Sanitizing netif-rename code

PATCH In review Matthias Schwarzott 3 participants 3 messages Jan 23, 2007 – Jan 24, 2007

lore

Share & tools

lore

Matthias Schwarzott

author

2007-01-23 16:12 UTC #0 lore raw

PATCH 1/1 Sanitizing netif-rename code

Hi there!

The attached patch comes origins from this Bugreport:
http://bugs.gentoo.org/show_bug.cgi?id=158531

The original code from udev_device.c / function rename_netif looks like this:

while (loop--) {
	retval = ioctl(sk, SIOCSIFNAME, &ifr);
	if (retval != 0) {
		if (errno != EEXIST) {
			err("error changing net interface name %s to %s: %s",
				ifr.ifr_name, ifr.ifr_newname, strerror(errno));
			break;
		}
		dbg("wait for netif '%s' to become free, loop=%i", udev->name, (30 * 20) -
			loop);
		usleep(1000 * 1000 / 20);
	}
}

In some loop the rename-syscall will succeed, and retval is 0.
In this case the loop is not left, but just retried the loop and executing the rename code over and over again.

Attached patch adds something similar to that (inside the loop):
if (retval == 0)

	break;

Greetings
Matthias

Signature · 1 line

Matthias Schwarzott (zzam)

udev_device.c +6 -4

diff -Nur udev-103old/udev_device.c udev-103/udev_device.c
@@ -122,16 +122,18 @@
 		strlcpy(ifr.ifr_newname, udev->name, IFNAMSIZ);
 		loop = 30 * 20;
 		while (loop--) {
-			retval = ioctl(sk, SIOCSIFNAME, &ifr);	
-			if (retval != 0) {
+			retval = ioctl(sk, SIOCSIFNAME, &ifr);
+			if (retval == 0) {
+				break;
+			} else {
 				if (errno != EEXIST) {
 					err("error changing net interface name %s to %s: %s",
 					    ifr.ifr_name, ifr.ifr_newname, strerror(errno));
 					break;
 				}
-				dbg("wait for netif '%s' to become free, loop=%i", udev->name, (30 * 20) - loop);
-				usleep(1000 * 1000 / 20);
 			}
+			dbg("wait for netif '%s' to become free, loop=%i", udev->name, (30 * 20) - loop);
+			usleep(1000 * 1000 / 20);
 		}
 	}

_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Kay Sievers

2007-01-24 09:51 UTC #1 lore raw

Quoted 3 lines · Matthias Schwarzott

In some loop the rename-syscall will succeed, and retval is 0.
In this case the loop is not left, but just retried the loop and executing the rename code over and over again.

Thanks! I've applied it without the "else" after the "break".

Kay

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

Alexander E. Patrakov

2007-01-24 11:58 UTC #2 lore raw

In reply to Matthias Schwarzott

Quoted 6 lines · Kay Sievers

Quoted 3 lines · Matthias Schwarzott
In some loop the rename-syscall will succeed, and retval is 0.
In this case the loop is not left, but just retried the loop and executing the rename code over and over again.
Thanks! I've applied it without the "else" after the "break".

This also fixes the "Udev sometimes forgets to RUN a program when renaming
network interface" bug I reported earlier. Thanks!

Signature · 11 lines

Alexander E. Patrakov

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CIDÞVDEV
_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel