Re: [REGRESSION] `lis3lv02d_i2c_suspend()` causes `unbalanced disables for regulator-dummy` and `Failed to disable Vdd_IO: -EIO`

[Hans de Goede]

Hi,

On 2/14/24 16:34, Linux regression tracking (Thorsten Leemhuis) wrote:
> On 08.02.24 15:34, Mark Brown wrote:
>> On Fri, Feb 02, 2024 at 12:38:04PM +0100, Paul Menzel wrote:
>>
>>> Testing commit 2f189493ae32 (i2c: i801: Add lis3lv02d for Dell XPS 15 7590)
>>> [1], it’s very likely this commit, it turns out, that Linux logs the warning
>>> below during ACPI S3 suspend:
> 
> Mark, many thx for your reply:
> 
>> The driver is just buggy here AFAICT,
> 
> /me reads that as "buggy even before 2f189493ae32, that commit just
> exposed the problem"
> 
>> it's powering off the device in
>> both runtime suspend and runtime resume so if the device is runtime
>> suspended when system suspend happens then it'll power off the device
>> again.  The runtime suspend and system suspend need to talk to each
>> other here.
> 
> Well, that was 6 days ago and nothing further happened. Makes me wonder:
> Who will look into this? And when?

I have a laptop with a lis3lv02d where I can try to reproduced Paul's
problem. I expect this to reproduce on any Dell with the lis3lv02d 
so reverting 2f189493ae32 won't help, it fixes the problem on
Paul's model but not on other Dell models which use the lis3lv02d
driver.

Anways I plan to look at this sometime next week.

Regards,

Hans