It doesn't actually make any sense for a kms-only ddx. And it causes problems with non-PCI devices, because some xservers will skip trying to claim the device as a regular device if it is not a PCI device. The following two xserver patches should fix this:
http://patchwork.freedesktop.org/patch/28494/ http://patchwork.freedesktop.org/patch/28495/ but since a display-only non-display device makes no sense, let's just nuke that code to avoid being a problem. Signed-off-by: Rob Clark <[email protected]> --- src/driver.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/src/driver.c b/src/driver.c index d99b027..46ff0da 100644 --- a/src/driver.c +++ b/src/driver.c @@ -351,13 +351,9 @@ ms_platform_probe(DriverPtr driver, { ScrnInfoPtr scrn = NULL; const char *path = xf86_get_platform_device_attrib(dev, ODEV_ATTRIB_PATH); - int scr_flags = 0; - - if (flags & PLATFORM_PROBE_GPU_SCREEN) - scr_flags = XF86_ALLOCATE_GPU_SCREEN; if (probe_hw(path, dev)) { - scrn = xf86AllocateScreen(driver, scr_flags); + scrn = xf86AllocateScreen(driver, 0); xf86AddEntityToScreen(scrn, entity_num); scrn->driverName = "modesetting"; -- 1.9.3 _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
