From: tiancyin <[email protected]> [why] When page fault happens, it could lead to sdma hang is RESP_MODE = 0 for non-PRT case.
[how] Setting SDMAx_UTCL1_CNTL.RESP_MODE to 0b011 to avoid SDMA halt. Reviewed-by: Jack Xiao <[email protected]> Signed-off-by: tiancyin <[email protected]> Signed-off-by: Alex Deucher <[email protected]> --- drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c index 083f81611e24..32b48c9ca051 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c @@ -658,7 +658,8 @@ static int sdma_v5_0_gfx_resume(struct amdgpu_device *adev) /* Set up RESP_MODE to non-copy addresses */ temp = RREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL)); - temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 2); + temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, RESP_MODE, 3); + temp = REG_SET_FIELD(temp, SDMA0_UTCL1_CNTL, REDO_DELAY, 9); WREG32(sdma_v5_0_get_reg_offset(adev, i, mmSDMA0_UTCL1_CNTL), temp); /* program default cache read and write policy */ -- 2.20.1 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
