io: deal with blocking/non-blocking fds

[PATCH 02/10] char-socket: rework tcp_chr_recv()

Vladimir Sementsov-Ogievskiy posted 10 patches 3 days, 4 hours ago

Download mbox

Maintainers: "Marc-André Lureau" <marcandre.lureau@redhat.com>, Paolo Bonzini <pbonzini@redhat.com>, "Michael S. Tsirkin" <mst@redhat.com>, Gerd Hoffmann <kraxel@redhat.com>, Gustavo Romero <gustavo.romero@linaro.org>, Elena Ufimtseva <elena.ufimtseva@oracle.com>, Jagannathan Raman <jag.raman@oracle.com>, John Levon <john.levon@nutanix.com>, Thanos Makatos <thanos.makatos@nutanix.com>, "Cédric Le Goater" <clg@redhat.com>, Stefano Garzarella <sgarzare@redhat.com>, "Daniel P. Berrangé" <berrange@redhat.com>, Peter Xu <peterx@redhat.com>, Fabiano Rosas <farosas@suse.de>, Jason Wang <jasowang@redhat.com>, Michael Roth <michael.roth@amd.com>, Kostiantyn Kostiuk <kkostiuk@redhat.com>, Fam Zheng <fam@euphon.net>, Alexander Bulekov <alxndr@bu.edu>, Bandan Das <bsd@redhat.com>, Stefan Hajnoczi <stefanha@redhat.com>, Darren Kenny <darren.kenny@oracle.com>, Qiuhao Li <Qiuhao.Li@outlook.com>, Laurent Vivier <lvivier@redhat.com>, Stefan Berger <stefanb@linux.vnet.ibm.com>, Stefan Weil <sw@weilnetz.de>, Coiby Xu <Coiby.Xu@gmail.com>

Expand all Fold all

[PATCH 02/10] char-socket: rework tcp_chr_recv()

Posted by Vladimir Sementsov-Ogievskiy 3 days, 4 hours ago

First, qio_channel_readv_full() already guarantees BLOCKING and
CLOEXEC states for incoming descriptors, no reason call extra
ioctls.

Second, current implementation calls _set_block() and _set_cloexec()
again on old descriptors on failure path - we fix this too.

Finally, handling errors exactly after qio_channel_readv_full() call
looks more readable.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
---
 chardev/char-socket.c | 37 +++++++++++++------------------------
 1 file changed, 13 insertions(+), 24 deletions(-)

diff --git a/chardev/char-socket.c b/chardev/char-socket.c
index 1e8313915b..5b9b19ba8b 100644
--- a/chardev/char-socket.c
+++ b/chardev/char-socket.c
@@ -293,6 +293,18 @@ static ssize_t tcp_chr_recv(Chardev *chr, char *buf, size_t len)
                                      0, &err);
     }
 
+    if (ret == QIO_CHANNEL_ERR_BLOCK) {
+        errno = EAGAIN;
+        return -1;
+    } else if (ret == -1) {
+        trace_chr_socket_recv_err(chr, chr->label, error_get_pretty(err));
+        error_free(err);
+        errno = EIO;
+        return -1;
+    }
+
+    assert(ret >= 0);
+
     if (msgfds_num) {
         /* close and clean read_msgfds */
         for (i = 0; i < s->read_msgfds_num; i++) {
@@ -307,30 +319,7 @@ static ssize_t tcp_chr_recv(Chardev *chr, char *buf, size_t len)
         s->read_msgfds_num = msgfds_num;
     }
 
-    for (i = 0; i < s->read_msgfds_num; i++) {
-        int fd = s->read_msgfds[i];
-        if (fd < 0) {
-            continue;
-        }
-
-        /* O_NONBLOCK is preserved across SCM_RIGHTS so reset it */
-        qemu_socket_set_block(fd);
-
-#ifndef MSG_CMSG_CLOEXEC
-        qemu_set_cloexec(fd);
-#endif
-    }
-
-    if (ret == QIO_CHANNEL_ERR_BLOCK) {
-        errno = EAGAIN;
-        ret = -1;
-    } else if (ret == -1) {
-        trace_chr_socket_recv_err(chr, chr->label, error_get_pretty(err));
-        error_free(err);
-        errno = EIO;
-    } else if (ret == 0) {
-        trace_chr_socket_recv_eof(chr, chr->label);
-    }
+    trace_chr_socket_recv_eof(chr, chr->label);
 
     return ret;
 }
-- 
2.48.1