From: Kees Cook Subject: Re: [BISECTED] 4943ba16 ("include crypto- module prefix") breaks wifi Date: Fri, 5 Jun 2015 16:07:29 -0700 Message-ID: References: <20150430110051.32143.qmail@ns.horizon.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Cc: Mathias Krause , Herbert Xu , linux-crypto To: George Spelvin Return-path: Received: from mail-vn0-f44.google.com ([209.85.216.44]:37045 "EHLO mail-vn0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750836AbbFEXHa (ORCPT ); Fri, 5 Jun 2015 19:07:30 -0400 Received: by vnbg62 with SMTP id g62so10972807vnb.4 for ; Fri, 05 Jun 2015 16:07:29 -0700 (PDT) In-Reply-To: <20150430110051.32143.qmail@ns.horizon.com> Sender: linux-crypto-owner@vger.kernel.org List-ID: On Thu, Apr 30, 2015 at 4:00 AM, George Spelvin wrote: > Sorry for the long silence; the last e-mails arrived as I went on a trip, > and the packet got lost. > > I just upgraded my laptop to 4.0.1 and had to remember the magic > incantation to get the wireless working. ("modprobe ctr") > >> George, any updates on this? > > It turns out that I found the problem. An odd bit of pilot error, but > NOT a kernel problem. See bottom. > >> Also, could you please provide the output of "depmod -n | grep >> crypto-"? There should be lines for crypto-ccm and crypto-ctr if you >> build them as modules. > > alias crypto-twofish-asm twofish_i586 > alias crypto-twofish twofish_i586 > alias crypto-salsa20-asm salsa20_i586 > alias crypto-salsa20 salsa20_i586 > alias crypto-serpent serpent_sse2_i586 > alias crypto-cmac cmac > alias crypto-xcbc xcbc > alias crypto-md4 md4 > alias crypto-sha256-generic sha256_generic > alias crypto-sha256 sha256_generic > alias crypto-sha224-generic sha256_generic > alias crypto-sha224 sha256_generic > alias crypto-ecb ecb > alias crypto-lrw lrw > alias crypto-xts xts > alias crypto-ctr ctr > alias crypto-rfc3686 ctr > alias crypto-gcm gcm > alias crypto-rfc4543 gcm > alias crypto-rfc4106 gcm > alias crypto-gcm_base gcm > alias crypto-ccm ccm > alias crypto-rfc4309 ccm > alias crypto-ccm_base ccm > alias crypto-cryptd cryptd > alias crypto-twofish-generic twofish_generic > alias crypto-twofish twofish_generic > alias crypto-serpent-generic serpent_generic > alias crypto-serpent serpent_generic > alias crypto-tnepres serpent_generic > alias crypto-salsa20-generic salsa20_generic > alias crypto-salsa20 salsa20_generic > alias crypto-michael_mic michael_mic > alias crypto-crc32 crc32 > alias crypto-ansi_cprng ansi_cprng > alias crypto-stdrng ansi_cprng > alias crypto-ghash-generic ghash_generic > alias crypto-ghash ghash_generic > > > Anyway, the problem was a long time ago, in /etc/modprobe.d/local.conf, > I had blacklisted several unwanted crypto modules in order to suppress > some mysterious urge my system had to load every loadable module at > boot time. > > "Hey, bozo! The point of making them modules is that I *don't* want them > taking up unswappable memory all the time!" So I hit it with a hammer. > > One line written years ago was "blacklist ctr". The light dawns. > > It turns out that this doesn's stop an explicit "modprobe ctr" from > working, but *does* stop alias processing that resolves to ctr. > Thus, the kernel change broke my strange kmod (mis-)configuration. > > This fits the observed symptoms, and I apologize for wasting your time. No worries! Thanks for digging into it. I was really scratching my head. :) -Kees -- Kees Cook Chrome OS Security