[PATCH] drm/amd/amdgpu: Add rev_id workaround logic for SRIOV setup

Bokun Zhang Bokun.Zhang at amd.com
Tue Oct 27 22:01:48 UTC 2020


- When we are under SRIOV setup, the rev_id cannot be read
  properly. Therefore, we will return default value for it

Change-Id: I188d8e1b77f97c2eb29ef01aaf9ff9ea396a51c2
Signed-off-by: Bokun Zhang <Bokun.Zhang at amd.com>
---
 drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
index e0048806afaa..04c152843601 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v2_3.c
@@ -53,8 +53,20 @@ static void nbio_v2_3_remap_hdp_registers(struct amdgpu_device *adev)
 
 static u32 nbio_v2_3_get_rev_id(struct amdgpu_device *adev)
 {
-	u32 tmp = RREG32_SOC15(NBIO, 0, mmRCC_DEV0_EPF0_STRAP0);
+	u32 tmp;
+
+	if (amdgpu_sriov_vf(adev)) {
+		/* workaround on rev_id for sriov
+		* guest vm gets 0xffffffff when reading RCC_DEV0_EPF0_STRAP0,
+		* as a consequence, the rev_id and external_rev_id are wrong.
+		*
+		* workaround it by hardcoding the rev_id to 0,
+		* (which is the default value)
+		*/
+		return 0;
+	}
 
+	tmp = RREG32_SOC15(NBIO, 0, mmRCC_DEV0_EPF0_STRAP0);
 	tmp &= RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0_MASK;
 	tmp >>= RCC_DEV0_EPF0_STRAP0__STRAP_ATI_REV_ID_DEV0_F0__SHIFT;
 
-- 
2.25.1



More information about the amd-gfx mailing list