Bug 583522 – sys-kernel/gentoo-sources-4.4.6 Oops in the i915 driver

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Garry Filakhtov 2016-05-19 20:05:42 UTC]

Current stable kernel version (gentoo-sources-4.4.6) is affected by https://bugs.freedesktop.org/show_bug.cgi?id=93822.

Proposed patch to be backported to fix the issue is here: https://patchwork.freedesktop.org/patch/67152/

Reproducible: Sometimes

Steps to Reproduce:
1. Using i915 driver
1. OOPS is triggered by simply switching back and forth between Xserver and framebuffer console
Actual Results:  
Kernel OOPS, black screen and system freeze.

Expected Results:  
Seamless switching between graphical X env and fb console.

CPU: Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz
MB: AsRock H87 Pro4

[Comment 1 Tomáš Mózes 2016-05-19 21:01:57 UTC]

Hello Igor, can you please test version 4.4.10? If it fixes your problem, we could eventually stabilize it.

[Comment 2 Garry Filakhtov 2016-05-20 15:14:42 UTC]

Hi Tomáš,
I've just tried 4.4.10 and got:

May 20 17:03:47 ecam-ws046 kernel: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008
May 20 17:03:47 ecam-ws046 kernel: IP: [<ffffffffc04367e0>] intel_fb_obj_invalidate+0x10/0xf0 [i915]
May 20 17:03:47 ecam-ws046 kernel: PGD 3fc637067 PUD 3fc636067 PMD 0 
May 20 17:03:47 ecam-ws046 kernel: Oops: 0000 [#1] PREEMPT SMP 
May 20 17:03:47 ecam-ws046 kernel: Modules linked in: drbg ecb fuse bridge stp llc tun usbhid arc4 iwldvm mac80211 snd
May 20 17:03:47 ecam-ws046 kernel: CPU: 0 PID: 2623 Comm: X Not tainted 4.4.10-gentoo-bonny #1
May 20 17:03:47 ecam-ws046 kernel: Hardware name: LENOVO 10A7000PGE/SHARKBAY, BIOS FBKT72AUS 01/26/2014
May 20 17:03:47 ecam-ws046 kernel: task: ffff8800d31b3400 ti: ffff8803fc61c000 task.ti: ffff8803fc61c000
May 20 17:03:47 ecam-ws046 kernel: RIP: 0010:[<ffffffffc04367e0>]  [<ffffffffc04367e0>] intel_fb_obj_invalidate+0x10/0
May 20 17:03:47 ecam-ws046 kernel: RSP: 0018:ffff8803fc61fac8  EFLAGS: 00010246
May 20 17:03:47 ecam-ws046 kernel: RAX: ffff8800d31b3400 RBX: ffff8804087d1000 RCX: 0000000000bb5dc0
May 20 17:03:47 ecam-ws046 kernel: RDX: ffff8804087e7cc0 RSI: 0000000000000000 RDI: 0000000000000000
May 20 17:03:47 ecam-ws046 kernel: RBP: ffff8803fc61faf0 R08: 00000000000177c0 R09: ffffffffc027ec31
May 20 17:03:47 ecam-ws046 kernel: R10: ffffea000fccbdc0 R11: ffffffffc035a89e R12: ffff8804077f7800
May 20 17:03:47 ecam-ws046 kernel: R13: 0000000000000000 R14: ffff8804077f7860 R15: 0000000000000080
May 20 17:03:47 ecam-ws046 kernel: FS:  00007f2fe7c6f940(0000) GS:ffff88041ea00000(0000) knlGS:0000000000000000
May 20 17:03:47 ecam-ws046 kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
May 20 17:03:47 ecam-ws046 kernel: CR2: 0000000000000008 CR3: 0000000402fb2000 CR4: 00000000001406f0
May 20 17:03:47 ecam-ws046 kernel: DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
May 20 17:03:47 ecam-ws046 kernel: DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
May 20 17:03:47 ecam-ws046 kernel: Stack:
May 20 17:03:47 ecam-ws046 kernel:  ffff8804087d1000 ffff8804077f7800 0000000000200001 ffff8804077f7860
May 20 17:03:47 ecam-ws046 kernel:  0000000000000080 ffff8803fc61fb10 ffffffffc043f89e 00000000081c5b60
May 20 17:03:47 ecam-ws046 kernel:  ffff8803fc61fc68 ffff8803fc61fc48 ffffffff9689959f ffffffffc04a2d80
May 20 17:03:47 ecam-ws046 kernel: Call Trace:
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffffc043f89e>] intel_fbdev_set_par+0x3e/0x60 [i915]
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff9689959f>] fb_set_var+0x19f/0x3e0
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff96677498>] ? check_preempt_wakeup+0x158/0x250
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff96675327>] ? __enqueue_entity+0x67/0x70
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff96678551>] ? enqueue_entity+0x331/0xc40
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff96896aba>] fbcon_blank+0x20a/0x2e0
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff96884965>] do_unblank_screen+0xa5/0x1a0
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff9687b324>] complete_change_console+0x54/0xd0
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff9687c4c9>] vt_ioctl+0x1129/0x1300
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff9666f40d>] ? get_parent_ip+0xd/0x50
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff9686fa52>] tty_ioctl+0x3d2/0xc20
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff96733ec5>] ? handle_mm_fault+0x875/0x14a0
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff96776e35>] do_vfs_ioctl+0x2b5/0x490
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff96780372>] ? __fget+0x72/0xa0
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff9677704c>] SyS_ioctl+0x3c/0x70
May 20 17:03:47 ecam-ws046 kernel:  [<ffffffff969e059b>] entry_SYSCALL_64_fastpath+0x16/0x6e
May 20 17:03:47 ecam-ws046 kernel: Code: 48 00 00 44 89 f2 89 de 4c 89 e7 e8 eb f7 ff ff 5b 41 5c 41 5d 41 5e 41 5f 5d
May 20 17:03:47 ecam-ws046 kernel: RIP  [<ffffffffc04367e0>] intel_fb_obj_invalidate+0x10/0xf0 [i915]
May 20 17:03:47 ecam-ws046 kernel:  RSP <ffff8803fc61fac8>
May 20 17:03:47 ecam-ws046 kernel: CR2: 0000000000000008
May 20 17:03:47 ecam-ws046 kernel: ---[ end trace 03efcb8e9718d92e ]---

[Comment 3 Pacho Ramos 2016-05-21 11:18:36 UTC]

Created attachment 434832 [details, diff]
1.patch

Can you try this patch? It comes from a openSuSE maintainer that looks to be packaging 4.4 kernel for them :/

[Comment 4 Garry Filakhtov 2016-05-21 19:28:06 UTC]

Hi Pacho,

Tried to patch 4.4.6 with patch provided.
Had to manually apply code changes, because source is already a bit different in 4.4.6

Unfortunately I still have the same issue.

Any other suggestions?

[Comment 5 Pacho Ramos 2016-05-22 10:33:55 UTC]

No, apart of trying to provide any help upstream could request in upstream report :(

I added a comment for trying to get their attention but...:
https://bugs.freedesktop.org/show_bug.cgi?id=93822#c9

[Comment 6 Garry Filakhtov 2016-05-22 14:44:31 UTC]

Tested with 4.4.11, fails so far... Let's hope for upstream...

[Comment 7 Mike Pagano 2017-03-03 00:51:21 UTC]

The patch from the upstream bug is present in the following kernels:

>= v4.5.0

Can you try a later kernel >4.5 and let us know if this is working?

[Comment 8 Garry Filakhtov 2017-03-03 00:58:08 UTC]

Hi Mike,

Thanks for update. I can confirm that at least kernel 4.8 and later are working fine for me. Been running them for awhile now.

We can close this bug as resolved now, for good!

[Comment 9 Mike Pagano 2017-03-03 12:07:49 UTC]

Thanks for the update, Garry.