On Wed, Nov 08, 2017 at 02:28:23PM +0100, Daniel Vetter wrote:
On Wed, Nov 08, 2017 at 03:26:15PM +0200, Ville Syrjälä wrote:
On Wed, Nov 08, 2017 at 02:11:46PM +0100, Daniel Vetter wrote:
On Wed, Nov 08, 2017 at 03:04:58PM +0200, Ville Syrjälä wrote:
On Wed, Nov 08, 2017 at 01:57:50PM +0100, Daniel Vetter wrote:
Two bits:
check actual atomic state, the legacy stuff can only be looked at from within the atomic_commit_tail function, since it's only protected by ordering and not by any locks.
Make sure we don't wreak the work an ongoing nonblocking commit is doing.
I still think we need to move the retraining to the hotplug work.
Why? One of the patch series Maarten mentioned says there's a deadlock with dp aux, but it seems to be all just fine when CI retrains.
This retraining here would be not because of the training failing in the commit but when we get a short pulse right so when the link was already up and running?
I guess the deadlock is possible only with MST, maybe. Currently MST has it's own copy of the retraining code without the locks.
At one point I started to rewrite the entire sink irq handling into a much nicer shape, also unifying the approach between MST and SST. IIRC I did still make the mistake of having some kind of higher level MST vs. SST split, but I think the proper solution is to combine the two almost entirely. I think we should even be using the ESI registers with SST for DPCD 1.2+. Currently we use ESI for MST and non-ESI for SST. Sadly I've not found the time to continue that work (same story with the "shutdown displays prior to rebooting to keep Dell MST monitors working" thing).
Yeah, merging sst and mst more definitely sounds like a good idea. I've also been toying with it since forever.
And even if there is a deadlock, it's there already, so not sure why we need to block this bugfix on a different bugfix. Which seems to be what you're suggesting here (but it's a bit sparse).
I guess what I'm really saying is that we shouldn't stop here.
Fully agreed. -Daniel
-Daniel
v2: We need the crtc lock too, because a plane update might change it without having to acquire the connection_mutex (Maarten). Use Maarten's changes for this locking, while keeping the logic that uses the connection->commit->hw_done signal for syncing with nonblocking commits.
Cc: Manasi Navare manasi.d.navare@intel.com Cc: Maarten Lankhorst maarten.lankhorst@linux.intel.com Cc: Ville Syrjälä ville.syrjala@linux.intel.com References: https://bugs.freedesktop.org/show_bug.cgi?id=103336 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99272 Signed-off-by: Daniel Vetter daniel.vetter@intel.com
drivers/gpu/drm/i915/intel_dp.c | 56 ++++++++++++++++++++++++++++++++++++----- 1 file changed, 50 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index d27c0145ac91..7cd7ab4fb431 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -4319,6 +4319,8 @@ static void intel_dp_check_link_status(struct intel_dp *intel_dp) { struct intel_encoder *intel_encoder = &dp_to_dig_port(intel_dp)->base;
I dont think we need intel_encoder, can we remove this?
Manasi
- struct drm_connector_state *conn_state =
intel_dp->attached_connector->base.state;@@ -4329,10 +4331,15 @@ intel_dp_check_link_status(struct intel_dp *intel_dp) return; }
- if (!intel_encoder->base.crtc)
- if (!conn_state->crtc) return;
- if (!to_intel_crtc(intel_encoder->base.crtc)->active)
WARN_ON(!drm_modeset_is_locked(&conn_state->crtc->mutex));
if (!conn_state->crtc->state->active)
return;if (!try_wait_for_completion(&conn_state->commit->hw_done)) return;
/*
@@ -4368,7 +4375,6 @@ intel_dp_check_link_status(struct intel_dp *intel_dp) static bool intel_dp_short_pulse(struct intel_dp *intel_dp) {
- struct drm_device *dev = intel_dp_to_dev(intel_dp); struct intel_encoder *intel_encoder = &dp_to_dig_port(intel_dp)->base; u8 sink_irq_vector = 0; u8 old_sink_count = intel_dp->sink_count;
@@ -4408,9 +4414,8 @@ intel_dp_short_pulse(struct intel_dp *intel_dp) DRM_DEBUG_DRIVER("CP or sink specific irq unhandled\n"); }
- drm_modeset_lock(&dev->mode_config.connection_mutex, NULL); intel_dp_check_link_status(intel_dp);
- drm_modeset_unlock(&dev->mode_config.connection_mutex);
- if (intel_dp->compliance.test_type == DP_TEST_LINK_TRAINING) { DRM_DEBUG_KMS("Link Training Compliance Test requested\n"); /* Send a Hotplug Uevent to userspace to start modeset */
@@ -4860,8 +4865,19 @@ intel_dp_detect(struct drm_connector *connector, connector->base.id, connector->name);
/* If full detect is not performed yet, do a full detect */
- if (!intel_dp->detect_done)
if (!intel_dp->detect_done) {
struct drm_crtc *crtc;int ret;crtc = connector->state->crtc;if (crtc) {ret = drm_modeset_lock(&crtc->mutex, ctx);if (ret)return ret;}status = intel_dp_long_pulse(intel_dp->attached_connector);
}
intel_dp->detect_done = false;
@@ -5146,10 +5162,38 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd) }
if (!intel_dp->is_mst) {
struct drm_modeset_acquire_ctx ctx;struct drm_connector *connector = &intel_dp->attached_connector->base;struct drm_crtc *crtc;int iret;drm_modeset_acquire_init(&ctx, 0);+retry:
iret = drm_modeset_lock(&dev->mode_config.connection_mutex, &ctx);if (iret)goto err;crtc = connector->state->crtc;if (crtc) {iret = drm_modeset_lock(&crtc->mutex, &ctx);if (iret)goto err;}- if (!intel_dp_short_pulse(intel_dp)) { intel_dp->detect_done = false; goto put_power; }
+err:
if (iret == -EDEADLK) {drm_modeset_backoff(&ctx);goto retry;}drm_modeset_drop_locks(&ctx);drm_modeset_acquire_fini(&ctx);WARN(iret, "Acquiring modeset locks failed with %i\n", iret);}
ret = IRQ_HANDLED;
-- 2.15.0
-- Ville Syrjälä Intel OTC
-- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch
-- Ville Syrjälä Intel OTC
-- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch