drm/amdgpu: move iommu_resume before ip init/resume
authorJames Zhu <James.Zhu@amd.com>
Tue, 7 Sep 2021 15:32:22 +0000 (11:32 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 18 Nov 2021 18:16:09 +0000 (19:16 +0100)
[ Upstream commit 9cec53c18a3170c7e5673c414da56aeecee94832 ]

Separate iommu_resume from kfd_resume, and move it before
other amdgpu ip init/resume.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=211277
Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c

index b8d9004fb16357388ea83c4b90ebeb391eddf8da..5b88c873c8a8982d96d659b522dc665bb91eebf7 100644 (file)
@@ -2394,6 +2394,10 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
        if (r)
                goto init_failed;
 
+       r = amdgpu_amdkfd_resume_iommu(adev);
+       if (r)
+               goto init_failed;
+
        r = amdgpu_device_ip_hw_init_phase1(adev);
        if (r)
                goto init_failed;