On Sun 12-05-19 01:48:29, Qian Cai wrote:
> The linux-next commit ("x86, numa: always initialize all possible
> nodes") introduced a crash below during boot for systems with a
> memory-less node. This is due to CPUs that get onlined during SMP boot,
> but that onlining triggers a page fault in bus_add_device() during
> device registration:
>
> error = sysfs_create_link(&bus->p->devices_kset->kobj,
>
> bus->p is NULL. That "p" is the subsys_private struct, and it should
> have been set in,
>
> postcore_initcall(register_node_type);
>
> but that happens in do_basic_setup() after smp_init().
>
> The old code had set this node online via alloc_node_data(), so when it
> came time to do_cpu_up() -> try_online_node(), the node was already up
> and nothing happened.
>
> Now, it attempts to online the node, which registers the node with
> sysfs, but that can't happen before the 'node' subsystem is registered.
>
> Since kernel_init() is running by a kernel thread that is in
> SYSTEM_SCHEDULING state, fixed this by skipping registering with sysfs
> during the early boot in __try_online_node().
Relying on SYSTEM_SCHEDULING looks really hackish. Why cannot we simply
drop try_online_node from do_cpu_up? Your v2 remark below suggests that
we need to call node_set_online because something later on depends on
that. Btw. why do we even allocate a pgdat from this path? This looks
really messy.
> Call Trace:
> device_add+0x43e/0x690
> device_register+0x107/0x110
> __register_one_node+0x72/0x150
> __try_online_node+0x8f/0xd0
> try_online_node+0x2b/0x50
> do_cpu_up+0x46/0xf0
> cpu_up+0x13/0x20
> smp_init+0x6e/0xd0
> kernel_init_freeable+0xe5/0x21f
> kernel_init+0xf/0x180
> ret_from_fork+0x1f/0x30
>
> Reported-by: Barret Rhoden <brho@xxxxxxxxxx>
> Signed-off-by: Qian Cai <cai@xxxxxx>
> ---
>
> v2: Set the node online as it have CPUs. Otherwise, those memory-less nodes will
> end up being not in sysfs i.e., /sys/devices/system/node/.
>
> mm/memory_hotplug.c | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index b236069ff0d8..6eb2331fa826 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -1037,6 +1037,18 @@ static int __try_online_node(int nid, u64 start, bool set_node_online)
> if (node_online(nid))
> return 0;
>
> + /*
> + * Here is called by cpu_up() to online a node without memory from
> + * kernel_init() which guarantees that "set_node_online" is true which
> + * will set the node online as it have CPUs but not ready to call
> + * register_one_node() as "node_subsys" has not been initialized
> + * properly yet.
> + */
> + if (system_state == SYSTEM_SCHEDULING) {
> + node_set_online(nid);
> + return 0;
> + }
> +
> pgdat = hotadd_new_pgdat(nid, start);
> if (!pgdat) {
> pr_err("Cannot online node %d due to NULL pgdat\n", nid);
> --
> 2.20.1 (Apple Git-117)
--
Michal Hocko
SUSE Labs
