linux-stable.git/drivers/input/rmi4, branch v6.2.2

Input: rmi_smbus - Convert to i2c's .probe_new()

2022-12-02T22:53:43+00:00

The probe function doesn't make use of the i2c_device_id * parameter so it
can be trivially converted.

Signed-off-by: Uwe Kleine-König 
Link: https://lore.kernel.org/r/20221118224540.619276-225-uwe@kleine-koenig.org
Signed-off-by: Dmitry Torokhov

Input: rmi_i2c - Convert to i2c's .probe_new()

2022-12-02T22:53:43+00:00

The probe function doesn't make use of the i2c_device_id * parameter so it
can be trivially converted.

Signed-off-by: Uwe Kleine-König 
Link: https://lore.kernel.org/r/20221118224540.619276-224-uwe@kleine-koenig.org
Signed-off-by: Dmitry Torokhov

Input: synaptics-rmi4 - convert to use sysfs_emit() APIs

2022-09-27T16:00:41+00:00

Follow the advice of the Documentation/filesystems/sysfs.rst and show()
should only use sysfs_emit() or sysfs_emit_at() when formatting the value
to be returned to user space.

Reported-by: Zeal Robot 
Signed-off-by: zhang songyi 
Link: https://lore.kernel.org/r/20220927070936.258300-1-zhang.songyi@zte.com.cn
Signed-off-by: Dmitry Torokhov

Input: synaptics-rmi4 - drop useless gotos in rmi_f34v7_do_reflash()

2022-09-23T20:00:14+00:00

Returning directly makes the code clearer.

Signed-off-by: Matthias Schiffer 
Reviewed-by: Lyude Paul 
Link: https://lore.kernel.org/r/20220608124808.51402-10-matthias.schiffer@ew.tq-group.com
Signed-off-by: Dmitry Torokhov

Input: synaptics-rmi4 - simplify rmi_f34v7_start_reflash()

2022-09-23T20:00:13+00:00

rmi_f34v7_enter_flash_prog() already enables IRQs and checks the flash
status - there's no need for rmi_f34v7_start_reflash() to do the same just
before calling rmi_f34v7_enter_flash_prog().

Signed-off-by: Matthias Schiffer 
Reviewed-by: Lyude Paul 
Link: https://lore.kernel.org/r/20220608124808.51402-9-matthias.schiffer@ew.tq-group.com
Signed-off-by: Dmitry Torokhov

Input: synaptics-rmi4 - remove unneeded struct register_offset

2022-09-23T20:00:13+00:00

All register offsets are fixed, and a number of places even read or
write multiple registers as a block, so there is no way to support
reordering them without move involved changes. Remove the unneeded level
of indirection in the register access.

Signed-off-by: Matthias Schiffer 
Reviewed-by: Lyude Paul 
Link: https://lore.kernel.org/r/20220608124808.51402-8-matthias.schiffer@ew.tq-group.com
Signed-off-by: Dmitry Torokhov

Input: synaptics-rmi4 - make rmi_f34v7_erase_all() use the "erase all" command

2022-09-23T20:00:13+00:00

A full erase is required to recover from error conditions like "Bad
Partition Table". Various individual partition erase commands can be
(and need to be) omitted, as they will fail until a new partition table
has been written.

Signed-off-by: Matthias Schiffer 
Reviewed-by: Lyude Paul 
Link: https://lore.kernel.org/r/20220608124808.51402-7-matthias.schiffer@ew.tq-group.com
Signed-off-by: Dmitry Torokhov

Input: synaptics-rmi4 - reset after writing partition table

2022-09-23T20:00:13+00:00

When recovering from a bad partition table (for example after an
interrupted update), a reset is necessary for the new partition table to
become effective. Without this reset, writing the core code partition
will fail with status 0x03 (Invalid Command).

Signed-off-by: Matthias Schiffer 
Reviewed-by: Lyude Paul 
Link: https://lore.kernel.org/r/20220608124808.51402-6-matthias.schiffer@ew.tq-group.com
Signed-off-by: Dmitry Torokhov

Input: synaptics-rmi4 - rewrite partition table unconditionally

2022-09-23T20:00:13+00:00

Preparation for use of the "erase application" command, which is required
to recover from a bad partition table error condition. Rather than adding
complex fallback error paths for such errors, it seems more robust to do
the full erase unconditionally.

Signed-off-by: Matthias Schiffer 
Reviewed-by: Lyude Paul 
Link: https://lore.kernel.org/r/20220608124808.51402-5-matthias.schiffer@ew.tq-group.com
Signed-off-by: Dmitry Torokhov

Input: synaptics-rmi4 - fix command completion check for bootloader v7/v8

2022-09-23T20:00:13+00:00

The command register is reset to 0 when a command has completed. Check
for this condition instead of the error status, which will not accurately
reflect completion. In particular, the incorrect condition caused every
command error to be reported as a timeout.

Signed-off-by: Matthias Schiffer 
Reviewed-by: Lyude Paul 
Link: https://lore.kernel.org/r/20220608124808.51402-4-matthias.schiffer@ew.tq-group.com
Signed-off-by: Dmitry Torokhov