On Thu, Aug 13, 2020 at 11:17:11PM -0400, Alex Deucher wrote: > The non-atomic modesetting code in amdgpu now spews warnings[1] with > this patch applied. I haven't really paged the legacy locking stuff > into my head in quite a while. Thoughts on how to proceed? Uh that mess. So the problem (and it's really the same for atomic as for legacy) is that we have shared state between connector probe and modeset code. Generally that's solved by just copying stuff (like drm_connector->display_info) and hoping we don't race badly. All the probe state is protected by dev->mode_config.mutex, while all the modeset stuff is protected by the various drm_modeset_lock. I think the simplest duct-tape is to reapply the old locking scheme to the new macro, but only for non-atomic drivers. I'll type up something. But I'll go on vacations for 2 weeks, so if it doesn't work out you have to respin (and also merge). Cheers, Daniel > > Thanks, > > Alex > > commit 9bcaa3fe58ab7559e71df798bcff6e0795158695 > Author: Michal Orzel <michalorzel.eng@xxxxxxxxx> > Date: Tue Apr 28 19:10:04 2020 +0200 > > drm: Replace drm_modeset_lock/unlock_all with DRM_MODESET_LOCK_ALL_* helpers > > As suggested by the TODO list for the kernel DRM subsystem, replace > the deprecated functions that take/drop modeset locks with new helpers. > > Signed-off-by: Michal Orzel <michalorzel.eng@xxxxxxxxx> > Signed-off-by: Daniel Vetter <daniel.vetter@xxxxxxxx> > Link: https://patchwork.freedesktop.org/patch/msgid/1588093804-30446-1-git-send-email-michalorzel.eng@xxxxxxxxx > > > [1] https://gitlab.freedesktop.org/drm/amd/-/issues/1224 > _______________________________________________ > dri-devel mailing list > dri-devel@xxxxxxxxxxxxxxxxxxxxx > https://lists.freedesktop.org/mailman/listinfo/dri-devel -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
