Skip to contentSkip to conversation

drm: set DRM_RENDER_ALLOW flag on DRM_IOCTL_MODE_CREATE/DESTROY_DUMB ioctls

PATCH In review Dongwon Kim 3 participants 4 messages Jun 10, 2021 – Jun 11, 2021

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm: set DRM_RENDER_ALLOW flag on DRM_IOCTL_MODE_CREATE/DESTROY_DUMB ioctls

Render clients should be able to create/destroy dumb object to import and use it as render buffer in case the default DRM device is different from the render device (i.e. kmsro).

---
 drivers/gpu/drm/drm_ioctl.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/drm_ioctl.c +2 -2
index 98ae00661656..f2f72e132741 100644
@@ -685,9 +685,9 @@ static const struct drm_ioctl_desc drm_ioctls[] = {
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb_ioctl, 0),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_MAP_DUMB, drm_mode_mmap_dumb_ioctl, 0),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_GETPROPERTIES, drm_mode_obj_get_properties_ioctl, 0),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_SETPROPERTY, drm_mode_obj_set_property_ioctl, DRM_MASTER),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR2, drm_mode_cursor2_ioctl, DRM_MASTER),
Signature · 1 line
2.20.1
#1 lore raw
Quoted 5 lines · Dongwon Kim

Render clients should be able to create/destroy dumb object to import and use it as render buffer in case the default DRM device is different from the render device (i.e. kmsro).

Uh no.

Well I know everyone just hacks around this, but the idea behind dumb buffer objects is that they're for kms scanout only. Furthermore on many drivers they allocate a limited resource like CMA memory. Handing that out like candy isn't a great idea.

And it's exactly those drivers that kmsro currently is used for where the display driver needs special memory.
-Daniel

Quoted 23 lines
---
 drivers/gpu/drm/drm_ioctl.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/drm_ioctl.c +2 -2
index 98ae00661656..f2f72e132741 100644
@@ -685,9 +685,9 @@ static const struct drm_ioctl_desc drm_ioctls[] = {
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb_ioctl, 0),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CREATE_DUMB, drm_mode_create_dumb_ioctl, DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_MAP_DUMB, drm_mode_mmap_dumb_ioctl, 0),
-	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, 0),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_GETPROPERTIES, drm_mode_obj_get_properties_ioctl, 0),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_SETPROPERTY, drm_mode_obj_set_property_ioctl, DRM_MASTER),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR2, drm_mode_cursor2_ioctl, DRM_MASTER),
Signature · 1 line
2.20.1
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
#2 lore raw
Quoted 10 lines · Daniel Vetter
Quoted 5 lines · Dongwon Kim

Render clients should be able to create/destroy dumb object to import and use it as render buffer in case the default DRM device is different from the render device (i.e. kmsro).

Uh no.

Quoted 7 lines

Well I know everyone just hacks around this, but the idea behind dumb buffer objects is that they're for kms scanout only. Furthermore on many drivers they allocate a limited resource like CMA memory. Handing that out like candy isn't a great idea.

And it's exactly those drivers that kmsro currently is used for where the display driver needs special memory.

Couldn't agree more. Perhaps we should add an inline comment and/or reference to a thread why?

-Emil

#3 lore raw

Understood. I saw weston client apps were failing to create render buffers from kmsro driver and found it was because they were not allowed to create and destroy dumb objects then I came up with this patch. I just thought it's the simplest solution. I didn't know it violates the rule. I think I should look into kmsro to make the client app to get the render buffer from ro-device instead. Thanks

Quoted 23 lines · Emil Velikov
Quoted 10 lines · Daniel Vetter
Quoted 5 lines · Dongwon Kim

Render clients should be able to create/destroy dumb object to import and use it as render buffer in case the default DRM device is different from the render device (i.e. kmsro).

Uh no.

Quoted 7 lines

Well I know everyone just hacks around this, but the idea behind dumb buffer objects is that they're for kms scanout only. Furthermore on many drivers they allocate a limited resource like CMA memory. Handing that out like candy isn't a great idea.

And it's exactly those drivers that kmsro currently is used for where the display driver needs special memory.

Couldn't agree more. Perhaps we should add an inline comment and/or reference to a thread why?

-Emil