spi: amd: fix Wvoid-pointer-to-enum-cast warning

'version' is an enum, thus cast of pointer on 64-bit compile test with W=1
causes:

spi-amd.c:401:21: error: cast to smaller integer type 'enum amd_spi_versions' from 'const void *' [-Werror,-Wvoid-pointer-to-enum-cast]

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20230810091247.70149-3-krzysztof.kozlowski@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: switch to use modern name

Change legacy name master to modern name host or controller.

No functional changed.

Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20230728093221.3312026-2-yangyingliang@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: Convert to SPI_CONTROLLER_HALF_DUPLEX

Convert the users under SPI subsystem to SPI_CONTROLLER_HALF_DUPLEX.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20230710154932.68377-15-andriy.shevchenko@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: Replace all spi->chip_select and spi->cs_gpiod references with function call

Supporting multi-cs in spi drivers would require the chip_select & cs_gpiod
members of struct spi_device to be an array. But changing the type of these
members to array would break the spi driver functionality. To make the
transition smoother introduced four new APIs to get/set the
spi->chip_select & spi->cs_gpiod and replaced all spi->chip_select and
spi->cs_gpiod references with get or set API calls.
While adding multi-cs support in further patches the chip_select & cs_gpiod
members of the spi_device structure would be converted to arrays & the
"idx" parameter of the APIs would be used as array index i.e.,
spi->chip_select[idx] & spi->cs_gpiod[idx] respectively.

Signed-off-by: Amit Kumar Mahapatra <amit.kumar-mahapatra@amd.com>
Acked-by: Heiko Stuebner <heiko@sntech.de> # Rockchip drivers
Reviewed-by: Michal Simek <michal.simek@amd.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org> # Aspeed driver
Reviewed-by: Dhruva Gole <d-gole@ti.com> # SPI Cadence QSPI
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com> # spi-stm32-qspi
Acked-by: William Zhang <william.zhang@broadcom.com> # bcm63xx-hsspi driver
Reviewed-by: Serge Semin <fancer.lancer@gmail.com> # DW SSI part
Link: https://lore.kernel.org/r/167847070432.26.15076794204368669839@mailman-core.alsa-project.org
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Fix SPI_SPD7 value

According to data sheet SPI_SPD7 should be set to 7.

Signed-off-by: Vitaly Rodionov <vitalyr@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20221104100637.13376-1-vitalyr@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Fix speed selection

If the current speed is equal to the requested speed by the device
then return success.
This patch fixes a bug introduced by the commit 3fe26121dc3a
("spi: amd: Configure device speed") which checks speed_hz instead
of amd_spi->speed_hz.

Fixes: 3fe26121dc3a ("spi: amd: Configure device speed")
Signed-off-by: Shreeya Patel <shreeya.patel@collabora.com>
Link: https://lore.kernel.org/r/20220830093607.45484-1-shreeya.patel@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Configure device speed

Number of clock frequencies are supported by AMD controller
which are mentioned in the amd_spi_freq structure table.

Create mechanism to configure device clock frequency such
that it is strictly less than the requested frequency.

Give priority to the device transfer speed and in case
it is not set then use the max clock speed supported
by the device.

Signed-off-by: Lucas Tanure <tanureal@opensource.cirrus.com>
Co-developed-by: Shreeya Patel <shreeya.patel@collabora.com>
Signed-off-by: Shreeya Patel <shreeya.patel@collabora.com>
Link: https://lore.kernel.org/r/20220825143132.253224-1-shreeya.patel@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Setup all xfers before opcode execution

The AMD SPI controller hardware seems to expect the FIFO buffer to be
fully setup with the details of all transfers in the SPI message before
it is able to start processing the data in a reliable way.

Furthermore, it imposes a strict ordering restriction, in the sense that
all TX transfers must be handled prior any RX transfer.

Hence, let's ensure amd_spi_execute_opcode() is called only once, after
all TX transfers have been setup, and process any remaining RX transfers
afterwards, in a second iteration.

Additionally, get rid of the unnecessary AMD_SPI_XFER_TX/RX defines and
improve error handling.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20220818010059.403776-1-cristian.ciocaltea@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Add struct and enum kernel-doc comments

Provide documentation comments in the kernel-doc format
for enum amd_spi_versions and struct amd_spi.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20220706100626.1234731-6-cristian.ciocaltea@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Drop io_base_addr member from struct amd_spi

The io_base_addr member of struct amd_spi is not referenced anywhere
in the driver implementation and there is no indication that it could
be used in the future, hence drop it.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20220706100626.1234731-5-cristian.ciocaltea@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Make use of dev_err_probe()

Simplify the error handling in probe function by switching from
dev_err() to dev_err_probe().

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20220706100626.1234731-4-cristian.ciocaltea@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Make use of devm_spi_alloc_master()

Make use of the devm variant of spi_alloc_master() in order to cleanup
and simplify the error handling in the probe function by getting rid
of the goto statements.

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20220706100626.1234731-3-cristian.ciocaltea@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Limit max transfer and message size

Enabling the SPI CS35L41 audio codec driver for Steam Deck [1]
revealed a problem with the current AMD SPI controller driver
implementation, consisting of an unrecoverable system hang.

The issue can be prevented if we ensure the max transfer size
and the max message size do not exceed the FIFO buffer size.

According to the implementation of the downstream driver, the
AMD SPI controller is not able to handle more than 70 bytes per
transfer, which corresponds to the size of the FIFO buffer.

Hence, let's fix this by setting the SPI limits mentioned above.

[1] https://lore.kernel.org/r/20220621213819.262537-1-cristian.ciocaltea@collabora.com

Reported-by: Anastasios Vacharakis <vacharakis@o2mail.de>
Fixes: bbb336f39efc ("spi: spi-amd: Add AMD SPI controller driver support")
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20220706100626.1234731-2-cristian.ciocaltea@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Fix building without ACPI enabled

Commit 209043554915 ("spi: amd: Add support for version AMDI0062")
removed the cast ACPI_PTR() for no good reason. This wrapper is
important to make sure that the driver can be compiled with or without
CONFIG_ACPI enabled, useful for compiling test. Give back the cast so
compilation works again.

Fixes: 209043554915 ("spi: amd: Add support for version AMDI0062")
Signed-off-by: André Almeida <andrealmeid@collabora.com>
Link: https://lore.kernel.org/r/20220216162719.116062-1-andrealmeid@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Add support for version AMDI0062

Add support for the AMD SPI controller version AMDI0062. Do this in a
modular way where's easy to add new versions.

Signed-off-by: André Almeida <andrealmeid@collabora.com>
Link: https://lore.kernel.org/r/20220211143155.75513-4-andrealmeid@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Remove needless rom_addr variable

rom_addr is not used in the code, so we can just drop it from struct
amd_spi.

Signed-off-by: André Almeida <andrealmeid@collabora.com>
Link: https://lore.kernel.org/r/20220211143155.75513-3-andrealmeid@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Use iopoll for busy waiting

Instead of implementing a custom IO busy wait function, just use
readl_poll_timeout().

Signed-off-by: André Almeida <andrealmeid@collabora.com>
Link: https://lore.kernel.org/r/20220211143155.75513-2-andrealmeid@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Don't wait for a write-only transfer to finish

Return from a write-only transfer without waiting for
it to finish
But wait before a new transfer as the previous may
still happening and also wait before reading the data
from the FIFO

Signed-off-by: Lucas Tanure <tanureal@opensource.cirrus.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20210910111529.12539-4-tanureal@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Remove unneeded variable

Remove internal cs from amd_spi

Signed-off-by: Lucas Tanure <tanureal@opensource.cirrus.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20210910111529.12539-3-tanureal@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Refactor amd_spi_busy_wait

Use amd_spi_readreg32 to read 32 bits registers

Signed-off-by: Lucas Tanure <tanureal@opensource.cirrus.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20210910111529.12539-2-tanureal@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Refactor code to use less spi_master_get_devdata

Get master data in the start and then just use struct amd_spi
as it has the needed variable

Signed-off-by: Lucas Tanure <tanureal@opensource.cirrus.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20210910111529.12539-1-tanureal@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Use devm_platform_ioremap_resource() in amd_spi_probe

Simplify this function implementation by using a known wrapper function.

Signed-off-by: Qing Zhang <zhangqing@loongson.cn>
Link: https://lore.kernel.org/r/1605930231-19448-1-git-send-email-zhangqing@loongson.cn
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: spi-amd: Do not define 'struct acpi_device_id' when !CONFIG_ACPI

Since ACPI_PTR() is used to NULLify the value when !CONFIG_ACPI,
struct 'spi_acpi_match' becomes defined but unused.

Fixes the following W=1 kernel build warning(s):

drivers/spi/spi-amd.c:297:36: warning: ‘spi_acpi_match’ defined but not used [-Wunused-const-variable=]
297 | static const struct acpi_device_id spi_acpi_match[] = {
| ^~~~~~~~~~~~~~

Signed-off-by: Lee Jones <lee.jones@linaro.org>
Cc: Sanjay R Mehta <sanju.mehta@amd.com>
Link: https://lore.kernel.org/r/20200717135424.2442271-15-lee.jones@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Drop superfluous member from struct amd_spi

The AMD SPI driver stores a pointer to the spi_master in struct amd_spi
so that it can get from the latter to the former in amd_spi_fifo_xfer().

It's simpler to just pass the pointer from the sole caller
amd_spi_master_transfer() and drop the pointer from struct amd_spi.

Signed-off-by: Lukas Wunner <lukas@wunner.de>
Link: https://lore.kernel.org/r/a088b684ad292faf3bd036e51529e608e5c94638.1588590210.git.lukas@wunner.de
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Fix refcount underflow on remove

The AMD SPI driver calls spi_master_put() in its ->remove() hook even
though the preceding call to spi_unregister_master() already drops a
ref, thus leading to a refcount underflow. Drop the superfluous call
to spi_master_put().

This only leaves the call to spi_unregister_master() in the ->remove()
hook, so it's safe to change the ->probe() hook to use the devm version
of spi_register_master() and drop the ->remove() hook altogether.

Signed-off-by: Lukas Wunner <lukas@wunner.de>
Link: https://lore.kernel.org/r/5e53ccdf1eecd4e015dba99d0d77389107f8a2e3.1588590210.git.lukas@wunner.de
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Drop duplicate driver data assignments

The AMD SPI driver calls platform_set_drvdata() on probe even though
it's already been set by __spi_alloc_controller(). Likewise, it calls
platform_set_drvdata() on remove even though it's going to be set by
__device_release_driver(). Drop the duplicate assignments.

Signed-off-by: Lukas Wunner <lukas@wunner.de>
Link: https://lore.kernel.org/r/499f8ad4759c2ff0f586e0459fb9a293faecff6d.1588590210.git.lukas@wunner.de
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Pass probe errors back to driver core

If probing fails, the AMD SPI driver pretends success to the driver core
by returning 0. Return the errno instead.

Signed-off-by: Lukas Wunner <lukas@wunner.de>
Link: https://lore.kernel.org/r/689f29a359718dab4f5de9ee66c02ea97b3bd9e8.1588590210.git.lukas@wunner.de
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: amd: Fix duplicate iounmap in error path

The AMD SPI driver uses devm_ioremap_resource() to map its registers, so
they're automatically unmapped via device_release() when the last ref on
the SPI controller is dropped. The additional iounmap() in the ->probe()
error path is thus unnecessary.

Signed-off-by: Lukas Wunner <lukas@wunner.de>
Link: https://lore.kernel.org/r/497cc38ae2beb7900ae05a1463eb83ff96e2770e.1588590210.git.lukas@wunner.de
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: spi-amd: Fix a NULL vs IS_ERR() check in amd_spi_probe()

In case of error, the function devm_ioremap_resource() returns ERR_PTR()
and never returns NULL. The NULL test in the return value check should
be replaced with IS_ERR().

Fixes: bbb336f39efc ("spi: spi-amd: Add AMD SPI controller driver support")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Link: https://lore.kernel.org/r/20200429025426.167664-1-weiyongjun1@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: spi-amd: fix warning

remove unused variable "opcode"

Signed-off-by: Sanjay R Mehta <sanju.mehta@amd.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/1588049801-37995-1-git-send-email-sanju.mehta@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: spi-amd: Add AMD SPI controller driver support

This driver supports SPI Controller for AMD SOCs.This driver
supports SPI operations using FIFO mode of transfer.

Signed-off-by: Sanjay R Mehta <sanju.mehta@amd.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/1587844788-33997-1-git-send-email-sanju.mehta@amd.com
Signed-off-by: Mark Brown <broonie@kernel.org>