From: Herbert Poetzl (herbert_at_13thfloor.at)
Date: Sat 20 Dec 2003 - 02:33:32 GMT
Hi!
just wanted to find out how far the rabbit hole goes ...
browsing the source, I found the folowing
locations for possible s_info races ...
forget_original_parent()
goodness()
schedule()
update_process_times()
vc_new_s_context()
and of course similar is true for ip_info :(
there are two 'general' solutions to this
issue, besides the obvious one (adding tons
of lock/unlock pairs)
* replace alloc/dealloc with get/put and
change if (x->s_info) ... to if (.._get(x->s_info))
* add a dummy s_info/ip_info for the host
context (this can result in data corruption)
so I guess we are going for the get/put approach
best,
Herbert
_______________________________________________
Vserver mailing list
Vserver_at_list.linux-vserver.org
http://list.linux-vserver.org/mailman/listinfo/vserver