* Use elv_add_request() instead of __elv_add_request() in ide_do_drive_cmd().

* ide_do_drive_cmd() is used only in ide-{atapi,cd}.c so inline it there.

There should be no functional changes caused by this patch.
Acked-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Move ide_dma_timeout_retry() to ide-dma.c and add static inline
version for CONFIG_BLK_DEV_IDEDMA=n.
Acked-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Move drive_is_ready() to ide-io.c, then make it static.

Also make some minor CodingStyle fixups while at it.
Acked-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

There are no modular ide_devset_execute() users left so unexport it.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Then make it static and remove 'dma' argument.
Acked-by: NSergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Signed-off-by: NStanislaw Gruszka <stf_xl@wp.pl>
Cc: Andrew Victor <linux@maxim.org.za>
[bart: minor checkpatch.pl / CodingStyle fixups]
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

bart:
It seems like the bug could cause insanely long timeouts for:
- ATA_DMA_ERR error in dma_timer_expiry()
- commands without ->expiry in tc86c001_timer_expiry()
  (TC86C001 IDE controller only)
Signed-off-by: NRoel Kluin <roel.kluin@gmail.com>
Cc: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
[bart: port it to the current tree]
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

This one fixes http://bugzilla.kernel.org/show_bug.cgi?id=12320.
Signed-off-by: NBorislav Petkov <petkovbb@gmail.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* ->cur_dev should now be always valid if ->handler is set so
  remove redundant checks from ide_intr() and ide_timer_expiry().

* Apply CodingStyle fixups in ide_timer_expiry() while at it.

There should be no functional changes caused by this patch.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

There should be no functional changes caused by this patch.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

If ->handler is set while it shouldn't be it indicates deep problems
so BUG_ON()-ning and preventing further damage is much more appropriate
than merely printing an error message.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Remove (now superfluous) ->error method from struct ide_driver.

* Unexport __ide_error() and make it static.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

While at it:
- s/struct ide_driver_s/struct ide_driver/
- use to_ide_driver() macro in ide-proc.c
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Merge ide_hwgroup_t with ide_hwif_t.

* Cleanup init_irq() accordingly, then remove no longer needed
  ide_remove_port_from_hwgroup() and ide_ports[].

* Remove now unused HWGROUP() macro.

While at it:

* ide_dump_ata_error() fixups

v2:
* Fix ->quirk_list check in do_ide_request()
  (s/hwif->cur_dev/prev_port->cur_dev).
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Add ->host_busy field to struct ide_host and use it's first bit
  together with lock bitops to provide new ports serialization method.

* Convert core IDE code to use new ide_[un]lock_host() helpers.

  This removes the need for taking hwgroup->lock if host is already
  busy on serialized hosts and makes it possible to merge ide_hwgroup_t
  into ide_hwif_t (done in the later patch).

* Remove no longer needed ide_hwgroup_t.busy and ide_[un]lock_hwgroup().

* Update do_ide_request() documentation.

v2:
* ide_release_lock() should be called inside IDE_HFLAG_SERIALIZE check.

* Add ide_hwif_t.busy flag and ide_[un]lock_port() for serializing
  devices on a port.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Add 'int port_count' field to ide_hwgroup_t to keep the track
  of the number of ports in the hwgroup.  Then update init_irq()
  and ide_remove_port_from_hwgroup() to use it.

* Remove no longer needed hwgroup->hwif, {drive,hwif}->next,
  ide_add_drive_to_hwgroup() and ide_remove_drive_from_hwgroup()
  (hwgroup->drive now only denotes the currently active device
   in the hwgroup).

* Update locking documentation in <linux/ide.h>.

While at it:

* Rename ->drive field in ide_hwgroup_t to ->cur_dev.

* Use __func__ in ide_timer_expiry().
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Use hwif instead of hwgroup as {request,free}_irq()'s cookie,
teach ide_intr() to return early for non-active serialized ports,
modify unexpected_intr() accordingly and then use per-port IRQ
handlers instead of per-hwgroup ones.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Pass 'ide_hwif_t *' instead of 'ide_hwgroup_t *' to unexpected_intr().

* Cache pointer to the port currently being serviced in ->cur_port
  and use it instead of hwif->hwgroup on serialized hosts.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Fix do_ide_request() to operate on previous device / port instead of
the current one.  The original code was wrong since at least Feb 2002
(2.4.0 timeframe).
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

commit 295f0004 ("ide: don't execute
the next queued command from the hard-IRQ context") overlooked that
ide_do_drive_cmd() (used for REQUEST SENSE command handling) may still
invoke do_ide_request() (->request_fn) in the hard-IRQ context through
blk_start_queueing().  This resulted in a LOCKDEP warning after commit
b599bc7a1199419e122cb2e9ec6b0fa2cfbbc17b ("ide: use per-device request
queue locks (v2)").

Since calling blk_start_queuing() in ide_do_drive_cmd() doesn't make
much sense as the port is already marked as busy (so the execution of
the new command will be deferred anyway) then just remove it fixing
LOCKDEP warning and saving some CPU cycles at the same time.
Reported-and-tested-by: NLarry Finger <Larry.Finger@lwfinger.net>
Reported-by: NGrissiom <chaos.proton@gmail.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Signed-off-by: NBorislav Petkov <petkovbb@gmail.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Move hack for flush requests from choose_drive() to do_ide_request().

* Add ide_plug_device() helper and convert core IDE code from using
  per-hwgroup lock as a request lock to use the ->queue_lock instead.

* Remove no longer needed:
  - choose_drive() function
  - WAKEUP() macro
  - 'sleeping' flag from ide_hwif_t
  - 'service_{start,time}' fields from ide_drive_t

This patch results in much simpler and more maintainable code
(besides being a scalability improvement).

v2:
* Fixes/improvements based on review from Elias:
  - take as many requests off the queue as possible
  - remove now redundant BUG_ON()

Cc: Elias Oltmanns <eo@nebensachen.de>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Add ide_[un]lock_hwgroup() inline helpers for obtaining exclusive
access to the given hwgroup and update the core code accordingly.

[ This change besides making code saner results in more efficient
  use of ide_{get,release}_lock(). ]

Cc: Michael Schmitz <schmitz@biophys.uni-duesseldorf.de>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Elias Oltmanns <eo@nebensachen.de>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Remove "paranoia" checks for hwgroup->busy from ide_timer_expiry()
and ide_intr().  This is a preparation for future changes.

Cc: Michael Schmitz <schmitz@biophys.uni-duesseldorf.de>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Elias Oltmanns <eo@nebensachen.de>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

We now tell block layer that there is still work to do using
blk_plug_device() so hack for IDE Power Management can be removed
(it was buggy for hwgroups having more than 4 devices anyway).
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Tell the block layer that we are not done handling requests by using
  blk_plug_device() in ide_do_request() (request handling function)
  and ide_timer_expiry() (timeout handler) if the queue is not empty.

* Remove optimization which directly calls ide_do_request() for the next
  queued command from the ide_intr() (IRQ handler) and ide_timer_expiry().

* Remove no longer needed IRQ masking from ide_do_request() - in case of
  IDE ports needing serialization disable_irq_nosync()/enable_irq() was
  used for the (possibly shared) IRQ of the other IDE port.

* Put the misplaced comment in the right place in ide_do_request().

* Drop no longer needed 'int masked_irq' argument from ide_do_request().

* Merge ide_do_request() into do_ide_request().

* Remove no longer needed IDE_NO_IRQ define.

While at it:

* Don't use HWGROUP() macro in do_ide_request().

* Use __func__ in ide_intr().

This patch reduces IRQ hadling latency for IDE and improves the system-wide
handling of shared IRQs (which should result in more timeout resistant and
stable IDE systems).  It also makes it possible to do some further changes
later (i.e. replace some busy-waiting delays with sleeping equivalents).

v2:
Changes per review from Elias Oltmanns:
- fix wrong goto statement in 'if (startstop == ide_stopped)' block
- use spin_unlock_irq()
- don't use obsolete HWIF() macro

Cc: Elias Oltmanns <eo@nebensachen.de>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

There should be no functional changes caused by this patch.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

rq->errors is overwritten later in this function.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Set nIEN for previous port/device in ide_do_request()
  also if port uses a non-shared IRQ.

* Remove no longer needed ide_hwif_t.sharing_irq.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Fix nIEN quirk check to also omit quirky devices using pdc202xx_{new,old}
host drivers for which ->quirk_list == 2.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Both choose_drive() and PM handling code make sure that the queue
is unplugged so no need to check it again.

Cc: Elias Oltmanns <eo@nebensachen.de>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Now that (almost) all host drivers have been fixed not to abuse ide_lock
and core code usage of ide_lock has been sanitized we may safely replace
ide_lock by per-hwgroup locks.

This patch is partially based on earlier patch from Ravikiran G Thirumalai.

While at it:
- don't use deprecated HWIF() and HWGROUP() macros
- update locking documentation in ide.h

v2:
Add missing spin_lock_init(&hwgroup->lock).  (Noticed by Elias Oltmanns)

Cc: Vaibhav V. Nivargi <vaibhav.nivargi@gmail.com>
Cc: Alok N. Kataria <alokk@calsoftinc.com>
Cc: Shai Fultheim <shai@scalex86.org>
Signed-off-by: NRavikiran Thirumalai <kiran@scalex86.org>
Cc: Elias Oltmanns <eo@nebensachen.de>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

This is just a preparation for future changes and there should be no
functional changes caused by this patch since ide_lock is currently
also used as queue lock.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Use blk_end_request() instead of ide_lock + __blk_end_request()
in cdrom_end_request(), cdrom_newpc_intr(), __ide_end_request(),
ide_complete_pm_request() and ide_end_drive_cmd().

[ ide_lock is currently also used as queue lock ]
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

__ide_end_request() needs ide_lock only for __blk_end_request()
call so push ide_lock taking inside __ide_end_requests().
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

While at it:
- no need to check for hwgroup presence in ide_dump_opcode()
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Just a preparation for future changes.

There should be no functional changes caused by this patch.
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

This reverts commit e9eb8388 since
it could break resume (thanks to Paul Collins for the report).

I'll look into sorting this out properly for 2.6.29
but for 2.6.28 it is the best to just revert my patch.
Reported-by: NPaul Collins <paul@burly.ondioline.org>
Cc: rjw@sisk.pl
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Also while at it:

* Drop unused arguments from ide_complete_power_step().

* Move DEBUG_PM printk() from ide_end_drive_cmd() to
  ide_complete_power_step().
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>