| Message ID | 20230330194644.64628-1-hdegoede@redhat.com (mailing list archive) |
|---|---|
| State | Accepted, archived |
| Headers | show |
| Series | platform/x86: ideapad-laptop: Stop sending KEY_TOUCHPAD_TOGGLE | expand |
On Thu, 30 Mar 2023 at 21:46:44 +0200, Hans de Goede wrote: > Commit 5829f8a897e4 ("platform/x86: ideapad-laptop: Send > KEY_TOUCHPAD_TOGGLE on some models") made ideapad-laptop send > KEY_TOUCHPAD_TOGGLE when we receive an ACPI notify with VPC event bit 5 set > and the touchpad-state has not been changed by the EC itself already. > > This was done under the assumption that this would be good to do to make > the touchpad-toggle hotkey work on newer models where the EC does not > toggle the touchpad on/off itself (because it is not routed through > the PS/2 controller, but uses I2C). > > But it turns out that at least some models, e.g. the Yoga 7-15ITL5 the EC > triggers an ACPI notify with VPC event bit 5 set on resume, which would > now cause a spurious KEY_TOUCHPAD_TOGGLE on resume to which the desktop > environment responds by disabling the touchpad in software, breaking > the touchpad (until manually re-enabled) on resume. Oh gosh, the touchpad toggle on Ideapads is so much broken, I wonder how the Windows driver deals with all this variety of different behaviors (unless it's broken too :D). I'll test the patch on Z570, but as I see, it shouldn't change anything for Z570. > It was never confirmed that sending KEY_TOUCHPAD_TOGGLE actually improves > things on new models and at least some new models like the Yoga 7-15ITL5 > don't have a touchpad on/off toggle hotkey at all, while still sending > ACPI notify events with VPC event bit 5 set. > > So it seems best to revert the change to send KEY_TOUCHPAD_TOGGLE when > receiving an ACPI notify events with VPC event bit 5 and the touchpad > state as reported by the EC has not changed. > > Note this is not a full revert the code to cache the last EC touchpad > state is kept to avoid sending spurious KEY_TOUCHPAD_ON / _OFF events > on resume. > > Fixes: 5829f8a897e4 ("platform/x86: ideapad-laptop: Send KEY_TOUCHPAD_TOGGLE on some models") > Link: https://bugzilla.kernel.org/show_bug.cgi?id=217234 > Cc: stable@vger.kernel.org > Signed-off-by: Hans de Goede <hdegoede@redhat.com> > --- > drivers/platform/x86/ideapad-laptop.c | 23 ++++++++++------------- > 1 file changed, 10 insertions(+), 13 deletions(-) > > diff --git a/drivers/platform/x86/ideapad-laptop.c b/drivers/platform/x86/ideapad-laptop.c > index b5ef3452da1f..35c63cce0479 100644 > --- a/drivers/platform/x86/ideapad-laptop.c > +++ b/drivers/platform/x86/ideapad-laptop.c > @@ -1170,7 +1170,6 @@ static const struct key_entry ideapad_keymap[] = { > { KE_KEY, 65, { KEY_PROG4 } }, > { KE_KEY, 66, { KEY_TOUCHPAD_OFF } }, > { KE_KEY, 67, { KEY_TOUCHPAD_ON } }, > - { KE_KEY, 68, { KEY_TOUCHPAD_TOGGLE } }, > { KE_KEY, 128, { KEY_ESC } }, > > /* > @@ -1526,18 +1525,16 @@ static void ideapad_sync_touchpad_state(struct ideapad_private *priv, bool send_ > if (priv->features.ctrl_ps2_aux_port) > i8042_command(¶m, value ? I8042_CMD_AUX_ENABLE : I8042_CMD_AUX_DISABLE); > > - if (send_events) { > - /* > - * On older models the EC controls the touchpad and toggles it > - * on/off itself, in this case we report KEY_TOUCHPAD_ON/_OFF. > - * If the EC did not toggle, report KEY_TOUCHPAD_TOGGLE. > - */ > - if (value != priv->r_touchpad_val) { > - ideapad_input_report(priv, value ? 67 : 66); > - sysfs_notify(&priv->platform_device->dev.kobj, NULL, "touchpad"); > - } else { > - ideapad_input_report(priv, 68); > - } > + /* > + * On older models the EC controls the touchpad and toggles it on/off > + * itself, in this case we report KEY_TOUCHPAD_ON/_OFF. Some models do > + * an acpi-notify with VPC bit 5 set on resume, so this function get > + * called with send_events=true on every resume. Therefor if the EC did > + * not toggle, do nothing to avoid sending spurious KEY_TOUCHPAD_TOGGLE. > + */ > + if (send_events && value != priv->r_touchpad_val) { > + ideapad_input_report(priv, value ? 67 : 66); > + sysfs_notify(&priv->platform_device->dev.kobj, NULL, "touchpad"); > } > > priv->r_touchpad_val = value; > -- > 2.39.1 >
Hi, On 3/30/23 21:46, Hans de Goede wrote: > Commit 5829f8a897e4 ("platform/x86: ideapad-laptop: Send > KEY_TOUCHPAD_TOGGLE on some models") made ideapad-laptop send > KEY_TOUCHPAD_TOGGLE when we receive an ACPI notify with VPC event bit 5 set > and the touchpad-state has not been changed by the EC itself already. > > This was done under the assumption that this would be good to do to make > the touchpad-toggle hotkey work on newer models where the EC does not > toggle the touchpad on/off itself (because it is not routed through > the PS/2 controller, but uses I2C). > > But it turns out that at least some models, e.g. the Yoga 7-15ITL5 the EC > triggers an ACPI notify with VPC event bit 5 set on resume, which would > now cause a spurious KEY_TOUCHPAD_TOGGLE on resume to which the desktop > environment responds by disabling the touchpad in software, breaking > the touchpad (until manually re-enabled) on resume. > > It was never confirmed that sending KEY_TOUCHPAD_TOGGLE actually improves > things on new models and at least some new models like the Yoga 7-15ITL5 > don't have a touchpad on/off toggle hotkey at all, while still sending > ACPI notify events with VPC event bit 5 set. > > So it seems best to revert the change to send KEY_TOUCHPAD_TOGGLE when > receiving an ACPI notify events with VPC event bit 5 and the touchpad > state as reported by the EC has not changed. > > Note this is not a full revert the code to cache the last EC touchpad > state is kept to avoid sending spurious KEY_TOUCHPAD_ON / _OFF events > on resume. > > Fixes: 5829f8a897e4 ("platform/x86: ideapad-laptop: Send KEY_TOUCHPAD_TOGGLE on some models") > Link: https://bugzilla.kernel.org/show_bug.cgi?id=217234 > Cc: stable@vger.kernel.org > Signed-off-by: Hans de Goede <hdegoede@redhat.com> I've applied this patch to my fixes branch now: https://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86.git/log/?h=fixes Note it will show up in my fixes branch once I've pushed my local branch there, which might take a while. I will include this patch in my next fixes pull-req to Linus for the current kernel development cycle. Regards, Hans > --- > drivers/platform/x86/ideapad-laptop.c | 23 ++++++++++------------- > 1 file changed, 10 insertions(+), 13 deletions(-) > > diff --git a/drivers/platform/x86/ideapad-laptop.c b/drivers/platform/x86/ideapad-laptop.c > index b5ef3452da1f..35c63cce0479 100644 > --- a/drivers/platform/x86/ideapad-laptop.c > +++ b/drivers/platform/x86/ideapad-laptop.c > @@ -1170,7 +1170,6 @@ static const struct key_entry ideapad_keymap[] = { > { KE_KEY, 65, { KEY_PROG4 } }, > { KE_KEY, 66, { KEY_TOUCHPAD_OFF } }, > { KE_KEY, 67, { KEY_TOUCHPAD_ON } }, > - { KE_KEY, 68, { KEY_TOUCHPAD_TOGGLE } }, > { KE_KEY, 128, { KEY_ESC } }, > > /* > @@ -1526,18 +1525,16 @@ static void ideapad_sync_touchpad_state(struct ideapad_private *priv, bool send_ > if (priv->features.ctrl_ps2_aux_port) > i8042_command(¶m, value ? I8042_CMD_AUX_ENABLE : I8042_CMD_AUX_DISABLE); > > - if (send_events) { > - /* > - * On older models the EC controls the touchpad and toggles it > - * on/off itself, in this case we report KEY_TOUCHPAD_ON/_OFF. > - * If the EC did not toggle, report KEY_TOUCHPAD_TOGGLE. > - */ > - if (value != priv->r_touchpad_val) { > - ideapad_input_report(priv, value ? 67 : 66); > - sysfs_notify(&priv->platform_device->dev.kobj, NULL, "touchpad"); > - } else { > - ideapad_input_report(priv, 68); > - } > + /* > + * On older models the EC controls the touchpad and toggles it on/off > + * itself, in this case we report KEY_TOUCHPAD_ON/_OFF. Some models do > + * an acpi-notify with VPC bit 5 set on resume, so this function get > + * called with send_events=true on every resume. Therefor if the EC did > + * not toggle, do nothing to avoid sending spurious KEY_TOUCHPAD_TOGGLE. > + */ > + if (send_events && value != priv->r_touchpad_val) { > + ideapad_input_report(priv, value ? 67 : 66); > + sysfs_notify(&priv->platform_device->dev.kobj, NULL, "touchpad"); > } > > priv->r_touchpad_val = value;
On Fri, 31 Mar 2023 at 12:51:30 +0300, Maxim Mikityanskiy wrote: > On Thu, 30 Mar 2023 at 21:46:44 +0200, Hans de Goede wrote: > > Commit 5829f8a897e4 ("platform/x86: ideapad-laptop: Send > > KEY_TOUCHPAD_TOGGLE on some models") made ideapad-laptop send > > KEY_TOUCHPAD_TOGGLE when we receive an ACPI notify with VPC event bit 5 set > > and the touchpad-state has not been changed by the EC itself already. > > > > This was done under the assumption that this would be good to do to make > > the touchpad-toggle hotkey work on newer models where the EC does not > > toggle the touchpad on/off itself (because it is not routed through > > the PS/2 controller, but uses I2C). > > > > But it turns out that at least some models, e.g. the Yoga 7-15ITL5 the EC > > triggers an ACPI notify with VPC event bit 5 set on resume, which would > > now cause a spurious KEY_TOUCHPAD_TOGGLE on resume to which the desktop > > environment responds by disabling the touchpad in software, breaking > > the touchpad (until manually re-enabled) on resume. > > Oh gosh, the touchpad toggle on Ideapads is so much broken, I wonder how > the Windows driver deals with all this variety of different behaviors > (unless it's broken too :D). > > I'll test the patch on Z570, but as I see, it shouldn't change anything > for Z570. Tested the kernel from your branch on Z570, the touchpad button still works fine. > > It was never confirmed that sending KEY_TOUCHPAD_TOGGLE actually improves > > things on new models and at least some new models like the Yoga 7-15ITL5 > > don't have a touchpad on/off toggle hotkey at all, while still sending > > ACPI notify events with VPC event bit 5 set. > > > > So it seems best to revert the change to send KEY_TOUCHPAD_TOGGLE when > > receiving an ACPI notify events with VPC event bit 5 and the touchpad > > state as reported by the EC has not changed. > > > > Note this is not a full revert the code to cache the last EC touchpad > > state is kept to avoid sending spurious KEY_TOUCHPAD_ON / _OFF events > > on resume. > > > > Fixes: 5829f8a897e4 ("platform/x86: ideapad-laptop: Send KEY_TOUCHPAD_TOGGLE on some models") > > Link: https://bugzilla.kernel.org/show_bug.cgi?id=217234 > > Cc: stable@vger.kernel.org > > Signed-off-by: Hans de Goede <hdegoede@redhat.com> > > --- > > drivers/platform/x86/ideapad-laptop.c | 23 ++++++++++------------- > > 1 file changed, 10 insertions(+), 13 deletions(-) > > > > diff --git a/drivers/platform/x86/ideapad-laptop.c b/drivers/platform/x86/ideapad-laptop.c > > index b5ef3452da1f..35c63cce0479 100644 > > --- a/drivers/platform/x86/ideapad-laptop.c > > +++ b/drivers/platform/x86/ideapad-laptop.c > > @@ -1170,7 +1170,6 @@ static const struct key_entry ideapad_keymap[] = { > > { KE_KEY, 65, { KEY_PROG4 } }, > > { KE_KEY, 66, { KEY_TOUCHPAD_OFF } }, > > { KE_KEY, 67, { KEY_TOUCHPAD_ON } }, > > - { KE_KEY, 68, { KEY_TOUCHPAD_TOGGLE } }, > > { KE_KEY, 128, { KEY_ESC } }, > > > > /* > > @@ -1526,18 +1525,16 @@ static void ideapad_sync_touchpad_state(struct ideapad_private *priv, bool send_ > > if (priv->features.ctrl_ps2_aux_port) > > i8042_command(¶m, value ? I8042_CMD_AUX_ENABLE : I8042_CMD_AUX_DISABLE); > > > > - if (send_events) { > > - /* > > - * On older models the EC controls the touchpad and toggles it > > - * on/off itself, in this case we report KEY_TOUCHPAD_ON/_OFF. > > - * If the EC did not toggle, report KEY_TOUCHPAD_TOGGLE. > > - */ > > - if (value != priv->r_touchpad_val) { > > - ideapad_input_report(priv, value ? 67 : 66); > > - sysfs_notify(&priv->platform_device->dev.kobj, NULL, "touchpad"); > > - } else { > > - ideapad_input_report(priv, 68); > > - } > > + /* > > + * On older models the EC controls the touchpad and toggles it on/off > > + * itself, in this case we report KEY_TOUCHPAD_ON/_OFF. Some models do > > + * an acpi-notify with VPC bit 5 set on resume, so this function get > > + * called with send_events=true on every resume. Therefor if the EC did > > + * not toggle, do nothing to avoid sending spurious KEY_TOUCHPAD_TOGGLE. > > + */ > > + if (send_events && value != priv->r_touchpad_val) { > > + ideapad_input_report(priv, value ? 67 : 66); > > + sysfs_notify(&priv->platform_device->dev.kobj, NULL, "touchpad"); > > } > > > > priv->r_touchpad_val = value; > > -- > > 2.39.1 > >
Hi, On 4/3/23 22:51, Maxim Mikityanskiy wrote: > On Fri, 31 Mar 2023 at 12:51:30 +0300, Maxim Mikityanskiy wrote: >> On Thu, 30 Mar 2023 at 21:46:44 +0200, Hans de Goede wrote: >>> Commit 5829f8a897e4 ("platform/x86: ideapad-laptop: Send >>> KEY_TOUCHPAD_TOGGLE on some models") made ideapad-laptop send >>> KEY_TOUCHPAD_TOGGLE when we receive an ACPI notify with VPC event bit 5 set >>> and the touchpad-state has not been changed by the EC itself already. >>> >>> This was done under the assumption that this would be good to do to make >>> the touchpad-toggle hotkey work on newer models where the EC does not >>> toggle the touchpad on/off itself (because it is not routed through >>> the PS/2 controller, but uses I2C). >>> >>> But it turns out that at least some models, e.g. the Yoga 7-15ITL5 the EC >>> triggers an ACPI notify with VPC event bit 5 set on resume, which would >>> now cause a spurious KEY_TOUCHPAD_TOGGLE on resume to which the desktop >>> environment responds by disabling the touchpad in software, breaking >>> the touchpad (until manually re-enabled) on resume. >> >> Oh gosh, the touchpad toggle on Ideapads is so much broken, I wonder how >> the Windows driver deals with all this variety of different behaviors >> (unless it's broken too :D). >> >> I'll test the patch on Z570, but as I see, it shouldn't change anything >> for Z570. > > Tested the kernel from your branch on Z570, the touchpad button still > works fine. Thank you for testing. Regards, Hans
diff --git a/drivers/platform/x86/ideapad-laptop.c b/drivers/platform/x86/ideapad-laptop.c index b5ef3452da1f..35c63cce0479 100644 --- a/drivers/platform/x86/ideapad-laptop.c +++ b/drivers/platform/x86/ideapad-laptop.c @@ -1170,7 +1170,6 @@ static const struct key_entry ideapad_keymap[] = { { KE_KEY, 65, { KEY_PROG4 } }, { KE_KEY, 66, { KEY_TOUCHPAD_OFF } }, { KE_KEY, 67, { KEY_TOUCHPAD_ON } }, - { KE_KEY, 68, { KEY_TOUCHPAD_TOGGLE } }, { KE_KEY, 128, { KEY_ESC } }, /* @@ -1526,18 +1525,16 @@ static void ideapad_sync_touchpad_state(struct ideapad_private *priv, bool send_ if (priv->features.ctrl_ps2_aux_port) i8042_command(¶m, value ? I8042_CMD_AUX_ENABLE : I8042_CMD_AUX_DISABLE); - if (send_events) { - /* - * On older models the EC controls the touchpad and toggles it - * on/off itself, in this case we report KEY_TOUCHPAD_ON/_OFF. - * If the EC did not toggle, report KEY_TOUCHPAD_TOGGLE. - */ - if (value != priv->r_touchpad_val) { - ideapad_input_report(priv, value ? 67 : 66); - sysfs_notify(&priv->platform_device->dev.kobj, NULL, "touchpad"); - } else { - ideapad_input_report(priv, 68); - } + /* + * On older models the EC controls the touchpad and toggles it on/off + * itself, in this case we report KEY_TOUCHPAD_ON/_OFF. Some models do + * an acpi-notify with VPC bit 5 set on resume, so this function get + * called with send_events=true on every resume. Therefor if the EC did + * not toggle, do nothing to avoid sending spurious KEY_TOUCHPAD_TOGGLE. + */ + if (send_events && value != priv->r_touchpad_val) { + ideapad_input_report(priv, value ? 67 : 66); + sysfs_notify(&priv->platform_device->dev.kobj, NULL, "touchpad"); } priv->r_touchpad_val = value;
Commit 5829f8a897e4 ("platform/x86: ideapad-laptop: Send KEY_TOUCHPAD_TOGGLE on some models") made ideapad-laptop send KEY_TOUCHPAD_TOGGLE when we receive an ACPI notify with VPC event bit 5 set and the touchpad-state has not been changed by the EC itself already. This was done under the assumption that this would be good to do to make the touchpad-toggle hotkey work on newer models where the EC does not toggle the touchpad on/off itself (because it is not routed through the PS/2 controller, but uses I2C). But it turns out that at least some models, e.g. the Yoga 7-15ITL5 the EC triggers an ACPI notify with VPC event bit 5 set on resume, which would now cause a spurious KEY_TOUCHPAD_TOGGLE on resume to which the desktop environment responds by disabling the touchpad in software, breaking the touchpad (until manually re-enabled) on resume. It was never confirmed that sending KEY_TOUCHPAD_TOGGLE actually improves things on new models and at least some new models like the Yoga 7-15ITL5 don't have a touchpad on/off toggle hotkey at all, while still sending ACPI notify events with VPC event bit 5 set. So it seems best to revert the change to send KEY_TOUCHPAD_TOGGLE when receiving an ACPI notify events with VPC event bit 5 and the touchpad state as reported by the EC has not changed. Note this is not a full revert the code to cache the last EC touchpad state is kept to avoid sending spurious KEY_TOUCHPAD_ON / _OFF events on resume. Fixes: 5829f8a897e4 ("platform/x86: ideapad-laptop: Send KEY_TOUCHPAD_TOGGLE on some models") Link: https://bugzilla.kernel.org/show_bug.cgi?id=217234 Cc: stable@vger.kernel.org Signed-off-by: Hans de Goede <hdegoede@redhat.com> --- drivers/platform/x86/ideapad-laptop.c | 23 ++++++++++------------- 1 file changed, 10 insertions(+), 13 deletions(-)