Re: [REGRESSION] hp-wmi fails to probe on HP OMEN (board 8D41)

Re: [REGRESSION] hp-wmi fails to probe on HP OMEN (board 8D41)

[Thorsten Leemhuis]

On 5/3/26 09:26, Krishna Chomal wrote:
> On Thu, Apr 30, 2026 at 12:59:58PM +0800, yue li wrote:
>> [...]
>> Regression:
>> The hp-wmi driver works correctly on kernel 6.18.6-zen1-1-zen,
>> but fails to probe on 7.0.2-zen1-1-zen with error -22.
>>
>> Issue:
>> On the affected kernel, loading the module fails:
> [...]
> Thanks for reporting. This is a known issue [1]. There are 2 failures in
> your dmesg output:
> 1. Unknown EC layout: The fix for this has been sent recently.
>    Currently, it is in the review-ilpo-fixes branch [2].
> 2. Probe failed: This is due to a failure in parsing the fan table. The
>    fix for this has been merged into 7.1-rc1 [3].
> 
> Using kernel 7.1 (or applying the above mentioned 2 patch manually to
> your current build) should restore full support for your device.

You you then maybe be so kind and submit backports to the 7.0.y stable
series to prevent more people from running into this?

Ciao, Thorsten

> [1] https://lore.kernel.org/platform-driver-x86/4ae4d7f9-f1db-4394-b0d9-
> c75dc39afcf8@gmail.com
> [2] https://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-
> drivers-x86.git/commit/?h=review-ilpo-
> fixes&id=c2d4b76458c9ebf81eae2916970320f1f61ad002
> [3] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/
> commit/drivers/platform/x86/hp/hp-wmi.c?h=v7.1-
> rc1&id=9d317a54e46d3b6420567dc5b63e9d7ff5c064a3

Re: [REGRESSION] hp-wmi fails to probe on HP OMEN (board 8D41)

[Thorsten Leemhuis]

[+Ilpo]

On 5/4/26 09:22, Thorsten Leemhuis wrote:
> On 5/3/26 09:26, Krishna Chomal wrote:
>> On Thu, Apr 30, 2026 at 12:59:58PM +0800, yue li wrote:
>>> [...]
>>> Regression:
>>> The hp-wmi driver works correctly on kernel 6.18.6-zen1-1-zen,
>>> but fails to probe on 7.0.2-zen1-1-zen with error -22.
>>>
>>> Issue:
>>> On the affected kernel, loading the module fails:
>> [...]
>> Thanks for reporting. This is a known issue [1]. There are 2 failures in
>> your dmesg output:
>> 1. Unknown EC layout: The fix for this has been sent recently.
>>    Currently, it is in the review-ilpo-fixes branch [2].

It[1] afaics is sitting there for a week already afaics. Wondering if it
might be worth resubmitting the patch with a Fixes tag as well as a
Link/Closes tag pointing to this thread to make the "this is fixing a
regression" aspect obvious to ensure the fix is merged this cycle?
Adding a stable tag would be great, too.

Ciao, Thorsten

[1]
https://lore.kernel.org/all/20260429180953.129885-1-krishna.chomal108@gmail.com/

>> 2. Probe failed: This is due to a failure in parsing the fan table. The
>>    fix for this has been merged into 7.1-rc1 [3].
>>
>> Using kernel 7.1 (or applying the above mentioned 2 patch manually to
>> your current build) should restore full support for your device.
> 
> You you then maybe be so kind and submit backports to the 7.0.y stable
> series to prevent more people from running into this?

Side note: s/You you/Could you/ #uhhps #sorry

>> [1] https://lore.kernel.org/platform-driver-x86/4ae4d7f9-f1db-4394-b0d9-
>> c75dc39afcf8@gmail.com
>> [2] https://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-
>> drivers-x86.git/commit/?h=review-ilpo-
>> fixes&id=c2d4b76458c9ebf81eae2916970320f1f61ad002
>> [3] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/
>> commit/drivers/platform/x86/hp/hp-wmi.c?h=v7.1-
>> rc1&id=9d317a54e46d3b6420567dc5b63e9d7ff5c064a3
> 
> 

Re: [REGRESSION] hp-wmi fails to probe on HP OMEN (board 8D41)

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 3949 bytes --]

On Thu, 7 May 2026, Thorsten Leemhuis wrote:

> [+Ilpo]
> 
> On 5/4/26 09:22, Thorsten Leemhuis wrote:
> > On 5/3/26 09:26, Krishna Chomal wrote:
> >> On Thu, Apr 30, 2026 at 12:59:58PM +0800, yue li wrote:
> >>> [...]
> >>> Regression:
> >>> The hp-wmi driver works correctly on kernel 6.18.6-zen1-1-zen,
> >>> but fails to probe on 7.0.2-zen1-1-zen with error -22.
> >>>
> >>> Issue:
> >>> On the affected kernel, loading the module fails:
> >> [...]
> >> Thanks for reporting. This is a known issue [1]. There are 2 failures in
> >> your dmesg output:
> >> 1. Unknown EC layout: The fix for this has been sent recently.
> >>    Currently, it is in the review-ilpo-fixes branch [2].
> 
> It[1] afaics is sitting there for a week already afaics.

Hi,

I'm sorry to be a disappointment here, but I don't find a week a 
particularly long time.

First, a patch sits in the review-ilpo-fixes branch (think of it as a 
staging branch) until LKP builds it to avoid totally breaking anyone's 
next builds if prematurely advancing anything to the fixes branch. My 
feel is LKP hasn't been particularly swift recently.

As not many will have a look on patches in review-ilpo-fixes, it will have 
to sit some more in the fixes branch to be exposed to wider scale of 
testing and checking before doing PR. It takes one day for getting it 
included into linux-next in the first place. Then the remaining question 
is, how many days you expect Dan (or others) to need to do his checker 
runs and come up with a problem report or correction patch.

I don't see how I could make the time much shorter than a week without 
effectively bypassing some of the quality control steps. But perhaps you 
have some suggestions how we can make Dan work even faster than he already
does... :-)

Besides, the change you brought up is on my scale the least critical 
change among the ones queued in the fixes branch. Without the other two,
I'd delay making the PR so it would be sitting there yet another week. ;-)

> Wondering if it
> might be worth resubmitting the patch with a Fixes tag as well as a
> Link/Closes tag pointing to this thread to make the "this is fixing a
> regression" aspect obvious to ensure the fix is merged this cycle?
> Adding a stable tag would be great, too.

I'm not sure what to add and where, wasn't the fan table fix already 
merged earlier? And it had Fixes tag.

This ec layout thing, on the other hand, doesn't seem that critical or 
even a serious issue. We cannot really "fix" it anyway without platform 
providing us the current profile information so it's all about silencing 
the warning which is hardly worth all this fuzz, IMO.


With all this being said, I still do appreciate the work you do, my 
intention is to not antagonize regression tracking you do in any way.

-- 
 i.

> Ciao, Thorsten
> 
> [1]
> https://lore.kernel.org/all/20260429180953.129885-1-krishna.chomal108@gmail.com/
> 
> >> 2. Probe failed: This is due to a failure in parsing the fan table. The
> >>    fix for this has been merged into 7.1-rc1 [3].
> >>
> >> Using kernel 7.1 (or applying the above mentioned 2 patch manually to
> >> your current build) should restore full support for your device.
> > 
> > You you then maybe be so kind and submit backports to the 7.0.y stable
> > series to prevent more people from running into this?
> 
> Side note: s/You you/Could you/ #uhhps #sorry
> 
> >> [1] https://lore.kernel.org/platform-driver-x86/4ae4d7f9-f1db-4394-b0d9-
> >> c75dc39afcf8@gmail.com
> >> [2] https://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-
> >> drivers-x86.git/commit/?h=review-ilpo-
> >> fixes&id=c2d4b76458c9ebf81eae2916970320f1f61ad002
> >> [3] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/
> >> commit/drivers/platform/x86/hp/hp-wmi.c?h=v7.1-
> >> rc1&id=9d317a54e46d3b6420567dc5b63e9d7ff5c064a3
> > 
> > 
> 

Re: [REGRESSION] hp-wmi fails to probe on HP OMEN (board 8D41)

[Thorsten Leemhuis]

On 5/7/26 11:55, Ilpo Järvinen wrote:
> On Thu, 7 May 2026, Thorsten Leemhuis wrote:
>> On 5/4/26 09:22, Thorsten Leemhuis wrote:
>>> On 5/3/26 09:26, Krishna Chomal wrote:
>>>> On Thu, Apr 30, 2026 at 12:59:58PM +0800, yue li wrote:
>>>>> [...]
>>>>> Regression:
>>>>> The hp-wmi driver works correctly on kernel 6.18.6-zen1-1-zen,
>>>>> but fails to probe on 7.0.2-zen1-1-zen with error -22.
>>>>>
>>>>> Issue:
>>>>> On the affected kernel, loading the module fails:
>>>> [...]
>>>> Thanks for reporting. This is a known issue [1]. There are 2 failures in
>>>> your dmesg output:
>>>> 1. Unknown EC layout: The fix for this has been sent recently.
>>>>    Currently, it is in the review-ilpo-fixes branch [2].
>>
>> It[1] afaics is sitting there for a week already afaics.
> 
> I'm sorry to be a disappointment here, but I don't find a week a 
> particularly long time.

For a normal patch, I fully agree. But for a regression fix, it is a
long time[1]. And to me it looked like the patch "platform/x86: hp-wmi:
silence unknown board warning for 8D41"[2] was a the second and
necessary part of a fix for the regression Yue Li reported at the start
of this thread[3] -- without making it obvious in the patch description
that it was one, so you couldn't know, which is why I CCed you here.

But from looking at Krishna's earlier reply in this thread again, it now
seems to me like I misinterpreted the "applying the above mentioned 2
patch" statement, as 9d317a54e46d3b ("platform/x86: hp-wmi: fix fan
table parsing") [already in v7.1-rc1 and hopefully backported soon]
seems to be the main fix and the other one ("platform/x86: hp-wmi:
silence unknown board warning for 8D41"[2] just silences a warning that
is not causing trouble.

Sorry, my deepest apologies for causing trouble.

Ciao, Thorsten

[1]https://www.kernel.org/doc/html/latest/process/handling-regressions.html#on-how-quickly-regressions-should-be-fixed
[2]https://lore.kernel.org/all/20260429180953.129885-1-krishna.chomal108@gmail.com/

[3]https://lore.kernel.org/all/CABd86uaQY9-ZSU-_PHKtZZyWnDFHLEsuaM+NBV5nOLWA+YSX7A@mail.gmail.com/

Re: [REGRESSION] hp-wmi fails to probe on HP OMEN (board 8D41)

[Krishna Chomal]

On Thu, May 07, 2026 at 12:59:41PM +0200, Thorsten Leemhuis wrote:
>On 5/7/26 11:55, Ilpo Järvinen wrote:
>> On Thu, 7 May 2026, Thorsten Leemhuis wrote:
>>> On 5/4/26 09:22, Thorsten Leemhuis wrote:
>>>> On 5/3/26 09:26, Krishna Chomal wrote:
>>>>> On Thu, Apr 30, 2026 at 12:59:58PM +0800, yue li wrote:
>>>>>> [...]
>>>>>> Regression:
>>>>>> The hp-wmi driver works correctly on kernel 6.18.6-zen1-1-zen,
>>>>>> but fails to probe on 7.0.2-zen1-1-zen with error -22.
>>>>>>
>>>>>> Issue:
>>>>>> On the affected kernel, loading the module fails:
>>>>> [...]
>>>>> Thanks for reporting. This is a known issue [1]. There are 2 failures in
>>>>> your dmesg output:
>>>>> 1. Unknown EC layout: The fix for this has been sent recently.
>>>>>    Currently, it is in the review-ilpo-fixes branch [2].
>>>
>>> It[1] afaics is sitting there for a week already afaics.
>>
>> I'm sorry to be a disappointment here, but I don't find a week a
>> particularly long time.
>
>For a normal patch, I fully agree. But for a regression fix, it is a
>long time[1]. And to me it looked like the patch "platform/x86: hp-wmi:
>silence unknown board warning for 8D41"[2] was a the second and
>necessary part of a fix for the regression Yue Li reported at the start
>of this thread[3] -- without making it obvious in the patch description
>that it was one, so you couldn't know, which is why I CCed you here.
>
>But from looking at Krishna's earlier reply in this thread again, it now
>seems to me like I misinterpreted the "applying the above mentioned 2
>patch" statement, as 9d317a54e46d3b ("platform/x86: hp-wmi: fix fan
>table parsing") [already in v7.1-rc1 and hopefully backported soon]
>seems to be the main fix and the other one ("platform/x86: hp-wmi:
>silence unknown board warning for 8D41"[2] just silences a warning that
>is not causing trouble.
>

Yes that's right. The fan table fix allows the driver to be loaded and
fixes the regression, while the follow-up patch is only to silence the
dmesg warning. I will be more explicit from next time, sorry for any
confusion that I might have created.

>Sorry, my deepest apologies for causing trouble.
>
>Ciao, Thorsten
>
>[1]https://www.kernel.org/doc/html/latest/process/handling-regressions.html#on-how-quickly-regressions-should-be-fixed
>[2]https://lore.kernel.org/all/20260429180953.129885-1-krishna.chomal108@gmail.com/
>
>[3]https://lore.kernel.org/all/CABd86uaQY9-ZSU-_PHKtZZyWnDFHLEsuaM+NBV5nOLWA+YSX7A@mail.gmail.com/