Re: [RFC PATCH] netlink: Safer deletion of sk_bind_node

[Harish Jenny Kandiga Nagaraj <harish_kandiga@mentor.com>]

If that is the case , then subscriptions of netlink_sock should have been updated after netlink_remove or netlink_release.
I don't see that happening.

On Wednesday 03 September 2014 12:22 AM, David Miller wrote:
> From: Harish Jenny Kandiga Nagaraj <harish_kandiga@mentor.com>
> Date: Tue, 2 Sep 2014 14:14:38 +0530
>
>> In one of our random test runs we observed the crash mentioned in the previous mail.
>>
>> After debugging we found out that the call flow of the inline and static functions were
>> netlink_release
>> -----netlink_remove
>> ---------__sk_del_bind_node
>> --------------__hlist_del
>>
>> *pprev was NULL in __hlist_del function while deleting &sk->sk_bind_node hlist_node. Hence the patch was given.
>>
>> In netlink_remove function , first the sk_del_node_init function will be called. This internally calls __sk_del_node_init function. While deleting &sk->sk_node hlist_node using __sk_del_node function there is a NULL check with sk_hashed function.
>>
>> Why there is no NULL check for *pprev while deleting &sk->sk_bind_node ?
> Because if ->subscriptions is non-zero, it must be on a list, and therefore
> pprev must be non-NULL.