Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 20 Mar 2012 12:02:48 +0300
From:      Sergey Kandaurov <pluknet@freebsd.org>
To:        John Baldwin <jhb@freebsd.org>
Cc:        svn-src-stable@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org, svn-src-stable-8@freebsd.org
Subject:   Re: svn commit: r233199 - in stable/8/sys: i386/conf netinet6
Message-ID:  <CAE-mSOL=qEHEi4vN18JuhfoihsjaLDOrnpbaNoBdGKhZB9FLFw@mail.gmail.com>
In-Reply-To: <201203192015.q2JKFItl084480@svn.freebsd.org>
References:  <201203192015.q2JKFItl084480@svn.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 20 March 2012 00:15, John Baldwin <jhb@freebsd.org> wrote:
> Author: jhb
> Date: Mon Mar 19 20:15:18 2012
> New Revision: 233199
> URL: http://svn.freebsd.org/changeset/base/233199
>
> Log:
> =A0MFC 225096:
> =A0Fix if_addr_mtx recursion in mld6.

Ehh.. Thank you for doing this.
As I mentioned earlier I could not reproduce this problem myself
in RELENG_8, that's why it was not MFCed so far...

>
> =A0mld_set_version() is called only from mld_v1_input_query() and
> =A0mld_v2_input_query() both holding the if_addr_mtx lock, and then calli=
ng
> =A0into mld_v2_cancel_link_timers() acquires it the second time, which re=
sults
> =A0in mtx recursion. To avoid that, delay if_addr_mtx acquisition until a=
fter
> =A0mld_set_version() is called; while here, further reduce locking scope
> =A0to protect only the needed pieces: if_multiaddrs, in6m_lookup_locked()=
.

[...]

--=20
wbr,
pluknet



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAE-mSOL=qEHEi4vN18JuhfoihsjaLDOrnpbaNoBdGKhZB9FLFw>