Re: [GIT PULL] Re: REGRESSION: Performance regressions from switching anon_vma->lock to mutex

From: Linus Torvalds
Date: Wed Jun 15 2011 - 16:47:59 EST




"Paul E. McKenney" <paulmck@xxxxxxxxxxxxxxxxxx> wrote:
>
>It would be much lower risk to make the current code always use softirq
>if !RCU_BOOST -- last time I attempted the revert, it was quite hairy.

I don't care if it's a real revert or not, but I want the threads gone. Entirely. Not just the patch that uses softirqs for some things, and threads for the callbacks. No, I don't want the threads to show up or exist at all.

And to be sure, I'd like the code to set up and use the threads to actually compile away statically, so that there clearly isn't some way it's partially enabled.

Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/