Re: [PATCH RFC v2 17/18] drm/vkms: Introduce config for connector EDID

[Louis Chauvet]

Le 25 novembre 2024 09:45:04 UTC, Jani Nikula <jani.nikula at linux.intel.com> a écrit :
>On Fri, 22 Nov 2024, Louis Chauvet <louis.chauvet at bootlin.com> wrote:
>> To properly test the EDID reading without using the DRM override, add an
>> option to configure the EDID for a connector.
>>
>> Signed-off-by: Louis Chauvet <louis.chauvet at bootlin.com>
>> ---
>>  drivers/gpu/drm/vkms/vkms_config.c |  1 +
>>  drivers/gpu/drm/vkms/vkms_config.h |  2 ++
>>  drivers/gpu/drm/vkms/vkms_output.c | 37 ++++++++++++++++++++++++++++++++++---
>>  3 files changed, 37 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/vkms/vkms_config.c b/drivers/gpu/drm/vkms/vkms_config.c
>> index ac1a9658c5075c118d59da965ca3392355ccb2b2..1a1234d4f10fa8e5ea6bd649139ecc10c991f875 100644
>> --- a/drivers/gpu/drm/vkms/vkms_config.c
>> +++ b/drivers/gpu/drm/vkms/vkms_config.c
>> @@ -199,6 +199,7 @@ struct vkms_config_connector *vkms_config_create_connector(struct vkms_config *v
>>  	xa_init_flags(&vkms_config_connector->possible_encoders, XA_FLAGS_ALLOC);
>>  	vkms_config_connector->type = DRM_MODE_CONNECTOR_VIRTUAL;
>>  	vkms_config_connector->status = connector_status_unknown;
>> +	vkms_config_connector->edid_blob_len = 0;
>>  
>>  	return vkms_config_connector;
>>  }
>> diff --git a/drivers/gpu/drm/vkms/vkms_config.h b/drivers/gpu/drm/vkms/vkms_config.h
>> index bba56c9d8aeceac97a4339ef42ab663c5dc54e65..1220b16f6c98d1ebb0ae55d662a84fe25e1a6a02 100644
>> --- a/drivers/gpu/drm/vkms/vkms_config.h
>> +++ b/drivers/gpu/drm/vkms/vkms_config.h
>> @@ -112,6 +112,8 @@ struct vkms_config_connector {
>>  	struct xarray possible_encoders;
>>  	int type;
>>  	enum drm_connector_status status;
>> +	char edid_blob[PAGE_SIZE];
>> +	int edid_blob_len;
>>  
>>  	/* Internal usage */
>>  	struct drm_connector *connector;
>> diff --git a/drivers/gpu/drm/vkms/vkms_output.c b/drivers/gpu/drm/vkms/vkms_output.c
>> index fc6a0cdade0739b94820ed4e0924cf355137fe79..56590afb33d75465971d10a282040690840cdbee 100644
>> --- a/drivers/gpu/drm/vkms/vkms_output.c
>> +++ b/drivers/gpu/drm/vkms/vkms_output.c
>> @@ -31,13 +31,44 @@ static const struct drm_connector_funcs vkms_connector_funcs = {
>>  	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
>>  };
>>  
>> +static int vkms_connector_read_block(void *context, u8 *buf, unsigned int block, size_t len)
>> +{
>> +	struct vkms_config_connector *config = context;
>> +
>> +	if (block * len + len > config->edid_blob_len)
>
>The parameters to the read block function are a bit weird for historical
>reasons. The start offset is indicated by block number, length by
>len. The start byte offset is thus block * EDID_LENGTH! There's no
>smaller granularity for start offset. However len can be < EDID_LENGTH!
>
>So the above should be (block * EDID_LENGTH + len > edid_blob_len)
>
>> +		return 1;
>> +	memcpy(buf, &config->edid_blob[block * len], len);
>
>And this should be &config->edid_blob[block * EDID_LENGTH].
>
>(Your patch would work, but just by coincidence due to the way the read
>block function is currently called.)

Thanks for those clarifications! 

>> +	return 0;
>> +}
>> +
>>  static int vkms_conn_get_modes(struct drm_connector *connector)
>>  {
>> +	const struct drm_edid *drm_edid = NULL;
>>  	int count;
>> +	struct vkms_config_connector *connector_cfg;
>> +	struct vkms_device *vkmsdev = drm_device_to_vkms_device(connector->dev);
>> +	struct vkms_config_connector *context = NULL;
>> +
>> +	list_for_each_entry(connector_cfg, &vkmsdev->config->connectors, link) {
>> +		if (connector_cfg->connector == connector) {
>> +			context = connector_cfg;
>> +			break;
>> +		}
>> +	}
>> +	if (context)
>> +		drm_edid = drm_edid_read_custom(connector, vkms_connector_read_block, context);
>
>Thanks for using drm_edid_read_custom() for this btw!
>
>> +
>> +	/*
>> +	 * Unconditionally update the connector. If the EDID was read
>> +	 * successfully, fill in the connector information derived from the
>> +	 * EDID. Otherwise, if the EDID is NULL, clear the connector
>> +	 * information.
>> +	 */
>> +	drm_edid_connector_update(connector, drm_edid);
>> +
>> +	count = drm_edid_connector_add_modes(connector);
>>  
>> -	/* Use the default modes list from DRM */
>> -	count = drm_add_modes_noedid(connector, XRES_MAX, YRES_MAX);
>> -	drm_set_preferred_mode(connector, XRES_DEF, YRES_DEF);
>
>I don't really know anything about your use case, but don't you want to
>fall back to the above for drm_edid == NULL? *shrug*

You are right, I will probably fallback to noedid version, at least for VIRTUAL connectors.

Or maybe I will add something in VKMS configuration to use "empty edid", "default modes_noedid" or "custom_edid".

This way you will be able to test all the scenarios for your userspace.

Thanks for your review,
Louis Chauvet

>BR,
>Jani.
>
>> +	drm_edid_free(drm_edid);
>>  
>>  	return count;
>>  }
>