Re: Running the network stack without Giant -- change in default coming

From: Jun Kuriyama (kuriyama_at_imgsrc.co.jp)
Date: 08/25/04

  • Next message: Giorgos Keramidas: "Re: IPv4 checksum oddness (gcc compiler bug?)" 
    Date: Wed, 25 Aug 2004 10:39:38 +0900
To: Robert Watson <rwatson@FreeBSD.org>

    At Tue, 24 Aug 2004 14:41:20 +0000 (UTC),
    Robert Watson wrote:
    > - We've focussed primarily on getting mainstream network configurations to
    > run without Giant: this means that less mainstream subsystems (parts of
    > IPv6, some netgraph nodes, IPX, etc) are currently unsafe without the
    > Giant lock turned on.

    I'm interesting to use without Giant. But (as you know :-)) I'm using
    IPv6 usually. How can I help to test this?

    I'm sorry I don't understand how it is difficult, but is it possible
    to protect whole IPv6 code only with Giant without network stack
    Giant? 

    -- 
Jun Kuriyama <kuriyama@imgsrc.co.jp> // IMG SRC, Inc.
             <kuriyama@FreeBSD.org> // FreeBSD Project
_______________________________________________
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org"
  • Next message: Giorgos Keramidas: "Re: IPv4 checksum oddness (gcc compiler bug?)"

    Relevant Pages

    • Re: cvs commit: src/sys/conf options src/sys/sys kernel.h src/sys/net netisr.c
      ... >> will cause the network stack to operate without the Giant lock by ... > until the locking of IPv6 is done? ... so I think it is probably reasonable to run w/o Giant ... Regarding IPv6: significant parts of IPv6 are safe in an MPSAFE ...
      (freebsd-current)
    • Re: Postgresql performance profiling
      ... This is currently done via a Giant-locked sysctl so there is enormous contention for Giant. ... You really want to avoid using IPv6 for transport. ... It is likely quite easy to put subsystem locks around System V IPC subsystems. ...
      (freebsd-performance)
    • Re: Running the network stack without Giant -- change in default coming
      ... > protect whole IPv6 code only with Giant without network stack Giant? ... but I think a better strategy would be for us to complete the locking work ... - We need to review TCP/UDP locking, which was mirrored from the IPv4 ...
      (freebsd-current)
    • Re: locking & iovecs
      ... > Brooks Davis wrote: ... the code is not fully thread-safe yet (we want to drop GIANT)? ... I believe IPv4 is pretty close other then NIC drivers, but IPv6 is ...
      (freebsd-net)