Skip to contentSkip to conversation

platform/x86: redmi-wmi: report EC state change events

PATCH In review Musaev Ibragim 2 participants 4 messages Jul 14 – Aug 18

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 platform/x86: redmi-wmi: report EC state change events

The Redmibook EC/firmware fully handles the keyboard backlight cycle, the OEM preset power mode (Fn+K) and the Fn lock toggle by itself, and sends a WMI event carrying the resulting state in the third payload byte. These events are currently swallowed with KE_IGNORE, so userspace never learns that the state changed and cannot give the user any feedback (OSD), even though the WMI event is the only notification channel for these EC-driven changes.

Report them as key presses instead:

- keyboard backlight cycle -> KEY_KBDILLUMTOGGLE
- OEM preset power mode -> KEY_PERFORMANCE
- Fn lock toggle -> KEY_FN_ESC

Desktops that only look at the keycode get the usual hotkey behaviour; since sparse-keymap emits MSC_SCAN with the raw payload before the key event, an OSD daemon can additionally recover the exact new state from byte 2 (e.g. backlight Off/Low/High/Auto is 0x00/0x05/0x0a/0x80).

Note that the power mode event must keep being read from the WMI device in any case: on the TM2209 the ACPI event handler (EV20) applies the mode change as a side effect of building the event payload for _WED.

Tested on Redmi Book Pro 15 2023 (TM2209).

---
A related problem noticed while testing this: bitland-mifs-wmi lists the same event GUID (46C93E13-EE9B-4262-8488-563BCA757FEF, BITLAND_EVENT_GUID) in its wmi_device_id table, so on a Redmibook both drivers race for the event device at boot and whichever probes first wins. When bitland-mifs-wmi wins, the "Redmibook WMI keys" input device never appears, and since the MIFS control methods don't work on this machine either (different SMM interface — platform_profile writes fail, the kbd_backlight LED always reads 0), the laptop is left with no working Fn events at all until bitland_mifs_wmi is blacklisted. Should bitland-mifs-wmi gate its binding on DMI, or is another way to resolve the overlap preferred? Happy to test patches on TM2209.

 drivers/platform/x86/redmi-wmi.c | 26 +++++++++++++-------------
 1 file changed, 13 insertions(+), 13 deletions(-)
drivers/platform/x86/redmi-wmi.c +13 -13
index 5889863..cc82ef5 100644
@@ -29,24 +29,24 @@ static const struct key_entry redmi_wmi_keymap[] = {
 	{KE_KEY, 0x00011801,	{KEY_ASSISTANT}},
 	{KE_KEY, 0x00011901,	{KEY_ASSISTANT}},
 
-	/* Keyboard backlight */
-	{KE_IGNORE, 0x00000501, {}},
-	{KE_IGNORE, 0x00800501, {}},
-	{KE_IGNORE, 0x00050501, {}},
-	{KE_IGNORE, 0x000a0501, {}},
+	/* Keyboard backlight: Off / Auto / Low / High (new state in byte 2) */
+	{KE_KEY, 0x00000501,	{KEY_KBDILLUMTOGGLE}},
+	{KE_KEY, 0x00800501,	{KEY_KBDILLUMTOGGLE}},
+	{KE_KEY, 0x00050501,	{KEY_KBDILLUMTOGGLE}},
+	{KE_KEY, 0x000a0501,	{KEY_KBDILLUMTOGGLE}},
 
 	/* Xiaomi G Command Center */
 	{KE_KEY, 0x00010a01,	{KEY_VENDOR}},
 
-	/* OEM preset power mode */
-	{KE_IGNORE, 0x00011601, {}},
-	{KE_IGNORE, 0x00021601, {}},
-	{KE_IGNORE, 0x00031601, {}},
-	{KE_IGNORE, 0x00041601, {}},
+	/* OEM preset power mode: 1=Balanced 2=Silent 3=Turbo 4=Full speed */
+	{KE_KEY, 0x00011601,	{KEY_PERFORMANCE}},
+	{KE_KEY, 0x00021601,	{KEY_PERFORMANCE}},
+	{KE_KEY, 0x00031601,	{KEY_PERFORMANCE}},
+	{KE_KEY, 0x00041601,	{KEY_PERFORMANCE}},
 
-	/* Fn Lock state */
-	{KE_IGNORE, 0x00000701, {}},
-	{KE_IGNORE, 0x00010701, {}},
+	/* Fn Lock state: 1=on 0=off */
+	{KE_KEY, 0x00000701,	{KEY_FN_ESC}},
+	{KE_KEY, 0x00010701,	{KEY_FN_ESC}},
 
 	/* Fn+`/1/2/3/4 */
 	{KE_KEY, 0x00011101, {KEY_F13}},
Signature · 1 line
2.54.0
maintainer
#1 lore raw
Quoted 26 lines · Musaev Ibragim

The Redmibook EC/firmware fully handles the keyboard backlight cycle, the OEM preset power mode (Fn+K) and the Fn lock toggle by itself, and sends a WMI event carrying the resulting state in the third payload byte. These events are currently swallowed with KE_IGNORE, so userspace never learns that the state changed and cannot give the user any feedback (OSD), even though the WMI event is the only notification channel for these EC-driven changes.

Report them as key presses instead:

- keyboard backlight cycle -> KEY_KBDILLUMTOGGLE
- OEM preset power mode -> KEY_PERFORMANCE
- Fn lock toggle -> KEY_FN_ESC

Desktops that only look at the keycode get the usual hotkey behaviour; since sparse-keymap emits MSC_SCAN with the raw payload before the key event, an OSD daemon can additionally recover the exact new state from byte 2 (e.g. backlight Off/Low/High/Auto is 0x00/0x05/0x0a/0x80).

Note that the power mode event must keep being read from the WMI device in any case: on the TM2209 the ACPI event handler (EV20) applies the mode change as a side effect of building the event payload for _WED.

Tested on Redmi Book Pro 15 2023 (TM2209).

Quoted 12 lines

---
A related problem noticed while testing this: bitland-mifs-wmi lists the same event GUID (46C93E13-EE9B-4262-8488-563BCA757FEF, BITLAND_EVENT_GUID) in its wmi_device_id table, so on a Redmibook both drivers race for the event device at boot and whichever probes first wins. When bitland-mifs-wmi wins, the "Redmibook WMI keys" input device never appears, and since the MIFS control methods don't work on this machine either (different SMM interface — platform_profile writes fail, the kbd_backlight LED always reads 0), the laptop is left with no working Fn events at all until bitland_mifs_wmi is blacklisted. Should bitland-mifs-wmi gate its binding on DMI, or is another way to resolve the overlap preferred? Happy to test patches on TM2209.

Hi,

Like you've now noticed (I think), there's another series attempting to fix the GUID problem.

I'd primarily prefer taking the GUID problem fixing series (but it should make sure this too is addressed).

If the GUID problem series not finished in this cycle, I can consider this patch instead towards the end of the cycle.

Is that fine with you?

Signature · 1 line
 i.
Quoted 46 lines
 drivers/platform/x86/redmi-wmi.c | 26 +++++++++++++-------------
 1 file changed, 13 insertions(+), 13 deletions(-)
drivers/platform/x86/redmi-wmi.c +13 -13
index 5889863..cc82ef5 100644
@@ -29,24 +29,24 @@ static const struct key_entry redmi_wmi_keymap[] = {
 	{KE_KEY, 0x00011801,	{KEY_ASSISTANT}},
 	{KE_KEY, 0x00011901,	{KEY_ASSISTANT}},
 
-	/* Keyboard backlight */
-	{KE_IGNORE, 0x00000501, {}},
-	{KE_IGNORE, 0x00800501, {}},
-	{KE_IGNORE, 0x00050501, {}},
-	{KE_IGNORE, 0x000a0501, {}},
+	/* Keyboard backlight: Off / Auto / Low / High (new state in byte 2) */
+	{KE_KEY, 0x00000501,	{KEY_KBDILLUMTOGGLE}},
+	{KE_KEY, 0x00800501,	{KEY_KBDILLUMTOGGLE}},
+	{KE_KEY, 0x00050501,	{KEY_KBDILLUMTOGGLE}},
+	{KE_KEY, 0x000a0501,	{KEY_KBDILLUMTOGGLE}},
 
 	/* Xiaomi G Command Center */
 	{KE_KEY, 0x00010a01,	{KEY_VENDOR}},
 
-	/* OEM preset power mode */
-	{KE_IGNORE, 0x00011601, {}},
-	{KE_IGNORE, 0x00021601, {}},
-	{KE_IGNORE, 0x00031601, {}},
-	{KE_IGNORE, 0x00041601, {}},
+	/* OEM preset power mode: 1=Balanced 2=Silent 3=Turbo 4=Full speed */
+	{KE_KEY, 0x00011601,	{KEY_PERFORMANCE}},
+	{KE_KEY, 0x00021601,	{KEY_PERFORMANCE}},
+	{KE_KEY, 0x00031601,	{KEY_PERFORMANCE}},
+	{KE_KEY, 0x00041601,	{KEY_PERFORMANCE}},
 
-	/* Fn Lock state */
-	{KE_IGNORE, 0x00000701, {}},
-	{KE_IGNORE, 0x00010701, {}},
+	/* Fn Lock state: 1=on 0=off */
+	{KE_KEY, 0x00000701,	{KEY_FN_ESC}},
+	{KE_KEY, 0x00010701,	{KEY_FN_ESC}},
 
 	/* Fn+`/1/2/3/4 */
 	{KE_KEY, 0x00011101, {KEY_F13}},
#2 lore raw
Quoted 10 lines · Ilpo Järvinen

Like you've now noticed (I think), there's another series attempting to fix the GUID problem.

I'd primarily prefer taking the GUID problem fixing series (but it should make sure this too is addressed).

If the GUID problem series not finished in this cycle, I can consider this patch instead towards the end of the cycle.

Is that fine with you?

Yes, that's fine with me -- landing the GUID-fix series first makes sense. I've already reviewed the Redmibook side of it on my TM2209 and left comments on patch 2/4, including a request to keep these EC state events observable rather than KE_IGNOREd. Happy to test v2 on this hardware.

If the series misses the cycle, picking this patch up at the end of it works for me.

Thanks,
Ibragim

maintainer
#3 lore raw
In reply to Musaev Ibragim
Quoted 9 lines · Musaev Ibragim

The Redmibook EC/firmware fully handles the keyboard backlight cycle, the OEM preset power mode (Fn+K) and the Fn lock toggle by itself, and sends a WMI event carrying the resulting state in the third payload byte. These events are currently swallowed with KE_IGNORE, so userspace never learns that the state changed and cannot give the user any feedback (OSD), even though the WMI event is the only notification channel for these EC-driven changes.

[...]

Thank you for your contribution, it has been applied to my local review-ilpo-next branch. Note it will show up in the public platform-drivers-x86/review-ilpo-next branch only once I've pushed my local branch there, which might take a while.

FYI [if applicable to your patch], as per Linus' policy change, also fixes are mostly routed through for-next unless the fix is for a commit introduced in the most recent cycle or is clearly a regression fix.

The list of commits applied:
[1/1] platform/x86: redmi-wmi: report EC state change events

      commit: e0d6312578e1fd03738fc2ac8ac21c8bd84e965e

--
i.