Skip to contentSkip to conversation

thinkpad_acpi: Add support for X1 Yoga (2016) Tablet Mode

PATCH Superseded Lyude 3 participants 7 messages Oct 25, 2016 – Nov 5, 2016 x86 platform drivers

v1 of 5 Superseded Latest is v5 →

lore
Share & tools
lore

Lyude

author
#0 lore raw

PATCH 1/1 thinkpad_acpi: Add support for X1 Yoga (2016) Tablet Mode

For whatever reason, the X1 Yoga doesn't support the normal method of querying for tablet mode. Instead of providing the MHKG method under the hotkey handle, we're instead given the CMMD method under the EC handle. Values on this handle are either 0x1, laptop mode, or 0x6, tablet mode.

---
So HOPEFULLY I actually did this right. There was some discussion here previously about supporting this, and it was previously mentioned that the IOST method was responsible for holding the current tablet status. However, it really doesn't seem like this is the case:

    Method (_Q2E, 0, NotSerialized)  // _Qxx: EC Query
    {
	Local0 = \_SB.PCI0.LPC.EC.CMMD
	If ((Local0 != 0x00) && (Local0 <= 0x06))
	{
	    If (Local0 != \PMMD)
	    {
		\PMMD = Local0
		\_SB.PCI0.LPC.EC.HKEY.MHKQ (0x60C0)
	    }
	}

	If ((Local0 == 0x00) || (Local0 >= 0x07))
	{
	    Local1 = 0x01
	}
	ElseIf ((Local0 == 0x02) || (Local0 == 0x03))
	{
	    Local1 = 0x02
	}
	Else
	{
	    Local1 = 0x03
	}

	If (Local1 != \LVMD)
	{
	    \LVMD = Local1
	    Sleep (0x0A)
	    \_SB.PCI0.LPC.EC.HKEY.MHKQ (0x60F0)
	}
    }

IOST seems to be present in the decompiled DSDT from my Yoga, however it's quite far from the spot that actually sends the 0x60C0 hotkey event. Let me know if you guys want to verify this against the full decompiled DSDT for this machine.

 drivers/platform/x86/thinkpad_acpi.c | 36 +++++++++++++++++++++++++++++++++---
 1 file changed, 33 insertions(+), 3 deletions(-)
drivers/platform/x86/thinkpad_acpi.c +33 -3
index b65ce75..8082dc9 100644
@@ -190,6 +190,9 @@ enum tpacpi_hkey_event_t {
 	TP_HKEY_EV_LID_OPEN		= 0x5002, /* laptop lid opened */
 	TP_HKEY_EV_TABLET_TABLET	= 0x5009, /* tablet swivel up */
 	TP_HKEY_EV_TABLET_NOTEBOOK	= 0x500a, /* tablet swivel down */
+	TP_HKEY_EV_TABLET_CHANGED	= 0x60c0, /* X1 Yoga (2016):
+						   * enter/leave tablet mode
+						   */
 	TP_HKEY_EV_PEN_INSERTED		= 0x500b, /* tablet pen inserted */
 	TP_HKEY_EV_PEN_REMOVED		= 0x500c, /* tablet pen removed */
 	TP_HKEY_EV_BRGHT_CHANGED	= 0x5010, /* backlight control event */
@@ -303,6 +306,7 @@ static struct {
 	u32 hotkey_mask:1;
 	u32 hotkey_wlsw:1;
 	u32 hotkey_tablet:1;
+	u32 hotkey_tablet_cmmd:1;
 	u32 kbdlight:1;
 	u32 light:1;
 	u32 light_status:1;
@@ -2059,6 +2063,8 @@ static void hotkey_poll_setup(const bool may_warn);
 
 /* HKEY.MHKG() return bits */
 #define TP_HOTKEY_TABLET_MASK (1 << 3)
+/* ThinkPad X1 Yoga (2016) */
+#define TP_EC_CMMD_TABLET_MODE 0x6
 
 static int hotkey_get_wlsw(void)
 {
@@ -2083,10 +2089,18 @@ static int hotkey_get_tablet_mode(int *status)
 {
 	int s;
 
-	if (!acpi_evalf(hkey_handle, &s, "MHKG", "d"))
-		return -EIO;
+	if (tp_features.hotkey_tablet_cmmd) {
+		if (!acpi_evalf(ec_handle, &s, "CMMD", "d"))
+			return -EIO;
+
+		*status = (s == TP_EC_CMMD_TABLET_MODE);
+	} else {
+		if (!acpi_evalf(hkey_handle, &s, "MHKG", "d"))
+			return -EIO;
+
+		*status = ((s & TP_HOTKEY_TABLET_MASK) != 0);
+	}
 
-	*status = ((s & TP_HOTKEY_TABLET_MASK) != 0);
 	return 0;
 }
 
@@ -3475,6 +3489,18 @@ static int __init hotkey_init(struct ibm_init_struct *iibm)
 				&dev_attr_hotkey_tablet_mode.attr);
 	}
 
+	/* For X1 Yoga (2016) */
+	if (!res && acpi_evalf(ec_handle, &status, "CMMD", "qd")) {
+		tp_features.hotkey_tablet = 1;
+		tp_features.hotkey_tablet_cmmd = 1;
+		tabletsw_state = (status == TP_EC_CMMD_TABLET_MODE);
+
+		pr_info("Possible tablet mode switch found; ThinkPad in %s mode\n",
+			(tabletsw_state) ? "tablet" : "laptop");
+		res = add_to_attr_set(hotkey_dev_attributes,
+				      &dev_attr_hotkey_tablet_mode.attr);
+	}
+
 	if (!res)
 		res = register_attr_set_with_sysfs(
 				hotkey_dev_attributes,
@@ -3899,6 +3925,10 @@ static bool hotkey_notify_6xxx(const u32 hkey,
 		*ignore_acpi_ev = true;
 		return true;
 
+	case TP_HKEY_EV_TABLET_CHANGED:
+		tpacpi_input_send_tabletsw();
+		break;
+
 	default:
 		pr_warn("unknown possible thermal alarm or keyboard event received\n");
 		known = false;
Signature · 1 line
2.7.4
author
#1 lore raw

as well, can someone confirm this patch made it to the ibm-acpi-devel list? When I originally sent this I realized I wasn't subscribed to the list, so I'm guessing I might need to resend.

Quoted 157 lines · Lyude

For whatever reason, the X1 Yoga doesn't support the normal method of querying for tablet mode. Instead of providing the MHKG method under the
hotkey handle, we're instead given the CMMD method under the EC handle.
Values on this handle are either 0x1, laptop mode, or 0x6, tablet mode.

---
So HOPEFULLY I actually did this right. There was some discussion here
previously about supporting this, and it was previously mentioned that the IOST
method was responsible for holding the current tablet status. However, it
really doesn't seem like this is the case:

    Method (_Q2E, 0, NotSerialized)  // _Qxx: EC Query

    {
	Local0 = \_SB.PCI0.LPC.EC.CMMD
	If ((Local0 != 0x00) && (Local0 <= 0x06))
	{
	    If (Local0 != \PMMD)
	    {
		\PMMD = Local0
		\_SB.PCI0.LPC.EC.HKEY.MHKQ (0x60C0)
	    }
	}

	If ((Local0 == 0x00) || (Local0 >= 0x07))
	{
	    Local1 = 0x01
	}
	ElseIf ((Local0 == 0x02) || (Local0 == 0x03))
	{
	    Local1 = 0x02
	}
	Else
	{
	    Local1 = 0x03
	}

	If (Local1 != \LVMD)
	{
	    \LVMD = Local1
	    Sleep (0x0A)
	    \_SB.PCI0.LPC.EC.HKEY.MHKQ (0x60F0)
	}

    }

IOST seems to be present in the decompiled DSDT from my Yoga, however it's
quite far from the spot that actually sends the 0x60C0 hotkey event. Let me
know if you guys want to verify this against the full decompiled DSDT for this
machine.

 drivers/platform/x86/thinkpad_acpi.c | 36
+++++++++++++++++++++++++++++++++---

 1 file changed, 33 insertions(+), 3 deletions(-)
drivers/platform/x86/thinkpad_acpi.c +0 -0
index b65ce75..8082dc9 100644
@@ -190,6 +190,9 @@ enum tpacpi_hkey_event_t {

  TP_HKEY_EV_LID_OPEN = 0x5002, /* laptop lid
opened */
  TP_HKEY_EV_TABLET_TABLET = 0x5009, /* tablet swivel
up */
  TP_HKEY_EV_TABLET_NOTEBOOK = 0x500a, /* tablet swivel
down */

drivers/platform/x86/thinkpad_acpi.c +1 -0
+	TP_HKEY_EV_TABLET_CHANGED	= 0x60c0, /* X1 Yoga

(2016):

drivers/platform/x86/thinkpad_acpi.c +1 -0
+						   * enter/leave

tablet mode

drivers/platform/x86/thinkpad_acpi.c +1 -0
+						   */

  TP_HKEY_EV_PEN_INSERTED = 0x500b, /* tablet
pen inserted */
  TP_HKEY_EV_PEN_REMOVED = 0x500c, /* tablet
pen removed */
  TP_HKEY_EV_BRGHT_CHANGED = 0x5010, /* backlight
control event */

drivers/platform/x86/thinkpad_acpi.c +1 -0
@@ -303,6 +306,7 @@ static struct {
  	u32 hotkey_mask:1;
  	u32 hotkey_wlsw:1;
  	u32 hotkey_tablet:1;
+	u32 hotkey_tablet_cmmd:1;
  	u32 kbdlight:1;
  	u32 light:1;
  	u32 light_status:1;
@@ -2059,6 +2063,8 @@ static void hotkey_poll_setup(const bool
 may_warn);


 /* HKEY.MHKG() return bits */
 #define TP_HOTKEY_TABLET_MASK (1 << 3)

drivers/platform/x86/thinkpad_acpi.c +2 -0
+/* ThinkPad X1 Yoga (2016) */
+#define TP_EC_CMMD_TABLET_MODE 0x6


 static int hotkey_get_wlsw(void)
 {

drivers/platform/x86/thinkpad_acpi.c +0 -0
@@ -2083,10 +2089,18 @@ static int hotkey_get_tablet_mode(int
 *status)
  {
  	int s;

drivers/platform/x86/thinkpad_acpi.c +11 -2
-	if (!acpi_evalf(hkey_handle, &s, "MHKG", "d"))
-		return -EIO;
+	if (tp_features.hotkey_tablet_cmmd) {
+		if (!acpi_evalf(ec_handle, &s, "CMMD", "d"))
+			return -EIO;
+
+		*status = (s == TP_EC_CMMD_TABLET_MODE);
+	} else {
+		if (!acpi_evalf(hkey_handle, &s, "MHKG", "d"))
+			return -EIO;
+
+		*status = ((s & TP_HOTKEY_TABLET_MASK) != 0);
+	}

drivers/platform/x86/thinkpad_acpi.c +0 -1
-	*status = ((s & TP_HOTKEY_TABLET_MASK) != 0);

  return 0;
 }

drivers/platform/x86/thinkpad_acpi.c +0 -0
@@ -3475,6 +3489,18 @@ static int __init hotkey_init(struct
 ibm_init_struct *iibm)
  				&dev_attr_hotkey_tablet_mode.attr);
  	}

drivers/platform/x86/thinkpad_acpi.c +7 -0
+	/* For X1 Yoga (2016) */
+	if (!res && acpi_evalf(ec_handle, &status, "CMMD", "qd")) {
+		tp_features.hotkey_tablet = 1;
+		tp_features.hotkey_tablet_cmmd = 1;
+		tabletsw_state = (status == TP_EC_CMMD_TABLET_MODE);
+
+		pr_info("Possible tablet mode switch found; ThinkPad

in %s mode\n",

drivers/platform/x86/thinkpad_acpi.c +3 -0
+			(tabletsw_state) ? "tablet" : "laptop");
+		res = add_to_attr_set(hotkey_dev_attributes,
+				      &dev_attr_hotkey_tablet_mode.a

ttr);

drivers/platform/x86/thinkpad_acpi.c +2 -0
+	}
+

  if (!res)
  res = register_attr_set_with_sysfs( hotkey_dev_attributes,

drivers/platform/x86/thinkpad_acpi.c +0 -0
@@ -3899,6 +3925,10 @@ static bool hotkey_notify_6xxx(const u32 hkey,
  		*ignore_acpi_ev = true;
  		return true;

drivers/platform/x86/thinkpad_acpi.c +4 -0
+	case TP_HKEY_EV_TABLET_CHANGED:
+		tpacpi_input_send_tabletsw();
+		break;
+
Quoted 4 lines

  default:
  pr_warn("unknown possible thermal alarm or keyboard event received\n");
  known = false;

------------------------------------------------------------------------------
The Command Line: Reinvented for Modern Developers Did the resurgence of CLI tooling catch you by surprise? Reconnect with the command line and become more productive. Learn the new .NET and ASP.NET CLI. Get your free copy!
http://sdm.link/telerik
_______________________________________________
ibm-acpi-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/ibm-acpi-devel

#2 lore raw
Quoted 3 lines · Lyude Paul

as well, can someone confirm this patch made it to the ibm-acpi-devel list? When I originally sent this I realized I wasn't subscribed to the list, so I'm guessing I might need to resend.

I received it. I just didn't have the time to go over it in detail.

At first look, it seems correct...

Quoted 5 lines
Quoted 5 lines
drivers/platform/x86/thinkpad_acpi.c +5 -0
+	/* For X1 Yoga (2016) */
+	if (!res && acpi_evalf(ec_handle, &status, "CMMD", "qd")) {
+		tp_features.hotkey_tablet = 1;
+		tp_features.hotkey_tablet_cmmd = 1;
+		tabletsw_state = (status == TP_EC_CMMD_TABLET_MODE);
Quoted 8 lines

Review

Quoted 8 lines
(quoted diff) +2 -0
+
+		pr_info("Possible tablet mode switch found; ThinkPad

in %s mode\n",

drivers/platform/x86/thinkpad_acpi.c +3 -0
+			(tabletsw_state) ? "tablet" : "laptop");
+		res = add_to_attr_set(hotkey_dev_attributes,
+				      &dev_attr_hotkey_tablet_mode.a

ttr);

(quoted diff) +1 -0
+	}

Ehh... the code doesn't act as if it were a "possible" switch, but rather that it _did_ find a x1_yoga-style switch, so the debug message should be assertive to match the code:

pr_info("Tablet mode switch found (X1 Yoga style), Thinkpad is in...")

or something to that effect.

If the code is wrong, and it has false possitives, we enhance the detection (possibly by using a quirk table if we must).

Care to respin with that change?

Signature · 1 line
  Henrique Holschuh
author
#3 lore raw

PATCH v2 1/1 thinkpad_acpi: Add support for X1 Yoga (2016) Tablet Mode

For whatever reason, the X1 Yoga doesn't support the normal method of querying for tablet mode. Instead of providing the MHKG method under the hotkey handle, we're instead given the CMMD method under the EC handle. Values on this handle are either 0x1, laptop mode, or 0x6, tablet mode.

Changes since v1:
- Clarify kernel output when finding the tablet mode switch

---
 drivers/platform/x86/thinkpad_acpi.c | 36 +++++++++++++++++++++++++++++++++---
 1 file changed, 33 insertions(+), 3 deletions(-)
drivers/platform/x86/thinkpad_acpi.c +33 -3
index b65ce75..d9f9956 100644
@@ -190,6 +190,9 @@ enum tpacpi_hkey_event_t {
 	TP_HKEY_EV_LID_OPEN		= 0x5002, /* laptop lid opened */
 	TP_HKEY_EV_TABLET_TABLET	= 0x5009, /* tablet swivel up */
 	TP_HKEY_EV_TABLET_NOTEBOOK	= 0x500a, /* tablet swivel down */
+	TP_HKEY_EV_TABLET_CHANGED	= 0x60c0, /* X1 Yoga (2016):
+						   * enter/leave tablet mode
+						   */
 	TP_HKEY_EV_PEN_INSERTED		= 0x500b, /* tablet pen inserted */
 	TP_HKEY_EV_PEN_REMOVED		= 0x500c, /* tablet pen removed */
 	TP_HKEY_EV_BRGHT_CHANGED	= 0x5010, /* backlight control event */
@@ -303,6 +306,7 @@ static struct {
 	u32 hotkey_mask:1;
 	u32 hotkey_wlsw:1;
 	u32 hotkey_tablet:1;
+	u32 hotkey_tablet_cmmd:1;
 	u32 kbdlight:1;
 	u32 light:1;
 	u32 light_status:1;
@@ -2059,6 +2063,8 @@ static void hotkey_poll_setup(const bool may_warn);
 
 /* HKEY.MHKG() return bits */
 #define TP_HOTKEY_TABLET_MASK (1 << 3)
+/* ThinkPad X1 Yoga (2016) */
+#define TP_EC_CMMD_TABLET_MODE 0x6
 
 static int hotkey_get_wlsw(void)
 {
@@ -2083,10 +2089,18 @@ static int hotkey_get_tablet_mode(int *status)
 {
 	int s;
 
-	if (!acpi_evalf(hkey_handle, &s, "MHKG", "d"))
-		return -EIO;
+	if (tp_features.hotkey_tablet_cmmd) {
+		if (!acpi_evalf(ec_handle, &s, "CMMD", "d"))
+			return -EIO;
+
+		*status = (s == TP_EC_CMMD_TABLET_MODE);
+	} else {
+		if (!acpi_evalf(hkey_handle, &s, "MHKG", "d"))
+			return -EIO;
+
+		*status = ((s & TP_HOTKEY_TABLET_MASK) != 0);
+	}
 
-	*status = ((s & TP_HOTKEY_TABLET_MASK) != 0);
 	return 0;
 }
 
@@ -3475,6 +3489,18 @@ static int __init hotkey_init(struct ibm_init_struct *iibm)
 				&dev_attr_hotkey_tablet_mode.attr);
 	}
 
+	/* For X1 Yoga (2016) */
+	if (!res && acpi_evalf(ec_handle, &status, "CMMD", "qd")) {
+		tp_features.hotkey_tablet = 1;
+		tp_features.hotkey_tablet_cmmd = 1;
+		tabletsw_state = (status == TP_EC_CMMD_TABLET_MODE);
+
+		pr_info("Tablet mode switch found (X1 Yoga style); ThinkPad in %s mode\n",
+			(tabletsw_state) ? "tablet" : "laptop");
+		res = add_to_attr_set(hotkey_dev_attributes,
+				      &dev_attr_hotkey_tablet_mode.attr);
+	}
+
 	if (!res)
 		res = register_attr_set_with_sysfs(
 				hotkey_dev_attributes,
@@ -3899,6 +3925,10 @@ static bool hotkey_notify_6xxx(const u32 hkey,
 		*ignore_acpi_ev = true;
 		return true;
 
+	case TP_HKEY_EV_TABLET_CHANGED:
+		tpacpi_input_send_tabletsw();
+		break;
+
 	default:
 		pr_warn("unknown possible thermal alarm or keyboard event received\n");
 		known = false;
Signature · 1 line
2.7.4
#4 lore raw
Quoted 7 lines · Lyude

For whatever reason, the X1 Yoga doesn't support the normal method of querying for tablet mode. Instead of providing the MHKG method under the hotkey handle, we're instead given the CMMD method under the EC handle. Values on this handle are either 0x1, laptop mode, or 0x6, tablet mode.

Changes since v1:
- Clarify kernel output when finding the tablet mode switch

These two lines go below the --- below (they aren't meant to be part of the permanent commit message). Please see Documentation/SubmittingPatches.

I've queued this to testing as it addresses Henrique's change request. Henrique, I'll add your Reviewed-by if you send it along, or drop it if you have more reservations.

Thanks,

Signature · 2 lines
Darren Hart
Intel Open Source Technology Center
#5 lore raw
Quoted 9 lines · Darren Hart
Quoted 7 lines · Lyude

For whatever reason, the X1 Yoga doesn't support the normal method of querying for tablet mode. Instead of providing the MHKG method under the hotkey handle, we're instead given the CMMD method under the EC handle. Values on this handle are either 0x1, laptop mode, or 0x6, tablet mode.

Changes since v1:
- Clarify kernel output when finding the tablet mode switch

Quoted 6 lines

These two lines go below the --- below (they aren't meant to be part of the permanent commit message). Please see Documentation/SubmittingPatches.

I've queued this to testing as it addresses Henrique's change request. Henrique, I'll add your Reviewed-by if you send it along, or drop it if you have more reservations.

No reservations, since it was tested to not regress previous thinkpads.

Signature · 1 line
  Henrique Holschuh
#6 lore raw
In reply to Darren Hart
Quoted 9 lines · Darren Hart
Quoted 7 lines · Lyude

For whatever reason, the X1 Yoga doesn't support the normal method of querying for tablet mode. Instead of providing the MHKG method under the hotkey handle, we're instead given the CMMD method under the EC handle. Values on this handle are either 0x1, laptop mode, or 0x6, tablet mode.

Changes since v1:
- Clarify kernel output when finding the tablet mode switch

Quoted 6 lines

These two lines go below the --- below (they aren't meant to be part of the permanent commit message). Please see Documentation/SubmittingPatches.

I've queued this to testing as it addresses Henrique's change request. Henrique, I'll add your Reviewed-by if you send it along, or drop it if you have more reservations.

Scratch that, I see there is a new series with this patch in it. Will respond there.

Signature · 2 lines
Darren Hart
Intel Open Source Technology Center