PSARC 2006/366 IP Instances

6289221 RFE: Need virtualized ip-stack for each local zone
6512601 panic in ipsec_in_tag - allocation failure
6514637 error message from dhcpagent: add_pkt_opt: option type 60 is missing required value
6364643 RFE: allow persistent setting of interface flags per zone
6307539 RFE: Invalid network address causes zone boot failure
5041214 Allow IPMP configuration with zones
5005887 RFE: zoneadmd should support plumbing an interface via DHCP
4991139 RFE: zones should provide a mechanism to configure a defaultrouter for a zone
6218378 zoneadmd doesn't set the netmask for non-loopback addresses hosted on lo0
4963280 zones: need to virtualize the IPv6 default address selection mechanism
4963285 zones: need support of stateless address autoconfiguration for IPv6
5048068 zones don't boot if one of its interfaces has failed
5057154 RFE: ability to change interface status from within a zone
4963287 zones should support the plumbing of the first (and only) logical interface
4978517 TCP privileged port space should be partitioned per zone
5023347 zones don't work well with network routes other than default
4963372 investigate whether global zone can act as a router for local zones
6378364 RFE: Allow each zone to have its own virtual IPFilter

1 files changed, 50 insertions, 13 deletions

diff --git a/usr/src/uts/common/inet/tcp.h b/usr/src/uts/common/inet/tcp.h
index a586064d3b..6937e3e9e0 100644
--- a/usr/src/uts/common/inet/tcp.h
+++ b/usr/src/uts/common/inet/tcp.h
@@ -36,9 +36,16 @@ extern "C" {
 #include <sys/inttypes.h>
 #include <netinet/ip6.h>
 #include <netinet/tcp.h>
-#include <inet/tcp_sack.h>
 #include <sys/socket.h>
 #include <sys/multidata.h>
+#include <sys/md5.h>
+#include <inet/common.h>
+#include <inet/ip.h>
+#include <inet/ip6.h>
+#include <inet/mi.h>
+#include <inet/mib2.h>
+#include <inet/tcp_stack.h>
+#include <inet/tcp_sack.h>
 #include <inet/kssl/ksslapi.h>
 
 /*
@@ -142,12 +149,14 @@ struct conn_s;
 
 typedef struct tcp_s {
 				/* Pointer to previous bind hash next. */
-	struct tcp_s *tcp_time_wait_next;
+	struct tcp_s	*tcp_time_wait_next;
 				/* Pointer to next T/W block */
-	struct tcp_s *tcp_time_wait_prev;
+	struct tcp_s	*tcp_time_wait_prev;
 				/* Pointer to previous T/W next */
-	clock_t	tcp_time_wait_expire;
-	struct conn_s *tcp_connp;
+	clock_t		tcp_time_wait_expire;
+
+	struct conn_s	*tcp_connp;
+	tcp_stack_t	*tcp_tcps;	/* Shortcut via conn_netstack */
 
 	int32_t	tcp_state;
 	int32_t	tcp_rcv_ws;		/* My window scale power */
@@ -603,16 +612,48 @@ typedef struct tcp_s {
 #define	TCP_DEBUG_GETPCSTACK(buffer, depth)
 #endif
 
+/*
+ * Track a reference count on the tcps in order to know when
+ * the tcps_g_q can be removed. As long as there is any
+ * tcp_t, other that the tcps_g_q itself, in the tcp_stack_t we
+ * need to keep tcps_g_q around so that a closing connection can
+ * switch to using tcps_g_q as part of it closing.
+ */
+#define	TCPS_REFHOLD(tcps) {					\
+	atomic_add_32(&(tcps)->tcps_refcnt, 1);			\
+	ASSERT((tcps)->tcps_refcnt != 0);			\
+	DTRACE_PROBE1(tcps__refhold, tcp_stack_t, tcps);	\
+}
+
+/*
+ * Decrement the reference count on the tcp_stack_t.
+ * In architectures e.g sun4u, where atomic_add_32_nv is just
+ * a cas, we need to maintain the right memory barrier semantics
+ * as that of mutex_exit i.e all the loads and stores should complete
+ * before the cas is executed. membar_exit() does that here.
+ */
+#define	TCPS_REFRELE(tcps) {					\
+	ASSERT((tcps)->tcps_refcnt != 0);			\
+	membar_exit();						\
+	DTRACE_PROBE1(tcps__refrele, tcp_stack_t, tcps);	\
+	if (atomic_add_32_nv(&(tcps)->tcps_refcnt, -1) == 0 &&	\
+	    (tcps)->tcps_g_q != NULL) {				\
+		/* Only tcps_g_q left */			\
+		tcp_g_q_inactive(tcps);				\
+	}							\
+}
+
 extern void 	tcp_free(tcp_t *tcp);
-extern void	tcp_ddi_init(void);
-extern void	tcp_ddi_destroy(void);
+extern void	tcp_ddi_g_init(void);
+extern void	tcp_ddi_g_destroy(void);
+extern void	tcp_g_q_inactive(tcp_stack_t *);
 extern void	tcp_xmit_listeners_reset(mblk_t *mp, uint_t ip_hdr_len,
-		    zoneid_t zoneid);
+    zoneid_t zoneid, tcp_stack_t *);
 extern void	tcp_conn_request(void *arg, mblk_t *mp, void *arg2);
 extern void	tcp_conn_request_unbound(void *arg, mblk_t *mp, void *arg2);
 extern void 	tcp_input(void *arg, mblk_t *mp, void *arg2);
 extern void	tcp_rput_data(void *arg, mblk_t *mp, void *arg2);
-extern void 	*tcp_get_conn(void *arg);
+extern void 	*tcp_get_conn(void *arg, tcp_stack_t *);
 extern void	tcp_time_wait_collector(void *arg);
 extern int	tcp_snmp_get(queue_t *, mblk_t *);
 extern int	tcp_snmp_set(queue_t *, int, int, uchar_t *, int len);
@@ -627,15 +668,11 @@ extern mblk_t	*tcp_xmit_mp(tcp_t *tcp, mblk_t *mp, int32_t max_to_send,
  *
  * The listener and acceptor hash queues are lists of tcp_t.
  */
-
 /* listener hash and acceptor hash queue head */
 typedef struct tf_s {
 	tcp_t		*tf_tcp;
 	kmutex_t	tf_lock;
 } tf_t;
-
-extern mib2_tcp_t tcp_mib;
-
 #endif	/* (defined(_KERNEL) || defined(_KMEMUSER)) */
 
 /* Contract private interface between TCP and Clustering. */