Bug 17357

Created attachment 18577 [details]
Screenshot of desktop showing video thumbnail without stripes in the thumbnail image

Created attachment 18578 [details]
Video playing stripes visible in video frame

Created attachment 18579 [details]
Video playing no stripes visible in this frame

Created attachment 18581 [details]
/var/log/messages

Created attachment 18582 [details]
/var/log/Xorg.0.log

Additional data points;

1. If I go to the website this video is from http://www.nimportequi.com/ and watch the original "VIDEO STAR DECATHLON - PEKIN 2008" which plays with flash player 9 under firefox 3.0.1 there are no stripes in the video at all.

2. If I play the same file on a Windows XP SP2 machine (via samba) with Media Player 10 there are no stripes in the video at all.

3. Similar artefacts, smaller height and more of (12 rows of), observed in playback of mpeg-2 streams captured from the hauppage video capture card in this machine (520x576).

Original file is 320x240 and appers to have 2 rows of artefacts.


Conclusion, the artefacts are not a result of the conversion from flash to windows media 9 format so the file is ok, it is not a problem with playback of the windows media 9 format as similar issue present in mpeg-2 format playback, flash player appears to use a different video playback mechanism to totem not triggering the problem.

I think on your card the texture adaptor is the default, can you try playing the vid with mplayer and compare the texture adaptor with the blitter and see if the problem remains?(mplayer allows you to specify ports/adaptors, read the man page, use xvinfo to see which adaptors are available, and what is their base port).

danny

Danny, 

nouveau driver version has now cnaged, ROAF pushed an update.

I stopped vmware from starting on boot.

xvinfo reports I have 

  Adaptor #0: "NV40 texture adapter"
    number of ports: 32
    port base: 128

  Adaptor #1: "NV40 high quality adapter"
    number of ports: 32
    port base: 160

  Adaptor #2: "NV Video Blitter"
    number of ports: 16
    port base: 192

mplayer Home-made-olympics.wmv -vo xv:port={128, 160, 192}

All result in the same artifacts in the video.

dmesg reports

[   64.891532] Bluetooth: RFCOMM TTY layer initialized
[   64.891534] Bluetooth: RFCOMM ver 1.8
[   66.026122] [drm] Initialized drm 1.1.0 20060810
[   66.142744] ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16
[   66.142755] PCI: Setting latency timer of device 0000:01:00.0 to 64
[   66.142829] [drm] Detected an NV40 generation card (0x04b300a2)
[   66.142833] [drm] Initialized nouveau 0.0.11 drm-2.3.0-1592-gcd7d71f19cf2a1652 on minor 0
[   66.611961] [drm] Allocating FIFO number 0
[   66.613996] [drm] nouveau_fifo_alloc: initialised FIFO 0
[   66.614398] [drm] Allocating FIFO number 1
[   66.616272] [drm] nouveau_fifo_alloc: initialised FIFO 1
[   67.169154] eth0: no IPv6 routers present
[ 1017.330778] [drm] no space left after collision
[ 1017.330784] [drm:nouveau_ramht_insert] *ERROR* RAMHT space exhausted. ch=1
[ 1017.330789] [drm:nouveau_notifier_alloc] *ERROR* Error referencing notifier ctxdma: -12

/var/log/messages notes

Sep 12 20:45:19 hereford kernel: [   64.891532] Bluetooth: RFCOMM TTY layer initialized
Sep 12 20:45:19 hereford kernel: [   64.891534] Bluetooth: RFCOMM ver 1.8
Sep 12 20:45:20 hereford kernel: [   66.026122] [drm] Initialized drm 1.1.0 20060810
Sep 12 20:45:20 hereford kernel: [   66.142744] ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16
Sep 12 20:45:20 hereford kernel: [   66.142829] [drm] Detected an NV40 generation card (0x04b300a2)
Sep 12 20:45:20 hereford kernel: [   66.142833] [drm] Initialized nouveau 0.0.11 drm-2.3.0-1592-gcd7d71f19cf2a1652 on minor 0
Sep 12 20:45:20 hereford kernel: [   66.611961] [drm] Allocating FIFO number 0
Sep 12 20:45:20 hereford kernel: [   66.613996] [drm] nouveau_fifo_alloc: initialised FIFO 0
Sep 12 20:45:20 hereford kernel: [   66.614398] [drm] Allocating FIFO number 1
Sep 12 20:45:20 hereford kernel: [   66.616272] [drm] nouveau_fifo_alloc: initialised FIFO 1
Sep 12 21:01:12 hereford kernel: [ 1017.330778] [drm] no space left after collision

Even with the collision entry and the RAMHT space exhausted video continues to play, will attach updated dmesg, /var/log/Xorg.0.log and /var/log/messages 

Created attachment 18839 [details]
/var/log/messages

Created attachment 18840 [details]
/var/log/Xorg.0.log

That's interesting. Since it happens even with the blitter, and doesn't happen with flashplayer I wonder if this is not a bug in mplayer (or some other component on your system) instead. Do you have the problem with nv or the binary nvidia driver?

I never saw those drm errors before, maybe someone else can comment on that, and whether or not it can be related. 

danny

When using nv no artifacts in video, will attach Xorg.0.log from nv usage

Created attachment 18914 [details]
Xorg log using the nv driver

so the collision message is 100% related to the video artifacts?

You need to load the drm with debug enabled to see what kind of collisions happen and why. But I am not sure it is necessary. Perhaps somebody who understands that gpu object creating code can offer some insight:)

danny

havent done the drm debug=1 thing as yet, but note that the same behaviour persists in ubuntu 8.10 

kernel 2.6.27
xorg xserver 1.5.2

[  125.225131] Symbol init_mm is marked as UNUSED, however this module is using it.
[  125.225139] This symbol will go away in the future.
[  125.225142] Please evalute if this is the right api to use and if it really is, submit a report the linux kernel mailinglist together with submitting your code for inclusion.
[  125.241684] [drm] Initialized drm 1.1.0 20060810
[  125.257724] nouveau 0000:01:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[  125.257734] nouveau 0000:01:00.0: setting latency timer to 64
[  125.259034] [drm] Detected an NV40 generation card (0x04b300a2)
[  125.259042] [drm] Initialized nouveau 0.0.11 libdrm-2.4.1 on minor 0
[  125.686046] [drm] Allocating FIFO number 0
[  125.688809] [drm] nouveau_fifo_alloc: initialised FIFO 0
[  125.690011] [drm] Allocating FIFO number 1
[  125.692352] [drm] nouveau_fifo_alloc: initialised FIFO 1

Have upgraded to Ubuntu 9.10 and tried the Nvidia 185 binary driver, the Nvidia driver displays the same stripes I see with nouveau, closing as nouveau is RE´ed from the binary driver so nouveau is not going to be able to fix.   May also be a hardware issue ..... closing