2018-04-27 1:34 GMT+08:00 Rob Herring <r...@kernel.org>: > On Thu, Apr 26, 2018 at 11:56 AM, Emil Velikov <emil.l.veli...@gmail.com> > wrote: >> On 26 April 2018 at 03:30, Chih-Wei Huang <cwhu...@linux.org.tw> wrote: >>> 2018-04-25 19:55 GMT+08:00 Robert Foss <robert.f...@collabora.com>: >>>> Hey Emil & Chih-Wei, >>>> >>>> On 04/24/2018 01:59 PM, Emil Velikov wrote: >>>>> >>>>> On 24 April 2018 at 12:28, Emil Velikov <emil.l.veli...@gmail.com> wrote: >>>>> >>>>>> On the topic of keeping the old code behind a #define or just removing >>>>>> it, it'll be great if interested parties can reach a consensus. >>>>>> >>>>> Actually one can simply drop this code and drm_gralloc users can add a >>>>> drm_ioctl_permit() hack. >>>>> Namely: loosen the restrictions to consider render nodes identical to >>>>> primary/card ones. >>>>> >>>>> Yes, it's a nasty hack, yet no worse than the existing one that >>>>> removes the auth :-\ >>>> >>>> I'm fine with adding a #define. >>>> Chih-Wei: Do you have any objections? >>> >>> "keeping the old code behind a #define"? >>> Sounds good to me. Thank you! >> >> Having a look at the Android-x86 kernel: >> The current hack in the permission check (see drm_ioctl_permit in [1]) >> effectively threats render nodes and primary nodes as the same thing. >> >> Thus, on top of the current patch the !dri2_dpy->is_render_node check >> should be removed, for Android-x86. >> There's no need to keep the flink code around ;-) > > Except I don't think drm_gralloc fills in a dma-buf fd in it's > handle... There was an AOSP version that did IIRC. > > Or maybe I'm missing something.
IIRC, AOSP's modified drm_gralloc requires drm_hwcomposer. Then we loop back to the old problem: drm_hwcomposer doesn't work for most x86 GPUs. _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev