User: Password:
|
|
Subscribe / Log in / New account

KS2009: Realtime preemption

KS2009: Realtime preemption

Posted Oct 23, 2009 22:27 UTC (Fri) by dvhart (subscriber, #19636)
Parent article: KS2009: Realtime preemption

"Significantly, nobody questioned the overall value of merging the realtime code into the mainline."

Did this include sleeping spinlocks? My understanding has been that these are the ugly duckling of the PREEMPT_RT tree that the rest of the kernel developers would rather not see in mainline.


(Log in to post comments)

Ugly duckling?

Posted Oct 28, 2009 10:29 UTC (Wed) by dion (guest, #2764) [Link]

The ugly duckling was a swan, not a duck and it was ostracized for being ugly when viewed by the standards of ducklings.

... so are you saying that the "sleeping spinlocks" feature is the most wonderful change of all the RT changes and that people simply doesn't understand it?

KS2009: Realtime preemption

Posted Oct 29, 2009 20:35 UTC (Thu) by joib (subscriber, #8541) [Link]

AFAICS the "adaptive mutexes" patch went into 2.6.30 ( http://lwn.net/Articles/325921/ ). Well, sleeping spinlock, spinning mutex, adaptive mutex, or whatever you want to call it.

Or did you mean the part of the -rt patch that automagically turns most spinlocks into such adaptive mutexes (w/ PI)? Yeah, that might be more difficult to get into mainline?


Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds