Convert the Renesas SDHI SD/SDIO controller driver using SYS-DMAC from
an open-coded dev_pm_ops structure to DEFINE_RUNTIME_DEV_PM_OPS() and
pm_ptr(). This simplifies the code, and reduces kernel size in case
CONFIG_PM is disabled.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
This patch describes the two MMC controllers of the Loongson-2K0300 SoC,
one providing an eMMC interface and the other exporting an SD/SDIO
interface.
Its hardware design is similar to that of the Loongson-2K2000, but it
suffers from hardware defects such as missing CMD48 interrupts.
Signed-off-by: Binbin Zhou <zhoubinbin@loongson.cn>
Reviewed-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
More Loongson SoCs will be added, gathering all SoC private data
(`loongson2_mmc_pdata`) together to make the code clearer.
No functional change intended.
Signed-off-by: Binbin Zhou <zhoubinbin@loongson.cn>
Reviewed-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Add "loongson,ls2k0300-mmc" dedicated compatible to represent the
eMMC/SD/SDIO controller interface of the Loongson-2K0300 chip.
Its hardware design is similar to that of the Loongson-2K2000, but it
suffers from hardware defects such as missing CMD48 interrupts.
Signed-off-by: Binbin Zhou <zhoubinbin@loongson.cn>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Reviewed-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
In sdhci_bst_alloc_bounce_buffer(), if dma_alloc_coherent() fails, the
function immediately returns -ENOMEM without releasing the reserved
memory, which results in a memory leak.
Add the missing of_reserved_mem_device_release() call before returning
-ENOMEM to properly clean up the reserved memory.
Fixes: 695824f456 ("mmc: sdhci: add Black Sesame Technologies BST C1200 controller driver")
Signed-off-by: Felix Gu <ustc.gu@gmail.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The mux branch contains updates to the mux core along with some
corresponding changes for a couple of consumer drivers, including an mmc
driver. Let's merge it into the next branch to get it tested and queued for
the next release.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Some hardware designs route data or control signals through a mux to
support multiple devices on a single sdhi controller.
In particular SolidRun RZ/G2L/G2LC/V2L System on Module use a mux for
switching between soldered eMMC and an optional microSD on a carrier
board, e.g. for development or provisioning.
SD/SDIO/eMMC are not well suited for runtime switching between different
cards, however boot-time selection is possible and useful - in
particular considering dt overlays.
Add support for an optional SD/SDIO/eMMC mux defined in dt, and select
it during probe.
Similar functionality already exists in other places, e.g. i2c-omap.
Signed-off-by: Josua Mayer <josua@solid-run.com>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Tested-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The multiplexer subsystem was initially designed to be completely
hidden, relying on consumers to "select MULTIPLEXER" explicitly.
Drivers implementing multiplexers depend on this hidden symbol.
This prevents users from manually enabling both the mux core and any of
the multiplexer drivers.
All multiplexer drivers in drivers/mux/ can operate standalone without a
consumer. This is particularly useful in a device-tree, where a default
state can be set through the idle-state property.
Over time, several drivers have added "select MULTIPLEXER" dependencies,
some of which require a mux and some consider it optional. v7.0-rc1
shows 15 such occurrences in Kconfig files, in a variety of subsystems.
The natural step forward to allow enabling mux core and drivers would be
adding a prompt and help text to the existing symbol.
This violates the general Kbuild advice to avoid selecting visible
symbols for all existing consumers of the mux core.
Add the new config symbol MUX_CORE with a prompt and help text as a
wrapper for users to enable manually. This avoids existing consumers
automatically selecting a visible symbol.
Change the MULTIPLEXER symbol from tristate to bool. This avoids complex
dependencies if users were to attempt a configuration where the mux is a
module but one of its consumers is built-in, as well as difficulties
keeping the state of visible and invisible symbols in sync.
Further convert the "menu ... depends on ..." structure to "if ... menu
... endmenu endif". These are functionally equivalent, but the new
structure is more efficient and can support future source statements
within the conditional block.
Signed-off-by: Josua Mayer <josua@solid-run.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Add mux controller support for data or control lines that are muxed
between a host and multiple cards.
There are several devices supporting a choice of eMMC or SD on a single
board by both dip switch and gpio, e.g. Renesas RZ/G2L SMARC SoM and
SolidRun RZ/G2L SoM.
In-tree dts for the Renesas boards currently rely on preprocessor macros
and gpio hogs to describe the respective cards.
By adding mux-states property to sdhi controller description, boards can
correctly describe the mux that already exists in hardware - and drivers
can coordinate between mux selection and probing for cards.
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Josua Mayer <josua@solid-run.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Multiplexer subsystem has added generic helper functions for getting an
already selected mux-state object.
Replace existing logic in probe with the equivalent helper function.
There is a functional difference in that the mux is now automatically
deselected on release, replacing the explicit mux_state_deselect call.
This change is only compile-tested.
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Andreas Kemnade <andreas@kemnade.info>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Josua Mayer <josua@solid-run.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Multiplexer subsystem has now added helpers for getting managed optional
mux-state.
Switch to the new devm_mux_state_get_optional_selected helper.
This change is only compile-tested.
Signed-off-by: Josua Mayer <josua@solid-run.com>
Acked-by: Vinod Koul <vkoul@kernel.org>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
In-tree phy-can-transceiver and phy_rcar_gen3_usb2 have already
implemented local versions of devm_mux_state_get_optional.
The omap-i2c driver gets and selects an optional mux in its probe
function without using any helper.
Add new helper functions covering both aforementioned use-cases:
- mux_control_get_optional:
Get a mux-control if specified in dt, return NULL otherwise.
- devm_mux_state_get_optional:
Get a mux-state if specified in dt, return NULL otherwise.
- devm_mux_state_get_selected:
Get and select a mux-state specified in dt, return error otherwise.
- devm_mux_state_get_optional_selected:
Get and select a mux-state if specified in dt, return error or NULL.
Existing mux_get helper function is changed to take an extra argument
indicating whether the mux is optional.
In this case no error is printed, and NULL returned in case of ENOENT.
Calling code is adapted to handle NULL return case, and to pass optional
argument as required.
To support automatic deselect for _selected helper, a new structure is
created storing an exit pointer similar to clock core which is called on
release.
To facilitate code sharing between optional/mandatory/selected helpers,
a new internal helper function is added to handle quiet (optional) and
verbose (mandatory) errors, as well as storing the correct callback for
devm release: __devm_mux_state_get
Due to this structure devm_mux_state_get_*_selected can no longer print
a useful error message when select fails. Instead callers should print
errors where needed.
Commit e153fdea9d ("phy: can-transceiver: Re-instate "mux-states"
property presence check") noted that "mux_get() always prints an error
message in case of an error, including when the property is not present,
confusing the user."
The first error message covers the case that a mux name is not matched
in dt. The second error message is based on of_parse_phandle_with_args
return value.
In optional case no error is printed and NULL is returned.
This ensures that the new helper functions will not confuse the user
either.
With the addition of optional helper functions it became clear that
drivers should compile and link even if CONFIG_MULTIPLEXER was not enabled.
Add stubs for all symbols exported by mux core.
Acked-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Josua Mayer <josua@solid-run.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Merge the mmc fixes for v7.0-rc[n] into the next branch, to allow them to
get tested together with the mmc changes that are targeted for the next
release.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The GL9750 SD host controller has intermittent data corruption during
DMA write operations. The GM_BURST register's R_OSRC_Lmt field
(bits 17:16), which limits outstanding DMA read requests from system
memory, is not being cleared during initialization. The Windows driver
sets R_OSRC_Lmt to zero, limiting requests to the smallest unit.
Clear R_OSRC_Lmt to match the Windows driver behavior. This eliminates
write corruption verified with f3write/f3read tests while maintaining
DMA performance.
Cc: stable@vger.kernel.org
Fixes: e51df6ce66 ("mmc: host: sdhci-pci: Add Genesys Logic GL975x support")
Closes: https://lore.kernel.org/linux-mmc/33d12807-5c72-41ce-8679-57aa11831fad@linux.dev/
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Matthew Schwartz <matthew.schwartz@linux.dev>
Reviewed-by: Ben Chuang <ben.chuang@genesyslogic.com.tw>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The binding defines tuple-style reset-names items for some
compatibles, which implicitly enforces a fixed array length
via JSON Schema.
Defining global maxItems for resets and reset-names causes these
constraints to be intersected via allOf, resulting in an effective
minItems equal to the global maxItems. This leads to dtbs_check
failures reporting reset arrays as too short, even when the DTS
provides the correct number of entries.
Fixes: 30009a21f2 ("dt-bindings: mmc: sdhci-of-dwcmshc: Add Eswin EIC7700")
Co-developed-by: Pritesh Patel <pritesh.patel@einfochips.com>
Signed-off-by: Pritesh Patel <pritesh.patel@einfochips.com>
Signed-off-by: Huan He <hehuan1@eswincomputing.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Now that the Realtek PCIe SDMMC host implements the card_busy() callback,
the MMC core performs the required DAT[3:0] validation during the signal
voltage switch sequence.
As a result, the driver-side voltage stabilization helpers
sd_wait_voltage_stable_1() and sd_wait_voltage_stable_2() become redundant.
These helpers duplicate DAT line checks that are now handled centrally by
the MMC core via card_busy().
Remove the two stabilization helpers and simplify sdmmc_switch_voltage()
accordingly, keeping only the minimal clock control needed around the
voltage switch.
No functional changes and intended.
Signed-off-by: Ricky Wu <ricky_wu@realtek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Merge the mmc fixes for v7.0-rc[n] into the next branch, to allow them to
get tested together with the mmc changes that are targeted for the next
release.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
This driver currently only supports builds against a PIC32 target. Now
that commit d6618d277c ("mmc: sdhci-pic32: update include to use
pic32.h from platform_data") is merged, it's possible to compile this
driver on other architectures.
To avoid future breakage of this driver in the future, let's update the
Kconfig so that it can be built with COMPILE_TEST enabled on all
architectures.
Signed-off-by: Brian Masney <bmasney@redhat.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Add compatible strings for Axiado AX3000 SoC eMMC controller which
is based on Arasan eMMC controller.
Signed-off-by: SriNavmani A <srinavmani@axiado.com>
Signed-off-by: Tzu-Hao Wei <twei@axiado.com>
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Three MMC drivers have unnecessary module_init and module_exit functions
that are empty or just print a message. Remove them. Note that if a
module_init function exists, a module_exit function must also exist;
otherwise, the module cannot be unloaded.
Signed-off-by: Ethan Nelson-Moore <enelsonmoore@gmail.com>
Acked-by: Shawn Lin <shawn.lin@rock-chips.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Move claimed and retune control flags out of the bitfield word to
avoid unrelated RMW side effects in asynchronous contexts.
The host->claimed bit shared a word with retune flags. Writes to claimed
in __mmc_claim_host() or retune_now in mmc_mq_queue_rq() can overwrite
other bits when concurrent updates happen in other contexts, triggering
spurious WARN_ON(!host->claimed). Convert claimed, can_retune,
retune_now and retune_paused to bool to remove shared-word coupling.
Fixes: 6c0cedd1ef ("mmc: core: Introduce host claiming by context")
Fixes: 1e8e55b670 ("mmc: block: Add CQE support")
Cc: stable@vger.kernel.org
Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Penghe Geng <pgeng@nvidia.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Kingston eMMC IY2964 and IB2932 takes a fixed ~2 seconds for each secure
erase/trim operation regardless of size - that is, a single secure
erase/trim operation of 1MB takes the same time as 1GB. With default
calculated 3.5MB max discard size, secure erase 1GB requires ~300 separate
operations taking ~10 minutes total.
Add a card quirk, MMC_QUIRK_FIXED_SECURE_ERASE_TRIM_TIME, to set maximum
secure erase size for those devices. This allows 1GB secure erase to
complete in a single operation, reducing time from 10 minutes to just 2
seconds.
Signed-off-by: Luke Wang <ziniu.wang_1@nxp.com>
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Since the platform power management structure is now shared with the PCI
front-end, there is a link failure if only the PCI variant is enabled:
arm-linux-gnueabi/bin/arm-linux-gnueabi-ld: drivers/mmc/host/dw_mmc-pci.o:(.data+0x74): undefined reference to `dw_mci_pltfm_pmops'
This could be fixed by selecting the platform driver from the PCI one
in Kconfig, or by reverting the change to the PCI driver, but since this
is now used by all dw_mmc variants, just move the structure into the core
code as well.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Remove the long-deprecated mshc alias support, as the mmc core already
provides alias functionality through the standard mmc alias. This eliminates
the redundant dual-alias system. The driver now obtains the controller ID
from struct mmc_host::index(supplied by mmc alias) instead of the legacy mshc
alias.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Remove the long-deprecated mshc alias support, as the mmc core already
provides alias functionality through the standard mmc alias. This eliminates
the redundant dual-alias system. The driver now obtains the controller ID
from struct mmc_host::index(supplied by mmc alias) instead of the legacy mshc
alias.
dw_mci_hi6220_parse_dt() which parse mshc alias is used for hi6220, hi3660 and
hi3670. Given hi6220 never assigned mshc alias on the DTS files, and hi3360
and hi3670 share the same code, so with it removed, add a return value to
dw_mci_hs_set_timing() and let dw_mci_hi3660_init() check if index exceeds
TIMING_MODE in the firs place to bail out early, the same as before.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Add the MMC device tree binding and driver files to the existing
ARM/BST SOC SUPPORT maintainer entry.
Signed-off-by: Albert Yang <yangzh0906@thundersoft.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Add SDHCI controller driver for Black Sesame Technologies C1200 SoC.
This driver supports the DWCMSHC SDHCI controller with BST-specific
enhancements including:
- Custom clock management and tuning
- Power management support
- BST-specific register configurations
- Support for eMMC and SD card interfaces
- Hardware limitation workaround for 32-bit DMA addressing
The driver addresses specific hardware constraints where:
- System memory uses 64-bit bus, eMMC controller uses 32-bit bus
- eMMC controller cannot access memory through SMMU due to hardware bug
- All system DRAM is configured outside 4GB boundary (ZONE_DMA32)
- Uses SRAM-based bounce buffer within 32-bit address space
Signed-off-by: Ge Gordon <gordon.ge@bst.ai>
Signed-off-by: Albert Yang <yangzh0906@thundersoft.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Allow platform drivers to pre-allocate bounce buffer by checking if
host->bounce_buffer is already set before attempting allocation.
This enables platforms with specific DMA constraints (such as 32-bit
DMA on controllers that cannot access high memory) to use their own
reserved memory regions for the bounce buffer.
Suggested-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Albert Yang <yangzh0906@thundersoft.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Add device tree bindings for the Black Sesame Technologies
DWCMSHC SDHCI controller used in C1200 SoC.
The binding describes a Synopsys DesignWare Cores Mobile Storage
Host Controller with BST-specific extensions including:
- Two register regions (core SDHCI and CRM registers)
- Optional memory-region for bounce buffer support
- Fixed clock input
Signed-off-by: Ge Gordon <gordon.ge@bst.ai>
Signed-off-by: Albert Yang <yangzh0906@thundersoft.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The SDHCI controller found on SpacemiT K3 SoC share the same IP with K1
generation and introduce a compatible data to denote the change that broken
64BIT DMA issue has been fixed.
Signed-off-by: Yixun Lan <dlan@kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The SDHCI controller found on SpacemiT K3 SoC share the same IP with
K1 generation, while fixed the broken 64BIT DMA issue. Introduce a
compatible string to enable support for it.
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Yixun Lan <dlan@kernel.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
As much as we have grown used to seeing this message on
every kernel boot, it does not add any technical value.
Drop all messages from sdhci_drv_init(), and drop the
module_init() and module_exit() calls as well since they
now become empty.
The modules becomes a pure library module, meaning it will
get pulled in by modprobe() when the symbols inside it
are needed, or compiled in if any users are compiled in.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Add NXP's SDIO vendor ID (0x0471) and IW61x device ID (0x0205) to
sdio_ids.h for future support of NXP Wi-Fi chips over SDIO.
Signed-off-by: Jeff Chen <jeff.chen_1@nxp.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Use the modern PM macros for the suspend and resume functions to be
automatically dropped by the compiler when CONFIG_PM or CONFIG_PM_SLEEP
are disabled, without having to use #ifdef guards.
When this's done, remove all variant drivers' local pm definition and
replace it with dw_mci_pltfm_pmops.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
[Ulf: Fixed conflict by dropping changes for dw_mmc-rockchip]
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
PCIe endpoint framework defines standard BAR enum for EP drivers
to use, so remove PCI_BAR_NO here.
Then, pcim_iomap_regions has been deprecated for a long time, replace
it with pcim_iomap_region();
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The mci_readw/mci_writew/mci_readq/mci_writeq macros were added
to provide 16-bit and 64-bit register access operations, but they
have remained unused since their introduction. Remove these dead code.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The current dw_mci_get_cd() implementation maintains a DW_MMC_CARD_PRESENT
flag primarily for logging purposes, which adds unnecessary complexity.
Additionally, the if-else-elif control flow does not align with the Linux
kernel coding style.
This commit simplifies the function by:
- Removing the redundant card presence flag
- Replacing the conditional chain with a cleaner implementation
- Improving code readability while maintaining functionality
The change reduces code complexity without affecting the actual card
detection behavior.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The header file linux/mmc/core.h, which is already included in dw_mmc.h,
contains the forward declaration of struct mmc_data. There's no need to
redeclare it here.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The only user of dw_mci_board is dw_pci-pci, now we can provide
the caps from dw_mci_drv_data, so we could let dw_pci-pci use
dw_mci_drv_data and remove caps from struct dw_mci_board.
With that, struct dw_mci_board is no longer needed, we can remove
it. Then we should check all settings in dw_mci_parse_dt in order
not to overwrite them if provided from variant drivers.
Also, without CONFIG_OF support, dw_mmc doesn' work as host->pdata
is always ERR_PTR(-EINVAL), we could remove it together.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
struct dw_mci already keeps one, reuse it. Now, as dw_mmc-pci still
provide struct dw_mci_board, so host->bus_hz will not be overwritten.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Geert Uytterhoeven
Binbin Zhou
Felix Gu
Ulf Hansson
Josua Mayer
Matthew Schwartz
Huan He
Ricky Wu
Brian Masney
Shawn Lin
Fabio Estevam
SriNavmani A
Ethan Nelson-Moore
Penghe Geng
Luke Wang
Arnd Bergmann
Andre Korol
Albert Yang
Yixun Lan
Linus Walleij
Jeff Chen