Yes, and that's what I tested it with. On Wed, Jun 24, 2026, 11:54 AM Harry Wentland <[email protected]> wrote:
> > > On 2026-06-19 12:17, Deucher, Alexander wrote: > > Public > > > > Adding a few display folks to review. > > > > Thanks, > > > > Alex > > > >> -----Original Message----- > >> From: amd-gfx <[email protected]> On Behalf Of > >> Charles Munger > >> Sent: Friday, June 19, 2026 12:59 AM > >> To: [email protected] > >> Cc: Charles Munger <[email protected]> > >> Subject: [PATCH] drm/amdgpu: Add Paradetech OUI to VRR compat list > >> > >> There's at least two firmware versions on the CalDigit DP20-HDMI21 > adapter, > >> which is powered by the ParadeTech PS196G chip. The first version has a > bug > >> where it incorrectly neglects to set the DOWN_STREAM_PORT_PRESENT field > >> in the DPCD: > >> > >> $ sudo dpcd_reg read --offset=0x2205 --count=1 --device=0 > >> 0x2205: 00 > >> $ sudo dpcd_reg read --offset=0x0005 --count=1 --device=0 > >> 0x0005: 00 > >> > >> The newer one fixes this, and with this patch correctly negotiates VRR. > >> > >> [~]$ sudo dpcd_reg read --offset=0x5 --count=1 --device=2 > >> 0x0005: 1d > >> [~]$ sudo dpcd_reg read --offset=0x2205 --count=1 --device=2 > >> 0x2205: 1d > >> > >> So if you're using a Paradetech adapter and it doesn't negotiate VRR, > check > >> that it has a firmware that produces the correct DPCD. > >> > >> Signed-off-by: Charles Munger <[email protected]> > >> --- > >> drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c | 1 + > >> 1 file changed, 1 insertion(+) > >> > >> diff --git > >> a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c > >> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c > >> index 6c827e6703a0..f463495d1c87 100644 > >> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c > >> +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c > >> @@ -1410,6 +1410,7 @@ bool dm_helpers_is_vrr_pcon_allowed(const > > Looks like this is based on the "drm/amd: VRR fixes, HDMI Gaming Features" > series, not directly on amd-staging-drm-next. I'll have to take another > closer look at that series. > > Harry > > >> struct dc_link *link, const struct drm > >> case DP_BRANCH_DEVICE_ID_00E04C: > >> case DP_BRANCH_DEVICE_ID_90CC24: > >> case DP_BRANCH_DEVICE_ID_2B02F0: > >> + case DP_BRANCH_DEVICE_ID_001CF8: > >> return true; > >> } > >> > >> -- > >> 2.54.0 > > > >
