linux-kernel - Re: [PATCH] firewire: Fix ohci free

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:	Sat, 2 Feb 2013 10:16:00 -0500 (EST)
From:	Alan Stern <stern@...land.harvard.edu>
To:	Stefan Richter <stefanr@...6.in-berlin.de>
cc:	Mark Einon <mark.einon@...il.com>,
	Peter Hurley <peter@...leysoftware.com>,
	<linux1394-devel@...ts.sourceforge.net>,
	<linux-kernel@...r.kernel.org>, <linux-pm@...r.kernel.org>
Subject: Re: [PATCH] firewire: Fix ohci free_irq() warning

On Sat, 2 Feb 2013, Stefan Richter wrote:

> On Feb 01 Mark Einon wrote:
> > On 1 February 2013 21:09, Peter Hurley <peter@...leysoftware.com> wrote:
> > >>>> On Jan 29 Alan Stern wrote:
> > >>>>> Why does the pci_suspend routine call free_irq() at all?  As far as I 
> > >>>>> know, it's not supposed to do that.  Won't the device continue to use 
> > >>>>> the same IRQ after it is resumed?
> 
> As far as I can tell, it happened to be done that way as a side effect of
> how the probe() and resume() methods share code.  It has remained like
> this since the initial implementation:
> http://git.kernel.org/linus/2aef469a35a2

At one point, quite a few years ago, Linus complained about drivers the 
release IRQs during suspend only to reacquire them during resume.  A 
little refactoring should be able to separate out resource 
acquisition/release (done only during probe and remove) from activation 
and shutdown (also done during resume and suspend).

> Still, at this point I would like to learn whether .suspend() followed
> by .remove() is a valid order of sequence which drivers must support
> before I prepare myself to get comfortable with a refactoring of
> firewire-ohci's .probe()/.resume()/suspend()/remove().  Obviously, so far
> my assumption was that a successful .suspend() can only ever be followed
> by .resume().

It depends on the subsystem.  Some subsystems do have suspend -> remove
transitions and others don't.  In general, it's a good idea for drivers
to be prepared for removal while the system is asleep.  Presumably any
hot-unpluggable bus (which includes most of the important buses these
days) would have to support it.

Alan Stern

--
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/