If the display is not enable()d, then we aren't holding a runtime PM reference here. Thus, it's easy to accidentally cause a hang, if user space is poking around at /dev/drm_dp_aux0 at the "wrong" time.
Let's get a runtime PM reference, and check that we "see" the panel. Don't force any panel power-up, etc., because that can be intrusive, and that's not what other drivers do (see drivers/gpu/drm/bridge/ti-sn65dsi86.c and drivers/gpu/drm/bridge/parade-ps8640.c.)
Fixes: 0d97ad03f422 ("drm/bridge: analogix_dp: Remove duplicated code") Cc: stable@vger.kernel.org Cc: Tomeu Vizoso tomeu.vizoso@collabora.com Signed-off-by: Brian Norris briannorris@chromium.org ---
Changes in v3: - Avoid panel power-up; just check for HPD state, and let the rest happen "as-is" (e.g., time out, if the caller hasn't prepared things properly)
Changes in v2: - Fix spelling in Subject - DRM_DEV_ERROR() -> drm_err() - Propagate errors from un-analogix_dp_prepare_panel()
drivers/gpu/drm/bridge/analogix/analogix_dp_core.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c index b7d2e4449cfa..16be279aed2c 100644 --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c @@ -1632,8 +1632,19 @@ static ssize_t analogix_dpaux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) { struct analogix_dp_device *dp = to_dp(aux); + int ret; + + pm_runtime_get_sync(dp->dev); + + ret = analogix_dp_detect_hpd(dp); + if (ret) + goto out;
- return analogix_dp_transfer(dp, msg); + ret = analogix_dp_transfer(dp, msg); +out: + pm_runtime_put(dp->dev); + + return ret; }
struct analogix_dp_device *
DP AUX transactions can consist of many short operations. There's no need to power things up/down in short intervals.
I pick an arbitrary 100ms; for the systems I'm testing (Rockchip RK3399), runtime-PM transitions only take a few microseconds.
Signed-off-by: Brian Norris briannorris@chromium.org ---
Changes in v3: - New in v3
drivers/gpu/drm/bridge/analogix/analogix_dp_core.c | 6 ++++-- 1 file changed, 4 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 16be279aed2c..d82a4ddf44e7 100644 --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c @@ -1121,7 +1121,7 @@ static int analogix_dp_get_modes(struct drm_connector *connector)
pm_runtime_get_sync(dp->dev); edid = drm_get_edid(connector, &dp->aux.ddc); - pm_runtime_put(dp->dev); + pm_runtime_put_autosuspend(dp->dev); if (edid) { drm_connector_update_edid_property(&dp->connector, edid); @@ -1642,7 +1642,7 @@ static ssize_t analogix_dpaux_transfer(struct drm_dp_aux *aux,
ret = analogix_dp_transfer(dp, msg); out: - pm_runtime_put(dp->dev); + pm_runtime_put_autosuspend(dp->dev);
return ret; } @@ -1775,6 +1775,8 @@ int analogix_dp_bind(struct analogix_dp_device *dp, struct drm_device *drm_dev) if (ret) return ret;
+ pm_runtime_use_autosuspend(dp->dev); + pm_runtime_set_autosuspend_delay(dp->dev, 100); pm_runtime_enable(dp->dev);
ret = analogix_dp_create_bridge(drm_dev, dp);
Hi,
On Thu, Feb 17, 2022 at 2:42 PM Brian Norris briannorris@chromium.org wrote:
DP AUX transactions can consist of many short operations. There's no need to power things up/down in short intervals.
I pick an arbitrary 100ms; for the systems I'm testing (Rockchip RK3399), runtime-PM transitions only take a few microseconds.
Signed-off-by: Brian Norris briannorris@chromium.org
Changes in v3:
- New in v3
drivers/gpu/drm/bridge/analogix/analogix_dp_core.c | 6 ++++-- 1 file changed, 4 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 16be279aed2c..d82a4ddf44e7 100644 --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c @@ -1121,7 +1121,7 @@ static int analogix_dp_get_modes(struct drm_connector *connector)
pm_runtime_get_sync(dp->dev); edid = drm_get_edid(connector, &dp->aux.ddc);
pm_runtime_put(dp->dev);
pm_runtime_put_autosuspend(dp->dev);
So I think you can fully get rid of these ones now and rely on the ones in the aux transfer, right?
if (edid) { drm_connector_update_edid_property(&dp->connector, edid);@@ -1642,7 +1642,7 @@ static ssize_t analogix_dpaux_transfer(struct drm_dp_aux *aux,
ret = analogix_dp_transfer(dp, msg);out:
pm_runtime_put(dp->dev);
pm_runtime_put_autosuspend(dp->dev); return ret;} @@ -1775,6 +1775,8 @@ int analogix_dp_bind(struct analogix_dp_device *dp, struct drm_device *drm_dev) if (ret) return ret;
pm_runtime_use_autosuspend(dp->dev);pm_runtime_set_autosuspend_delay(dp->dev, 100);
It's explicitly listed in the Documentation that you need the corresponding pm_runtime_dont_use_autosuspend(). Specifically, it says:
Drivers in ->remove() callback should undo the runtime PM changes done in ->probe(). Usually this means calling pm_runtime_disable(), pm_runtime_dont_use_autosuspend() etc.
Not that it's very common to see anyone actually get it right, but I seem to remember running into an issue when I didn't do it. I think ti-sn65dsi86 still has it wrong since I found out about this later. Need to write a patch up for that... Basically you want to put it right before the two calls in your driver to pm_runtime_disable().
-Doug
On Tue, Feb 22, 2022 at 2:10 PM Doug Anderson dianders@chromium.org wrote:
On Thu, Feb 17, 2022 at 2:42 PM Brian Norris briannorris@chromium.org wrote:
--- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c @@ -1121,7 +1121,7 @@ static int analogix_dp_get_modes(struct drm_connector *connector)
pm_runtime_get_sync(dp->dev); edid = drm_get_edid(connector, &dp->aux.ddc);
pm_runtime_put(dp->dev);
pm_runtime_put_autosuspend(dp->dev);So I think you can fully get rid of these ones now and rely on the ones in the aux transfer, right?
Yep, good catch.
if (edid) { drm_connector_update_edid_property(&dp->connector, edid);@@ -1642,7 +1642,7 @@ static ssize_t analogix_dpaux_transfer(struct drm_dp_aux *aux,
ret = analogix_dp_transfer(dp, msg);out:
pm_runtime_put(dp->dev);
pm_runtime_put_autosuspend(dp->dev); return ret;} @@ -1775,6 +1775,8 @@ int analogix_dp_bind(struct analogix_dp_device *dp, struct drm_device *drm_dev) if (ret) return ret;
pm_runtime_use_autosuspend(dp->dev);pm_runtime_set_autosuspend_delay(dp->dev, 100);It's explicitly listed in the Documentation that you need the corresponding pm_runtime_dont_use_autosuspend(). Specifically, it says:
Drivers in ->remove() callback should undo the runtime PM changes done in ->probe(). Usually this means calling pm_runtime_disable(), pm_runtime_dont_use_autosuspend() etc.
Not that it's very common to see anyone actually get it right, but I seem to remember running into an issue when I didn't do it. I think ti-sn65dsi86 still has it wrong since I found out about this later. Need to write a patch up for that... Basically you want to put it right before the two calls in your driver to pm_runtime_disable().
Ack.
Speaking of API misfeatures that we missed: I've failed to call pm_runtime_mark_last_busy(). I'll add that in the next rev, for the cases where we weren't already calling *_put_sync() (i.e., presumably we don't really care to wait around for autosuspend).
*gripe* What a silly API. *gripe*
Brian
Hi,
On Thu, Feb 17, 2022 at 2:42 PM Brian Norris briannorris@chromium.org wrote:
If the display is not enable()d, then we aren't holding a runtime PM reference here. Thus, it's easy to accidentally cause a hang, if user space is poking around at /dev/drm_dp_aux0 at the "wrong" time.
Let's get a runtime PM reference, and check that we "see" the panel. Don't force any panel power-up, etc., because that can be intrusive, and that's not what other drivers do (see drivers/gpu/drm/bridge/ti-sn65dsi86.c and drivers/gpu/drm/bridge/parade-ps8640.c.)
Fixes: 0d97ad03f422 ("drm/bridge: analogix_dp: Remove duplicated code") Cc: stable@vger.kernel.org Cc: Tomeu Vizoso tomeu.vizoso@collabora.com Signed-off-by: Brian Norris briannorris@chromium.org
Changes in v3:
- Avoid panel power-up; just check for HPD state, and let the rest happen "as-is" (e.g., time out, if the caller hasn't prepared things properly)
Changes in v2:
- Fix spelling in Subject
- DRM_DEV_ERROR() -> drm_err()
- Propagate errors from un-analogix_dp_prepare_panel()
drivers/gpu/drm/bridge/analogix/analogix_dp_core.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-)
Reviewed-by: Douglas Anderson dianders@chromium.org
-Doug
dri-devel@lists.freedesktop.org
-
Brian Norris -
Doug Anderson