diff options
author | Linus Torvalds <torvalds@ppc970.osdl.org> | 2005-04-16 15:20:36 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@ppc970.osdl.org> | 2005-04-16 15:20:36 -0700 |
commit | 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 (patch) | |
tree | 0bba044c4ce775e45a88a51686b5d9f90697ea9d /Documentation/networking/slicecom.txt | |
download | linux-1da177e4c3f41524e886b7f1b8a0c1fc7321cac2.tar.gz linux-1da177e4c3f41524e886b7f1b8a0c1fc7321cac2.tar.bz2 linux-1da177e4c3f41524e886b7f1b8a0c1fc7321cac2.zip |
Linux-2.6.12-rc2v2.6.12-rc2
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
Diffstat (limited to 'Documentation/networking/slicecom.txt')
-rw-r--r-- | Documentation/networking/slicecom.txt | 369 |
1 files changed, 369 insertions, 0 deletions
diff --git a/Documentation/networking/slicecom.txt b/Documentation/networking/slicecom.txt new file mode 100644 index 000000000000..59cfd95121fb --- /dev/null +++ b/Documentation/networking/slicecom.txt @@ -0,0 +1,369 @@ + +SliceCOM adapter user's documentation - for the 0.51 driver version + +Written by Bartók István <bartoki@itc.hu> + +English translation: Lakatos György <gyuri@itc.hu> +Mon Dec 11 15:28:42 CET 2000 + +Last modified: Wed Aug 29 17:25:37 CEST 2001 + +----------------------------------------------------------------- + +Usage: + +Compiling the kernel: + +Code maturity level options + [*] Prompt for development and/or incomplete code/drivers + +Network device support + Wan interfaces + <M> MultiGate (COMX) synchronous + <M> Support for MUNICH based boards: SliceCOM, PCICOM (NEW) + <M> Support for HDLC and syncPPP... + + +Loading the modules: + +modprobe comx + +modprobe comx-proto-ppp # module for Cisco-HDLC and SyncPPP protocols + +modprobe comx-hw-munich # the module logs information by the kernel + # about the detected boards + + +Configuring the board: + +# This interface will use the Cisco-HDLC line protocol, +# the timeslices assigned are 1,2 (128 KiBit line speed) +# (the first data timeslice in the G.703 frame is no. 1) +# +mkdir /proc/comx/comx0.1/ +echo slicecom >/proc/comx/comx0.1/boardtype +echo hdlc >/proc/comx/comx0.1/protocol +echo 1 2 >/proc/comx/comx0.1/timeslots + + +# This interface uses SyncPPP line protocol, the assigned +# is no. 3 (64 KiBit line speed) +# +mkdir /proc/comx/comx0.2/ +echo slicecom >/proc/comx/comx0.2/boardtype +echo ppp >/proc/comx/comx0.2/protocol +echo 3 >/proc/comx/comx0.2/timeslots + +... + +ifconfig comx0.1 up +ifconfig comx0.2 up + +----------------------------------------------------------------- + +The COMX interfaces use a 10 packet transmit queue by default, however WAN +networks sometimes use bigger values (20 to 100), to utilize the line better +by large traffic (though the line delay increases because of more packets +join the queue). + +# ifconfig comx0 txqueuelen 50 + +This option is only supported by the ifconfig command of the later +distributions, which came with 2.2 kernels, such as RedHat 6.1 or Debian 2.2. + +You can download a newer netbase packet from +http://www.debian.org/~rcw/2.2/netbase/ for Debian 2.1, which has a new +ifconfig. You can get further information about using 2.2 kernel with +Debian 2.1 from http://www.debian.org/releases/stable/running-kernel-2.2 + +----------------------------------------------------------------- + +The SliceCom LEDs: + +red - on, if the interface is unconfigured, or it gets Remote Alarm-s +green - on, if the board finds frame-sync in the received signal + +A bit more detailed: + +red: green: meaning: + +- - no frame-sync, no signal received, or signal SNAFU. +- on "Everything is OK" +on on Recepion is ok, but the remote end sends Remote Alarm +on - The interface is unconfigured + +----------------------------------------------------------------- + +A more detailed description of the hardware setting options: + +The general and the protocol layer options described in the 'comx.txt' file +apply to the SliceCom as well, I only summarize the SliceCom hardware specific +settings below. + +The '/proc/comx' configuring interface: + +An interface directory should be created for every timeslot group with +'mkdir', e,g: 'comx0', 'comx1' etc. The timeslots can be assigned here to the +specific interface. The Cisco-like naming convention (serial3:1 - first +timeslot group of the 3rd. board) can't be used here, because these mean IP +aliasing in Linux. + +You can give any meaningful name to keep the configuration clear; +e.g: 'comx0.1', 'comx0.2', 'comx1.1', comx1.2', if you have two boards +with two interfaces each. + +Settings, which apply to the board: + +Neither 'io' nor 'irq' settings required, the driver uses the resources +given by the PCI BIOS. + +comx0/boardnum - board number of the SliceCom in the PC (using the 'natural' + PCI order) as listed in '/proc/pci' or the output of the + 'lspci' command, generally the slots nearer to the motherboard + PCI driver chips have the lower numbers. + + Default: 0 (the counting starts with 0) + +Though the options below are to be set on a single interface, they apply to the +whole board. The restriction, to use them on 'UP' interfaces, is because the +command sequence below could lead to unpredicable results. + + # echo 0 >boardnum + # echo internal >clock_source + # echo 1 >boardnum + +The sequence would set the clock source of board 0. + +These settings will persist after all the interfaces are cleared, but are +cleared when the driver module is unloaded and loaded again. + +comx0/clock_source - source of the transmit clock + Usage: + + # echo line >/proc/comx/comx0/clock_source + # echo internal >/proc/comx/comx0/clock_source + + line - The Tx clock is being decoded if the input data stream, + if no clock seen on the input, then the board will use it's + own clock generator. + + internal - The Tx clock is supplied by the builtin clock generator. + + Default: line + + Normally, the telecommunication company's end device (the HDSL + modem) provides the Tx clock, that's why 'line' is the default. + +comx0/framing - Switching CRC4 off/on + + CRC4: 16 PCM frames (The 32 64Kibit channels are multiplexed into a + PCM frame, nothing to do with HDLC frames) are divided into 2x8 + groups, each group has a 4 bit CRC. + + # echo crc4 >/proc/comx/comx0/framing + # echo no-crc4 >/proc/comx/comx0/framing + + Default is 'crc4', the Hungarian MATAV lines behave like this. + The traffic generally passes if this setting on both ends don't match. + +comx0/linecode - Setting the line coding + + # echo hdb3 >/proc/comx/comx0/linecode + # echo ami >/proc/comx/comx0/linecode + + Default a 'hdb3', MATAV lines use this. + + (AMI coding is rarely used with E1 lines). Frame sync may occur, if + this setting doesn't match the other end's, but CRC4 and data errors + will come, which will result in CRC errors on HDLC/SyncPPP level. + +comx0/reg - direct access to the board's MUNICH (reg) and FALC (lbireg) +comx0/lbireg circuit's registers + + # echo >reg 0x04 0x0 - write 0 to register 4 + # echo >reg 0x104 - write the contents of register 4 with + printk() to syslog + +WARNING! These are only for development purposes, messing with this will + result much trouble! + +comx0/loopback - Places a loop to the board's G.703 signals + + # echo none >/proc/comx/comx0/loopback + # echo local >/proc/comx/comx0/loopback + # echo remote >/proc/comx/comx0/loopback + + none - normal operation, no loop + local - the board receives it's own output + remote - the board sends the received data to the remote side + + Default: none + +----------------------------------------------------------------- + +Interface (channel group in Cisco terms) settings: + +comx0/timeslots - which timeslots belong to the given interface + + Setting: + + # echo '1 5 2 6 7 8' >/proc/comx/comx0/timeslots + + # cat /proc/comx/comx0/timeslots + 1 2 5 6 7 8 + # + + Finding a timeslot: + + # grep ' 4' /proc/comx/comx*/timeslots + /proc/comx/comx0/timeslots:1 3 4 5 6 + # + + The timeslots can be in any order, '1 2 3' is the same as '1 3 2'. + + The interface has to be DOWN during the setting ('ifconfig comx0 + down'), but the other interfaces could operate normally. + + The driver checks if the assigned timeslots are vacant, if not, then + the setting won't be applied. + + The timeslot values are treated as decimal numbers, not to misunderstand + values of 08, 09 form. + +----------------------------------------------------------------- + +Checking the interface and board status: + +- Lines beginning with ' ' (space) belong to the original output, the lines +which begin with '//' are the comments. + + papaya:~$ cat /proc/comx/comx1/status + Interface administrative status is UP, modem status is UP, protocol is UP + Modem status changes: 0, Transmitter status is IDLE, tbusy: 0 + Interface load (input): 978376 / 947808 / 951024 bits/s (5s/5m/15m) + (output): 978376 / 947848 / 951024 bits/s (5s/5m/15m) + Debug flags: none + RX errors: len: 22, overrun: 1, crc: 0, aborts: 0 + buffer overrun: 0, pbuffer overrun: 0 + TX errors: underrun: 0 + Line keepalive (value: 10) status UP [0] + +// The hardware specific part starts here: + Controller status: + No alarms + +// Alarm: +// +// No alarms - Everything OK +// +// LOS - Loss Of Signal - No signal sensed on the input +// AIS - Alarm Indication Signal - The remot side sends '11111111'-s, +// it tells, that there's an error condition, or it's not +// initialised. +// AUXP - Auxiliary Pattern Indication - 01010101.. received. +// LFA - Loss of Frame Alignment - no frame sync received. +// RRA - Receive Remote Alarm - the remote end's OK, but singnals error cond. +// LMFA - Loss of CRC4 Multiframe Alignment - no CRC4 multiframe sync. +// NMF - No Multiframe alignment Found after 400 msec - no such alarm using +// no-crc4 or crc4 framing, see below. +// +// Other possible error messages: +// +// Transmit Line Short - the board felt, that it's output is short-circuited, +// so it switched the transmission off. (The board can't definitely tell, +// that it's output is short-circuited.) + +// Chained list of the received packets, for debug purposes: + + Rx ring: + rafutott: 0 + lastcheck: 50845731, jiffies: 51314281 + base: 017b1858 + rx_desc_ptr: 0 + rx_desc_ptr: 017b1858 + hw_curr_ptr: 017b1858 + 06040000 017b1868 017b1898 c016ff00 + 06040000 017b1878 017b1e9c c016ff00 + 46040000 017b1888 017b24a0 c016ff00 + 06040000 017b1858 017b2aa4 c016ff00 + +// All the interfaces using the board: comx1, using the 1,2,...16 timeslots, +// comx2, using timeslot 17, etc. + + Interfaces using this board: (channel-group, interface, timeslots) + 0 comx1: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 + 1 comx2: 17 + 2 comx3: 18 + 3 comx4: 19 + 4 comx5: 20 + 5 comx6: 21 + 6 comx7: 22 + 7 comx8: 23 + 8 comx9: 24 + 9 comx10: 25 + 10 comx11: 26 + 11 comx12: 27 + 12 comx13: 28 + 13 comx14: 29 + 14 comx15: 30 + 15 comx16: 31 + +// The number of events handled by the driver during an interrupt cycle: + + Interrupt work histogram: + hist[ 0]: 0 hist[ 1]: 2 hist[ 2]: 18574 hist[ 3]: 79 + hist[ 4]: 14 hist[ 5]: 1 hist[ 6]: 0 hist[ 7]: 1 + hist[ 8]: 0 hist[ 9]: 7 + +// The number of packets to send in the Tx ring, when a new one arrived: + + Tx ring histogram: + hist[ 0]: 2329 hist[ 1]: 0 hist[ 2]: 0 hist[ 3]: 0 + +// The error counters of the E1 interface, according to the RFC2495, +// (similar to the Cisco "show controllers e1" command's output: +// http://www.cisco.com/univercd/cc/td/doc/product/software/ios11/rbook/rinterfc.htm#xtocid25669126) + +Data in current interval (91 seconds elapsed): + 9516 Line Code Violations, 65 Path Code Violations, 2 E-Bit Errors + 0 Slip Secs, 2 Fr Loss Secs, 2 Line Err Secs, 0 Degraded Mins + 0 Errored Secs, 0 Bursty Err Secs, 0 Severely Err Secs, 11 Unavail Secs +Data in Interval 1 (15 minutes): + 0 Line Code Violations, 0 Path Code Violations, 0 E-Bit Errors + 0 Slip Secs, 0 Fr Loss Secs, 0 Line Err Secs, 0 Degraded Mins + 0 Errored Secs, 0 Bursty Err Secs, 0 Severely Err Secs, 0 Unavail Secs +Data in last 4 intervals (1 hour): + 0 Line Code Violations, 0 Path Code Violations, 0 E-Bit Errors + 0 Slip Secs, 0 Fr Loss Secs, 0 Line Err Secs, 0 Degraded Mins + 0 Errored Secs, 0 Bursty Err Secs, 0 Severely Err Secs, 0 Unavail Secs +Data in last 96 intervals (24 hours): + 0 Line Code Violations, 0 Path Code Violations, 0 E-Bit Errors + 0 Slip Secs, 0 Fr Loss Secs, 0 Line Err Secs, 0 Degraded Mins + 0 Errored Secs, 0 Bursty Err Secs, 0 Severely Err Secs, 0 Unavail Secs + +----------------------------------------------------------------- + +Some unique options, (may get into the driver later): +Treat them very carefully, these can cause much trouble! + + modified CRC-4, for improved interworking of CRC-4 and non-CRC-4 + devices: (see page 107 and g706 Annex B) + lbireg[ 0x1b ] |= 0x08 + lbireg[ 0x1c ] |= 0xc0 + + - The NMF - 'No Multiframe alignment Found after 400 msec' alarm + comes into account. + + FALC - the line driver chip. + local loop - I hear my transmission back. + remote loop - I echo the remote transmission back. + + Something useful for finding errors: + + - local loop for timeslot 1 in the FALC chip: + + # echo >lbireg 0x1d 0x21 + + - Swithing the loop off: + + # echo >lbireg 0x1d 0x00 |