Dave Witbrodt wrote:

The next items on my list are:

A) Try booting with "radeon.modeset=0". [Simple idea, but it didn't occur to me until just now.]

[As mentioned in previous msg:]
Using "radeon.modeset=0" eliminates the pink line. This is not a desirable solution, since I lose KMS and (on Evergreen) my virtual terminals if I start X.


B) Pull the HD 5750 and install my HD 4850. The older card was working perfectly with KMS (even with a KVM + Dsub connection!) before I bought the new card to try out the Evergreen support. I can try VGA/Dsub, DVI, and DVI-to-HDMI cables with the old card and see which permutations (if any) cause the pink vertical line to disappear. If no pink line occurs at all, then the problem would seem to be Evergreen related.

OK. Just swapped in the HD 4850 in place of the Evergreen card. Same kernel (2.6.33 with drm-radeon-testing of 100223 [git 383be5d] merged) and same DVI-to-HDMI converter cable. Result: no pink line.

The problem is with the Evergreen card only. This means some hardware fault (doubtful) or something lacking in the current Evergreen support.

Please understand that none of these messages are a complaint. I am actually ecstatic about all of the efforts of the people working on open source driver support for ATI hardware. These messages are simply a report about a problem I'm having using my new monitor with drm-radeon-testing.


C) Try a real HDMI cable instead. I mentioned that I don't own one, but one is in the mail and should be here early next week. :)

Will be a few days before I can test this. Until then, I'll probably just leave the HD 4850 installed... and enjoy some 2D/3D acceleration once again! ;)

OK. HDMI cable arrived, and I have now installed the HD 5750 and rebooted. [Xorg.0.log (compressed) attached.]

Results:

1) Pink vertical line on left side of monitor still there.
2) Monitor OSD info still reports 1922x1200, a seemingly impossible video mode!
3) Screen size still altered (either by X server or by radeon):

    $ xdpyinfo | grep -A 2 "screen #0"
    screen #0:
      dimensions:    1920x1200 pixels (508x317 millimeters)
      resolution:    96x96 dots per inch

The screen size ought to be 593x371, and the DPI should be 82.

4) On Saturday, I received a message in reply to this thread from a HD 5870 user. I asked him to repeat what he said to me on this list, but so far he has not done so. I am going to defy etiquette/decorum, and copy the two sentences that he wrote to me (sorry, Andrew, for doing this without permission):

    I also got a (purple?) line with HDMI on the far left my 5870 w/
    kms.
    So you're not crazy  :)

Personal conclusions:

I believe that something is broken in the way that the Evergreen code deals with HDMI connections. My results are with HD 5750, and the other user has HD 5870, so it's not one single card (or even model) that is affected.

He and I are both using kernel 2.6.33 with drm-radeon-testing merged in. There are code conflicts which had to be resolved, and it is possible (maybe even likely?) that either or both of us made errors in the merge. Speaking for myself, I also have tried a kernel built directly from drm-radeon-testing (no merging, no modifications) with identical results. Noob-damaged merges cannot be the only problem, as it turns out.

Except for the impossible 1922 horizontal resolution reported (only 1920 is _physically_ possible with this hardware) and resulting 2 pixel-wide vertical line on the far left of the screen, the picture is just fine.

When I was first experimenting with HD 4850 in November, only software rendering was available. It is surprising how many programs are usable without acceleration! However, I became spoiled once I learned to build upstream packages of xf86-video-radeon, libdrm, and mesa. Since it only takes 5-10 minutes for me to swap the HD 5750 in, I think I'll stick with the HD 4850 until some more support starts to appear.

I will carry out any direct requests to perform testing with the HD 5750; I will also be doing testing of my own as I detect any new Evergreen support announced (or appearing in git repos). Thanks again for what has been done so far!


Dave W.

Attachment: HD5750-modeset=1-HDMI-Xorg.0.log.bz2
Description: application/bzip

_______________________________________________
xorg-driver-ati mailing list
[email protected]
http://lists.x.org/mailman/listinfo/xorg-driver-ati

Reply via email to