netdev - Re: DM9000 issue with mem resource management

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:	Fri, 6 Jun 2008 11:11:59 +0200
From:	Laurent Pinchart <laurentp@...-semaphore.com>
To:	Ben Dooks <ben-linux@...ff.org>
Cc:	netdev@...r.kernel.org
Subject: Re: DM9000 issue with mem resource management

Hi Ben,

On Thursday 05 June 2008 19:40, Ben Dooks wrote:
> On Thu, Jun 05, 2008 at 05:42:58PM +0200, Laurent Pinchart wrote:
> > Hi everybody,
> > 
> > I ran into a resource-related bug in the DM9000 driver.
> > 
> > When the platform device has only 2 resources, dm9000_probe() doesn't set 
> > db->irq_res, which results in a segfault when the pointer gets
> > dereferenced in dm9000_open().
> > 
> > I tried to fix the issue, and found out that the resource management code
> > is quite broken.
> 
> Personally, I'm thinking about just removing the case for 2, and making it
> three resources only. The following in-kernel machines use the following
> resources:
> 
> arch/arm/mach-at91/board-sam9261ek.c			3
> arch/arm/mach-pxa/cm-x270.c				3
> arch/arm/mach-pxa/em-x270.c				3
> arch/arm/mach-pxa/trizeps4.c				3
> arch/arm/mach-pxa/colibri.c				3
> arch/arm/mach-s3c2410/mach-bast.c			3
> arch/arm/mach-s3c2410/mach-vr1000.c			3
> arch/blackfin/mach-bf527/boards/ezkit.c			2
> arch/blackfin/mach-bf533/boards/H8606.c			2
> arch/blackfin/mach-bf533/boards/ip0x.c			3
> arch/blackfin/mach-bf537/boards/generic_board.c		2
> arch/blackfin/mach-bf537/boards/stamp.c			2
> 
> As you can see, the #3 outweigh the #2. Unless anyone else objects, I
> will add a patch to reduce this to the case where the driver expects
> 3 resources, and ask the users of #2 to submit changes for their
> bots.

I've just had a look at the #2 boards and there's something I don't get. My 
understanding is that the 2 resources case is used to pass ioremap()ed memory 
to the driver, while the 3 resources case is used to pass physical addresses. 
This is what the driver currently does. Does it work by sheer luck due to a 
global 1-to-1 mapping on arm ?

Best regards,

-- 
Laurent Pinchart
CSE Semaphore Belgium

Chaussee de Bruxelles, 732A
B-1410 Waterloo
Belgium

T +32 (2) 387 42 59
F +32 (2) 387 42 75

Content of type "application/pgp-signature" skipped