Bug 94870 - [BAT ILK] list_add corruption in test drv_module_reload_basic
Summary: [BAT ILK] list_add corruption in test drv_module_reload_basic
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: DRI git
Hardware: x86-64 (AMD64) Linux (All)
: high critical
Assignee: Imre Deak
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-04-08 14:14 UTC by Tim Gore
Modified: 2016-09-30 09:37 UTC (History)
1 user (show)

See Also:
i915 platform: ILK
i915 features:


Attachments

Description Tim Gore 2016-04-08 14:14:34 UTC
During BAT testing of a patch targeting gen9 only, igt test drv_module_reload_basic causes "list_add corruption" warning in dmesg log on the ilk-hp8440p target, although the test reports success.

Patch being tested is https://patchwork.freedesktop.org/series/5367/
This introduces a workaround in gen9_init_workarounds, which is also
conditional on
 (IS_SKL_REVID(dev, SKL_REVID_C0, REVID_FOREVER) || IS_BROXTON(dev))
So this patch should not be able to affect an ILK target.

Regression build. CI_DRM_1201

Dmesg:
[  200.736652] Setting dangerous option inject_load_failure - tainting kernel
[  201.274660] Setting dangerous option inject_load_failure - tainting kernel
[  201.827477] Setting dangerous option inject_load_failure - tainting kernel
[  202.341648] Setting dangerous option inject_load_failure - tainting kernel
[  205.869759] ------------[ cut here ]------------
[  205.869769] WARNING: CPU: 1 PID: 5982 at lib/list_debug.c:29 __list_add+0x65/0xc0
[  205.869771] list_add corruption. next->prev should be prev (ffff8800b6832ce8), but was           (null). (next=ffff8801207480f8).
[  205.869773] Modules linked in: snd_hda_intel(+) i915 ax88179_178a usbnet mii intel_powerclamp coretemp snd_hda_codec_hdmi crct10dif_pclmul crc32_pclmul snd_hda_codec_generic ghash_clmulni_intel snd_hda_codec snd_hwdep lpc_ich snd_hda_core snd_pcm mei_me mei sdhci_pci sdhci mmc_core e1000e ptp pps_core [last unloaded: i915]
[  205.869795] CPU: 1 PID: 5982 Comm: modprobe Tainted: G     U          4.6.0-rc2-gfxbench+ #1
[  205.869797] Hardware name: Hewlett-Packard HP EliteBook 8440p/172A, BIOS 68CCU Ver. F.24 09/13/2013
[  205.869798]  0000000000000000 ffff8800b497faa8 ffffffff81404e65 ffff8800b497faf8
[  205.869801]  0000000000000000 ffff8800b497fae8 ffffffff81079c7c 0000001d00000286
[  205.869804]  ffff88007ee8bb78 ffff8801207480f8 ffff8800b6832ce8 ffffffff81c52150
[  205.869807] Call Trace:
[  205.869811]  [<ffffffff81404e65>] dump_stack+0x67/0x92
[  205.869816]  [<ffffffff81079c7c>] __warn+0xcc/0xf0
[  205.869818]  [<ffffffff81079cea>] warn_slowpath_fmt+0x4a/0x50
[  205.869820]  [<ffffffff814212c5>] __list_add+0x65/0xc0
[  205.869824]  [<ffffffff811f4fc6>] d_alloc+0x56/0x70
[  205.869827]  [<ffffffff811e7ebe>] __lookup_hash+0x2e/0x50
[  205.869830]  [<ffffffff811e7fad>] lookup_one_len+0xcd/0x120
[  205.869833]  [<ffffffff812e6509>] start_creating+0x69/0xf0
[  205.869835]  [<ffffffff812e6891>] __create_dir+0x11/0xf0
[  205.869837]  [<ffffffff812e6a80>] tracefs_create_dir+0x10/0x20
[  205.869840]  [<ffffffff8113926e>] event_create_dir+0x2ae/0x4d0
[  205.869842]  [<ffffffff811394b1>] __trace_add_new_event+0x21/0x30
[  205.869844]  [<ffffffff81139714>] trace_module_notify+0x154/0x1a0
[  205.869848]  [<ffffffff8109e1b9>] notifier_call_chain+0x39/0xa0
[  205.869851]  [<ffffffff8109e5b8>] __blocking_notifier_call_chain+0x48/0x70
[  205.869853]  [<ffffffff8109e5f1>] blocking_notifier_call_chain+0x11/0x20
[  205.869857]  [<ffffffff8110bff1>] load_module+0x1a81/0x24b0
[  205.869860]  [<ffffffff811e109c>] ? kernel_read+0x3c/0x50
[  205.869862]  [<ffffffff8110cc39>] SyS_finit_module+0xa9/0xd0
[  205.869867]  [<ffffffff817d13a9>] entry_SYSCALL_64_fastpath+0x1c/0xac
[  205.869869] ---[ end trace 0dbd577555177c51 ]---
Comment 1 Jani Nikula 2016-04-08 14:59:00 UTC
(In reply to Tim Gore from comment #0)
> [  200.736652] Setting dangerous option inject_load_failure - tainting kernel

So the kernel load itself fails on purpose, but this shouldn't happen.
Comment 2 Jari Tahvanainen 2016-07-04 08:22:32 UTC
priority aligned for igt basic tests on gen7 to High+Critical
Comment 3 Imre Deak 2016-08-16 13:17:41 UTC
I couldn't reproduce this with the nightly kernel and haven't seen this in CI logs, so closing it assuming it's fixed.
Comment 4 Jari Tahvanainen 2016-09-30 09:37:57 UTC
Closing.


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.