linux-kernel - Re: [PATCH v2] mtd: spi-nor: keep lock bits if they are non-volatile

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: <1658390.DVVSh22Ze7@192.168.0.113>
Date:   Wed, 22 Jan 2020 06:58:38 +0000
From:   <Tudor.Ambarus@...rochip.com>
To:     <michael@...le.cc>
CC:     <vigneshr@...com>, <linux-mtd@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>, <richard@....at>,
        <boris.brezillon@...labora.com>, <miquel.raynal@...tlin.com>,
        <marex@...x.de>
Subject: Re: [PATCH v2] mtd: spi-nor: keep lock bits if they are non-volatile

Michael,

To be more explicit:

On Tuesday, January 21, 2020 8:53:20 PM EET Tudor Ambarus wrote:
> Yes but that was the whole idea of this patch. So if I get you correct
> 
> > it is
> > not possible to change that even if:
> > 
> > (1) it was never intended that way. Eg. the original patch(es) were
> > about
> > removing the volatile write protection (which makes perfectly sense,
> > even
> > during probe time) to be able to write to the flash. But it was never
> > intended
> > to disable the non-volatile write protection.

Even if this is true, we can't break backward compat.

> > 
> > (2) it might be even harmful. It is still an open question wether the
> > write
> > to the non-volatile bits (even if it is the same value) might wear them
> > out.
> > Unfortunately our FAE didn't answered yet..
> > 

We'll think about this when we know for sure.

> > (3) it makes the write protection utterly useless, because if you lock
> > the
> > flash it will be automatically unlocked after the next reboot. Even
> > worse, the
> > user likely won't notice it.

Even if this is true, we can't break backward compat.

> 
> Breaking backward compatibility and keeping the locking state of the spi-nor
> flashes at probe is a no-go, because there might be user space apps that
> expect that all the spi-nor flashes are by default unlocked. The unlocking
> of the flash at probe time was introduced 12 years ago, we definitely can't
> change this now.

Kconfig option or module param will fix this without breaking backward compat, 
we should focus on this direction.

Cheers,
ta