linux/drivers/platform/chrome, branch v5.10

platform/chrome: Use kobj_to_dev() instead of container_of()

2020-09-29T08:08:44Z

Use kobj_to_dev() instead of container_of(). Signed-off-by: Wang Qing Reviewed-by: Guenter Roeck Signed-off-by: Enric Balletbo i Serra

platform/chrome: cros_ec_proto: Drop cros_ec_cmd_xfer()

2020-09-21T08:51:58Z

Since cros_ec_cmd_xfer_status() now returns Linux error codes and all other files use that command, remove the now-unused function cros_ec_cmd_xfer(). Signed-off-by: Prashant Malani Signed-off-by: Enric Balletbo i Serra

platform/chrome: cros_ec_proto: Update cros_ec_cmd_xfer() call-sites

2020-09-21T08:51:58Z

Since all the other call-sites of cros_ec_cmd_xfer() have been converted to use cros_ec_cmd_xfer_status() instead, update the remaining call-sites to prepare for the merge of cros_ec_cmd_xfer() into cros_ec_cmd_xfer_status(). As part of this update, change the error handling inside cros_ec_get_sensor_count() such that the legacy LPC interface is tried on all error values, not just when msg->result != EC_RESULT_SUCCESS. Note that there is a slight change in API in cros_ec_get_sensor_count(): it will return a negative number of sensors when there are no sensors on arm platform when MOTIONSENSE_CMD_DUMP is not supported (typical for sensorless chromebook) instead of 0. However, this is not a problem when probing the EC as we ignore errors only looking for cros_ec_get_sensor_count() returning a positive number of sensors. Signed-off-by: Prashant Malani Reviewed-by: Gwendal Grignou Tested-by: Gwendal Grignou Signed-off-by: Enric Balletbo i Serra

platform/chrome: Kconfig: Remove the transitional MFD_CROS_EC config

2020-09-04T07:30:04Z

The MFD_CROS_EC config was a transitional Kconfig option to not break current defconfigs in the kernel. Now, this is not required anymore because all the defconfigs have been removed this option and migrated to enable the CrOS EC parts individually. Signed-off-by: Enric Balletbo i Serra Reviewed-by: Guenter Roeck Tested-by: Gwendal Grignou

platform/chrome: cros_ec_lightbar: Reduce ligthbar get version command

2020-09-04T07:30:04Z

By default, the lightbar commands are set to the biggest lightbar command and response. That length is greater than 128 bytes and may not work on all machines. But all EC are probed for lightbar by sending a get version request. Set that request size precisely. Before the command would be: cros_ec_cmd: version: 0, command: EC_CMD_LIGHTBAR_CMD, outsize: 194, insize: 128, result: 0 Afer: cros_ec_cmd: version: 0, command: EC_CMD_LIGHTBAR_CMD, outsize: 1, insize: 8, result: 0 Fixes: a841178445bb7 ("mfd: cros_ec: Use a zero-length array for command data") Signed-off-by: Gwendal Grignou Signed-off-by: Enric Balletbo i Serra

platform/chrome: cros_ec_trace: Add fields to command traces

2020-08-31T10:14:45Z

In ftrace, add more fields to the cros_ec command event: - Add size of commands to check if they are properly set. - Add offset (in case an EC is cascaded being another EC), to allow proper command output With: echo 1 > events/cros_ec/cros_ec_cmd/enable We now have (on samus) Invalid command for the sensor stack: ectool-6942 [002] .... 3082.783116: cros_ec_request_done: version: 3, offset: 0, command: EC_CMD_MOTION_SENSE_CMD, outsize: 2, insize: 19, ec result: EC_RES_INVALID_PARAM, retval: 0 Powerd accessing PD EC being the main EC: powerd-1272 [002] .... 40.644026: cros_ec_request_done: version: 0, offset: 1, command: EC_CMD_USB_PD_POWER_INFO, outsize: 1, insize: 16, ec result: EC_RES_SUCCESS, retval: 16 Signed-off-by: Gwendal Grignou Acked-by: Raul E Rangel Signed-off-by: Enric Balletbo i Serra

platform/chrome: cros_ec_typec: Re-order connector configuration steps

2020-08-24T10:45:56Z

As per USB Type-C Spec R2.0 section 4.5.1.2 (Connecting Sources and Sinks) and section 4.5.2.2 (Connection State Machine Requirements), the typical flow for configuring a device connected to a typeC port is as below: 1. Source/sink detection 2. Orientation 3. Data role 4. VCONN 5. VBUS (USB Type-C currents) 6. The connector is now configured. We can start the PD communication that should lead into configuration of the mux if we enter a mode. But in existing code data role was set after the connector and mux are already configured. So fix this by following the spec to set the data role before the connector and mux are configured. Signed-off-by: Azhar Shaikh Reviewed-by: Prashant Malani Reviewed-by: Heikki Krogerus Signed-off-by: Enric Balletbo i Serra

platform/chrome: cros_ec_typec: Avoid setting usb role twice during disconnect

2020-08-24T10:45:56Z

On disconnect port partner is removed and usb role is set to NONE. But then in cros_typec_port_update() the role is set again. Avoid this by moving usb_role_switch_set_role() to cros_typec_configure_mux(). Suggested-by: Prashant Malani Signed-off-by: Azhar Shaikh Reviewed-by: Heikki Krogerus Signed-off-by: Enric Balletbo i Serra

platform/chrome: cros_ec_typec: Send enum values to usb_role_switch_set_role()

2020-08-24T10:45:56Z

usb_role_switch_set_role() has the second argument as enum for usb_role. Currently depending upon the data role i.e. UFP(0) or DFP(1) is sent. This eventually translates to USB_ROLE_NONE in case of UFP and USB_ROLE_DEVICE in case of DFP. Correct this by sending correct enum values as USB_ROLE_DEVICE in case of UFP and USB_ROLE_HOST in case of DFP. Fixes: 7e7def15fa4b ("platform/chrome: cros_ec_typec: Add USB mux control") Signed-off-by: Azhar Shaikh Cc: Prashant Malani Reviewed-by: Prashant Malani Reviewed-by: Heikki Krogerus Signed-off-by: Enric Balletbo i Serra

platform/chrome: cros_ec_typec: USB4 support

2020-08-24T09:37:34Z

With USB4 mode the mux driver needs the Enter_USB Data Object (EUDO) that was used when the USB mode was entered. Though the object is not available in the driver, it is possible to construct it from the information we have. Signed-off-by: Heikki Krogerus Signed-off-by: Prashant Malani Signed-off-by: Enric Balletbo i Serra