When running the ARTPEC-8 DWMMC IP version, and a data error interrupt
comes during a data read transfer, there is no guarantee for the data
transfer over interrupt (DTO) to come within the specified data timeout.
This case is handled by the dto_timer handler which will complete the
request with the comment:
/*
* If DTO interrupt does NOT come in sending data state,
* we should notify the driver to terminate current transfer
* and report a data timeout to the core.
*/
But since the ARTPEC-8 DWMMC IP version, supports an extended TMOUT
register which allows longer timeouts than the non ARTPEC-8 version
does, waiting for the dto_timer to complete the request in error cases
may cause the request to take significantly longer time than necessary.
This is specifically true for the failing steps during tuning of a
device.
Fix this by completing the request when the error interrupt comes. Since
this fix is specific for the ARTPEC-8, a quirk is added.
Signed-off-by: Mårten Lindahl <marten.lindahl@axis.com>
Link: https://lore.kernel.org/r/20211220113026.21129-5-marten.lindahl@axis.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Current dw_mci driver supports a TMOUT register which consists of a 24
bit field (TMOUT[31:8]) for the DATA_TIMEOUT. The maximum value of this
field is 0xFFFFFF, which with a 200MHz clock will give a full DRTO of:
0xFFFFFF / 200000000 => ~84 ms
However, the ARTPEC-8 SoC DWMMC IP version has a TMOUT register with an
extended DATA_TIMEOUT field, which supports longer timers for the DRTO.
In this version the DATA_TIMEOUT field is split into two, which with the
same 200MHz clock as above will allow a maximum timeout of:
((TMOUT[10:8] -1) * 0xFFFFFF + TMOUT[31:11] * 8) / 200000000 => ~587 ms
Add driver callbacks for implementation specific data timeout, and
implement callback functions for the ARTPEC-8 SoC.
Signed-off-by: Mårten Lindahl <marten.lindahl@axis.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20211220113026.21129-4-marten.lindahl@axis.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The "0x" prefix is redundant when # flag is used. It prints "0x0x".
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Mårten Lindahl <marten.lindahl@axis.com>
Link: https://lore.kernel.org/r/20211217150348.GD16611@kili
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The caps field depends on the mshcN alias ID but for some devices this
is unnecessary as the capabilities are the same for all instances
sharing the same compatible.
Add a common_caps field for this case which updates the host's
capabilities without needing the mshcN alias ID.
Signed-off-by: John Keeping <john@metanate.com>
Tested-by: Nicolas Frattaroli <frattaroli.nicolas@gmail.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Link: https://lore.kernel.org/r/20211124184603.3897245-2-john@metanate.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The TMOUT register is always set with a full value for every transfer,
which (with a 200MHz clock) will give a full DRTO of ~84 milliseconds.
This is normally good enough to complete the request, but setting a full
value makes it impossible to test shorter timeouts, when for example
testing data read times on different SD cards.
Add a function to set any value smaller than the maximum of 0xFFFFFF.
Signed-off-by: Mårten Lindahl <marten.lindahl@axis.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Link: https://lore.kernel.org/r/20211119155337.14341-1-marten.lindahl@axis.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
If we get a data error during a block transfer command, a stop command
(CMD12) is normally initiated. But this does not work for the general
command (CMD56), but instead the action is ignored and an uninitialized
command struct is used for the stop action, with unexpected result.
Fix this by adding a check for GEN_CMD when preparing stop transmission.
Signed-off-by: Mårten Lindahl <marten.lindahl@axis.com>
Link: https://lore.kernel.org/r/20211103182716.28419-1-marten.lindahl@axis.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
For dw_mmc, the ->init_card() callback is being used to turn on/off
automatic internal clock gating for powersave, which is needed to properly
support SDIO irqs on DAT1.
However, using the ->init_card() comes with a drawback in this case, as it
means that the powersave feature becomes disabled, no matter whether the
SDIO irqs becomes turned on or not. To improve the behaviour, let's change
into using the ->enable_sdio_irq() callback instead. This works fine,
because dw_mmc uses sdio_signal_irq() to signal the irqs, thus the
->enable_sdio_irq() is never executed from within atomic context.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Link: https://lore.kernel.org/r/20211020102907.70195-1-ulf.hansson@linaro.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Only wait for DRTO on reads, otherwise the driver hangs.
The driver prevents sending CMD12 on response errors like CRCs. According
to the comment this is because some cards have problems with this during
the UHS tuning sequence. Unfortunately this workaround currently also
applies for any command with data. On reads this will set the drto timer,
which then triggers after a while. On writes this will not set any timer
and the tasklet will not be scheduled again.
I cannot test for the UHS workarounds need, but even if so, it should at
most apply to reads. I have observed many hangs when CMD25 response
contained a CRC error. This patch fixes this without touching the actual
UHS tuning workaround.
Signed-off-by: Christian Loehle <cloehle@hyperstone.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/af8f8b8674ba4fcc9a781019e4aeb72c@hyperstone.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The fault injection function can set EVENT_DATA_ERROR but skip the
setting of ->data_status to an error status if it hits just after a data
over interrupt. This confuses the tasklet which can later end up
triggering the WARN_ON(host->cmd || ..) in dw_mci_request_end() since
dw_mci_data_complete() would return success.
Prevent the fault injection function from doing this since this is not a
real case, and ensure that the fault injection doesn't race with a real
error either.
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Fixes: 2b8ac062f3 ("mmc: dw_mmc: Add data CRC error injection")
Link: https://lore.kernel.org/r/20210825114213.7429-1-vincent.whitchurch@axis.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Depending on the DMA driver being used, the struct dma_slave_config may
need to be initialized to zero for the unused data.
For example, we have three DMA drivers using src_port_window_size and
dst_port_window_size. If these are left uninitialized, it can cause DMA
failures.
For dw_mmc, this is probably not currently an issue but is still good to
fix though.
Fixes: 3fc7eaef44 ("mmc: dw_mmc: Add external dma interface support")
Cc: Shawn Lin <shawn.lin@rock-chips.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Cc: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Link: https://lore.kernel.org/r/20210810081644.19353-2-tony@atomide.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
This driver has had problems when handling data errors. Add fault
injection support so that the abort handling can be easily triggered and
regression-tested. A hrtimer is used to indicate a data CRC error at
various points during the data transfer.
Note that for the recent problem with hangs in the case of some data CRC
errors, a udelay(10) inserted at the start of send_stop_abort() greatly
helped in triggering the error, but I've not included this as part of
the fault injection support since it seemed too specific.
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Link: https://lore.kernel.org/r/20210701080534.23138-1-vincent.whitchurch@axis.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
When a Data CRC interrupt is received, the driver disables the DMA, then
sends the stop/abort command and then waits for Data Transfer Over.
However, sometimes, when a data CRC error is received in the middle of a
multi-block write transfer, the Data Transfer Over interrupt is never
received, and the driver hangs and never completes the request.
The driver sets the BMOD.SWR bit (SDMMC_IDMAC_SWRESET) when stopping the
DMA, but according to the manual CMD.STOP_ABORT_CMD should be programmed
"before assertion of SWR". Do these operations in the recommended
order. With this change the Data Transfer Over is always received
correctly in my tests.
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20210630102232.16011-1-vincent.whitchurch@axis.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
As of commit bb475230b8 ("reset: make optional functions really
optional"), the reset framework API calls use NULL pointers to describe
optional, non-present reset controls.
This allows to return errors from devm_reset_control_get_optional and to
call reset_control_(de)assert unconditionally.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Link: https://lore.kernel.org/r/20210305090724.18832-1-p.zabel@pengutronix.de
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The ->card_event callback isn't being used by any of the dw_mmc variants.
It's likely a leftover from an earlier change, hence let's just drop the
redundant call to it.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
This converts the driver to use the new tasklet API introduced in
commit 12cc923f1c ("tasklet: Introduce new initialization API")
Signed-off-by: Emil Renner Berthing <kernel@esmil.dk>
Link: https://lore.kernel.org/r/20210204151847.91353-4-kernel@esmil.dk
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Common pattern of handling deferred probe can be simplified with
dev_err_probe(). Less code, the error value gets printed and real error
from dw_mci_parse_dt() is passed further instead of fixed -EINVAL.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Link: https://lore.kernel.org/r/20200902193658.20539-10-krzk@kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The MMC_CAP_ERASE bit is no longer used by the mmc core as erase, discard
and trim operations are now always supported. Therefore, drop the bit and
move all mmc hosts away from using it.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Rui Miguel Silva <rmfrfs@gmail.com>
Link: https://lore.kernel.org/r/20200508112902.23575-1-ulf.hansson@linaro.org
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Patch all drivers which use mmc_regulator_set_vqmmc() and prepare them for
the fact that mmc_regulator_set_vqmmc() can return a value > 0, which would
happen if the signal voltage switch did NOT happen, because the voltage was
already set correctly.
Signed-off-by: Marek Vasut <marex@denx.de>
Link: https://lore.kernel.org/r/20200416163649.336967-1-marex@denx.de
[Ulf: Re-worked/simplified the code a bit]
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
dma_request_slave_channel() is a wrapper on top of dma_request_chan()
eating up the error code.
By using dma_request_chan() directly the driver can support deferred
probing against DMA.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20191217112656.30860-1-peter.ujfalusi@ti.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Here is the "big" set of driver core patches for 5.5-rc1
There's a few minor cleanups and fixes in here, but the majority of the
patches in here fall into two buckets:
- debugfs api cleanups and fixes
- driver core device link support for boot dependancy issues
The debugfs api cleanups are working to slowly refactor the debugfs apis
so that it is even harder to use incorrectly. That work has been
happening for the past few kernel releases and will continue over time,
it's a long-term project/goal
The driver core device link support missed 5.4 by just a bit, so it's
been sitting and baking for many months now. It's from Saravana Kannan
to help resolve the problems that DT-based systems have at boot time
with dependancy graphs and kernel modules. Turns out that no one has
actually tried to build a generic arm64 kernel with loads of modules and
have it "just work" for a variety of platforms (like a distro kernel)
The big problem turned out to be a lack of depandancy information
between different areas of DT entries, and the work here resolves that
problem and now allows devices to boot properly, and quicker than a
monolith kernel.
All of these patches have been in linux-next for a long time with no
reported issues.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
iG0EABECAC0WIQT0tgzFv3jCIUoxPcsxR9QN2y37KQUCXd6m6Q8cZ3JlZ0Brcm9h
aC5jb20ACgkQMUfUDdst+yntJQCcCqg6RQ7LTdHuZv1ETeefXlsfk00An1Jtean6
42bWGx52bGFvAcpjWy8R
=P7hq
-----END PGP SIGNATURE-----
Merge tag 'driver-core-5.5-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core
Pull driver core updates from Greg KH:
"Here is the "big" set of driver core patches for 5.5-rc1
There's a few minor cleanups and fixes in here, but the majority of
the patches in here fall into two buckets:
- debugfs api cleanups and fixes
- driver core device link support for boot dependancy issues
The debugfs api cleanups are working to slowly refactor the debugfs
apis so that it is even harder to use incorrectly. That work has been
happening for the past few kernel releases and will continue over
time, it's a long-term project/goal
The driver core device link support missed 5.4 by just a bit, so it's
been sitting and baking for many months now. It's from Saravana Kannan
to help resolve the problems that DT-based systems have at boot time
with dependancy graphs and kernel modules. Turns out that no one has
actually tried to build a generic arm64 kernel with loads of modules
and have it "just work" for a variety of platforms (like a distro
kernel). The big problem turned out to be a lack of dependency
information between different areas of DT entries, and the work here
resolves that problem and now allows devices to boot properly, and
quicker than a monolith kernel.
All of these patches have been in linux-next for a long time with no
reported issues"
* tag 'driver-core-5.5-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (68 commits)
tracing: Remove unnecessary DEBUG_FS dependency
of: property: Add device link support for interrupt-parent, dmas and -gpio(s)
debugfs: Fix !DEBUG_FS debugfs_create_automount
of: property: Add device link support for "iommu-map"
of: property: Fix the semantics of of_is_ancestor_of()
i2c: of: Populate fwnode in of_i2c_get_board_info()
drivers: base: Fix Kconfig indentation
firmware_loader: Fix labels with comma for builtin firmware
driver core: Allow device link operations inside sync_state()
driver core: platform: Declare ret variable only once
cpu-topology: declare parse_acpi_topology in <linux/arch_topology.h>
crypto: hisilicon: no need to check return value of debugfs_create functions
driver core: platform: use the correct callback type for bus_find_device
firmware_class: make firmware caching configurable
driver core: Clarify documentation for fwnode_operations.add_links()
mailbox: tegra: Fix superfluous IRQ error message
net: caif: Fix debugfs on 64-bit platforms
mac80211: Use debugfs_create_xul() helper
media: c8sectpfe: no need to check return value of debugfs_create functions
of: property: Add device link support for iommus, mboxes and io-channels
...
There are a couple of statements that are indented by an extra
space, clean this up by remove the extraneous spaces.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
"dw_mci.state" is an enum, which is compatible with u32, so there is no
need to cast its address, preventing further compiler checks.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20191025094130.26033-8-geert+renesas@glider.be
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
"dw_mci.pending_events" and "dw_mci.completed_events" are "unsigned
long", i.e. 32-bit or 64-bit, depending on the platform. Hence casting
their addresses to "u32 *", and calling debugfs_create_x32() breaks
operation on 64-bit platforms.
Fix this by using the new debugfs_create_xul() helper instead.
Fixes: f95f3850f7 ("mmc: dw_mmc: Add Synopsys DesignWare mmc host driver.")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20191025094130.26033-7-geert+renesas@glider.be
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In cases when SDIO IRQs have been enabled, runtime suspend is prevented by
the driver. However, this still means dw_mci_runtime_suspend|resume() gets
called during system suspend/resume, via pm_runtime_force_suspend|resume().
This means during system suspend/resume, the register context of the dw_mmc
device most likely loses its register context, even in cases when SDIO IRQs
have been enabled.
To re-enable the SDIO IRQs during system resume, the dw_mmc driver
currently relies on the mmc core to re-enable the SDIO IRQs when it resumes
the SDIO card, but this isn't the recommended solution. Instead, it's
better to deal with this locally in the dw_mmc driver, so let's do that.
Tested-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
In commit 46d179525a ("mmc: dw_mmc: Wait for data transfer after
response errors.") we fixed a tuning-induced hang that I saw when
stress testing tuning on certain SD cards. I won't re-hash that whole
commit, but the summary is that as a normal part of tuning you need to
deal with transfer errors and there were cases where these transfer
errors was putting my system into a bad state causing all future
transfers to fail. That commit fixed handling of the transfer errors
for me.
In downstream Chrome OS my fix landed and had the same behavior for
all SD/MMC commands. However, it looks like when the commit landed
upstream we limited it to only SD tuning commands. Presumably this
was to try to get around problems that Alim Akhtar reported on exynos
[1].
Unfortunately while stress testing reboots (and suspend/resume) on
some rk3288-based Chromebooks I found the same problem on the eMMC on
some of my Chromebooks (the ones with Hynix eMMC). Since the eMMC
tuning command is different (MMC_SEND_TUNING_BLOCK_HS200
vs. MMC_SEND_TUNING_BLOCK) we were basically getting back into the
same situation.
I'm hoping that whatever problems exynos was having in the past are
somehow magically fixed now and we can make the behavior the same for
all commands.
[1] https://lkml.kernel.org/r/CAGOxZ53WfNbaMe0_AM0qBqU47kAfgmPBVZC8K8Y-_J3mDMqW4A@mail.gmail.com
Fixes: 46d179525a ("mmc: dw_mmc: Wait for data transfer after response errors.")
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Alim Akhtar <alim.akhtar@gmail.com>
Cc: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
When calling debugfs functions, there is no need to ever check the
return value. The function can work or not, but the code logic should
never do something different based on this
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license as published by
the free software foundation either version 2 of the license or at
your option any later version
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-or-later
has been chosen to replace the boilerplate/reference in 3029 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190527070032.746973796@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Card write threshold control is supposed to be set since controller
version 2.80a for data write in HS400 mode and data read in
HS200/HS400/SDR104 mode. However the current code returns without
configuring it in the case of data writing in HS400 mode.
Meanwhile the patch fixes that the current code goes to
'disable' when doing data reading in HS400 mode.
Fixes: 7e4bf1bc95 ("mmc: dw_mmc: add the card write threshold for HS400 mode")
Signed-off-by: Qing Xia <xiaqing17@hisilicon.com>
Cc: stable@vger.kernel.org # v4.8+
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
It was found that in IDMAC mode after soft-reset driver switches
to PIO mode.
That's what happens in case of DTO timeout overflow calculation failure:
1. soft-reset is called
2. driver restarts dma
3. descriptors states are checked, one of descriptor is owned by the IDMAC.
4. driver can't use DMA and then switches to PIO mode.
Failure was already fixed in:
https://www.spinics.net/lists/linux-mmc/msg48125.html.
Behaviour while soft-reset is not something we except or
even want to happen. So we switch from dw_mci_idmac_reset
to dw_mci_idmac_init, so descriptors are cleaned before starting dma.
And while at it explicitly zero des0 which otherwise might
contain garbage as being allocated by dmam_alloc_coherent().
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Ulf Hansson <ulf.hansson@linaro.org>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Cc: Shawn Lin <shawn.lin@rock-chips.com>
Cc: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Cc: linux-snps-arc@lists.infradead.org
Cc: <stable@vger.kernel.org> # 4.4+
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
'num-slots' property had already deprecated.
Remove the 'nom-slots' property that is kept to maintain the compatibility.
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
'clock-freq-min-max' property had already deprecated.
Remove the 'clock-freq-min-max' property that is kept to maintain
the compatibility.
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Clang reports a compile warning:
drivers/mmc/host/dw_mmc.c:2124:5: warning: Value stored to 'prev_state'
is never read
By checking the code, prev_state and state assignment for
STATE_SENDING_CMD is indeed never used after jumping to unlock tag.
So remove it.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Use the newly added macro to simply to the code.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The commit 9d9491a7da ("mmc: dw_mmc: Fix the DTO timeout calculation")
and commit 4c2357f57d ("mmc: dw_mmc: Fix the CTO timeout calculation")
made changes, which cause multiply overflow for 32-bit systems. The broken
timeout calculations leads to unexpected ETIMEDOUT errors and causes
stacktrace splat (such as below) during normal data exchange with SD-card.
| Running : 4M-check-reassembly-tcp-cmykw2-rotatew2.out -v0 -w1
| - Info: Finished target initialization.
| mmcblk0: error -110 transferring data, sector 320544, nr 2048, cmd
| response 0x900, card status 0x0
DIV_ROUND_UP_ULL helps to escape usage of __udivdi3() from libgcc and so
code gets compiled on all 32-bit platforms as opposed to usage of
DIV_ROUND_UP when we may only compile stuff on a very few arches.
Lets cast this multiply to u64 type to prevent the overflow.
Fixes: 9d9491a7da ("mmc: dw_mmc: Fix the DTO timeout calculation")
Fixes: 4c2357f57d ("mmc: dw_mmc: Fix the CTO timeout calculation")
Tested-by: Vineet Gupta <Vineet.Gupta1@synopsys.com>
Reported-by: Vineet Gupta <Vineet.Gupta1@synopsys.com> # ARC STAR 9001306872 HSDK, sdio: board crashes when copying big files
Signed-off-by: Evgeniy Didin <Evgeniy.Didin@synopsys.com>
Cc: <stable@vger.kernel.org> # 4.14
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
Reviewed-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
cat /sys/kernel/debug/mmc0/regs will hang up the system since
it's in runtime suspended state, so the genpd and biu_clk is
off. This patch fixes this problem by calling pm_runtime_get_sync
to wake it up before reading the registers.
Fixes: e9ed8835e9 ("mmc: dw_mmc: add runtime PM callback")
Cc: <stable@vger.kernel.org>
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Add num_caps field for dw_mci_drv_data to validate the controller
id from DT alias and non-DT ways.
Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Fixes: 800d78bfcc ("mmc: dw_mmc: add support for implementation specific callbacks")
Cc: <stable@vger.kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Factor out dw_mci_init_slot_caps to consolidate parsing
all differents types of capabilities from host contrllers.
No functional change intended.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Fixes: 800d78bfcc ("mmc: dw_mmc: add support for implementation specific callbacks")
Cc: <stable@vger.kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
In preparation for unconditionally passing the struct timer_list pointer to
all timer callbacks, switch to using the new timer_setup() and from_timer()
to pass the timer pointer explicitly.
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Ulf Hansson <ulf.hansson@linaro.org>
Cc: linux-mmc@vger.kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The recent CTO timer introduced in commit 03de19212e ("mmc: dw_mmc:
introduce timer for broken command transfer over scheme") was causing
observable problems due to race conditions. Previous patches have
fixed those race conditions.
It can be observed that these same race conditions ought to be
theoretically possible with the DTO timer too though they are
massively less likely to happen because the data timeout is always set
to 0xffffff right now. That means even at a 200 MHz card clock we
were arming the DTO timer for 94 ms:
>>> (0xffffff * 1000. / 200000000) + 10
93.886075
We always also were setting the DTO timer _after_ starting the
transfer, unlike how the old code was seting the CTO timer.
In any case, even though the DTO timer is much less likely to have
races, it still makes sense to add code to handle it _just in case_.
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Just like the CTO timeout calculation introduced recently, the DTO
timeout calculation was incorrect. It used "bus_hz" but, as far as I
can tell, it's supposed to use the card clock. Let's account for the
div value, which is documented as 2x the value stored in the register,
or 1 if the register is 0.
NOTE: This was likely not terribly important until commit 16a34574c6
("mmc: dw_mmc: remove the quirks flags") landed because "DIV" is
documented on Rockchip SoCs (the ones that used to define the quirk)
to always be 0 or 1. ...and, in fact, it's documented to only be 1
with EMMC in 8-bit DDR52 mode. Thus before the quirk was applied to
everyone it was mostly OK to ignore the DIV value.
I haven't personally observed any problems that are fixed by this
patch but I also haven't tested this anywhere with a DIV other an 0.
AKA: this problem was found simply by code inspection and I have no
failing test cases that are fixed by it. Presumably this could fix
real bugs for someone out there, though.
Fixes: 16a34574c6 ("mmc: dw_mmc: remove the quirks flags")
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
This attempts to instill a bit of paranoia to the code dealing with
the CTO timer. It's believed that this will make the CTO timer more
robust in the case that we're having very long interrupt latencies.
Note that I originally thought that perhaps this patch was being
overly paranoid and wasn't really needed, but then while I was running
mmc_test on an rk3399 board I saw one instance of the message:
dwmmc_rockchip fe320000.dwmmc: Unexpected interrupt latency
I had debug prints in the CTO timer code and I found that it was
running CMD 13 at the time.
...so even though this patch seems like it might be overly paranoid,
maybe it really isn't?
Presumably the bad interrupt latency experienced was due to the fact
that I had serial console enabled as serial console is typically where
I place blame when I see absurdly large interrupt latencies. In this
particular case there was an (unrelated) printout to the serial
console just before I saw the "Unexpected interrupt latency" printout.
...and actually, I managed to even reproduce the problems by running
"iw mlan0 scan > /dev/null" while mmc_test was running. That not only
does a bunch of PCIe traffic but it also (on my system) outputs some
SELinux log spam.
Fixes: 03de19212e ("mmc: dw_mmc: introduce timer for broken command transfer over scheme")
Tested-by: Emil Renner Berthing <kernel@esmil.dk>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
In the commit 03de19212e ("mmc: dw_mmc: introduce timer for broken
command transfer over scheme") we tried to calculate the expected
hardware command timeout value. Unfortunately that calculation isn't
quite correct in all cases. It used "bus_hz" but, as far as I can
tell, it's supposed to use the card clock. Let's account for the div
value, which is documented as 2x the value stored in the register, or
1 if the register is 0.
NOTE: It's not expected that this will actually fix anything important
since the 10 ms margin added by the function will pretty much dwarf
any calculations. The card clock should be 100 kHz at minimum and:
1000 ms/s * (255 * 2) / 100000 Hz.
Gives us 5.1 ms.
...so really the point of this patch is just to make the code more
"correct" in case anyone ever tries to remove the 10 ms buffer.
Fixes: 03de19212e ("mmc: dw_mmc: introduce timer for broken command transfer over scheme")
Tested-by: Emil Renner Berthing <kernel@esmil.dk>
Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
