genirq: Move prio assignment into the newly created thread
authorThomas Gleixner <tglx@linutronix.de>
Tue, 10 Nov 2020 11:38:48 +0000 (12:38 +0100)
committerPeter Zijlstra <peterz@infradead.org>
Fri, 17 Sep 2021 13:08:22 +0000 (15:08 +0200)
commite739f98b4b11337a4e3865364b8922a9e5ad32b6
tree2ee755daa0014b41a0b33fc51121fb426dca8f4f
parente1a6af4b000c39148ab407362fcce3ab63b186f2
genirq: Move prio assignment into the newly created thread

With enabled threaded interrupts the nouveau driver reported the
following:

| Chain exists of:
|   &mm->mmap_lock#2 --> &device->mutex --> &cpuset_rwsem
|
|  Possible unsafe locking scenario:
|
|        CPU0                    CPU1
|        ----                    ----
|   lock(&cpuset_rwsem);
|                                lock(&device->mutex);
|                                lock(&cpuset_rwsem);
|   lock(&mm->mmap_lock#2);

The device->mutex is nvkm_device::mutex.

Unblocking the lockchain at `cpuset_rwsem' is probably the easiest
thing to do.  Move the priority assignment to the start of the newly
created thread.

Fixes: 710da3c8ea7df ("sched/core: Prevent race condition between cpuset and __sched_setscheduler()")
Reported-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
[bigeasy: Patch description]
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/a23a826af7c108ea5651e73b8fbae5e653f16e86.camel@gmx.de
kernel/irq/manage.c