summaryrefslogtreecommitdiffstats
path: root/Documentation/scsi
diff options
context:
space:
mode:
Diffstat (limited to 'Documentation/scsi')
-rw-r--r--Documentation/scsi/00-INDEX2
-rw-r--r--Documentation/scsi/aic7xxx_old.txt511
-rw-r--r--Documentation/scsi/scsi_eh.txt69
-rw-r--r--Documentation/scsi/scsi_mid_low_api.txt9
4 files changed, 46 insertions, 545 deletions
diff --git a/Documentation/scsi/00-INDEX b/Documentation/scsi/00-INDEX
index 9b0787f965e9..2044be565d93 100644
--- a/Documentation/scsi/00-INDEX
+++ b/Documentation/scsi/00-INDEX
@@ -42,8 +42,6 @@ aic79xx.txt
- Adaptec Ultra320 SCSI host adapters
aic7xxx.txt
- info on driver for Adaptec controllers
-aic7xxx_old.txt
- - info on driver for Adaptec controllers, old generation
arcmsr_spec.txt
- ARECA FIRMWARE SPEC (for IOP331 adapter)
dc395x.txt
diff --git a/Documentation/scsi/aic7xxx_old.txt b/Documentation/scsi/aic7xxx_old.txt
deleted file mode 100644
index ecfc474f36a8..000000000000
--- a/Documentation/scsi/aic7xxx_old.txt
+++ /dev/null
@@ -1,511 +0,0 @@
- AIC7xxx Driver for Linux
-
-Introduction
-----------------------------
-The AIC7xxx SCSI driver adds support for Adaptec (http://www.adaptec.com)
-SCSI controllers and chipsets. Major portions of the driver and driver
-development are shared between both Linux and FreeBSD. Support for the
-AIC-7xxx chipsets have been in the default Linux kernel since approximately
-linux-1.1.x and fairly stable since linux-1.2.x, and are also in FreeBSD
-2.1.0 or later.
-
- Supported cards/chipsets
- ----------------------------
- Adaptec Cards
- ----------------------------
- AHA-274x
- AHA-274xT
- AHA-2842
- AHA-2910B
- AHA-2920C
- AHA-2930
- AHA-2930U
- AHA-2930CU
- AHA-2930U2
- AHA-2940
- AHA-2940W
- AHA-2940U
- AHA-2940UW
- AHA-2940UW-PRO
- AHA-2940AU
- AHA-2940U2W
- AHA-2940U2
- AHA-2940U2B
- AHA-2940U2BOEM
- AHA-2944D
- AHA-2944WD
- AHA-2944UD
- AHA-2944UWD
- AHA-2950U2
- AHA-2950U2W
- AHA-2950U2B
- AHA-29160M
- AHA-3940
- AHA-3940U
- AHA-3940W
- AHA-3940UW
- AHA-3940AUW
- AHA-3940U2W
- AHA-3950U2B
- AHA-3950U2D
- AHA-3960D
- AHA-39160M
- AHA-3985
- AHA-3985U
- AHA-3985W
- AHA-3985UW
-
- Motherboard Chipsets
- ----------------------------
- AIC-777x
- AIC-785x
- AIC-786x
- AIC-787x
- AIC-788x
- AIC-789x
- AIC-3860
-
- Bus Types
- ----------------------------
- W - Wide SCSI, SCSI-3, 16bit bus, 68pin connector, will also support
- SCSI-1/SCSI-2 50pin devices, transfer rates up to 20MB/s.
- U - Ultra SCSI, transfer rates up to 40MB/s.
- U2- Ultra 2 SCSI, transfer rates up to 80MB/s.
- D - Differential SCSI.
- T - Twin Channel SCSI. Up to 14 SCSI devices.
-
- AHA-274x - EISA SCSI controller
- AHA-284x - VLB SCSI controller
- AHA-29xx - PCI SCSI controller
- AHA-394x - PCI controllers with two separate SCSI controllers on-board.
- AHA-398x - PCI RAID controllers with three separate SCSI controllers
- on-board.
-
- Not Supported Devices
- ------------------------------
- Adaptec Cards
- ----------------------------
- AHA-2920 (Only the cards that use the Future Domain chipset are not
- supported, any 2920 cards based on Adaptec AIC chipsets,
- such as the 2920C, are supported)
- AAA-13x Raid Adapters
- AAA-113x Raid Port Card
-
- Motherboard Chipsets
- ----------------------------
- AIC-7810
-
- Bus Types
- ----------------------------
- R - Raid Port busses are not supported.
-
- The hardware RAID devices sold by Adaptec are *NOT* supported by this
- driver (and will people please stop emailing me about them, they are
- a totally separate beast from the bare SCSI controllers and this driver
- cannot be retrofitted in any sane manner to support the hardware RAID
- features on those cards - Doug Ledford).
-
-
- People
- ------------------------------
- Justin T Gibbs gibbs@plutotech.com
- (BSD Driver Author)
- Dan Eischen deischen@iworks.InterWorks.org
- (Original Linux Driver Co-maintainer)
- Dean Gehnert deang@teleport.com
- (Original Linux FTP/patch maintainer)
- Jess Johnson jester@frenzy.com
- (AIC7xxx FAQ author)
- Doug Ledford dledford@redhat.com
- (Current Linux aic7xxx-5.x.x Driver/Patch/FTP maintainer)
-
- Special thanks go to John Aycock (aycock@cpsc.ucalgary.ca), the original
- author of the driver. John has since retired from the project. Thanks
- again for all his work!
-
- Mailing list
- ------------------------------
- There is a mailing list available for users who want to track development
- and converse with other users and developers. This list is for both
- FreeBSD and Linux support of the AIC7xxx chipsets.
-
- To subscribe to the AIC7xxx mailing list send mail to the list server,
- with "subscribe AIC7xxx" in the body (no Subject: required):
- To: majordomo@FreeBSD.ORG
- ---
- subscribe AIC7xxx
-
- To unsubscribe from the list, send mail to the list server with:
- To: majordomo@FreeBSD.ORG
- ---
- unsubscribe AIC7xxx
-
- Send regular messages and replies to: AIC7xxx@FreeBSD.ORG
-
- Boot Command line options
- ------------------------------
- "aic7xxx=no_reset" - Eliminate the SCSI bus reset during startup.
- Some SCSI devices need the initial reset that this option disables
- in order to work. If you have problems at bootup, please make sure
- you aren't using this option.
-
- "aic7xxx=reverse_scan" - Certain PCI motherboards scan for devices at
- bootup by scanning from the highest numbered PCI device to the
- lowest numbered PCI device, others do just the opposite and scan
- from lowest to highest numbered PCI device. There is no reliable
- way to autodetect this ordering. So, we default to the most common
- order, which is lowest to highest. Then, in case your motherboard
- scans from highest to lowest, we have this option. If your BIOS
- finds the drives on controller A before controller B but the linux
- kernel finds your drives on controller B before A, then you should
- use this option.
-
- "aic7xxx=extended" - Force the driver to detect extended drive translation
- on your controller. This helps those people who have cards without
- a SEEPROM make sure that linux and all other operating systems think
- the same way about your hard drives.
-
- "aic7xxx=scbram" - Some cards have external SCB RAM that can be used to
- give the card more hardware SCB slots. This allows the driver to use
- that SCB RAM. Without this option, the driver won't touch the SCB
- RAM because it is known to cause problems on a few cards out there
- (such as 3985 class cards).
-
- "aic7xxx=irq_trigger:x" - Replace x with either 0 or 1 to force the kernel
- to use the correct IRQ type for your card. This only applies to EISA
- based controllers. On these controllers, 0 is for Edge triggered
- interrupts, and 1 is for Level triggered interrupts. If you aren't
- sure or don't know which IRQ trigger type your EISA card uses, then
- let the kernel autodetect the trigger type.
-
- "aic7xxx=verbose" - This option can be used in one of two ways. If you
- simply specify aic7xxx=verbose, then the kernel will automatically
- pick the default set of verbose messages for you to see.
- Alternatively, you can specify the command as
- "aic7xxx=verbose:0xXXXX" where the X entries are replaced with
- hexadecimal digits. This option is a bit field type option. For
- a full listing of the available options, search for the
- #define VERBOSE_xxxxxx lines in the aic7xxx.c file. If you want
- verbose messages, then it is recommended that you simply use the
- aic7xxx=verbose variant of this command.
-
- "aic7xxx=pci_parity:x" - This option controls whether or not the driver
- enables PCI parity error checking on the PCI bus. By default, this
- checking is disabled. To enable the checks, simply specify pci_parity
- with no value afterwords. To reverse the parity from even to odd,
- supply any number other than 0 or 255. In short:
- pci_parity - Even parity checking (even is the normal PCI parity)
- pci_parity:x - Where x > 0, Odd parity checking
- pci_parity:0 - No check (default)
- NOTE: In order to get Even PCI parity checking, you must use the
- version of the option that does not include the : and a number at
- the end (unless you want to enter exactly 2^32 - 1 as the number).
-
- "aic7xxx=no_probe" - This option will disable the probing for any VLB
- based 2842 controllers and any EISA based controllers. This is
- needed on certain newer motherboards where the normal EISA I/O ranges
- have been claimed by other PCI devices. Probing on those machines
- will often result in the machine crashing or spontaneously rebooting
- during startup. Examples of machines that need this are the
- Dell PowerEdge 6300 machines.
-
- "aic7xxx=seltime:2" - This option controls how long the card waits
- during a device selection sequence for the device to respond.
- The original SCSI spec says that this "should be" 256ms. This
- is generally not required with modern devices. However, some
- very old SCSI I devices need the full 256ms. Most modern devices
- can run fine with only 64ms. The default for this option is
- 64ms. If you need to change this option, then use the following
- table to set the proper value in the example above:
- 0 - 256ms
- 1 - 128ms
- 2 - 64ms
- 3 - 32ms
-
- "aic7xxx=panic_on_abort" - This option is for debugging and will cause
- the driver to panic the linux kernel and freeze the system the first
- time the drivers abort or reset routines are called. This is most
- helpful when some problem causes infinite reset loops that scroll too
- fast to see. By using this option, you can write down what the errors
- actually are and send that information to me so it can be fixed.
-
- "aic7xxx=dump_card" - This option will print out the *entire* set of
- configuration registers on the card during the init sequence. This
- is a debugging aid used to see exactly what state the card is in
- when we finally finish our initialization routines. If you don't
- have documentation on the chipsets, this will do you absolutely
- no good unless you are simply trying to write all the information
- down in order to send it to me.
-
- "aic7xxx=dump_sequencer" - This is the same as the above options except
- that instead of dumping the register contents on the card, this
- option dumps the contents of the sequencer program RAM. This gives
- the ability to verify that the instructions downloaded to the
- card's sequencer are indeed what they are supposed to be. Again,
- unless you have documentation to tell you how to interpret these
- numbers, then it is totally useless.
-
- "aic7xxx=override_term:0xffffffff" - This option is used to force the
- termination on your SCSI controllers to a particular setting. This
- is a bit mask variable that applies for up to 8 aic7xxx SCSI channels.
- Each channel gets 4 bits, divided as follows:
- bit 3 2 1 0
- | | | Enable/Disable Single Ended Low Byte Termination
- | | En/Disable Single Ended High Byte Termination
- | En/Disable Low Byte LVD Termination
- En/Disable High Byte LVD Termination
-
- The upper 2 bits that deal with LVD termination only apply to Ultra2
- controllers. Furthermore, due to the current Ultra2 controller
- designs, these bits are tied together such that setting either bit
- enables both low and high byte LVD termination. It is not possible
- to only set high or low byte LVD termination in this manner. This is
- an artifact of the BIOS definition on Ultra2 controllers. For other
- controllers, the only important bits are the two lowest bits. Setting
- the higher bits on non-Ultra2 controllers has no effect. A few
- examples of how to use this option:
-
- Enable low and high byte termination on a non-ultra2 controller that
- is the first aic7xxx controller (the correct bits are 0011),
- aic7xxx=override_term:0x3
-
- Enable all termination on the third aic7xxx controller, high byte
- termination on the second aic7xxx controller, and low and high byte
- SE termination on the first aic7xxx controller
- (bits are 1111 0010 0011),
- aic7xxx=override_term:0xf23
-
- No attempt has been made to make this option non-cryptic. It really
- shouldn't be used except in dire circumstances, and if that happens,
- I'm probably going to be telling you what to set this to anyway :)
-
- "aic7xxx=stpwlev:0xffffffff" - This option is used to control the STPWLEV
- bit in the DEVCONFIG PCI register. Currently, this is one of the
- very few registers that we have absolutely *no* way of detecting
- what the variable should be. It depends entirely on how the chipset
- and external terminators were coupled by the card/motherboard maker.
- Further, a chip reset (at power up) always sets this bit to 0. If
- there is no BIOS to run on the chipset/card (such as with a 2910C
- or a motherboard controller with the BIOS totally disabled) then
- the variable may not get set properly. Of course, if the proper
- setting was 0, then that's what it would be after the reset, but if
- the proper setting is actually 1.....you get the picture. Now, since
- we can't detect this at all, I've added this option to force the
- setting. If you have a BIOS on your controller then you should never
- need to use this option. However, if you are having lots of SCSI
- reset problems and can't seem to get them knocked out, this may help.
-
- Here's a test to know for certain if you need this option. Make
- a boot floppy that you can use to boot your computer up and that
- will detect the aic7xxx controller. Next, power down your computer.
- While it's down, unplug all SCSI cables from your Adaptec SCSI
- controller. Boot the system back up to the Adaptec EZ-SCSI BIOS
- and then make sure that termination is enabled on your adapter (if
- you have an Adaptec BIOS of course). Next, boot up the floppy you
- made and wait for it to detect the aic7xxx controller. If the kernel
- finds the controller fine, says scsi : x hosts and then tries to
- detect your devices like normal, up to the point where it fails to
- mount your root file system and panics, then you're fine. If, on
- the other hand, the system goes into an infinite reset loop, then
- you need to use this option and/or the previous option to force the
- proper termination settings on your controller. If this happens,
- then you next need to figure out what your settings should be.
-
- To find the correct settings, power your machine back down, connect
- back up the SCSI cables, and boot back into your machine like normal.
- However, boot with the aic7xxx=verbose:0x39 option. Record the
- initial DEVCONFIG values for each of your aic7xxx controllers as
- they are listed, and also record what the machine is detecting as
- the proper termination on your controllers. NOTE: the order in
- which the initial DEVCONFIG values are printed out is not guaranteed
- to be the same order as the SCSI controllers are registered. The
- above option and this option both work on the order of the SCSI
- controllers as they are registered, so make sure you match the right
- DEVCONFIG values with the right controllers if you have more than
- one aic7xxx controller.
-
- Once you have the detected termination settings and the initial
- DEVCONFIG values for each controller, then figure out what the
- termination on each of the controllers *should* be. Hopefully, that
- part is correct, but it could possibly be wrong if there is
- bogus cable detection logic on your controller or something similar.
- If all the controllers have the correct termination settings, then
- don't set the aic7xxx=override_term variable at all, leave it alone.
- Next, on any controllers that go into an infinite reset loop when
- you unplug all the SCSI cables, get the starting DEVCONFIG value.
- If the initial DEVCONFIG value is divisible by 2, then the correct
- setting for that controller is 0. If it's an odd number, then
- the correct setting for that controller is 1. For any other
- controllers that didn't have an infinite reset problem, then reverse
- the above options. If DEVCONFIG was even, then the correct setting
- is 1, if not then the correct setting is 0.
-
- Now that you know what the correct setting was for each controller,
- we need to encode that into the aic7xxx=stpwlev:0x... variable.
- This variable is a bit field encoded variable. Bit 0 is for the first
- aic7xxx controller, bit 1 for the next, etc. Put all these bits
- together and you get a number. For example, if the third aic7xxx
- needed a 1, but the second and first both needed a 0, then the bits
- would be 100 in binary. This then translates to 0x04. You would
- therefore set aic7xxx=stpwlev:0x04. This is fairly standard binary
- to hexadecimal conversions here. If you aren't up to speed on the
- binary->hex conversion then send an email to the aic7xxx mailing
- list and someone can help you out.
-
- "aic7xxx=tag_info:{{8,8..},{8,8..},..}" - This option is used to disable
- or enable Tagged Command Queueing (TCQ) on specific devices. As of
- driver version 5.1.11, TCQ is now either on or off by default
- according to the setting you choose during the make config process.
- In order to en/disable TCQ for certain devices at boot time, a user
- may use this boot param. The driver will then parse this message out
- and en/disable the specific device entries that are present based upon
- the value given. The param line is parsed in the following manner:
-
- { - first instance indicates the start of this parameter values
- second instance is the start of entries for a particular
- device entry
- } - end the entries for a particular host adapter, or end the entire
- set of parameter entries
- , - move to next entry. Inside of a set of device entries, this
- moves us to the next device on the list. Outside of device
- entries, this moves us to the next host adapter
- . - Same effect as , but is safe to use with insmod.
- x - the number to enter into the array at this position.
- 0 = Enable tagged queueing on this device and use the default
- queue depth
- 1-254 = Enable tagged queueing on this device and use this
- number as the queue depth
- 255 = Disable tagged queueing on this device.
- Note: anything above 32 for an actual queue depth is wasteful
- and not recommended.
-
- A few examples of how this can be used:
-
- tag_info:{{8,12,,0,,255,4}}
- This line will only effect the first aic7xxx card registered. It
- will set scsi id 0 to a queue depth of 8, id 1 to 12, leave id 2
- at the default, set id 3 to tagged queueing enabled and use the
- default queue depth, id 4 default, id 5 disabled, and id 6 to 4.
- Any not specified entries stay at the default value, repeated
- commas with no value specified will simply increment to the next id
- without changing anything for the missing values.
-
- tag_info:{,,,{,,,255}}
- First, second, and third adapters at default values. Fourth
- adapter, id 3 is disabled. Notice that leading commas simply
- increment what the first number effects, and there are no need
- for trailing commas. When you close out an adapter, or the
- entire entry, anything not explicitly set stays at the default
- value.
-
- A final note on this option. The scanner I used for this isn't
- perfect or highly robust. If you mess the line up, the worst that
- should happen is that the line will get ignored. If you don't
- close out the entire entry with the final bracket, then any other
- aic7xxx options after this will get ignored. So, in general, be
- sure of what you are entering, and after you have it right, just
- add it to the lilo.conf file so there won't be any mistakes. As
- a means of checking this parser, the entire tag_info array for
- each card is now printed out in the /proc/scsi/aic7xxx/x file. You
- can use that to verify that your options were parsed correctly.
-
- Boot command line options may be combined to form the proper set of options
- a user might need. For example, the following is valid:
-
- aic7xxx=verbose,extended,irq_trigger:1
-
- The only requirement is that individual options be separated by a comma or
- a period on the command line.
-
- Module Loading command options
- ------------------------------
- When loading the aic7xxx driver as a module, the exact same options are
- available to the user. However, the syntax to specify the options changes
- slightly. For insmod, you need to wrap the aic7xxx= argument in quotes
- and replace all ',' with '.'. So, for example, a valid insmod line
- would be:
-
- insmod aic7xxx aic7xxx='verbose.irq_trigger:1.extended'
-
- This line should result in the *exact* same behaviour as if you typed
- it in at the lilo prompt and the driver was compiled into the kernel
- instead of being a module. The reason for the single quote is so that
- the shell won't try to interpret anything in the line, such as {.
- Insmod assumes any options starting with a letter instead of a number
- is a character string (which is what we want) and by switching all of
- the commas to periods, insmod won't interpret this as more than one
- string and write junk into our binary image. I consider it a bug in
- the insmod program that even if you wrap your string in quotes (quotes
- that pass the shell mind you and that insmod sees) it still treats
- a comma inside of those quotes as starting a new variable, resulting
- in memory scribbles if you don't switch the commas to periods.
-
-
- Kernel Compile options
- ------------------------------
- The various kernel compile time options for this driver are now fairly
- well documented in the file drivers/scsi/Kconfig. In order to
- see this documentation, you need to use one of the advanced configuration
- programs (menuconfig and xconfig). If you are using the "make menuconfig"
- method of configuring your kernel, then you would simply highlight the
- option in question and hit the ? key. If you are using the "make xconfig"
- method of configuring your kernel, then simply click on the help button
- next to the option you have questions about. The help information from
- the Configure.help file will then get automatically displayed.
-
- /proc support
- ------------------------------
- The /proc support for the AIC7xxx can be found in the /proc/scsi/aic7xxx/
- directory. That directory contains a file for each SCSI controller in
- the system. Each file presents the current configuration and transfer
- statistics (enabled with #define in aic7xxx.c) for each controller.
-
- Thanks to Michael Neuffer for his upper-level SCSI help, and
- Matthew Jacob for statistics support.
-
- Debugging the driver
- ------------------------------
- Should you have problems with this driver, and would like some help in
- getting them solved, there are a couple debugging items built into
- the driver to facilitate getting the needed information from the system.
- In general, I need a complete description of the problem, with as many
- logs as possible concerning what happens. To help with this, there is
- a command option aic7xxx=panic_on_abort. This option, when set, forces
- the driver to panic the kernel on the first SCSI abort issued by the
- mid level SCSI code. If your system is going to reset loops and you
- can't read the screen, then this is what you need. Not only will it
- stop the system, but it also prints out a large amount of state
- information in the process. Second, if you specify the option
- "aic7xxx=verbose:0x1ffff", the system will print out *SOOOO* much
- information as it runs that you won't be able to see anything.
- However, this can actually be very useful if your machine simply
- locks up when trying to boot, since it will pin-point what was last
- happening (in regards to the aic7xxx driver) immediately prior to
- the lockup. This is really only useful if your machine simply can
- not boot up successfully. If you can get your machine to run, then
- this will produce far too much information.
-
- FTP sites
- ------------------------------
- ftp://ftp.redhat.com/pub/aic/
- - Out of date. I used to keep stuff here, but too many people
- complained about having a hard time getting into Red Hat's ftp
- server. So use the web site below instead.
- ftp://ftp.pcnet.com/users/eischen/Linux/
- - Dan Eischen's driver distribution area
- ftp://ekf2.vsb.cz/pub/linux/kernel/aic7xxx/ftp.teleport.com/
- - European Linux mirror of Teleport site
-
- Web sites
- ------------------------------
- http://people.redhat.com/dledford/
- - My web site, also the primary aic7xxx site with several related
- pages.
-
-Dean W. Gehnert
-deang@teleport.com
-
-$Revision: 3.0 $
-
-Modified by Doug Ledford 1998-2000
-
diff --git a/Documentation/scsi/scsi_eh.txt b/Documentation/scsi/scsi_eh.txt
index 6ff16b620d84..a0c85110a07e 100644
--- a/Documentation/scsi/scsi_eh.txt
+++ b/Documentation/scsi/scsi_eh.txt
@@ -42,20 +42,14 @@ discussion.
Once LLDD gets hold of a scmd, either the LLDD will complete the
command by calling scsi_done callback passed from midlayer when
-invoking hostt->queuecommand() or SCSI midlayer will time it out.
+invoking hostt->queuecommand() or the block layer will time it out.
[1-2-1] Completing a scmd w/ scsi_done
For all non-EH commands, scsi_done() is the completion callback. It
-does the following.
-
- 1. Delete timeout timer. If it fails, it means that timeout timer
- has expired and is going to finish the command. Just return.
-
- 2. Link scmd to per-cpu scsi_done_q using scmd->en_entry
-
- 3. Raise SCSI_SOFTIRQ
+just calls blk_complete_request() to delete the block layer timer and
+raise SCSI_SOFTIRQ
SCSI_SOFTIRQ handler scsi_softirq calls scsi_decide_disposition() to
determine what to do with the command. scsi_decide_disposition()
@@ -64,10 +58,12 @@ with the command.
- SUCCESS
scsi_finish_command() is invoked for the command. The
- function does some maintenance choirs and notify completion by
- calling scmd->done() callback, which, for fs requests, would
- be HLD completion callback - sd:sd_rw_intr, sr:rw_intr,
- st:st_intr.
+ function does some maintenance chores and then calls
+ scsi_io_completion() to finish the I/O.
+ scsi_io_completion() then notifies the block layer on
+ the completed request by calling blk_end_request and
+ friends or figures out what to do with the remainder
+ of the data in case of an error.
- NEEDS_RETRY
- ADD_TO_MLQUEUE
@@ -86,33 +82,45 @@ function
1. invokes optional hostt->eh_timed_out() callback. Return value can
be one of
- - EH_HANDLED
- This indicates that eh_timed_out() dealt with the timeout. The
- scmd is passed to __scsi_done() and thus linked into per-cpu
- scsi_done_q. Normal command completion described in [1-2-1]
- follows.
+ - BLK_EH_HANDLED
+ This indicates that eh_timed_out() dealt with the timeout.
+ The command is passed back to the block layer and completed
+ via __blk_complete_requests().
+
+ *NOTE* After returning BLK_EH_HANDLED the SCSI layer is
+ assumed to be finished with the command, and no other
+ functions from the SCSI layer will be called. So this
+ should typically only be returned if the eh_timed_out()
+ handler raced with normal completion.
- - EH_RESET_TIMER
+ - BLK_EH_RESET_TIMER
This indicates that more time is required to finish the
command. Timer is restarted. This action is counted as a
retry and only allowed scmd->allowed + 1(!) times. Once the
- limit is reached, action for EH_NOT_HANDLED is taken instead.
+ limit is reached, action for BLK_EH_NOT_HANDLED is taken instead.
- *NOTE* This action is racy as the LLDD could finish the scmd
- after the timeout has expired but before it's added back. In
- such cases, scsi_done() would think that timeout has occurred
- and return without doing anything. We lose completion and the
- command will time out again.
-
- - EH_NOT_HANDLED
- This is the same as when eh_timed_out() callback doesn't exist.
+ - BLK_EH_NOT_HANDLED
+ eh_timed_out() callback did not handle the command.
Step #2 is taken.
+ 2. If the host supports asynchronous completion (as indicated by the
+ no_async_abort setting in the host template) scsi_abort_command()
+ is invoked to schedule an asynchrous abort. If that fails
+ Step #3 is taken.
+
2. scsi_eh_scmd_add(scmd, SCSI_EH_CANCEL_CMD) is invoked for the
command. See [1-3] for more information.
+[1-3] Asynchronous command aborts
+
+ After a timeout occurs a command abort is scheduled from
+ scsi_abort_command(). If the abort is successful the command
+ will either be retried (if the number of retries is not exhausted)
+ or terminated with DID_TIME_OUT.
+ Otherwise scsi_eh_scmd_add() is invoked for the command.
+ See [1-4] for more information.
-[1-3] How EH takes over
+[1-4] How EH takes over
scmds enter EH via scsi_eh_scmd_add(), which does the following.
@@ -320,7 +328,8 @@ scmd->allowed.
<<scsi_eh_abort_cmds>>
- This action is taken for each timed out command.
+ This action is taken for each timed out command when
+ no_async_abort is enabled in the host template.
hostt->eh_abort_handler() is invoked for each scmd. The
handler returns SUCCESS if it has succeeded to make LLDD and
all related hardware forget about the scmd.
diff --git a/Documentation/scsi/scsi_mid_low_api.txt b/Documentation/scsi/scsi_mid_low_api.txt
index 2b06aba4fa0f..d6a9bdeee7f2 100644
--- a/Documentation/scsi/scsi_mid_low_api.txt
+++ b/Documentation/scsi/scsi_mid_low_api.txt
@@ -882,8 +882,11 @@ Details:
*
* Calling context: kernel thread
*
- * Notes: Invoked from scsi_eh thread. No other commands will be
- * queued on current host during eh.
+ * Notes: If 'no_async_abort' is defined this callback
+ * will be invoked from scsi_eh thread. No other commands
+ * will then be queued on current host during eh.
+ * Otherwise it will be called whenever scsi_times_out()
+ * is called due to a command timeout.
*
* Optionally defined in: LLD
**/
@@ -1257,6 +1260,8 @@ of interest:
address space
use_clustering - 1=>SCSI commands in mid level's queue can be merged,
0=>disallow SCSI command merging
+ no_async_abort - 1=>Asynchronous aborts are not supported
+ 0=>Timed-out commands will be aborted asynchronously
hostt - pointer to driver's struct scsi_host_template from which
this struct Scsi_Host instance was spawned
hostt->proc_name - name of LLD. This is the driver name that sysfs uses