home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
SGI Origin & Onyx2 Patches 1998 May
/
Origin and Onyx2 System Disk Patches May 1998.img
/
dist
/
patchSG0002839.idb
/
var
/
sysgen
/
system
/
irix.sm.z
/
irix.sm
Wrap
Text File
|
1998-04-01
|
19KB
|
596 lines
**************************************************************************
* *
* Copyright (C) 1986-1996 Silicon Graphics, Inc. *
* *
* These coded instructions, statements, and computer programs contain *
* unpublished proprietary information of Silicon Graphics, Inc., and *
* are protected by Federal copyright law. They may not be disclosed *
* to third parties or copied or duplicated in any form, in whole or *
* in part, without the prior written consent of Silicon Graphics, Inc. *
* *
**************************************************************************
*
* This file is used to control the configuration of the operating system.
* Its is read by the lboot(1) command which builds the /1 file. The
* syntax of this file is described in system(4). For more information on
* system configuration, see autoconfig(1), lboot(1), system(4), master(4),
* and systune(1).
*
* $Revision: 1.738 $
*
*****************************************************************************
*
* The following statements direct lboot as it generates a kernel.
* There are 2 styles of VECTOR line. The preferred style is defined
* first and in depth in the driver writers guide and the system
* file man page. Also look within this file for examples.
*
* VECTOR directives must be before all other lboot directives which
* refer to modules, such as USE, INCLUDE, and EXCLUDE.
*
* VECTOR lines must occur in the opposite order that the devices
* should be probed. Do not change the order of the VECTOR
* lines in this file, unless you understand the implications.
*
* VECTOR: bustype module adapter [ipl [vector]] [ctlr] iospace iospace2
* iospace3 probe_space exprobe_space
*
* bustype: bus the device sits on.
* module: boot module
* adapter: identifies which bus of type bustype
* ipl: VME interrupt priority level
* ctlr: identifying number for driver
* iospace: base address of card
* iospace2,3: additional addresses passed to driver edtinit
* probe_space: address and size to be read to determine existence
* of card (not specified)
* exprobe_space: extended probe which allows for reading and value
* checking and writing of values as part of the probe.
*
* The old style VECTOR line is defined as:
*
* VECTOR: module [intr] [vector ipl unit] [base] [probe [probe_size]]
* [base2] [base3]
*
* module: boot module
* intr: interrupt function (default: module_prefix|"intr")
* vector: vme interrupt vector
* ipl: (vme or local IRQx) interrupt priority level
* do not specify it unless it is needed
* base: base address of card
* probe: address read to determine existence of card
* (no probe address implies existence)
* probe_size: size of probe (default 4 bytes)
* base2, base3: additional addresses passed to driver edtinit
* routine via edt structure
*
*****************************************************************************
NOPROBE: 0
*
*
* lboot/system-file version
VERSION: 3
*
*
* Network Interfaces
*
* Primary interface is determined by network script via "primary"
* flag for "ifconfig" regardless of the order in lboot drv list.
*
USE: if_ef
USE: if_rns
*
* Madge Tokenring driver
*
VECTOR: module=if_mtr
*
* XPI FDDI board
*
*
* parallel port
*
VECTOR: module=ecpplp
*
* SCSI host adapter driver
*
EXCLUDE: wd93 wd95
VECTOR: module=ql
* Q-logic SCSI controller custom configuration
* Note: All variables must be given in decimal.
*
* These DEVICE_ADMIN directives are used to custom configure
* q-logic controllers.
*
* These directives take the form:
* DEVICE_ADMIN: hw_graph_path variable_name=variable_value
*
* The following Q-logic driver variables can be set with the
* DEVICE_ADMIN directive. NOTE, some variables can be applied to all
* targets, or to only a particular target, while others apply to
* only to the controller. In such cases, the hw_graph_path supplied
* must be that of a target and of a controller respectively.
*
* e.g. controller path
* /hw/node/xtalk/9/pci/0/scsi_ctlr/0
* or
* /hw/module/1/slot/io5/mscsi/pci/3/scsi_ctlr/0
*
* e.g. target path
* /hw/node/xtalk/9/pci/0/scsi_ctlr/0/target/3
* or
* /hw/module/1/slot/io5/mscsi/pci/3/scsi_ctlr/0/target/3
*
* applicable to controller applicable to target
* ------------------------- ---------------------
* ql_request_queue_depth ql_disconnect_allowed
* ql_response_queue_depth ql_ctq_enable
* ql_hostid ql_sync_enable
* ql_selection_timeout ql_wide_enable
* ql_disconnect_allowed ql_sync_period
* ql_ctq_enable ql_force_sync_negotiation
* ql_sync_enable
* ql_wide_enable
* ql_sync_period
* ql_sync_offset
*
* ql_request_queue_depth: default value=2048 (controller path)
* (131k ring size based on a 64 byte request block)
* You may want to increase the size of the request ring
* if you start getting messages in the SYSLOG that tell
* that you are you are out of continuation entries.
* You may want to decrease the size of the ring for a bus
* that is not being used. This will save kernel memory
* the minimum value is 256 (one page).
*
*
* examples on an Origin 2000 machine
* on board controllers
* DEVICE_ADMIN: /hw/module/1/slot/io1/baseio/pci/0/scsi_ctlr/0 ql_request_queue_depth=1024
* DEVICE_ADMIN: /hw/module/1/slot/io1/baseio/pci/1/scsi_ctlr/0 ql_request_queue_depth=4096
*
* 4 channel scsi card example (Origin 2000)
* DEVICE_ADMIN: /hw/module/1/slot/io5/mscsi/pci/0/scsi_ctlr/0 ql_request_queue_depth=4096
* DEVICE_ADMIN: /hw/module/1/slot/io5/mscsi/pci/1/scsi_ctlr/0 ql_request_queue_depth=4096
* DEVICE_ADMIN: /hw/module/1/slot/io5/mscsi/pci/2/scsi_ctlr/0 ql_request_queue_depth=4096
* DEVICE_ADMIN: /hw/module/1/slot/io5/mscsi/pci/3/scsi_ctlr/0 ql_request_queue_depth=4096
* example on an Origin 200 machine
* DEVICE_ADMIN: /hw/module/1/slot/MotherBoard/node/xtalk/8/pci/0/scsi_ctlr/0 ql_request_queue_depth=4096
* DEVICE_ADMIN: /hw/module/1/slot/MotherBoard/node/xtalk/8/pci/1/scsi_ctlr/0 ql_request_queue_depth=4096
*
*
* ql_response_queue_depth: default value=256 (16k ring size)
* (16k ring size based on a 64 byte response block)
* this in the maximum number of responses the driver can
* handle per controller completion interrupt.
* the minimum value is 256 (one page)
*
* ql_disconnect_allowed: default value=1
* by default disconnects are enabled.
* set ql_disconnect_allowed = 0 to disable disconnects
* for the controller or for a target.
*
* ql_ctq_enable: default value=1
* by default command tag queuing is enabled.
* set ql_ctq_enable=0 to disable command tag queuing for
* the controller or for a target.
*
* ql_sync_enable: default value=1
* by default command synchronous operation is enabled
* set ql_sync_enable=0 to disable synchronous operation
* for the controller or for a target.
*
* ql_sync_period: default value=12
* the following are valid ql_sync_period values
*
* value period
* 12 50ns - ULTRA SCSI
* 25 100ns - FAST
* 50 200ns - SLOW
* 100 400ns
*
* ql_sync_offset: default value=8
* valid values for the synchronous offset or 0 -> 12
* 8 has been chosen for optimal performance
*
* ql_wide_enable: default value=1
* reset ql_wide_enable to zero to disable wide(16bit) operation
* for the controller or a target.
*
* ql_force_sync_negotiation: default value=0
* some SCSI-1 devices don't set bit 4 byte 7 (sync capable) and
* consequently the QL driver will not attempt sync
* negotiation. This variable may be used to force sync negotiation
* in such cases.
*
* note that on Origin 200 machines the default value is zero for
* /hw/module/1/slot/n1/node/xtalk/8/pci/1/scsi_ctlr/0 and cannot
* be set to one.
*
* ql_hostid: default value=0
* For maximum throughput, set the host adapter id to 7.
* For maximum fairness, set it to 0.
* SCSI devices that have bus id numbers lower than the
* host adapter, and have two other devices with higher
* numbers on the bus, may not get the SCSI bus for
* extended periods of time when the two higher numbered
* devices are busy.
*
* For example, if the host adapter id number is 5, and
* there are devices at 6 and 4, devices numbered 3 and
* lower may not get the SCSI bus when * they need it.
*
* If there is a device with the same address as the host
* adapter, nothing on that SCSI bus will work.
*
* ql_selection_timeout: default value=250
*
* Allowed values for selection timeout are
* 1 sec - 1000
* 750 mS - 750
* 500 mS - 500
* 250 mS - 250
* 100 uS - 100
* 75 uS - 75
* 50 uS - 50
* 25 uS - 25
EXCLUDE: wd93
*
* SCSI modules
* These cannot be probed for so must always be manually included.
* dksc - Hard Disk and CD-ROM
* tpsc - Tape
* smfd - Floppy
* ds - Generic SCSI support (devscsi)
*
USE: cdl, scsi, dksc, tpsc, ds, scsiha
USE: smfd
*
* XLV Logical volume pseudo-device driver core
*
USE: xlv
*
* XLV Logical volume plexing support
*
USE: xlv_plex
*
* FAILOVER Dual path HA failover support
*
USE: failover
USE: cpr
*
* File systems
* efs - Local Disk filesystem
* xfs - XFS Local filesystem
* grio - guaranteed rate I/O for xfs
* xfsrt - grio support in xfs
* dmi - data management interfaces
* nfs,snfs - Network File System
* nfs3 - Network File System version 3 client
* proc - Debug (/proc) pseudo-filesystem, see proc(4)
* socket - Sockets, socket(2)
* uds - Unix Domain Socket, see Unix(7F)
* fifofs - Streams-based pipes and fifos
* pipefs - Old style, uni-directional pipes, see pipe(2)
* specfs - Character and block device interface
* fd - File descriptor (/dev/fd) pseudo-filesystem, see fd(4)
* namefs - Mountable streams, see fattach(3C)
* cfs - Nexus Cell File System
*
* The /dev/fd filesystem must be marked USE to allow for the execution
* of set-user-ID/set-group-ID interpreter scripts (see exec(2)).
*
INCLUDE: specfs, pipefs, proc, fifofs, hwgfs
USE: fd, namefs
USE: efs, socket, uds
USE: nfs, snfs, lockd, nfs3
*
* onc3 modules
*
USE: autofs, cachefs, lofs
*
* Disk quota subsystem (see quotaon(1M) command)
*
USE: quotas
USE: xfsquotas
*
* XFS related subsystems
*
USE: dmi xfsrt
*
* NOTE: xfs USE line must come *after* its 'dependents' (eg. xfsrt,dmi)
* or those dependents may not get linked in.
*
USE: xfs
*
* Hardware graph subsystem
*
USE: graph
*
* Frame Scheduler
*
USE: frs
*
* libnsl (TLI) & libsocket helper modules
*
USE: sockmod, timod, tirdwr,
*
* TPI loopback device drivers
*
USE: ticlts, ticots, ticotsord
*
* TPI Streams device drivers for sockets networking
*
USE: tpisocket, tpitcp, tpiudp, tpirawip, tpiicmp
*
* SVR4 tty compatibility modules
*
USE: ldterm, ptem
*
* Streams based pipe support
*
USE: pipemod
USE: connld
*
* Streams event logger (see strerr(1M))
*
USE: log
*
* Streams administrative device
*
USE: sad
*
* Internationalization Support (see kbdset(1))
*
* alp - Algorithm Pool
* kbd - Keyboard mapping and string translation
*
USE: kbd, alp
*
* IP Filter (see ipfilterd(1M))
*
USE: ipfilter
*
* IP Multicast routing (see mrouted(1M))
*
USE: ip_mroute
*
* Serial Line IP (see slip(1M))
*
USE: if_sl
*
* Point to Point Protocol (see ppp(1M))
*
USE: if_ppp
*
* swIPe Protocol
*
USE: if_sw
*
* Filesystem Control (see fsr(1M))
*
USE: fsctl
*
* Inode Monitor (see famd(1M))
*
USE: imon
*
* Security Audit Trail (see satconfig(1M))
* USE is correct for sat
* There used to be an "if defined(sat-enabled)"
*
USE: sat
*
* XFS attribute unlocked fetching
*
* Traditional SuperUser
USE: suser
*
* process extended attributes.
USE: eag
*
* Kernel Asyncio for Informix
*
USE: kaio
*
* Post Wait driver
* The 'postwait_enabled' allows easy inclusion via klocaldefs
*
USE: postwait
*
* kernel listio (Oracle Performance)
* The 'klistio_enabled' allows easy inclusion via klocaldefs
*
USE: klistio
*
* Stubs
*
USE: dnstubs, gfxstubs, prfstubs, quotasstubs, idbgstubs, nfsstubs
USE: pipefsstubs, vidstubs, ng1stubs, gr2stubs, frsstubs, griostubs
USE: efsstubs, xfsstubs, xfsquotasstubs, xlvstubs, tpisocketstubs, bpqueuestubs, dmistubs
USE: macstubs, aclstubs, capstubs, satstubs, sesmgrstubs
USE: xlvplexstubs
USE: failoverstubs
USE: cachefsstubs, svckudpstubs, nlmstubs, lockdstubs, mgrasstubs, crimestubs
USE: xfsrtstubs, frustubs, cachefsidbgstubs, kaiostubs, devicestubs, mgvstubs
USE: lnshmstubs, dfsstubs, uncstubs
USE: partitionstubs
*
* Spin locks - both kernel and user level
* Kernel spinlocks used only on MultiProcessor systems
*
INCLUDE: hardlocks
*
* Tie a VME level to a processor by adding a line of the form
* "IPL: <level> <cpuid>".
* Interrupts from VME levels that are not specifically bound
* to a processor, interrupt from VME levels that are bound to
* non-existant or disabled processors, and all non-VME interrupts
* are assigned CPUs at boot time, using a "spray" heuristic.
* Example: In order to send all VME level 3 interrupts to CPU1, use:
* IPL: 3 1
*
* Declare a CPU that is not suitable for "sprayed" interrupts
* (usually, because they will be used for real-time activities) using
* the NOINTR directive.
* Example: to declare that cpu 3 should not accept sprayed
* interrupts, use: "NOINTR: 3". You CAN tie a VME interrupt
* to a processor that accepts no sprayed interrupts using the IPL
* directive, above. You may NOT restrict CPU 0 from receiving
* interrupts. You can specify multiple cpus on the NOINTR line.
*
* One can also use the NOINTR device administration directive to
* disable interrupts on a particular CPU. For example, on Origin:
* DEVICE_ADMIN: /hw/module/1/slot/n3/node/cpu/a NOINTR=1
*
* One can use the INTR_TARGET device administration directive to
* direct interrupts from a device to a particular cpu in the system.
* For example, on Origin:
* DEVICE_ADMIN: /hw//hw/module/1/slot/n3/node/xtalk/8/pci/1 INTR_TARGET=/hw/cpunum/2
* DEVICE_ADMIN: /hw/module/1/slot/n3/node/xtalk/8/pci/1 INTR_TARGET=/hw/module/1/slot/n4/node/cpu/b
*
* One can use the INTR_SWLEVEL device administration directive to
* specify the thread priority level at which the ithread for
* handling the device interrupts is supposed to run.
* For example, on Origin:
* DEVICE_ADMIN: /hw/module/1/slot/n3/node/xtalk/8/pci/1 INTR_SWLEVEL=231
*
* Priority IO support library
*
INCLUDE: prio
*
*
* Required kernel modules
*
* kernel - machine specific
* os - system calls, vm, etc..
* disp - Process scheduler
* numa - Numa Memory Management
* ksync - kernel synchronization routines (mutex_lock, sv_wait, psema...)
* or
* ksync_metered - metered kernel synchronization routines
* mem - /dev/mem, /dev/null, /dev/mmem, /dev/kmem
* zero - /dev/zero
* ipc - support for sem, shm, and msg
* sem - System V semaphores (see semop(2))
* msg - Interprocess messages (see msgop(2))
* klog - Kernel message logger (see klog(7))
* usema - User semaphores (see usema(7))
* usync - usync module
* psema - posix named semaphore module
*
KERNEL: kernel
INCLUDE: numa, os, disp, mem, zero
INCLUDE: ksync
EXCLUDE: ksync_metered
USE: ipc, msg, sem
USE: psema, usync
USE: klog
USE: usema
*
* General I/O modules
*
* io - VME interrupt handler
* gentty - Generic tty device (/dev/tty)
* ptc - Pseudo terminals
* sduart - Serial port driver
* cn - Virtual console driver
*
INCLUDE: io, gentty, ptc, sduart, cn, stream
*
* include standard serialio upper modules
*
INCLUDE: cserialio
INCLUDE: sio_i2c i2c elsc
*
* Keyboard/mouse driver
*
INCLUDE: ioc3_pckm
*
* Bus Registrations
*
*
* PCI devices
*
INCLUDE: ioc3
INCLUDE: ioc3ei
*
* PCI infrastructure
*
INCLUDE: pciio
INCLUDE: usrpci
*
* VMEbus service provider
*
USE: xiovme vmeio universe usrvme ude
*
* XTALK devices
*
INCLUDE: menet
INCLUDE: mscsi
INCLUDE: io6
INCLUDE: mio
INCLUDE: kitchen_sync
INCLUDE: pcibr
INCLUDE: pciiox
*
* XTALK infrastructure
*
INCLUDE: xbmon
INCLUDE: xbow
INCLUDE: xswitch
INCLUDE: xtalk
*
* Hub special space interface driver: Specific to 1 only.
*
USE: hubspc
USE: sn0drv
USE: fru
EXCLUDE: partition
*
* Number-in-a-can kernel interface
*
INCLUDE: nic
*
* Device specifications
*
* The ROOTDEV specification can be customized with the format:
* /dev/dsk/dks?d?s?
*
ROOTDEV: /dev/root
SWAPDEV: /dev/swap 0 0
DUMPDEV: /dev/swap
* Kernel debugging tools (see profiler(1M) and idbg(1M))
*
* for most users, only the first line below needs to be changed from
* EXCLUDE to INCLUDE
* when configuring the kernel debugger.
EXCLUDE: idbg, disp_idbg, mloadidbg, bsdidbg, graphidbg
EXCLUDE: klidbg
EXCLUDE: vme_idbg
EXCLUDE: dmiidbg, grioidbg, xfsidbg, xlvidbg
EXCLUDE: asidbg, dp_idbg, prioidbg, cachefsidbg
USE: prf
*
* Compilation and load flags
* To load a kernel that can be co-resident with symmon
* (for breakpoint debugging) replace LDOPTS
* with the following. You must also install eoe.sw.kdebug
* and eoe.s
