From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH net-next v2 3/5] net: filter: get rid of sock_fprog_kern Date: Thu, 24 Apr 2014 16:02:37 -0400 (EDT) Message-ID: <20140424.160237.1682834218902091375.davem@davemloft.net> References: <1398321927-8845-1-git-send-email-dborkman@redhat.com> <1398321927-8845-4-git-send-email-dborkman@redhat.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: ast@plumgrid.com, netdev@vger.kernel.org To: dborkman@redhat.com Return-path: Received: from shards.monkeyblade.net ([149.20.54.216]:33913 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751177AbaDXUCj (ORCPT ); Thu, 24 Apr 2014 16:02:39 -0400 In-Reply-To: <1398321927-8845-4-git-send-email-dborkman@redhat.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Daniel Borkmann Date: Thu, 24 Apr 2014 08:45:25 +0200 > It is actually cleaner to just get rid of sock_fprog_kern structure. > It's not really useful as we can just use sock_fprog structure as > we do elsewhere in the kernel, this could throw some sparse false > positives though, but getting rid of the structure duplication is > probably the better way to go. > > Signed-off-by: Daniel Borkmann > Acked-by: Alexei Starovoitov But this is the whole point of sock_fprog_kern. It's so that we can always have the correct type for the pointers, and we can show cleanly (without lots of ugly casts) to sparse that whether we expect a pointer to be a user or a kernel one. I really don't like this change, sorry.