xen/arm: bootfdt: Fix device tree memory node probing
PATCH v2 1/1 xen/arm: bootfdt: Fix device tree memory node probing
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.
---
Changes in v2:
- improve the length check
--- xen/arch/arm/bootfdt.c | 28 +++++++++++++++++++++++++++- 1 file changed, 27 insertions(+), 1 deletion(-)
xen/arch/arm/bootfdt.c +27 -1
@@ -83,6 +83,32 @@ static bool __init device_tree_node_compatible(const void *fdt, int node,return false; } +/* +* Check if a node is a proper /memory node according to Devicetree +* Specification v0.4, chapter 3.4. +*/ +static bool __init device_tree_is_memory_node (const void *fdt,int node, +int depth) +{ +const char *type; +int len; + +if ( depth !=1 ) +return false; + +if ( !device_tree_node_matches (fdt, node,"memory" ) ) +return false; + + type =fdt_getprop (fdt, node,"device_type" , &len); +if ( !type ) +return false; + +if ( (len <=strlen ("memory" )) ||strcmp (type,"memory" ) ) +return false; + +return true; +} +void __init device_tree_get_reg (const __be32 **cell,uint32_t address_cells,uint32_t size_cells,paddr_t *start,paddr_t *size)
@@ -448,7 +474,7 @@ static int __init early_scan_node(const void *fdt,* populated. So we should skip the parsing. */ if ( !efi_enabled (EFI_BOOT ) && -device_tree_node_matches (fdt, node,"memory" ) ) +device_tree_is_memory_node (fdt, node, depth) ) rc =process_memory_node (fdt, node, name, depth, address_cells, size_cells,bootinfo_get_mem ());else if ( depth ==1 && !dt_node_cmp (name,"reserved-memory" ) )
Signature · 1 line
2.25.1
Hi,
Quoted 11 lines · Michal Orzel
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 introducingdevice_tree_is_memory_node()to perform all the required checks to assess if a node is a proper /memory node.
Quoted 4 lines
Fixes 3e99c95ba1c8 ("arm, device tree: parse the DTB for RAM location and size")Signed-off-by Michal Orzel <[email protected]>Reviewed-by Luca Fancellu <[email protected]>Tested-by Luca Fancellu <[email protected]>
Cheers,
Signature · 1 line
Julien Grall
Hi,
Quoted 14 lines · Julien Grall
Hi,
Quoted 11 lines · Michal Orzel
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 introducingdevice_tree_is_memory_node()to perform all the required checks to assess if a node is a proper /memory node.
Quoted 6 lines
Quoted 4 lines
Fixes 3e99c95ba1c8 ("arm, device tree: parse the DTB for RAM location and size")Signed-off-by Michal Orzel <[email protected]>Reviewed-by Luca Fancellu <[email protected]>Tested-by Luca Fancellu <[email protected]>Reviewed-by Julien Grall <[email protected]>
This is now committed.
Cheers,
Signature · 1 line
Julien Grall