Skip to contentSkip to conversation

Linux 2.6.16.7

PATCH In review Greg KH 9 participants 19 messages Apr 18, 2006 – May 8, 2006

Subsystems (5)
lore
Share & tools
lore

Greg KH

author
#0 lore raw

We (the -stable team) are announcing the release of the 2.6.16.7 kernel.

The diffstat and short summary of the fixes are below.

I'll also be replying to this message with a copy of the patch between 2.6.16.6 and 2.6.16.7, as it is small enough to do so.

The updated 2.6.16.y git tree can be found at:
rsync://rsync.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.16.y.git
and can be browsed at the normal kernel.org git web browser:
www.kernel.org/git/

Note, there seems to be a bit of a discrepancy in the assigning of CVE numbers to a specific issue that was solved in this release. The original author and I both asked for CVE numbers for different patches and seem to have gotten the same one. Please note that these are two different issues, yet, the same CVE number. Hopefully it gets fixed eventually for those who like to track these kinds of things...

thanks,

greg k-h

--------

 Makefile     |    2 +-
 mm/madvise.c |    3 +++
 2 files changed, 4 insertions(+), 1 deletion(-)

Summary of changes from v2.6.16.6 to v2.6.16.7
==============================================

Greg Kroah-Hartman:

      Linux 2.6.16.7

Hugh Dickins:
fix MADV_REMOVE vulnerability (CVE-2006-1524 for real this time)

author
#1 lore raw
Makefile +1 -1
index 7959505..06a8926 100644
@@ -1,7 +1,7 @@
 VERSION = 2
 PATCHLEVEL = 6
 SUBLEVEL = 16
-EXTRAVERSION = .6
+EXTRAVERSION = .7
 NAME=Sliding Snow Leopard
 
 # *DOCUMENTATION*
mm/madvise.c +3 -0
index af3d573..4e19615 100644
@@ -168,6 +168,9 @@ static long madvise_remove(struct vm_are
 			return -EINVAL;
 	}
 
+	if ((vma->vm_flags & (VM_SHARED|VM_WRITE)) != (VM_SHARED|VM_WRITE))
+		return -EACCES;
+
 	mapping = vma->vm_file->f_mapping;
 
 	offset = (loff_t)(start - vma->vm_start)
#2 lore raw

Makefile line 4

Quoted 2 lines · Greg KH
Makefile +1 -1
-EXTRAVERSION = .6
+EXTRAVERSION = .7

Hello, I would like to know if there is a plan to include this in the next -stable update?

http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blobdiff;h=2731570eba5b35a21c311dd587057c39805082f1;hp=dfb62998866ae2e298139164a85ec0757b7f3fc7;hb=9469d458b90bfb9117cbb488cfa645d94c3921b1;f=net/core/dev.c

Signature · 1 line
Alexander E. Patrakov
#3 lore raw
Quoted 8 lines · Alexander E. Patrakov

Makefile line 4

Quoted 2 lines · Greg KH
Makefile +1 -1
-EXTRAVERSION = .6
+EXTRAVERSION = .7

Hello, I would like to know if there is a plan to include this in the next -stable update?

http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blobdiff;h=2731570eba5b35a21c311dd587057c39805082f1;hp=dfb62998866ae2e298139164a85ec0757b7f3fc7;hb=9469d458b90bfb9117cbb488cfa645d94c3921b1;f=net/core/dev.c

No one has submitted it to the [email protected] mail address from what I can see, so no, it is not in the queue. If you think otherwise, please send it.

thanks,

greg k-h

#4 lore raw
Quoted 7 lines

No one has submitted it to the [email protected] mail address from what I can see, so no, it is not in the queue. If you think otherwise, please send it.

thanks,

greg k-h

I thought the 2.6.xx.n series was the 'stable' (as per kernel.org)?

Nick

#5 lore raw
Quoted 9 lines · Nick Warne
Quoted 7 lines

No one has submitted it to the [email protected] mail address from what I can see, so no, it is not in the queue. If you think otherwise, please send it.

thanks,

greg k-h

I thought the 2.6.xx.n series was the 'stable' (as per kernel.org)?

Yes it is. Did I say otherwise in those sentences?

thanks,

greg k-h

#6 lore raw
Quoted 12 lines · Greg KH
Quoted 9 lines · Nick Warne
Quoted 7 lines

No one has submitted it to the [email protected] mail address from what I can see, so no, it is not in the queue. If you think otherwise, please send it.

thanks,

greg k-h

I thought the 2.6.xx.n series was the 'stable' (as per kernel.org)?

Yes it is. Did I say otherwise in those sentences?

Quoted 6 lines · I see... Alexander
Quoted 2 lines · Greg KH

-EXTRAVERSION = .6
+EXTRAVERSION = .7

Hello, I would like to know if there is a plan to include this in the next -stable update?

I misread the ref. to the different link/patch referred too...

Sorry for the noise.

Nick

Signature · 2 lines
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
#7 lore raw
In reply to Greg KH
Quoted 10 lines · Greg KH
Quoted 8 lines · Alexander E. Patrakov

Review

Quoted 3 lines · Greg KH
(quoted diff) +1 -1
-EXTRAVERSION = .6
+EXTRAVERSION = .7

Hello, I would like to know if there is a plan to include this in the next -stable update?

Quoted 7 lines
Quoted 2 lines

http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blobdiff;h=2731570eba5b35a21c311dd587057c39805082f1;hp=dfb62998866ae2e298139164a85ec0757b7f3fc7;hb=9469d458b90bfb9117cbb488cfa645d94c3921b1;f=net/core/dev.c

No one has submitted it to the [email protected] mail address from what I can see, so no, it is not in the queue. If you think otherwise, please send it.

Without that patch, there is a race when registering network interfaces and renaming it with udev rules, because initially the "address" in sysfs doesn't contain useful data. See http://marc.theaimsgroup.com/?t=114460338900002&r=1&w=2

Breaking the recommended way of assigning persistent network interface names is, IMHO, a bug serious enough to be fixed in -stable.

---

net/core/dev.c +0 -0
@@ -2932,11 +2932,11 @@
 
		switch(dev->reg_state) {
		case NETREG_REGISTERING:
+			dev->reg_state = NETREG_REGISTERED;
			err = netdev_register_sysfs(dev);
			if (err)
				printk(KERN_ERR "%s: failed sysfs registration (%d)\n",
				       dev->name, err);
-			dev->reg_state = NETREG_REGISTERED;
			break;

		case NETREG_UNREGISTERING:
Signature · 1 line
Alexander E. Patrakov
maintainer
#8 lore raw

From: "Alexander E. Patrakov" <[email protected]>
Date: Wed, 19 Apr 2006 10:22:39 +0600

Quoted 9 lines

Without that patch, there is a race when registering network interfaces and renaming it with udev rules, because initially the "address" in sysfs doesn't contain useful data. See
http://marc.theaimsgroup.com/?t=114460338900002&r=1&w=2

Breaking the recommended way of assigning persistent network interface names is, IMHO, a bug serious enough to be fixed in -stable.

I've been waiting to see if there is any negative fallout from this change, but I guess it's OK for stable.

Applied
#9 lore raw
Quoted 8 lines · David S. Miller

From: "Alexander E. Patrakov" <[email protected]>
Date: Wed, 19 Apr 2006 10:22:39 +0600

Quoted 5 lines

Without that patch, there is a race when registering network interfaces and renaming it with udev rules, because initially the "address" in sysfs doesn't contain useful data. See
http://marc.theaimsgroup.com/?t=114460338900002&r=1&w=2

Quoted 7 lines
Quoted 4 lines

Breaking the recommended way of assigning persistent network interface names is, IMHO, a bug serious enough to be fixed in -stable.

I've been waiting to see if there is any negative fallout from this change, but I guess it's OK for stable.

Thanks for the ack.

Queued to -stable, thanks.

greg k-h

#10 lore raw

Hotplug race on name change

In reply to Alexander E. Patrakov

This:

Quoted 17 lines

Without that patch, there is a race when registering network interfaces and renaming it with udev rules, because initially the "address" in sysfs doesn't contain useful data. See
http://marc.theaimsgroup.com/?t=114460338900002&r=1&w=2

Breaking the recommended way of assigning persistent network interface names is, IMHO, a bug serious enough to be fixed in -stable.

---

net/core/dev.c +0 -0
@@ -2932,11 +2932,11 @@
 
 		switch(dev->reg_state) {

net/core/dev.c line 2942

Quoted 10 lines
net/core/dev.c +1 -1
 		case NETREG_REGISTERING:
+			dev->reg_state = NETREG_REGISTERED;
 			err = netdev_register_sysfs(dev);
 			if (err)
 				printk(KERN_ERR "%s: failed sysfs registration (%d)\n",
 				       dev->name, err);
-			dev->reg_state = NETREG_REGISTERED;
 			break;
 
 		case NETREG_UNREGISTERING:

Introduces new races in netdev_register_sysfs if the name changes, because netdev_register_sysfs runs without RTNL at this point. So if some application gets in and changes the device name while netdev_register_sysfs is running, then the class_dev->class_id would end up not matching the netdevice->name.

Not a big issue since, hotplug doesn't get run until the device is registered. Ideally, it would be possible to create the groups in the class device before it was registered. It won't work with existing class device interface.

I am working on a patch to extend class_device to allow the creation of groups to be atomic (like the attributes are).

#11 lore raw

Re: Hotplug race on name change

Quoted 5 lines · Stephen Hemminger

Introduces new races in netdev_register_sysfs if the name changes, because netdev_register_sysfs runs without RTNL at this point. So if some application gets in and changes the device name while netdev_register_sysfs is running, then the class_dev->class_id would end up not matching the netdevice->name.

Indeed. It also seems to be possible now for someone to do an unregister_netdevice while netdev_register_sysfs is still going.

Cheers,

Signature · 4 lines
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
#12 lore raw

PATCH 1/1 netdev: hotplug napi race cleanup

This follows after the earlier two patches.

Change the initialization of the class device portion of the net device to be done earlier, so that any races before registration completes are harmless. Add a mutex to avoid changes to netdevice during the class device registration.

net/core/dev.c +9 -2
@@ -203,10 +203,12 @@
 
 #ifdef CONFIG_SYSFS
 extern int netdev_sysfs_init(void);
-extern int netdev_register_sysfs(struct net_device *);
-extern void netdev_unregister_sysfs(struct net_device *);
+extern void netdev_init_classdev(struct net_device *);
+#define netdev_register_sysfs(dev)	class_device_add(&(dev->class_dev))
+#define	netdev_unregister_sysfs(dev)	class_device_del(&(dev->class_dev))
 #else
 #define netdev_sysfs_init()	 	(0)
+#define netdev_init_classdev(dev)	do { } while(0)
 #define netdev_register_sysfs(dev)	(0)
 #define	netdev_unregister_sysfs(dev)	do { } while(0)
 #endif
@@ -2870,6 +2872,8 @@
 
 	set_bit(__LINK_STATE_PRESENT, &dev->state);
 
+	netdev_init_classdev(dev);
+
 	dev->next = NULL;
 	dev_init_scheduler(dev);
 	write_lock_bh(&dev_base_lock);
@@ -3047,7 +3051,10 @@
 			 * this is a delayed call after register_netdevice
 			 * so no way to tell device driver what is wrong.
 			 */
+			rtnl_lock();
 			err = netdev_register_sysfs(dev);
+			__rtnl_unlock();
+
 			if (err) {
 				printk(KERN_ERR "%s: failed sysfs registration (%d)\n",
 				       dev->name, err);
net/core/net-sysfs.c +2 -9
@@ -443,13 +443,8 @@
 #endif
 };
 
-void netdev_unregister_sysfs(struct net_device * net)
-{
-	class_device_del(&(net->class_dev));
-}
-
-/* Create sysfs entries for network device. */
-int netdev_register_sysfs(struct net_device *net)
+/* Setup class device */
+void netdev_init_classdev(struct net_device *net)
 {
 	struct class_device *class_dev = &(net->class_dev);
 	struct attribute_group **groups = net->sysfs_groups;
@@ -470,8 +465,6 @@
 	    || (net->wireless_handlers && net->wireless_handlers->get_wireless_stats))
 		*groups++ = &wireless_group;
 #endif
-
-	return class_device_add(class_dev);
 }
 
 int netdev_sysfs_init(void)
#13 lore raw

Re: [PATCH] netdev: hotplug napi race cleanup

Quoted 8 lines · Stephen Hemminger

This follows after the earlier two patches.

Change the initialization of the class device portion of the net device to be done earlier, so that any races before registration completes are harmless. Add a mutex to avoid changes to netdevice during the class device registration.

I had to fix up a reject in here.

net/core/dev.c line 3060

Quoted 11 lines
net/core/dev.c +3 -0
@@ -3047,7 +3051,10 @@
 			 * this is a delayed call after register_netdevice
 			 * so no way to tell device driver what is wrong.
 			 */
+			rtnl_lock();
 			err = netdev_register_sysfs(dev);
+			__rtnl_unlock();
+
 			if (err) {
 				printk(KERN_ERR "%s: failed sysfs registration (%d)\n",
 				       dev->name, err);

bix:/usr/src/25> grep 'this is a delayed call after' net/core/*.c patches/*.patch bix:/usr/src/25>

I cannot find that comment anywhere.

maintainer
#14 lore raw

Re: [PATCH] netdev: hotplug napi race cleanup

In reply to Stephen Hemminger

From: Stephen Hemminger <[email protected]>
Date: Mon, 24 Apr 2006 15:23:41 -0700

Quoted 8 lines

This follows after the earlier two patches.

Change the initialization of the class device portion of the net device to be done earlier, so that any races before registration completes are harmless. Add a mutex to avoid changes to netdevice during the class device registration.

Once Greg KH puts in the necessary infrastructure patches, I'll put this one in too.

maintainer
#15 lore raw

Re: [PATCH] netdev: hotplug napi race cleanup

In reply to Stephen Hemminger

From: Stephen Hemminger <[email protected]>
Date: Mon, 24 Apr 2006 15:23:41 -0700

Quoted 8 lines

This follows after the earlier two patches.

Change the initialization of the class device portion of the net device to be done earlier, so that any races before registration completes are harmless. Add a mutex to avoid changes to netdevice during the class device registration.

I'm not going to apply this patch and instead request that we think about why this problem exists in the first place.

This patch is even stronger evidence that doing the sysfs registry in the todo list processing is wrong. If you can legally do this while holding the rtnl semaphore, you can just as equally do it inside of register_netdevice() which is where it truly belongs.

Then you can handle errors properly, unwind the state, and return the error to the caller instead of just losing the error and leaving the device in a half-registered state.

#16 lore raw

Re: [PATCH] netdev: hotplug napi race cleanup

Quoted 15 lines · On Sat, 06 May 2006 18:09:47 -0700 (PDT) "David S. Miller"…

From: Stephen Hemminger <[email protected]>
Date: Mon, 24 Apr 2006 15:23:41 -0700

Quoted 8 lines

This follows after the earlier two patches.

Change the initialization of the class device portion of the net device to be done earlier, so that any races before registration completes are harmless. Add a mutex to avoid changes to netdevice during the class device registration.

I'm not going to apply this patch and instead request that we think about why this problem exists in the first place.

Quoted 8 lines

This patch is even stronger evidence that doing the sysfs registry in the todo list processing is wrong. If you can legally do this while holding the rtnl semaphore, you can just as equally do it inside of register_netdevice() which is where it truly belongs.

Then you can handle errors properly, unwind the state, and return the error to the caller instead of just losing the error and leaving the device in a half-registered state.

The issue is are there network devices that can't sleep during register_netdevice?

maintainer
#17 lore raw

Re: [PATCH] netdev: hotplug napi race cleanup

From: Stephen Hemminger <[email protected]>
Date: Mon, 8 May 2006 09:54:58 -0700

Quoted 2 lines

The issue is are there network devices that can't sleep during register_netdevice?

Oh right, I forgot about that.

#18 lore raw

Re: [PATCH] netdev: hotplug napi race cleanup

Quoted 7 lines · On Mon, 08 May 2006 11:37:31 -0700 (PDT) "David S. Miller"…

From: Stephen Hemminger <[email protected]>
Date: Mon, 8 May 2006 09:54:58 -0700

Quoted 2 lines

The issue is are there network devices that can't sleep during register_netdevice?

Oh right, I forgot about that.

We could do something like this in register_netdevice()

	if (in_atomic() || irqs_disabled())
		net_set_todo(dev);
	else {
		dev->reg_state = NETREG_REGISTERED;
		ret = netdev_register_sysfs(dev);
		if (ret) {
			... 
	}

It seems a bit grotty, and might cause pain later.