From owner-freebsd-x11@freebsd.org Mon Nov 13 19:03:21 2017 Return-Path: Delivered-To: freebsd-x11@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id AEB6DDC1B38 for ; Mon, 13 Nov 2017 19:03:21 +0000 (UTC) (envelope-from pete@nomadlogic.org) Received: from vps-mail.nomadlogic.org (mail.nomadlogic.org [IPv6:2607:f2f8:a098::2]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 985B77917B for ; Mon, 13 Nov 2017 19:03:21 +0000 (UTC) (envelope-from pete@nomadlogic.org) Received: from creek.local (cpe-75-82-192-14.socal.res.rr.com [75.82.192.14]) by vps-mail.nomadlogic.org (OpenSMTPD) with ESMTPSA id aae886e6 TLS version=TLSv1.2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128 verify=NO; Mon, 13 Nov 2017 11:03:14 -0800 (PST) Subject: Re: Intel graphics, GPU Hung To: Ben Lavery-Griffiths , freebsd-x11@freebsd.org References: From: Pete Wright Message-ID: <2461a3eb-d754-3621-a59d-f77075a0ca0a@nomadlogic.org> Date: Mon, 13 Nov 2017 11:03:18 -0800 User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:52.0) Gecko/20100101 Thunderbird/52.4.0 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Content-Language: en-US X-BeenThere: freebsd-x11@freebsd.org X-Mailman-Version: 2.1.25 Precedence: list List-Id: X11 on FreeBSD -- maintaining and support List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 13 Nov 2017 19:03:21 -0000 On 11/12/17 9:32 AM, Ben Lavery-Griffiths wrote: > Hi all, > > Since upgrading to FreeBSD-RELEASE 11.1 I have noticed an intermittent issue where XDM isn’t displayed after booting. A mouse pointer usually appears in the middle of the screen and I’m able to control it for a short while (less than a minute). Switching back to the console I am greeted with the following message: > >> error: [drm:pid12:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung >> info: [drm] capturing error event; look for more information in sysctl hw.dri.0.info.i915_error_state >> error: [drm:pid0:i915_write64] *ERROR* Unknown unclaimed register before writing to 100000 >> error: [drm:pid908:i915_write32] *ERROR* Unknown unclaimed register before writing to 20a8 >> error: [drm:pid12:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung >> error: [drm:pid0:__gen6_gt_force_wake_mt_get] *ERROR* Timed out waiting for forcewake old ack to clear. >> error: [drm:pid12:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung >> error: [drm:pid0:i915_reset] *ERROR* GPU hanging too fast, declaring wedged! >> error: [drm:pid0:i915_reset] *ERROR* Failed to reset chip. > I have uploaded the following files to a gist on Github found here: https://gist.github.com/forquare/2a9e1c3e6a0ec32c3a27a768f1b50371 > > - dmesg.boot > - loader.conf > - messages > - rc.conf > - sysctl.conf > - Xorg.0.log > - Output of `sysctl hw.dri.0.info.i915_error_state` > > I’ve seen a forum post (https://forums.freebsd.org/threads/62831/) stating that it is an ‘issue’, but haven’t managed to find anything else. Is there some setting which is causing this? Or a regression that will likely be remain unfixed for a time? I don't have a 11.1 system handy at the moment, but one think that i think helped on my current system which ran into a similar issue was to define this in /boot/loader.conf: compat.linuxkpi.enable_hangcheck=0 kinda a shot in the dark, so I'm not too sure. -pete