Re: [PATCH v8 1/3] platform/chrome: cros_ec_uart: Add cros-ec-uart transport layer

From: Prashant Malani
Date: Thu Dec 01 2022 - 16:12:35 EST


Hi,

On Wed, Nov 30, 2022 at 12:13 PM Mark Hasemeyer <markhas@xxxxxxxxxxxx> wrote:
>
> This patch does following:
> 1. Adds a new cros-ec-uart driver. This driver can send EC requests on
> UART and process response packets received on UART transport.
> 2. Once probed, this driver will initialize the serdev device based on
> the underlying information in the ACPI resource. After serdev device
> properties are set, this driver will register itself cros-ec.
> 3. High level driver can use this implementation to talk to ChromeOS
> Embedded Controller device in case it supports UART as transport.
> 4. When cros-ec driver initiates a request packet, outgoing message is
> processed in buffer and sent via serdev. Once bytes are sent, driver
> enables a wait_queue.
> 5. Since ChromeOS EC device sends response asynchronously, AP's TTY
> driver accumulates response bytes and calls the registered callback.
> TTY driver can send multiple callback for bytes ranging from 1 to MAX
> bytes supported by EC device.
> 6. Driver waits for EC_MSG_DEADLINE_MS to collect and process received
> bytes. It wakes wait_queue if expected bytes are received or else
> wait_queue timeout. Based on the error condition, driver returns
> data_len or error to cros_ec.
>
> Signed-off-by: Bhanu Prakash Maiya <bhanumaiya@xxxxxxxxxxxx>
> Signed-off-by: Mark Hasemeyer <markhas@xxxxxxxxxxxx>
> ---
>
> Changes in v8:
> - Combine for loops used for xsum calculation
> - Use ec_dev->dev instead of serdev->dev
> - Sort includes
> - Swap struct var order to match doc order
> - fix whitespace
> - Update contact information in MAINTAINERS
> - Update copyright date
>
> Changes in v7:
> - Use ACPI ID matching
> - Change MODULE_LICENSE from GPL v2 to GPL
>
> Changes in v6:
> - Tuning EC_MSG_DEADLINE_MS from 1 sec to 500 msec. Since the EC's UART
> response buffer was tuned and redeuced to 256 byte, 500 msec timeout
> should give enough time for 256 bytes to be recieved from SoC UART
> buffer.
>
> Changes in v5:
> - Fixed checkpatch.pl warnings.
>
> Changes in v4:
> - Change EC_MSG_DEADLINE_MS from 3 secs to 1 sec. 3 secs timeout prevents
> driver from intercepting EC in RO stage. In case of request failure, 1
> sec allows driver to retry and send HC when EC is in RO.
> - Add more information in commit message.
>
> Changes in v3:
> - checkpatch.pl script warns about char len 80 even though we have
> relaxed view guideline for line below 100 chars. Currently sticking
> with 80 chars in v3.
> - Fixed style issues
>
> Changes in v2:
> - Fixed build error on v1.
> - Changed EC timeout for response packet to 3 Sec and added comments.
> - Fixed cros_ec_uart_rx_bytes function to handle rx buffer < size of response header.
>
> ---
> MAINTAINERS | 7 +
> drivers/platform/chrome/Kconfig | 10 +
> drivers/platform/chrome/Makefile | 1 +
> drivers/platform/chrome/cros_ec_uart.c | 411 +++++++++++++++++++++++++
> 4 files changed, 429 insertions(+)
> create mode 100644 drivers/platform/chrome/cros_ec_uart.c
>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 2585e7edc3358..8ed08316fd11e 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -4918,6 +4918,13 @@ S: Maintained
> F: Documentation/devicetree/bindings/sound/google,cros-ec-codec.yaml
> F: sound/soc/codecs/cros_ec_codec.*
>
> +CHROMEOS EC UART DRIVER
> +M: Bhanu Prakash Maiya <bhanumaiya@xxxxxxxxxxxx>
> +R: Benson Leung <bleung@xxxxxxxxxx>
> +R: Tzung-Bi Shih <tzungbi@xxxxxxxxxx>
> +S: Maintained
> +F: drivers/platform/chrome/cros_ec_uart.c
> +
> CHROMEOS EC SUBDRIVERS
> M: Benson Leung <bleung@xxxxxxxxxxxx>
> R: Guenter Roeck <groeck@xxxxxxxxxxxx>
> diff --git a/drivers/platform/chrome/Kconfig b/drivers/platform/chrome/Kconfig
> index 6b954c5acadb1..d4a95ec261ed3 100644
> --- a/drivers/platform/chrome/Kconfig
> +++ b/drivers/platform/chrome/Kconfig
> @@ -119,6 +119,16 @@ config CROS_EC_SPI
> response time cannot be guaranteed, we support ignoring
> 'pre-amble' bytes before the response actually starts.
>
> +config CROS_EC_UART
> + tristate "ChromeOS Embedded Controller (UART)"
> + depends on CROS_EC && ACPI && SERIAL_DEV_BUS
> + help
> + If you say Y here, you get support for talking to the ChromeOS EC
> + through a UART, using a byte-level protocol.
> +
> + To compile this driver as a module, choose M here: the
> + module will be called cros_ec_uart.
> +
> config CROS_EC_LPC
> tristate "ChromeOS Embedded Controller (LPC)"
> depends on CROS_EC && ACPI && (X86 || COMPILE_TEST)
> diff --git a/drivers/platform/chrome/Makefile b/drivers/platform/chrome/Makefile
> index 2950610101f15..6f5b6f01940db 100644
> --- a/drivers/platform/chrome/Makefile
> +++ b/drivers/platform/chrome/Makefile
> @@ -15,6 +15,7 @@ obj-$(CONFIG_CROS_EC_ISHTP) += cros_ec_ishtp.o
> obj-$(CONFIG_CROS_TYPEC_SWITCH) += cros_typec_switch.o
> obj-$(CONFIG_CROS_EC_RPMSG) += cros_ec_rpmsg.o
> obj-$(CONFIG_CROS_EC_SPI) += cros_ec_spi.o
> +obj-$(CONFIG_CROS_EC_UART) += cros_ec_uart.o
> cros_ec_lpcs-objs := cros_ec_lpc.o cros_ec_lpc_mec.o
> obj-$(CONFIG_CROS_EC_TYPEC) += cros_ec_typec.o
> obj-$(CONFIG_CROS_EC_LPC) += cros_ec_lpcs.o
> diff --git a/drivers/platform/chrome/cros_ec_uart.c b/drivers/platform/chrome/cros_ec_uart.c
> new file mode 100644
> index 0000000000000..ef9ac7ace04b6
> --- /dev/null
> +++ b/drivers/platform/chrome/cros_ec_uart.c
> @@ -0,0 +1,411 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * UART interface for ChromeOS Embedded Controller
> + *
> + * Copyright 2020-2022 Google LLC.
> + */
> +
> +#include <linux/acpi.h>
> +#include <linux/delay.h>
> +#include <linux/errno.h>
> +#include <linux/init.h>
> +#include <linux/kernel.h>
> +#include <linux/module.h>
> +#include <linux/platform_data/cros_ec_commands.h>
> +#include <linux/platform_data/cros_ec_proto.h>
> +#include <linux/serdev.h>
> +#include <linux/slab.h>
> +#include <uapi/linux/sched/types.h>
> +
> +#include "cros_ec.h"
> +
> +/*
> + * EC sends contiguous bytes of response packet on UART AP RX.
> + * TTY driver in AP accumulates incoming bytes and calls the registered callback
> + * function. Byte count can range from 1 to MAX bytes supported by EC.
> + * This driver should wait for long time for all callbacks to be processed.
> + * Considering the worst case scenario, wait for 500 msec. This timeout should
> + * account for max latency and some additional guard time.
> + * Best case: Entire packet is received in ~200 ms, wait queue will be released
> + * and packet will be processed.
> + * Worst case: TTY driver sends bytes in multiple callbacks. In this case this
> + * driver will wait for ~1 sec beyond which it will timeout.
> + * This timeout value should not exceed ~500 msec because in case if
> + * EC_CMD_REBOOT_EC sent, high level driver should be able to intercept EC
> + * in RO.
> + */
> +#define EC_MSG_DEADLINE_MS 500
> +
> +/**
> + * struct response_info - Encapsulate EC response related
> + * information for passing between function
> + * cros_ec_uart_pkt_xfer() and cros_ec_uart_rx_bytes()
> + * callback.
> + * @data: Copy the data received from EC here.
> + * @max_size: Max size allocated for the @data buffer. If the
> + * received data exceeds this value, we log an error.
> + * @size: Actual size of data received from EC. This is also
> + * used to accumulate byte count with response is received
> + * in dma chunks.
> + * @exp_len: Expected bytes of response from EC including header.
> + * @error: 0 for success, negative error code for a failure.
> + * @received: Set to true on receiving a valid EC response.
> + * @wait_queue: Wait queue EC response where the cros_ec sends request
> + * to EC and waits
> + */
> +struct response_info {
> + void *data;
> + size_t max_size;
> + size_t size;
> + size_t exp_len;
> + int error;
> + bool received;
> + wait_queue_head_t wait_queue;
> +};
> +
> +/**
> + * struct cros_ec_uart - information about a uart-connected EC
> + *
> + * @serdev_device: serdev uart device we are connected to.
> + * @baudrate: UART baudrate of attached EC device.
> + * @flowcontrol: UART flowcontrol of attached device.
> + * @irq: Linux IRQ number of associated serial device.
> + * @response: Response info passing between cros_ec_uart_pkt_xfer()
> + * and cros_ec_uart_rx_bytes()
> + */
> +struct cros_ec_uart {
> + struct serdev_device *serdev;
> + u32 baudrate;
> + u8 flowcontrol;
> + u32 irq;
> + struct response_info response;
> +};
> +
> +static int cros_ec_uart_rx_bytes(struct serdev_device *serdev,
> + const u8 *data,
> + size_t count)
> +{
> + struct ec_host_response *response;
> + struct cros_ec_device *ec_dev = serdev_device_get_drvdata(serdev);
> + struct cros_ec_uart *ec_uart = ec_dev->priv;
> +
> + /* Check if bytes were sent out of band */
> + if (!ec_uart->response.data)
> + /* Discard all bytes */
> + return count;
> +
> + /*
> + * Check if incoming bytes + response.size are less than allocated
> + * buffer in din by cros_ec. This will ensure that if EC sends more
> + * bytes than max_size, waiting process will be notified with an error.
> + */

Please try to avoid over-commenting (here and elsewhere in the file)
[1]; it shouldn't
be necessary to comment every block of code, where the intention can be surmised
purely by looking at the code.

> + if (ec_uart->response.size + count <= ec_uart->response.max_size) {
> + /* Copy bytes in data in buffer */
This comment seems unnecessary.
> + memcpy((void *)ec_uart->response.data + ec_uart->response.size,
> + (void *)data, count);
> +
> + /* Add incoming bytes in size */
> + ec_uart->response.size += count;
> +
> + /*
> + * Read data_len if we received response header and if exp_len
> + * was not read before.
> + */
> + if (ec_uart->response.size >= sizeof(*response) &&
> + ec_uart->response.exp_len == 0) {
> + /* Get expected response length from response header */
> + response = (struct ec_host_response *)
> + ec_uart->response.data;
> +
> + ec_uart->response.exp_len = response->data_len +
> + sizeof(*response);
> + }
> +
> + /*
> + * If driver received response header and payload from EC,
> + * Wake up the wait queue.
> + */
> + if (ec_uart->response.size >= sizeof(*response) &&
> + ec_uart->response.size == ec_uart->response.exp_len) {
> + /* Set flag before waking up the caller */
> + ec_uart->response.received = true;
> +
> + /* Wake the calling thread */
> + wake_up_interruptible(&ec_uart->response.wait_queue);
> + }

You can save 1 level of indentation by inverting the if block and making it a
guard clause:
if (ec_uart->response_size + count > ec_uart->response_max_size) {
ec_uart->response.error = -EMSGSIZE;
wake_up_interruptible(...);
return;
}
memcpy(...);
ec_uart->response.size += count;

> + } else {
> + /* Received bytes are more the allocated buffer*/
> + ec_uart->response.error = -EMSGSIZE;
> +
> + /* Wake the calling thread */

Same issue as earlier. There is no need to add a comment for single lines
like these where what is happening is obvious.

> + wake_up_interruptible(&ec_uart->response.wait_queue);
> + }
> +
> + return count;
> +}
> +
> +static int cros_ec_uart_pkt_xfer(struct cros_ec_device *ec_dev,
> + struct cros_ec_command *ec_msg)
> +{
> + struct cros_ec_uart *ec_uart = ec_dev->priv;
> + struct serdev_device *serdev = ec_uart->serdev;
> + struct ec_host_response *response;
> + unsigned int len;
> + int ret, i;
> + u8 sum = 0;
> +
> + /* Prepare an outgoing message in the output buffer */

Superfluous comment.

> + len = cros_ec_prepare_tx(ec_dev, ec_msg);
> + dev_dbg(ec_dev->dev, "Prepared len=%d\n", len);
> +
> + /* Setup for incoming response */
> + ec_uart->response.data = ec_dev->din;
> + ec_uart->response.max_size = ec_dev->din_size;
> + ec_uart->response.size = 0;
> + ec_uart->response.error = 0;
> + ec_uart->response.exp_len = 0;
> + ec_uart->response.received = false;
> +
> + /* Write serial device buffer */
Superfluous comment.
> + ret = serdev_device_write_buf(serdev, ec_dev->dout, len);
> + if (ret < len) {
> + dev_err(ec_dev->dev, "Unable to write data");
> +
> + /* Return EIO as controller had issues writing buffer */
> + ret = -EIO;
> + goto exit;
> + }
> +
> + /* Once request is successfully sent to EC, wait to wait_queue */
> + wait_event_interruptible_timeout(ec_uart->response.wait_queue,
> + ec_uart->response.received,
> + msecs_to_jiffies(EC_MSG_DEADLINE_MS));
wait_event_interruptible_timeout() returns a value indicating if a
timeout occurred
and/or if the condition evaluated to true[3]. Can that not be used here, instead
of re-writing the timeout/error check mechanism?

> +
> + /* Check if wait_queue was interrupted due to an error */
> + if (ec_uart->response.error < 0) {
> + dev_warn(ec_dev->dev, "Response error detected.\n");
> +
> + ret = ec_uart->response.error;
> + goto exit;
> + }
> +
> + /* Check if valid response was received or there was a timeout */
> + if (!ec_uart->response.received) {
> + dev_warn(ec_dev->dev, "EC failed to respond in time.\n");
> +
> + ret = -ETIMEDOUT;
> + goto exit;
> + }
> +
> + /* Check response error code */
> + response = (struct ec_host_response *)ec_dev->din;
> + ec_msg->result = response->result;
> +
> + /* Check if received response is longer than expected */
> + if (response->data_len > ec_msg->insize) {
> + dev_err(ec_dev->dev, "Resp too long (%d bytes, expected %d)",
> + response->data_len,
> + ec_msg->insize);
> + ret = -ENOSPC;
> + goto exit;
> + }
> +
> + /* Validate checksum */
> + for (i = 0; i < sizeof(*response) + response->data_len; i++)
> + sum += ec_dev->din[i];
> +
> + if (sum) {
> + dev_err(ec_dev->dev,
> + "Bad packet checksum calculated %x\n",
> + sum);
Line length allowances are 100 chars now [2]. I think this should fit on 1 line.
> + ret = -EBADMSG;
> + goto exit;
> + }
> +
> + /* Copy response packet to ec_msg data buffer */
Superfluous comment.
> + memcpy(ec_msg->data,
> + ec_dev->din + sizeof(*response),
> + response->data_len);
> +
> + /* Return data_len to cros_ec */
> + ret = response->data_len;
> +
> +exit:
> + /* Reset ec_uart */
> + ec_uart->response.data = NULL;
> + ec_uart->response.max_size = 0;
> + ec_uart->response.size = 0;
> + ec_uart->response.error = 0;
> + ec_uart->response.exp_len = 0;
> + ec_uart->response.received = false;
Sorry if I'm missing something, but are you expecting some other access to these
variables (before the next call to cros_ec_uart_pkt_xfer()) ?

If not, we shouldn't need to reset everything here (re-initializing it
on the next
call to cros_ec_uart_pkt_xfer() should be sufficient)

> +
> + if (ec_msg->command == EC_CMD_REBOOT_EC)
> + msleep(EC_REBOOT_DELAY_MS);
> +
> + return ret;
> +}
> +
> +static int cros_ec_uart_resource(struct acpi_resource *ares, void *data)
> +{
> + struct cros_ec_uart *ec_uart = data;
> + struct acpi_resource_uart_serialbus *sb;
> +
> + switch (ares->type) {
> + case ACPI_RESOURCE_TYPE_SERIAL_BUS:
> + sb = &ares->data.uart_serial_bus;
> + if (sb->type == ACPI_RESOURCE_SERIAL_TYPE_UART) {
> + ec_uart->baudrate = sb->default_baud_rate;
> + dev_dbg(&ec_uart->serdev->dev, "Baudrate %d\n",
> + ec_uart->baudrate);
> +
> + ec_uart->flowcontrol = sb->flow_control;
> + dev_dbg(&ec_uart->serdev->dev, "Flow control %d\n",
> + ec_uart->flowcontrol);
> + }
> + break;
> + default:
> + break;
> + }
Is a switch necessary here? An if statement could accomplish the same thing
struct cros_ec_uart *ec_uart = data;
struct acpi_resource_uart_serialbus *sb = &ares->data.uart_serial_bus;
if (ares->type == ACPI_RESOURCE_TYPEC_SERIAL_BUS &&
sb->type == ACPI_RESOURCE_SERIAL_TYPE_UART) {
ec_uart->baudrate = sb->default_baud_rate;
....
}
> +
> + return 0;
> +}
> +
> +static int cros_ec_uart_acpi_probe(struct cros_ec_uart *ec_uart)
> +{
> + LIST_HEAD(resources);
> + struct acpi_device *adev = ACPI_COMPANION(&ec_uart->serdev->dev);
> + int ret;
It's not a strict requirement, but it's always nice to list the
initializer list in decreasing (or
increasing) order of line length;
> +
> + /* Retrieve UART ACPI info */
Unnecessary comment.
> + ret = acpi_dev_get_resources(adev, &resources,
> + cros_ec_uart_resource, ec_uart);
> + if (ret < 0)
> + return ret;
> +
> + acpi_dev_free_resource_list(&resources);
> +
> + /* Retrieve GpioInt and translate it to Linux IRQ number */
> + ret = acpi_dev_gpio_irq_get(adev, 0);
> + if (ret < 0)
> + return ret;
> +
> + ec_uart->irq = ret;
> + dev_dbg(&ec_uart->serdev->dev, "IRQ number %d\n", ec_uart->irq);
> +
> + return 0;
> +}
> +
> +static const struct serdev_device_ops cros_ec_uart_client_ops = {
> + .receive_buf = cros_ec_uart_rx_bytes,
> +};
> +
> +static int cros_ec_uart_probe(struct serdev_device *serdev)
> +{
> + struct device *dev = &serdev->dev;
> + struct cros_ec_device *ec_dev;
> + struct cros_ec_uart *ec_uart;
> + int ret;
> +
> + ec_uart = devm_kzalloc(dev, sizeof(*ec_uart), GFP_KERNEL);
> + if (!ec_uart)
> + return -ENOMEM;
> +
> + ec_dev = devm_kzalloc(dev, sizeof(*ec_dev), GFP_KERNEL);
> + if (!ec_dev)
> + return -ENOMEM;
> +
> + ec_uart->serdev = serdev;
> +
> + /* Open the serial device */
> + ret = devm_serdev_device_open(dev, ec_uart->serdev);
> + if (ret) {
> + dev_err(dev, "Unable to open UART device");
> + return ret;
> + }
> +
> + serdev_device_set_drvdata(serdev, ec_dev);
> +
Unnecessary newline.
> + serdev_device_set_client_ops(serdev, &cros_ec_uart_client_ops);
> +
> + /* Initialize wait queue */
> + init_waitqueue_head(&ec_uart->response.wait_queue);
> +
> + ret = cros_ec_uart_acpi_probe(ec_uart);
> + if (ret < 0) {
> + dev_err(dev, "Failed to get ACPI info (%d)", ret);
> + return ret;
> + }
> +
> + /* Set baud rate of serial device */
> + ret = serdev_device_set_baudrate(serdev, ec_uart->baudrate);
> + if (ret < 0) {
> + dev_err(dev, "Failed to set up host baud rate (%d)", ret);
> + return ret;
> + }
> +
> + /* Set flow control of serial device */
> + serdev_device_set_flow_control(serdev, ec_uart->flowcontrol);
> +
> + /* Initialize ec_dev for cros_ec */
> + ec_dev->phys_name = dev_name(dev);
> + ec_dev->dev = dev;
> + ec_dev->priv = ec_uart;
> + ec_dev->irq = ec_uart->irq;
> + ec_dev->cmd_xfer = NULL;
> + ec_dev->pkt_xfer = cros_ec_uart_pkt_xfer;
> + ec_dev->din_size = sizeof(struct ec_host_response) +
> + sizeof(struct ec_response_get_protocol_info);
> + ec_dev->dout_size = sizeof(struct ec_host_request);
> +
> + /* Register a new cros_ec device */
> + return cros_ec_register(ec_dev);
> +}
> +
> +static void cros_ec_uart_remove(struct serdev_device *serdev)
> +{
> + struct cros_ec_device *ec_dev = serdev_device_get_drvdata(serdev);
> +
> + cros_ec_unregister(ec_dev);
> +};
> +
> +static int __maybe_unused cros_ec_uart_suspend(struct device *dev)
> +{
> + struct cros_ec_device *ec_dev = dev_get_drvdata(dev);
> +
> + return cros_ec_suspend(ec_dev);
> +}
> +
> +static int __maybe_unused cros_ec_uart_resume(struct device *dev)
> +{
> + struct cros_ec_device *ec_dev = dev_get_drvdata(dev);
> +
> + return cros_ec_resume(ec_dev);
> +}
> +
> +static SIMPLE_DEV_PM_OPS(cros_ec_uart_pm_ops, cros_ec_uart_suspend,
> + cros_ec_uart_resume);
> +
> +#ifdef CONFIG_ACPI
> +static const struct acpi_device_id cros_ec_uart_acpi_id[] = {
> + { "GOOG0019", 0 },
> + { /* sentinel */ }
> +};
> +
> +MODULE_DEVICE_TABLE(acpi, cros_ec_uart_acpi_id);
> +#endif
> +
> +static struct serdev_device_driver cros_ec_uart_driver = {
> + .driver = {
> + .name = "cros-ec-uart",
> + .acpi_match_table = ACPI_PTR(cros_ec_uart_acpi_id),
> + .pm = &cros_ec_uart_pm_ops,
> + },
> + .probe = cros_ec_uart_probe,
> + .remove = cros_ec_uart_remove,
> +};
> +
> +module_serdev_device_driver(cros_ec_uart_driver);
> +
> +MODULE_LICENSE("GPL");
> +MODULE_DESCRIPTION("UART interface for ChromeOS Embedded Controller");
> +MODULE_AUTHOR("Bhanu Prakash Maiya <bhanumaiya@xxxxxxxxxxxx>");
> --
> 2.38.1.584.g0f3c55d4c2-goog
>
>

Best regards,

[1] https://www.kernel.org/doc/html/latest/process/coding-style.html#commenting
[2] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=bdc48fa11e46f867ea4d75fa59ee87a7f48be144
[3] https://elixir.bootlin.com/linux/v6.1-rc7/source/include/linux/wait.h#L510