netdev - RE: [PATCH V2 2/4] can: flexcan: try to exit stop mode during probe stage

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: <DB7PR04MB46188AAE07C0B032301FE474E65D0@DB7PR04MB4618.eurprd04.prod.outlook.com>
Date:   Wed, 4 Dec 2019 09:58:09 +0000
From:   Joakim Zhang <qiangqing.zhang@....com>
To:     Marc Kleine-Budde <mkl@...gutronix.de>,
        "sean@...nix.com" <sean@...nix.com>,
        "linux-can@...r.kernel.org" <linux-can@...r.kernel.org>
CC:     dl-linux-imx <linux-imx@....com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: RE: [PATCH V2 2/4] can: flexcan: try to exit stop mode during probe
 stage


> -----Original Message-----
> From: Marc Kleine-Budde <mkl@...gutronix.de>
> Sent: 2019年12月4日 16:45
> To: Joakim Zhang <qiangqing.zhang@....com>; sean@...nix.com;
> linux-can@...r.kernel.org
> Cc: dl-linux-imx <linux-imx@....com>; netdev@...r.kernel.org
> Subject: Re: [PATCH V2 2/4] can: flexcan: try to exit stop mode during probe
> stage
> 
> On 12/4/19 3:22 AM, Joakim Zhang wrote:
> >
> >> -----Original Message-----
> >> From: Marc Kleine-Budde <mkl@...gutronix.de>
> >> Sent: 2019年12月4日 2:15
> >> To: Joakim Zhang <qiangqing.zhang@....com>; sean@...nix.com;
> >> linux-can@...r.kernel.org
> >> Cc: dl-linux-imx <linux-imx@....com>; netdev@...r.kernel.org
> >> Subject: Re: [PATCH V2 2/4] can: flexcan: try to exit stop mode
> >> during probe stage
> >>
> >> On 11/27/19 6:56 AM, Joakim Zhang wrote:
> >>> CAN controller could be stucked in stop mode once it enters stop
> >>> mode
> >>                           ^^^^^^^ stuck
> >>> when suspend, and then it fails to exit stop mode when resume.
> >>
> >> How can this happen?
> >
> > I am also confused how can this happen, as I asked Sean, only CAN
> > enter stop mode when suspend, then system hang,
> How do you recover the system when suspended?
RTC wakeup or TTY wakeup.

> > it could let CAN
> > stuck in stop mode. However, Sean said this indeed happen at his side,
> > @sean@...nix.com, could you explain how this happen in details?
> That would be good.
>
> >>> Only code reset can get CAN out of stop mode,
> >>
> >> What is "code reset"?
> >
> > As I know, "code reset" is to press the POWER KEY from the board. At
> > my side, reboot command from OS also can get CAN out of stop mode.
> Do you mean "cold reset", also known as Power-On-Reset, POR or power
> cycle?
Should be Power-On-Reset.

> What does pressing the POWER KEY do? A power cycle of the system or
> toggling the reset line of the imx?
I think it toggles the reset line of imx. I am so sorry that I am not familiar with system reset :(.
 
> We need to describe in detail, as not everyone has the same board as you, and
> these boards might not even have a power key :)
Yes.

> > Below is experiment I did:
> > 	Firstly, do a hacking to let CAN stuck into stop mode, then:
> 
> You mean you put the CAN into stop mode without keeping track in the CAN
> driver that the CAN-IP is in stop mode, e.g. by hacking the driver.
Yes, you can add flexcan_enter_stop_mode() at the last of driver probe. After probe, CAN has been stuck in stop mode.
Or you can enable CAN wakeup, then comment out flexcan_exit_stop_mode() in flexcan_resume(), do suspend then wakeup system, CAN has been stuck in stop mode.

> Then you try several methods to recover:
> 
> > 	(1) press power on/off key, get CAN out of stop mode;
> > 	(2) reboot command from console, get CAN out of stop mode;
> > 	(3) unbind/bind driver, cannot get CAN out of stop mode;
> > 	(4) remod/insmod module, cannot get CAN out of stop mode;
> 
> (2) resets the complete imx, including the CAN-IP core, (1) probably, too.
Yes, since stop mode enter/exit request at a chip level, need reset completely, such as a "code reset", would get CAN out stop mode.
"Soft reset" cannot get CAN out of stop mode.

> (3) and (4) fail to recover the CAN core, as the IP core is still powered off by
> some upstream component. So the question why this happens in the first place
> is IMHO as important as trying to wake up the core. I think if we discover this
> situation (CAN Core is in stop-mode in probe) we should print a warning
> message, but try to recover.
We really need figure out why CAN could be stuck in stop mode. As I know, enter stop mode in flexcan_suspend(), and then exit stop mode in flexcan_resume(), it could be impossible.
Hope Sean can explain it in details, then we can discuss how to fix it more reasonable.

Thanks Marc.

Best Regards,
Joakim Zhang
> >>> so add stop mode remove request during probe stage for other
> >>> methods(soft reset from chip level, unbind/bind driver, etc) to let
> >>         ^^^ please add a space
> >>> CAN active again.
> >>
> >> Can you rephrase the sentence after "so add stop mode remove request
> >> during probe stage". I'm not completely sure what you want to tell.
> >
> > Sure.
> 
> tnx,
> Marc
> 
> --
> Pengutronix e.K.                 | Marc Kleine-Budde           |
> Embedded Linux                   | https://www.pengutronix.de  |
> Vertretung West/Dortmund         | Phone: +49-231-2826-924     |
> Amtsgericht Hildesheim, HRA 2686 | Fax:   +49-5121-206917-5555 |