netdev - Re: network driver takedown at reboof -f ?

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: <0893c6fa-84f7-887d-784b-f27abf8868c6@gmail.com>
Date:   Thu, 14 Jul 2022 10:26:09 -0700
From:   Florian Fainelli <f.fainelli@...il.com>
To:     Joakim Tjernlund <Joakim.Tjernlund@...inera.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: network driver takedown at reboof -f ?

On 7/14/22 10:16, Joakim Tjernlund wrote:
> On Thu, 2022-07-14 at 07:58 -0700, Florian Fainelli wrote:
>>
>> On 7/14/2022 7:21 AM, Joakim Tjernlund wrote:
>>> Doing a fast reboot -f I notice that the ethernet I/Fs are NOT shutdown/stopped.
>>> Is this expected? I sort of expected kernel to do ifconfig down automatically.
>>>
>>> Is there some function in netdev I can hook into to make reboot shutdown my eth I/Fs?
>>
>> If you want that to happen you typically have to implement a ->shutdown
>> callback in your network driver registered via platform/pci/bus, if
>> nothing else to turn off all DMAs and prevent, e.g.: a kexec'd kernel to
>> be corrupted by a wild DMA engine still running.
>>
>> There is no generic provision in the network stack to deal with those cases.
>>
>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgit.kernel.org%2Fpub%2Fscm%2Flinux%2Fkernel%2Fgit%2Ftorvalds%2Flinux.git%2Fcommit%2F%3Fid%3Dd9f45ab9e671166004b75427f10389e1f70cfc30&amp;data=05%7C01%7CJoakim.Tjernlund%40infinera.com%7Cf107ea4c5731411a813d08da65a96a4d%7C285643de5f5b4b03a1530ae2dc8aaf77%7C1%7C0%7C637934075577842203%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&amp;sdata=QerJo10uvNJW9aI4PaREvoWsZKAzn8SGx6HAP1Zr%2Bt0%3D&amp;reserved=0
> 
> Exactly what I need, thank you :)
> 
> Trying to add a shutdown I have a problem getting at priv ptr:
> static void ccip_shutdown(struct platform_device *pdev)
> {
> 	struct device *dev = &pdev->dev;
> 	struct net_device *ndev = (void*)dev->parent;
> 	struct ccip_priv *priv = netdev_priv(ndev);;
> 
> Above does not work, my probe has:
> 	ndev = alloc_netdev(sizeof(struct ccip_priv), if_name,
> 			    NET_NAME_UNKNOWN, ccip_init_dev);
> 	priv = netdev_priv(ndev);
>          SET_NETDEV_DEV(ndev, &pdev->dev);

How about platform_set_drvdata(&pdev->dev, priv) in probe, and in remove:

struct ccip_priv *priv = platform_get_drvdata(&pdev->dev);

does that work better?
-- 
Florian