Currently the driver imposes a limit of 256 total minor numbers,
apparently based on the historic Unix/Linux limit.  This is quite
restrictive, particularly if we raise the maximum number of
partitions per card to 256 to match sd.

In order to make the full minor number space available we would
have to replace the static dev_use and name_use arrays with struct
ida.  But we can at least allow use of 256 cards rather than just
256 minors, with only a small change.
Signed-off-by: NBen Hutchings <ben@decadent.org.uk>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

By using mmc_get_ext_csd() in favor of mmc_send_ext_csd, we decrease
code duplication.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

In most of the cases mmc_get|set_drvdata() didn't simplify code, which
should be the primary reason for such macros.

Let's remove them and convert to the common device_driver macros,
dev_set|get_drvdata() instead.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

The struct mmc_driver adds an extra layer on top of the struct
device_driver. That would be fine, if there were a good reason, but
that's not the case.

Let's simplify code by converting to the common struct device_driver
instead and thus also removing superfluous overhead.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Instead of having specific mmc system PM callbacks for the mmc driver,
let's convert to use the common ones.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Make force_ro consistent with other sysfs entries.

Fixes: 371a689f ('mmc: MMC boot partitions support')
Cc: Andrei Warkentin <andrey.warkentin@gmail.com>
Signed-off-by: NBaruch Siach <baruch@tkos.co.il>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Use the much more common pr_warn instead of pr_warning.

Other miscellanea:

o Coalesce formats
o Realign arguments
o Remove extra spaces when coalescing formats
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Stop command errors are not fatal to the transfer since we make sure
that the card returns to the transfer state and check the card status.
Change an unnecessary error to an info.
Signed-off-by: NJohan Rudholm <johanru@axis.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

It seems very unlikely that eMMC devices would hold a standard
partitiontable in one of it's boot areas. Therefore, let's prevent
them from being scanned.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Now, mmc framework uses multi_io_quirk
for I/O HW bug workaround.
MMC_CAP2_NO_MULTI_READ flag is no longer needed
Signed-off-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Historically, we have been using MMC_CAP* to handle host HW issues and
currently the block layer uses MMC_CAP2_NO_MULTI_READ flag for a multi
I/O HW bug workaround.

There are a few tweaks needed to make MMC_CAP2_NO_MULTI_READ suite all
situations. Therefore let's add an optional host ops callback to enable
host drivers to return the number of blocks it allows per request.

In a future patch and when host drivers have converted to the new
callback, MMC_CAP2_NO_MULTI_READ shall be removed.
Signed-off-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

Currently when the device secure discard implementation is
blacklisted (MMC_QUIRK_SEC_ERASE_TRIM_BROKEN quirk is set)
instead of secure discard we're going to do normal discard,
which is wrong.

When the secure discard is known to be broken we should just
disallow it entirely and not advertise this functionality to
the user. Fix it.

Also move mmc_fixup_device() in from of mmc_blk_alloc() so we
can get quirks set before we attempt to set queue information.
Signed-off-by: NLukas Czerner <lczerner@redhat.com>
Acked-by: NJaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>

When sending a stop command at the recovery path, use a R1B response
when the failing data request are a WRITE. Thus we also care about the
busy detection completion in this case.

For a failing READ request, we use a R1 response for the stop command,
since we don't need to care about busy detection in this case.

To align behavior between hosts supporting MMC_CAP_WAIT_WHILE_BUSY and
those who are not, we add a CMD13 polling method for the card's status.

We also respect whether the host has specified the max_busy_timeout,
which means we may fallback to CMD13 polling if the timeout is greater
than what the host are able to support.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <chris@printf.net>

Currently for write request we don't trust the hw busy detection to be
fully handled by host, thus we also poll the card's status until we see
it's gets out of the busy state.

Still there are scenarios where it will a benefit to trust the hw busy
detection done by the host, since no additional polling is needed.
Let's prepare card_busy_detect() to be able to handle this.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <chris@printf.net>

To complete a data write request we poll for the card's status register
by sending CMD13. The are other scenarios when this polling method are
needed, which is why we here moves this code to it's own function. No
functional change.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <chris@printf.net>

While using open ended transmission and thus ending the transfer by
sending a stop command, we shall use R1B only for writes and R1 shall
be used for reads. Previously R1B were used in both cases.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <chris@printf.net>

There is no need for keeping a host cap for MMC_CAP2_SANITIZE, instead
we just make the feature default available.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <chris@printf.net>

Under function mmc_blk_issue_rq, after an MMC discard operation,
the MMC request data structure may be freed in memory. Later in
the same function, the check of req->cmd_flags & MMC_REQ_SPECIAL_MASK
is dangerous and invalid. It causes the MMC host not to be released
when it should.

This patch fixes the issue by marking the special request down before
the discard/flush operation.

Reported by: Harold (SoonYeal) Yang <haroldsy@broadcom.com>
Signed-off-by: NRay Jui <rjui@broadcom.com>
Reviewed-by: NSeungwon Jeon <tgih.jun@samsung.com>
Acked-by: NSeungwon Jeon <tgih.jun@samsung.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

By adding a card state that records if it is suspended or resumed, we
can accept asyncronus suspend/resume requests for the mmc and sd
bus_ops.

MMC_CAP_AGGRESSIVE_PM, will at request inactivity through the runtime
bus_ops callbacks, execute a suspend of the the card. In the state were
this has been done, we can receive a suspend request for the mmc bus,
which for sd and mmc forced the card to active state by a
pm_runtime_get_sync. In other words, the card was resumed and then
immediately suspended again, completely unnecessary.

Since the suspend/resume bus_ops callbacks for sd and mmc are now
capable of handling asynchronous requests, we no longer need to force
the card to active state before executing suspend. Evidently preventing
the above sequence for MMC_CAP_AGGRESSIVE_PM.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

A previous commit (fdfa20c1) reordered the shutdown sequence
in mmc_blk_remove_req. However, mmc_cleanup_queue is now called before
we get the card pointer, and mmc_cleanup_queue sets mq->card to NULL.

This patch moves the card pointer assignment before mmc_cleanup_queue.
Signed-off-by: NFranck Jullien <franck.jullien@gmail.com>
Signed-off-by: NChris Ball <cjb@laptop.org>

Current MMC driver doesn't handle generic error (bit19 of device
status) in write sequence. As a result, write data gets lost when
generic error occurs. For example, a generic error when updating a
filesystem management information causes a loss of write data and
corrupts the filesystem. In the worst case, the system will never
boot.

This patch includes the following functionality:
  1. To enable error checking for the response of CMD12 and CMD13
     in write command sequence
  2. To retry write sequence when a generic error occurs

Messages are added for v2 to show what occurs.
Signed-off-by: NKOBAYASHI Yoshitake <yoshitake.kobayashi@toshiba.co.jp>
Signed-off-by: NChris Ball <cjb@laptop.org>

Considering shutdown of the card, the responsibility to initate this
sequence shall be driven from the mmc_bus.

This patch enables the mmc_bus to handle this sequence properly. A new
.shutdown callback is added in the mmc_driver struct which is used to
shutdown the blk device.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

Inside the routine mmc_blk_ioctl_cmd() the sanitize command is
identified according to the value of bits 16-23 of the argument.

but what happens if a different command is sent, and only by
chance, bits 16-23 contain the value of SANITIZE command ?
In that case a SANITIZE command will be falsely identified.
In order to prevent such a case, the condition is expanded and
now it also checks the opcode itself, and verifies that it is an
MMC_SWITCH opcode.
Signed-off-by: NYaniv Gardi <ygardi@codeaurora.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

We had a multi-partition SD-Card with two ext2 file systems. The partition
table was getting overwritten by a race between the card removal and
the unmount of the 2nd ext2 partition.

What was observed:
1. Suspend/resume would call to remove the device. The clearing
   of the device information is done asynchronously.
2. A request is made to unmount the file system (this is called
   after the removal has started).
3. The remapping table was cleared by the asynchronous part of
   the device removal.
4. A write request to the super block (block 0 of the partition)
   was sent down and instead of being remapped to the partition
   offset, it was remapped to block 0 of the device which is where
   the partition table is located.
5. Write was queued and written resulting in the overwriting
   of the partition table with the ext2 super block.
6. The mmc_queue is cleaned up.

The mmc card device driver used to access SD cards, was calling del_gendisk
before calling mmc_cleanup-queue. The comment in the mmc_blk_remove_req
code indicated that it expected del_gendisk to block all further requests
from being queued but it doesn't. The mmc driver uses the presences of the
mmc_queue to determine if the request should be queued.

The fix was to clean up the mmc_queue before the rest of the
the delete partition code is called.

This prevents the overwriting of the partition table.

However, the umount gets an error trying to write the super block.
The umount should be issued before the device is removed but that
is not always possible. The umount is still needed to cleanup other
data structures.

Addresses the problem described in http://crbug.com/240815Signed-off-by: NPaul Taysom <taysom@chromium.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

Once the mmc blkdevice is being probed, runtime pm will be enabled.
By using runtime autosuspend, the power save operations can be done
when request inactivity occurs for a certain time. Right now the
selected timeout value is set to 3 s. Obviously this value will likely
need to be configurable somehow since it needs to be trimmed depending
on the power save algorithm.

For SD-combo cards, we are still leaving the enablement of runtime PM
to the SDIO init sequence since it depends on the capabilities of the
SDIO func driver.

Moreover, when the blk device is being suspended, we make sure the device
will be runtime resumed. The reason for doing this is that we want the
host suspend sequence to be unaware of any runtime power save operations
done for the card in this phase. Thus it can just handle the suspend as
the card is fully powered from a runtime perspective.

Finally, this patch prepares to make it possible to move BKOPS handling
into the runtime callbacks for the mmc bus_ops. Thus IDLE BKOPS can be
accomplished.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

The sanitize support is added as a user-app ioctl call, and
was removed from the block-device request, since its purpose is
to be invoked not via File-System but by a user.

This feature deletes the unmap memory region of the eMMC card,
by writing to a specific register in the EXT_CSD.

unmap region is the memory region that was previously deleted
(by erase, trim or discard operation).

In order to avoid timeout when sanitizing large-scale cards,
the timeout for sanitize operation is 240 seconds.
Signed-off-by: NYaniv Gardi <ygardi@codeaurora.org>
Signed-off-by: NMaya Erez <merez@codeaurora.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

The value passed is 0 in all but "it can never happen" cases (and those
only in a couple of drivers) *and* it would've been lost on the way
out anyway, even if something tried to pass something meaningful.
Just don't bother.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

For normal request mmc_blk_issue_rq is called twice with asynchronous
transfer(cur and prev). Host's claim and release can be done in each
mmc_blk_issue_rq. However, Special request is currently excluded in
asynchronous transfer. After special request is finished, if there is
no new request, mmc_release_host won't be called in mmc_blk_issue_rq.
The problem is founded during mmc_suspend.

[<c0541124>] (__schedule+0x0/0x78c) from [<c05419e8>] (schedule+0x38/0x78)
[<c05419b0>] (schedule+0x0/0x78) from [<c03a843c>] (__mmc_claim_host+0xac/0x1b4)
[<c03a8390>] (__mmc_claim_host+0x0/0x1b4) from [<c03ac98c>] (mmc_suspend+0x28/0x9c)
[<c03ac964>] (mmc_suspend+0x0/0x9c) from [<c03aad24>] (mmc_suspend_host+0xb4/0x194)
...
Reported-by: NJohan Rudholm <jrudholm@gmail.com>
Signed-off-by: NSeungwon Jeon <tgih.jun@samsung.com>
Tested-by: NJohan Rudholm <johan.rudholm@stericsson.com>
Signed-off-by: NChris Ball <cjb@laptop.org>

This patch supports packed write command of eMMC4.5 devices.  Several
writes can be grouped in packed command and all data of the individual
commands can be sent in a single transfer on the bus. Large amounts of
data in one transfer rather than several data of small size are
effective for eMMC write internally.  As a result, packed command help
write throughput be improved.  The following tables show the results
of packed write.

Type A:
test     none |  packed
iozone   25.8 |  31
tiotest  27.6 |  31.2
lmdd     31.2 |  35.4

Type B:
test     none |  packed
iozone   44.1 |  51.1
tiotest  47.9 |  52.5
lmdd     51.6 |  59.2

Type C:
test     none |  packed
iozone   19.5 |  32
tiotest  19.9 |  34.5
lmdd     22.8 |  40.7
Signed-off-by: NSeungwon Jeon <tgih.jun@samsung.com>
Reviewed-by: NMaya Erez <merez@codeaurora.org>
Reviewed-by: NNamjae Jeon <linkinjeon@gmail.com>
Signed-off-by: NChris Ball <cjb@laptop.org>

It's not necessary to start a new request while error handling if
the card was removed.
Signed-off-by: NSeungwon Jeon <tgih.jun@samsung.com>
Acked-by: NJaehoon Chung <jh80.chung@samsung.com>
Tested-by: NKonstantin Dorfman <kdorfman@codeaurora.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

When current request is running on the bus and if next request fetched
by mmcqd is NULL, mmc context (mmcqd thread) gets blocked until the
current request completes. This means that if new request comes in while
the mmcqd thread is blocked, this new request can not be prepared in
parallel to current ongoing request. This may result in delaying the new
request execution and increase it's latency.

This change allows to wake up the MMC thread on new request arrival.
Now once the MMC thread is woken up, a new request can be fetched and
prepared in parallel to the current running request which means this new
request can be started immediately after the current running request
completes.

With this change read throughput is improved by 16%.
Signed-off-by: NKonstantin Dorfman <kdorfman@codeaurora.org>
Reviewed-by: NSeungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: NChris Ball <cjb@laptop.org>

When current request is running on the bus and if next request fetched
by mmcqd is NULL, mmc context (mmcqd thread) gets blocked until the
current request completes. This means that if new request comes in while
the mmcqd thread is blocked, this new request can not be prepared in
parallel to current ongoing request. This may result in delaying the new
request execution and increase it's latency.

This change allows to wake up the MMC thread on new request arrival.
Now once the MMC thread is woken up, a new request can be fetched and
prepared in parallel to the current running request which means this new
request can be started immediately after the current running request
completes.

With this change read throughput is improved by 16%.
Signed-off-by: NKonstantin Dorfman <kdorfman@codeaurora.org>
Reviewed-by: NSeungwon Jeon <tgih.jun@samsung.com>
Signed-off-by: NChris Ball <cjb@laptop.org>

RPMB partition is accessing though /dev/block/mmcXrpmb device
User callers can read and write entire data frame(s) as defined
by JEDEC Standard JESD84-A441, using standard IOCTL interface.
Signed-off-by: NAlex Macro <alex.macro@stericsson.com>
Signed-off-by: NLoic Pallardy <loic.pallardy@stericsson.com>
Reviewed-by: NNamjae Jeon <linkinjeon@gmail.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NKrishna Konda <kkonda@codeaurora.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

Do not scan rpmb partitions for "soft" partitions, since the rpmb
partition contains protected data. Silences the following message
during boot:

 mmcblkXRPMB: unknown partition table
Signed-off-by: NJohan Rudholm <johan.rudholm@stericsson.com>
Reviewed-by: NNamjae Jeon <linkinjeon@gmail.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NKrishna Konda <kkonda@codeaurora.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

There are infinite loops in the mmc code that can be caused by bad
hardware.  The code will loop forever if the device never comes back
from program mode, R1_STATE_PRG, and it is not ready for data,
R1_READY_FOR_DATA.

A long timeout is added to prevent the code from looping forever.
The timeout will occur if the device never comes back from program
state or the device never becomes ready for data.

It's not clear whether the timeout will do more than log a pr_err()
and then start a fresh hang all over again.  We may need to extend
this patch later to perform some kind of reset of the device (is
that possible?) or rejection of new I/O to the device.
Signed-off-by: NTrey Ramsay <tramsay@linux.vnet.ibm.com>
Signed-off-by: NChris Ball <cjb@laptop.org>

For several MoviNAND eMMC parts, there are known issues with secure
erase and secure trim.  For these specific MoviNAND devices, we skip
these operations.

Specifically, there is a bug in the eMMC firmware that causes
unrecoverable corruption when the MMC is erased with MMC_CAP_ERASE
enabled.

References:

http://forum.xda-developers.com/showthread.php?t=1644364
https://plus.google.com/111398485184813224730/posts/21pTYfTsCkB#111398485184813224730/posts/21pTYfTsCkBSigned-off-by: NIan Chen <ian.cy.chen@samsung.com>
Reviewed-by: NNamjae Jeon <linkinjeon@gmail.com>
Acked-by: NJaehoon Chung <jh80.chung@samsung.com>
Reviewed-by: NLinus Walleij <linus.walleij@linaro.org>
Cc: stable <stable@vger.kernel.org> [3.0+]
Signed-off-by: NChris Ball <cjb@laptop.org>

For completing any block request, MMC block driver is calling:
	spin_lock_irq(queue)
	__blk_end_request()
	spin_unlock_irq(queue)

But if we analyze the sources of latency in kernel using ftrace,
__blk_end_request() function at times may take up to 6.5ms with
spinlock held and irq disabled.

__blk_end_request() calls couple of functions and ftrace output
shows that blk_update_bidi_request() function is almost taking 6ms.
There are 2 function to end the current request: ___blk_end_request()
and blk_end_request(). Both these functions do same thing except
that blk_end_request() function doesn't take up the spinlock
while calling the blk_update_bidi_request().

This patch replaces all __blk_end_request() calls with
blk_end_request() and __blk_end_request_all() calls with
blk_end_request_all().

Testing done: 20 process concurrent read/write on sd card
and eMMC. Ran this test for almost a day on multicore system
and no errors observed.

This change is not meant for improving MMC throughput; it's basically
about becoming fair to other threads/interrupts in the system. By
holding spin lock and interrupts disabled for longer duration, we
won't allow other threads/interrupts to run at all.  Actually slight
performance degradation at file system level can be expected as we
are not holding the spin lock during blk_update_bidi_request() which
means our mmcqd thread may get preempted for other high priority
thread or any interrupt in the system.

These are performance numbers (100MB file write) with eMMC running
in DDR mode:

Without this patch:
	Name of the Test   Value   Unit
	LMDD Read Test     53.79   MBPS
	LMDD Write Test    18.86   MBPS
	IOZONE  Read Test  51.65   MBPS
	IOZONE  Write Test 24.36   MBPS

With this patch:
	Name of the Test    Value  Unit
	LMDD Read Test      52.94  MBPS
	LMDD Write Test     16.70  MBPS
	IOZONE  Read Test   52.08  MBPS
	IOZONE  Write Test  23.29  MBPS

Read numbers are fine. Write numbers are bit down (especially LMDD
write), may be because write requests normally have large transfer
size and which means there are chances that while mmcq is executing
blk_update_bidi_request(), it may get interrupted by interrupts or
other high priority thread.
Signed-off-by: NSubhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: NNamjae Jeon <linkinjeon@gmail.com>
Signed-off-by: NChris Ball <cjb@laptop.org>

If multi block write operation fails for SD card, during
error handling we send the SD_APP_SEND_NUM_WR_BLKS (ACMD22)
to know how many blocks were already programmed by card.

But mmc_sd_num_wr_blocks() function which sends the ACMD22
calculates the data timeout value from csd.tacc_ns and
csd.tacc_clks parameters which will be 0 for block addressed
(>2GB cards) SD card. This would result in timeout_ns and
timeout_clks being 0 in the mmc_request passed to host driver.
This means host controller would program its data timeout timer
value with 0 which could result in DATA TIMEOUT errors from
controller.

To fix this issue, mmc_sd_num_wr_blocks() should instead
just call the mmc_set_data_timeout() to calculate the
data timeout value. mmc_set_data_timeout() function
ensures that non zero timeout value is set even for
block addressed SD cards.
Signed-off-by: NSubhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: NVenkatraman S <svenkatr@ti.com>
Signed-off-by: NChris Ball <cjb@laptop.org>

This patch adds support for large sector size of 4KB by disabling
emulation.  This patch passes eMMC DATA_SECTOR_SIZE as the logical
block size during mmc_blk_alloc_req.

In order to use this patch for 4KB sector size, ensure that
USE_NATIVE_SECTOR is enabled, partition table is 4KB sector size
aligned and file system block and sector size are 4KB multiples.
Signed-off-by: NSaugata Das <saugata.das@linaro.org>
Reviewed-by: NSubhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: NNamjae Jeon <linkinjeon@gmail.com>
Reviewed-by: NSubhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: NChris Ball <cjb@laptop.org>

After the null check on md the code jumped to cmd_done, which then
will dereference md in mmc_blk_put. This patch avoids the possible
null pointer dereference in that case.
Signed-off-by: NPhilippe De Swert <philippedeswert@gmail.com>
Reviewed-by: NNamjae Jeon <linkinjeon@gmail.com>
Signed-off-by: NChris Ball <cjb@laptop.org>