linux-kernel - Re: [RFC PATCH 2/2] leds: trigger: implement block trigger

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] [day] [month] [year] [list]

Date:   Tue, 4 May 2021 17:43:56 +0200
From:   Pavel Machek <pavel@....cz>
To:     Enzo Matsumiya <ematsumiya@...e.de>
Cc:     Hannes Reinecke <hare@...e.de>, linux-leds@...r.kernel.org,
        linux-block@...r.kernel.org, u.kleine-koenig@...gutronix.de,
        Jens Axboe <axboe@...nel.dk>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH 2/2] leds: trigger: implement block trigger

On Mon 2021-05-03 13:56:15, Enzo Matsumiya wrote:
> On 05/03, Pavel Machek wrote:
> > > As already commented on, this for_each_blk() construct is not a good idea.
> > > Infact, I guess it would be better if you could invert the logic:
> > > Not having the block trigger enumerating all devices, but rather let the
> > > devices register with the block trigger.
> > > That would have the benefit that one could choose which block device should
> > > be handled by the LED trigger subsystem, _and_ you would avoid the need for
> > > a for_each_blk() construct.
> > > Thing is, I don't think that all block devices should be handled by the LED
> > > trigger; eg for things like 'loop' or 'ramdisk' it is very
> > > >questionable.
> > 
> > > Downside is that you would need to modify the drivers, but realistically
> > > there are only very few drivers which should be modified; I would go for
> > > nvme-pci and the sd driver for starters. Maybe floppy, but arguably that can
> > > omitted as one has a very good audio indicator for floppy accesses
> > > :-)
> > 
> > And we already have disk activity trigger. Maybe NVMe and SD needs to
> > be modified to use it?
> 
> TBH I haven't thought of that. My initial idea was to actually offer
> maximum flexibility to the user, so exposing all block devices on the
> system [*], being able to set any LED available as an indicator for each
> of those.
> 
> But, indeed, just using ledtrig-disk in NVMe and SD might just be
> simpler.
> 
> 
> [*] - again, I see now this was a bad idea and will be changed in a
> possible next version

Sounds like there should be no new version. Modify NVMe/SD if
required, instead.

Oh and disk-activity LED trigger blinks when disk is fully loaded. I
believe that is a bug and I'd not mind if it was fixed. I probably
have local patch that needs cleaning up somewhere.

Best regards,
								Pavel

--