Message ID | 20240215220442.1343152-1-martin.blumenstingl@googlemail.com (mailing list archive) |
---|---|
State | New, archived |
Delegated to: | Neil Armstrong |
Headers | show |
Series | drm/meson: Don't remove bridges which are created by other drivers | expand |
Tested-by: Steve Morvai <stevemorvai@hotmail.com>
Just to confirm that the correction addresses the failure of the HDMI initialization issue as reported. Thanks.
On 15/02/2024 23:04, Martin Blumenstingl wrote: > Stop calling drm_bridge_remove() for bridges allocated/managed by other > drivers in the remove paths of meson_encoder_{cvbs,dsi,hdmi}. > drm_bridge_remove() unregisters the bridge so it cannot be used > anymore. Doing so for bridges we don't own can lead to the video > pipeline not being able to come up after -EPROBE_DEFER of the VPU > because we're unregistering a bridge that's managed by another driver. > The other driver doesn't know that we have unregistered it's bridge > and on subsequent .probe() we're not able to find those bridges anymore > (since nobody re-creates them). > > This fixes probe errors on Meson8b boards with the CVBS outputs enabled. > > Fixes: 09847723c12f ("drm/meson: remove drm bridges at aggregate driver unbind time") > Fixes: 42dcf15f901c ("drm/meson: add DSI encoder") > Cc: stable@vger.kernel.org > Reported-by: Steve Morvai <stevemorvai@hotmail.com> > Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com> > --- > This issue was reported by Steve off-list to me (thanks again for your > patience and sorry it took so long)! > The Meson8b VPU driver is not upstream, but the problematic code is. > Meaning: This issue can also appear on SoCs which are supported > upstream if the meson DRM driver probe has to be re-tried (with > -EPROBE_DEFER). That's why I chose to Cc the stable list. > > > drivers/gpu/drm/meson/meson_encoder_cvbs.c | 1 - > drivers/gpu/drm/meson/meson_encoder_dsi.c | 1 - > drivers/gpu/drm/meson/meson_encoder_hdmi.c | 1 - > 3 files changed, 3 deletions(-) > > diff --git a/drivers/gpu/drm/meson/meson_encoder_cvbs.c b/drivers/gpu/drm/meson/meson_encoder_cvbs.c > index 3f73b211fa8e..3407450435e2 100644 > --- a/drivers/gpu/drm/meson/meson_encoder_cvbs.c > +++ b/drivers/gpu/drm/meson/meson_encoder_cvbs.c > @@ -294,6 +294,5 @@ void meson_encoder_cvbs_remove(struct meson_drm *priv) > if (priv->encoders[MESON_ENC_CVBS]) { > meson_encoder_cvbs = priv->encoders[MESON_ENC_CVBS]; > drm_bridge_remove(&meson_encoder_cvbs->bridge); > - drm_bridge_remove(meson_encoder_cvbs->next_bridge); > } > } > diff --git a/drivers/gpu/drm/meson/meson_encoder_dsi.c b/drivers/gpu/drm/meson/meson_encoder_dsi.c > index 3f93c70488ca..311b91630fbe 100644 > --- a/drivers/gpu/drm/meson/meson_encoder_dsi.c > +++ b/drivers/gpu/drm/meson/meson_encoder_dsi.c > @@ -168,6 +168,5 @@ void meson_encoder_dsi_remove(struct meson_drm *priv) > if (priv->encoders[MESON_ENC_DSI]) { > meson_encoder_dsi = priv->encoders[MESON_ENC_DSI]; > drm_bridge_remove(&meson_encoder_dsi->bridge); > - drm_bridge_remove(meson_encoder_dsi->next_bridge); > } > } > diff --git a/drivers/gpu/drm/meson/meson_encoder_hdmi.c b/drivers/gpu/drm/meson/meson_encoder_hdmi.c > index 25ea76558690..c4686568c9ca 100644 > --- a/drivers/gpu/drm/meson/meson_encoder_hdmi.c > +++ b/drivers/gpu/drm/meson/meson_encoder_hdmi.c > @@ -474,6 +474,5 @@ void meson_encoder_hdmi_remove(struct meson_drm *priv) > if (priv->encoders[MESON_ENC_HDMI]) { > meson_encoder_hdmi = priv->encoders[MESON_ENC_HDMI]; > drm_bridge_remove(&meson_encoder_hdmi->bridge); > - drm_bridge_remove(meson_encoder_hdmi->next_bridge); > } > } Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org> Thanks a lot for the patch! Neil
Hi, On Thu, 15 Feb 2024 23:04:42 +0100, Martin Blumenstingl wrote: > Stop calling drm_bridge_remove() for bridges allocated/managed by other > drivers in the remove paths of meson_encoder_{cvbs,dsi,hdmi}. > drm_bridge_remove() unregisters the bridge so it cannot be used > anymore. Doing so for bridges we don't own can lead to the video > pipeline not being able to come up after -EPROBE_DEFER of the VPU > because we're unregistering a bridge that's managed by another driver. > The other driver doesn't know that we have unregistered it's bridge > and on subsequent .probe() we're not able to find those bridges anymore > (since nobody re-creates them). > > [...] Thanks, Applied to https://anongit.freedesktop.org/git/drm/drm-misc.git (drm-misc-fixes) [1/1] drm/meson: Don't remove bridges which are created by other drivers https://cgit.freedesktop.org/drm/drm-misc/commit/?id=bd915ae73a2d78559b376ad2caf5e4ef51de2455
diff --git a/drivers/gpu/drm/meson/meson_encoder_cvbs.c b/drivers/gpu/drm/meson/meson_encoder_cvbs.c index 3f73b211fa8e..3407450435e2 100644 --- a/drivers/gpu/drm/meson/meson_encoder_cvbs.c +++ b/drivers/gpu/drm/meson/meson_encoder_cvbs.c @@ -294,6 +294,5 @@ void meson_encoder_cvbs_remove(struct meson_drm *priv) if (priv->encoders[MESON_ENC_CVBS]) { meson_encoder_cvbs = priv->encoders[MESON_ENC_CVBS]; drm_bridge_remove(&meson_encoder_cvbs->bridge); - drm_bridge_remove(meson_encoder_cvbs->next_bridge); } } diff --git a/drivers/gpu/drm/meson/meson_encoder_dsi.c b/drivers/gpu/drm/meson/meson_encoder_dsi.c index 3f93c70488ca..311b91630fbe 100644 --- a/drivers/gpu/drm/meson/meson_encoder_dsi.c +++ b/drivers/gpu/drm/meson/meson_encoder_dsi.c @@ -168,6 +168,5 @@ void meson_encoder_dsi_remove(struct meson_drm *priv) if (priv->encoders[MESON_ENC_DSI]) { meson_encoder_dsi = priv->encoders[MESON_ENC_DSI]; drm_bridge_remove(&meson_encoder_dsi->bridge); - drm_bridge_remove(meson_encoder_dsi->next_bridge); } } diff --git a/drivers/gpu/drm/meson/meson_encoder_hdmi.c b/drivers/gpu/drm/meson/meson_encoder_hdmi.c index 25ea76558690..c4686568c9ca 100644 --- a/drivers/gpu/drm/meson/meson_encoder_hdmi.c +++ b/drivers/gpu/drm/meson/meson_encoder_hdmi.c @@ -474,6 +474,5 @@ void meson_encoder_hdmi_remove(struct meson_drm *priv) if (priv->encoders[MESON_ENC_HDMI]) { meson_encoder_hdmi = priv->encoders[MESON_ENC_HDMI]; drm_bridge_remove(&meson_encoder_hdmi->bridge); - drm_bridge_remove(meson_encoder_hdmi->next_bridge); } }
Stop calling drm_bridge_remove() for bridges allocated/managed by other drivers in the remove paths of meson_encoder_{cvbs,dsi,hdmi}. drm_bridge_remove() unregisters the bridge so it cannot be used anymore. Doing so for bridges we don't own can lead to the video pipeline not being able to come up after -EPROBE_DEFER of the VPU because we're unregistering a bridge that's managed by another driver. The other driver doesn't know that we have unregistered it's bridge and on subsequent .probe() we're not able to find those bridges anymore (since nobody re-creates them). This fixes probe errors on Meson8b boards with the CVBS outputs enabled. Fixes: 09847723c12f ("drm/meson: remove drm bridges at aggregate driver unbind time") Fixes: 42dcf15f901c ("drm/meson: add DSI encoder") Cc: stable@vger.kernel.org Reported-by: Steve Morvai <stevemorvai@hotmail.com> Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com> --- This issue was reported by Steve off-list to me (thanks again for your patience and sorry it took so long)! The Meson8b VPU driver is not upstream, but the problematic code is. Meaning: This issue can also appear on SoCs which are supported upstream if the meson DRM driver probe has to be re-tried (with -EPROBE_DEFER). That's why I chose to Cc the stable list. drivers/gpu/drm/meson/meson_encoder_cvbs.c | 1 - drivers/gpu/drm/meson/meson_encoder_dsi.c | 1 - drivers/gpu/drm/meson/meson_encoder_hdmi.c | 1 - 3 files changed, 3 deletions(-)