Bug 62944

Summary: [915gm 3.8 regression] pfit disabled -- fullscreen games only occupy top-left quarter of screen
Product: DRI Reporter: AndrzejL <AndrzejL>
Component: DRM/IntelAssignee: Intel GFX Bugs mailing list <intel-gfx-bugs>
Status: CLOSED FIXED QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: normal    
Priority: high    
Version: unspecified   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:

Description AndrzejL 2013-03-30 18:23:11 UTC
Hi folks.

I am running ArchLinux 2013 on a older Acer Travelmate 2420 laptop.

[root@wishmacer andrzejl]# inxi -v5
System:    Host: wishmacer.loc Kernel: 3.7.9-2-ARCH i686 (32 bit, gcc: 4.7.2) 
           Desktop: KDE 4.10.1 (Qt 4.8.4) Distro: Arch Linux
Machine:   System: Acer product: TravelMate 2420 version: 0100 serial: LXTB20504360808B79KS00 
           Mobo: Acer model: Garda-910 version: Rev serial: LXTB20504360808B79KS00
           Bios: Phoenix version: V1.01 date: 12/15/2005
CPU:       Single core Intel Celeron M (-UP-) cache: 1024 KB flags: (nx pae sse sse2) bmips: 2993.39 clocked at 1496.350 MHz 
Graphics:  Card: Intel Mobile 915GM/GMS/910GML Express Graphics Controller bus-ID: 00:02.0 
           X.org: 1.14.0 driver: intel tty size: 123x34 Advanced Data: N/A for root 
Audio:     Card: Intel 82801FB/FBM/FR/FW/FRW (ICH6 Family) AC'97 Audio Controller 
           driver: snd_intel8x0 ports: 1c00 18c0 bus-ID: 00:1e.2 
           Sound: Advanced Linux Sound Architecture ver: k3.7.9-2-ARCH
Network:   Card-1: Realtek RTL-8139/8139C/8139C+ driver: 8139too ver: 0.9.28 port: 3000 bus-ID: 06:07.0
           IF: eth0 state: unknown speed: 100 Mbps duplex: full mac: 00:0a:e4:f6:d4:8f
           Card-2: Atheros AR5416 Wireless Network Adapter [AR5008 802.11(a)bgn] driver: ath9k bus-ID: 06:05.0
           IF: wlan0 state: down mac: 00:1e:52:7a:46:5d
Drives:    HDD Total Size: 474.0GB (42.2% used) 1: id: /dev/sda model: SAMSUNG_HM160HC size: 160.0GB temp: 32C 
           2: USB id: /dev/sdb model: ST950212A size: 50.0GB temp: 0C 3: USB id: /dev/sdc model: WD600UE size: 60.0GB temp: 0C 
           4: USB id: /dev/sdd model: 6Y200P0 size: 203.9GB temp: 0C 
           Optical: /dev/sr0 model: N/A rev: N/A dev-links: cdrom,cdrw,dvd,dvdrw
Partition: ID: / size: 25G used: 7.9G (34%) fs: ext4 dev: /dev/sda1 
           label: N/A uuid: 1d57d4c0-ecdc-4f85-b997-bbaac4313b16
           ID: /home size: 9.9G used: 7.5G (80%) fs: ext4 dev: /dev/sda6 
           label: N/A uuid: b0cd3c12-a326-45b6-9a18-96c5af876a7e
           ID: swap-1 size: 2.15GB used: 0.00GB (0%) fs: swap dev: /dev/sda5 
           label: N/A uuid: f84bd38a-3ab1-48c7-bc87-8251c8603d85
Sensors:   System Temperatures: cpu: 53.0C mobo: 30.0C 
           Fan Speeds (in rpm): cpu: N/A 
Info:      Processes: 129 Uptime: 2:35 Memory: 599.4/2011.0MB Gcc sys: 4.7.2 Client: Shell (bash 4.2.45) inxi: 1.8.45 
[root@wishmacer andrzejl]# 

I have noticed (in a stable repository) a kernel upgrade sometime ago from 3.7.9 to 3.8.3 and I have run into a little snag after upgrading. All the games (latest crossover for Linux / Crossover 10 Games for Linux) and also games with native Linux clients / installers (Quake 3 Arena / Neverwinter Nights) are having problems with full screen.

I was fully upgraded without using testing / unstable sections. Some time has passed and there were few other kernel upgrades (3.8.4 in stable section / 3.8.5 since last night in testing section). I have tested them all and the issue is not solved.

What is the GPU?

00:02.0 VGA compatible controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 03)
Subsystem: Acer Incorporated [ALI] Device 006a
Kernel driver in use: i915

What is the issue?

All games look like this:

http://s6.postimage.org/b05uducnk/18032013228.jpg
http://s6.postimage.org/iu6fz8kgg/18032013229.jpg
http://s6.postimage.org/va35szdsg/18032013230.jpg
http://s6.postimage.org/sh9y8ydg0/18032013231.jpg

After downgrading kernel and ONLY kernel and rebooting

http://s6.postimage.org/r8m2nblht/Andrzej_L_Arch_Linux_Snapshot_038.png

the games were back to normal.

http://s6.postimage.org/tku2kwy34/18032013232.jpg
http://s6.postimage.org/ah0r4kl8w/18032013233.jpg
http://s6.postimage.org/ksd3x8cy8/18032013234.jpg
http://s6.postimage.org/o0hla9z80/18032013235.jpg

Sorry about the photo quality. I took them with the mobile phone coz the screenshots were not really working the way I planned ;).

The problem must be Intel specific or maybe even module specific as it does not happens on a twin install on Dell Latitude D600.

01:00.0 VGA compatible controller: Advanced Micro Devices [AMD] nee ATI Radeon RV250 [Mobility FireGL 9000] (rev 01)
Subsystem: Dell Device 011d
Kernel driver in use: radeon

I have reported it using the ArchLinux bug-reporting tool:

https://bugs.archlinux.org/task/34355

but after waiting for a longer while now I got no luck at all so I went to #intel-gfx channel to seek help:

14:20 -!- Irssi: Join to #intel-gfx was synced in 3 secs
14:20 < AndrzejL> guys... any known issues with the 3.8. kernels and i915 cards? I am having trouble with full screen games
14:20 < AndrzejL> they are not full screen but only 1/3 of the screen or so
14:20 < AndrzejL> let me show You...
14:21 < AndrzejL> https://bugs.archlinux.org/task/34355
14:22 < AndrzejL> it was all perfect with 3.7.9 kernel and then BOOM kernel 3.8.x messed it up :)
14:24 < AndrzejL> any ideas other then "Buy new gear"? :D
14:50 < ickle> AndrzejL: I would had said a modesetting failure, but the log doesn't even show it attempt to change modes
14:50 < ickle> AndrzejL: can you grab an Xorg.log / dmesg from both 3.8 and 3.7?
15:26 < AndrzejL> ickle: absolutely - no problem
15:26 < AndrzejL> give me a bit of time and I will
15:40 < AndrzejL> ickle: http://pastebin.com/wZmfUPnv
15:41 < AndrzejL> let me know if I can provide any more info
15:43 < AndrzejL> And again - downgrading the kernel and the kernel only fixed the issue :)
16:43 < ickle> AndrzejL: ok, I don't see any change in userspace behaviour, so it looks to be an internal issue in the kernel

Diagnose: Internal issue in the kernel...

Ickle also asked me if I was willing to do a bisection of the kernel I was willing at first but then it turns out it was to much of a hustle for a few reasons (download ~ 700 megs - I am on a 20 gigs monthly download quota limit and I don't have much of it left this month, time - the machine has single core Celeron cpu and it's my only / everyday use machine - I cannot just leave it for few days to compile / dissect the kernel coz I am left with no gear to work on).

I was told that upstreaming the bug could be useful and I was pointed here so...

I am not sure what other info could be provided. Please do not hesitate to ask additional questions. I will reply as soon as possible.

Regards.

Andrzej
Comment 1 Daniel Vetter 2013-04-02 08:27:01 UTC
Can you please test whether 3.9-rc kernels work, or whether those are broken, too?

Also, can you please grab intel-gpu-tools (either from git or if your distro provides packages, that should work, too) and then attach the output of intel_reg_dump for both working and broken kernels? It is important that you run this dumper _while_ the game is running. If you don't have ssh, you can use a trick like

# sleep 30; intel_reg_dumper > dump

in a vt console and then switch back to the game/X.
Comment 2 AndrzejL 2013-04-02 14:05:03 UTC
Hi Daniel.

About the 3.9 RC kernels - I will be able to test it when it hits the ArchLinux testing repository. Never compiled kernel from the source in my life and I know it would take a long time on this single core celeron CPU and it would render it useless / unusable during the process. Also my connection has a 20 gigs per month download limit and downloading kernel source from git repository would take a nice bite out of it. I hope it's ok.

Installed the intel-gpu-tools as asked and downgraded kernel to the working 3.7.9-2 one and rebooted:

[root@wishmacer ~]# pacman -Ss intel-gpu-tools
community/intel-gpu-tools 1.3-2
    Tools for development and testing of the Intel DRM driver
[root@wishmacer ~]# pacman -S intel-gpu-tools
resolving dependencies...
looking for inter-conflicts...

Targets (1): intel-gpu-tools-1.3-2

Total Download Size:    0.09 MiB
Total Installed Size:   0.70 MiB

Proceed with installation? [Y/n] y
:: Retrieving packages from community...
 intel-gpu-tools-1.3-2-i686                       96.6 KiB  91.1K/s 00:01 [##########################################] 100%
(1/1) checking package integrity                                          [##########################################] 100%
(1/1) loading package files                                               [##########################################] 100%
(1/1) checking for file conflicts                                         [##########################################] 100%
(1/1) checking available disk space                                       [##########################################] 100%
(1/1) installing intel-gpu-tools                                          [##########################################] 100%
[root@wishmacer ~]# pacman -U /media/PKG_BackUp/pkg/linux-3.7.9-2-i686.pkg.tar.xz 
loading packages...
warning: downgrading package linux (3.8.4-1 => 3.7.9-2)
resolving dependencies...
looking for inter-conflicts...

Targets (1): linux-3.7.9-2

Total Installed Size:   61.20 MiB
Net Upgrade Size:       -0.93 MiB

Proceed with installation? [Y/n] y
(1/1) checking package integrity                                          [##########################################] 100%
(1/1) loading package files                                               [##########################################] 100%
(1/1) checking for file conflicts                                         [##########################################] 100%
(1/1) checking available disk space                                       [##########################################] 100%
(1/1) upgrading linux                                                     [##########################################] 100%
>>> Updating module dependencies. Please wait ...
>>> Generating initial ramdisk, using mkinitcpio.  Please wait...
==> Building image from preset: 'default'
  -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux.img
==> Starting build: 3.7.9-2-ARCH
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [autodetect]
  -> Running build hook: [block]
  -> Running build hook: [lvm2]
  -> Running build hook: [filesystems]
  -> Running build hook: [keyboard]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating gzip initcpio image: /boot/initramfs-linux.img
==> Image generation successful
==> Building image from preset: 'fallback'
  -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-fallback.img -S autodetect
==> Starting build: 3.7.9-2-ARCH
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [block]
  -> Running build hook: [lvm2]
  -> Running build hook: [filesystems]
  -> Running build hook: [keyboard]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating gzip initcpio image: /boot/initramfs-linux-fallback.img
==> Image generation successful
[root@wishmacer ~]# reboot

Now about the command that You asked me to test... There is a problem I think...  Dump file comes up empty? Am I doing it wrong?

WORKING KERNEL

[andrzejl@wishmacer ~]$ uname -a
Linux wishmacer.loc 3.7.9-2-ARCH #1 SMP PREEMPT Mon Feb 25 12:29:24 CET 2013 i686 GNU/Linux
[andrzejl@wishmacer ~]$ su
Password: 
[root@wishmacer andrzejl]# sleep 30; intel_reg_dumper > dump
Gen2/3 Ranges are not supported. Please use unsafe access.Aborted (core dumped)
[root@wishmacer andrzejl]# ls --full ./dump
-rw-r--r-- 1 andrzejl users 0 2013-04-02 13:17:36.629652278 +0100 ./dump
[root@wishmacer andrzejl]# du -h ./dump
0       ./dump
[root@wishmacer andrzejl]# 

I went to the #intel-gfx channel and was told:

14:16 < AndrzejL> I was told to provide intel_req_dumper dump
14:17 < AndrzejL> but it says I need to enable the unsafe access but there is no "--unsafe-access" like switch
14:17 < danvet> AndrzejL, have you tried the latest version from git?
14:17 < danvet> iirc we've broken a few things ...
14:24 < AndrzejL> would I compile git stuff or is there another option?
14:28 < danvet> AndrzejL, compiling from git is the preferred option

So... now I am running this version:

[andrzejl@wishmacer ~]$ su
Password: 
[root@wishmacer andrzejl]# pacman -U /home/andrzejl/.makepkgs/intel-gpu-tools-git/intel-gpu-tools-git-20130402-1-i686.pkg.tar.xz 
loading packages...
resolving dependencies...
looking for inter-conflicts...
:: intel-gpu-tools-git and intel-gpu-tools are in conflict. Remove intel-gpu-tools? [y/N] y

Targets (2): intel-gpu-tools-1.3-2 [removal]  intel-gpu-tools-git-20130402-1

Total Installed Size:   0.98 MiB
Net Upgrade Size:       0.28 MiB

Proceed with installation? [Y/n] y
(1/1) checking package integrity                                          [##########################################] 100%
(1/1) loading package files                                               [##########################################] 100%
(1/1) checking for file conflicts                                         [##########################################] 100%
(2/2) checking available disk space                                       [##########################################] 100%
(1/1) removing intel-gpu-tools                                            [##########################################] 100%
(1/1) installing intel-gpu-tools-git                                      [##########################################] 100%
[root@wishmacer andrzejl]# 

Which has replaced the previous version from stable repo.

It gives me the same error message:

[root@wishmacer andrzejl]# intel_reg_dumper > dump 
Gen2/3 Ranges are not supported. Please use unsafe access.Aborted (core dumped)
[root@wishmacer andrzejl]# du -h ./dump0       ./dump
[root@wishmacer andrzejl]# 

and again a 0 bytes sized dump file.

I had to modify the source code...

debug-rdata.c
intel_reg_dumper.c


- intel_register_access_init(pci_dev, 1);
+ intel_register_access_init(pci_dev, 0); 

as I read here: 

https://bugs.freedesktop.org/show_bug.cgi?id=57365#c8

AND... after reinstalling the package:

[root@wishmacer andrzejl]# pacman -U /home/andrzejl/.makepkgs/intel-gpu-tools-git/intel-gpu-tools-git-20130402-1-i686.pkg.tar.xz 
loading packages...
warning: intel-gpu-tools-git-20130402-1 is up to date -- reinstalling
resolving dependencies...
looking for inter-conflicts...

Targets (1): intel-gpu-tools-git-20130402-1

Total Installed Size:   1.02 MiB
Net Upgrade Size:       0.04 MiB

Proceed with installation? [Y/n] y
(1/1) checking package integrity                                          [##########################################] 100%
(1/1) loading package files                                               [##########################################] 100%
(1/1) checking for file conflicts                                         [##########################################] 100%
(1/1) checking available disk space                                       [##########################################] 100%
(1/1) upgrading intel-gpu-tools-git                                       [##########################################] 100%

I get the exact same error...

[root@wishmacer andrzejl]# intel_reg_dumper > dump 
Gen2/3 Ranges are not supported. Please use unsafe access.Aborted (core dumped)
[root@wishmacer andrzejl]# 

I am getting frustrated... Kernel modesetting seems broken. I cannot debug it for you coz debug tool seems broken...

Must be my lucky day...

Cheers.

Andrzej
Comment 3 Daniel Vetter 2013-04-02 22:36:50 UTC
Latest intel-gpu-tools git should now be fixed for gen2/3 again. Sorry for the mess.
Comment 4 AndrzejL 2013-04-02 23:57:57 UTC
This is a truly wonderful news. I have git-ed and makepkg-ed the app again and this time it did worked as expected - or so it seems ;D.

Working properly kernel:

http://andrzejl.cyryl.net/AndrzejL/Problem%20Explanations/Intel_kernel_3.8.x_regression/dump_3.7.9-2

Stable not working properly kernel:

http://andrzejl.cyryl.net/AndrzejL/Problem%20Explanations/Intel_kernel_3.8.x_regression/dump_3.8.4-1

Testing not working properly kernel:

http://andrzejl.cyryl.net/AndrzejL/Problem%20Explanations/Intel_kernel_3.8.x_regression/dump_3.8.5-1

The dumps were taken (err that does not sounds right...) with the sleep 30; trick while Plants vs. Zombies was running.

Regards and big thank You for taking care of this issue.

Andrzej
Comment 5 Chris Wilson 2013-04-03 10:50:13 UTC
The important difference between 3.7 and 3.8 is that the panel fitter is disabled.
Comment 6 AndrzejL 2013-04-03 11:08:25 UTC
Can I enable it somehow or is it something that demands kernel to be recompiled? Is this a default setting in kernels 3.8.x and above or was it something that was configured by ArchLinux developers?

Regards.

Andrzej
Comment 7 Daniel Vetter 2013-04-03 15:33:35 UTC
It's a bug in our code. I'll look into it.
Comment 8 AndrzejL 2013-04-03 17:46:27 UTC
I am doing the happy dance just about now ;).

Glad I could find / report it Dudes. Waiting for the fixed code in ArchLinux repos :).

Cheers.

Andrzej
Comment 9 Daniel Vetter 2013-04-07 19:03:26 UTC
This /should/ work on 3.9-rc kernels, so testing those would be really appreciated.
Comment 10 AndrzejL 2013-04-07 20:30:44 UTC
Undrestood Daniel. Thank You very much. As soon as I will see kernel 3.9.x available for testing in ArchLinux repository I will test it and report back.

Regards.

Andrzej
Comment 11 Daniel Vetter 2013-04-07 21:34:11 UTC
Compiling from sources would be much better, since otherwise we'll have a horrible test-turn-around of months from now on :( Also, I can't ask you to test any patches ...
Comment 12 AndrzejL 2013-04-07 21:49:20 UTC
I am very sorry but I am no power user... I am using precompiled packages from the repo... I would love to help but to do so I would also need a test machine and the one that's problematic is my main machine...

Regards.

Andrzej
Comment 13 Daniel Vetter 2013-04-22 09:12:49 UTC
Note to self: We need to backport the following commit to stable kernels:

87476d632c5041cfb962630e618d787243c81ed8 drm/i915: Fixup pfit disabling for gen2/3

Note to Andrew: If you're filing a bug in upstream (as opposed to against your distro), there's the reasonable expectation that you're willing (with the help of a bunch of howtos and links we can provide) to test stuff from upstream ... Without that an upstream bug report is a but useless for us.
Comment 14 AndrzejL 2013-04-22 12:59:10 UTC
Sounds like unreasonable expectations to me. I was using a kernel 3.7.9 and I could play my fav games no problem. I was happy. All of the sudden something got broken - it took me a lot of time to find the right place to report this problem - and when I do so I am asked to wreck my stable and perfectly working system just so I can test a patch that may or may not work? Heh... Guys - it's a bug in Your code not my fault that's one. I am the one that is waiting for almost a month to fire up any of my games that's two. Cmon - why did You not added / backported the patch to the stable systems? It's not working now... Are You trying to tell me that the patch can break it even more? Even with a bunch of howtos You're providing making my machine to run for several hours full speed just to test one patch sounds like a waste of a lot of electricity and time... Let me tell You this - unless this patch is going to be added to a stable kernel - You cannot expect anything of me. I was the one that lost functionality of my system after You had a bug in Your code. I found You and asked You to fix it and You demand me compiling a... That's awesome but in this case it's not the bug report that is useless... This way I can brake more things then fix them...

Cheers.

Andrzej
Comment 15 Jani Nikula 2013-04-22 14:01:34 UTC
Andrzej, it's a catch-22. Generally speaking, we can't just backport commits to stable kernels unless they are proven to fix an issue. If everyone waited for the fixes to land in stable, they never would, because they need to be tested first.

This particular commit seems likely to be eventually backported anyway, because it fixes an issue for someone else. Whether it fixes the issue for you we don't know for sure. If it doesn't, and you only find out after waiting for a stable release to waterfall through your distro, it will have taken quite a while.

It would help your own cause to try the upstream kernel.
Comment 16 AndrzejL 2013-04-22 14:34:54 UTC
Thank You for Your reply.

Please forgive me if my previous post seems a bit harsh it wasn't my intention I should have used a better wording. Trust me I understand what You are saying and if I wasn't in such a crappy situation I would try the upstream kernel long time ago using ArchLinux pkgbuild and report if the patch works or not but there are way to many cons for me when I look at it at the moment one of them being the limitations of my stupid internet (20 gigs per month shared between 3 of us...)... There is a lot going on here for quite a while now and on top of that I am moving from one house to another in the next 2 days - like I had not enough on my head recently... anyway... New house will probably get me a better internet solutions / different ISP that will allow me to clone git the upstream kernel and possibly leave the machine to compile it overnight... Cannot promise You anything but there is hope altho it won't happen very soon. I know those are only excuses - sorry.

Regards.

Andrzej
Comment 17 AndrzejL 2013-05-01 18:47:06 UTC
[root@wishmacer andrzejl]# uname -r
3.9.0-2-ARCH
[root@wishmacer andrzejl]# pacman -Q | grep -i linux
archlinux-keyring 20130406-1
linux 3.9-2
linux-api-headers 3.8.4-1
linux-firmware 20130430-1
linux-headers 3.9-2
linux-lts-headers 3.0.75-1
util-linux 2.22.2-2
[root@wishmacer andrzejl]# sleep 30; intel_reg_dumper > dump
[root@wishmacer andrzejl]# cat ./dump 
                           DCC: 0x000f0401 (dual channel asymmetric, XOR randomization: disabled, XOR bit: 11)
                     CHDECMISC: 0x11987820 (XOR bank/rank, ch2 enh disabled, ch1 enh disabled, ch0 enh disabled, flex disabled, ep not present)
                        C0DRB0: 0x000f0401 (0x0401)
                        C0DRB1: 0x0000000f (0x000f)
                        C0DRB2: 0x00000000 (0x0000)
                        C0DRB3: 0x04000000 (0x0000)
                        C1DRB0: 0x05040302 (0x0302)
                        C1DRB1: 0x07060504 (0x0504)
                        C1DRB2: 0x09080706 (0x0706)
                        C1DRB3: 0x0b0a0908 (0x0908)
                       C0DRA01: 0x01000400 (0x0400)
                       C0DRA23: 0x02000100 (0x0100)
                       C1DRA01: 0x0d0c0b0a (0x0b0a)
                       C1DRA23: 0x0f0e0d0c (0x0d0c)
                    PGETBL_CTL: 0x7ffc0001
             VCLK_DIVISOR_VGA0: 0x00031108 (n = 3, m1 = 17, m2 = 8)
             VCLK_DIVISOR_VGA1: 0x00031406 (n = 3, m1 = 20, m2 = 6)
                 VCLK_POST_DIV: 0x00800080 (vga0 p1 = 2, p2 = 4, vga1 p1 = 2, p2 = 2)
                     DPLL_TEST: 0x00010001 ()
                  CACHE_MODE_0: 0x00006820
                       D_STATE: 0x0000000b
                 DSPCLK_GATE_D: 0x00001000 (clock gates disabled: DPLUNIT)
                RENCLK_GATE_D1: 0x00000000
                RENCLK_GATE_D2: 0x00000000
                         SDVOB: 0x00480000 (disabled, pipe A, stall disabled, not detected, SDVO mult 1)
                         SDVOC: 0x00480000 (disabled, pipe A, stall disabled, not detected, SDVO mult 1)
                       SDVOUDI: 0x0000004f
                        DSPARB: 0x00001d9c
                        DSPFW1: 0x00000000
                        DSPFW2: 0x00000000
                        DSPFW3: 0x00000000
                          ADPA: 0x00000c00 (disabled, pipe A, -hsync, -vsync)
                          LVDS: 0xc0308300 (enabled, pipe B, 18 bit, 1 channel)
                          DVOA: 0x00000000 (disabled, pipe A, no stall, -hsync, -vsync)
                          DVOB: 0x00480000 (disabled, pipe A, no stall, -hsync, -vsync)
                          DVOC: 0x00480000 (disabled, pipe A, no stall, -hsync, -vsync)
                   DVOA_SRCDIM: 0x00000000
                   DVOB_SRCDIM: 0x00000000
                   DVOC_SRCDIM: 0x00000000
                   BLC_PWM_CTL: 0x00000002
                  BLC_PWM_CTL2: 0x00000000
                    PP_CONTROL: 0xabcd0001 (power target: on)
                     PP_STATUS: 0xc0000008 (on, ready, sequencing idle)
                  PP_ON_DELAYS: 0x00c807d0
                 PP_OFF_DELAYS: 0x01900fa0
                    PP_DIVISOR: 0x00270f04
                  PFIT_CONTROL: 0x00000000
               PFIT_PGM_RATIOS: 0x00000000
               PORT_HOTPLUG_EN: 0x00000000
             PORT_HOTPLUG_STAT: 0x00000000
                      DSPACNTR: 0xd9000000 (enabled, pipe B)
                    DSPASTRIDE: 0x00002000 (8192 bytes)
                       DSPAPOS: 0x00000000 (0, 0)
                      DSPASIZE: 0x02ff03ff (1024, 768)
                      DSPABASE: 0x00800000
                      DSPASURF: 0x00000000
                   DSPATILEOFF: 0x00000000
                     PIPEACONF: 0x00000000 (disabled, single-wide)
                      PIPEASRC: 0x04ff03ff (1280, 1024)
                     PIPEASTAT: 0x00000000 (status:)
             PIPEA_GMCH_DATA_M: 0x00000000
             PIPEA_GMCH_DATA_N: 0x00000000
               PIPEA_DP_LINK_M: 0x00000000
               PIPEA_DP_LINK_N: 0x00000000
                 CURSOR_A_BASE: 0x00000000
              CURSOR_A_CONTROL: 0x00000000
             CURSOR_A_POSITION: 0x00000000
                          FPA0: 0x00020f03 (n = 2, m1 = 15, m2 = 3)
                          FPA1: 0x00020f03 (n = 2, m1 = 15, m2 = 3)
                        DPLL_A: 0x14020000 (disabled, non-dvo, default clock, DAC/serial mode, p1 = 2, p2 = 10)
                     DPLL_A_MD: 0x00000000
                      HTOTAL_A: 0x06af04ff (1280 active, 1712 total)
                      HBLANK_A: 0x06af04ff (1280 start, 1712 end)
                       HSYNC_A: 0x05d70557 (1368 start, 1496 end)
                      VTOTAL_A: 0x044f03ff (1024 active, 1104 total)
                      VBLANK_A: 0x044f03ff (1024 start, 1104 end)
                       VSYNC_A: 0x04090402 (1027 start, 1034 end)
                     BCLRPAT_A: 0x00000000
                  VSYNCSHIFT_A: 0x00000000
                      DSPBCNTR: 0x58000000 (disabled, pipe A)
                    DSPBSTRIDE: 0x00001400 (5120 bytes)
                       DSPBPOS: 0x00000000 (0, 0)
                      DSPBSIZE: 0x03ff04ff (1280, 1024)
                      DSPBBASE: 0x07900000
                      DSPBSURF: 0x00000000
                   DSPBTILEOFF: 0x00000000
                     PIPEBCONF: 0x80000000 (enabled, single-wide)
                      PIPEBSRC: 0x03ff02ff (1024, 768)
                     PIPEBSTAT: 0x00020000 (status: VBLANK_INT_ENABLE)
             PIPEB_GMCH_DATA_M: 0x00000000
             PIPEB_GMCH_DATA_N: 0x00000000
               PIPEB_DP_LINK_M: 0x00000000
               PIPEB_DP_LINK_N: 0x00000000
                 CURSOR_B_BASE: 0x00000000
              CURSOR_B_CONTROL: 0x10000000
             CURSOR_B_POSITION: 0x018001fb
                          FPB0: 0x00020c05 (n = 2, m1 = 12, m2 = 5)
                          FPB1: 0x00020c05 (n = 2, m1 = 12, m2 = 5)
                        DPLL_B: 0x98026000 (enabled, non-dvo, spread spectrum clock, LVDS mode, p1 = 2, p2 = 14)
                     DPLL_B_MD: 0x00000000
                      HTOTAL_B: 0x057f04ff (1280 active, 1408 total)
                      HBLANK_B: 0x057f04ff (1280 start, 1408 end)
                       HSYNC_B: 0x05340514 (1301 start, 1333 end)
                      VTOTAL_B: 0x032f031f (800 active, 816 total)
                      VBLANK_B: 0x032f031f (800 start, 816 end)
                       VSYNC_B: 0x03270323 (804 start, 808 end)
                     BCLRPAT_B: 0x00000000
                  VSYNCSHIFT_B: 0x00000000
             VCLK_DIVISOR_VGA0: 0x00031108
             VCLK_DIVISOR_VGA1: 0x00031406
                 VCLK_POST_DIV: 0x00800080
                      VGACNTRL: 0x80000000 (disabled)
                        TV_CTL: 0x000c0c00
                        TV_DAC: 0x70000000
                      TV_CSC_Y: 0x0332012d
                     TV_CSC_Y2: 0x07d30104
                      TV_CSC_U: 0x0733052d
                     TV_CSC_U2: 0x05c70200
                      TV_CSC_V: 0x0340030c
                     TV_CSC_V2: 0x06d00200
                  TV_CLR_KNOBS: 0x00606000
                  TV_CLR_LEVEL: 0x010b00e1
                    TV_H_CTL_1: 0x00400359
                    TV_H_CTL_2: 0x80480022
                    TV_H_CTL_3: 0x007c0344
                    TV_V_CTL_1: 0x00f01415
                    TV_V_CTL_2: 0x00060607
                    TV_V_CTL_3: 0x80120001
                    TV_V_CTL_4: 0x000900f0
                    TV_V_CTL_5: 0x000a00f0
                    TV_V_CTL_6: 0x000900f0
                    TV_V_CTL_7: 0x000a00f0
                   TV_SC_CTL_1: 0xc1710087
                   TV_SC_CTL_2: 0x6b405140
                   TV_SC_CTL_3: 0x00000000
                    TV_WIN_POS: 0x00360024
                   TV_WIN_SIZE: 0x02640198
               TV_FILTER_CTL_1: 0x800010bb
               TV_FILTER_CTL_2: 0x00028283
               TV_FILTER_CTL_3: 0x00014141
                 TV_CC_CONTROL: 0x00000000
                    TV_CC_DATA: 0x00000000
                   TV_H_LUMA_0: 0xb1403000
                  TV_H_LUMA_59: 0x0000b060
                 TV_H_CHROMA_0: 0xb1403000
                TV_H_CHROMA_59: 0x0000b060
                  FBC_CFB_BASE: 0x00000000
                   FBC_LL_BASE: 0x00000000
                   FBC_CONTROL: 0x00000000
                   FBC_COMMAND: 0x00000000
                    FBC_STATUS: 0x20000000
                  FBC_CONTROL2: 0x00000000
                 FBC_FENCE_OFF: 0x00000000
                   FBC_MOD_NUM: 0x00000000
                       MI_MODE: 0x00000000
                  MI_ARB_STATE: 0x00000840
                MI_RDRET_STATE: 0x00000000
                       ECOSKPD: 0x00000306
                          DP_B: 0x00000000
                DPB_AUX_CH_CTL: 0x00000000
              DPB_AUX_CH_DATA1: 0x00000000
              DPB_AUX_CH_DATA2: 0x00000000
              DPB_AUX_CH_DATA3: 0x00000000
              DPB_AUX_CH_DATA4: 0x00000000
              DPB_AUX_CH_DATA5: 0x00000000
                          DP_C: 0x00000000
                DPC_AUX_CH_CTL: 0x00000000
              DPC_AUX_CH_DATA1: 0x00000000
              DPC_AUX_CH_DATA2: 0x00000000
              DPC_AUX_CH_DATA3: 0x00000000
              DPC_AUX_CH_DATA4: 0x00000000
              DPC_AUX_CH_DATA5: 0x00000000
                          DP_D: 0x00000000
                DPD_AUX_CH_CTL: 0x00000000
              DPD_AUX_CH_DATA1: 0x00000000
              DPD_AUX_CH_DATA2: 0x00000000
              DPD_AUX_CH_DATA3: 0x00000000
              DPD_AUX_CH_DATA4: 0x00000000
              DPD_AUX_CH_DATA5: 0x00000000
                    AUD_CONFIG: 0x00000000
              AUD_HDMIW_STATUS: 0x00000000
                AUD_CONV_CHCNT: 0x00000000
                 VIDEO_DIP_CTL: 0x00000000
                 AUD_PINW_CNTR: 0x00000000
                   AUD_CNTL_ST: 0x00000000
                   AUD_PIN_CAP: 0x00000000
                  AUD_PINW_CAP: 0x00000000
            AUD_PINW_UNSOLRESP: 0x00000000
              AUD_OUT_DIG_CNVT: 0x00000000
                 AUD_OUT_CWCAP: 0x00000000
                   AUD_GRP_CAP: 0x00000000
                      FENCE  0: 0x00800341 (enabled, X tiled, 8192 pitch, 0x00800000 - 0x01000000 (8192kb))
                      FENCE  1: 0x0d000231 (enabled, X tiled, 4096 pitch, 0x0d000000 - 0x0d400000 (4096kb))
                      FENCE  2: 0x03300001 (enabled, X tiled,  512 pitch, 0x03300000 - 0x03400000 (1024kb))
                      FENCE  3: 0x04300001 (enabled, X tiled,  512 pitch, 0x04300000 - 0x04400000 (1024kb))
                      FENCE  4: 0x09900001 (enabled, X tiled,  512 pitch, 0x09900000 - 0x09a00000 (1024kb))
                      FENCE  5: 0x04600001 (enabled, X tiled,  512 pitch, 0x04600000 - 0x04700000 (1024kb))
                      FENCE  6: 0x0a000331 (enabled, X tiled, 4096 pitch, 0x0a000000 - 0x0a800000 (8192kb))
                      FENCE  7: 0x04400001 (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                      FENCE  8: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                      FENCE  9: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                     FENCE  10: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                     FENCE  11: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                     FENCE  12: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                     FENCE  13: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                     FENCE  14: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                     FENCE  15: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 0: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 0: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 1: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 1: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 2: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 2: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 3: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 3: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 4: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 4: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 5: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 5: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 6: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 6: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 7: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 7: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 8: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 8: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                 FENCE START 9: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                   FENCE END 9: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                FENCE START 10: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                  FENCE END 10: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                FENCE START 11: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                  FENCE END 11: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                FENCE START 12: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                  FENCE END 12: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                FENCE START 13: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                  FENCE END 13: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                FENCE START 14: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                  FENCE END 14: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                FENCE START 15: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                  FENCE END 15: 0xffffffff (enabled, X tiled,  512 pitch, 0x04400000 - 0x04500000 (1024kb))
                       INST_PM: 0x00000000
pipe A dot 108000 n 2 m1 15 m2 3 p1 2 p2 10
pipe B dot 68750 n 2 m1 12 m2 5 p1 2 p2 14
[root@wishmacer andrzejl]# 

I have just noticed upgrade to the kernel in the testing repository. I am still in mess caused by moving houses but decided to let You know that upgrading to this version of the kernel did not fixed the issue.

Regards.

Andrzej
Comment 18 Mika Kuoppala 2013-05-03 08:28:36 UTC
Hi Andrzej,

Thanks for the register dump. It seems that your display is
1280x800. And your game is running on 1024x768 resolution.
Panel fitter should be enabled in this case, but it isn't.

As i don't have those games, could you test if you can provoke
the bug just by using xrandr? So plain X desktop, do:

xrandr --output LVDS1 --mode 1024x768

And report your findings. intel_reg_dumper output as attachment
would be nice also. Thanks.
Comment 19 Daniel Vetter 2013-05-03 09:27:47 UTC
Ok, we've found a few more pfit bugs, but they should be all fixed in 3.10 (drm merge just landed in Linus' git repo). Can you please test latest upstream git or 3.10-rc1 once it shows up?

pfit stuff on gen2/3 was a bit a mess in 3.9 and it took us a while to figure it out. But once -rc1 is out and we have some test coverage that it doesn't blow up again we'll start backporting to stable kernels. Your test feedback would obviously accelerate this.
Comment 20 Daniel Vetter 2013-05-31 20:30:45 UTC
Presumed fixed, please reopen with the test results supplied if not.

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.