svn commit: r333744 - in head/sys: kern sys
Matt Macy
mmacy at FreeBSD.org
Thu May 17 17:59:36 UTC 2018
Author: mmacy
Date: Thu May 17 17:59:35 2018
New Revision: 333744
URL: https://svnweb.freebsd.org/changeset/base/333744
Log:
AF_UNIX: make unix socket locking finer grained
This change moves to using a reference count across lock drop / reacquire
to guarantee liveness.
Currently sends on unix sockets contend heavily on read locking the list lock.
unix1_processes in will-it-scale peaks at 6 processes and then declines.
With this change I get a substantial improvement in number of operations per second
with 96 processes:
x before
+ after
N Min Max Median Avg Stddev
x 11 1688420 1696389 1693578 1692766.3 2971.1702
+ 10 63417955 71030114 70662504 69576423 2374684.6
Difference at 95.0% confidence
6.78837e+07 +/- 1.49463e+06
4010.22% +/- 88.4246%
(Student's t, pooled s = 1.63437e+06)
And even for 2 processes shows a ~18% improvement.
"Small" iron changes (1, 2, and 4 processes):
x before1
+ after1.2
+------------------------------------------------------------------------+
| + |
| x + |
| x + |
| x + |
| x ++ |
| xx ++ |
|x x xx ++ |
| |__________________A_____M_____AM____||
+------------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 10 1131648 1197750 1197138.5 1190369.3 20651.839
+ 10 1203840 1205056 1204919 1204827.9 353.27404
Difference at 95.0% confidence
14458.6 +/- 13723
1.21463% +/- 1.16683%
(Student's t, pooled s = 14605.2)
x before2
+ after2.2
+------------------------------------------------------------------------+
| +|
| +|
| +|
| +|
| +|
| +|
| x +|
| x +|
| x xx +|
|x xxxx +|
| |___AM_| A|
+------------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 10 1972843 2045866 2038186.5 2030443.8 21367.694
+ 10 2400853 2402196 2401043.5 2401172.7 385.40024
Difference at 95.0% confidence
370729 +/- 14198.9
18.2585% +/- 0.826943%
(Student's t, pooled s = 15111.7)
x before4
+ after4.2
N Min Max Median Avg Stddev
x 10 3986994 3991728 3990137.5 3989985.2 1300.0164
+ 10 4799990 4806664 4806116.5 4805194 1990.6625
Difference at 95.0% confidence
815209 +/- 1579.64
20.4314% +/- 0.0421713%
(Student's t, pooled s = 1681.19)
Tested by: pho
Reported by: mjg
Approved by: sbruno
Sponsored by: Limelight Networks
Differential Revision: https://reviews.freebsd.org/D15430
Modified:
head/sys/kern/uipc_usrreq.c
head/sys/sys/unpcb.h
Modified: head/sys/kern/uipc_usrreq.c
==============================================================================
--- head/sys/kern/uipc_usrreq.c Thu May 17 17:57:41 2018 (r333743)
+++ head/sys/kern/uipc_usrreq.c Thu May 17 17:59:35 2018 (r333744)
@@ -191,13 +191,41 @@ SYSCTL_INT(_net_local, OID_AUTO, deferred, CTLFLAG_RD,
/*
* Locking and synchronization:
*
- * Two types of locks exist in the local domain socket implementation: a
- * a global linkage rwlock and per-unpcb mutexes. The linkage lock protects
- * the socket count, global generation number, stream/datagram global lists and
- * interconnection of unpcbs, the v_socket and unp_vnode pointers, and can be
- * held exclusively over the acquisition of multiple unpcb locks to prevent
- * deadlock.
+ * Three types of locks exist in the local domain socket implementation: a
+ * a global linkage rwlock, the mtxpool lock, and per-unpcb mutexes.
+ * The linkage lock protects the socket count, global generation number,
+ * and stream/datagram global lists.
*
+ * The mtxpool lock protects the vnode from being modified while referenced.
+ * Lock ordering requires that it be acquired before any unpcb locks.
+ *
+ * The unpcb lock (unp_mtx) protects all fields in the unpcb. Of particular
+ * note is that this includes the unp_conn field. So long as the unpcb lock
+ * is held the reference to the unpcb pointed to by unp_conn is valid. If we
+ * require that the unpcb pointed to by unp_conn remain live in cases where
+ * we need to drop the unp_mtx as when we need to acquire the lock for a
+ * second unpcb the caller must first acquire an additional reference on the
+ * second unpcb and then revalidate any state (typically check that unp_conn
+ * is non-NULL) upon requiring the initial unpcb lock. The lock ordering
+ * between unpcbs is the conventional ascending address order. Two helper
+ * routines exist for this:
+ *
+ * - unp_pcb_lock2(unp, unp2) - which just acquires the two locks in the
+ * safe ordering.
+ *
+ * - unp_pcb_owned_lock2(unp, unp2, freed) - the lock for unp is held
+ * when called. If unp is unlocked and unp2 is subsequently freed
+ * freed will be set to 1.
+ *
+ * The helper routines for references are:
+ *
+ * - unp_pcb_hold(unp): Can be called any time we currently hold a valid
+ * reference to unp.
+ *
+ * - unp_pcb_rele(unp): The caller must hold the unp lock. If we are
+ * releasing the last reference, detach must have been called thus
+ * unp->unp_socket be NULL.
+ *
* UNIX domain sockets each have an unpcb hung off of their so_pcb pointer,
* allocated in pru_attach() and freed in pru_detach(). The validity of that
* pointer is an invariant, so no lock is required to dereference the so_pcb
@@ -210,16 +238,9 @@ SYSCTL_INT(_net_local, OID_AUTO, deferred, CTLFLAG_RD,
* to the unpcb is held. Typically, this reference will be from the socket,
* or from another unpcb when the referring unpcb's lock is held (in order
* that the reference not be invalidated during use). For example, to follow
- * unp->unp_conn->unp_socket, you need unlock the lock on unp, not unp_conn,
- * as unp_socket remains valid as long as the reference to unp_conn is valid.
+ * unp->unp_conn->unp_socket, you need to hold a lock on unp_conn to guarantee
+ * that detach is not run clearing unp_socket.
*
- * Fields of unpcbss are locked using a per-unpcb lock, unp_mtx. Individual
- * atomic reads without the lock may be performed "lockless", but more
- * complex reads and read-modify-writes require the mutex to be held. No
- * lock order is defined between unpcb locks -- multiple unpcb locks may be
- * acquired at the same time only when holding the linkage rwlock
- * exclusively, which prevents deadlocks.
- *
* Blocking with UNIX domain sockets is a tricky issue: unlike most network
* protocols, bind() is a non-atomic operation, and connect() requires
* potential sleeping in the protocol, due to potentially waiting on local or
@@ -257,13 +278,19 @@ static struct mtx unp_defers_lock;
#define UNP_DEFERRED_LOCK() mtx_lock(&unp_defers_lock)
#define UNP_DEFERRED_UNLOCK() mtx_unlock(&unp_defers_lock)
+#define UNP_REF_LIST_LOCK() UNP_DEFERRED_LOCK();
+#define UNP_REF_LIST_UNLOCK() UNP_DEFERRED_UNLOCK();
+
#define UNP_PCB_LOCK_INIT(unp) mtx_init(&(unp)->unp_mtx, \
"unp_mtx", "unp_mtx", \
- MTX_DUPOK|MTX_DEF|MTX_RECURSE)
+ MTX_DUPOK|MTX_DEF)
#define UNP_PCB_LOCK_DESTROY(unp) mtx_destroy(&(unp)->unp_mtx)
#define UNP_PCB_LOCK(unp) mtx_lock(&(unp)->unp_mtx)
+#define UNP_PCB_TRYLOCK(unp) mtx_trylock(&(unp)->unp_mtx)
#define UNP_PCB_UNLOCK(unp) mtx_unlock(&(unp)->unp_mtx)
+#define UNP_PCB_OWNED(unp) mtx_owned(&(unp)->unp_mtx)
#define UNP_PCB_LOCK_ASSERT(unp) mtx_assert(&(unp)->unp_mtx, MA_OWNED)
+#define UNP_PCB_UNLOCK_ASSERT(unp) mtx_assert(&(unp)->unp_mtx, MA_NOTOWNED)
static int uipc_connect2(struct socket *, struct socket *);
static int uipc_ctloutput(struct socket *, struct sockopt *);
@@ -289,6 +316,75 @@ static int unp_externalize_fp(struct file *);
static struct mbuf *unp_addsockcred(struct thread *, struct mbuf *);
static void unp_process_defers(void * __unused, int);
+
+static void
+unp_pcb_hold(struct unpcb *unp)
+{
+ MPASS(unp->unp_refcount);
+ refcount_acquire(&unp->unp_refcount);
+}
+
+static int
+unp_pcb_rele(struct unpcb *unp)
+{
+ int freed;
+
+ UNP_PCB_LOCK_ASSERT(unp);
+ MPASS(unp->unp_refcount);
+ if ((freed = refcount_release(&unp->unp_refcount))) {
+ /* we got here with having detached? */
+ MPASS(unp->unp_socket == NULL);
+ UNP_PCB_UNLOCK(unp);
+ UNP_PCB_LOCK_DESTROY(unp);
+ uma_zfree(unp_zone, unp);
+ }
+ return (freed);
+}
+
+static void
+unp_pcb_lock2(struct unpcb *unp, struct unpcb *unp2)
+{
+ UNP_PCB_UNLOCK_ASSERT(unp);
+ UNP_PCB_UNLOCK_ASSERT(unp2);
+ if ((uintptr_t)unp2 > (uintptr_t)unp) {
+ UNP_PCB_LOCK(unp);
+ UNP_PCB_LOCK(unp2);
+ } else {
+ UNP_PCB_LOCK(unp2);
+ UNP_PCB_LOCK(unp);
+ }
+}
+
+static __noinline void
+unp_pcb_owned_lock2_slowpath(struct unpcb *unp, struct unpcb **unp2p, int *freed)
+
+{
+ struct unpcb *unp2;
+
+ unp2 = *unp2p;
+ unp_pcb_hold((unp2));
+ UNP_PCB_UNLOCK((unp));
+ UNP_PCB_LOCK((unp2));
+ UNP_PCB_LOCK((unp));
+ *freed = unp_pcb_rele((unp2));
+ if (*freed)
+ *unp2p = NULL;
+}
+
+#define unp_pcb_owned_lock2(unp, unp2, freed) do { \
+ freed = 0; \
+ UNP_PCB_LOCK_ASSERT((unp)); \
+ UNP_PCB_UNLOCK_ASSERT((unp2)); \
+ if (__predict_true(UNP_PCB_TRYLOCK((unp2)))) \
+ break; \
+ else if ((uintptr_t)(unp2) > (uintptr_t)(unp)) \
+ UNP_PCB_LOCK((unp2)); \
+ else { \
+ unp_pcb_owned_lock2_slowpath((unp), &(unp2), &freed); \
+ } \
+} while (0)
+
+
/*
* Definitions of protocols supported in the LOCAL domain.
*/
@@ -344,17 +440,16 @@ uipc_abort(struct socket *so)
unp = sotounpcb(so);
KASSERT(unp != NULL, ("uipc_abort: unp == NULL"));
+ UNP_PCB_UNLOCK_ASSERT(unp);
- UNP_LINK_WLOCK();
UNP_PCB_LOCK(unp);
unp2 = unp->unp_conn;
if (unp2 != NULL) {
- UNP_PCB_LOCK(unp2);
+ unp_pcb_hold(unp2);
+ UNP_PCB_UNLOCK(unp);
unp_drop(unp2);
- UNP_PCB_UNLOCK(unp2);
- }
- UNP_PCB_UNLOCK(unp);
- UNP_LINK_WUNLOCK();
+ } else
+ UNP_PCB_UNLOCK(unp);
}
static int
@@ -551,14 +646,12 @@ restart:
ASSERT_VOP_ELOCKED(vp, "uipc_bind");
soun = (struct sockaddr_un *)sodupsockaddr(nam, M_WAITOK);
- UNP_LINK_WLOCK();
UNP_PCB_LOCK(unp);
VOP_UNP_BIND(vp, unp);
unp->unp_vnode = vp;
unp->unp_addr = soun;
unp->unp_flags &= ~UNP_BINDING;
UNP_PCB_UNLOCK(unp);
- UNP_LINK_WUNLOCK();
VOP_UNLOCK(vp, 0);
vn_finished_write(mp);
free(buf, M_TEMP);
@@ -585,9 +678,7 @@ uipc_connect(struct socket *so, struct sockaddr *nam,
int error;
KASSERT(td == curthread, ("uipc_connect: td != curthread"));
- UNP_LINK_WLOCK();
error = unp_connect(so, nam, td);
- UNP_LINK_WUNLOCK();
return (error);
}
@@ -598,9 +689,7 @@ uipc_connectat(int fd, struct socket *so, struct socka
int error;
KASSERT(td == curthread, ("uipc_connectat: td != curthread"));
- UNP_LINK_WLOCK();
error = unp_connectat(fd, so, nam, td);
- UNP_LINK_WUNLOCK();
return (error);
}
@@ -609,26 +698,41 @@ uipc_close(struct socket *so)
{
struct unpcb *unp, *unp2;
struct vnode *vp = NULL;
-
+ struct mtx *vplock;
+ int freed;
unp = sotounpcb(so);
KASSERT(unp != NULL, ("uipc_close: unp == NULL"));
- UNP_LINK_WLOCK();
+
+ vplock = NULL;
+ if ((vp = unp->unp_vnode) != NULL) {
+ vplock = mtx_pool_find(mtxpool_sleep, vp);
+ mtx_lock(vplock);
+ }
UNP_PCB_LOCK(unp);
- unp2 = unp->unp_conn;
- if (unp2 != NULL) {
- UNP_PCB_LOCK(unp2);
- unp_disconnect(unp, unp2);
- UNP_PCB_UNLOCK(unp2);
+ if (vp && unp->unp_vnode == NULL) {
+ mtx_unlock(vplock);
+ vp = NULL;
}
- if (SOLISTENING(so) && ((vp = unp->unp_vnode) != NULL)) {
+ if (vp != NULL) {
VOP_UNP_DETACH(vp);
unp->unp_vnode = NULL;
}
- UNP_PCB_UNLOCK(unp);
- UNP_LINK_WUNLOCK();
- if (vp)
+ unp2 = unp->unp_conn;
+ unp_pcb_hold(unp);
+ if (unp2 != NULL) {
+ unp_pcb_hold(unp2);
+ unp_pcb_owned_lock2(unp, unp2, freed);
+ unp_disconnect(unp, unp2);
+ if (unp_pcb_rele(unp2) == 0)
+ UNP_PCB_UNLOCK(unp2);
+ }
+ if (unp_pcb_rele(unp) == 0)
+ UNP_PCB_UNLOCK(unp);
+ if (vp) {
+ mtx_unlock(vplock);
vrele(vp);
+ }
}
static int
@@ -637,17 +741,14 @@ uipc_connect2(struct socket *so1, struct socket *so2)
struct unpcb *unp, *unp2;
int error;
- UNP_LINK_WLOCK();
unp = so1->so_pcb;
KASSERT(unp != NULL, ("uipc_connect2: unp == NULL"));
- UNP_PCB_LOCK(unp);
unp2 = so2->so_pcb;
KASSERT(unp2 != NULL, ("uipc_connect2: unp2 == NULL"));
- UNP_PCB_LOCK(unp2);
+ unp_pcb_lock2(unp, unp2);
error = unp_connect2(so1, so2, PRU_CONNECT2);
UNP_PCB_UNLOCK(unp2);
UNP_PCB_UNLOCK(unp);
- UNP_LINK_WUNLOCK();
return (error);
}
@@ -655,6 +756,7 @@ static void
uipc_detach(struct socket *so)
{
struct unpcb *unp, *unp2;
+ struct mtx *vplock;
struct sockaddr_un *saved_unp_addr;
struct vnode *vp;
int freeunp, local_unp_rights;
@@ -669,49 +771,77 @@ uipc_detach(struct socket *so)
LIST_REMOVE(unp, unp_link);
unp->unp_gencnt = ++unp_gencnt;
--unp_count;
+ UNP_LINK_WUNLOCK();
+
+ UNP_PCB_UNLOCK_ASSERT(unp);
+ restart:
+ if ((vp = unp->unp_vnode) != NULL) {
+ vplock = mtx_pool_find(mtxpool_sleep, vp);
+ mtx_lock(vplock);
+ }
UNP_PCB_LOCK(unp);
- if ((unp->unp_flags & UNP_NASCENT) != 0)
+ if ((unp2 = unp->unp_conn) != NULL) {
+ unp_pcb_owned_lock2(unp, unp2, freeunp);
+ if (freeunp)
+ unp2 = NULL;
+ }
+ if (unp->unp_vnode != vp &&
+ unp->unp_vnode != NULL) {
+ mtx_unlock(vplock);
+ UNP_PCB_UNLOCK(unp);
+ if (unp2)
+ UNP_PCB_UNLOCK(unp2);
+ goto restart;
+ }
+ if ((unp->unp_flags & UNP_NASCENT) != 0) {
+ if (unp2)
+ UNP_PCB_UNLOCK(unp2);
goto teardown;
-
+ }
if ((vp = unp->unp_vnode) != NULL) {
VOP_UNP_DETACH(vp);
unp->unp_vnode = NULL;
}
- unp2 = unp->unp_conn;
+ unp_pcb_hold(unp);
if (unp2 != NULL) {
- UNP_PCB_LOCK(unp2);
+ unp_pcb_hold(unp2);
unp_disconnect(unp, unp2);
- UNP_PCB_UNLOCK(unp2);
+ if (unp_pcb_rele(unp2) == 0)
+ UNP_PCB_UNLOCK(unp2);
}
-
- /*
- * We hold the linkage lock exclusively, so it's OK to acquire
- * multiple pcb locks at a time.
- */
+ UNP_PCB_UNLOCK(unp);
+ UNP_REF_LIST_LOCK();
while (!LIST_EMPTY(&unp->unp_refs)) {
struct unpcb *ref = LIST_FIRST(&unp->unp_refs);
- UNP_PCB_LOCK(ref);
+ unp_pcb_hold(ref);
+ UNP_REF_LIST_UNLOCK();
+
+ MPASS(ref != unp);
+ UNP_PCB_UNLOCK_ASSERT(ref);
unp_drop(ref);
- UNP_PCB_UNLOCK(ref);
+ UNP_REF_LIST_LOCK();
}
+
+ UNP_REF_LIST_UNLOCK();
+ UNP_PCB_LOCK(unp);
+ freeunp = unp_pcb_rele(unp);
+ MPASS(freeunp == 0);
local_unp_rights = unp_rights;
teardown:
- UNP_LINK_WUNLOCK();
unp->unp_socket->so_pcb = NULL;
saved_unp_addr = unp->unp_addr;
unp->unp_addr = NULL;
- unp->unp_refcount--;
- freeunp = (unp->unp_refcount == 0);
+ unp->unp_socket = NULL;
+ freeunp = unp_pcb_rele(unp);
if (saved_unp_addr != NULL)
free(saved_unp_addr, M_SONAME);
- if (freeunp) {
- UNP_PCB_LOCK_DESTROY(unp);
- uma_zfree(unp_zone, unp);
- } else
+ if (!freeunp)
UNP_PCB_UNLOCK(unp);
- if (vp)
+ if (vp) {
+ mtx_unlock(vplock);
vrele(vp);
+ }
if (local_unp_rights)
taskqueue_enqueue_timeout(taskqueue_thread, &unp_gc_task, -1);
}
@@ -720,20 +850,28 @@ static int
uipc_disconnect(struct socket *so)
{
struct unpcb *unp, *unp2;
+ int freed;
unp = sotounpcb(so);
KASSERT(unp != NULL, ("uipc_disconnect: unp == NULL"));
- UNP_LINK_WLOCK();
UNP_PCB_LOCK(unp);
- unp2 = unp->unp_conn;
- if (unp2 != NULL) {
- UNP_PCB_LOCK(unp2);
- unp_disconnect(unp, unp2);
- UNP_PCB_UNLOCK(unp2);
+ if ((unp2 = unp->unp_conn) == NULL) {
+ UNP_PCB_UNLOCK(unp);
+ return (0);
}
- UNP_PCB_UNLOCK(unp);
- UNP_LINK_WUNLOCK();
+ unp_pcb_owned_lock2(unp, unp2, freed);
+ if (__predict_false(freed)) {
+ UNP_PCB_UNLOCK(unp);
+ return (0);
+ }
+ unp_pcb_hold(unp2);
+ unp_pcb_hold(unp);
+ unp_disconnect(unp, unp2);
+ if (unp_pcb_rele(unp) == 0)
+ UNP_PCB_UNLOCK(unp);
+ if (unp_pcb_rele(unp2) == 0)
+ UNP_PCB_UNLOCK(unp2);
return (0);
}
@@ -852,13 +990,35 @@ uipc_rcvd(struct socket *so, int flags)
}
static int
+connect_internal(struct socket *so, struct sockaddr *nam, struct thread *td)
+{
+ int error;
+ struct unpcb *unp;
+
+ unp = so->so_pcb;
+ if (unp->unp_conn != NULL)
+ return (EISCONN);
+ error = unp_connect(so, nam, td);
+ if (error)
+ return (error);
+ UNP_PCB_LOCK(unp);
+ if (unp->unp_conn == NULL) {
+ UNP_PCB_UNLOCK(unp);
+ if (error == 0)
+ error = ENOTCONN;
+ }
+ return (error);
+}
+
+
+static int
uipc_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
struct mbuf *control, struct thread *td)
{
struct unpcb *unp, *unp2;
struct socket *so2;
u_int mbcnt, sbcc;
- int error = 0;
+ int freed, error;
unp = sotounpcb(so);
KASSERT(unp != NULL, ("%s: unp == NULL", __func__));
@@ -866,49 +1026,66 @@ uipc_send(struct socket *so, int flags, struct mbuf *m
so->so_type == SOCK_SEQPACKET,
("%s: socktype %d", __func__, so->so_type));
+ freed = error = 0;
if (flags & PRUS_OOB) {
error = EOPNOTSUPP;
goto release;
}
if (control != NULL && (error = unp_internalize(&control, td)))
goto release;
- if ((nam != NULL) || (flags & PRUS_EOF))
- UNP_LINK_WLOCK();
- else
- UNP_LINK_RLOCK();
+
+ unp2 = NULL;
switch (so->so_type) {
case SOCK_DGRAM:
{
const struct sockaddr *from;
- unp2 = unp->unp_conn;
if (nam != NULL) {
- UNP_LINK_WLOCK_ASSERT();
- if (unp2 != NULL) {
- error = EISCONN;
+ /*
+ * We return with UNP_PCB_LOCK_HELD so we know that
+ * the reference is live if the pointer is valid.
+ */
+ if ((error = connect_internal(so, nam, td)))
break;
- }
- error = unp_connect(so, nam, td);
- if (error)
- break;
+ MPASS(unp->unp_conn != NULL);
unp2 = unp->unp_conn;
- }
+ } else {
+ UNP_PCB_LOCK(unp);
+ /*
+ * Because connect() and send() are non-atomic in a sendto()
+ * with a target address, it's possible that the socket will
+ * have disconnected before the send() can run. In that case
+ * return the slightly counter-intuitive but otherwise
+ * correct error that the socket is not connected.
+ */
+ if ((unp2 = unp->unp_conn) == NULL) {
+ UNP_PCB_UNLOCK(unp);
+ error = ENOTCONN;
+ break;
+ }
+ }
+ unp_pcb_owned_lock2(unp, unp2, freed);
+ if (__predict_false(freed)) {
+ UNP_PCB_UNLOCK(unp);
+ error = ENOTCONN;
+ break;
+ }
/*
- * Because connect() and send() are non-atomic in a sendto()
- * with a target address, it's possible that the socket will
- * have disconnected before the send() can run. In that case
- * return the slightly counter-intuitive but otherwise
- * correct error that the socket is not connected.
+ * The socket referencing unp2 may have been closed
+ * or unp may have been disconnected if the unp lock
+ * was dropped to acquire unp2.
*/
- if (unp2 == NULL) {
+ if (__predict_false(unp->unp_conn == NULL) ||
+ unp2->unp_socket == NULL) {
+ UNP_PCB_UNLOCK(unp);
+ if (unp_pcb_rele(unp2) == 0)
+ UNP_PCB_UNLOCK(unp2);
error = ENOTCONN;
break;
}
- /* Lockless read. */
if (unp2->unp_flags & UNP_WANTCRED)
control = unp_addsockcred(td, control);
- UNP_PCB_LOCK(unp);
if (unp->unp_addr != NULL)
from = (struct sockaddr *)unp->unp_addr;
else
@@ -924,12 +1101,9 @@ uipc_send(struct socket *so, int flags, struct mbuf *m
SOCKBUF_UNLOCK(&so2->so_rcv);
error = ENOBUFS;
}
- if (nam != NULL) {
- UNP_LINK_WLOCK_ASSERT();
- UNP_PCB_LOCK(unp2);
+ if (nam != NULL)
unp_disconnect(unp, unp2);
- UNP_PCB_UNLOCK(unp2);
- }
+ UNP_PCB_UNLOCK(unp2);
UNP_PCB_UNLOCK(unp);
break;
}
@@ -938,42 +1112,37 @@ uipc_send(struct socket *so, int flags, struct mbuf *m
case SOCK_STREAM:
if ((so->so_state & SS_ISCONNECTED) == 0) {
if (nam != NULL) {
- UNP_LINK_WLOCK_ASSERT();
- error = unp_connect(so, nam, td);
- if (error)
- break; /* XXX */
- } else {
+ if ((error = connect_internal(so, nam, td)))
+ break;
+ } else {
error = ENOTCONN;
break;
}
- }
-
- /* Lockless read. */
- if (so->so_snd.sb_state & SBS_CANTSENDMORE) {
+ } else if ((unp2 = unp->unp_conn) == NULL) {
+ error = ENOTCONN;
+ break;
+ } else if (so->so_snd.sb_state & SBS_CANTSENDMORE) {
error = EPIPE;
break;
+ } else {
+ UNP_PCB_LOCK(unp);
+ if ((unp2 = unp->unp_conn) == NULL) {
+ UNP_PCB_UNLOCK(unp);
+ error = ENOTCONN;
+ break;
+ }
}
-
- /*
- * Because connect() and send() are non-atomic in a sendto()
- * with a target address, it's possible that the socket will
- * have disconnected before the send() can run. In that case
- * return the slightly counter-intuitive but otherwise
- * correct error that the socket is not connected.
- *
- * Locking here must be done carefully: the linkage lock
- * prevents interconnections between unpcbs from changing, so
- * we can traverse from unp to unp2 without acquiring unp's
- * lock. Socket buffer locks follow unpcb locks, so we can
- * acquire both remote and lock socket buffer locks.
- */
- unp2 = unp->unp_conn;
- if (unp2 == NULL) {
+ unp_pcb_owned_lock2(unp, unp2, freed);
+ UNP_PCB_UNLOCK(unp);
+ if (__predict_false(freed)) {
error = ENOTCONN;
break;
}
- so2 = unp2->unp_socket;
- UNP_PCB_LOCK(unp2);
+ if ((so2 = unp2->unp_socket) == NULL) {
+ UNP_PCB_UNLOCK(unp2);
+ error = ENOTCONN;
+ break;
+ }
SOCKBUF_LOCK(&so2->so_rcv);
if (unp2->unp_flags & UNP_WANTCRED) {
/*
@@ -1046,12 +1215,6 @@ uipc_send(struct socket *so, int flags, struct mbuf *m
unp_shutdown(unp);
UNP_PCB_UNLOCK(unp);
}
-
- if ((nam != NULL) || (flags & PRUS_EOF))
- UNP_LINK_WUNLOCK();
- else
- UNP_LINK_RUNLOCK();
-
if (control != NULL && error != 0)
unp_dispose_mbuf(control);
@@ -1124,12 +1287,10 @@ uipc_shutdown(struct socket *so)
unp = sotounpcb(so);
KASSERT(unp != NULL, ("uipc_shutdown: unp == NULL"));
- UNP_LINK_WLOCK();
UNP_PCB_LOCK(unp);
socantsendmore(so);
unp_shutdown(unp);
UNP_PCB_UNLOCK(unp);
- UNP_LINK_WUNLOCK();
return (0);
}
@@ -1333,16 +1494,11 @@ unp_connectat(int fd, struct socket *so, struct sockad
char buf[SOCK_MAXADDRLEN];
struct sockaddr *sa;
cap_rights_t rights;
- int error, len;
+ int error, len, freed;
+ struct mtx *vplock;
if (nam->sa_family != AF_UNIX)
return (EAFNOSUPPORT);
-
- UNP_LINK_WLOCK_ASSERT();
-
- unp = sotounpcb(so);
- KASSERT(unp != NULL, ("unp_connect: unp == NULL"));
-
if (nam->sa_len > sizeof(struct sockaddr_un))
return (EINVAL);
len = nam->sa_len - offsetof(struct sockaddr_un, sun_path);
@@ -1351,12 +1507,12 @@ unp_connectat(int fd, struct socket *so, struct sockad
bcopy(soun->sun_path, buf, len);
buf[len] = 0;
+ unp = sotounpcb(so);
UNP_PCB_LOCK(unp);
if (unp->unp_flags & UNP_CONNECTING) {
UNP_PCB_UNLOCK(unp);
return (EALREADY);
}
- UNP_LINK_WUNLOCK();
unp->unp_flags |= UNP_CONNECTING;
UNP_PCB_UNLOCK(unp);
@@ -1389,11 +1545,8 @@ unp_connectat(int fd, struct socket *so, struct sockad
unp = sotounpcb(so);
KASSERT(unp != NULL, ("unp_connect: unp == NULL"));
- /*
- * Lock linkage lock for two reasons: make sure v_socket is stable,
- * and to protect simultaneous locking of multiple pcbs.
- */
- UNP_LINK_WLOCK();
+ vplock = mtx_pool_find(mtxpool_sleep, vp);
+ mtx_lock(vplock);
VOP_UNP_CONNECT(vp, &unp2);
if (unp2 == NULL) {
error = ECONNREFUSED;
@@ -1404,8 +1557,7 @@ unp_connectat(int fd, struct socket *so, struct sockad
error = EPROTOTYPE;
goto bad2;
}
- UNP_PCB_LOCK(unp);
- UNP_PCB_LOCK(unp2);
+ unp_pcb_lock2(unp, unp2);
if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
if (so2->so_options & SO_ACCEPTCONN) {
CURVNET_SET(so2->so_vnet);
@@ -1418,7 +1570,9 @@ unp_connectat(int fd, struct socket *so, struct sockad
goto bad3;
}
unp3 = sotounpcb(so2);
- UNP_PCB_LOCK(unp3);
+ UNP_PCB_UNLOCK(unp);
+ unp_pcb_owned_lock2(unp2, unp3, freed);
+ MPASS(!freed);
if (unp2->unp_addr != NULL) {
bcopy(unp2->unp_addr, sa, unp2->unp_addr->sun_len);
unp3->unp_addr = (struct sockaddr_un *) sa;
@@ -1445,6 +1599,8 @@ unp_connectat(int fd, struct socket *so, struct sockad
unp3->unp_flags |= UNP_WANTCRED;
UNP_PCB_UNLOCK(unp2);
unp2 = unp3;
+ unp_pcb_owned_lock2(unp2, unp, freed);
+ MPASS(!freed);
#ifdef MAC
mac_socketpeer_set_from_socket(so, so2);
mac_socketpeer_set_from_socket(so2, so);
@@ -1459,12 +1615,12 @@ bad3:
UNP_PCB_UNLOCK(unp2);
UNP_PCB_UNLOCK(unp);
bad2:
- UNP_LINK_WUNLOCK();
+ mtx_unlock(vplock);
bad:
- if (vp != NULL)
+ if (vp != NULL) {
vput(vp);
+ }
free(sa, M_SONAME);
- UNP_LINK_WLOCK();
UNP_PCB_LOCK(unp);
unp->unp_flags &= ~UNP_CONNECTING;
UNP_PCB_UNLOCK(unp);
@@ -1482,7 +1638,6 @@ unp_connect2(struct socket *so, struct socket *so2, in
unp2 = sotounpcb(so2);
KASSERT(unp2 != NULL, ("unp_connect2: unp2 == NULL"));
- UNP_LINK_WLOCK_ASSERT();
UNP_PCB_LOCK_ASSERT(unp);
UNP_PCB_LOCK_ASSERT(unp2);
@@ -1490,10 +1645,13 @@ unp_connect2(struct socket *so, struct socket *so2, in
return (EPROTOTYPE);
unp2->unp_flags &= ~UNP_NASCENT;
unp->unp_conn = unp2;
-
+ unp_pcb_hold(unp2);
+ unp_pcb_hold(unp);
switch (so->so_type) {
case SOCK_DGRAM:
+ UNP_REF_LIST_LOCK();
LIST_INSERT_HEAD(&unp2->unp_refs, unp, unp_reflink);
+ UNP_REF_LIST_UNLOCK();
soisconnected(so);
break;
@@ -1517,31 +1675,48 @@ unp_connect2(struct socket *so, struct socket *so2, in
static void
unp_disconnect(struct unpcb *unp, struct unpcb *unp2)
{
- struct socket *so;
+ struct socket *so, *so2;
+ int rele, freed;
KASSERT(unp2 != NULL, ("unp_disconnect: unp2 == NULL"));
- UNP_LINK_WLOCK_ASSERT();
UNP_PCB_LOCK_ASSERT(unp);
UNP_PCB_LOCK_ASSERT(unp2);
+ if (unp->unp_conn == NULL && unp2->unp_conn == NULL)
+ return;
+
+ MPASS(unp->unp_conn == unp2);
unp->unp_conn = NULL;
+ rele = 0;
+ so = unp->unp_socket;
+ so2 = unp2->unp_socket;
switch (unp->unp_socket->so_type) {
case SOCK_DGRAM:
+ UNP_REF_LIST_LOCK();
LIST_REMOVE(unp, unp_reflink);
- so = unp->unp_socket;
- SOCK_LOCK(so);
- so->so_state &= ~SS_ISCONNECTED;
- SOCK_UNLOCK(so);
+ UNP_REF_LIST_UNLOCK();
+ if (so) {
+ SOCK_LOCK(so);
+ so->so_state &= ~SS_ISCONNECTED;
+ SOCK_UNLOCK(so);
+ }
break;
case SOCK_STREAM:
case SOCK_SEQPACKET:
- soisdisconnected(unp->unp_socket);
+ if (so)
+ soisdisconnected(so);
+ MPASS(unp2->unp_conn == unp);
unp2->unp_conn = NULL;
- soisdisconnected(unp2->unp_socket);
+ if (so2)
+ soisdisconnected(so2);
break;
}
+ freed = unp_pcb_rele(unp);
+ MPASS(freed == 0);
+ freed = unp_pcb_rele(unp2);
+ MPASS(freed == 0);
}
/*
@@ -1625,7 +1800,7 @@ unp_pcblist(SYSCTL_HANDLER_ARGS)
continue;
}
unp_list[i++] = unp;
- unp->unp_refcount++;
+ unp_pcb_hold(unp);
}
UNP_PCB_UNLOCK(unp);
}
@@ -1637,8 +1812,9 @@ unp_pcblist(SYSCTL_HANDLER_ARGS)
for (i = 0; i < n; i++) {
unp = unp_list[i];
UNP_PCB_LOCK(unp);
- unp->unp_refcount--;
- if (unp->unp_refcount != 0 && unp->unp_gencnt <= gencnt) {
+ freeunp = unp_pcb_rele(unp);
+
+ if (freeunp == 0 && unp->unp_gencnt <= gencnt) {
xu->xu_len = sizeof *xu;
xu->xu_unpp = unp;
/*
@@ -1665,14 +1841,8 @@ unp_pcblist(SYSCTL_HANDLER_ARGS)
sotoxsocket(unp->unp_socket, &xu->xu_socket);
UNP_PCB_UNLOCK(unp);
error = SYSCTL_OUT(req, xu, sizeof *xu);
- } else {
- freeunp = (unp->unp_refcount == 0);
+ } else if (freeunp == 0)
UNP_PCB_UNLOCK(unp);
- if (freeunp) {
- UNP_PCB_LOCK_DESTROY(unp);
- uma_zfree(unp_zone, unp);
- }
- }
}
free(xu, M_TEMP);
if (!error) {
@@ -1709,7 +1879,6 @@ unp_shutdown(struct unpcb *unp)
struct unpcb *unp2;
struct socket *so;
- UNP_LINK_WLOCK_ASSERT();
UNP_PCB_LOCK_ASSERT(unp);
unp2 = unp->unp_conn;
@@ -1726,22 +1895,28 @@ unp_drop(struct unpcb *unp)
{
struct socket *so = unp->unp_socket;
struct unpcb *unp2;
+ int freed;
- UNP_LINK_WLOCK_ASSERT();
- UNP_PCB_LOCK_ASSERT(unp);
-
/*
* Regardless of whether the socket's peer dropped the connection
* with this socket by aborting or disconnecting, POSIX requires
* that ECONNRESET is returned.
*/
- so->so_error = ECONNRESET;
+ /* acquire a reference so that unp isn't freed from underneath us */
+
+ UNP_PCB_LOCK(unp);
+ if (so)
+ so->so_error = ECONNRESET;
unp2 = unp->unp_conn;
- if (unp2 == NULL)
- return;
- UNP_PCB_LOCK(unp2);
- unp_disconnect(unp, unp2);
- UNP_PCB_UNLOCK(unp2);
+ if (unp2 != NULL) {
+ unp_pcb_hold(unp2);
+ unp_pcb_owned_lock2(unp, unp2, freed);
+ unp_disconnect(unp, unp2);
+ if (unp_pcb_rele(unp2) == 0)
+ UNP_PCB_UNLOCK(unp2);
+ }
+ if (unp_pcb_rele(unp) == 0)
+ UNP_PCB_UNLOCK(unp);
}
static void
@@ -1881,7 +2056,7 @@ unp_init(void)
return;
#endif
unp_zone = uma_zcreate("unpcb", sizeof(struct unpcb), NULL, NULL,
- NULL, NULL, UMA_ALIGN_PTR, 0);
+ NULL, NULL, UMA_ALIGN_CACHE, 0);
if (unp_zone == NULL)
panic("unp_init");
uma_zone_set_max(unp_zone, maxsockets);
@@ -2464,13 +2639,15 @@ vfs_unp_reclaim(struct vnode *vp)
{
struct unpcb *unp;
int active;
+ struct mtx *vplock;
ASSERT_VOP_ELOCKED(vp, "vfs_unp_reclaim");
KASSERT(vp->v_type == VSOCK,
("vfs_unp_reclaim: vp->v_type != VSOCK"));
active = 0;
- UNP_LINK_WLOCK();
+ vplock = mtx_pool_find(mtxpool_sleep, vp);
+ mtx_lock(vplock);
VOP_UNP_CONNECT(vp, &unp);
if (unp == NULL)
goto done;
@@ -2481,8 +2658,8 @@ vfs_unp_reclaim(struct vnode *vp)
active = 1;
}
UNP_PCB_UNLOCK(unp);
-done:
- UNP_LINK_WUNLOCK();
+ done:
+ mtx_unlock(vplock);
if (active)
vunref(vp);
}
Modified: head/sys/sys/unpcb.h
==============================================================================
--- head/sys/sys/unpcb.h Thu May 17 17:57:41 2018 (r333743)
+++ head/sys/sys/unpcb.h Thu May 17 17:59:35 2018 (r333744)
@@ -69,23 +69,25 @@ typedef uint64_t unp_gen_t;
LIST_HEAD(unp_head, unpcb);
struct unpcb {
- LIST_ENTRY(unpcb) unp_link; /* glue on list of all PCBs */
- struct socket *unp_socket; /* pointer back to socket */
- struct file *unp_file; /* back-pointer to file for gc. */
- struct vnode *unp_vnode; /* if associated with file */
- ino_t unp_ino; /* fake inode number */
+ /* Cache line 1 */
+ struct mtx unp_mtx; /* mutex */
struct unpcb *unp_conn; /* control block of connected socket */
- struct unp_head unp_refs; /* referencing socket linked list */
- LIST_ENTRY(unpcb) unp_reflink; /* link in unp_refs list */
- struct sockaddr_un *unp_addr; /* bound address of socket */
- unp_gen_t unp_gencnt; /* generation count of this instance */
+ volatile u_int unp_refcount;
short unp_flags; /* flags */
short unp_gcflag; /* Garbage collector flags. */
+ struct sockaddr_un *unp_addr; /* bound address of socket */
+ struct socket *unp_socket; /* pointer back to socket */
+ /* Cache line 2 */
+ struct vnode *unp_vnode; /* if associated with file */
struct xucred unp_peercred; /* peer credentials, if applicable */
- u_int unp_refcount;
+ LIST_ENTRY(unpcb) unp_reflink; /* link in unp_refs list */
+ LIST_ENTRY(unpcb) unp_link; /* glue on list of all PCBs */
+ struct unp_head unp_refs; /* referencing socket linked list */
+ unp_gen_t unp_gencnt; /* generation count of this instance */
+ struct file *unp_file; /* back-pointer to file for gc. */
u_int unp_msgcount; /* references from message queue */
- struct mtx unp_mtx; /* mutex */
-};
+ ino_t unp_ino; /* fake inode number */
+} __aligned(CACHE_LINE_SIZE);
*** DIFF OUTPUT TRUNCATED AT 1000 LINES ***
More information about the svn-src-all
mailing list