linux-kernel - RE: [char-misc-next v2 1/5] mei: set parent for char device

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]

Message-ID: <CY5PR11MB6366B3733FB8F09D7F66ABCDED56A@CY5PR11MB6366.namprd11.prod.outlook.com>
Date: Wed, 16 Jul 2025 08:09:13 +0000
From: "Usyskin, Alexander" <alexander.usyskin@...el.com>
To: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
CC: "Abliyev, Reuven" <reuven.abliyev@...el.com>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [char-misc-next v2 1/5] mei: set parent for char device

> Subject: Re: [char-misc-next v2 1/5] mei: set parent for char device
> 
> On Wed, Jul 16, 2025 at 07:44:28AM +0000, Usyskin, Alexander wrote:
> > > Subject: Re: [char-misc-next v2 1/5] mei: set parent for char device
> > >
> > > On Mon, Jun 30, 2025 at 01:02:12PM +0000, Usyskin, Alexander wrote:
> > > > > Subject: Re: [char-misc-next v2 1/5] mei: set parent for char device
> > > > >
> > > > > On Mon, Jun 30, 2025 at 12:19:38PM +0300, Alexander Usyskin
> wrote:
> > > > > > Connect char device to parent device to avoid
> > > > > > parent device unload while char device is
> > > > > > still held open by user-space.
> > > > >
> > > > > No, that's not what cdev_set_parent() does.
> > > > >
> > > > https://elixir.bootlin.com/linux/v6.16-rc3/source/fs/char_dev.c#L511
> > > > * cdev_set_parent() sets a parent kobject which will be referenced
> > > >  * appropriately so the parent is not freed before the cdev. This
> > > >  * should be called before cdev_add.
> > > >
> > > > This is what written there. Is my interpretation wrong?
> > >
> > > It has nothing to do with a char device being open or closed from
> > > userspace, it has to deal with properly placing the device in the
> > > kobject tree and the object references being correct.
> > >
> > > Surely the mei device had this set up before, right?  If not, how did
> > > sysfs look with these char devices with no real parent?
> > >
> >
> > Mei device never had this setup.
> 
> So what is the "parent" of a mei device today?
> 

PCI device in CSME case and auxiliary one for discrete graphics.

> > I've looked at other drivers - some of them do not have this setup.
> 
> Then they are broken, don't copy them :)

Yep, I'm trying to fix this mess.
I've found a way to use existing mei class device to anchor the mei_device.
With this we can use mei class device as parent for the cdev.
Pushed the series (v3), looking for reactions whatever this is the right way.

- - 
Thanks,
Sasha