Rockbox mail archive

• Next message: Paul van der Heu: "Re: How can we manage performance problems?"
• Previous message: Andreas Zwirtes: "Re: How can we manage performance problems?"
• In reply to: Andreas Zwirtes: "Re: How can we manage performance problems?"
• Next in thread: Andreas Zwirtes: "Re: How can we manage performance problems?"
• Reply: Andreas Zwirtes: "Re: How can we manage performance problems?"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]

Quoting Andreas Zwirtes <zwirtes_at_gmx.de>:
> From what I know, priority ceiling multithreading adresses problems
> like this exactly. With a high priority disk access and decoding, it
> wouldn't be possible to interrupt the playback. Since the implementation
> of real OSEK multithreading is very challenging, it is possible to use
> ready made (and certified) solutions for microcontrollers. Correct me if
> I'm wrong, please.

There is a reason for having such a simple scheduler that we have today, and that is because we want things simple. Adding priorities and even preemptive scheduling is something we definitely want to avoid, since it buys us a shitload of new problems.

/Linus

• Next message: Paul van der Heu: "Re: How can we manage performance problems?"
• Previous message: Andreas Zwirtes: "Re: How can we manage performance problems?"
• In reply to: Andreas Zwirtes: "Re: How can we manage performance problems?"
• Next in thread: Andreas Zwirtes: "Re: How can we manage performance problems?"
• Reply: Andreas Zwirtes: "Re: How can we manage performance problems?"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]