Root/Documentation/scsi/aic7xxx_old.txt

1                AIC7xxx Driver for Linux
2
3Introduction
4----------------------------
5The AIC7xxx SCSI driver adds support for Adaptec (http://www.adaptec.com)
6SCSI controllers and chipsets. Major portions of the driver and driver
7development are shared between both Linux and FreeBSD. Support for the
8AIC-7xxx chipsets have been in the default Linux kernel since approximately
9linux-1.1.x and fairly stable since linux-1.2.x, and are also in FreeBSD
102.1.0 or later.
11
12  Supported cards/chipsets
13  ----------------------------
14    Adaptec Cards
15    ----------------------------
16    AHA-274x
17    AHA-274xT
18    AHA-2842
19    AHA-2910B
20    AHA-2920C
21    AHA-2930
22    AHA-2930U
23    AHA-2930CU
24    AHA-2930U2
25    AHA-2940
26    AHA-2940W
27    AHA-2940U
28    AHA-2940UW
29    AHA-2940UW-PRO
30    AHA-2940AU
31    AHA-2940U2W
32    AHA-2940U2
33    AHA-2940U2B
34    AHA-2940U2BOEM
35    AHA-2944D
36    AHA-2944WD
37    AHA-2944UD
38    AHA-2944UWD
39    AHA-2950U2
40    AHA-2950U2W
41    AHA-2950U2B
42    AHA-29160M
43    AHA-3940
44    AHA-3940U
45    AHA-3940W
46    AHA-3940UW
47    AHA-3940AUW
48    AHA-3940U2W
49    AHA-3950U2B
50    AHA-3950U2D
51    AHA-3960D
52    AHA-39160M
53    AHA-3985
54    AHA-3985U
55    AHA-3985W
56    AHA-3985UW
57
58    Motherboard Chipsets
59    ----------------------------
60    AIC-777x
61    AIC-785x
62    AIC-786x
63    AIC-787x
64    AIC-788x
65    AIC-789x
66    AIC-3860
67
68    Bus Types
69    ----------------------------
70    W - Wide SCSI, SCSI-3, 16bit bus, 68pin connector, will also support
71        SCSI-1/SCSI-2 50pin devices, transfer rates up to 20MB/s.
72    U - Ultra SCSI, transfer rates up to 40MB/s.
73    U2- Ultra 2 SCSI, transfer rates up to 80MB/s.
74    D - Differential SCSI.
75    T - Twin Channel SCSI. Up to 14 SCSI devices.
76
77    AHA-274x - EISA SCSI controller
78    AHA-284x - VLB SCSI controller
79    AHA-29xx - PCI SCSI controller
80    AHA-394x - PCI controllers with two separate SCSI controllers on-board.
81    AHA-398x - PCI RAID controllers with three separate SCSI controllers
82               on-board.
83
84  Not Supported Devices
85  ------------------------------
86    Adaptec Cards
87    ----------------------------
88    AHA-2920 (Only the cards that use the Future Domain chipset are not
89              supported, any 2920 cards based on Adaptec AIC chipsets,
90          such as the 2920C, are supported)
91    AAA-13x Raid Adapters
92    AAA-113x Raid Port Card
93
94    Motherboard Chipsets
95    ----------------------------
96    AIC-7810
97
98    Bus Types
99    ----------------------------
100    R - Raid Port busses are not supported.
101
102    The hardware RAID devices sold by Adaptec are *NOT* supported by this
103    driver (and will people please stop emailing me about them, they are
104    a totally separate beast from the bare SCSI controllers and this driver
105    cannot be retrofitted in any sane manner to support the hardware RAID
106    features on those cards - Doug Ledford).
107    
108
109  People
110  ------------------------------
111    Justin T Gibbs gibbs@plutotech.com
112      (BSD Driver Author)
113    Dan Eischen deischen@iworks.InterWorks.org
114      (Original Linux Driver Co-maintainer)
115    Dean Gehnert deang@teleport.com
116      (Original Linux FTP/patch maintainer)
117    Jess Johnson jester@frenzy.com
118      (AIC7xxx FAQ author)
119    Doug Ledford dledford@redhat.com
120      (Current Linux aic7xxx-5.x.x Driver/Patch/FTP maintainer)
121    
122    Special thanks go to John Aycock (aycock@cpsc.ucalgary.ca), the original
123    author of the driver. John has since retired from the project. Thanks
124    again for all his work!
125    
126  Mailing list
127  ------------------------------
128    There is a mailing list available for users who want to track development
129    and converse with other users and developers. This list is for both
130    FreeBSD and Linux support of the AIC7xxx chipsets.
131
132    To subscribe to the AIC7xxx mailing list send mail to the list server,
133    with "subscribe AIC7xxx" in the body (no Subject: required):
134        To: majordomo@FreeBSD.ORG
135        ---
136        subscribe AIC7xxx
137
138    To unsubscribe from the list, send mail to the list server with:
139        To: majordomo@FreeBSD.ORG
140        ---
141        unsubscribe AIC7xxx
142
143    Send regular messages and replies to: AIC7xxx@FreeBSD.ORG
144    
145  Boot Command line options
146  ------------------------------
147    "aic7xxx=no_reset" - Eliminate the SCSI bus reset during startup.
148        Some SCSI devices need the initial reset that this option disables
149    in order to work. If you have problems at bootup, please make sure
150    you aren't using this option.
151    
152    "aic7xxx=reverse_scan" - Certain PCI motherboards scan for devices at
153        bootup by scanning from the highest numbered PCI device to the
154    lowest numbered PCI device, others do just the opposite and scan
155    from lowest to highest numbered PCI device. There is no reliable
156    way to autodetect this ordering. So, we default to the most common
157    order, which is lowest to highest. Then, in case your motherboard
158    scans from highest to lowest, we have this option. If your BIOS
159    finds the drives on controller A before controller B but the linux
160    kernel finds your drives on controller B before A, then you should
161    use this option.
162    
163    "aic7xxx=extended" - Force the driver to detect extended drive translation
164        on your controller. This helps those people who have cards without
165        a SEEPROM make sure that linux and all other operating systems think
166        the same way about your hard drives.
167
168    "aic7xxx=scbram" - Some cards have external SCB RAM that can be used to
169        give the card more hardware SCB slots. This allows the driver to use
170    that SCB RAM. Without this option, the driver won't touch the SCB
171    RAM because it is known to cause problems on a few cards out there
172    (such as 3985 class cards).
173    
174    "aic7xxx=irq_trigger:x" - Replace x with either 0 or 1 to force the kernel
175        to use the correct IRQ type for your card. This only applies to EISA
176        based controllers. On these controllers, 0 is for Edge triggered
177        interrupts, and 1 is for Level triggered interrupts. If you aren't
178        sure or don't know which IRQ trigger type your EISA card uses, then
179        let the kernel autodetect the trigger type.
180    
181    "aic7xxx=verbose" - This option can be used in one of two ways. If you
182        simply specify aic7xxx=verbose, then the kernel will automatically
183    pick the default set of verbose messages for you to see.
184    Alternatively, you can specify the command as
185    "aic7xxx=verbose:0xXXXX" where the X entries are replaced with
186    hexadecimal digits. This option is a bit field type option. For
187    a full listing of the available options, search for the
188    #define VERBOSE_xxxxxx lines in the aic7xxx.c file. If you want
189    verbose messages, then it is recommended that you simply use the
190    aic7xxx=verbose variant of this command.
191    
192    "aic7xxx=pci_parity:x" - This option controls whether or not the driver
193        enables PCI parity error checking on the PCI bus. By default, this
194        checking is disabled. To enable the checks, simply specify pci_parity
195        with no value afterwords. To reverse the parity from even to odd,
196        supply any number other than 0 or 255. In short:
197          pci_parity - Even parity checking (even is the normal PCI parity)
198          pci_parity:x - Where x > 0, Odd parity checking
199          pci_parity:0 - No check (default)
200        NOTE: In order to get Even PCI parity checking, you must use the
201        version of the option that does not include the : and a number at
202        the end (unless you want to enter exactly 2^32 - 1 as the number).
203    
204    "aic7xxx=no_probe" - This option will disable the probing for any VLB
205        based 2842 controllers and any EISA based controllers. This is
206    needed on certain newer motherboards where the normal EISA I/O ranges
207    have been claimed by other PCI devices. Probing on those machines
208    will often result in the machine crashing or spontaneously rebooting
209    during startup. Examples of machines that need this are the
210    Dell PowerEdge 6300 machines.
211
212    "aic7xxx=seltime:2" - This option controls how long the card waits
213        during a device selection sequence for the device to respond.
214    The original SCSI spec says that this "should be" 256ms. This
215    is generally not required with modern devices. However, some
216    very old SCSI I devices need the full 256ms. Most modern devices
217    can run fine with only 64ms. The default for this option is
218    64ms. If you need to change this option, then use the following
219    table to set the proper value in the example above:
220      0 - 256ms
221      1 - 128ms
222      2 - 64ms
223      3 - 32ms
224    
225    "aic7xxx=panic_on_abort" - This option is for debugging and will cause
226        the driver to panic the linux kernel and freeze the system the first
227    time the drivers abort or reset routines are called. This is most
228    helpful when some problem causes infinite reset loops that scroll too
229    fast to see. By using this option, you can write down what the errors
230    actually are and send that information to me so it can be fixed.
231    
232    "aic7xxx=dump_card" - This option will print out the *entire* set of
233        configuration registers on the card during the init sequence. This
234    is a debugging aid used to see exactly what state the card is in
235    when we finally finish our initialization routines. If you don't
236    have documentation on the chipsets, this will do you absolutely
237    no good unless you are simply trying to write all the information
238    down in order to send it to me.
239    
240    "aic7xxx=dump_sequencer" - This is the same as the above options except
241        that instead of dumping the register contents on the card, this
242    option dumps the contents of the sequencer program RAM. This gives
243    the ability to verify that the instructions downloaded to the
244    card's sequencer are indeed what they are supposed to be. Again,
245    unless you have documentation to tell you how to interpret these
246    numbers, then it is totally useless.
247    
248    "aic7xxx=override_term:0xffffffff" - This option is used to force the
249        termination on your SCSI controllers to a particular setting. This
250    is a bit mask variable that applies for up to 8 aic7xxx SCSI channels.
251    Each channel gets 4 bits, divided as follows:
252    bit 3 2 1 0
253          | | | Enable/Disable Single Ended Low Byte Termination
254          | | En/Disable Single Ended High Byte Termination
255          | En/Disable Low Byte LVD Termination
256          En/Disable High Byte LVD Termination
257
258    The upper 2 bits that deal with LVD termination only apply to Ultra2
259    controllers. Furthermore, due to the current Ultra2 controller
260    designs, these bits are tied together such that setting either bit
261    enables both low and high byte LVD termination. It is not possible
262    to only set high or low byte LVD termination in this manner. This is
263    an artifact of the BIOS definition on Ultra2 controllers. For other
264    controllers, the only important bits are the two lowest bits. Setting
265    the higher bits on non-Ultra2 controllers has no effect. A few
266    examples of how to use this option:
267
268    Enable low and high byte termination on a non-ultra2 controller that
269    is the first aic7xxx controller (the correct bits are 0011),
270    aic7xxx=override_term:0x3
271
272    Enable all termination on the third aic7xxx controller, high byte
273    termination on the second aic7xxx controller, and low and high byte
274    SE termination on the first aic7xxx controller
275    (bits are 1111 0010 0011),
276    aic7xxx=override_term:0xf23
277    
278    No attempt has been made to make this option non-cryptic. It really
279    shouldn't be used except in dire circumstances, and if that happens,
280    I'm probably going to be telling you what to set this to anyway :)
281
282    "aic7xxx=stpwlev:0xffffffff" - This option is used to control the STPWLEV
283        bit in the DEVCONFIG PCI register. Currently, this is one of the
284    very few registers that we have absolutely *no* way of detecting
285    what the variable should be. It depends entirely on how the chipset
286    and external terminators were coupled by the card/motherboard maker.
287    Further, a chip reset (at power up) always sets this bit to 0. If
288    there is no BIOS to run on the chipset/card (such as with a 2910C
289    or a motherboard controller with the BIOS totally disabled) then
290    the variable may not get set properly. Of course, if the proper
291    setting was 0, then that's what it would be after the reset, but if
292    the proper setting is actually 1.....you get the picture. Now, since
293    we can't detect this at all, I've added this option to force the
294    setting. If you have a BIOS on your controller then you should never
295    need to use this option. However, if you are having lots of SCSI
296    reset problems and can't seem to get them knocked out, this may help.
297
298    Here's a test to know for certain if you need this option. Make
299    a boot floppy that you can use to boot your computer up and that
300    will detect the aic7xxx controller. Next, power down your computer.
301    While it's down, unplug all SCSI cables from your Adaptec SCSI
302    controller. Boot the system back up to the Adaptec EZ-SCSI BIOS
303    and then make sure that termination is enabled on your adapter (if
304    you have an Adaptec BIOS of course). Next, boot up the floppy you
305    made and wait for it to detect the aic7xxx controller. If the kernel
306    finds the controller fine, says scsi : x hosts and then tries to
307    detect your devices like normal, up to the point where it fails to
308    mount your root file system and panics, then you're fine. If, on
309    the other hand, the system goes into an infinite reset loop, then
310    you need to use this option and/or the previous option to force the
311    proper termination settings on your controller. If this happens,
312    then you next need to figure out what your settings should be.
313
314    To find the correct settings, power your machine back down, connect
315    back up the SCSI cables, and boot back into your machine like normal.
316    However, boot with the aic7xxx=verbose:0x39 option. Record the
317    initial DEVCONFIG values for each of your aic7xxx controllers as
318    they are listed, and also record what the machine is detecting as
319    the proper termination on your controllers. NOTE: the order in
320    which the initial DEVCONFIG values are printed out is not guaranteed
321    to be the same order as the SCSI controllers are registered. The
322    above option and this option both work on the order of the SCSI
323    controllers as they are registered, so make sure you match the right
324    DEVCONFIG values with the right controllers if you have more than
325    one aic7xxx controller.
326
327    Once you have the detected termination settings and the initial
328    DEVCONFIG values for each controller, then figure out what the
329    termination on each of the controllers *should* be. Hopefully, that
330    part is correct, but it could possibly be wrong if there is
331    bogus cable detection logic on your controller or something similar.
332    If all the controllers have the correct termination settings, then
333    don't set the aic7xxx=override_term variable at all, leave it alone.
334    Next, on any controllers that go into an infinite reset loop when
335    you unplug all the SCSI cables, get the starting DEVCONFIG value.
336    If the initial DEVCONFIG value is divisible by 2, then the correct
337    setting for that controller is 0. If it's an odd number, then
338    the correct setting for that controller is 1. For any other
339    controllers that didn't have an infinite reset problem, then reverse
340    the above options. If DEVCONFIG was even, then the correct setting
341    is 1, if not then the correct setting is 0.
342
343    Now that you know what the correct setting was for each controller,
344    we need to encode that into the aic7xxx=stpwlev:0x... variable.
345    This variable is a bit field encoded variable. Bit 0 is for the first
346    aic7xxx controller, bit 1 for the next, etc. Put all these bits
347    together and you get a number. For example, if the third aic7xxx
348    needed a 1, but the second and first both needed a 0, then the bits
349    would be 100 in binary. This then translates to 0x04. You would
350    therefore set aic7xxx=stpwlev:0x04. This is fairly standard binary
351    to hexadecimal conversions here. If you aren't up to speed on the
352    binary->hex conversion then send an email to the aic7xxx mailing
353    list and someone can help you out.
354
355    "aic7xxx=tag_info:{{8,8..},{8,8..},..}" - This option is used to disable
356        or enable Tagged Command Queueing (TCQ) on specific devices. As of
357    driver version 5.1.11, TCQ is now either on or off by default
358    according to the setting you choose during the make config process.
359    In order to en/disable TCQ for certain devices at boot time, a user
360    may use this boot param. The driver will then parse this message out
361        and en/disable the specific device entries that are present based upon
362        the value given. The param line is parsed in the following manner:
363
364          { - first instance indicates the start of this parameter values
365              second instance is the start of entries for a particular
366              device entry
367          } - end the entries for a particular host adapter, or end the entire
368              set of parameter entries
369          , - move to next entry. Inside of a set of device entries, this
370              moves us to the next device on the list. Outside of device
371              entries, this moves us to the next host adapter
372          . - Same effect as , but is safe to use with insmod.
373          x - the number to enter into the array at this position.
374              0 = Enable tagged queueing on this device and use the default
375                  queue depth
376              1-254 = Enable tagged queueing on this device and use this
377                      number as the queue depth
378              255 = Disable tagged queueing on this device.
379              Note: anything above 32 for an actual queue depth is wasteful
380                    and not recommended.
381
382        A few examples of how this can be used:
383
384        tag_info:{{8,12,,0,,255,4}}
385          This line will only effect the first aic7xxx card registered. It
386          will set scsi id 0 to a queue depth of 8, id 1 to 12, leave id 2
387          at the default, set id 3 to tagged queueing enabled and use the
388          default queue depth, id 4 default, id 5 disabled, and id 6 to 4.
389          Any not specified entries stay at the default value, repeated
390          commas with no value specified will simply increment to the next id
391          without changing anything for the missing values.
392
393        tag_info:{,,,{,,,255}}
394          First, second, and third adapters at default values. Fourth
395          adapter, id 3 is disabled. Notice that leading commas simply
396      increment what the first number effects, and there are no need
397      for trailing commas. When you close out an adapter, or the
398      entire entry, anything not explicitly set stays at the default
399      value.
400
401        A final note on this option. The scanner I used for this isn't
402        perfect or highly robust. If you mess the line up, the worst that
403        should happen is that the line will get ignored. If you don't
404        close out the entire entry with the final bracket, then any other
405        aic7xxx options after this will get ignored. So, in general, be
406        sure of what you are entering, and after you have it right, just
407        add it to the lilo.conf file so there won't be any mistakes. As
408        a means of checking this parser, the entire tag_info array for
409        each card is now printed out in the /proc/scsi/aic7xxx/x file. You
410        can use that to verify that your options were parsed correctly.
411        
412    Boot command line options may be combined to form the proper set of options
413    a user might need. For example, the following is valid:
414    
415    aic7xxx=verbose,extended,irq_trigger:1
416    
417    The only requirement is that individual options be separated by a comma or
418    a period on the command line.
419        
420  Module Loading command options
421  ------------------------------
422    When loading the aic7xxx driver as a module, the exact same options are
423    available to the user. However, the syntax to specify the options changes
424    slightly. For insmod, you need to wrap the aic7xxx= argument in quotes
425    and replace all ',' with '.'. So, for example, a valid insmod line
426    would be:
427
428    insmod aic7xxx aic7xxx='verbose.irq_trigger:1.extended'
429
430    This line should result in the *exact* same behaviour as if you typed
431    it in at the lilo prompt and the driver was compiled into the kernel
432    instead of being a module. The reason for the single quote is so that
433    the shell won't try to interpret anything in the line, such as {.
434    Insmod assumes any options starting with a letter instead of a number
435    is a character string (which is what we want) and by switching all of
436    the commas to periods, insmod won't interpret this as more than one
437    string and write junk into our binary image. I consider it a bug in
438    the insmod program that even if you wrap your string in quotes (quotes
439    that pass the shell mind you and that insmod sees) it still treats
440    a comma inside of those quotes as starting a new variable, resulting
441    in memory scribbles if you don't switch the commas to periods.
442
443
444  Kernel Compile options
445  ------------------------------
446    The various kernel compile time options for this driver are now fairly
447    well documented in the file Documentation/Configure.help. In order to
448    see this documentation, you need to use one of the advanced configuration
449    programs (menuconfig and xconfig). If you are using the "make menuconfig"
450    method of configuring your kernel, then you would simply highlight the
451    option in question and hit the ? key. If you are using the "make xconfig"
452    method of configuring your kernel, then simply click on the help button
453    next to the option you have questions about. The help information from
454    the Configure.help file will then get automatically displayed.
455
456  /proc support
457  ------------------------------
458    The /proc support for the AIC7xxx can be found in the /proc/scsi/aic7xxx/
459    directory. That directory contains a file for each SCSI controller in
460    the system. Each file presents the current configuration and transfer
461    statistics (enabled with #define in aic7xxx.c) for each controller.
462
463    Thanks to Michael Neuffer for his upper-level SCSI help, and
464    Matthew Jacob for statistics support.
465
466  Debugging the driver
467  ------------------------------
468    Should you have problems with this driver, and would like some help in
469    getting them solved, there are a couple debugging items built into
470    the driver to facilitate getting the needed information from the system.
471    In general, I need a complete description of the problem, with as many
472    logs as possible concerning what happens. To help with this, there is
473    a command option aic7xxx=panic_on_abort. This option, when set, forces
474    the driver to panic the kernel on the first SCSI abort issued by the
475    mid level SCSI code. If your system is going to reset loops and you
476    can't read the screen, then this is what you need. Not only will it
477    stop the system, but it also prints out a large amount of state
478    information in the process. Second, if you specify the option
479    "aic7xxx=verbose:0x1ffff", the system will print out *SOOOO* much
480    information as it runs that you won't be able to see anything.
481    However, this can actually be very useful if your machine simply
482    locks up when trying to boot, since it will pin-point what was last
483    happening (in regards to the aic7xxx driver) immediately prior to
484    the lockup. This is really only useful if your machine simply can
485    not boot up successfully. If you can get your machine to run, then
486    this will produce far too much information.
487
488  FTP sites
489  ------------------------------
490    ftp://ftp.redhat.com/pub/aic/
491      - Out of date. I used to keep stuff here, but too many people
492        complained about having a hard time getting into Red Hat's ftp
493    server. So use the web site below instead.
494    ftp://ftp.pcnet.com/users/eischen/Linux/
495      - Dan Eischen's driver distribution area
496    ftp://ekf2.vsb.cz/pub/linux/kernel/aic7xxx/ftp.teleport.com/
497      - European Linux mirror of Teleport site
498
499  Web sites
500  ------------------------------
501    http://people.redhat.com/dledford/
502      - My web site, also the primary aic7xxx site with several related
503        pages.
504
505Dean W. Gehnert
506deang@teleport.com
507
508$Revision: 3.0 $
509
510Modified by Doug Ledford 1998-2000
511
512

Archive Download this file



interactive