Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750968AbdILSwQ (ORCPT ); Tue, 12 Sep 2017 14:52:16 -0400 Received: from mail-pg0-f68.google.com ([74.125.83.68]:36843 "EHLO mail-pg0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750742AbdILSwP (ORCPT ); Tue, 12 Sep 2017 14:52:15 -0400 X-Google-Smtp-Source: ADKCNb6Kpt05e75nidpk0y8IBD6cW2prOcijh7CidTJbqB28d6WaNWEiCu0gyb/hOB9cTkbJ6L5BGg== Date: Tue, 12 Sep 2017 11:52:11 -0700 From: Dmitry Torokhov To: Jakub Kicinski Cc: Greg Kroah-Hartman , LKML Subject: Re: [bisected] Re: Module removal-related regression? Message-ID: <20170912185211.GB6654@dtor-ws> References: <20170910000338.093aa04e@cakuba.netronome.com> <20170910162111.GA17387@dtor-ws> <20170910200010.4ad7a032@cakuba.netronome.com> <8D38F5B5-C5EA-4436-91CB-B302A42A78F0@gmail.com> <20170910190908.GA5164@kroah.com> <20170910211345.5da116ba@cakuba.netronome.com> <20170911152332.GC6246@kroah.com> <20170911182926.GA22514@dtor-ws> <20170912140041.1e960388@cakuba.netronome.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20170912140041.1e960388@cakuba.netronome.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3334 Lines: 113 On Tue, Sep 12, 2017 at 02:00:41PM +0200, Jakub Kicinski wrote: > On Mon, 11 Sep 2017 11:29:26 -0700, Dmitry Torokhov wrote: > > > > That is an option, but maybe we could have the patch below for a year or > > 2 instead? > > > > Jakub, can you try and see if that works for you? > > Unfortunately this doesn't seem to solve it :( > > # modprobe nfp; lsmod | grep nfp; modprobe -r nfp; lsmod | grep nfp > nfp 1101824 0 > nfp 1101824 1 Well, I should have tested it before sending out. How about this one? -- Dmitry driver core: suppress sending MODALIAS in UNBIND uevents From: Dmitry Torokhov The current udev rules cause modules to be loaded on all device events save for "remove". With the introduction of KOBJ_BIND/KOBJ_UNBIND this causes issues, as driver modules that have devices bound to their drivers get immediately reloaded, and it appears to the user that module unloading doe snot work. The standard udev matching rule is foillowing: ENV{MODALIAS}=="?*", RUN{builtin}+="kmod load $env{MODALIAS}" Given that MODALIAS data is not terribly useful for UNBIND event, let's zap it from the generated uevent environment until we get userspace updated with the correct udev rule that only loads modules on "add" event. Reported-by: Jakub Kicinski Signed-off-by: Dmitry Torokhov --- lib/kobject_uevent.c | 49 +++++++++++++++++++++++++++++++++++++++++-------- 1 file changed, 41 insertions(+), 8 deletions(-) diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c index e590523ea476..f237a09a5862 100644 --- a/lib/kobject_uevent.c +++ b/lib/kobject_uevent.c @@ -294,6 +294,26 @@ static void cleanup_uevent_env(struct subprocess_info *info) } #endif +static void zap_modalias_env(struct kobj_uevent_env *env) +{ + static const char modalias_prefix[] = "MODALIAS="; + int i; + + for (i = 0; i < env->envp_idx;) { + if (strncmp(env->envp[i], modalias_prefix, + sizeof(modalias_prefix) - 1)) { + i++; + continue; + } + + if (i != env->envp_idx - 1) + memmove(&env->envp[i], &env->envp[i + 1], + sizeof(env->envp[i]) * env->envp_idx - 1); + + env->envp_idx--; + } +} + /** * kobject_uevent_env - send an uevent with environmental data * @@ -409,16 +429,29 @@ int kobject_uevent_env(struct kobject *kobj, enum kobject_action action, } } - /* - * Mark "add" and "remove" events in the object to ensure proper - * events to userspace during automatic cleanup. If the object did - * send an "add" event, "remove" will automatically generated by - * the core, if not already done by the caller. - */ - if (action == KOBJ_ADD) + switch (action) { + case KOBJ_ADD: + /* + * Mark "add" event so we can make sure we deliver "remove" + * event to userspace during automatic cleanup. If + * the object did send an "add" event, "remove" will + * automatically generated by the core, if not already done + * by the caller. + */ kobj->state_add_uevent_sent = 1; - else if (action == KOBJ_REMOVE) + break; + + case KOBJ_REMOVE: kobj->state_remove_uevent_sent = 1; + break; + + case KOBJ_UNBIND: + zap_modalias_env(env); + break; + + default: + break; + } mutex_lock(&uevent_sock_mutex); /* we will send an event, so request a new sequence number */