4.19 stable kernel crash caused by vxlan testing

4.19 stable kernel crash caused by vxlan testing

[zhulei]

Hey all,

I recently used a testing program to test the 4.19 stable branch kernel and found that a crash occurred immediately. The test source code link is:
https://github.com/Backmyheart/src0358/blob/master/vxlan_fdb_destroy.c

The test command is as follows:
gcc vxlan_fdb_destroy.c -o vxlan_fdb_destroy -lpthread

According to its stack, upstream has relevant repair patch, the commit id is 7c31e54aeee517d1318dfc0bde9fa7de75893dc6.

May i ask if the 4.19 kernel will port this patch ?

Re: 4.19 stable kernel crash caused by vxlan testing

[Taehee Yoo]

On Tue, Apr 23, 2024 at 2:53 PM zhulei <zhulei_szu@163.com> wrote:
>
> Hey all,
>
> I recently used a testing program to test the 4.19 stable branch kernel and found that a crash occurred immediately. The test source code link is:
> https://github.com/Backmyheart/src0358/blob/master/vxlan_fdb_destroy.c
>
> The test command is as follows:
> gcc vxlan_fdb_destroy.c -o vxlan_fdb_destroy -lpthread
>
> According to its stack, upstream has relevant repair patch, the commit id is 7c31e54aeee517d1318dfc0bde9fa7de75893dc6.
>
> May i ask if the 4.19 kernel will port this patch ?

Hi zhulei,

The commit 7c31e54aeee5 ("vxlan: do not destroy fdb if
register_netdevice() is failed") was not backported to 4.19-stable
tree.
https://lore.kernel.org/stable/15641355392228@kroah.com/

So, you can request a backport.
Please check the URL:
https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html

Thanks a lot!
Taehee Yoo

Re: 4.19 stable kernel crash caused by vxlan testing

[Greg KH]

On Tue, Apr 23, 2024 at 01:52:40PM +0800, zhulei wrote:
> Hey all,
> 
> I recently used a testing program to test the 4.19 stable branch kernel and found that a crash occurred immediately. The test source code link is:
> https://github.com/Backmyheart/src0358/blob/master/vxlan_fdb_destroy.c
> 
> The test command is as follows:
> gcc vxlan_fdb_destroy.c -o vxlan_fdb_destroy -lpthread
> 
> According to its stack, upstream has relevant repair patch, the commit id is 7c31e54aeee517d1318dfc0bde9fa7de75893dc6.
> 
> May i ask if the 4.19 kernel will port this patch ?

Have you tested to verify that it does work?  If so, please provide the
working patch because as-is, it does not apply to 4.19.y at all.

thansk,

greg k-h

Re: 4.19 stable kernel crash caused by vxlan testing

[zhulei]

By making the following changes and using our own 4.19 kernel
verification, we can observe the output that passes the test
in dmesg.

When using the 4.19 stable branch kernel for verification, the
dmesg has no output, which looks more like the test program is not
running. However, the code of vxlan.c is the same.

May i ask if the changes made to the maintainer are reasonable?

---
 drivers/net/vxlan.c | 36 ++++++++++++++++++++++++++----------
 1 file changed, 26 insertions(+), 10 deletions(-)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 704db80df38b..d3000c58c0f2 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -657,6 +657,14 @@ static struct vxlan_fdb *vxlan_fdb_alloc(struct vxlan_dev *vxlan,
 	return f;
 }
 
+static void vxlan_fdb_insert(struct vxlan_dev *vxlan, const u8 *mac,
+			     __be32 src_vni, struct vxlan_fdb *f)
+{
+	++vxlan->addrcnt;
+	hlist_add_head_rcu(&f->hlist,
+			   vxlan_fdb_head(vxlan, mac, src_vni));
+}
+
 static int vxlan_fdb_create(struct vxlan_dev *vxlan,
 			    const u8 *mac, union vxlan_addr *ip,
 			    __u16 state, __be16 port, __be32 src_vni,
@@ -682,10 +690,6 @@ static int vxlan_fdb_create(struct vxlan_dev *vxlan,
 		return rc;
 	}
 
-	++vxlan->addrcnt;
-	hlist_add_head_rcu(&f->hlist,
-			   vxlan_fdb_head(vxlan, mac, src_vni));
-
 	*fdb = f;
 
 	return 0;
@@ -758,15 +762,15 @@ static int vxlan_fdb_update(struct vxlan_dev *vxlan,
 	if (notify) {
 		if (rd == NULL)
 			rd = first_remote_rtnl(f);
+		vxlan_fdb_insert(vxlan, mac, src_vni, f);
 		vxlan_fdb_notify(vxlan, f, rd, RTM_NEWNEIGH);
 	}
 
 	return 0;
 }
 
-static void vxlan_fdb_free(struct rcu_head *head)
+static void __vxlan_fdb_free(struct vxlan_fdb *f)
 {
-	struct vxlan_fdb *f = container_of(head, struct vxlan_fdb, rcu);
 	struct vxlan_rdst *rd, *nd;
 
 	list_for_each_entry_safe(rd, nd, &f->remotes, list) {
@@ -776,6 +780,13 @@ static void vxlan_fdb_free(struct rcu_head *head)
 	kfree(f);
 }
 
+static void vxlan_fdb_free(struct rcu_head *head)
+{
+	struct vxlan_fdb *f = container_of(head, struct vxlan_fdb, rcu);
+
+	__vxlan_fdb_free(f);
+}
+
 static void vxlan_fdb_destroy(struct vxlan_dev *vxlan, struct vxlan_fdb *f,
 			      bool do_notify)
 {
@@ -3265,9 +3276,12 @@ static int __vxlan_dev_create(struct net *net, struct net_device *dev,
 	if (err)
 		goto errout;
 
-	/* notify default fdb entry */
-	if (f)
+	if (f) {
+		vxlan_fdb_insert(vxlan, all_zeros_mac,
+				 vxlan->default_dst.remote_vni, f);
+		/* notify default fdb entry */
 		vxlan_fdb_notify(vxlan, f, first_remote_rtnl(f), RTM_NEWNEIGH);
+	}
 
 	list_add(&vxlan->next, &vn->vxlan_list);
 	return 0;
@@ -3278,9 +3292,11 @@ static int __vxlan_dev_create(struct net *net, struct net_device *dev,
 	 * destroy the entry by hand here.
 	 */
 	if (f)
-		vxlan_fdb_destroy(vxlan, f, false);
-	if (unregister)
+		__vxlan_fdb_free(f);
+	if (unregister) {
+		vxlan_fdb_destroy(vxlan, f, true);
 		unregister_netdevice(dev);
+	}
 	return err;
 }
 
-- 
2.27.0

Re: 4.19 stable kernel crash caused by vxlan testing

[Greg KH]

On Tue, May 07, 2024 at 05:20:49PM +0800, zhulei wrote:
> By making the following changes and using our own 4.19 kernel
> verification, we can observe the output that passes the test
> in dmesg.
> 
> When using the 4.19 stable branch kernel for verification, the
> dmesg has no output, which looks more like the test program is not
> running. However, the code of vxlan.c is the same.
> 
> May i ask if the changes made to the maintainer are reasonable?

Why is this only a 4.19.y thing?  Why is this not also relevent for
Linus's current tree?

thanks,

greg k-h