Rabin Vincent [Sun, 14 Jun 2015 17:26:11 +0000 (19:26 +0200)]
mmc: queue: prevent soft lockups on PREEMPT=n
On systems with CONFIG_PREEMPT=n, under certain circumstances, mmcqd
can continuously process requests for several seconds without blocking,
triggering the soft lockup watchdog. For example, this can happen if
mmcqd runs on the CPU which services the controller's interrupt, and
a process on a different CPU continuously writes to the MMC block
device.
NMI watchdog: BUG: soft lockup - CPU#0 stuck for 22s! [mmcqd/0:664]
CPU: 0 PID: 664 Comm: mmcqd/0 Not tainted 4.1.0-rc7+ #4
PC is at _raw_spin_unlock_irqrestore+0x24/0x28
LR is at mmc_start_request+0x104/0x134
...
[<805112a8>] (_raw_spin_unlock_irqrestore) from [<803db664>] (mmc_start_request+0x104/0x134)
[<803db664>] (mmc_start_request) from [<803dc008>] (mmc_start_req+0x274/0x394)
[<803dc008>] (mmc_start_req) from [<803eb2c4>] (mmc_blk_issue_rw_rq+0xd0/0xb98)
[<803eb2c4>] (mmc_blk_issue_rw_rq) from [<803ebe8c>] (mmc_blk_issue_rq+0x100/0x470)
[<803ebe8c>] (mmc_blk_issue_rq) from [<803ecab8>] (mmc_queue_thread+0xd0/0x170)
[<803ecab8>] (mmc_queue_thread) from [<8003fd14>] (kthread+0xe0/0xfc)
[<8003fd14>] (kthread) from [<8000f768>] (ret_from_fork+0x14/0x2c)
Fix it by adding a cond_resched() in the request handling loop so that
other processes get a chance to run.
Signed-off-by: Rabin Vincent <rabin.vincent@axis.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Ding Wang [Mon, 18 May 2015 12:14:15 +0000 (20:14 +0800)]
mmc: card: Fixup request missing in mmc_blk_issue_rw_rq
The current handler of MMC_BLK_CMD_ERR in mmc_blk_issue_rw_rq function
may cause new coming request permanent missing when the ongoing
request (previoulsy started) complete end.
The problem scenario is as follows:
(1) Request A is ongoing;
(2) Request B arrived, and finally mmc_blk_issue_rw_rq() is called;
(3) Request A encounters the MMC_BLK_CMD_ERR error;
(4) In the error handling of MMC_BLK_CMD_ERR, suppose mmc_blk_cmd_err()
end request A completed and return zero. Continue the error handling,
suppose mmc_blk_reset() reset device success;
(5) Continue the execution, while loop completed because variable ret
is zero now;
(6) Finally, mmc_blk_issue_rw_rq() return without processing request B.
The process related to the missing request may wait that IO request
complete forever, possibly crashing the application or hanging the system.
Fix this issue by starting new request when reset success.
Jiri Slaby [Fri, 12 Jun 2015 09:45:02 +0000 (11:45 +0200)]
mmc: sdhci: fix low memory corruption
When dma mapping (dma_map_sg) fails in sdhci_pre_dma_transfer, -EINVAL
is returned. There are 3 callers of sdhci_pre_dma_transfer:
* sdhci_pre_req and sdhci_adma_table_pre: handle negative return
* sdhci_prepare_data: handles 0 (error) and "else" (good) only
So teach sdhci_prepare_data to understand negative return values from
sdhci_pre_dma_transfer and disable DMA in that case, as well as for
zero.
It was introduced in 348487cb28e66b032bae1b38424d81bf5b444408 (mmc:
sdhci: use pipeline mmc requests to improve performance). The commit
seems to be suspicious also by assigning host->sg_count both in
sdhci_pre_dma_transfer and sdhci_adma_table_pre.
mmc: dw_mmc: insmod followed by rmmod will hung for eMMC
Remove module of dw_mmc driver will hung for eMMC devices if we follow the
steps which are listed below,
insmod dw_mmc.ko
insmod dw_mmc-pci.ko
rmmod dw_mmc-pci.ko
The root cause for this issue is, dw_mci_remove() will disable all the
interrupts by programming 0x0 to INTMASK register then it will call
dw_mci_cleanup_slot(). But dw_mci_cleanup_slot() is issuing CMD6 to
disable the eMMC boot partition and it is waiting for Command Complete
interrupt. Since INTMASK was already cleared by dw_mci_remove(), Command
Complete interrupt is not reaching the system. This leads to process hung.
Ulf Hansson [Fri, 5 Jun 2015 09:40:08 +0000 (11:40 +0200)]
mmc: sdhci: Restore behavior while creating OCR mask
Commit 3a48edc4bd68 ("mmc: sdhci: Use mmc core regulator infrastucture")
changed the behavior for how to assign the ocr_avail mask for the mmc
host. More precisely it started to mask the bits instead of assigning
them.
Restore the behavior, but also make it clear that an OCR mask created
from an external regulator overrides the other ones. The OCR mask is
determined by one of the following with this priority:
1. Supported ranges of external regulator if one supplies VDD
2. Host OCR mask if set by the driver (based on DT properties)
3. The capabilities reported by the controller itself
Fixes: 3a48edc4bd68 ("mmc: sdhci: Use mmc core regulator infrastucture") Cc: Tim Kryger <tim.kryger@gmail.com> Reported-by: Yangbo Lu <yangbo.lu@freescale.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Reviewed-by: Tim Kryger <tim.kryger@gmail.com>
MMC_PM_KEEP_POWER doesn't imply MMC_PM_WAKE_SDIO_IRQ, we should only
enable device wake up when MMC_PM_WAKE_SDIO_IRQ is set. And "pm_flags"
is the requested pm features, we should not set it in the host driver.
At the same time, device wakeup is disabled by default, so there's no
need to disable device wakeup explicitly.
This patch fixes the warning as following:
[ 64.616651] ------------[ cut here ]------------
[ 64.616665] WARNING: CPU: 0 PID: 79 at linux/kernel/irq/manage.c:603 irq_set_irq_wake+0xf0/0x11c()
[ 64.616667] Unbalanced IRQ 87 wake disable
Other subsystem buses attach PM domains during probe, but prior calling
the driver's ->probe() method. During the removal phase, detaching the PM
domain will be done after invoking the driver's ->remove() callback.
Convert the SDIO bus to follow this behavior and add error handling.
Eric Anholt [Fri, 29 May 2015 21:06:12 +0000 (14:06 -0700)]
mmc: sdhci-bcm2835: Actually enable the clock
We're currently using a fixed frequency clock specified in the DT, so
enabling is a no-op. However, the RPi firmware-based clocks driver
can actually disable unused clocks, so when switching to use it we
ended up losing our MMC clock once all devices were probed.
Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Yangbo Lu [Mon, 1 Jun 2015 05:47:12 +0000 (13:47 +0800)]
mmc: sdhci-of-esdhc: enable interrupt mode to detect card
Enable interrupt mode to detect card instead of polling mode
for P1020/P4080/P5020/P5040/T1040 by removing the quirk
SDHCI_QUIRK_BROKEN_CARD_DETECTION. This could improve data
transferring performance and avoid the call trace caused by
polling card status sometime.
Signed-off-by: Yangbo Lu <yangbo.lu@freescale.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Dong Aisheng [Wed, 27 May 2015 10:13:30 +0000 (18:13 +0800)]
mmc: sdhci-esdhc-imx: add imx6sx support
The imx6sx usdhc is derived from imx6sl, the difference is minor.
imx6sx have the errata ESDHC_FLAG_ERR004536 fixed.
So introduce a new compatible string for imx6sx to distinguish them.
Dong Aisheng [Wed, 27 May 2015 10:13:29 +0000 (18:13 +0800)]
mmc: sdhci-esdhc-imx: using specific compatible string in binding doc
Using specific compatible string in binding doc to make the binding
more clear.
It's also used to avoid checkpatch warning in the future like follows:
WARNING: DT compatible string "fsl,imx6sx-usdhc" appears un-documented --
check ./Documentation/devicetree/bindings/
+ { .compatible = "fsl,imx6sx-usdhc", .data = &usdhc_imx6sx_data, },
The uSDHC has an ADMA Length Mismatch errata ERR004536 which may
cause ADMA work abnormally. The errata has already been fixed for
i.MX6Q TO1.2 and i.MX6DL TO1.1 by enable the bit 7 in 0x6c register.
Unfortunately this fix is not included in i.MX6SL.
So we disable ADMA for i.MX6SL and use SDMA instead.
Dong Aisheng [Wed, 27 May 2015 10:13:26 +0000 (18:13 +0800)]
mmc: sdhci-esdhc-imx: merge the same register check into one place
In esdhc_writel_le() function, there's duplicated checking of the same
register as follows:
"if (unlikely(reg == SDHCI_INT_ENABLE || reg == SDHCI_SIGNAL_ENABLE))".
Merge them into one and remove the duplicated one.
Adrian Hunter [Fri, 6 Feb 2015 12:12:56 +0000 (14:12 +0200)]
mmc: core: Record card drive strength
In preparation for adding drive strength support
for eMMC, add drive_strength to struct mmc_card
to record the card drive strength for UHS-I modes
and HS200 / HS400. For eMMC this will be needed
when switching between HS200 and HS400.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Adrian Hunter [Fri, 6 Feb 2015 12:12:54 +0000 (14:12 +0200)]
mmc: core: Add 'card' to drive strength selection callback
In preparation for supporting also eMMC drive strength,
add the 'card' as a parameter so that the callback can
distinguish different types of cards if necessary.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Adrian Hunter [Fri, 6 Feb 2015 12:12:53 +0000 (14:12 +0200)]
mmc: core: Simplify card drive strength mask
Card drive strength selection uses a callback to
which a mask of supported drive strengths is passed.
Currently, the bits are checked against the values
in the SD specifications. That is not necessary
because the callback will anyway match the mask
against a valid value. Simplify by taking the mask
as is but still ensuring that the default mandatory
value (type B) is always supported.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Adrian Hunter [Fri, 6 Feb 2015 12:12:52 +0000 (14:12 +0200)]
mmc: core: Allow card drive strength to be different to host
Initialization of UHS-I modes for SD and SDIO cards
employs a callback to allow the host driver to
choose a drive strength value. Currently that
assumes the card drive strength and host driver
type must be the same value. Change to let the
callback make that decision and return both the
card drive strength and host driver type.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Adrian Hunter [Fri, 6 Feb 2015 12:12:51 +0000 (14:12 +0200)]
mmc: core: Reset driver type to default
IO state variable drv_type could be set during card
initialization. Consequently, it must be reset to the
default value when setting the initial state.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
mmc: sh_mmcif: calculate best clock with parent clock
MMCIF IP on R-Car series has parent clock which can be set several rate,
and it was not implemented on old SH-Mobile series (= SH-Mobile series
parent clock was fixed rate) R-Car series MMCIF can use more high speed
access if it setups parent clock. This patch adds parent clock setup
method. It will be used if DT has "max-frequency", and then, this driver
assumes it is booted on R-Car Gen2 or later SoC. Because SH-Mobile series
(which doesn't boot from DT) and R-Car series (which boots from DT) have
different divider.
mmc: sh_mmcif: add sh_mmcif_host_to_dev() macro and use it.
Current sh_mmcif driver is directly using &host->pd->dev in all place.
It is not big problem, but it is unreadable, and it can be cause of
future bug. This patch adds new sh_mmcif_host_to_dev() and use it.
Doug Anderson [Tue, 12 May 2015 21:46:11 +0000 (14:46 -0700)]
mmc: core: Increase delay for voltage to stabilize from 3.3V to 1.8V
Since the regulator used for the SDMMC IO voltage is not expected to
draw a lot of current, most systems will probably use an inexpensive
LDO for it. LDO regulators apparently have the feature that they
don't actively drive the voltage down--they wait for other components
in the system to drag the voltage down. Thus they will transition
faster under heavy loads and slower under light loads.
During an SDMMC voltage change from 3.3V to 1.8V, we are almost
certainly under a light load. To be specific:
* The regulator is hooked through pulls to CMD0-3 and DAT. Probably
the CMD pulls are something like 47K and the DAT is something like
10K.
* The card is supposed to be driving DAT0-3 low during voltage change
which will draw _some_ current, but not a lot.
* The regulator is also provided to the SDMMC host controller, but the
SDMMC host controller is in open drain mode during the voltage
change and so shouldn't be drawing much current.
In order to keep the SDMMC host working properly (or for noise
reasons), there might also be a capacitor attached to the SDMMC IO
regulator. This also will have the effect of slowing down transitions
of the regulator, especially under light loads.
From experimental evidence, we've seen the voltage change fail if the
card doesn't detect that the voltage fell to less than about 2.3V when
we turn on the clock. On one device (that admittedly had a 47K CMD
pullup instead of a 10K CMD pullup) we saw that the voltage was just
about 2.3V after 5ms and thus the voltage change would sometimes fail.
Doubling the delay gave margin and made the voltage change work 100%
of the time, despite the slightly weaker CMD pull.
At the moment submitting this as an RFC patch since my problem _could_
be fixed by increasing the pull strength (or using a smaller
capacitor). However being a little bit more lenient to strange
hardware could also be a good thing.
Signed-off-by: Doug Anderson <dianders@chromium.org> Acked-by: Mark Brown <broonie@kernel.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
mmc: dw_mmc: Use core to handle absent write protect line
Use the new MMC_CAP2_NO_WRITE_PROTECT to let the core handle the case where
no write protect line is present instead of having custom driver code to
handle it.
dw_mci_of_get_slot_quirks() is slightly refactored to directly modify the
mmc_host capabilities instead of returning a quirk mask.
mmc: cast unsigned int to typeof(sector_t) to avoid unexpected error
card->csd.capacity is defined as "unsigned int", and sector_t is defined as
"u64" or "unsigned long" (depends on CONFIG_LBDAF). Thus, sector_t data
might have strange data (see below). This patch cast it to typeof(sector_t)
Special thanks to coverity <http://www.coverity.com>
mmc: cast u8 to unsigned long long to avoid unexpected error
card->ext_csd.enhanced_area_offset is defined as "unsigned long long",
and, ext_csd[] is defined as u8.
unsigned long long data might have strange data if first bit of ext_csd[]
was 1. this patch cast it to (unsigned long long)
Special thanks to coverity <http://www.coverity.com>
Fabio Estevam [Sat, 9 May 2015 12:57:09 +0000 (09:57 -0300)]
mmc: sdhci-esdhc-imx: Do not break platform data boards
The only user of this driver that has not been converted to fully
device tree is the i.MX35 SoC.
There is a i.MX35-based board (mach-pcm043.c) that uses platform data
to pass wp_gpio and cd_gpio information.
Commit 8d86e4fcccf61ba ("mmc: sdhci-esdhc-imx: Call mmc_of_parse()")
broke the platform data case by removing mmc_gpio_request_ro() and
mmc_gpio_request_cd(), so restore the functionality for the non-dt
case.
Also, restore the check for ESDHC_CD_CONTROLLER so that we can still
support the "fsl,cd-controller" property.
mmc: dt: Allow to specify that no write protect signal is present
Allow to specify in the device-tree that no physical write-protect signal
is connected to a particular instance of a MMC controller. Setting the
property will cause the core will assume that the SD card is always
read-write.
The name for the new property is 'disable-wp' and was chosen based on the
property with the same function from the Synopsys designware mobile storage
host controller DT bindings specification.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de> Cc: Rob Herring <robh+dt@kernel.org> Cc: Pawel Moll <pawel.moll@arm.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Ian Campbell <ijc+devicetree@hellion.org.uk> Cc: Kumar Gala <galak@codeaurora.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
mmc: Add support for disabling write-protect detection
It is not uncommon to see systems where there is no physical write-protect
signal (e.g. when using eMMC or microSD card slots). For some controllers,
which have a dedicated write-protection detection logic (like SDHCI
controllers), the get_ro() callback can return bogus data in such a case.
Instead of handling this on a per controller basis this patch adds a new
capability flag to the MMC core that can be set to specify that the result
of get_ro() is invalid. When the flag is set the core will not call
get_ro() and assume that the card is always read-write.
Adrian Hunter [Thu, 7 May 2015 10:10:22 +0000 (13:10 +0300)]
mmc: core: Flag re-tuning is needed on CRC errors
CRC errors could possibly be alleviated by
re-tuning so flag re-tuning needed in those cases.
Note this has no effect if re-tuning has not been
enabled.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Adrian Hunter [Thu, 7 May 2015 10:10:21 +0000 (13:10 +0300)]
mmc: sdhci: Change to new way of doing re-tuning
Make use of mmc core support for re-tuning instead
of doing it all in the sdhci driver.
This patch also changes to flag the need for re-tuning
always after runtime suspend when tuning has been used
at initialization. Previously it was only done if
the re-tuning timer was in use.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Adrian Hunter [Thu, 7 May 2015 10:10:20 +0000 (13:10 +0300)]
mmc: core: Add support for HS400 re-tuning
HS400 re-tuning must be done in HS200 mode. Add
the ability to switch from HS400 mode to HS200
mode before re-tuning and switch back to HS400
after re-tuning.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Adrian Hunter [Thu, 7 May 2015 10:10:14 +0000 (13:10 +0300)]
mmc: core: Add support for re-tuning before each request
At the start of each request, re-tune if needed and
then hold off re-tuning again until the request is done.
Note that though there is one function that starts
requests (mmc_start_request) there are two that wait for
the request to be done (mmc_wait_for_req_done and
mmc_wait_for_data_req_done). Also note that
mmc_wait_for_data_req_done can return even when the
request is not done (which allows the block driver
to prepare a newly arrived request while still
waiting for the previous request).
This patch ensures re-tuning is held for the duration
of a request. Subsequent patches will also hold
re-tuning at other times when it might cause a
conflict.
In addition, possibly a command is failing because
re-tuning is needed. Use mmc_retune_recheck() to check
re-tuning. At that point re-tuning is held, at least by
the request, so mmc_retune_recheck() flags host->retune_now
if the hold count is 1.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Adrian Hunter [Thu, 7 May 2015 10:10:13 +0000 (13:10 +0300)]
mmc: core: Enable / disable re-tuning
Enable re-tuning when tuning is executed and
disable re-tuning when card is no longer initialized.
In the case of SDIO suspend, the card can keep power.
In that case, re-tuning need not be disabled, but, if
a re-tuning timer is being used, ensure it is disabled
and assume that re-tuning will be needed upon resume
since it is not known how long the suspend will last.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Adrian Hunter [Thu, 7 May 2015 10:10:12 +0000 (13:10 +0300)]
mmc: host: Add facility to support re-tuning
Currently, there is core support for tuning during
initialization. There can also be a need to re-tune
periodically (e.g. sdhci) or to re-tune after the
host controller is powered off (e.g. after PM
runtime suspend / resume) or to re-tune in response
to CRC errors.
The main requirements for re-tuning are:
- ability to enable / disable re-tuning
- ability to flag that re-tuning is needed
- ability to re-tune before any request
- ability to hold off re-tuning if the card is busy
- ability to hold off re-tuning if re-tuning is in
progress
- ability to run a re-tuning timer
To support those requirements 7 members are added to struct
mmc_host:
unsigned int can_retune:1; /* re-tuning can be used */
unsigned int doing_retune:1; /* re-tuning in progress */
unsigned int retune_now:1; /* do re-tuning at next req */
int need_retune; /* re-tuning is needed */
int hold_retune; /* hold off re-tuning */
unsigned int retune_period; /* re-tuning period in secs */
struct timer_list retune_timer; /* for periodic re-tuning */
need_retune is an integer so it can be set without needing
synchronization. hold_retune is a integer to allow nesting.
Various simple functions are provided to set / clear those
variables.
Subsequent patches take those functions into use.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Weijun Yang [Mon, 27 Apr 2015 08:15:13 +0000 (08:15 +0000)]
mmc: sdhci-sirf: fix the tuning count in platform_execute_tuning
hardware has 16bit to record the tuning count, so fix it to 16384.
at the same time, tuned_phases[SIRF_TUNING_COUNT] is useless as the
array is never used, so move it to a variant.
Signed-off-by: Weijun Yang <Weijun.Yang@csr.com> Signed-off-by: Barry Song <Baohua.Song@csr.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
mmc: sh_mmcif: separate sh_mmcif_clk_update() into setup and prepare
Current sh_mmcif_clk_update() is called from probe() and set_ios(),
but, the purpose of later one is just clk_prepare_enable().
No need to setup mmc->f_max/f_min in many times.
This patch separe sh_mmcif_clk_update() into clk_prepare_enable()
and mmc->f_max/f_min setup.
mmc: sh_mmcif: remove unnecessary int clk from struct sh_mmcif_host
struct sh_mmcif_host has 1) int clk, 2) struct clock *hclk,
and host->clk = clk_get_rate(host->hclk).
This int clk is not necessary. Let's remove it.
And, current hclk is confusable naming. Let's rename it to clk.
Andreas Fenkart [Thu, 23 Apr 2015 11:15:42 +0000 (13:15 +0200)]
mmc: sdio: add reset callback to bus operations
Some drivers schedule automatic hw resets. An example is mwifiex,
which schedules a card reset if the command handler between driver
and card firmware becomes out of sync
Signed-off-by: Andreas Fenkart <afenkart@gmail.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Linus Torvalds [Sun, 24 May 2015 18:15:28 +0000 (11:15 -0700)]
Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI fixes from James Bottomley:
"This is a set of five fixes: Two MAINTAINER email updates (urgent
because the non-avagotech emails will start bouncing) an lpfc big
endian oops fix, a 256 byte sector hang fix (to eliminate 256 byte
sectors) and a storvsc fix which could cause test unit ready failures
on bringup"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
MAINTAINERS: Revise lpfc maintainers for Avago Technologies ownership of Emulex
MAINTAINERS, be2iscsi: change email domain
sd: Disable support for 256 byte/sector disks
lpfc: Fix breakage on big endian kernels
storvsc: Set the SRB flags correctly when no data transfer is needed
projects / karo-tx-linux.git / log