From: Konrad Rzeszutek Wilk Date: Fri, 18 Dec 2015 17:38:23 +0000 (-0500) Subject: Revert "xen/fb: allow xenfb initialization for hvm guests" X-Git-Tag: v4.1.12-92~208^2~1^2 X-Git-Url: https://www.infradead.org/git/?a=commitdiff_plain;h=716ff9b9463d973c16691988acdc0b6e47a9e903;p=users%2Fjedix%2Flinux-maple.git Revert "xen/fb: allow xenfb initialization for hvm guests" This reverts commit d9581c7dcac15c02ad4d47c60c60f4d8f197db55. which is: Author: Stefano Stabellini Date: Fri Jan 3 19:02:09 2014 +0000 xen/fb: allow xenfb initialization for hvm guests There is no reasons why an HVM guest shouldn't be allowed to use xenfb. as "Xend" exposes an PV and VGA (via QEMU) driver to the guests (either PV or HVM). The backed in this case is QEMU - and it only provides the emulation for the VGA backend (for HVM guests). Upstream QEMU can do both - but since we are using Xend - we only expose QEMU-traditional. Upstreamwise we hadn't reached a good decisions. Poking for XenStore keys to see if the backend is Xend vs xl seems odd. And the issue goes away in 'xl' as you can't define both PV and VGA drivers. Only one of them is allowed. The 'proper' fix would be to teach Xend to not expose PV FB for HVM guests but that code is pretty genric. So continuing on with this revert. OraBug: 20386370 - XENBUS_PROBE_FRONTEND: TIMEOUT CONNECTING TO DEVICE EROR WITH UEK4 Signed-off-by: Konrad Rzeszutek Wilk --- diff --git a/drivers/video/fbdev/xen-fbfront.c b/drivers/video/fbdev/xen-fbfront.c index 09dc44736c1ac..b35bbf6c490d1 100644 --- a/drivers/video/fbdev/xen-fbfront.c +++ b/drivers/video/fbdev/xen-fbfront.c @@ -694,7 +694,7 @@ static struct xenbus_driver xenfb_driver = { static int __init xenfb_init(void) { - if (!xen_domain()) + if (!xen_pv_domain()) return -ENODEV; /* Nothing to do if running in dom0. */