On Wed, May 23, 2018 at 02:49:19AM -0700, Vito Caputo wrote: > On Mon, May 21, 2018 at 02:57:18PM -0700, Vito Caputo wrote: > > On Mon, May 21, 2018 at 12:53:20PM -0700, Vito Caputo wrote: > > > Hello all, > > > > > > 4.17-rc4 (my latest kernel ATM) consistently fails to start xgalaga > > > without -window. I will try find time to build the latest rc this > > > evening. > > > > > > > ~$ xgalaga > > > > X Error of failed request: BadValue (integer parameter out of range > > > > for operation) > > > > Major opcode of failed request: 152 (XFree86-VidModeExtension) > > > > Minor opcode of failed request: 10 (XF86VidModeSwitchToMode) > > > > Value in failed request: 0x120004e > > > > Serial number of failed request: 199 > > > > Current serial number in output stream: 203 > > > > > > Haven't dug into this much yet, only did a perfunctory check by booting > > > into a > > > few older kernels (4.11, 4.12, 4.16) and the problem is absent on all of > > > them. > > > It appears to be a 4.17-specific regression right now. > > > > > > Also observed, though this is surely a different regression, the game > > > ran like molasses with -window, showing some prominent kworkers in top: > > > > > > 692 vc 20 0 312852 45884 20556 R 32.0 1.2 0:08.69 Xorg > > > > > > 102 root 20 0 0 0 0 R 11.2 0.0 0:01.43 > > > kworker/1:3 > > > 94 root 20 0 0 0 0 I 8.9 0.0 0:00.83 > > > kworker/0:2 > > > 696 vc 20 0 39948 4124 2912 S 1.0 0.1 0:05.57 vwm > > > > > > 902 vc 30 10 46372 4144 3500 S 0.7 0.1 0:00.08 > > > xgalaga > > > 891 vc 30 10 44924 3868 3156 R 0.3 0.1 0:00.09 top > > > > > > 903 vc 30 10 4180 1184 1100 S 0.3 0.0 0:00.01 > > > xgal.sndsrv.oss > > > > > > The windowed performance issue was observed on the older kernels tested > > > as well, though 4.11 felt better and didn't have the busy kworkers. > > > > > > I have not attempted to play xgalaga for ages, but it used to be perfectly > > > playable on this machine in windowed mode when I last did. > > > > > > Machine is the venerable Thinkpad X61s, 1.8Ghz, Debian 9, config attached. > > > > > > > Just built and booted v4.17-rc6, still broken. > > Bisected to: > > e995ca0b8139c5f6807095464e969931b443f55a is the first bad commit > commit e995ca0b8139c5f6807095464e969931b443f55a > Author: Ville Syrjälä <[email protected]> > Date: Tue Nov 14 20:32:58 2017 +0200 > > drm/i915: Provide a device level .mode_valid() hook > > We never support certain mode flags etc. Reject those early on in the > mode_config.mode_valid() hook. That allows us to remove some duplicated > checks from the connector .mode_valid() hooks, and it guarantees that > we never see those flags even from user mode as the > mode_config.mode_valid() hooks gets executed for those as well. > > Signed-off-by: Ville Syrjälä <[email protected]> > Link: > https://patchwork.freedesktop.org/patch/msgid/[email protected] > Reviewed-by: Daniel Vetter <[email protected]>
Hmm. I guess xgalaga passes some garbage in via xf86vidmode which the ddx doesn't validate before passing it on to the kernel. So far I can't reproduce the problem here unfortnately. Can you try the following patch and reproduce the problem with drm.debug=0xe passed to the kernel so that we can seewhat the bad modeline looks like? >From c8b4eaaf3ee8e796dbbb4a5080c1aec2b435d110 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= <[email protected]> Date: Wed, 23 May 2018 16:12:08 +0300 Subject: [PATCH] drm/modes: Dump the rejected user mode MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Print put the modeline When we reject a user mode. Cc: Vito Caputo <[email protected]> Signed-off-by: Ville Syrjälä <[email protected]> --- drivers/gpu/drm/drm_modes.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c index c78ca0e84ffd..ca50b12d9701 100644 --- a/drivers/gpu/drm/drm_modes.c +++ b/drivers/gpu/drm/drm_modes.c @@ -1739,8 +1739,11 @@ int drm_mode_convert_umode(struct drm_device *dev, } out->status = drm_mode_validate_driver(dev, out); - if (out->status != MODE_OK) + if (out->status != MODE_OK) { + DRM_DEBUG_KMS("Bad user mode:\n"); + drm_mode_debug_printmodeline(out); return -EINVAL; + } drm_mode_set_crtcinfo(out, CRTC_INTERLACE_HALVE_V); -- 2.16.1 -- Ville Syrjälä Intel
