linux-kernel - Re: [ 57/72] genirq: Unmask oneshot irqs when thread was not woken

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:	Tue, 6 Mar 2012 21:25:15 +0100
From:	"Stefan Lippers-Hollmann" <s.L-H@....de>
To:	Thomas Gleixner <tglx@...utronix.de>
Cc:	Linus Torvalds <torvalds@...ux-foundation.org>,
	Sven Joachim <svenjoac@....de>,
	Greg KH <gregkh@...uxfoundation.org>,
	LKML <linux-kernel@...r.kernel.org>, stable@...r.kernel.org,
	Andrew Morton <akpm@...ux-foundation.org>,
	Alan Cox <alan@...rguk.ukuu.org.uk>,
	Jonathan Nieder <jrnieder@...il.com>,
	linux-wireless@...r.kernel.org,
	Stefano Brivio <stefano.brivio@...imi.it>
Subject: Re: [ 57/72] genirq: Unmask oneshot irqs when thread was not woken

Hi

On Tuesday 06 March 2012, Thomas Gleixner wrote:
> On Tue, 6 Mar 2012, Thomas Gleixner wrote:
> > On Tue, 6 Mar 2012, Thomas Gleixner wrote:
> > 
> > > On Mon, 5 Mar 2012, Linus Torvalds wrote:
> > > 
> > > > Thomas, should we just plan on reverting that commit from mainline? It
> > > > clearly causes regressions.
> > > 
> > > Give me a day or two to figure out why it breaks stuff. I have no idea
> > > why it should wreckage anything.
> > 
> > Hmm. This is interesting. The b43 driver has a primary handler which
> > can return IRQ_NONE. So up to that change the interrupt line was kept
> > disabled when that happened. Possibly the driver relies on that
> > behaviour. Digging for a machine with a b43.
> 
> Does not reproduce. Now I was looking at the driver again, it does not
> use IRQ_ONESHOT anyway.
> 
> So for handle_fasteoi_irq() this patch is actually a NOOP. So the only
> affected handler would be handle_level_irq(). Still can't see how it
> changes the !IRQ_ONESHOT behaviour :(
> 
> Stephan, Sven: Can you please provide the output of /proc/interrupts ?

Full /proc/interrupts and gzipped dmesg attached for kernel 3.2.9 + 
current stable queue-3.2, no difference in the interrupt output between
working and broken state. Both outputs taken immediately after a cold
reboot.

*.broken is with "genirq: Unmask oneshot irqs when thread was not 
woken" applied, resulting in empty wlan scan results, no auth).

*.working with only this patch reverted (3 wlan cells visible, 
immediate auth.

Regards
	Stefan Lippers-Hollmann

Download attachment "dmesg.broken.gz" of type "application/x-gzip" (11242 bytes)

Download attachment "dmesg.working.gz" of type "application/x-gzip" (11513 bytes)

View attachment "interrupts.broken" of type "text/plain" (1219 bytes)

View attachment "interrupts.working" of type "text/plain" (1220 bytes)

Download attachment "signature.asc " of type "application/pgp-signature" (199 bytes)