Bug 25141

Summary: Unusably poor performance on Radeon M6 LY
Product: xorg Reporter: Martin Dratva <raquacontact>
Component: Driver/RadeonAssignee: xf86-video-ati maintainers <xorg-driver-ati>
Status: RESOLVED INVALID QA Contact: Xorg Project Team <xorg-team>
Severity: major    
Priority: medium CC: orion
Version: 7.4 (2008.09)   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
xorg.conf that worked well with Ubuntu 9.04
none
Testing XAA + RenderAccel false files
none
used xorg.conf
none
screenshot
none
Xorg log none

Description Martin Dratva 2009-11-17 03:29:05 UTC 
I switched from Ubuntu 9.04 to Ubuntu 9.10 and my Radeon Mobility M6 LY card performance dropped radically. 

Moving a window around or scrolling in text console makes CPU go to 100%, looks like HW acceleration is not working. 

Even before on 9.04 I had to tweak xorg.conf to switch back to XAA mode, because otherwise performance was poor. 
Now it seems that this workaround does not work in new Ubuntu version and I can not get it to work. 
Using some tweaks I was able to make it work at least a bit better, but I am really not seeing the performance I was used to. 

Please see also this Ubuntu bug report where I attached all the necessary files.
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/478685

I had the same problems with Mandirva 2010 Live CD I tried so I guess it is not sole Ubuntu bug.

If I can add any other info or try some configuration, let me know.
Comment 1 Michel Dänzer 2009-11-17 03:54:11 UTC 
You'd probably need to disable KMS to be able to use XAA again, e.g. by adding radeon.modeset=0 on the kernel command line.

I think it's mostly up to the drivers to deal with the limitations (in particular little VRAM) of your setup, reassigning.
Comment 2 Martin Dratva 2009-11-17 05:53:24 UTC 
I tried to disable KMS but it was even worse. According to Xorg.0.log XAA mode was loaded, but with a lot of artefacts on screen and problems. Tried my old xorg.conf file, which worked on previous Ubuntu version, but no luck. I tried to disable options I had there one by one but it did not help.
Comment 3 Alex Deucher 2009-11-17 07:55:28 UTC 
Does:
Option "RenderAccel" "False"
fix the XAA with UMS issues?  I think XAA render accel is broken now in newer X servers.
Comment 4 Martin Dratva 2009-11-17 08:17:57 UTC 
Created attachment 31261 [details]
xorg.conf that worked well with Ubuntu 9.04

Option "RenderAccel" "False" 

did not help. It made things even worse I'd say.

I'm attaching my last good xorg.conf, that I was using with Ubuntu 9.04. I tried to change "RenderAccel" to false in this one and I also tried to use only this option + XAA with the same result.
Comment 5 Alex Deucher 2009-11-17 08:30:02 UTC 
Can you attach your log with renderaccel set to false?
Comment 6 Martin Dratva 2009-11-17 08:44:15 UTC 
Created attachment 31264 [details]
Testing XAA + RenderAccel false files

Here are files I have tested with. And I passed a radeon.modeset=0 parameter to kernel. 
There is a lot of artifacts, bad repainting, etc. See attached screenshot.
Comment 7 Martin Dratva 2009-11-17 08:44:36 UTC 
Created attachment 31265 [details]
used xorg.conf
Comment 8 Martin Dratva 2009-11-17 08:44:53 UTC 
Created attachment 31266 [details]
screenshot
Comment 9 Orion Poplawski 2010-01-04 13:14:23 UTC 
Same problems on Fedora 12 (https://bugzilla.redhat.com/show_bug.cgi?id=539624).  Working with the latest builds it is mostly working if I keep external monitor at 1024x768 or under.
Comment 10 Orion Poplawski 2010-01-14 09:48:15 UTC 
Created attachment 32639 [details]
Xorg log

This is from Fedora 12 with:

kernel-2.6.32.3-24.fc12.i686
xorg-x11-drv-ati-6.13.0-0.20.20091221git4b05c47ac.fc12.i686
libdrm-2.4.17-1.fc12.i686
xorg-x11-server-Xorg-1.7.4-1.fc12.i686

Window redraws are still so slow as to be unusable.

xorg.conf:

Section "Device"
        Identifier  "Videocard0"
        Driver      "radeon"
#Only works in "nomodeset" mode
        Option      "IgnoreLidStatus" "false"
        Option      "Monitor-VGA-0" "Dell"
EndSection

Section "Monitor"
        Identifier   "Dell"
        VendorName   "Dell"
        ModelName    "2001FP"
        Option      "dpms"
        Option      "PreferredMode" "1600x1200"
EndSection

Section "Screen"
        Identifier "Screen0"
        #DefaultDepth 16
        SubSection "Display"
                Virtual 1600 1200
        EndSubSection
EndSection
Comment 11 Alex Deucher 2010-10-19 18:43:03 UTC 
Is this still an issue with a newer kernel and driver?
Comment 12 Orion Poplawski 2010-10-20 08:45:05 UTC 
Still an issue with:
2.6.35.6-46.fc14.i686
xorg-x11-server-Xorg-1.9.0-15.fc14.i686
xorg-x11-drv-ati-6.13.1-0.3.20100705git37b348059.fc14.i686

In fact doing xrandr --output LVDS --off appears to lockup the machine completely after a few seconds (and the X server immediately).

Oct 20 09:33:56 makani kernel: [  221.741700] BUG: unable to handle kernel paging request at 3432307d
Oct 20 09:33:56 makani kernel: [  221.741709] IP: [<c045e232>] do_raw_spin_lock+0xd/0x1e
Oct 20 09:33:56 makani kernel: [  221.741718] *pde = 00000000 
Oct 20 09:33:56 makani kernel: [  221.741722] Oops: 0002 [#1] SMP 
Oct 20 09:33:56 makani kernel: [  221.741726] last sysfs file: /sys/devices/pci0000:00/0000:00:1e.0/0000:02:02.0/rf_kill
Oct 20 09:33:56 makani kernel: [  221.741731] Modules linked in: nfs lockd fscache nfs_acl auth_rpcgss michael_mic arc4 ecb lib80211_crypt_tkip aes_i586 aes_generic lib80211_crypt_ccmp sunrpc 
ipv6 cpufreq_ondemand acpi_cpufreq mperf uinput snd_intel8x0 snd_intel8x0m snd_ac97_codec ac97_bus snd_seq snd_seq_device snd_pcm snd_timer thinkpad_acpi snd ipw2200 nsc_ircc e1000 iTCO_wdt snd_page_alloc soundcore i2c_i801 ppdev libipw cfg80211 rfkill iTCO_vendor_support parport_pc irda
 lib80211 crc_ccitt parport microcode firewire_ohci firewire_core video yenta_socket crc_itu_t output radeon ttm drm_kms_helper drm i2c_algo_bit i2c_core [last unloaded: scsi_wait_scan]
Oct 20 09:33:56 makani kernel: [  221.741786] 
Oct 20 09:33:56 makani kernel: [  221.741790] Pid: 1756, comm: X Not tainted 2.6.35.6-46.fc14.i686 #1 2884M5U/2884M5U
Oct 20 09:33:56 makani kernel: [  221.741794] EIP: 0060:[<c045e232>] EFLAGS: 00013292 CPU: 0
Oct 20 09:33:56 makani kernel: [  221.741798] EIP is at do_raw_spin_lock+0xd/0x1e
Oct 20 09:33:56 makani kernel: [  221.741801] EAX: 3432307d EBX: 34323031 ECX: 00000000 EDX: 00000100
Oct 20 09:33:56 makani kernel: [  221.741804] ESI: f1130a2c EDI: 00000001 EBP: f3afdd0c ESP: f3afdd0c
Oct 20 09:33:56 makani kernel: [  221.741807]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
Oct 20 09:33:56 makani kernel: [  221.741811] Process X (pid: 1756, ti=f3afc000 task=f3ba9940 task.ti=f3afc000)
Oct 20 09:33:56 makani kernel: [  221.741814] Stack:
Oct 20 09:33:56 makani kernel: [  221.741816]  f3afdd14 c07a6f84 f3afdd2c f78e81b5 00000000 f1130a00 f1130a00 f6a0c000
Oct 20 09:33:56 makani kernel: [  221.741823] <0> f3afdd48 f79c8be6 00000000 00000000 f6168000 f1130a00 f6a0c000 f3afdd98
Oct 20 09:33:56 makani kernel: [  221.741831] <0> f79c8ce5 f3afdd58 c042a21b f3afdd60 c042a232 f3afdd68 00000000 00000006
Oct 20 09:33:56 makani kernel: [  221.741839] Call Trace:
Oct 20 09:33:56 makani kernel: [  221.741848]  [<c07a6f84>] ? _raw_spin_lock+0xd/0xf
Oct 20 09:33:56 makani kernel: [  221.741864]  [<f78e81b5>] ? ttm_bo_reserve+0x25/0x73 [ttm]
Oct 20 09:33:56 makani kernel: [  221.741902]  [<f79c8be6>] ? radeon_bo_reserve.clone.3+0x20/0x61 [radeon]
Oct 20 09:33:56 makani kernel: [  221.741937]  [<f79c8ce5>] ? radeon_crtc_set_base+0xbe/0x43a [radeon]
Oct 20 09:33:56 makani kernel: [  221.741943]  [<c042a21b>] ? need_resched+0x19/0x23
Oct 20 09:33:56 makani kernel: [  221.741947]  [<c042a232>] ? should_resched+0xd/0x28
Oct 20 09:33:56 makani kernel: [  221.741951]  [<c07a6391>] ? mutex_lock+0x24/0x3e
Oct 20 09:33:56 makani kernel: [  221.741982]  [<f7897949>] ? drm_mode_object_find+0x52/0x5a [drm]
Oct 20 09:33:56 makani kernel: [  221.741995]  [<f78c9cdb>] ? drm_crtc_helper_set_config+0x476/0x673 [drm_kms_helper]
Oct 20 09:33:56 makani kernel: [  221.742004]  [<f78c9dd1>] ? drm_crtc_helper_set_config+0x56c/0x673 [drm_kms_helper]
Oct 20 09:33:56 makani kernel: [  221.742011]  [<f7897949>] ? drm_mode_object_find+0x52/0x5a [drm]
Oct 20 09:33:56 makani kernel: [  221.742011]  [<f7898d5b>] ? drm_mode_setcrtc+0x23e/0x266 [drm]
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c042a232>] ? should_resched+0xd/0x28
Oct 20 09:33:56 makani kernel: [  221.742011]  [<f788f80e>] ? drm_ioctl+0x28d/0x378 [drm]
Oct 20 09:33:56 makani kernel: [  221.742011]  [<f7898b1d>] ? drm_mode_setcrtc+0x0/0x266 [drm]
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c06864a2>] ? ahci_init_one+0x82c/0x88a
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c0583188>] ? file_has_perm+0x8d/0xa7
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c04e5aa4>] ? vfs_ioctl+0x33/0x91
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c06864a2>] ? ahci_init_one+0x82c/0x88a
Oct 20 09:33:56 makani kernel: [  221.742011]  [<f788f581>] ? drm_ioctl+0x0/0x378 [drm]
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c04e64a4>] ? do_vfs_ioctl+0x462/0x48d
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c06864a2>] ? ahci_init_one+0x82c/0x88a
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c058348f>] ? selinux_file_ioctl+0x43/0x46
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c06864a2>] ? ahci_init_one+0x82c/0x88a
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c04e6515>] ? sys_ioctl+0x46/0x69
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c06864a2>] ? ahci_init_one+0x82c/0x88a
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c07a737c>] ? syscall_call+0x7/0xb
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c06864a2>] ? ahci_init_one+0x82c/0x88a
Oct 20 09:33:56 makani kernel: [  221.742011]  [<c06864a2>] ? ahci_init_one+0x82c/0x88a
Oct 20 09:33:56 makani kernel: [  221.742011] Code: 55 89 e5 0f 1f 44 00 00 3e ff 00 5d c3 55 89 e5 0f 1f 44 00 00 3e 81 00 00 00 00 01 5d c3 55 89 e5 0f 1f 44 00 00 ba 00 01 00 00 <3e> 66 0f c1 10 38 f2 74 06 f3 90 8a 10 eb f6 5d c3 55 89 e5 0f 
Oct 20 09:33:56 makani kernel: [  221.742011] EIP: [<c045e232>] do_raw_spin_lock+0xd/0x1e SS:ESP 0068:f3afdd0c
Oct 20 09:33:56 makani kernel: [  221.742011] CR2: 000000003432307d
Oct 20 09:33:56 makani kernel: [  221.742242] ---[ end trace e439adf3dd44d13a ]---
Comment 13 Adam Jackson 2018-06-12 19:07:35 UTC 
Mass closure: This bug has been untouched for more than six years, and is not
obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases.

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.