Re: Deferred freeing of kauth_cred_ts

[Elad Efrat <elad%NetBSD.org@localhost>]

Hey,

On Fri, Dec 25, 2009 at 3:46 PM, Mindaugas Rasiukevicius
<rmind%netbsd.org@localhost> wrote:

> The question is - why freeing is performed from interrupt context?

I couldn't find where sofree() might be reached from interrupt
context, but I might be missing something. Of course it's possible (?)
that the path that called it now doesn't do so anymore. At the moment
I'm running locally with the original code in place (i.e. no deferred
freeing of any kind) in hope to trip on such a case... If it won't
happen, I'll go ahead and check in the changes that add credentials to
sockets so it gets wider testing on -current.

Should any problems arise, we can then analyze the trace and see about
fixing it. (And in the meantime I'll back out the changes, of course.)

Thanks,

-e.