USB: usblp: "Big cleanup" breaks O_NONBLOCK
authorPete Zaitcev <zaitcev@redhat.com>
Wed, 11 Jul 2007 03:09:58 +0000 (20:09 -0700)
committerGreg Kroah-Hartman <gregkh@suse.de>
Fri, 20 Jul 2007 00:46:03 +0000 (17:46 -0700)
I found the first regresson in the rewritten ("all dynamic" and "no races")
driver. If application uses O_NONBLOCK, I return -EAGAIN despite the URB
being submitted successfuly. This causes the application to resubmit the
same data erroneously.

The fix is to pretend that the transfer has succeeded even if URB was
merely queued. It is the same behaviour as with the old version.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
drivers/usb/class/usblp.c

index 9a1478972bf5aee1cf75c5f70434d4e68c5fa83c..80ec1039d670c928159652e9c3f18fd6210ea674 100644 (file)
@@ -741,10 +741,11 @@ static ssize_t usblp_write(struct file *file, const char __user *buffer, size_t
                 */
                rv = usblp_wwait(usblp, !!(file->f_flags&O_NONBLOCK));
                if (rv < 0) {
-                       /*
-                        * If interrupted, we simply leave the URB to dangle,
-                        * so the ->release will call usb_kill_urb().
-                        */
+                       if (rv == -EAGAIN) {
+                               /* Presume that it's going to complete well. */
+                               writecount += transfer_length;
+                       }
+                       /* Leave URB dangling, to be cleaned on close. */
                        goto collect_error;
                }