netdev - RE: ks8851: reset on RX failure and TX timeout

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:	Thu, 10 Apr 2014 09:16:20 +0000
From:	"Nekludov, Max" <Max.Nekludov@...ter.com>
To:	David Miller <davem@...emloft.net>
CC:	"torvalds@...ux-foundation.org" <torvalds@...ux-foundation.org>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"sboyd@...eaurora.org" <sboyd@...eaurora.org>,
	"mjr@...wisc.edu" <mjr@...wisc.edu>,
	"jiri@...nulli.us" <jiri@...nulli.us>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"trivial@...tcorp.com.au" <trivial@...tcorp.com.au>
Subject: RE: ks8851: reset on RX failure and TX timeout

I've tried to use ndo_tx_timeout. Actually it works. But it takes too long time after TX failure and before watchdog fires.

As I understand ndo_tx_timeout will be called only if tx queue is stopped. But ks8851 driver stops queue only after it's full.
So one have to  send a lot of data to reset device (~6KiB).

But I need to recover as fast as possible. Could you suggest how to achieve that?
________________________________________
From: David Miller <davem@...emloft.net>
Sent: Thursday, April 3, 2014 9:32 PM
To: Nekludov, Max
Cc: torvalds@...ux-foundation.org; netdev@...r.kernel.org; sboyd@...eaurora.org; mjr@...wisc.edu; jiri@...nulli.us; linux-kernel@...r.kernel.org; trivial@...tcorp.com.au
Subject: Re: ks8851: reset on RX failure and TX timeout

From: "Nekludov, Max" <Max.Nekludov@...ter.com>
Date: Wed, 2 Apr 2014 06:31:23 +0000

> Electromagnetic noise can make device to not send TX interrupts.
> As result outgoing queue will be suspended forever.
> Also EMI can raise RX interrupt with zero value in KS_RXFC register.
>
> Signed-off-by: Max Nekludov <Max.Nekludov@...ter.com>

We have a watchdog facility for TX timeouts, don't reinvent the wheel.

Implement ->ndo_tx_timeout() and set dev->watchdog_timeo appropriately.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html