Vanzetti_: hi, yesterday you said me radeon driver should work with my card, but not
Vanzetti_: v3400 doesn't work with radeon driver
udovdh: who is you?
udovdh: and does radeonhd do v3400?
airlied: Vanzetti_: you need the latest git master..
airlied: Vanzetti_: we just merged atombios support.
Vanzetti_: the radeonhd yes, but withotu acceleration
Vanzetti_: I got the git version
udovdh: there is no accelleration in radeonhd yet
udovdh: it is work in progress
airlied: Vanzetti_: what pciid?
Vanzetti_: udovdh: yes I know
Vanzetti_: airlied: I use dual screen
airlied: is v3400 a firegl?
Vanzetti_: 07:00.0 VGA compatible controller: ATI Technologies Inc Unknown device 71d2
Vanzetti_: 07:00.1 Display controller: ATI Technologies Inc Unknown device 71f2
Vanzetti_: X say this :
Vanzetti_: (WW) RADEON: No matching Device section for instance (BusID PCI:7:0:1) found
airlied: Vanzetti_: that bits fine..
airlied: 7:0:0 is the main device..
airlied: Vanzetti_: the master of git should work with that pci ..
airlied: Vanzetti_: pleast pastebin a log and jump to #radeon
Vanzetti_: ok, but how I configure my 2nd output ?
udovdh: howto for that is in the wiki (partly)?
udovdh: else I havesome other url for you
udovdh: btw: why does it say RADEON in stead of RADEONHD?
sonne: libv, sure that you fixed the color screwup bug?
sonne: because I still have it :(
sonne: libv, see http://nn7.de/debugging/Xorg.0.log_new2
egbert: sonne: a color screwup bug with suspend/resume was fixed. is this what you are seeing? maybe you've got a different problem...
sonne: egbert, I exactly have this bug... and it is still there :-)
egbert: have you checked that you are running the latest driver?
sonne: egbert, I was hoping that the log I posted above could tell me...
noop: anyone here?
noop: I get a question about R630 with radeon-git
udovdh: i am awake :-)
udovdh: and I have an rv630 (hd2600pro, you have an xt?)
noop: udovdh, back
noop: udovdh, what u use? radeonhd?
noop: I get a HD2400pro AGP.
udovdh: i use radeonhd
udovdh: the hd2600pro I have is pci-E
noop: have u ever try radeon?
noop: the radeon0 support R500 &R600 now.
noop: udovdh, I still got some problem when I using radeonhd
noop: sometimes it'll hangs.
udovdh: oh? then you must speak to a real developer I guess
noop: yeah, the developer is here.
udovdh: if you can, maybe start Xorg with -logverbose 7 or higher
udovdh: and maybe you know how to reproduce the hang?
udovdh: then more detail might be found about the cause
noop: still no idea. I dont have time check last time when I get hanged.
udovdh: be patient, developers do read here
udovdh: maybe try to write down each hang?
udovdh: and is it just Xorg or the whole machine?
udovdh: does it also hang when using vesa?
noop: but it'll happens when I using firefox, vmware-workstation and gaim(pidgin).
udovdh: only that combo?
noop: vesa? I never tried vesa.
udovdh: leave out vmware and try again?
udovdh: try to eliminate the cause
noop: I'm doing,firefox + xchat is okay.
noop: just the Xorg hangs. not the system.
noop: I think maybe not related to radeonhd ?
noop: It's a module, if radeonhd hangs , the whole system should hang.
noop: If there is a bug like double free or someting , the whole system should crash.
udovdh: yes, that's why you should try without vmware, or any other application to see if the hang does occur or not
udovdh: i.e.: eliminate the source
rbmorse: Congratulations on the 1.1 release. Installation from git on my Ubuntu 7.10 /R 580+ machine totally nominal. Works great!
libv: rbmorse: great :)
libv: rbmorse: we of course still have bugs open, but i think we fixed enough important ones already to wrap things up again
libv: too bad that we didn't manage a proper level of rv670 support though
udovdh: There is a 1.1?
sonne: libv, the color screwup bug is also not fixed...
sonne: at least I still have it after a suspend resume cycle
udovdh: when/how does it screw up?
udovdh: ah, oh
sonne: and yes I just recompiled...
udovdh: after a git pull?
sonne: udovdh http://nn7.de/shots/distortion.jpg
sonne: yes after a fresh pull, restart X 1 minute ago
udovdh: hmm. pretty bad...
udovdh: thanks for the pic
sonne: Xorg -logverbose 7 log here: http://nn7.de/debugging/Xorg.0.log_new2
udovdh: i hope the devs can find the bug
sonne: libv, ?
libv: sonne: egbert told me you were on a mac
libv: sonne: right?
libv: sonne: for some reason, macs are special
libv: sonne: for some reason, more than half the open bugs are macs :)
udovdh: differen bios?
sonne: great we can help ;-)
udovdh: same chip
libv: sonne: :)
sonne: well I only see the magic colors and the ddc problems so far...
libv: "only" :p
sonne: libv, do you want cold freezes and other great issues?
libv: sonne: no! no! thanks very much!
sonne: then go ahead `fix' the driver ;-)
libv: sonne: so this is still the suspend resume cycle, and then enabling the external display
libv: sonne: both tmds and vga?
libv: sonne: and the cursor is ok?
libv: or not?
sonne: actually regarding the ddc bug, I recognized that I can quite reliably get ddc to work on my display at home after I do a suspend to ram cycle. However here at work this *never* helped (different display) ... so It might be a timing thing...
libv: because last time the cursor was not ok, as gamma influences it directly
libv: sonne: yup, we know that trick already :)
libv: sonne: it is not a timing thing
udovdh: the knock on the door issue?
sonne: libv, exactly. I only tried tmds and yes the cursor is OK
libv: sonne: some io line is raised or dropped during suspend resume which suddenly makes it work i think
sonne: libv, misunderstanding, this trick does *not* work - never ever has here at work
libv: right, didn't read completely
sonne: but it works reliably on a tft at home...
libv: (too much all over the place)
libv: sonne: ooh, that is very interesting
sonne: (and on some beamer I tried recently)
libv: so yes, your conclusion seems very correct
sonne: (had to give a talk...)
libv: it does point at a timing issue then indeed
libv: sonne: can you explain this again in this bug: https://bugs.freedesktop.org/show_bug.cgi?id=13470
libv: sonne: i could add an explanation myself, but imho, since you dug this out, you should take the credit for it :)
libv: thing is... the ddc engine needs to get a certain clock scaling value
libv: so that it uses an acceptable bitrate
libv: now this clock scaling value is retrieved by reading in some atombios predefined value
libv: and i don't think this predefined value is changed dynamically
libv: so... after resume, this value might match the actual timing
libv: and then it magically works?
libv: so, in theory, if the prescale value is set to maximum, it should work
libv: sonne: let's open a new one for the colour stuff
libv: sonne: i'll do that and cc you :) since you're already documenting your discovery
libv: sonne: also, does xgamma fix the issue, like it did in the other cases?
udovdh: i'm gone for now
libv: heh, 5 out of 9 open bugs are apples :(
Fry-kun: can someone help me get my external monitor running with correct resolution?
Fry-kun: it has messed up EDID, I think