IRIX GSN error messages are written into the /var/adm/SYSLOG file or displayed at the terminal; some messages appear in both places. Within the SYSLOG file, each message is preceded by the date, time, hostname, name of the process that created the message, and process ID number, as illustrated in Figure 4-1 (process ID number not shown).
This chapter contains an alphabetic listing of the IRIX GSN product's error messages. Only the text of the error message (see the text of error message in Figure 4-1) is included in this list.
With each error message is a short description of the problems the message may indicate. The list contains only messages that indicate an error or problem; it does not contain informational messages that occur during normal operation.
Messages in this chapter are alphabetized according to the following rules:
Each message is alphabetized by the numerals (0–9) and letters (a–z) of the message's text. Numerals precede letters.
Nonletters (for example, - or %) and blank spaces are shown in the text of the message, but are ignored in alphabetization. For example, these messages are alphabetized as follows: gsn . . ., gsnnet . . ., gsn_open . . .
When an error message includes an item that the software specifies differently (fills in) for each instance of the message, this item is displayed in italic font and labeled with a generic name (for example, filename). The generic names are skipped for alphabetization purposes. For example, the error message goofy not responding is located under hostname not responding among the “n” listings. Common generic names used in this listing include hostname, interfacename, packet#, version#, userentry, reason, digit, filename, and hexnumeral.
If you cannot find an error message in the listing, identify potential fill-in words, then look up the message without those words.
Capitalization is not considered in alphabetization.
Note: The lists of error messages in this chapter cover only those unique to IRIX GSN. Standard system error messages, even when caused by the IRIX GSN code, are not covered.
The harpd daemon logs error conditions in two places: into its own log file, and into the system log.
Into its own logfile harpd writes all error conditions, as well as information pertaining to a wide variety of normal operational events. This logfile is named /var/adm/harpd.log by default, but this can be changed by harpd configuration file commands. Under normal operations, the system administrator need not be concerned with the contents of the harpd logfile, although it may be of use when inexplicable harpd error messages are detected in the system log. The harpd logfile's message formats are beyond the scope of this manual.
The harpd daemon writes all error messages that may be of interest to the system administrator into the system log. This section lists only the harpd error messages that are displayed in the system log.
harpd: invoked with invalid arguments
The harpd daemon was invoked with invalid arguments. For a list of valid arguments, see the harpd(1M) man page.
harpd: could not initialize
Global initialization of harpd did not occur. This message is preceded by more specific error messages indicating the cause of the error.
harpd: DOWN on already-down interface, index %d
The harpd daemon has received a "DOWN" command from the kernel for a GSN interface that is already down. This causes no problem, but should not occur.
harpd: got zero IP address from kernel
An ARP_Request request was received from the kernel that contains a zero IP address. The request is discarded; the error indicates that the kernel is misbehaving.
harpd: got zero ULA from kernel
An InARP_Request request was received from the kernel that contains a zero ULA. The request is discarded; the error indicates that the kernel is misbehaving.
harpd: unknown ioctl from kernel: %d
The kernel passed us a request that originated as a user-level ioctl() for some program running as root. However, the ioctl type is unrecognized. This indicates a problem with the originating problem, not with the kernel. The %d is the unrecognized ioctl command value.
harpd: reply failed: %s
The harpd daemon processed an ioctl request from the user, and has tried to send the return status of the ioctl back to the user, but the ioctl that harpd uses to communicate with the kernel has itself failed. The %s indicates the reason for the failure.
harpd: kernel ioctl error %d
The harpd daemon processed an ioctl request from the user, and has sent the return status of the ioctl back to the user. While this ioctl to the kernel itself succeeded, the kernel found an error in the data of the ioctl. The %d indicates the error.
harpd: invalid opcode in header
The kernel has sent a command that cannot be parsed. This causes harpd to abort. The command that caused us to fail is dumped in hexadecimal in the harpd log.
harpd: shutting down at superuser request
The harpd daemon has received a STOP request from the superuser. This is the preferred method of terminating harpd.
harpd: internal error with %s
The harpd daemon has detected a serious inconsistency in the data structures used for IP-to-ULA mapping. harpd continues, but such inconsistencies should be examined.
harpd: out of table space
The harpd daemon is being asked to record more IP-to-ULA mappings than it has space to hold. This error indicates that the harpd tables should be sized larger, or there are extraneous mapping entries that should be removed.
harpd: invalid index: %d
A request has come from the kernel that refers to an interface that is not a GSN interface. This indicates a critical misunderstanding between the kernel and harpd; harpd aborts.
harpd: ioctl op %d failure: %s
For some operations, the kernel's HARP driver entrusts harpd with mbuf pointers, which it passes back to the kernel when it is to dispose of them or transmit their contents. This error indicates that the kernel has found an error in a harpd command to operate on one of these mbufs. Since mbuf misbehavior is a critical problem, harpd aborts immediately after issuing this message.
harpd: ioctl op %d kernel reject: %s
For some operations, the kernel's HARP driver entrusts the harpd with mbuf pointers, which it passes back to the kernel when it is to dispose of them or transmit their contents. This error indicates that the kernel has found an error in a harpd command to operate on one of these mbufs. Since mbuf misbehavior is a critical problem, harpd aborts immediately after issuing this message.
harpd: caught signal %d
The superuser has sent harpd an unexpected signal. harpd aborts.
harpd: shutting down on SIGINT
The superuser has sent harpd a SIGINT. At this time, harpd calls exit(0) on the reception of this signal, but this signal has been reserved for harpd activity.
harpd: shutting down on SIGTERM
The superuser has sent harpd a SIGTERM. harpd calls exit(0) on the reception of this signal.
harpd: Error with signal %d sigaction: %s
During initialization, harpd calls sigaction() for all signals. This message indicates that the sigaction() call has failed for signal %d. The reason for the failure is indicated by %s. harpd calls exit(1) after printing this message.
harpd: Error on _daemonize: %s
During initialization, harpd calls _daemonize() to daemonize the process. This message is printed when this system call fails. The reason for the failure is indicated by %s. harpd calls exit(1) after printing this message.
harpd: Can't open interface file %s: %s
The harpd daemon was unable to open the HARP daemon device. The actual name of the device is given by the first %s; the reason for the failure is indicated by the second %s. The most common reason for this failure is “Resource busy," which normally indicates that an instance of harpd is already running, so a second instance cannot start. harpd calls exit(1) after printing this message.
harpd: can't lock memory: %s
During initialization, harpd calls plock() to lock memory. This message indicates that that call failed. The %s indicates the reason for the failure. harpd calls exit(1) after printing this message.
harpd: can't set priority: %s
During initialization, harpd calls setpriority() to lock memory. This message indicates that that call failed. The %s indicates the reason for the failure. harpd calls exit(1) after printing this message.
harpd: shared memory version mismatch: we want %5.3f, we found %5.3f
The HARP driver and harpd share a number of tables. For these two entities to share tables, they must agree upon the table structures. This message indicates that harpd is looking for a different table structure than the one being used by the HARP daemon. harpd calls exit(1) after printing this message.
harpd: INIT ioctl failed: %s
The harpd daemon is responsible for telling the HARP driver to create shared table space for their mutual use. This message is printed when the ioctl() used to send this request to the kernel fails. harpd calls exit(1) after printing this message.
harpd: INIT ioctl error: %s
The harpd daemon is responsible for telling the HARP driver to create shared table space for their mutual use. This message is printed when the creation request is rejected by the kernel. harpd calls exit(1) after printing this message.
harpd: cannot change entry count after first harpd run
After the kernel has reserved table space for the mutual use of the HARP driver and harpd, harpd cannot request a different table size. This message indicates that harpd has requested a different table size. harpd calls exit(1) after printing this message.
harpd: cannot change bucket count after first harpd run
After the kernel has reserved table space for the mutual use of the HARP driver and harpd, harpd cannot request a different table size. This message indicates that harpd has requested a different table size. harpd calls exit(1) after printing this message.
harpd: can't mmap harp tables: %s
The harpd daemon uses the mmap call to map the kernel-created shared mapping table into its address space. This message indicates that the mmap() call failed. The %s indicates the reason for the failure. harpd calls exit(1) after printing this message.
harpd: error loading interface data: %s
The harpd daemon was unable to load the kernel's interface table, which it needs to determine which GSN devices are valid. The %s indicates the reason for the failure. harpd calls exit(1) after printing this message.
harpd: no GSN interfaces found on system
The harpd daemon loaded the kernel's interface table, but found no configured GSN devices on the system. With no GSN devices, harpd cannot run. harpd calls exit(1) after printing this message.
harpd: config files errors
Errors were encountered while reading the harpd configuration file. This message is preceded by specific error messages indicating which errors occurred. harpd calls exit(1) after printing this message.
harpd: can't malloc mib
The harpd daemon was unable to use the malloc() command to allocate enough space for the system's routing table MIB. harpd calls exit(1) after printing this message.
harpd: can't malloc %d bytes
The harpd daemon was unable to use the malloc() command to allocate enough space for the system's routing table. harpd calls exit(1) after printing this message.
harpd: can't open logfile %s: %s
the harpd daemon was unable to open the logfile. The name of the logfile harpd tried to open is printed as the first %s; the reason for the open failure is specified in the second %s. harpd continues after printing this error, although (obviously) nothing is logged until the daemon can successfully open a logfile. If multiple, sequential logfile open() attempts fail, only the first of them results in a syslog error message.
harpd: can't write to logfile %s: %s
The harpd daemon was unable to write to the logfile. The name of the logfile harpd currently has open is printed as the first %s; the reason for the write failure is specified in the second %s. harpd continues after printing this error, although (obviously) nothing is logged until write operations begin working. If multiple, sequential logfile write() attempts fail, only the first of them results in a syslog error message.
harp_init: bad sizeof(uint64_t)
The size of the uint64_t type is not 64 bits. This indicates that there is a fundamental problem in the way the kernel was built; this condition causes a system panic.
harp_init: bad sizeof(Hmac)
The size of an Hmac is 64 bits. This indicates that there is a fundamental problem in the way the kernel was built; this condition causes a system panic.
harp_init: mbufs are too small for us!
A normal mbuf must be able to hold a HARP request. This indicates that there is a fundamental problem in the way the kernel was built; this condition causes a system panic.
harp_init: can't make hw device %s: error = %d
One of the required HARP devices cannot be made. Without the HARP devices, HARP (and any GSN devices) cannot function. This message indicates that the kernel was not built properly. The error number is the error returned from hwgraph_char_device_add().
harp_open: process %d opened invalid harp device %lx
The HARP driver was called to open a non-HARP device, which should never happen. The %d is the process number of the offending process; the %lx is the address of the invalid device.
harp_close: process %d closing invalid harp device %lx
The HARP driver was called to close a non-HARP device, which should never happen. The %d is the process number of the offending process; the %lx is the address of the invalid device.
harp_read: empty queue found, expected data
The HARP driver attempted to satisfy an outstanding read request from the harp daemon, but found that nothing was queued. This message indicates that the HARP driver is seriously confused.
harp_write: bad write on gsn%d
In attempting to transfer data from the user to the GSN device, the HARP driver received an error from the GSN driver. This indicates a GSN problem that is affecting the GSN driver. There will likely be syslog messages from the GSN driver when this message occurs. The HARP driver will print this message only one time for any sequence of write errors to a particular GSN. This avoids flooding the syslog. If the message appears multiple times in the syslog, it indicates that between successive messages, at least one transfer to the GSN driver was successful.
harp_ioctl: out of mpkt space
The HARP driver contains a pool of data pointers where it stores mbufs that are (partially) entrusted to the harp daemon. This message indicates that too many messages have been passed to the daemon without any instructions from the daemon about their disposal. The message indicates that the HARP daemon is not keeping up with the data flow. Messages for which mpkt space cannot be reserved are dropped. The HARP driver will print this message only one time for any sequence of mpkt space errors. This avoids flooding the syslog. If the message appears multiple times in the syslog, it indicates that between successive messages, at least one packet was successfully queued for the daemon.
harp_resolve called with IP = 0
The harp_resolve() routine is called by the GSN driver to find a ULA that matches some specific IP address. This message indicates that no IP address was provided in the request. The HARP driver will print this message only one time for any GSN interface. This avoids flooding the syslog.
harp_invresolve called with ULA = 0
The harp_invresolve() routine is called by the GSN driver to find an IP address that matches some specific ULA. This message indicates that no ULA was provided in the request. The HARP driver will print this message only one time for any GSN interface. This avoids flooding the syslog.
harp_input: no space in mbuf for Hmac
A HARP packet has been received that is so close to the maximum size of a default mbuf (MLEN) that there is no room for the 8-byte Hmac header. HARP packets should not be nearly this large; this error indicates that invalid HARP packets are being received from the network.
harp_request: out of mpkt space
The HARP driver contains a pool of data pointers where it stores mbufs that are (partially) entrusted to the harp daemon. This message indicates that too many messages have been passed to the daemon without any instructions from the daemon about their disposal. The message indicates that the HARP daemon is not keeping up with the data flow. Message for which mpkt space cannot be reserved are dropped. The HARP driver will only print this message one time for any sequence of mpkt space errors; this avoids flooding the syslog. If the message appears multiple times in the syslog, it indicates that between successive messages at least one packet was successfully queued for the daemon.
harp_cb_append: queue too big, tossing packets
To limit the problems a misbehaving harp daemon can cause, there is a limit to the number of messages that the HARP driver will queue awaiting ingestion by the daemon. This message indicates that this upper limit has been reached. The message will only appear once in the logfile for any number of sequential queue-full errors; only after the daemon has read messages off the queue and the HARP driver once again reaches the maximum limit will the message be rewritten.
harp_send_mpkts: bad packet family: %d
The HARP driver might have to hold data packets while an IP-to-ULA mapping is being resolved. This message indicates that one of these messages, which looked fine when it was saved, does not look so fine after the address resolution has taken place. Specifically, the address family of the socket which owns the message is no longer understood. This message should never occur; it indicates that there is a serious data inconsistency within the HARP driver.
harp_get: max retries hit
At the core of the HARP driver is its IP-to-ULA mapping table and the harp_get() routine, which finds the entry that matches a given mapping request. The harp_get() routine is multi-threaded and is totally asynchronous from all other HARP driver or HARP daemon processing. As a result, it is possible (although extremely improbable) that an IP-to-ULA mapping will change at the very instant that it is being resolved. To recover from this unlikely event, harp_get() retries the mapping a large number of times. If it happens that harp_get() fails for every single one of these attempts, the HARP driver prints this error message and gives up, returning a "no match found" error to its caller.
harp_request_init: can't alloc harp table
The first time the HARP daemon starts, it allocates kernel space for the various HARP kernel-resident tables. This error indicates that the HARP driver could not allocate this table space. The HARP daemon (and all GSN devices) cannot function without this table space.
This section lists gsncntl error messages displayed in the user's window.
Bad MAC address string
The MAC address entered is not formatted correctly. The error was detected while parsing the MAC address for the setmac command.
Cannot enable/disable watchdog
The GSN_WATCHDOG_ENABLE ioctl failed.
Cannot get firmware error log mask
The GSN_GET_FWLOG_MASK ioctl failed.
Cannot set firmware error log mask
The GSN_SET_FWLOG_MASK ioctl failed.
Couldn't get GSN IC Port histogram information
The GSN_GET_IC_HIST ioctl returned an error.
Couldn't get GSN SSRAM Port histogram information
The GSN_GET_SSRAM_HIST ioctl returned an error.
Couldn't get GSN ULA
The GSN_GET_DEV_ULA ioctl returned an error.
Couldn't open GSN device: devicename
An open system call on the specified device failed.
Couldn't set debug value
The GSN_SET_DEBUG ioctl returned an error
Couldn't set GSN MAC address
The GSN_SET_PROM_ULA ioctl retuned an error.
Couldn't set IC Port histogram value
The GSN_SET_IC_HIST_CTL ioctl returned an error.
Error ioctl call failed
The GSN_SHUTDOWN ioctl failed.
Error ioctl call failed
The GSN_STARTUP ioctl failed.
Error reading from SUMAC
The GSN_ADMIN_WRRD ioctl returned an error.
Error writing to SUMAC
The write system call returned an error while writing to the GSN device.
GSN adapter already has a MAC address programmed
The device has an existing MAC address that is not the same as the one entered.
GSN_GET_FLASH failed
The GSN_GET_FLASH ioctl returned an error.
GSN_GET_VERSIONS failed
The GSN_GET_VERSIONS ioctl returned an error.
GSN_PGM_FLASH failed
The GSN_PGM_FLASH ioctl returned an error.
GSN production firmware magic header is wrong
The magic number in the PROM firmware to be downloaded is incorrect. It is probably not SHAC PROM firmware.
Gsngetstate: Error ioctl
The GSN_GET_LAST_STATE ioctl failed.
Gsngetstate: Error ioctl
The GSN_GET_STATE ioctl failed.
Illegal destination timeout value
An illegal timeout value was passed for setting the destination timeout value (dtimeo command). The value must be less than or equal to 1023.
Illegal prom sector specified: sector
The number of the specified sector exceeds the maximum allowed.
Illegal source timeout value
The stimeo timeout value exceeds the maximum of 1023.
Insufficient number of arguments
An insufficient number of arguments were passed to the command option.
Invalid MAC address: must start with 8:0:0x69
An SGI MAC address must begin with 8:0:0x69.
Invalid number of arguments
An incorrect number of arguments were entered for this command.
Ioctl call failed
The ioctl GSN_GET_STATS system call for the GSN device returned an error.
Mask must be a hex number that starts with 0x
The firmware mask should be a hex format number.
Mode is out of range: number
The mode exceeds the maximum value allowed.
Packet PID does not match
The packet data returned does not match what was sent out.
Trouble with GSN_ERASE_FLASH might need to shutdown device
The GSN_ERASE_FLASH ioctl returned an error.
Unable to read Sumac Global Parm register
In an attempt to read the SUMAC GLOBAL_PARMS register, the driver returned an error.
Unable to write Sumac Global Parm register
In an attempt to write the SUMAC GLOBAL_PARMS register, the driver returned an error.
You'll run past end of EEPROM!!
The boot code plus firmware exceeds the physical size of EEPROM.
You'll run past end of sector 0 and clobber the MAC address!!
The SHAC PROM boot code is larger than a PROM sector and will overrun the next sector.
This section lists gsntest error messges displayed in the user's window.
Accept error
The accept system call returned an error to the loopback receive process.
clientname: Unspecified client name
No target host address was specified. Using default SERV_HOST_ADDR defined in gsntest.h.
Data miscompare Exp: hexnumber Act: hexnumber
A data miscompare occurred on the test while in loopback mode.
Error opening device_name for dev access
The open system call on the specified device returned an error. Check for the existence of the device file.
Error writing to socket
The write system call returned an error while sending data in loopback mode.
Illegal protocol specified number, test->proto
The protocol specified was not PROTO_TCP, PROTO_UDP, or PROTO_ST and is not allowed.
Link state is unknown - gsn boot sequence has not completed
Could not determine the link state. The state was not internal loopback, external loopback, point-to-point, or switch.
option: unknown option - character
The specified option is not recognized. See allowable options listed in the usage message.
Ping error on hop number (status == hexnumber)
An error was detected after attempting to send an admin packet out the specified number of hop counts. Check other error messages for timeouts or unexpected packet responses
Read time-out
A SIGALRM was received by the loopback receive process, indicating a timeout on a read operation.
Sock open failed --
The socket system call returned an error. Unable to get a socket for this protocol.
Trouble writing
The write from the client side did not write out the requested number of bytes. The write failed.
Unable to bind to local address
The bind system call returned an error to the loopback receive process.
Unable to connect
The connect system call returned an error on this socket.
Unable to get host name
The gethostname call failed.
Unable to open fd/socket
The socket system call returned an error. Unable to get a socket for this protocol.
Unable to start timer
The gsntest command was unable to start the timer.
This section lists the error messages displayed on the system console by the IRIX GSN utilities and driver and echoed to the SYSLOG file.
Admin pkt NG: cmd = command, status = status in state
An admin command or response packet was received but is being dropped due to a detected error condition. Possible reasons include bad packet status or incorrect virtual channel used.
Allocation of xtalk interrupt
A kernel xtalk utility function was unable to allocate interrupt resources for the xtalk interrupt vector.
Attempting to program flash EEPROM
Informative message that the EEPROM for this device is going to be flashed in the next step.
Bad ADMIN_READ from SUMAC
An admin read register request was sent to the GSN SUMAC ASIC but the response from the SUMAC was not correct.
Bad READ_RESPONSE not valid for SUMAC
A read response admin packet was received but is not valid for the current GSN state.
Bad WRITE_RESPONSE from the SUMAC
An admin write register request was sent to the GSN SUMAC ASIC but the response from the SUMAC was not correct.
Cable problem
A cable was not detected.
Cannot alloc memory for admin rx
Out of receive descriptor entries for this node.
Cannot get memory for FLUSH PAGE.
An attempt to allocate kernel memory for GSN to use for bufx flush token target pages failed.
Can't malloc HARP HW table
The HARP module was unable to allocate memory from the kernel for the HARP hardware entry table.
Can't malloc HARP IP table
The HARP module was unable to allocate memory from the kernel for the HARP IP table.
Check_inventory_info(): hwgraph_inventory_add(number) err: error
A hardware graph utility function returned an error while trying to add dual xtalk entries to the hardware inventory.
Check_inventory_info(): xp1->xp0 hwgraph_edge_add(numbers) err: error
A hardware graph utility function returned an error while trying to add an edge. It was replacing a single xtalk with dual xtalks and cross linking the hardware graph edges.
Check_inventory_info(): xp0->xp1 hwgraph_edge_add(numbers) err: error
A hardware graph utility function returned an error while trying to add an edge. It was replacing a single xtalk with dual xtalks and cross linking the hardware graph edges.
Check_inventory_info(): xp0 hwgraph_edge_remove(number) err: error
A hardware graph utility function returned an error while trying to remove an edge. It was replacing a single xtalk with dual xtalks and cross linking the hardware graph edges.
Check_inventory_info(): xp1 hwgraph_edge_remove(number) err: error
A hardware graph utility function returned an error while trying to remove an edge. It was replacing a single xtalk with dual xtalks and cross linking the hardware graph edges.
Check_inventory_info(): hwgraph_inventory_remove (hexnumber) err: error
A hardware graph utility function returned an error while trying to remove an xtalk entry from the hardware inventory.
CLR_BUFX bad port numb number flags flags
During an attempt to clear transmit bufx table entries, the port table entry number passed into the current function was not in the range allowed by SHAC (0-2047).
Could not initialize warm reset propagation mask in SUMAC
An admin write register request was sent to the GSN SUMAC ASIC to initialize a register but the operation failed.
Desc_dup for number
An error was returned from an attempt to duplicate a device descriptor. Possible memory allocation problem.
Device_desc_dup() for device failed
While setting up a new xtalk device, a new device descriptor could not be created. An error was returned during an attempt to create a descriptor from the current xtalk vertex handle. An EIO error is returned to the caller.
Did not get RX flush token. f_offset number token number bft hex number
An RX flush token was sent but not received. The maximum number of flush retires was attempted.
Driver in UNKNOWN STATE.
The GSN link bring up received an admin packet during which the driver was in an unknown state.
ENOMEM for new_xtalk
The kernel memory allocator is unable to allocate memory for a new xtalk context.
ENOMEM for qid: number
The kernel memory allocator could not allocate memory pages to contain this QID area.
Erasing flash EEPROM
Informative message that the EEPROM is being erased.
Erasing flash EEPROM MAC sector
The GSN driver is processing an ioctl to erase the GSN EEPROM MAC sector for this device. The flash prom will then be written and re-read to check for errors.
Failed hwgraph add inventory number: under number: number
The hardware graph susbsystem returned an error when the driver attempted to add the xtalk vertex to the hardware inventory.
Failed to add convenience edge string to string
A hardware graph utility function returned an error while trying to add a graph edge from the convenience edge (/hw/gsn/<devnum>) to the physical device.
Failed to allocate QID number interrupt
The xtalk interrupt handler could not be allocated for this QID for this interface. An error was returned by a kernel xtalk interrupt management function.
Failed to cfg i/f:
An error was detected during an attempt to configure the interface. Check possible preceding messages for details on the specific error.
Failed to create string: under number
The hardware graph subsystem could not add this GSN character device node to the /hw tree and returned an error.
Failed to start timer
An attempt to queue a timeout event to the kernel timeout handler failed.
Failed to start timer
The kernel timeout utility returned an error when called by GSN to set a link timeout entry for this interface.
Failed to traverse into string: under number
A hardware graph utility function returned an error trying to traverse the graph from the /hw root to the convenience vertex in the message.
Firmware processor did not finish booting
The firmware processor portion of the GSN adaptor did not finish booting.
Firmware size or offset = 0
One of the tunables (gsn_firmware_size or gsn_firmware_offset) is zero.
FW BOOTING bit set: FW completely booted
The SHAC control and status register indicates that the SHAC successfully booted its firmware.
FW BOOTING bit set: FW entered reset
The SHAC control and status register indicates that the SHAC is still booting its firmware.
FW BOOTING bit set: FW successfully tested
The SHAC control and status register indicates that the SHAC successfully tested its firmware.
FW did not come out of SSRAM
The EEPROM was reflashed and an attempt was made to reboot the processor but it did not reboot successfully after waiting past a timeout period.
Get_num(hexnumber)
An invalid number of SHAC SSRAM FIFO entries was requested during configuration of the interface.
GSN adapter dynamic loopback unsupported at this time
This is a currently unsupported GSN option.
GSN adapter shutdown unsupported at this time
This is a currently unsupported GSN option.
GSN adapter startup unsupported at this time
This is a currently unsupported GSN option.
GSN cannot alloc memory for admin rx
There are no more receive descriptor entries for this node.
Gsn device: can't allocate device driver structure
One of several possible errors occurred while trying to set up the hardware graph information for the device driver.
Gsn: intr_connect() number. number. number : number!
An error was encountered while trying to connect the GSN error interrupt vector.
Gsn_admin_cmd(cmd: command)
An attempt was made to send an admin packet with an invalid cmd code or an admin packet was received but contains an invalid cmd field (cmd > ADMIN_MAX_COMMANDS). The check was made while updating the statistics counters.
Gsn_admin_desc_entries: entries
The tunable parameter, gsn_admin_desc_entries, exceeds the maximum of MAX_GSN_ADMIN_DESC_ENTRIES or is not a power of two.
Gsn_char_timeout(): number
An attempt was made to start a timeout event for this driver but one is active already.
Gsn_char_timeout_handler():timeout_id!
The GSN timeout handler was called but the timeout id was zero. There was no valid timeout event queued.
Gsn_flash_req cmd=command cmd
A GSN function called this function with an unrecognized command option (cmd).
Gsn_flash_req cmd= hex number???
A GSN function called this function with an unrecognized command option (cmd).
Gsn_get_dev(): error hexnumber
gsn_get_dev() was called with an illegal flag. The flag is not recognized.
Gsn_ioc_fwlog_mask cmd = hex number???cmd
An unrecognized flag was passed to function gsn_ioc_fwlog_mask while trying to perform an ioctl command on the firmware processor.
Gsn_ioc_gfx_credits cmd = hex number???cmd
An unrecognized flag was passed to function gsn_ioc_fwlog_mask while trying to perform a GFX hub credit ioctl.
Gsn_mac_req cmd=command cmd
A GSN function called this function with an unrecognized command option (cmd).
Gsn_open: add_exit err number
An open of the gsn device failed.
Gsn_qid_per_node: number
The tunable parameter, gsn_qids_per_node, exceeds the limit of MAX_CPUS_PER_NODE.
Gsn_register_t size: number.number!
An internal driver structure limit has been exceeded. This is a driver software error.
Gsn_st_if_input: Failed to allocate mbuf for receive descriptor!
The ST portion of the GSN driver failed to allocate an mbuf from the kernel mbuf allocator.
Gsn_st_if_output: Non VC2,3 Data Op specified
An ST data operation was specified but the virtual channel number is not correct. IRIX ST data operations must use VC2 or VC3.
Gsn_st_if_txdone: Global credits exceeded number on port number VC number token hex number
The total system wide transmit credits for this VC have been exceeded.
Gsn_st_if_txdone: received bypass txdone
A txdone interrupt was received for an OS bypass operation.
Gsn_st_set_bufx:Misformed Node Group
A memory striped bufx scheme is being used but it was detected that xtalk ports were not being alternated for each bufx as required. Correcting this now.
Gsn_st_set_port: string number Could not satisfy bypass tx credit request
This function could not satisfy the request for tx credits from the credit pools. One or more tx credit pools for the VCs are empty.
Gsn_tx_admin out of tx credits... sending anyway
An admin packet is being prepared to be transmitted but the interface is out of admin transmit queue credits. Sending anyway.
Gsn_ula_req cmd=command, cmd ioctl
This GSN ioctl is not supported.
HARP server error: downing interface
An error occurred during an attempt to set up the HARP table for this interface.
Harp_update: invalid PERM case
An error was detected for a HARP permanent entry during a HARP update operation.
Hwgraph_edge_add (number, number, string) err
The hardware graph subsystem could not add this edge vertex to the xtalk sister device and returned an error.
Hwgraph_vertex_name_get: name
The hardware graph subsystem returned an error while retrieving the canonical name of the xtalk vertex from the vertex handle.
IDESC len == 0
An interrupt descriptor shows an mbuf chain of length zero, which is an error. This was detected during an IP receive interrupt.
IDESC status hex number
Illegal status found in an interrupt packet.
Ifconfig gsnnumber UP
An error occurred during an attempt to set up the HARP table for this interface.
If_gsn_net_output() unsupported sa_family number
An unrecognized or unsupported address family was specified during GSN output for this interface. Only AF_INET is supported.
If_gsn_st_input: a ddq entry was missed (seq=number) (ddq_idx=number)
The current ST receive descriptor does not match the current sequence number for the port. The next receive descriptor does not match either. Cannot find the correct receive descriptor for this interrupt. Returning an error to the ifnet layer.
If_gsn_st_input: a ddq entry was skipped (seq=number) (ddq_idx=number)
An ST receive descriptor was skipped because it didn't match the current sequence number for the receiving port. The next descriptor matches and will be used. The result is that one receive descriptor slot will be unprocessed. This is a workaround for a SHAC edge condition.
Ignoring admin pkt cmd: command, daddr=address
An admin packet was received but will be ignored because of an error condition.
In error state.
The interface adapter specified has detected an error and is changing to an error state.
Incorrect manufacturing number: number
The SHAC XT_ID field shows an unrecognized manufacturer's number.
Incorrect xt_id: partnumber
The SHAC XT_ID field shows an unrecognized widget part number. Cannot determine the widget type.
Init Node Array: Xtalk device not attached
An attempt was made to configure an xtalk but the GSN device structure (gsn_dev_s) for this interface does not find it connected.
Interface has more aliases then the st_ifnet can handle - limiting.
The maximum number of interface aliases has been exceeded. The actual number of aliases will be limited. The maximum number is compiled into the kernel and cannot be changed.
Interface usable.
The driver state has transitioned to GSN_STATE_LINK_USABLE.
Intr_connect() numbers
A kernel xtalk utility function was unable to connect previously allocated xtalk interrupt resources with its handler.
Invalid config: SSRAM mbuf entries sm = value, lg = value
One of the tunables does not fall into the allowable ranges. Both values will be set to the defaults.
if_gsn_small_mbuf_entries must be between zero and MAX_SSRAM_IP_RX_SMALL_ENTRIES, inclusively.
or
if_gsn_large_mbuf_entries must between zero and MAX_SSRAM_IP_RX_LARGE_ENTRIES, inclusively.
Invalid config: SSRAM mbuf offsets,value1, value2
One of the tunables exceeds the limit. Either if_gsn_small_mbuf_offset is greater than GSN_SSRAM_UPPER_BOUND or if_gsn_large_mbuf_offset is greater than GSN_SSRAM_UPPER_BOUND. Both values will be set to the defaults.
Link startup error
An error occurred during an attempt to set up the HARP table for this interface.
Link Timeout with switch negotiate
The interface went into the BRING_UP_LINK state and tried to do a ULA_REQUESTED operation on a switch but timed out.
Local interface not usable - VC1 flow controlled due remote interface.ifconfig down/up when remote problem is cleared.
The number of entries in the admin packet transmit queue has been exceeded. A likely cause is VC1 being flow controlled by the remote interface. An ifconfig up/down may clear the problem when the remote interface recovers.
Misconfig of h2f fifo, offset = offset, entries = entries
Because one of the tunables (gsn_h2f_offset or gsn_h2f_entries) is zero or an illegal value, H2F FIFO configuration is skipped.
No firmware will be configured
The tunable parameter, gsn_firmware_offset, exceeds the limit. It must be less than GSN_SSRAM_UPPER_BOUND. Firmware offset and size are being set to null.
No H2F will be configured
The tunable parameter, gsn_h2f_offset, exceeds the limit. It must be less than GSN_SSRAM_UPPER_BOUND. Host to firmware FIFO configuration values are being set to null.
No IP TX FIFO as number
Because one of the tunables (if_gsn_ip_tx_desc_offset or if_gsn_ip_tx_desc_entries) is zero, IP will not be set up.
No read interface.
There is no direct read interface available for this device.
No ST vc configured
The ST virtual channel has not been configured. SSRAM FIFO setup is skipped.
No write interface.
There is no direct write interface available for this device.
Not compiled w/ GSN_DEBUG, ignoring
A GSN ioctl was issued for a debug capability that requires that the GSN driver be compiled with the GSN_DEBUG debug flag. The ioctl will be ignored.
Other end inactive. Waiting for other side to come up
An admin packet was received but indicates that the other side is inactive.
Part number not defined in PROM
GSN board part numbers known to the driver were not found in the PROM.
Pio_base
The base address pointer of the PIO mapped SHAC address area is null. The error was found during verification of PIO mappings for this interface.
Processed more ST than Total idesc???: number. number!
An unknown interrupt descriptor type was detected while handling a GSN interrupt.
Received gsn_admin_cmd(cmd: number)
An admin packet was received but contains an invalid cmd field (cmd > ADMIN_MAX_COMMANDS). The check was made while updating the statistics counters.
Remote end report UNDEFINDED ELEMENT. Using Point <-> Point.
The remote endpoint responded during a negotiate link process but the element type is neither a switch nor a link end element. The endpoint type is an undefined type.
Remote end reports UNKNOWN ELEMENT TYPE. Using Point <-> Point.
The remote endpoint responded during a negotiate link process but the element type is neither a switch nor a link endpoint. The remote element type is of type UNKNOWN_ELEMENT.
Remote END-POINT not responding
The interface went into the BRING_UP_LINK state and tried to find a link end point but failed.
Remote hardware not responding
The interface went into the BRING_UP_LINK state and tried to find a HIPPI end point but failed.
Remote says it is INACTIVE
An admin packet was received but indicates that the remote unit is inactive.
Response timeout
The GSN interface state is being changed to GSN_STATE_TIMEOUT.
RX_FLUSH bad port numb number flags flags
During an attempt to clear receive bufx table entries, the port table entry number passed into the current function was not in the range allowed by SHAC (0-2047).
RX not flushing: base hex number token number f_offset number bft hex number mask hex number port number cpu number
The maximum number of bufx flush token tries has been attempted without success. The ST driver was not successful in flushing in-flight incoming data.
Setting ST vc[number] size=0,was number, offset=hexnumber, was hexnumber
An initialization error was detected during driver bootup while addresses and entry sizes were being assigned for the ST transmit FIFO for this virtual channel. Assignments for this virtual channel are being set to zero.
Set_unit_number failed!
An error was returned from a GSN driver utility function used to set the adaptor unit number.
Setup_fifo_regs
An error was discovered while checking SSRAM address alignment for FIFO entries during interface configuration.
Shac 1.0 is installed.
An old revision of the SHAC ASIC in the GSN adapter that is no longer supported has been detected.
SSRAM already initialized
An attempt was made to initialize SSRAM more than once.
string string timeout - Resetting Sumac
There was a VC full or RX tail timeout error detected. Resetting the SUMAC to recover.
Sumac was not brought out of string chip_reset by fw
The SUMAC ASIC in the GSN adaptor was not reset by the firmware processor portion of the adaptor, indicating a possible problem.
The tunable gsn_ssram_profile is outside the legal range 0 to number. Setting gsn_ssram_profile = 0.
The gsn_ssram_profile tunable is an invalid number. The SSRAM profile used will be the default (profile 0).
Too many refcnt: number
An attempt was made to delete a hardware graph vertex but the graph was not accessible after several retries. GRAPH_IN_USE was returned from the hardware graph subsystem.
Transition from state to unknown state state
The GSN driver attempted to change to an undefined or illegal state.
TX flush did no progress on VC vc Port port Aborting tx flush
During an attempt to clear and flush a range of transmit bufx table entries, the driver detected no progress after a period of time.
TX flush timeout on VC vc Port port Aborting tx flush
During an attempt to clear and flush a range of transmit bufx table entries, the timer for this event expired after multiple tries.
Unable to select CPU from master node for error interrupt
An error occurred during an attempt to configure the error interrupt handler for SHAC errors on this interface. An error was returned from the hardware graph subsystem during an attempt to traverse the graph from the I/O vertex to a CPU.
Unable to select CPU from master node for error interrupt
A kernel hardware graph function returned an error while trying to traverse the hardware graph from the master node vertex to a CPU. No CPU vertex could be found on that node.
Unable to send Admin Reset packet
While the interface was being configured, an error was encountered when an admin reset packet was sent to the SUMAC.
Unable to shutdown RX cleanly.
An error was encountered while trying to shut down receive and transmit activity. The driver was in the process of error recovery and tried to shut down receive processing.
Unexpected part number, Expected: string Got: string There might be a mismatch between the driver and hardware
An unexpected GSN adaptor part number was encountered while probing the adaptor. Check for mismatches between the GSN adaptor hardware revision number on the driver revision.
Unknown device: device
An attempt was made to close a GSN device with a dev_t structure that did not contain a good GSN_DEV_TAG. This dev_t struct does not point to a GSN device.
Unknown ioctl: ioctl
This GSN ioctl is unrecognized. Returning an EINVAL error.
Unrecognized i-desc type: number
A packet was received by this interface with an unrecognized i-desc type.
Verify_pio_addr_offset pio: hex number!
GSN utility function gsn_verify_pio_addr_offset returned an error indicating that there was an xtalk related address that was not correctly set up.
Vertex(device): no pio_map
An error was encountered during an attempt to set up PIO mapping for an xtalk device for this interface. An EIO error is returned to the caller.
Vertex(number.number): no pio addr!
A sanity check showed that the PIO address for an xtalk was not set up correctly.
Watchdog timeout: number
The watchdog timer for this interface has exceeded the maximum tries to determine that the link is in an up or usable state. Resetting the state to reflect a watchdog timeout.
XT 0 node has no QID!
This interface has no interrupt queue available. The error was found during setup of interface for receiving and transmitting of admin packets.
Xwidget_driver_register
An attempt to register a driver on an xtalk port failed.
Xwidget_error_intr(hexnumber, number, hexnumber, hexnumber)
An error interrupt was received by one of the widget error interrupt handlers. This particular widget error interrupt handler receives errors reported by the xbow or hub.
Xwidgetnum: numbers for number
While setting up interrupt handling for the interface an error was detected. The xtalk widget number in the interrupt resource handle does not match the destination widget number for this node.GSNadmin_3.ps