Power Management Bugs Hold Up Some Linux Laptops Due To Regulatory Requirements

[Old Grouch]

Ok, I see now where the misunderstanding lies. You're talking about the shutdown process getting stuck and the system remaining visibly on, which I've certainly experienced my fair share of. But here the system looks like it's actually turned off properly, but somehow internally isn't and continues consuming significant power.

OK. I don't think I've experienced that...yet. So we have another new! exciting! failure mode to add to the collection. Forgive me if my celebrations are too noisy.

It illustrates the importance of standard interfaces with well-documented and well-implemented consistent behaviour: which I know is severely lacking when talking about ACPI. For obvious reasons, once something works 'well enough' with Microsoft Windows, it's good for shipping; and one never knows about undocumented quirks and non-standard behaviour.

I'm not sure how things can be improved, as the process has remained broken for years now.

[reba]

- As a side note, for the platforms in question we are working on a FW workaround, as it will take some time to deliver the proper fix upstream, especially with the holidays on us It was really important to understand root cause though.

Mark

Hey Mark,

thank you for the informations and for stopping by at Phoronix!

As you said there are firmware updates necessary to fix this, can you give any means to see for oneselves if one happens to have one of the affected systems? Like, specifiy firmware vesions, specifiy hardware configurtaions, etc. ?
I have an 20YM000B, which is an amazing machine and has extremely good Linux support (just the fingerprint sensor isn'tworking - which I never intended to use, perfect) and I am unsure if Lenovo still makes firmware updates for it.

[superm1]

As you said there are firmware updates necessary to fix this, can you give any means to see for oneselves if one happens to have one of the affected systems? Like, specifiy firmware vesions, specifiy hardware configurtaions, etc. ?

It's a generic problem with Linux. It doesn't put PCIe devices into D3cold at shutdown (S5). So you're certainly affected; but the degree of how severe it is would need to be measured.
In most laptops the amount of power that we're talking here is relatively insignificant. The worst offenders that I observed which lead to this patch have about a 20 percent improvement over S5 power consumption.

[oiaohm]

Mistakes happen. But even if this is merely a mistake, they can edit their own website..

No this is the USA EPA we are talking about those 7 defective documents before getting to one that correct with pictures is published over 20 years. Text and tables is very good for working on ground water contamination spread.

I don't know if these mistakes are intentional. But I can tell you with the documentation revision system the USA EPA uses its not as straight forwards to fix a mistake. They will publish another page in some random area with corrections to the page they never apply reasonable quickly or sometimes at all.

Some of this I suspect the USA EPA documentation handling system has been design that errors in documentation production are not detected early due to lack of editors and are insanely hard to fix after published. This is way of when USA EPA have regulations that work and can be documented as working that could be made stricter and work better end up looking like a laughing stock because of errors in the documents so meaning Industry does not have to obey as much regulation on emissions and the like as they should.. You also have documents from the USA EPA that are 100 pages long with 200 pages of corrections yet they will not do a new reviesion because they don't have funding to create new revision..

I don't say that USA EPA lies as such. USA EPA has a document/website production system of absolute trash that means any human mistake when creating the document/website end up stuck in it. Mixed with lots of human mistakes is a lot of valid and correct research. I am very clear to say USA EPA because the other EPA around the world all have very good documentation production systems with very effect mistake removal procedures.

As I said mistakes like the one you pointed to cause people to through the baby out with the bathwater. The valid and correctly made research of the USA EPA is the baby. The documents that research is in has a lot of bath water of human mistakes including personal bias in write ups that the USA FDA documentation system is not designed to prevent.

Having to do formal land management has really made me hate the USA FDA documentation production system due to how much time you waste making sure you have the document and all it correction notes.

[oiaohm]

It's a generic problem with Linux. It doesn't put PCIe devices into D3cold at shutdown (S5). So you're certainly affected; but the degree of how severe it is would need to be measured.
In most laptops the amount of power that we're talking here is relatively insignificant. The worst offenders that I observed which lead to this patch have about a 20 percent improvement over S5 power consumption.

For the users effected by the 20% cases that a lot. Another question is could this failure to D3cold at shutdown be why some laptops that ship windows only that are then converted to Linux seam fine when you boot into Linux after Windows but are absolute dog breakfast with issues if you shutdown Linux than reboot back into Linux. If this is the case then the problem is a little bit more serous than reduced battery. Modern day laptops where you cannot pull out battery has made it a lot harder to troubleshoot issues like this.

[superm1]

For the users effected by the 20% cases that a lot. Another question is could this failure to D3cold at shutdown be why some laptops that ship windows only that are then converted to Linux seam fine when you boot into Linux after Windows but are absolute dog breakfast with issues if you shutdown Linux than reboot back into Linux. If this is the case then the problem is a little bit more serous than reduced battery. Modern day laptops where you cannot pull out battery has made it a lot harder to troubleshoot issues like this.

I wouldn't say it's impossible to be caused by the same thing. Try my patches and see!