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  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:   Thu, 26 Nov 2020 16:12:42 +0100
From:   Benjamin Block <bblock@...ux.ibm.com>
To:     Qinglang Miao <miaoqinglang@...wei.com>
Cc:     Cornelia Huck <cohuck@...hat.com>,
        Steffen Maier <maier@...ux.ibm.com>,
        Heiko Carstens <hca@...ux.ibm.com>,
        Vasily Gorbik <gor@...ux.ibm.com>,
        Christian Borntraeger <borntraeger@...ibm.com>,
        linux-s390@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-scsi@...r.kernel.org
Subject: Re: [PATCH] scsi: zfcp: fix use-after-free in zfcp_unit_remove

On Thu, Nov 26, 2020 at 08:07:32PM +0800, Qinglang Miao wrote:
> 在 2020/11/26 17:42, Benjamin Block 写道:
> > On Thu, Nov 26, 2020 at 09:13:53AM +0100, Cornelia Huck wrote:
> > > On Thu, 26 Nov 2020 09:27:41 +0800
> > > Qinglang Miao <miaoqinglang@...wei.com> wrote:
> > > > 在 2020/11/26 1:06, Benjamin Block 写道:
> > > > > On Fri, Nov 20, 2020 at 03:48:54PM +0800, Qinglang Miao wrote:
....
> > Let's go by example. If we assume the reference count of `unit->dev` is
> > R, and the function starts with R = 1 (otherwise the deivce would've
> > been freed already), we get:
> > 
> >      int zfcp_unit_remove(struct zfcp_port *port, u64 fcp_lun)
> >      {
> >      	struct zfcp_unit *unit;
> >      	struct scsi_device *sdev;
> >      	write_lock_irq(&port->unit_list_lock);
> > // unit->dev (R = 1)
> >      	unit = _zfcp_unit_find(port, fcp_lun);
> > // get_device(&unit->dev)
> > // unit->dev (R = 2)
> >      	if (unit)
> >      		list_del(&unit->list);
> >      	write_unlock_irq(&port->unit_list_lock);
> >      	if (!unit)
> >      		return -EINVAL;
> >      	sdev = zfcp_unit_sdev(unit);
> >      	if (sdev) {
> >      		scsi_remove_device(sdev);
> >      		scsi_device_put(sdev);
> >      	}
> > // unit->dev (R = 2)
> >      	put_device(&unit->dev);
> > // unit->dev (R = 1)
> >      	device_unregister(&unit->dev);
> > // unit->dev (R = 0)
> >      	return 0;
> >      }
> > 
> > If we now apply this patch, we'd end up with R = 1 after
> > `device_unregister()`, and the device would not be properly removed.
> > 
> > If you still think that's wrong, then you'll need to better explain why.
> > 
> Hi Banjamin and Cornelia,
> 
> Your replies make me reliaze that I've been holding a mistake understanding
> of put_device() as well as reference count.
> 
> Thanks for you two's patient explanation !!
> 
> BTW, should I send a v2 on these two patches to move the position of
> put_device()?

Feel free to do so.

I think having the `put_device()` call after `device_unregister()` in
both `zfcp_unit_remove()` and `zfcp_sysfs_port_remove_store()` is more
natural, because it ought to be the last time we touch the object in
both functions.


-- 
Best Regards, Benjamin Block  / Linux on IBM Z Kernel Development / IBM Systems
IBM Deutschland Research & Development GmbH    /    https://www.ibm.com/privacy
Vorsitz. AufsR.: Gregor Pillen         /        Geschäftsführung: Dirk Wittkopp
Sitz der Gesellschaft: Böblingen / Registergericht: AmtsG Stuttgart, HRB 243294

Powered by blists - more mailing lists