Date: Tue, 4 Mar 2014 18:17:57 -0500 From: George Neville-Neil <gnn@freebsd.org> To: John Baldwin <jhb@FreeBSD.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r262727 - head/sys/net Message-ID: <FD8EBD57-749C-4247-8CA2-E846E046476F@freebsd.org> In-Reply-To: <201403041214.44230.jhb@freebsd.org> References: <201403040509.s2459lou017310@svn.freebsd.org> <201403041214.44230.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mar 4, 2014, at 12:14 , John Baldwin <jhb@FreeBSD.org> wrote: > On Tuesday, March 04, 2014 12:09:47 am George V. Neville-Neil wrote: >> Author: gnn >> Date: Tue Mar 4 05:09:46 2014 >> New Revision: 262727 >> URL: http://svnweb.freebsd.org/changeset/base/262727 >>=20 >> Log: >> Naming consistency fix. The routing code defines >> RADIX_NODE_HEAD_LOCK as grabbing the write lock, >> but RADIX_NODE_HEAD_LOCK_ASSERT as checking the read lock. >=20 > Actually, that isn't what RA_LOCKED means. RA_LOCKED means that it is > either read- or write-locked. Note that you have now made=20 > RADIX_NODE_HEAD_LOCK_ASSERT() a redundant copy of=20 > RADIX_NODE_HEAD_WLOCK_ASSERT(). You should revert that part in some > way (either remove HEAD_LOCK_ASSERT() entirely leaving just = RLOCK_ASSERT() and=20 > WLOCK_ASSERT(), or restore HEAD_LOCK_ASSERT() to using RA_LOCKED if = there are=20 > places that want to assert that the lock is held, but don't care if it = is read=20 > or write). Actually I=92ll revert the whole thing and to back to Vijay to rework = this. Best, George
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?FD8EBD57-749C-4247-8CA2-E846E046476F>