On Tue, May 31, 2016 at 12:49:03PM -0400, Lyude wrote: > Unfortunately we've never really made use of creating/destroying connectors > on the fly like we have with MST, so 4.6 ended up showing a lot of various > bugs with hotplugging MST displays, booting with MST displays, etc. Most of > these bugs are very likely to panic the kernel, and a couple of them end up > even doing out of bounds memory accesses causing all sorts of other issues. > > The proper fix for these issues is Dave's connector lifetime patch series > in 4.7-rc1[1], however backporting those patches would be too big of a fix > to submit for stable. This patch series is a much smaller set of changes to > workaround this issue. > > As another note: the first two patches in this series are already upstream for > 4.7-rc1, however since we have the connector ref lifetime patches in 4.7-rc1 > they don't fix any kernel panics there, only a few inconsistencies in > i915/drm's code. They do however, fix kernel panics for 4.6 since connectors > getting destroyed can make dev->mode_config.num_connector have a different > value from fb_helper->connector_count. > > [1]: 0552f7651bc233e5407ab06ba97a9d7c25e19580 in master In case it's not clear: There's no way we're going to backport the connector refcounting stuff from 4.7, way too risky. But we still need some way to make current stable kernels happy, and the 2 small hacks (plus 2 backports) seem like an adequate solution. Hence acked from my side for applying to all supported stable kernels. Cheers, Daniel > > Lyude (4): > drm/i915/fbdev: Fix num_connector references in > intel_fb_initial_config() > drm/fb_helper: Fix references to dev->mode_config.num_connector > drm/i915: Discard previous atomic state on resume if connectors change > drm/atomic: Verify connector->funcs != NULL when clearing states > > drivers/gpu/drm/drm_atomic.c | 2 +- > drivers/gpu/drm/drm_fb_helper.c | 5 ++--- > drivers/gpu/drm/i915/intel_display.c | 12 ++++++++++++ > drivers/gpu/drm/i915/intel_fbdev.c | 6 +++--- > 4 files changed, 18 insertions(+), 7 deletions(-) > > -- > 2.5.5 > > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@xxxxxxxxxxxxxxxxxxxxx > https://lists.freedesktop.org/mailman/listinfo/intel-gfx -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch -- To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html