^^MAg^^: Hi, what's the current status of radeonhd? any date for new release?
^^MAg^^: configure.ac:18: error: possibly undefined macro: AC_DISABLE_STATIC
^^MAg^^: what am I misssing?
^^MAg^^: installing libtool fixed this
agd5f: R3xx 3D register guide now available
michaellarabel: ooo nice :)
michaellarabel: Is this the re-release of the previous NDA guide from a few years ago or is it mostly rewritten?
agd5f: michaellarabel: there was no NDA guide for 3D
agd5f: so it's all new
michaellarabel: Oops, misread the 3 :)
agd5f: :)
michaellarabel: Is 6xx 3D guide still planned for this month?
agd5f: I suspect it will slip
michaellarabel: Any progress with tcore?
agd5f: good question. I'll check
marcheu: agd5f: there was reg headers and the windows stuff :)
agd5f: marcheu: true, this this has descriptions! ;)
udovdh: hello!
bewst: As long as radeonhd has no 2D or 3D acceleration is there any reason to use it instead of vesa? Does it have any other capabilities vesa doesn't have?
agd5f: bewst: dualhead
bewst: Oh, that's important. vesa doesn't support dualhead?
agd5f: nope
bewst: Well, I tried to get radeonhd working with my FireGL v5200, but shortly after X started the screen would end up with some weird "vertical slices" that echoed nearby bits of screen image every 64 pixels or so.
bewst: Is that a known issue?
Zhenech: i saw such things in early versions
Zhenech: but not with recent (>0.0.4) ones
Zhenech: so probably just try the latest and greatest from git
Zhenech: this is usual the best thing with radeonhd, as it kinda quick in develpment
udovdh: Zhenech, very true!
libv: bewst: were you using the drm driver at the same time?
bewst: Sorry, what is the drm driver?
libv: bewst: did you reboot after running fglrx or radeon?
bewst: definitely rebooted.
libv: bewst: because either might throw some wrench in the wheels
bewst: Had been running fglrx, but rebooted to make sure the module was unloaded.
libv: ok... this striping was something you've seen even with our git driver, right?
bewst: Yes, I built/installed from git.
libv: bewst: ok, so it isn't scaling either
libv: bewst: what hardware were you running this on?
libv: bewst: what resolution?
bewst: Thinkpad T60p FireGL v5200 1600x1200 native laptop panel and 2x 1600x1200 native external LCD monitors
bewst: Naturally I wasn't trying to run all 3 displays at once ;-)
libv: :)
libv: so this machine really has 1600x1200, full 4:3?
libv: cool :)
Zhenech: libv t60p has it, yeah
Zhenech: but kinda fuckin expensive
dmb: wow, amd is on a role with releasing these specs!
bewst: libv: picking up from where I left off -- yes, T60p had the 1600x1200 native if you ordered it that way. I don't think you can buy it 4:3 anymore though :(
dmb: pats his 1920x1200 laptop display
dmb: i think its 4:3, don't know, does 4:3 mean widescreen?
bewst: dmb: no, 4:3 is the opposite
bewst: some of us need the height more than the width
dmb: oh
dmb: anyway, people think i'm crazing for using 1920x1200 on a 17 inch laptop monitor
dmb: whenever this one friend comes over to read something on my screen, i have to increase the font size
bewst: I've been that guy. Furthermore, a 15" 1600x1200 (what I have now) has the same size pixels.
bewst: It just doesn't break your back to carry it across town
yangman: I have a 14" T60 at 1400x1050
Death_Syn: i wish I had 1920x1200 on my notebook
yangman: fits my usage a lot better than my old, red R51
yangman: although I miss the screen on that thing :(
yangman: s/red/dead/
yangman: back in the days when IBM still shipped non-TN screens in their high-end laptops
dmb: yeh, 1920x1200 is awsom
dmb: i think i began taking it for granted though
dmb: i have a hard time using any resolution smaller
Death_Syn: i can barely manage on 1680x1050
dmb: i goto my friend which prefers 1024x768 and i just can't use it
dmb: i wonder if 1920x1200 is bad for your vision
dmb: on a 17inch lcd monitor
Tigerchen: i loved 1600x1200 on my notebook
Tigerchen: and that was 15"
Zhenech: dmb, 1024x768 is okay on 12"
Zhenech: :)
Zhenech: eh, he's gone
yangman: eh, the resolution doesn't matter if what you read is big enough
yangman: of course, having a higher resolution makes reading CJK *a lot* easier
Zhenech: never need that :P
yangman: it's difficult to represent 14+ stroke characters in a 12-pixel high block ;)
Zhenech: uhm, the most complicated asian character I can write is a no, so don't think this will be needed here soon
Zhenech: but I can see your point
libv: i use virtual everywhere... the wms kind of work against me all the time, but it makes up somewhat for smaller screen resolutions
metalth: is there a way to set dpi resolution size in xorg.conf? DisplaySize option in monitor section seems to be 'ignored'...
El_Angelo: there is something on the faq about that metalth
El_Angelo: i think so at least
Tigerchen: libv: you know a time when the brightness-key-blocking will be removed?
libv: Tigerchen: shouldn't take that much work, but the mileage will vary from laptop to laptop :(
the-me: libv, maybe I could help out with informations, I've got this issue on an HP compaq notebook
Tigerchen: libv: you told me once you do this locking on purpose, so removing the lock shouldn't be the problem
Tigerchen: and strangely the current version I'm running also doesn't allow me to switch to console to change the brightnes there, it stayt the same anyway
Tigerchen: and 1 more thing: did you know if egbert did come any further with my bug i files yesterday?
libv: the-me: we have two different hp laptops in offices :)
libv: the-me: this is where it will work on :)
libv: Tigerchen: he's rather worried about this issue...
libv: Tigerchen: but he leaves for italy for a week tomorrow morning
the-me: libv, oh that's realy nice ;)
libv: i will dig into this more deeply, but... i am smelling the prey on the restoration issue already, so i am going to hunt that down further first
libv: Tigerchen: you didn't have this issue at all with our earlier driver right...
Tigerchen: libv: the segfault? nope!
libv: you had a list of sha-ids iirc
libv: let me see whether i still have that in my browser
Tigerchen: libv: i think i stioll have the broken ids
libv: Tigerchen: keep them around, in case we lose them, otherwise a lot of work might be lost
Tigerchen: but as i said, one of them introduced the bug, and all of them segfaulted at startup
libv: right
libv: there was an issue with the code for a bit there... we debated whether we should've reordered one fixup commit before the push upstream
Tigerchen: no 2 of them started but produced a black screen
libv: we probably should've
Tigerchen: ^^
libv: i've added the ids to the bug now
libv: it is highly related to what yaneti was seeing so this should actually get the highest priority
libv: but i cannot seem to reproduce it here
libv: i probably need a brandnew xserver for that
Tigerchen: need any versions?
libv: Tigerchen: i think you used a pretty young 1.4.something
Tigerchen: 1.4.0.90-r3
libv: right
Tigerchen: gentoo testing ^^
libv: so i should go with xorg head
libv: :)
Tigerchen: that's really bloody ^^
Tigerchen: i think i could also build a binary-pkg for you if you'd like
richiefrich: udovdh: you here?
richiefrich: udovdh: nvm i got it :)
richiefrich: is now on dual DVI-D's big desktop view
Zhenech: hm
Zhenech: wanna sponsor me some screens too?
richiefrich: but then i would be back to one :\
Zhenech: and I would need some dvi capable graphics
Zhenech: Radeon Mobility M6 LY isn't really new
Tigerchen: i hate the thought that with the next notebook i will again be limited to a closed source driver :(
Zhenech: Tigerchen, why?
Tigerchen: Zhenech: because silly lenovo switched to nvidia
Zhenech: heh
Zhenech: has his thinkpad with ati :P
Tigerchen: dito
Zhenech: but they still offer intel, do they?
Tigerchen: sure, but not with i nice display
Zhenech: hm, there was somethin like that on linux-thinkpad
Tigerchen: i read it
oninoshiko: doesnt mind closed drivers as long as they work
oninoshiko: waits to be booted
Tigerchen: but they don't
Zhenech: oninoshiko, the problem is the drivers usually work as in "display a picture" but fail on new kernels or xorg which is a pitty
oninoshiko: if on my kernel/xorg copy i have accelleration im happy
siimo: hi anyone know if i buy a Radeon HD 3450 256MB, will i be able to run it in linux
dmb: i'm confused as to what the big difference between radeon and radeonhd now and in the future
dmb: they both seem to have r5xx and r6xx support
siimo: the new cards are called radeon HD thats why? :/
dmb: i mean the drivers
Zhenech: dmb: radeon is an "old-style" driver
Zhenech: written and extended for a long time now
dmb: oh
Zhenech: radeonhd is new and fresh
dmb: so radeonhd is expected to replace radeon then for r5xx and r6xx cards?
michaellarabel: dmb: The big difference at this point comes down to their use of AtomBIOS vs. banging the registers directly and then the Radeon code-base just being extended.
Zhenech: probably better designed and more focused on the new chips and their capabilities
dmb: oh
dmb: i'm guess radeonhd will be the first to get 3d support?
dmb: guessing*
ajax: except for rv620, apparently, since they both use atombios for that.
Zhenech: dmb, the first (i guess) will be radeon
egbert: ajax: only for for the uniphy transmitter.
dmb: does the radeon have drm/dri support yet for r5xx?
ajax: dmb: drm, yes. r500 2d accel in radeon uses the drm.
ajax: dmb: dri, not yet.
dmb: what exactly does drm improve?
Zhenech: ajax, does it need kernel support for drm?
ajax: Zhenech: drm _is_ kernel support. so, yes.
dmb: does that make it use the video cards memory instead of the main memory?
Zhenech: uh, THATs why radeon is so slow on my x1400
Zhenech: :)
airlied: radeonhd need to leverage a lot more code from radeon to get to having DRM/CP aupport
ajax: dmb: drm for 2d accel mostly means that you use the command processor to build blocks of commands rather than banging them directly. so you get better decoupling of CPU from GPU, and thus better performance.
airlied: support even
airlied: then the 3D driver is going to be in mesa in any case.
Tigerchen: but what do i need to rebuild with radeon to use the newest version? the radeon-module? mesa? there are 3 different spots to type "make"
Tigerchen: and that sucks
ajax: egbert: interesting. can i ask what motivated that?
dmb: someone told me that radeonhd would work with drm right now, is that true?
ajax: radeonhd does not interface with the drm right now, correct.
dmb: oh, so it doesn't
airlied: Tigerchen: no open source driver will ship a 3D driver outside of Mesa.
airlied: Tigerchen: you can't really..
Tigerchen: airlied: maybe, but a simple 1. 2. 3. to update the driver would be nice
ajax: 1) build libdrm
ajax: 2) build radeon 2d driver
ajax: 3) build mesa
airlied: Tigerchen: when we get a 3D driver for r500 I'm sure someone will do one :)
Tigerchen: ah, thanks, that would be nice
airlied: but really everyone should just run Fedora :)
libv: airlied: sure.
Tigerchen: airlied: why's that? I'm quite happy with the configuration-options gentoo provides for me
airlied: Tigerchen: because we ship the cutting edge stuff as its written :)
airlied: saves people typing make :-P
ajax: not that we're biased or anything
airlied: ajax: never :)
airlied: Tigerchen: but gentoo should have ebuilds for all the drivers stuff.
airlied: Tigerchen: I'm nearly sure you can build git server/mesa/libdrm etc.
airlied: Tigerchen: with various ebuilds/overlays
Tigerchen: airlied: I'm sure of that
Tigerchen: but then still I'd have to know the order in which to build them
ajax: points above...
Tigerchen: i know
libv: Tigerchen: i keep on getting distracted away from code now... all sorts of stuff to deal with here.
Tigerchen: perhaps i should really search for a overlay containing git-ebuilds
Tigerchen: libv: no problem
libv: Tigerchen: i should probably first do a thorough review of the code in those commits and see if i can spot something
libv: Tigerchen: egbert wrote the code, and as always, a fresh pair of eyes might spot it
libv: will take some time, but it will probably be worth it
libv: that is, if it really is an issue with our driver on its own
Tigerchen: yeah i know that, I hate reading somebody elses code
libv: because it kind of seems like it is a mix of both the recent X and our driver.
Tigerchen: maybe. but i think the only thing i did was updating the driver
Tigerchen: and i don't know what exactly trigegrs it
libv: sure, but we are not immediately seeing it here with slightly older and stable Xen
libv: so it might not be easy to spot outright too
Tigerchen: might be
egbert: ajax: i didn
egbert: t get unuphy specs in time...
libv: egbert: if Tigerchens bisect range is really where this happens, then my gut says that it is 56a264f8
libv: Tigerchen: 6742f79e8 cannot be ok for you
libv: Tigerchen: you will need to use the NoRandR option for that one
libv: heh, my assumption higher up would mean that NoRandR would fix this magically too ...
libv: this being the actual bug
Tigerchen: libv: you mean with NoRandr i should not get a segfault when starting urxvt?
libv: Tigerchen: did you test this already?
Tigerchen: libv: no but i might just now
libv: i know yaneti didn't have a change with NoRandR
libv: still had a horrible crash
Tigerchen: I'll try
libv: ok :)
libv: try head first, then try the commit i listed
Tigerchen: will do
libv: because the commit i listed is where i broke the RandR path :)
Tigerchen: i hope i won't get the black-screen issue with that
libv: but the fix for this is 42d3201b
Tigerchen: i might bisect again with NoRandr in my conf if that might fix something
Tigerchen: we
Tigerchen: 'll see
libv: if HEAD still is broken with that, then it might give you a slightly larger range
Tigerchen: Option "NoRandr" "true"
libv: this is the trouble with a huge codedrop though :)
Tigerchen: right so
Tigerchen: ?
libv: yup
Tigerchen: ok, brb
Tigerchen: ok tested current head with norandr:
Tigerchen: 1. resolution sucks ^^
libv: :p
libv: but it is solid?
Tigerchen: 2. segafaults when starting urxvt with -depth 32
libv: prff
Tigerchen: should i bisect again?
Tigerchen: or don't you think that it will bring any results?
libv: this is an m56... so 52f04a7f8f will still cause trouble
Tigerchen: just tell me what to test and I'll do
libv: Tigerchen: does this also happen when you set option "Accelmethod" "false" ?
libv: err
libv: s/false/none/
Tigerchen: I'll try
Tigerchen: yes it does
libv: hrm...
Tigerchen: i have a logverbose 7 log will that perhaps help?
Tigerchen: libv: this is a M54, not M56, just to clarify
libv: Tigerchen: sure, but this one doesn't seem that hw specific :)
Tigerchen: k, just noticed the difference
Tigerchen: should i perhaps try 56a264f8 as you said before?
libv: 56a264f8 is where i fear the trouble is
libv: but...
Tigerchen: then perhaps i should test one before that?
libv: does your laptop list TMDSA as an output in its connectortable?
Tigerchen: how do i find out?
libv: ah, just a minute, i can look at your log too :p
Soul_keeper: this is almost as exciting as dri was on my radeon 7500 a few years ago :)
libv: ah, urgh...
Tigerchen: Soul_keeper: ^^
libv: you have this docking station dvi :)
Soul_keeper: i get my 3870 on tuesday
Tigerchen: libv: i don't have a docking station, but i theory yes
libv: (II) RADEONHD(0): Connector[2] {RHD_CONNECTOR_DVI, "DVI-I DFP1", RHD_DDC_1, RHD_HPD_0, { RHD_OUTPUT_TMDSA, RHD_OUTPUT_NONE } }
Tigerchen: in theory
libv: so it will try to do load detection and segfault
Tigerchen: Soul_keeper: like my firegl t2 about 3 years ago
dmb: i seriously think the radeonhd driver in git is a lot more stable then the fglrx driver
Tigerchen: dmb: welcome to the club ^^
dmb: just can't wait for 3d/dri :D
dmb: 3d with fgrlx is a disaster for me
Tigerchen: I'd like to play ppracer again
dmb: no matter what game i try to use, at some point in the game it hard locks the computer
dmb: i can't do anything but hard reboot
dmb: i wish there was a way to kill x when that happens
Tigerchen: try to ssh to your box
dmb: not easy if you don't have another computer nearby
dmb: someone told me something may be possible with a SysRq combo
Tigerchen: has no experience with sysrq
dmb: Tigerchen: don't know if you know, but why do things like the keyboard stop functioning when the display locks?
dmb: i should be able to go into a VT and type killall xorg without seeing the monitor
dmb: but keyboard stops functioning
Tigerchen: i think the game tries to grab the keyboard-input, and the signals won't reach X
Tigerchen: or perhaps the maching dies a thermal-shutdown
Tigerchen: machine
dmb: well, the cpu is still working
Tigerchen: have you tried pinging/sshing the box?
dmb: i'v done that is the past, i can still ssh into it
libv: Tigerchen: ok, nothing for it... i will have to torture myself with egberts buildtool :p
dmb: for example,i was playing doom3 today, and all the sudden the display crashes, the music still goes on
libv: Tigerchen: as if looking at his code was not bad enough :ppp
dmb: just the keyboard and monitor are completely locked
Tigerchen: libv: mein beileid
libv: (i will so get kicked up the rear when egbert wakes up :p)
libv: Tigerchen: :)
Tigerchen: libv: because you said something about his coding-style?
libv: uh-oh...
libv: he's still awake :p
libv: egbert: i was just joking!
Tigerchen: somehow uh-oh makes me think of icq
libv: Tigerchen: makes me think of the teletubbies to be honest
dmb: makes me think of an old ftp client for windows
dmb: that made that sound with an error
TinkyWinky: libv: still awake?
libv: Tigerchen: urgh
libv: i had overlooked it sorry...
libv: Tigerchen: that was rather bad
Tigerchen: ?
libv: Tigerchen: i now feel less inclined to fix this bug of yours :ppp
Tigerchen: :confused
libv: there should be #radeonhd channel rules...
Tigerchen: doesn't get what he did wrong
libv: if claim to be a ridculous purple kids figure, and you're wearing a handbag, your bug will not get fixed :p
libv: if you claim even
Tigerchen: oh, /me will never do so again, but i wonder why you know so much about these *beeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeep* "creatures"
libv: well...
libv: erm...
Tigerchen: gotcha
libv: damn :p
Tigerchen: :þ
libv: oh well, i should've known that those awful skeletons wouldn't have stayed in my closet for too long :p
libv: still...
libv: it could've been worse
libv: it could've been bananaman
Tigerchen: wtf is bananaman?
libv: youtube is your friend...
libv: or nemesis in this case
Tigerchen: i just startet google picture search
Tigerchen: and i don't like what i see
Tigerchen: libv: any idea
Tigerchen: ?
libv: suse != debian... no debootstrap :)
Tigerchen: hmm, the first think i could've imagined
libv: Tigerchen: i'm installing a base system now, then starting a build, to see if i can reproduce it with upstream
Tigerchen: but i never used any of both
Tigerchen: ah, k
libv: the trouble it, for a proper bisect, we have two segfaults throwing a wrench in the wheels, so it is probably better to reproduce it first and see from there
Tigerchen: sure
Tigerchen: 3 problems at once suck, because 1 stops one from finding the other
libv: yup
libv: we should've put quick fixes right after the issues before we pushed this upstream
libv: other option is, i hand you a special branch with the fixes moved..
libv: while this thing is installing and building
Tigerchen: you could do that, I'm willing to try
dmb: if you want, i can test it with debian testing
dmb: and see if i can reproduce it
dmb: Tigerchen: what version of xorg are you using?
Tigerchen: 1.4.0.90-r3
dmb: Tigerchen: thats xorg server?
Tigerchen: yes
Tigerchen: libv: would ssh-access to my box help you?
dmb: i'm confused, debian says 1:7.2-5 is the latest
dmb: http://packages.debian.org/testing/x11/xserver-xorg
Tigerchen: x11-base/xorg-x11
sytse: mm, ssh access, huh
Tigerchen: 7.3
sytse: ^_^
sytse: ah, gentoo
dmb: http://packages.debian.org/testing/x11/xserver-xorg-core
dmb: ah
dmb: testing only has 1.3
dmb: lets see what sid has
dmb: when did 1.4 get released?
Tigerchen: no idea
Tigerchen: ages ago
dmb: Tigerchen: your not using ~x86?
Tigerchen: i am
sytse: $ git log xorg-server-1.4.0.90 |grep Date |head -n1
sytse: Date: Wed Dec 12 21:57:59 2007 +0200
sytse: $ git log xorg-server-1.4 |grep Date |head -n1
sytse: Date: Thu Sep 6 01:42:43 2007 -0700
dmb: mmm
sytse: so, 'ages' for a gentoo user indeed ;-)
dmb: not ages for debian based
sytse: heh
sytse: indeed
dmb: looks for a sid cd
dmb: oh wait, i am such a retard
dmb: i was going to try to test it on virtualbox :P
dmb: i wonder what happens if i run xorg in a chroot
Tigerchen: something would jump out of your box and hit you on the head? ^^
dmb: i was going to test it, but now i don't think i can
dmb: did any of you ever notice that with fglrx, when using gdm, the dots when entering in the password are misaligned?
dmb: but every other driver they're fine
dmb: ?
Tigerchen: nice effect
dmb: Tigerchen: you have noticed it?
Tigerchen: dmb: i haven't used fglrx on this notebook at all, and never in my life gdm
Tigerchen: so... no
dmb: oh
Tigerchen: wonders if libv has died
dmb: well if he did, he had a nice life as a graphics driver developer :P
marcheu: don't worry he probably came back home :)
dmb: Tigerchen: just remember, he died while helping you!
marcheu: he might reincarnate as an r700
Tigerchen: marcheu: LOL
Tigerchen: dmb: i don't know if i could stand that thought
libv: hrm... did i just die
Tigerchen: UAHHH, ITS ALIVE *runs around in circles*
libv: Tigerchen: you wondered if i died
libv: and then the network at my appartment dropped off :)
Tigerchen: ah
marcheu: libv: so you had to go back to work ? :)
libv: marcheu: yeah, sure, in 2 minutes i made it back to work :p
Tigerchen: libv: perhaps you live just next door
libv: nah, just building X here, requires slightly too much work to create the git branch from which Tigerchen can bisect properly
Tigerchen: ah, k, again the offer: would ssh to my box help?
dmb: libv: thats nice of you to do work even though your not at work :P
libv: Tigerchen: hrm... yeah, but i am not sure that i am going to continue this for much longer today :)
Tigerchen: libv: no problem, i just set up a port-fw if you want to try
libv: ok, i'll poke you for the details when i'm back :)
Tigerchen: k
libv: probably not this weekend any more though, i'm going to spend most of the weekend on my bicycle
marcheu: don't push it too hard, it's sports :)
libv: :)
Tigerchen: cool. I'm just sitting around, waiting for my girlfriend to come back from work
Tigerchen: that's why I'm on and annoying you
libv: Tigerchen: oh, you're not annoying at all, it's good that you're this persistent about these things and that you're providing real solid datapoints on these issues
Tigerchen: thanks. I'd like to commit something like code myself, since i like coding, but haven't had any experiance with hardware-programming or C at all
Tigerchen: last thing i tried with C was a pkcs11-login-plugin which i wondered that it worked for ANYONE AT ALL, because there were some points where it just spat out false instead of testing or doing something
libv: digging out bugs bugs is not about spewing out code :) it's fixing the code that was spewed out before :)
Tigerchen: libv: i know, but one has to understand the code someone else wrote, aber i dont think i would get through the pointers, registers and stuff
Tigerchen: (haven't been able to remember which one does what of * and &
dmb: reading someone elses code is a challenge
libv: Tigerchen: our driver is rather nicely structured, should make a lot of sense, but yeah, it takes time to get into any project, even when you're proficient in C
dmb: one of the issues is that the schools an colleges only seem to teach c++ and java now
dmb: no more low level c
Tigerchen: one of the funniest things i think is a void** which a normal thinking human would translate into a pointer to a pointer to a nothing
dmb: mmm
Tigerchen: dmb: only java, scheme and misp-asm here
dmb: x86 asm and c++ for me
Tigerchen: and school? BASIC is standard at school
dmb: although i have to force myself to get into c
dmb: Tigerchen: i guess more of colleges then
Tigerchen: TUD atm, some 3 or 4 terms to go
libv: tud? delft?
libv: ah, no darmstadt :)
Tigerchen: technical university darmstadt
libv: Tigerchen: egbert lives in the same town btw :)
Tigerchen: really? i live in eschborn near frankfurt
Tigerchen: at least public transport is nearly free for students
libv: suse had some office floor in eschborn for a while, and egbert had his office there
Tigerchen: trying my BSc next term, hope everything works out as i like it to
Tigerchen: noting you two on my list of "contacts" for a possible job round here ^^
werdz: hooray DCU \o/, x86 asm, mips asm, c, c++, java, even some haskell and prolog... by the end of second year..
libv: yikes... almost 600mb got pulled in from git
Tigerchen: oh right, i forgot prolog and datalog
Tigerchen: libv: nice
dmb: werdz: sounds like our opl class
dmb: which i'm not taking to next year
werdz: heh
dmb: organization of programming languages
dmb: java, prolog, c++, lisp and some other weird languages we are forced to learn
dmb: well, weird as in lisp and prolog
Tigerchen: lisp (or scheme) is kind of funny, especially if you try to understand the environment-model
Tigerchen: i find prolog not THAT weird
Tigerchen: interestingly prolog is needed for abitur in computer sciences in hessen
werdz: prolog just reminds me of our crazy german logic lecturer
Tigerchen: and the best language of all: P
Tigerchen: a purely theoretical language operationg on all natural numbers
Tigerchen: something we needed to use for our lecture on computability
the-me: Tigerchen, if I understood libv rightly in some sentences, this backplane gamma thing is working on some laptops but they completly locked it, because it makes trouble on some other nbs?
Tigerchen: the-me: i think so
the-me: hmmm
the-me: Would be interesting which patch includes this lock :)
Tigerchen: the weird thing is that it worked here when i switched to a VT but since i had to reinstall my nb this workaround doesn't work anymore
the-me: I need my notebook in business, so on this isn't a solution for me, that's weird
Tigerchen: well, switching so console to make the change and then switching back to x doesn't take that much time
libv: the-me: no, we stop it everywhere
libv: the-me: at least we have consistent behaviour :)
the-me: hmm where could I see the current roadmap?
libv: the-me: we can set some flags in some scratch registers, and then the bioses will not be interfering too much anymore... but this depends on the bioses
Tigerchen: libv: would it be a possibility to control this behaviour via a parameter for the autogen script? like "please don't report any bugs if you --enable-brightness , disable it first and test again"
the-me: Tigerchen, if they for example implement it in the next week this would be far useless
the-me: but if it would take something like a year, then it would be nice ;)
Tigerchen: the-me: well, 3D should come in some months brightness i think would be before that
the-me: and how is it with this new commits and that RV690 (aka x1250) is now supported? It worked here for me before it, too :)
libv: Tigerchen: we could have some xorg.conf options for this i guess
libv: the-me: rs690 needs the newly discovered ddia block still
libv: the-me: egbert is close to having this coded up, but he will not be doing work the next week
dmb: i sometimes have an issue where if i switch to a VT, the screen is all jarbled
dmb: like the look of a misaligned vhold in the old crts
dmb: i guess thats a bug
libv: dmb: definitely
the-me: libv, what are the pros if it supports this? How I said, the driver worked here before this commits, too
the-me: (dont know what ddia does)
libv: dmb: since when are you seeing this?
dmb: libv: how do i debug that (i'll probably wait for a business day)
dmb: libv: well, today
libv: the-me: this is for HDMI/DVI out right?
dmb: libv: i'm not sure, but it may of happened in earlier versions
libv: the-me: plus... were you seeing issues before, or only in the last few days?
the-me: libv, ah okay, nothing for me ;) I just need my nb tft :)
libv: dmb: ok... tigerchen has some really nasty issue now which i'm trying to reproduce, so this kind of makes things harder for us in the coming days
Tigerchen: libv: well i just remember that the flashing colors are also still there sometimes (every 2nd time?) i switch to console
libv: Tigerchen: flashing during the switch?
dmb: Tigerchen: maybe your having the same issue as me
libv: or after the switch?
dmb: its like lines going up and distortion
dmb: after the switch to console
Tigerchen: libv: nope, flashing of colored (especially red) letters
Tigerchen: maybe i could make a vid
libv: in the VT... after a switch back...
libv: hrm...
libv: palette not restored correctly?
libv: or, no, wrong...
libv: font not restored correctly
libv: Tigerchen: this is with VGA textmode right, no vesafb involved?
dmb: hmm
Tigerchen: narf, not reproduceable atm
Tigerchen: libv: yes
libv: right, font problem indeed
libv: Tigerchen: were there complaints in the log about not being able to access the VGA FB?
dmb: i can't seem to reproduce my issue now either
Tigerchen: libv: wait I'll look
libv: Tigerchen: it is some issue with storing/restoring the bit of memory the VGA emulation engine uses
dmb: the vga console is emulated?
dmb: i thought it was native to the card
libv: dmb: there's a whole engine to emulate vga
dmb: interesting
libv: dmb: it's quite smart and nice and even handles most of a modeset for you
libv: i haven't tried using it directly yet, but it looks rather nice
dmb-vga: ok, i am reproducing the issue right now
dmb-vga: its like the screen isn't stable
libv: our driver tries to save as much as it can, disable it, and then restore and reenable it, but apart from that, we do not want to touch it and keep it as is
libv: ok....
dmb-vga: text from the bottom of the screen flashes towards the top and vice verca
dmb-vga: isn there anything i should do to debug?
Tigerchen: lib I'm uploading an vid
libv: dmb: what do you mean, flashes towards the top?
libv: dmb: the text flashes on and off?
libv: dmb: this usually points at a font with this flashing bit set
dmb-vga: its kind of hard to explain, but everything is all flashy and unstable, i'll try to make a video with my cell phone
libv: dmb: which means bad memory restore
dmb-vga: could be
libv: Tigerchen: can you open a bug for this as well?
libv: Tigerchen: and add the log when it happens, dmb, could you add a log as well?
Tigerchen: lib ok
Tigerchen: but i saw nothing in the log except one of the MC not idle ( but i dont know if thats related
dmb: libv: what do you mean log?
dmb: xorg log?
Tigerchen: libv: http://rapidshare.com/files/99622737/mvi_4096.avi.html
libv: mc not idle... ok... even on an M5x...
dmb: yours is nowhere near as crazy as mine
Tigerchen: bug opened https://bugs.freedesktop.org/show_bug.cgi?id=15020
Tigerchen: including xorg.log
libv: Tigerchen: weird... doesn't immediately look like vga memory restore issue :(
Tigerchen: libv: noticed this for a long time, but didn't bother since i normally doesn't use console
Tigerchen: the end of the vid is a bug of my cam, don't wonder
Tigerchen: sadly my good old digital ixus v3 slowly reaches it's end of life
Tigerchen: :(
libv: so red flashes to green and back all the time...
dmb: libv: http://output.da4.org/vidbug1.mpg and http://output.da4.org/vidbug2.mpg
dmb: poor resolution but i think it conveys what its doing
dmb: (i only have a cell phone vid cam)
Tigerchen: libv: i think there might also be pink to blue
Tigerchen: and vice versa
libv: and this happening all the time...
Tigerchen: no not everytime
Tigerchen: i think about 50%
libv: all the colours move at pretty much the same time...
libv: Tigerchen: on which outputs?
libv: panel only?
dmb: libv: did you see my video (you might have an issue playing it)
Tigerchen: i can only test panel
libv: Tigerchen: right
libv: so we do not properly restore some panel registers
libv: dmb: this looks like alaptop as well
libv: dmb: so same issue indeed
dmb: it is
dmb: mobility x1400
dmb: libv: it may be related to the same theme
libv: right, it is the same bug, please attach yourself to it
dmb: argh, did i miss something, i accidently killed x
libv: ok guys... hope to at least fix some of these issues soon, but not tonight... X is building nicely, i need to get up in 7h, and i have a 20minute walk home and then i will have to find out what killed my internet
dmb:
dmb: libv: have fun
libv: libv> right, it is the same bug, please attach yourself to it
libv: https://bugs.freedesktop.org/show_bug.cgi?id=15020
dmb: ok
libv: thanks for your help :)
dmb: np
dmb: Tigerchen: this happens randomly for you right?
Tigerchen: thank YOU for providing such a nice piece of work
Tigerchen: dmb: rigth
dmb: Tigerchen: and switching back to x then back to the console with sometimes fix it?
Tigerchen: yes
dmb: yup, probably same as mine then
dmb: Tigerchen: what vid card is this?
Tigerchen: i guess it happens in about 50% of the switches
Tigerchen: x1400
Tigerchen: should add this to the bug
dmb: Tigerchen: well, i have the same video card
dmb: so 90% its the same issue
Tigerchen: done
Tigerchen: yoeah think so
partymola: hello everyone
partymola: I have a "ATI Technologies Inc RS482 [Radeon Xpress 200]" graphics card
partymola: if i am not wrong, it isn't supported by radeonhd
partymola: am I wrong? Is there any other decent driver i can use for it?
marcheu: you are right, but double negations are difficult to parse
partymola: sorry, i am a spaniard... ;)
partymola: thing is
partymola: is my card supported?
Tigerchen: perhaps radeon might work son't know
partymola: and, if not, is there any alternative?
partymola: ok
marcheu: yeah radeon should work for you
dmb: ati probably
dmb: yeh
dmb: i got a mach64 card working earlier, and thats older then that :D
partymola: hhmmm... weird
marcheu: dmb: that card is r300 grade..
partymola: it says (EE) RADEON(0): No valid modes
Tigerchen: partymola: try latest git-radeon
dmb: marcheu: i don't know about that... its from a really old mac laptop
dmb: marcheu: and has less then 15 megs of video memory i think
Tigerchen: partymola: perhaps someone in #radeon might know
partymola: ok, thanks
marcheu: dmb: it's r300 without a vertex processor
dmb: marcheu: even though its not a radeon?
marcheu: dmb: hmm ? r300 is a kind of radeon...
dmb: marcheu: i mean the mach64 card i have, i don't think its a radeon
marcheu: ahhh the mach64
marcheu: yeah the mach64 is surely not a radeon :)
Tigerchen: hmm, mach64 has it's own video_cards flag on gentoo
marcheu: Tigerchen: it's a separate driver...
Tigerchen: hmm, seems to only pull xf86-video-ati
marcheu: (mach64, r128, radeon) are all wrapped within a big "ati" driver
marcheu: you can put either "ati" or "mach64" in your config file (provided you have a mach64) and you get the same result
marcheu: well technically it's a bit more complex than that, but I'm tired
dmb: mach64 card doesn't work very well with compiz or aiglx i learned :P
partymola: #radeon guys are not replying... :(
marcheu: partymola: what is it we didn't answer here ?
Tigerchen: partymola: http://dri.freedesktop.org/wiki/Building
dmb: Tigerchen: ugh oh, i might of realized something
dmb: Tigerchen: i never unloaded the fglrx kernel module
Tigerchen: dmb: i haven't even installed it
dmb: ok, so i guess that could be taken out of the picture
partymola: Tigerchen: building all that will screw up my Ubuntu installation... i think :/
Tigerchen: partymola: i don
marcheu: it will also fail to solve an issue like (EE) RADEON(0): No valid modes
Tigerchen: marcheu: it could well
marcheu: Tigerchen: great, please go ahead and explain me how
Tigerchen: marcheu: i think if no valid modes are found that might be e.g. because of incorrect parsing, a bug in xrandr(, the phase of the moon), which could be fixed with a current version
marcheu: Tigerchen: sure, you realise this page is about building the 3D driver, and his error is from the 2D driver, right ?
Tigerchen: marcheu: i built my radeon for testing like said on this page
Tigerchen: and it worked
Tigerchen: it handles the whole bunch: libdrm, mesa, radeon
marcheu: and how does replacing the 3D driver fix the 2D driver ?
Tigerchen: marcheu: this should also replace the 2d part i think, at least it has in my case, because i don't think that 6.7.197 already supports r500
marcheu: please read the page correctly. it does not build the 2D driver.
Tigerchen: marcheu: ok sry. then I'm totally lost wrt radeon
dmb: hmm, what exactly is a hardware cursor in x?
marcheu: dmb: on that's drawn by the card as an overlay, i.e. doesn't modify the underlying memory as it moves
dmb: hmm, interesting
dmb: marcheu: the cursor includes the mouse right?
marcheu: the cursor is the picture
marcheu: the little arrow or whatever
dmb: yeh, thats what i meant
dmb: the thing that moves :D
dmb: Tigerchen: have you ever had any non-software related issues from your x1400?
Tigerchen: dmb: don't think so
dmb: oh
Tigerchen: what do you mean exactly?
dmb: Tigerchen: are you able to go into suspend or hibernation?
dmb: Tigerchen: don't know, my card seems to act a little strange, not sure if it was because of fglrx driver, or the card
Tigerchen: dmb: yes s2ram or echo mem > /sys/power/state works perfect, tuxonice-suspend2ram used to work, but fails atm
Tigerchen: mann ich hab das gefühl im tv läuft überall nur pseudo-sex-scheisse
dmb: can't read german :D
Soul_keeper: nein, nicht kanst
dmb: Tigerchen: i can't go into it at all
dmb: s2ram at least
dmb: the display stays off
dmb: even if i'm not running X
Soul_keeper: tv ist kerplunk ...
Tigerchen: dmb: try just echo for a change (and there might help the kernel param acpi_sleep=s3_bios
dmb: i'll try that in a little
dmb: probably not going to work, but might as well try
dmb: found another intersting bug i think
dmb: sometimes when i type, i see weird artifacts on the side
Tigerchen: thats interesting, i also found another one
Tigerchen: after running with radeon for a test radeonhd is not able to rednder the display correctly
Tigerchen: mousecursor is a rectangular pinkysh whatever
Tigerchen: test is all stripes
Tigerchen: text
dmb: Tigerchen: yeh, you have to reboot after using radeon i think
Tigerchen: yes
Tigerchen: sucks
Tigerchen: radeon-bug or radeonhd-bug?
dmb: Tigerchen: i'm guessing radeon bug, because it forgets to do something i think, fglrx does the same where you have to reboot after using it
Tigerchen: hmm, perhaps radeonhd can do something about it, when they say its not their problem I'll just dismiss it ^^