linux-kernel - Re: [PATCH 03/10] mfd / platform: cros_ec: Miscellaneous character device to talk with the EC

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: <20190605084002.GP4797@dell>
Date:   Wed, 5 Jun 2019 09:40:02 +0100
From:   Lee Jones <lee.jones@...aro.org>
To:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc:     Guenter Roeck <groeck@...gle.com>,
        Ezequiel Garcia <ezequiel@...labora.com>,
        Enric Balletbo i Serra <enric.balletbo@...labora.com>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        Gwendal Grignou <gwendal@...omium.org>,
        Guenter Roeck <groeck@...omium.org>,
        Benson Leung <bleung@...omium.org>, kernel@...labora.com,
        Dmitry Torokhov <dtor@...omium.org>,
        Gustavo Pimentel <gustavo.pimentel@...opsys.com>,
        Randy Dunlap <rdunlap@...radead.org>,
        Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        linux-doc@...r.kernel.org, Enno Luebbers <enno.luebbers@...el.com>,
        Guido Kiener <guido@...ner-muenchen.de>,
        Thomas Gleixner <tglx@...utronix.de>,
        Kishon Vijay Abraham I <kishon@...com>,
        Jonathan Corbet <corbet@....net>, Wu Hao <hao.wu@...el.com>,
        Kate Stewart <kstewart@...uxfoundation.org>,
        Tycho Andersen <tycho@...ho.ws>,
        Gerd Hoffmann <kraxel@...hat.com>,
        Jilayne Lovejoy <opensource@...ayne.com>
Subject: Re: [PATCH 03/10] mfd / platform: cros_ec: Miscellaneous character
 device to talk with the EC

On Wed, 05 Jun 2019, Greg Kroah-Hartman wrote:

> On Wed, Jun 05, 2019 at 07:48:39AM +0100, Lee Jones wrote:
> > On Tue, 04 Jun 2019, Greg Kroah-Hartman wrote:
> > > On Tue, Jun 04, 2019 at 11:39:21AM -0700, Guenter Roeck wrote:
> > > > On Tue, Jun 4, 2019 at 11:35 AM Greg Kroah-Hartman
> > > > <gregkh@...uxfoundation.org> wrote:
> > > > >
> > > > > On Tue, Jun 04, 2019 at 01:58:38PM -0300, Ezequiel Garcia wrote:
> > > > > > Hey Greg,
> > > > > >
> > > > > > > > + dev_info(&pdev->dev, "Created misc device /dev/%s\n",
> > > > > > > > +          data->misc.name);
> > > > > > >
> > > > > > > No need to be noisy, if all goes well, your code should be quiet.
> > > > > > >
> > > > > >
> > > > > > I sometimes wonder about this being noise or not, so I will slightly
> > > > > > hijack this thread for this discussion.
> > > > > >
> > > > > > >From a kernel developer point-of-view, or even from a platform
> > > > > > developer or user with a debugging hat point-of-view, having
> > > > > > a "device created" or "device registered" message is often very useful.
> > > > >
> > > > > For you, yes.  For someone with 30000 devices attached to their system,
> > > > > it is not, and causes booting to take longer than it should be.
> > 
> > Who has 30,000 devices attached to their systems?
> 
> More than you might imagine.
> 
> > I would argue that
> > in these special corner-cases, they should knock the log-level *down*
> > a notch.  For the rest of us who run normal platforms, an extra second
> > of boot time renders a more forthcoming/useful system than if each of
> > our devices initialised silently.
> > 
> > Personally I like to know what devices I have on my system, and the
> > kernel log is the first place I look.  As far as I'm concerned, for
> > the most part, if it's not in the kernel log, I don't have it.
> 
> Then you "do not have" lots of devices, as we have been removing these
> messages for a number of years now :)
> 
> >  "Oh wow, I didn't know I had XXX functionality on this platform."
> > 
> > In my real job, I am currently enabling some newly released AArch64
> > based laptops for booting with ACPI.  I must have wasted a day whilst
> > enabling some of the devices the system relies upon, just to find
> > out that 90% of them were actually probing semi-fine (at least probe()
> > was succeeding), just silently. *grumble*
> 
> Yup, that's normal.  If you want to see what devices are in the system,
> look in /sys/devices/ as that is what it is for, not the kernel log.

My guess is that less than 1% of Linux users use /sys/devices in this
way.  It's a very unfriendly interface.  Besides, when enabling a new
platform, access to sysfs comes too far down the line to be useful in
the majority of cases.

-- 
Lee Jones [李琼斯]
Linaro Services Technical Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog