[Bug 99028] New: [KVM][GVT-d] [BDW & SKL ]Ubuntu 16.04 guest boot up with GPU Hang with drm-intel kernel

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=99028

            Bug ID: 99028
           Summary: [KVM][GVT-d] [BDW & SKL ]Ubuntu 16.04 guest boot up
                    with GPU Hang with drm-intel kernel
           Product: DRI
           Version: DRI git
          Hardware: x86-64 (AMD64)
                OS: Linux (All)
            Status: NEW
          Severity: major
          Priority: medium
         Component: DRM/Intel
          Assignee: intel-gfx-bugs at lists.freedesktop.org
          Reporter: terrence.xu at intel.com
        QA Contact: intel-gfx-bugs at lists.freedesktop.org
                CC: gordon.jin at intel.com,
                    intel-gfx-bugs at lists.freedesktop.org,
                    jani.saarinen at intel.com, zhiyuan.lv at intel.com
     i915 platform: BDW, SKL

System Environment
=======
Host kernel repo: kvm.git
Host commit: master-813ae37e
Guest repo: drm-intel.git
Guest commit: drm-intel-next-queued-312c3c46

Regression?
=======
No

Bug detailed description
=======
The guest boot up with the drm-intel 4.8.0-rc2+ kernel, with GPU hang be found.
The guest boot up with the latest drm-intel 4.9.0-rc4+ kernel with kernel panic
(as https://bugs.freedesktop.org/show_bug.cgi?id=99025), but if we try to
remove the kernel panic error handling manually, the GPU hang issue also won
existed.
This is KVM GVT-d environment issue.

Reproduce Steps
==============
Boot up Ubuntu 16.04 guest with the drm-intel kernel, the command as below:
qemu-system-x86_64 --enable-kvm -m 2048 -smp 4 -hda /root/ubuntu-16.04.img -usb
-usbdevice tablet -device virtio-net-pci,netdev=nic0,mac=00:16:3e:60:0a:50
-netdev tap,id=nic0,script=/etc/kvm/qemu-ifup -serial stdio

Expected Result
=============
Guest boot up without any issues.

Actual Result
===========
Guest boot up with GPU hang.

Analysis & Root Cause
===================
[    6.939377] [drm] GPU HANG: ecode 8:0:0xfffffffe, reason: No progress on
rend er ring, action: reset
[    6.939378] [drm] GPU hangs can indicate a bug anywhere in the entire gfx
stack, including userspace.
[    6.939379] [drm] Please file a _new_ bug report on bugs.freedesktop.org
against DRI -> DRM/Intel
[    6.939379] [drm] drm/i915 developers can then reassign to the right compone
t if it's not a kernel issue.
[    6.939380] [drm] The gpu crash dump is required to analyze gpu hangs, so
please always attach it.
[    6.939380] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[    6.939429] drm/i915: Resetting chip after gpu hang
[    7.869346] EXT4-fs (sda1): mounted filesystem with ordered data mode. Opts:
   (null)
[    7.964728] systemd[1]: systemd 229 running in system mode. (+PAM +AUDIT
+SELINUX +IMA +APPARMOR +SMACK +SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS
+ACL +                                                                         
                                           XZ -LZ4 +SECCOMP +BLKID +ELFUTILS
+KMOD -IDN)
[    7.964756] systemd[1]: Detected virtualization qemu.
[    7.964758] systemd[1]: Detected architecture x86-64.
[    7.965211] systemd[1]: Set hostname to <gvt-ub16>.
[    8.092829] systemd[1]: Listening on fsck to fsckd communication Socket.
[    8.092875] systemd[1]: Reached target User and Group Name Lookups.
[    8.092882] systemd[1]: Reached target Remote File Systems (Pre).
[    8.092902] systemd[1]: Listening on udev Kernel Socket.
[    8.092945] systemd[1]: Created slice System Slice.
[    8.092976] systemd[1]: Listening on Syslog Socket.
[    8.278376] EXT4-fs (sda1): re-mounted. Opts: errors=remount-ro
[    8.349290] systemd-journald[205]: Received request to flush runtime journal
from PID 1
[    8.481575] random: crng init done
[    8.889146] drm/i915: Resetting chip after gpu hang
[    9.060874] Adding 4000764k swap on /dev/sda5.  Priority:-1 extents:1
across:  4000764k
[   10.364525] e1000: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control:
RX
[   10.873269] drm/i915: Resetting chip after gpu hang
[   12.923317] drm/i915: Resetting chip after gpu hang
[   14.905157] drm/i915: Resetting chip after gpu hang
[   16.900984] drm/i915: Resetting chip after gpu hang
[   18.885876] drm/i915: Resetting chip after gpu hang
[   20.921190] drm/i915: Resetting chip after gpu hang
[   22.905154] drm/i915: Resetting chip after gpu hang

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20161208/71b216c0/attachment.html>