When compiled with PREEMPT_RT, spinlock_t is sleepable. While I did not observe this causing any lockups on my system, it did cause warnings to be emitted when compiled with lock debugging. This series fixes some instances where spinlock_t is used in non-sleepable contexts, though it almost certainly does not find every such instance. An example of the warning raised by lockdep: ============================= [BUG: Invalid wait context ] 6.5.0-rc1+ #16 Not tainted ----------------------------- swapper/0/1 is trying to lock: ffffa05a014d64c0 (&channel→sched_lock) {...}-{3:3}, at: vmbus_isr+0x179/0x320 other info that might help us debug this: context-{2:2} 2 locks held by swapper/0/1: #0: ffffffff909a9c70 (misc_mtx){+.+.}-{4:4}, at: misc_register+0x34/0x180 #1: ffffffff9079b4c8 (rcu_read_lock) {...}-{1:3}, at: rcu_lock_acquire+0x0/0x40 stack backtrace: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 6.5.0-rc1+ #16 Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS Hyper-V UEFI Release v4.1 05/09/2022 Signed-off-by: Mitchell Levy <levymitchell0@xxxxxxxxx> --- Mitchell Levy (2): Use raw_spinlock_t for vmbus_channel.sched_lock Use raw_spinlock_t in vmbus_requestor drivers/hv/channel.c | 6 +++--- drivers/hv/channel_mgmt.c | 2 +- drivers/hv/vmbus_drv.c | 4 ++-- drivers/pci/controller/pci-hyperv.c | 6 +++--- include/linux/hyperv.h | 8 ++++---- 5 files changed, 13 insertions(+), 13 deletions(-) --- base-commit: 14f9643dc90adea074a0ffb7a17d337eafc6a5cc change-id: 20230807-b4-rt_preempt-fix-35a65c90c6c9 Best regards, -- Mitchell Levy <levymitchell0@xxxxxxxxx>