Re: kpi socket filter: Safe to block in callbacks?
Re: kpi socket filter: Safe to block in callbacks?
- Subject: Re: kpi socket filter: Safe to block in callbacks?
- From: Brian Bergstrand <email@hidden>
- Date: Thu, 26 May 2005 22:47:02 -0500
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
I'm thinking worse case scenario here. It would be possible for us to
issue our own callback to be serviced on another thread, but I'm
afraid of lag issues with that model.
This brings up a question: why are there no try variants for R/W and
mutex locks in the new lock KPI (lck_*)?
On May 26, 2005, at 8:36 PM, Vincent Lubet wrote:
> It is safe up to the point it takes a long time to resolve the
> contention because some events are notified on the input thread and
> you don't want to hold up all the incoming network traffic.
>
> Vincent
>
> On May 25, 2005, at 3:05 PM, Josh Graessley wrote:
>
>
>>
>> That should be safe.
>>
>> -josh
>>
>> On May 25, 2005, at 2:38 PM, Brian Bergstrand wrote:
>>
>>
>>
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>> Is it safe to block on a R/W lock inside a socket filter callback?
>>> I'm specifically interested in sf_notify_func(). I doubt lock
>>> contention will be an issue for us at the point the notify callback
>>> happens, but it's always possible.
>
Brian Bergstrand
<http://www.bergstrand.org/brian/> PGP Key ID: 0xB6C7B6A2
-----BEGIN PGP SIGNATURE-----
Version: PGP Desktop 9.0.1 (Build 2130)
iQA/AwUBQpaYO3nR2Fu2x7aiEQIHQACeNtwDFQa5Z0SKSYDy8/3SuhRaposAni8a
A0/z2ta9SnKxiJ64wb4cbSed
=cyv+
-----END PGP SIGNATURE-----
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden