linux-kernel - Re: [PATCH/RFC RESEND] leds: Use set_brightness_work for brightness

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]

Date:	Tue, 30 Jun 2015 15:55:36 +0300
From:	Stas Sergeev <stsp@...t.ru>
To:	Jacek Anaszewski <j.anaszewski@...sung.com>
CC:	Stas Sergeev <stsp@...rs.sourceforge.net>,
	linux-leds@...r.kernel.org, linux-kernel@...r.kernel.org,
	Bryan Wu <cooloney@...il.com>,
	Richard Purdie <rpurdie@...ys.net>,
	Pavel Machek <pavel@....cz>,
	Sakari Ailus <sakari.ailus@...ux.intel.com>,
	Andreas Werner <andreas.werner@....de>,
	Andrew Lunn <andrew@...n.ch>,
	Antonio Ospite <ospite@...denti.unina.it>,
	Atsushi Nemoto <anemo@....ocn.ne.jp>,
	Ben Dooks <ben@...tec.co.uk>, Chris Boot <bootc@...tc.net>,
	Dan Murphy <dmurphy@...com>,
	Daniel Jeong <daniel.jeong@...com>,
	Daniel Mack <daniel@...que.org>,
	"David S. Miller" <davem@...emloft.net>,
	Fabio Baltieri <fabio.baltieri@...il.com>,
	Felipe Balbi <balbi@...com>,
	Florian Fainelli <florian@...nwrt.org>,
	"G.Shark Jeong" <gshark.jeong@...il.com>,
	Guennadi Liakhovetski <g.liakhovetski@....de>,
	Ingi Kim <ingi2.kim@...sung.com>,
	Jan-Simon Moeller <dl9pf@....de>,
	Johan Hovold <johan@...nel.org>, John Lenz <lenz@...wisc.edu>,
	Jonas Gorski <jogo@...nwrt.org>,
	Kim Kyuwon <q1.kim@...sung.com>,
	Kristian Kielhofner <kris@...sk.org>,
	Kristoffer Ericson <kristoffer.ericson@...il.com>,
	Linus Walleij <linus.walleij@...aro.org>,
	Mark Brown <broonie@...nel.org>,
	Michael Hennerich <michael.hennerich@...log.com>,
	Milo Kim <milo.kim@...com>,
	Márton Németh <nm127@...email.hu>,
	Nate Case <ncase@...-inc.com>, NeilBrown <neilb@...e.de>,
	Nick Forbes <nick.forbes@...epta.com>,
	Paul Parsons <lost.distance@...oo.com>,
	Peter Meerwald <p.meerwald@...-electronic.com>,
	Phil Sutter <n0-1@...ewrt.org>,
	Philippe Retornaz <philippe.retornaz@...l.ch>,
	Raphael Assenat <raph@...com>,
	Richard Purdie <rpurdie@...nedhand.com>,
	Rod Whitby <rod@...tby.id.au>, Dave Hansen <dave@...1.net>,
	Rodolfo Giometti <giometti@...ux.it>,
	"Sebastian A. Siewior" <bigeasy@...utronix.de>,
	Shuah Khan <shuahkhan@...il.com>,
	Simon Guinot <sguinot@...ie.com>,
	Álvaro Fernández Rojas <noltari@...il.com>
Subject: Re: [PATCH/RFC RESEND] leds: Use set_brightness_work for brightness_set
 ops that can sleep

30.06.2015 15:41, Jacek Anaszewski пишет:
> On 06/30/2015 01:41 PM, Stas Sergeev wrote:
>> 30.06.2015 11:27, Jacek Anaszewski пишет:
>>> On 06/29/2015 05:17 PM, Stas Sergeev wrote:> 29.06.2015 17:05, Jacek Anaszewski пишет:
>>>>> +         * If need to disable soft blinking delegate this to the
>>>>> +         * work queue task to avoid problems in case we are
>>>>> +         * called from hard irq context.
>>>>> +         */
>>>>> +        led_cdev->flags |= LED_BLINK_DISABLE;
>>>> Wouldn't it be better to just enforce the callers
>>>> to explicitly disable software blink, so that it to
>>>> never happen from irq context? Something like in this
>>>> patch:
>>>> https://lkml.org/lkml/2015/5/13/491
>>>>
>>>
>>> Blinking can be disabled not only by removing trigger explicitly,
>>> but also by setting brightness to 0 and led_set_brightness
>>> can be called from hard irq context. set_brightness_work
>>> was originally introduced exactly for this use case.
>> Could you please describe where does this happen?
> This in fact doesn't take place in the mainline kernel,
> however there are some out of tree users apparently [1].
Oh, IMHO the hooks that are needed only for out-of-tree code
require the appropriate comment, at least. I was entirely
confused by its existence. IIRC in the past there was even
the policy to not include any hooks for out-of-tree code at
all.

> Modifications I am proposing also need set_brightness_work,
> so there is almost no cost of keeping the support for calling led_set_brightness from hard irq context intact.
I wonder if there are possible races, eg you schedule disabling
of the softblink timer, and someone else at the same time also
disables it (and probably also re-enables).
Well if you think that code is safe, I would agree that the cost
is now very small with your patch, so maybe it is not a big deal
any more.
--
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/