Bay Trail Linux C-State bug

[powerarmour]

Is there any news on when (or if!) this pretty major bug will be addressed?

https://bugs.freedesktop.org/show_bug.cgi?id=88012 Bug 88012 – [bisected BYT] complete freeze after: drm/i915/vlv: WA for Turbo and RC6 to work together

[VGerr]

Hi,

Thank you for your feedback. I am very happy to read that it helped you. Can you share your hardware and software specifics that are applicable so we can get an idea of what is affected and fixed?

I asked the original creator (Jochen Hein) to post it on the kernel mailing list, which he did.

It may need some cleaning up, I am keeping an eye on it.

Looking forward to at least processor and kernel you tested and anything more relevant you can share.

Thanks for contributing!

[uk]

hello Vincent,

I've been following this bug on bugzilla.com. I really appreciate your enthusiasm. I've put up the output of 3 commands ( uname -a, lshw, lscpu ) here.... http://pastebin.com/raw/nYBF9QCL http://pastebin.com/raw/nYBF9QCL

Let me know if you want anything else. I'd be glad to help.

Thank You

[VGerr]

hi,

That is great to read . As you may have noticed the discussion is still going on there.

I think it is a good idea to try and contribute there as much as possible in the line of bug report messages.The discussion regarding the patch is going at the moment and Len Brown who is from intel is involved in it and he made some suggestions.While the patch seems to fix the issue it may not be the best way to do it, so the discussion is still going on.

Would you be able to try a 4.9 kernel before and after the latest patch he posted?

This auto-demotion thing may actually also be a solution.If that patch works, it will most likely end up in the main line soon.

To Esteban: why have you been silent? Would it be possible to involve the right intel engineer next time in a timely manner, as in, to learn from this rather bad way the issue was picked up?

It seems in the end that it is still intel who has the best information on this. I understand the company is big, but I hope you do know a bit on who does kernel work and open source things?

thanks

[VGerr]

hi,

For every one still reading here, please read/subscribe to the bug:

https://bugzilla.kernel.org/show_bug.cgi?id=109051 https://bugzilla.kernel.org/show_bug.cgi?id=109051

Kernel developer Len Brown is posting patches and tools to try and track this down.

One promising patch is one that re-enables auto demotion.

I have supplied prebuilt kernels with those patches here:

https://www.dropbox.com/sh/c39et4hr6tgp60q/AAC35c56aOEOwkmhjdvtG6dsa?dl=0 Dropbox - ubuntu x86_64 4.8 and 4.9 Bay trail freeze patched kernel packages

To concentrate communication, it will be great if anyone reading this takes the time to read the bug and contribute by testing patches.

As mentioned by Len, the specific bug will be closed if we do no longer need the kernel parameter on Bay trail.

Bay trail processors are:http://ark.intel.com/products/codename/55844/Bay-Trail?q=bay%20trail# @All http://ark.intel.com/products/codename/55844/Bay-Trail?q=bay%20trail# @All

http://ark.intel.com/products/codename/55844/Bay-Trail?q=bay%20trail# @All Products (Formerly Bay Trail)

Note the remarks regarding Cherry trail that Hans de Goede is working on and the request to file a NEW bug if you are having freezes on another platform. It will be of great help if you contribute!

So call to all Bay trail users to contribute by trying the auto demotion patches on Bay trail and report about:

- difference in stability if any

- effect of nano sleep / output of turbo stat

- a way to reproduce the issue in a very short time (minute(s) instead of hours would be very usefull).

Please refer to the bug report and thanks everyone for your help, the more people help, the more likely this issue can be found and fixed!

Vincent

[AMukh6]

Hi Vincent,

I'm following the kernel.org bug you mentioned, and just finished reading the whole thread. However, where is the mention of Hans de Goede working on Cherry Trail? (I'm asking because I'm seeing the problem on a bunch of Cherry Trail devices (Atom Z8350) and I'm wondering if there's another bug I should be watching too. Thanks...

[VGerr]

Hi,

Maybe I wasn't specific .

Some of his patches can be found here:https://bugzilla.kernel.org/show_bug.cgi?id=155241 Bug 155241 – i2c_designware 808622C1:06: punit semaphore timed out, resetting

https://bugzilla.kernel.org/show_bug.cgi?id=155241 https://bugzilla.kernel.org/show_bug.cgi?id=155241

and here:

https://patchwork.ozlabs.org/patch/704783/ https://patchwork.ozlabs.org/patch/704783/

If you type up in the last url (end in 4,5 etc) you get the consecutive patches, which need to be applied in order.

I think the best is to see if the mentioned bug report applies to you, find another one that does match if it doesn't and file a new one if you can't find anything existing.I know that the patches Hans made gave him stability on a Cherry trail based tablet, so it may work for yours too.

good luck!

Vincent

https://patchwork.ozlabs.org/patch/704783/ https://patchwork.ozlabs.org/patch/704783/