Hi

Am 31.03.20 um 10:50 schrieb Daniel Vetter:
> On Tue, Mar 31, 2020 at 10:11:30AM +0200, Thomas Zimmermann wrote:
>> Calling the VRAM helper's prepare_fb() helper now sets the plane's
>> fence object. The helper still has to synchronize with other users
>> of the GEM object. Leave a related TODO comment in the code.
>>
>> This will be useful for PRIME support. VRAM helpers don't support
>> buffer sharing ATM, so there are no drivers requiring this change.
>>
>> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
>> ---
>>  drivers/gpu/drm/drm_gem_vram_helper.c | 18 +++++++++++++++---
>>  1 file changed, 15 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
>> index b3201a70cbfcb..d4e4f80d3a6c1 100644
>> --- a/drivers/gpu/drm/drm_gem_vram_helper.c
>> +++ b/drivers/gpu/drm/drm_gem_vram_helper.c
>> @@ -5,6 +5,7 @@
>>  #include <drm/drm_drv.h>
>>  #include <drm/drm_file.h>
>>  #include <drm/drm_framebuffer.h>
>> +#include <drm/drm_gem_framebuffer_helper.h>
>>  #include <drm/drm_gem_ttm_helper.h>
>>  #include <drm/drm_gem_vram_helper.h>
>>  #include <drm/drm_mode.h>
>> @@ -670,9 +671,9 @@ EXPORT_SYMBOL(drm_gem_vram_driver_dumb_mmap_offset);
>>   * @plane:	a DRM plane
>>   * @new_state:	the plane's new state
>>   *
>> - * During plane updates, this function pins the GEM VRAM
>> - * objects of the plane's new framebuffer to VRAM. Call
>> - * drm_gem_vram_plane_helper_cleanup_fb() to unpin them.
>> + * During plane updates, this function sets the plane's fence and
>> + * pins the GEM VRAM objects of the plane's new framebuffer to VRAM.
>> + * Call drm_gem_vram_plane_helper_cleanup_fb() to unpin them.
>>   *
>>   * Returns:
>>   *	0 on success, or
>> @@ -689,6 +690,13 @@ drm_gem_vram_plane_helper_prepare_fb(struct drm_plane *plane,
>>  	if (!new_state->fb)
>>  		return 0;
>>  
>> +	/*
>> +	 * TODO: Synchronize with other users of the buffer. Buffers
>> +	 *       cannot be pinned to VRAM while they are in use by other
>> +	 *       drivers for DMA. We should probably wait for each GEM
>> +	 *       object's fence before attempting to pin the buffer.
>> +	 *       There are currently no users of this functionality.
>> +	 */
> 
> Not sure this is warranted, we have lots of drivers with these kind of
> restrictions ... It's a big can of worms, I'd just leave this all out.

Well, OK. Adding PRIME support would require a rework of these helpers
anyway.

Best regards
Thomas

> 
>>  	for (i = 0; i < ARRAY_SIZE(new_state->fb->obj); ++i) {
>>  		if (!new_state->fb->obj[i])
>>  			continue;
>> @@ -698,6 +706,10 @@ drm_gem_vram_plane_helper_prepare_fb(struct drm_plane *plane,
>>  			goto err_drm_gem_vram_unpin;
>>  	}
>>  
>> +	ret = drm_gem_fb_prepare_fb(plane, new_state);
>> +	if (ret)
>> +		goto err_drm_gem_vram_unpin;
> 
> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> 
>> +
>>  	return 0;
>>  
>>  err_drm_gem_vram_unpin:
>> -- 
>> 2.26.0
>>
> 

-- 
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Felix Imendörffer