diff -r c7137074391a -r dec9340cfc71 sys/netinet/cc/cc.c --- a/sys/netinet/cc/cc.c Sun Nov 14 15:46:25 2010 +1100 +++ b/sys/netinet/cc/cc.c Sun Nov 14 17:15:06 2010 +1100 @@ -190,10 +190,7 @@ cc_deregister_algo(struct cc_algo *remove_cc) { struct cc_algo *funcs, *tmpfuncs; - struct tcpcb *tp; - struct inpcb *inp; int err; - VNET_ITERATOR_DECL(vnet_iter); err = ENOENT; @@ -220,53 +217,14 @@ } CC_LIST_WUNLOCK(); - if (!err) { + if (!err) /* - * Check all active control blocks across all network stacks and - * change any that are using this algorithm back to newreno. If - * the algorithm that was in use requires cleanup code to be - * run, call it. - * - * New connections already part way through being initialised - * with the CC algo we're removing will not race with this code - * because the INP_INFO_WLOCK is held during initialisation. - * We therefore don't enter the loop below until the connection - * list has stabilised. + * XXXLAS: + * - We may need to handle non-zero return values in future. + * - If we add CC framework support for protocols other than + * TCP, we may want a more generic way to handle this step. */ - VNET_LIST_RLOCK(); - VNET_FOREACH(vnet_iter) { - CURVNET_SET(vnet_iter); - INP_INFO_RLOCK(&V_tcbinfo); - LIST_FOREACH(inp, &V_tcb, inp_list) { - INP_WLOCK(inp); - /* Important to skip tcptw structs. */ - if (!(inp->inp_flags & INP_TIMEWAIT) && - (tp = intotcpcb(inp)) != NULL) { - /* - * By holding INP_WLOCK here, we are - * assured that the connection is not - * currently executing inside the CC - * module's functions i.e. it is safe - * to make the switch back to newreno. - */ - if (CC_ALGO(tp) == remove_cc) { - tmpfuncs = CC_ALGO(tp); - /* - * Newreno does not - * require any init. - */ - CC_ALGO(tp) = &newreno_cc_algo; - if (tmpfuncs->cb_destroy != NULL) - tmpfuncs->cb_destroy(tp->ccv); - } - } - INP_WUNLOCK(inp); - } - INP_INFO_RUNLOCK(&V_tcbinfo); - CURVNET_RESTORE(); - } - VNET_LIST_RUNLOCK(); - } + tcp_ccalgounload(remove_cc); return (err); } diff -r c7137074391a -r dec9340cfc71 sys/netinet/tcp_subr.c --- a/sys/netinet/tcp_subr.c Sun Nov 14 15:46:25 2010 +1100 +++ b/sys/netinet/tcp_subr.c Sun Nov 14 17:15:06 2010 +1100 @@ -708,6 +708,67 @@ } /* + * Switch the congestion control algorithm back to NewReno for any active + * control blocks using an algorithm which is about to go away. + * This ensures the CC framework can allow the unload to proceed without leaving + * any dangling pointers which would trigger a panic. + * Returning non-zero would inform the CC framework that something went wrong + * and it would be unsafe to allow the unload to proceed. However, there is no + * way for this to occur with this implementation so we always return zero. + */ +int +tcp_ccalgounload(struct cc_algo *unload_algo) +{ + struct cc_algo *tmpalgo; + struct inpcb *inp; + struct tcpcb *tp; + VNET_ITERATOR_DECL(vnet_iter); + + /* + * Check all active control blocks across all network stacks and change + * any that are using "unload_algo" back to NewReno. If "unload_algo" + * requires cleanup code to be run, call it. + * + * New connections already part way through being initialised with the + * CC algo we're removing will not race with this code because the + * INP_INFO_WLOCK is held during initialisation. We therefore don't + * enter the loop below until the connection list has stabilised. + */ + VNET_LIST_RLOCK(); + VNET_FOREACH(vnet_iter) { + CURVNET_SET(vnet_iter); + INP_INFO_RLOCK(&V_tcbinfo); + LIST_FOREACH(inp, &V_tcb, inp_list) { + INP_WLOCK(inp); + /* Important to skip tcptw structs. */ + if (!(inp->inp_flags & INP_TIMEWAIT) && + (tp = intotcpcb(inp)) != NULL) { + /* + * By holding INP_WLOCK here, we are assured + * that the connection is not currently + * executing inside the CC module's functions + * i.e. it is safe to make the switch back to + * newreno. + */ + if (CC_ALGO(tp) == unload_algo) { + tmpalgo = CC_ALGO(tp); + /* Newreno does not require any init. */ + CC_ALGO(tp) = &newreno_cc_algo; + if (tmpalgo->cb_destroy != NULL) + tmpalgo->cb_destroy(tp->ccv); + } + } + INP_WUNLOCK(inp); + } + INP_INFO_RUNLOCK(&V_tcbinfo); + CURVNET_RESTORE(); + } + VNET_LIST_RUNLOCK(); + + return (0); +} + +/* * Drop a TCP connection, reporting * the specified error. If connection is synchronized, * then send a RST to peer. diff -r c7137074391a -r dec9340cfc71 sys/netinet/tcp_var.h --- a/sys/netinet/tcp_var.h Sun Nov 14 15:46:25 2010 +1100 +++ b/sys/netinet/tcp_var.h Sun Nov 14 17:15:06 2010 +1100 @@ -605,6 +605,7 @@ #define V_tcp_ecn_maxretries VNET(tcp_ecn_maxretries) int tcp_addoptions(struct tcpopt *, u_char *); +int tcp_ccalgounload(struct cc_algo *unload_algo); struct tcpcb * tcp_close(struct tcpcb *); void tcp_discardcb(struct tcpcb *);