linux-kernel - Re: [PATCH] drm/bridge: analogix_dp: Fix connector status detection for bridges

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

Message-ID: <3572997.QJadu78ljV@diego>
Date: Fri, 10 Oct 2025 01:42:11 +0200
From: Heiko Stübner <heiko@...ech.de>
To: Dmitry Baryshkov <dmitry.baryshkov@....qualcomm.com>
Cc: damon.ding@...k-chips.com, m.szyprowski@...sung.com,
 andrzej.hajda@...el.com, neil.armstrong@...aro.org, rfoss@...nel.org,
 linux-kernel@...r.kernel.org, dri-devel@...ts.freedesktop.org,
 linux-arm-kernel@...ts.infradead.org, linux-samsung-soc@...r.kernel.org,
 linux-rockchip@...ts.infradead.org
Subject:
 Re: [PATCH] drm/bridge: analogix_dp: Fix connector status detection for
 bridges

Hi Dmitry,

Am Freitag, 10. Oktober 2025, 00:30:11 Mitteleuropäische Sommerzeit schrieb Dmitry Baryshkov:
> On Thu, Oct 09, 2025 at 09:30:28PM +0200, Heiko Stuebner wrote:
> > Right now if there is a next bridge attached to the analogix-dp controller
> > the driver always assumes this bridge is connected to something, but this
> > is of course not always true, as that bridge could also be a hotpluggable
> > dp port for example.
> > 
> > On the other hand, as stated in commit cb640b2ca546 ("drm/bridge: display-
> > connector: don't set OP_DETECT for DisplayPorts"), "Detecting the monitor
> > for DisplayPort targets is more complicated than just reading the HPD pin
> > level" and we should be "letting the actual DP driver perform detection."
> > 
> > So use drm_bridge_detect() to detect the next bridge's state but ignore
> > that bridge if the analogix-dp is handling the hpd-gpio.
> > 
> > Signed-off-by: Heiko Stuebner <heiko@...ech.de>
> > ---
> > As this patch stands, it would go on top of v6 of Damon's bridge-connector
> > work, but could very well be also integrated into one of the changes there.
> > 
> > I don't know yet if my ordering and/or reasoning is the correct one or if
> > a better handling could be done, but with that change I do get a nice
> > hotplug behaviour on my rk3588-tiger-dp-carrier board, where the
> > Analogix-DP ends in a full size DP port.
> > 
> >  drivers/gpu/drm/bridge/analogix/analogix_dp_core.c | 8 ++++++--
> >  1 file changed, 6 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
> > index c04b5829712b..cdc56e83b576 100644
> > --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
> > +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c
> > @@ -983,8 +983,12 @@ analogix_dp_bridge_detect(struct drm_bridge *bridge, struct drm_connector *conne
> >  	struct analogix_dp_device *dp = to_dp(bridge);
> >  	enum drm_connector_status status = connector_status_disconnected;
> >  
> > -	if (dp->plat_data->next_bridge)
> > -		return connector_status_connected;
> > +	/*
> > +	 * An optional next bridge should be in charge of detection the
> > +	 * connection status, except if we manage a actual hpd gpio.
> > +	 */
> > +	if (dp->plat_data->next_bridge && !dp->hpd_gpiod)
> > +		return drm_bridge_detect(dp->plat_data->next_bridge, connector);
> 
> And it's also not correct because the next bridge might be a retimer
> with the bridge next to it being a one with the actual detection
> capabilities. drm_bridge_connector solves that in a much better way. See
> the series at [1]
> 
> [1] https://lore.kernel.org/dri-devel/41c2a141-a72e-4780-ab32-f22f3a2e0179@samsung.com/

Hence my comment above about that possibly not being the right variant.
Sort of asking for direction :-) .

I am working on top of Damon's drm-bridge-connector series as noted above,
but it looks like the detect function still is called at does then stuff.

My board is the rk3588-tiger-displayport-carrier [0], with a dp-connector
which is the next bridge, so _without_ any changes, the analogix-dp
always assumes "something" is connected and I end up with

[    9.869198] [drm:analogix_dp_bridge_atomic_enable] *ERROR* failed to get hpd single ret = -110
[    9.980422] [drm:analogix_dp_bridge_atomic_enable] *ERROR* failed to get hpd single ret = -110
[   10.091522] [drm:analogix_dp_bridge_atomic_enable] *ERROR* failed to get hpd single ret = -110
[   10.202419] [drm:analogix_dp_bridge_atomic_enable] *ERROR* failed to get hpd single ret = -110
[   10.313651] [drm:analogix_dp_bridge_atomic_enable] *ERROR* failed to get hpd single ret = -110

when no display is connected.

With this change I do get the expected hotplug behaviour, so something is
missing still even with the bridge-connector series.


Heiko


[0] v3: https://lore.kernel.org/r/20250812083217.1064185-3-heiko@sntech.de
    v4: https://lore.kernel.org/r/20251009225050.88192-3-heiko@sntech.de
    (moved hpd-gpios from dp-connector back to analogix-dp per dp-connector
    being not able to detect dp-monitors)
> 
> >  
> >  	if (!analogix_dp_detect_hpd(dp))
> >  		status = connector_status_connected;
> 
>