stack smashing detected with 'nvme sanitize-log /dev/nvme0'

[Keith Busch]

On Mon, Aug 28, 2023 at 11:20:38AM +0200, Christoph Hellwig wrote:
> On Mon, Aug 21, 2023 at 09:11:38AM -0600, Keith Busch wrote:
> > I don't think we want to bounce to kernel memory for the device to
> > overwrite it. I suggest just change nvme-cli's stack allocated santize
> > log to a use page aligned and sized buffer.
> 
> That assumes it actually overwrites it in that case and doesn't just
> have a PRP parsing bug when there is not enough alignment.
> 
> We should be able to find out by enabling KASAN and then requiring the
> larger alignment before re-running the reproducer.

Good point. I assumed it was a simple buffer overrun regardless of the
starting offset, but bad PRP parsing sounds plausible.

If you don't want to enable a kernel with kasan, we can just align user
space buffers with padding at different offsets and see what happens.