Enforce access rules where appropriate.

If the compiler is smart enough, this may buy us an optimization or two
as a side effect.

This is needed for full AMD and VIA drivers and possibly more. Functions
to turn actual clocking and cycle timings into register values. Also to
merge shared timings to compute an optimal timing set.

Built from the drivers/ide version by Vojtech Pavlik
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

problem:
      id[53-58] might be changed after initializing device CHS settings.

changes:
    - call ata_dev_reread_id() to reread the identify device info,
      after initializing device CHS settings.
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>

============
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

     - merge ata_prot_to_cmd() and ata_dev_set_protocol() as
       ata_rwcmd_protocol()
     - pave road for read/write multiple support
     - remove usage of pre-cached command and protocol values and call
       ata_rwcmd_protocol() instead
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>

==============
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

   move the initialization of taskfile LBA flags
     "ATA_TFLAG_LBA" and "ATA_TFLAG_LBA48 flags"
   to the SCSI translation functions
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>

=============
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

A few code shuffles, to make merging future code easier.

Add (DRIVER_SENSE << 24) to certain result codes, as noted by Douglas
Gilbert.

No content changes, just moving code around.

This redoes the n_ports logic I proposed before as a bitmask.
ata_pci_init_native_mode is now used with a mask allowing for mixed mode
stuff later on. ata_pci_init_legacy_port is called with port number and
does one port now not two. Instead it is called twice by the ata init
logic which cleans both of them up.

There are stil limits in the original code left over

- IRQ/port mapping for legacy mode should be arch specific values
- You can have one legacy mode IDE adapter per PCI root bridge on some systems
- Doesn't handle mixed mode devices yet (but is now a lot closer to it)

Replace SCSI's legacy "bang at the door" method of probing with one
directly controlled by the underlying ATA transport layer.

We now only call scsi_scan_target() for devices we find, rather than
probing every possible channel/id within a certain range.

Changes:
s/PIO_ST_/HSM_ST_/ and s/pio_task_state/hsm_task_state/.
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Make sure we that completion is the final action we take; prior to this
change, another CPU may have changed ap->pio_task_state before we tested
it a final time.

Spotted by, and original patch by Albert Lee @ IBM.

Also includes a minor optimization:  eliminate a ton of unnecessary
queue_work() calls, simply by jumping to the beginning of the FSM
function ata_pio_task().

Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Most serious is fixing include/sound/pcm.h, which breaks the DocBook
build.

The other stuff is just filling in things that cause warnings.

ATAPI is getting close to being ready.  To increase exposure, we enable
the code in the upstream kernel, but default it to off (present
behavior).  Users must pass atapi_enabled=1 as a module option (if
module) or on the kernel command line (if built in) to turn on
discovery of their ATAPI devices.

- changes license of all code from OSL+GPL to plain ole GPL
  - except for NVIDIA, who hasn't yet responded about sata_nv
  - copyright holders were already contacted privately

- adds info in each driver about where hardware/protocol docs may be
  obtained

- where I have made major contributions, updated copyright dates

Needed for a few PATA drivers.

Also fix up a wrong comment.
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

You spelt heuristic wrongly. Also reformatted to 80 columns,
ignore the diff and fix the typo if you prefer that.
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Wrap ata_qc_complete() calls in EH context in spinlocks, to prevent
races (mainly in ATAPI code paths).

- bump versions where necessary
- remove two duplicated+outdated doc comments
- add MODULE_VERSION() to AHCI driver

[PATCH libata-dev-2.6:upstream] implement ata_poll_qc_complete and use it in polling functions

 Previously, libata polling functions turned irq back on and completed
qc commands without holding host lock.  This creates a race condition
between the polling task and interrupts from other ports on the same
host set or spurious interrupt from itself.

 This patch implements ata_poll_qc_complete which enables irq and
completes qc atomically and convert all polling functions.

 Note: atapi_packet_task() didn't use to turn irq back on or clear
ATA_FLAG_NOINTR on error exits.  This patch makes it use
ata_poll_qc_complete which does both.

 Note: With this change, ALL invocations of ata_qc_complete() are now
done under host_set lock.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Interrupts from devices sharing the same IRQ could cause
ata_host_intr to finish commands being processed by atapi_packet_task
if the commands are using ATA_PROT_ATAPI_NODATA or ATA_PROT_ATAPI_DMA
protocol.  This is because libata interrupt handler is unaware that
interrupts are not expected during that period.  This patch adds
ATA_FLAG_NOINTR flag to tell the interrupt handler that we're not
expecting interrupts.

 Note that once proper HSM is implemented for interrupt-driven PIO,
this should be merged into it and this flag will be removed.

 ahci.c is a different kind of beast, so it's left alone.

* The following drivers use ata_qc_issue_prot and ata_interrupt, so
  changes in libata core will do.

  ata_piix sata_sil sata_svw sata_via sata_sis sata_uli

* The following drivers use ata_qc_issue_prot and custom intr handler.
  They need this change to work correctly.

  sata_nv sata_vsc

* The following drivers use custom issue function and intr handler.
  Currently all custom issue functions don't support ATAPI, so this
  change is irrelevant, updated for consistency and to avoid later
  mistakes.

  sata_promise sata_qstor sata_sx4
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Description:
  After calling the completion callback, the libata error handler might be
running and getting atapi sense data. Clearing the ATA_QCFLAG_ACTIVE flag
at this point might interfere with the libata error handler.

Changes:
   - Clear the ATA_QCFLAG_ACTIVE flag before calling the completion callback
     (and also before the error handler)
   - Add some comment
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

PATCH 2/2:  handle the case when device returns/needs extra data

Description:
   Sometimes the device returns/needs extra data than expected.

Changes:
   Modify __atapi_pio_bytes() to handle the case where device returns/needs extra data.
     - for read case, discard trailing data from the device
     - for write case, padding zero data to the device
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

PATCH 1/2: ata_data_xfer() fix

Changes:
   - Modify ata_mmio_data_xfer() and ata_pio_data_xfer() to handle odd-lengthed buffer.
   - Add some function comments

This patch does not reuse ap->pad as alignment buffer since
using local variable seems good enough.
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Also, fixup a tabs-to-spaces block of code in ata_piix.

This patch lengthens the delay between DET setting and clearing for
COMMRESET from 400us to 1ms.  I couldn't find any requiremen regarding
the duration of COMMRESET in SATA I/II specs but AHCI-1.1 10.4.2
states that it should be at least 1ms.
Signed-off-by: NTejun Heo <htejun@gmail.com>

Contributions from Alan Cox and maximilian attems.

Just opencoded access to eh_eflags, it's much more readable anyway.
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

scsi_add_host is the proper place to set the device, but people copy
the scsi_set_device usage from older drivers again and again.

note that this leaves some legacy drivers like qlogicisp/qlogicfc
without pci association in sysfs, but they're scheduled to go away soon
anyway.
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Problem:
Incorrect md5sum when using ATAPI PIO mode to verify a distro CD.

Root cause:  sg traverse problem.
In __atapi_pio_bytes(), if qc->cursg++ is increased and "goto
next_page" is executed, then sg is not updated to the new qc->cursg
and the old sg is overwritten with the new data.

Changes:
- Replace "goto next_page" with "goto next_sg" to make sg updated.
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>

Document recently-added ata_port_operations hooks.

Fill several doc stubs in libata-core.c.

Jens Axboe pointed out that the iounmap() call in libata was occurring
too early, and some drivers (ahci, probably others) were using ioremap'd
memory after it had been unmapped.

The patch should address that problem by way of improving the libata
driver API:

* move ->host_stop() call after all ->port_stop() calls have occurred.

* create default helper function ata_host_stop(), and move iounmap()
call there.

* add ->host_stop_prewalk() hook, use it in sata_qstor.c (hi Mark).
sata_qstor appears to require the host-stop-before-port-stop ordering
that existed prior to applying the attached patch.

When testing ATAPI PIO data transfer on the ppc64 platform,  __atapi_pio_bytes() got zero when
sg_dma_len() is used. I checked the <asm-ppc64/scatterlish.h>, the struct scatterlist is defined as:

struct scatterlist {
	struct page *page;
	unsigned int offset;
	unsigned int length;

	/* For TCE support */
	u32 dma_address;
	u32 dma_length;
};

#define sg_dma_address(sg)	((sg)->dma_address)
#define sg_dma_len(sg)		((sg)->dma_length)

So, if the scatterlist is not DMA mapped, sg_dma_len() will return zero on ppc64.
The same problem should occur on the x86-64 platform.
On the i386 platform, sg_dma_len() returns sg->length, that's why the problem does not occur on an i386.

Changes:
- Use sg->length if the scatterlist is not DMA mapped (yet).
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>