On 2022. 08. 30. 17:21, Andrew Lunn wrote:

You are not taking a mutex, you are taking a spinlock. You can do that
in atomic context. Can you protect everything which needs protecting
with a spinlock? And avoid sleeping...

You are correct, the issue Marc experienced turns out to be caused by a 
`mutex_lock()` introduced in 6a4d7234ae9a3bb31181f348ade9bbdb55aeb5c5...

On 2022. 08. 31. 5:41, Richard Cochran wrote:
 > Just replace the mutex with a spinlock.

Will do. Disregard this patch and see "[PATCH] Use a spinlock to guard 
`fep->ptp_clk_on`" instead

Bence