* RAOF should just bite the bullet and add an interface for drivers to say “Thanks X, but kms has already set up a perfectly acceptable mode. Use it. Kthxbye’ | 04:44 | |
ScottK | Any experience on how "Integrated Intel® 3000 HD" is working on Natty? | 05:25 |
---|---|---|
* ScottK is considering a laptop purchase. | 05:25 | |
bryceh | ScottK, HD 3000 == sandybridge | 05:45 |
ScottK | bryceh: That's bad? | 05:45 |
bryceh | ScottK, well you've probably seen all the discussion bouncing around about it lately | 05:46 |
ScottK | bryceh: I've managed to avoid it. | 05:46 |
bryceh | ScottK, synopsis is it's currently a bit buggy, but fast. Stick it in a closet until oneirik is out and you'll probably love it at that point | 05:47 |
ScottK | Ah. | 05:47 |
bryceh | maybe we'll get all the bugs sorted by N; we'll see | 05:47 |
bryceh | +1 tester can't hurt | 05:48 |
bryceh | ScottK, sandybridge is one of sarvatt's main foci right now, he knows all the patches what's worth having for SB | 05:48 |
ScottK | Considering the not so great state of video on my current i945GM laptop, how buggy is buggy? | 05:49 |
bryceh | standard Intel level of buggy | 05:49 |
bryceh | freezes once and a while, etc. | 05:49 |
ScottK | The current i945 doesn't freeze, but it has all kinds of wierd display effects. | 05:50 |
ScottK | (the one I have anyway) | 05:50 |
ScottK | Perhaps Sarvatt will have an opinion then. | 05:50 |
bryceh | the main sb bug I know about is when running amd64 + ia32-libs is a bad combo | 05:52 |
bryceh | so sticking i386 should avoid that issue (which might be resolved by release if ia32-libs gets updated) | 05:52 |
bryceh | ScottK, besides you know you'll be running OO in a few weeks, and that'll probably have a lot better SB support anyway | 05:54 |
ScottK | I run i386 on desktops, so not a problem. | 05:54 |
hyperair | sb levelclear -level clientcrap,crap,joins,parts,quits,nicks,clientnotice | 05:57 |
ScottK | bryceh: Thanks for the info. | 06:07 |
RAOF | Aah! *That's* why my IRC bouncer occasionally dies! The screensaver cycles through to gltext and gltext OOMs. | 06:28 |
bryceh | hah | 06:28 |
RAOF | ScottK: Have I asked you whether libdrm 2.24 resolves your problems? | 06:28 |
ScottK | RAOF: You haven't, but my current laptop is still on Maverick. | 06:29 |
ScottK | Netbook, on Natty, seems to be working reasonably well. | 06:29 |
RAOF | Oh, my. Maverick? So long ago :) | 06:29 |
ScottK | Yeah. | 06:29 |
ScottK | Well I've got a major project due to the customer on Thursday. Once it's done I'll probably take the plunge and upgrade the laptop too. | 06:30 |
codemagician | hi bjsnider | 13:58 |
bjsnider | hi | 13:58 |
codemagician | you might remember me, the guy who switched from Nvidia to the Nouveau drivers | 13:59 |
codemagician | im still getting the constant hanging on the x server | 13:59 |
codemagician | http://pastebin.com/TMmggJnN | 13:59 |
codemagician | "EQ overflowing. The server is probably stuck in an infinite loop." | 13:59 |
codemagician | i wondered how I might go about finding a solution for this? And also, is the problem and X problem or a driver problem? | 14:00 |
tjaalton | there is no solution, other than disabling the acceleration | 14:00 |
codemagician | tjaalton, is that in response to my question? | 14:00 |
tjaalton | yes | 14:00 |
codemagician | tjaalton, I don't believe I have anything like that enabled | 14:01 |
tjaalton | it's accelerated by default | 14:01 |
codemagician | It "Visual Effects" menu it is set to "None". is this the same thing? | 14:01 |
tjaalton | no | 14:01 |
codemagician | how to I go about disabling the acceleration? | 14:02 |
tjaalton | put "nouveau.noaccel=1" to kernel options | 14:02 |
codemagician | i've never done anything like this, which command will put this option to the kernal? | 14:02 |
codemagician | i am familiar with the command line | 14:02 |
codemagician | might be worth noting I had the same hanging problem with the Nvidia drivers also | 14:03 |
tjaalton | GRUB_CMDLINE_LINUX in /etc/default/grub, then run update-grub2 i guess | 14:03 |
tjaalton | yes, drivers have bugs that seem to be the same, but arent' | 14:04 |
codemagician | tjaalton, are you familiar with this as a bug? | 14:05 |
tjaalton | with nouvea, yes | 14:05 |
tseliot | codemagician: in nvidia you can try adding this option: Option "NoAccel" | 14:05 |
tjaalton | +u | 14:05 |
codemagician | whats +u ? | 14:06 |
tjaalton | me fixing a typo | 14:06 |
codemagician | aha | 14:06 |
bjsnider | tseliot, i tested the feature that is supposed to stop the nvidia-installer if the pre-install script is present. it doesn't stop the installer, it only adds a yes/no prompt | 14:07 |
codemagician | so in Nvidia that option is in the same place? | 14:07 |
tjaalton | nvidia doesn't support KMS, so it's different | 14:07 |
tseliot | bjsnider: really? I don't think that was the plan... | 14:07 |
codemagician | KMS? | 14:07 |
tjaalton | kernel modesetting | 14:07 |
tseliot | kernel mode setting | 14:07 |
bjsnider | tseliot, right, i thought you could bug aplattner about it | 14:08 |
codemagician | is this is a type of way of configuring the devices via the kernal itself? | 14:08 |
tseliot | bjsnider: what's the exact text that you're getting? | 14:08 |
tjaalton | codemagician: you need to tell the nouveau kernel module it's options before it's loaded.. | 14:08 |
bjsnider | tseliot, i can't give you the exact text without testing it again, but it was something like "pre-install script error: continue yes/no?" | 14:09 |
bjsnider | yes is highlighted so hitting return continues | 14:09 |
bjsnider | in effect all it does is make the user hit return an extra time | 14:09 |
codemagician | tjaalton, so to edit the kernel mode settings, is there an easy command tool to use? | 14:10 |
tjaalton | whatever text editor you use | 14:10 |
codemagician | tseliot, i don't seem to even have a /etc/default directory | 14:11 |
tseliot | bjsnider: I should definitely talk to Aaron about this | 14:11 |
bjsnider | tjaalton, he's 100% n00b | 14:11 |
codemagician | tjaalton, , i don't seem to even have a /etc/default directory | 14:11 |
codemagician | tseliot, sorry wrong person | 14:11 |
tjaalton | codemagician: and this is ubuntu? doubt it | 14:11 |
bjsnider | tseliot, yes, i thought you'd said you wanted it to immediately exit the installer, which is not at all what's happening | 14:11 |
codemagician | tjaalton, using 10.10 | 14:11 |
tjaalton | bjsnider: got it ;) | 14:11 |
codemagician | tjaalton, sorry | 14:11 |
codemagician | tjaalton, i had some weird character in the line | 14:12 |
bjsnider | tjaalton, he's a mac refugee | 14:12 |
codemagician | ha ha | 14:12 |
tseliot | bjsnider: right, that was the plan. Only passing an additional parameter would have caused the installer to ignore that error | 14:12 |
tjaalton | codemagician: 'nano' is probably a good editor candidate then.. | 14:12 |
codemagician | tjaalton, I see GRUB_CMDLINE_LINUX="" at the moment | 14:13 |
tjaalton | codemagician: so put nouveau.noaccel=1 inside the quotes | 14:13 |
codemagician | tjaalton, should I insert the nouveau.noaccel=1 in there? | 14:13 |
codemagician | tjaalton, right I ran update-grub2 | 14:14 |
codemagician | tjaalton, so now when I reboot how to I confirm that this option took effect? | 14:14 |
tjaalton | yep | 14:15 |
codemagician | tjaalton, it there a visual cue to show that acceleration is switched off? | 14:15 |
tseliot | you'll notice... | 14:15 |
tjaalton | right.. | 14:16 |
codemagician | will I cry as my windows take 10 mins to draw ;-) | 14:16 |
codemagician | so, this issue is a driver problem or within x? | 14:17 |
tjaalton | driver | 14:17 |
bjsnider | officially, there are no x driver bugs. x is bug-free | 14:17 |
codemagician | will there be a driver fix? | 14:17 |
tjaalton | eventually, maybe | 14:18 |
codemagician | is this issue across all Nvidia chipsets? | 14:18 |
tjaalton | not aiui | 14:18 |
bjsnider | codemagician, i'm using the same chip you are and i have no issues | 14:18 |
codemagician | do you think its my card, ASUS EN210 Silent | 14:19 |
codemagician | Can other hardware effect this too? I have a Intel X-25 SSD disk and use Marvel controllers for 6GB interal Western Digital 1TB drive. Are those things all separate? | 14:20 |
codemagician | *6GB SATA III | 14:20 |
tjaalton | separate | 14:20 |
bjsnider | my board also has marvell controllers | 14:20 |
codemagician | does the pastbin output prove its a driver issue? | 14:21 |
codemagician | http://pastebin.com/CKK6Zzed | 14:23 |
codemagician | here is the full X log on the session that hung | 14:23 |
codemagician | any suggestions or other things amiss would be welcome | 14:24 |
codemagician | back in a moment, I'll reboot and try the non-accelerated gfx machine | 14:25 |
codemagician | back, but can't say I noticed any difference so far | 14:27 |
codemagician | tjaalton, should I try running a particular type of application to check the acceleration is off? | 14:28 |
tjaalton | dmesg | grep nouveau | 14:29 |
mdeslaur | tjaalton: so, is it too late to get accel quirked off by default for the problematic nouveau chipsets for natty? | 14:33 |
codemagician | BOOT_IMAGE=/vmlinuz-2.6.35-28-generic root=UUID=5acb5583-1369-4c00-b1b2-b059a2ad0fe5 ro nouveau.noaccel=1 quiet splash | 14:33 |
codemagician | http://pastebin.com/zLuEL517 | 14:33 |
codemagician | (output from dmesg | grep nouveau ) | 14:34 |
codemagician | tjaalton, how does it look? | 14:37 |
tjaalton | codemagician: like it should work | 14:38 |
codemagician | tjaalton, ok... I use it. Usually it hangs my machine 2 times per day | 14:38 |
tjaalton | mdeslaur: i guess it's not too late, but I'd have to ask.. apw? | 14:38 |
codemagician | tjaalton, fingers crossed | 14:38 |
bjsnider | ricotz, i''m not clear on smething. if i want to use all of gnome 3 in natty + the ppa after both are released stable, is that going to be possible? or is it not known yet? | 18:22 |
kklimonda | bjsnider: what do you mean by possible? | 18:27 |
bjsnider | right now it apparently doesn't work due to instability | 18:28 |
kklimonda | yeah, it's not that stable but I think it's mostly an upstream problem. | 18:29 |
bjsnider | i thought they said natty woudln't have everything that gnome 3 needs | 18:29 |
kklimonda | we may not have everything updated, but the goal is to update as much as possible in the ppa | 18:29 |
kklimonda | I do remember raof saying that we won't get something in X that gnome-shell requires, but it hasn't been yet released in upstream and he has no idea why are g-s folks using it ;) | 18:30 |
kklimonda | RAOF: btw, do you have any leads on nvidia making everything use tons of memory? | 18:31 |
ricotz | bjsnider, hi, the gtk3 releated packages in natty arent used for the official release, and if you are using the gnome3 packages will break gnome2 desktop currently | 18:39 |
ricotz | gnome3 ppa* | 18:39 |
bjsnider | what about by the end of april? | 18:40 |
ricotz | i am not sure if it is possible, probably not | 18:41 |
bjsnider | so then what happens to my system if i fully install the gnome 3 ppa when it and natty are stable? | 18:42 |
bjsnider | will it work in that situation if i choose to use only the gnome 3 desktop? | 18:43 |
ricotz | yeah only using the gnome3 desktop should work and i hope it gets stable ;) | 18:43 |
ricotz | but currently there are still some packages missing | 18:44 |
ricotz | so you should be careful with testing if you need a stable system | 18:45 |
ricotz | for me it is running with a usable stability (working around the problems) | 18:45 |
ricotz | kklimonda, are you member of gnome3-team? | 18:48 |
kklimonda | ricotz: not really, but I've been asking similar questions recently. | 18:49 |
ricotz | kklimonda, ok ;) | 18:49 |
kklimonda | I've tried working on it | 18:50 |
kklimonda | was interested in updating some packages | 18:50 |
kklimonda | but I can't use it ;/ | 18:50 |
kklimonda | something between gnome3 and the beta nvidia drivers make it much more unstable then it should be. | 18:50 |
ricotz | what packages? | 18:50 |
kklimonda | ricotz: eog, libseed and libpeas | 18:51 |
ricotz | i am using nvidia blob with xorg-edgers ppa and gnome3 ppa | 18:51 |
kklimonda | ricotz: i've tried to fix ftbfs. | 18:51 |
kklimonda | ricotz: yeah, the new nvidia blob seems to work for some, just not for me ;) | 18:51 |
ricotz | kklimonda, ah, ok, this is in progress | 18:51 |
kklimonda | I can't even use compiz recently | 18:51 |
ricotz | eog needs still some packaging fixes for introspection | 18:52 |
bjsnider | ricotz, you don't sound confident that upstream is going to make gnome 3 stable | 18:52 |
kklimonda | it does seem like a lot of work to make it stable at this point :) | 18:52 |
kklimonda | ricotz: I'd love to help, I'll give it a shot after nvidia releases new drivers - maybe it will make it work better | 18:52 |
ricotz | bjsnider, not really, it should get stable upstream, but updating gdm and gnome-session to their gtk3 builds would be some effort | 18:53 |
ricotz | kklimonda, you are using 270.30? on what card? | 18:54 |
kklimonda | ricotz: quadro 140M | 18:54 |
kklimonda | (it's 8400M or maybe 8600M) | 18:54 |
ricotz | ok, dont have problems here with a gtx460 | 18:55 |
bjsnider | maybe because your card is like a ferrari to his pinto? | 18:58 |
kklimonda | :D | 18:58 |
kklimonda | bjsnider: you've just made my day :P | 18:58 |
bjsnider | ricotz, when you do a glxinfo does it say your opengl version is 4? | 18:58 |
bjsnider | kklimonda, that was not an accurate analogy though, because your chip is at least workstation quality. it won't explode if you touch it, unlike a pinto | 19:00 |
kklimonda | bjsnider: ahaha, that would be funnier if my chipset wasn't the one from the faulty serie and did not actually explode (or at least stop working) at some random date ;) | 19:00 |
kklimonda | it's a damn time bomb | 19:01 |
bjsnider | well, nvidia's had some engineering problems in the past | 19:03 |
Sarvatt | kklimonda: my 8400m gs system died recently, first symptoms were that I couldn't use compiz with it because it would hang, then I had to use nouveau because the blob would always hang, then it stopped turning on completely :) | 19:07 |
kklimonda | Sarvatt: it doesn't hang and rather crashes | 19:07 |
kklimonda | Sarvatt: but yeah, I'm getting used to the idea that it's dying | 19:07 |
ricotz | bjsnider, yeah, it says 4.1.0 | 19:58 |
bjsnider | cool | 19:58 |
RAOF | kklimonda: Fedora are shipping ajax's pointer barrier stuff, which is what gnome-shell is using. Leads for nvidia memory useage: the blob does something when it's loaded that dirties memory. There's nothing more I can investigate without a decompiler ;). | 20:18 |
bjsnider | RAOF, you mean with gnome 3 on nvidia memory usage would be huge? | 20:19 |
RAOF | bjsnider: No. The problem is that the blob causes anything that links to it to use a small but significant amount of non-sharable memory. With cairo-gl, *anything* GTK based linked to libGL, so everything had increased memory usage. | 20:20 |
RAOF | bjsnider: Unless gnome-shell requires GTK to link to GL you're not going to see the same problem. | 20:21 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!