User talk:Dsimic

From PINE64
Revision as of 23:16, 25 April 2021 by Dsimic (talk | contribs) (→‎Changes to PinePhone: Replied again)
Jump to navigation Jump to search

Please, feel free to leave a message, but remember to sign it with -- ~~~~. You may also want to contact me as "dsimic" on the Pine64 forum.

Changes to PinePhone

I've reverted some changes made to PinePhone due to the direction they're going. Especially one larger addition is moved into your user section (see https://wiki.pine64.org/wiki/User:Dsimic/PinePhone_Screen_Issue) due to being unconfirmed. Please get a confirmation for the claimed issue and causes via a dev. In my experience it is more likely that the thermal emergency shutdown function is broken again (see http://xnux.eu/log/#017), which should trigger under considerably lower temperatures, before any damage can occur. So claiming that certain settings can "fix" this is highly misleading, if temperature is an issue at all. If it's not the thermal emergency function failing again, it can also simply be bad apples. I recommend to contact the devs who look into this, such as megi, via the pine-dev channel. Regards, --Fxc (talk) 12:06, 22 April 2021 (UTC)

On second thought, you're probably right; the information on screen delamination should be verified further before becoming included into the PinePhone article. Regarding the thermal emergency shutdown, my own experience clearly shows that the upper third of the PinePhone screen gets very hot, even under light CPU load, unless more restrictive CPU thermal throttling is applied. Thus, the primary concern of the thermal emergency shutdown (as described by Ondrej) is the temperature of the battery, not the temperature of the A64 SoC.
Also, I would highly suggest that a brief description of out-of-tree kernel patches, which you also removed, becomes part of the article again. Not mentioning out-of-tree kernel patches is highly misleading. Surely, the end goal is to have everything upstreamed, but the reality now is that additional kernel patches are required. As we know, many Linux distributions for PinePhone use Ondrej's kernel source, which contains a number of patches. -- Dsimic (talk) 15:59, 25 April 2021 (UTC)
After a lengthy chat session with Ondrej, the conclusion is to include a brief description of the issue in the wiki page. We'll also ask current PinePhone owners to test out different values for the throttling thresholds; based on the results and reports, changes will be made in the Linux kernel tree maintained by Ondrej, so the more restrictive CPU throttling eventually becomes the default. This issue has also been brought to the attention of Pine64, thanks to Xalius, and the PinePhone production team will investigate the issue.
In a few words, it's better to be on the safe side. This issue might be simply a result of a bad batch of phones, which used bad or weak glue, but it's much better to have more restrictive CPU throttling configured on hundreds of phones, as a preventative measure, than to have dozens of screens damaged. Maybe the heat isn't the root cause, but changing software settings is cheap, while replacing damaged screens isn't. -- Dsimic (talk) 23:15, 25 April 2021 (UTC)