From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH net] bpf: fix regression on verifier pruning wrt map lookups Date: Sat, 17 Dec 2016 10:51:57 -0500 (EST) Message-ID: <20161217.105157.1777770626254774920.davem@davemloft.net> References: Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: ast@fb.com, jbacik@fb.com, tgraf@suug.ch, netdev@vger.kernel.org To: daniel@iogearbox.net Return-path: Received: from shards.monkeyblade.net ([184.105.139.130]:40812 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752560AbcLQPv7 (ORCPT ); Sat, 17 Dec 2016 10:51:59 -0500 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: From: Daniel Borkmann Date: Thu, 15 Dec 2016 01:30:06 +0100 > Commit 57a09bf0a416 ("bpf: Detect identical PTR_TO_MAP_VALUE_OR_NULL > registers") introduced a regression where existing programs stopped > loading due to reaching the verifier's maximum complexity limit, > whereas prior to this commit they were loading just fine; the affected > program has roughly 2k instructions. > > What was found is that state pruning couldn't be performed effectively > anymore due to mismatches of the verifier's register state, in particular > in the id tracking. It doesn't mean that 57a09bf0a416 is incorrect per > se, but rather that verifier needs to perform a lot more work for the > same program with regards to involved map lookups. ... > Fixes: 57a09bf0a416 ("bpf: Detect identical PTR_TO_MAP_VALUE_OR_NULL registers") > Signed-off-by: Daniel Borkmann > Acked-by: Thomas Graf > Acked-by: Alexei Starovoitov Applied.