Re: [PATCH] drm/amd/display:Check dp_128b_132b_supported_link_rates.raw before link_rate setting

[Jackie Dong <xy-jackie@139.com>]

On Tue, 2026-06-30 at 15:11 +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential
> issue(s) to consider:
> - [High] Checking `dp_128b_132b_supported_link_rates.raw` before it
> is read from the DPCD register completely disables DP 2.0 UHBR
> support.
> --
> 
> commit 11d06153280ec5dae6682a402606818b4f84b79d
> Author: Jackie Dong <xy-jackie@139.com>
> 
> drm/amd/display:Check dp_128b_132b_supported_link_rates.raw before
> link_rate setting
> 
> This patch modifies the DP 128b/132b capability check to also verify
> that
> dp_128b_132b_supported_link_rates.raw is non-zero before attempting
> to read
> and parse the specific UHBR link rates.
> 
> > diff --git
> > a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.
> > c
> > b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.
> > c
> > ---
> > a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.
> > c
> > +++
> > b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.
> > c
> > @@ -2079,7 +2079,8 @@ static bool retrieve_link_cap(struct dc_link
> > *link)
> >  	if (!dpcd_read_sink_ext_caps(link))
> >  		link->dpcd_sink_ext_caps.raw = 0;
> >  
> > -	if (link-
> > >dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED) {
> > +	if (link-
> > >dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED &&
> > +			link-
> > >dpcd_caps.dp_128b_132b_supported_link_rates.raw) {
> 
> [Severity: High]
> Will this change completely disable DP 2.0 UHBR support? 
> 
> In retrieve_link_cap() in
> drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c,
> dp_128b_132b_supported_link_rates.raw is now checked before it is
> actually 
> read from the DPCD register. Since the dc_link struct is zero-
> initialized and
> the read hasn't occurred yet, won't this raw value always evaluate to
> 0 upon
> connection of any DP 2.0 monitor? 
> 
> This would cause the if condition to always fail, bypassing the
> actual read
> block below and permanently preventing UHBR feature detection.
> 
Hi Sashiko,
 I have verified it as below. To add
dp_128b_132b_supported_link_rates.raw  as a new judgment doesn's impact
the value of dp_128b_132b_supported_link_rates.raw. If the monitor can
support UHBR and it can be into feature detection and assigned the
right UHBR rate.

root@yoga-Yoga-Pro-7-15ASH11:/media/yoga/Data/git/linux-502d801f0ab03-
20260622# grep -B23 "Invalid RX 128b_132b_supported_link_rates"
drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
 
	//if (link-
>dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED) {
	link->dpcd_caps.dp_128b_132b_supported_link_rates.raw = 0x10;
	pr_info("Jackie %x is SUPPORTED, %x: RX
128b_132b_supported_link_rates.raw\n", link-
>dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED, link-
>dpcd_caps.dp_128b_132b_supported_link_rates.raw) ;
	//pr_info("Jackie  %x: RX
128b_132b_supported_link_rates.raw\n", link-
>dpcd_caps.dp_128b_132b_supported_link_rates.raw) ;
	//pr_info("Jackie RX 128b_132b_supported_link_rates.raw\n") ;
	if ( (link-
>dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED) && (link-
>dpcd_caps.dp_128b_132b_supported_link_rates.raw) ) {
		DC_LOG_DP2("128b/132b encoding is supported at link
%d", link->link_index);
 
		/* Read 128b/132b suppoerted link rates */
		core_link_read_dpcd(link,
				DP_128B132B_SUPPORTED_LINK_RATES,
&link->dpcd_caps.dp_128b_132b_supported_link_rates.raw,
				sizeof(link-
>dpcd_caps.dp_128b_132b_supported_link_rates.raw));
		if (link-
>dpcd_caps.dp_128b_132b_supported_link_rates.bits.UHBR20) {
			pr_info("Jackie RX
128b_132b_supported_link_rates is UHBR20\n") ;
			link->reported_link_cap.link_rate =
LINK_RATE_UHBR20;
		} else if (link-
>dpcd_caps.dp_128b_132b_supported_link_rates.bits.UHBR13_5)
			link->reported_link_cap.link_rate =
LINK_RATE_UHBR13_5;
		else if (link-
>dpcd_caps.dp_128b_132b_supported_link_rates.bits.UHBR10)
			link->reported_link_cap.link_rate =
LINK_RATE_UHBR10;
		else
			//link->reported_link_cap.link_rate =
LINK_RATE_UHBR10;//Jackie Add
			dm_error("%s: Invalid RX
128b_132b_supported_link_rates\n", __func__);
root@yoga-Yoga-Pro-7-15ASH11:/media/yoga/Data/git/linux-502d801f0ab03-
20260622# dmesg |grep -i "fail\|error\|Jackie"
[    0.025954] KHO: Failed to reserve lowmem scratch buffer
[    0.025956] KHO: Failed to reserve scratch area, disabling kexec
handover
[    1.731523] RAS: Correctable Errors collector initialized.
[    7.629667] Jackie 1 is SUPPORTED, 10: RX
128b_132b_supported_link_rates.raw
[    7.630067] [drm:retrieve_link_cap [amdgpu]] *ERROR*
retrieve_link_cap: Invalid RX 128b_132b_supported_link_rates

Above is my verified source code and test result, FYR.

Thanks for your comments.

> >  		DC_LOG_DP2("128b/132b encoding is supported at
> > link %d", link->link_index);
> >  
> >  		/* Read 128b/132b suppoerted link rates */
> >  		core_link_read_dpcd(link,
> >  				DP_128B132B_SUPPORTED_LINK_RATES,
> >  				&link-
> > >dpcd_caps.dp_128b_132b_supported_link_rates.raw,
>