https://bugs.freedesktop.org/show_bug.cgi?id=73200
--- Comment #7 from Emil Velikov <emil.l.veli...@gmail.com> --- (In reply to comment #6) > Created attachment 93520 [details] [review] > cache dri fd > > Based on the radeon version, cache inode major/minor to only create a single > screen. > > Does this fix things? Haven't tested it, but noticed a couple of interesting things * util_hash_table_create can fail fd_tab = util_hash_table_create(hash_fd, compare_fd); + assert(fd_tab); * if screen fails, we are at the mercy of the compiler if we null ptr deref or not + if (screen) + util_hash_table_set(fd_tab, intptr_to_pointer(fd), screen); + return &screen->base; + if (!screen) + return NULL; + util_hash_table_set(fd_tab, intptr_to_pointer(fd), screen); + return &screen->base; * nouveau_drm_screen_unref might need a prototype to build. * drop the fd_tab check in nouveau_drm_screen_unref. it must be non NULL at this point * fold nouveau_drm_screen_unref uses into nouveau_screen_fini() -- You are receiving this mail because: You are the assignee for the bug.
_______________________________________________ Nouveau mailing list Nouveau@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/nouveau