lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20130328180643.GB10420@lnx-rg>
Date:	Thu, 28 Mar 2013 19:06:45 +0100
From:	Richard GENOUD <richard.genoud@...il.com>
To:	Stephen Warren <swarren@...dotorg.org>
Cc:	Linus Walleij <linus.walleij@...aro.org>,
	Stephen Warren <swarren@...dia.com>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 4/4] pinctrl: re-enable old state in case of error in
 pinctrl_select_state

On [jeu., 28.03.2013 11:38:23], Stephen Warren wrote:
> On 03/28/2013 11:34 AM, Richard GENOUD wrote:
> > On [mer., 27.03.2013 17:55:45], Stephen Warren wrote:
> >> On 03/25/2013 08:47 AM, Richard Genoud wrote:
> >>> If a new state is applied, the groups configured in the old state but
> >>> not in the new state are disabled.
> >>> If something goes wrong and the new state can't be applied, we have to
> >>> re-enable those groups.
> >>
> >> What is the use-case for this? I wonder if it isn't better to simply
> >> undo the partial selection of the new state (as patch 3/4 attempts to
> >> do) and then leave p->state==NULL, indicating that no state is actively
> >> selected. IIRC, this would be the same as right after the initial
> >> pinctrl_select().
> >>
> >> I wonder if it's likely that attempting to re-apply the old state would
> >> actually work, given that applying something just failed.
> >>
> >> Finally, this recovery code doesn't:
> >>
> >> a) Process anything except MUX_GROUP; any pin config settings in the old
> >> state aren't restored.
> >>
> >> b) (I think) Apply any mux settings that don't involve groups that are
> >> referenced by both the old and new states; given that patch 3/4 attempts
> >> to undo everything in the failed application of the new state, I think
> >> this "re-apply the old state" code should simple run through everything
> >> in the old state any apply it unconditionally.
> >
> > So, if I understand correctly, it could be as simple as that:
> >  	}
> >  
> > -	if (old_state) {
> > -		list_for_each_entry(setting, &old_state->settings, node) {
> > -			bool found = false;
> > -			if (setting->type != PIN_MAP_TYPE_MUX_GROUP)
> > -				continue;
> > -			list_for_each_entry(setting2, &state->settings, node) {
> > -				if (setting2->type != PIN_MAP_TYPE_MUX_GROUP)
> > -					continue;
> > -				if (setting2->data.mux.group ==
> > -						setting->data.mux.group) {
> > -					found = true;
> > -					break;
> > -				}
> > -			}
> > -			if (!found)
> > -				pinmux_enable_setting(setting);
> > -		}
> > -	}
> > -
> >  	p->state = old_state;
> 
> I think you want to remove that line too, so that p->state == NULL in
> the error case, so that if the pinctrl_select_state_locked() call below
> also fails to restore the old state, then (!old_state) will be true
> inside the recursive call, so it doesn't recurse into itself forever.
> 
> > +	if (old_state)
> > +		pinctrl_select_state_locked(p, NULL);
> 
> You want to pass old_state rather than NULL there, I think.
Hum, I'm starting to code nonsense. It's time to go home !

But yes, we clearly don't want to have p->state != NULL in the recursive
call (I've seen that, and an evil force made me wrote the contrary :))

So, just:
+	if (old_state)
+		pinctrl_select_state_locked(p, old_state);
Will do the trick.
It can't loop forever as p->state is null.
If it fails, p->state will be still null.
If it succeeds, p->state will be == old_state, and that's what we want.

Got it !(finally)

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ