[PATCH] drm/edid: drop the dcc probe before drm_do_get_edid

[PATCH] drm/edid: drop the dcc probe before drm_do_get_edid

[Lubomir Rintel]

Now that drm_do_get_edid() handles override and firmware EDIDs it makes
no sense to conditionalize it with a DCC probe. On the contrary -- the
overrides are useful specifically when DCC is not functioning.

drm_do_get_edid() already bails out when DCC fails, there's no need for
an extra check. It seems entirely safe to just drop this.

Working around the issue by forcing the connector on is not a good
solution because that way the desktop shell can't tell when a monitor is
plugged in or removed.

Cc: Jani Nikula <jani.nikula@intel.com>
Fixes: 53fd40a ("drm: handle override and firmware EDID at drm_do_get_edid() level")
---
 drivers/gpu/drm/drm_edid.c | 18 ------------------
 1 file changed, 18 deletions(-)

diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index cb487148359a..72169d84256d 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -1642,21 +1642,6 @@ struct edid *drm_do_get_edid(struct drm_connector *connector,
 }
 EXPORT_SYMBOL_GPL(drm_do_get_edid);
 
-/**
- * drm_probe_ddc() - probe DDC presence
- * @adapter: I2C adapter to probe
- *
- * Return: True on success, false on failure.
- */
-bool
-drm_probe_ddc(struct i2c_adapter *adapter)
-{
-	unsigned char out;
-
-	return (drm_do_probe_ddc_edid(adapter, &out, 0, 1) == 0);
-}
-EXPORT_SYMBOL(drm_probe_ddc);
-
 /**
  * drm_get_edid - get EDID data, if available
  * @connector: connector we're probing
@@ -1675,9 +1660,6 @@ struct edid *drm_get_edid(struct drm_connector *connector,
 	if (connector->force == DRM_FORCE_OFF)
 		return NULL;
 
-	if (connector->force == DRM_FORCE_UNSPECIFIED && !drm_probe_ddc(adapter))
-		return NULL;
-
 	edid = drm_do_get_edid(connector, drm_do_probe_ddc_edid, adapter);
 	if (edid)
 		drm_get_displayid(connector, edid);
-- 
2.14.3

Re: [PATCH] drm/edid: drop the dcc probe before drm_do_get_edid

[Lubomir Rintel]

On Sat, 2018-02-10 at 19:03 +0100, Lubomir Rintel wrote:
> Now that drm_do_get_edid() handles override and firmware EDIDs it makes
> no sense to conditionalize it with a DCC probe. On the contrary -- the
> overrides are useful specifically when DCC is not functioning.
> 
> drm_do_get_edid() already bails out when DCC fails, there's no need for
> an extra check. It seems entirely safe to just drop this.
> 
> Working around the issue by forcing the connector on is not a good
> solution because that way the desktop shell can't tell when a monitor is
> plugged in or removed.
> 
> Cc: Jani Nikula <jani.nikula@intel.com>
> Fixes: 53fd40a ("drm: handle override and firmware EDID at drm_do_get_edid() level")

Please disregard this, should have given it more thought. Sorry.

It removes a symbol other modules use, and typoes DDC/DCC in the commit
message. Moreover it doesn't actually seem to solve my problem -- will
try to follow up with a better fix.

Lubo

Re: [PATCH] drm/edid: drop the dcc probe before drm_do_get_edid

[Jani Nikula]

On Sun, 11 Feb 2018, Lubomir Rintel <lkundrak@v3.sk> wrote:
> On Sat, 2018-02-10 at 19:03 +0100, Lubomir Rintel wrote:
>> Now that drm_do_get_edid() handles override and firmware EDIDs it makes
>> no sense to conditionalize it with a DCC probe. On the contrary -- the
>> overrides are useful specifically when DCC is not functioning.
>> 
>> drm_do_get_edid() already bails out when DCC fails, there's no need for
>> an extra check. It seems entirely safe to just drop this.
>> 
>> Working around the issue by forcing the connector on is not a good
>> solution because that way the desktop shell can't tell when a monitor is
>> plugged in or removed.
>> 
>> Cc: Jani Nikula <jani.nikula@intel.com>
>> Fixes: 53fd40a ("drm: handle override and firmware EDID at drm_do_get_edid() level")
>
> Please disregard this, should have given it more thought. Sorry.
>
> It removes a symbol other modules use, and typoes DDC/DCC in the commit
> message. Moreover it doesn't actually seem to solve my problem -- will
> try to follow up with a better fix.

I think the ddc probe was originally added for a reason. What that
reason was, I have no idea. I suggest you go on a bit of git blame/log
history digging to find what it was before removing.

Anyway, as you can see from the code, you can still use connector
forcing to bypass the ddc check.

BR,
Jani.


-- 
Jani Nikula, Intel Open Source Technology Center