With the recent addition of locking checks in
commit
62ff94a5492175759546f8bc61383189d6b49122
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
AuthorDate: Thu Jan 23 22:18:47 2014 +0100
drm/crtc-helper: remove LOCKING from kerneldoc
drm_add_edid_modes started to WARN about the mode_config.mutex not
being held in the lvds and dp initialization code.
Now since this is init code locking is fairly redudant if it wouldn't
be for the drm core registering sysfs files a bit early. And the
locking WARNINGs nicely enforce that indeed all access to the mode
lists are properly protected. And a full audit shows that only i915
and gma500 touch the modes lists at init time.
Hence I've opted to wrap up this entire mode detection sequence for
fixed panels with the mode_config mutex for both lvds and edp outputs.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
/* We now know it's not a ghost, init power sequence regs. */
intel_dp_init_panel_power_sequencer_registers(dev, intel_dp, power_seq);
+ mutex_lock(&dev->mode_config.mutex);
edid = drm_get_edid(connector, &intel_dp->aux.ddc);
if (edid) {
if (drm_add_edid_modes(connector, edid)) {
if (fixed_mode)
fixed_mode->type |= DRM_MODE_TYPE_PREFERRED;
}
+ mutex_unlock(&dev->mode_config.mutex);
intel_panel_init(&intel_connector->panel, fixed_mode, NULL);
intel_panel_setup_backlight(connector);
* Attempt to get the fixed panel mode from DDC. Assume that the
* preferred mode is the right one.
*/
+ mutex_lock(&dev->mode_config.mutex);
edid = drm_get_edid(connector, intel_gmbus_get_adapter(dev_priv, pin));
if (edid) {
if (drm_add_edid_modes(connector, edid)) {
goto failed;
out:
+ mutex_unlock(&dev->mode_config.mutex);
+
lvds_encoder->is_dual_link = compute_is_dual_link_lvds(lvds_encoder);
DRM_DEBUG_KMS("detected %s-link lvds configuration\n",
lvds_encoder->is_dual_link ? "dual" : "single");
return;
failed:
+ mutex_unlock(&dev->mode_config.mutex);
+
DRM_DEBUG_KMS("No LVDS modes found, disabling.\n");
drm_connector_cleanup(connector);
drm_encoder_cleanup(encoder);