Initial import

6 files changed, 2624 insertions, 0 deletions

diff --git a/Documentation/ioctl/00-INDEX b/Documentation/ioctl/00-INDEX
new file mode 100644
index 000000000..c1a925787
--- /dev/null
+++ b/Documentation/ioctl/00-INDEX
@@ -0,0 +1,12 @@
+00-INDEX
+	- this file
+botching-up-ioctls.txt
+	- how to avoid botching up ioctls
+cdrom.txt
+	- summary of CDROM ioctl calls
+hdio.txt
+	- summary of HDIO_ ioctl calls
+ioctl-decoding.txt
+	- how to decode the bits of an IOCTL code
+ioctl-number.txt
+	- how to implement and register device/driver ioctl calls
diff --git a/Documentation/ioctl/botching-up-ioctls.txt b/Documentation/ioctl/botching-up-ioctls.txt
new file mode 100644
index 000000000..45fe78c58
--- /dev/null
+++ b/Documentation/ioctl/botching-up-ioctls.txt
@@ -0,0 +1,219 @@
+(How to avoid) Botching up ioctls
+=================================
+
+From: http://blog.ffwll.ch/2013/11/botching-up-ioctls.html
+
+By: Daniel Vetter, Copyright © 2013 Intel Corporation
+
+One clear insight kernel graphics hackers gained in the past few years is that
+trying to come up with a unified interface to manage the execution units and
+memory on completely different GPUs is a futile effort. So nowadays every
+driver has its own set of ioctls to allocate memory and submit work to the GPU.
+Which is nice, since there's no more insanity in the form of fake-generic, but
+actually only used once interfaces. But the clear downside is that there's much
+more potential to screw things up.
+
+To avoid repeating all the same mistakes again I've written up some of the
+lessons learned while botching the job for the drm/i915 driver. Most of these
+only cover technicalities and not the big-picture issues like what the command
+submission ioctl exactly should look like. Learning these lessons is probably
+something every GPU driver has to do on its own.
+
+
+Prerequisites
+-------------
+
+First the prerequisites. Without these you have already failed, because you
+will need to add a a 32-bit compat layer:
+
+ * Only use fixed sized integers. To avoid conflicts with typedefs in userspace
+   the kernel has special types like __u32, __s64. Use them.
+
+ * Align everything to the natural size and use explicit padding. 32-bit
+   platforms don't necessarily align 64-bit values to 64-bit boundaries, but
+   64-bit platforms do. So we always need padding to the natural size to get
+   this right.
+
+ * Pad the entire struct to a multiple of 64-bits - the structure size will
+   otherwise differ on 32-bit versus 64-bit. Having a different structure size
+   hurts when passing arrays of structures to the kernel, or if the kernel
+   checks the structure size, which e.g. the drm core does.
+
+ * Pointers are __u64, cast from/to a uintprt_t on the userspace side and
+   from/to a void __user * in the kernel. Try really hard not to delay this
+   conversion or worse, fiddle the raw __u64 through your code since that
+   diminishes the checking tools like sparse can provide.
+
+
+Basics
+------
+
+With the joys of writing a compat layer avoided we can take a look at the basic
+fumbles. Neglecting these will make backward and forward compatibility a real
+pain. And since getting things wrong on the first attempt is guaranteed you
+will have a second iteration or at least an extension for any given interface.
+
+ * Have a clear way for userspace to figure out whether your new ioctl or ioctl
+   extension is supported on a given kernel. If you can't rely on old kernels
+   rejecting the new flags/modes or ioctls (since doing that was botched in the
+   past) then you need a driver feature flag or revision number somewhere.
+
+ * Have a plan for extending ioctls with new flags or new fields at the end of
+   the structure. The drm core checks the passed-in size for each ioctl call
+   and zero-extends any mismatches between kernel and userspace. That helps,
+   but isn't a complete solution since newer userspace on older kernels won't
+   notice that the newly added fields at the end get ignored. So this still
+   needs a new driver feature flags.
+
+ * Check all unused fields and flags and all the padding for whether it's 0,
+   and reject the ioctl if that's not the case. Otherwise your nice plan for
+   future extensions is going right down the gutters since someone will submit
+   an ioctl struct with random stack garbage in the yet unused parts. Which
+   then bakes in the ABI that those fields can never be used for anything else
+   but garbage.
+
+ * Have simple testcases for all of the above.
+
+
+Fun with Error Paths
+--------------------
+
+Nowadays we don't have any excuse left any more for drm drivers being neat
+little root exploits. This means we both need full input validation and solid
+error handling paths - GPUs will die eventually in the oddmost corner cases
+anyway:
+
+ * The ioctl must check for array overflows. Also it needs to check for
+   over/underflows and clamping issues of integer values in general. The usual
+   example is sprite positioning values fed directly into the hardware with the
+   hardware just having 12 bits or so. Works nicely until some odd display
+   server doesn't bother with clamping itself and the cursor wraps around the
+   screen.
+
+ * Have simple testcases for every input validation failure case in your ioctl.
+   Check that the error code matches your expectations. And finally make sure
+   that you only test for one single error path in each subtest by submitting
+   otherwise perfectly valid data. Without this an earlier check might reject
+   the ioctl already and shadow the codepath you actually want to test, hiding
+   bugs and regressions.
+
+ * Make all your ioctls restartable. First X really loves signals and second
+   this will allow you to test 90% of all error handling paths by just
+   interrupting your main test suite constantly with signals. Thanks to X's
+   love for signal you'll get an excellent base coverage of all your error
+   paths pretty much for free for graphics drivers. Also, be consistent with
+   how you handle ioctl restarting - e.g. drm has a tiny drmIoctl helper in its
+   userspace library. The i915 driver botched this with the set_tiling ioctl,
+   now we're stuck forever with some arcane semantics in both the kernel and
+   userspace.
+
+ * If you can't make a given codepath restartable make a stuck task at least
+   killable. GPUs just die and your users won't like you more if you hang their
+   entire box (by means of an unkillable X process). If the state recovery is
+   still too tricky have a timeout or hangcheck safety net as a last-ditch
+   effort in case the hardware has gone bananas.
+
+ * Have testcases for the really tricky corner cases in your error recovery code
+   - it's way too easy to create a deadlock between your hangcheck code and
+   waiters.
+
+
+Time, Waiting and Missing it
+----------------------------
+
+GPUs do most everything asynchronously, so we have a need to time operations and
+wait for oustanding ones. This is really tricky business; at the moment none of
+the ioctls supported by the drm/i915 get this fully right, which means there's
+still tons more lessons to learn here.
+
+ * Use CLOCK_MONOTONIC as your reference time, always. It's what alsa, drm and
+   v4l use by default nowadays. But let userspace know which timestamps are
+   derived from different clock domains like your main system clock (provided
+   by the kernel) or some independent hardware counter somewhere else. Clocks
+   will mismatch if you look close enough, but if performance measuring tools
+   have this information they can at least compensate. If your userspace can
+   get at the raw values of some clocks (e.g. through in-command-stream
+   performance counter sampling instructions) consider exposing those also.
+
+ * Use __s64 seconds plus __u64 nanoseconds to specify time. It's not the most
+   convenient time specification, but it's mostly the standard.
+
+ * Check that input time values are normalized and reject them if not. Note
+   that the kernel native struct ktime has a signed integer for both seconds
+   and nanoseconds, so beware here.
+
+ * For timeouts, use absolute times. If you're a good fellow and made your
+   ioctl restartable relative timeouts tend to be too coarse and can
+   indefinitely extend your wait time due to rounding on each restart.
+   Especially if your reference clock is something really slow like the display
+   frame counter. With a spec laywer hat on this isn't a bug since timeouts can
+   always be extended - but users will surely hate you if their neat animations
+   starts to stutter due to this.
+
+ * Consider ditching any synchronous wait ioctls with timeouts and just deliver
+   an asynchronous event on a pollable file descriptor. It fits much better
+   into event driven applications' main loop.
+
+ * Have testcases for corner-cases, especially whether the return values for
+   already-completed events, successful waits and timed-out waits are all sane
+   and suiting to your needs.
+
+
+Leaking Resources, Not
+----------------------
+
+A full-blown drm driver essentially implements a little OS, but specialized to
+the given GPU platforms. This means a driver needs to expose tons of handles
+for different objects and other resources to userspace. Doing that right
+entails its own little set of pitfalls:
+
+ * Always attach the lifetime of your dynamically created resources to the
+   lifetime of a file descriptor. Consider using a 1:1 mapping if your resource
+   needs to be shared across processes -  fd-passing over unix domain sockets
+   also simplifies lifetime management for userspace.
+
+ * Always have O_CLOEXEC support.
+
+ * Ensure that you have sufficient insulation between different clients. By
+   default pick a private per-fd namespace which forces any sharing to be done
+   explictly. Only go with a more global per-device namespace if the objects
+   are truly device-unique. One counterexample in the drm modeset interfaces is
+   that the per-device modeset objects like connectors share a namespace with
+   framebuffer objects, which mostly are not shared at all. A separate
+   namespace, private by default, for framebuffers would have been more
+   suitable.
+
+ * Think about uniqueness requirements for userspace handles. E.g. for most drm
+   drivers it's a userspace bug to submit the same object twice in the same
+   command submission ioctl. But then if objects are shareable userspace needs
+   to know whether it has seen an imported object from a different process
+   already or not. I haven't tried this myself yet due to lack of a new class
+   of objects, but consider using inode numbers on your shared file descriptors
+   as unique identifiers - it's how real files are told apart, too.
+   Unfortunately this requires a full-blown virtual filesystem in the kernel.
+
+
+Last, but not Least
+-------------------
+
+Not every problem needs a new ioctl:
+
+ * Think hard whether you really want a driver-private interface. Of course
+   it's much quicker to push a driver-private interface than engaging in
+   lengthy discussions for a more generic solution. And occasionally doing a
+   private interface to spearhead a new concept is what's required. But in the
+   end, once the generic interface comes around you'll end up maintainer two
+   interfaces. Indefinitely.
+
+ * Consider other interfaces than ioctls. A sysfs attribute is much better for
+   per-device settings, or for child objects with fairly static lifetimes (like
+   output connectors in drm with all the detection override attributes). Or
+   maybe only your testsuite needs this interface, and then debugfs with its
+   disclaimer of not having a stable ABI would be better.
+
+Finally, the name of the game is to get it right on the first attempt, since if
+your driver proves popular and your hardware platforms long-lived then you'll
+be stuck with a given ioctl essentially forever. You can try to deprecate
+horrible ioctls on newer iterations of your hardware, but generally it takes
+years to accomplish this. And then again years until the last user able to
+complain about regressions disappears, too.
diff --git a/Documentation/ioctl/cdrom.txt b/Documentation/ioctl/cdrom.txt
new file mode 100644
index 000000000..59df81c8d
--- /dev/null
+++ b/Documentation/ioctl/cdrom.txt
@@ -0,0 +1,966 @@
+		Summary of CDROM ioctl calls.
+		============================
+
+		Edward A. Falk <efalk@google.com>
+
+		November, 2004
+
+This document attempts to describe the ioctl(2) calls supported by
+the CDROM layer.  These are by-and-large implemented (as of Linux 2.6)
+in drivers/cdrom/cdrom.c and drivers/block/scsi_ioctl.c
+
+ioctl values are listed in <linux/cdrom.h>.  As of this writing, they
+are as follows:
+
+	CDROMPAUSE		Pause Audio Operation
+	CDROMRESUME		Resume paused Audio Operation
+	CDROMPLAYMSF		Play Audio MSF (struct cdrom_msf)
+	CDROMPLAYTRKIND		Play Audio Track/index (struct cdrom_ti)
+	CDROMREADTOCHDR		Read TOC header (struct cdrom_tochdr)
+	CDROMREADTOCENTRY	Read TOC entry (struct cdrom_tocentry)
+	CDROMSTOP		Stop the cdrom drive
+	CDROMSTART		Start the cdrom drive
+	CDROMEJECT		Ejects the cdrom media
+	CDROMVOLCTRL		Control output volume (struct cdrom_volctrl)
+	CDROMSUBCHNL		Read subchannel data (struct cdrom_subchnl)
+	CDROMREADMODE2		Read CDROM mode 2 data (2336 Bytes)
+					   (struct cdrom_read)
+	CDROMREADMODE1		Read CDROM mode 1 data (2048 Bytes)
+					   (struct cdrom_read)
+	CDROMREADAUDIO		(struct cdrom_read_audio)
+	CDROMEJECT_SW		enable(1)/disable(0) auto-ejecting
+	CDROMMULTISESSION	Obtain the start-of-last-session
+				  address of multi session disks
+				  (struct cdrom_multisession)
+	CDROM_GET_MCN		Obtain the "Universal Product Code"
+				   if available (struct cdrom_mcn)
+	CDROM_GET_UPC		Deprecated, use CDROM_GET_MCN instead.
+	CDROMRESET		hard-reset the drive
+	CDROMVOLREAD		Get the drive's volume setting
+					  (struct cdrom_volctrl)
+	CDROMREADRAW		read data in raw mode (2352 Bytes)
+					   (struct cdrom_read)
+	CDROMREADCOOKED		read data in cooked mode
+	CDROMSEEK		seek msf address
+	CDROMPLAYBLK		scsi-cd only, (struct cdrom_blk)
+	CDROMREADALL		read all 2646 bytes
+	CDROMGETSPINDOWN	return 4-bit spindown value
+	CDROMSETSPINDOWN	set 4-bit spindown value
+	CDROMCLOSETRAY		pendant of CDROMEJECT
+	CDROM_SET_OPTIONS	Set behavior options
+	CDROM_CLEAR_OPTIONS	Clear behavior options
+	CDROM_SELECT_SPEED	Set the CD-ROM speed
+	CDROM_SELECT_DISC	Select disc (for juke-boxes)
+	CDROM_MEDIA_CHANGED	Check is media changed
+	CDROM_DRIVE_STATUS	Get tray position, etc.
+	CDROM_DISC_STATUS	Get disc type, etc.
+	CDROM_CHANGER_NSLOTS	Get number of slots
+	CDROM_LOCKDOOR		lock or unlock door
+	CDROM_DEBUG		Turn debug messages on/off
+	CDROM_GET_CAPABILITY	get capabilities
+	CDROMAUDIOBUFSIZ	set the audio buffer size
+	DVD_READ_STRUCT		Read structure
+	DVD_WRITE_STRUCT	Write structure
+	DVD_AUTH		Authentication
+	CDROM_SEND_PACKET	send a packet to the drive
+	CDROM_NEXT_WRITABLE	get next writable block
+	CDROM_LAST_WRITTEN	get last block written on disc
+
+
+The information that follows was determined from reading kernel source
+code.  It is likely that some corrections will be made over time.
+
+
+
+
+
+
+
+General:
+
+	Unless otherwise specified, all ioctl calls return 0 on success
+	and -1 with errno set to an appropriate value on error.  (Some
+	ioctls return non-negative data values.)
+
+	Unless otherwise specified, all ioctl calls return -1 and set
+	errno to EFAULT on a failed attempt to copy data to or from user
+	address space.
+
+	Individual drivers may return error codes not listed here.
+
+	Unless otherwise specified, all data structures and constants
+	are defined in <linux/cdrom.h>
+
+
+
+
+CDROMPAUSE			Pause Audio Operation
+
+	usage:
+
+	  ioctl(fd, CDROMPAUSE, 0);
+
+	inputs:		none
+
+	outputs:	none
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+
+
+CDROMRESUME			Resume paused Audio Operation
+
+	usage:
+
+	  ioctl(fd, CDROMRESUME, 0);
+
+	inputs:		none
+
+	outputs:	none
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+
+
+CDROMPLAYMSF			Play Audio MSF (struct cdrom_msf)
+
+	usage:
+
+	  struct cdrom_msf msf;
+	  ioctl(fd, CDROMPLAYMSF, &msf);
+
+	inputs:
+	  cdrom_msf structure, describing a segment of music to play
+
+	outputs:	none
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+
+	notes:
+	  MSF stands for minutes-seconds-frames
+	  LBA stands for logical block address
+
+	  Segment is described as start and end times, where each time
+	  is described as minutes:seconds:frames.  A frame is 1/75 of
+	  a second.
+
+
+CDROMPLAYTRKIND			Play Audio Track/index (struct cdrom_ti)
+
+	usage:
+
+	  struct cdrom_ti ti;
+	  ioctl(fd, CDROMPLAYTRKIND, &ti);
+
+	inputs:
+	  cdrom_ti structure, describing a segment of music to play
+
+	outputs:	none
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+
+	notes:
+	  Segment is described as start and end times, where each time
+	  is described as a track and an index.
+
+
+
+CDROMREADTOCHDR			Read TOC header (struct cdrom_tochdr)
+
+	usage:
+
+	  cdrom_tochdr header;
+	  ioctl(fd, CDROMREADTOCHDR, &header);
+
+	inputs:
+	  cdrom_tochdr structure
+
+	outputs:
+	  cdrom_tochdr structure
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+
+
+
+CDROMREADTOCENTRY		Read TOC entry (struct cdrom_tocentry)
+
+	usage:
+
+	  struct cdrom_tocentry entry;
+	  ioctl(fd, CDROMREADTOCENTRY, &entry);
+
+	inputs:
+	  cdrom_tocentry structure
+
+	outputs:
+	  cdrom_tocentry structure
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+	  EINVAL	entry.cdte_format not CDROM_MSF or CDROM_LBA
+	  EINVAL	requested track out of bounds
+	  EIO		I/O error reading TOC
+
+	notes:
+	  TOC stands for Table Of Contents
+	  MSF stands for minutes-seconds-frames
+	  LBA stands for logical block address
+
+
+
+CDROMSTOP			Stop the cdrom drive
+
+	usage:
+
+	  ioctl(fd, CDROMSTOP, 0);
+
+	inputs:		none
+
+	outputs:	none
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+
+	notes:
+	  Exact interpretation of this ioctl depends on the device,
+	  but most seem to spin the drive down.
+
+
+CDROMSTART			Start the cdrom drive
+
+	usage:
+
+	  ioctl(fd, CDROMSTART, 0);
+
+	inputs:		none
+
+	outputs:	none
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+
+	notes:
+	  Exact interpretation of this ioctl depends on the device,
+	  but most seem to spin the drive up and/or close the tray.
+	  Other devices ignore the ioctl completely.
+
+
+CDROMEJECT			Ejects the cdrom media
+
+	usage:
+
+	  ioctl(fd, CDROMEJECT, 0);
+
+	inputs:		none
+
+	outputs:	none
+
+	error returns:
+	  ENOSYS	cd drive not capable of ejecting
+	  EBUSY		other processes are accessing drive, or door is locked
+
+	notes:
+	  See CDROM_LOCKDOOR, below.
+
+
+
+CDROMCLOSETRAY			pendant of CDROMEJECT
+
+	usage:
+
+	  ioctl(fd, CDROMCLOSETRAY, 0);
+
+	inputs:		none
+
+	outputs:	none
+
+	error returns:
+	  ENOSYS	cd drive not capable of closing the tray
+	  EBUSY		other processes are accessing drive, or door is locked
+
+	notes:
+	  See CDROM_LOCKDOOR, below.
+
+
+
+CDROMVOLCTRL			Control output volume (struct cdrom_volctrl)
+
+	usage:
+
+	  struct cdrom_volctrl volume;
+	  ioctl(fd, CDROMVOLCTRL, &volume);
+
+	inputs:
+	  cdrom_volctrl structure containing volumes for up to 4
+	  channels.
+
+	outputs:	none
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+
+
+
+CDROMVOLREAD			Get the drive's volume setting
+					  (struct cdrom_volctrl)
+
+	usage:
+
+	  struct cdrom_volctrl volume;
+	  ioctl(fd, CDROMVOLREAD, &volume);
+
+	inputs:		none
+
+	outputs:
+	  The current volume settings.
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+
+
+
+CDROMSUBCHNL			Read subchannel data (struct cdrom_subchnl)
+
+	usage:
+
+	  struct cdrom_subchnl q;
+	  ioctl(fd, CDROMSUBCHNL, &q);
+
+	inputs:
+	  cdrom_subchnl structure
+
+	outputs:
+	  cdrom_subchnl structure
+
+	error return:
+	  ENOSYS	cd drive not audio-capable.
+	  EINVAL	format not CDROM_MSF or CDROM_LBA
+
+	notes:
+	  Format is converted to CDROM_MSF on return
+
+
+
+CDROMREADRAW			read data in raw mode (2352 Bytes)
+					   (struct cdrom_read)
+
+	usage:
+
+	  union {
+	    struct cdrom_msf msf;		/* input */
+	    char buffer[CD_FRAMESIZE_RAW];	/* return */
+	  } arg;
+	  ioctl(fd, CDROMREADRAW, &arg);
+
+	inputs:
+	  cdrom_msf structure indicating an address to read.
+	  Only the start values are significant.
+
+	outputs:
+	  Data written to address provided by user.
+
+	error return:
+	  EINVAL	address less than 0, or msf less than 0:2:0
+	  ENOMEM	out of memory
+
+	notes:
+	  As of 2.6.8.1, comments in <linux/cdrom.h> indicate that this
+	  ioctl accepts a cdrom_read structure, but actual source code
+	  reads a cdrom_msf structure and writes a buffer of data to
+	  the same address.
+
+	  MSF values are converted to LBA values via this formula:
+
+	    lba = (((m * CD_SECS) + s) * CD_FRAMES + f) - CD_MSF_OFFSET;
+
+
+
+
+CDROMREADMODE1			Read CDROM mode 1 data (2048 Bytes)
+					   (struct cdrom_read)
+
+	notes:
+	  Identical to CDROMREADRAW except that block size is
+	  CD_FRAMESIZE (2048) bytes
+
+
+
+CDROMREADMODE2			Read CDROM mode 2 data (2336 Bytes)
+					   (struct cdrom_read)
+
+	notes:
+	  Identical to CDROMREADRAW except that block size is
+	  CD_FRAMESIZE_RAW0 (2336) bytes
+
+
+
+CDROMREADAUDIO			(struct cdrom_read_audio)
+
+	usage:
+
+	  struct cdrom_read_audio ra;
+	  ioctl(fd, CDROMREADAUDIO, &ra);
+
+	inputs:
+	  cdrom_read_audio structure containing read start
+	  point and length
+
+	outputs:
+	  audio data, returned to buffer indicated by ra
+
+	error return:
+	  EINVAL	format not CDROM_MSF or CDROM_LBA
+	  EINVAL	nframes not in range [1 75]
+	  ENXIO		drive has no queue (probably means invalid fd)
+	  ENOMEM	out of memory
+
+
+CDROMEJECT_SW			enable(1)/disable(0) auto-ejecting
+
+	usage:
+
+	  int val;
+	  ioctl(fd, CDROMEJECT_SW, val);
+
+	inputs:
+	  Flag specifying auto-eject flag.
+
+	outputs:	none
+
+	error return:
+	  ENOSYS	Drive is not capable of ejecting.
+	  EBUSY		Door is locked
+
+
+
+
+CDROMMULTISESSION		Obtain the start-of-last-session
+				  address of multi session disks
+				  (struct cdrom_multisession)
+	usage:
+
+	  struct cdrom_multisession ms_info;
+	  ioctl(fd, CDROMMULTISESSION, &ms_info);
+
+	inputs:
+	  cdrom_multisession structure containing desired
+	  format.
+
+	outputs:
+	  cdrom_multisession structure is filled with last_session
+	  information.
+
+	error return:
+	  EINVAL	format not CDROM_MSF or CDROM_LBA
+
+
+CDROM_GET_MCN			Obtain the "Universal Product Code"
+				   if available (struct cdrom_mcn)
+
+	usage:
+
+	  struct cdrom_mcn mcn;
+	  ioctl(fd, CDROM_GET_MCN, &mcn);
+
+	inputs:		none
+
+	outputs:
+	  Universal Product Code
+
+	error return:
+	  ENOSYS	Drive is not capable of reading MCN data.
+
+	notes:
+	  Source code comments state:
+
+	    The following function is implemented, although very few
+	    audio discs give Universal Product Code information, which
+	    should just be the Medium Catalog Number on the box.  Note,
+	    that the way the code is written on the CD is /not/ uniform
+	    across all discs!
+
+
+
+
+CDROM_GET_UPC			CDROM_GET_MCN  (deprecated)
+
+	Not implemented, as of 2.6.8.1
+
+
+
+CDROMRESET			hard-reset the drive
+
+	usage:
+
+	  ioctl(fd, CDROMRESET, 0);
+
+	inputs:		none
+
+	outputs:	none
+
+	error return:
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  ENOSYS	Drive is not capable of resetting.
+
+
+
+
+CDROMREADCOOKED			read data in cooked mode
+
+	usage:
+
+	  u8 buffer[CD_FRAMESIZE]
+	  ioctl(fd, CDROMREADCOOKED, buffer);
+
+	inputs:		none
+
+	outputs:
+	  2048 bytes of data, "cooked" mode.
+
+	notes:
+	  Not implemented on all drives.
+
+
+
+
+CDROMREADALL			read all 2646 bytes
+
+	Same as CDROMREADCOOKED, but reads 2646 bytes.
+
+
+
+CDROMSEEK			seek msf address
+
+	usage:
+
+	  struct cdrom_msf msf;
+	  ioctl(fd, CDROMSEEK, &msf);
+
+	inputs:
+	  MSF address to seek to.
+
+	outputs:	none
+
+
+
+CDROMPLAYBLK			scsi-cd only, (struct cdrom_blk)
+
+	usage:
+
+	  struct cdrom_blk blk;
+	  ioctl(fd, CDROMPLAYBLK, &blk);
+
+	inputs:
+	  Region to play
+
+	outputs:	none
+
+
+
+CDROMGETSPINDOWN
+
+	usage:
+
+	  char spindown;
+	  ioctl(fd, CDROMGETSPINDOWN, &spindown);
+
+	inputs:		none
+
+	outputs:
+	  The value of the current 4-bit spindown value.
+
+
+
+
+CDROMSETSPINDOWN
+
+	usage:
+
+	  char spindown
+	  ioctl(fd, CDROMSETSPINDOWN, &spindown);
+
+	inputs:
+	  4-bit value used to control spindown (TODO: more detail here)
+
+	outputs:	none
+
+
+
+
+
+CDROM_SET_OPTIONS		Set behavior options
+
+	usage:
+
+	  int options;
+	  ioctl(fd, CDROM_SET_OPTIONS, options);
+
+	inputs:
+	  New values for drive options.  The logical 'or' of:
+	    CDO_AUTO_CLOSE	close tray on first open(2)
+	    CDO_AUTO_EJECT	open tray on last release
+	    CDO_USE_FFLAGS	use O_NONBLOCK information on open
+	    CDO_LOCK		lock tray on open files
+	    CDO_CHECK_TYPE	check type on open for data
+
+	outputs:
+	  Returns the resulting options settings in the
+	  ioctl return value.  Returns -1 on error.
+
+	error return:
+	  ENOSYS	selected option(s) not supported by drive.
+
+
+
+
+CDROM_CLEAR_OPTIONS		Clear behavior options
+
+	Same as CDROM_SET_OPTIONS, except that selected options are
+	turned off.
+
+
+
+CDROM_SELECT_SPEED		Set the CD-ROM speed
+
+	usage:
+
+	  int speed;
+	  ioctl(fd, CDROM_SELECT_SPEED, speed);
+
+	inputs:
+	  New drive speed.
+
+	outputs:	none
+
+	error return:
+	  ENOSYS	speed selection not supported by drive.
+
+
+
+CDROM_SELECT_DISC		Select disc (for juke-boxes)
+
+	usage:
+
+	  int disk;
+	  ioctl(fd, CDROM_SELECT_DISC, disk);
+
+	inputs:
+	  Disk to load into drive.
+
+	outputs:	none
+
+	error return:
+	  EINVAL	Disk number beyond capacity of drive
+
+
+
+CDROM_MEDIA_CHANGED		Check is media changed
+
+	usage:
+
+	  int slot;
+	  ioctl(fd, CDROM_MEDIA_CHANGED, slot);
+
+	inputs:
+	  Slot number to be tested, always zero except for jukeboxes.
+	  May also be special values CDSL_NONE or CDSL_CURRENT
+
+	outputs:
+	  Ioctl return value is 0 or 1 depending on whether the media
+	  has been changed, or -1 on error.
+
+	error returns:
+	  ENOSYS	Drive can't detect media change
+	  EINVAL	Slot number beyond capacity of drive
+	  ENOMEM	Out of memory
+
+
+
+CDROM_DRIVE_STATUS		Get tray position, etc.
+
+	usage:
+
+	  int slot;
+	  ioctl(fd, CDROM_DRIVE_STATUS, slot);
+
+	inputs:
+	  Slot number to be tested, always zero except for jukeboxes.
+	  May also be special values CDSL_NONE or CDSL_CURRENT
+
+	outputs:
+	  Ioctl return value will be one of the following values
+	  from <linux/cdrom.h>:
+
+	    CDS_NO_INFO		Information not available.
+	    CDS_NO_DISC
+	    CDS_TRAY_OPEN
+	    CDS_DRIVE_NOT_READY
+	    CDS_DISC_OK
+	    -1			error
+
+	error returns:
+	  ENOSYS	Drive can't detect drive status
+	  EINVAL	Slot number beyond capacity of drive
+	  ENOMEM	Out of memory
+
+
+
+
+CDROM_DISC_STATUS		Get disc type, etc.
+
+	usage:
+
+	  ioctl(fd, CDROM_DISC_STATUS, 0);
+
+	inputs:		none
+
+	outputs:
+	  Ioctl return value will be one of the following values
+	  from <linux/cdrom.h>:
+	    CDS_NO_INFO
+	    CDS_AUDIO
+	    CDS_MIXED
+	    CDS_XA_2_2
+	    CDS_XA_2_1
+	    CDS_DATA_1
+
+	error returns:	none at present
+
+	notes:
+	  Source code comments state:
+
+	    Ok, this is where problems start.  The current interface for
+	    the CDROM_DISC_STATUS ioctl is flawed.  It makes the false
+	    assumption that CDs are all CDS_DATA_1 or all CDS_AUDIO, etc.
+	    Unfortunately, while this is often the case, it is also
+	    very common for CDs to have some tracks with data, and some
+	    tracks with audio.	Just because I feel like it, I declare
+	    the following to be the best way to cope.  If the CD has
+	    ANY data tracks on it, it will be returned as a data CD.
+	    If it has any XA tracks, I will return it as that.	Now I
+	    could simplify this interface by combining these returns with
+	    the above, but this more clearly demonstrates the problem
+	    with the current interface.  Too bad this wasn't designed
+	    to use bitmasks...	       -Erik
+
+	    Well, now we have the option CDS_MIXED: a mixed-type CD.
+	    User level programmers might feel the ioctl is not very
+	    useful.
+			---david
+
+
+
+
+CDROM_CHANGER_NSLOTS		Get number of slots
+
+	usage:
+
+	  ioctl(fd, CDROM_CHANGER_NSLOTS, 0);
+
+	inputs:		none
+
+	outputs:
+	  The ioctl return value will be the number of slots in a
+	  CD changer.  Typically 1 for non-multi-disk devices.
+
+	error returns:	none
+
+
+
+CDROM_LOCKDOOR			lock or unlock door
+
+	usage:
+
+	  int lock;
+	  ioctl(fd, CDROM_LOCKDOOR, lock);
+
+	inputs:
+	  Door lock flag, 1=lock, 0=unlock
+
+	outputs:	none
+
+	error returns:
+	  EDRIVE_CANT_DO_THIS	Door lock function not supported.
+	  EBUSY			Attempt to unlock when multiple users
+	  			have the drive open and not CAP_SYS_ADMIN
+
+	notes:
+	  As of 2.6.8.1, the lock flag is a global lock, meaning that
+	  all CD drives will be locked or unlocked together.  This is
+	  probably a bug.
+
+	  The EDRIVE_CANT_DO_THIS value is defined in <linux/cdrom.h>
+	  and is currently (2.6.8.1) the same as EOPNOTSUPP
+
+
+
+CDROM_DEBUG			Turn debug messages on/off
+
+	usage:
+
+	  int debug;
+	  ioctl(fd, CDROM_DEBUG, debug);
+
+	inputs:
+	  Cdrom debug flag, 0=disable, 1=enable
+
+	outputs:
+	  The ioctl return value will be the new debug flag.
+
+	error return:
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+
+
+
+CDROM_GET_CAPABILITY		get capabilities
+
+	usage:
+
+	  ioctl(fd, CDROM_GET_CAPABILITY, 0);
+
+	inputs:		none
+
+	outputs:
+	  The ioctl return value is the current device capability
+	  flags.  See CDC_CLOSE_TRAY, CDC_OPEN_TRAY, etc.
+
+
+
+CDROMAUDIOBUFSIZ		set the audio buffer size
+
+	usage:
+
+	  int arg;
+	  ioctl(fd, CDROMAUDIOBUFSIZ, val);
+
+	inputs:
+	  New audio buffer size
+
+	outputs:
+	  The ioctl return value is the new audio buffer size, or -1
+	  on error.
+
+	error return:
+	  ENOSYS	Not supported by this driver.
+
+	notes:
+	  Not supported by all drivers.
+
+
+
+DVD_READ_STRUCT			Read structure
+
+	usage:
+
+	  dvd_struct s;
+	  ioctl(fd, DVD_READ_STRUCT, &s);
+
+	inputs:
+	  dvd_struct structure, containing:
+	    type		specifies the information desired, one of
+	    			DVD_STRUCT_PHYSICAL, DVD_STRUCT_COPYRIGHT,
+				DVD_STRUCT_DISCKEY, DVD_STRUCT_BCA,
+				DVD_STRUCT_MANUFACT
+	    physical.layer_num	desired layer, indexed from 0
+	    copyright.layer_num	desired layer, indexed from 0
+	    disckey.agid
+
+	outputs:
+	  dvd_struct structure, containing:
+	    physical		for type == DVD_STRUCT_PHYSICAL
+	    copyright		for type == DVD_STRUCT_COPYRIGHT
+	    disckey.value	for type == DVD_STRUCT_DISCKEY
+	    bca.{len,value}	for type == DVD_STRUCT_BCA
+	    manufact.{len,valu}	for type == DVD_STRUCT_MANUFACT
+
+	error returns:
+	  EINVAL	physical.layer_num exceeds number of layers
+	  EIO		Received invalid response from drive
+
+
+
+DVD_WRITE_STRUCT		Write structure
+
+	Not implemented, as of 2.6.8.1
+
+
+
+DVD_AUTH			Authentication
+
+	usage:
+
+	  dvd_authinfo ai;
+	  ioctl(fd, DVD_AUTH, &ai);
+
+	inputs:
+	  dvd_authinfo structure.  See <linux/cdrom.h>
+
+	outputs:
+	  dvd_authinfo structure.
+
+	error return:
+	  ENOTTY	ai.type not recognized.
+
+
+
+CDROM_SEND_PACKET		send a packet to the drive
+
+	usage:
+
+	  struct cdrom_generic_command cgc;
+	  ioctl(fd, CDROM_SEND_PACKET, &cgc);
+
+	inputs:
+	  cdrom_generic_command structure containing the packet to send.
+
+	outputs:	none
+	  cdrom_generic_command structure containing results.
+
+	error return:
+	  EIO		command failed.
+	  EPERM		Operation not permitted, either because a
+			write command was attempted on a drive which
+			is opened read-only, or because the command
+			requires CAP_SYS_RAWIO
+	  EINVAL	cgc.data_direction not set
+
+
+
+CDROM_NEXT_WRITABLE		get next writable block
+
+	usage:
+
+	  long next;
+	  ioctl(fd, CDROM_NEXT_WRITABLE, &next);
+
+	inputs:		none
+
+	outputs:
+	  The next writable block.
+
+	notes:
+	  If the device does not support this ioctl directly, the
+	  ioctl will return CDROM_LAST_WRITTEN + 7.
+
+
+
+CDROM_LAST_WRITTEN		get last block written on disc
+
+	usage:
+
+	  long last;
+	  ioctl(fd, CDROM_LAST_WRITTEN, &last);
+
+	inputs:		none
+
+	outputs:
+	  The last block written on disc
+
+	notes:
+	  If the device does not support this ioctl directly, the
+	  result is derived from the disc's table of contents.  If the
+	  table of contents can't be read, this ioctl returns an
+	  error.
diff --git a/Documentation/ioctl/hdio.txt b/Documentation/ioctl/hdio.txt
new file mode 100644
index 000000000..18eb98c44
--- /dev/null
+++ b/Documentation/ioctl/hdio.txt
@@ -0,0 +1,1071 @@
+		Summary of HDIO_ ioctl calls.
+		============================
+
+		Edward A. Falk <efalk@google.com>
+
+		November, 2004
+
+This document attempts to describe the ioctl(2) calls supported by
+the HD/IDE layer.  These are by-and-large implemented (as of Linux 2.6)
+in drivers/ide/ide.c and drivers/block/scsi_ioctl.c
+
+ioctl values are listed in <linux/hdreg.h>.  As of this writing, they
+are as follows:
+
+    ioctls that pass argument pointers to user space:
+
+	HDIO_GETGEO		get device geometry
+	HDIO_GET_UNMASKINTR	get current unmask setting
+	HDIO_GET_MULTCOUNT	get current IDE blockmode setting
+	HDIO_GET_QDMA		get use-qdma flag
+	HDIO_SET_XFER		set transfer rate via proc
+	HDIO_OBSOLETE_IDENTITY	OBSOLETE, DO NOT USE
+	HDIO_GET_KEEPSETTINGS	get keep-settings-on-reset flag
+	HDIO_GET_32BIT		get current io_32bit setting
+	HDIO_GET_NOWERR		get ignore-write-error flag
+	HDIO_GET_DMA		get use-dma flag
+	HDIO_GET_NICE		get nice flags
+	HDIO_GET_IDENTITY	get IDE identification info
+	HDIO_GET_WCACHE		get write cache mode on|off
+	HDIO_GET_ACOUSTIC	get acoustic value
+	HDIO_GET_ADDRESS	get sector addressing mode
+	HDIO_GET_BUSSTATE	get the bus state of the hwif
+	HDIO_TRISTATE_HWIF	execute a channel tristate
+	HDIO_DRIVE_RESET	execute a device reset
+	HDIO_DRIVE_TASKFILE	execute raw taskfile
+	HDIO_DRIVE_TASK		execute task and special drive command
+	HDIO_DRIVE_CMD		execute a special drive command
+	HDIO_DRIVE_CMD_AEB	HDIO_DRIVE_TASK
+
+    ioctls that pass non-pointer values:
+
+	HDIO_SET_MULTCOUNT	change IDE blockmode
+	HDIO_SET_UNMASKINTR	permit other irqs during I/O
+	HDIO_SET_KEEPSETTINGS	keep ioctl settings on reset
+	HDIO_SET_32BIT		change io_32bit flags
+	HDIO_SET_NOWERR		change ignore-write-error flag
+	HDIO_SET_DMA		change use-dma flag
+	HDIO_SET_PIO_MODE	reconfig interface to new speed
+	HDIO_SCAN_HWIF		register and (re)scan interface
+	HDIO_SET_NICE		set nice flags
+	HDIO_UNREGISTER_HWIF	unregister interface
+	HDIO_SET_WCACHE		change write cache enable-disable
+	HDIO_SET_ACOUSTIC	change acoustic behavior
+	HDIO_SET_BUSSTATE	set the bus state of the hwif
+	HDIO_SET_QDMA		change use-qdma flag
+	HDIO_SET_ADDRESS	change lba addressing modes
+
+	HDIO_SET_IDE_SCSI	Set scsi emulation mode on/off
+	HDIO_SET_SCSI_IDE	not implemented yet
+
+
+The information that follows was determined from reading kernel source
+code.  It is likely that some corrections will be made over time.
+
+
+
+
+
+
+
+General:
+
+	Unless otherwise specified, all ioctl calls return 0 on success
+	and -1 with errno set to an appropriate value on error.
+
+	Unless otherwise specified, all ioctl calls return -1 and set
+	errno to EFAULT on a failed attempt to copy data to or from user
+	address space.
+
+	Unless otherwise specified, all data structures and constants
+	are defined in <linux/hdreg.h>
+
+
+
+HDIO_GETGEO			get device geometry
+
+	usage:
+
+	  struct hd_geometry geom;
+	  ioctl(fd, HDIO_GETGEO, &geom);
+
+
+	inputs:		none
+
+	outputs:
+
+	  hd_geometry structure containing:
+
+	    heads	number of heads
+	    sectors	number of sectors/track
+	    cylinders	number of cylinders, mod 65536
+	    start	starting sector of this partition.
+
+
+	error returns:
+	  EINVAL	if the device is not a disk drive or floppy drive,
+	  		or if the user passes a null pointer
+
+
+	notes:
+
+	  Not particularly useful with modern disk drives, whose geometry
+	  is a polite fiction anyway.  Modern drives are addressed
+	  purely by sector number nowadays (lba addressing), and the
+	  drive geometry is an abstraction which is actually subject
+	  to change.  Currently (as of Nov 2004), the geometry values
+	  are the "bios" values -- presumably the values the drive had
+	  when Linux first booted.
+
+	  In addition, the cylinders field of the hd_geometry is an
+	  unsigned short, meaning that on most architectures, this
+	  ioctl will not return a meaningful value on drives with more
+	  than 65535 tracks.
+
+	  The start field is unsigned long, meaning that it will not
+	  contain a meaningful value for disks over 219 Gb in size.
+
+
+
+
+HDIO_GET_UNMASKINTR		get current unmask setting
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_GET_UNMASKINTR, &val);
+
+	inputs:		none
+
+	outputs:
+	  The value of the drive's current unmask setting
+
+
+
+HDIO_SET_UNMASKINTR		permit other irqs during I/O
+
+	usage:
+
+	  unsigned long val;
+	  ioctl(fd, HDIO_SET_UNMASKINTR, val);
+
+	inputs:
+	  New value for unmask flag
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 1]
+	  EBUSY		Controller busy
+
+
+
+
+HDIO_GET_MULTCOUNT		get current IDE blockmode setting
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_GET_MULTCOUNT, &val);
+
+	inputs:		none
+
+	outputs:
+	  The value of the current IDE block mode setting.  This
+	  controls how many sectors the drive will transfer per
+	  interrupt.
+
+
+
+HDIO_SET_MULTCOUNT		change IDE blockmode
+
+	usage:
+
+	  int val;
+	  ioctl(fd, HDIO_SET_MULTCOUNT, val);
+
+	inputs:
+	  New value for IDE block mode setting.  This controls how many
+	  sectors the drive will transfer per interrupt.
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range supported by disk.
+	  EBUSY		Controller busy or blockmode already set.
+	  EIO		Drive did not accept new block mode.
+
+	notes:
+
+	  Source code comments read:
+
+	    This is tightly woven into the driver->do_special cannot
+	    touch.  DON'T do it again until a total personality rewrite
+	    is committed.
+
+	  If blockmode has already been set, this ioctl will fail with
+	  EBUSY
+
+
+
+HDIO_GET_QDMA			get use-qdma flag
+
+	Not implemented, as of 2.6.8.1
+
+
+
+HDIO_SET_XFER			set transfer rate via proc
+
+	Not implemented, as of 2.6.8.1
+
+
+
+HDIO_OBSOLETE_IDENTITY		OBSOLETE, DO NOT USE
+
+	Same as HDIO_GET_IDENTITY (see below), except that it only
+	returns the first 142 bytes of drive identity information.
+
+
+
+HDIO_GET_IDENTITY		get IDE identification info
+
+	usage:
+
+	  unsigned char identity[512];
+	  ioctl(fd, HDIO_GET_IDENTITY, identity);
+
+	inputs:		none
+
+	outputs:
+
+	  ATA drive identity information.  For full description, see
+	  the IDENTIFY DEVICE and IDENTIFY PACKET DEVICE commands in
+	  the ATA specification.
+
+	error returns:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  ENOMSG	IDENTIFY DEVICE information not available
+
+	notes:
+
+	  Returns information that was obtained when the drive was
+	  probed.  Some of this information is subject to change, and
+	  this ioctl does not re-probe the drive to update the
+	  information.
+
+	  This information is also available from /proc/ide/hdX/identify
+
+
+
+HDIO_GET_KEEPSETTINGS		get keep-settings-on-reset flag
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_GET_KEEPSETTINGS, &val);
+
+	inputs:		none
+
+	outputs:
+	  The value of the current "keep settings" flag
+
+	notes:
+
+	  When set, indicates that kernel should restore settings
+	  after a drive reset.
+
+
+
+HDIO_SET_KEEPSETTINGS		keep ioctl settings on reset
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_SET_KEEPSETTINGS, val);
+
+	inputs:
+	  New value for keep_settings flag
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 1]
+	  EBUSY		Controller busy
+
+
+
+HDIO_GET_32BIT			get current io_32bit setting
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_GET_32BIT, &val);
+
+	inputs:		none
+
+	outputs:
+	  The value of the current io_32bit setting
+
+	notes:
+
+	  0=16-bit, 1=32-bit, 2,3 = 32bit+sync
+
+
+
+HDIO_GET_NOWERR			get ignore-write-error flag
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_GET_NOWERR, &val);
+
+	inputs:		none
+
+	outputs:
+	  The value of the current ignore-write-error flag
+
+
+
+HDIO_GET_DMA			get use-dma flag
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_GET_DMA, &val);
+
+	inputs:		none
+
+	outputs:
+	  The value of the current use-dma flag
+
+
+
+HDIO_GET_NICE			get nice flags
+
+	usage:
+
+	  long nice;
+	  ioctl(fd, HDIO_GET_NICE, &nice);
+
+	inputs:		none
+
+	outputs:
+
+	  The drive's "nice" values.
+
+	notes:
+
+	  Per-drive flags which determine when the system will give more
+	  bandwidth to other devices sharing the same IDE bus.
+	  See <linux/hdreg.h>, near symbol IDE_NICE_DSC_OVERLAP.
+
+
+
+
+HDIO_SET_NICE			set nice flags
+
+	usage:
+
+	  unsigned long nice;
+	  ...
+	  ioctl(fd, HDIO_SET_NICE, nice);
+
+	inputs:
+	  bitmask of nice flags.
+
+	outputs:	none
+
+	error returns:
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EPERM		Flags other than DSC_OVERLAP and NICE_1 set.
+	  EPERM		DSC_OVERLAP specified but not supported by drive
+
+	notes:
+
+	  This ioctl sets the DSC_OVERLAP and NICE_1 flags from values
+	  provided by the user.
+
+	  Nice flags are listed in <linux/hdreg.h>, starting with
+	  IDE_NICE_DSC_OVERLAP.  These values represent shifts.
+
+
+
+
+
+HDIO_GET_WCACHE			get write cache mode on|off
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_GET_WCACHE, &val);
+
+	inputs:		none
+
+	outputs:
+	  The value of the current write cache mode
+
+
+
+HDIO_GET_ACOUSTIC		get acoustic value
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_GET_ACOUSTIC, &val);
+
+	inputs:		none
+
+	outputs:
+	  The value of the current acoustic settings
+
+	notes:
+
+	  See HDIO_SET_ACOUSTIC
+
+
+
+HDIO_GET_ADDRESS
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_GET_ADDRESS, &val);
+
+	inputs:		none
+
+	outputs:
+	  The value of the current addressing mode:
+	    0 = 28-bit
+	    1 = 48-bit
+	    2 = 48-bit doing 28-bit
+	    3 = 64-bit
+
+
+
+HDIO_GET_BUSSTATE		get the bus state of the hwif
+
+	usage:
+
+	  long state;
+	  ioctl(fd, HDIO_SCAN_HWIF, &state);
+
+	inputs:		none
+
+	outputs:
+	  Current power state of the IDE bus.  One of BUSSTATE_OFF,
+	  BUSSTATE_ON, or BUSSTATE_TRISTATE
+
+	error returns:
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+
+
+
+
+HDIO_SET_BUSSTATE		set the bus state of the hwif
+
+	usage:
+
+	  int state;
+	  ...
+	  ioctl(fd, HDIO_SCAN_HWIF, state);
+
+	inputs:
+	  Desired IDE power state.  One of BUSSTATE_OFF, BUSSTATE_ON,
+	  or BUSSTATE_TRISTATE
+
+	outputs:	none
+
+	error returns:
+	  EACCES	Access denied:  requires CAP_SYS_RAWIO
+	  EOPNOTSUPP	Hardware interface does not support bus power control
+
+
+
+
+HDIO_TRISTATE_HWIF		execute a channel tristate
+
+	Not implemented, as of 2.6.8.1.  See HDIO_SET_BUSSTATE
+
+
+
+HDIO_DRIVE_RESET		execute a device reset
+
+	usage:
+
+	  int args[3]
+	  ...
+	  ioctl(fd, HDIO_DRIVE_RESET, args);
+
+	inputs:		none
+
+	outputs:	none
+
+	error returns:
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  ENXIO		No such device:	phy dead or ctl_addr == 0
+	  EIO		I/O error:	reset timed out or hardware error
+
+	notes:
+
+	  Execute a reset on the device as soon as the current IO
+	  operation has completed.
+
+	  Executes an ATAPI soft reset if applicable, otherwise
+	  executes an ATA soft reset on the controller.
+
+
+
+HDIO_DRIVE_TASKFILE		execute raw taskfile
+
+	Note:  If you don't have a copy of the ANSI ATA specification
+	handy, you should probably ignore this ioctl.
+
+	Execute an ATA disk command directly by writing the "taskfile"
+	registers of the drive.  Requires ADMIN and RAWIO access
+	privileges.
+
+	usage:
+
+	  struct {
+	    ide_task_request_t req_task;
+	    u8 outbuf[OUTPUT_SIZE];
+	    u8 inbuf[INPUT_SIZE];
+	  } task;
+	  memset(&task.req_task, 0, sizeof(task.req_task));
+	  task.req_task.out_size = sizeof(task.outbuf);
+	  task.req_task.in_size = sizeof(task.inbuf);
+	  ...
+	  ioctl(fd, HDIO_DRIVE_TASKFILE, &task);
+	  ...
+
+	inputs:
+
+	  (See below for details on memory area passed to ioctl.)
+
+	  io_ports[8]	values to be written to taskfile registers
+	  hob_ports[8]	high-order bytes, for extended commands.
+	  out_flags	flags indicating which registers are valid
+	  in_flags	flags indicating which registers should be returned
+	  data_phase	see below
+	  req_cmd	command type to be executed
+	  out_size	size of output buffer
+	  outbuf	buffer of data to be transmitted to disk
+	  inbuf		buffer of data to be received from disk (see [1])
+
+	outputs:
+
+	  io_ports[]	values returned in the taskfile registers
+	  hob_ports[]	high-order bytes, for extended commands.
+	  out_flags	flags indicating which registers are valid (see [2])
+	  in_flags	flags indicating which registers should be returned
+	  outbuf	buffer of data to be transmitted to disk (see [1])
+	  inbuf		buffer of data to be received from disk
+
+	error returns:
+	  EACCES	CAP_SYS_ADMIN or CAP_SYS_RAWIO privilege not set.
+	  ENOMSG	Device is not a disk drive.
+	  ENOMEM	Unable to allocate memory for task
+	  EFAULT	req_cmd == TASKFILE_IN_OUT (not implemented as of 2.6.8)
+	  EPERM		req_cmd == TASKFILE_MULTI_OUT and drive
+	  		multi-count not yet set.
+	  EIO		Drive failed the command.
+
+	notes:
+
+	  [1] READ THE FOLLOWING NOTES *CAREFULLY*.  THIS IOCTL IS
+	  FULL OF GOTCHAS.  Extreme caution should be used with using
+	  this ioctl.  A mistake can easily corrupt data or hang the
+	  system.
+
+	  [2] Both the input and output buffers are copied from the
+	  user and written back to the user, even when not used.
+
+	  [3] If one or more bits are set in out_flags and in_flags is
+	  zero, the following values are used for in_flags.all and
+	  written back into in_flags on completion.
+
+	   * IDE_TASKFILE_STD_IN_FLAGS | (IDE_HOB_STD_IN_FLAGS << 8)
+	     if LBA48 addressing is enabled for the drive
+	   * IDE_TASKFILE_STD_IN_FLAGS
+	     if CHS/LBA28
+
+	  The association between in_flags.all and each enable
+	  bitfield flips depending on endianness; fortunately, TASKFILE
+	  only uses inflags.b.data bit and ignores all other bits.
+	  The end result is that, on any endian machines, it has no
+	  effect other than modifying in_flags on completion.
+
+	  [4] The default value of SELECT is (0xa0|DEV_bit|LBA_bit)
+	  except for four drives per port chipsets.  For four drives
+	  per port chipsets, it's (0xa0|DEV_bit|LBA_bit) for the first
+	  pair and (0x80|DEV_bit|LBA_bit) for the second pair.
+
+	  [5] The argument to the ioctl is a pointer to a region of
+	  memory containing a ide_task_request_t structure, followed
+	  by an optional buffer of data to be transmitted to the
+	  drive, followed by an optional buffer to receive data from
+	  the drive.
+
+	  Command is passed to the disk drive via the ide_task_request_t
+	  structure, which contains these fields:
+
+	    io_ports[8]		values for the taskfile registers
+	    hob_ports[8]	high-order bytes, for extended commands
+	    out_flags		flags indicating which entries in the
+	    			io_ports[] and hob_ports[] arrays
+				contain valid values.  Type ide_reg_valid_t.
+	    in_flags		flags indicating which entries in the
+	    			io_ports[] and hob_ports[] arrays
+				are expected to contain valid values
+				on return.
+	    data_phase		See below
+	    req_cmd		Command type, see below
+	    out_size		output (user->drive) buffer size, bytes
+	    in_size		input (drive->user) buffer size, bytes
+
+	  When out_flags is zero, the following registers are loaded.
+
+	    HOB_FEATURE		If the drive supports LBA48
+	    HOB_NSECTOR		If the drive supports LBA48
+	    HOB_SECTOR		If the drive supports LBA48
+	    HOB_LCYL		If the drive supports LBA48
+	    HOB_HCYL		If the drive supports LBA48
+	    FEATURE
+	    NSECTOR
+	    SECTOR
+	    LCYL
+	    HCYL
+	    SELECT		First, masked with 0xE0 if LBA48, 0xEF
+				otherwise; then, or'ed with the default
+				value of SELECT.
+
+	  If any bit in out_flags is set, the following registers are loaded.
+
+	    HOB_DATA		If out_flags.b.data is set.  HOB_DATA will
+				travel on DD8-DD15 on little endian machines
+				and on DD0-DD7 on big endian machines.
+	    DATA		If out_flags.b.data is set.  DATA will
+				travel on DD0-DD7 on little endian machines
+				and on DD8-DD15 on big endian machines.
+	    HOB_NSECTOR		If out_flags.b.nsector_hob is set
+	    HOB_SECTOR		If out_flags.b.sector_hob is set
+	    HOB_LCYL		If out_flags.b.lcyl_hob is set
+	    HOB_HCYL		If out_flags.b.hcyl_hob is set
+	    FEATURE		If out_flags.b.feature is set
+	    NSECTOR		If out_flags.b.nsector is set
+	    SECTOR		If out_flags.b.sector is set
+	    LCYL		If out_flags.b.lcyl is set
+	    HCYL		If out_flags.b.hcyl is set
+	    SELECT		Or'ed with the default value of SELECT and
+				loaded regardless of out_flags.b.select.
+
+	  Taskfile registers are read back from the drive into
+	  {io|hob}_ports[] after the command completes iff one of the
+	  following conditions is met; otherwise, the original values
+	  will be written back, unchanged.
+
+	    1. The drive fails the command (EIO).
+	    2. One or more than one bits are set in out_flags.
+	    3. The requested data_phase is TASKFILE_NO_DATA.
+
+	    HOB_DATA		If in_flags.b.data is set.  It will contain
+				DD8-DD15 on little endian machines and
+				DD0-DD7 on big endian machines.
+	    DATA		If in_flags.b.data is set.  It will contain
+				DD0-DD7 on little endian machines and
+				DD8-DD15 on big endian machines.
+	    HOB_FEATURE		If the drive supports LBA48
+	    HOB_NSECTOR		If the drive supports LBA48
+	    HOB_SECTOR		If the drive supports LBA48
+	    HOB_LCYL		If the drive supports LBA48
+	    HOB_HCYL		If the drive supports LBA48
+	    NSECTOR
+	    SECTOR
+	    LCYL
+	    HCYL
+
+	  The data_phase field describes the data transfer to be
+	  performed.  Value is one of:
+
+	    TASKFILE_IN
+	    TASKFILE_MULTI_IN
+	    TASKFILE_OUT
+	    TASKFILE_MULTI_OUT
+	    TASKFILE_IN_OUT
+	    TASKFILE_IN_DMA
+	    TASKFILE_IN_DMAQ		== IN_DMA (queueing not supported)
+	    TASKFILE_OUT_DMA
+	    TASKFILE_OUT_DMAQ		== OUT_DMA (queueing not supported)
+	    TASKFILE_P_IN		unimplemented
+	    TASKFILE_P_IN_DMA		unimplemented
+	    TASKFILE_P_IN_DMAQ		unimplemented
+	    TASKFILE_P_OUT		unimplemented
+	    TASKFILE_P_OUT_DMA		unimplemented
+	    TASKFILE_P_OUT_DMAQ		unimplemented
+
+	  The req_cmd field classifies the command type.  It may be
+	  one of:
+
+	    IDE_DRIVE_TASK_NO_DATA
+	    IDE_DRIVE_TASK_SET_XFER	unimplemented
+	    IDE_DRIVE_TASK_IN
+	    IDE_DRIVE_TASK_OUT		unimplemented
+	    IDE_DRIVE_TASK_RAW_WRITE
+
+	  [6] Do not access {in|out}_flags->all except for resetting
+	  all the bits.  Always access individual bit fields.  ->all
+	  value will flip depending on endianness.  For the same
+	  reason, do not use IDE_{TASKFILE|HOB}_STD_{OUT|IN}_FLAGS
+	  constants defined in hdreg.h.
+
+
+
+HDIO_DRIVE_CMD			execute a special drive command
+
+	Note:  If you don't have a copy of the ANSI ATA specification
+	handy, you should probably ignore this ioctl.
+
+	usage:
+
+	  u8 args[4+XFER_SIZE];
+	  ...
+	  ioctl(fd, HDIO_DRIVE_CMD, args);
+
+	inputs:
+
+	  Commands other than WIN_SMART
+	    args[0]	COMMAND
+	    args[1]	NSECTOR
+	    args[2]	FEATURE
+	    args[3]	NSECTOR
+
+	  WIN_SMART
+	    args[0]	COMMAND
+	    args[1]	SECTOR
+	    args[2]	FEATURE
+	    args[3]	NSECTOR
+
+	outputs:
+
+	  args[] buffer is filled with register values followed by any
+	  data returned by the disk.
+	    args[0]	status
+	    args[1]	error
+	    args[2]	NSECTOR
+	    args[3]	undefined
+	    args[4+]	NSECTOR * 512 bytes of data returned by the command.
+
+	error returns:
+	  EACCES	Access denied:  requires CAP_SYS_RAWIO
+	  ENOMEM	Unable to allocate memory for task
+	  EIO		Drive reports error
+
+	notes:
+
+	  [1] For commands other than WIN_SMART, args[1] should equal
+	  args[3].  SECTOR, LCYL and HCYL are undefined.  For
+	  WIN_SMART, 0x4f and 0xc2 are loaded into LCYL and HCYL
+	  respectively.  In both cases SELECT will contain the default
+	  value for the drive.  Please refer to HDIO_DRIVE_TASKFILE
+	  notes for the default value of SELECT.
+
+	  [2] If NSECTOR value is greater than zero and the drive sets
+	  DRQ when interrupting for the command, NSECTOR * 512 bytes
+	  are read from the device into the area following NSECTOR.
+	  In the above example, the area would be
+	  args[4..4+XFER_SIZE].  16bit PIO is used regardless of
+	  HDIO_SET_32BIT setting.
+
+	  [3] If COMMAND == WIN_SETFEATURES && FEATURE == SETFEATURES_XFER
+	  && NSECTOR >= XFER_SW_DMA_0 && the drive supports any DMA
+	  mode, IDE driver will try to tune the transfer mode of the
+	  drive accordingly.
+
+
+
+HDIO_DRIVE_TASK			execute task and special drive command
+
+	Note:  If you don't have a copy of the ANSI ATA specification
+	handy, you should probably ignore this ioctl.
+
+	usage:
+
+	  u8 args[7];
+	  ...
+	  ioctl(fd, HDIO_DRIVE_TASK, args);
+
+	inputs:
+
+	  Taskfile register values:
+	    args[0]	COMMAND
+	    args[1]	FEATURE
+	    args[2]	NSECTOR
+	    args[3]	SECTOR
+	    args[4]	LCYL
+	    args[5]	HCYL
+	    args[6]	SELECT
+
+	outputs:
+
+	  Taskfile register values:
+	    args[0]	status
+	    args[1]	error
+	    args[2]	NSECTOR
+	    args[3]	SECTOR
+	    args[4]	LCYL
+	    args[5]	HCYL
+	    args[6]	SELECT
+
+	error returns:
+	  EACCES	Access denied:  requires CAP_SYS_RAWIO
+	  ENOMEM	Unable to allocate memory for task
+	  ENOMSG	Device is not a disk drive.
+	  EIO		Drive failed the command.
+
+	notes:
+
+	  [1] DEV bit (0x10) of SELECT register is ignored and the
+	  appropriate value for the drive is used.  All other bits
+	  are used unaltered.
+
+
+
+HDIO_DRIVE_CMD_AEB		HDIO_DRIVE_TASK
+
+	Not implemented, as of 2.6.8.1
+
+
+
+HDIO_SET_32BIT			change io_32bit flags
+
+	usage:
+
+	  int val;
+	  ioctl(fd, HDIO_SET_32BIT, val);
+
+	inputs:
+	  New value for io_32bit flag
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 3]
+	  EBUSY		Controller busy
+
+
+
+
+HDIO_SET_NOWERR			change ignore-write-error flag
+
+	usage:
+
+	  int val;
+	  ioctl(fd, HDIO_SET_NOWERR, val);
+
+	inputs:
+	  New value for ignore-write-error flag.  Used for ignoring
+	  WRERR_STAT
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 1]
+	  EBUSY		Controller busy
+
+
+
+HDIO_SET_DMA			change use-dma flag
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_SET_DMA, val);
+
+	inputs:
+	  New value for use-dma flag
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 1]
+	  EBUSY		Controller busy
+
+
+
+HDIO_SET_PIO_MODE		reconfig interface to new speed
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_SET_PIO_MODE, val);
+
+	inputs:
+	  New interface speed.
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 255]
+	  EBUSY		Controller busy
+
+
+
+HDIO_SCAN_HWIF			register and (re)scan interface
+
+	usage:
+
+	  int args[3]
+	  ...
+	  ioctl(fd, HDIO_SCAN_HWIF, args);
+
+	inputs:
+	  args[0]	io address to probe
+	  args[1]	control address to probe
+	  args[2]	irq number
+
+	outputs:	none
+
+	error returns:
+	  EACCES	Access denied:  requires CAP_SYS_RAWIO
+	  EIO		Probe failed.
+
+	notes:
+
+	  This ioctl initializes the addresses and irq for a disk
+	  controller, probes for drives, and creates /proc/ide
+	  interfaces as appropriate.
+
+
+
+HDIO_UNREGISTER_HWIF		unregister interface
+
+	usage:
+
+	  int index;
+	  ioctl(fd, HDIO_UNREGISTER_HWIF, index);
+
+	inputs:
+	  index		index of hardware interface to unregister
+
+	outputs:	none
+
+	error returns:
+	  EACCES	Access denied:  requires CAP_SYS_RAWIO
+
+	notes:
+
+	  This ioctl removes a hardware interface from the kernel.
+
+	  Currently (2.6.8) this ioctl silently fails if any drive on
+	  the interface is busy.
+
+
+
+HDIO_SET_WCACHE			change write cache enable-disable
+
+	usage:
+
+	  int val;
+	  ioctl(fd, HDIO_SET_WCACHE, val);
+
+	inputs:
+	  New value for write cache enable
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 1]
+	  EBUSY		Controller busy
+
+
+
+HDIO_SET_ACOUSTIC		change acoustic behavior
+
+	usage:
+
+	  int val;
+	  ioctl(fd, HDIO_SET_ACOUSTIC, val);
+
+	inputs:
+	  New value for drive acoustic settings
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 254]
+	  EBUSY		Controller busy
+
+
+
+HDIO_SET_QDMA			change use-qdma flag
+
+	Not implemented, as of 2.6.8.1
+
+
+
+HDIO_SET_ADDRESS		change lba addressing modes
+
+	usage:
+
+	  int val;
+	  ioctl(fd, HDIO_SET_ADDRESS, val);
+
+	inputs:
+	  New value for addressing mode
+	    0 = 28-bit
+	    1 = 48-bit
+	    2 = 48-bit doing 28-bit
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 2]
+	  EBUSY		Controller busy
+	  EIO		Drive does not support lba48 mode.
+
+
+HDIO_SET_IDE_SCSI
+
+	usage:
+
+	  long val;
+	  ioctl(fd, HDIO_SET_IDE_SCSI, val);
+
+	inputs:
+	  New value for scsi emulation mode (?)
+
+	outputs:	none
+
+	error return:
+	  EINVAL	(bdev != bdev->bd_contains) (not sure what this means)
+	  EACCES	Access denied:  requires CAP_SYS_ADMIN
+	  EINVAL	value out of range [0 1]
+	  EBUSY		Controller busy
+
+
+
+HDIO_SET_SCSI_IDE
+
+	Not implemented, as of 2.6.8.1
+
+
diff --git a/Documentation/ioctl/ioctl-decoding.txt b/Documentation/ioctl/ioctl-decoding.txt
new file mode 100644
index 000000000..e35efb0ce
--- /dev/null
+++ b/Documentation/ioctl/ioctl-decoding.txt
@@ -0,0 +1,24 @@
+To decode a hex IOCTL code:
+
+Most architectures use this generic format, but check
+include/ARCH/ioctl.h for specifics, e.g. powerpc
+uses 3 bits to encode read/write and 13 bits for size.
+
+ bits    meaning
+ 31-30	00 - no parameters: uses _IO macro
+	10 - read: _IOR
+	01 - write: _IOW
+	11 - read/write: _IOWR
+
+ 29-16	size of arguments
+
+ 15-8	ascii character supposedly
+	unique to each driver
+
+ 7-0	function #
+
+
+So for example 0x82187201 is a read with arg length of 0x218,
+character 'r' function 1. Grepping the source reveals this is:
+
+#define VFAT_IOCTL_READDIR_BOTH         _IOR('r', 1, struct dirent [2])
diff --git a/Documentation/ioctl/ioctl-number.txt b/Documentation/ioctl/ioctl-number.txt
new file mode 100644
index 000000000..ec7c81b4f
--- /dev/null
+++ b/Documentation/ioctl/ioctl-number.txt
@@ -0,0 +1,332 @@
+Ioctl Numbers
+19 October 1999
+Michael Elizabeth Chastain
+<mec@shout.net>
+
+If you are adding new ioctl's to the kernel, you should use the _IO
+macros defined in <linux/ioctl.h>:
+
+    _IO    an ioctl with no parameters
+    _IOW   an ioctl with write parameters (copy_from_user)
+    _IOR   an ioctl with read parameters  (copy_to_user)
+    _IOWR  an ioctl with both write and read parameters.
+
+'Write' and 'read' are from the user's point of view, just like the
+system calls 'write' and 'read'.  For example, a SET_FOO ioctl would
+be _IOW, although the kernel would actually read data from user space;
+a GET_FOO ioctl would be _IOR, although the kernel would actually write
+data to user space.
+
+The first argument to _IO, _IOW, _IOR, or _IOWR is an identifying letter
+or number from the table below.  Because of the large number of drivers,
+many drivers share a partial letter with other drivers.
+
+If you are writing a driver for a new device and need a letter, pick an
+unused block with enough room for expansion: 32 to 256 ioctl commands.
+You can register the block by patching this file and submitting the
+patch to Linus Torvalds.  Or you can e-mail me at <mec@shout.net> and
+I'll register one for you.
+
+The second argument to _IO, _IOW, _IOR, or _IOWR is a sequence number
+to distinguish ioctls from each other.  The third argument to _IOW,
+_IOR, or _IOWR is the type of the data going into the kernel or coming
+out of the kernel (e.g.  'int' or 'struct foo').  NOTE!  Do NOT use
+sizeof(arg) as the third argument as this results in your ioctl thinking
+it passes an argument of type size_t.
+
+Some devices use their major number as the identifier; this is OK, as
+long as it is unique.  Some devices are irregular and don't follow any
+convention at all.
+
+Following this convention is good because:
+
+(1) Keeping the ioctl's globally unique helps error checking:
+    if a program calls an ioctl on the wrong device, it will get an
+    error rather than some unexpected behaviour.
+
+(2) The 'strace' build procedure automatically finds ioctl numbers
+    defined with _IO, _IOW, _IOR, or _IOWR.
+
+(3) 'strace' can decode numbers back into useful names when the
+    numbers are unique.
+
+(4) People looking for ioctls can grep for them more easily when
+    this convention is used to define the ioctl numbers.
+
+(5) When following the convention, the driver code can use generic
+    code to copy the parameters between user and kernel space.
+
+This table lists ioctls visible from user land for Linux/x86.  It contains
+most drivers up to 2.6.31, but I know I am missing some.  There has been
+no attempt to list non-X86 architectures or ioctls from drivers/staging/.
+
+Code  Seq#(hex)	Include File		Comments
+========================================================
+0x00	00-1F	linux/fs.h		conflict!
+0x00	00-1F	scsi/scsi_ioctl.h	conflict!
+0x00	00-1F	linux/fb.h		conflict!
+0x00	00-1F	linux/wavefront.h	conflict!
+0x02	all	linux/fd.h
+0x03	all	linux/hdreg.h
+0x04	D2-DC	linux/umsdos_fs.h	Dead since 2.6.11, but don't reuse these.
+0x06	all	linux/lp.h
+0x09	all	linux/raid/md_u.h
+0x10	00-0F	drivers/char/s390/vmcp.h
+0x10	10-1F	arch/s390/include/uapi/sclp_ctl.h
+0x10	20-2F	arch/s390/include/uapi/asm/hypfs.h
+0x12	all	linux/fs.h
+		linux/blkpg.h
+0x1b	all	InfiniBand Subsystem	<http://infiniband.sourceforge.net/>
+0x20	all	drivers/cdrom/cm206.h
+0x22	all	scsi/sg.h
+'#'	00-3F	IEEE 1394 Subsystem	Block for the entire subsystem
+'$'	00-0F	linux/perf_counter.h, linux/perf_event.h
+'&'	00-07	drivers/firewire/nosy-user.h
+'1'	00-1F	<linux/timepps.h>	PPS kit from Ulrich Windl
+					<ftp://ftp.de.kernel.org/pub/linux/daemons/ntp/PPS/>
+'2'	01-04	linux/i2o.h
+'3'	00-0F	drivers/s390/char/raw3270.h	conflict!
+'3'	00-1F	linux/suspend_ioctls.h	conflict!
+		and kernel/power/user.c
+'8'	all				SNP8023 advanced NIC card
+					<mailto:mcr@solidum.com>
+';'	64-7F	linux/vfio.h
+'@'	00-0F	linux/radeonfb.h	conflict!
+'@'	00-0F	drivers/video/aty/aty128fb.c	conflict!
+'A'	00-1F	linux/apm_bios.h	conflict!
+'A'	00-0F	linux/agpgart.h		conflict!
+		and drivers/char/agp/compat_ioctl.h
+'A'	00-7F	sound/asound.h		conflict!
+'B'	00-1F	linux/cciss_ioctl.h	conflict!
+'B'	00-0F	include/linux/pmu.h	conflict!
+'B'	C0-FF				advanced bbus
+					<mailto:maassen@uni-freiburg.de>
+'C'	all	linux/soundcard.h	conflict!
+'C'	01-2F	linux/capi.h		conflict!
+'C'	F0-FF	drivers/net/wan/cosa.h	conflict!
+'D'	all	arch/s390/include/asm/dasd.h
+'D'	40-5F	drivers/scsi/dpt/dtpi_ioctl.h
+'D'	05	drivers/scsi/pmcraid.h
+'E'	all	linux/input.h		conflict!
+'E'	00-0F	xen/evtchn.h		conflict!
+'F'	all	linux/fb.h		conflict!
+'F'	01-02	drivers/scsi/pmcraid.h	conflict!
+'F'	20	drivers/video/fsl-diu-fb.h	conflict!
+'F'	20	drivers/video/intelfb/intelfb.h	conflict!
+'F'	20	linux/ivtvfb.h		conflict!
+'F'	20	linux/matroxfb.h	conflict!
+'F'	20	drivers/video/aty/atyfb_base.c	conflict!
+'F'	00-0F	video/da8xx-fb.h	conflict!
+'F'	80-8F	linux/arcfb.h		conflict!
+'F'	DD	video/sstfb.h		conflict!
+'G'	00-3F	drivers/misc/sgi-gru/grulib.h	conflict!
+'G'	00-0F	linux/gigaset_dev.h	conflict!
+'H'	00-7F	linux/hiddev.h		conflict!
+'H'	00-0F	linux/hidraw.h		conflict!
+'H'	01	linux/mei.h		conflict!
+'H'	00-0F	sound/asound.h		conflict!
+'H'	20-40	sound/asound_fm.h	conflict!
+'H'	80-8F	sound/sfnt_info.h	conflict!
+'H'	10-8F	sound/emu10k1.h		conflict!
+'H'	10-1F	sound/sb16_csp.h	conflict!
+'H'	10-1F	sound/hda_hwdep.h	conflict!
+'H'	40-4F	sound/hdspm.h		conflict!
+'H'	40-4F	sound/hdsp.h		conflict!
+'H'	90	sound/usb/usx2y/usb_stream.h
+'H'	A0	uapi/linux/usb/cdc-wdm.h
+'H'	C0-F0	net/bluetooth/hci.h	conflict!
+'H'	C0-DF	net/bluetooth/hidp/hidp.h	conflict!
+'H'	C0-DF	net/bluetooth/cmtp/cmtp.h	conflict!
+'H'	C0-DF	net/bluetooth/bnep/bnep.h	conflict!
+'H'	F1	linux/hid-roccat.h	<mailto:erazor_de@users.sourceforge.net>
+'H'	F8-FA	sound/firewire.h
+'I'	all	linux/isdn.h		conflict!
+'I'	00-0F	drivers/isdn/divert/isdn_divert.h	conflict!
+'I'	40-4F	linux/mISDNif.h		conflict!
+'J'	00-1F	drivers/scsi/gdth_ioctl.h
+'K'	all	linux/kd.h
+'L'	00-1F	linux/loop.h		conflict!
+'L'	10-1F	drivers/scsi/mpt2sas/mpt2sas_ctl.h	conflict!
+'L'	E0-FF	linux/ppdd.h		encrypted disk device driver
+					<http://linux01.gwdg.de/~alatham/ppdd.html>
+'M'	all	linux/soundcard.h	conflict!
+'M'	01-16	mtd/mtd-abi.h		conflict!
+		and drivers/mtd/mtdchar.c
+'M'	01-03	drivers/scsi/megaraid/megaraid_sas.h
+'M'	00-0F	drivers/video/fsl-diu-fb.h	conflict!
+'N'	00-1F	drivers/usb/scanner.h
+'N'	40-7F	drivers/block/nvme.c
+'O'     00-06   mtd/ubi-user.h		UBI
+'P'	all	linux/soundcard.h	conflict!
+'P'	60-6F	sound/sscape_ioctl.h	conflict!
+'P'	00-0F	drivers/usb/class/usblp.c	conflict!
+'Q'	all	linux/soundcard.h
+'R'	00-1F	linux/random.h		conflict!
+'R'	01	linux/rfkill.h		conflict!
+'R'	C0-DF	net/bluetooth/rfcomm.h
+'S'	all	linux/cdrom.h		conflict!
+'S'	80-81	scsi/scsi_ioctl.h	conflict!
+'S'	82-FF	scsi/scsi.h		conflict!
+'S'	00-7F	sound/asequencer.h	conflict!
+'T'	all	linux/soundcard.h	conflict!
+'T'	00-AF	sound/asound.h		conflict!
+'T'	all	arch/x86/include/asm/ioctls.h	conflict!
+'T'	C0-DF	linux/if_tun.h		conflict!
+'U'	all	sound/asound.h		conflict!
+'U'	00-CF	linux/uinput.h		conflict!
+'U'	00-EF	linux/usbdevice_fs.h
+'U'	C0-CF	drivers/bluetooth/hci_uart.h
+'V'	all	linux/vt.h		conflict!
+'V'	all	linux/videodev2.h	conflict!
+'V'	C0	linux/ivtvfb.h		conflict!
+'V'	C0	linux/ivtv.h		conflict!
+'V'	C0	media/davinci/vpfe_capture.h	conflict!
+'V'	C0	media/si4713.h		conflict!
+'W'	00-1F	linux/watchdog.h	conflict!
+'W'	00-1F	linux/wanrouter.h	conflict!		(pre 3.9)
+'W'	00-3F	sound/asound.h		conflict!
+'X'	all	fs/xfs/xfs_fs.h		conflict!
+		and fs/xfs/linux-2.6/xfs_ioctl32.h
+		and include/linux/falloc.h
+		and linux/fs.h
+'X'	all	fs/ocfs2/ocfs_fs.h	conflict!
+'X'	01	linux/pktcdvd.h		conflict!
+'Y'	all	linux/cyclades.h
+'Z'	14-15	drivers/message/fusion/mptctl.h
+'['	00-07	linux/usb/tmc.h		USB Test and Measurement Devices
+					<mailto:gregkh@linuxfoundation.org>
+'a'	all	linux/atm*.h, linux/sonet.h	ATM on linux
+					<http://lrcwww.epfl.ch/>
+'a'	00-0F	drivers/crypto/qat/qat_common/adf_cfg_common.h	conflict! qat driver
+'b'	00-FF				conflict! bit3 vme host bridge
+					<mailto:natalia@nikhefk.nikhef.nl>
+'c'	all	linux/cm4000_cs.h	conflict!
+'c'	00-7F	linux/comstats.h	conflict!
+'c'	00-7F	linux/coda.h		conflict!
+'c'	00-1F	linux/chio.h		conflict!
+'c'	80-9F	arch/s390/include/asm/chsc.h	conflict!
+'c'	A0-AF   arch/x86/include/asm/msr.h	conflict!
+'d'	00-FF	linux/char/drm/drm/h	conflict!
+'d'	02-40	pcmcia/ds.h		conflict!
+'d'	F0-FF	linux/digi1.h
+'e'	all	linux/digi1.h		conflict!
+'e'	00-1F	drivers/net/irda/irtty-sir.h	conflict!
+'f'	00-1F	linux/ext2_fs.h		conflict!
+'f'	00-1F	linux/ext3_fs.h		conflict!
+'f'	00-0F	fs/jfs/jfs_dinode.h	conflict!
+'f'	00-0F	fs/ext4/ext4.h		conflict!
+'f'	00-0F	linux/fs.h		conflict!
+'f'	00-0F	fs/ocfs2/ocfs2_fs.h	conflict!
+'g'	00-0F	linux/usb/gadgetfs.h
+'g'	20-2F	linux/usb/g_printer.h
+'h'	00-7F				conflict! Charon filesystem
+					<mailto:zapman@interlan.net>
+'h'	00-1F	linux/hpet.h		conflict!
+'h'	80-8F	fs/hfsplus/ioctl.c
+'i'	00-3F	linux/i2o-dev.h		conflict!
+'i'	0B-1F	linux/ipmi.h		conflict!
+'i'	80-8F	linux/i8k.h
+'j'	00-3F	linux/joystick.h
+'k'	00-0F	linux/spi/spidev.h	conflict!
+'k'	00-05	video/kyro.h		conflict!
+'k'	10-17	linux/hsi/hsi_char.h	HSI character device
+'l'	00-3F	linux/tcfs_fs.h		transparent cryptographic file system
+					<http://web.archive.org/web/*/http://mikonos.dia.unisa.it/tcfs>
+'l'	40-7F	linux/udf_fs_i.h	in development:
+					<http://sourceforge.net/projects/linux-udf/>
+'m'	00-09	linux/mmtimer.h		conflict!
+'m'	all	linux/mtio.h		conflict!
+'m'	all	linux/soundcard.h	conflict!
+'m'	all	linux/synclink.h	conflict!
+'m'	00-19	drivers/message/fusion/mptctl.h	conflict!
+'m'	00	drivers/scsi/megaraid/megaraid_ioctl.h	conflict!
+'m'	00-1F	net/irda/irmod.h	conflict!
+'n'	00-7F	linux/ncp_fs.h and fs/ncpfs/ioctl.c
+'n'	80-8F	linux/nilfs2_fs.h	NILFS2
+'n'	E0-FF	linux/matroxfb.h	matroxfb
+'o'	00-1F	fs/ocfs2/ocfs2_fs.h	OCFS2
+'o'     00-03   mtd/ubi-user.h		conflict! (OCFS2 and UBI overlaps)
+'o'     40-41   mtd/ubi-user.h		UBI
+'o'     01-A1   linux/dvb/*.h		DVB
+'p'	00-0F	linux/phantom.h		conflict! (OpenHaptics needs this)
+'p'	00-1F	linux/rtc.h		conflict!
+'p'	00-3F	linux/mc146818rtc.h	conflict!
+'p'	40-7F	linux/nvram.h
+'p'	80-9F	linux/ppdev.h		user-space parport
+					<mailto:tim@cyberelk.net>
+'p'	A1-A5	linux/pps.h		LinuxPPS
+					<mailto:giometti@linux.it>
+'q'	00-1F	linux/serio.h
+'q'	80-FF	linux/telephony.h	Internet PhoneJACK, Internet LineJACK
+		linux/ixjuser.h		<http://web.archive.org/web/*/http://www.quicknet.net>
+'r'	00-1F	linux/msdos_fs.h and fs/fat/dir.c
+'s'	all	linux/cdk.h
+'t'	00-7F	linux/ppp-ioctl.h
+'t'	80-8F	linux/isdn_ppp.h
+'t'	90	linux/toshiba.h
+'u'	00-1F	linux/smb_fs.h		gone
+'u'	20-3F	linux/uvcvideo.h	USB video class host driver
+'v'	00-1F	linux/ext2_fs.h		conflict!
+'v'	00-1F	linux/fs.h		conflict!
+'v'	00-0F	linux/sonypi.h		conflict!
+'v'	C0-FF	linux/meye.h		conflict!
+'w'	all				CERN SCI driver
+'y'	00-1F				packet based user level communications
+					<mailto:zapman@interlan.net>
+'z'	00-3F				CAN bus card	conflict!
+					<mailto:hdstich@connectu.ulm.circular.de>
+'z'	40-7F				CAN bus card	conflict!
+					<mailto:oe@port.de>
+'z'	10-4F	drivers/s390/crypto/zcrypt_api.h	conflict!
+'|'	00-7F	linux/media.h
+0x80	00-1F	linux/fb.h
+0x89	00-06	arch/x86/include/asm/sockios.h
+0x89	0B-DF	linux/sockios.h
+0x89	E0-EF	linux/sockios.h		SIOCPROTOPRIVATE range
+0x89	E0-EF	linux/dn.h		PROTOPRIVATE range
+0x89	F0-FF	linux/sockios.h		SIOCDEVPRIVATE range
+0x8B	all	linux/wireless.h
+0x8C	00-3F				WiNRADiO driver
+					<http://www.winradio.com.au/>
+0x90	00	drivers/cdrom/sbpcd.h
+0x92	00-0F	drivers/usb/mon/mon_bin.c
+0x93	60-7F	linux/auto_fs.h
+0x94	all	fs/btrfs/ioctl.h
+0x95	all	uapi/linux/kdbus.h	kdbus IPC driver
+0x97	00-7F	fs/ceph/ioctl.h		Ceph file system
+0x99	00-0F				537-Addinboard driver
+					<mailto:buk@buks.ipn.de>
+0xA0	all	linux/sdp/sdp.h		Industrial Device Project
+					<mailto:kenji@bitgate.com>
+0xA2	00-0F	arch/tile/include/asm/hardwall.h
+0xA3	80-8F	Port ACL		in development:
+					<mailto:tlewis@mindspring.com>
+0xA3	90-9F	linux/dtlk.h
+0xAB	00-1F	linux/nbd.h
+0xAC	00-1F	linux/raw.h
+0xAD	00	Netfilter device	in development:
+					<mailto:rusty@rustcorp.com.au>
+0xAE	all	linux/kvm.h		Kernel-based Virtual Machine
+					<mailto:kvm@vger.kernel.org>
+0xAF	00-1F	linux/fsl_hypervisor.h	Freescale hypervisor
+0xB0	all	RATIO devices		in development:
+					<mailto:vgo@ratio.de>
+0xB1	00-1F	PPPoX			<mailto:mostrows@styx.uwaterloo.ca>
+0xB3	00	linux/mmc/ioctl.h
+0xC0	00-0F	linux/usb/iowarrior.h
+0xCA	00-0F	uapi/misc/cxl.h
+0xCB	00-1F	CBM serial IEC bus	in development:
+					<mailto:michael.klein@puffin.lb.shuttle.de>
+0xCD	01	linux/reiserfs_fs.h
+0xCF	02	fs/cifs/ioctl.c
+0xDB	00-0F	drivers/char/mwave/mwavepub.h
+0xDD	00-3F	ZFCP device driver	see drivers/s390/scsi/
+					<mailto:aherrman@de.ibm.com>
+0xEC	00-01	drivers/platform/chrome/cros_ec_dev.h	ChromeOS EC driver
+0xF3	00-3F	drivers/usb/misc/sisusbvga/sisusb.h	sisfb (in development)
+					<mailto:thomas@winischhofer.net>
+0xF4	00-1F	video/mbxfb.h		mbxfb
+					<mailto:raph@8d.com>
+0xF6	all	LTTng			Linux Trace Toolkit Next Generation
+					<mailto:mathieu.desnoyers@efficios.com>
+0xFD	all	linux/dm-ioctl.h