asus-wmi: add support for ROG XFlow tablet mode
Luke D. Jones
maintainerv2 cover asus-wmi: add support for ROG XFlow tablet mode
This patch series refactors a part of the tablet/lid-flip code to use an enum to set the different behaviours, and adds support for the ASUS ROG XFlow (X13) 2-in-1 laptop.
Changelog:
- V2:
+ Refactor the base handling paths of tablet modes to use enum
+ Add support for ROG XFlow using the refactored code
This obsoletes a previous patch I submitted:
- https://lkml.org/lkml/2022/8/3/94
Luke D. Jones (2):
asus-wmi: Adjust tablet/lidflip handling to use enum
asus-wmi: Add support for ROG X13 tablet mode
drivers/platform/x86/asus-nb-wmi.c | 30 +++++--- drivers/platform/x86/asus-wmi.c | 89 ++++++++++++++++++---- drivers/platform/x86/asus-wmi.h | 10 ++- include/linux/platform_data/x86/asus-wmi.h | 1 + 4 files changed, 103 insertions(+), 27 deletions(-)
Signature · 1 line
2.37.1
Luke D. Jones
maintainerPATCH v2 1/2 asus-wmi: Adjust tablet/lidflip handling to use enum
Due to multiple types of tablet/lidflip, the existing code for handlign these events is refactored to use an enum for each type.
--- drivers/platform/x86/asus-nb-wmi.c | 13 +++----- drivers/platform/x86/asus-wmi.c | 53 +++++++++++++++++++++--------- drivers/platform/x86/asus-wmi.h | 9 +++-- 3 files changed, 49 insertions(+), 26 deletions(-)
drivers/platform/x86/asus-nb-wmi.c +5 -8
@@ -115,12 +115,12 @@ static struct quirk_entry quirk_asus_forceals = { };static struct quirk_entry quirk_asus_use_kbd_dock_devid = { - .use_kbd_dock_devid = true, + .tablet_switch_mode = asus_wmi_kbd_dock_devid, };static struct quirk_entry quirk_asus_use_lid_flip_devid = { .wmi_backlight_set_devstate = true, - .use_lid_flip_devid = true, + .tablet_switch_mode = asus_wmi_lid_flip_devid, };static int dmi_matched (const struct dmi_system_id *dmi)
@@ -492,16 +492,13 @@ static void asus_nb_wmi_quirks(struct asus_wmi_driver *driver)switch (tablet_mode_sw) {case 0 : - quirks->use_kbd_dock_devid = false; - quirks->use_lid_flip_devid = false; + quirks->tablet_switch_mode = asus_wmi_no_tablet_switch;break ;case 1 : - quirks->use_kbd_dock_devid = true; - quirks->use_lid_flip_devid = false; + quirks->tablet_switch_mode = asus_wmi_kbd_dock_devid;break ;case 2 : - quirks->use_kbd_dock_devid = false; - quirks->use_lid_flip_devid = true; + quirks->tablet_switch_mode = asus_wmi_lid_flip_devid;break ; }
drivers/platform/x86/asus-wmi.c +37 -16
98 lines not on this page · raw message
drivers/platform/x86/asus-wmi.h +7 -2
@@ -25,6 +25,12 @@ struct module;struct key_entry ;struct asus_wmi ; +enum asus_wmi_tablet_switch_mode { + asus_wmi_no_tablet_switch, + asus_wmi_kbd_dock_devid, + asus_wmi_lid_flip_devid, +}; +struct quirk_entry {bool hotplug_wireless;bool scalar_panel_brightness;
@@ -33,8 +39,7 @@ struct quirk_entry {bool wmi_backlight_native;bool wmi_backlight_set_devstate;bool wmi_force_als_set; -bool use_kbd_dock_devid; -bool use_lid_flip_devid; +enum asus_wmi_tablet_switch_mode tablet_switch_mode;int wapf;/* * For machines with AMD graphic chips, it will send out WMI event
Signature · 1 line
2.37.1
Re: [PATCH v2 1/2] asus-wmi: Adjust tablet/lidflip handling to use enum
Quoted 3 lines · Luke D. Jones
Due to multiple types of tablet/lidflip, the existing code for handlign these events is refactored to use an enum for each type.
handling
Can you run a spell checker for your commit messages?
...
To the switch-cases, please add a "default" case to each of them.
Signature · 2 lines
With Best Regards, Andy Shevchenko
Re: [PATCH v2 1/2] asus-wmi: Adjust tablet/lidflip handling to use enum
Hi,
Quoted 5 lines · Andy Shevchenko
Quoted 3 lines · Luke D. Jones
Due to multiple types of tablet/lidflip, the existing code for handlign these events is refactored to use an enum for each type.
Quoted 7 lines
handling
Can you run a spell checker for your commit messages?
...
To the switch-cases, please add a "default" case to each of them.
The switch-cases are on an enum type, so adding a default is not necessary and adding one will actually loose the useful compiler warning about unhandled enum values.
Regards,
Hans
Re: [PATCH v2 1/2] asus-wmi: Adjust tablet/lidflip handling to use enum
Quoted 2 lines · Hans de Goede
Quoted 1 line · Andy Shevchenko
On Mon, Aug 8, 2022 at 5:12 AM Luke D. Jones <[email protected]> wrote:
...
Quoted 5 lines
Quoted 1 line
To the switch-cases, please add a "default" case to each of them.
The switch-cases are on an enum type, so adding a default is not necessary and adding one will actually loose the useful compiler warning about unhandled enum values.
It's good if you can cover all enum values, which usually you can't. enum according to the standard should be located in the type that is enough to keep it and be compatible to a char. This means that the code somewhere else may assign anything to enum (actually enum values are type of int) and without default you can't see the difference here and the compiler probably will be happy. That said, I doubt the usefulness of such a warning. But it's up to you.
Signature · 2 lines
With Best Regards, Andy Shevchenko
Re: [PATCH v2 1/2] asus-wmi: Adjust tablet/lidflip handling to use enum
Hi,
Quoted 12 lines · Andy Shevchenko
Quoted 2 lines · Hans de Goede
Quoted 1 line · Andy Shevchenko
On Mon, Aug 8, 2022 at 5:12 AM Luke D. Jones <[email protected]> wrote:
...
Quoted 5 lines
Quoted 1 line
To the switch-cases, please add a "default" case to each of them.
The switch-cases are on an enum type, so adding a default is not necessary and adding one will actually loose the useful compiler warning about unhandled enum values.
Quoted 7 lines
It's good if you can cover all enum values, which usually you can't. enum according to the standard should be located in the type that is enough to keep it and be compatible to a char. This means that the code somewhere else may assign anything to enum (actually enum values are type of int) and without default you can't see the difference here and the compiler probably will be happy. That said, I doubt the usefulness of such a warning. But it's up to you.
I would prefer to not introduce a default label in this case; in the unexpected case that the value gets set out of the enum range then the switch-case will be a no-op and any added default would also be a no-op, so adding a default gains us nothing.
Regards,
Hans
Luke D. Jones
maintainerPATCH v2 2/2 asus-wmi: Add support for ROG X13 tablet mode
In reply to Luke D. JonesAdd quirk for ASUS ROG X13 Flow 2-in-1 to enable tablet mode with lid flip (all screen rotations).
--- drivers/platform/x86/asus-nb-wmi.c | 17 +++++++++- drivers/platform/x86/asus-wmi.c | 36 ++++++++++++++++++++++ drivers/platform/x86/asus-wmi.h | 1 + include/linux/platform_data/x86/asus-wmi.h | 1 + 4 files changed, 54 insertions(+), 1 deletion(-)
drivers/platform/x86/asus-nb-wmi.c +16 -1
@@ -123,6 +123,11 @@ static struct quirk_entry quirk_asus_use_lid_flip_devid = { .tablet_switch_mode = asus_wmi_lid_flip_devid, }; +static struct quirk_entry quirk_asus_tablet_mode = { + .wmi_backlight_set_devstate = true, + .tablet_switch_mode = asus_wmi_lid_flip_rog_devid, +}; +static int dmi_matched (const struct dmi_system_id *dmi) {pr_info ("Identified laptop model '%s'\n" , dmi->ident);
@@ -471,6 +476,15 @@ static const struct dmi_system_id asus_quirks[] = { }, .driver_data = &quirk_asus_use_lid_flip_devid, }, + { + .callback = dmi_matched, + .ident = "ASUS ROG FLOW X13", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "ASUSTeK COMPUTER INC."), + DMI_MATCH(DMI_PRODUCT_NAME, "GV301Q"), + }, + .driver_data = &quirk_asus_tablet_mode, + }, {}, };
@@ -574,7 +588,8 @@ static const struct key_entry asus_nb_wmi_keymap[] = { { KE_KEY, 0xC4, { KEY_KBDILLUMUP } }, { KE_KEY, 0xC5, { KEY_KBDILLUMDOWN } }, { KE_IGNORE, 0xC6, }, /* Ambient Light Sensor notification */ - { KE_KEY, 0xFA, { KEY_PROG2 } }, /* Lid flip action */ + { KE_KEY, 0xFA, { KEY_PROG2 } }, /* Lid flip action */ + { KE_KEY, 0xBD, { KEY_PROG2 } }, /* Lid flip action on ROG xflow laptops */ { KE_END, 0}, };
drivers/platform/x86/asus-wmi.c +36 -0
@@ -68,6 +68,7 @@ module_param(fnlock_default, bool, 0444);#define NOTIFY_KBD_FBM 0x99#define NOTIFY_KBD_TTP 0xae#define NOTIFY_LID_FLIP 0xfa +#define NOTIFY_LID_FLIP_ROG 0xbd#define ASUS_WMI_FNLOCK_BIOS_DISABLED BIT(0)
@@ -529,6 +530,19 @@ static int asus_wmi_input_init(struct asus_wmi *asus)pr_err ("Error checking for lid-flip: %d\n" , result); }break ; +case asus_wmi_lid_flip_rog_devid: + result =asus_wmi_get_devstate_simple (asus,ASUS_WMI_DEVID_LID_FLIP_ROG ); +if (result <0 ) + asus->driver->quirks->tablet_switch_mode = asus_wmi_no_tablet_switch; +if (result >=0 ) { +input_set_capability (asus->inputdev,EV_SW ,SW_TABLET_MODE ); +input_report_switch (asus->inputdev,SW_TABLET_MODE , result); + }else if (result == -ENODEV ) { +pr_err ("This device has lid-flip-rog quirk but got ENODEV checking it. This is a bug." ); + }else { +pr_err ("Error checking for lid-flip: %d\n" , result); + } +break ; } err =input_register_device (asus->inputdev);
@@ -562,6 +576,16 @@ static void lid_flip_tablet_mode_get_state(struct asus_wmi *asus) } } +static void lid_flip_rog_tablet_mode_get_state (struct asus_wmi *asus) +{ +int result =asus_wmi_get_devstate_simple (asus,ASUS_WMI_DEVID_LID_FLIP_ROG ); + +if (result >=0 ) { +input_report_switch (asus->inputdev,SW_TABLET_MODE , result); +input_sync (asus->inputdev); + } +} +/* dGPU ********************************************************************/ static int dgpu_disable_check_present (struct asus_wmi *asus) {
@@ -3106,6 +3130,12 @@ static void asus_wmi_handle_event_code(int code, struct asus_wmi *asus)return ; }break ; +case asus_wmi_lid_flip_rog_devid: +if (code ==NOTIFY_LID_FLIP_ROG ) { +lid_flip_rog_tablet_mode_get_state (asus); +return ; + } +break ; }if (asus->fan_boost_mode_available && code ==NOTIFY_KBD_FBM ) {
@@ -3747,6 +3777,9 @@ static int asus_hotk_resume(struct device *device)case asus_wmi_lid_flip_devid:lid_flip_tablet_mode_get_state (asus);break ; +case asus_wmi_lid_flip_rog_devid: +lid_flip_rog_tablet_mode_get_state (asus); +break ; }return 0 ;
@@ -3795,6 +3828,9 @@ static int asus_hotk_restore(struct device *device)case asus_wmi_lid_flip_devid:lid_flip_tablet_mode_get_state (asus);break ; +case asus_wmi_lid_flip_rog_devid: +lid_flip_rog_tablet_mode_get_state (asus); +break ; }return 0 ;
drivers/platform/x86/asus-wmi.h +1 -0
@@ -29,6 +29,7 @@ enum asus_wmi_tablet_switch_mode { asus_wmi_no_tablet_switch, asus_wmi_kbd_dock_devid, asus_wmi_lid_flip_devid, + asus_wmi_lid_flip_rog_devid, }; struct quirk_entry {
include/linux/platform_data/x86/asus-wmi.h +1 -0
@@ -64,6 +64,7 @@#define ASUS_WMI_DEVID_PANEL_OD 0x00050019#define ASUS_WMI_DEVID_CAMERA 0x00060013#define ASUS_WMI_DEVID_LID_FLIP 0x00060062 +#define ASUS_WMI_DEVID_LID_FLIP_ROG 0x00060077/* Storage */ #define ASUS_WMI_DEVID_CARDREADER 0x00080013
Signature · 1 line
2.37.1
Re: [PATCH v2 2/2] asus-wmi: Add support for ROG X13 tablet mode
Quoted 3 lines · Luke D. Jones
Add quirk for ASUS ROG X13 Flow 2-in-1 to enable tablet mode with lid flip (all screen rotations).
...
drivers/platform/x86/asus-nb-wmi.c static const struct key_entry asus_nb_wmi_keymap[] = { · line 591
Quoted 2 lines
drivers/platform/x86/asus-nb-wmi.c +1
-1- { KE_KEY, 0xFA, { KEY_PROG2 } }, /* Lid flip action */ + { KE_KEY, 0xFA, { KEY_PROG2 } }, /* Lid flip action */
Have maintainers asked you about this? Otherwise it is irrelevant change.
...
Quoted 1 line
+ pr_err("This device has lid-flip-rog quirk but got ENODEV checking it. This is a bug.");
dev_err() ?
...
Quoted 1 line
+
pr_err("Error checking for lid-flip: %d\n", result);
Ditto.
...
drivers/platform/x86/asus-wmi.c static void lid_flip_tablet_mode_get_state(struct asus_wmi *asus) · line 583
Quoted 5 lines
drivers/platform/x86/asus-wmi.c +5
-0+static void lid_flip_rog_tablet_mode_get_state (struct asus_wmi *asus) +{ +int result =asus_wmi_get_devstate_simple (asus,ASUS_WMI_DEVID_LID_FLIP_ROG ); + +if (result >=0 ) {
First of all, it's better to decouple assignment and definition, and move assignment closer to its user. This is usual pattern.
int result;
result = ...
if (result...)
drivers/platform/x86/asus-wmi.c static void lid_flip_tablet_mode_get_state(struct asus_wmi *asus) · line 586
Quoted 3 lines
drivers/platform/x86/asus-wmi.c +3
-0+input_report_switch (asus->inputdev,SW_TABLET_MODE , result); +input_sync (asus->inputdev); + }
Second, it will look better with standard pattern of checking for errors, i.e.
int result;
if (result < 0)
return;
...
Quoted 1 line
+}
Signature · 2 lines
With Best Regards, Andy Shevchenko
Luke Jones
maintainerRe: [PATCH v2 2/2] asus-wmi: Add support for ROG X13 tablet mode
Hi Andy,
Quoted 7 lines
Quoted 4 lines
- {
KE_KEY,0xFA, {KEY_PROG2} }, /* Lid flip
action */
+ {KE_KEY,0xFA, {KEY_PROG2} }, /* Lid flip action */Have maintainers asked you about this? Otherwise it is irrelevant change.
Fixed
Quoted 6 lines
...
Quoted 2 lines
+ pr_err("This device has lid-flip-rog quirk
but got ENODEV checking it. This is a bug.");
dev_err()?
Okay, changed here and in previous patch to match it.
So that I'm clearer on dev_err(), this doesn't do something like exit the module does it? It's just a more detailed error print?
Quoted 9 lines
Quoted 8 lines
+static void lid_flip_rog_tablet_mode_get_state(struct
asus_wmi
*asus)
+{
+ int result = asus_wmi_get_devstate_simple(asus,ASUS_WMI_DEVID_LID_FLIP_ROG);
+
+ if (result >= 0) {
Quoted 3 lines
First of all, it's better to decouple assignment and definition, and move assignment closer to its user. This is usual pattern.
I don't fully understand why you would want the separation given how short these two blocks are (I'll change in this and previous patch of course, I just don't personally understand it).
Cheers,
Luke.
Quoted 1 line
Re: [PATCH v2 2/2] asus-wmi: Add support for ROG X13 tablet mode
On Tue, Aug 9, 2022 at 5:26 AM Luke Jones <[email protected]> wrote:
...
Quoted 9 lines
Quoted 4 lines
Quoted 2 lines
+ pr_err("This device has lid-flip-rog quirk
but got ENODEV checking it. This is a bug.");
dev_err()?Okay, changed here and in previous patch to match it.
So that I'm clearer on
dev_err(), this doesn't do something like exit the module does it? It's just a more detailed error print?
Yes, it's more specific when the user sees it. The pr_err() is global and anonymous (you can only point to the driver, and not the instance of the device bound to it), while dev_err() is device specific and the user will immediately see which device instance is failing. Yet it's not a problem for this particular driver, because I don't believe one may have two, but it's a good coding practice in general.
(Note the last sentence: "good coding practice")
...
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
+static void lid_flip_rog_tablet_mode_get_state(struct
asus_wmi
*asus)
+{
+ int result = asus_wmi_get_devstate_simple(asus,ASUS_WMI_DEVID_LID_FLIP_ROG);
+
+ if (result >= 0) {
Quoted 6 lines
Quoted 2 lines
First of all, it's better to decouple assignment and definition, and move assignment closer to its user. This is usual pattern.
I don't fully understand why you would want the separation given how short these two blocks are (I'll change in this and previous patch of course, I just don't personally understand it).
See above, "good coding practice". Why?
Imagine your code to be in hypothetical v5.10:
int x = foo(param1, param2, ...);
if (x)
return Y;
Now, at v5.12 somebody adds a new feature which touches your code:
int x = foo(param1, param2, ...); struct bar *baz;
if (we_have_such_feature_disabled)
return Z;
if (x)
return Y;
baz = ...
And then somebody else in v5.13 does another feature:
int x = foo(param1, param2, ...); struct bar *baz;
if (we_have_such_feature_disabled)
return Z;
/* parameter 1 can be NULL, check it */
if (!param1)
return -EINVAL;
if (x)
return Y;
baz = ...
Do you see now an issue? If you emulate this as a sequence of Git changes the last one is easily missing subtle detail. That's why "good coding practice".
Signature · 2 lines
With Best Regards, Andy Shevchenko
Luke Jones
maintainerRe: [PATCH v2 2/2] asus-wmi: Add support for ROG X13 tablet mode
Quoted 84 lines · Andy Shevchenko
On Tue, Aug 9, 2022 at 5:26 AM Luke Jones <[email protected]> wrote:
...
Quoted 11 lines
Quoted 1 line
Quoted 1 line
+ pr_err("This device has lid-flip-rog
quirk
Quoted 3 lines
Quoted 1 line
but got ENODEV checking it. This is a bug.");
dev_err()?Okay, changed here and in previous patch to match it.
So that I'm clearer on
dev_err(), this doesn't do something like
exit
the module does it? It's just a more detailed error print?Yes, it's more specific when the user sees it. The
pr_err()is global and anonymous (you can only point to the driver, and not the instance of the device bound to it), whiledev_err()is device specific and the user will immediately see which device instance is failing. Yet it's not a problem for this particular driver, because I don't believe one may have two, but it's a good coding practice in general.(Note the last sentence: "good coding practice")
...
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
+static void lid_flip_rog_tablet_mode_get_state(struct
asus_wmi
*asus)
+{
+ int result = asus_wmi_get_devstate_simple(asus,ASUS_WMI_DEVID_LID_FLIP_ROG);
+
+ if (result >= 0) {Quoted 8 lines
Quoted 1 line
First of all, it's better to decouple assignment and definition,
and
Quoted 1 line
move assignment closer to its user. This is usual pattern.
I don't fully understand why you would want the separation given how short these two blocks are (I'll change in this and previous patch
of
course, I just don't personally understand it).See above, "good coding practice". Why?
Imagine your code to be in hypothetical v5.10:
int x = foo(param1, param2, ...);if (x)
return Y;Now, at v5.12 somebody adds a new feature which touches your code:
int x = foo(param1, param2, ...); struct bar *baz;if (
we_have_such_feature_disabled)return Z;if (x)
return Y;baz = ...
And then somebody else in v5.13 does another feature:
int x = foo(param1, param2, ...); struct bar *baz;if (
we_have_such_feature_disabled)return Z;/* parameter 1 can be NULL, check it */
if (!param1)return -EINVAL;if (x)
return Y;baz = ...
Quoted 7 lines
Do you see now an issue? If you emulate this as a sequence of Git changes the last one is easily missing subtle detail. That's why "good coding practice".
--
With Best Regards,
Andy Shevchenko
That's a great example! Thanks mate, really appreciate it.
Quoted 5 lines · Luke D. Jones
This patch series refactors a part of the tablet/lid-flip code to use an enum to set the different behaviours, and adds support for the ASUS ROG XFlow (X13) 2-in-1 laptop.
Quoted 7 lines
Changelog:
- V2:
+ Refactor the base handling paths of tablet modes to use enum
+ Add support for ROG XFlow using the refactored codeThis obsoletes a previous patch I submitted:
- https://lkml.org/lkml/2022/8/3/94
Good written cover letter, thanks!
Signature · 2 lines
With Best Regards, Andy Shevchenko