Google Finally Shifting To "Upstream First" Linux Kernel Approach For Android Features

[MastaG]

Imagine they can actually pull this off and also make Qualcomm switch their BSP to the Mesa drivers instead of the shitty proprietary blobs..

Then every phone shipping from 2023 or 2024 would have their drivers and graphics stack in upstream kernel/Mesa already.. meaning you can theoretically keep updating until the end of time.

[kpedersen]

Imagine they can actually pull this off and also make Qualcomm switch their BSP to the Mesa drivers instead of the shitty proprietary blobs..

Then every phone shipping from 2023 or 2024 would have their drivers and graphics stack in upstream kernel/Mesa already.. meaning you can theoretically keep updating until the end of time.

The Samsung, Huawei, Google and every other phone you can actually buy will still be full of blobs, spyware and bloatware bollocks as usual however. The consumer is still screwed with whatever garbage the actual downstream vender churns out.

[CommunityMember]

Imagine they can actually pull this off and also make Qualcomm switch their BSP to the Mesa drivers instead of the shitty proprietary blobs..

Google themselves likely no longer cares about Qualcomm's choices going forward (that is, however, the point with modularization/hooks of those proprietary implementations), as Google themselves are moving towards their "Tensor" SoC for new devices, which is based on the ARM reference designs, and no longer Qualcomm derived designs. Other vendors will, as always, continue to do what they choose to do.

[CommunityMember]

Awesome plan, but I fear they won't be able to execute it 100% as planned, specially having to follow upstream release cadence and code review 🤔

I do not believe anyone believes 100% reliable execution in practice, but given the long lead times for most new designs/products, and the vast resources Google has available to it in order to succeed, getting the essential parts in the upstream kernel should be a reasonably achievable result (look towards Intel who tends to get new platform enablements into the kernel long before a new product ships, with the acknowledged occasional stumbles).

[stormcrow]

Bold to assume anyone but some nerds like us care about what's the running kernel... For PR to work you need someone interested.



Besides, no, because most of the kernel changes aren't really at the ABI level and applications can't see them.

Naw, this isn't a PR push to the general public. Most people don't even understand why they should even care about security and how device security works. They don't care about version numbers. They only care if their battery dead and if they can connect to Facebook or play their new special game. This is PR for Android developers and a notice to OEMs.

Android's Achilles heel is the version fragmentation and OEMs that either won't or can't keep their software updated. This not only makes the Android market harder to support, but it also makes the general Android user community unsafe - even if the average consumer is completely unaware of the problem. Tossing developers a bone to gnaw on for 12 at least gives them something to look forward too... in about 3-5 years most of the rest of the market will have caught up to 12's features due to device attrition.

[chocolate]

Cynical, moi?

Not at all, this is 100% what's happening.

[sinepgib]

Imagine they can actually pull this off and also make Qualcomm switch their BSP to the Mesa drivers instead of the shitty proprietary blobs..

Then every phone shipping from 2023 or 2024 would have their drivers and graphics stack in upstream kernel/Mesa already.. meaning you can theoretically keep updating until the end of time.

That won't happen because programmed obsolescence is the fabric of their businesses. There's just not only a lack of incentive but an actual pressing reason for them to avoid it.

Also the bloatware kpedersen mentioned, your data is part of their revenue and making it so you can easily switch the running system is damaging for them.

Naw, this isn't a PR push to the general public. Most people don't even understand why they should even care about security and how device security works. They don't care about version numbers. They only care if their battery dead and if they can connect to Facebook or play their new special game. This is PR for Android developers...

What exactly do Android developers care about here?

... and a notice to OEMs.

Android's Achilles heel is the version fragmentation and OEMs that either won't or can't keep their software updated. This not only makes the Android market harder to support, but it also makes the general Android user community unsafe - even if the average consumer is completely unaware of the problem. Tossing developers a bone to gnaw on for 12 at least gives them something to look forward too... in about 3-5 years most of the rest of the market will have caught up to 12's features due to device attrition.

That's not PR, but an actual technical reason. This kind of notice, intended to pressure OEMs, is not precisely something I'd call PR. If anything it makes them look a bit mob-ish.

[Quackdoc]

I haven't seen the talk yet. but it's safe to assume Google doesn't want to use nor maintain the android kernel. it is a financial and technical burden they likely just don't want. that being said, android still needs to run their own kernel. as they still have modifications that they can't upstream yet. but the end goal would be eventually to run via upstream (to be clear, android CAN boot from upstream kernel last i checked.) but proprietary vendor shit will always be an issue. but thankfully you should be able to boot android on most modern devices without it, but you will be missing support for sensors, camera's etc. since most of these need proprietary firmware from vendor.img. the overall goal is to minimize the issues.

[brad0]

Why not switching to mainline Linux kernel entirely?

Same reason Linux based OS's don't ship unpatched kernels. Google has tons of patches that would need to be commited to make that happen. A lot of that stuff for various reasons is very unlikely to ever go upstream.

[tildearrow]

Why not switching to mainline Linux kernel entirely?

Because that would upset every device manufacturer and greatly delay their plans by requiring strict evaluation prior to the addition of a new device/platform.