drm/tegra: Fix device/module refs for DP
cover drm/tegra: Fix device/module refs for DP
This patch series fixes a regression that was introduced by one of the changes I made to when Tegra registers its AUX adapters, along with fixing some reference leaks that I found along the way.
!!!NOTE!!!
There's one thing I'm not entirely sure about, which is the use of module references (e.g. try_module_get()) here. If I'm understanding how this code worked previously: since the get_device call in tegra_sor_probe() was previously the i2c adapter for the AUX channel - which itself is initialized in drm_dp_aux_register() - then I -think- that the module owner for the DDC adapter would likely have been drm_kms_helper. With these changes, if I'm understanding things correctly we're now just grabbing a module reference for ourselves - something which might not be the best idea?
If anyone could confirm if I need to fix this or not that'd be appreciated, along with reviews of course :P
Lyude Paul (2):
drm/tegra: Get ref for DP AUX channel, not its ddc adapter drm/tegra: Fix DP AUX channel reference leaks
drivers/gpu/drm/tegra/sor.c | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-)
Signature · 6 lines
2.30.2 _______________________________________________ dri-devel mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/dri-devel
PATCH 1/2 drm/tegra: Get ref for DP AUX channel, not its ddc adapter
While we're taking a reference of the DDC adapter for a DP AUX channel in tegra_sor_probe() because we're going to be using that adapter with the SOR, now that we've moved where AUX registration happens the actual device structure for the DDC adapter isn't initialized yet. Which means that we can't really take a reference from it to try to keep it around anymore.
This should be fine though, because we can just take a reference of its parent instead.
--- drivers/gpu/drm/tegra/sor.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/gpu/drm/tegra/sor.c +3 -3
@@ -3739,11 +3739,11 @@ static int tegra_sor_probe(struct platform_device *pdev)if (!sor->aux)return -EPROBE_DEFER ; -if (get_device (&sor->aux->ddc.dev)) { -if (try_module_get (sor->aux->ddc.owner)) +if (get_device (sor->aux->dev)) { +if (try_module_get (sor->aux->dev->driver->owner)) sor->output.ddc = &sor->aux->ddc;else -put_device (&sor->aux->ddc.dev); +put_device (sor->aux->dev); } }
Signature · 1 line
2.30.2
Re: [PATCH 1/2] drm/tegra: Get ref for DP AUX channel, not its ddc adapter
Quoted 28 lines · Lyude Paul
While we're taking a reference of the DDC adapter for a DP AUX channel in
tegra_sor_probe()because we're going to be using that adapter with the SOR, now that we've moved where AUX registration happens the actual device structure for the DDC adapter isn't initialized yet. Which means that we can't really take a reference from it to try to keep it around anymore.This should be fine though, because we can just take a reference of its parent instead.
Signed-off-by Lyude Paul <[email protected]>Fixes 39c17ae60ea9 ("drm/tegra: Don't register DP AUX channels before connectors")Cc Lyude Paul <[email protected]>Cc Thierry Reding <[email protected]>Cc Jonathan Hunter <[email protected]>--- drivers/gpu/drm/tegra/sor.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)drivers/gpu/drm/tegra/sor.c +0
-0@@ -3739,11 +3739,11 @@ static int tegra_sor_probe(struct platform_device *pdev)if (!sor->aux)return -EPROBE_DEFER ;
drivers/gpu/drm/tegra/sor.c static int tegra_sor_probe(struct platform_device *pdev) · line 3748
Quoted 10 lines
drivers/gpu/drm/tegra/sor.c +3
-3-if (get_device (&sor->aux->ddc.dev)) { -if (try_module_get (sor->aux->ddc.owner)) +if (get_device (sor->aux->dev)) { +if (try_module_get (sor->aux->dev->driver->owner)) sor->output.ddc = &sor->aux->ddc;else -put_device (&sor->aux->ddc.dev); +put_device (sor->aux->dev); } }
Unfortunately, I think it's a bit more subtle than that. The reason for this get_device()/try_module_get() dance was to mirror the behaviour of of_get_i2c_adapter_by_node() so that when we call i2c_put_adapter() in tegra_output_remove() we correctly decrease the reference count.
The above will increase the reference on the I2C adapter's parent while i2c_put_adapter() will then only decrease the reference on the I2C adapter, so I think effectively we'd be leaking a reference to the I2C adapter's parent.
Also, since we didn't take a reference on the I2C adapter explicitly, releasing that reference in tegra_output_remove() might free the I2C adapter too early.
I wonder if perhaps it'd be easier to get rid of the struct tegra_output abstraction altogether and push this down into the individual drivers, even if that means a bit more code duplication. That's not the kind of quick fix to resolve this current situation, so perhaps as a stop-gap we just need to sprinkle a few more conditionals throughout tegra_output code. We could, for example, avoid calling i2c_put_adapter() in tegra_output_remove() for the DisplayPort cases and instead manually release the reference to the I2C adapter's parent in tegra_sor_remove(). On top of your patch above that /should/ fix things properly for now.
Thierry
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
Re: [PATCH 1/2] drm/tegra: Get ref for DP AUX channel, not its ddc adapter
Quoted 56 lines · Thierry Reding
Quoted 32 lines · Lyude Paul
While we're taking a reference of the DDC adapter for a DP AUX channel in
tegra_sor_probe()because we're going to be using that adapter with the SOR, now that we've moved where AUX registration happens the actual device structure for the DDC adapter isn't initialized yet. Which means that we can't really take a reference from it to try to keep it around anymore.This should be fine though, because we can just take a reference of its parent instead.
Signed-off-by Lyude Paul <[email protected]>Fixes 39c17ae60ea9 ("drm/tegra: Don't register DP AUX channels beforeconnectors")
Cc: Lyude Paul <[email protected]>
Cc: Thierry Reding <[email protected]>
Cc: Jonathan Hunter <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
drivers/gpu/drm/tegra/sor.c | 6 +++---1 file changed, 3 insertions(+), 3 deletions(-)drivers/gpu/drm/tegra/sor.c +0
-0@@ -3739,11 +3739,11 @@ static int tegra_sor_probe(struct platform_device *pdev) if (!sor->aux) return -EPROBE_DEFER;
- if (get_device(&sor->aux->ddc.dev)) { - if (try_module_get(sor->aux->ddc.owner))Quoted 8 lines
+ if (
get_device(sor->aux->dev)) {
+ if (try_module_get(sor->aux->dev->driver->owner))sor->output.ddc=&sor->aux->ddc; else
- put_device(&sor->aux->ddc.dev); + put_device(sor->aux->dev);
}
}Unfortunately, I think it's a bit more subtle than that. The reason for this get_device()/try_module_get() dance was to mirror the behaviour of
of_get_i2c_adapter_by_node()so that when we calli2c_put_adapter()integra_output_remove()we correctly decrease the reference count.The above will increase the reference on the I2C adapter's parent while
i2c_put_adapter()will then only decrease the reference on the I2C adapter, so I think effectively we'd be leaking a reference to the I2C adapter's parent.Also, since we didn't take a reference on the I2C adapter explicitly, releasing that reference in
tegra_output_remove()might free the I2C adapter too early.
Quoted 9 lines
I wonder if perhaps it'd be easier to get rid of the struct
tegra_outputabstraction altogether and push this down into the individual drivers, even if that means a bit more code duplication. That's not the kind of quick fix to resolve this current situation, so perhaps as a stop-gap we just need to sprinkle a few more conditionals throughouttegra_outputcode. We could, for example, avoid callingi2c_put_adapter()integra_output_remove()for the DisplayPort cases and instead manually release the reference to the I2C adapter's parent integra_sor_remove(). On top of your patch above that /should/ fix things properly for now.
Alright - I will try to get to this tomorrow
Quoted 2 lines
Thierry
Signature · 3 lines
Cheers, Lyude Paul (she/her) Software Engineer at Red Hat
PATCH v2 1/1 drm/tegra: Get ref for DP AUX channel, not its ddc adapter
In reply to Lyude PaulWhile we're taking a reference of the DDC adapter for a DP AUX channel in tegra_sor_probe() because we're going to be using that adapter with the SOR, now that we've moved where AUX registration happens the actual device structure for the DDC adapter isn't initialized yet. Which means that we can't really take a reference from it to try to keep it around anymore.
This should be fine though, because we can just take a reference of its parent instead.
v2:
* Avoid calling i2c_put_adapter() in tegra_output_remove() for eDP/DP cases
--- drivers/gpu/drm/tegra/output.c | 5 ++++- drivers/gpu/drm/tegra/sor.c | 6 +++--- 2 files changed, 7 insertions(+), 4 deletions(-)
drivers/gpu/drm/tegra/output.c +4 -1
@@ -180,10 +180,13 @@ int tegra_output_probe(struct tegra_output *output)void tegra_output_remove (struct tegra_output *output) { +int connector_type = output->connector.connector_type; +if (output->hpd_gpio)free_irq (output->hpd_irq, output); -if (output->ddc) +if (connector_type != DRM_MODE_CONNECTOR_eDP && + connector_type != DRM_MODE_CONNECTOR_DisplayPort && output->ddc)i2c_put_adapter (output->ddc); }
drivers/gpu/drm/tegra/sor.c +3 -3
@@ -3739,11 +3739,11 @@ static int tegra_sor_probe(struct platform_device *pdev)if (!sor->aux)return -EPROBE_DEFER ; -if (get_device (&sor->aux->ddc.dev)) { -if (try_module_get (sor->aux->ddc.owner)) +if (get_device (sor->aux->dev)) { +if (try_module_get (sor->aux->dev->driver->owner)) sor->output.ddc = &sor->aux->ddc;else -put_device (&sor->aux->ddc.dev); +put_device (sor->aux->dev); } }
Signature · 1 line
2.31.1
PATCH 2/2 drm/tegra: Fix DP AUX channel reference leaks
In reply to Lyude PaulNoticed while fixing the regression I introduced in Tegra that Tegra seems to actually never release the device or module references it's grabbing for the DP AUX channel. So, let's fix that by dropping them when appropriate.
--- drivers/gpu/drm/tegra/sor.c | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-)
drivers/gpu/drm/tegra/sor.c +15 -4
@@ -3772,12 +3772,13 @@ static int tegra_sor_probe(struct platform_device *pdev) err =tegra_sor_parse_dt (sor);if (err <0 ) -return err; + goto put_aux; err =tegra_output_probe (&sor->output); -if (err <0 ) -return dev_err_probe (&pdev->dev, err, -"failed to probe output\n" ); +if (err <0 ) { + err =dev_err_probe (&pdev->dev, err,"failed to probe output\n" ); + goto put_aux; + }if (sor->ops && sor->ops->probe) { err = sor->ops->probe (sor);
@@ -3966,6 +3967,11 @@ static int tegra_sor_probe(struct platform_device *pdev)pm_runtime_disable (&pdev->dev); remove:tegra_output_remove (&sor->output); +put_aux: +if (sor->aux && sor->output.ddc) { +module_put (sor->aux->dev->driver->owner); +put_device (sor->aux->dev); + }return err; }
@@ -3985,6 +3991,11 @@ static int tegra_sor_remove(struct platform_device *pdev)tegra_output_remove (&sor->output); +if (sor->aux && sor->output.ddc) { +module_put (sor->aux->dev->driver->owner); +put_device (sor->aux->dev); + } +return 0 ; }
Signature · 1 line
2.30.2