On 31.01.22 07:23, Oscar Salvador wrote:

On Fri, Jan 28, 2022 at 04:15:40PM +0100, David Hildenbrand wrote:

quoted

... and call node_dev_init() after memory_dev_init() from driver_init(),
so before any of the existing arch/subsys calls. All online nodes should
be known at that point.

This is in line with memory_dev_init(), which initializes the memory
device subsystem and creates all memory block devices.

Similar to memory_dev_init(), panic() if anything goes wrong, we don't
want to continue with such basic initialization errors.

The important part is that node_dev_init() gets called after
memory_dev_init() and after cpu_dev_init(), but before any of the
relevant archs call register_cpu() to register the new cpu device under
the node device. The latter should be the case for the current users
of topology_init().

Hi Oscar,

So, before this change we had something like this:

do_basic_setup
 driver_init
  memory_dev_init
 do_init_calls
  ...
   topology_init
    register_nodes/register_one_node

And after the patch all happens in driver_init()

driver_init
 memory_dev_init
 node_dev_init

I guess this is fine as we do not have any ordering problems (aka: none
of the functions we used to call before expect the nodes not to be
there for some weird reason).

So, no functional change, right?

Right, and the idea is that the online state of nodes (+ node/zone
ranges) already has to be known at that point in time, because
otherwise, we'd be in bigger trouble.

Thanks!


-- 
Thanks,

David / dhildenb