match that which the software expects
the size of the stored data does not match that which the software expects
the checksum stored in the EEPROM is not internally consistent
Solution: Run setmon -S to rewrite the EEPROM.
The Multi-Channel Option VME address (<address>) as specified by RealityEngine device driver is unusable.
Explanation: The RealityEngine device driver specified a bad address for MCO.
Solution: Contact technical support.
Unable to find number of graphics boards during MCO initialization: reason
Explanation: The operating system prevented initialization from determining the number of graphics boards.
Solution: Check software revisions.
Unable to load filename in venice_get_vs2_xilinx_config() during MCO initialization: reason
Explanation: The specified format file could not be loaded for the reason detailed.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
Unable to map MCO during board probe (reason), assuming it is not present
Explanation: The initialization software was unable to map the MCO board for the specified reason.
Solution: Correct the specified problem. Check whether the file /usr/sysgen/master.d still contains the entries describing the MCO address (shipped by default).
unable to malloc() for getboardinfo() during MCO initialization
Explanation: An internal error has kept initialization from proceeding.
Solution: Report this error to technical support.
Unable to open /dev/graphics during MCO initialization: reason
Explanation: The operating system prevented the software from manipulating the RealityEngine graphics board.
Solution: Check software revisions.
Unable to open device during initialization of MCO - reason
Explanation: The specified device is the means by which MCO initialization software accesses the MCO hardware. The device could not be opened for the reason specified.
Solution: Correct the problem as specified in the reason. If you cannot correct the situation, report this to technical support.
Unable to set diagnostic flag in routine-name - reason
Explanation: The operating system prevented initialization from determining information about or initializing the graphics system.
Solution: Ensure that graphics have stopped before you attempt whatever caused this error. Check software revisions.
venice_get_vs2_xilinx_config(): Error reading xilinx config file filename
Explanation: The initialization software could not load the specified microcode file.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
venice_get_vs2_xilinx_config: unknown xilinx type type
Explanation: The initialization software found a microcode file of the wrong type.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
vs2_loaducode(): can't open vof file, filename
Explanation: The specified format file could not be opened.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
vs2_loaducode(): internal error in filename
Explanation: The specified MCO format file was corrupt.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.
vs2_loaducode(): unexpected EOF in filename
Explanation: The specified MCO format file was not readable.
Solution: Identify and repair the file. If necessary, install the Multi-Channel Option software again.