linux-kernel - RE: Problem with /proc/iomem on ARM

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]

Date:	Wed, 8 Jul 2009 18:09:51 -0400
From:	"H Hartley Sweeten" <hartleys@...ionengravers.com>
To:	"Kay Sievers" <kay.sievers@...y.org>
Cc:	<linux-kernel@...r.kernel.org>
Subject: RE: Problem with /proc/iomem on ARM

On Wednesday, July 01, 2009 11:47 AM, Kay Sievers wrote:
> On Wed, Jul 1, 2009 at 20:24, H Hartley Sweeten wrote:
>> On Wednesday, July 01, 2009 11:13 AM, Kay Sievers wrote:
>>>> The serial ports on my system are amba bus devices.
>>>>
>>>>> After all, it seems like this should be fixed in arm somewhere,
>>>>> not to copy and store internal driver core pointers, but use the
>>>>> device itself to retrieve the values.
>>>>
>>>> Maybe the problem is in drivers/amba/bus.c? amba_device_register()
>>>> does:
>>>>
>>>>        dev->res.name = dev_name(&dev->dev);
>>>
>>> Looks like. If you get the name directly from dev->init_name. Does
>>> that work?
>>>
>>
>> The following patch does fix /proc/iomem for the amba uarts on my
>> system.  Is this the correct approach to fixing the issue?
>
> Looks fine to work around the issue that statically allocated struct
> devices can cause. The proper fix would be to convert them to dynamic
> objects and get rid of .init_name entirely.

Disregard.

Russell King added a patch that fixed the issue.  I just showed up.

commit 557dca5f48a45df88a73e69ee0700cfd4e2358c9
Author: Russell King <rmk@...-67.arm.linux.org.uk>
Date:   Sun Jul 5 22:39:08 2009 +0100

    [ARM] amba: fix amba device resources

    AMBA device resources were being reported as:

    10004000-10004fff : <BAD>

    This is because dev_name() was returning NULL prior to device_register.
    Ensure that the struct device is properly initialized, and the name is
    set before adding it to the device tree.

    Signed-off-by: Russell King <rmk+kernel@....linux.org.uk>

Thanks,
Hartley