Add ata_pack_xfermask(), ata_xfer_mask2mode(), ata_xfer_mode2mask(),
ata_xfer_mode2shift() and ata_id_xfermask().  These functions will be
used by following patches to simplify xfer_mask handling.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Add ATA_BITS_*, ATA_MASK_* macros and reorder xfer_mask fields such
that higher transfer mode is placed at higher order bit.  As thie
reordering breaks ata_mode_string(), this patch also rewrites
ata_mode_string().
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Rename ATA_FLAG_FLUSH_PIO_TASK to ATA_FLAG_FLUSH_PORT_TASK.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Kill unused pio_task and packet_task.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Make pio_task and atapi_packet_task use port_task.
atapi_packet_task() is moved upward such that it's right after
ata_pio_task().  This position is more natural and makes adding
prototype for ata_qc_issue_prot() unnecessary.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Implement port_task.  LLDD's can schedule a function to be executed
with context after specified delay.  libata core takes care of
synchronization against EH.  This is generalized form of pio_task and
packet_task which are tied to PIO hsm implementation.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Revalidate device after transfer mode configuration.  This also makes
dev->id up-to-date.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

ata_dev_revalidate() re-reads IDENTIFY PAGE of the given device and
makes sure it's the same device as the configured one.  Once it's
verified that it's the same device, @dev is configured according to
newly read IDENTIFY PAGE.  Note that revalidation currently doesn't
invoke transfer mode reconfiguration.

Criteria for 'same device'

* same class (of course)
* same model string
* same serial string
* if ATA, same n_sectors (to catch geometry parameter changes)
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Add @print_info argument to ata_dev_configure().  Details of
configured device is printed only when @pinfo_info is non-zero.  This
patch also reorganizes device info printing for LBA case to simplify
code (necessary as @print_info adds extra nesting around it).
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

In ata_dev_configure(), reinitialize parameters before configuring.
This change is for revalidation and hotplug.  As ata_dev_configure()
can be entered multiple times, parameters need to be reinitialized.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Now that reset and configure are converted such that they don't modify
or disable libata core data structures, reorganize ata_bus_probe() to
reflect this change.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

ata_dev_config() needs to be done everytime a device is configured.
Fold it into ata_dev_configure().
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Separate out ata_dev_configure() from ata_dev_identify() such that
ata_dev_configure() only configures @dev according to passed in @id.
The function now does not disable device on failure, it just returns
appropirate error code.

As this change leaves ata_dev_identify() with only reading ID, calling
configure and disabling devices according to the results, this patch
also kills ata_dev_identify() and inlines the logic into
ata_bus_probe().
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Convert dev->id from array to pointer.  This is to accomodate
revalidation.  During revalidation, both old and new IDENTIFY pages
should be accessible and single ->id array doesn't cut it.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Kill kfree(id) in failure path of ata_dev_read_id().  id is not
dynamically allocated yet.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Cc: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: NJeff Garzik <jeff@garzik.org>

Until problems are sorted.

Kill now-unused ata_dev_reread_id().
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Separate out ata_dev_read_id() from ata_dev_identify().  This is the
first half of splitting ata_dev_identify().  ata_dev_read_id() will
also be used for revalidation.  This patch does not make any behavior
change.

ata_dev_read_id() doesn't modify any of libata-internal data
structures.  It simply reads IDENTIFY page and returns error code on
failure.  INIT_DEV_PARAMS and EDD wrong class code are also handled by
this function.

Re-reading IDENTIFY after INIT_DEV_PARAMS is performed by jumping to
retry: instead of calling ata_dev_reread_id().  This is done because
1. there's retry label anyway 2. ata_dev_reread_id() cannot be used
anywhere else so there's no reason to keep it.

This function is probably the place to set transfer mode to PIO0
before IDENTIFY.  However, reset -> identify -> init_dev_params order
should be kept for pre-ATA4 devices so we cannot set transfer mode
before IDENTIFY for them.  How do we know if a device is post-ATA4
before IDENTIFY?
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

This patch makes ata_sg_setup_one() trim sg entry (thus making
qc->n_elem zero) if padding results in zero length sg entry.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

For ATAPI commands, padding can reduce qc->n_elem by one and thus to
zero making assert(qc->n_elem > 0)'s in ata_fill_sg() and qs_fill_sg()
fail for legal commands.  This patch fixes the assert()'s to take
qc->pad_len into account.

Although the condition check seems a bit excessive, as this part of
code isn't still stable yet, I think it's worth to keep those.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

The comment above ata_std_postreset() specified that setting cable
type is the responsibility of postreset(), which isn't possible /
optimal depending on controller / driver.  This patch kills the
comment.  Setting cable type is responsibility of ->probe_reset.
libata doesn't care whether it's done in probeinit, reset or
postreset.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Update ata_dev_init_params() such that it doesn't disable port
directly but return with appropriate error mask on failure.  This is
preparation for splitting ata_dev_identify().  Note that this patch
changes behavior of dev_init_params failure such that only failing
devices are taken offline not the whole port.  This change is
intended.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

This patch renames ata_dev_id_[c_]string() to ata_id_[c_]string().
All other functions which read data from ATA ID data start with ata_id
and those two function names were getting too long.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

 - Fix the array index value in ata_rwcmd_protocol() for the added FUA commands.
 - Filter out ATAPI packet command error messages in ata_pio_error()
Signed-off-by: NAlbert Lee <albertcc@tw.ibm.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

If a low level driver wants to control max_sectors, it had to adjust
ap->host->max_sectors and set ATA_DFLAG_LOCK_SECTORS to tell
ata_scsi_slave_config not to override the limit.  This is not only
cumbersome but also incorrect for hosts which support more than one
devices per port.

This patch adds per-device ->max_sectors.  If the field is unset
(zero), libata core layer will adjust ->max_sectors according to
default rules.  If the field is set, libata honors the setting.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

cdb_len is per-device property.  Sharing cdb_len on ap results in
inaccurate configuration on revalidation and hotplugging.  This patch
makes cdb_len per-device.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

ata_dev_knobble() unconditionally used the first device of the port to
determine whether a device is bridged or not.  This causes bridge
limit to be incorrectly applied or unapplied for hosts with slave
devices (e.g. ata_piix).
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

EDD is never used with ->probe_reset.  Don't handle EDD special case
in ata_dev_identify if ->probe_reset is in use.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Make ata_dump_id() take @id instead of @dev.  This is preparation for
splitting ata_dev_identify().
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Separate out ATA major version calculation from ata_dev_identify()
into ata_id_major_version().  It's preparation for splitting
ata_dev_identify().
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Separate out n_sectors calculation into ata_id_n_sectors() from
ata_dev_identify().  This will be used by revalidation.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Use ata_dev_id_c_string()
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

ata_dev_id_c_string() reads ATA string from the specified offset of
the given IDENTIFY PAGE and puts it in the specified buffer in trimmed
and NULL-terminated form.  The caller must supply a buffer which is
one byte larger than the maximum size of the target ID string.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

This patch makes ata_bus_probe() normalize classes[] returned by
->probe_reset such that ->probe_reset can return ATA_DEV_UNKNOWN.
This eases implementation of ->probe_reset's which don't directly use
ata_drive_probe_reset().
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

In an effort to kill libata-specific assert() and use generic
WARN_ON(), this patch converts all assert(X)'s in libata core layer to
WARN_ON(!X)'s.  Most conversions are straight-forward logical negation
exception for the followings.

* In libata-core.c:ata_fill_sg(),
  assert(qc->n_elem > 0) is converted to WARN_ON(qc->n_elem == 0) because
  qc->n_elem is unsigned and unsigned <= 0 is weird.

* In libata-scsi.c:ata_gen_ata_desc/fixed_sense(),
  assert(NULL != qc->ap->ops->tf_read) is converted to
  WARN_ON(qc->ap->ops->tf_read == NULL), as there are no other users of
  'constant cond var' style in libata.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

This patch inlines ata_qc_complete() and uninlines __ata_qc_complete()
as suggested by Jeff Garzik.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Don't clear SError in sata_std_hardreset().  This makes hardreset act
identically to ->phy_reset register-wise.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

This patch makes std component operations act identical to ->phy_reset
register-wise except for SError clearing on sata_std_hardreset.

Note that if a driver only implements/uses hardreset, it should not
use ata_std_probeinit() to avoid extra sata_phy_resume() and
ata_busy_sleep() compared to ->phy_reset.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

->eng_timeout cannot be invoked with NULL qc anymore.  Add an
assertion in ata_scsi_error() and kill NULL qc handling from all
->eng_timeout callbacks.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>

Implement ata_scsi_timed_out(), to be used as
scsi_host_template->eh_timed_out callback for all libata drivers.
Without this function, the following race exists.

If a qc completes after SCSI timer expires but before libata EH kicks
in, the qc gets completed but the scsicmd still gets passed to libata
EH resulting in ->eng_timeout invocation with NULL qc, which none is
handling properly.

This patch makes sure that scmd and qc share the same lifetime.
Original idea from Jeff Garzik <jgarzik@pobox.com>.
Signed-off-by: NTejun Heo <htejun@gmail.com>
Signed-off-by: NJeff Garzik <jgarzik@pobox.com>