Re: [PULL v2 36/38] hw/block/nvme: support namespace attachment command

[Peter Maydell]

On Tue, 9 Mar 2021 at 11:46, Klaus Jensen <its@irrelevant.dk> wrote:
>
> From: Minwoo Im <minwoo.im.dev@gmail.com>
>
> This patch supports Namespace Attachment command for the pre-defined
> nvme-ns device nodes.  Of course, attach/detach namespace should only be
> supported in case 'subsys' is given.  This is because if we detach a
> namespace from a controller, somebody needs to manage the detached, but
> allocated namespace in the NVMe subsystem.
>
> As command effect for the namespace attachment command is registered,
> the host will be notified that namespace inventory is changed so that
> host will rescan the namespace inventory after this command.  For
> example, kernel driver manages this command effect via passthru IOCTL.

> diff --git a/hw/block/nvme.h b/hw/block/nvme.h
> index 85a7b5a14f4e..1287bc2cd17a 100644
> --- a/hw/block/nvme.h
> +++ b/hw/block/nvme.h
> @@ -235,6 +235,11 @@ static inline void nvme_ns_attach(NvmeCtrl *n, 
> NvmeNamespace *ns)
>      n->namespaces[nvme_nsid(ns) - 1] = ns;
>  }
>
> +static inline void nvme_ns_detach(NvmeCtrl *n, NvmeNamespace *ns)
> +{
> +    n->namespaces[nvme_nsid(ns) - 1] = NULL;
> +}

Hi; Coverity complains about a possible array overflow both here
in nvme_ns_detach() (CID 1450757) and in nvme_ns_attach() (CID 1450758):
nvme_nsid() can return -1, but this code does not check for that.

If these functions both assume that their ns argument is non-NULL,
then adding an "assert(ns)" would probably placate Coverity and also
would mean that any bugs elsewhere resulting in accidentally passing
a NULL pointer would result in a clean assertion failure rather than
memory corruption. (Or you could directly assert that the array index
is in-bounds, I guess.)

thanks
-- PMM