blob: f93462c5db25f7828a17897424604bf7d84e3ea2 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001===========================================================================
2 HVCS
3 IBM "Hypervisor Virtual Console Server" Installation Guide
4 for Linux Kernel 2.6.4+
5 Copyright (C) 2004 IBM Corporation
6
7===========================================================================
8NOTE:Eight space tabs are the optimum editor setting for reading this file.
9===========================================================================
10
11 Author(s) : Ryan S. Arnold <rsa@us.ibm.com>
12 Date Created: March, 02, 2004
13 Last Changed: August, 24, 2004
14
15---------------------------------------------------------------------------
16Table of contents:
17
18 1. Driver Introduction:
19 2. System Requirements
20 3. Build Options:
21 3.1 Built-in:
22 3.2 Module:
23 4. Installation:
24 5. Connection:
25 6. Disconnection:
26 7. Configuration:
27 8. Questions & Answers:
28 9. Reporting Bugs:
29
30---------------------------------------------------------------------------
311. Driver Introduction:
32
33This is the device driver for the IBM Hypervisor Virtual Console Server,
34"hvcs". The IBM hvcs provides a tty driver interface to allow Linux user
35space applications access to the system consoles of logically partitioned
36operating systems (Linux and AIX) running on the same partitioned Power5
37ppc64 system. Physical hardware consoles per partition are not practical
38on this hardware so system consoles are accessed by this driver using
39firmware interfaces to virtual terminal devices.
40
41---------------------------------------------------------------------------
422. System Requirements:
43
44This device driver was written using 2.6.4 Linux kernel APIs and will only
45build and run on kernels of this version or later.
46
47This driver was written to operate solely on IBM Power5 ppc64 hardware
48though some care was taken to abstract the architecture dependent firmware
49calls from the driver code.
50
51Sysfs must be mounted on the system so that the user can determine which
52major and minor numbers are associated with each vty-server. Directions
53for sysfs mounting are outside the scope of this document.
54
55---------------------------------------------------------------------------
563. Build Options:
57
58The hvcs driver registers itself as a tty driver. The tty layer
59dynamically allocates a block of major and minor numbers in a quantity
60requested by the registering driver. The hvcs driver asks the tty layer
61for 64 of these major/minor numbers by default to use for hvcs device node
62entries.
63
64If the default number of device entries is adequate then this driver can be
65built into the kernel. If not, the default can be over-ridden by inserting
66the driver as a module with insmod parameters.
67
68---------------------------------------------------------------------------
693.1 Built-in:
70
71The following menuconfig example demonstrates selecting to build this
72driver into the kernel.
73
74 Device Drivers --->
75 Character devices --->
76 <*> IBM Hypervisor Virtual Console Server Support
77
78Begin the kernel make process.
79
80---------------------------------------------------------------------------
813.2 Module:
82
83The following menuconfig example demonstrates selecting to build this
84driver as a kernel module.
85
86 Device Drivers --->
87 Character devices --->
88 <M> IBM Hypervisor Virtual Console Server Support
89
90The make process will build the following kernel modules:
91
92 hvcs.ko
93 hvcserver.ko
94
95To insert the module with the default allocation execute the following
96commands in the order they appear:
97
98 insmod hvcserver.ko
99 insmod hvcs.ko
100
101The hvcserver module contains architecture specific firmware calls and must
102be inserted first, otherwise the hvcs module will not find some of the
103symbols it expects.
104
105To override the default use an insmod parameter as follows (requesting 4
106tty devices as an example):
107
108 insmod hvcs.ko hvcs_parm_num_devs=4
109
110There is a maximum number of dev entries that can be specified on insmod.
111We think that 1024 is currently a decent maximum number of server adapters
112to allow. This can always be changed by modifying the constant in the
113source file before building.
114
115NOTE: The length of time it takes to insmod the driver seems to be related
116to the number of tty interfaces the registering driver requests.
117
118In order to remove the driver module execute the following command:
119
120 rmmod hvcs.ko
121
122The recommended method for installing hvcs as a module is to use depmod to
123build a current modules.dep file in /lib/modules/`uname -r` and then
124execute:
125
126modprobe hvcs hvcs_parm_num_devs=4
127
128The modules.dep file indicates that hvcserver.ko needs to be inserted
129before hvcs.ko and modprobe uses this file to smartly insert the modules in
130the proper order.
131
132The following modprobe command is used to remove hvcs and hvcserver in the
133proper order:
134
135modprobe -r hvcs
136
137---------------------------------------------------------------------------
1384. Installation:
139
140The tty layer creates sysfs entries which contain the major and minor
141numbers allocated for the hvcs driver. The following snippet of "tree"
142output of the sysfs directory shows where these numbers are presented:
143
144 sys/
145 |-- *other sysfs base dirs*
146 |
147 |-- class
148 | |-- *other classes of devices*
149 | |
150 | `-- tty
151 | |-- *other tty devices*
152 | |
153 | |-- hvcs0
154 | | `-- dev
155 | |-- hvcs1
156 | | `-- dev
157 | |-- hvcs2
158 | | `-- dev
159 | |-- hvcs3
160 | | `-- dev
161 | |
162 | |-- *other tty devices*
163 |
164 |-- *other sysfs base dirs*
165
166For the above examples the following output is a result of cat'ing the
167"dev" entry in the hvcs directory:
168
169 Pow5:/sys/class/tty/hvcs0/ # cat dev
170 254:0
171
172 Pow5:/sys/class/tty/hvcs1/ # cat dev
173 254:1
174
175 Pow5:/sys/class/tty/hvcs2/ # cat dev
176 254:2
177
178 Pow5:/sys/class/tty/hvcs3/ # cat dev
179 254:3
180
181The output from reading the "dev" attribute is the char device major and
182minor numbers that the tty layer has allocated for this driver's use. Most
183systems running hvcs will already have the device entries created or udev
184will do it automatically.
185
186Given the example output above, to manually create a /dev/hvcs* node entry
187mknod can be used as follows:
188
189 mknod /dev/hvcs0 c 254 0
190 mknod /dev/hvcs1 c 254 1
191 mknod /dev/hvcs2 c 254 2
192 mknod /dev/hvcs3 c 254 3
193
194Using mknod to manually create the device entries makes these device nodes
195persistent. Once created they will exist prior to the driver insmod.
196
197Attempting to connect an application to /dev/hvcs* prior to insertion of
198the hvcs module will result in an error message similar to the following:
199
200 "/dev/hvcs*: No such device".
201
202NOTE: Just because there is a device node present doesn't mean that there
203is a vty-server device configured for that node.
204
205---------------------------------------------------------------------------
2065. Connection
207
208Since this driver controls devices that provide a tty interface a user can
209interact with the device node entries using any standard tty-interactive
210method (e.g. "cat", "dd", "echo"). The intent of this driver however, is
211to provide real time console interaction with a Linux partition's console,
212which requires the use of applications that provide bi-directional,
213interactive I/O with a tty device.
214
215Applications (e.g. "minicom" and "screen") that act as terminal emulators
216or perform terminal type control sequence conversion on the data being
217passed through them are NOT acceptable for providing interactive console
218I/O. These programs often emulate antiquated terminal types (vt100 and
219ANSI) and expect inbound data to take the form of one of these supported
220terminal types but they either do not convert, or do not _adequately_
221convert, outbound data into the terminal type of the terminal which invoked
222them (though screen makes an attempt and can apparently be configured with
223much termcap wrestling.)
224
225For this reason kermit and cu are two of the recommended applications for
226interacting with a Linux console via an hvcs device. These programs simply
227act as a conduit for data transfer to and from the tty device. They do not
228require inbound data to take the form of a particular terminal type, nor do
229they cook outbound data to a particular terminal type.
230
231In order to ensure proper functioning of console applications one must make
232sure that once connected to a /dev/hvcs console that the console's $TERM
233env variable is set to the exact terminal type of the terminal emulator
234used to launch the interactive I/O application. If one is using xterm and
235kermit to connect to /dev/hvcs0 when the console prompt becomes available
236one should "export TERM=xterm" on the console. This tells ncurses
237applications that are invoked from the console that they should output
238control sequences that xterm can understand.
239
240As a precautionary measure an hvcs user should always "exit" from their
241session before disconnecting an application such as kermit from the device
242node. If this is not done, the next user to connect to the console will
243continue using the previous user's logged in session which includes
244using the $TERM variable that the previous user supplied.
245
246Hotplug add and remove of vty-server adapters affects which /dev/hvcs* node
247is used to connect to each vty-server adapter. In order to determine which
248vty-server adapter is associated with which /dev/hvcs* node a special sysfs
249attribute has been added to each vty-server sysfs entry. This entry is
250called "index" and showing it reveals an integer that refers to the
251/dev/hvcs* entry to use to connect to that device. For instance cating the
252index attribute of vty-server adapter 30000004 shows the following.
253
254 Pow5:/sys/bus/vio/drivers/hvcs/30000004 # cat index
255 2
256
257This index of '2' means that in order to connect to vty-server adapter
25830000004 the user should interact with /dev/hvcs2.
259
260It should be noted that due to the system hotplug I/O capabilities of a
261system the /dev/hvcs* entry that interacts with a particular vty-server
Matt LaPlantea2ffd272006-10-03 22:49:15 +0200262adapter is not guaranteed to remain the same across system reboots. Look
Linus Torvalds1da177e2005-04-16 15:20:36 -0700263in the Q & A section for more on this issue.
264
265---------------------------------------------------------------------------
2666. Disconnection
267
268As a security feature to prevent the delivery of stale data to an
269unintended target the Power5 system firmware disables the fetching of data
270and discards that data when a connection between a vty-server and a vty has
271been severed. As an example, when a vty-server is immediately disconnected
272from a vty following output of data to the vty the vty adapter may not have
273enough time between when it received the data interrupt and when the
274connection was severed to fetch the data from firmware before the fetch is
275disabled by firmware.
276
277When hvcs is being used to serve consoles this behavior is not a huge issue
278because the adapter stays connected for large amounts of time following
279almost all data writes. When hvcs is being used as a tty conduit to tunnel
280data between two partitions [see Q & A below] this is a huge problem
281because the standard Linux behavior when cat'ing or dd'ing data to a device
282is to open the tty, send the data, and then close the tty. If this driver
283manually terminated vty-server connections on tty close this would close
284the vty-server and vty connection before the target vty has had a chance to
285fetch the data.
286
287Additionally, disconnecting a vty-server and vty only on module removal or
288adapter removal is impractical because other vty-servers in other
289partitions may require the usage of the target vty at any time.
290
291Due to this behavioral restriction disconnection of vty-servers from the
292connected vty is a manual procedure using a write to a sysfs attribute
293outlined below, on the other hand the initial vty-server connection to a
294vty is established automatically by this driver. Manual vty-server
295connection is never required.
296
297In order to terminate the connection between a vty-server and vty the
298"vterm_state" sysfs attribute within each vty-server's sysfs entry is used.
299Reading this attribute reveals the current connection state of the
300vty-server adapter. A zero means that the vty-server is not connected to a
301vty. A one indicates that a connection is active.
302
303Writing a '0' (zero) to the vterm_state attribute will disconnect the VTERM
304connection between the vty-server and target vty ONLY if the vterm_state
305previously read '1'. The write directive is ignored if the vterm_state
306read '0' or if any value other than '0' was written to the vterm_state
307attribute. The following example will show the method used for verifying
308the vty-server connection status and disconnecting a vty-server connection.
309
310 Pow5:/sys/bus/vio/drivers/hvcs/30000004 # cat vterm_state
311 1
312
313 Pow5:/sys/bus/vio/drivers/hvcs/30000004 # echo 0 > vterm_state
314
315 Pow5:/sys/bus/vio/drivers/hvcs/30000004 # cat vterm_state
316 0
317
318All vty-server connections are automatically terminated when the device is
319hotplug removed and when the module is removed.
320
321---------------------------------------------------------------------------
3227. Configuration
323
324Each vty-server has a sysfs entry in the /sys/devices/vio directory, which
325is symlinked in several other sysfs tree directories, notably under the
326hvcs driver entry, which looks like the following example:
327
328 Pow5:/sys/bus/vio/drivers/hvcs # ls
329 . .. 30000003 30000004 rescan
330
331By design, firmware notifies the hvcs driver of vty-server lifetimes and
332partner vty removals but not the addition of partner vtys. Since an HMC
333Super Admin can add partner info dynamically we have provided the hvcs
334driver sysfs directory with the "rescan" update attribute which will query
335firmware and update the partner info for all the vty-servers that this
336driver manages. Writing a '1' to the attribute triggers the update. An
337explicit example follows:
338
339 Pow5:/sys/bus/vio/drivers/hvcs # echo 1 > rescan
340
341Reading the attribute will indicate a state of '1' or '0'. A one indicates
342that an update is in process. A zero indicates that an update has
343completed or was never executed.
344
345Vty-server entries in this directory are a 32 bit partition unique unit
346address that is created by firmware. An example vty-server sysfs entry
347looks like the following:
348
349 Pow5:/sys/bus/vio/drivers/hvcs/30000004 # ls
David Brownell0b405a02005-05-12 12:06:27 -0700350 . current_vty devspec name partner_vtys
351 .. index partner_clcs vterm_state
Linus Torvalds1da177e2005-04-16 15:20:36 -0700352
353Each entry is provided, by default with a "name" attribute. Reading the
354"name" attribute will reveal the device type as shown in the following
355example:
356
357 Pow5:/sys/bus/vio/drivers/hvcs/30000003 # cat name
358 vty-server
359
360Each entry is also provided, by default, with a "devspec" attribute which
361reveals the full device specification when read, as shown in the following
362example:
363
364 Pow5:/sys/bus/vio/drivers/hvcs/30000004 # cat devspec
365 /vdevice/vty-server@30000004
366
367Each vty-server sysfs dir is provided with two read-only attributes that
368provide lists of easily parsed partner vty data: "partner_vtys" and
369"partner_clcs".
370
371 Pow5:/sys/bus/vio/drivers/hvcs/30000004 # cat partner_vtys
372 30000000
373 30000001
374 30000002
375 30000000
376 30000000
377
378 Pow5:/sys/bus/vio/drivers/hvcs/30000004 # cat partner_clcs
379 U5112.428.103048A-V3-C0
380 U5112.428.103048A-V3-C2
381 U5112.428.103048A-V3-C3
382 U5112.428.103048A-V4-C0
383 U5112.428.103048A-V5-C0
384
385Reading partner_vtys returns a list of partner vtys. Vty unit address
386numbering is only per-partition-unique so entries will frequently repeat.
387
388Reading partner_clcs returns a list of "converged location codes" which are
389composed of a system serial number followed by "-V*", where the '*' is the
390target partition number, and "-C*", where the '*' is the slot of the
391adapter. The first vty partner corresponds to the first clc item, the
392second vty partner to the second clc item, etc.
393
394A vty-server can only be connected to a single vty at a time. The entry,
395"current_vty" prints the clc of the currently selected partner vty when
396read.
397
398The current_vty can be changed by writing a valid partner clc to the entry
399as in the following example:
400
401 Pow5:/sys/bus/vio/drivers/hvcs/30000004 # echo U5112.428.10304
402 8A-V4-C0 > current_vty
403
404Changing the current_vty when a vty-server is already connected to a vty
405does not affect the current connection. The change takes effect when the
406currently open connection is freed.
407
408Information on the "vterm_state" attribute was covered earlier on the
409chapter entitled "disconnection".
410
411---------------------------------------------------------------------------
4128. Questions & Answers:
413===========================================================================
414Q: What are the security concerns involving hvcs?
415
416A: There are three main security concerns:
417
418 1. The creator of the /dev/hvcs* nodes has the ability to restrict
419 the access of the device entries to certain users or groups. It
420 may be best to create a special hvcs group privilege for providing
421 access to system consoles.
422
423 2. To provide network security when grabbing the console it is
424 suggested that the user connect to the console hosting partition
425 using a secure method, such as SSH or sit at a hardware console.
426
427 3. Make sure to exit the user session when done with a console or
428 the next vty-server connection (which may be from another
429 partition) will experience the previously logged in session.
430
431---------------------------------------------------------------------------
432Q: How do I multiplex a console that I grab through hvcs so that other
433people can see it:
434
435A: You can use "screen" to directly connect to the /dev/hvcs* device and
436setup a session on your machine with the console group privileges. As
437pointed out earlier by default screen doesn't provide the termcap settings
438for most terminal emulators to provide adequate character conversion from
439term type "screen" to others. This means that curses based programs may
440not display properly in screen sessions.
441
442---------------------------------------------------------------------------
443Q: Why are the colors all messed up?
444Q: Why are the control characters acting strange or not working?
445Q: Why is the console output all strange and unintelligible?
446
447A: Please see the preceding section on "Connection" for a discussion of how
448applications can affect the display of character control sequences.
449Additionally, just because you logged into the console using and xterm
450doesn't mean someone else didn't log into the console with the HMC console
451(vt320) before you and leave the session logged in. The best thing to do
452is to export TERM to the terminal type of your terminal emulator when you
453get the console. Additionally make sure to "exit" the console before you
454disconnect from the console. This will ensure that the next user gets
455their own TERM type set when they login.
456
457---------------------------------------------------------------------------
458Q: When I try to CONNECT kermit to an hvcs device I get:
459"Sorry, can't open connection: /dev/hvcs*"What is happening?
460
461A: Some other Power5 console mechanism has a connection to the vty and
462isn't giving it up. You can try to force disconnect the consoles from the
463HMC by right clicking on the partition and then selecting "close terminal".
464Otherwise you have to hunt down the people who have console authority. It
465is possible that you already have the console open using another kermit
466session and just forgot about it. Please review the console options for
467Power5 systems to determine the many ways a system console can be held.
468
469OR
470
471A: Another user may not have a connectivity method currently attached to a
472/dev/hvcs device but the vterm_state may reveal that they still have the
473vty-server connection established. They need to free this using the method
474outlined in the section on "Disconnection" in order for others to connect
475to the target vty.
476
477OR
478
479A: The user profile you are using to execute kermit probably doesn't have
480permissions to use the /dev/hvcs* device.
481
482OR
483
484A: You probably haven't inserted the hvcs.ko module yet but the /dev/hvcs*
485entry still exists (on systems without udev).
486
487OR
488
489A: There is not a corresponding vty-server device that maps to an existing
490/dev/hvcs* entry.
491
492---------------------------------------------------------------------------
493Q: When I try to CONNECT kermit to an hvcs device I get:
494"Sorry, write access to UUCP lockfile directory denied."
495
496A: The /dev/hvcs* entry you have specified doesn't exist where you said it
497does? Maybe you haven't inserted the module (on systems with udev).
498
499---------------------------------------------------------------------------
500Q: If I already have one Linux partition installed can I use hvcs on said
501partition to provide the console for the install of a second Linux
502partition?
503
504A: Yes granted that your are connected to the /dev/hvcs* device using
505kermit or cu or some other program that doesn't provide terminal emulation.
506
507---------------------------------------------------------------------------
508Q: Can I connect to more than one partition's console at a time using this
509driver?
510
511A: Yes. Of course this means that there must be more than one vty-server
512configured for this partition and each must point to a disconnected vty.
513
514---------------------------------------------------------------------------
515Q: Does the hvcs driver support dynamic (hotplug) addition of devices?
516
517A: Yes, if you have dlpar and hotplug enabled for your system and it has
518been built into the kernel the hvcs drivers is configured to dynamically
519handle additions of new devices and removals of unused devices.
520
521---------------------------------------------------------------------------
522Q: For some reason /dev/hvcs* doesn't map to the same vty-server adapter
523after a reboot. What happened?
524
525A: Assignment of vty-server adapters to /dev/hvcs* entries is always done
526in the order that the adapters are exposed. Due to hotplug capabilities of
527this driver assignment of hotplug added vty-servers may be in a different
528order than how they would be exposed on module load. Rebooting or
529reloading the module after dynamic addition may result in the /dev/hvcs*
530and vty-server coupling changing if a vty-server adapter was added in a
531slot inbetween two other vty-server adapters. Refer to the section above
532on how to determine which vty-server goes with which /dev/hvcs* node.
533Hint; look at the sysfs "index" attribute for the vty-server.
534
535---------------------------------------------------------------------------
536Q: Can I use /dev/hvcs* as a conduit to another partition and use a tty
537device on that partition as the other end of the pipe?
538
539A: Yes, on Power5 platforms the hvc_console driver provides a tty interface
540for extra /dev/hvc* devices (where /dev/hvc0 is most likely the console).
541In order to get a tty conduit working between the two partitions the HMC
542Super Admin must create an additional "serial server" for the target
543partition with the HMC gui which will show up as /dev/hvc* when the target
544partition is rebooted.
545
546The HMC Super Admin then creates an additional "serial client" for the
547current partition and points this at the target partition's newly created
548"serial server" adapter (remember the slot). This shows up as an
549additional /dev/hvcs* device.
550
551Now a program on the target system can be configured to read or write to
552/dev/hvc* and another program on the current partition can be configured to
553read or write to /dev/hvcs*. Now you have a tty conduit between two
554partitions.
555
556---------------------------------------------------------------------------
5579. Reporting Bugs:
558
559The proper channel for reporting bugs is either through the Linux OS
560distribution company that provided your OS or by posting issues to the
Stephen Rothwell1d049812006-03-22 11:26:58 +1100561PowerPC development mailing list at:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700562
Stephen Rothwell1d049812006-03-22 11:26:58 +1100563linuxppc-dev@ozlabs.org
Linus Torvalds1da177e2005-04-16 15:20:36 -0700564
565This request is to provide a documented and searchable public exchange
566of the problems and solutions surrounding this driver for the benefit of
567all users.