Messages in this thread | | | Date | Fri, 6 Nov 2015 12:28:56 +0800 | Subject | Re: Improve spinlock performance by moving work to one core | From | Ling Ma <> |
| |
Longman
Thanks for your suggestion. We will look for real scenario to test, and could you please introduce some benchmarks on spinlock ?
Regards Ling
> > Your new spinlock code completely change the API and the semantics of the > existing spinlock calls. That requires changes to thousands of places > throughout the whole kernel. It also increases the size of the spinlock from > 4 bytes to 32 bytes. It is basically a no-go. > > However, if you can improve the performance within the existing API and > locking semantics, it will be much more acceptable. > > Cheers, > Longman
| |