When submitting an identical spi_message multiple times via spi_sync
the spi_message.frame_length does not get reset to 0 in __spi_validate
before adding up all spi_transfer.len resulting in
frame_length > actual_length on all but the first spi_sync call.

Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Signed-off-by: Mark Brown <broonie@kernel.org>

Handle EPROBE_DEFER explicitly so that we ensure that we get the DMA
channel specified in the device tree, instead of depending on the DMA
controller getting probed before us.

Signed-off-by: Rabin Vincent <rabin.vincent@axis.com>
Signed-off-by: Mark Brown <broonie@kernel.org>

With a 64-bit resource_size_t, we get a build warning on bcm63xx_spi_probe:

drivers/spi/spi-bcm63xx.c:565:16: warning: format '%x' expects argument of type 'unsigned int', but argument 3 has type 'resource_size_t {aka long long unsigned int}' [-Wformat=]

As we are printing a resource, we can just use the %pr format
specifier that pretty-prints the address and avoids the warning.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>

When only one device is present, it is not necessary to specify
cs_gpios, as the CS line can be controlled by the hardware
module.

Without this patch, older device tree bindings used before
37457607 "spi: mediatek: mt8173 spi multiple devices support"
would cause a panic on boot. This fixes the crash, and
re-introduces backward compatibility.

Signed-off-by: Nicolas Boichat <drinkcat@chromium.org>
Acked-by: Leilk Liu <leilk.liu@mediatek.com>
Signed-off-by: Mark Brown <broonie@kernel.org>

Commit ca5d24854210 ("spi: Add THIS_MODULE to spi_driver in SPI core")
adds the new __spi_register_driver() function, but keeps the kerneldoc
for the spi_register_driver() function in place and forgets to add the
description for the new owner parameter.

Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>