| Message ID | 20240527-cros_ec-hwmon-v3-1-e5cd5ab5ba37@weissschuh.net (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | ChromeOS Embedded controller hwmon driver | expand |
On Mon, May 27, 2024 at 10:58:31PM +0200, Thomas Weißschuh wrote: > +/** > + * cros_ec_cmd_readmem - Read from EC memory. > + * > + * @ec_dev: EC device > + * @offset: Is within EC_LPC_ADDR_MEMMAP region. > + * @size: Number of bytes to read. zero means "read a string" (including > + * the trailing '\0'). The behavior is cros_ec_lpc_readmem() only but not for cros_ec_cmd().
On 2024-05-28 06:39:03+0000, Tzung-Bi Shih wrote: > On Mon, May 27, 2024 at 10:58:31PM +0200, Thomas Weißschuh wrote: > > +/** > > + * cros_ec_cmd_readmem - Read from EC memory. > > + * > > + * @ec_dev: EC device > > + * @offset: Is within EC_LPC_ADDR_MEMMAP region. > > + * @size: Number of bytes to read. zero means "read a string" (including > > + * the trailing '\0'). > > The behavior is cros_ec_lpc_readmem() only but not for cros_ec_cmd(). Indeed. I thought I checked for this specifically, but got it wrong. I'll drop the docs and add a if (!size) return -EINVAL; to make sure nobody starts relying on that behaviour when writing a driver against an LPC EC.
diff --git a/drivers/platform/chrome/cros_ec_proto.c b/drivers/platform/chrome/cros_ec_proto.c index 945b1b15a04c..4b48fa1fe3e0 100644 --- a/drivers/platform/chrome/cros_ec_proto.c +++ b/drivers/platform/chrome/cros_ec_proto.c @@ -1035,3 +1035,30 @@ int cros_ec_cmd(struct cros_ec_device *ec_dev, return ret; } EXPORT_SYMBOL_GPL(cros_ec_cmd); + +/** + * cros_ec_cmd_readmem - Read from EC memory. + * + * @ec_dev: EC device + * @offset: Is within EC_LPC_ADDR_MEMMAP region. + * @size: Number of bytes to read. zero means "read a string" (including + * the trailing '\0'). + * Caller must ensure that the buffer is large enough for the + * result when reading a string. + * @dest: EC command output data + * + * Return: >= 0 on success, negative error number on failure. + */ +int cros_ec_cmd_readmem(struct cros_ec_device *ec_dev, u8 offset, u8 size, void *dest) +{ + struct ec_params_read_memmap params; + + if (ec_dev->cmd_readmem) + return ec_dev->cmd_readmem(ec_dev, offset, size, dest); + + params.offset = offset; + params.size = size; + return cros_ec_cmd(ec_dev, 0, EC_CMD_READ_MEMMAP, + ¶ms, sizeof(params), dest, size); +} +EXPORT_SYMBOL_GPL(cros_ec_cmd_readmem); diff --git a/include/linux/platform_data/cros_ec_proto.h b/include/linux/platform_data/cros_ec_proto.h index 8865e350c12a..1ddc52603f9a 100644 --- a/include/linux/platform_data/cros_ec_proto.h +++ b/include/linux/platform_data/cros_ec_proto.h @@ -261,6 +261,8 @@ int cros_ec_get_sensor_count(struct cros_ec_dev *ec); int cros_ec_cmd(struct cros_ec_device *ec_dev, unsigned int version, int command, const void *outdata, size_t outsize, void *indata, size_t insize); +int cros_ec_cmd_readmem(struct cros_ec_device *ec_dev, u8 offset, u8 size, void *dest); + /** * cros_ec_get_time_ns() - Return time in ns. *
To read from the EC memory different mechanism are possible. ECs connected via LPC expose their memory via a ->cmd_readmem operation. Other protocols require the usage of EC_CMD_READ_MEMMAP, which on the other hand is not implemented by LPC ECs. Provide a helper that automatically selects the correct mechanism. Signed-off-by: Thomas Weißschuh <linux@weissschuh.net> --- drivers/platform/chrome/cros_ec_proto.c | 27 +++++++++++++++++++++++++++ include/linux/platform_data/cros_ec_proto.h | 2 ++ 2 files changed, 29 insertions(+)