linux-kernel - Re: [PATCH RFC 2/2] mtd: ubi: add support for protecting critical volumes

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: <Zvr-hOM2DfveaqeC@makrotopia.org>
Date: Mon, 30 Sep 2024 20:39:48 +0100
From: Daniel Golle <daniel@...rotopia.org>
To: Richard Weinberger <richard@....at>
Cc: chengzhihao1 <chengzhihao1@...wei.com>,
	Miquel Raynal <miquel.raynal@...tlin.com>,
	Vignesh Raghavendra <vigneshr@...com>, robh <robh@...nel.org>,
	Krzysztof Kozlowski <krzk+dt@...nel.org>,
	Conor Dooley <conor+dt@...nel.org>, John Crispin <john@...ozen.org>,
	linux-mtd <linux-mtd@...ts.infradead.org>,
	devicetree <devicetree@...r.kernel.org>,
	linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH RFC 2/2] mtd: ubi: add support for protecting critical
 volumes

On Mon, Sep 30, 2024 at 08:43:40PM +0200, Richard Weinberger wrote:
> ----- Ursprüngliche Mail -----
> > Von: "chengzhihao1" <chengzhihao1@...wei.com>
> >>> Von: "Daniel Golle" <daniel@...rotopia.org>
> >>> Allow the boot firmware to define volumes which are critical for the
> >>> system to boot, such as the bootloader itself if stored inside a UBI
> >>> volume. Protect critical volumes by preventing the user from removing,
> >>> resizing or writing to them, and also prevent the UBI device from
> >>> being detached if a critical volume is present.
> >> 
> >> I agree with the doubts raised in patch 1/2, if userspace is so hostile
> >> to delete system partitions, there is little hope.
> >> But I'm still open for discussion.
> > 
> > Yes, I agree that it is meaningful to prevent user from operating
> > volumes accidently. How about doing that by some existing methods? Eg.
> > selinux(Design sepolicy for ioctl cmd).
> 
> Another thought, do we really need to enforce this in kernel space?
> Teaching ubi-tools to be super careful with some volumes is also an option.
> 
> like a ubirmvol ... --i-know-what-im-doing.

True, enforcement doesn't need to happen in kernel (though I think it's
nicer, but really just a matter of taste, I guess). ubi-tools would still
need to be able to recognize critical volumes somehow, and that could be
done by checking if the 'volume-is-critical' property is present in
/sys/class/ubi/ubi*_*/of_node/

If you prefer going down that road instead I will work on patches for
git.infradead.org/mtd-utils.git instead.