linux-kernel - Re: [E1000-devel] [PATCH RESEND 3/3] e1000e: fix accessing to suspended device

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]

Message-ID: <5131EFFE.4010700@openvz.org>
Date:	Sat, 02 Mar 2013 16:26:38 +0400
From:	Konstantin Khlebnikov <khlebnikov@...nvz.org>
To:	Ben Hutchings <bhutchings@...arflare.com>
CC:	"Waskiewicz Jr, Peter P" <peter.p.waskiewicz.jr@...el.com>,
	linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
	e1000-devel@...ts.sourceforge.net,
	"Rafael J. Wysocki" <rafael.j.wysocki@...el.com>,
	Bruce Allan <bruce.w.allan@...el.com>
Subject: Re: [E1000-devel] [PATCH RESEND 3/3] e1000e: fix accessing to suspended
 device

Ben Hutchings wrote:
> On Tue, 2013-02-26 at 14:03 +0400, Konstantin Khlebnikov wrote:
>> Waskiewicz Jr, Peter P wrote:
>>> On 2/24/2013 9:19 PM, Konstantin Khlebnikov wrote:
>>>> This patch fixes some annoying messages like 'Error reading PHY register' and
>>>> 'Hardware Erorr' and saves several seconds on reboot.
>>>
>>> Any networking-related patches should also include netdev@...r.kernel.org.
>>
>> Yeah, I forgot about this, since I came here from PCI-bus side, not from the network =)
>>
>>>
>>> I'm also a bit confused how the changes below match the patch description.
>>   >  Elaborating a bit more how the changes suppress the messages might be a good thing.
>>
>> Patch eliminates reason of these errors -- now driver will wake up
>> the device before accessing to its registers.
> [...]
>
> But e1000e calls netif_device_detach() when entering run-time suspend.
> That currently prevents any ethtool operations from running until it
> resumes.

This seems racy. dev_ethtool() runs under rtnl-lock, but suspend-resume
run asynchronously. Runtime-suspend shoudn't call netif_device_detach(),
logically device isn't dead it just take a nap for a while..

>
> (This is definitely a misfeature of either the ethtool core or e1000e.
> It is absolutely ridiculous that I can't check the PHY settings of an
> e1000e - or even driver information! - when the link is down.)

Ok link is down, but device still alive and I not see anything ridiculous
in checking its registers.

--
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/