arm64: irq: set the correct node for VMAP stack
authorHuang Shijie <shijie@os.amperecomputing.com>
Fri, 24 Nov 2023 03:15:13 +0000 (11:15 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Mon, 5 Feb 2024 20:14:14 +0000 (20:14 +0000)
commit40a5dce893f0ea46a208733984c90e6803bf2ed3
treea5b6af833fdb16950847b92c7c7d0816f63610bc
parent145febd85c3bcc5c74d87ef9a598fc7d9122d532
arm64: irq: set the correct node for VMAP stack

[ Upstream commit 75b5e0bf90bffaca4b1f19114065dc59f5cc161f ]

In current code, init_irq_stacks() will call cpu_to_node().
The cpu_to_node() depends on percpu "numa_node" which is initialized in:
     arch_call_rest_init() --> rest_init() -- kernel_init()
--> kernel_init_freeable() --> smp_prepare_cpus()

But init_irq_stacks() is called in init_IRQ() which is before
arch_call_rest_init().

So in init_irq_stacks(), the cpu_to_node() does not work, it
always return 0. In NUMA, it makes the node 1 cpu accesses the IRQ stack which
is in the node 0.

This patch fixes it by:
  1.) export the early_cpu_to_node(), and use it in the init_irq_stacks().
  2.) change init_irq_stacks() to __init function.

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Huang Shijie <shijie@os.amperecomputing.com>
Link: https://lore.kernel.org/r/20231124031513.81548-1-shijie@os.amperecomputing.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
arch/arm64/kernel/irq.c
drivers/base/arch_numa.c
include/asm-generic/numa.h