paul`: does anyone know more about dual-head users of the radeonhd driver having the wrong colours on the second output in Hardy Heron?
tulcod: paul`: "the wrong colours" sounds like a bug
tulcod: paul`: which version of radeonhd?\
tulcod: paul`: and... which gpu, of course?
paul`: well, the top of 'git log' says
paul`: commit 6f34607b6b1b0d700c68f8b19de4445a547c34a4
paul`: Author: Egbert Eich
paul`: Date: Mon Apr 21 19:47:27 2008 +0200
paul`: Fix mapping flags for FB when using libpciaccess.
paul`: Setting the map range to both WC and cachable doesn't make sense.
paul`: The cachable flag which has precedence causes stability problems
paul`: on VT switch.
paul`: Also more reading of FbIntAddress to the MC setup code.
tulcod: paul`: please
tulcod: don't paste here
paul`: anyway lspci says:
paul`: Display controller: ATI Technologies Inc RV516 [Radeon X1300 Pro] (Secondary)
paul`: it worked mostly OK in Gutsy Gibbon
tulcod: paul`: well, it's a git version, you can't exactly expect it to work right :)
tulcod: paul`: define "wrong colours" though
paul`: what version should I use?
tulcod: cause it should work most of hte time
tulcod: paul`: if you want to use a git version, use a git version which works for you ;)
tulcod: paul`: if you want a release, just the latest, which is 1.2.1
tulcod: paul`: but please define "wrong colours"
paul`: grey shows up as white
paul`: some light blue colours also show up as white
tulcod: so some colours are replaced by white?
paul`: i can make it some kind of blue with the 'xgamma' utility
paul`: but then the other output also gets darker
paul`: and the letters become blurred
paul`: it seems to work OK with DVI monitors
tulcod: paul`: well... i'm no gpu driver developer, so i can't tell exactly what the problem is
paul`: but not with analog monitors
tulcod: but you might want to try the latest git
paul`: but you just said not to use the git version?
tulcod: i didn't :p
tulcod: paul`: 1.2.1 *should* work for you, have you tried it?
paul`: is 1.2.1 in Hardy Heron?
tulcod: i have no idea
paul`: I started using the git version in Gutsy because the drivers that came with Ubuntu didn't work
tulcod: paul`: don't you have a package manager or something?
tulcod: i mean, i knew ubuntu lacked clarity for a long time, but doesn't it tell you the version of packages it has?
paul`: well, I see there's a package called xserver-xorg-video-radeonhd
paul`: the version is 1.1.0-1
tulcod: paul`: well that's not 1.2.1 is it?
tulcod: weird ubuntu guys...
paul`: hi, I'm back
paul`: I've just swapped monitors with my colleague
paul`: now I have DVI monitors and it works fine
tulcod: paul`: eh, okay...
tulcod: paul`: well, gratz then :)
paul`: so, do you think I should use Hardy's driver, the 1.2.1 driver, the version now in git, or just keep it as is now?
tulcod: paul`: if it works ok for you, then stay where you are
paul`: OK, I'll do that
tulcod: paul`: if a new feature appears in which you're interested, wait for a release
ndim: Interesting. The http://lenovoblogs.com/insidethebox/ reports that my Advanced MiniDock CAN support two simultaneous external displays for the ThinkPad...
ndim: ...but the graphics chip can only control two independent screens :-(
tulcod: ndim: lol, sue somebody
ndim: tulcod: lol, nah.
ndim: But it means that there might be laptops which do support three outputs total.
agd5f: ndim: two external displays if you don't use the internal one
agd5f: in theory 3 if you clone 2 of the outputs
agd5f: but only two can have separate timings
ndim: agd5f: With the current hardware I have, right. It might be possible to have three independents if you build the right hardware into the laptop, but I doubt anybody does that.
ndim: Typical software can only use one display anyway.
ndim: So there might be a business case for one+1, but not for one+2.
ndim: I'm not aware even of presentation software which displays the slides on the external projector and the slide, speaker notes, and a running clock on the laptop's internal display for the speaker.
agd5f: ndim: openoffice supports that iirc
libv: hrm... many laptops support three outputs.
libv: lvds panel of course, and then commonly VGA + either of s-video or dvi/hdmi
ndim: libv: With different content on all three displays? Nice.
libv: ndim: no, but three different outputs distributed over 2crts
libv: crtcs evem
libv: for igps, unless you go for embedded or semiembedded, buying laptops is usually the cheapest way (these days) to get at least three outputs in one go
libv: luckily things are kind of changing... many microatx motherboards with rs690 and i think even rs780 come with at least a dac and tmds used
libv: and sometimes even s-video
agd5f: emmes: does teh dri work if you disable shadowfb? I suspect that may be the problem
crow: where should i look for underscan option, i dont know if update to radeonhd driver will bring me to right resolution egg not underscaned/overscaned...
yangman: crow: underscanning is a TV-specific thing. since TV-out isn't supported yet, it's not relevant for radeonhd
yangman: crow: with digital signals, it's something you'd have to adjust on the TV, not the source
crow: yangman its ATI Xpress 1250 (intel) with HDMI out, wich is connect on LCD TV (HDMI in)
yangman: right. so, tweak the TV
crow: yangman in winxp there is no such need to tweak tv, just turn underscan to 0
crow: its something like in this post from ubuntu http://ubuntuforums.org/showthread.php?t=667335
yangman: I don't know what XP does
crow: yangman well xp nothing, but ATI driver in xp
agd5f: crow: the driver probably just reduces the visible area in that case.
crow: agd5f yea right, its underscaned, i just somehow need to put underscan to 0...
agd5f: your tv probably has an option to turn off scaling or overscan
crow: doesnt have to do anything with tv, its doing 1:1 maping
yangman: would you try adjusting the TV anyway, for kicks?
yangman: the driver doesn't do any scaling at all atm, underscanning related or not
yangman: there's nothing to tweak on the driver side
crow: yangman well i dont know what to adjust, egg when i install ATI driver in XP/Vista i just adjust it using slide bar in CCC, and i got perfect 1:1 mapping,
yangman: does your TV have an "aspect" button?
agd5f: lots of tv expect the input to be a cable box or whatever so it adds overscan so there are no black borders
crow: yangman ok then will try to play with tv set
yangman: set the aspect to "native" or "full" or something like that
crow: k thnx to both, will try to play with tv set then.., i just know ati driver for winz adjust that using slide bar in CCC
yangman: it helps to understand what's going on in the background
yangman: with a digital signal, the PC will always be outputting "true" 1080p, or whatever the resolution is set at
yangman: the TV underscans that, because of historical reasons
yangman: when you compensate using software, the PC effectively scales the picture down, and puts a black border around it so that only some of the entire 1080p picture contains content
yangman: the TV shows only that scaled portion, and everything appears okay
crow: yea that is, i have 1920x1080 resolution but underscaned with black border
yangman: but you're effectively losing clarity because the image is downscaled then upscaled again
yangman: wait, hold on...
yangman: are you actually seeing the black border on the TV?
yangman: with radeonhd?
crow: nop my question was if i update to radeonhd will i be able to turn underscan off
crow: i am using ati driver and have black border
yangman: that's completely different, then :p
yangman: technically "no", but in the sense that radeonhd can't do underscanning at all
yangman: so, there's no option to turn off. it's not underscanned all the time ;)
crow: well it mean then that will give what it have?
crow: egg no underscan/overscan at all
crow: then i need to try it out :P
yangman: it'll just treat the TV like an ordinary HD monitor
yangman: works pretty well, actually
crow: well i am using HDMI cabel, and not VGA, but my tv can accept also over VGA 1920x1080
crow: main debug: adding playlist item `http://mrdvb.ath.cx:8008/vlc.html?channel=S19.2E-133-17-14&async=1' ( http://mrdvb.ath.cx:8008/vlc.html?channel=S19.2E-133-17-14&async=1 )
crow: main debug: creating new input thread
crow: main debug: waiting for thread completion
crow: main debug: thread 4080 (input) created at priority 1 (input/input.c:265)
crow: ops sorry
yangman: iirc, it'll be 1080i over VGA, as it doesn't have the bandwith for 1080p
yangman: unless something's changed recently, interlaced mode isn't supported yet
yangman: I've tried it with DVI->HDMI to a 1080p TV
yangman: works wonderfully
yangman: and this was back before 1.0 release
crow: http://www.avforums.com/forums/showthread.php?t=511377 <-- this is my TV (only 40")
agd5f: crow: ati and radeonhd will output the same signal
agd5f: barring slight variances in the clock
agd5f: neither supports under/overscan
crow: agd5f then its my tv then.. dont know why ... will try to play with it then
yangman: if you're only seeing part of the image, then the TV isn't doing true 1:1
yangman: if you're seeing black borders, then something's messed up because that's technically not possible with the linux drivers
yangman: "I'm not gonna pay you, but, just so you know, I for some strage reason, I often leave money"
yangman: and, that is for the wrong channel :P
libv: dmb: ?
libv: dmb: why blah :)
dmb: just wanted to see if my connection was working :D
dmb: /whoising yourself is a better method
dmb: libv: so whats this thing your going to be announcing :D
dmb: enhancement or whatever
libv: i think rs780 is close to being useful soon, most pressing pll issues have been fixed
dmb: libv: you think that issue with the memory restore is fixed yet in git (i can try)
dmb: i saw a lot of restore commits
libv: not sure whether this was aimed at an issue you were seeing...
libv: but yeah, if things segfaulted in the VGA save/restore, then this probably is fixed now
dmb: that one
libv: ah, this is not fixed yet
libv: not sure what is happening there yet
dmb: do you have a way to reproduce it?
libv: some synchronisation error between the output and the crtc
libv: this specific issue, not really
dmb: if you ever need to debug/test/do whatever just shout
dmb: i think i sent my bios to egbert, don't know if he got it
libv: i still need to find the time to fix another issue i am seeing on r5xx laptop panels: where it doesn't come up properly again after the crtc was assigned to another output first
libv: hrm... has it been a month already since we found out that the m54 mc is like the rv515?
libv: seems like just last week :(
dmb: i haven't been in the world of computing in a while, so it seems like a week for me also (univ does that kind of stuff to you)