slirp: switch to GPollFD

Slirp uses rfds/wfds/xfds more extensively than other QEMU components.

The rarely-used out-of-band TCP data feature is used.  That means we
need the full table of select(2) to g_poll(3) events:

  rfds -> G_IO_IN | G_IO_HUP | G_IO_ERR
  wfds -> G_IO_OUT | G_IO_ERR
  xfds -> G_IO_PRI

I came up with this table by looking at Linux fs/select.c which maps
select(2) to poll(2) internally.

Another detail to watch out for are the global variables that reference
rfds/wfds/xfds during slirp_select_poll().  sofcantrcvmore() and
sofcantsendmore() use these globals to clear fd_set bits.  When
sofcantrcvmore() is called, the wfds bit is cleared so that the write
handler will no longer be run for this iteration of the event loop.

This actually seems buggy to me since TCP connections can be half-closed
and we'd still want to handle data in half-duplex fashion.  I think the
real intention is to avoid running the read/write handler when the
socket has been fully closed.  This is indicated with the SS_NOFDREF
state bit so we now check for it before invoking the TCP write handler.
Note that UDP/ICMP code paths don't care because they are
connectionless.

Note that slirp/ has a lot of tabs and sometimes mixed tabs with spaces.
I followed the style of the surrounding code.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Message-id: 1361356113-11049-6-git-send-email-stefanha@redhat.com
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

5 files changed, 82 insertions, 69 deletions

diff --git a/libslirp.h b/libslirp.h
index 40a5568..925a207 100644
--- a/libslirp.h
+++ b/libslirp.h
@@ -17,11 +17,9 @@ Slirp *slirp_init(int restricted, struct in_addr vnetwork,
 void slirp_cleanup(Slirp *slirp);
 
 void slirp_update_timeout(uint32_t *timeout);
-void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
-                       fd_set *xfds);
+void slirp_pollfds_fill(GArray *pollfds);
 
-void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds,
-                       int select_error);
+void slirp_pollfds_poll(GArray *pollfds, int select_error);
 
 void slirp_input(Slirp *slirp, const uint8_t *pkt, int pkt_len);
 
diff --git a/main.h b/main.h
index b1f94c6..939b4c6 100644
--- a/main.h
+++ b/main.h
@@ -31,7 +31,6 @@ extern int ctty_closed;
 extern char *slirp_tty;
 extern char *exec_shell;
 extern u_int curtime;
-extern fd_set *global_readfds, *global_writefds, *global_xfds;
 extern struct in_addr loopback_addr;
 extern unsigned long loopback_mask;
 extern char *username;
diff --git a/slirp.c b/slirp.c
index e15b98a..e82b7c9 100644
--- a/slirp.c
+++ b/slirp.c
@@ -38,9 +38,6 @@ static const uint8_t special_ethaddr[ETH_ALEN] = { 0x52, 0x55, 0x00,
 
 static const uint8_t zero_ethaddr[ETH_ALEN] = { 0, 0, 0, 0, 0, 0 };
 
-/* XXX: suppress those select globals */
-fd_set *global_readfds, *global_writefds, *global_xfds;
-
 u_int curtime;
 static u_int time_fasttimo, last_slowtimo;
 static int do_slowtimo;
@@ -262,9 +259,6 @@ void slirp_cleanup(Slirp *slirp)
     (((so)->so_state & (SS_FCANTSENDMORE | SS_ISFCONNECTED)) == SS_ISFCONNECTED)
 #define CONN_CANFRCV(so) \
     (((so)->so_state & (SS_FCANTRCVMORE | SS_ISFCONNECTED)) == SS_ISFCONNECTED)
-#define UPD_NFDS(x) \
-    if (nfds < (x)) \
-    nfds = (x)
 
 void slirp_update_timeout(uint32_t *timeout)
 {
@@ -273,23 +267,15 @@ void slirp_update_timeout(uint32_t *timeout)
     }
 }
 
-void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
-                       fd_set *xfds)
+void slirp_pollfds_fill(GArray *pollfds)
 {
     Slirp *slirp;
     struct socket *so, *so_next;
-    int nfds;
 
     if (QTAILQ_EMPTY(&slirp_instances)) {
         return;
     }
 
-    /* fail safe */
-    global_readfds = NULL;
-    global_writefds = NULL;
-    global_xfds = NULL;
-
-    nfds = *pnfds;
     /*
      * First, TCP sockets
      */
@@ -304,8 +290,12 @@ void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
                         (&slirp->ipq.ip_link != slirp->ipq.ip_link.next));
 
         for (so = slirp->tcb.so_next; so != &slirp->tcb; so = so_next) {
+            int events = 0;
+
             so_next = so->so_next;
 
+            so->pollfds_idx = -1;
+
             /*
              * See if we need a tcp_fasttimo
              */
@@ -325,8 +315,12 @@ void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
              * Set for reading sockets which are accepting
              */
             if (so->so_state & SS_FACCEPTCONN) {
-                FD_SET(so->s, readfds);
-                UPD_NFDS(so->s);
+                GPollFD pfd = {
+                    .fd = so->s,
+                    .events = G_IO_IN | G_IO_HUP | G_IO_ERR,
+                };
+                so->pollfds_idx = pollfds->len;
+                g_array_append_val(pollfds, pfd);
                 continue;
             }
 
@@ -334,8 +328,12 @@ void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
              * Set for writing sockets which are connecting
              */
             if (so->so_state & SS_ISFCONNECTING) {
-                FD_SET(so->s, writefds);
-                UPD_NFDS(so->s);
+                GPollFD pfd = {
+                    .fd = so->s,
+                    .events = G_IO_OUT | G_IO_ERR,
+                };
+                so->pollfds_idx = pollfds->len;
+                g_array_append_val(pollfds, pfd);
                 continue;
             }
 
@@ -344,8 +342,7 @@ void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
              * we have something to send
              */
             if (CONN_CANFSEND(so) && so->so_rcv.sb_cc) {
-                FD_SET(so->s, writefds);
-                UPD_NFDS(so->s);
+                events |= G_IO_OUT | G_IO_ERR;
             }
 
             /*
@@ -354,9 +351,16 @@ void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
              */
             if (CONN_CANFRCV(so) &&
                 (so->so_snd.sb_cc < (so->so_snd.sb_datalen / 2))) {
-                FD_SET(so->s, readfds);
-                FD_SET(so->s, xfds);
-                UPD_NFDS(so->s);
+                events |= G_IO_IN | G_IO_HUP | G_IO_ERR | G_IO_PRI;
+            }
+
+            if (events) {
+                GPollFD pfd = {
+                    .fd = so->s,
+                    .events = events,
+                };
+                so->pollfds_idx = pollfds->len;
+                g_array_append_val(pollfds, pfd);
             }
         }
 
@@ -366,6 +370,8 @@ void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
         for (so = slirp->udb.so_next; so != &slirp->udb; so = so_next) {
             so_next = so->so_next;
 
+            so->pollfds_idx = -1;
+
             /*
              * See if it's timed out
              */
@@ -389,8 +395,12 @@ void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
              * (XXX <= 4 ?)
              */
             if ((so->so_state & SS_ISFCONNECTED) && so->so_queued <= 4) {
-                FD_SET(so->s, readfds);
-                UPD_NFDS(so->s);
+                GPollFD pfd = {
+                    .fd = so->s,
+                    .events = G_IO_IN | G_IO_HUP | G_IO_ERR,
+                };
+                so->pollfds_idx = pollfds->len;
+                g_array_append_val(pollfds, pfd);
             }
         }
 
@@ -400,6 +410,8 @@ void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
         for (so = slirp->icmp.so_next; so != &slirp->icmp; so = so_next) {
             so_next = so->so_next;
 
+            so->pollfds_idx = -1;
+
             /*
              * See if it's timed out
              */
@@ -413,17 +425,18 @@ void slirp_select_fill(int *pnfds, fd_set *readfds, fd_set *writefds,
             }
 
             if (so->so_state & SS_ISFCONNECTED) {
-                FD_SET(so->s, readfds);
-                UPD_NFDS(so->s);
+                GPollFD pfd = {
+                    .fd = so->s,
+                    .events = G_IO_IN | G_IO_HUP | G_IO_ERR,
+                };
+                so->pollfds_idx = pollfds->len;
+                g_array_append_val(pollfds, pfd);
             }
         }
     }
-
-    *pnfds = nfds;
 }
 
-void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds,
-                       int select_error)
+void slirp_pollfds_poll(GArray *pollfds, int select_error)
 {
     Slirp *slirp;
     struct socket *so, *so_next;
@@ -433,10 +446,6 @@ void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds,
         return;
     }
 
-    global_readfds = readfds;
-    global_writefds = writefds;
-    global_xfds = xfds;
-
     curtime = qemu_get_clock_ms(rt_clock);
 
     QTAILQ_FOREACH (slirp, &slirp_instances, entry) {
@@ -461,12 +470,16 @@ void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds,
              * Check TCP sockets
              */
             for (so = slirp->tcb.so_next; so != &slirp->tcb; so = so_next) {
+                int revents;
+
                 so_next = so->so_next;
 
-                /*
-                 * FD_ISSET is meaningless on these sockets
-                 * (and they can crash the program)
-                 */
+                revents = 0;
+                if (so->pollfds_idx != -1) {
+                    revents = g_array_index(pollfds, GPollFD, so->pollfds_idx)
+                                  .revents;
+                }
+
                 if (so->so_state & SS_NOFDREF || so->s == -1) {
                     continue;
                 }
@@ -474,15 +487,15 @@ void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds,
                 /*
                  * Check for URG data
                  * This will soread as well, so no need to
-                 * test for readfds below if this succeeds
+                 * test for G_IO_IN below if this succeeds
                  */
-                if (FD_ISSET(so->s, xfds)) {
+                if (revents & G_IO_PRI) {
                     sorecvoob(so);
                 }
                 /*
                  * Check sockets for reading
                  */
-                else if (FD_ISSET(so->s, readfds)) {
+                else if (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR)) {
                     /*
                      * Check for incoming connections
                      */
@@ -501,7 +514,8 @@ void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds,
                 /*
                  * Check sockets for writing
                  */
-                if (FD_ISSET(so->s, writefds)) {
+                if (!(so->so_state & SS_NOFDREF) &&
+                    (revents & (G_IO_OUT | G_IO_ERR))) {
                     /*
                      * Check for non-blocking, still-connecting sockets
                      */
@@ -585,9 +599,18 @@ void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds,
              * Incoming UDP data isn't buffered either.
              */
             for (so = slirp->udb.so_next; so != &slirp->udb; so = so_next) {
+                int revents;
+
                 so_next = so->so_next;
 
-                if (so->s != -1 && FD_ISSET(so->s, readfds)) {
+                revents = 0;
+                if (so->pollfds_idx != -1) {
+                    revents = g_array_index(pollfds, GPollFD, so->pollfds_idx)
+                                  .revents;
+                }
+
+                if (so->s != -1 &&
+                    (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR))) {
                     sorecvfrom(so);
                 }
             }
@@ -596,9 +619,18 @@ void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds,
              * Check incoming ICMP relies.
              */
             for (so = slirp->icmp.so_next; so != &slirp->icmp; so = so_next) {
+                int revents;
+
                 so_next = so->so_next;
 
-                if (so->s != -1 && FD_ISSET(so->s, readfds)) {
+                revents = 0;
+                if (so->pollfds_idx != -1) {
+                    revents = g_array_index(pollfds, GPollFD, so->pollfds_idx)
+                                  .revents;
+                }
+
+                if (so->s != -1 &&
+                    (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR))) {
                     icmp_receive(so);
                 }
             }
@@ -606,15 +638,6 @@ void slirp_select_poll(fd_set *readfds, fd_set *writefds, fd_set *xfds,
 
         if_start(slirp);
     }
-
-    /* clear global file descriptor sets.
-     * these reside on the stack in vl.c
-     * so they're unusable if we're not in
-     * slirp_select_fill or slirp_select_poll.
-     */
-    global_readfds = NULL;
-    global_writefds = NULL;
-    global_xfds = NULL;
 }
 
 static void arp_input(Slirp *slirp, const uint8_t *pkt, int pkt_len)
diff --git a/socket.c b/socket.c
index bc2d112..9f7c1a6 100644
--- a/socket.c
+++ b/socket.c
@@ -685,9 +685,6 @@ static void sofcantrcvmore(struct socket *so)
 {
     if ((so->so_state & SS_NOFDREF) == 0) {
         shutdown(so->s, 0);
-        if (global_writefds) {
-            FD_CLR(so->s, global_writefds);
-        }
     }
     so->so_state &= ~(SS_ISFCONNECTING);
     if (so->so_state & SS_FCANTSENDMORE) {
@@ -702,12 +699,6 @@ static void sofcantsendmore(struct socket *so)
 {
     if ((so->so_state & SS_NOFDREF) == 0) {
         shutdown(so->s, 1); /* send FIN to fhost */
-        if (global_readfds) {
-            FD_CLR(so->s, global_readfds);
-        }
-        if (global_xfds) {
-            FD_CLR(so->s, global_xfds);
-        }
     }
     so->so_state &= ~(SS_ISFCONNECTING);
     if (so->so_state & SS_FCANTRCVMORE) {
diff --git a/socket.h b/socket.h
index 7f67581..a40aa66 100644
--- a/socket.h
+++ b/socket.h
@@ -20,6 +20,8 @@ struct socket {
 
     int s; /* The actual socket */
 
+    int pollfds_idx; /* GPollFD GArray index */
+
     Slirp *slirp; /* managing slirp instance */
 
     /* XXX union these with not-yet-used sbuf params */