linux-kernel - Re: irq mask swapping during suspend/resume

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:	Thu, 18 Sep 2014 13:32:06 -0700 (PDT)
From:	Thomas Gleixner <tglx@...utronix.de>
To:	Dmitry Torokhov <dtor@...gle.com>
cc:	Eric Caruso <ejcaruso@...gle.com>, rjw@...ysocki.net,
	linux-kernel@...r.kernel.org, Benson Leung <bleung@...gle.com>
Subject: Re: irq mask swapping during suspend/resume

On Wed, 17 Sep 2014, Dmitry Torokhov wrote:
> Hi Thomas,
> 
> On Wednesday, September 17, 2014 12:05:42 PM Thomas Gleixner wrote:
> > On Tue, 16 Sep 2014, Eric Caruso wrote:
> > > We would like to be able to set different irq masks for triggers during
> > > normal operation and for waking up the system. For example, while a laptop
> > > is awake, closing the lid and opening the lid should both fire an
> > > interrupt, but when the system is asleep, we would like to stay asleep
> > > when
> > > closing the lid.
> > > 
> > > We are thinking about stashing the irq mask used specifically for waking
> > > the system up in the irq_desc struct, and then swapping it during
> > > enable_irq_wake and disable_irq_wake calls. Devices that do not specify a
> > > different wake mask will use their normal trigger mask for both
> > > situations.
> > > 
> > > Is this acceptable?
> > 
> > Not really. Why should irq_desc provide storage for random
> > configurations and bind them to some random system state?
> > 
> > What's wrong with calling
> > 
> >        irq_set_type(irq, B);
> >        enable_irq_wake(irq);
> > 
> >        disable_irq_wake(irq);
> >        irq_set_type(irq, A);
> 
> The desire is to avoid doing it in [every] driver but rather have it done
> centrally by device/PM core. It does not have to be irq_desc though,
> maybe you can suggest a better place for it (aside of the individual driver
> code that is)?

Well, if it should be done by the device/pm core then you want to
store that information in the device related data structure. 

struct dev_pm_info might be the right place for it, but that's up to
Rafael.

So driver would set

   dev->power.update_wakeirq_type = true;
   dev->power.irq_type_normal = IRQ_TYPE_EDGE_BOTH;
   dev->power.irq_type_sleep = IRQ_TYPE_EDGE_LOW;

And the dev/PM core can issue the calls on suspend/resume.

Thanks,

	tglx
--
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/