linux-kernel - Re: [PATCH] drm/i915/dp: Use slow and wide link training for everything

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <YIBxduMptaKAFOUq@intel.com>
Date:   Wed, 21 Apr 2021 21:39:50 +0300
From:   Ville Syrjälä <ville.syrjala@...ux.intel.com>
To:     Kai-Heng Feng <kai.heng.feng@...onical.com>
Cc:     jani.nikula@...ux.intel.com, joonas.lahtinen@...ux.intel.com,
        rodrigo.vivi@...el.com, David Airlie <airlied@...ux.ie>,
        Daniel Vetter <daniel@...ll.ch>,
        Imre Deak <imre.deak@...el.com>,
        Uma Shankar <uma.shankar@...el.com>,
        Manasi Navare <manasi.d.navare@...el.com>,
        Ankit Nautiyal <ankit.k.nautiyal@...el.com>,
        Gwan-gyeong Mun <gwan-gyeong.mun@...el.com>,
        Lucas De Marchi <lucas.demarchi@...el.com>,
        Sean Paul <seanpaul@...omium.org>,
        intel-gfx@...ts.freedesktop.org,
        "open list:DRM DRIVERS" <dri-devel@...ts.freedesktop.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] drm/i915/dp: Use slow and wide link training for
 everything

On Wed, Apr 21, 2021 at 01:20:31PM +0800, Kai-Heng Feng wrote:
> Screen flickers on Innolux eDP 1.3 panel when clock rate 540000 is in use.
> 
> According to the panel vendor, though clock rate 540000 is advertised,
> but the max clock rate it really supports is 270000.
> 
> Ville Syrjälä mentioned that fast and narrow also breaks some eDP 1.4
> panel, so use slow and wide training for all panels to resolve the
> issue.
> 
> User also confirmed that the new strategy doesn't introduce any
> regression on XPS 9380.
> 
> v2:
>  - Use slow and wide for everything.
> 
> Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/3384
> References: https://gitlab.freedesktop.org/drm/intel/-/issues/272
> Signed-off-by: Kai-Heng Feng <kai.heng.feng@...onical.com>

Thanks. Pushed to drm-intel-next.

I did a quick scan of a few CI logs and noticed that at least cml-u2
changed behaviour:
- [CONNECTOR:95:eDP-1] Link Training passed at link rate = 432000, lane count = 1, at DPRX
+ [CONNECTOR:95:eDP-1] Link Training passed at link rate = 216000, lane count = 2, at DPRX

But it still appears to work, and 2.16Gbps is also the link rate chosen
by the BIOS, which is reassuring.

> --- a/drivers/gpu/drm/i915/display/intel_dp.c
> +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> @@ -1095,44 +1095,6 @@ intel_dp_compute_link_config_wide(struct intel_dp *intel_dp,
>  	return -EINVAL;
>  }
>  
> -/* Optimize link config in order: max bpp, min lanes, min clock */
> -static int
> -intel_dp_compute_link_config_fast(struct intel_dp *intel_dp,
> -				  struct intel_crtc_state *pipe_config,
> -				  const struct link_config_limits *limits)
> -{
> -	const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
> -	int bpp, clock, lane_count;
> -	int mode_rate, link_clock, link_avail;
> -
> -	for (bpp = limits->max_bpp; bpp >= limits->min_bpp; bpp -= 2 * 3) {
> -		int output_bpp = intel_dp_output_bpp(pipe_config->output_format, bpp);
> -
> -		mode_rate = intel_dp_link_required(adjusted_mode->crtc_clock,
> -						   output_bpp);
> -
> -		for (lane_count = limits->min_lane_count;
> -		     lane_count <= limits->max_lane_count;
> -		     lane_count <<= 1) {
> -			for (clock = limits->min_clock; clock <= limits->max_clock; clock++) {
> -				link_clock = intel_dp->common_rates[clock];
> -				link_avail = intel_dp_max_data_rate(link_clock,
> -								    lane_count);
> -
> -				if (mode_rate <= link_avail) {
> -					pipe_config->lane_count = lane_count;
> -					pipe_config->pipe_bpp = bpp;
> -					pipe_config->port_clock = link_clock;
> -
> -					return 0;
> -				}
> -			}
> -		}
> -	}
> -
> -	return -EINVAL;
> -}
> -
>  static int intel_dp_dsc_compute_bpp(struct intel_dp *intel_dp, u8 dsc_max_bpc)
>  {
>  	int i, num_bpc;
> @@ -1382,22 +1344,11 @@ intel_dp_compute_link_config(struct intel_encoder *encoder,
>  	    intel_dp_can_bigjoiner(intel_dp))
>  		pipe_config->bigjoiner = true;
>  
> -	if (intel_dp_is_edp(intel_dp))
> -		/*
> -		 * Optimize for fast and narrow. eDP 1.3 section 3.3 and eDP 1.4
> -		 * section A.1: "It is recommended that the minimum number of
> -		 * lanes be used, using the minimum link rate allowed for that
> -		 * lane configuration."
> -		 *
> -		 * Note that we fall back to the max clock and lane count for eDP
> -		 * panels that fail with the fast optimal settings (see
> -		 * intel_dp->use_max_params), in which case the fast vs. wide
> -		 * choice doesn't matter.
> -		 */
> -		ret = intel_dp_compute_link_config_fast(intel_dp, pipe_config, &limits);
> -	else
> -		/* Optimize for slow and wide. */
> -		ret = intel_dp_compute_link_config_wide(intel_dp, pipe_config, &limits);
> +	/*
> +	 * Optimize for slow and wide for everything, because there are some
> +	 * eDP 1.3 and 1.4 panels don't work well with fast and narrow.
> +	 */
> +	ret = intel_dp_compute_link_config_wide(intel_dp, pipe_config, &limits);
>  
>  	/* enable compression if the mode doesn't fit available BW */
>  	drm_dbg_kms(&i915->drm, "Force DSC en = %d\n", intel_dp->force_dsc_en);
> -- 
> 2.30.2

-- 
Ville Syrjälä
Intel