Re: [TEST] netgraph callout interface

From: Julian Elischer <julian_at_elischer.org>
Date: Tue, 07 Dec 2004 15:06:34 -0800
Gleb Smirnoff wrote:

>  The patch at 
>
>	http://people.freebsd.org/~glebius/totest/netgraph_callout
>
>has been updated. Now it covers the following nodes:
>
>	ng_bridge.c
>	ng_l2tp.c
>	ng_lmi.c
>	ng_pppoe.c
>	ng_pptpgre.c
>
>One more time I ask for testing. The patch can be applied both to HEAD
>and RELENG_5.
>
the bits I am familiar with look reasonable.

>	
>On Thu, Dec 02, 2004 at 06:10:48PM +0300, Gleb Smirnoff wrote:
>T>   Dear collegues,
>T> 
>T>   we are working on making netgraph ISR mpsafe. To do it we need to
>T> fix all (ab)users of bare timeout(9) in src/sys/netgraph. These
>T> timeout calls are running in synch with netgraph now because timeout(9)
>T> is Giant-locked. As soon as we mark ISR mpsafe, they are going to break.
>T> 
>T> This patch semi-mechanically changes all timeout(9) calls to ng_callout(9),
>T> which runs scheduled callouts in netgraph context:
>T> 
>T> 	http://people.freebsd.org/~glebius/totest/netgraph_callout
>T> 
>T> It can be applied to HEAD or RELENG_5 (not 5.3-RELEASE). It patches
>T> the following nodes:
>T> 
>T> 	ng_bridge.c
>T> 	ng_l2tp.c
>T> 	ng_lmi.c
>T> 	ng_ppp.c
>T> 	ng_pppoe.c
>T> 	ng_tty.c
>T> 
>T> If you are using at least one of them, then I'm asking you to test the
>T> patch and respond. Thanks in advance!
>T> 
>T> P.S. Sorry for crossposting. The target users are both RELENG_5 and CURRENT.
>
>  
>
Received on Tue Dec 07 2004 - 22:06:35 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:24 UTC