linux-kernel - Re: Issue seen with FET re-enable during auto discharge time

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]

Message-ID: <20131025101613.GS18506@sirena.org.uk>
Date:	Fri, 25 Oct 2013 11:16:13 +0100
From:	Mark Brown <broonie@...nel.org>
To:	Laxman Dewangan <ldewangan@...dia.com>
Cc:	Stephen Warren <swarren@...dia.com>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: Issue seen with FET re-enable during auto discharge time

On Thu, Oct 10, 2013 at 06:35:42PM +0530, Laxman Dewangan wrote:
> Hi Mark,

Please send messages to advertised maintainer addresses; for me upstream
things sent to my work address will not get dealt with so quickly.

> In one of our system, we are using TPS65090 which has FET switches
> for  power control. This has also the auto discharger resistance for
> turning -off.

> We observed that when we disable the FET and re-enable before it
> completely off (during power discharge time), it does not get
> enabled. It enable only if wait for it to completely off.

> Does this mean we should also provide the disable time for tuning
> off (optional) so that disable_regulator() should return after
> actually tuning off?
> In downstream, we stressed this after putting delay and it worked fine.

This is the first time I've heard of hardware with that sort of
breakage, you probably want to interview the hardware engineers...  in
any case I'd not implement this just as a delay, it seems better to
return immediately on disable and instead store or set a timer for the
earliest time that we can reenable the regulator.  The delay can then be
implemented on enable if required.  This will be more consistent with
the normal behaviour so will avoid surprises for consumers.  Of course
this is pretty uncommon so perhaps it's as well to start off with the
trivial implementation...

Download attachment "signature.asc" of type "application/pgp-signature" (837 bytes)