* Fix ide_init_sg_cmd() setup for non-fs requests.

* Convert ide_pc_intr() to use ide_pio_bytes() for floppy media.

* Remove no longer needed ide_io_buffers() and sg/sg_cnt fields
  from struct ide_atapi_pc.

* Remove partial completions; kill idefloppy_update_buffers(), as a
  result.

* Add some more debugging statements.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Signed-off-by: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Return number of bytes left to transfer from idetape_{in,out}put_buffers()
  and number of bytes done from ide_tape_io_buffers().

* Fix padding for PIO transfers in ide_pc_intr() so read/write buffers are
  always completely processed and then the transfer is padded if necessary.

* Remove invalid error messages.

* Remove now superfluous padding from ide{_io_buffers,tape_input_buffers}().

While at it:

* Set pc->bh to NULL in idetape_input_buffers() after all bh-s are done.

* Cache !!(pc->flags & PC_FLAG_WRITING) in local variable in ide_pc_intr().

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Nowadays it is not worth having a separate config option for
Amiga IDE Doubler support so always include it (it still needs
to be explicitly enabled by module parameter).

Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

struct ide_atapi_pc is often allocated on the stack and size of ->pc_buf
size is 256 bytes.  However since only ide_floppy_create_read_capacity_cmd()
and idetape_create_inquiry_cmd() require such size allocate buffers for
these pc-s explicitely and decrease ->pc_buf size to 64 bytes.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Request queue cleanup should happen before freeing drive->id
  and marking device as non-present.  Fix it.

* Remove superfluous hwif->lock acquiring/releasing.

Cc: Stanislaw Gruszka <stf_xl@wp.pl>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

There is no need for a separate ide_arm host driver nowadays
so merge it into ide_generic one.

While at it:
- return -EBUSY from ide_generic_init() if I/O resources are busy
- scale down ide_generic_check_pci_legacy_iobases() for CONFIG_PCI=n

Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Alexander Schulz <alex@shark-linux.de>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Nowadays we have "ide_generic.probe_mask=" module parameter
and ide_platform host driver so sysfs interface for adding
IDE interfaces is no longer needed.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Move ide_map_sg() calls out from ide_build_sglist()
  to ide_dma_prepare().

* Pass command to ide_destroy_dmatable().

* Rename ide_build_sglist() to ide_dma_map_sg()
  and ide_destroy_dmatable() to ide_dma_unmap_sg().

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

Set/clear drive->waiting_for_dma flag in the core code
instead of in ->dma_setup and ->dma_end methods.

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

Move ide_map_sg() call from ->dma_setup implementations and
ide_destroy_dmatable() one from *_build_dmatable() to ide_dma_prepare().

There should be no functional changes caused by this patch.

Sergei:
Removed 'use_pio_instead' labels and replaced 'goto' with 'return 0' --
that required no changes to the follow-up patches...
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Add (an optional) ->dma_check method for checking if DMA can be
  used for a given command and fail DMA setup in ide_dma_prepare()
  if necessary.

* Convert alim15x3 and trm290 host drivers to use ->dma_check.

* Rename ali15x3_dma_setup() to ali_dma_check() while at it.

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

Use custom ->dma_{start,end} methods to handle trm290_prepare_drive()
there instead of in ->dma_setup method.

There should be no functional changes caused by this patch
(DMA support is disabled currently in trm290.c).
Acked-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Use custom ->dma_{start,end} methods to handle ns87415_prepare_drive()
there instead of in ->dma_setup method.
Acked-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Add ide_dma_prepare() helper.

* Convert ide_issue_pc() and do_rw_taskfile() to use it.

* Make ide_build_sglist() static.

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

Move ide_destroy_dmatable() call out from ->dma_end method to
{ide_pc,cdrom_newpc,ide_dma}_intr(), ide_dma_timeout_retry()
and sgiioc4_resetproc().

This causes minor/safe behavior changes w.r.t.:
* cmd64x.c::cmd64{8,x}_dma_end()
* cs5536.c::cs5536_dma_end()
* icside.c::icside_dma_end()
* it821x.c::it821x_dma_end()
* scc_pata.c::__scc_dma_end()
* sl82c105.c::sl82c105_dma_end()
* tx4939ide.c::tx4939ide_dma_end()

v2:
* Fix build for CONFIG_BLK_DEV_IDEDMA=n (reported by Randy Dunlap).

Cc: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Since ide_dma_timeout() is only used by ide_dma_timeout_retry()
inline it there.

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

All custom ->dma_timeout implementations call the generic one thus it is
possible to have only an optional method for resetting DMA engine instead:

* Add ->dma_clear method and convert hpt366, pdc202xx_old and sl82c105
  host drivers to use it.

* Always use ide_dma_timeout() in ide_dma_timeout_retry() and remove
 ->dma_timeout method.

* Make ide_dma_timeout() static.

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

Cache drive->queue in local variable and use max().

There should be no functional changes caused by this patch.
Acked-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Acked-by: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* 'thislen' is always <= cmd->nleft for non-fs requests so the transfer
  padding inside the 'while (thislen > 0)' loop can happen only for fs
  requests -- then move it out of the loop and unify with the transfer
  padding for non-fs requests ('thislen' == 'len' for fs requests).

* blk_dump_rq_flags() dumps all request flags so it is enough to pass
  only the function name to it.

* Update my Copyrights while at it.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Use the following facts:

- rq->nr_sectors should now be always equal to (non-zero)
  rq->hard_nr_sectors for fs requests

- REQ_TYPE_ATA_PC requests have never bio attached to them

- rq->hard_nr_sectors == 0 for REQ_TYPE_ATA_PC requests

- DMA is used only for fs, pc and REQ_TYPE_ATA_PC requests

- 'uptodate' is ignored for pc requests ('rc == 0' case)

and use the common completion path also for DMA requests.

There should be no functional changes caused by this patch

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Convert ide-cd to use scatterlists for PIO transfers and get rid of
partial completions (except on error) also for non-fs requests.

v2:
Do not map dataless commands to an sg since it oopses on the virt_to_page()
translation check when DEBUG_VIRTUAL is enabled.  (from Borislav Petkov,
reported/bisected-by Tetsuo Handa).

Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Acked-by: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

We now support arbitrary number of bytes per-IRQ also for fs requests
so remove ide_cd_check_transfer_size() and IDE_AFLAG_LIMIT_NFRAMES.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Export ide_pio_bytes().

* Add ->last_xfer_len field to struct ide_cmd.

* Add ide_cd_error_cmd() helper to ide-cd.

* Convert ide-cd to use scatterlists also for PIO transfers (fs requests
  only for now) and get rid of partial completions (except when the error
  happens -- which is still subject to change later because looking at
  ATAPI spec it seems that the device is free to error the whole transfer
  with setting the Error bit only on the last transfer chunk).

* Update ide_cd_{prepare_rw,restore_request,do_request}() accordingly.

* Inline ide_cd_restore_request() into cdrom_start_rw().

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Untangling cdrom_end_request() uncovered an error in completing
failed requests in ide_cd_do_request().  Fix it.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Inline cdrom_end_request() into cdrom_newpc_intr()
and ide_cd_do_request().

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Move cdrom_end_request() calls from cdrom_decode_status()
  and ide_cd_check_ireason() to cdrom_newpc_intr().

* Unify cdrom_newpc_intr() exit paths.

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Move setting REQ_FAILED flag out from 'end_request' exit path in
cdrom_newpc_intr() and also rename 'end_request' to 'out_end'.

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

* Move cdrom_end_request() calls from cdrom_start_rw()
  and ide_cd_prepare_rw_request() to ide_cd_do_request().

* Unify ide_cd_do_request() exit paths.

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

ide-scsi is gone...

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

It makes no sense to check for BSY bit being set as earlier OK_STAT()
check in cdrom_end_request() makes sure that BSY bit is cleared.

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Use ide_end_rq() also for failed non-fs requests on completion
of REQUEST SENSE requests + use blk_rq_bytes() while at it.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Use PageHighMem() instead of ifdefs in ide_pio_bytes()
(=> local IRQs won't be disabled when not necessary).

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Add support for arbitrary transfer lengths to ide_pio_bytes()
and then inline ide_pio_multi() into ide_pio_datablock().

There should be no functional changes caused by this patch.

Cc: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Apparently¹, some ATAPI devices want to see the packet command first
before enabling DMA otherwise they simply hang indefinitely. Reorder the
two steps and start DMA only after having issued the command first.

[1] http://marc.info/?l=linux-kernel&m=123835520317235&w=2Signed-off-by: Borislav Petkov <petkovbb@gmail.com>
Reported-by: Michael Roth <mroth@nessie.de>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Fix some incorrect IDE_FTFLAG_* changes which slipped in commit
"ide: add "flagged" taskfile flags to struct ide_taskfile (v2)"
(commit 19710d25) few days ago.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>