| === ripps_ is now known as ripps | ||
| === seb128_ is now known as seb128 | ||
| === seb128 is now known as seb128_ | ||
| === seb128_ is now known as seb128 | ||
| apw | anyone know whether there is a how-to on getting different settings (such as keyboard layout) per input device in X ? | 08:56 |
|---|---|---|
| tseliot | apw: yes, you can rely on the name and/or the id of the input device (which you can get from the "xinput list" command). | 09:08 |
| tseliot | as regards keyboard layout you can use the device id | 09:08 |
| tseliot | for example setxkbmap -device 2 -print | 09:09 |
| tseliot | or something like that | 09:09 |
| apw | tseliot, and is there somewhere you add that to get it to persist, as it gets lost on suspend | 09:09 |
| * tseliot hasn't played much with setxkbmap | 09:09 | |
| tseliot | apw: in Gnome I *guess* the gnome-settings-daemon would take care of it | 09:12 |
| tseliot | jcristau: ^^ | 09:13 |
| apw | tseliot, does xmodmap work per device too? | 09:13 |
| apw | that is a poor question ... i want to change the layout, xsetxkbmap is good for that. i also want to swap two keys on the new keyboard, can i do that on a per device basis | 09:14 |
| tjaalton | I don't think there's gui support for it yet | 09:14 |
| apw | happy to use command line for this for now | 09:14 |
| tjaalton | I meant for g-s-d | 09:15 |
| apw | ahh | 09:15 |
| tjaalton | the tools are lacking | 09:15 |
| tseliot | apw: I don't think xmodmap allows you to have per device settings | 09:17 |
| bryce | tseliot, if you're around, there is a conf call with AMD today in about an hour | 16:09 |
| tseliot | bryce: thanks but I don't think I'll be around in an hour. Maybe you can forward the report to me? | 16:11 |
| tseliot | I mean the report on the call | 16:16 |
| bryce | actually I don't really have anything to discuss with them, so was hoping the call would be of more usefulness to you | 16:17 |
| tseliot | bryce: oh, nothing comes to my mind at the moment but maybe I can ask them next time (I'm quite busy with the new house) | 16:22 |
| === mac_v_ is now known as mac_v | ||
| lyhana8 | hi, I try to use ppa-purge, but I´m running a Linux Mint 8 aka Helena and got plenty of error messages | 17:17 |
| bryce | lyhana8, sorry I don't think it is supported on linux mint | 17:19 |
| lyhana8 | here is my log : http://pastebin.com/d53d34889 | 17:20 |
| lyhana8 | LM is ubuntu-based, I think it's only the distro name that trigger the problem | 17:20 |
| Sarvatt | yeah | 17:21 |
| Sarvatt | is it based on a ubuntu release? | 17:22 |
| lyhana8 | yep | 17:22 |
| lyhana8 | http://www.linuxmint.com/about.php | 17:23 |
| Sarvatt | you can try DIST="whatever" ppa-purge xorg-edgers | 17:23 |
| Sarvatt | whatever the one you're using is based on | 17:23 |
| lyhana8 | http://en.wikipedia.org/wiki/Linux_Mint | 17:23 |
| lyhana8 | ok | 17:23 |
| Sarvatt | i meant are those packages from the ubuntu archives instead of linux mint archives, no idea how they handle the dist thing | 17:24 |
| lyhana8 | I added the xorg-edgers repo manually | 17:24 |
| lyhana8 | well doesn´t work : Warning: Could not find package list for PPA: xorg-edgers ppa | 17:26 |
| lyhana8 | how can I do it manually ? | 17:26 |
| Sarvatt | thats cus it removed it last time it failed | 17:26 |
| Sarvatt | try uncommenting it from your sources then apt-get update then try again | 17:26 |
| Sarvatt | you want to sudo apt-get install package/karmic package2/karmic to downgrade manually | 17:28 |
| Sarvatt | or uncomment the edgers ppa in your apt sources and apt-get update to refresh it and try again with DIST="karmic" ppa-purge xorg-edgers | 17:29 |
| tjaalton | bryce: I'll merge the xserver so it's (more) ready for upload once the prereqs are in | 17:30 |
| lyhana8 | Sarvatt: isn´t removing the ppa and reinstalling the xorg-server package enough ? | 17:30 |
| Sarvatt | nope | 17:31 |
| lyhana8 | same error, he look for an helena archive for each package | 17:32 |
| lyhana8 | did the package use lsb_release or something la that to know the context ? | 17:33 |
| bryce | tjaalton, excellent | 17:33 |
| Sarvatt | yeah it did | 17:33 |
| Sarvatt | DIST=$(lsb_release -c -s) | 17:34 |
| lyhana8 | Sarvatt: so it override my DIST ? | 17:35 |
| tjaalton | kees: is libaudit-dev finally in main (lucid) ?-) | 17:35 |
| Sarvatt | did you use sudo somewhere in there or did you pass DIST="karmic" as root? | 17:35 |
| lyhana8 | I use the sudo commadn | 17:36 |
| lyhana8 | I try to remove the ppa version cause I think it make my firefox buggy | 17:38 |
| Sarvatt | just run this command and bypass ppa-purge, http://pastebin.com/d6f09bb34 :D | 17:39 |
| lyhana8 | XD | 17:39 |
| lyhana8 | any idea if your driver can cause issue to firefox (flash) ? | 17:40 |
| Sarvatt | yeah it sure could | 17:41 |
| lyhana8 | or this ppa : launchpad-weyland/xserver-nobackfill/ubuntu | 17:41 |
| lyhana8 | oh~ | 17:41 |
| lyhana8 | Sarvatt: downgrade ok, i'll logout later. Thanks | 17:44 |
| mac_v | hmm , even with kernel .32-6 ATI KMS performance is slow :( ... | 17:44 |
| mac_v | anyone know how to effectively measure the fps/speed reduction with KMS ? | 17:47 |
| * mac_v was told glxgears isnt a good indicator | 17:49 | |
| tjaalton | try openarena or similar | 17:50 |
| mac_v | hehe , i have to learn to play a new game :) | 17:51 |
| mac_v | has been "on the wagon" for nearly 2yrs... ;) | 17:52 |
| tjaalton | or run phoronix-test-suite | 17:52 |
| mac_v | ah , great , ok | 17:52 |
| mac_v | tjaalton: basically i wanted to help in testing , will those results help , or rather are they needed/essential? | 17:54 |
| * mac_v realizes he asked the questions first :) | 17:55 | |
| tjaalton | we know it's slower | 17:57 |
| tjaalton | phoronix has an article about it | 17:57 |
| bryce | what is launchpad-weyland ? Did someone package wayland? | 17:59 |
| tjaalton | nah, the author is mathias weyland :) | 18:00 |
| bryce | mac_v, what's needed is to locate patches upstream that haven't filtered down to us that sound like they would improve performance, patch your driver, and test *that* | 18:01 |
| bryce | mac_v, and then if you find something which looks really good, flag it for our attention so we can look into incorporating it | 18:01 |
| * mac_v is unfortuantely not good/technical enough to help bryce with patching and testing :( | 18:02 | |
| * mac_v is more a noob , who can test ppa and whine when things are broken ;p | 18:02 | |
| bryce | heh, we have an oversupply of that at the moment ;-) | 18:03 |
| mac_v | lol ;) | 18:03 |
| bryce | one other thing we could use is regular performance testing | 18:05 |
| bryce | do an upgrade each day to latest lucid, run the test suite, plot the data point on a graph | 18:06 |
| bryce | and then make the graph available to us | 18:06 |
| bryce | this would enable us to spot performance regressions (or improvements) faster | 18:06 |
| Sarvatt | i'm surprised how good gallium is right now on i915 at least, simple things like qgears2 and glxgears are 2x faster and openarena is within 10% of the normal driver. the old build i did of it in july is 4-6x slower in the same things (or didnt even run them at all for the most part) | 18:09 |
| mac_v | bryce: that i can do :) ... right now , what i have done is I'm running the Karmic with lucid kernels , in a few days once alpha 1 is released I'll fuly install Lucid on another partition and start doing that... | 18:11 |
| Sarvatt | wish I knew of a way to make phoronix-test-suite not run some tests 3x so I wouldnt have to babysit it and manually exit the 2nd and 3rd runs, takes out my machine for 3 hours running openarena and smoking guns tests when i just want a rough idea of the speed | 18:19 |
| bryce | Sarvatt, maybe yank out just that one test and package it to run in isolation? | 18:40 |
| tjaalton | bryce: looks like the "remaining changes" -part of the xserver changelog was a mess, partly because of me :) | 19:39 |
| tjaalton | it had obsolete information and nothing new, because I haven't been updating it when merging | 19:40 |
| tjaalton | the diff, apart from patches, is pretty small | 19:40 |
| virtuald | I thought the random dpms offs were fixed | 19:43 |
| bryce | virtuald, should be yeah | 19:52 |
| bryce | tjaalton, ok | 19:52 |
| bryce | tjaalton, yeah this would be a good opportunity to do some housecleaning on xserver | 19:52 |
| bryce | btw, I revamped the Xorg.0.log timing patch | 19:53 |
| bryce | been testing it on my laptop, I can check it in, although I don't think there's much need to enable it | 19:53 |
| tjaalton | ok, nice | 19:55 |
| virtuald | bryce: i'm on edgers now, and a while after resume from ram my monitors turned off. i have an rv570. | 20:09 |
| bryce | virtuald, let me know if you find a patch that fixes it | 20:14 |
| bryce | although I'm more focused at the moment on what's in lucid than what's in edgers, but I'm open to looking at patches either way | 20:14 |
| tjaalton | anyone else have issues loading http://cvs.fedoraproject.org on karmic (firefox), with intel gfx or other? | 20:19 |
| tjaalton | it makes ffox take 100% cpu and takes several minutes.. | 20:20 |
| tjaalton | umm, http://cvs.fedoraproject.org/viewvc/devel | 20:21 |
| virtuald | i don't know where or how to find such a patch, and i'll try to remember to try to reproduce on lucid | 20:38 |
| tjaalton | bryce: should we still autoload poulsbo? | 20:39 |
| bryce | tjaalton, do we? thought that was defaulted to vesa in our xserver | 20:40 |
| tjaalton | bryce: so it seems. anyway, the patch fails now | 20:40 |
| tjaalton | hmm, done differently upstream | 20:41 |
| tjaalton | } else if (dev->device_id == 0x8108) { | 20:41 |
| tjaalton | break; /* "hooray" for poulsbo */ | 20:41 |
| tjaalton | that should fall back to vesa then | 20:42 |
| bryce | so that causes it to fall back to vesa? | 20:42 |
| tjaalton | yes | 20:42 |
| jcristau | yes | 20:42 |
| bryce | should be fine | 20:42 |
| tjaalton | echo :) | 20:42 |
| bryce | I'd defer to alberto, he's more clued in on what the OEM team wants | 20:42 |
| bryce | but they can patch it themselves for the psb hardware they ship | 20:42 |
| tjaalton | they can change it later if needed | 20:42 |
| tjaalton | that's the fourth patch I've tried to apply, and fourth to delete :) | 20:43 |
| bryce | I don't think we have a -psb at present that'll work on lucid with xserver 1.7 anyway. probably needs to go to vesa. | 20:43 |
| tjaalton | two of the patches were from fedora, and dropped there. I doubt we should be shipping stuff the "upstream" doesn't | 20:44 |
| bryce | tjaalton, just please give a mention of the patches being dropped and why. Sometimes in troubleshooting a regression I have to figure out why a given patch was dropped so it would be helpful info. | 20:44 |
| tjaalton | bryce: sure | 20:44 |
| jcristau | i haven't looked at what f12 has on top of 1.7.x recently. | 20:45 |
| jcristau | just cvs upped, though :) | 20:45 |
| jcristau | ah there's the glx 1.4 thing | 20:47 |
| jcristau | should probably steal that | 20:48 |
| bryce | sconklin, btw do you know if the kms pageflip patch is on the radar for the 2.6.32 kernel? http://www.phoronix.com/scan.php?page=news_item&px=NzY5OA | 21:04 |
| sconklin | bryce: I wasn't familar with it, I can see | 21:05 |
| bryce | if not, would be nice to have it in a ppa for testing purposes | 21:05 |
| sconklin | bryce: according to the page you linked, it didn't make .32 - I'll try dropping it on top of our karmic and see what happens. | 21:07 |
| jcristau | right it's queued for .33 upstream | 21:08 |
| sconklin | bryce: I'm just now uploading a karmic kernel + the moblin speedup test patches to a PPA for Alberto to have a look at. | 21:08 |
| bryce | sweet | 21:09 |
| bryce | sconklin, I tested apw's karmic kernel with those patches, but on an SSD didn't spot changes to X's boot time (didn't look at kernel boot time) | 21:09 |
| bryce | I'll test your kernel too | 21:10 |
| sconklin | and andy is doing research with the sme patches for Lucid. Did you see Andy's email to kernel-list about accounting for the differences in moblin kernel speed? | 21:10 |
| sconklin | He's accounted for all of the differences without the patches we're looking at | 21:10 |
| apw | bryce, thats odd as tseliot claimed 2s with some x patches included | 21:10 |
| sconklin | bryce: apw's kernel doesn't have all the patches that mine will, but they still may not make much difference | 21:13 |
| bryce | sconklin, ok no I'm not on the kernel list | 21:19 |
| bryce | apw, he thought it might be due to me running on SSD and him not on SSD | 21:19 |
| apw | bryce, may be so indeed | 21:19 |
| Sarvatt | do you use compiz jcristau? i just wrote down the time of every flicker for the past 30 minutes and when i turned on compiz halfway it went from having 80 seconds exactly between flickers alot of the times to 20 seconds | 21:19 |
| bryce | apw, my X boots in <2 sec on SSD already, so a 2 sec savings would be pretty astounding ;-) | 21:19 |
| apw | bryce, yeah it would | 21:20 |
| bryce | I've a i945 laptop with a much longer X boot, I could try on that | 21:20 |
| bryce | but since the target hardware is SSD, I'm sort of more interested in that case | 21:20 |
| apw | yep. testing here on the same | 21:20 |
| jcristau | Sarvatt: no | 21:21 |
| sconklin | bryce: basically, turning off initramfs and ISA-PNP results in a kernel init time of about 0.8S | 21:21 |
| bryce | mm | 21:21 |
| jcristau | Sarvatt: so that seems to match what you see | 21:22 |
| Sarvatt | metacity compositing maybe? i didnt test that | 21:22 |
| jcristau | no compositing for me. plain old fluxbox. | 21:22 |
| jcristau | oh i misread what you said | 21:22 |
| jcristau | sorry | 21:22 |
| apw | bryce, you done any testing on lucid yet.. | 21:22 |
| jcristau | thought you said it went 80 -> 20s by turning compiz *off* | 21:23 |
| * apw thinks he just had a hang on there on the 10v | 21:23 | |
| apw | bryce, oh i forgot ... did you note ATI KMS is now on by default | 21:23 |
| Sarvatt | its weird, if i stared at gnome terminal with nothing else open it'd have chunks of time where it'd go off 80 seconds exactly, if i opened a new program it'd almost always trigger it not long after | 21:24 |
| bryce | apw, yeah I've got both my laptops on lucid now. | 21:24 |
| apw | brave :) | 21:24 |
| jcristau | Sarvatt: jesse suspected the fifo watermarks settings. i haven't had time to play with that yet. | 21:24 |
| bryce | I ran into an issue where upstart isn't starting gdm, unless I remove some line about kernel tty7 from grub.cfg | 21:25 |
| apw | bryce, i think i saw a flicker on my 10v too, like a jump left an inch for a frame | 21:25 |
| bryce | filed a bug | 21:25 |
| apw | bryce, ouch | 21:25 |
| bryce | apw, not so brave... I still keep my desktop on karmic for now ;-) | 21:25 |
| jcristau | apw: what chip is the 10v? | 21:25 |
| bryce | but I'm not traveling for the next couple months so if laptop breaks it's not a problem | 21:25 |
| bryce | jcristau, i965 iirc | 21:26 |
| apw | 00:02.0 VGA compatible controller: Intel Corporation Mobile 945GME Express Integrated Graphics Controller (rev 03) | 21:26 |
| bryce | ah 945 | 21:26 |
| jcristau | could be the same bug Sarvatt and i are seeing | 21:26 |
| Sarvatt | thats what I have, i get the flickering after a suspend/resume cycle every time on 2.6.32 | 21:26 |
| apw | bah ... and we are close to final on the kernel | 21:27 |
| * jcristau gets it after boot, no suspend needed | 21:27 | |
| Sarvatt | also i cant suspend with a SD card mounted on 2.6.32 | 21:27 |
| Sarvatt | and its not a ubuntu patch problem, happens when i compile a vanilla kernel | 21:27 |
| apw | Sarvatt, we had a patch for that in karmic and i think its still in lucid | 21:28 |
| apw | may not work of course | 21:28 |
| Sarvatt | hmm wonder if i suspend without a dpms off from pm-utils if it'll still flicker | 21:29 |
| Sarvatt | or lock up on the solid color | 21:30 |
| apw | Sarvatt, on the SD card thing, it hangs hard on suspend yes? | 21:30 |
| Sarvatt | yeah, the screen wont turn off but the system freezes | 21:30 |
| apw | as you can reproduce it, perhaps you could boot with 'no_console_suspend' on the kernel commad line, and try a suspend from VT1 using sudo pm-suspend | 21:30 |
| apw | and see if you get a panic | 21:30 |
| Sarvatt | i'll do that right now | 21:31 |
| Sarvatt | same issue, /var/log/pm-suspend.log shows a success but it was frozen with the screen still on and unresponsive | 21:38 |
| Sarvatt | http://paste.ubuntu.com/333422/ | 21:40 |
| sconklin | is it worth sending announcements of test kernels like the moblin patches on Karmic to the ubuntu-x list? | 21:40 |
| tjaalton | bryce: preliminary merge done, there are a bunch of patches that need a review and/or a refresh | 22:04 |
| tjaalton | we should probably replace patch 140 by disabling acpi altogether | 22:06 |
| tjaalton | since it's not used by anything | 22:07 |
| tjaalton | http://cvs.fedoraproject.org/viewvc/devel/xorg-x11-server/xserver-1.6.0-less-acpi-brokenness.patch?revision=1.1&view=markup | 22:07 |
| tjaalton | one way to do it :) | 22:07 |
| Sarvatt | hmm, looking like a good excuse to go splurge on an SSD and just shutdown instead of suspending :D | 22:35 |
| jcristau | tjaalton: patch 140 is obsolete anyway | 22:35 |
| jcristau | (commit 7a05c8b1e70680ddd3b3e09ad448788f8d70a428) | 22:36 |
| Sarvatt | 140 174 178 179 180 fix-dga-removal 181 182 183 185 187 Add-libgcrypt-as-an-option-for-sha1 were all upstream last i looked, but i was looking at master not 1.7 branch | 22:41 |
| jcristau | Sarvatt: the libgcrypt one is only on master | 22:42 |
| === seb128_ is now known as seb128 | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!