[Libguestfs] Re: [nbdkit PATCH v2 1/2] server: Fix off-by-one for maximum block_status length [CVE-XXX]
On Wed, Apr 23, 2025 at 10:19:53PM +0100, Richard W.M. Jones wrote:
>
> The problem can only be triggered by a client request for 2**32-1
> bytes; anything smaller is immune. The problam also does not occur
s/problam/problem/
(and seeing this made me think to run codespell and fix a number of
other typos)
> +cat eval-extents.out
> +cat <<EOF | diff -u - eval-extents.out
> +[4294967294, 1, 1073741824, 2]
> +[4294967295, 1]
> +[4294967295, 1]
> +EOF
^^ I really need to use this 'diff -' trick in a few other places ...
Useless use of cat. It can be further simplified to:
diff -u - eval-extents.out <<EOF
expected...
EOF
Also, I've noticed that we are inconsistent on whether we use 'diff
expected actual' or 'diff actual expected'. Personally, I find
'expected actual' to produce nicer output (if expected is wrong
because of a false positive, such as due to an intentional change in
an error message, the output shows you how to patch the expected so
that re-running will work the new desired state).
Reviewed-by: Richard W.M. Jones <rjones(a)redhat.com>
Thanks; I'll include that before pushing. I'm still awaiting
resolution on whether this gets a CVE assigned.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc.
Virtualization: qemu.org | libguestfs.org