platform/kernel/linux-starfive.git
6 years agodt-bindings: mmci: add stm32 sdmmc variant
Ludovic Barre [Mon, 8 Oct 2018 12:08:54 +0000 (14:08 +0200)]
dt-bindings: mmci: add stm32 sdmmc variant

This patch adds properties for stm32 sdmmc variant.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add stm32 sdmmc registers
Ludovic Barre [Mon, 8 Oct 2018 12:08:53 +0000 (14:08 +0200)]
mmc: mmci: add stm32 sdmmc registers

This patch adds stm32 sdmmc specific registers.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add clock divider for stm32 sdmmc
Ludovic Barre [Mon, 8 Oct 2018 12:08:52 +0000 (14:08 +0200)]
mmc: mmci: add clock divider for stm32 sdmmc

The STM32 sdmmc variant has a different clock divider.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add optional reset property
Ludovic Barre [Mon, 8 Oct 2018 12:08:51 +0000 (14:08 +0200)]
mmc: mmci: add optional reset property

This patch adds a optional reset management.
STM32 sdmmc variant needs to reset hardware block
during the power cycle procedure (for re-initialization).

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agodt-bindings: mmci: add optional reset property
Ludovic Barre [Mon, 8 Oct 2018 12:08:50 +0000 (14:08 +0200)]
dt-bindings: mmci: add optional reset property

This patch adds a optional reset management.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add variant property to not read datacnt
Ludovic Barre [Mon, 8 Oct 2018 12:08:49 +0000 (14:08 +0200)]
mmc: mmci: add variant property to not read datacnt

This patch adds a boolean property to not read datacnt register.
Needed to support the STM32 sdmmc variant. MMCIDATACNT
register should be read only after the data transfer is completed.
When reading after an error event the read data count value may be
different from the real number of data bytes transferred.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add variant property to write datactrl before command
Ludovic Barre [Mon, 8 Oct 2018 12:08:48 +0000 (14:08 +0200)]
mmc: mmci: add variant property to write datactrl before command

This patch adds a boolean property to allow to write datactrl
before to send command, whatever the command type (read or write).
Needed to support the STM32 sdmmc variant.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add variant property to define irq pio mask
Ludovic Barre [Mon, 8 Oct 2018 12:08:47 +0000 (14:08 +0200)]
mmc: mmci: add variant property to define irq pio mask

This patch allows to define specific pio mask for variants.
Needed to support the STM32 sdmmc variant which has some bits
with different meaning (bits: 21,20,13,12,9)

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add variant property to define dpsm bit
Ludovic Barre [Mon, 8 Oct 2018 12:08:46 +0000 (14:08 +0200)]
mmc: mmci: add variant property to define dpsm bit

This patch adds datactrl variant property to define
dpsm enable bit. Needed to support the STM32 variant
(STM32 has no dpsm enable bit).

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add variant properties to define cpsm & cmdresp bits
Ludovic Barre [Mon, 8 Oct 2018 12:08:45 +0000 (14:08 +0200)]
mmc: mmci: add variant properties to define cpsm & cmdresp bits

This patch adds command variant properties to define
cpsm enable bit and responses.
Needed to support the STM32 variant (shift of cpsm bit,
specific definition of commands response).

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: expand startbiterr to irqmask and error check
Ludovic Barre [Mon, 8 Oct 2018 12:08:44 +0000 (14:08 +0200)]
mmc: mmci: expand startbiterr to irqmask and error check

All variants don't pretend to have a startbiterr.
-While data error check, if status register return an error
(like  MCI_DATACRCFAIL) we must avoid to check MCI_STARTBITERR
(if not desired).
-expand start_err to MCI_IRQENABLE to avoid to set this bit by default.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add datactrl block size variant property
Ludovic Barre [Mon, 8 Oct 2018 12:08:43 +0000 (14:08 +0200)]
mmc: mmci: add datactrl block size variant property

This patch allows to define a datactrl block size
by variant, requested by STM32 sdmmc variant.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add set_clk/pwrreg callbacks
Ludovic Barre [Mon, 8 Oct 2018 12:08:42 +0000 (14:08 +0200)]
mmc: mmci: add set_clk/pwrreg callbacks

This patch adds set_clkreg and set_pwrreg callbacks
at mmci_host_ops to allow to call specific variant.
extends visibility of mmci_write_clk/pwrreg functions
to be used into specific file variant.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add validate_data callback
Ludovic Barre [Mon, 8 Oct 2018 12:08:41 +0000 (14:08 +0200)]
mmc: mmci: add validate_data callback

This patch adds validate_data callback at mmci_host_ops
to check specific constraints of variant.
Move mmci_validate_data function to regroup mmci_host_ops interfaces.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add dma_error callback
Ludovic Barre [Mon, 8 Oct 2018 12:08:40 +0000 (14:08 +0200)]
mmc: mmci: add dma_error callback

This patch adds dma_error callback at mmci_host_ops
to allow to call specific variant.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add dma_finalize callback
Ludovic Barre [Mon, 8 Oct 2018 12:08:39 +0000 (14:08 +0200)]
mmc: mmci: add dma_finalize callback

This patch adds dma_finalize callback at mmci_host_ops
to allow to call specific variant.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add dma_start callback
Ludovic Barre [Mon, 8 Oct 2018 12:08:38 +0000 (14:08 +0200)]
mmc: mmci: add dma_start callback

This patch adds dma_start callback to mmci_host_ops.
Create a generic mmci_dma_start function which regroup
common action between variant.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add get_next_data callback
Ludovic Barre [Mon, 8 Oct 2018 12:08:37 +0000 (14:08 +0200)]
mmc: mmci: add get_next_data callback

This patch adds get_next_data callback to mmci_host_ops.
Generic mmci_get_next_data factorizes next_cookie check and
the host ops call.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: add prepare/unprepare_data callbacks
Ludovic Barre [Mon, 8 Oct 2018 12:08:36 +0000 (14:08 +0200)]
mmc: mmci: add prepare/unprepare_data callbacks

This patch adds prepare/unprepare callbacks to mmci_host_ops.
Like this mmci_pre/post_request can be generic, mmci_prepare_data
and mmci_unprepare_data provide common next_cookie management.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: merge prepare data functions
Ludovic Barre [Mon, 8 Oct 2018 12:08:35 +0000 (14:08 +0200)]
mmc: mmci: merge prepare data functions

This patch merges the prepare data functions.
This allows to define a single access to prepare data service.
This prepares integration for mmci host ops.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: introduce dma_priv pointer to mmci_host
Ludovic Barre [Mon, 8 Oct 2018 12:08:34 +0000 (14:08 +0200)]
mmc: mmci: introduce dma_priv pointer to mmci_host

-Introduces dma_priv pointer to define specific
needs for each dma engine. This patch is needed to prepare
sdmmc variant with internal dma which not use dmaengine API.
-Moves next cookie to mmci host structure to share same cookie
management between all variants.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: create common mmci_dma_setup/release
Ludovic Barre [Mon, 8 Oct 2018 12:08:33 +0000 (14:08 +0200)]
mmc: mmci: create common mmci_dma_setup/release

This patch creates a common mmci_dma_setup/release which calls
dma_setup/release callbacks of mmci_host_ops and manages
common features like use_dma... If there is a fallbacks to
pio mode, dma functions must check use_dma.

error management:
-mmci_dmae_setup fail if Tx and Rx dma channels are not defined
-qcom_dma_setup fail if one of both dma channels is not defined,
 Qcom has no specific resource to release, just mmci dmae resource.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci-of-arasan: Add Support for AM654 MMC and PHY
Faiz Abbas [Thu, 4 Oct 2018 11:14:50 +0000 (16:44 +0530)]
mmc: sdhci-of-arasan: Add Support for AM654 MMC and PHY

The current arasan sdhci PHY configuration isn't compatible
with the PHY on TI's AM654 devices.

Therefore, add a new compatible, AM654 specific quirks
and a new AM654 specific set_clock function which
configures the PHY in a sane way.

Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci-of-arasan: Add a single data structure to incorporate pdata and soc_ctl_map
Faiz Abbas [Thu, 4 Oct 2018 11:14:49 +0000 (16:44 +0530)]
mmc: sdhci-of-arasan: Add a single data structure to incorporate pdata and soc_ctl_map

Currently, the driver passes platform data as a global structure
and uses the .data of of_device_id to pass the soc_ctl_map. To
make the implementation more flexible add a single data structure
that incorporates both of the above and pass it in the .data of
of_device_id.

Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agodt-bindings: mmc: sdhci-of-arasan: Add new compatible for AM654 MMC PHY
Faiz Abbas [Thu, 4 Oct 2018 11:14:48 +0000 (16:44 +0530)]
dt-bindings: mmc: sdhci-of-arasan: Add new compatible for AM654 MMC PHY

Add a new compatible to use the host controller driver with the
MMC PHY on TI's AM654 SOCs

Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mediatek: add bus_clk control
Chaotian Jing [Sat, 29 Sep 2018 02:29:55 +0000 (10:29 +0800)]
mmc: mediatek: add bus_clk control

when gate MSDC0_HCLK, access register will hang, even the MSDC driver
will never accessing register after HCLK was gated, but for safety, need
gate the bus_clk(which used to access register) too.

Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: dt-bindings: add "bus-clk" for MT2712
Chaotian Jing [Sat, 29 Sep 2018 02:29:54 +0000 (10:29 +0800)]
mmc: dt-bindings: add "bus-clk" for MT2712

On MT2712 MSDC0/3, HCLK/bus-clk need gate/ungate together,
or will hang when access MSDC register.

Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: dw_mmc-exynos: Add tuning for sdr and ddr timing for USH-I mode
Anand Moon [Thu, 27 Sep 2018 14:07:38 +0000 (14:07 +0000)]
mmc: dw_mmc-exynos: Add tuning for sdr and ddr timing for USH-I mode

Add tuning for sdr and ddr timing for USH-I mode sdr104/sdr50/ddr50
for host controller.

Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Anand Moon <linux.amoon@gmail.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: meson-mx-sdio: mark expected switch fall-through
Gustavo A. R. Silva [Fri, 5 Oct 2018 10:09:03 +0000 (12:09 +0200)]
mmc: meson-mx-sdio: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Notice that in this particular case, I replaced the "fall-through:"
comment with a proper "fall through", which is what GCC is
expecting to find.

Addresses-Coverity-ID: 1373880 ("Missing break in switch")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tifm_sd: Mark expected switch fall-through
Gustavo A. R. Silva [Fri, 5 Oct 2018 09:54:57 +0000 (11:54 +0200)]
mmc: tifm_sd: Mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Notice that in this particular case, I replaced the
"deliberate fall-through" comment with a proper "fall through"
at the bottom of the case, which is what GCC is expecting to find.

Addresses-Coverity-ID: 1373887 ("Missing break in switch")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: Change struct members from bool to u8
Ludovic Barre [Tue, 2 Oct 2018 12:09:03 +0000 (14:09 +0200)]
mmc: mmci: Change struct members from bool to u8

Recent versions of checkpatch have a new warning based on a documented
preference of Linus to not use bool in structures due to wasted space and
the size of bool is implementation dependent.  For more information, see
the email thread at https://lkml.org/lkml/2017/11/21/384

fix checkpatch --strict issues:
-CHECK: Avoid using bool structure members because of possible alignment
 issues - see: https://lkml.org/lkml/2017/11/21/384
-WARNING: Avoid using bool as bitfield. Prefer bool bitfields as
 unsigned int or u<8|16|32>

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: internalize dma_inprogress into mmci dma functions
Ludovic Barre [Fri, 21 Sep 2018 09:45:56 +0000 (11:45 +0200)]
mmc: mmci: internalize dma_inprogress into mmci dma functions

This patch internalizes the dma_inprogress into mmci dma interfaces.
This allows to simplify and prepare the next dma callbacks
for mmci host ops. dma_inprogress is called in mmci_dma_data_error
and mmci_dma_finalize.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: internalize dma map/unmap into mmci dma functions
Ludovic Barre [Fri, 21 Sep 2018 09:45:55 +0000 (11:45 +0200)]
mmc: mmci: internalize dma map/unmap into mmci dma functions

This patch internalizes the management of dma map/unmap into
mmci dma interfaces. This allows to simplify and prepare the next dma
callbacks for mmci host ops.
mmci_dma_unmap was called in mmci_data_irq & mmci_cmd_irq functions
and can be integrated in mmci_dma_data_error.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agodt-bindings: mmc: sh_mmcif: Document r8a7744 DT bindings
Biju Das [Tue, 25 Sep 2018 17:27:24 +0000 (18:27 +0100)]
dt-bindings: mmc: sh_mmcif: Document r8a7744 DT bindings

Add support for r8a7744 SoC. Renesas RZ/G1N (R8A7744) MMCIF is identical
to the R-Car Gen2 family.

Signed-off-by: Biju Das <biju.das@bp.renesas.com>
Reviewed-by: Chris Paterson <Chris.Paterson2@renesas.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agodt-bindings: mmc: renesas_sdhi: Add r8a7744 support
Biju Das [Tue, 25 Sep 2018 17:23:07 +0000 (18:23 +0100)]
dt-bindings: mmc: renesas_sdhi: Add r8a7744 support

Add support for r8a7744 SoC. Renesas RZ/G1N (R8A7744) SDHI is identical
to the R-Car Gen2 family.

Signed-off-by: Biju Das <biju.das@bp.renesas.com>
Reviewed-by: Chris Paterson <Chris.Paterson2@renesas.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci-pci-o2micro: Add quirk for O2 Micro dev 0x8620 rev 0x01
Yu Zhao [Sun, 23 Sep 2018 20:39:24 +0000 (14:39 -0600)]
mmc: sdhci-pci-o2micro: Add quirk for O2 Micro dev 0x8620 rev 0x01

This device reports SDHCI_CLOCK_INT_STABLE even though it's not
ready to take SDHCI_CLOCK_CARD_EN. The symptom is that reading
SDHCI_CLOCK_CONTROL after enabling the clock shows absence of the
bit from the register (e.g. expecting 0x0000fa07 = 0x0000fa03 |
SDHCI_CLOCK_CARD_EN but only observed the first operand).

mmc1: Timeout waiting for hardware cmd interrupt.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr:  0x00000000 | Version:  0x00000603
mmc1: sdhci: Blk size:  0x00000000 | Blk cnt:  0x00000000
mmc1: sdhci: Argument:  0x00000000 | Trn mode: 0x00000000
mmc1: sdhci: Present:   0x01ff0001 | Host ctl: 0x00000001
mmc1: sdhci: Power:     0x0000000f | Blk gap:  0x00000000
mmc1: sdhci: Wake-up:   0x00000000 | Clock:    0x0000fa03
mmc1: sdhci: Timeout:   0x00000000 | Int stat: 0x00000000
mmc1: sdhci: Int enab:  0x00ff0083 | Sig enab: 0x00ff0083
mmc1: sdhci: AC12 err:  0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps:      0x25fcc8bf | Caps_1:   0x00002077
mmc1: sdhci: Cmd:       0x00000000 | Max curr: 0x005800c8
mmc1: sdhci: Resp[0]:   0x00000000 | Resp[1]:  0x00000000
mmc1: sdhci: Resp[2]:   0x00000000 | Resp[3]:  0x00000000
mmc1: sdhci: Host ctl2: 0x00000008
mmc1: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: 0x00000000
mmc1: sdhci: ============================================

The problem happens during wakeup from S3. Adding a delay quirk
after power up reliably fixes the problem.

Signed-off-by: Yu Zhao <yuzhao@google.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: omap_hsmmc: Delete platform data GPIO CD and WP
Linus Walleij [Mon, 24 Sep 2018 11:30:51 +0000 (13:30 +0200)]
mmc: omap_hsmmc: Delete platform data GPIO CD and WP

The OMAP HSMMC driver has some elaborate and hairy handling for
passing GPIO card detect and write protect lines from a boardfile
into the driver: the machine defines a struct omap2_hsmmc_info
that is copied into struct omap_hsmmc_platform_data by
omap_hsmmc_pdata_init() in arch/arm/mach-omap2/hsmmc.c.

However the .gpio_cd and .gpio_wp fields are not copied from
omap2_hsmmc_info to omap_hsmmc_platform_data by
omap_hsmmc_pdata_init() so they remain unused. The only platform
defining omap2_hsmmc_info also define both to -1, unused.

It turn out there are no boardfiles passing any valid GPIO
lines into the OMAP HSMMC driver at all. And since we are not
going to add any more OMAP2 boardfiles, we can delete this
card detect and write protect handling altogether.

This seems to also fix a bug: the card detect callback
mmc_gpio_get_cd() in the slot GPIO core needs to be called
by drivers utilizing slot GPIO. It appears the the boardfile
quirks were not doing this right, so this would only get
called for boardfiles, i.e. since no boardfile was using it,
never.

Just assign mmc_gpio_get_cd() unconditionally to omap_hsmmc_ops
.get_cd() so card detects from the device tree works.
AFAICT card detect with GPIO lines assigned from
mmc_of_parse() are not working at the moment, but that is
no regression since it probably never worked.

Cc: Tony Lindgren <tony@atomide.com>
Cc: linux-omap@vger.kernel.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: omap_hsmmc: Kill off cover detection
Linus Walleij [Mon, 24 Sep 2018 11:30:50 +0000 (13:30 +0200)]
mmc: omap_hsmmc: Kill off cover detection

Cover detection appears to be a feature protecting the SD
card on mobile phones with a slide-cover, such as some Nokia
phones. The idea seems to be to not allow access to the
SD card when the cover is open.

It is only usable with platform data from board files, but
no board file in the kernel is using it, yet it takes up
a sizeable chunk of code in the OMAP HSMMC driver.

Since we do not add new board files for the OMAPs any target
that need this should anyway reimplement it properly using
the device tree, so delete this legacy code.

The driver is marked as orphan in MAINTAINERS by the way.

Cc: Tony Lindgren <tony@atomide.com>
Cc: linux-omap@vger.kernel.org
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Tony Lindgren <tony@atomide.com>
Acked-by: Kishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: core: Allow building PWRSEQ_SD8787 with LIBERTAS_SDIO
Lubomir Rintel [Mon, 24 Sep 2018 08:56:32 +0000 (10:56 +0200)]
mmc: core: Allow building PWRSEQ_SD8787 with LIBERTAS_SDIO

The sd8686 "libertas" SDIO adapter's power is controlled with WLAN_RST
and WLAN_PD pins -- pretty much the same way as sd8787. Allow building
the power sequencing driver along with the libertas Wi-Fi driver.

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: sirf: Use the slot GPIO descriptor
Linus Walleij [Mon, 24 Sep 2018 08:02:33 +0000 (10:02 +0200)]
mmc: sdhci: sirf: Use the slot GPIO descriptor

This driver is complicating things for no reason: the "cd"
GPIO can easily be retrieved from the device tree if present
using just mmc_gpiod_request_cd(), which will fetch the
descriptor from the device tree using the standard binding
just fine.

If the retrieveal is successful, we also request the IRQ.

As a result the private subdriver data can be removed
entirely.

Cc: Weijun Yang <york.yang@csr.com>
Cc: Barry Song <baohua@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: pxav3: Delete GPIO handling
Linus Walleij [Sun, 23 Sep 2018 07:03:21 +0000 (09:03 +0200)]
mmc: sdhci: pxav3: Delete GPIO handling

The platform data for the PXAv3 driver allows passing a card
detect GPIO, but this code is not used in the kernel.

In order to not encourage the use of the old global GPIO
numberspace we need to remove this.

Card detect (and write protect) GPIO can easily be added into
the driver using machine descriptor tables instead, and the
descriptor-based (gpiod) variants of the slot GPIO APIs.

Cc: Jisheng Zhang <jszhang@marvell.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: spear: Use the slot GPIO descriptor
Linus Walleij [Sun, 23 Sep 2018 06:33:20 +0000 (08:33 +0200)]
mmc: sdhci: spear: Use the slot GPIO descriptor

This driver is complicating things for no reason: the "cd"
GPIO can easily be retrieved from the device tree if present
using just mmc_gpiod_request_cd(), which will fetch the
descriptor from the device tree using the standard binding
just fine.

Cc: Viresh Kumar <vireshk@kernel.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mmci: Drop support for pdata GPIO numbers
Linus Walleij [Thu, 20 Sep 2018 23:01:10 +0000 (16:01 -0700)]
mmc: mmci: Drop support for pdata GPIO numbers

All the machines using the MMCI are passing GPIOs for the
card detect and write protect using the device tree or
descriptor table (one single case, Integrator/AP IM-PD1).

Drop support for passing global GPIO numbers through
platform data, noone is using it.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: fix __sdhci_adma_write_desc
Jisheng Zhang [Mon, 17 Sep 2018 05:30:41 +0000 (13:30 +0800)]
mmc: sdhci: fix __sdhci_adma_write_desc

If hosts provides ops->adma_write_desc, we should not fall back to the
general sdhci_adma_write_desc().

Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: mxcmmc: replace spin_lock_irqsave with spin_lock in ISR
jun qian [Tue, 11 Sep 2018 14:47:01 +0000 (07:47 -0700)]
mmc: mxcmmc: replace spin_lock_irqsave with spin_lock in ISR

As you are already in ISR, it is unnecessary to call spin_lock_irqsave.

Signed-off-by: jun qian <hangdianqj@163.com>
Reviewed-by: Vladimir Zapolskiy <vz@mleia.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: renesas_sdhi_internal_dmac: set scatter/gather max segment size
Niklas Söderlund [Thu, 13 Sep 2018 14:47:08 +0000 (16:47 +0200)]
mmc: renesas_sdhi_internal_dmac: set scatter/gather max segment size

Fix warning when running with CONFIG_DMA_API_DEBUG_SG=y by allocating a
device_dma_parameters structure and filling in the max segment size. The
size used is the result of a discussion with Renesas hardware engineers
and unfortunately not found in the datasheet.

  renesas_sdhi_internal_dmac ee140000.sd: DMA-API: mapping sg segment
  longer than device claims to support [len=126976] [max=65536]

Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
[wsa: simplified some logic after validating intended dma_parms life cycle
      and added comment]
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sunxi: Use new timing mode for A64 eMMC controller
Chen-Yu Tsai [Thu, 6 Sep 2018 15:33:04 +0000 (23:33 +0800)]
mmc: sunxi: Use new timing mode for A64 eMMC controller

The eMMC controller is also a new timing mode controller, but it doesn't
have the timing mode switch. It does however have signal delay and
calibration controls, typical of Allwinner MMC controllers that support
the new timing mode.

Enable the new timing mode setting for the A64 eMMC controller. This
also enables MMC HS-DDR modes, which gives higher throughput for eMMC
chips that support it, and can deliver such throughput.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sunxi: Clarify new timing mode usage and implementation
Chen-Yu Tsai [Thu, 6 Sep 2018 15:31:07 +0000 (23:31 +0800)]
mmc: sunxi: Clarify new timing mode usage and implementation

Newer sunxi mmc controller variants support what they call the "new
timing mode". Support for this was implemented in two ways, according
to the hardware that was seen at the time.

The first type retained the old timing mode, and both the clock and mmc
controllers had switches to select which mode was used. Both switches
had to be set to the same setting. This variant was denoted with the
.has_timings_switch field in the sunxi_mmc_cfg structure. This hardware
is only seen on the A83T.

The second type did away with the old timing mode. The clock controller
no longer had the mode selection or clock delay setting bits. In some
cases the mmc controller retained its mode selection bit, but this
always needed to be set to the new mode, or instabilities would occur.
In a few cases, such as the A64 and H6 eMMC controller, the mode
selection bit is gone, but the controller still behaves like the new
timing mode, requiring the module clock to be double the card clock
in DDR transfer modes. This variant is denoted with the
.needs_new_timings field.

This patch adds more comments explaining the two fields, as well as
the possibly nonexistent mode switch in the mmc controller.

The .has_timings_switch is renamed to .ccu_has_timings_switch to clarify
its meaning.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci-sprd: Add Spreadtrum's initial host controller
Chunyan Zhang [Thu, 30 Aug 2018 08:21:44 +0000 (16:21 +0800)]
mmc: sdhci-sprd: Add Spreadtrum's initial host controller

This patch adds the initial support of Secure Digital Host Controller
Interface compliant controller found in some latest Spreadtrum chipsets.
This patch has been tested on the version of SPRD-R11 controller.

R11 is a variant based on SD v4.0 specification.

With this driver, R11 mmc can be initialized, can be mounted, read and
written.

Original-by: Billows Wu <billows.wu@unisoc.com>
Signed-off-by: Chunyan Zhang <chunyan.zhang@unisoc.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agodt-bindings: sdhci-sprd: Add bindings for the sdhci-sprd controller
Chunyan Zhang [Thu, 30 Aug 2018 08:21:45 +0000 (16:21 +0800)]
dt-bindings: sdhci-sprd: Add bindings for the sdhci-sprd controller

This patch adds the device-tree binding documentation for Spreadtrum
SDHCI driver.

Signed-off-by: Chunyan Zhang <chunyan.zhang@unisoc.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: SDMA may use Auto-CMD23 in v4 mode
Chunyan Zhang [Thu, 30 Aug 2018 08:21:43 +0000 (16:21 +0800)]
mmc: sdhci: SDMA may use Auto-CMD23 in v4 mode

When Host Version 4 Enable is set to 1, SDMA uses ADMA System Address
register (05Fh-058h) instead of using register (000h-004h) to indicate
its system address of data location. The register (000h-004h) is
re-assigned to 32-bit Block Count and Auto CMD23 argument, so then SDMA
may use Auto CMD23.

Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: Add Auto CMD Auto Select support
Chunyan Zhang [Thu, 30 Aug 2018 08:21:42 +0000 (16:21 +0800)]
mmc: sdhci: Add Auto CMD Auto Select support

As SD Host Controller Specification v4.10 documents:
Host Controller Version 4.10 defines this "Auto CMD Auto Select" mode.
Selection of Auto CMD depends on setting of CMD23 Enable in the Host
Control 2 register which indicates whether card supports CMD23. If CMD23
Enable =1, Auto CMD23 is used and if CMD23 Enable =0, Auto CMD12 is
used. In case of Version 4.10 or later, use of Auto CMD Auto Select is
recommended rather than use of Auto CMD12 Enable or Auto CMD23
Enable.

This patch add this new mode support.

Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: Add 32-bit block count support for v4 mode
Chunyan Zhang [Thu, 30 Aug 2018 08:21:41 +0000 (16:21 +0800)]
mmc: sdhci: Add 32-bit block count support for v4 mode

Host Controller Version 4.10 re-defines SDMA System Address register
as 32-bit Block Count for v4 mode, and SDMA uses ADMA System
Address register (05Fh-058h) instead if v4 mode is enabled. Also
when using 32-bit block count, 16-bit block count register need
to be set to zero.

Since using 32-bit Block Count would cause problems for auto-cmd23,
it can be chosen via host->quirk2.

Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: Add ADMA2 64-bit addressing support for V4 mode
Chunyan Zhang [Thu, 30 Aug 2018 08:21:40 +0000 (16:21 +0800)]
mmc: sdhci: Add ADMA2 64-bit addressing support for V4 mode

ADMA2 64-bit addressing support is divided into V3 mode and V4 mode.
So there are two kinds of descriptors for ADMA2 64-bit addressing
i.e. 96-bit Descriptor for V3 mode, and 128-bit Descriptor for V4
mode. 128-bit Descriptor is aligned to 8-byte.

For V4 mode, ADMA2 64-bit addressing is enabled via Host Control 2
register.

Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
[Ulf: Fixed conflict while applying]
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: Change SDMA address register for v4 mode
Chunyan Zhang [Thu, 30 Aug 2018 08:21:39 +0000 (16:21 +0800)]
mmc: sdhci: Change SDMA address register for v4 mode

According to the SD host controller specification version 4.10, when
Host Version 4 is enabled, SDMA uses ADMA System Address register
(05Fh-058h) instead of using SDMA System Address register to
support both 32-bit and 64-bit addressing.

Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: Add sd host v4 mode
Chunyan Zhang [Thu, 30 Aug 2018 08:21:38 +0000 (16:21 +0800)]
mmc: sdhci: Add sd host v4 mode

For SD host controller version 4.00 or later ones, there're two
modes of implementation - Version 3.00 compatible mode or
Version 4 mode.  This patch introduced an interface to enable
v4 mode.

Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: Add version V4 definition
Chunyan Zhang [Thu, 30 Aug 2018 08:21:37 +0000 (16:21 +0800)]
mmc: sdhci: Add version V4 definition

Added definitions for v400, v410, v420.

Signed-off-by: Chunyan Zhang <zhang.chunyan@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: fix inconsistent IS_ERR and PTR_ERR
YueHaibing [Tue, 4 Sep 2018 02:59:09 +0000 (10:59 +0800)]
mmc: tegra: fix inconsistent IS_ERR and PTR_ERR

Fix inconsistent IS_ERR and PTR_ERR in tegra_sdhci_init_pinctrl_info,
the proper pointer to be passed as argument is 'pinctrl_state_1v8'

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Reviewed-by: Aapo Vienamo <aapo.vienamo@iki.fi>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Implement periodic pad calibration
Aapo Vienamo [Mon, 20 Aug 2018 09:23:33 +0000 (12:23 +0300)]
mmc: tegra: Implement periodic pad calibration

Rerun the pad calibration procedure before sdhci_request() if
the 100 ms recalibration interval has been exceeded.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Implement HS400 delay line calibration
Aapo Vienamo [Fri, 10 Aug 2018 18:14:01 +0000 (21:14 +0300)]
mmc: tegra: Implement HS400 delay line calibration

Implement HS400 specific delay line calibration procedure. This is a
Tegra specific procedure and has to be performed regardless whether
enhanced strobe or HS400 tuning is used.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Implement HS400 enhanced strobe
Aapo Vienamo [Fri, 10 Aug 2018 18:14:00 +0000 (21:14 +0300)]
mmc: tegra: Implement HS400 enhanced strobe

Implement eMMC HS400 enhanced strobe. Enhanced strobe is an alternative
mechanism to the HS400 tuning procedure.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Parse and program DQS trim value
Aapo Vienamo [Fri, 10 Aug 2018 18:13:59 +0000 (21:13 +0300)]
mmc: tegra: Parse and program DQS trim value

Parse and program the HS400 DQS trim value from DT. Program a fallback
value in case the property is missing.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agodt-bindings: mmc: Add DQS trim value to Tegra SDHCI
Aapo Vienamo [Fri, 10 Aug 2018 18:13:58 +0000 (21:13 +0300)]
dt-bindings: mmc: Add DQS trim value to Tegra SDHCI

Document HS400 DQS trim value device tree property.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Enable UHS and HS200 modes for Tegra186
Aapo Vienamo [Thu, 30 Aug 2018 15:06:28 +0000 (18:06 +0300)]
mmc: tegra: Enable UHS and HS200 modes for Tegra186

Set nvquirks to enable higher speed modes.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Enable UHS and HS200 modes for Tegra210
Aapo Vienamo [Thu, 30 Aug 2018 15:06:27 +0000 (18:06 +0300)]
mmc: tegra: Enable UHS and HS200 modes for Tegra210

Set nvquirks to enable higher speed modes.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Disable card clock during tuning cmd on Tegra210
Aapo Vienamo [Thu, 30 Aug 2018 15:06:26 +0000 (18:06 +0300)]
mmc: tegra: Disable card clock during tuning cmd on Tegra210

Implement tegra210_sdhci_writew() to disable card clock and issue a
reset when the tuning command is sent. This is done to prevent an
intermittent hang with around 10 % failure rate during tuning.

Add tegra186_sdhci_ops because this workaround is specific to Tegra210.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Remove tegra_sdhci_writew() from tegra210_sdhci_ops
Aapo Vienamo [Thu, 30 Aug 2018 15:06:25 +0000 (18:06 +0300)]
mmc: tegra: Remove tegra_sdhci_writew() from tegra210_sdhci_ops

tegra_sdhci_writew() defers the write to SDHCI_TRANSFER_MODE until
SDHCI_COMMAND is written. This is not necessary on Tegra210 and Tegra186
and it breaks read-modify-write operations on SDHCI_TRANSFER_MODE
because writes to SDHCI_TRANSFER_MODE aren't visible until SDHCI_COMMAND
has been written to. This results in tuning failures on Tegra210.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Use standard SDHCI tuning on Tegra210 and Tegra186
Aapo Vienamo [Thu, 30 Aug 2018 15:06:24 +0000 (18:06 +0300)]
mmc: tegra: Use standard SDHCI tuning on Tegra210 and Tegra186

Add a new sdhci_ops struct for Tegra210 and Tegra186 which doesn't
set the custom tuning callback used on previous SoC generations.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Configure default trim value on reset
Aapo Vienamo [Thu, 30 Aug 2018 15:06:23 +0000 (18:06 +0300)]
mmc: tegra: Configure default trim value on reset

Program the outbound sampling trim value in tegra_sdhci_reset(). Unlike
the outbound tap value this does not depend on the signaling mode and
needs to be only programmed once.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Configure default tap values
Aapo Vienamo [Thu, 30 Aug 2018 15:06:22 +0000 (18:06 +0300)]
mmc: tegra: Configure default tap values

Set the default inbound timing adjustment tap value on reset and on
non-tunable modes.

The default tap value is not programmed on tunable modes because the
tuning sequence is used instead to determine the tap value.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Parse default trim and tap from dt
Aapo Vienamo [Thu, 30 Aug 2018 15:06:21 +0000 (18:06 +0300)]
mmc: tegra: Parse default trim and tap from dt

Parse the default inbound and outbound sampling trimmer values from
the device tree.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Add a workaround for tap value change glitch
Aapo Vienamo [Thu, 30 Aug 2018 15:06:20 +0000 (18:06 +0300)]
mmc: tegra: Add a workaround for tap value change glitch

Add quirk to disable the card clock during configuration of the tap
value in tegra_sdhci_set_tap() and issue sdhci_reset() after value
change. This is a workaround to avoid propagation of a potential
glitch caused by setting the tap value.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Enable pad calibration on Tegra210 and Tegra186
Aapo Vienamo [Thu, 30 Aug 2018 15:06:19 +0000 (18:06 +0300)]
mmc: tegra: Enable pad calibration on Tegra210 and Tegra186

Set NVQUIRK_HAS_PADCALIB on Tegra210 and Tegra186 to enable automatic
pad drive strength calibration.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Perform pad calibration after voltage switch
Aapo Vienamo [Thu, 30 Aug 2018 15:06:18 +0000 (18:06 +0300)]
mmc: tegra: Perform pad calibration after voltage switch

Run the automatic pad calibration after voltage switching if
tegra_host->pad_calib_required is set.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Program pad autocal offsets from dt
Aapo Vienamo [Thu, 30 Aug 2018 15:06:17 +0000 (18:06 +0300)]
mmc: tegra: Program pad autocal offsets from dt

Parse the pad drive strength calibration offsets from the device tree.
Program the calibration offsets in accordance with the current signaling
mode.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Disable card clock during pad calibration
Aapo Vienamo [Thu, 30 Aug 2018 15:06:16 +0000 (18:06 +0300)]
mmc: tegra: Disable card clock during pad calibration

Disable the card clock during automatic pad drive strength calibration
and re-enable it afterwards.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Power on the calibration pad
Aapo Vienamo [Thu, 30 Aug 2018 15:06:15 +0000 (18:06 +0300)]
mmc: tegra: Power on the calibration pad

Automatic pad drive strength calibration is performed on a separate pad
identical to the ones used for driving the actual bus. Power on the
calibration pad during the calibration procedure and power it off
afterwards to save power.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Set calibration pad voltage reference
Aapo Vienamo [Thu, 30 Aug 2018 15:06:14 +0000 (18:06 +0300)]
mmc: tegra: Set calibration pad voltage reference

Configure the voltage reference used by the automatic pad drive strength
calibration procedure. The value is a magic number from the TRM.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Poll for calibration completion
Aapo Vienamo [Thu, 30 Aug 2018 15:06:13 +0000 (18:06 +0300)]
mmc: tegra: Poll for calibration completion

Implement polling with 10 ms timeout for automatic pad drive strength
calibration.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tegra: Reconfigure pad voltages during voltage switching
Aapo Vienamo [Thu, 30 Aug 2018 15:06:12 +0000 (18:06 +0300)]
mmc: tegra: Reconfigure pad voltages during voltage switching

Parse the pinctrl state and nvidia,only-1-8-v properties from the device
tree. Validate the pinctrl and regulator configuration before unmasking
UHS modes. Implement pad voltage state reconfiguration in the mmc
start_signal_voltage_switch() callback. Add NVQUIRK_NEEDS_PAD_CONTROL
and add set it for Tegra210 and Tegra186.

The pad configuration is done in the mmc callback because the order of
pad reconfiguration and sdhci voltage switch depend on the voltage to
which the transition occurs.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agodt-bindings: mmc: Add Tegra SDHCI sampling trimmer values
Aapo Vienamo [Thu, 30 Aug 2018 15:06:05 +0000 (18:06 +0300)]
dt-bindings: mmc: Add Tegra SDHCI sampling trimmer values

Document the Tegra SDHCI inbound and outbound sampling trimmer values.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agodt-bindings: Add Tegra SDHCI pad pdpu offset bindings
Aapo Vienamo [Thu, 30 Aug 2018 15:06:04 +0000 (18:06 +0300)]
dt-bindings: Add Tegra SDHCI pad pdpu offset bindings

Add bindings documentation for pad pull up and pull down offset values to be
programmed before executing automatic pad drive strength calibration.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agodt-bindings: mmc: tegra: Add pad voltage control properties
Aapo Vienamo [Thu, 30 Aug 2018 15:06:03 +0000 (18:06 +0300)]
dt-bindings: mmc: tegra: Add pad voltage control properties

Document the pinctrl bindings used by the SDHCI driver to reconfigure
pad voltages on controllers supporting multiple voltage levels.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tmio: remove now unused variable
Wolfram Sang [Thu, 30 Aug 2018 12:16:03 +0000 (14:16 +0200)]
mmc: tmio: remove now unused variable

This variable is unused now after some refactoring.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tmio: more concise clk calculation
Wolfram Sang [Thu, 30 Aug 2018 12:14:38 +0000 (14:14 +0200)]
mmc: tmio: more concise clk calculation

Concise, but still readable.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: tmio: Fix SCC error detection
Masaharu Hayakawa [Wed, 29 Aug 2018 23:32:07 +0000 (01:32 +0200)]
mmc: tmio: Fix SCC error detection

SDR104, HS200 and HS400 need to check for SCC error. If SCC error is
detected, retuning is necessary.

Signed-off-by: Masaharu Hayakawa <masaharu.hayakawa.ry@renesas.com>
[Niklas: update commit message]
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
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>
6 years agommc: renesas_sdhi: skip SCC error check when retuning
Masaharu Hayakawa [Wed, 29 Aug 2018 23:32:06 +0000 (01:32 +0200)]
mmc: renesas_sdhi: skip SCC error check when retuning

Checking for SCC error during retuning is unnecessary.

Signed-off-by: Masaharu Hayakawa <masaharu.hayakawa.ry@renesas.com>
[Niklas: fix small style issue]
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
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>
6 years agommc: core: add helper to see if a host is doing a retune
Niklas Söderlund [Wed, 29 Aug 2018 23:32:05 +0000 (01:32 +0200)]
mmc: core: add helper to see if a host is doing a retune

Add a helper to allow host drivers checking if a retune is in progress.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
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>
6 years agommc: tmio: refactor CLK_CTL bit calculation
Masahiro Yamada [Thu, 23 Aug 2018 04:44:20 +0000 (13:44 +0900)]
mmc: tmio: refactor CLK_CTL bit calculation

for (clk = 0x80000080; new_clock >= (clock << 1); clk >>= 1)
          clock <<= 1;

... is too tricky, hence I replaced with

  roundup_pow_of_two(divisor) >> 2

'(clk >> 22) & 0x1' is the bit test for the 1/1 divisor, but
it is not clear.  'divisor <= 1' is easier to understand.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
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>
6 years agommc: renesas_sdhi: merge clk_{start,stop} functions to set_clock
Masahiro Yamada [Thu, 23 Aug 2018 04:44:19 +0000 (13:44 +0900)]
mmc: renesas_sdhi: merge clk_{start,stop} functions to set_clock

renesas_sdhi_clk_start() and renesas_sdhi_clk_stop() are now only
called from renesas_sdhi_set_clock().  Merge them.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
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>
6 years agommc: sdhci-of-dwcmshc: solve 128MB DMA boundary limitation
Jisheng Zhang [Tue, 28 Aug 2018 09:48:14 +0000 (17:48 +0800)]
mmc: sdhci-of-dwcmshc: solve 128MB DMA boundary limitation

When using DMA, if the DMA addr spans 128MB boundary, we have to split
the DMA transfer into two so that each one doesn't exceed the boundary.

Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: introduce adma_write_desc() hook to struct sdhci_ops
Jisheng Zhang [Tue, 28 Aug 2018 09:47:23 +0000 (17:47 +0800)]
mmc: sdhci: introduce adma_write_desc() hook to struct sdhci_ops

Add this hook so that it can be overridden with driver specific
implementations. We also let the original sdhci_adma_write_desc()
accept &desc so that the function can set its new value. Then export
the function so that it could be reused by driver's specific
implementations.

Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: add adma_table_cnt member to struct sdhci_host
Jisheng Zhang [Tue, 28 Aug 2018 09:46:35 +0000 (17:46 +0800)]
mmc: sdhci: add adma_table_cnt member to struct sdhci_host

This patch adds adma_table_cnt member to struct sdhci_host to give more
flexibility to drivers to control the ADMA table count.

Default value of adma_table_cnt is set to (SDHCI_MAX_SEGS * 2 + 1).

Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: Convert to using %pOFn instead of device_node.name
Rob Herring [Tue, 28 Aug 2018 01:52:33 +0000 (20:52 -0500)]
mmc: Convert to using %pOFn instead of device_node.name

In preparation to remove the node name pointer from struct device_node,
convert printf users to use the %pOFn format specifier.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Hu Ziji <huziji@marvell.com>
Cc: Ulf Hansson <ulf.hansson@linaro.org>
Cc: linux-mmc@vger.kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: Export sdhci_request()
Aapo Vienamo [Mon, 20 Aug 2018 09:23:32 +0000 (12:23 +0300)]
mmc: sdhci: Export sdhci_request()

Allow SDHCI drivers to hook code before and after sdhci_request() by
making it externally visible.

Signed-off-by: Aapo Vienamo <avienamo@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agosdhci: acpi: add qcom sdhci host reset quirk fix
Wang Dongsheng [Thu, 16 Aug 2018 04:48:43 +0000 (12:48 +0800)]
sdhci: acpi: add qcom sdhci host reset quirk fix

After host requests RESET_FOR_ALL action, the hardware output an
interrupt for OS and waiting for the OS to approve.

Before writing this fix, ACPI GED has handled the interrupt. But
the ACPI GED belongs to a slow process, and sometimes the handling
process time is more than 100ms(Mutex wait more than 100ms). So
drop the GED solution and add this quirk fix.

Signed-off-by: Wang Dongsheng <dongsheng.wang@hxt-semitech.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agosdhci: acpi: add free_slot callback
Wang Dongsheng [Thu, 16 Aug 2018 04:48:42 +0000 (12:48 +0800)]
sdhci: acpi: add free_slot callback

The device specific resource can be free in free_slot after
removing host controller.

Signed-off-by: Wang Dongsheng <dongsheng.wang@hxt-semitech.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci-of-esdhc: add erratum A008171 support
Yinbo Zhu [Thu, 23 Aug 2018 08:48:32 +0000 (16:48 +0800)]
mmc: sdhci-of-esdhc: add erratum A008171 support

In tuning mode of operation, when TBCTL[TB_EN] is set, eSDHC may report
one of the following errors :
1)Tuning error while running tuning operation where SYSCTL2[SAMPCLKSEL]
will not get set even when SYSCTL2[EXTN] is reset. OR
2)Data transaction error (e.g. IRQSTAT[DCE], IRQSTAT[DEBE]) during data
transaction errors.
This issue occurs when the data window sampled within eSDHC is in full
cycle. So, in that case, eSDHC is not able to find out the start and
end points of the data window and sets the sampling pointer at default
location (which is middle of the internal SD clock). If this sampling
point coincides with the data eye boundary, then it can result in the
above mentioned errors. Impact: Tuning mode of operation for SDR50,
SDR104 or HS200 speed modes may not work properly
Workaround: In case eSDHC reports tuning error or data errors in tuning
mode of operation, by add the erratum A008171 support to fix the issue.

Signed-off-by: Yinbo Zhu <yinbo.zhu@nxp.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: sdhci: add tuning error codes
Yinbo Zhu [Thu, 23 Aug 2018 08:48:31 +0000 (16:48 +0800)]
mmc: sdhci: add tuning error codes

This patch is to add tuning error codes to
judge tuning state

Signed-off-by: Yinbo Zhu <yinbo.zhu@nxp.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
6 years agommc: uniphier-sd: add UniPhier SD/eMMC controller driver
Masahiro Yamada [Thu, 23 Aug 2018 04:44:18 +0000 (13:44 +0900)]
mmc: uniphier-sd: add UniPhier SD/eMMC controller driver

Here is another TMIO MMC variant found in Socionext UniPhier SoCs.

As commit b6147490e6aa ("mmc: tmio: split core functionality, DMA and
MFD glue") said, these MMC controllers use the IP from Panasonic.

However, the MMC controller in the TMIO (Toshiba Mobile IO) MFD chip
was the first upstreamed user of this IP.  The common driver code
for this IP is now called 'tmio-mmc-core' in Linux although it is a
historical misnomer.

Anyway, this driver select's MMC_TMIO_CORE to borrow the common code
from tmio-mmc-core.c

Older UniPhier SoCs (LD4, Pro4, sLD8) support the external DMA engine
like renesas_sdhi_sys_dmac.c.  The difference is UniPhier SoCs use a
single DMA channel whereas Renesas chips request separate channels for
RX and TX.

Newer UniPhier SoCs (Pro5 and later) support the internal DMA engine
like renesas_sdhi_internal_dmac.c  The register map is almost the same,
so I guess Renesas and Socionext use the same internal DMA hardware.
The main difference is, the register offsets are doubled for Renesas.

                        Renesas      Socionext
                        SDHI         UniPhier
  DM_CM_DTRAN_MODE      0x820        0x410
  DM_CM_DTRAN_CTRL      0x828        0x414
  DM_CM_RST             0x830        0x418
  DM_CM_INFO1           0x840        0x420
  DM_CM_INFO1_MASK      0x848        0x424
  DM_CM_INFO2           0x850        0x428
  DM_CM_INFO2_MASK      0x858        0x42c
  DM_DTRAN_ADDR         0x880        0x440
  DM_DTRAN_ADDREX        ---         0x444

This comes from the difference of host->bus_shift; 2 for Renesas SoCs,
and 1 for UniPhier SoCs.  Also, the datasheet for UniPhier SoCs defines
DM_DTRAN_ADDR and DM_DTRAN_ADDREX as two separate registers.

It could be possible to factor out the DMA common code by introducing
some hooks to cope with platform quirks, but this patch does not touch
that for now.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>