[PATCH] scsi: set dma direction to NONE for TEST UNIT READY

to return error. Fix that, by explicitly setting dma direction to NONE for TEST UNIT READY, and restoring it back DMA_FROM_DEVICE for the following READ CAPACITY. Signed-off-by: Nikita Yushchenko --- drivers/scsi/scsi.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/scsi/scsi.c b

Re: [PATCH 1/2] reset: introduce reset_get_bulk_optional()

+int reset_get_bulk_optional(struct udevice *dev, struct reset_ctl_bulk *bulk) +{ + int ret = reset_get_bulk(dev, bulk); + + if (ret == -ENODEV) Are you sure about this return type? I think it would be -ENODATA if I'm not wrong. It shall be -ENOENT on master. Oops.

[PATCH 2/2] spi: cadence_qspi: use reset_get_bulk_optional()

It is a valid setup to have no resets, e.g. TI J721E uses such setup. Accept it without error messages. And, error out if there is a real issue with getting resets. Signed-off-by: Nikita Yushchenko --- drivers/spi/cadence_qspi.c | 11 ++- 1 file changed, 6 insertions(+), 5 deletions

[PATCH 1/2] reset: introduce reset_get_bulk_optional()

This is a version of reset_get_bulk() that does not treat no resets definition as an error. Signed-off-by: Nikita Yushchenko --- drivers/reset/reset-uclass.c | 10 ++ include/reset.h | 21 + 2 files changed, 31 insertions(+) diff --git a/drivers/reset

[PATCH 0/2] add and use reset_get_bulk_optional()

Nikita Yushchenko (2): reset: introduce reset_get_bulk_optional() spi: cadence_qspi: use reset_get_bulk_optional() drivers/reset/reset-uclass.c | 10 ++ drivers/spi/cadence_qspi.c | 11 ++- include/reset.h | 21 + 3 files changed, 37

[PATCH] spi: spi-mem: fix always-true condition in spi_mem_exec_op()

Indeed fallback to generic implementation if driver's exec_op() returns -ENOTSUPP. Signed-off-by: Nikita Yushchenko --- drivers/spi/spi-mem.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/spi/spi-mem.c b/drivers/spi/spi-mem.c index 9c1ede1b61..7f9e65e870 100644

[PATCH] ti: i2c: fix probe_chip() return value

Per documentation, dm_i2c_ops.probe_chip() shall return -EREMOTEIO if probe fails. Currently, omap_i2c_probe_chip() returns 1 instead. Fix that. Signed-off-by: Nikita Yushchenko --- drivers/i2c/omap24xx_i2c.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/i2c

[PATCH] i2c: fix always-true condition in i2c_probe_chip()

Per dm_i2c_ops.probe_chip documentation, i2c_probe_chip() shall fallback to default probe method when .probe_chip() returns -ENOSYS. Signed-off-by: Nikita Yushchenko --- drivers/i2c/i2c-uclass.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/i2c/i2c-uclass.c b