<html> <head> <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1"> <meta name="Generator" content="Microsoft Exchange Server"> <style></style> </head> <body> <div> <div dir="auto">Well this is still a NAK. <div dir="auto"><br> </div> <div dir="auto">As stated previously please just don't remove DRM_AUTH and keep the functionality as it is.</div> <div dir="auto"><br> </div> <div dir="auto">I absolutely don't see the point to add a new flag to remove the same functionality a different flag provides.</div> <div dir="auto"><br> </div> <div dir="auto">Christian.</div> </div> <div class="x_gmail_extra"><br> <div class="x_gmail_quote">Am 03.07.2019 15:30 schrieb Emil Velikov <emil.l.velikov@gmail.com>:<br type="attribution"> </div> </div> </div> <font size="2"><span style="font-size:11pt;"> <div class="PlainText">From: Emil Velikov <emil.velikov@collabora.com><br> <br> With earlier commits we've removed DRM_AUTH for driver ioctls annotated<br> with DRM_AUTH | DRM_RENDER_ALLOW, as the protection it introduces is<br> effectively not existent.<br> <br> With next commit, we'll effectively do the same for DRM core.<br> <br> Yet the AMD developers have voiced concerns that by doing so, developers<br> working on the closed source user-space driver might remove render node<br> support.<br> <br> Since we do _not_ want that to happen, add workaround for those two<br> drivers<br> <br> Cc: Alex Deucher <alexander.deucher@amd.com><br> Cc: Christian König <christian.koenig@amd.com><br> Cc: amd-gfx@lists.freedesktop.org<br> Cc: Daniel Vetter <daniel@ffwll.ch><br> Signed-off-by: Emil Velikov <emil.velikov@collabora.com><br> ---<br> Christian, Alex this is the cleaner way to handle AMDGPU/radeon although<br> if you prefer alternative methods let me know.<br> <br> Review, acks and others are appreciated, since I'd like to get this<br> through the drm-misc tree.<br> <br> Thanks<br> Emil<br> <br> Unrelated:<br> The USE_AGP flag in AMDGPU should be nuked. While for radeon, one can<br> copy in the driver the 10-20 lines worth of agp_init/release and also<br> drop the flag.<br> <br> Bonus points of agp_init code gets a LEGACY check alongside the USE_AGP<br> one.<br> ---<br> drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 2 +-<br> drivers/gpu/drm/radeon/radeon_drv.c | 2 +-<br> include/drm/drm_drv.h | 10 ++++++++++<br> 3 files changed, 12 insertions(+), 2 deletions(-)<br> <br> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c<br> index 8e1b269351e8..cfc2ef11330c 100644<br> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c<br> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c<br> @@ -1307,7 +1307,7 @@ amdgpu_get_crtc_scanout_position(struct drm_device *dev, unsigned int pipe,<br> <br> static struct drm_driver kms_driver = {<br> .driver_features =<br> - DRIVER_USE_AGP | DRIVER_ATOMIC |<br> + DRIVER_USE_AGP | DRIVER_ATOMIC | DRIVER_FORCE_AUTH |<br> DRIVER_GEM |<br> DRIVER_RENDER | DRIVER_MODESET | DRIVER_SYNCOBJ,<br> .load = amdgpu_driver_load_kms,<br> diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c<br> index 4403e76e1ae0..5a1bfad1ad5e 100644<br> --- a/drivers/gpu/drm/radeon/radeon_drv.c<br> +++ b/drivers/gpu/drm/radeon/radeon_drv.c<br> @@ -538,7 +538,7 @@ radeon_get_crtc_scanout_position(struct drm_device *dev, unsigned int pipe,<br> <br> static struct drm_driver kms_driver = {<br> .driver_features =<br> - DRIVER_USE_AGP | DRIVER_GEM | DRIVER_RENDER,<br> + DRIVER_USE_AGP | DRIVER_GEM | DRIVER_RENDER | DRIVER_FORCE_AUTH,<br> .load = radeon_driver_load_kms,<br> .open = radeon_driver_open_kms,<br> .postclose = radeon_driver_postclose_kms,<br> diff --git a/include/drm/drm_drv.h b/include/drm/drm_drv.h<br> index b33f2cee2099..5fb2846396bc 100644<br> --- a/include/drm/drm_drv.h<br> +++ b/include/drm/drm_drv.h<br> @@ -92,6 +92,16 @@ enum drm_driver_feature {<br> * synchronization of command submission.<br> */<br> DRIVER_SYNCOBJ_TIMELINE = BIT(6),<br> + /**<br> + * @DRIVER_FORCE_AUTH:<br> + *<br> + * Driver mandates that DRM_AUTH is honoured, even if the same ioctl<br> + * is exposed via the render node - aka any of an "authentication" is<br> + * a fallacy.<br> + *<br> + * Used only by amdgpu and radeon. Do not use.<br> + */<br> + DRIVER_FORCE_AUTH = BIT(7),<br> <br> /* IMPORTANT: Below are all the legacy flags, add new ones above. */<br> <br> -- <br> 2.21.0<br> <br> </div> </span></font> </body> </html>