Skip to contentSkip to conversation

toshiba_acpi: Avoid registering input device on WMI event laptops

PATCH Superseded Azael Avalos 2 participants 5 messages Jul 16, 2015 – Jul 22, 2015

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Azael Avalos

author maintainer
#0 lore raw

PATCH 1/1 toshiba_acpi: Avoid registering input device on WMI event laptops

Commit f11f999e9890 ("toshiba_acpi: Refuse to load on machines with buggy INFO implementations") denied loading on laptops with a WMI Event GUID given that such laptops manage the hotkeys via that interface, however, such laptops have a working Toshiba Configuration Interface (TCI), and thus, such commit denied several supported features.

This patch avoids registering the input device and ignores all hotkey events on laptops with such WMI Event GUI, making the supported features found in those laptops to work.

---
 drivers/platform/x86/toshiba_acpi.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)
drivers/platform/x86/toshiba_acpi.c +13 -8
index 15d8f0d..a3c6ea2 100644
@@ -2397,6 +2397,11 @@ static int toshiba_acpi_setup_keyboard(struct toshiba_acpi_dev *dev)
 	u32 hci_result;
 	int error;
 
+	if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID)) {
+		pr_info("WMI event detected, hotkeys won't be monitored\n");
+		return 0;
+	}
+
 	error = toshiba_acpi_enable_hotkeys(dev);
 	if (error)
 		return error;
@@ -2730,6 +2735,14 @@ static void toshiba_acpi_notify(struct acpi_device *acpi_dev, u32 event)
 
 	switch (event) {
 	case 0x80: /* Hotkeys and some system events */
+		/*
+		 * Machines with this WMI guid aren't supported due to bugs in
+		 * their AML.
+		 *
+		 * Return silently to avoid triggering a netlink event.
+		 */
+		if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
+			return;
 		toshiba_acpi_process_hotkeys(dev);
 		break;
 	case 0x81: /* Dock events */
@@ -2816,14 +2829,6 @@ static int __init toshiba_acpi_init(void)
 {
 	int ret;
 
-	/*
-	 * Machines with this WMI guid aren't supported due to bugs in
-	 * their AML. This check relies on wmi initializing before
-	 * toshiba_acpi to guarantee guids have been identified.
-	 */
-	if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
-		return -ENODEV;
-
 	toshiba_proc_dir = proc_mkdir(PROC_TOSHIBA, acpi_root_dir);
 	if (!toshiba_proc_dir) {
 		pr_err("Unable to create proc dir " PROC_TOSHIBA "\n");
Signature · 1 line
2.4.3
#1 lore raw
Quoted 8 lines · Azael Avalos

Commit f11f999e9890 ("toshiba_acpi: Refuse to load on machines with buggy INFO implementations") denied loading on laptops with a WMI Event GUID given that such laptops manage the hotkeys via that interface, however, such laptops have a working Toshiba Configuration Interface (TCI), and thus, such commit denied several supported features.

This patch avoids registering the input device and ignores all hotkey events on laptops with such WMI Event GUI, making the supported

GUID

Quoted 12 lines

features found in those laptops to work.

---
 drivers/platform/x86/toshiba_acpi.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)
drivers/platform/x86/toshiba_acpi.c +0 -0
index 15d8f0d..a3c6ea2 100644

drivers/platform/x86/toshiba_acpi.c static int toshiba_acpi_setup_keyboard(struct toshiba_acpi_dev *dev) · line 2401

Quoted 5 lines
drivers/platform/x86/toshiba_acpi.c +2 -0
@@ -2397,6 +2397,11 @@ static int toshiba_acpi_setup_keyboard(struct toshiba_acpi_dev *dev)
 	u32 hci_result;
 	int error;
 
+	if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID)) {
+		pr_info("WMI event detected, hotkeys won't be monitored\n");

It's best practice to avoid contractions in any form of technical writing. (See what I did there? ;-)

While this is good for comments, I feel it's more important for kernel messages.

Quoted 22 lines
drivers/platform/x86/toshiba_acpi.c +11 -0
+		return 0;
+	}
+
 	error = toshiba_acpi_enable_hotkeys(dev);
 	if (error)
 		return error;
@@ -2730,6 +2735,14 @@ static void toshiba_acpi_notify(struct acpi_device *acpi_dev, u32 event)
 
 	switch (event) {
 	case 0x80: /* Hotkeys and some system events */
+		/*
+		 * Machines with this WMI guid aren't supported due to bugs in
+		 * their AML.
+		 *
+		 * Return silently to avoid triggering a netlink event.
+		 */
+		if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
+			return;
 		toshiba_acpi_process_hotkeys(dev);
 		break;
 	case 0x81: /* Dock events */

drivers/platform/x86/toshiba_acpi.c static int __init toshiba_acpi_init(void) · line 2820

Quoted 5 lines
drivers/platform/x86/toshiba_acpi.c +0 -2
@@ -2816,14 +2829,6 @@ static int __init toshiba_acpi_init(void)
 {
 	int ret;
 
-	/*
-	 * Machines with this WMI guid aren't supported due to bugs in

Good idea to capitalize acronyms like GUID.

Quoted 13 lines
drivers/platform/x86/toshiba_acpi.c +0 -6
-	 * their AML. This check relies on wmi initializing before
-	 * toshiba_acpi to guarantee guids have been identified.
-	 */
-	if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
-		return -ENODEV;
-
 	toshiba_proc_dir = proc_mkdir(PROC_TOSHIBA, acpi_root_dir);
 	if (!toshiba_proc_dir) {
 		pr_err("Unable to create proc dir " PROC_TOSHIBA "\n");
Signature · 1 line
2.4.3

Functional content looks good. No need to resend, just please keep the above in mind for the future. I'll make the minor updates and merge all three after your response to the first regarding user:kernel interface - unless of course you choose to resend the 3 at the same time, and I'll pick those up.

Thanks Azael,

Signature · 2 lines
Darren Hart
Intel Open Source Technology Center
author maintainer
#2 lore raw

Hi Darren,

2015-07-20 15:55 GMT-06:00 Darren Hart <[email protected]>:

Quoted 79 lines · Azael Avalos
Quoted 8 lines

Commit f11f999e9890 ("toshiba_acpi: Refuse to load on machines with buggy INFO implementations") denied loading on laptops with a WMI Event GUID given that such laptops manage the hotkeys via that interface, however, such laptops have a working Toshiba Configuration Interface (TCI), and thus, such commit denied several supported features.

This patch avoids registering the input device and ignores all hotkey events on laptops with such WMI Event GUI, making the supported

GUID

Quoted 12 lines

features found in those laptops to work.

---
 drivers/platform/x86/toshiba_acpi.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)
drivers/platform/x86/toshiba_acpi.c +0 -0
index 15d8f0d..a3c6ea2 100644

drivers/platform/x86/toshiba_acpi.c static int toshiba_acpi_setup_keyboard(struct toshiba_acpi_dev *dev) · line 2401

Quoted 5 lines
drivers/platform/x86/toshiba_acpi.c +2 -0
@@ -2397,6 +2397,11 @@ static int toshiba_acpi_setup_keyboard(struct toshiba_acpi_dev *dev)
      u32 hci_result;
      int error;
 
+     if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID)) {
+             pr_info("WMI event detected, hotkeys won't be monitored\n");

It's best practice to avoid contractions in any form of technical writing. (See what I did there? ;-)

While this is good for comments, I feel it's more important for kernel messages.

Quoted 22 lines
drivers/platform/x86/toshiba_acpi.c +11 -0
+             return 0;
+     }
+
      error = toshiba_acpi_enable_hotkeys(dev);
      if (error)
              return error;
@@ -2730,6 +2735,14 @@ static void toshiba_acpi_notify(struct acpi_device *acpi_dev, u32 event)
 
      switch (event) {
      case 0x80: /* Hotkeys and some system events */
+             /*
+              * Machines with this WMI guid aren't supported due to bugs in
+              * their AML.
+              *
+              * Return silently to avoid triggering a netlink event.
+              */
+             if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
+                     return;
              toshiba_acpi_process_hotkeys(dev);
              break;
      case 0x81: /* Dock events */

drivers/platform/x86/toshiba_acpi.c static int __init toshiba_acpi_init(void) · line 2820

Quoted 5 lines
drivers/platform/x86/toshiba_acpi.c +0 -2
@@ -2816,14 +2829,6 @@ static int __init toshiba_acpi_init(void)
 {
      int ret;
 
-     /*
-      * Machines with this WMI guid aren't supported due to bugs in

Good idea to capitalize acronyms like GUID.

Quoted 10 lines
drivers/platform/x86/toshiba_acpi.c +0 -7
-      * their AML. This check relies on wmi initializing before
-      * toshiba_acpi to guarantee guids have been identified.
-      */
-     if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
-             return -ENODEV;
-
      toshiba_proc_dir = proc_mkdir(PROC_TOSHIBA, acpi_root_dir);
      if (!toshiba_proc_dir) {
              pr_err("Unable to create proc dir " PROC_TOSHIBA "\n");
--
Quoted 3 lines

2.4.3

Quoted 4 lines

Functional content looks good. No need to resend, just please keep the above in mind for the future. I'll make the minor updates and merge all three after your response to the first regarding user:kernel interface - unless of course you choose to resend the 3 at the same time, and I'll pick those up.

OK, will keep those things in mind, will wait for comments on first patch and send a v2.

Quoted 6 lines

Thanks Azael,

--
Darren Hart
Intel Open Source Technology Center

Cheers
Azael

Signature · 1 line
-- El mundo apesta y vosotros apestais tambien --
#3 lore raw
Quoted 82 lines · Azael Avalos

Hi Darren,

2015-07-20 15:55 GMT-06:00 Darren Hart <[email protected]>:

Quoted 79 lines · Azael Avalos
Quoted 8 lines

Commit f11f999e9890 ("toshiba_acpi: Refuse to load on machines with buggy INFO implementations") denied loading on laptops with a WMI Event GUID given that such laptops manage the hotkeys via that interface, however, such laptops have a working Toshiba Configuration Interface (TCI), and thus, such commit denied several supported features.

This patch avoids registering the input device and ignores all hotkey events on laptops with such WMI Event GUI, making the supported

GUID

Quoted 12 lines

features found in those laptops to work.

---
 drivers/platform/x86/toshiba_acpi.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)
drivers/platform/x86/toshiba_acpi.c +0 -0
index 15d8f0d..a3c6ea2 100644

drivers/platform/x86/toshiba_acpi.c static int toshiba_acpi_setup_keyboard(struct toshiba_acpi_dev *dev) · line 2401

Quoted 5 lines
drivers/platform/x86/toshiba_acpi.c +2 -0
@@ -2397,6 +2397,11 @@ static int toshiba_acpi_setup_keyboard(struct toshiba_acpi_dev *dev)
      u32 hci_result;
      int error;
 
+     if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID)) {
+             pr_info("WMI event detected, hotkeys won't be monitored\n");

It's best practice to avoid contractions in any form of technical writing. (See what I did there? ;-)

While this is good for comments, I feel it's more important for kernel messages.

Quoted 22 lines
drivers/platform/x86/toshiba_acpi.c +11 -0
+             return 0;
+     }
+
      error = toshiba_acpi_enable_hotkeys(dev);
      if (error)
              return error;
@@ -2730,6 +2735,14 @@ static void toshiba_acpi_notify(struct acpi_device *acpi_dev, u32 event)
 
      switch (event) {
      case 0x80: /* Hotkeys and some system events */
+             /*
+              * Machines with this WMI guid aren't supported due to bugs in
+              * their AML.
+              *
+              * Return silently to avoid triggering a netlink event.
+              */
+             if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
+                     return;
              toshiba_acpi_process_hotkeys(dev);
              break;
      case 0x81: /* Dock events */

drivers/platform/x86/toshiba_acpi.c static int __init toshiba_acpi_init(void) · line 2820

Quoted 5 lines
drivers/platform/x86/toshiba_acpi.c +0 -2
@@ -2816,14 +2829,6 @@ static int __init toshiba_acpi_init(void)
 {
      int ret;
 
-     /*
-      * Machines with this WMI guid aren't supported due to bugs in

Good idea to capitalize acronyms like GUID.

Quoted 13 lines
drivers/platform/x86/toshiba_acpi.c +0 -7
-      * their AML. This check relies on wmi initializing before
-      * toshiba_acpi to guarantee guids have been identified.
-      */
-     if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
-             return -ENODEV;
-
      toshiba_proc_dir = proc_mkdir(PROC_TOSHIBA, acpi_root_dir);
      if (!toshiba_proc_dir) {
              pr_err("Unable to create proc dir " PROC_TOSHIBA "\n");
--

2.4.3

Quoted 8 lines
Quoted 4 lines

Functional content looks good. No need to resend, just please keep the above in mind for the future. I'll make the minor updates and merge all three after your response to the first regarding user:kernel interface - unless of course you choose to resend the 3 at the same time, and I'll pick those up.

OK, will keep those things in mind, will wait for comments on first patch and send a v2.

I believe I've responded to all the patches in my queue from you. If I'm missing one, please let me know which one. I have 4 pending from you. If you send a v2, would you just send all 4 to avoid any confusion?

Thanks,

Signature · 2 lines
Darren Hart
Intel Open Source Technology Center
author maintainer
#4 lore raw

Hi Darren,

2015-07-22 14:54 GMT-06:00 Darren Hart <[email protected]>:

Quoted 88 lines · Azael Avalos
Quoted 87 lines

Hi Darren,

2015-07-20 15:55 GMT-06:00 Darren Hart <[email protected]>:

Quoted 83 lines · Azael Avalos
Quoted 8 lines

Commit f11f999e9890 ("toshiba_acpi: Refuse to load on machines with buggy INFO implementations") denied loading on laptops with a WMI Event GUID given that such laptops manage the hotkeys via that interface, however, such laptops have a working Toshiba Configuration Interface (TCI), and thus, such commit denied several supported features.

This patch avoids registering the input device and ignores all hotkey events on laptops with such WMI Event GUI, making the supported

GUID

Quoted 12 lines

features found in those laptops to work.

---
 drivers/platform/x86/toshiba_acpi.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)
drivers/platform/x86/toshiba_acpi.c +0 -0
index 15d8f0d..a3c6ea2 100644

drivers/platform/x86/toshiba_acpi.c static int toshiba_acpi_setup_keyboard(struct toshiba_acpi_dev *dev) · line 2401

Quoted 5 lines
drivers/platform/x86/toshiba_acpi.c +2 -0
@@ -2397,6 +2397,11 @@ static int toshiba_acpi_setup_keyboard(struct toshiba_acpi_dev *dev)
      u32 hci_result;
      int error;
 
+     if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID)) {
+             pr_info("WMI event detected, hotkeys won't be monitored\n");

It's best practice to avoid contractions in any form of technical writing. (See what I did there? ;-)

While this is good for comments, I feel it's more important for kernel messages.

Quoted 22 lines
drivers/platform/x86/toshiba_acpi.c +11 -0
+             return 0;
+     }
+
      error = toshiba_acpi_enable_hotkeys(dev);
      if (error)
              return error;
@@ -2730,6 +2735,14 @@ static void toshiba_acpi_notify(struct acpi_device *acpi_dev, u32 event)
 
      switch (event) {
      case 0x80: /* Hotkeys and some system events */
+             /*
+              * Machines with this WMI guid aren't supported due to bugs in
+              * their AML.
+              *
+              * Return silently to avoid triggering a netlink event.
+              */
+             if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
+                     return;
              toshiba_acpi_process_hotkeys(dev);
              break;
      case 0x81: /* Dock events */

drivers/platform/x86/toshiba_acpi.c static int __init toshiba_acpi_init(void) · line 2820

Quoted 5 lines
drivers/platform/x86/toshiba_acpi.c +0 -2
@@ -2816,14 +2829,6 @@ static int __init toshiba_acpi_init(void)
 {
      int ret;
 
-     /*
-      * Machines with this WMI guid aren't supported due to bugs in

Good idea to capitalize acronyms like GUID.

Quoted 10 lines
drivers/platform/x86/toshiba_acpi.c +0 -7
-      * their AML. This check relies on wmi initializing before
-      * toshiba_acpi to guarantee guids have been identified.
-      */
-     if (wmi_has_guid(TOSHIBA_WMI_EVENT_GUID))
-             return -ENODEV;
-
      toshiba_proc_dir = proc_mkdir(PROC_TOSHIBA, acpi_root_dir);
      if (!toshiba_proc_dir) {
              pr_err("Unable to create proc dir " PROC_TOSHIBA "\n");
--
Quoted 3 lines

2.4.3

Functional content looks good. No need to resend, just please keep the above in mind for the future. I'll make the minor updates and merge all three after your response to the first regarding user:kernel interface - unless of course you choose to resend the 3 at the same time, and I'll pick those up.

Quoted 7 lines
Quoted 3 lines

OK, will keep those things in mind, will wait for comments on first patch and send a v2.

I believe I've responded to all the patches in my queue from you. If I'm missing one, please let me know which one. I have 4 pending from you. If you send a v2, would you just send all 4 to avoid any confusion?

I was waiting for some feedback on the first patch, tho' if there is no problem with
it, I'll send v2 in a few for you to queue, plus a few others for-review.

Quoted 6 lines

Thanks,

--
Darren Hart
Intel Open Source Technology Center

Cheers
Azael

Signature · 1 line
-- El mundo apesta y vosotros apestais tambien --