Resolve keyboard not working out of the box for Dell XPS 9345 13" codenamed 'tributo'. X1E80100-based laptop's initial support is currently being upstreamed [1]. In present state, keyboard is succesfully initialized, however attempt to type anything throws 'incomplete report' errors. When utilizing I2C_HID_QUIRK_BAD_INPUT_SIZE quirk the error is gone, however raw data coming from the keyboard is always the same, no matter the key pressed. Issue 'resolves' itself when suspending and resuming the device. It appears that calling power on command one more time after device initialization before finishing off the probing fixes this weird behavior, and keyboard works right away. Introduce a new quirk for such behaviour, and enable it for particular keyboard. Vendor is shown as 'QTEC', however device id is reported as 0000. Given that vendor was not present before, using HID_ANY_ID to match the device should be okay in this case. [1] https://lore.kernel.org/all/20240921163455.12577-1-alex.vinarskis@gmail.com/ Aleksandrs Vinarskis (2): HID: i2c-hid: introduce re-power-on quirk HID: i2c-hid: introduce qtec vendor, enable re-power-on quirk drivers/hid/hid-ids.h | 2 ++ drivers/hid/i2c-hid/i2c-hid-core.c | 12 +++++++++++- 2 files changed, 13 insertions(+), 1 deletion(-) -- 2.43.0 It appears some keyboards from vendor 'QTEC' will not work properly until suspend & resume. Empirically narrowed down to solution of re-sending power on command _after_ initialization was completed before the end of initial probing. Signed-off-by: Aleksandrs Vinarskis <alex.vinarskis@gmail.com> --- drivers/hid/i2c-hid/i2c-hid-core.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/hid/i2c-hid/i2c-hid-core.c b/drivers/hid/i2c-hid/i2c-hid-core.c index XXXXXXX..XXXXXXX 100644 --- a/drivers/hid/i2c-hid/i2c-hid-core.c +++ b/drivers/hid/i2c-hid/i2c-hid-core.c @@ -XXX,XX +XXX,XX @@ #define I2C_HID_QUIRK_BAD_INPUT_SIZE BIT(3) #define I2C_HID_QUIRK_NO_WAKEUP_AFTER_RESET BIT(4) #define I2C_HID_QUIRK_NO_SLEEP_ON_SUSPEND BIT(5) +#define I2C_HID_QUIRK_RE_POWER_ON BIT(6) /* Command opcodes */ #define I2C_HID_OPCODE_RESET 0x01 @@ -XXX,XX +XXX,XX @@ static int i2c_hid_core_register_hid(struct i2c_hid *ihid) return ret; } - return 0; + /* At least some QTEC devices need this after initialization */ + if (ihid->quirks & I2C_HID_QUIRK_RE_POWER_ON) + ret = i2c_hid_set_power(ihid, I2C_HID_PWR_ON); + + return ret; } static int i2c_hid_core_probe_panel_follower(struct i2c_hid *ihid) -- 2.43.0 This solves keyboard not working until suspend&resume issue on Dell XPS 9345 13" (codenamed 'tributo'). Signed-off-by: Aleksandrs Vinarskis <alex.vinarskis@gmail.com> --- drivers/hid/hid-ids.h | 2 ++ drivers/hid/i2c-hid/i2c-hid-core.c | 5 +++++ 2 files changed, 7 insertions(+) diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index XXXXXXX..XXXXXXX 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -XXX,XX +XXX,XX @@ #define USB_VENDOR_ID_PRODIGE 0x05af #define USB_DEVICE_ID_PRODIGE_CORDLESS 0x3062 +#define I2C_VENDOR_ID_QTEC 0x6243 + #define USB_VENDOR_ID_QUANTA 0x0408 #define USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH 0x3000 #define USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH_3001 0x3001 diff --git a/drivers/hid/i2c-hid/i2c-hid-core.c b/drivers/hid/i2c-hid/i2c-hid-core.c index XXXXXXX..XXXXXXX 100644 --- a/drivers/hid/i2c-hid/i2c-hid-core.c +++ b/drivers/hid/i2c-hid/i2c-hid-core.c @@ -XXX,XX +XXX,XX @@ static const struct i2c_hid_quirks { I2C_HID_QUIRK_BAD_INPUT_SIZE }, { I2C_VENDOR_ID_CIRQUE, I2C_PRODUCT_ID_CIRQUE_1063, I2C_HID_QUIRK_NO_SLEEP_ON_SUSPEND }, + /* + * Without additional power on command, at least some QTEC devices send garbage + */ + { I2C_VENDOR_ID_QTEC, HID_ANY_ID, + I2C_HID_QUIRK_RE_POWER_ON }, /* * Sending the wakeup after reset actually break ELAN touchscreen controller */ -- 2.43.0 Resolve keyboard not working out of the box for Dell XPS 9345 13" codenamed 'tributo'. X1E80100-based laptop's initial support is currently being upstreamed [1]. In present state, keyboard is succesfully initialized, however attempt to type anything throws 'incomplete report' errors. When utilizing I2C_HID_QUIRK_BAD_INPUT_SIZE quirk the error is gone, however raw data coming from the keyboard is always the same, no matter the key pressed. Issue 'resolves' itself when suspending and resuming the device. It appears that calling power on command one more time after device initialization before finishing off the probing fixes this weird behavior, and keyboard works right away. Introduce a new quirk for such behaviour, and enable it for particular keyboard. Vendor is shown as 'QTEC', however device id is reported as 0000. Given that vendor was not present before, using HID_ANY_ID to match the device should be okay in this case. In v1 it was suggested to make a dedicated i2c-of-qtec driver, but I was not sure how to proceed at the time. I have now drafted a dedicated driver, and it really is just probe method being extended to call powerup command again. Given that a similarly 'ugly' quirk was just merged to i2c-hid-core.c for a Goodix device [2], and that (IMO) creating a dedicated driver for such a small change without any plan on extending it will be just polluting, I am asking you to consider this change again. Alternatively, if it is yet still strongly preferred to have a dedicated driver to include this quirk, please let me know so I can proceed accordingly. [1] https://lore.kernel.org/all/20241003211139.9296-1-alex.vinarskis@gmail.com/ [2] https://lore.kernel.org/all/20241007222629.172047-1-marynczakbartlomiej@gmail.com/ -------- Changes to V1: * Rebase on top of latest linux-next * Update coverletter's reasoning and links * link: https://lore.kernel.org/all/20240925100303.9112-1-alex.vinarskis@gmail.com/ Aleksandrs Vinarskis (2): HID: i2c-hid: introduce re-power-on quirk HID: i2c-hid: introduce qtec vendor, enable re-power-on quirk drivers/hid/hid-ids.h | 2 ++ drivers/hid/i2c-hid/i2c-hid-core.c | 12 +++++++++++- 2 files changed, 13 insertions(+), 1 deletion(-) -- 2.45.2 It appears some keyboards from vendor 'QTEC' will not work properly until suspend & resume. Empirically narrowed down to solution of re-sending power on command _after_ initialization was completed before the end of initial probing. Signed-off-by: Aleksandrs Vinarskis <alex.vinarskis@gmail.com> --- drivers/hid/i2c-hid/i2c-hid-core.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/hid/i2c-hid/i2c-hid-core.c b/drivers/hid/i2c-hid/i2c-hid-core.c index XXXXXXX..XXXXXXX 100644 --- a/drivers/hid/i2c-hid/i2c-hid-core.c +++ b/drivers/hid/i2c-hid/i2c-hid-core.c @@ -XXX,XX +XXX,XX @@ #define I2C_HID_QUIRK_NO_WAKEUP_AFTER_RESET BIT(4) #define I2C_HID_QUIRK_NO_SLEEP_ON_SUSPEND BIT(5) #define I2C_HID_QUIRK_DELAY_WAKEUP_AFTER_RESUME BIT(6) +#define I2C_HID_QUIRK_RE_POWER_ON BIT(7) /* Command opcodes */ #define I2C_HID_OPCODE_RESET 0x01 @@ -XXX,XX +XXX,XX @@ static int i2c_hid_core_register_hid(struct i2c_hid *ihid) return ret; } - return 0; + /* At least some QTEC devices need this after initialization */ + if (ihid->quirks & I2C_HID_QUIRK_RE_POWER_ON) + ret = i2c_hid_set_power(ihid, I2C_HID_PWR_ON); + + return ret; } static int i2c_hid_core_probe_panel_follower(struct i2c_hid *ihid) -- 2.45.2 This solves keyboard not working until suspend&resume issue on Dell XPS 9345 13" (codenamed 'tributo'). Signed-off-by: Aleksandrs Vinarskis <alex.vinarskis@gmail.com> --- drivers/hid/hid-ids.h | 2 ++ drivers/hid/i2c-hid/i2c-hid-core.c | 5 +++++ 2 files changed, 7 insertions(+) diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index XXXXXXX..XXXXXXX 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -XXX,XX +XXX,XX @@ #define USB_VENDOR_ID_PRODIGE 0x05af #define USB_DEVICE_ID_PRODIGE_CORDLESS 0x3062 +#define I2C_VENDOR_ID_QTEC 0x6243 + #define USB_VENDOR_ID_QUANTA 0x0408 #define USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH 0x3000 #define USB_DEVICE_ID_QUANTA_OPTICAL_TOUCH_3001 0x3001 diff --git a/drivers/hid/i2c-hid/i2c-hid-core.c b/drivers/hid/i2c-hid/i2c-hid-core.c index XXXXXXX..XXXXXXX 100644 --- a/drivers/hid/i2c-hid/i2c-hid-core.c +++ b/drivers/hid/i2c-hid/i2c-hid-core.c @@ -XXX,XX +XXX,XX @@ static const struct i2c_hid_quirks { I2C_HID_QUIRK_BAD_INPUT_SIZE }, { I2C_VENDOR_ID_CIRQUE, I2C_PRODUCT_ID_CIRQUE_1063, I2C_HID_QUIRK_NO_SLEEP_ON_SUSPEND }, + /* + * Without additional power on command, at least some QTEC devices send garbage + */ + { I2C_VENDOR_ID_QTEC, HID_ANY_ID, + I2C_HID_QUIRK_RE_POWER_ON }, /* * Sending the wakeup after reset actually break ELAN touchscreen controller */ -- 2.45.2