simplexe: airlied: ok, i look now branch drm-next and didn't find any information about PM
airlied: simplexe: the patches are mostly on dri-devel
airlied: under discussion
airlied: also thread "Power management in KMS"
simplexe: airlied: in mailing list?
airlied: yes
simplexe: airlied: thx
bridgman: nanonyme; the 1D and 2D names don't mean what you think
bridgman: 1D is single level tiling, 2D is "tiles within tiles"
bridgman: 1D is a lot easier to figure out when writing the entile/detile code ;)
bridgman: 1D is sort of like macro linear / micro tile on 3xx-5xx, 2D is sort of like macro tile / micro tile
bridgman: sth like that
MostAwesomeDude: Could be worse. nVidia's got something called the "swizzle."
MostAwesomeDude: It's a space-filling fractal curve.
MostAwesomeDude: Of tiles.
MostAwesomeDude: Cool as hell because pixel distance on the plane is related to pixel distance in memory, but it's really weird.
kcodyjr: why isn't that a surprise, coming from the company that did quadratic primitives at the very dawn of desktop 3d hardware
kcodyjr: if only they weren't such a bunch of horses arses about specs.
bridgman: yeah, engineers always find a way to sneak in their favorite idea
kcodyjr: it's kind of a good one. drawing curves is a bastard of a thing to ask in a universe defined by squares.
bridgman: sounds like that would have been fun to RE
kcodyjr: yeah it would have. wonder if you can even still get them on ebay, i don't remember for sure but i think those chips died around the same time vlb did
MostAwesomeDude: bridgman: I have *no* idea how ymanton did it, but he's the one.
airlied: should get someone to do the r100
kcodyjr: i thought the r100 was a thoroughly shot up target by now
airlied: well the tiling on it isn't figured out
kcodyjr: maybe this winter i'll try digging mine up. but i suspect all the tiling an r100 will ever do, will require nails and a wall.
bridgman: ok, didn't know that; is there any specific part that's not known or just "everything" ?
airlied: bridgman: macro tiling format
airlied: also r200
airlied: I think Alex has looked but hadn't much lcuk
kcodyjr: not to slight efforts. but i can't imagine too many of them are still experiencing power flow.
airlied: I have an r100 here
bridgman: my main desktop machine at work has an r100
airlied: lots of T42 laptops still exist, people liked them
bridgman: that's two ;)
bridgman: actually my laptop has an rv280/m9
kcodyjr: you guys are developers, it's kinda normal you'd have a collection of cards available. but how many r100's in the wild are likely to get a driver update before they get recycled?
MostAwesomeDude: bridgman: http://en.wikipedia.org/wiki/Z-order_%28curve%29 is the wiki page on it; http://nouveau.freedesktop.org/wiki/Surface_Layouts if you're interested in ymanton's take on it.
kcodyjr: t42? dammit, why did i have to inherit a t43 with i915...
airlied: all the ones I get emails from when I break shit ;-)
airlied: also embedded markets still use r100s
kcodyjr: hahaha, ok, guess you would know. really? that's dedication man, that was two cards ago and i'm one cheap sob about buying new hardware
bridgman: T41, I think
kcodyjr: either way, not the one i've got.
bridgman: kcodyjr; I'm not a developer, I'm a pointy haired boss ;)
bridgman: wonders why secret features have wiki pages
kcodyjr: cranial embellishments notwithstanding, you do a good job hiding it.
kcodyjr: and, sounds like they're no longer secret
bridgman: I guess ;)
kcodyjr: i actually had one of those... i mean for real, the hairstyle might have been stolen from him
bridgman: give me 5 years
kcodyjr: actually, another recent boss shaved his head when that was imminent
simplexe: airlied: ok i found it
Nightwulf|work: hi all
dileX: hi Nightwulf|work
Nightwulf|work: hi dileX
dileX: new drm-next commits. yesterdays kernel-build was a bit freaky.
airlied: agd5f: I'm getting as far as fbcon restore on resume now, just can't convince the accel to come back
mjt: Hello. I've an AMD780g internal graphics which shows an.. interesting cursor shape on a regular basis. That is, most of the time the cursor looks ok, but sometimes it switches to some.. mode where looks very different from the normal.
mjt: Here are 3 examples: http://www.corpit.ru/mjt/cursor.jpg
mjt: it can be fixed by rebooting or it fixes by its own in about half a hour or so.
mjt: anyone know what's going on?
mjt: on the pics: left - should be normal mouse arrow; middle - text cursor; right - window resize arrow.
hlukk: mjt: did you try to put line 'Option "SWcursor" "on"' to xorg.conf's Device section?
mjt: i tried, long ago. Bit I didn't like the result at all: the cursor becomes flickering when changing windows, this is very annoing
mjt: it gets redrawn several times when I type each character in a text field for example.
yangman: mjt: known problem. no solution yet
mjt: btw, this thing - "interesting" cursor - happens with either radeon or radeonhd
mjt: yangman: the garbled cursor you mean?
mjt: or flickering? :)
yangman: mjt: cursor
mjt: maybe there's some way to reset/reinit it forcibly?
yangman: no
mjt: 'hwell... :(
mjt: at least it does not happen every day. I'm an optimist ;)
yangman: it's a hardware quirk that we don't know the workaround for
bridgman: yangman; ping
yangman: bridgman: pong
bridgman: the cursor corruption; I assume someone has already checked the cursor image to make sure it still has the right contents ?
bridgman: I think we were over-writing the cursor image in fglrx; some obscure memory manager bug
bridgman: fix memory manager => no corruption
yangman: yup, memory content is fine. registers are all sane
bridgman: ok, just checking, thanks
EruditeHermit: bridgman, hi
airlied: agd5f, glisse: drm-r600-sr branch of my drm tree has where I'm at now
bridgman: presumably we need to find and document the "scramble" bit
bridgman: maybe tiling or something
bridgman: EruditeHermit; hi
yangman: that's what I remember anyway. I don't think I've ever had it happen on one of my machines
vehemens: run an app that writes to the screen, and corruption goes away.
EruditeHermit: bridgman, do you think new graphics processors for laptops will be out soon with upcoming evergreen? or do they lag behind and if so roughly how long?
bridgman: the chips themselves become available at roughly the same time (say within a month) but the design/qual cycles for laptops are a lot longer than for plug-in cards
EruditeHermit: so how long before you estimate vendors start shipping laptops with them
bridgman: IIRC the chips usually show up in laptops ~4-ish months after desktop
EruditeHermit: =(
EruditeHermit: I sort of need a laptop now, and my only options for the weight I want are nvidia
bridgman: our chips are too heavy ?
EruditeHermit: it seems so
EruditeHermit: =p
airlied: I think my r600 card weighs more than my laptop
EruditeHermit: they end up on 15inch or heavier laptops
vehemens: needs to get out more
EruditeHermit: you won't find any 13 or 14inch laptops with ATI cards on Dell
bridgman: vehemens; if people got out nothing would get done
EruditeHermit: the chips themselves may not be heavier, but they are used on heavier machines
EruditeHermit: atleast for r6/7xx generations
vehemens: power issue?
EruditeHermit: maybe
EruditeHermit: it seems NV 9400M is very popular
EruditeHermit: however, nouveau doesn't support it
EruditeHermit: which makes it difficult
EruditeHermit: airlied, did you see my post about the attached display on resume?
EruditeHermit: palimpsest says my HD is failing =(
vehemens: i thought the main problem with nvidia gpu's is that they die an early death
bridgman: ahh, 9400M is IGP
EruditeHermit: that too
EruditeHermit: nvidia GPUs die early
EruditeHermit: I really would like to buy an ATI one
EruditeHermit: simply because of gallium goodness that is to come
zhasha: NO ES BUENO!
EruditeHermit: zhasha, ?
zhasha: sorry, just tired
EruditeHermit: bridgman, I guess AMD wouldn't make ATI chipsets for Intel CPUs
EruditeHermit: bridgman, and some vendors like Dell and Apple don't do the AMD CPU thing
EruditeHermit: so no AMD IGPs
zhasha: EruditeHermit: there are socket specifications, now I don't know anymore but there used to be some convention so that any cpu using some socket would work with any chipset taking that socket
EruditeHermit: zhasha, yeah, but each CPU vendor has their own sockets these days
zhasha: how ugly
EruditeHermit: its all proprietary
EruditeHermit: no standards anymore as far as that goes
zhasha: I would still buy nvidia (because their linux driver seems to work better) and r300g is still way off
EruditeHermit: classic radeon is good enough
vehemens: better in what way?
EruditeHermit: but nouveau doesn't support IGPs
zhasha: vehemens: I'm not sure anymore since I can't use catalyst on my card ever since they dropped r500
zhasha: but before that at least the nvidia driver didn;'
zhasha: didn't cause kernel panics when turning on external monitors*
EruditeHermit: lol
EruditeHermit: catalyst doesn't do that now either
vehemens: zhasha: got a plan on getting amd to fix your problems -> http://techblog.dallasnews.com/mob.jpg
zhasha: vehemens: but they pawmised they would look at gallium
bridgman: pff... no pitchforks
bridgman: they don't scare me
Lutz_Ifer: i'm in for that plan
EruditeHermit: me too
EruditeHermit: they have an office near me too
zhasha: I'm in Denmark
bridgman: not my office I hope; don't make me have to wear a suit
bridgman: ok, that's a fair drive
zhasha: I'm assuming I need to invent a sailing or flying car first
bridgman: it's surprising how few pictures of mobs with pitchforks there are on the internet
MostAwesomeDude: bridgman: They're usually wearing Guy Fawkes masks these days.
vehemens: it's even hard to find a place to buy a pitch fork
zhasha: puts on his Guy Fawkes mask
zhasha: I have the mask, wig and hat :P
bridgman: vehemens; not if you live in the sticks
zhasha: never used any of it, but in case there is a protest, I would like to be there with a "Down with this sort of thing" sign :P
bridgman: I can drive down to TSC (was Tractor Supply Canada) and they have 7 or 8 different models to choose from
bridgman: you can get all kinds of scary stuff there
bridgman: makes sense, you don't want to commit until you see how it's going ;)
zhasha: bridgman: http://www.youtube.com/watch?v=xOobK3mRNeM
bridgman: now here's a mob : http://2.bp.blogspot.com/_O2JHv7M-9RE/Sck0keARK7I/AAAAAAAAAEU/7DIbdUOz4nI/s320/Angry+mob+with+pitchforks.jpg
MostAwesomeDude: zhasha: Never forgive, never forget?
bridgman: zhasha; I can't really do youtube at home, will check it tomorrow
bridgman: ooh, bumper stickers : http://rlv.zcache.com/torch_pitchfork_t8802_i_am_the_mob_bumper_sticker-p128601534447264600trl0_400.jpg
bridgman: I guess we're going off topic again, maybe time for zzz ;)
Lutz_Ifer: just realises that copying >500gb takes some time...
zhasha: MostAwesomeDude: nah, it's just illegal to wear masks and there's a good chance someone will laugh
bridgman: what about ugly people ?
zhasha: I know right? but no, it's illegal to wear masks in public
MostAwesomeDude: zhasha: What. That's, um.
MostAwesomeDude: zhasha: So you guys don't really do Halloween?
zhasha: I think that's the exception
bridgman: he's right : http://www.dumblaws.com/laws/denmark
bridgman: (scroll down)
zhasha: kids don't usually wear masks though
zhasha: "When driving, you must have someone in front of your car with a flag to warn horse drawn carriages that a motorcar is coming." <- that one was a serious suggestion from a newspaper article some 100 years ago
EruditeHermit: zhasha, would you really get in trouble for wearing a mask though?
zhasha: but was never an actual law
zhasha: EruditeHermit: at a protest, probably not as there will likely be people throwing large rocks at the police
zhasha: walking down the street, maybe
EruditeHermit: lol
MostAwesomeDude: EruditeHermit: It could be used to shut down protests where protesters use masks to hide their identities.
zhasha: probably depends on how uneasy it makes the policeman
EruditeHermit: I was going to dress up as V
EruditeHermit: in V for Vendetta
zhasha: EruditeHermit: that would be the guy fawkes mask we just talked about
EruditeHermit: yeah
EruditeHermit: the mask itself doesn't do it though
zhasha: the wig doesn't work too well on me though as I have fairly long hair
EruditeHermit: you need the rest of the outfit
bridgman: what does Danish law say about swords ?
EruditeHermit: there are more stupid US laws
zhasha: any blade longer than 4cm may not be on your person unless you're in a legitimate scouts outfit
vehemens: what's considered legitimate?
zhasha: a real scouts uniform
vehemens: but you don't have to be a real scout?
zhasha: you do but how will they distinguish?
MostAwesomeDude: There's no concealed-carry laws over there?
EruditeHermit: bridgman, Comic books which depict any illegal acts are banned. in Canada
zhasha: I'm not sure
zhasha: don't think so though
EruditeHermit: that basically bans every superhero comic book
zhasha: guns are illegal too
EruditeHermit: I would hope so
MostAwesomeDude: enjoys the possibility of carrying a long knife or gun with permit
vehemens: still got rocks over there don't you?
EruditeHermit: MostAwesomeDude, do you really?
EruditeHermit: MostAwesomeDude, if you can carry one, some half wit can too
zhasha: enjoys knowing that he will probably not meet a man carrying a long knife and/or gun
EruditeHermit: and quite frankly, I don't trust you to carry one either =p
MostAwesomeDude: EruditeHermit: "If you make something illegal, only criminals will do it." I'd like to be able to defend myself, and I *do* know how to use a knife.
bridgman: Oregon is a right-to-carry state, isn't it ?
zhasha: pepper spray is legal
EruditeHermit: MostAwesomeDude, that sentence is a truism
bridgman: pepper spray is prohibited in Canada, like short barrel handguns
bridgman: highly illegal
MostAwesomeDude: bridgman: Yep. You need permits for the bigger stuff, of course.
bridgman: bear repellent, however, is fine (pepper spray in large cans)
EruditeHermit: MostAwesomeDude, if you made breathing illegal, then only criminals would breathe
bridgman: whenever anyone is offering permits, I apply in case they change their mind later
EruditeHermit: is also a true statement
MostAwesomeDude: EruditeHermit: Pretend I'm a methhead. I don't mind breaking into people's homes and taking their things; I also don't mind carrying illegal weapons.
zhasha: oh here's a funny one
MostAwesomeDude: "Criminal," in that statement, applies to people breaking *other* laws.
zhasha: It wasn't on dumblaws evidently
zhasha: If someone breaks into your home, you still may not do him any harm
zhasha: there has been a case of a thief suing the man whose house he broke into because he stepped on a knife coming through the window
EruditeHermit: zhasha, thats true of sweden too
EruditeHermit: and many other countries
zhasha: and another case of a thief suing the man who beat the living crap out of him after having broken into his home... oh and one of the things he sued for was "lost commission"
zhasha: as in "I made no money by stealing so I'm suing you for preventing me"
bridgman: yeah, self defence is frowned on in Canada as well
EruditeHermit: MostAwesomeDude, if you make guns illegal and hard to get, then no one will have them
MostAwesomeDude: EruditeHermit: Except people who intend to do illegal things anyway.
EruditeHermit: MostAwesomeDude, but if they are hard to get, then not as many criminals will have them
MostAwesomeDude: EruditeHermit: s/guns/meth/ and you've got Oregon; s/guns/marijuana/ and you've got the entire West Coast.
EruditeHermit: MostAwesomeDude, the way it stands, every petty criminal has one here
EruditeHermit: MostAwesomeDude, there is a difference, guns are used to kill others. Marijuana doesn't hurt others directly
EruditeHermit: you don't use a gun on yourself
zhasha: in both the above mentioned cases, the thieves won
MostAwesomeDude: Yep.
EruditeHermit: atleast I hope not
MostAwesomeDude: EruditeHermit: Meth's pretty horrible.
EruditeHermit: true
EruditeHermit: but its done to oneself
EruditeHermit: and Meth is illegal
EruditeHermit: why are guns legal
MostAwesomeDude: At any rate, if it makes you feel better, only the cops have the guns.
MostAwesomeDude: Cops and criminals.
EruditeHermit: well in the US if things go wrong, there is a shooting
nanonyme: bridgman: That's kinda what I said?
EruditeHermit: i.e. a protest or something
EruditeHermit: in EUrope, they throw stones
vehemens: cops don't have to protect you. it's the law.
nanonyme: I think.
bridgman: EruditeHermit; guns are legal because all the evidence is that making them illegal doesn't help
EruditeHermit: bridgman, violent crime with guns is far less in Europe
bridgman: presence of illegal guns seems to track level of gang activity more than anything else
nanonyme: bridgman: Or rather that it didnd't mean what he thought but I have no idea what it actually means. I assumed it had to do with array dimensions but that was an assumption, not a fact.
EruditeHermit: bridgman, thats because they manufacture and sell them so easily here
bridgman: violent crime with guns is less in Canada as well, but per capita gun ownership is nearly as high as the US
zhasha: bridgman: in Denmark, if a person gets killed, or is even just missing, it always makes the news
vehemens: bridgman: this is your fault for dropping r500 support
zhasha: in the US, so many people get killed that it's just not interesting anymore
nanonyme: :P
bridgman: hey, I wanted to keep r500
nanonyme: vehemens: When you can't blame Canada, blame ATi? ;)
zhasha: I'm not saying it's because of guns, I just think it's a result of paranoid americans + guns
EruditeHermit: bridgman, who didn't want to keep r500?
MostAwesomeDude: bridgman: You guys have easier access to weed. Moar weed -> less shootings.
EruditeHermit: bridgman, who should we stone?
EruditeHermit: =p
Zajec: does sb know what is "ACPI I/F"?
bridgman: the people who had to code and test for it... and also work on win7... and evergreen... and the next generation of both... .etc...
airlied: Zajec: acpi interface I assume
EruditeHermit: slackers
nanonyme: thinks he would prefer them not to be stoned, might want to use Windows with new ATi cards some day still
bridgman: MostAwesomeDude; the big grow ops are out west
EruditeHermit: =p
yangman: yeah, they're all over here >.>
bridgman: we took most of them out here with thermal imagers; you can see the extra heat from the lights pretty easily with a helicopter
bridgman: predictably, it ends up being "stealing electricity" laws that get people; the drug laws are a lot harder to enforce
nanonyme: bridgman: Btw, how does me saying "probably related to something else" contradict with your explanation? ;)
yangman: like this one: http://www.cbc.ca/canada/british-columbia/story/2009/09/14/bc-grow-op-headquarters.html
MostAwesomeDude: bridgman: Funny story. There were a few stoners that bought an old school bus, drove it out to the woods, and buried it underground.
bridgman: nanonyme; I'm grasping for context
EruditeHermit: yangman, you in BC?
yangman: EruditeHermit: yup
MostAwesomeDude: Couldn't be seen from the aerial thermal imagers.
EruditeHermit: are you an AMDer?
yangman: no
bridgman: MostAwesomeDude; cool, I was just reading about making root cellars out of abandoned vehicles; a grow op is a great idea
MostAwesomeDude: Filled it full of lights and plants. It was out there for years.
zhasha: srsly.. I don't get why weed is illegal
MostAwesomeDude: zhasha: Big Tobacco in the US. Dunno about over there.
EruditeHermit: because tobacco is legal
nanonyme: bridgman: You rushed in to explain the 1D and that I had the wrong idea what they were. Thanks for the explanation but I didn't have any idea what they were. I just said it's pretty intuitive what they are not. ;)
bridgman: ahh
bridgman: I may have missed the "not" ;)
zhasha: In Denmark it's just general idiocy and people refuting scientific evidence
EruditeHermit: science is frowned upon here
bridgman: our best grow-op around here was an abandoned brewery right on a major highway
EruditeHermit: you have to prove things with god
bridgman: the building looked empty but inside the tanks they had huge grow facilities
zhasha: Tobacco? sure it can give you lung cancer which will end up costing the state a fortune in medical bills, but that's only around 70% of the time
bridgman: and only in countries where the state pays medical bills
zhasha: Marijuana? THAT STUFF IS EVIL. IT WILL KILL YOU AND YOUR ENTIRE FAMILY
nanonyme: EruditeHermit: Well, at least you don't have to prove things to God. That might be a bit tricky if God didn't play fair. :)
bridgman: yeah, I saw "Reefer Madness"
yangman: supposedly it's legal to own up to 10 marjiuana plants here
EruditeHermit: nanonyme, he'll take away your $10
zhasha: at least we have a secular government :P take that 'murrica!
bridgman: ahh, so you just explain to the nice policeman that the plants belong to you and 250 friends
nanonyme: EruditeHermit: Or just change the past so you never got it.
yangman: heh
EruditeHermit: yangman, so you're a volunteer?
zhasha: and the general european consensus is "believe in god all you want, but keep him out of the state"
yangman: EruditeHermit: yeah. haven't done much lately. been busy with with another project and trying to get employed
EruditeHermit: zhasha, this country was founded on the freedom of religion
EruditeHermit: the freedom to believe in god
zhasha: EruditeHermit: yeah but jesus also said he wanted a secular government, which you really don't have
nanonyme: The Freedom to Believe (tm) in the Christian God in all its forms? ;)
MostAwesomeDude: zhasha: It's not tooo wacky over here. Mostly, the government's a bit to the right, so naturally most other democracies appear to be SOCIALIST DENS OF SIN in comparison.
zhasha: MostAwesomeDude: I lived in south dakota which believe it or not was not very religious, so I know ;)
zhasha: what I don't get is all the people protesting things like free hospitals :/
MostAwesomeDude: People are strange.
MostAwesomeDude: goes to see if he can flame more Phoronix forumgoers
EruditeHermit: MostAwesomeDude, is sale of alcohol allowed on Sunday in Oregon?
zhasha: lol @ phoronix. good joke
EruditeHermit: blah
zhasha: what is with that law?!
EruditeHermit: what law?
MostAwesomeDude: Meh, nope.
zhasha: sale of alcohol on sundays
EruditeHermit: its not allowed in Oregon?
EruditeHermit: dude wtf
EruditeHermit: I thought Oregon was cool
MostAwesomeDude: That one guy intractably believes in NX as the savior.
MostAwesomeDude: Yeah, we know. You can't buy strong liquor anywhere but the liquor store, and they can't be open on Sundays.
EruditeHermit: zhasha, alcohol isn't allowed on Sundays in TX
EruditeHermit: TX being Texas
nanonyme: MostAwesomeDude: All alcohol or just strong liquor?
MostAwesomeDude: OTOH I only go there for Everclear for rockets, so whatever.
EruditeHermit: I don't drink
EruditeHermit: but I find it funny anyway
zhasha: also, to me it's just ass backwards that you have to be 14 to drive (SD) and 21 to buy alcohol (also SD)
nanonyme: Selling strong liquor is banned all throughout weekend in Finland so I wouldn't find such a limitation weird.
MostAwesomeDude: nanonyme: Just strong stuff. I think the cutoff's around 90 proof or so. You can still buy wine and beer at the mart.
nanonyme: MostAwesomeDude: We've even stricter limitations here then.
zhasha: I very much prefer to have my teens drink their brains out, get tired of it, and THEN get a drivers license
EruditeHermit: zhasha, well they send off 16 year olds here to die in war
EruditeHermit: or 18
nanonyme: MostAwesomeDude: Oh, wait. Actually not. You just can't buy it in bottles during weekend here. :3
EruditeHermit: and they can't vote about it till 21
zhasha: EruditeHermit: is that legal? O_o
bridgman: what, you have to bring a bag ?
nanonyme: Restaurants are still allowed to sell.
EruditeHermit: bridgman, bring a bag?
nanonyme: Just can't buy a bottle of wine or stronger anywhere and take it home.
zhasha: I've been able to buy alcohol for 4 years here, but in the US, I still can't T_T
nanonyme: I think I'm over the age limit in pretty much all countries. I feel old. :(
EruditeHermit: lol
bridgman: EruditeHermit; there's a big push for reuseable packaging here, so you're encouraged to bring your own bag for everything, even bulk foods
bridgman: naturally I thought...
zhasha: yeah but I only just turned 20
EruditeHermit: bridgman, that was random. What was the context for that?
yangman: http://blog.nanobryg.com/2006/09/doggy-bag-beer.htm
EruditeHermit: uses his own bags all the time for shopping
nanonyme: will turn 24 next year...
MostAwesomeDude: is 21
nanonyme: Quarterlife crisis approaching. :o
bridgman: "you just can't buy it in bottles during weekend here"
MostAwesomeDude: Oh noes!
nanonyme: bridgman: Yeah. Implying you can buy drinks and glasses of wine in bars and restaurants still.
zhasha: In DK it's legal to purchase when you're 16, but you can drink at any age your parents will let you
EruditeHermit: bridgman, I got disconnected for a bit
EruditeHermit: I think I lost part of the convo
EruditeHermit: well here, its illegal till 21, but everyone has done it by 18
zhasha: is running a Mac OS X program ported for Windows in WINE
EruditeHermit: a lot by 16 or 17
bridgman: yangman; yeah, that's what I was thinking
bridgman: nanonyme; can you pour the drinks into a bag ?
nanonyme: I guess they think that if Finns were able to buy wine or stronger home during the weekend, they wouldn't stop drinking during the weekend at all. :p
EruditeHermit: lol at beer bags
bridgman: EruditeHermit; see yangman's link
EruditeHermit: yeah I just did
nanonyme: bridgman: Taking drinks to outside restaurant areas is illegal.
bridgman: geez, it's a good think they restrict guns; we wouldn't stand for that ;)
zhasha: nanonyme: remember when Danske Bank bought Sampo Pakki? (I think that's how it's spelled)
nanonyme: You drink what you drink in restaurant, then leave.
bridgman: or you stock up during the week
nanonyme: bridgman: They still sell alcohol on ships.
bridgman: this seems like a law passed by liquor stores wanting to save money by not having to pay staff on the weekends
EruditeHermit: lol
nanonyme: zhasha: Yeah, thank them very damn much for obliterating a perfectly working bank system.
EruditeHermit: I doubt that
bridgman: nanonyme; I figured there would be an exception; in Ontario all the beer and liquor stores are closed on holidays except up in cottage country
nanonyme: bridgman: Keep in mind Finland was one of the very few countries that actually banned alcohol altogether once.
bridgman: I think they're classed as an "essential service"
zhasha: nanonyme: remember that one guy whose account was emptied, to which he responded not with a gun, but with a fucking AXE to the desk of the teller?
EruditeHermit: lol
nanonyme: zhasha: Cute.
zhasha: and I can't help but think "That's so Finnish"
EruditeHermit: axe =p
bridgman: nanonyme; I know some Finns; what did they drink when alcohol was banned ?
bridgman: is that where the awful pine-tar-liquor-stuff came from ?
nanonyme: bridgman: It was a long ago. It led into widespread sales of illegal alcohol.
EruditeHermit: nanonyme, perhaps thats why they are such good drivers
nanonyme: Lots of smuggling and home-made booze.
MostAwesomeDude: Huh. Sounds familiar.
EruditeHermit: they did that in the US too
nanonyme: The alcohol ban obviously was removed since it was of no use and reduced taxes.
bridgman: and pushed alcohol production out to other countries, like Canada
nanonyme: MostAwesomeDude: Yeah, as far as I've heard USA and Finland are the only countries that have ever done that.
bridgman: I wish the US would ban alcohol again; we made a ton of money last time
nanonyme: :D
EruditeHermit: I was shocked that alcohol sale is banned on Sundays
EruditeHermit: in some States
EruditeHermit: its so weird
nanonyme: No big deal imo. You hoarde alcohol during the week anyway.
EruditeHermit: but its just so weird
EruditeHermit: the reason is because its the Sabbath
nanonyme: It's not as if they could ban drinking.
EruditeHermit: right
EruditeHermit: but its the reason behind it
boghog: drinks about 2 beers a month :D
MostAwesomeDude: EruditeHermit: That's the reason behind banks being closed Sunday and only half-open Saturday, but nobody seems to mind (often.)
EruditeHermit: who is your bank?
nanonyme: MostAwesomeDude: Well, we do have web interfaces to the banks nowadays.
EruditeHermit: mine is open on Sunday
nanonyme: And there's debit and credit cards.
zhasha: MostAwesomeDude: you think banks are bad? you should look at "citizen service" as they call it here
MostAwesomeDude: Didn't say it was a problem.
nanonyme: And I guess same with cheques. So no big deal. :)
EruditeHermit: and also its not so much about the Sabbath, its about giving workers a break
nanonyme: Yups.
zhasha: which is basically a fancy word for "people who manage your money"
MostAwesomeDude: Just marking the history.
zhasha: they used to be open only on weekdays, and only 2 hours a day
EruditeHermit: MostAwesomeDude, grocery stores are open but can't sell alcohol on Sunday in some States
nanonyme: MostAwesomeDude: Big stores have to close down during the weekend here by law iirc.
zhasha: from 10 to 12
zhasha: except for the "long" day thursday, where they remained open from 10 to 15
nanonyme: It was some maximum square meter size your store could be for you to be allowed to be open.
MostAwesomeDude: Man. Laws, man. Like, dude.
zhasha: Dude
EruditeHermit: I say we should all get together and form a country
EruditeHermit: =p
zhasha: Most non-triumphant
MostAwesomeDude: I'm gonna sleep; gotta go talk to my bosses tomorrow. Take it easy, guys.
nanonyme: G'night.
zhasha: Bogus dude
EruditeHermit: gnight
bridgman: zzz
zhasha: nite MAD man
EruditeHermit: zhasha, how hard is it to get into Denmark?
EruditeHermit: what level of education do you need?
nanonyme: o.O
nanonyme: EruditeHermit: Do you mean as in to get a citizenship or what?
EruditeHermit: yeah
EruditeHermit: or move there and get a job
nanonyme: Where do you live?
EruditeHermit: and start process towards getting citizenship
EruditeHermit: US
nanonyme: Egh.
EruditeHermit: i live in the best part of it though
EruditeHermit: the least scary
nanonyme: That'd be more of a bureaucratic Hell then. Assuming they've similar system to ours, you'd need some sort of a work permit and that means you already have to have a job there.
EruditeHermit: they don't like US immigrants?
nanonyme: Just asked because naturally anyone living in a EU country can go live in any other EU country.
EruditeHermit: well yeah
nanonyme: Naw, the USA just hasn't bothered negotiating better terms with the EU.
EruditeHermit: well its not harder than any other non EU nation right?
nanonyme: Shouldn't be, no.
EruditeHermit: thats fine
EruditeHermit: standard fare
suokko: airlied: We need a texture that has all pixels with different values and in some good gradient order so finding their location from tiled data is easier
airlied: suokko: yeah I lost my tile tester code, stupid git clean -f
airlied: it wasn't great since I suck at maths
suokko: git branch+git commit ;)
suokko: But I have kernel panic in my code :/
suokko: Some radeon_object_kmap call causes panic in some where deeper in kernel
suokko: And stupid log file doesn't have any info about panic even when it spammed my console with many backtraces
suokko: (backtrace was also so long that it didn't fit to my console)
EruditeHermit: suokko, my FPS increased quite a bit on benchmarks after the last few days worth of code
EruditeHermit: suokko, did you notice that too on xmoto benchmark?
suokko: It has been r600 only improvements ;)
suokko: My r200 is even slower than before because I have a lot extra safety code to stop some font corruption from happening
suokko: (and I don't want to disable DFS/UTS to fix the problem)
nanonyme: suokko: Is it just one that's broken or both, btw?
suokko: nanonyme: They use same code for blitting
EruditeHermit: suokko, maybe it was the upgrade from jaunty to karmic alpha that did it then
EruditeHermit: other things got faster
suokko: goes to cause kernel panic
suokko: failure :P
suokko: aha. I did boot to wrong kernel so it can't work :/
taiu: i think we could easily support invariant vp's on r600 also: http://pastebin.ca/raw/1566613
taiu: makes arbvptorus work and not hang the card
nanonyme: Neat. :)
rah: No package 'xtst' found
rah: wtf is this?
jcristau: libXtst
rah: Changing dependencies is not allowed.
rah: Except with express permission from Me.
jcristau: …
nanonyme: jcristau: Got that mapped to some key on your keyboard?
jcristau: the compose file.
nanonyme: Ah, ok.
zhasha: EruditeHermit: sorry, missed that. To get into Denmark and work, I don't think that's too hard, however to gain citizenship you have to be a certified genius
nanonyme: Funky.
EruditeHermit: zhasha, how so?
zhasha: It's not that you have to be smart, you just have to remember so many ridiculous things like our previous kings, famous soccer players and other stuff half the population does know or care about
suokko: airlied: Do you have M7?
suokko: Does suspend work with it?
nanonyme: zhasha: Half the Danish population, right?
zhasha: yeah
suokko: goes to correct kernel to cause a kernel panic this time
zhasha: or the vast majority of the worlds population
nanonyme: zhasha: Sounds like what we have then. It's hilarious that most of the ones who do have a citizenship wouldn't pass the tests you need to get a citizenship if you aren't born in the country. :3
Lutz_Ifer: quite normal
Lutz_Ifer: except for germany and their planned test. questions are like "your son disbehave. it is correct to: a) rape him b) stone him to death c)beat the shit out of him d) tell him, that this was no appropriate action and ask him to think about what he has done"
nanonyme: Let me guess: c is the right answer? ;P
Lutz_Ifer: nope, but you need just about 50% of the right answers, to just come here and get your free citizenship
nanonyme: *shrug*
nanonyme: Not that I'd in any way want the citizenship. ;)
nanonyme: wonders what drm-r600-sr is
zhasha: nanonyme: I'd fail our test with flying colors
taiu: nanonyme: suspend/resume i guess
nanonyme: Ah.
nanonyme: Would make sense.
xiando: I have the r600
nanonyme: So you do.
xiando: the driver won't let me set primary screen. :-((( other than that it works alright with the radeon driver
nanonyme: isn't sure if "primary display" is a well-defined thing
xiando: nanonyme: display 0, 1, etc. Primary is where some windows who want to be put on the "first" display are placed.
xiando: There is actually a X option in the latest which is Option "Primary" "1" in the Monitor section.
xiando: the ati driver seems to ignore it
jcristau: the driver doesn't care
xiando: (WW) RADEON(0): Option "Primary" is not used
nanonyme: Imo programs that want to be put on the "first" display are incorrectly designed but that's just my option. :p
jcristau: show the config
zhasha: huh, riddle me this. Spotify works on my netbook, but not on this pc
suokko: xiando: Can yo uattach xorg.conf and full log to bug report?
jcristau: i'm guessing you just used it wrong
xiando: https://bugs.freedesktop.org/show_bug.cgi?id=23952
zhasha: It used to work, now it only plays fragments
suokko: How hard it would to add bugzilla autodetect to notice conf and log are text?
xiando: I hate to admit that the main reason I would like it fixed is that sdlmame pops up on the wrong monitor
xiando: gdm also
xiando: and sometimes dialog boxes like "Want ot accept the browser cookie" goes there too - when the browser is on the other monitor
nanonyme: Browser bug?
suokko: xiando: http://wiki.debian.org/XStrikeForce/HowToRandR12
suokko: xiando: Your configuration doesn't look correct to me
zhasha: suokko: persistent cache, how goes the "Xserver config from within Xserver that stays the same after restart"
adamk: Any thoughts on this: http://bugs.freedesktop.org/show_bug.cgi?id=23901
suokko: xserver would need some persistent config storage for runtime configuration and better runtime configuration interface but that is a lot of work to do. Maybe something sqlite based might work well.
airlied: adamk: drm.debug=4 logs might help
adamk: I'll give that a shot.
adamk: airlied, When I pass that on the kernel command line, I'm not noticing any more debug information. This is with drm.debug=4: http://pastebin.com/m633da970 and without: http://bugs.freedesktop.org/attachment.cgi?id=29484
rah: I notice lots of .debug=
rah: are there recommended options for filing bug reports?
suokko: We need wiki page for debugging. That would probably help many to find the problem if it was linked enough to show in google
rah: I'm about to reboot with drm-next
rah: any recommendations?
suokko: rah: If there is problems then some specific value for drm.debug might be useful but it would spam the dmesg and slow a system in normal use
rah: ko, derf
airlied: adamk: try =0xf then
airlied: it should print out the mode de bugging
adamk: drm.debug=0xf ?
airlied: how it picks them etc
suokko: airlied: Is it possible to run KMS in UML kernel boot?
airlied: adamk: yup
airlied: suokko: not really since it needs hw
suokko: too bad :/ I would like to have way to debug my buggy code but I can't find any that would work for me
airlied: welcome to the fun of gpus
adamk_: airlied: http://pastebin.com/m138a514
airlied: hmm I wonder do we dpms it off by mistake
airlied: can you looks in /sys/class/drm/card0*
suokko: Is there some kernel panic handler that writes the backtrace to disk?
airlied: suokko: don't think so, since using a disk driver after the kernel has paniced could do anything
airlied: netconsole or serial console are the usually suspects
adamk_: airlied: dpms is "On" ?
adamk_: airlied: But, then again, it's "On" for the other monitor too.
adamk_: I think I'll have to continue this later. I'm not actually in front of the machine at the moment. :-)
adamk_: airlied: When I do get in front of the machine, the dpms file the correct one to check, right?
airlied: adamk_: yup, no worries I'll try and get some time to play with dual-head this week
airlied: but it'll probably be next week knowing my scheduling
adamk_: That's fine. I actually have another machine I'm going to test with shortly, with two LCDs and an x1900.
adamk_: I'll see the problem exists there too, or if it's just specific to my home box.
AStorm: hello
AStorm: is the S/R branch safe? (as in, no exploding cards or anything)
airlied: it might, its just a dump of whats on my pc
AStorm: seems I got rid of one hang by a mainboard BIOS upgrade (not sr one)
AStorm: airlied: but it looks fairly ok
AStorm: except one FIXME bit
airlied: it doesn't actually do much yet
airlied: well it can restore fbcon if X isn't running
AStorm: it does reinit the pipe and gfx
airlied: any accel kills it
AStorm: yes
AStorm: but 2D X might work
airlied: no
airlied: any accel kills it
AStorm: hm? is shader pin not enough?
airlied: any accel kills it
airlied: reiniting the gpu is busted somewhere
AStorm: huh, I wonder why wouldn't that be enough
AStorm: ah.
airlied: we have an issue if you rmmod/insmod I think it mgiht be same thing
airlied: need to try running ib test on resume also
AStorm: airlied: btw, how can I speed up that firefox scrolling vs xorg-server master and pixman master?
AStorm: you mentioned it's some commented out code
airlied: AStorm: look for MIXED_PIXMAPS in r600_exa.c
airlied: it mgiht help not sure
suokko: airlied: http://nopaste.org/p/atS1IrCT0 Is my binary broken?
suokko: Maybe that just means ttm->swap is not filled correctly
AStorm: hey
AStorm: any idea about that font corruption?
AStorm: I managed to hit that again
AStorm: real easy - open 100 gvim windows with text inside
AStorm: they start to get bad after a while
AStorm: closing them doesn't help
AStorm: :)
AStorm: also, only small glyphs get corrupt at first
AStorm: then progressively larger ones
suokko: Is that r600?
suokko: I have even worse bug now :/ [ 208.079853] Thread overran stack, or stack corrupted
legume: glisse: 4aac047323e3082d0866b8ad3784236632105af4 breaks kms for me ring test fails. rv670 agp with agpmode=-1.
legume: glisse: I boot with nomodesetting then modprobe -r radeon && modprobe radeon modeset=1 agpmode=-1. It works OK with current + that commit reverted.
AStorm: legume: no, modprobe -r is broken with KMS
AStorm: don't do that.
AStorm: (although airlied is working on it as well as glisse)
legume: AStorm: Ok, but I have to as the agpmode=-1 doesn't work in grub. It was OK until recently for me (starting with nomodesetting) so I could be seeing a different issue.
suokko: legume: radeon.agpmode=-1
suokko: in grub
AStorm: hm.
legume: doesn't work (well didn't I haven't tried for a couple of days).
suokko: All radeon parameters in grub has to be in format radeon.
AStorm: dammit
AStorm: suokko, fix your connection ;)
legume: suokko: I know I was doing that but it didn't work, I should have written radeon.agpmode=-1 above.
legume: suokko: I saw someone else on here with the same problem told to try nomodesetting which does work.
franjva: i was trying drm-r600-sr...
franjva: it works in console, but I get millions of
franjva: [drm] copying object ffff8800bf875a00 6 from 0 to 2: ttm: (null) cp 1
franjva: (maybe they were already there in drm-2.6, i'm not sure)
franjva: ok. no, I see the message was added in the latest commit. forget it :P
suokko: franjva: Are you passing drm.debug kernel paramter?
franjva: no
franjva: look at the last part of the 66bfbc7bf942eb7bfa53aedcdfcddc10e23986c0 commit
franjva: those are the spamming lines
taiu: franjva: seems ok then
franjva: taiu: yeah, i've removed them now from the patch. /var/log/messages is growing too fast
franjva: anyway, just wanted to say that s/r works on r600 for me now (even though I have to stop X) :D
AStorm: suokko: yes, I'm on RV670
AStorm: btw, the corruption "fixed" itself
AStorm: somehow
AStorm: weird.
AStorm: after another 50 gvim windows
AStorm: at least mostly
AStorm: suokko: the easiest way to get this reproduced is to run compositor
AStorm: e.g. xcompmgr
AStorm: but it does happen very rarely w/o it too
MrCooper: AStorm: most likely some pixmaps get corrupted when the backing BOs are migrated into / out of VRAM - is this an AGP card?
AStorm: no
AStorm: PCI-e
suokko1: I don't understand how I can cause stack overflow :/
suokko1: Maybe some pointer is unitialized ...
AStorm: suokko1: no, stack overflow means many function call
AStorm: *s
suokko1: char buf[2]; memset(buf, 0, 50);
suokko1: That is buffer overflow but something stupid has over written my stack
AStorm: try -fstack-protector kernel option?
AStorm: that should give you a backtrace
AStorm: (and for userspace driver too)
suokko1: is not going to recompile the kernel
AStorm: heh
suokko1: It takes over 2 hours
AStorm: huhhow?
AStorm: suokko1: trim your kernel then
suokko1: That takes even longer :P
AStorm: I don't know what could you put in there for the build to take 2h
AStorm: but it's one-time.
suokko1: AStorm: Have you heard about single core cpus?
AStorm: sure yes
AStorm: it doesn't change the fact that there are fast single core cpus
AStorm: hope you're not talking about Atom or Celeron-D
AStorm: *Celeron-M
suokko1: max 1.7GHz athlon mobility ;)
suokko1: So new atoms are faster than my cpu
AStorm: no, they're not.
AStorm: really.
AStorm: unless you're talking about that funky new dual-core Atom.
adamk: airlied, Well, both monitors work with KMS enabled at work, x1900, two viewsonic LCDs.
adamk: Interestingly, though, the mouse cursor disappears on the right monitor, though.
rnoland: agd5f: ping
agd5f: rnoland: pong
rnoland: agd5f: is it known that colors get screwed up in engine?
agd5f: rnoland: not that I know of
nanonyme: napping++
rnoland: agd5f: ok, run engine and hit m a couple of times
rnoland: agd5f: they start out correct, but after going to wiremesh and back they become somewhat metallic looking...
agd5f: rnoland: looks like the textured variant ends up for both the textured and non-textured options
rnoland: agd5f: so is that a bug in engine?
agd5f: rnoland: I think it used to work. probably needs to be bisected
adamk: On the plus side, at least neither of the monitors power off after switching from X -> console -> X, as was happening here before.
nanonyme: wonders if agd5f's z test fix also had an impact on any realworld applications (and indeed whether realworld applications suffered of the bug)
rnoland: agd5f: bisect mesa or engine?
agd5f: rnoland: mesa
rnoland: agd5f: k, will try and get that shortly.
kdekorte: adamk, there is a bug for the missing mouse cursor http://bugs.freedesktop.org/show_bug.cgi?id=23234
adamk: Yeah, I thought I remembered seeing something about it.
adamk: Oddly enough, it's reappeared now.
adamk: Did it ever do that for you?
kdekorte: adamk, there is a sequence that you can use to make it reappear using xrandr... see the report
kdekorte: I don't think it ever reappeared for me
kdekorte: adamk, although I could see how switching to console and back could cause it to reappear
adamk: Oh, that must have been it.
adamk: I didn't do anything with xrandr, but I did drop to console.
adamk: And loved how beautifully that went.
adamk: Now, unfortunately, all 3D applications are empty windows.
kdekorte: adamk, kms is just a little to crashy for me, I had a ton of lockups yesterday just doing normal desktop things, so I have disabled it for the moment
marvin24: does anyone also see a kernel memleak with current drm-next?
marvin24: mmh - will try to bisect it then
AStorm: hmm
AStorm: it seems that KMS sometimes doesn't init my RV670 PCI-e correctly
AStorm: I get a black screen
AStorm: airlied, glisse: that's w/ that r600-sr branch, but it happened before too
nanonyme: AStorm: Monitor goes to powersave mode?
AStorm: not sure
AStorm: might be.
AStorm: it's off certainly
AStorm: system is working fine otherwise
kdekorte: what is the r600-sr branch?
nanonyme: For me monitor light blank => power off, yellow => powersave => green online.
AStorm: nanonyme: my laptop panel doesn't have that luxury
nanonyme: And some crashes end it up in yellow.
AStorm: it's not a crash
nanonyme: Right...
AStorm: it seems like it fails to init
AStorm: happens with about 20% probability
AStorm: it is switching the screen mode (blinking)
AStorm: but after that, a black screen w/o backlight
nanonyme: Right...
MrCooper: marvin24: I'll be submitting a fix for a huge leak hidden in commit 9f022ddfb23793b475ff7e57ac08a766dd5d31bd shortly
nanonyme: loops
AStorm: MrCooper: :)
nanonyme: AStorm: Btw, how was it again: do you have the dword issue or not?
AStorm: does a backflip
AStorm: nanonyme: none yet, we'll see
AStorm: it's very rare
nanonyme: As in, in dmesg.
nanonyme: Hmm, right.
AStorm: nopes.
AStorm: I do have some debug msgs
nanonyme: Usually happens for me sooner or later if I casually use the computer.
AStorm: since I added extra drm debugging
adamk_: Yeah, once I disable KMS, opengl applications show up with content in their windows.
nanonyme: Ahh.
AStorm: tons of spam about copying objects
AStorm: have to filter it or it'll fill up my drive ;p
nanonyme: Heh.
kdekorte: nanonyme, are you still running DRI2? or were there to many lockups for you?
nanonyme: I am atm, yeah.
marvin24: MrCooper: thanks! will try it out.
marvin24: I guess you mean the alloc in *_cs_parse ...
MrCooper: right
MrCooper: cost me a couple of hours to bisect it :(
kdekorte: MrCooper, does it affect all cards or just a subset?
MrCooper: all of them AFAICT
MrCooper: every CS emit ioctl call allocates a sizeable chunk of kernel memory that is never freed
MrCooper: for me it only survived about two runs of x11perf -range aa10text,aaftext -repeat 1 -time 3
MrCooper: patches sent
kdekorte: MrCooper, could that be what is causing lockups in dri2?
agd5f: MrCooper: we ought to just get rid of the radeon_native_mode struct and use a drm_mode
MrCooper: kdekorte: possibly, in particular if there's a slowdown before the lockup
MrCooper: agd5f: if you say so :)
kdekorte: MrCooper, yes, I see stuttering in my card occasionally.. btw is the patch in git or only on a mailing list?
MrCooper: kdekorte: dri-devel list
MrCooper: 'drm/radeon/kms: Free CS parser state tracking memory.'
rnoland: lalala... bisect would go much faster if i wasn't rebuilding ports at the same time....
kdekorte: MrCooper, well I do see the leak now... building the new module now and will test in a minute
kdekorte: damn.. machine hardlocked logging into gnome
MrCooper: kdekorte: I plead not guilty ;)
kdekorte: MrCooper, had this problem quite a bit
kdekorte: if I swtich to dri1, I don't see it
agd5f: kdekorte: currently with kms, unloading and reloading doesn't work reliably. the engine doesn't get reset properly
agd5f: at least on r6xx/r7xx
agd5f: unloading and reloading the drm modules that is
kdekorte: I don't think that I am unloading and reloading unless it is somewhere in startup
kdekorte: agd5f, and it doesn't always do it
agd5f: kdekorte: ok. I wasn't sure if you were rebooting or just reloading when you were trying new modules
kdekorte: agd5f, no full reboot when I install new modules
kdekorte: MrCooper, well the leaking seems to be fixed... so +1 on that patch
kdekorte: MrCooper, can you fix the cursor issue in http://bugs.freedesktop.org/show_bug.cgi?id=23234
MrCooper: afraid not
rnoland: agd5f: e6ad286a80eadd3f38105bf3643e13db83c5b40e is first bad commit
rnoland: commit e6ad286a80eadd3f38105bf3643e13db83c5b40e
rnoland: Author: Alex Deucher
rnoland: Date: Tue Sep 8 16:03:25 2009 -0400
rnoland: r600: fix dri2 clipping
rnoland: :040000 040000 99380a5061845718ccb9415868858a451d0277be 1aaf168878772e10df0186951156d44b9cae8468 M src
agd5f: rnoland: weird. that commit only affects the dri2 path
rnoland: agd5f: yeah, i was just looking at that...
rnoland: unless i did something to screw up the bisect
rnoland: i.e. the bad commit is outside of r600
rnoland: let me step the bisect back to pull in all of drivers...
nanonyme: Yeah, maybe shared code...
agd5f: rnoland: could be. bisecting just r600 found the wrong commit for me as well for a previous issue
suokko: That why you should always bisect without path in mesa. You maybe win many rebuilds with limited bisect
MrCooper: rnoland: you could just bisect the whole tree between the good/bad r600 commits
rnoland: MrCooper: yeah, except i already started with drivers/dri
rnoland: though that would have been better...
rnoland: hrm, still got me to the same commit...
rnoland: now for the full tree...
rnoland: oooh, a bpaul commit
rnoland: bcb62ae78a9d2f4d08001e9f207b6f1291443968 is first bad commit
rnoland: commit bcb62ae78a9d2f4d08001e9f207b6f1291443968
rnoland: Author: Brian Paul
rnoland: Date: Thu Sep 3 21:27:06 2009 -0600
rnoland: mesa: _mesa_meta_bitmap() function
rnoland: :040000 040000 59187a621eb3c63005926ba958fa0ad610ddbb88 eb4da44a186e7f90c07b64a0abf7e1307e015c25 M src
rnoland: agd5f: hrm, that patch exceeds my ability to review...
nanonyme: agd5f: Btw, are the dword sizes for blits documented somewhere? Didn't spot a docco that mentioned them.
nanonyme: As in, r6xx/r7xx.
agd5f: nanonyme: the number of dwords you write to the ring
agd5f: commands
agd5f: nanonyme: depends on the command stream
nanonyme: Hmm.
nanonyme: Just mostly wondering where the extra 2 dwords in rv6xx come from, I guess, and where it's stated.
agd5f: rnoland: can you file a bug so we can track it?
agd5f: nanonyme: rv6xx needs to additional dwords when setting the cb base
agd5f: *two
agd5f: nanonyme: see set_render_target in r600_blit_kms.c
nanonyme: Ah.
nanonyme: Thanks.
agd5f: rnoland: brianp may have an idea. but might be a similar type of issue as 23657. presumably the old behavior was hiding a bug in the r600 driver
AStorm: I wonder if xf86-video-fbdev works with KMS radeon
AStorm: (slow, but it should)
nanonyme: Hmm...
AStorm: time to try that for a change
nanonyme: wonders why on Earth one would write 2 << 0 instead of 2
nanonyme: Not that it matters.
franjva: hi. has anyone compiled libdrm and mesa 32bit in x86-64?
AStorm: sure, why not
franjva: is there any way to disable cairo?
AStorm: just pass -m32 in CFLAGS
AStorm: what does this have to do with cairo?
AStorm: ah, the demos
franjva: my ./configure is detecting cairo.pc
AStorm: yes, there is a configure option
franjva: but I only have the 64bit version and fails at make
kdekorte: franjva, yes I have
franjva: the configure only talks about --disable-{udev,radeon-experimental,nouvau-experimental}
nanonyme: agd5f: So... PACKET3(PACKET3_SURFACE_BASE_UPDATE, 0) takes one dword (checked from registry guide, seems to conform; 2-1 = 1) and the other comes from writing 2 << 0 then. Can't sadly see anything obviously wrong but that's the way you count it, right?
franjva: --disable-cairo doesn't work
franjva: i still get checking for CAIRO... yes
franjva: and it's not there, it's just the 64 bit version that is there
kdekorte: franjva, can you install a 32bit cairo-devel package
agd5f: nanonyme: each radeon_write_ring writes one dword
nanonyme: Ah.
franjva: kdekorte: sure, I could, but I don't want to pull all the emul-linux-x86-crap libraries that come with it :)
franjva: i use gentoo
nanonyme: agd5f: Doesn't depend on packet3 size?
kdekorte: franjva, you could also try --without-cairo
nanonyme: Okay then.
kdekorte: or --without-demos
agd5f: nanonyme: each command has some number of dwords associated with it
nanonyme: kdekorte: There's no such flags in configure --help
kdekorte: nanonyme, there is a --without-PACKAGE
nanonyme: franjva: Make the cairo.pc unreadable during configure, should do the trick.
franjva: kdekorte: no luck. still detects cairo
franjva: nanonyme: yeah, that's what I was doing, but I wanted to make it "cleaner"
franjva: I wanted to avoid renaming cairo.pc everytime I updated libdrm-32
franjva: same with libudev
agd5f: franjva: just don't build the tests
kdekorte: franjva, there are also CAIRO_CFLAGS and CAIRO_LIBS
franjva: (I don't know where is libudev in emul-linux-x86-xxxx)
agd5f: edit the drm Makefile and remove the tests build
kdekorte: could try setting those to bogus values...
agd5f: you don't need the tests just to build the libs
franjva: agd5f: thanks, thats an option. just a sed -i Makefile -e '/^SUBDIRS/s~tests~~' did the trick
nanonyme: Hmm, rounding errors should be pretty negligible too. *shrug*
AStorm: franjva: btw, it's the demos and examples that want cairo
AStorm: just --disable them
franjva: AStorm: but there is no option for that, I think
AStorm: there is
AStorm: --disable-examples --disable-demos
agd5f: AStorm: libdmr
agd5f: libdrm
AStorm: hmmh
AStorm: why would libdrm want cairo?
franjva: yeah, that's for mesa
agd5f: AStorm: for the tests
franjva: for libdrm there is no options
AStorm: franjva: so --disable-tests?
agd5f: not for libdrm itself
franjva: it's ok, I've just added the sed line I posted before to the script and everything's ok
franjva: now I can automatically update libdrm and mesa
franjva: (gentoo versions for 32 bit are quite outdated)
nanonyme: Hmm...
nanonyme: I think the debugging information in drm-next for radeon_ring_write could be just a *tiny* big more helpful. ;)
AStorm: and less repeteteteteete...titive
franjva: is it normal that the 32bit version of glxgears is slower than the 64bit version?
nanonyme: AStorm: That might not be easily done.
rnoland: agd5f: ok, so should I report on dri-devel?
AStorm: like, add printk_limit
AStorm: nanonyme: very easily, kernel has that already
AStorm: message limiting
nanonyme: Hmm.
nanonyme: *shrug*
[Enrico]: franjva: glxgrars is not a benchmark, so it is worth to use something else to test performance
[Enrico]: glxgears*
nanonyme: Well, I don't know how, just rather more interested atm in pulling information out on the dword counts.
franjva: [Enrico]: yeah, usually, but being exactly the same machine...
franjva: I get 2400 vs 3200
AStorm: I'm more interested in limiting "copying object" message
franjva: and it seems... choppy
AStorm: franjva: hmmh.
[Enrico]: franjva: mhm well try with some 3d/opengl game to see if it is choppy too
franjva: the only 32bit games I have are darwinia and doom3
franjva: neither one starts
AStorm: so get enemy-territory
AStorm: :)
AStorm: that's q3 engine, will work
simplexe: warsow - q3 engine?
AStorm: no idea
AStorm: if so, highly tweaked
franjva: wait, I can emerge quake3-bin to get a 32bit version
AStorm: true
AStorm: or rather, quake3-demo
nanonyme: AStorm: Btw, you've DRM_DEBUG_CODE enabled, aight?
AStorm: unless you have Q3 data files
AStorm: nanonyme: ayup
AStorm: that message needs some limiting
nanonyme: Yeah, you should really be getting this too...
AStorm: or I'll set dynamic_printk to disable it
nanonyme: I've suddendly getting suspicions about the usage of radeon_ring_write...
nanonyme: Oh, never mind. Just misunderstood.
nanonyme: Okay, now I stoppped understanding again. Maybe I should eat something.
dileX: dcc brain-food to nanonyme
nanonyme: Seriously, it's as if things were going out of order here. :p
nanonyme: As in, in wrong order.
franjva: ok, I don't get this. I've removed my compiled mesa and libdrm
franjva: reverted to gentoo's versions (from 2008)
dileX: booted in non-kms env
franjva: and glxinfo 32bit says OpenGL version string: 1.4 Mesa 7.7-devel
AStorm: it reads that from X server
AStorm: I guess.
franjva: and quake3-bin is fast
franjva: I thought a 32bit version of recent libdrm/mesa was needed to get acceleration in 32 bit games
nanonyme: Ahhh.
adamk: franjva, If the libraries aren't available, it could be falling back to indirect rendering.
adamk: franjva, What is the output of 'glxinfo | grep render' ?
nanonyme: So radeon_ring_lock can actually return (due to an error) without having changed rdev->cp.count_dw to the value of the dwords in blit. radeon_ring_write, on the other paw, warns if rdev->cp.count_dw is 0 or less and then decrements rdev->cp.count_dw and does its stuff.
franjva: adamk: 64 bit=yes. 32 bit=no :). what does that mean?
_Groo_: hi/2 all
adamk: franjva, Exactly what I said. 32 bits apps can't find the 32-bit GL library and driver, so they are reverting back to indirect rendering.
suokko: nanonyme: also that ring lock function is evil when it calls wait fence in a mutex
adamk: franjva, Which is accelerated, so many apps will run, just slower than they could.
_Groo_: im having a strange crash when using bespin with the experimental ARGB support, X crashes, could any dev take a look t the x dump? i wanna know if its a radeon driver problem or a X related one
nanonyme: suokko: That looked like the only place that could cause the warning that floods my dmesg, yes.
suokko: nanonyme: and DRM_ERRO to that error condition
suokko: +R
_Groo_: see pastebin http://pastebin.ca/1567283
nanonyme: The scenario when rdev->cp.count_dw isn't reset is exactly and might be negative is exactly when radeon_fence_wait_next fails.
nanonyme: That sentence made little sense.
nanonyme: As in, the exact scenario where it could be negative is exactly blah blah.
agd5f: rnoland: file a bug on bugzilla
_Groo_: agd5f: hi akex
agd5f: _Groo_: hi
_Groo_: agd5f: alex*
_Groo_: agd5f: could you take a quick look at http://pastebin.ca/1567283 and tell me if the segfault is X or radeon related?
_Groo_: agd5f: so i can open a proper bug
agd5f: _Groo_: I added non-native lvds modes by default in kms
franjva: adamk: ok, I get it
suokko: nanonyme: if ring is full and wait fails that would trigger GPU reset which would dead lock
nanonyme: Egh.
agd5f: _Groo_: looks like a crash in pixman
nanonyme: suokko: That would either imply I'm wrong or that reset is very good since I don't experience many crashes.
_Groo_: agd5f: yeah, im using latest 1.6.0, but although it looks like pixman, its actually exa/composite related..
suokko: nanonyme: You would see the GPU reset in log
nanonyme: suokko: Good point. If I silence out the dword stuff, might just work.
_Groo_: agd5f: looks like bespin is calling some broken function in composite that crashes X.
nanonyme: suokko: The problem is that if I'm right, on each time GPU resets, 78 warnings due to dwords get written in dmesg.
nanonyme: Should be able to workaround by silencing it though.
suokko: nanonyme: http://nopaste.com/p/ahA2KcKqnb
suokko: Try that instead
nanonyme: Good point.
nanonyme: Very nice. :)
_Groo_: agd5f: ddx could die if an argba exa call is wrongly called?
Vinky: is it a known problem in latest git that X crashes when trying to run compiz or other apps using 3d acc ?
nanonyme: suokko: Doesn't apply cleanly but it's trivial to copy over anyway. :)
agd5f: _Groo_: if there's a bug in exa or pixman that's causing a problem like a segfault then yes
_Groo_: agd5f: so whats the proper bug report channel? generic X?
suokko: nanonyme: You probably forgot to click the plain text link ;)
nanonyme: Hmm, maybe...
agd5f: _Groo_: looks like pixman
agd5f: I'd say pixman or exa
nanonyme: Guess I'll better retry, having an unbootable kernel due to a typo != fun. ;)
_Groo_: agd5f: ok alex, thanks...
suokko: nanonyme: compiler should find the typos
nanonyme: Not if I do some in macros. :)
nanonyme: (or not necessarily anyway)
nanonyme: suokko: How are __FUNCTION__ and __LINE__ defined, btw?
suokko: gcc builtin
nanonyme: Heh, just got an error on compile. ^^
suokko: What error?
suokko: typo in macro paramter name
suokko: parameter
nanonyme: I think we've nopaste.com to blame.
suokko: rdeb/rdev
nanonyme: Compiler didn't like the macro wrapping on two lines.
nanonyme: Yeah, that's a point too. :D
suokko: so line wrapping and a typo only ;)
nanonyme: Well, saved a little time not having to remake kernel.
nanonyme: (that is, after unsuccesful reboot)
nanonyme: Now just initrd and reboot.
suokko: nanonyme: You don't need to reboot
suokko: You could just reload radeon module
nanonyme: Rebooting is fast enough.
suokko: But if you just reload you don't have to recreate initrd ;)
nanonyme: Oops...
AStorm: initrd? who needs a stinking initrd ;)
nanonyme: I guess that wasn't a good idea, failed to init KMS.
AStorm: initramfs ftw
AStorm: yes, KMS is incompatible with reloading
AStorm: suokko: see.
suokko: AStorm: it works for me
nanonyme: No, I didn't mean that.
nanonyme: The changes caused KMS to fail init.
suokko: nanonyme: How it fails the init?
nanonyme: Unless...
nanonyme: Ah, my mistake.
nanonyme: Will be rebooting soonly.
nanonyme: It always helps if you haven't erased the whole drm-next pull...
stikonas: after latest upgrades to git versions, my kernel can no longer load firmware...
stikonas: can it be that I need to run make clean?
stikonas: on my kernel
AStorm: stikonas: no.
AStorm: forgot make modules_install
AStorm: :)
frojnd: Hello there
frojnd: I just wanna know how can I know if my graphic card has 128MB memory and how much shared memmory.
bpaterni: why do I need firmware?
frojnd: I know there was a fancy command but I can't recall it.
stikonas: AStorm: I use make-kpkg script, strange
frojnd: lspic -v just won't tell enough
AStorm: frojnd: /var/log/Xorg.0.log will tell
frojnd: AStorm: I'm looking for what line? I know there are many of lines :D
AStorm: one that says VRAM.
frojnd: VRAM for virtual ram?
AStorm: Video RAM.
frojnd: (II) RADEON(0): Start of VRAM area used by Firmware: 0x7ffb000
AStorm: e.g. with KMS here:
AStorm: (II) RADEON(0): Front buffer size: 9600K
frojnd: (II) RADEON(0): AtomBIOS requests 20kB of VRAM scratch space
AStorm: (II) RADEON(0): Remaining VRAM size (used for pixmaps): 510560K
frojnd: (II) RADEON(0): AtomBIOS VRAM scratch base: 0x7ffb000
AStorm: no, further down.
AStorm: :)
frojnd: 3 liens that includes VRAM
frojnd: oh.. down :)
agd5f: AStorm: the bios uses part of vram for a scratch area
AStorm: hm.
AStorm: agd5f: yes.
AStorm: but that will be only some KB, irrelevant
frojnd: (II) RADEON(0): Detected total video RAM=131072K, accessible=262144K (PCI BAR=262144K)
AStorm: frojnd: that, yes :)
AStorm: now you know.
frojnd: how many MB is this :)
AStorm: ...
frojnd: also I don't know how many actual ram and how many "shared"
AStorm: 128 MB VRAM, accessible 256 MB
frojnd: if any..
AStorm: shared is accessible - VRAM
agd5f: AStorm: no
AStorm: no?
AStorm: well, almost
agd5f: accessible is just the same of the PCI BAR
frojnd: AStorm: so that mean that I have like 378MB of ram for VIDEO?
AStorm: frojnd: no.
AStorm: it certainly means that you have 128 MB VRAM.
frojnd: AStorm: aha
frojnd: Than I don't know what mean "shared" or accessable
agd5f: frojnd: with KMS we use a 512 MB gart aperture. pre-kms we use 32 MB
agd5f: well 32 MB on r3xx+ and 8 MB on r1xx/r2xx
frojnd: How can I check how fast is gpu?
AStorm: frojnd: hmm, not easily
AStorm: you could run a benchmark, like quake3-demo
agd5f: frojnd: play a game like quake
frojnd: agd5f: and see output where?
frojnd: I already have nexuiz installed
frojnd: if that helps
AStorm: frojnd: press ~, write cl_showfps 1
AStorm: that's for nexuiz
AStorm: :)
AStorm: afaicr
frojnd: ok I'll wrote cl_showfps1
frojnd: brb
frojnd: Hope it doesn't freeze :D
AStorm: ate the space
AStorm: ...
legume: glisse: I recompiles kernel with CONFIG_DRM_RADEON_KMS so I don't need to modprobe -r radeon and I still need to revert 4aac047323e3082d0866b8ad3784236632105af4 to avoud a ring test error when I modprobe radeon modeset=1 agpmode=-1.
legume: glisse: I mean without CONFIG_DRM_RADEON_KMS
frojnd: cl_showfps 1 command not found AStorm
AStorm: frojnd: hmh, then check nexuiz page.
agd5f: frojnd: maybe just showfps 1
legume: frojnd: nexuiz has a show fps setting on the gui settings under misc IIRC
frojnd: AStorm: showfps 1
frojnd: and I saw maximum of 25fp/S
frojnd: does that make my graphic card crappy?
AStorm: not too much, assuming proper settings in nexuiz
frojnd: We were trying to determine how fast is gpu in my radeon mobilitx x1400
suokko: frojnd: wikipedia has the nice specs for all the cards
agd5f: frojnd: you need latest mesa for best performance on your chip
legume: frojnd: I found nexuiz fps quite sensitive to settings under effects.
frojnd: agd5f: I'm waiting for packager to submit the lates version of radeon driver.
rnoland: agd5f: hrm, looks like brian committed a fix before my bug mail has come out...
rnoland: ah, there it is..
agd5f: rnoland: hopefully that fixes it
nanonyme: C'est la vie... When you put in the additional debugging info, problem disappears.
rnoland: ooo, heisenbug
adamk: airlied, 'cat /sys/class/drm/card0-*/dpms' shows "On" for both monitors, the one that's not receiving a signal and the one that is.
rnoland: agd5f: yep, all fixed
agd5f: rnoland: cool
legume: rnoland: does the lighting on the pistons look slightly wrong to you (not a new issue or anything to do with "m")
rnoland: legume: yes, something still seems a little odd, but nothing i can really point to...
rnoland: it's like some flickering in the lighting on the upper pistons of the v6
nanonyme: Seriously, this is ridiculous.
legume: rnoland: maybe curved surfaces are a problem, gloss also has some issues (no reflective bands) for me.
legume: non relective even
rnoland: i don't notice it except with the default texture...
legume: rnoland: yea it's the default startup settings that I see it on.
moe61: hello. Yesterday I got a update of my debian testing to xorg 7.4. Now I have some problems with my Xorg configuration. When I start my xorg without any xorg.conf, my radeon hd3200 seams to bea fully initialized because I have a xv-screen and glx acceleration. But the resolution seems to be false, my input devices are not fully configured (dead keys on keyborad, no mouse). When I try to start with my old xorg config I my input devices are
moe61: initialized but my radeon is not. xrandr says it hat the right resolution ov 1280*800 but the lower taskbar of my kde is cut of and the dpi seems to be to big. How do I configure my radeon device right?
moe61: hello. Yesterday I got a update of my debian testing to xorg 7.4. Now I have some problems with my Xorg configuration. When I start my xorg without any xorg.conf, my radeon hd3200 seams to bea fully initialized because I have a xv-screen and glx acceleration. But the resolution seems to be false, my input devices are not fully configured (dead keys on keyborad, no mouse). When I try to start with my old xorg config I my input devices are init
moe61: radeon is not. xrandr says it hat the right resolution ov 1280*800 but the lower taskbar of my kde is cut of and the dpi seems to be to big. How do I configure my radeon device right?
AStorm: hello again
AStorm: what do I have to patch to get faster KMS/firefox scrolling w/ xorg master?
AStorm: (and pixman master too)
nanonyme: Hmm, do GL_ARB_shader_objects, GL_ARB_shading_language_100 and GL_ARB_fragment_shader require GLSL?
suokko: nanonyme: They are GLSL ;)
nanonyme: Right...
nanonyme: So we reach the conclusion that Adobe Flash does require GLSL. :)
nanonyme: (for acceleration, that is)
soreau: nanonyme: I think they will fall back to sw rendering which is slow
agd5f: moe61: not sure about the input stuff. but for video: http://wiki.debian.org/XStrikeForce/HowToRandR12
nanonyme: suokko: Btw, I haven't managed to get any errors after I put in your debug code. :D
suokko: Did you forgot to compile with drm debug enabled?
nanonyme: I don't think I changed any config variables. Hrm.
agd5f: nanonyme: load drm.ko with debug=1
nanonyme: But it worked before with that. :p
agd5f: or echo 1 > /sys/.../drm/parameters/debug
nanonyme: Ah. ^^
nanonyme: I'm still baffled as to how it could have been disabled by itself though.
nanonyme: Since I didn't really change anything.
agd5f: nanonyme: it's always disabled by default
nanonyme: It was enabled with an earlier kernel and I didn't change anything. :)
suokko: #if DRM_DEBUG_CODE might be problem
agd5f: nanonyme: some messages are printed even without debug enabled
agd5f: DRM_ERROR stuff for example
nanonyme: suokko: Yes *but* it worked before.
nanonyme: And it's not clearly obvious why.
suokko: But you did probably upgrade to new commit
nanonyme: I did, yes. I don't think there were any new ones in my local drm-next tree.
nanonyme: Just that radeon_ring_write wrapper.
nanonyme: But if debug mode goes randomly on and off for me, no surprise I only got the stuff in dmesg on some boots. ;)
nanonyme: Hrm, dammit.
rnoland: yes, DRM_INFO and DRM_ERROR always print
nanonyme: Me fail.
nanonyme: Now I got the message and noticed I had wiped the debug info. Reeetry. :(
rnoland: although linux just changed some of that around, so i dunno anymore...
nanonyme: rnoland: As suokko said, the DRM_ERROR is inside a DRM_DEBUG_CODE block.
rnoland: hrm, that seems like a bug then...
nanonyme: radeon.h -> radeon_ring_write inline function
rnoland: hrm, i haven't imported that i guess...
agd5f: nanonyme: DRM_DEBUG_CODE is separate from the module option
nanonyme: Ah, right.
nanonyme: agd5f: Is DRM_DEBUG_CODE nowadays always on?
agd5f: nanonyme: I don't know off hand
nanonyme: Hmm, I guess I should try grepping then.
nanonyme: I guess it's safe to assume though that it is.
agd5f: nanonyme: or just find out
nanonyme: Yeah. ;)
nanonyme: Well, I know it is at least *part* of the time on.
nanonyme: But whether it's been on with all of my kernel builds, hard to say.
rnoland: looking at drm-next, it should be enabled
rnoland: it is defined in include/drmP.h
kdekorte: agd5f, is drm-next still the preferred branch for KMS testing or should we be using another right now?
rnoland: though i don't see anything saying why it is defined to 2, or what different values might do...
rnoland: it is only referenced 3 places for radeon though.
moe61: hello again
moe61: my setup is now better. It has the right resolution and the right dpi but no xv and no 3d acceleration
moe61: How do I get this on with my hd3200?
nanonyme: Which kernel do you have?
moe61: debian 2.6.30-1
nanonyme: Try installing linux-firmware.
moe61: is already
nanonyme: Oh, hmm.
nanonyme: What does your xorg log look like?
moe61: [drm] Initialized radeon 1.30.0 20080528 for 0000:01:05.0 on minor 0
moe61: where can I drop my log?
nanonyme: nopaste.com works
moe61: http://nopaste.com/p/aWQAGdBeq
suokko: moe61: Some of xserver modules are not loading correctly. You should reinstall them
soreau: moe61: I think you need aiglx on
nanonyme: Shouldn't affect Xv.
agd5f: soreau: aiglx isn't needed
suokko: moe61: Also I think that you kernel doesn't have 3D support for r600 cards
soreau: agd5f: It isn't?
suokko: But xv should work
nanonyme: (EE) module ABI major version (0) doesn't match the server's version (2)
soreau: agd5f: For 3D?
moe61: yes and 3d also works
nanonyme: That is bad for libdri.
moe61: without xorg.conf
nanonyme: *Very* bad.
agd5f: soreau: nope. only needed for indirect rendering
moe61: but i need a xorg.conf for my inputs
soreau: agd5f: Ahh
soreau: moe61: How many input devices do you have man? ;)
nanonyme: Doesn't matter much...
moe61: keyboard mouse touchpad touchscreeen
nanonyme: Main thing is that his X.org stack is borked.
suokko: I suspect fglrx
moe61: is deinstalled
nanonyme: suokko: That and a partial upgrade are the most likely causes, yes.
suokko: But after that you need to reinstall a lot of stuff to make sure that broken fails are fixed
suokko: I think that there should be easier way for distros to make fglrx install without breaking oss drivers :/
nanonyme: suokko: Some restructuring in X.org stack and Mesa would help a lot.
moe61: any suggestion what to do?
suokko: moe61: reinstall all x packages and mesa
nanonyme: As in, such that you wouldn't end up replacing anything in the first place to use non-X.org or non-Mesa bits.
suokko: That is the best start
nanonyme: (but no, I'm not seriously suggesting anything like that, especially not just after the last restructuring)
rnoland: agd5f: so the plan is for mesa 7.6 to have r600 then?
agd5f: rnoland: maybe
rnoland: agd5f: would be nice... ;)
suokko: nanonyme: It would be simple with runtime loading and requiring blobs and mesa to install librarires to /usr/lib/
rnoland: agd5f: especially if the font corruption issue were fixed.
suokko: No need to add anything more complex than just way to load correct library t runtime
moe61: aptitude reinstall worked; thanks a lot
moe61: xvinfo and glxinfo seems good
nanonyme: suokko: Yeah but it might require a lot of restructuring to do that sanely.
moe61: good night and thanks for help again
suokko: I think wrapper libgl.so should do the trick nd then xserver should know to load the modules from multiple locations in runtime
nanonyme: Same for libglx.so?
suokko: Ok. Selecting what to load in runtime is the complex part but some easy way to distros configure iti n /etc would be nice
airlied: libGL is already a wrapper
suokko: nanonyme: btw, THat would work without any wrappers too
airlied: they still all replace it
rnoland: agd5f: one additional thing that i've noticed... in nexuiz, the enemies don't show up.
suokko: ld load path would work too and requiring blobs and mesa to install to subdirectories would help
nanonyme: airlied: And define extra symbols as far as I've seen.
agd5f: rnoland: bug in nexuiz
rnoland: weapons do, but can't see the oponents... not that i play nexuiz...
rnoland: agd5f: ok, cool
agd5f: fixed in newer versions
airlied: really closed drivers are going to do the least amount of work necessary to take over the system
rnoland: agd5f: ah, haven't updated lately...
airlied: nothing we can ever do to fix that
rnoland: agd5f: so from my perspective the only issue that i don't classify as a minor aggravation is the font issue....
agd5f: rnoland: yeah. been trying to sort that one out, but no luck so far
rnoland: and i'm not sure that i know of any minor aggravation issues right now either...
suokko: airlied: Except we can make distros change it so that distro packages will install them to /usr/lib/fglrx and /usr/lib/nvidia
airlied: suokko: some distro do that alreay
airlied: some don't, and packages direct from upstream don't
suokko: yes
rnoland: performance is quite acceptable now....
nanonyme: airlied: I thought even the ones that do still have icky scripts that you have to use to change between the stuff and have wrappers in /usr/lib/. If this was the intended behaviour and you'd have runtime loading, just changing driver in xorg.conf would suffice.
nanonyme: (maybe)
airlied: nanonyme: it doesn;t fix client side
nanonyme: Hmm, true.
suokko: Also kernel modules conflict so you have to anyway reboot
nanonyme: Also KMS is gonna cause trouble with all non-Linus-tree-drivers that have a Linus tree equivalent so you have to tamper with bootloader anyway. :3
scott_ino2: hello, i have an ati mobil hd3400 and im wodering whether or not i should use radeon or radeonhd drivers.
scott_ino2: I've been battling with drivers for quite sometime now, and I'm just fed up with the proprietary drivers and no suspend, and don't need compiz. just want good Xvideo and no tearing
scott_ino2: plus i know compiz will come with time for the radeon/radeonhd drivers
stikonas: scott_ino2: you are welcome to use radeon, 3D and XVideo and suspend2disk works for me
scott_ino2: 3d for hd3400 r620chipset??
stikonas: the same one
stikonas: 3D is supported on all chipsets
Ampheus: hd4570 working here with kms+dri2 :)
nanonyme: 3D only in development trees though.
suokko: There is still bugs in 3D
stikonas: bus suspend2ram is WIP
scott_ino2: i don't mind bugs
scott_ino2: and don't need 3d as i have other bigger issues
Ampheus: just played warzone2100 which runs fine
scott_ino2: primarily VLC not working with Xvideo using the radeon driver currently
nanonyme: You should be getting nice Xv with either driver.
scott_ino2: vlc wont use Xv even when set and i don't understand why
AStorm: if there's any ATI CUDA-hacker here... please please implement this:
scott_ino2: do you have any suggestions
AStorm: http://ldesoras.free.fr/doc/articles/resampler-en.pdf
AStorm: it just asks for a gfx card to run it
AStorm: it's the same algorithm as used in latest Sox version.
scott_ino2: scratch that, Xv now works after restard
scott_ino2: ahhhh i think i know what happened, so how do i enable 3d on hd3400 (again, don't mind bugs)
AStorm: a MIP-mapped (multiscale) oversampled FIR filter.
scott_ino2: stikonas, suspend2ram works for me under radeon driver
AStorm: that is almost infinitely parallelizable
scott_ino2: which is what i wanted over the ati driver
nanonyme: scott_ino2: I'd wait for 2.6.32 RC1 if you aren't in too much of a rush.
scott_ino2: nanonyme, no problem i can wait, just glad my Xvideo is working now
nanonyme: It's going to have 3D for your card.
scott_ino2: trying to move away from the proprietary driver
nanonyme: Well, you do get Xv with any 2.6.30 or newer. :)
AStorm: I wonder how well is this algorithm applicable to textures.
nanonyme: It's mostly that if you wait until 2.6.32 RC1, you don't need to play around with development trees.
AStorm: I suspect very well.
scott_ino2: is there a pseudo timeframe on that, again in no hurry, just wanna know if it's a month or 6 months
nanonyme: At most have to compile Mesa.
airlied: rc1 is prob not the time to jump on
nanonyme: airlied: Oh?
scott_ino2: I don't mind compiling, or testing, so long as it's "somewhat" usable
nanonyme: scott_ino2: Right, ignore me and listen to airlied, he ought to know how it's like. ;)
airlied: 2.6.31 + drm-next pulled into it at least has stable non-drm bits
scott_ino2: lol
airlied: -rc1 will have unstable *
nanonyme: Right.
scott_ino2: Xvideo is sooo much friendlier on my cpu
scott_ino2: thank god
scott_ino2: Second question, is there an easy way to configure multiple displays say if im using a projector or TV hookup (vga) with radeon? as I was using the catalyst tool to configure secondary screens
scott_ino2: And does radeon support HDMI sound output
scott_ino2: that hsould be all my questions ;-)
scott_ino2: thanks again i realize i just come in here and start asking questions but seriously graphics on my machine have been driving me insane
suokko: scott_ino2: http://wiki.debian.org/XStrikeForce/HowToRandR12
scott_ino2: yes i can do it that way, however I was wondering if the default "display" utility would detect the second monitor
suokko: And no HDMI audio yet with radeon but Radeonhd has HDMI audio
suokko: scott_ino2: It does
scott_ino2: ty suokko, any disadvantages of using radeonhd over radeon
suokko: radeonhd doesn't have KMS support
rnoland: sigh... i may just port the audio setup code from radeonhd....
rnoland: i can't believe noone else has yet...
nanonyme: rnoland: To KMS? :o
suokko: scott_ino2: default dispaly utility does use xrandr to communicate with xerver so that link is usefull if yo uhave problems and want to check what is going behind the utility
rnoland: nanonyme: no, from radeonhd to radeon
rnoland: the hdmi codec setup
nanonyme: rnoland: Well, I've the impresseion everyone's expecting someone to port the HDMI audio stuff to KMS.
nanonyme: Impression even.
suokko: rnoland: There was patch in mailing list but none did the cleaning up for it
scott_ino2: suokko, ty, i'll test and will be able to figure out for external so im not worried about that
rnoland: suokko: ah, missed that...
suokko: because plan is to have it in kernel
scott_ino2: what is KMS support
nanonyme: Kernel modesetting.
rnoland: i might just make our hdac controller do it then...
suokko: That means fast VT switching, full resolution console, working oepngl with composite and future support for advanced opengl features is easier
rnoland: though i wonder if i'll get into resource sharing issues if i do that...
rnoland: things that make you go, hrm....
spreeuw: ah, so the terminal framebuffers will use that too?
agd5f: spreeuw: yup
spreeuw: that should make things easier for people who like to have both around
scott_ino2: I guess what i want to know is whether I should use radeon or radeonHD, I'd like to be able to use HDMI for video/audio
nanonyme: For now radeonhd.
scott_ino2: nanonyme, ty
suokko: scott_ino2: You can switch between them by only changing the dirver in xorg.conf and restarting xserver if you install both
scott_ino2: that was my next question
suokko: So jsut tets which one works better for you and use it and have another for back up :)
nanonyme: suokko: Assuming he doesn't use KMS?
scott_ino2: if i have radeon intalled now can I install radeonHD through synaptic and have both installed at the same time
scott_ino2: nanonyme, im reading up on KMS and trying to understand it ;-)
nanonyme: I doubt radeonhd would start that well if you boot your kernel with KMS on.
scott_ino2: graphics drivers are seriously something i know nothing about
suokko: yes. nanonymehas point that if you jump in to KMS world with 2.6.32-rc kernel radeonhd will stop from working
scott_ino2: suokko, nanonyme noted not to use KMS with radeonHD and 2.6.32 ;-)
suokko: scott_ino2: http://xorg.freedesktop.org/wiki/RadeonFeature for status of driver :)
nanonyme: Anyway, as soon as KMS gets HDMI audio support (I don't think anyone's working on it, rnoland apparently might be interested), radeon is the better call.
airlied: needs to get a laptop with hdmi r600, my tv is too far away from my desktop
rnoland: nanonyme: i'm not on linux and so kms support will be somewhat different for me...
nanonyme: Right...
nanonyme: Never mind then... Well, anyway. I guess no one's working on it then. :p
Luzipher: airlied: i could plug my tv (hdmi) into my rv770 if that would help
scott_ino2: so for ubuntu would I have to manually create a "driver" section for xorg since it doesn't use that anymore?
airlied: also rhd is missing support for some chips I think
rnoland: it just happens to be the only feature that i'm missing from the radeon driver that is in radeonhd
airlied: someone did send some patches
airlied: they just needed some cleaning up I think
airlied: at least I thought someone did
nanonyme: Yeah, someone did send a patch.
scott_ino2: brb gonna do some testing thanks again for all your help
rnoland: i'll search around... i don't remember seeing them though... but i get a sh@#load of mail....
rnoland: any clue where they might have gone?
nanonyme: rnoland: http://lists.x.org/archives/xorg-driver-ati/2009-July/010721.html
nanonyme: Maybe?
nanonyme: I haven't kep that big of a track on the issue myself either.
nanonyme: It's not like I've any HDMI monitors.
scott_ino2: is there an easy way to tell if im using radeon or radeonhd for ubuntu since it doesn't have a driver section in xorg?
nanonyme: Xorg log.
suokko: scott_ino2: xorg.log
nanonyme: It either says RADEON or RADEONHD, I think.
scott_ino2: in /etc/X11?
rnoland: hrm, i had never looked at this code...
rnoland: i don't think i need 90+% of it... my audio driver will do most all of this...
scott_ino2: nvm ive had a hectic day
scott_ino2: not sure what i was thinking im good
riri: Hi, Just saw there're big improvements with git version of the driver
riri: Does someone knows if AGP cards support is better too or if did not changed?
scott_ino2: well... I think im good for today..... glad im using the open source drivers im seriously not going back to the ATI ones just caused too much trouble for me
scott_ino2: provided everything else works ;-)
scott_ino2: however my xorg.log doesn't seem to be shedding any light on this... is there a specific section I should be looking at
nanonyme: scott_ino2: You may pastebin the whole xorg log.
suokko: scott_ino2: It should be starting around 200-300 lines spamming a lot of messages about RADEON or RADEONHD
scott_ino2: k.. i'll pastebin to confirm
scott_ino2: http://pastebin.com/d551e999b
nanonyme: Yeah, it's radeon driver. Also apparently manages to setup EXA quite nicely. :)
scott_ino2: hmmm i want radeonHD though for HDMI
scott_ino2: easy way to do that without getting rid of the radeon driver?
suokko: scott_ino2: try sudo dpkg-recongfigure -phigh xserver-xorg and the add driver to device section
scott_ino2: k let's see if we can start the magic
scott_ino2: fyi you can do lsmod |grep radeon
scott_ino2: or radeonhd
scott_ino2: i believe
suokko: no. They both use same kernel module
scott_ino2: ahh
scott_ino2: i see
suokko: scott_ino2: You want to add to device section Driver "radeonhd"
legume: riri: What AGP problems do you have & what card?
riri: legume: must emulate PCIE bus type, otherwise xorg hangs (solution given by suokko)
scott_ino2: be back in a sec
riri: hthe problem is with that, I've artefacts in place of the cursor and any drawing with alpha
suokko: riri: none yet. That is hard problem to solve
riri: the solution (again by suokko :-)), is to add the options EXANoUploadToScreen and EXANoDownloadFromScreen
riri: this's quiet good, but this seem to slow down graphics - at least Gtk+ (but improves Xorg)
legume: riri: If it's a r6xx card setting Aperture size to 256 in bios may help, also modprobe radeon no_wb=1
nanonyme: suokko: Wonder if it's the fault of the AGP controller or the driver.
riri: legume: r6xx card yes, Paerture size already to 256
scott_ino2: k all appears to be well
riri: no_wb =1, well, I'll try
legume: riri: Yea some things are slower with no dfs/uts, though kms seems better in PCIE mode.
nanonyme: (erm, rather of the AGP kernel driver or the userspace driver; assumably AGP controller is at fault anyway if userspace driver is attempting to follow the standard)
scott_ino2: new pastebin: which i already know is running radeonhd now: http://pastebin.com/d7d8321b5
riri: AGP and PCI are so different buses?
scott_ino2: but wanted to know if things seem right
riri: just wondering
scott_ino2: does radeon or radeonhd have powersaving features?
stikonas: http://www.x.org/wiki/RadeonFeature
suokko: scott_ino2: both have some basic options. KMS will provide better powersaving
legume: riri: I think AGP is a problem because different mobos don't work the same way, using PCIE lets the driver set some things up.
suokko: man radeon and man radeonhd will give you list of xorg.conf options
scott_ino2: suokko, ty
suokko: There is options to enable low powermode so card is set to low power mode for whole runtime and. DynamicPM which changes to low power mode only for display power saving period
riri: legume: k - if you have any other ideas on bios configuration quirks, they're welcome :)
legume: riri: not sure, you could try 512 Aperture if it's possible, also in xorg.conf you can set the speed with -
airlied: MrCooper: rv350 depth not changing bug? (just want to make aure its not atom)
riri: 512, not sure my mobo accept it
legume: riri: Option "AGPMode" "4" in the device section (assuming it is using 8)
MrCooper: airlied: yeah, my PowerBook with RV350
riri: no_wb:Disable AGP writeback -- well, what does it changes?
airlied: ah I think I see it, set_crtc_timing plays with format
airlied: probably need to rework that so it happens in set_Base
legume: riri: for me it makes some 3D apps work, but I could still startx OK without it (it gets disabled in PCIE mode anyway)
scott_ino2: hmmmm my Xvideo is now gone
riri: :)
riri: bless its soul
suokko: scott_ino2: You might need Option "DRI" "on" for radeonhd
mihaita9999: hello
riri: legume: ok, I'll make a hard test on several comibnaisons, trying with AGP mode
scott_ino2: suokko, under device section in xorg?
suokko: scott_ino2: You can see what is going wrong in xorg.log
suokko: yes
scott_ino2: here's my xorg.log http://pastebin.com/d7d8321b5
riri: thnks for all
mihaita9999: I have ATI Mobility Radeon HD3450 on my laptop. When connected to my TV (S-video to Scart) I can see only in black and white on TV. Please help me, I strived all day
riri: bye
suokko: scott_ino2: You will probably need also Option "AccelMethod" "EXA"
suokko: radeonhd in Ubuntu is old so it does need some extra configuration and might not be as good as radeon
scott_ino2: k i added those two anyting else?
scott_ino2: I mean i suppose i could update, just wouldn't want other stuff to break
scott_ino2: brb let me restart
agd5f: mihaita9999: you need Option "ATOMTvOut" "TRUE" in teh device section of your config
suokko: scott_ino2: https://launchpad.net/~ubuntu-x-swat/+archive/x-updates
agd5f: you may also have to play with the tv standard
mihaita9999: agd5f: where is that Atom TV out? In Catalyst?
legume: mihaita9999: If you are using fglrx then this is the wrong channel #ati. My TV only takes s-video through one of the 2 scarts, so maybe try swapping.
agd5f: mihaita9999: tha's fo the open source driver
mihaita9999: I use Win xp on my laptop (Ubuntu on desktop only). I need laptop for TV. TV is older, sony trinitron, classic Scart
kkuno: lol
scott_ino2: so it appears as though things are very very broken after adding DRI on and accellmethod
scott_ino2: let me get my xorg.log
scott_ino2: http://pastebin.com/d6ed03c72
scott_ino2: if someone could take a look
suokko: scott_ino2: I think you need radeonhd 1.2.5
yangman: scott_ino2: (WW) RADEONHD(0): DRI conflicts with SHADOWFB
scott_ino2: suokko, yeah i was using 1.2.4 I think... what's easiest way to upgrade since ubuntu has older version
yangman: scott_ino2: and 1.2.4 is old. try 1.2.5 or newer
suokko: scott_ino2: https://launchpad.net/~ubuntu-x-swat/+archive/x-updates
scott_ino2: suokko, thanks
scott_ino2: so I can just add the ppa correct?
scott_ino2: and do an update that way
suokko: yes
scott_ino2: k, ty upgrading now
scott_ino2: annnnd restarting x... wish me luck
scott_ino2: should those additions I added earlier hold in xorg
rehabdoll: sweet, ioquake3 runs great on r7xx. only a few artifacts
suokko: scott_ino2: They are what you want so yes
scott_ino2: yeah all appears to be well, here's latest pastebin mind having a look? http://pastebin.com/d37d30cb5
scott_ino2: now I just need 3d ;-) but this can tide me over for a month or two
scott_ino2: appreciate all your help
shane2peru: hey I have the radeon hd 3100 in my toshiba laptop, nothing but headaches, can anyone suggest anything? I currently have Jaunty installed, and Suse11.1
shane2peru: I have downloaded the driver and installed it in jaunty, and it didn't work, I had to uninstall all fglrx stuff to even get it to load x
shane2peru: without the proprietary drivers it runs very hot
shane2peru: is anyone in here actually in here?
shane2peru: hi PSiL0
shane2peru: I need help. :) and no one else seems to be here
PSiL0: what type of help do you need?
shane2peru: I have the dreaded radeon hd 3100 graphics card in my laptop, and cannot seem to get anything to work with it.
shane2peru: if I don't install the proprietary drivers it runs very hot and overheats shutting down
shane2peru: and now I can't even get the proprietary drivers installed
shane2peru: I have tried about several different distros, and PCLinuxOS is about the only one that comes close to working ok
shane2peru: I'm about to throw the laptop out the window.
PSiL0: hmmm, if you want opensource drivers (since you are in the opensource channel), here's a discussion on phoronix on how to get it working
PSiL0: http://www.phoronix.com/forums/showthread.php?t=17603
shane2peru: ok, I'm willing to try anything at this point, what is phoronix?
PSiL0: shane2peru: My 2 year old laptop is running with mobility radeon x2300, a generation lower than yours
PSiL0: shane2peru: a website detailing the latest development in opensource and proprietary drivers for ati, nvidia, etc
shane2peru: PSiL0: yes, I really hate my ati graphics card
PSiL0: shane2peru: I know how you feel.. what distro are you using?
shane2peru: PSiL0: I will be very carefull never to buy another one
shane2peru: Suse and Ubuntu
shane2peru: atm
PSiL0: what method did you use to try to install fglrx for your drive?
shane2peru: but I'm very flexible, I have tried PCLinuxOS, and Debian, and another debian variant
PSiL0: if you want to install fglrx, we should move this discussion to #ati
shane2peru: well, I just removed all fglrx to get the proprietary drivers working, and it was sort of working, but kept running up the cpu and jumping the temp
shane2peru: PSiL0: now I just installed fglrx and borked the system again, about the 3rd time today
shane2peru: PSiL0: I know how to get it back, and remove the proprietary driver, but running the fglrx driver runs it at about 88 degrees C
shane2peru: So, what are my options? Is any distro better than the other for this thing?
PSiL0: 88C with the proprietary drivers?
shane2peru: PSiL0: no with opensource
shane2peru: proprietary drops it down to about 65C
PSiL0: ahh, well the opensource drivers, afaik, does not support powerplay options yet
shane2peru: PSiL0: but I can't seem to get them installed now
PSiL0: hmmm
shane2peru: and opensource overheats the cpu to shutdown point, about 104c
PSiL0: I usually install my fglrx via the terminal
PSiL0: and I have had no problems so far other than slight video tearing when using kwin
shane2peru: I have really grown to hate this card
PSiL0: however, playing avi's and using my s-video out show no problems
shane2peru: it expressly seems to be the hd 3100 radeon card, the 3200 integrated didn't have as many probs
shane2peru: PSiL0: I was trying to process video
PSiL0: shane2peru: hmmm
shane2peru: my wifes box has ati, and have never had this much problem, matter of fact, I never blinked an eye installing
PSiL0: shane2peru: well, I've been using kubuntu the past year or so, and I haven't had any problems shoehorning fglrx (catalyst 9.3) in intrepid and jaunty. However, it could be with your card
PSiL0: shane2peru: could be card specific
shane2peru: PSiL0: it is really driving me crazy, if I could reconcile with my conscience, I would destroy it and collect my warranty. :)
PSiL0: shane2peru: for ubuntu varients, what would do is to download the latest supported proprietary drivers from ati's website
shane2peru: I have used Linux for a while and installed on about a 12 machines, and never had this much problems with any computer ever.
shane2peru: ok, done venting for the moment, I downloaded the latest proprietary drivers and can't get them installed, if I had the older one I would install that
PSiL0: shane2peru: build the packages for your distro and install it via that way
shane2peru: PSiL0: it worked about a month ok
shane2peru: PSiL0: what do you mean?
shane2peru: about a month ago, I meant, and I installed something else trying to be able to process my videos without over heating
PSiL0: shane2peru: sh ./ati-(drivername).run --listpkg
shane2peru: PSiL0: hmm, I think I remember seeing something about that
shane2peru: can you point me to some documentation on that? I have to go back and remove fglrx stuff because now x won't load
PSiL0: shane2peru: then find your distro, i.e. mine: sh ./ati-blahblah.run --buildpkg Ubuntu/jaunty
shane2peru: PSiL0: ok, mine is the same
shane2peru: booting to terminal now
PSiL0: shane2peru: here's my install script after building those packages: http://pastebin.com/m5fd3bac1
PSiL0: shane2peru: this is for my x86_64 kubuntu install
shane2peru: should be about the same, except I'm runing ubuntu, not kde. :)
PSiL0: shane2peru: yes, it should be the same
shane2peru: PSiL0: so should I leave the fglrx stuff installed??
PSiL0: shane2peru: here's the correct install script: http://pastebin.com/m704c7990
PSiL0: shane2peru: if you are running ubuntu, you should sudo dpkg -P fglrx* xorg-driver-fglrx
shane2peru: PSiL0: ok, Ubuntu is my prefered OS, I have turned to others just trying to get things running right, to no avail
PSiL0: if everything works, then you have to tweak /etc/X11/xorg.conf to get it working just right..
shane2peru: hmm, it didn't like the dpkg -P fglrx* xorg-driver-fglrx
shane2peru: I ran apt-get remove fglrx* and that did it
PSiL0: or just apt-get remove --purge fglrx* to remove all vestiges of it
shane2peru: PSiL0: If I don't purge it is that a problem?
PSiL0: shane2peru: it shouldn't be... I just do it to be thorough.
shane2peru: PSiL0: it is better practice
PSiL0: shane2peru: I guess so, especially when removing a pesky program/driver
shane2peru: ok, package build failed
shane2peru: I should probably remove my old attempt at installing it?
PSiL0: shane2peru: sure, what was the message?
shane2peru: it is complaingin about dpkg-buildpackage
shane2peru: probably need to install build-essentials?
PSiL0: ahhh
PSiL0: shane2peru: https://help.ubuntu.com/community/BinaryDriverHowto/ATI
PSiL0: shane2peru: sudo apt-get install dpkg-dev debhelper libstdc++5 dkms build-essential cdbs fakeroot
shane2peru: ahh, ok, thanks!
PSiL0: after you build your packages, you follow my install script (the correct one) listed above
shane2peru: PSiL0: will do, have to reboot, no web, prehaps I will have gui and web this time and install those pacakges
shane2peru: if not I can download them here and put them on the laptop
PSiL0: shane2peru: yeah, it is probably best..
PSiL0: shane2peru: I have to leave to grab some dinner.. brb in 30mins... pm me if you have any issues after half an hour
shane2peru: PSiL0: ok, thanks, I will probably be here plugging away at this
shane2peru: PSiL0: aaaaaaaahhhhhhhhh, arrrrrgggggg, it overheated trying to generate the packages! hit 105c
shane2peru: ok, took a breather, rebooted with recovery dropped to root shell to try to build packages without the loading x, should keep it running cooler
shane2peru: PSiL0: I don't know how to PM, after installing the packages, rebooted, it runs in low-graphics mode? this was where I was before when installing the proprietary drivers
shane2peru: > PSiL0 testing PM
shane2peru: < PSiL0 testing PM