Re: [PATCH for-4.19(?)] xen/arm: bootfdt: Fix device tree memory node probing

[Stefano Stabellini <sstabellini@xxxxxxxxxx>]

On Wed, 26 Jun 2024, Julien Grall wrote:
> Hi Michal,
> 
> On 26/06/2024 09:04, Michal Orzel wrote:
> > Memory node probing is done as part of early_scan_node() that is called
> > for each node with depth >= 1 (root node is at depth 0). According to
> > Devicetree Specification v0.4, chapter 3.4, /memory node can only exists
> > as a top level node. However, Xen incorrectly considers all the nodes with
> > unit node name "memory" as RAM. This buggy behavior can result in a
> > failure if there are other nodes in the device tree (at depth >= 2) with
> > "memory" as unit node name. An example can be a "memory@xxx" node under
> > /reserved-memory. Fix it by introducing device_tree_is_memory_node() to
> > perform all the required checks to assess if a node is a proper /memory
> > node.
> > 
> > Fixes: 3e99c95ba1c8 ("arm, device tree: parse the DTB for RAM location and
> > size")
> > Signed-off-by: Michal Orzel <michal.orzel@xxxxxxx>
> > ---
> > 4.19: This patch is fixing a possible early boot Xen failure (before main
> > console is initialized). In my case it results in a warning "Shattering
> > superpage is not supported" and panic "Unable to setup the directmap
> > mappings".
> > 
> > If this is too late for this patch to go in, we can backport it after the
> > tree
> > re-opens.
> 
> The code looks correct to me, but I am not sure about merging it to 4.19.
> 
> This is not a new bug (in fact has been there since pretty much Xen on Arm was
> created) and we haven't seen any report until today. So in some way it would
> be best to merge it after 4.19 so it can get more testing.

First it was found on a new board, but then the issue appeared also on
an old board (the Ultrascale+). I think the reason is that a
reserved-memory node was added triggering the bug.


> In the other hand, I guess this will block you. Is this a new platform? Is it
> available?

Yes the platform is available so I would be more concerned about Xen
4.19 not booting on newer Ultrascale+ device trees. That said, we can
also backport the fix later to staging-4.19. I'll leave the decision to
you.