System Environment: -------------------------- Platform: SugarBay(only) Libdrm: (master)2.4.23-8-g36d4939343d8789d9066f7245fa2d4fe69119dd8 Mesa: (7.10)cc1636b6db85604510f97f8a37d7fd0ecf453866 Xserver: (master)xorg-server-1.9.99.901-197-gea1ffd3e60bdcedbec5a6f28929f8677bf45d450 Xf86_video_intel: (master)2.14.0-26-g23f9b14df7c102c1036134835dd5d1a508059858 Cairo: (master)d866bb0405bc989666f886e2c0e293a1fe718d74 Kernel: (drm-intel-fixes)04dbff52600719017598f7439bf42e5a72e7de3b Bug detailed description: ------------------------- System hangs when open Terminal([Applications]-[System Tools]-[Terminal])in gnome-desktop on SugarBay(DO:i5-2500 D2 (id=0x0102, rev 09), Q67 B1 (rev 03) and Host Bridge id=0x0100 (rev09) ). In particular, 1.It works fine on HuronRiver and other SugarBay(DH:i7-2600 D2 (id=0x0102, rev 09), H67 B1 (Intel DH67CL rev 03), and Host Bridge id=0x0100 (rev09)). 2.It's kernel regression. In drm-intel-fixes branch, 3c0eee3fe6a3a1c745379547c7e7c904aa64f6d5 is the good commit, and 24b0b53ca89ba3a568d53065cc336061029a7321 is the bad one. 3.Message in console when system hangs: [mi] EQ overflowing. The server is probably stuck in an infinite loop. Backtrace: 0: X (xorg_backtrace+0x28) [0x457868] 1: X (mieqEnqueue+0x1f4) [0x4576e4] 2: X (xf86PostMotionEventM+0x97) [0x4753c7 3: /opt/X11R7/lib/xorg/modules/input/evdev_drv.so (0x7fbf6b178000+0x5411) [0x7fbf6b17d411] 4: X (0x400000+0x686c7) [0x4686c7] 5: X (0x400000+0x117b33) [0x517b33] 6: /lib64/libpthread.so.0 (0x3dbfe00000+0xf440) [0x3dbfe0f440] 7: /opt/X11R7/lib/libpixman-1.so.0 (0x7fbf6daf8000+0x57646) [0x7fbf6db4f646] 8: /opt/X11R7/lib/libpixman-1.so.0 (pixman_image_composite32+0x434) [0x7fbf6db2b2b4] 9: /opt/X11R7/lib/xorg/modules/libfb.so (fbComposite+0x220) [0x7fbf6c1a3560] 10: /opt/X11R7/lib/xorg/modules/drivers/intel_drv.so (0x7fbf6c5b9000+0x3043c) [0x7fbf6c5e943c] 11: /opt/X11R7/lib/xorg/modules/drivers/intel_drv.so (0x7fbf6c5b9000+0x2ee0e) [0x7fbf6c5e7e0e] 12: X (0x400000+0xdd230) [0x4dd230] 13: X (0x400000+0xd4f59) [0x4d4f59] 14: X (0x400000+0x54691) [0x454691] 15: X (0x400000+0x2163b) [0x42163b] 16: /lib64/libc.so.6 (__libc_start_main+0xfd) [0x3dbf61ec5d] 17: X (0x400000+0x211c9) [0x4211c9] ** (deja-dup-monitor:2964): DEBUG: monitor.vala:263: Invalid next run date. Not scheduling a backup. Reproduce steps: ---------------- 1.gnome-session 2.open and close Terminal
Puzzled by the bisection range drm-intel-fixes/drm-intel-next mix up? Where's the dmesg and i915_error_state? The backtrace would imply a GPU hang...
1.Sometimes system hangs,and sometimes only GPU hangs. "Backtrace" is in Terminal by remote when system or GPU hangs. 2.When GPU hangs,test with eb43f4af7ecb7d51ba44f5e96bf74eedf1c27d62 1)Only in dmesg: DRHD: handling fault status reg 3 DMAR:[DMA Read] Request device [00:02.0] fault addr 1c818000 DMAR:[fault reason 06] PTE Read access is not set 2)i915_error_state,pls see attachment. 3.eb43f4af7ecb7d51ba44f5e96bf74eedf1c27d62 is the first bad commit. commit eb43f4af7ecb7d51ba44f5e96bf74eedf1c27d62 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Wed Dec 8 17:32:24 2010 +0000 drm/i915: Terminate the FORCE WAKE after we have finished reading Once we have read the value out of the GT power well, we need to remove the FORCE WAKE bit to allow the system to auto-power down. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Created attachment 43593 [details] i915_error_state when GPU hangs with the commit eb43f4a
The bisection implies that allowing the system to enter rc6 is the cause. So can you try the patch on -staging: commit f2fedc2dc18d8556b99ae8d34558d44522c7ceb1 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Mon Feb 21 12:15:38 2011 +0000 drm/i915: Disable RC6 (render power states) for SandyBridge by default I'm receiving various regression reports in which render P-states appear to be implicated, so disable them by default. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> to see if that improves stability. However, we also have an interesting dmar error being reported, the first I'm aware of for SNB - i.e. there is more than one bug at play here.
Test with (drm-intel-staging)f2fedc2dc18d8556b99ae8d34558d44522c7ceb1, the problem exists too.System hangs 6 times,GPU only hangs 2 times when testing with the commit 8 times.Pls see attachments about dmesg and i915_error_state. Only in /var/log/message Feb 20 03:44:10 x-sgb2 kernel: DMAR:[fault reason 06] PTE Read access is not set Feb 20 03:44:10 x-sgb2 kernel: DRHD: handling fault status reg 3 Feb 20 03:44:10 x-sgb2 kernel: DMAR:[DMA Read] Request device [00:02.0] fault addr 777690
Created attachment 43635 [details] dmesg when GPU hangs with the commit f2fed
Created attachment 43636 [details] i915_error_state when GPU hangs with the commit f2fed
Have we added some new GT power well read/writes outside of a force wake block perhaps? Also, does the hang go away if you disable VT-d in the BIOS setup?
(In reply to comment #8) > Also, does the hang go away if you disable VT-d in the BIOS setup? Test in (drm-intel-fixes)811779a0fbd23035872d1fc8f2d1baf550a94f10 when disable VT-d in the BIOS, GPU still hangs.
*** This bug has been marked as a duplicate of bug 34059 ***
*** This bug has been marked as a duplicate of bug 34056 ***
Verified with 91355834646328e7edc6bd25176ae44bcd7386c7,it works fines.
Closing old verified.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.