1 This driver is for Compaq's SMART Array Controllers.
6 This driver is known to work with the following cards:
15 * SA 6400 U320 Expansion Module
20 If nodes are not already created in the /dev/cciss directory
24 Where ctlrs is the number of controllers you have (defaults to 1 if not
30 You need some entries in /dev for the cciss device. The mkdev.cciss script
31 can make device nodes for you automatically. Currently the device setup
41 b7 b6 b5 b4 b3 b2 b1 b0
42 |----+----| |----+----|
44 | +-------- Partition ID (0=wholedev, 1-15 partition)
46 +-------------------- Logical Volume number
48 The suggested device naming scheme is:
49 /dev/cciss/c0d0 Controller 0, disk 0, whole device
50 /dev/cciss/c0d0p1 Controller 0, disk 0, partition 1
51 /dev/cciss/c0d0p2 Controller 0, disk 0, partition 2
52 /dev/cciss/c0d0p3 Controller 0, disk 0, partition 3
54 /dev/cciss/c1d1 Controller 1, disk 1, whole device
55 /dev/cciss/c1d1p1 Controller 1, disk 1, partition 1
56 /dev/cciss/c1d1p2 Controller 1, disk 1, partition 2
57 /dev/cciss/c1d1p3 Controller 1, disk 1, partition 3
59 Support for more than 8 controllers
60 -----------------------------------
61 Originally the driver only supports 8 controllers in the system,
62 and this is due to the major numbers assigned to the driver
65 The driver can now support up to 32 controllers in the system.
67 For the ninth controller and beyond, the major numbers will be
68 assigned dynamically by the kernel when it is discovered,
69 and there is no guarantee what the major number you will get,
70 but most likely it will start from 254 and goes down from there.
72 You can check the assigned major numbers by typing
74 And look for cciss controllers
76 Once you have this, you need to create device nodes in
77 /dev/cciss directory. The nodes for the first 8 controllers
78 should already be created by mkdev.cciss script or
79 /etc/makedev.d/cciss script. You can add the major number(s)
80 in those scripts, or create the nodes manually by using
83 You can also use mknod_dyn.cciss and rmnod_dyn.cciss scripts
84 to create or remove nodes easily. These scripts can be found
85 in the Documentation directory.
87 Then you can mount the devices and create partitions
88 (You also need to make nodes for these partitions).
90 As for the minor number, the disk device will have a minor
91 number divisible by 16 (e.g: 0, 16, 32 etc), and the
92 partitions on those disk devices will have the minor number
93 of the disk device plus the partition number (1-15).
94 For example, disk d2 will have minor number 32, and its
95 partitions 1 and 2 will have minor numbers 33 and 34.
97 Look at the mkdev.cciss script for example.
100 In 2.4 kernel, partition names are hard coded in
101 /usr/src/linux/fs/partitions/check.c
102 and only for the first 8 cciss controllers. The rest
103 will be reported as ccissXX. This should not affect
104 I/O operation or performance. Please apply the
105 cciss_2.4_partition_name.patch to address this. This
106 will not be an issue under 2.5 kernel, since partition
107 names will be handled by the driver.
109 SCSI tape drive and medium changer support
110 ------------------------------------------
112 SCSI sequential access devices and medium changer devices are supported and
113 appropriate device nodes are automatically created. (e.g.
114 /dev/st0, /dev/st1, etc. See the "st" man page for more details.)
115 You must enable "SCSI tape drive support for Smart Array 5xxx" and
116 "SCSI support" in your kernel configuration to be able to use SCSI
117 tape drives with your Smart Array 5xxx controller.
119 Additionally, note that the driver will not engage the SCSI core at init
120 time. The driver must be directed to dynamically engage the SCSI core via
121 the /proc filesystem entry which the "block" side of the driver creates as
122 /proc/driver/cciss/cciss* at runtime. This is because at driver init time,
123 the SCSI core may not yet be initialized (because the driver is a block
124 driver) and attempting to register it with the SCSI core in such a case
125 would cause a hang. This is best done via an initialization script
126 (typically in /etc/init.d, but could vary depending on distibution).
129 for x in /proc/driver/cciss/cciss[0-9]*
131 echo "engage scsi" > $x
134 Once the SCSI core is engaged by the driver, it cannot be disengaged
135 (except by unloading the driver, if it happens to be linked as a module.)
137 Note also that if no sequential access devices or medium changers are
138 detected, the SCSI core will not be engaged by the action of the above
141 Hot plug support for SCSI tape drives
142 -------------------------------------
144 Hot plugging of SCSI tape drives is supported, with some caveats.
145 The cciss driver must be informed that changes to the SCSI bus
146 have been made, in addition to and prior to informing the SCSI
147 mid layer. This may be done via the /proc filesystem. For example:
149 echo "rescan" > /proc/scsi/cciss0/1
151 This causes the adapter to query the adapter about changes to the
152 physical SCSI buses and/or fibre channel arbitrated loop and the
153 driver to make note of any new or removed sequential access devices
154 or medium changers. The driver will output messages indicating what
155 devices have been added or removed and the controller, bus, target and
156 lun used to address the device. Once this is done, the SCSI mid layer
157 can be informed of changes to the virtual SCSI bus which the driver
158 presents to it in the usual way. For example:
160 echo scsi add-single-device 3 2 1 0 > /proc/scsi/scsi
162 to add a device on controller 3, bus 2, target 1, lun 0. Note that
163 the driver makes an effort to preserve the devices positions
164 in the virtual SCSI bus, so if you are only moving tape drives
165 around on the same adapter and not adding or removing tape drives
166 from the adapter, informing the SCSI mid layer may not be necessary.
168 Note that the naming convention of the /proc filesystem entries
169 contains a number in addition to the driver name. (E.g. "cciss0"
170 instead of just "cciss" which you might expect.) This is because
171 of changes to the 2.4 kernel PCI interface related to PCI hot plug
172 that imply the driver must register with the SCSI mid layer once per
173 adapter instance rather than once per driver.
175 Note: ONLY sequential access devices and medium changers are presented
176 as SCSI devices to the SCSI mid layer by the cciss driver. Specifically,
177 physical SCSI disk drives are NOT presented to the SCSI mid layer. The
178 physical SCSI disk drives are controlled directly by the array controller
179 hardware and it is important to prevent the OS from attempting to directly
180 access these devices too, as if the array controller were merely a SCSI
181 controller in the same way that we are allowing it to access SCSI tape drives.
186 For multipath configurations (acheived via a higher level driver, such
187 as the "md" driver) it is important that failure of a controller is detected.
188 Ordinarily, the driver is entirely interrupt driven. If a failure occurs
189 in such a way that the processor cannot receive interrupts from an adapter,
190 the driver will wait forever for i/o's to complete. In a multipath
191 configuration this is undesirable, as the md driver relies on i/o's being
192 reported as failed by the low level driver to trigger failing over to an
193 alternate controller. The monitor threads allow the driver to detect such
194 situations and report outstanding i/o's as having failed so that recovery
195 actions such switching to an alternate controller can occur. The monitor
196 threads periodically sends a trivial "no-operation" command down to
197 the controllers and expect them to complete within a a reasonable (short)
198 time period. The firmware on the adapter is designed such that no matter
199 how busy the adapter is serving i/o, it can respond quickly to a
200 "no-operation" command. In the event that a deadline elapses before a no
201 operation command completes, all outstanding commands on that controller
202 are reported back to the upper layers as having failed, and any new commands
203 sent to the controller are immediately reported back as failed.
205 To enable the monitor threads, the compile time option must be enabled
206 (via the usual linux kernel configuration) and the monitor thread must
207 be enabled at runtime as well. A system may have many adapters, but
208 perhaps only a single pair operating in a multipath configuration.
209 In this way, it is possible to run monitoring threads only for those
210 adapters which require it.
212 To start a monitoring thread on the first cciss adapter, "cciss0" with
213 a polling interval of 30 seconds, execute the following command:
215 echo "monitor 30" > /proc/driver/cciss/cciss0
217 To change the polling interval, to say, 60 seconds:
219 echo "monitor 60" > /proc/driver/cciss/cciss0
221 (Note, the change will not take effect until the previous polling
224 To disable the monitoring thread, set the polling interval to 0 seconds:
226 echo "monitor 0" > /proc/driver/cciss/cciss0
228 (Again, the monitoring thread will not exit until the previous polling
231 The minimum monitoring period is 10 seconds, and the maximum monitoring
232 period is 3600 seconds (1 hour). The no-operation command must complete
233 with 5 seconds of submission in all cases or the controller will be presumed