Bug Report - Kernel Branch 4.4.y - asus-wmi.c

Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Bob Funk]

Hello,

I am contacting the stable branch maintainers with a bug report concerning
the asus-wmi kernel driver in the 4.4 kernel branch. I had initially 
contacted
maintainers for the specific driver and received a response stating that I
should contact the stable branch maintainers about the issue instead. Their
opinion was that the patch in question should be reverted rather than
debugged. I will append my initial report here and let you decide what to do
with the bug.


Original Bug Report:

The 2019-01-26 commit to the asus-wmi.c driver code in the 4.4 kernel
branch has introduced a bug with several known models of EeePC netbooks.

Description of Problem:
The bug occurs during boot, where the screen (possibly backlight?) will
shut off and display hotkeys are unable to bring it back on. The problem
is present on all kernels since the 2019-01-26 commit. There have been
several reports of the issue in the slackware forum at linuxquestions.org

Corrective actions taken so far:
Appending acpi_osi=Linux will circumvent the issue and keep the screen
on, but this causes several error messages
in the boot log about eeepc_wmi "failing to load both WMI and and legacy
ATKD devices", and warns not to use acpi_osi=Linux.

Appending acpi_backlight=vendor also prevents the screen from shutting
off during boot. However, pressing the brightness hotkeys
causes the system to hang.

Reversing the asus-wmi.c patch that was committed on 2019-01-26 and
rebuilding the 4.4 series module also fixes the problem, and brightness
hotkeys work normally. The commit in question is
0c4a25cc6f2934f3aa99a0bbfd20b71949bcad25

Model I have tested this on: ASUS EeePC 1000H (Slackware 14.2, kernels
4.4.201, 4.4.202)
Additional models reporting this issue: ASUS Eee PC 1005HAB, ASUS Eee PC
1225b, ASUS Eee PC 1025c (Slackware 14.2, various kernels from 4.4.172
and higher)

Additional Notes:
This problem seems to have been corrected in the 4.19 kernel branch, as
reported by several users in the slackware forum.
I attempted to test some of the fixes from the 4.19 code as patches to
the 4.4 code but had no success. There have been multiple
changes in that branch and I am unsure what exactly has corrected the
bug in that version.

If there is any additional information that I can provide, please let me
know.

Regards,

Bob Funk

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Willy Tarreau]

Hi,

On Fri, Nov 22, 2019 at 08:33:46PM -0600, Bob Funk wrote:
> Hello,
> 
> I am contacting the stable branch maintainers with a bug report concerning
> the asus-wmi kernel driver in the 4.4 kernel branch. I had initially
> contacted
> maintainers for the specific driver and received a response stating that I
> should contact the stable branch maintainers about the issue instead. Their
> opinion was that the patch in question should be reverted rather than
> debugged. I will append my initial report here and let you decide what to do
> with the bug.
> 
> 
> Original Bug Report:
> 
> The 2019-01-26 commit to the asus-wmi.c driver code in the 4.4 kernel
> branch has introduced a bug with several known models of EeePC netbooks.
> 
> Description of Problem:
> The bug occurs during boot, where the screen (possibly backlight?) will
> shut off and display hotkeys are unable to bring it back on. The problem
> is present on all kernels since the 2019-01-26 commit. There have been
> several reports of the issue in the slackware forum at linuxquestions.org
> 
> Corrective actions taken so far:
> Appending acpi_osi=Linux will circumvent the issue and keep the screen
> on, but this causes several error messages
> in the boot log about eeepc_wmi "failing to load both WMI and and legacy
> ATKD devices", and warns not to use acpi_osi=Linux.
> 
> Appending acpi_backlight=vendor also prevents the screen from shutting
> off during boot. However, pressing the brightness hotkeys
> causes the system to hang.
> 
> Reversing the asus-wmi.c patch that was committed on 2019-01-26 and
> rebuilding the 4.4 series module also fixes the problem, and brightness
> hotkeys work normally. The commit in question is
> 0c4a25cc6f2934f3aa99a0bbfd20b71949bcad25
> 
> Model I have tested this on: ASUS EeePC 1000H (Slackware 14.2, kernels
> 4.4.201, 4.4.202)
> Additional models reporting this issue: ASUS Eee PC 1005HAB, ASUS Eee PC
> 1225b, ASUS Eee PC 1025c (Slackware 14.2, various kernels from 4.4.172
> and higher)
> 
> Additional Notes:
> This problem seems to have been corrected in the 4.19 kernel branch, as
> reported by several users in the slackware forum.
> I attempted to test some of the fixes from the 4.19 code as patches to
> the 4.4 code but had no success. There have been multiple
> changes in that branch and I am unsure what exactly has corrected the
> bug in that version.
> 
> If there is any additional information that I can provide, please let me
> know.

I suspect that this is caused by missing commit 401fee81, which fixes
78f3ac76d9e5, was backported to 4.19 but not to 4.4. However you will
have to backport it by hand as it doesn't apply due to context
differences, but it trivial to do.

I am also running updated kernels on a 1025C and do not experience such
issues (5.4-rc8 now, but I used to run 4.4 till 4.4.79, which didn't
contain the backport of 78f3ac76d9e5).

Regards,
Willy

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Willy Tarreau]

[-- Attachment #1: Type: text/plain, Size: 405 bytes --]

On Sat, Nov 23, 2019 at 07:14:46AM +0100, Willy Tarreau wrote:
> I suspect that this is caused by missing commit 401fee81, which fixes
> 78f3ac76d9e5, was backported to 4.19 but not to 4.4. However you will
> have to backport it by hand as it doesn't apply due to context
> differences, but it trivial to do.

Please try the attached patch. I haven't even tried to compile it but
I think it's OK.

Willy


[-- Attachment #2: 0001-platform-x86-asus-wmi-Only-Tell-EC-the-OS-will-handl.patch --]
[-- Type: text/plain, Size: 3312 bytes --]

From cfc80c044499f17e82c3ede448a4e1462572ae39 Mon Sep 17 00:00:00 2001
From: Hans de Goede <hdegoede@redhat.com>
Date: Wed, 12 Jun 2019 09:02:02 +0200
Subject: platform/x86: asus-wmi: Only Tell EC the OS will handle display
 hotkeys from asus_nb_wmi
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit

commit 401fee8195d401b2b94dee57383f627050724d5b upstream.

Commit 78f3ac76d9e5 ("platform/x86: asus-wmi: Tell the EC the OS will
handle the display off hotkey") causes the backlight to be permanently off
on various EeePC laptop models using the eeepc-wmi driver (Asus EeePC
1015BX, Asus EeePC 1025C).

The asus_wmi_set_devstate(ASUS_WMI_DEVID_BACKLIGHT, 2, NULL) call added
by that commit is made conditional in this commit and only enabled in
the quirk_entry structs in the asus-nb-wmi driver fixing the broken
display / backlight on various EeePC laptop models.

Cc: João Paulo Rechi Vita <jprvita@endlessm.com>
Fixes: 78f3ac76d9e5 ("platform/x86: asus-wmi: Tell the EC the OS will handle the display off hotkey")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
[wt: dropped changes for missing quirks]
Signed-off-by: Willy Tarreau <w@1wt.eu>
---
 drivers/platform/x86/asus-nb-wmi.c | 4 ++++
 drivers/platform/x86/asus-wmi.c    | 2 +-
 drivers/platform/x86/asus-wmi.h    | 1 +
 3 files changed, 6 insertions(+), 1 deletion(-)

diff --git a/drivers/platform/x86/asus-nb-wmi.c b/drivers/platform/x86/asus-nb-wmi.c
index a284a2b..bbea208 100644
--- a/drivers/platform/x86/asus-nb-wmi.c
+++ b/drivers/platform/x86/asus-nb-wmi.c
@@ -57,6 +57,7 @@ static struct quirk_entry *quirks;
 
 static struct quirk_entry quirk_asus_unknown = {
 	.wapf = 0,
+	.wmi_backlight_set_devstate = true,
 };
 
 /*
@@ -67,15 +68,18 @@ static struct quirk_entry quirk_asus_unknown = {
 static struct quirk_entry quirk_asus_x55u = {
 	.wapf = 4,
 	.wmi_backlight_power = true,
+	.wmi_backlight_set_devstate = true,
 	.no_display_toggle = true,
 };
 
 static struct quirk_entry quirk_asus_wapf4 = {
 	.wapf = 4,
+	.wmi_backlight_set_devstate = true,
 };
 
 static struct quirk_entry quirk_asus_x200ca = {
 	.wapf = 2,
+	.wmi_backlight_set_devstate = true,
 };
 
 static int dmi_matched(const struct dmi_system_id *dmi)
diff --git a/drivers/platform/x86/asus-wmi.c b/drivers/platform/x86/asus-wmi.c
index 7c1defa..c4386ae 100644
--- a/drivers/platform/x86/asus-wmi.c
+++ b/drivers/platform/x86/asus-wmi.c
@@ -2084,7 +2084,7 @@ static int asus_wmi_add(struct platform_device *pdev)
 		err = asus_wmi_backlight_init(asus);
 		if (err && err != -ENODEV)
 			goto fail_backlight;
-	} else
+	} else if (asus->driver->quirks->wmi_backlight_set_devstate)
 		err = asus_wmi_set_devstate(ASUS_WMI_DEVID_BACKLIGHT, 2, NULL);
 
 	status = wmi_install_notify_handler(asus->driver->event_guid,
diff --git a/drivers/platform/x86/asus-wmi.h b/drivers/platform/x86/asus-wmi.h
index 4da4c8b..0f565cc 100644
--- a/drivers/platform/x86/asus-wmi.h
+++ b/drivers/platform/x86/asus-wmi.h
@@ -42,6 +42,7 @@ struct quirk_entry {
 	bool scalar_panel_brightness;
 	bool store_backlight_power;
 	bool wmi_backlight_power;
+	bool wmi_backlight_set_devstate;
 	int wapf;
 	/*
 	 * For machines with AMD graphic chips, it will send out WMI event
-- 
2.9.0

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Greg KH]

On Fri, Nov 22, 2019 at 08:33:46PM -0600, Bob Funk wrote:
> Hello,
> 
> I am contacting the stable branch maintainers with a bug report concerning
> the asus-wmi kernel driver in the 4.4 kernel branch. I had initially
> contacted
> maintainers for the specific driver and received a response stating that I
> should contact the stable branch maintainers about the issue instead. Their
> opinion was that the patch in question should be reverted rather than
> debugged. I will append my initial report here and let you decide what to do
> with the bug.

Any reason you can not just use 5.3 or newer for this hardware?  Why are
you stuck on 4.4?

thanks,

greg k-h

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Bob Funk]

I compiled kernel with the patch you attached and it does indeed correct
the problem as well. Perhaps backporting this patch to the 4.4 kernel
would be a better alternative than reverting as I had originally suggested.

Regards,
Bob Funk


On 11/23/19 12:20 AM, Willy Tarreau wrote:
> On Sat, Nov 23, 2019 at 07:14:46AM +0100, Willy Tarreau wrote:
>> I suspect that this is caused by missing commit 401fee81, which fixes
>> 78f3ac76d9e5, was backported to 4.19 but not to 4.4. However you will
>> have to backport it by hand as it doesn't apply due to context
>> differences, but it trivial to do.
> Please try the attached patch. I haven't even tried to compile it but
> I think it's OK.
>
> Willy
>

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Bob Funk]

Hi Greg,
I have no personal issues against upgrading to a newer kernel myself. The
reason I am working with the 4.4 kernel is that it is the supported 
kernel from
my distro's stable release (slackware). Unfortunately slackware's stable 
release is
several years old and consequently based on the older 4.4 LTS kernel branch.

The result of the distro packaging this older kernel is multiple users 
reporting this
bug to the slackware community so I came to the kernel maintainers to 
see if this
can be fixed upstream. I could also advocate for the slackware 
maintainer to
apply the patch to his releases if you think that's a better route to take.

For the record, the patch that Willy offered does fix the issue on my 
affected
system. That might be a better choice than my request to revert as per the
original email.

Please let me know what you decide.

Thanks,

Bob Funk


On 11/23/19 3:27 AM, Greg KH wrote:
> On Fri, Nov 22, 2019 at 08:33:46PM -0600, Bob Funk wrote:
>> Hello,
>>
>> I am contacting the stable branch maintainers with a bug report concerning
>> the asus-wmi kernel driver in the 4.4 kernel branch. I had initially
>> contacted
>> maintainers for the specific driver and received a response stating that I
>> should contact the stable branch maintainers about the issue instead. Their
>> opinion was that the patch in question should be reverted rather than
>> debugged. I will append my initial report here and let you decide what to do
>> with the bug.
> Any reason you can not just use 5.3 or newer for this hardware?  Why are
> you stuck on 4.4?
>
> thanks,
>
> greg k-h

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Willy Tarreau]

On Sat, Nov 23, 2019 at 01:08:03PM -0600, Bob Funk wrote:
> For the record, the patch that Willy offered does fix the issue on my
> affected
> system. That might be a better choice than my request to revert as per the
> original email.

Greg, FWIW I did nothing more than a regular backport so that you can take
it as-is. I think you dropped it from 4.4 because it did not apply well
and was not worth the hassle, but given that it fixes a regression caused
by another backport I think it makes sense to take it, at least so that
some users do not stop updating. The fix was only merged into 4.19, not
4.4/4.9/4.14.

The backports for 4.9 and 4.14 are easy to do, if you're willing to take
the patches I can do them as well, just let me know.

Cheers,
Willy

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Sasha Levin]

On Sat, Nov 23, 2019 at 08:22:44PM +0100, Willy Tarreau wrote:
>On Sat, Nov 23, 2019 at 01:08:03PM -0600, Bob Funk wrote:
>> For the record, the patch that Willy offered does fix the issue on my
>> affected
>> system. That might be a better choice than my request to revert as per the
>> original email.
>
>Greg, FWIW I did nothing more than a regular backport so that you can take
>it as-is. I think you dropped it from 4.4 because it did not apply well
>and was not worth the hassle, but given that it fixes a regression caused
>by another backport I think it makes sense to take it, at least so that
>some users do not stop updating. The fix was only merged into 4.19, not
>4.4/4.9/4.14.
>
>The backports for 4.9 and 4.14 are easy to do, if you're willing to take
>the patches I can do them as well, just let me know.

Let's try something like this:

For 4.14 and 4.9 I'll also grab db2582afa744 ("platform/x86:
asus-nb-wmi: Support ALS on the Zenbook UX430UQ") which makes 401fee819
apply cleanly.

For 4.4, I'll grab this long list:

92a505e8055f ("platform/x86: asus-wmi: add SERIO_I8042 dependency")
db2582afa744 ("platform/x86: asus-nb-wmi: Support ALS on the Zenbook UX430UQ")
71f38c11cdb8 ("platform/x86: asus-wmi: Set specified XUSB2PR value for X550LB")
999d4376c628 ("platform/x86: asus-wmi: fix asus ux303ub brightness issue")
a961a285b479 ("asus-wmi: Add quirk_no_rfkill_wapf4 for the Asus X456UF")
a977e59c0c67 ("asus-wmi: Create quirk for airplane_mode LED")
6b7ff2af5286 ("asus-wmi: Add quirk_no_rfkill for the Asus Z550MA")
02db9ff7af18 ("asus-wmi: Add quirk_no_rfkill for the Asus U303LB")
2d735244b798 ("asus-wmi: Add quirk_no_rfkill for the Asus N552VW")
b5643539b825 ("platform/x86: asus-wmi: Filter buggy scan codes on ASUS Q500A")
7c1c184bb571 ("platform/x86: asus-wmi: try to set als by default")
aca234f63788 ("asus-wmi: provide access to ALS control")

Which looks scary, but it's all quirks for laptops folks are actually
using with this kernel. Then 401fee819 also applies cleanly on 4.4.

-- 
Thanks,
Sasha

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Willy Tarreau]

On Mon, Nov 25, 2019 at 08:30:48AM -0500, Sasha Levin wrote:
> On Sat, Nov 23, 2019 at 08:22:44PM +0100, Willy Tarreau wrote:
> > On Sat, Nov 23, 2019 at 01:08:03PM -0600, Bob Funk wrote:
> > > For the record, the patch that Willy offered does fix the issue on my
> > > affected
> > > system. That might be a better choice than my request to revert as per the
> > > original email.
> > 
> > Greg, FWIW I did nothing more than a regular backport so that you can take
> > it as-is. I think you dropped it from 4.4 because it did not apply well
> > and was not worth the hassle, but given that it fixes a regression caused
> > by another backport I think it makes sense to take it, at least so that
> > some users do not stop updating. The fix was only merged into 4.19, not
> > 4.4/4.9/4.14.
> > 
> > The backports for 4.9 and 4.14 are easy to do, if you're willing to take
> > the patches I can do them as well, just let me know.
> 
> Let's try something like this:
> 
> For 4.14 and 4.9 I'll also grab db2582afa744 ("platform/x86:
> asus-nb-wmi: Support ALS on the Zenbook UX430UQ") which makes 401fee819
> apply cleanly.
> 
> For 4.4, I'll grab this long list:
> 
> 92a505e8055f ("platform/x86: asus-wmi: add SERIO_I8042 dependency")
> db2582afa744 ("platform/x86: asus-nb-wmi: Support ALS on the Zenbook UX430UQ")
> 71f38c11cdb8 ("platform/x86: asus-wmi: Set specified XUSB2PR value for X550LB")
> 999d4376c628 ("platform/x86: asus-wmi: fix asus ux303ub brightness issue")
> a961a285b479 ("asus-wmi: Add quirk_no_rfkill_wapf4 for the Asus X456UF")
> a977e59c0c67 ("asus-wmi: Create quirk for airplane_mode LED")
> 6b7ff2af5286 ("asus-wmi: Add quirk_no_rfkill for the Asus Z550MA")
> 02db9ff7af18 ("asus-wmi: Add quirk_no_rfkill for the Asus U303LB")
> 2d735244b798 ("asus-wmi: Add quirk_no_rfkill for the Asus N552VW")
> b5643539b825 ("platform/x86: asus-wmi: Filter buggy scan codes on ASUS Q500A")
> 7c1c184bb571 ("platform/x86: asus-wmi: try to set als by default")
> aca234f63788 ("asus-wmi: provide access to ALS control")
> 
> Which looks scary, but it's all quirks for laptops folks are actually
> using with this kernel. Then 401fee819 also applies cleanly on 4.4.

Ah the joy of dealing with stable branches :-) I agree that all of them
are sufficiently small and isolated to be easy to deal with, and if they
can avoid bug reports that's always a bonus.

Willy

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Bob Funk]

I tested out the patch that was announced today for the 4.4.204-stable 
review and
all is working well here. Thanks for addressing this.

Regards,

Bob Funk


On 11/25/19 7:30 AM, Sasha Levin wrote:
> On Sat, Nov 23, 2019 at 08:22:44PM +0100, Willy Tarreau wrote:
>> On Sat, Nov 23, 2019 at 01:08:03PM -0600, Bob Funk wrote:
>>> For the record, the patch that Willy offered does fix the issue on my
>>> affected
>>> system. That might be a better choice than my request to revert as 
>>> per the
>>> original email.
>>
>> Greg, FWIW I did nothing more than a regular backport so that you can 
>> take
>> it as-is. I think you dropped it from 4.4 because it did not apply well
>> and was not worth the hassle, but given that it fixes a regression 
>> caused
>> by another backport I think it makes sense to take it, at least so that
>> some users do not stop updating. The fix was only merged into 4.19, not
>> 4.4/4.9/4.14.
>>
>> The backports for 4.9 and 4.14 are easy to do, if you're willing to take
>> the patches I can do them as well, just let me know.
>
> Let's try something like this:
>
> For 4.14 and 4.9 I'll also grab db2582afa744 ("platform/x86:
> asus-nb-wmi: Support ALS on the Zenbook UX430UQ") which makes 401fee819
> apply cleanly.
>
> For 4.4, I'll grab this long list:
>
> 92a505e8055f ("platform/x86: asus-wmi: add SERIO_I8042 dependency")
> db2582afa744 ("platform/x86: asus-nb-wmi: Support ALS on the Zenbook 
> UX430UQ")
> 71f38c11cdb8 ("platform/x86: asus-wmi: Set specified XUSB2PR value for 
> X550LB")
> 999d4376c628 ("platform/x86: asus-wmi: fix asus ux303ub brightness 
> issue")
> a961a285b479 ("asus-wmi: Add quirk_no_rfkill_wapf4 for the Asus X456UF")
> a977e59c0c67 ("asus-wmi: Create quirk for airplane_mode LED")
> 6b7ff2af5286 ("asus-wmi: Add quirk_no_rfkill for the Asus Z550MA")
> 02db9ff7af18 ("asus-wmi: Add quirk_no_rfkill for the Asus U303LB")
> 2d735244b798 ("asus-wmi: Add quirk_no_rfkill for the Asus N552VW")
> b5643539b825 ("platform/x86: asus-wmi: Filter buggy scan codes on ASUS 
> Q500A")
> 7c1c184bb571 ("platform/x86: asus-wmi: try to set als by default")
> aca234f63788 ("asus-wmi: provide access to ALS control")
>
> Which looks scary, but it's all quirks for laptops folks are actually
> using with this kernel. Then 401fee819 also applies cleanly on 4.4.
>

Re: Bug Report - Kernel Branch 4.4.y - asus-wmi.c

[Greg KH]

On Wed, Nov 27, 2019 at 08:17:25PM -0600, Bob Funk wrote:
> I tested out the patch that was announced today for the 4.4.204-stable
> review and
> all is working well here. Thanks for addressing this.

Wonderful, thanks for reporting this and for testing.

greg k-h