Merge tag 'drm-fixes-2024-04-12' of https://gitlab.freedesktop.org/drm/kernel

Pull drm fixes from Dave Airlie:
 "Looks like everyone woke up after holidays, this weeks pull has a
  bunch of stuff all over, 2 weeks worth of amdgpu is a lot of it, then
  i915/xe have a few, a bunch of msm fixes, then some scattered driver
  fixes.

  I expect things will settle down for rc5.

  client:
   - Protect connector modes with mode_config mutex

  ast:
   - Fix soft lockup

  host1x:
   - Do not setup DMA for virtual addresses

  ivpu:
   - Fix deadlock in context_xa
   - PCI fixes
   - Fixes to error handling

  nouveau:
   - gsp: Fix OOB access
   - Fix casting

  panfrost:
   - Fix error path in MMU code

  qxl:
   - Revert "drm/qxl: simplify qxl_fence_wait"

  vmwgfx:
   - Enable DMA for SEV mappings

  i915:
   - Couple CDCLK programming fixes
   - HDCP related fix
   - 4 Bigjoiner related fixes
   - Fix for a circular locking around GuC on reset+wedged case

  xe:
   - Fix double display mutex initializations
   - Fix u32 -> u64 implicit conversions
   - Fix RING_CONTEXT_CONTROL not marked as masked

  msm:
   - DP refcount leak fix on disconnect
   - Add missing newlines to prints in msm_fb and msm_kms
   - fix dpu debugfs entry permissions
   - Fix the interface table for the catalog of X1E80100
   - fix irq message printing
   - Bindings fix to add DP node as child of mdss for mdss node
   - Minor typo fix in DP driver API which handles port status change
   - fix CHRASHDUMP_READ()
   - fix HHB (highest bank bit) for a619 to fix UBWC corruption

  amdgpu:
   - GPU reset fixes
   - Fix some confusing logging
   - UMSCH fix
   - Aborted suspend fix
   - DCN 3.5 fixes
   - S4 fix
   - MES logging fixes
   - SMU 14 fixes
   - SDMA 4.4.2 fix
   - KASAN fix
   - SMU 13.0.10 fix
   - VCN partition fix
   - GFX11 fixes
   - DWB fixes
   - Plane handling fix
   - FAMS fix
   - DCN 3.1.6 fix
   - VSC SDP fixes
   - OLED panel fix
   - GFX 11.5 fix

  amdkfd:
   - GPU reset fixes
   - fix ioctl integer overflow"

* tag 'drm-fixes-2024-04-12' of https://gitlab.freedesktop.org/drm/kernel: (65 commits)
  amdkfd: use calloc instead of kzalloc to avoid integer overflow
  drm/xe: Label RING_CONTEXT_CONTROL as masked
  drm/xe/xe_migrate: Cast to output precision before multiplying operands
  drm/xe/hwmon: Cast result to output precision on left shift of operand
  drm/xe/display: Fix double mutex initialization
  drm/amdgpu: differentiate external rev id for gfx 11.5.0
  drm/amd/display: Adjust dprefclk by down spread percentage.
  drm/amd/display: Set VSC SDP Colorimetry same way for MST and SST
  drm/amd/display: Program VSC SDP colorimetry for all DP sinks >= 1.4
  drm/amd/display: fix disable otg wa logic in DCN316
  drm/amd/display: Do not recursively call manual trigger programming
  drm/amd/display: always reset ODM mode in context when adding first plane
  drm/amdgpu: fix incorrect number of active RBs for gfx11
  drm/amd/display: Return max resolution supported by DWB
  amd/amdkfd: sync all devices to wait all processes being evicted
  drm/amdgpu: clear set_q_mode_offs when VM changed
  drm/amdgpu: Fix VCN allocation in CPX partition
  drm/amd/pm: fix the high voltage issue after unload
  drm/amd/display: Skip on writeback when it's not applicable
  drm/amdgpu: implement IRQ_STATE_ENABLE for SDMA v4.4.2
  ...

1 files changed, 15 insertions, 11 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index aa16d51dd842..7753a2e64d41 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4135,18 +4135,22 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 					adev->ip_blocks[i].status.hw = true;
 				}
 			}
+		} else if (amdgpu_ip_version(adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 10) &&
+				   !amdgpu_device_has_display_hardware(adev)) {
+					r = psp_gpu_reset(adev);
 		} else {
-			tmp = amdgpu_reset_method;
-			/* It should do a default reset when loading or reloading the driver,
-			 * regardless of the module parameter reset_method.
-			 */
-			amdgpu_reset_method = AMD_RESET_METHOD_NONE;
-			r = amdgpu_asic_reset(adev);
-			amdgpu_reset_method = tmp;
-			if (r) {
-				dev_err(adev->dev, "asic reset on init failed\n");
-				goto failed;
-			}
+				tmp = amdgpu_reset_method;
+				/* It should do a default reset when loading or reloading the driver,
+				 * regardless of the module parameter reset_method.
+				 */
+				amdgpu_reset_method = AMD_RESET_METHOD_NONE;
+				r = amdgpu_asic_reset(adev);
+				amdgpu_reset_method = tmp;
+		}
+
+		if (r) {
+		  dev_err(adev->dev, "asic reset on init failed\n");
+		  goto failed;
 		}
 	}