Re: [PATCH 08/13] drm/vkms: Allow to configure multiple CRTCs

[Louis Chauvet <louis.chauvet@bootlin.com>]

Le 11/02/2025 à 11:44, José Expósito a écrit :
> On Thu, Jan 30, 2025 at 02:48:20PM +0100, Louis Chauvet wrote:
>> On 29/01/25 - 12:00, José Expósito wrote:
>>> Add a list of CRTCs to vkms_config and helper functions to add and
>>> remove as many CRTCs as wanted.
>>>
>>> For backwards compatibility, add one CRTC to the default configuration.
>>>
>>> A future patch will allow to attach planes and CRTCs, but for the
>>> moment there are no changes in the way the output is configured.
>>>
>>> Signed-off-by: Louis Chauvet <louis.chauvet@bootlin.com>
>>> Signed-off-by: José Expósito <jose.exposito89@gmail.com>
>>
>> Co-developped-by: Louis Chauvet <louis.chauvet@bootlin.com>
>> Signed-off-by: Louis Chauvet <louis.chauvet@bootlin.com>
>> Signed-off-by: José Expósito <jose.exposito89@gmail.com>
>>
>> [...]
>>
>>> diff --git a/drivers/gpu/drm/vkms/tests/vkms_config_test.c b/drivers/gpu/drm/vkms/tests/vkms_config_test.c
>>
>> [...]
>>
>>> +static void vkms_config_test_valid_crtc_number(struct kunit *test)
>>> +{
>>> +	struct vkms_config *config;
>>> +	struct vkms_config_crtc *crtc_cfg;
>>> +	int n;
>>> +
>>> +	config = vkms_config_default_create(false, false, false);
>>> +	KUNIT_ASSERT_NOT_ERR_OR_NULL(test, config);
>>> +
>>> +	/* Invalid: No CRTCs */
>>> +	crtc_cfg = list_first_entry(&config->crtcs, typeof(*crtc_cfg), link);
>>> +	vkms_config_destroy_crtc(config, crtc_cfg);
>>> +	KUNIT_EXPECT_FALSE(test, vkms_config_is_valid(config));
>>> +
>>> +	/* Invalid: Too many CRTCs */
>>> +	for (n = 0; n <= 32; n++)
>>> +		vkms_config_add_crtc(config);
>>> +
>>> +	KUNIT_EXPECT_FALSE(test, vkms_config_is_valid(config));
>>> +
>>> +	vkms_config_destroy(config);
>>> +}
>>
>> Same as before, can you rename the fonction to
>> vkms_config_test_invalid_crtc_number
>>
>> [...]
>>
>>> diff --git a/drivers/gpu/drm/vkms/vkms_config.c b/drivers/gpu/drm/vkms/vkms_config.c
>>
>> [...]
>>
>>> +struct vkms_config_crtc **vkms_config_get_crtcs(const struct vkms_config *config,
>>> +						size_t *out_length)
>>> +{
>>> +	struct vkms_config_crtc **array;
>>> +	struct vkms_config_crtc *crtc_cfg;
>>> +	size_t length;
>>> +	int n = 0;
>>> +
>>> +	length = list_count_nodes((struct list_head *)&config->crtcs);
>>> +	if (length == 0) {
>>> +		*out_length = length;
>>> +		return NULL;
>>> +	}
>>> +
>>> +	array = kmalloc_array(length, sizeof(*array), GFP_KERNEL);
>>> +	if (!array)
>>> +		return ERR_PTR(-ENOMEM);
>>> +
>>> +	list_for_each_entry(crtc_cfg, &config->crtcs, link) {
>>> +		array[n] = crtc_cfg;
>>> +		n++;
>>> +	}
>>> +
>>> +	*out_length = length;
>>> +	return array;
>>> +}
>>
>> Same as before, can't we use an iterator?
>>
>> [...]
>>
>>> +static bool valid_crtc_number(struct vkms_config *config)
>>> +{
>>> +	size_t n_crtcs;
>>> +
>>> +	n_crtcs = list_count_nodes(&config->crtcs);
>>> +	if (n_crtcs <= 0 || n_crtcs >= 32) {
>>> +		pr_err("The number of CRTCs must be between 1 and 31\n");
>>
>> I agree we need some logs, but I think pr_err is too agressive (i.e may
>> be considered as an error by some test tools).
>>
>> I think we should at least:
>> - lower to warn/notice/info
>> - use drm variants of the macro
>>
>>> +		return false;
>>> +	}
>>> +
>>> +	return true;
>>> +}
>>> +
>>
>> [...]
>>
>>> +struct vkms_config_crtc *vkms_config_add_crtc(struct vkms_config *config)
>>> +{
>>> +	struct vkms_config_crtc *crtc_cfg;
>>> +
>>> +	crtc_cfg = kzalloc(sizeof(*crtc_cfg), GFP_KERNEL);
>>> +	if (!crtc_cfg)
>>> +		return ERR_PTR(-ENOMEM);
>>> +
>>> +	vkms_config_crtc_set_writeback(crtc_cfg, false);
>>> +
>>> +	list_add_tail(&crtc_cfg->link, &config->crtcs);
>>> +
>>> +	return crtc_cfg;
>>> +}
>>> +
>>> +void vkms_config_destroy_crtc(struct vkms_config *config,
>>> +			      struct vkms_config_crtc *crtc_cfg)
>>> +{
>>> +	list_del(&crtc_cfg->link);
>>> +	kfree(crtc_cfg);
>>> +}
>>
>> Same as before, the pair add/destroy seems strange.
>>
>>> +++ b/drivers/gpu/drm/vkms/vkms_drv.c
>>> @@ -181,7 +181,8 @@ static int vkms_create(struct vkms_config *config)
>>>   		goto out_devres;
>>>   	}
>>>   
>>> -	ret = drm_vblank_init(&vkms_device->drm, 1);
>>> +	ret = drm_vblank_init(&vkms_device->drm,
>>> +			      vkms_config_get_num_crtcs(config));
>>
>> At this point we only create one crtc, can you move this change in the
>> commit where you create multiple crtc?
> 
> Since the code to add more than one CRTCs is unused but technically present, I
> think that this is the right patch to use this function.

I don't totally agree: you can create multiple vkms_config_crtc, but the 
code only creates one drm_crtc.

For me it is more logical to keep 1 here, as the rest of the code only 
creates one CRTC. With the next patch, the code will create more than 
one CRTC, so it makes sense to use vkms_config_get_num_crtcs.

It is not a strong blocking point, but if a v3 is needed, could you 
change it?

> Anyway, at the moment it'll always return 1, so it is a no-op.

The current user is only default_config, so yes I agree, it is always 1.

> I didn't change it in v2, let me know if it works for you.
> 
> Thanks,
> Jose
> 
>>>   	if (ret) {
>>>   		DRM_ERROR("Failed to vblank\n");
>>>   		goto out_devres;
>>> -- 
>>> 2.48.1
>>>

-- 
Louis Chauvet, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com