Hacker News new | past | comments | ask | show | jobs | submit login
What’s up with ARM (linuxfoundation.org)
124 points by pmjordan on May 15, 2011 | hide | past | web | favorite | 6 comments

Some context:

Gaah. Guys, this whole ARM thing is a f-cking pain in the ass https://lkml.org/lkml/2011/3/17/492

Let ARM rot in the mainline. I really don't care anymore. https://lwn.net/Articles/441384/

It'll be interesting to see what direction this goes. Considering also the forks for the Android kernel and the different directions ARM development wants to go compared to Intel.

Yeah, this isn't really new to anyone who reads the LWN (which is edited by the author of the OP, BTW). The LWN is really very good, and subscribing is worth it for those hackers who care about the larger Linux/POSIX ecosystem.

This is very welcome.

ARM has tended to be somewhat difficult to port to - vendors have multiple versions of the instruction set out there depending on device size/power requirements, with different floating point units (NEON, VFP) etc.

The result is we end up with situations like OpenEmbedded supporting multiple kernel trees in it's build engine just to take care of the wide range of hardware it supports.

Hopefully this will make the situation somewhat better.

Anything that increases functionality/LoC in the kernel is good. In any complex software project - and the Linux kernel is about as complex as sanity and present technology will allow - you have to do periodical codebase clean-ups. This is one case where the pain to maintain is forcing a cleanup.

And, as always, it's discussed openly, so, whoever depends on it doesn't get surprised by the next release.

What exactly is the "this" you are referring to? This article is pretty much an official acknowledgment that a problem exists and needs an investment of developer resources. The situation won't get better until/unless stakeholders make an investment of resources. Clearly someone is going to have to pay for a lot of hours of time spent by developers who are competent to work on arm arch stuff in a "global" way. I'm not too familiar with the details of the various multi-stakeholder alliances and how they get their funding and decide who to hire to do what, but it sounds like the Linux Foundation (or whoever) needs to support a few more developers with a job description more focused on the kernel itself rather than just relying on the companies contributing what they did to get specific devices ready for shipping.

This biggest problem with ARM devices is, that the only devices out there for usage are embedded ones. Up to now there is only one Notebook out there (alwaysinnovating) which could be used as a computer on a daily basis. But IMHO 10'' is too small to do serious stuff. 13'' is minimum. Hope we will see them soon in the wild.

Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact