Re: [RFC PATCH 13/13] HACK: schedule fence return on main AIO context

2023-04-25 Thread Stefan Hajnoczi
On Fri, 21 Apr 2023 at 19:21, Gurchetan Singh wrote: > > On Fri, Apr 21, 2023 at 9:00 AM Stefan Hajnoczi wrote: > > > > On Thu, 20 Apr 2023 at 21:13, Gurchetan Singh > > wrote: > > > > > > gfxstream and both cross-domain (and even newer versions > > > virglrenderer: see VIRGL_RENDERER_ASYNC_FENC

Re: [RFC PATCH 13/13] HACK: schedule fence return on main AIO context

2023-04-21 Thread Gurchetan Singh
On Fri, Apr 21, 2023 at 9:00 AM Stefan Hajnoczi wrote: > > On Thu, 20 Apr 2023 at 21:13, Gurchetan Singh > wrote: > > > > gfxstream and both cross-domain (and even newer versions > > virglrenderer: see VIRGL_RENDERER_ASYNC_FENCE_CB) like to signal > > fence completion on threads ("callback thread

Re: [RFC PATCH 13/13] HACK: schedule fence return on main AIO context

2023-04-21 Thread Stefan Hajnoczi
On Thu, 20 Apr 2023 at 21:13, Gurchetan Singh wrote: > > gfxstream and both cross-domain (and even newer versions > virglrenderer: see VIRGL_RENDERER_ASYNC_FENCE_CB) like to signal > fence completion on threads ("callback threads") that are > different from the thread that processes the command qu

[RFC PATCH 13/13] HACK: schedule fence return on main AIO context

2023-04-20 Thread Gurchetan Singh
gfxstream and both cross-domain (and even newer versions virglrenderer: see VIRGL_RENDERER_ASYNC_FENCE_CB) like to signal fence completion on threads ("callback threads") that are different from the thread that processes the command queue ("main thread"). This is generally possible with locking, a