Merge tag 'drm-intel-next-2020-07-15' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

drm/i915 features for v5.9, batch #2

Highlights:
- Very early DG1 enabling (Abdiel, Lucas, Anusha)

Gem/GT:
- Fix spinlock recursion on signaling a signaled request (Chris)
- Perf: Use GTT when saving/restoring engine GPR (Umesh Nerlige Ramappa)

- SSEU refactoring, debugfs move under gt/ (Daniele, Venkata Sandeep Dhanalakota)
- Various GT refactoring and cleanup, preparation for future changes (Daniele)
- Adjust HuC state accordingly after GuC fetch error (Michał Winiarski)
- UC debugfs updates (Michał Winiarski)
- Only revoke the GGTT mmappings on aperture detiling changes (Chris)
- Only revoke mmap handlers if active (Chris)
- Split the context's obj:vma lut into its own mutex (Chris)
- Various memory, mmap and performance optimisations (Chris)
- Improve system stability in case of false CS events (Chris)
- Various refactorings and cleanup (Chris)
- Always reset the engine on execlist failures (Chris)
- Trace placement of timeline HWSP (Chris)
- Update dma-attributes for our sg DMA (Chris)

Display:
- TGL CDCLK workaround tweaks to unbreak 8K display support (Stanislav)
- A number of FBC fixes, along with i865 FBC enabling (Ville)
- Validate MST modes against PBN limits (Lyude, Shawn Lee)
- Do not access non-existing swizzle registers (Lucas)
- Revert GEN11+ HBR3 rate fix that caused issues on TGL (Matt Atwood)
- Update TGL+ combo phy initialization to match spec update (José)
- Fix HDCP Content Protection property state machine (Anshuman)
- Fix HDCP revoked keys handling (Ram)
- Improve DDI BUF status checks and waits (Manasi)
- Various SDVO+HDMI+DVI fixes around colorimetry, clocking, pixel repeat etc. (Ville)
- DP voltage swing function refactoring (José)
- WARN if max vswing/pre-emphasis violates the DP spec (Ville)

Other:
- Add new EHL PCI IDs (José)
- Unify struct intel_digital_port variable naming (Lucas)
- Various taint updates to aid debugging and improve CI (Michał Winiarski)
- Straggler conversions to new mmio register accessors (Daniele)

Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/87a70029vz.fsf@intel.com

2 files changed, 50 insertions, 0 deletions

diff --git a/include/drm/drm_modeset_helper_vtables.h b/include/drm/drm_modeset_helper_vtables.h
index 421a30f08463..4efec30f8bad 100644
--- a/include/drm/drm_modeset_helper_vtables.h
+++ b/include/drm/drm_modeset_helper_vtables.h
@@ -968,6 +968,48 @@ struct drm_connector_helper_funcs {
 	 */
 	enum drm_mode_status (*mode_valid)(struct drm_connector *connector,
 					   struct drm_display_mode *mode);
+
+	/**
+	 * @mode_valid_ctx:
+	 *
+	 * Callback to validate a mode for a connector, irrespective of the
+	 * specific display configuration.
+	 *
+	 * This callback is used by the probe helpers to filter the mode list
+	 * (which is usually derived from the EDID data block from the sink).
+	 * See e.g. drm_helper_probe_single_connector_modes().
+	 *
+	 * This function is optional, and is the atomic version of
+	 * &drm_connector_helper_funcs.mode_valid.
+	 *
+	 * To allow for accessing the atomic state of modesetting objects, the
+	 * helper libraries always call this with ctx set to a valid context,
+	 * and &drm_mode_config.connection_mutex will always be locked with
+	 * the ctx parameter set to @ctx. This allows for taking additional
+	 * locks as required.
+	 *
+	 * Even though additional locks may be acquired, this callback is
+	 * still expected not to take any constraints into account which would
+	 * be influenced by the currently set display state - such constraints
+	 * should be handled in the driver's atomic check. For example, if a
+	 * connector shares display bandwidth with other connectors then it
+	 * would be ok to validate the minimum bandwidth requirement of a mode
+	 * against the maximum possible bandwidth of the connector. But it
+	 * wouldn't be ok to take the current bandwidth usage of other
+	 * connectors into account, as this would change depending on the
+	 * display state.
+	 *
+	 * Returns:
+	 * 0 if &drm_connector_helper_funcs.mode_valid_ctx succeeded and wrote
+	 * the &enum drm_mode_status value to @status, or a negative error
+	 * code otherwise.
+	 *
+	 */
+	int (*mode_valid_ctx)(struct drm_connector *connector,
+			      struct drm_display_mode *mode,
+			      struct drm_modeset_acquire_ctx *ctx,
+			      enum drm_mode_status *status);
+
 	/**
 	 * @best_encoder:
 	 *
diff --git a/include/drm/i915_pciids.h b/include/drm/i915_pciids.h
index bc989de2aac2..96e408b4bdc9 100644
--- a/include/drm/i915_pciids.h
+++ b/include/drm/i915_pciids.h
@@ -588,7 +588,11 @@
 	INTEL_VGA_DEVICE(0x4551, info), \
 	INTEL_VGA_DEVICE(0x4541, info), \
 	INTEL_VGA_DEVICE(0x4E71, info), \
+	INTEL_VGA_DEVICE(0x4557, info), \
+	INTEL_VGA_DEVICE(0x4555, info), \
 	INTEL_VGA_DEVICE(0x4E61, info), \
+	INTEL_VGA_DEVICE(0x4E57, info), \
+	INTEL_VGA_DEVICE(0x4E55, info), \
 	INTEL_VGA_DEVICE(0x4E51, info)
 
 /* TGL */
@@ -614,4 +618,8 @@
 	INTEL_VGA_DEVICE(0x4C90, info), \
 	INTEL_VGA_DEVICE(0x4C9A, info)
 
+/* DG1 */
+#define INTEL_DG1_IDS(info) \
+	INTEL_VGA_DEVICE(0x4905, info)
+
 #endif /* _I915_PCIIDS_H */