that provided by the block layer

ATA requires that all DMA transfers begin and end on word boundaries.
Because of this, a large amount of machinery grew up in ide to adjust
scatterlists on this basis.  However, as of 2.5, the block layer has a
dma_alignment variable which ensures both the beginning and length of a
DMA transfer are aligned on the dma_alignment boundary.  Although the
block layer does adjust the beginning of the transfer to ensure this
happens, it doesn't actually adjust the length, it merely makes sure
that space is allocated for transfers beyond the declared length.  The
upshot of this is that scatterlists may be padded to any size between
the actual length and the length adjusted to the dma_alignment safely
knowing that memory is allocated in this region.

Right at the moment, SCSI takes the default dma_aligment which is on a
512 byte boundary.  Note that this aligment only applies to transfers
coming in from user space.  However, since all kernel allocations are
automatically aligned on a minimum of 32 byte boundaries, it is safe to
adjust them in this manner as well.

tj: * Adjusting sg after padding is done in block layer.  Make libata
      set queue alignment correctly for ATAPI devices and drop broken
      sg mangling from ata_sg_setup().
    * Use request->raw_data_len for ATAPI transfer chunk size.
    * Killed qc->raw_nbytes.
    * Separated out killing qc->n_iter.
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJens Axboe <jens.axboe@oracle.com>

For misc ATAPI commands which transfer variable length data to the
host, overflow can occur due to application or hardware bug.  Such
overflows can be ignored safely as long as overflow data is properly
drained.  libata HSM implementation has this implemented in
__atapi_pio_bytes() and recently updated for 2.6.24-rc but it requires
further improvements.  Improve drain logic such that...

* Report overflow errors using ehi desc mechanism instead of printing
  directly.

* Properly calculate the number of bytes to be drained considering
  actual number of consumed bytes for partial draining.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Acked-by: NAlbert Lee <albertcc@tw.ibm.com>
Signed-off-by: NJens Axboe <jens.axboe@oracle.com>

Signed-off-by: NMikko Rapeli <mikko.rapeli@teleca.com>
Tested-by: NBart Champagne <bart@as35701.net>
Cc: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <bzolnier@gmail.com>

rc is used to test the return value and possibly return an error.
No need to redeclare inside the loop.

drivers/ata/libata-core.c:7089:7: warning: symbol 'rc' shadows an earlier one
drivers/ata/libata-core.c:7030:9: originally declared here
Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Some controllers (VIA CX700) raise device error on SETXFER even after
mode configuration succeeded.  Update ata_dev_set_mode() such that
device error is ignored if transfer mode is configured correctly.  To
implement this, device is revalidated even after device error on
SETXFER.

This fixes kernel bugzilla bug 8563.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

The HITACHI HDS7250SASUN500G and HITACHI HDS7225SBSUN250 drives
do not need to be blacklisted, the NCQ problem has been resolved
with the "sata_nv: fix for completion handling" patch.

Signed-off-by David Milburn <dmilburn@redhat.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NAdrian Bunk <bunk@kernel.org>

The documentation for ata_data_xfer and ata_data_xfer_noirq had the 'rw'
parameter named 'write_data'.
Signed-off-by: NLinus Nilsson <lajnold@acc.umu.se>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Factor out ata_pci_activate_sff_host() from ata_pci_one().  This does
about the same thing as ata_host_activate() but needs to be separate
because SFF controllers use different and multiple IRQs in legacy
mode.

This will be used to make SFF LLD initialization more flexible.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

ata_port_queue_task() served a single user:  ata_pio_task()

Rename to ata_pio_queue_task() and un-export it, as nobody outside of
libata-core.c uses it.
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Some it821x RAID firmwares return 0 for the err return off both devices.
A similar issue occurs with the slave returning 0 not 1 if you plug a
gigabyte sata ramdisk into a controller that fakes two SATA ports as
master/slave on an SFF channel.

The patch does the following

- Allow the 'failed diagnostics' case on both master and slave
- Move the HORKAGE_DIAGNOSTIC check after ->dev_config

This second change also allows IT821x to fix up a problem where we report
drive diagnostic failures when in fact the drive is fine but the
microcontroller firmware doesn't appear to get it right. IT821x clears
the flag again to avoid giving the user bogus warnings about their disk.

The other IT821x change is a bit ugly, we slightly abuse the cable type
hook to fiddle with the identify data for the devices. We could add a new
hook for this but as we have only one offender and no more seeming likely
it seems better to keep libata-core clean.

Please let this sit in -mm briefly, just in case the relaxed checking
breaks some other emulated interface.
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

qc->nbytes didn't use to include extra buffers setup by libata core
layer and my be odd.  This patch makes qc->nbytes include any extra
buffers setup by libata core layer and guaranteed to be aligned on 4
byte boundary.

This value is to be used to program the host controller.  As this
represents the actual length of buffer available to the controller and
the controller must be able to deal with short transfers for ATAPI
commands which can transfer variable length, this shouldn't break any
controllers while making problems like rounding-down and controllers
choking up on odd transfer bytes much less likely.

The unmodified value is stored in new field qc->raw_nbytes.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

libata used private sg iterator to handle padding sg.  Now that sg can
be chained, padding can be handled using standard sg ops.  Convert to
chained sg.

* s/qc->__sg/qc->sg/

* s/qc->pad_sgent/qc->extra_sg[]/.  Because chaining consumes one sg
  entry.  There need to be two extra sg entries.  The renaming is also
  for future addition of other extra sg entries.

* Padding setup is moved into ata_sg_setup_extra() which is organized
  in a way that future addition of other extra sg entries is easy.

* qc->orig_n_elem is unused and removed.

* qc->n_elem now contains the number of sg entries that LLDs should
  map.  qc->mapped_n_elem is added to carry the original number of
  mapped sgs for unmapping.

* The last sg of the original sg list is used to chain to extra sg
  list.  The original last sg is pointed to by qc->last_sg and the
  content is stored in qc->saved_last_sg.  It's restored during
  ata_sg_clean().

* All sg walking code has been updated.  Unnecessary assertions and
  checks for conditions the core layer already guarantees are removed.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Cc: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

ATA_QCFLAG_DMAMAP was a bit peculiar in that it got set during qc
initialization and cleared if DMA mapping wasn't necessary.  Make it
more straight forward by making the following changes.

* Don't set it during initialization.  Set it after DMA is actually
  mapped.

* Add BUG_ON() to guarantee that there is data to transfer if DMAMAP
  is set.  This always holds for the current code.  The BUG_ON() is
  for docummentation and sanity check.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

With atapi_request_sense() converted to use sg, there's no user of
non-sg interface.  Kill non-sg interface.

* ATA_QCFLAG_SINGLE and ATA_QCFLAG_SG are removed.  ATA_QCFLAG_DMAMAP
  is used instead.  (this way no LLD change is necessary)

* qc->buf_virt is removed.

* ata_sg_init_one() and ata_sg_setup_one() are removed.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Cc: Rusty Russel <rusty@rustcorp.com.au>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Depending on how many bytes are transferred as a unit, PIO data
transfer may consume more bytes than requested.  Knowing how much
data is consumed is necessary to determine how much is left for
draining.  This patch update ->data_xfer such that it returns the
number of consumed bytes.

While at it, it also makes the following changes.

* s/adev/dev/
* use READ/WRITE constants for rw indication
* misc clean ups
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

ATA_PROT_ATAPI_* are ugly and naming schemes between ATA_PROT_* and
ATA_PROT_ATAPI_* are inconsistent causing confusion.  Rename them to
ATAPI_PROT_* and make them consistent with ATA counterpart.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@redhat.com>

Treat zero xfer length as HSM violation.  While at it, add
unlikely()'s to ATAPI ireason and transfer length checks.

tj: Formatted patch and added unlikely()'s.
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

I missed one while converting to ata_is_*() protocol test helpers.
Convert it.  Pointed out by Jeff Garzik.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

libata-acpi is using separate timing tables for transfer modes
although libata-core has the complete ata_timing table.  Implement
ata_timing_cycle2mode() to look for matching mode given transfer type
and cycle duration and use it in libata-acpi and pata_acpi to replace
private timing tables.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

ATA_CBL_PATA_UNK indicates that the cable type can't be determined
from the host side and might be either 80c or 40c.  libata applies
drive or other generic limit in this case.  However, there are
controllers where both host and drive side detections are
misimplemented and the driver has to rely solely on private method -
peeking BIOS or ACPI configuration or using some other private
mechanism.

This patch adds ATA_CBL_PATA_IGN which tells libata to ignore the
cable type completely and just let the LLD determine the transfer mode
via host transfer mode masks and ->mode_filter().
Signed-off-by: NTejun Heo <htejun@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Jeff says xfer_mask is unsigned long not unsigned int.  Convert all
xfermask fields and handling functions to deal with unsigned longs.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

ata_id_to_dma_mode() isn't quite generic.  The function is basically
privately implemented ata_id_xfermask() combined with hardcoded mode
printing and configuration which are specific to ata_generic.

Kill the function and open code it in generic_set_mode() using generic
xfermode handling functions.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

* s/ATA_BITS_(PIO|MWDMA|UDMA)/ATA_NR_\1_MODES/g

* Consistently use 0xff to indicate invalid transfer mode (0x00 is
  valid for PIO_SLOW).

* Make ata_xfer_mode2mask() return proper mode mask instead of just
  the highest bit.

* Sort ata_timing table in increasing xfermode order and update
  ata_timing_find_mode() accordingly.

This patch doesn't introduce any behavior change.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Export the following xfermode related functions.

* ata_pack_xfermask()
* ata_unpack_xfermask()
* ata_xfer_mask2mode()
* ata_xfer_mode2mask()
* ata_xfer_mode2shift()
* ata_mode_string()
* ata_id_xfermask()
* ata_timing_find_mode()

These functions will be used later by LLD updates.  While at it,
change unsigned short @speed to u8 @xfer_mode in
ata_timing_find_mode() for consistency.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

ATA_DFLAG_DUBIOUS_XFER is set whenever data transfer speed or method
changes and gets cleared when data transfer command succeeds in the
newly configured transfer mode.

This will be used to improve speed down logic.

Signed-off-by: Tejun Heo <htejun@gmail.com<
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Move ata_set_mode() to libata-eh.c.  ata_set_mode() is surely an EH
action and will be more tightly coupled with the rest of error
handling.  Move it to libata-eh.c.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Implement protocol tests - ata_is_atapi(), ata_is_nodata(),
ata_is_pio(), ata_is_dma(), ata_is_ncq() and ata_is_data() and use
them to replace is_atapi_taskfile() and hard coded protocol tests.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

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

I believe this version meets all Sergei's objections

Correct the logic for when we issue a set features for transfer mode

- If the device has IORDY and the controller has IORDY - set the mode
- If the device has IORDY and the controller does not - turn IORDY off
- If neither has IORDY do nothing
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Historically word 48 in the identify data was used to mean 32bit I/O
was supported for VLB IDE etc. ATA8 reassigns this word to the Trusted
Computing Group, where it is used for TCG features. This means that
an ATA8 TCG drive is going to trigger 32bit I/O on some systems which
will be funny.

Anyway we need to sort this out ready for ATA8 so:
- Reorder the ata.h header a bit so the ata_version function occurs early
  in it
- Make dword_io check the ATA version
- Add an ATA8 version checking TCG presence test

While we are at it the current drafts have a flaw where it may not be
possible to disable TCG features at boot (and opt out of the trusted
model) as TCG intends because it relies on presence of a different
optional feature (DCS). Handle this in software by refusing the TCG
commands if libata.allow_tpm is not set. (We must make it possible
as some environments such as proprietary VDR devices will doubtless
want to use it to lock up content)

Finally as with CPRM print a warning so that the user knows they may
not be able to full access and use the device.
Signed-off-by: NAlan Cox <alan@redhat.com>

Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Tejun Heo <htejun@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

With ATAPI transfer chunk size properly programmed, libata PIO HSM
should be able to handle full spurious data chunks.  Also, it's a good
idea to suppress trailing data warning for misc ATAPI commands as
there can be many of them per command - for example, if the chunk size
is 16 and the drive tries to transfer 510 bytes, there can be 31
trailing data messages.

This patch makes the following updates to libata ATAPI PIO HSM
implementation.

* Make it drain full spurious chunks.

* Suppress trailing data warning message for misc commands.

* Put limit on how many bytes can be drained.

* If odd, round up consumed bytes and the number of bytes to be
  drained.  This gets the number of bytes to drain right for drivers
  which do 16bit PIO.

This patch is partial backport of improve-ATAPI-data-xfer patchset
pending for #upstream.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Add two hooks - ata_acpi_dissociate() which is called during driver
detach after the whole host is shutdown and ata_acpi_on_disable()
which is called when a device is disabled.
Signed-off-by: NTejun heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

ata_port_detach() calls ata_dev_disable() with host lock held but
ata_dev_disable() should be called from EH context.  ata_port_detach()
steals EH context by setting ATA_PFLAG_UNLOADAING and flushing EH.
Drop locking around ata_dev_disable() and note that ata_port_detach()
owns EH context at that point.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Like ST380817AS / 3.42, ST3160023AS / 3.42 times out commands if NCQ
is used.  Blacklist it.  This is reported by Matheus Izvekov in the
following thread.

  http://thread.gmane.org/gmane.linux.ide/24202Signed-off-by: NTejun Heo <htejun@gmail.com>
Cc: Matheus Izvekov <mizvekov@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

link->eh_info.serror is used to cache SError for controllers which
need it cleared from interrupt handler to clear IRQ.  It also should
be cleared after reset just like SError itself.

Make ata_std_postreset() clear link->eh_info.serror too and update
sata_sil such that it doesn't care about bookkeeping the value.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Spurious NCQ completion detection implemented in ahci was incorrect.
On AHCI receving and processing FISes and raising interrupts are not
interlocked and spurious interrupts are expected.

For example, if an interrupt occurs while interrupt handler is running
and the running interrupt handler handles the event the new IRQ
indicated, after IRQ handler finishes, it will be executed again
because IRQ pending bit is set by the new interrupt but there won't be
anything to process.

Please read the following message for more information.

  http://article.gmane.org/gmane.linux.ide/26012

This patch...

* Removes all spurious IRQ whining from ahci.  Spurious NCQ completion
  detection was completely wrong.  Spurious D2H Register FIS taught us
  that some early drives send spurious D2H Register FIS with I bit set
  while NCQ commands are in progress but none of recent drives does
  that and even the ones which show such behavior can do NCQ fine.

* Kills all NCQ blacklist entries which were added because of spurious
  NCQ completions.  I tracked down each commit and verified all
  removed ones are actually added because of spurious completions.

  WD740ADFD-00NLR1 wasn't deleted but moved upward because the drive
  not only had spurious NCQ completions but also is slow on sequential
  data transfers if NCQ is enabled.

  Maxtor 7V300F0 was added by 0e3dbc01
  from Alan Cox.  I can only find evidences that the drive only had
  troubles with spuruious completions by searching the mailing list.
  This entry needs to be verified and removed if it doesn't have other
  NCQ related problems.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

On Fri, 30 Nov 2007 14:34:11 +0200 (EET)
Meelis Roos <mroos@linux.ee> wrote:

> > Can you stick a stack trace in at that point ? That would help diagnose
> > it a great deal quicker.
>
> Finally done - found out hard way that BUG() is too bad and
> dump_st5ack() suits me better.

Thanks. This should fix the real cause, and also allow for port start to
fail politely with -ENODEV.
Signed-off-by: NAlan Cox <alan@redhat.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>