netdev - Re: [PATCH 1/1] can: sja1000: clear interrupts on start

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:	Wed, 11 Nov 2015 11:06:34 +0100
From:	Marc Kleine-Budde <mkl@...gutronix.de>
To:	Mirza Krak <mirza.krak@...tmobility.com>
Cc:	wg@...ndegger.com, andri.yngvason@...el.com,
	"linux-can@...r.kernel.org" <linux-can@...r.kernel.org>,
	netdev@...r.kernel.org, linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/1] can: sja1000: clear interrupts on start

On 11/11/2015 10:54 AM, Mirza Krak wrote:
> 2015-11-11 9:04 GMT+01:00 Marc Kleine-Budde <mkl@...gutronix.de>:
>> This problem occurs only an hardware, where the SJA1000 is powered
>> during system suspend?
> 
> Correct, and the controller is in UP state.
> 
>> Doesn't the SJA1000 trigger an interrupt that is detected after
>> resuming? You add the fix to the open() function, which is triggered
>> during $(ifconfig up), how is related to suspend/resume? Does the
>> network layer call $(ifconfig down) during shutdown?
> 
> It does not trigger another interrupt.

Can you put a scope on the interrupt pin and see if it's a problem of
the SoC and/or Linux or the SJA1000 not pulling the IRQ line.

> In our tests we have seen that IE and EPI flags are set in the IR
> registered when we resume the system. If we do not clear these two
> flags it does not produce any other interrupts. Setting the controller
> in to reset mode does not clear the IE and EPI flags, I can understand
> that IE flag is not cleared as this is stated in the data-sheet, can
> not explain why EPI flag is not cleared though as data-sheet states 0
> (reset) for both hardware reset and SETTING MOD.0 BY SOFTWARE (which
> is the reset-mode)
> 
> Reason I put it in open() is so at least a DOWN/UP procedure returns
> the controller to an operating state. Also if we do an UP we should
> clear any earlier states that might exist in the registers like we do
> with error counters and error code capture.

Yes - makes sense to have it in open().

> Network layer does not call $(ifconfig down) during shutdown.
> 
>> There isn't any suspend/resume code in the sja1000 driver, for me it
>> seems that we should add resume code that handles this problem.
> 
> Yes, resume code should be implemented to handle this and other
> problems (receive data). But still a DOWN/UP procedure should clear
> any previous state that could exist in the controller registers.

I'll add your patch to can/master and add stable on Cc. Looking forward
for some suspend/resume code :)

Thanks,
Marc

-- 
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |


Download attachment "signature.asc" of type "application/pgp-signature" (456 bytes)