Bug 71031

Summary: [PNV]boot system has [ 1.567808] [drm:i915_stolen_to_physical] *ERROR* conflict detected with stolen region: [0x7f800000 - 0x80000000]
Product: DRI Reporter: lu hua <huax.lu>
Component: DRM/IntelAssignee: Daniel Vetter <daniel>
Status: CLOSED FIXED QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: normal    
Priority: medium CC: account-disabled-20180731, intel-gfx-bugs, p, yi.sun
Version: unspecified   
Hardware: All   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
dmesg
none
iomem
none
iomem on 93825ce151e8
none
lspci -vvnn on latest kernel
none
stolen debug patch
none
dmesg with stolen debug patch
none
dmesg
none
dmesg with stolen debug patch
none
drm/i915: Trim stolen region for reserved pages and conflicting BIOS setups
none
dmesg
none
off-by-one fix
none
dmesg(patch)
none
iomem
none
new debug patch
none
dmesg(patch 96515)
none
New patch to test none

Description lu hua 2013-10-30 06:09:17 UTC
Created attachment 88339 [details]
dmesg

System Environment:
--------------------------
Platform: Pineview
Kernel:	(drm-intel-nightly)d3c493af99141b2ca1f189bce56ba7e8c2257aa6

Bug detailed description:
-------------------------
Boot system, [drm:i915_stolen_to_physical] *ERROR* conflict detected with stolen region: [0x7f800000 - 0x80000000] appears in dmesg. It happens on pineview with -queued -fixes and -nightly kernel.
I can't catch a good commit. It has Bug 66726, It also happens on commit ee75f952998e8365da90300cf1d421fc6c7dafa2(https://bugs.freedesktop.org/show_bug.cgi?id=66726#c28).

Reproduce steps:
----------------------------
1. clean boot system
2. dmesg -r | egrep "<[1-6]>" |grep drm
Comment 1 lu hua 2013-10-30 06:09:53 UTC
Created attachment 88340 [details]
iomem
Comment 2 Daniel Vetter 2013-10-30 09:18:36 UTC
Can you please also upload iomem from a kernel before ee75f952998e8365da90300cf1d421fc6c7dafa2 ?

I just wonder where     7f700000-7f8fffff : PCI Bus 0000:01 is from.

Also please upload the output of lspci -vvnn (from a recent kernel with that patch).
Comment 3 Chris Wilson 2013-10-30 09:24:44 UTC
Note I have a similar issue that has cropped up on my byt-t - which would be a pita to bisect.
Comment 4 lu hua 2013-10-31 08:15:57 UTC
Created attachment 88398 [details]
iomem on 93825ce151e8
Comment 5 lu hua 2013-10-31 08:16:35 UTC
Created attachment 88399 [details]
lspci -vvnn on latest kernel
Comment 6 Chris Wilson 2014-01-10 18:32:14 UTC
Can you confirm whether the error message still occurs on -nightly?
Comment 7 lu hua 2014-01-13 05:16:55 UTC
It still happens on latest -nightly kernel.
Comment 8 Daniel Vetter 2014-01-14 12:38:13 UTC
Created attachment 92036 [details] [review]
stolen debug patch

Please apply this debug patch on top of latest drm-intel-nightly, boot with drm.debug=0xe and attach the complete dmesg (early boot message are really important here).
Comment 9 lu hua 2014-01-15 06:44:06 UTC
Created attachment 92114 [details]
dmesg with stolen debug patch
Comment 10 Chris Wilson 2014-01-15 18:14:08 UTC
No error message there.
Comment 11 lu hua 2014-01-22 05:59:38 UTC
It still happens on latest -nightly kernel.
Comment 12 Chris Wilson 2014-01-22 08:46:46 UTC
The previous dmesg did not have any error message. Perhaps you might like to attach your current dmesg when reopening the bug.
Comment 13 lu hua 2014-01-24 01:44:36 UTC
Created attachment 92706 [details]
dmesg
Comment 14 Chris Wilson 2014-01-24 08:46:27 UTC
*sighs.

My system prints an INFO for the arch/x86/kernel/early-quirks.c stolen reservation.

Please apply the debug patch and try again.
Comment 15 lu hua 2014-01-26 02:29:05 UTC
(In reply to comment #14)
> *sighs.
> 
> My system prints an INFO for the arch/x86/kernel/early-quirks.c stolen
> reservation.
> 
> Please apply the debug patch and try again.

Apply stolen debug patch, this error goes away.
Comment 16 lu hua 2014-01-26 02:29:46 UTC
Created attachment 92793 [details]
dmesg with stolen debug patch
Comment 17 Chris Wilson 2014-01-26 09:43:11 UTC
That makes no sense as the debug patch just adds a line to dmesg and does not explain the behaviour you are reporting. Please, please, double check.
Comment 18 Chris Wilson 2014-03-05 12:02:49 UTC
Created attachment 95162 [details] [review]
drm/i915: Trim stolen region for reserved pages and conflicting BIOS setups

Please test this patch.
Comment 19 lu hua 2014-03-06 06:49:19 UTC
(In reply to comment #18)
> Created attachment 95162 [details] [review] [review]
> drm/i915: Trim stolen region for reserved pages and conflicting BIOS setups
> 
> Please test this patch.

Apply this patch, build fail.

make[6]: *** [drivers/gpu/drm/i915/i915_gpu_error.o] Error 1
make[5]: *** [drivers/gpu/drm/i915] Error 2
make[4]: *** [drivers/gpu/drm] Error 2
make[3]: *** [drivers/gpu] Error 2
make[2]: *** [drivers] Error 2
make[1]: *** [binrpm-pkg] Error 2
make: *** [binrpm-pkg] Error 2
Comment 20 Daniel Vetter 2014-03-10 09:25:03 UTC
Please retest on latest -nightly, it contains the light version of the stolen range frobnicator.
Comment 21 lu hua 2014-03-11 06:10:48 UTC
Created attachment 95571 [details]
dmesg

It still occurs on latest -nightly kernel(0cfec5).
Comment 22 Daniel Vetter 2014-03-27 21:32:15 UTC
Created attachment 96494 [details] [review]
off-by-one fix

Please test with this patch on top of latest -nightly.

If it still fails please attach dmesg _and_ /proc/iomem
Comment 23 lu hua 2014-03-28 03:29:03 UTC
(In reply to comment #22)
> Created attachment 96494 [details] [review] [review]
> off-by-one fix
> 
> Please test with this patch on top of latest -nightly.
> 
> If it still fails please attach dmesg _and_ /proc/iomem

Test this patch, It still occurs.
Comment 24 lu hua 2014-03-28 03:29:39 UTC
Created attachment 96507 [details]
dmesg(patch)
Comment 25 lu hua 2014-03-28 03:29:59 UTC
Created attachment 96508 [details]
iomem
Comment 26 Daniel Vetter 2014-03-28 07:38:48 UTC
Created attachment 96515 [details] [review]
new debug patch

Please apply and attach dmesg.
Comment 27 lu hua 2014-03-31 02:36:38 UTC
Created attachment 96635 [details]
dmesg(patch 96515)

Test this patch, It still has this error.
Comment 28 Daniel Vetter 2014-04-05 10:49:57 UTC
Created attachment 96943 [details] [review]
New patch to test

Please test this patch and then attach the contents of /sys/kernel/debug/dri/0/i915_gem_stolen You need debugfs enabled&mounted for that ofc.
Comment 29 lu hua 2014-04-08 06:03:53 UTC
(In reply to comment #28)
> Created attachment 96943 [details] [review] [review]
> New patch to test
> 
> Please test this patch and then attach the contents of
> /sys/kernel/debug/dri/0/i915_gem_stolen You need debugfs enabled&mounted for
> that ofc.

Apply this patch, It works well.
/sys/kernel/debug/dri/0/i915_gem_stolen:
Stolen:
   f587b500:   g     1252KiB 41 00 0 0 0 uncached (pinned x 0) (ggtt offset: 00000000, size: 00139000) (stolen: 00000000)
   f587b600: p g      128KiB 40 40 0 0 0 uncached dirty (pinned x 1) (ggtt offset: 0013c000, size: 00020000) (stolen: 00139000) (p mappable)
   f587b700: p g        4KiB 40 40 0 0 0 uncached dirty (pinned x 1) (ggtt offset: 0015c000, size: 00001000) (stolen: 00159000) (p mappable)
   f63aae00: p g     2400KiB 41 00 0 0 0 uncached (pinned x 1) (display) (ggtt offset: 00160000, size: 00258000) (stolen: 0015a000) (p mappable)
Total 4 objects, 3874816 bytes, 3874816 GTT size
Comment 30 emailofchris 2014-07-10 04:56:11 UTC
[   11.155637] [drm:i915_stolen_to_physical] *ERROR* conflict detected with stolen region: [0x3f800000 - 0x400000000]

Ran into this on Arch Linux on my Eee PC 900 at the end of the boot process, when it tried to start LMDE.
Comment 31 emailofchris 2014-07-29 05:22:07 UTC
Just updated, it's fixed for me now!
Comment 32 lu hua 2014-07-30 06:32:15 UTC
It still happens on latest drm-intel-nightly(eb2af97968f0) kernel. 
[root@x-pnv2 ~]# dmesg -r | egrep "<[1-3]>" |grep drm
<3>[    2.017276] [drm:i915_stolen_to_physical] *ERROR* conflict detected with stolen region: [0x7f800000 - 0x80000000]
[root@x-pnv2 ~]# uname -a
Linux x-pnv2 3.16.0-rc6_drm-intel-nightly_eb2af9_20140730+
Comment 33 lu hua 2014-11-13 06:55:38 UTC
Test on latest -nightly kernel, it still happens. 
Run ./drv_module_reload, it also has this error.
[root@x-pnv2 tests]# ./drv_module_reload
unbinding /sys/class/vtconsole/vtcon0/: (M) frame buffer device
module successfully unloaded
module successfully loaded again
[root@x-pnv2 tests]# dmesg -r|egrep "<[1-4]>"|grep drm
<3>[  163.125482] [drm:i915_gem_init_stolen [i915]] *ERROR* conflict detected with stolen region: [0x7f800000 - 0x80000000]
<3>[  163.439274] [drm:check_crtc_state [i915]] *ERROR* mismatch in adjusted_mode.crtc_clock (expected 45000, found -2857)
<4>[  163.439353] WARNING: CPU: 1 PID: 1157 at drivers/gpu/drm/i915/intel_display.c:10625 check_crtc_state+0xa07/0xa35 [i915]()
<4>[  163.439411] Modules linked in: i915 drm_kms_helper drm ip6table_filter ip6_tables ebtable_nat ebtables nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 iptable_mangle xt_tcpudp iptable_filter ip_tables x_tables bridge stp llc ipv6 dm_mod iTCO_wdt iTCO_vendor_support msi_laptop sparse_keymap rfkill pcspkr i2c_i801 snd_hda_codec_realtek lpc_ich mfd_core snd_hda_codec_generic snd_hda_controller snd_hda_codec snd_hwdep snd_pcm snd_timer snd r8169 soundcore wmi battery ac acpi_cpufreq joydev button video cfbfillrect cfbimgblt cfbcopyarea [last unloaded: drm]
<4>[  163.439418] CPU: 1 PID: 1157 Comm: kworker/u4:4 Tainted: G        W      3.18.0-rc3_drm-intel-nightly_de6d6c_20141111+ #1386
<4>[  163.439901]  [<f90d02e1>] ? drm_mode_set_config_internal+0x39/0x97 [drm]
<4>[  163.439910]  [<f93c8958>] ? restore_fbdev_mode+0x93/0xa8 [drm_kms_helper]
<4>[  163.439918]  [<f93c8982>] ? drm_fb_helper_restore_fbdev_mode_unlocked+0x15/0x23 [drm_kms_helper]
<4>[  163.439926]  [<f93c90c8>] ? drm_fb_helper_set_par+0x2c/0x42 [drm_kms_helper]
<4>[  163.440083]  [<f93c9303>] ? drm_fb_helper_initial_config+0x225/0x2a4 [drm_kms_helper]
Comment 34 Jani Nikula 2014-12-08 12:50:14 UTC
commit ee86e80a6f976e76bcdb5f2071a9d07f5fd63c61
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Fri Apr 11 15:55:17 2014 +0200

    drm/i915: Don't complain about stolen conflicts on gen3
Comment 35 lu hua 2014-12-09 02:48:34 UTC
Reboot 5 rounds on latest -nightly kernel, I don't meet this error.
Comment 36 Jarkko K 2016-07-27 08:31:00 UTC
[    1.425266] [drm:i915_gem_init_stolen [i915]] *ERROR* conflict detected with stolen region: [0x88000000 - 0xc8000000]


4.7.0-996-generic #201607222201
Comment 37 jc.francois 2016-08-12 07:13:26 UTC
[drm:i915_gem_init_stolen [i915]] *ERROR* conflict detected with stolen region: [0xb6000000 - 0xb8000000]
Comment 38 Jani Nikula 2016-08-12 07:59:30 UTC
(In reply to Jarkko K from comment #36)
> [    1.425266] [drm:i915_gem_init_stolen [i915]] *ERROR* conflict detected
> with stolen region: [0x88000000 - 0xc8000000]
> 
> 
> 4.7.0-996-generic #201607222201

(In reply to jc.francois from comment #37)
> [drm:i915_gem_init_stolen [i915]] *ERROR* conflict detected with stolen
> region: [0xb6000000 - 0xb8000000]

Both of these are missing any details on the platform, which is almost certainly not PNV as in the original bug. This bug is fixed, and has been for years. Please do not open old bugs to report new ones.

Please file a new bug, add drm.debug=14 module parameter and attach full dmesg on that bug.

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.