Re: [Xen-devel] [PATCH v2 14/16] Xen: EFI: Parse DT parameters for Xen specific UEFI

[Mark Rutland <mark.rutland-5wv7dgnIgG8@public.gmane.org>]

On Tue, Jan 19, 2016 at 09:19:05PM +0800, Shannon Zhao wrote:
> 
> 
> On 2016/1/18 23:41, Stefano Stabellini wrote:
> >CC'ing Matt Fleming
> >
> >On Fri, 15 Jan 2016, Shannon Zhao wrote:
> >>From: Shannon Zhao <shannon.zhao-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
> >>@@ -520,15 +531,28 @@ static int __init fdt_find_uefi_params(unsigned long node, const char *uname,
> >>  				       int depth, void *data)
> >>  {
> >>  	struct param_info *info = data;
> >>+	struct params *dt_params;
> >>+	unsigned int size;
> >>  	const void *prop;
> >>  	void *dest;
> >>  	u64 val;
> >>  	int i, len;
> >>
> >>-	if (depth != 1 || strcmp(uname, "chosen") != 0)
> >>-		return 0;
> >>+	if (efi_enabled(EFI_PARAVIRT)) {
> >>+		if (depth != 2 || strcmp(uname, "uefi") != 0)
> >
> >You are already introducing this check in the previous patch when
> >setting EFI_PARAVIRT, why do this again now?  But if we need to do this
> >check again, then, like Mark suggested, it should be done against the
> >full path.
> >
> This check just wants to confirm that the current node is the "uefi"
> node and we can parse it with xen_fdt_params now.

There is no single "uefi" node as many nodes can share that name. There
is at most a single, /hypervisor/uefi node, as that is qualified by a
full path.

Checking the leaf node name, as above, is insufficient. For example, the
below will be accepted:

* /chosen/uefi
* /foo/uefi
* /not-a-hypervisor/uefi

Any of these could exist in addition to a /hypervisor/uefi node, and
could appear ealier or later in the DTB.

There may be reasons to add such nodes in future, and regardless we
should not read properties from an invalid/wrong node.

Thanks,
Mark.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html