drm: Fix locking around accurate vblank counts.

- Make drm_crtc_accurate_vblank_count require the caller to hold the
event lock, rather than take it internally.

- Fix locking around drm_crtc_accurate_vblank_count and related
operations in amdgpu and nouveau interrupt handlers.

- Use drm_crtc_vblank_put_locked, not drm_crtc_vblank_put, when we
already hold the event lock.

PR kern/56557

drm: Must hold event_lock around wait on event_queue.

We must not have hit this path before if it never crashed!

drm: mutex_init/destroy and spin_lock_init/destroy audit

drm: Use drm_vblank_get_locked/unlocked.

drm: Merge vbl_lock into event_lock.

These aren't functionally very different, and most uses of vbl_lock
were inserted as local changes to make proper condition variables
anyway. Requiring both locks made cv_wait difficult because it only
unlocks and relocks one at a time, and this also led to annoying lock
order reversal.

drm: Eliminate hrtimer_destroy and teardown_timer.

These are non-Linux shims added to the Linux API to pair with
hrtimer_setup and setup_timer, but they really only serve to call
callout_destroy, which is not strictly necessary in the NetBSD API.
So although we lose a little bit of diagnostic help this way, we also
reduce diffs a bit and save some trouble that bit us last week.

drm: Simplify condvar conditionals in drm_vblank

drm: Fix lock order reversal in merge botch.

Add a note reminding myself what the lock order is.

drm: Restore drm_crtc_vblank_put_locked

drm: drm_update_vblank_count requires vbl_lock.

drm: Fix vblank locking.

Appropriate ifdef goop to make this file build for us.

Use the netbsd-ified versions.


Author: Maya Rashish <maya@NetBSD.org>

New seqlock code requires seqlock_destroy.

XXX missing cases in drm56

64-bit printf, include for do_div64, omit unused variable.

Merge linux-drm-v5-6-rc3

branches: 1.1.1;
Initial revision