linux-kernel - Re: [patch] ohci1394: don't leave interrupts enabled 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] [day] [month] [year] [list]

Message-Id: <200812111702.57999.elendil@planet.nl>
Date:	Thu, 11 Dec 2008 17:02:57 +0100
From:	Frans Pop <elendil@...net.nl>
To:	Stefan Richter <stefanr@...6.in-berlin.de>
Cc:	Robert Hancock <hancockr@...w.ca>,
	"Rafael J. Wysocki" <rjw@...k.pl>,
	linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [patch] ohci1394: don't leave interrupts enabled during suspend/resume

On Wednesday 10 December 2008, Stefan Richter wrote:
> Then I have no other idea what to try.

OK and thanks for your efforts to help investigate this.

As the patch I originally submitted still does fix the issue for me, I 
would personally be happy enough if that gets included.

> Whether the Ricoh R5C832 FireWire controller or another device issues
> the interrupt is IMO still not clear.  If it was the FireWire
> controller, then it would be fundamentally broken. 
> ohci1394_pci_suspend performs actually two register writes which both
> independently disable interrupts on the controller, and I have a hard
> time to believe that the controller gets both of these straight-forward
> things wrong.

I have tried one last thing: a suspend/resume with debugging enabled for 
the driver. Attached the results of a 'dmesg | grep 1394' split for boot 
up to suspend and for resume for easier comparison.

One possibly significant difference is that on boot there are two lines
"irq_handler: Bus reset requested" and "Cancel request received", which 
are not there during resume. That's even though in both cases 
ohci_initialize() is called which includes the trigger for the bus reset.

Significant?

View attachment "1394.boot" of type "text/plain" (3634 bytes)

View attachment "1394.resume" of type "text/plain" (2955 bytes)