Teach the MMC/SD/SDIO core about using SPI mode.

 - Use mmc_host_is_spi() so enumeration works through SPI signaling
   and protocols, not just the native versions.

 - Provide the SPI response type flags with each request issued,
   including requests from the new lock/unlock code.

 - Understand that cmd->resp[0] and mmc_get_status() results for SPI
   return different values than for "native" MMC/SD protocol; this
   affects resetting, checking card lock status, and some others.

 - Understand that some commands act a bit differently ... notably:
     * OP_COND command doesn't return the OCR
     * APP_CMD status doesn't have an R1_APP_CMD analogue

Those changes required some new and updated primitives:

 - Provide utilities to access two SPI-only requests, and one
   request that wasn't previously needed:
     * mmc_spi_read_ocr() ... SPI only
     * mmc_spi_set_crc() ... SPI only (override by module parm)
     * mmc_send_cid() ... for use without broadcast mode

 - Updated internal routines:
     * Previous mmc_send_csd() modified into mmc_send_cxd_native();
       it uses native "R2" responses, which include 16 bytes of data.
     * Previous mmc_send_ext_csd() becomes new mmc_send_cxd_data()
       helper for command-and-data access
     * Bugfix to that mmc_send_cxd_data() code:  dma-to-stack is
       unsafe/nonportable, so kmalloc a bounce buffer instead.

 - Modified mmc_send_ext_csd() now uses mmc_send_cxd_data() helper

 - Modified mmc_send_csd(), and new mmc_spi_send_cid(), routines use
   those helper routines based on whether they're native or SPI

The newest categories of cards supported by the MMC stack aren't expected
to work yet with SPI:  MMC or SD cards with over 4GB data, and SDIO.
All those cards support SPI mode, so eventually they should work too.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Increase delay for power up in order to support some slower boards.

Also add some comments about why the delays are there.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Replace all cases of BUG_ON with WARN_ON where there is a chance
(with varying degrees of slim) that the kernel can continue without
incidence.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

SDIO doesn't have a CSD so it uses different timeout values than
SD memory.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

It is sometimes necessary to give up on trying to claim the host lock,
especially if that happens in a thread that has to be stopped.

While at it, fix the description for mmc_claim_host() which was wrong.
Signed-off-by: NNicolas Pitre <npitre@mvista.com>
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

The problem is that the sdio_bus must be registered before any SDIO
drivers are registered against it otherwise the kernel sulks.  Because
the sdio_bus registration happens through module_init (equivalent to
device_initcall), then any SDIO
drivers linked before the SDIO core code in the kernel will be initialized
first.

Upcoming SDIO function drivers are likely to be located outside the
drivers/mmc directory as it is common practice to group drivers according
to their function rather than the bus they use.  SDIO drivers are therefore
likely to appear at random location in the kernel link.

To make sure the sdio_bus is always initialized before any SDIO drivers,
let's move the MMC init to the subsys_initcall level.
Signed-off-by: NNicolas Pitre <npitre@mvista.com>
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Add the sdio bus type and basic device handling.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Really basic init sequence for SDIO cards.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

The write parameter in mmc_set_data_timeout() is redundant as the
data structure contains information about the direction of the
transfer.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Convert the MMC layer to use standard error codes and not its own,
incompatible values.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

This patch makes the following needlessly global functions static:
- sd_ops.c: mmc_app_cmd()
- core.c: __mmc_release_bus()
- core.c: mmc_start_request()
Signed-off-by: NAdrian Bunk <bunk@stusta.de>
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Make sure that the debugging output in the core is complete.
This should allow us to clean up all the extra debug output
that each and every other host driver seems to contain.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

In the normal case, the host lock can be claimed directly.
When it cannot, the caller will sleep. Make sure we don't
have any latent bugs by always calling might_sleep().
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Make sure the kerneldoc comments are up to date and relevant.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Move basic host class device handling to its own file for clarity.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Move bus operations to its own file for the sake of clarity. Also
delegate sysfs attributes to bus handlers in preparation for other
more exotic types.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Cc: Pierre Ossman <drzeus@drzeus.cx>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

As mmc_detect_change() can be called from irq context, using
claim (which can sleep) is inherently unsafe. Use the host
spinlock instead, which also is faster.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Since many have the system root on MMC/SD we must allow some foot
shooting when it comes to resume.

We cannot detect if a card is removed and reinserted during suspend,
so the safe approach would be to assume it was, avoiding potential
filesystem corruption. This will of course not work if you cannot
release the card before suspend.

This commit adds a compile time option that makes the MMC layer
assume the card wasn't touched if it is redetected upon resume.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Delegate protocol handling to "bus handlers". This allows the core to
just handle the task of arbitrating the bus. Initialisation and
pampering of cards is now done by the different bus handlers.

This design also allows MMC and SD (and later SDIO) to be more cleanly
separated, allowing easier maintenance.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Move protocol operations and definitions into their own files
in an effort to separate protocol handling and bus
arbitration more clearly.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Create a "core" subdirectory to house the central bus handling
functions.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

The classic MMC bus was defined as multi card bus
system, which is reflected in the design in the MMC
layer.

When SD showed up, the bus topology was abandoned
and a star topology (one card per host) was mandated.
MMC version 4 has followed this, officially deprecating
the bus topology.

As we do not have any known users of the bus
topology we can remove support for it. This will
simplify the code and rectify some incorrect
assumptions in the newer additions.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Suspending MMC/SD cards (versus removing and readding them) is an
inherently unsafe operation and has even been broken for some time.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Make sure we kill of any pending detection runs when the host
is removed instead of when it is freed. Also add some debugging
to make sure the driver doesn't queue up more detection after it
has removed the host.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

The "present" state indicates that the card is a registered device, so
it is more clear to put it together with the actual registration.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Split out the type of card into its own field as it hardly
qualifies as a state.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Support for MMC 4.2 sector based cards. This tweaks the init a
bit and reads a new field out of the EXT_CSD.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Fix copy-n-paste error.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Now that we've fixed our only offender when it comes to strange sg
list, add a check so that future users keep the sg list proper with
regard to transfer size.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

The new high-speed timings are similar to each other and the old
system, but not identical. And although things "just work" most of
the time, sometimes it does not. So we need to start marking which
hosts are known to fully comply with the new timings.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

MMC high-speed, wide bus support and SD high-speed
are functions that aren't critical for correct
operation of the card. As such, they shouldn't mark
the card as bad or dead when there is a failure
activating these features.

This is needed in particular on some really stupid
hardware (e.g. Winbond's) where not all data transfer
commands are supported.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Many controllers have an upper limit on the number of blocks that can be
transferred in one request. Allow the host drivers to specify this and make
sure we avoid hitting this limit.

Also change the max_sectors field to avoid confusion. This makes it map
less directly to the block layer limits, but as they didn't apply directly
on MMC cards anyway, this isn't a great loss.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Most controllers have an upper limit on the block size. Allow the host
drivers to specify this and make sure we avoid hitting this limit.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Thanks to the generous donation of an SDHC card by John Gilmore, and
the surprisingly enlightened decision by the SD Card Association to
publish useful specs, I've been able to bash out support for SDHC. The
changes are not too profound:

i) Add a card flag indicating the card uses block level addressing and
check it in the block driver. As we never took advantage of byte-level
addressing, this simply involves skipping the block -> byte
translation when sending commands.

ii) The layout of the CSD is changed - a set of fields are discarded
to make space for a larger C_SIZE. We did not reference any of the
discarded fields except those related to the C_SIZE.

iii) Read and write timeouts are fixed values and not calculated from
CSD values.

iv) Before invoking SEND_APP_OP_COND, we must invoke the new
SEND_IF_COND to inform the card we support SDHC.
Signed-off-by: NPhilipl Langdale <philipl@overt.org>
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

As card_busy was only used to indicate if the host was exclusively
claimed and not really used to identify a particular card, replacing
it with just a boolean makes things a lot more easily understandable.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Modern SD cards support a clock speed of 50 MHz. Make sure we test for
this capability and do the song and dance required to activate it.

Activating high speed support actually modifies the TRAN_SPEED field
of the CSD. But as the spec says that the cards must report 50 MHz,
we might as well skip re-reading the CSD.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>

Several fixes for mmc_delay():

 * Repair if-clause that was supposed to detect sub-hz delays.
 * Change yield() to cond_resched() as yield() no longer has the
   semantics we desire.
 * mmc_delay() is used to guarantee protocol delays, so we cannot
   return early (i.e. use _interruptable).

Based on patch by Amol Lad.
Signed-off-by: NPierre Ossman <drzeus@drzeus.cx>