platform/kernel/linux-starfive.git
4 years agommc: sdhci-of-esdhc: exit HS400 properly before setting any speed mode
Yangbo Lu [Fri, 22 May 2020 03:12:56 +0000 (11:12 +0800)]
mmc: sdhci-of-esdhc: exit HS400 properly before setting any speed mode

The eSDHC HS400 timing requires many specific registers setting,
unlike other speed modes which need to set only host controller 2
register. When driver needs to downgrade HS400 mode to other speed
mode, the controller have to exit HS400 timing properly first.
This patch is to support the procedure of HS400 exiting at the
beginning of esdhc_set_uhs_signaling.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20200522031256.856-1-yangbo.lu@nxp.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-msm: dump vendor specific registers during error
Sarthak Garg [Fri, 22 May 2020 09:32:30 +0000 (15:02 +0530)]
mmc: sdhci-msm: dump vendor specific registers during error

Implement dump_vendor_registers host operation to print the
vendor specific registers in addition to standard SDHC
register during error conditions.

Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
Signed-off-by: Sarthak Garg <sartgarg@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1590139950-7288-9-git-send-email-sartgarg@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-msm: Introduce new ops to dump vendor specific registers
Sarthak Garg [Fri, 22 May 2020 09:32:29 +0000 (15:02 +0530)]
mmc: sdhci-msm: Introduce new ops to dump vendor specific registers

Introduce new sdhci ops to dump vendor specific registers in the
sdhci_dumpregs during error.

Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
Signed-off-by: Sarthak Garg <sartgarg@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1590139950-7288-8-git-send-email-sartgarg@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-msm: Read and use DLL Config property from device tree file
Sarthak Garg [Fri, 22 May 2020 09:32:28 +0000 (15:02 +0530)]
mmc: sdhci-msm: Read and use DLL Config property from device tree file

Certain platforms require different settings in the
SDCC_HC_REG_DLL_CONFIG register. This setting can change from platform
to platform. So the driver should check whether a particular platform
require a different setting by reading the DT file and use it.

Also use msm_cm_dll_set_freq only when DLL not supplied.

Signed-off-by: Bao D. Nguyen <nguyenb@codeaurora.org>
Signed-off-by: Sarthak Garg <sartgarg@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1590139950-7288-7-git-send-email-sartgarg@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-msm: Update DDR_CONFIG as per device tree file
Sarthak Garg [Fri, 22 May 2020 09:32:27 +0000 (15:02 +0530)]
mmc: sdhci-msm: Update DDR_CONFIG as per device tree file

Certain platforms require different settings in the
SDCC_HC_REG_DDR_CONFIG register. This setting can change from platform
to platform. So the driver should check whether a particular platform
require a different setting by reading the device tree file and use it.

Signed-off-by: Bao D. Nguyen <nguyenb@codeaurora.org>
Signed-off-by: Sarthak Garg <sartgarg@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1590139950-7288-6-git-send-email-sartgarg@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-msm: Update dll_config_3 as per HSR
Sarthak Garg [Fri, 22 May 2020 09:32:26 +0000 (15:02 +0530)]
mmc: sdhci-msm: Update dll_config_3 as per HSR

Update dll_config_3 as per the host clock frequency as specified in the
DLL Hardware Reference Guide.

Signed-off-by: Sarthak Garg <sartgarg@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1590139950-7288-5-git-send-email-sartgarg@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: host: sdhci-msm: Configure dll-user-control in dll init sequence
Veerabhadrarao Badiganti [Fri, 22 May 2020 09:32:25 +0000 (15:02 +0530)]
mmc: host: sdhci-msm: Configure dll-user-control in dll init sequence

With SDCC v5.1.0, additional setting needed for enabling DLL output.
The dll-user-control register need to be configured during dll
initialization for getting proper dll output.

Without this configuration, we don't get the DLL lock status properly.
Also update the DLL register settings according to the SDCC Hardware
Programming Guide.

Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
Signed-off-by: Sarthak Garg <sartgarg@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1590139950-7288-4-git-send-email-sartgarg@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agodt-bindings: mmc: Add information for DLL register properties
Sarthak Garg [Fri, 22 May 2020 09:32:24 +0000 (15:02 +0530)]
dt-bindings: mmc: Add information for DLL register properties

Add information regarding DLL register properties for getting board
specific configurations. These DLL register settings may vary from
board to board.

Signed-off-by: Sarthak Garg <sartgarg@codeaurora.org>
Link: https://lore.kernel.org/r/1590139950-7288-3-git-send-email-sartgarg@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agodt-bindings: mmc: Add new compatible string for sm8250 target
Sarthak Garg [Fri, 22 May 2020 09:32:23 +0000 (15:02 +0530)]
dt-bindings: mmc: Add new compatible string for sm8250 target

Add new compatible string for sm8250 target.

Signed-off-by: Sarthak Garg <sartgarg@codeaurora.org>
Link: https://lore.kernel.org/r/1590139950-7288-2-git-send-email-sartgarg@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agosdhci: tegra: Avoid reading autocal timeout values when not applicable
Sowjanya Komatineni [Wed, 20 May 2020 20:08:57 +0000 (13:08 -0700)]
sdhci: tegra: Avoid reading autocal timeout values when not applicable

When auto calibration timeouts, calibration is disabled and fail-safe
drive strength values are programmed based on the signal voltage.

Different fail-safe drive strength values based on voltage are
applicable only for SoCs supporting 3V3 and 1V8 pad controls.

So, this patch avoids reading these properties from the device tree
for SoCs not using pad controls and the warning of missing properties
will not show up on these SoC platforms.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Tested-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1590005337-1087-1-git-send-email-skomatineni@nvidia.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: renesas_sdhi: remove manual clk handling
Wolfram Sang [Tue, 19 May 2020 16:42:51 +0000 (18:42 +0200)]
mmc: renesas_sdhi: remove manual clk handling

The SDHI driver en-/disabled its main clock on its own, e.g. during
probe() and remove(). Now, we leave all handling to RPM.

clk_summary before:
sd0                   1        1        0    12480000          0     0  50000
   sdif0              2        2        0    12480000          0     0  50000

clk_summary after:
sd0                   1        1        0    12480000          0     0  50000
   sdif0              1        1        0    12480000          0     0  50000

Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20200519164251.5430-1-wsa+renesas@sang-engineering.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
4 years agommc: tmio: Make sure the PM domain is 'started' while probing
Ulf Hansson [Tue, 19 May 2020 15:24:45 +0000 (17:24 +0200)]
mmc: tmio: Make sure the PM domain is 'started' while probing

If the tmio device is attached to a genpd (PM domain), that genpd may have
->start|stop() callback assigned to it. To make sure the device is
accessible during ->probe(), genpd's ->start() callback must be invoked,
which is currently managed by tmio_mmc_host_probe(). However, it's likely
that may be too late for some cases, as registers may be read and written
way before that point.

To fix the behaviour, let's move the call to dev_pm_domain_start() from
tmio_mmc_host_probe() into those clients that needs it. From discussions at
linux-mmc mailing list, it turned out that it should be sufficient to do
this for the SDHI renesas variants, hence the call is move to
renesas_sdhi_probe().

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Tested-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20200519152445.6922-1-ulf.hansson@linaro.org
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
4 years agommc: tmio: Further fixup runtime PM management at remove
Ulf Hansson [Tue, 19 May 2020 15:24:34 +0000 (17:24 +0200)]
mmc: tmio: Further fixup runtime PM management at remove

Before calling tmio_mmc_host_probe(), the caller is required to enable
clocks for its device, as to make it accessible when reading/writing
registers during probe.

Therefore, the responsibility to disable these clocks, in the error path of
->probe() and during ->remove(), is better managed outside
tmio_mmc_host_remove(). As a matter of fact, callers of
tmio_mmc_host_remove() already expects this to be the behaviour.

However, there's a problem with tmio_mmc_host_remove() when the Kconfig
option, CONFIG_PM, is set. More precisely, tmio_mmc_host_remove() may then
disable the clock via runtime PM, which leads to clock enable/disable
imbalance problems, when the caller of tmio_mmc_host_remove() also tries to
disable the same clocks.

To solve the problem, let's make sure tmio_mmc_host_remove() leaves the
device with clocks enabled, but also make sure to disable the IRQs, as we
normally do at ->runtime_suspend().

Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Tested-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20200519152434.6867-1-ulf.hansson@linaro.org
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
4 years agoMAINTAINERS: add myself to maintain M5441X mmc host driver
Angelo Dureghello [Mon, 18 May 2020 19:17:42 +0000 (21:17 +0200)]
MAINTAINERS: add myself to maintain M5441X mmc host driver

Since actively working on Freescale ColdFire M5441X, adding
myself as a maintainer of this driver.

Signed-off-by: Angelo Dureghello <angelo.dureghello@timesys.com>
Link: https://lore.kernel.org/r/20200518191742.1251440-4-angelo.dureghello@timesys.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: host: add Coldfire esdhc support
Angelo Dureghello [Mon, 18 May 2020 19:17:41 +0000 (21:17 +0200)]
mmc: host: add Coldfire esdhc support

This driver has been developed as a separate module starting
from the similar sdhci-esdhc-imx.c.

Reasons for a separate sdchi-esdhc-mcf driver:

- m68K architecture does not support devicetrees, so modifying
sdhci-of-esdhc.c that is devicetree-related adding platform data
seems not appropriate,
- clock-related part, has to be implemented specifically for
mcf5441x family (see esdhc_mcf_pltfm_set_clock()),
- this is a big endian cpu accessing a big endian controller,
but about sdma, this controller does not support hw swap, which
needs to be handled with specific code,
- some other minor differences but mainly to avoid risks on
tweaking inside largely used imx driver. Adding just a small
size ColdFire-specific driver, with benefits in a further less
risky maintenance.

Signed-off-by: Angelo Dureghello <angelo.dureghello@timesys.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20200518191742.1251440-3-angelo.dureghello@timesys.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: add quirks for be to le byte swapping
Angelo Dureghello [Mon, 18 May 2020 19:17:40 +0000 (21:17 +0200)]
mmc: sdhci: add quirks for be to le byte swapping

Some controller as the ColdFire eshdc may require an endianness
byte swap, because DMA read endianness is not configurable.

Facilitate using the bounce buffer for this by adding
->copy_to_bounce_buffer().

Signed-off-by: Angelo Dureghello <angelo.dureghello@timesys.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20200518191742.1251440-2-angelo.dureghello@timesys.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agom68k: mcf5441x: add support for esdhc mmc controller
Angelo Dureghello [Mon, 18 May 2020 19:17:39 +0000 (21:17 +0200)]
m68k: mcf5441x: add support for esdhc mmc controller

Add support for sdhci-edshc mmc controller.

Signed-off-by: Angelo Dureghello <angelo.dureghello@timesys.com>
Acked-by: Greg Ungerer <gerg@linux-m68k.org>
Link: https://lore.kernel.org/r/20200518191742.1251440-1-angelo.dureghello@timesys.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agodt-bindings: mmc: renesas,sdhi: Document r8a7742 support
Lad Prabhakar [Fri, 15 May 2020 15:08:44 +0000 (16:08 +0100)]
dt-bindings: mmc: renesas,sdhi: Document r8a7742 support

Document SDHI controller for RZ/G1H (R8A7742) SoC, which is compatible
with R-Car Gen2 SoC family.

Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Marian-Cristian Rotariu <marian-cristian.rotariu.rb@bp.renesas.com>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/1589555337-5498-5-git-send-email-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: meson-mx-sdhc: Don't use literal 0 to initialize structs
Martin Blumenstingl [Sun, 17 May 2020 22:29:07 +0000 (00:29 +0200)]
mmc: meson-mx-sdhc: Don't use literal 0 to initialize structs

Kbuild test robot reports the following warning in lines 56 and 87 of
drivers/mmc/host/meson-mx-sdhc-clkc.c:
  Using plain integer as NULL pointer

Drop the integer value from the struct initialization to fix that
warning. This will still ensure that the compiler will zero out the
struct so it's in a well-defined state.

Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://lore.kernel.org/r/20200517222907.1277787-2-martin.blumenstingl@googlemail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: meson-mx-sdhc: Fix manual RX FIFO flushing
Martin Blumenstingl [Sun, 17 May 2020 22:29:06 +0000 (00:29 +0200)]
mmc: meson-mx-sdhc: Fix manual RX FIFO flushing

For Meson8 and Meson8b SoCs the vendor driver follows the following
pattern:
- for eMMC and SD cards in .set_pdma it sets:
  pdma->rxfifo_manual_flush = 1;
- for SDIO cards in .set_pdma it sets:
  pdma->rxfifo_manual_flush = 0;
- before syncing the DMA read buffer is sets:
  pdma->rxfifo_manual_flush |= 0x02;

Set the second bit of MESON_SDHC_PDMA_RXFIFO_MANUAL_FLUSH without
clearing the first bit before syncing the DMA read buffer. This fixes a
problem where Meson8 and Meson8b SoCs would read random garbage from SD
cards. It is not clear why it worked for eMMC cards. This manifested in
the following errors when plugging in an SD card:
  unrecognised SCR structure version <random number>

Cc: Thomas Graichen <thomas.graichen@gmail.com>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://lore.kernel.org/r/20200517222907.1277787-1-martin.blumenstingl@googlemail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-esdhc-imx: Add HS400 support for i.MX6SLL
Haibo Chen [Fri, 15 May 2020 07:28:23 +0000 (15:28 +0800)]
mmc: sdhci-esdhc-imx: Add HS400 support for i.MX6SLL

i.MX6SLL support MMC up to V5.0, which means support HS400 mode.

Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1589527703-19108-1-git-send-email-haibo.chen@nxp.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: meson-mx-sdhc: Fix building with CONFIG_MMC_MESON_MX_SDHC=m
Martin Blumenstingl [Mon, 18 May 2020 06:08:11 +0000 (08:08 +0200)]
mmc: meson-mx-sdhc: Fix building with CONFIG_MMC_MESON_MX_SDHC=m

For an x86_64 allmodconfig build Stephen reports that building
meson-mx-sdhc-clkc.o warns that MODULE_LICENSE is missing and when
linking meson_mx_sdhc_register_clkc cannot be found.

Compile the MMC controller driver together with the build-in clock
controller driver into one module rather than using two separate
modules to fix these issues.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://lore.kernel.org/r/20200518060811.1499962-1-martin.blumenstingl@googlemail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-of-dwcmshc: add suspend/resume support
Jisheng Zhang [Fri, 15 May 2020 06:19:26 +0000 (14:19 +0800)]
mmc: sdhci-of-dwcmshc: add suspend/resume support

Add dwcmshc specific system-level suspend and resume support.

Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Link: https://lore.kernel.org/r/20200515141926.52e088fe@xhacker.debian
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-of-dwcmshc: implement specific set_uhs_signaling
Jisheng Zhang [Wed, 13 May 2020 10:26:02 +0000 (18:26 +0800)]
mmc: sdhci-of-dwcmshc: implement specific set_uhs_signaling

We need a different set_uhs_signaling implementation for
MMC_TIMING_MMC_HS and MMC_TIMING_MMC_HS400.

Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Link: https://lore.kernel.org/r/20200513182602.3636a551@xhacker.debian
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: host: meson-mx-sdhc: new driver for the Amlogic Meson SDHC host
Martin Blumenstingl [Tue, 12 May 2020 20:41:47 +0000 (22:41 +0200)]
mmc: host: meson-mx-sdhc: new driver for the Amlogic Meson SDHC host

The SDHC MMC host controller on Amlogic SoCs provides an eMMC and MMC
card interface with 1/4/8-bit bus width.
It supports eMMC spec 4.4x/4.5x including HS200 (up to 100MHz clock).

The public S805 datasheet [0] contains a short documentation about the
registers. Unfortunately it does not describe how to use the registers
to make the hardware work. Thus this driver is based on reading (and
understanding) the Amlogic 3.10 GPL kernel code.

Some hardware details are not easy to see. Jianxin Pan was kind enough
to answer my questions:
The hardware has built-in busy timeout support. The maximum timeout is
30 seconds. This is only documented in Amlogic's internal
documentation.

The controller only works with very specific clock configurations. The
details are not part of the public datasheet. In my own words the
supported configurations are:
- 399.812kHz: clkin =  850MHz div = 2126 sd_rx_phase = 63
- 1MHz: clkin =  850MHz div = 850  sd_rx_phase = 55
- 5.986MHz: clkin =  850MHz div = 142  sd_rx_phase = 24
- 25MHz: clkin =  850MHz div = 34   sd_rx_phase = 15
- 47.222MHz: clkin =  850MHz div = 18   sd_rx_phase = 11/15 (SDR50/HS)
- 53.125MHz: clkin =  850MHz div = 16   sd_rx_phase = (tuning)
- 70.833MHz: clkin =  850MHz div = 12   sd_rx_phase = (tuning)
- 85MHz: clkin =  850MHz div = 10   sd_rx_phase = (tuning)
- 94.44MHz: clkin =  850MHz div = 9    sd_rx_phase = (tuning)
- 106.25MHz: clkin =  850MHz div = 8    sd_rx_phase = (tuning)
- 127.5MHz:     clkin = 1275MHz div = 10   sd_rx_phase = (tuning)
- 141.667MHz:   clkin =  850MHz div = 6    sd_rx_phase = (tuning)
- 159.375MHz: clkin = 1275MHz div = 8    sd_rx_phase = (tuning)
- 212.5MHz: clkin = 1275MHz div = 6    sd_rx_phase = (tuning)
- (sd_tx_phase is always 1, 94.44MHz is not listed in the datasheet
   but this is what the 3.10 BSP kernel on Odroid-C1 actually uses)

NOTE: CMD23 support is disabled for now because it results in command
timeouts and thus decreases read performance.

Tested-by: Wei Wang <lnykww@gmail.com>
Tested-by: Xin Yin <yinxin_1989@aliyun.com>
Reviewed-by: Xin Yin <yinxin_1989@aliyun.com>
Tested-by: Anand Moon <linux.amoon@gmail.com>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Reviewed-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20200512204147.504087-3-martin.blumenstingl@googlemail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agodt-bindings: mmc: Document the Amlogic Meson SDHC MMC host controller
Martin Blumenstingl [Tue, 12 May 2020 20:41:46 +0000 (22:41 +0200)]
dt-bindings: mmc: Document the Amlogic Meson SDHC MMC host controller

This documents the devicetree bindings for the SDHC MMC host controller
found in Meson6, Meson8, Meson8b and Meson8m2 SoCs. It can use a
bus-width of 1/4/8-bit and it supports eMMC spec 4.4x/4.5x including
HS200 mode (up to 100MHz clock). It embeds an internal clock controller
which outputs four clocks (mod_clk, sd_clk, tx_clk and rx_clk) and is
fed by four external input clocks (clkin[0-3]). "pclk" is the module
register clock, it has to be enabled to access the registers.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20200512204147.504087-2-martin.blumenstingl@googlemail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: fix compilation of user API
Jérôme Pouiller [Mon, 11 May 2020 16:19:02 +0000 (18:19 +0200)]
mmc: fix compilation of user API

The definitions of MMC_IOC_CMD  and of MMC_IOC_MULTI_CMD rely on
MMC_BLOCK_MAJOR:

    #define MMC_IOC_CMD       _IOWR(MMC_BLOCK_MAJOR, 0, struct mmc_ioc_cmd)
    #define MMC_IOC_MULTI_CMD _IOWR(MMC_BLOCK_MAJOR, 1, struct mmc_ioc_multi_cmd)

However, MMC_BLOCK_MAJOR is defined in linux/major.h and
linux/mmc/ioctl.h did not include it.

Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20200511161902.191405-1-Jerome.Pouiller@silabs.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: use FIELD_GET/PREP for current capabilities bit masks
Masahiro Yamada [Mon, 11 May 2020 06:28:27 +0000 (15:28 +0900)]
mmc: sdhci: use FIELD_GET/PREP for current capabilities bit masks

Use FIELD_GET and FIELD_PREP to get access to the register fields. Delete
the shift macros and use GENMASK() for the touched macros.

Note that, this has the side-effect of changing the constants to 64-bit on
64-bit platforms.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Link: https://lore.kernel.org/r/20200511062828.1791484-1-yamada.masahiro@socionext.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: uniphier-sd: call devm_request_irq() after tmio_mmc_host_probe()
Masahiro Yamada [Mon, 11 May 2020 06:21:58 +0000 (15:21 +0900)]
mmc: uniphier-sd: call devm_request_irq() after tmio_mmc_host_probe()

Currently, tmio_mmc_irq() handler is registered before the host is
fully initialized by tmio_mmc_host_probe(). I did not previously notice
this problem.

The boot ROM of a new Socionext SoC unmasks interrupts (CTL_IRQ_MASK)
somehow. The handler is invoked before tmio_mmc_host_probe(), then
emits noisy call trace.

Move devm_request_irq() below tmio_mmc_host_probe().

Fixes: 3fd784f745dd ("mmc: uniphier-sd: add UniPhier SD/eMMC controller driver")
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20200511062158.1790924-1-yamada.masahiro@socionext.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-pci-gli: Add Genesys Logic GL9763E support
Ben Chuang [Fri, 8 May 2020 06:41:54 +0000 (14:41 +0800)]
mmc: sdhci-pci-gli: Add Genesys Logic GL9763E support

GL9763E supports High Speed SDR, High Speed DDR, HS200, HS400, Enhanced
Strobe in HS400 mode, 1/4/8 bits data bus and 3.3/1.8V.

Signed-off-by: Ben Chuang <ben.chuang@genesyslogic.com.tw>
Link: https://lore.kernel.org/r/20200508064154.13473-1-benchuanggli@gmail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: host: Drop redundant MMC_CAP_ERASE
Ulf Hansson [Fri, 8 May 2020 11:29:02 +0000 (13:29 +0200)]
mmc: host: Drop redundant MMC_CAP_ERASE

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>
4 years agommc: core: Enable erase/discard/trim support for all mmc hosts
Ulf Hansson [Fri, 8 May 2020 11:28:53 +0000 (13:28 +0200)]
mmc: core: Enable erase/discard/trim support for all mmc hosts

Step by step, mmc host drivers and the mmc core have been improved in
regards to support erase/discard/trim operations. We have now reached a
point when it no longer seems reasonable to use an opt-in approach to
enable the functionality. Therefore, let's switch to make the operations
always supported.

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200508112853.23525-1-ulf.hansson@linaro.org
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
4 years agommc: sdricoh_cs: Respect the cmd->busy_timeout from the mmc core
Ulf Hansson [Fri, 8 May 2020 09:52:28 +0000 (11:52 +0200)]
mmc: sdricoh_cs: Respect the cmd->busy_timeout from the mmc core

Using a fixed 1s polling timeout for all commands is a bit problematic.

For some commands it means waiting longer than needed for the polling to be
aborted, which may not a big issue, but still. For other commands, like for
an erase (CMD38) that uses a R1B response, may require longer timeouts than
1s. In these cases, we may end up treating the command as it failed, while
it just needed some more time to complete successfully.

Fix the problem by respecting the cmd->busy_timeout, which is provided by
the mmc core.

Note that, even if the sdricoh_cs driver may currently not support HW busy
detection on DAT0, some comments in the code refer to that the HW may
support it. Therefore, it seems better to be proactive in this case.

Cc: Sascha Sommer <saschasommer@freenet.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200508095228.14230-1-ulf.hansson@linaro.org
4 years agommc: sdricoh_cs: Throttle polling rate for commands
Ulf Hansson [Fri, 8 May 2020 09:52:18 +0000 (11:52 +0200)]
mmc: sdricoh_cs: Throttle polling rate for commands

Rather than to poll in a busy-loop, let's convert into using
read_poll_timeout() and insert a small delay between each polling attempts.
In particular, this avoids hogging the CPU.

Additionally, to convert to read_poll_timeout() we also need to switch from
using a specific number of polling attempts, into a specific timeout in us
instead. The previous 100000 attempts, is translated into a total timeout
of total 1s, as that seemed like reasonable value to pick.

Cc: Sascha Sommer <saschasommer@freenet.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200508095218.14177-1-ulf.hansson@linaro.org
4 years agommc: sdricoh_cs: Throttle polling rate for data transfers
Ulf Hansson [Fri, 8 May 2020 09:52:10 +0000 (11:52 +0200)]
mmc: sdricoh_cs: Throttle polling rate for data transfers

Rather than to poll in a busy-loop, let's convert into using
read_poll_timeout() and insert a small delay between each polling attempts.
In particular, this avoids hogging the CPU.

Additionally, to convert to read_poll_timeout() we also need to switch from
using a specific number of polling attempts, into a specific timeout in us
instead. The previous 100000 attempts, is translated into a total timeout
of total 1s, as that seemed like reasonable value to pick.

Cc: Sascha Sommer <saschasommer@freenet.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200508095210.14123-1-ulf.hansson@linaro.org
4 years agommc: cb710: Inform the mmc core about the maximum busy timeout
Ulf Hansson [Tue, 14 Apr 2020 16:14:00 +0000 (18:14 +0200)]
mmc: cb710: Inform the mmc core about the maximum busy timeout

Some commands uses R1B responses, which means the card may assert the DAT0
line to signal busy for a period of time, after it has received the
command. The mmc core normally specifies the busy period for the command in
the cmd->busy_timeout. Ideally the driver should respect it, but that
requires quite some update of the code, so let's defer that to someone with
the HW at hand.

Instead, let's inform the mmc core about the maximum supported busy timeout
in ->max_busy_timeout during ->probe(). This value corresponds to the fixed
~2s timeout of the polling loop, implemented in cb710_wait_for_event(). In
this way, we let the mmc core validate the needed timeout, which may lead
to that it converts from a R1B into a R1 response and then use CMD13 to
poll for busy completion.

In other words, this change enables support for commands with longer busy
periods than 2s, like erase (CMD38) for example.

Cc: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-7-ulf.hansson@linaro.org
4 years agomemstick: Replace zero-length array with flexible-array
Gustavo A. R. Silva [Thu, 7 May 2020 19:22:18 +0000 (14:22 -0500)]
memstick: Replace zero-length array with flexible-array

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20200507192218.GA16315@embeddedor
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: meson-mx-sdio: trigger a soft reset after a timeout or CRC error
Martin Blumenstingl [Sun, 3 May 2020 22:28:05 +0000 (00:28 +0200)]
mmc: meson-mx-sdio: trigger a soft reset after a timeout or CRC error

The vendor driver (from the 3.10 kernel) triggers a soft reset every
time before starting a new command. While this fixes a problem where
SDIO cards are not detected at all (because all commands simply
timed out) this hurts SD card read performance a bit (in my tests
between 10% to 20%).

Trigger a soft reset after we got a CRC error or if the previous command
timed out (just like the vendor driver from the same 3.10 kernel for the
newer SDHC controller IP does). This fixes detection of SDIO cards and
doesn't hurt SD card read performance at the same time.

With this patch the initialization of an RTL8723BS SDIO card looks like
this:
  req done (CMD52): -110: 00000000 00000000 00000000 00000000
  clock 400000Hz busmode 2 powermode 2 cs 1 Vdd 21 width 1 timing 0
  starting CMD0 arg 00000000 flags 000000c0
  req done (CMD0): 0: 00000000 00000000 00000000 00000000
  clock 400000Hz busmode 2 powermode 2 cs 0 Vdd 21 width 1 timing 0
  starting CMD8 arg 000001aa flags 000002f5
  req done (CMD8): -110: 00000000 00000000 00000000 00000000
  starting CMD5 arg 00000000 flags 000002e1
  req done (CMD5): 0: 90ff0000 00000000 00000000 00000000
  starting CMD5 arg 00200000 flags 000002e1
  req done (CMD5): 0: 90ff0000 00000000 00000000 00000000
  starting CMD3 arg 00000000 flags 00000075
  req done (CMD3): 0: 00010000 00000000 00000000 00000000
  starting CMD7 arg 00010000 flags 00000015
  req done (CMD7): 0: 00001e00 00000000 00000000 00000000
  starting CMD52 arg 00000000 flags 00000195
  req done (CMD52): 0: 00001032 00000000 00000000 00000000
  [... more CMD52 omitted ...]
  clock 400000Hz busmode 2 powermode 2 cs 0 Vdd 21 width 1 timing 2
  clock 50000000Hz busmode 2 powermode 2 cs 0 Vdd 21 width 1 timing 2
  starting CMD52 arg 00000e00 flags 00000195
  req done (CMD52): 0: 00001000 00000000 00000000 00000000
  starting CMD52 arg 80000e02 flags 00000195
  req done (CMD52): 0: 00001002 00000000 00000000 00000000
  clock 50000000Hz busmode 2 powermode 2 cs 0 Vdd 21 width 4 timing 2
  starting CMD52 arg 00020000 flags 00000195
  req done (CMD52): 0: 00001007 00000000 00000000 00000000
  [... more CMD52 omitted ...]
  new high speed SDIO card at address 0001

Fixes: ed80a13bb4c4c9 ("mmc: meson-mx-sdio: Add a driver for the Amlogic Meson8 and Meson8b SoCs")
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://lore.kernel.org/r/20200503222805.2668941-1-martin.blumenstingl@googlemail.com
Tested-by: Tobias Baumann <017623705678@o2online.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agodt-bindings: mmc: renesas,mmcif: Document r8a7742 DT bindings
Lad Prabhakar [Wed, 6 May 2020 19:51:27 +0000 (20:51 +0100)]
dt-bindings: mmc: renesas,mmcif: Document r8a7742 DT bindings

Add support for r8a7742 SoC. Renesas RZ/G1H (R8A7742) MMCIF is identical
to the R-Car Gen2 family.

Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Marian-Cristian Rotariu <marian-cristian.rotariu.rb@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/1588794695-27852-2-git-send-email-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-esdhc: update contact email
Wolfram Sang [Sat, 2 May 2020 14:28:25 +0000 (16:28 +0200)]
mmc: sdhci-esdhc: update contact email

The 'pengutronix' address is defunct for years. Use the proper contact
address.

Signed-off-by: Wolfram Sang <wsa@kernel.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20200502142840.19418-1-wsa@kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: core: expose info about enhanced rpmb support
Krishna Konda [Fri, 1 May 2020 13:53:01 +0000 (19:23 +0530)]
mmc: core: expose info about enhanced rpmb support

Following eMMC JEDEC JESD84-B51 standard, an enhanced form of
rpmb is supported. What this enhanced mode supports is in addition
to be able to write one rpmb or two rpmb frames at a time,
32 frames can be written at a time.

Expose this information present in ext csd field so that the
user space application that wants to make use of this can do so.

Signed-off-by: Krishna Konda <kkonda@codeaurora.org>
Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
Link: https://lore.kernel.org/r/1588341189-4371-1-git-send-email-vbadigan@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdio: Align the initialization commands in retry path for UHS-I
Ulf Hansson [Thu, 30 Apr 2020 09:16:40 +0000 (11:16 +0200)]
mmc: sdio: Align the initialization commands in retry path for UHS-I

According to the comment in mmc_sdio_reinit_card(), some SDIO cards may
require a "[CMD5,5,3,7] init sequence", which isn't always obeyed in
mmc_sdio_init_card(). Especially, when we end up retrying the UHS-I
specific initialization, there is a missing CMD5.

Let's update the code to make the behaviour consistent and let's also take
the opportunity to clean up the code a bit, to avoid open coding.

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200430091640.455-5-ulf.hansson@linaro.org
4 years agommc: sdio: Re-use negotiated OCR mask when re-sending CMD8
Ulf Hansson [Thu, 30 Apr 2020 09:16:39 +0000 (11:16 +0200)]
mmc: sdio: Re-use negotiated OCR mask when re-sending CMD8

While initializing an SDIO card in mmc_sdio_init_card(), we may need to
retry the UHS-I specific initialization, in case the first attempt fails.

This leads to resending a CMD8, but also to restart from scratch with the
so called OCR mask negotiations. This is unnecessary as we already have a
negotiated OCR mask, so let's use that instead. In this way, the behaviour
also becomes more consistent with other similar paths.

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200430091640.455-4-ulf.hansson@linaro.org
4 years agommc: sdio: Fix several potential memory leaks in mmc_sdio_init_card()
Ulf Hansson [Thu, 30 Apr 2020 09:16:38 +0000 (11:16 +0200)]
mmc: sdio: Fix several potential memory leaks in mmc_sdio_init_card()

Over the years, the code in mmc_sdio_init_card() has grown to become quite
messy. Unfortunate this has also lead to that several paths are leaking
memory in form of an allocated struct mmc_card, which includes additional
data, such as initialized struct device for example.

Unfortunate, it's a too complex task find each offending commit. Therefore,
this change fixes all memory leaks at once.

Cc: <stable@vger.kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200430091640.455-3-ulf.hansson@linaro.org
4 years agommc: sdio: Fix potential NULL pointer error in mmc_sdio_init_card()
Ulf Hansson [Thu, 30 Apr 2020 09:16:37 +0000 (11:16 +0200)]
mmc: sdio: Fix potential NULL pointer error in mmc_sdio_init_card()

During some scenarios mmc_sdio_init_card() runs a retry path for the UHS-I
specific initialization, which leads to removal of the previously allocated
card. A new card is then re-allocated while retrying.

However, in one of the corresponding error paths we may end up to remove an
already removed card, which likely leads to a NULL pointer exception. So,
let's fix this.

Fixes: 5fc3d80ef496 ("mmc: sdio: don't use rocr to check if the card could support UHS mode")
Cc: <stable@vger.kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200430091640.455-2-ulf.hansson@linaro.org
4 years agommc: sdhci-msm: Fix error handling for dev_pm_opp_of_add_table()
Rajendra Nayak [Tue, 28 Apr 2020 13:32:57 +0000 (19:02 +0530)]
mmc: sdhci-msm: Fix error handling for dev_pm_opp_of_add_table()

Even though specifying OPP's in device tree is optional, ignoring all errors
reported by dev_pm_opp_of_add_table() means we can't distinguish between a
missing OPP table and a wrong/buggy OPP table. While missing OPP table
(dev_pm_opp_of_add_table() returns a -ENODEV in such case) can be ignored,
a wrong/buggy OPP table in device tree should make the driver error out.

while we fix that, lets also fix the variable names for opp/opp_table to
avoid confusion and name them opp_table/has_opp_table instead.

Suggested-by: Matthias Kaehlcke <matthias@chromium.org>
Signed-off-by: Rajendra Nayak <rnayak@codeaurora.org>
Link: https://lore.kernel.org/r/1588080785-6812-10-git-send-email-rnayak@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: renesas_sdhi: Avoid bad TAP in HS400
Takeshi Saito [Thu, 23 Apr 2020 13:04:32 +0000 (15:04 +0200)]
mmc: renesas_sdhi: Avoid bad TAP in HS400

With R-Car Gen3, CRC error occue at the following TAPs.

H3, M3W 1.3, M3N... TAP=2,3,6,7
M3W 3.0 ... TAP=1,3,5,7

(Note: for 4tap SoCs, the numbers get divided by 2)

Do not use these TAPs in HS400, and also don't use auto correction but
manual correction.

We check for bad taps in two places:

1) After tuning HS400: Then, we select a neighbouring TAP. One of them
   must be good, because there are never three bad taps in a row.
   Retuning won't help because we just finished tuning.

2) After a manual correction request: Here, we can't switch to the
   requested TAP. But we can retune (if the HS200 tuning was good)
   because the environment might have changed since the last tuning.
   If not, we stay on the same TAP.

Signed-off-by: Takeshi Saito <takeshi.saito.xv@renesas.com>
[wsa: refactored to match upstream driver, reworded commit msg]
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://lore.kernel.org/r/20200423130432.9990-3-wsa+renesas@sang-engineering.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: renesas_sdhi: handle M3-W ES1.2 and 1.3 revisions
Wolfram Sang [Thu, 23 Apr 2020 13:04:31 +0000 (15:04 +0200)]
mmc: renesas_sdhi: handle M3-W ES1.2 and 1.3 revisions

For ES1.2, add a comment explaining the situation. For ES1.3 (and
later, although unlikely), add a new entry defining it as 4tap.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://lore.kernel.org/r/20200423130432.9990-2-wsa+renesas@sang-engineering.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: mmc_spi: Respect the cmd->busy_timeout from the mmc core
Ulf Hansson [Tue, 14 Apr 2020 16:14:12 +0000 (18:14 +0200)]
mmc: mmc_spi: Respect the cmd->busy_timeout from the mmc core

Using a fixed 3s polling timeout for all commands with R1B responses is a
bit problematic.

For some commands it means waiting longer than needed for the polling to be
aborted, which may not a big issue, but still. For other commands, like for
an erase (CMD38), may require longer timeouts than 3s. In these cases, we
may end up treating the command as it failed, while it just needed some
more time to complete successfully.

Fix the problem by respecting the cmd->busy_timeout, which is provided by
the mmc core.

Cc: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-19-ulf.hansson@linaro.org
4 years agommc: mmc_spi: Add/rename defines for timeouts
Ulf Hansson [Tue, 14 Apr 2020 16:14:11 +0000 (18:14 +0200)]
mmc: mmc_spi: Add/rename defines for timeouts

Clarify the use of r1b_timeout, by renaming it to MMC_SPI_R1B_TIMEOUT_MS
and by dropping the corresponding confusing comment about it.

Additionally, let's also add a new define, MMC_SPI_INIT_TIMEOUT_MS and use
it during the initialization. Even if these two defines are given the same
value, the split makes it easier to understand them.

Cc: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-18-ulf.hansson@linaro.org
4 years agommc: via-sdmmc: Respect the cmd->busy_timeout from the mmc core
Ulf Hansson [Tue, 14 Apr 2020 16:14:10 +0000 (18:14 +0200)]
mmc: via-sdmmc: Respect the cmd->busy_timeout from the mmc core

Using a fixed 1s timeout for all commands (and data transfers) is a bit
problematic.

For some commands it means waiting longer than needed for the timer to
expire, which may not a big issue, but still. For other commands, like for
an erase (CMD38) that uses a R1B response, may require longer timeouts than
1s. In these cases, we may end up treating the command as it failed, while
it just needed some more time to complete successfully.

Fix the problem by respecting the cmd->busy_timeout, which is provided by
the mmc core.

Cc: Bruce Chang <brucechang@via.com.tw>
Cc: Harald Welte <HaraldWelte@viatech.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-17-ulf.hansson@linaro.org
4 years agommc: tifm_sd: Inform the mmc core about the maximum busy timeout
Ulf Hansson [Tue, 14 Apr 2020 16:14:09 +0000 (18:14 +0200)]
mmc: tifm_sd: Inform the mmc core about the maximum busy timeout

Some commands uses R1B responses, which means the card may assert the DAT0
line to signal busy for a period of time, after it has received the
command. The mmc core normally specifies the busy period for the command in
the cmd->busy_timeout. Ideally the driver should respect it, but that
requires quite some update of the code, so let's defer that to someone with
the HW at hand.

Instead, let's inform the mmc core about the maximum supported busy timeout
in ->max_busy_timeout during ->probe(). This value corresponds to the fixed
1s timeout used by tifm_sd. In this way, we let the mmc core validate the
needed timeout, which may lead to that it converts from a R1B into a R1
response and then use CMD13 to poll for busy completion.

In other words, this change enables support for commands with longer busy
periods than 1s, like erase (CMD38) for example.

Cc: Alex Dubov <oakad@yahoo.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-16-ulf.hansson@linaro.org
4 years agommc: sdricoh_cs: Drop redundant in-parameter to sdricoh_query_status()
Ulf Hansson [Tue, 14 Apr 2020 16:14:05 +0000 (18:14 +0200)]
mmc: sdricoh_cs: Drop redundant in-parameter to sdricoh_query_status()

The in-parameter timeout is always set to TRANSFER_TIMEOUT by the callers
of sdricoh_query_status(), hence let's drop it.

Cc: Sascha Sommer <saschasommer@freenet.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-12-ulf.hansson@linaro.org
4 years agommc: sdricoh_cs: Move MMC_APP_CMD handling to sdricoh_mmc_cmd()
Ulf Hansson [Tue, 14 Apr 2020 16:14:04 +0000 (18:14 +0200)]
mmc: sdricoh_cs: Move MMC_APP_CMD handling to sdricoh_mmc_cmd()

Move MMC_APP_CMD specific handling to be managed by sdricoh_mmc_cmd(), as
this makes the code a bit cleaner.

Cc: Sascha Sommer <saschasommer@freenet.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-11-ulf.hansson@linaro.org
4 years agommc: sdricoh_cs: Use MMC_APP_CMD rather than a hardcoded number
Ulf Hansson [Tue, 14 Apr 2020 16:14:03 +0000 (18:14 +0200)]
mmc: sdricoh_cs: Use MMC_APP_CMD rather than a hardcoded number

Cc: Sascha Sommer <saschasommer@freenet.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-10-ulf.hansson@linaro.org
4 years agommc: sdricoh_cs: Drop unused defines
Ulf Hansson [Tue, 14 Apr 2020 16:14:02 +0000 (18:14 +0200)]
mmc: sdricoh_cs: Drop unused defines

Cc: Sascha Sommer <saschasommer@freenet.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-9-ulf.hansson@linaro.org
4 years agommc: owl-mmc: Respect the cmd->busy_timeout from the mmc core
Ulf Hansson [Tue, 14 Apr 2020 16:14:01 +0000 (18:14 +0200)]
mmc: owl-mmc: Respect the cmd->busy_timeout from the mmc core

For commands that doesn't involve to prepare a data transfer, owl-mmc is
using a fixed 30s response timeout. This is a bit problematic.

For some commands it means waiting longer than needed for the completion to
expire, which may not a big issue, but still. For other commands, like for
an erase (CMD38) that uses a R1B response, may require longer timeouts than
30s. In these cases, we may end up treating the command as it failed, while
it just needed some more time to complete successfully.

Fix the problem by respecting the cmd->busy_timeout, which is provided by
the mmc core.

Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-8-ulf.hansson@linaro.org
4 years agommc: jz4740: Inform the mmc core about the maximum busy timeout
Ulf Hansson [Tue, 14 Apr 2020 16:13:58 +0000 (18:13 +0200)]
mmc: jz4740: Inform the mmc core about the maximum busy timeout

Some commands uses R1B responses, which means the card may assert the DAT0
line to signal busy for a period of time, after it has received the
command. The mmc core normally specifies the busy period for the command in
the cmd->busy_timeout. Ideally the driver should respect it, but that
requires quite some update of the code, so let's defer that to someone with
the HW at hand.

Instead, let's inform the mmc core about the maximum supported busy timeout
in ->max_busy_timeout during ->probe(). This value corresponds to the fixed
5s timeout used by jz4740. In this way, we let the mmc core validate the
needed timeout, which may lead to that it converts from a R1B into a R1
response and then use CMD13 to poll for busy completion.

In other words, this change enables support for commands with longer busy
periods than 5s, like erase (CMD38) for example.

Cc: Paul Cercueil <paul@crapouillou.net>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200414161413.3036-5-ulf.hansson@linaro.org
4 years agommc: sdhci-pci-o2micro: Make some symbols static
Zou Wei [Thu, 23 Apr 2020 06:43:19 +0000 (14:43 +0800)]
mmc: sdhci-pci-o2micro: Make some symbols static

Fix the following sparse warning:

drivers/mmc/host/sdhci-pci-o2micro.c:497:6: warning: symbol
'sdhci_pci_o2_set_clock' was not declared. Should it be static?
drivers/mmc/host/sdhci-pci-o2micro.c:512:5: warning: symbol
'sdhci_pci_o2_probe_slot' was not declared. Should it be static?
drivers/mmc/host/sdhci-pci-o2micro.c:581:5: warning: symbol
'sdhci_pci_o2_probe' was not declared. Should it be static?
drivers/mmc/host/sdhci-pci-o2micro.c:786:5: warning: symbol
'sdhci_pci_o2_resume' was not declared. Should it be static?

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zou Wei <zou_wei@huawei.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1587624199-96926-1-git-send-email-zou_wei@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: core: Use DEFINE_DEBUGFS_ATTRIBUTE instead of DEFINE_SIMPLE_ATTRIBUTE
Zou Wei [Thu, 23 Apr 2020 09:15:19 +0000 (17:15 +0800)]
mmc: core: Use DEFINE_DEBUGFS_ATTRIBUTE instead of DEFINE_SIMPLE_ATTRIBUTE

Fixes coccicheck warning:

drivers/mmc/core/debugfs.c:222:0-23: WARNING: mmc_clock_fops should be
defined with DEFINE_DEBUGFS_ATTRIBUTE

Use DEFINE_DEBUGFS_ATTRIBUTE rather than DEFINE_SIMPLE_ATTRIBUTE
for debugfs files

Fixes: 703aae3d09a4b ("mmc: add a file to debugfs for changing host clock at runtime")
Fixes: a04c50aaa916f ("mmc: core: no need to check return value of debugfs_create functions")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zou Wei <zou_wei@huawei.com>
Link: https://lore.kernel.org/r/1587633319-19835-1-git-send-email-zou_wei@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-of-at91: make MMC_SDHCI_OF_AT91 depend on HAVE_CLK
Masahiro Yamada [Wed, 22 Apr 2020 15:34:01 +0000 (00:34 +0900)]
mmc: sdhci-of-at91: make MMC_SDHCI_OF_AT91 depend on HAVE_CLK

If sdhci-of-at91.c is compiled without CONFIG_HAVE_CLK, the line

  caps1 |= FIELD_PREP(SDHCI_CLOCK_MUL_MASK, clk_mul);

... emits "FIELD_PREP: value too large for the field" warning.

The compiler seems to decide clk_mul is constant (unsigned int)-1,
because clk_get_rate() returns 0 when CONFIG_HAVE_CLK is disabled.

Add HAVE_CLK to the depenency since this driver does not work without
the clock APIs anyway.

Reported-by: Randy Dunlap <rdunlap@infradead.org>
Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Link: https://lore.kernel.org/r/20200422153401.7913-1-yamada.masahiro@socionext.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
4 years agommc: renesas_sdhi: shorten types after refactorization
Wolfram Sang [Mon, 20 Apr 2020 17:02:30 +0000 (19:02 +0200)]
mmc: renesas_sdhi: shorten types after refactorization

After TAP refactorization, we can use 'unsigned int' for two more
variables because all the calculations work on this type now.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Link: https://lore.kernel.org/r/20200420170230.9091-1-wsa+renesas@sang-engineering.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: mmci_sdmmc: fix power on issue due to pwr_reg initialization
Ludovic Barre [Mon, 20 Apr 2020 16:18:31 +0000 (18:18 +0200)]
mmc: mmci_sdmmc: fix power on issue due to pwr_reg initialization

This patch fix a power-on issue, and avoid to retry the power sequence.

In power off sequence: sdmmc must set pwr_reg in "power-cycle" state
(value 0x2), to prevent the card from being supplied through the signal
lines (all the lines are driven low).

In power on sequence: when the power is stable, sdmmc must set pwr_reg
in "power-off" state (value 0x0) to drive all signal to high before to
set "power-on".

To avoid writing the same value to the power register several times, this
register is cached by the pwr_reg variable. At probe pwr_reg is initialized
to 0 by kzalloc of mmc_alloc_host.

Like pwr_reg value is 0 at probing, the power on sequence fail because
the "power-off" state is not writes (value 0x0) and the lines
remain drive to low.

This patch initializes "pwr_reg" variable with power register value.
This it done in sdmmc variant init to not disturb default mmci behavior.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Link: https://lore.kernel.org/r/20200420161831.5043-1-ludovic.barre@st.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-msm: Use OPP API to set clk/perf state
Rajendra Nayak [Fri, 17 Apr 2020 14:04:31 +0000 (19:34 +0530)]
mmc: sdhci-msm: Use OPP API to set clk/perf state

On some qualcomm SoCs we need to vote on a performance state of a power
domain depending on the clock rates. Hence move to using OPP api to set
the clock rate and performance state specified in the OPP table.
On platforms without an OPP table, dev_pm_opp_set_rate() is eqvivalent to
clk_set_rate()

Signed-off-by: Rajendra Nayak <rnayak@codeaurora.org>
Link: https://lore.kernel.org/r/1587132279-27659-10-git-send-email-rnayak@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agostaging: greybus: sdio: Respect the cmd->busy_timeout from the mmc core
Ulf Hansson [Tue, 14 Apr 2020 16:14:13 +0000 (18:14 +0200)]
staging: greybus: sdio: Respect the cmd->busy_timeout from the mmc core

Using a fixed 1s timeout for all commands is a bit problematic.

For some commands it means waiting longer than needed for the timeout to
expire, which may not a big issue, but still. For other commands, like for
an erase (CMD38) that uses a R1B response, may require longer timeouts than
1s. In these cases, we may end up treating the command as it failed, while
it just needed some more time to complete successfully.

Fix the problem by respecting the cmd->busy_timeout, which is provided by
the mmc core.

Cc: Rui Miguel Silva <rmfrfs@gmail.com>
Cc: Johan Hovold <johan@kernel.org>
Cc: Alex Elder <elder@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: greybus-dev@lists.linaro.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rui Miguel Silva <rmfrfs@gmail.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/r/20200414161413.3036-20-ulf.hansson@linaro.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: usdhi6rol0: Inform the mmc core about the maximum busy timeout
Ulf Hansson [Tue, 14 Apr 2020 16:13:59 +0000 (18:13 +0200)]
mmc: usdhi6rol0: Inform the mmc core about the maximum busy timeout

Some commands uses R1B responses, which means the card may assert the DAT0
line to signal busy for a period of time, after it has received the
command. The mmc core normally specifies the busy period for the command in
the cmd->busy_timeout. Ideally the driver should respect it, but that
requires quite some update of the code, so let's defer that to someone with
the HW at hand.

Instead, let's inform the mmc core about the maximum supported busy timeout
in ->max_busy_timeout during ->probe(). This value corresponds to the fixed
4s timeout used by usdhi6rol0. In this way, we let the mmc core validate
the needed timeout, which may lead to that it converts from a R1B into a R1
response and then use CMD13 to poll for busy completion.

In other words, this change enables support for commands with longer busy
periods than 4s, like erase (CMD38) for example.

Cc: Jesper Nilsson <jesper.nilsson@axis.com>
Cc: Lars Persson <lars.persson@axis.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Jesper Nilsson <jesper.nilsson@axis.com>
Link: https://lore.kernel.org/r/20200414161413.3036-6-ulf.hansson@linaro.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: atmel-mci: Respect the cmd->busy_timeout from the mmc core
Ulf Hansson [Tue, 14 Apr 2020 16:13:57 +0000 (18:13 +0200)]
mmc: atmel-mci: Respect the cmd->busy_timeout from the mmc core

Using a fixed 2s timeout for all commands is a bit problematic.

For some commands it means waiting longer than needed for the timer to
expire, which may not a big issue, but still. For other commands, like for
an erase (CMD38) that uses a R1B response, may require longer timeouts than
2s. In these cases, we may end up treating the command as it failed, while
it just needed some more time to complete successfully.

Fix the problem by respecting the cmd->busy_timeout, which is provided by
the mmc core.

Cc: Ludovic Desroches <ludovic.desroches@microchip.com>
Cc: Nicolas Ferre <nicolas.ferre@microchip.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200414161413.3036-4-ulf.hansson@linaro.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: atmel-mci: Set the timer per command rather than per request
Ulf Hansson [Tue, 14 Apr 2020 16:13:56 +0000 (18:13 +0200)]
mmc: atmel-mci: Set the timer per command rather than per request

Setting the timer on a per request basis, is rather limiting as the timer
really depends on what commands that is to be sent as part of the request.

Therefore improve the behaviour by programming the timer per command basis
instead.

Cc: Ludovic Desroches <ludovic.desroches@microchip.com>
Cc: Nicolas Ferre <nicolas.ferre@microchip.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200414161413.3036-3-ulf.hansson@linaro.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: atmel-mci: Keep timer enabled when queuing a next request
Ulf Hansson [Tue, 14 Apr 2020 16:13:55 +0000 (18:13 +0200)]
mmc: atmel-mci: Keep timer enabled when queuing a next request

When atmci_request_end() is about to finish a request for one slot, there
is a possibility that there is new request queued for another slot. If this
turns out to be the case, the new request is started and the timer is
re-programmed for it.

Although, a few lines below in atmci_request_end(), this timer becomes
deleted, likely corresponding to the other recently completed request. This
looks wrong, so let's fix it.

Cc: Ludovic Desroches <ludovic.desroches@microchip.com>
Cc: Nicolas Ferre <nicolas.ferre@microchip.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
Link: https://lore.kernel.org/r/20200414161413.3036-2-ulf.hansson@linaro.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-msm: Enable ADMA length mismatch error interrupt
Veerabhadrarao Badiganti [Mon, 20 Apr 2020 06:20:25 +0000 (11:50 +0530)]
mmc: sdhci-msm: Enable ADMA length mismatch error interrupt

ADMA_ERR_SIZE_EN bit of VENDOR_SPECIFIC_FUNC register controls
ADMA length mismatch error interrupt. Enable it by default.

And update all bit shift defines with BIT macro.

Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1587363626-20413-4-git-send-email-vbadigan@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-msm: Set SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12 quirk
Veerabhadrarao Badiganti [Mon, 20 Apr 2020 06:20:24 +0000 (11:50 +0530)]
mmc: sdhci-msm: Set SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12 quirk

sdhci-msm can support auto cmd12.
So enable SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12 quirk.

Signed-off-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1587363626-20413-3-git-send-email-vbadigan@codeaurora.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: add spin lock for sdhci_set_default_irqs in sdhci_init
Haibo Chen [Wed, 15 Apr 2020 09:00:55 +0000 (17:00 +0800)]
mmc: sdhci: add spin lock for sdhci_set_default_irqs in sdhci_init

When use one SDIO wifi which enable the runtime PM feature on i.MX6SX,
we meet system hang. This hang happened during the usdhc runtime resume,
in sdhci_init(), when call the sdhci_set_default_irqs. One interrupt
(SDHCI_INT_CARD_INT) triggered just after the host->ier update and before
the write of register SDHCI_SIGNAL_ENABLE. So in sdhci_irq, it will skip
the call of sdio_signal_irq() because current host->ier do not set the
SDHCI_INT_CARD_INT. So this SDIO wifi interrupt always keep triggered,
let the system stuck in irq handle, can't response any other thread.

This patch add spin lock for the sdhci_set_default_irqs to fix this issue.

Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1586941255-9237-1-git-send-email-haibo.chen@nxp.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: host: sdhci-sprd: Implement the request_atomic() API
Baolin Wang [Mon, 13 Apr 2020 02:46:05 +0000 (10:46 +0800)]
mmc: host: sdhci-sprd: Implement the request_atomic() API

Implement the request_atomic() API for nonremovable cards, that means
we can submit next request in the irq hard handler context to reduce
latency.

Moreover factor out the AUTO CMD23 checking into a separate function
to reduce duplicate code.

Signed-off-by: Baolin Wang <baolin.wang7@gmail.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/60142fe6c6c1dbba2696e775564ae2166786f0bc.1586744073.git.baolin.wang7@gmail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: host: sdhci: Implement the request_atomic() API
Baolin Wang [Mon, 13 Apr 2020 02:46:04 +0000 (10:46 +0800)]
mmc: host: sdhci: Implement the request_atomic() API

Implement the request_atomic() ops for the sdhci driver to process
one request in the atomic context if the card is nonremovable.

Moreover, we should return BUSY flag if controller has not released
the inhibit bits to allow HSQ trying to send request again in non-atomic
context.

Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Baolin Wang <baolin.wang7@gmail.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/9ed34afa9fb42e0c234065cac5401d7826942b55.1586744073.git.baolin.wang7@gmail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: host: Introduce the request_atomic() for the host
Baolin Wang [Mon, 13 Apr 2020 02:46:03 +0000 (10:46 +0800)]
mmc: host: Introduce the request_atomic() for the host

The SD host controller can process one request in the atomic context if
the card is nonremovable, which means we can submit next request in the
irq hard handler when using the MMC host software queue to reduce the
latency. Thus this patch adds a new API request_atomic() for the host
controller, as well as adding support for host software queue to submit
a request by the new request_atomic() API.

Moreover there is an unusual case that the card is busy when trying to
send a command, and we can not polling the card status in interrupt
context by using request_atomic() to dispatch requests. Thus we should
queue a work to try again in the non-atomic context in case the host
releases the busy signal later.

Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Baolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/a344e27e506cb2329073cbd5cf65e15cc3cbeba9.1586744073.git.baolin.wang7@gmail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: Reduce maximum time under spinlock in sdhci_send_command()
Adrian Hunter [Sun, 12 Apr 2020 09:03:49 +0000 (12:03 +0300)]
mmc: sdhci: Reduce maximum time under spinlock in sdhci_send_command()

Spending time under spinlock increases IRQ latencies and also
response times because preemption is disabled.

sdhci_send_command() waits up to 10 ms under spinlock for inhibit bits
to clear. In general inhibit bits will not be set, but there may be
corner cases, especially in the face of errors, where waiting helps.
There might also be dysfunctional hardware that needs the waiting. So
retain the legacy behaviour but do not wait for inhibit bits while under
spinlock. Instead adjust the logic to enable waiting while not under
spinlock. That is mostly straight forward, but in the interrupt handler
it requires deferring an "inhibited" command to the IRQ thread where
sleeping is allowed.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Tested-by: Baolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/20200412090349.1607-6-adrian.hunter@intel.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: Tidy sdhci_request() a bit
Adrian Hunter [Sun, 12 Apr 2020 09:03:48 +0000 (12:03 +0300)]
mmc: sdhci: Tidy sdhci_request() a bit

In preparation for further changes, tidy sdhci_request() a bit.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Tested-by: Baolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/20200412090349.1607-5-adrian.hunter@intel.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: Remove unneeded forward declaration of sdhci_finish_data()
Adrian Hunter [Sun, 12 Apr 2020 09:03:47 +0000 (12:03 +0300)]
mmc: sdhci: Remove unneeded forward declaration of sdhci_finish_data()

sdhci_finish_data() is defined before it is referenced, so forward
declaration is not necessary.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Tested-by: Baolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/20200412090349.1607-4-adrian.hunter@intel.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: Stop exporting sdhci_send_command()
Adrian Hunter [Sun, 12 Apr 2020 09:03:46 +0000 (12:03 +0300)]
mmc: sdhci: Stop exporting sdhci_send_command()

sdhci_send_command() has not been used outside of sdhci.c for many
years. Stop exporting it.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Tested-by: Baolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/20200412090349.1607-3-adrian.hunter@intel.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: Add helpers for the auto-CMD23 flag
Adrian Hunter [Sun, 12 Apr 2020 09:03:45 +0000 (12:03 +0300)]
mmc: sdhci: Add helpers for the auto-CMD23 flag

Add 2 helper functions to make the use of the auto-CMD23 flag more
readable.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Tested-by: Baolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/20200412090349.1607-2-adrian.hunter@intel.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sd: use HIGH_SPEED_BUS_SPEED in mmc_sd_switch_hs()
Jisheng Zhang [Fri, 10 Apr 2020 06:57:12 +0000 (14:57 +0800)]
mmc: sd: use HIGH_SPEED_BUS_SPEED in mmc_sd_switch_hs()

Use the well defined HIGH_SPEED_BUS_SPEED macro in mmc_sd_switch_hs()
to make code more readable.

Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Link: https://lore.kernel.org/r/20200410145643.630b0731@xhacker.debian
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: mmci: Switch to mmc_regulator_set_vqmmc()
Marek Vasut [Thu, 16 Apr 2020 16:36:49 +0000 (18:36 +0200)]
mmc: mmci: Switch to mmc_regulator_set_vqmmc()

Instead of reimplementing the logic in mmc_regulator_set_vqmmc(), use the
mmc code function directly.

This also allows us to fix a related issue on STM32MP1, when a voltage
switch of 1.8V is done for the eMMC, but the current level is already set
to 1.8V. More precisely, in this scenario the call to the
->post_sig_volt_switch() hangs, indefinitely waiting for the voltage switch
to complete. Fix this problem by checking if mmc_regulator_set_vqmmc()
returned 1 and then skip invoking the callback.

Signed-off-by: Marek Vasut <marex@denx.de>
Link: https://lore.kernel.org/r/20200416163649.336967-3-marex@denx.de
[Ulf: Updated the commit message]
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: core: Return 1 from mmc_regulator_set_vqmmc() if switch skipped
Marek Vasut [Thu, 16 Apr 2020 16:36:48 +0000 (18:36 +0200)]
mmc: core: Return 1 from mmc_regulator_set_vqmmc() if switch skipped

Adjust mmc_regulator_set_vqmmc() to return 1 if the voltage switch was
skipped because the regulator voltage was already correct. This allows
drivers to detect such condition and possibly skip various voltage
switching extras.

Signed-off-by: Marek Vasut <marex@denx.de>
Link: https://lore.kernel.org/r/20200416163649.336967-2-marex@denx.de
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: host: Prepare host drivers for mmc_regulator_set_vqmmc() returning > 0
Marek Vasut [Thu, 16 Apr 2020 16:36:47 +0000 (18:36 +0200)]
mmc: host: Prepare host drivers for mmc_regulator_set_vqmmc() returning > 0

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>
4 years agommc: sdhci-of-arasan: Remove uninitialized ret variables
Nathan Chancellor [Thu, 16 Apr 2020 18:24:02 +0000 (11:24 -0700)]
mmc: sdhci-of-arasan: Remove uninitialized ret variables

Clang warns:

drivers/mmc/host/sdhci-of-arasan.c:784:9: warning: variable 'ret' is
uninitialized when used here [-Wuninitialized]
        return ret;
               ^~~
drivers/mmc/host/sdhci-of-arasan.c:738:9: note: initialize the variable
'ret' to silence this warning
        int ret;
               ^
                = 0
drivers/mmc/host/sdhci-of-arasan.c:860:9: warning: variable 'ret' is
uninitialized when used here [-Wuninitialized]
        return ret;
               ^~~
drivers/mmc/host/sdhci-of-arasan.c:810:9: note: initialize the variable
'ret' to silence this warning
        int ret;
               ^
                = 0
2 warnings generated.

This looks like a copy paste error. Neither function has handling that
needs ret so just remove it and return 0 directly.

Link: https://github.com/ClangBuiltLinux/linux/issues/996
Reported-by: kernelci.org bot <bot@kernelci.org>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Link: https://lore.kernel.org/r/20200416182402.16858-1-natechancellor@gmail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: renesas_sdhi: simplify summary output
Wolfram Sang [Wed, 8 Apr 2020 14:22:52 +0000 (16:22 +0200)]
mmc: renesas_sdhi: simplify summary output

No need to call platform_get_resource twice when we still have the
pointer from before. Also, use '%pa' for a resource_size_t pointer.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Link: https://lore.kernel.org/r/20200408142252.21958-1-wsa+renesas@sang-engineering.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
4 years agommc: renesas_sdhi: improve TAP selection if all TAPs are good
Wolfram Sang [Wed, 8 Apr 2020 09:46:38 +0000 (11:46 +0200)]
mmc: renesas_sdhi: improve TAP selection if all TAPs are good

When tuning HS400, if all TAPS are good, we can utilize the SMPCMP
register to select the optimal TAP. For that, we populate a second
bitmap with SMPCMP results and query it in case the regular bitmap is
full (= all good).

Signed-off-by: Masaharu Hayakawa <masaharu.hayakawa.ry@renesas.com>
Signed-off-by: Takeshi Saito <takeshi.saito.xv@renesas.com>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Tested-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://lore.kernel.org/r/20200408094638.10375-4-wsa+renesas@sang-engineering.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: renesas_sdhi: clarify handling of selecting TAPs
Wolfram Sang [Wed, 8 Apr 2020 09:46:37 +0000 (11:46 +0200)]
mmc: renesas_sdhi: clarify handling of selecting TAPs

The comment and the define about how TAPs are selected were confusing to
me because the good TAP was only valid if it was bigger than a *_MAX_*
value. Rename the define and adapt the comment to what really happens.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://lore.kernel.org/r/20200408094638.10375-3-wsa+renesas@sang-engineering.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: renesas_sdhi: refactor calculation of best TAP
Wolfram Sang [Wed, 8 Apr 2020 09:46:36 +0000 (11:46 +0200)]
mmc: renesas_sdhi: refactor calculation of best TAP

To select the best TAP, we need to find the longest stream of set bits
in a bit field. There is now a helper function for bitmaps which
iterates over all region of set bits. Using it makes the code much more
concise and easier to understand. Double so, because we need to handle
two bitmaps in the near future. Remove a superfluous comment while here.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://lore.kernel.org/r/20200408094638.10375-2-wsa+renesas@sang-engineering.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: use FIELD_GET/PREP for capabilities bit masks
Masahiro Yamada [Wed, 8 Apr 2020 07:21:05 +0000 (16:21 +0900)]
mmc: sdhci: use FIELD_GET/PREP for capabilities bit masks

Use FIELD_GET and FIELD_PREP to get access to the register fields. Delete
the shift macros and use GENMASK() for the touched macros.

Note that, this has the side-effect of changing the constants to 64-bit on
64-bit platforms.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20200408072105.422-2-yamada.masahiro@socionext.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci: move SDHCI_CAPABILITIES_1 to a more suitable place
Masahiro Yamada [Wed, 8 Apr 2020 07:21:04 +0000 (16:21 +0900)]
mmc: sdhci: move SDHCI_CAPABILITIES_1 to a more suitable place

In the SDHCI specification, the Capabilities Register (Offset 0x40h)
is the 64-bit width register, but in Linux, it is represented as two
registers, SDHCI_CAPABILITIES and SDHCI_CAPABILITIES_1 so that drivers
can use 32-bit register accessors.

The upper 32-bit field is associated with SDHCI_CAPABILITIES_1.

Move the definition of SDHCI_CAPABILITIES_1 to the correct place.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20200408072105.422-1-yamada.masahiro@socionext.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-of-arasan: Fix kernel-doc warnings
Manish Narani [Mon, 6 Apr 2020 17:43:35 +0000 (23:13 +0530)]
mmc: sdhci-of-arasan: Fix kernel-doc warnings

Modify code to fix the warnings reported by kernel-doc for better
documentation.

Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Link: https://lore.kernel.org/r/1586195015-128992-7-git-send-email-manish.narani@xilinx.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-of-arasan: Modify clock operations handling
Manish Narani [Mon, 6 Apr 2020 17:43:34 +0000 (23:13 +0530)]
mmc: sdhci-of-arasan: Modify clock operations handling

The SDHCI clock operations are platform specific. So it better to define
them separately for particular platform. This will prevent multiple
if..else conditions and will make it easy for user to add their own
clock operations handlers.

Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Link: https://lore.kernel.org/r/1586195015-128992-6-git-send-email-manish.narani@xilinx.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-of-arasan: Rearrange the platform data structs for modularity
Manish Narani [Mon, 6 Apr 2020 17:43:33 +0000 (23:13 +0530)]
mmc: sdhci-of-arasan: Rearrange the platform data structs for modularity

Existing driver code has the platform specific structures scattered
throughout the driver code. Rearrange the platform specific data
structures for more modularity and readability. This will help in adding
new static functions with more ease.

Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Link: https://lore.kernel.org/r/1586195015-128992-5-git-send-email-manish.narani@xilinx.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: sdhci-of-arasan: Rename sdhci_arasan_data to avoid confusion
Manish Narani [Mon, 6 Apr 2020 17:43:32 +0000 (23:13 +0530)]
mmc: sdhci-of-arasan: Rename sdhci_arasan_data to avoid confusion

There is 'struct sdhci_arasan_data' but also
'struct sdhci_arasan_of_data sdhci_arasan_data'. Rename the latter to
avoid confusion with the name of the struct.

Reported-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Link: https://lore.kernel.org/r/1586195015-128992-4-git-send-email-manish.narani@xilinx.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agosdhci: arasan: Add support for Versal Tap Delays
Manish Narani [Mon, 6 Apr 2020 17:43:31 +0000 (23:13 +0530)]
sdhci: arasan: Add support for Versal Tap Delays

Add support to set tap delays for Xilinx Versal SD controller. The tap
delay registers have moved to SD controller space in Versal. Make the
changes accordingly.

Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1586195015-128992-3-git-send-email-manish.narani@xilinx.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agodt-bindings: mmc: arasan: Document 'xlnx,versal-8.9a' controller
Manish Narani [Mon, 6 Apr 2020 17:43:30 +0000 (23:13 +0530)]
dt-bindings: mmc: arasan: Document 'xlnx,versal-8.9a' controller

Add documentation for 'xlnx,versal-8.9a' SDHCI controller followed by
example.

Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/1586195015-128992-2-git-send-email-manish.narani@xilinx.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
4 years agommc: wbsd: Replace hardcoded command numbers with existing defines
Ulf Hansson [Thu, 9 Apr 2020 12:54:22 +0000 (14:54 +0200)]
mmc: wbsd: Replace hardcoded command numbers with existing defines

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20200409125422.21842-1-ulf.hansson@linaro.org
4 years agommc: s3cmci: Drop redundant code in s3cmci_setup_data()
Ulf Hansson [Tue, 7 Apr 2020 14:39:03 +0000 (16:39 +0200)]
mmc: s3cmci: Drop redundant code in s3cmci_setup_data()

The in-parameter struct mmc_data *data is never NULL, because the caller
always provides a valid pointer. Hence drop the corresponding redundant
code.

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Krzysztof Kozlowski <krzk@kernel.org>
Link: https://lore.kernel.org/r/20200407143903.22477-1-ulf.hansson@linaro.org
4 years agommc: au1xmmc: Drop redundant code in au1xmmc_send_command()
Ulf Hansson [Mon, 6 Apr 2020 11:43:37 +0000 (13:43 +0200)]
mmc: au1xmmc: Drop redundant code in au1xmmc_send_command()

The in-parameter "wait" is always set to 0 by the caller, hence just drop
it and its corresponding code.

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Manuel Lauss <manuel.lauss@gmail.com>
Link: https://lore.kernel.org/r/20200406114337.8802-1-ulf.hansson@linaro.org