Libguestfs - May 2019 - [PATCH nbdkit 2/2] server: Disable Nagle's algorithm.

GCC is concerned that if we never go round the loop then fd will be
uninitialized.  By asserting that getaddrinfo set result != NULL we
can avoid this.

nbd.c: In function ‘nbd_open_handle’:
nbd.c:974:5: error: ‘fd’ may be used uninitialized in this function
[-Werror=maybe-uninitialized]
  974 |     close (fd);
      |     ^~~~~~~~~~
nbd.c:954:7: note: ‘fd’ was declared here
  954 |   int fd;
      |       ^~
cc1: all warnings being treated as errors
---
 plugins/nbd/nbd.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/plugins/nbd/nbd.c b/plugins/nbd/nbd.c
index dba46f1..4d0a67c 100644
--- a/plugins/nbd/nbd.c
+++ b/plugins/nbd/nbd.c
@@ -960,6 +960,8 @@ nbd_connect_tcp (void)
     return -1;
   }
 
+  assert (result != NULL);
+
   for (rp = result; rp; rp = rp->ai_next) {
     fd = socket (rp->ai_family, rp->ai_socktype, rp->ai_protocol);
     if (fd == -1)
-- 
2.21.0

Unlike the equivalent change on the client side which caused a
dramatic performance improvement, there is no noticable difference
from this patch in my testing.
---
 server/sockets.c | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/server/sockets.c b/server/sockets.c
index 2c71970..b25405c 100644
--- a/server/sockets.c
+++ b/server/sockets.c
@@ -37,13 +37,15 @@
 #include <stdbool.h>
 #include <string.h>
 #include <unistd.h>
+#include <poll.h>
+#include <errno.h>
+#include <assert.h>
 #include <sys/types.h>
 #include <sys/socket.h>
 #include <sys/un.h>
+#include <netinet/in.h>
+#include <netinet/tcp.h>
 #include <netdb.h>
-#include <poll.h>
-#include <errno.h>
-#include <assert.h>
 
 #ifdef HAVE_LIBSELINUX
 #include <selinux/selinux.h>
@@ -273,6 +275,7 @@ accept_connection (int listen_sock)
   pthread_t thread;
   struct thread_data *thread_data;
   static size_t instance_num = 1;
+  const int flag = 1;
 
   thread_data = malloc (sizeof *thread_data);
   if (!thread_data) {
@@ -293,6 +296,11 @@ accept_connection (int listen_sock)
     return;
   }
 
+  /* Disable Nagle's algorithm on this socket.  However we don't want
+   * to fail if this doesn't work.
+   */
+  setsockopt (thread_data->sock, IPPROTO_TCP, TCP_NODELAY, &flag, sizeof
flag);
+
   /* Start a thread to handle this connection.  Note we always do this
    * even for non-threaded plugins.  There are mutexes in plugins.c
    * which ensure that non-threaded plugins are handled correctly.
-- 
2.21.0

On 5/30/19 1:32 PM, Richard W.M. Jones wrote:
> Unlike the equivalent change on the client side which caused a
> dramatic performance improvement, there is no noticable difference
noticeable ?
> from this patch in my testing.
Still worth doing, though. ACK series.

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org