Martin Schwidefsky fb7d7518b0 s390/numa: move initial setup of node_to_cpumask_map
The numa_init_early initcall sets the node_to_cpumask_map[0] to the
full cpu_possible_mask. Unfortunately this early_initcall is too late,
the NUMA setup for numa=emu is done even earlier. The order of calls
is numa_setup() -> emu_update_cpu_topology(), then the early_initcalls(),
followed by sched_init_domains().

Starting with git commit 051f3ca02e46432c0965e8948f00c07d8a2f09c0
"sched/topology: Introduce NUMA identity node sched domain"
the incorrect node_to_cpumask_map[0] really screws up the domain
setup and the kernel panics with the follow oops:

Cc: <stable@vger.kernel.org> # v4.15+
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2018-08-01 07:48:33 +02:00
..
2018-04-27 06:51:31 +02:00
2018-07-04 08:35:58 +02:00
2018-04-16 10:29:34 +02:00
2018-04-16 09:10:22 +02:00
2018-07-25 09:07:20 +02:00