[PATCH V2] mm: fix NULL NODE_DATA dereference for memoryless nodes on boot

[PATCH V2] mm: fix NULL NODE_DATA dereference for memoryless nodes on boot

[Ming Lei]

Commit d49004c5f0c1 ("arch, mm: consolidate initialization of nodes,
zones and memory map") moved free_area_init() from setup_arch() to
mm_core_init_early(), which runs after setup_arch() returns.

This changed the ordering relative to init_cpu_to_node() on x86. Before
the commit, free_area_init() ran during paging_init() (called from
setup_arch()) *before* init_cpu_to_node(). After the commit, it runs
*after* init_cpu_to_node().

On machines with memoryless NUMA nodes (e.g., node 0 has CPUs but no
memory), this causes a NULL pointer dereference:

 1. numa_register_nodes() skips memoryless nodes: no alloc_node_data()
    and no node_set_online() for them.
 2. init_cpu_to_node() sets memoryless nodes online (they have CPUs)
    but does not allocate NODE_DATA.
 3. free_area_init() checks "if (!node_online(nid))" to decide whether
    to call alloc_offline_node_data(). Since the memoryless node is now
    online, the allocation is skipped, leaving NODE_DATA(nid) == NULL.
 4. The immediate "pgdat = NODE_DATA(nid)" dereferences NULL.

The crash happens before console_init(), so no output is visible without
earlyprintk. With earlyprintk enabled, the following panic is observed:

 BUG: unable to handle page fault for address: 000000000002a1e0
 Oops: Oops: 0000 [#1] SMP NOPTI
 RIP: 0010:free_area_init_node+0x3a/0x540
 Call Trace:
  <TASK>
  free_area_init+0x331/0x4e0
  start_kernel+0x69/0x4a0
  x86_64_start_reservations+0x24/0x30
  x86_64_start_kernel+0x125/0x130
  common_startup_64+0x13e/0x148
  </TASK>
 Kernel panic - not syncing: Attempted to kill the idle task!

Fix this by checking "if (!NODE_DATA(nid))" instead of
"if (!node_online(nid))". This directly tests whether the per-node data
structure needs to be allocated, regardless of the node's online status.
This change is also safe for non-x86 architectures as they all allocate
NODE_DATA for every node including memoryless ones, so the check simply
evaluates to false with no change in behavior.

Fixes: d49004c5f0c1 ("arch, mm: consolidate initialization of nodes, zones and memory map")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
---
V2:
	- add commit log for non-x86 arch
	- add comment for code change

 mm/mm_init.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/mm/mm_init.c b/mm/mm_init.c
index 61d983d23f55..df34797691bd 100644
--- a/mm/mm_init.c
+++ b/mm/mm_init.c
@@ -1896,7 +1896,11 @@ static void __init free_area_init(void)
 	for_each_node(nid) {
 		pg_data_t *pgdat;
 
-		if (!node_online(nid))
+		/*
+		 * If an architecture has not allocated node data for
+		 * this node, presume the node is memoryless or offline.
+		 */
+		if (!NODE_DATA(nid))
 			alloc_offline_node_data(nid);
 
 		pgdat = NODE_DATA(nid);
-- 
2.53.0

Re: [PATCH V2] mm: fix NULL NODE_DATA dereference for memoryless nodes on boot

[Mike Rapoport]

On Sun, Feb 22, 2026 at 07:57:02PM +0800, Ming Lei wrote:
> Commit d49004c5f0c1 ("arch, mm: consolidate initialization of nodes,
> zones and memory map") moved free_area_init() from setup_arch() to
> mm_core_init_early(), which runs after setup_arch() returns.
> 
> This changed the ordering relative to init_cpu_to_node() on x86. Before
> the commit, free_area_init() ran during paging_init() (called from
> setup_arch()) *before* init_cpu_to_node(). After the commit, it runs
> *after* init_cpu_to_node().
> 
> On machines with memoryless NUMA nodes (e.g., node 0 has CPUs but no
> memory), this causes a NULL pointer dereference:
> 
>  1. numa_register_nodes() skips memoryless nodes: no alloc_node_data()
>     and no node_set_online() for them.
>  2. init_cpu_to_node() sets memoryless nodes online (they have CPUs)
>     but does not allocate NODE_DATA.
>  3. free_area_init() checks "if (!node_online(nid))" to decide whether
>     to call alloc_offline_node_data(). Since the memoryless node is now
>     online, the allocation is skipped, leaving NODE_DATA(nid) == NULL.
>  4. The immediate "pgdat = NODE_DATA(nid)" dereferences NULL.
> 
> The crash happens before console_init(), so no output is visible without
> earlyprintk. With earlyprintk enabled, the following panic is observed:
> 
>  BUG: unable to handle page fault for address: 000000000002a1e0
>  Oops: Oops: 0000 [#1] SMP NOPTI
>  RIP: 0010:free_area_init_node+0x3a/0x540
>  Call Trace:
>   <TASK>
>   free_area_init+0x331/0x4e0
>   start_kernel+0x69/0x4a0
>   x86_64_start_reservations+0x24/0x30
>   x86_64_start_kernel+0x125/0x130
>   common_startup_64+0x13e/0x148
>   </TASK>
>  Kernel panic - not syncing: Attempted to kill the idle task!
> 
> Fix this by checking "if (!NODE_DATA(nid))" instead of
> "if (!node_online(nid))". This directly tests whether the per-node data
> structure needs to be allocated, regardless of the node's online status.
> This change is also safe for non-x86 architectures as they all allocate
> NODE_DATA for every node including memoryless ones, so the check simply
> evaluates to false with no change in behavior.

This kinda means that x86 does something odd, but that's a matter for
additional rework and audit of node allocations.
 
> Fixes: d49004c5f0c1 ("arch, mm: consolidate initialization of nodes, zones and memory map")
> Signed-off-by: Ming Lei <ming.lei@redhat.com>

Reviewed-by: Mike Rapoport (Microsoft) <rppt@kernel.org>

> ---
> V2:
> 	- add commit log for non-x86 arch
> 	- add comment for code change
> 
>  mm/mm_init.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index 61d983d23f55..df34797691bd 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -1896,7 +1896,11 @@ static void __init free_area_init(void)
>  	for_each_node(nid) {
>  		pg_data_t *pgdat;
>  
> -		if (!node_online(nid))
> +		/*
> +		 * If an architecture has not allocated node data for
> +		 * this node, presume the node is memoryless or offline.
> +		 */
> +		if (!NODE_DATA(nid))
>  			alloc_offline_node_data(nid);
>  
>  		pgdat = NODE_DATA(nid);
> -- 
> 2.53.0
> 

-- 
Sincerely yours,
Mike.