netdev - Re: [PATCH] sh_eth: ensure pm_runtime cannot suspend the device during init

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: <2182534.rxjU11HcCL@avalon>
Date:	Mon, 17 Mar 2014 22:34:49 +0100
From:	Laurent Pinchart <laurent.pinchart@...asonboard.com>
To:	Sergei Shtylyov <sergei.shtylyov@...entembedded.com>
Cc:	Ben Dooks <ben.dooks@...ethink.co.uk>, linux-sh@...r.kernel.org,
	netdev@...r.kernel.org
Subject: Re: [PATCH] sh_eth: ensure pm_runtime cannot suspend the device during init

Hi Sergei,

On Tuesday 18 March 2014 01:30:26 Sergei Shtylyov wrote:
> On 03/17/2014 11:23 PM, Laurent Pinchart wrote:
> >> The pm_rumtime work queue is causing the device to be suspended during
> >> initialisation, thus the initialisation may not be able to access
> >> registers properly. As the code is called from a work queue, it is
> >> possible that this is not seen from certain configurations/builds due to
> >> the asynchronos nature of the code.
> >> 
> >> Use pm_runtime_get_sync() and pm_runtime_put_sync() to ensure that the
> >> pm system does not suspend it during the probe() call and remove the
> >> now unnecessary pm_runtime_resume() call.
> >> 
> >> This fixes the external abort that can cause /sbin/init or other such
> >> init processed to die.
> >> 
> >> Signed-off-by: Ben Dooks <ben.dooks@...ethink.co.uk>
> >> Tested-by: Geert Uytterhoeven <geert@...ux-m68k.org>
> >> ---
> >> 
> >>   drivers/net/ethernet/renesas/sh_eth.c | 3 ++-
> >>   1 file changed, 2 insertions(+), 1 deletion(-)
> >> 
> >> diff --git a/drivers/net/ethernet/renesas/sh_eth.c
> >> b/drivers/net/ethernet/renesas/sh_eth.c index 4f76b5e..f1cfd64 100644
> >> --- a/drivers/net/ethernet/renesas/sh_eth.c
> >> +++ b/drivers/net/ethernet/renesas/sh_eth.c
> >> @@ -2869,7 +2869,7 @@ static int sh_eth_drv_probe(struct platform_device
> >> *pdev) spin_lock_init(&mdp->lock);
> >> 
> >>   	mdp->pdev = pdev;
> >>   	pm_runtime_enable(&pdev->dev);
> >> 
> >> -	pm_runtime_resume(&pdev->dev);
> >> +	pm_runtime_get_sync(&pdev->dev);
> > 
> > Now that we've found out that the problem is caused by registering the
> > network device before registering the mdio bus, shouldn't the proper
> > solution be to register the network device last in the probe function ?
> 
> Unfortunately, it's not easy to do, as sh_mdio_init() uses net_device::dev.
> Probably could get rid of that use by partly reverting my managed device API
> patch since the only user seems to be devm_kzalloc() in that function, at
> least I hope so... still looking into this.

What about using pdev->dev instead ?

> BTW, quite many drivers have the same problem, doing different things after
> register_netdev() call, many of them probing MDIO as well.

Great, that means more opportunities to fix bugs :-D

-- 
Regards,

Laurent Pinchart

--
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