Op 15-06-17 om 10:25 schreef Chris Wilson: > During cleanup we release the VMA of the previous framebuffer. This > requires taking a struct_mutex, and potential recursion when handling a > reset. A simple device here is to move that locking into its own work > and we can avoid blocking on it for the reset by waiting on the flip > completion (and not cleanup completion) instead. Ideally this reduces > the duration of a blocking KMS operation by offloading the cleanup. Once > again we dream of the legendary vblank worker. > > Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> > Cc: Maarten Lankhorst <maarten.lankhorst@xxxxxxxxxxxxxxx> > --- > drivers/gpu/drm/i915/intel_display.c | 53 +++++++++++++++++++++++++++++------- > 1 file changed, 43 insertions(+), 10 deletions(-) Will this still pass the kms_cursor_legacy tests? ~Maarten _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
