Date: Wed, 11 Aug 2004 18:30:29 -0400 (EDT) From: Garrett Wollman <wollman@khavrinen.lcs.mit.edu> To: Andre Oppermann <andre@FreeBSD.ORG> Cc: cvs-all@FreeBSD.ORG Subject: Re: cvs commit: src/sys/netinet ip_divert.c raw_ip.ctcp_hostcache.c tcp_subr.c tcp_syncache.c udp_usrreq.c Message-ID: <200408112230.i7BMUTfr077252@khavrinen.lcs.mit.edu> In-Reply-To: <411A60D4.7C25D34E@freebsd.org> References: <20040811175052.GA37093@freefall.freebsd.org> <411A60D4.7C25D34E@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
<<On Wed, 11 Aug 2004 20:09:24 +0200, Andre Oppermann <andre@FreeBSD.ORG> said: > In all cases the zone is used as kmalloc replacement. It always > initializes a newly allocated object. It always unlinks the object > from any lists before it uma_zfree's it. All such operations are > covered by mutexes, no concurrent access should happen. If there > is any use-after-free then it is a bug that needs to be fixed. No, it is an intentional design choice, not a bug. Lock-free synchronization is good, particularly when you're copying out a giant wodge of PCBs to userland (or doing anything else that might take a really long time). Locks should not be held during this process. -GAWollman
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200408112230.i7BMUTfr077252>