Re: CVE-2021-47188: scsi: ufs: core: Improve SCSI abort handling

From: Greg Kroah-Hartman
Date: Tue Aug 06 2024 - 02:41:45 EST


On Mon, Aug 05, 2024 at 07:48:13PM +0300, Cengiz Can wrote:
> Hello,
>
> I'm trying to figure out the security impact here:
>
> > That warning is triggered by the following statement:
> >
> > WARN_ON(lrbp->cmd);
>
> This is just a fix to silence a warning. How is this worthy of a CVE? What was
> the criteria here?
>
> If there are security implications of not nullifying `lrbp->cmd`, shouldn't they
> be noted in the CVE description?

CVE descriptions come directly from the kernel changelog text. If you
wish them to say something else, please submit a patch against that text
and we can apply that.

> If this just a fix to the warning, this CVE should be rejected.

If userspace can trigger a WARN_ON() then the machine can be rebooted if
panic-on-warn is enabled, and so it requires a CVE assignment.

thanks

greg k-h