Here's an array-based implementation.

ARM arch

m68k arch

SuperH arch

CRIS and x86-64 arches

The SYMBOL_NAME macro (and variations) have been obsolete since 2.1.0,
when the option to compile the kernel in a.out format was removed.  This
patch starts the process of removing these macros, starting with x86.

Originally by Christoph Hellwig back in February.
It recieved no objections when posted to l-k & Jens.

o remove DEVICE_REQUEST definitions - never used in blk.h itself.
o remove DEVICE_ON() - never used at all.
o define LOCAL_END_REQUEST when we do not want end_request() instead
  of other hacks.
o remove DEVICE_OFF() - only used in floppy driver, thus one now has
  a private end_request().
o use private end_request() functions for drivers not providing
  randomness.
o remove TIMEOUT_VALUE - only ever used in hd.c

Some that got lost along the way...

Uses the error codes, but doesn't include err.h

Originally from Manfred Spraul.

* dynamically grow the LDT
Every app that's linked against libpthread right now allocates a full 64
kB LDT, without proper error handling, and always from the vmalloc area

- Nuke unnecessary include.
- More infortmational debug info

Various bits mostly from 2.4
Also fix indentation of various entries to match the rest of the file.

Posted multiple times to l-k by Paul Fulghum

Posted multiple times to l-k by Paul Fulghum.

- Drop Intel P4 bits
- Update other parts in line with latest i386 bluesmoke.c

Missing loop increment.

This device returns a wrong class code which we work around at boot up time.

- printk loglevels
- request region checks.

Try to get config file distro placed in boot.

Some SiS boards won't boot without pci=bios forcing their method.
The reason for this is in the probe method, we try the BIOS method, and it
succeeds and we store the pci_ops. We then try the direct access, it fails
and we trash the old pci_ops.
The patch stores the pci_ops when BIOS method succeeds and restores it
if the direct hardware access method fails.

From Stelian, already in 2.4

We now have so many CONFIG_ options we overflow a buffer in xconfig.
Doubling the size should see us until someone either writes a replacement,
or can be bothered to make it dynamically allocate.

o   P4 thermal throttling is now compile time option
o   ifdefs cleaned up due to above
o   Only poke the LVT if thermal throttling is enabled.
o   Remove bogus cache flushing as per previous discussion.
o   Replace hard coded timer values with MCE_RATE
o   Change default polling frequency from 5 to 15 seconds
o   SMP fixes. (Don't readd timer)
o   Compile time warning fixes.
o   Add config helptext

The A20 code needs a slight tweak to work on AMD Elans.
Original patch from Robert Schwebel

This is a patch from Andreas Gruenbacher (the ext2/ext3 EA patch
maintainer) which converts the extended attribute VFS code from
using the BKL to instead use the inode semaphore; please apply.

Here's a patch for SCSI REPORT LUN scanning, including Douglas Gilbert's
recent changes to support a short INQUIRY followed by a longer INQUIRY.

Please apply, or let me know if you think it needs any modifications.

It's against linux-2.5.4. It does _not_ change the size of the linux lun.

A description of the 8 byte LUN layout can be found on page 35 of:

ftp://ftp.t10.org/t10/drafts/scc2/scc2r04.pdf

The above is a draft, but matches the layout seen on most disk arrays
(including EMC, IBM, LSI, and Hitachi). Later drafts (post SCSI-3) have
this information in the SCSI Architectural Model.

Patch description:

Adds REPORT LUN scanning.

Adds Douglas Gilbert's INQUIRY modification so broken devices that cannot handlean INQUIRY
of more than 36 bytes can be black-listed, plus saving the INQUIRY
result in Scsi_Device.

Adds scan_scsis_target function, replacing code in scan_scsis and parts
of scan_scsis_single. This cleans up the scanning code, and removes a
really ugly for loop. It would be difficult to add REPORT LUN scanning
without this change.

Adds missing scsi_release_commandblocks().

No longer sets max_dev_lun out of bounds for BLIST_FORCELUN devices.

Fixes scanning past LUN 7 for SCSI-3 devices (the patch in 2.4.17
for that fix will not cleanly apply against this code).

-- Patrick Mansfield

    Changes since 3.5.24 (20020319)
        - use Scsi_Request::upper_private_data
        - zero buffers for non-root users
    Changes since 3.5.23 (20011231)
        - change EACCES to EPERM when O_RDONLY is insufficient
        - suppress newlines in host string
        - fix xfer direction, old interface, short reply_len [Travers Carter]

Hello,

	this ALSA patch contains:

- fixed sound_core.c - devfs names
- added ARM driver for H3x00 iPaq handhelds from Compaq
- added RME Digi32 driver
- increased number of RawMidi devices per card to 8
- removed unimplemented IPC definitions from asequencer.h
- compilation fixes for snd_printk/snd_printd
- added snd_midi_event_no_status() function to seq_midi_event.c
  - used in OSS emulation code to follow OSS/Lite behaviour
- fixed snd_pcm_playback_silence() function (possible memory leak)
- fixed endless loop in snd_pcm_playback_drain()
  - OSS applications using mmap() should end correctly now
- fixed autoloading of sequencer clients (oops)
- fixed virmidi code (emu10k1)
- improved OPL4 chip detection
- fixed deadlock in sound/i2c.h
- improved cs8427 chip support (proper initialization)
- improved CMI8330 detection
- improved S/PDIF detection in ENS1371 driver
- fixed possible NULL pointer dereference in FM801 driver
- added ICH4 to intel8x0 driver
- fixed port names allocation in intel8x0 driver
- fixed rate detection in intel8x0 driver
- AC97 code
  - better initialization for wolfson codecs
  - improved CS4205 support (special S/PDIF handling)
- emu10k1/audigy updates

						Jaroslav

Reduce the "rain forest" of pointers a bit.

- Use the channel info from the drive instead of from hwgroup where possible.
   Since we have now a direct association between a channel and irq, we can fix
   the handler for unexpected interrupts to don't have to iterate over multiple
   possibilities.

- Finally it was just possible to remove the circular list of channels
   associated with a hwif. The hwif member of struct hwgroup_s as well as
   next member if struct ata_channel are gone.

- Unify ide_build_sglist() and raw_build_sglist() in to one function.  Use the
   queue associated with a request in the case of a CMD request there instead of
   referencing the device minor number. This will be pushed even further later.

Some of the following is a bit of sensitive... but anyway it has to be done.

- Scarp "portability" macros ide_request_irq() and ide_free_irq().  What a
   great of a surprise in the face of a kernel nowadays ported to nearly any
   major CPU out there, but apparently no platform needs something special here.
   All of them are just using the normal request_irq() and free_irq(). If some
   unusual platform (read obsolete door blocker) I have overlooked intentionally
   needs something special - please just hack the platform specific
   request_irq() and free_irq()-  they are *far off* from being performance
   relevant. (Hint: You could use cooked default irq values to detect there
   whatever the caller was the ATA driver to decide what to do about it.)

- Kill ide_check_region() ide_request_region() and ide_release_region()
   altogether. Add comments explaining the code there.  Same arguments as above
   apply. Additionally we want to be able at some time to do for example mmapped
   PCI address space access on IA32 instead of port IO.

- Rewrite the initialization code to don't use check_region at all. Add some
   comments there, which explain what's actually happening.

- Fix creeping in of a c99 construct.

- Make the main interrupt handler take a channel as data instead of the
   hwgroup. This is the entity an interrupt is physically associated with. PCI
   interrupt sharing is handled in a generically anyway. Analogously the main
   data entity of a device request queue (q->queuedata) is the device and not
   the hwgroup.  However right now we are only able to push it down to the
   channel level. The intention is to advance the choose_drive() function to
   just check for equal interrupt values for serialization of request and
   thus to get rid of the hwgroup altogether.

- Streamline the usage of sector_t over the strategy routines in question a
   bit. Streamline the do_request code in ide-disk.c.

- Improve the readability of start_request in ide.c.

- Remove obsolete/nowhere used stuff from hdreg.h.

- Splitup special_t into classical flag field.

- Use only a single field to determine the capacity of a drive.  Make this
   field and the code paths it follows as far as possible use the sector_t
   instead of a hard coded integer types.  This increases the chances that at
   some distant point in time we will indeed be able to use 64 bit wide sector_t
   entities. (Disks are getting huge those times now...)

- Initialise the per-request_queue readahead parameter properly,
  rather than the dopey "if it's zero you get the deafult"
  approach.

- Permit zero-length readahead.

- 80-columnify mm/readahead.c