[07:35] good morning [07:36] Good morning, intrepid psb explorer. [07:39] :) [07:39] more like "sucker who loved the specs of a netbook and failed to check driver status before buying it" [10:05] anyone still working on gma500? [10:06] hi... not testing atm [10:06] installed bernardo's debs last night and X loads.. [10:06] yesterday was able to get x with bernardo deb [10:06] but crashes almost directly.. [10:06] yes [10:07] did see the desktop, but the top panel was misformed [10:07] the same here [10:07] and back to gdm [10:08] do yo know how to build debug packages? [10:10] libdrm and lib-va have debug package.. don't know how to get for the other packages [10:10] ok.. I'll check those debian/ files to see if I can find out.. [10:11] ok.. good luck! [10:11] let me know [10:16] @fkefer: are you working on gma500? [10:28] outlaw45: sorry i was out of office for a few minutes [10:28] in general yes [10:28] but now right now [10:28] (no network cable to connect my laptop with) [10:28] ook [10:28] it seems as if my box is even harder to get running than most others [10:29] ok... [10:29] as i do not see a thing when i install the driver, just video-mode-change then blank screen [10:29] (also the vesa-1366x768 trick didn't work for me at all) [10:29] the same here.. [10:29] ah, ok [10:29] "good" (sort of ;-) ) [10:29] did get into X last night, but not at the moment.. [10:30] the problem is that all this is such a mess, interconnected with each other that it's everything but easy to track down the problem [10:30] or even just test one module at a time [10:30] yeah indeed, but I don't have the skills to make it better [10:31] me neither, esp as long as i lack the time and lack the understanding of what is _really_ going on there [10:33] yeah me too.. just trying to get some video performance with what we got.. [10:35] RAOF, about? [10:35] same here [10:38] apw: ok, i give up: RAOF? you asking what we're talking about? [10:38] fkefer, nope ROAF is a nick, i was pinging him [10:39] ah, ok, stupid me [10:39] if i'd said ping it might have been less confusing [10:39] looked a little like yet another acronym [10:40] apw: hm, ok, that laugh is on me :-) [10:41] outlaw45: thing is, i didn't get this box for 3d gaming or high-end-apps or even video watching, all i want is a nice box to carry with me and do basic stuff and have network and a console [10:41] now what i'm stuck with is a lightweight brick in white... [10:42] yeah know what you mean.. got a nice tablet pc (Archos 9) and 10.04 is performing much better dan 9.10 [10:42] just want to be able to browse the internet and maybe watch some flash video and things like that [10:43] btw.. even if we get the video to work there's still the issue of the touchscreen, but I'll see if I get to that.. [10:43] and Windows 7 Starter just sucks big time :) [10:44] i know the windows 7 issue on mine too; it came preinstalled with a windows vista; that was so slow, you couldn't even fire up solitaire [10:44] then 9,04 worked like a charm [10:44] 9.10 well enough [10:45] netbook launcher in 9.10 is a bit slow [10:45] 10.04 looks pretty fast actually except for the graphics [10:45] yeah indeed [10:45] can't wait to get the video working :p [10:45] once my workmate leaves his desk for today i'll put my machine back up :-) [10:45] fkefer: at least it's a cheap lightweight brick in white [10:46] but at the moment batteling with connecting to wireless network from the command line.. [10:46] jcristau: *lol* thing is, i could have spent ~70€ more and gotten the MSI X340 (no the MSI X320) and would NOT face this issue now [10:46] as there's a _real_ processor in it [10:46] no atom/gma500 thing [10:47] I do think the gma500 is a pretty good chipset, it just sucks it's created by powervr.. [10:47] fkefer: yeah. bad choice there. [10:47] in general intel has pretty good linux support [10:48] outlaw45: it doesn't matter how good the chipset is if there's no driver. [10:48] that's what i thought too, and at the time i bought it the device was supported [10:48] usually it's rather the other way round, you buy something new and after some time it just works [10:48] psb was never really supported [10:48] @jcristau: good point ;) [10:49] jcristau: true, but it worked in ubuntu first, then there was packages up that worked - now i'm in with those people trying to get packages... [10:50] it worked in ubuntu because canonical thinks it's a good idea to ship unsupportable drivers. joke's on them, i guess. [10:50] or on you :/ [10:53] outlaw45: have you found debug packages? [10:53] no not yet [10:53] still trying to set up ssh [10:54] but can't connect to wireless so can't install the package... [10:54] it could really help a ssh connection [10:56] it's a shame that i can't set up my machine here [10:56] my router isn't setup to allow internet connections... [11:03] I was trying to debug last night before I left [11:03] but I couldn't make it break in gdb [11:04] it was hanging when run within gdb [11:04] and the stacktrace in the log files is only of limited use [11:04] too bad [11:05] I'm almost convinced there's something wrong in the kernel space parts though [11:08] damn.. can't get the wireless up... [11:09] awoodland: any ideas how to unit test this? i don't even know what exactly they're here for!# [11:10] it's pretty ugly stuff to test [11:10] and I haven't seen any automated testing for it [11:11] I'm trying to eliminate individually kernel module(s), 2d driver, 3d driver, in that order [11:11] but the boundaries aren't terribly clean [11:11] and at some point I should do my paid work :) [11:34] awoodland: for one yes, i should get back to my paid work too [11:35] awoodland: second it's a shame that tha boundaries are blurry [11:46] well I've got a valgrind log now with some errors in it and line numbers for psb bits [11:46] but they mostly seem to be ioctl params [11:46] which probably aren't the problem [11:50] awoodland: then you're further than us all still... [11:52] I can share the log if people want [11:53] all I did was manually install the un stripped shared objects to replace the stripped ones provided in the .debs [11:53] should probably do a -dbg package with dh_strip [11:53] and when it runs in valgrind [11:53] it seems to "work" except that I get no output on the screen [11:54] but X is clearly running and taking mouse/keyboard inputs [12:17] awoodland: sorry, work is distracting :-) [12:17] awoodland: what do you mean with mouse/keyboard inputs are taken? [12:17] you see in the code, but nothing on the screen? [12:18] that makes it even _harder_ to test, right? [12:18] yeah I can see clear evidence that the server is running and responding [12:18] but not visual evidence :) [12:25] awoodland: ouch... [12:26] do we know if the kernel modules work as expected? how could we test that? start bottom up... [12:32] I was trying to test just framebuffer console stuff [12:33] on my 751h at least if I use no_fb=0 when loading the psb module [12:33] then after X starts the text consolse are corrupted too [12:33] and I meant to have a fiddle [12:33] and see if they worked ok without X in the way [12:34] (in karmic no_fb=0 worked fine) [12:34] (for me at least) [12:35] Going for lunch/meetings. What I have is at: http://users.aber.ac.uk/ajw/valgrind_x.log [12:35] I'm pretty sure it'st the ones after the last "select returned 0" that matter [12:36] the segfault at the end looks like it's in a clean up routine [12:37] my money at the moment is on DRICreateContext or DRICreateDrawable [12:37] anyway I'll see what I can find out later [13:05] am trying to get my wireless working for a few hours now [13:06] still no go.. [13:08] seems like the wireless kernel module conflicts with psb kernel module.. [13:08] but I'm not sure... [13:10] outlaw45: shame that is [13:11] i always used wired connection so far, so i don't know about my rtXXXXX crap [13:11] no wired connection present :'( [13:23] hm [13:23] none here either, and no wireless, i'm still @ work [13:47] I'll try to find a usb to ethernet converter.. [13:47] back in a bit.. [14:44] that could work... === Bernardo|away is now known as Bernardo [15:19] hi [15:21] hi [15:21] hi [15:23] what do you have to do to create debug packages? [15:23] how are things going with the gma500? I see awoodland did some progress [15:23] with your packages I'm able to get in to gdm [15:24] outlaw45: check the debian/rules file for libdrm, I was going to copy the install part to the debian/rules of xserver-xorg-video-psb [15:24] of course, changing the dh-strip line accordingly [15:24] and adding the relevant .files, .install, etc. under debian (like we have in libdrm-poulsbo) [15:25] if you can do that, I'd appreciate [15:26] dh_strip -plibdrm-poulsbo1 --dbg-package=libdrm-poulsbo1-dbg [15:26] that line? [15:27] I'll try.. [15:27] yes, that one [15:27] but you need to copy all the other install stages [15:31] line 67 - 83 from libdrm/debian/rules? [15:34] wait [15:36] 65-83 if you want to include the comment,but you should also edit the binary-indep and the binary: lines in the psb driver to look like those in libdrm [15:37] ok.. [15:37] but a line like 'dh_install --sourcedir=debian/tmp -Xlibdrm.la --list-missing' [15:37] should it just be 'dh_install --sourcedir=debian/tmp' [15:37] ? [15:40] probably... :) [15:40] but in the psb driver the install seems to be done under install-stamp, so I'd edit that one, adding the lines 67-83 [15:41] I really am out of my depth here, need to check online the rules file syntax [15:50] building xserver-xorg-video-psb [15:50] now.. [15:51] have to switch machines.. tried compiling on my laptop but don't have all the dependencies [17:21] Bernardo, I am following your work on poulsbo driver for Lucid... how far are you guys? [17:22] I am trying to get at least xorg working using vesa mode.. still stumped. [17:36] can't build xserver-xorg-video-psb [17:36] fails at configure with: [17:36] checking for DRI... configure: error: Package requirements (libdrm >= 2.0 xf86driproto) were not met: [17:37] No package 'libdrm' found [17:37] you need to install libdrm-dev [17:37] libdrm-dev and x11proto-xf86dri-dev [17:37] are installed [17:38] libdrm-poulsbo-dev also [17:40] do you have a /usr/lib/pkgconfig/libdrm.pc [17:41] no.. [17:41] but libdrm.pc.libdrmdev does [17:42] hm, something seems to have diverted it [17:42] ckeck /usr/lib/pkgconfig/libdrm.pc [17:42] eh [17:42] dpkg-divert --list /usr/lib/pkgconfig/libdrm.pc* [17:43] diversion of /usr/lib/pkgconfig/libdrm.pc to /usr/lib/pkgconfig/libdrm.pc.libdrmdev by libdrm-poulsbo-dev [17:44] but libdrm-poulsbo-dev doesn't actually ship a libdrm.pc? [17:44] (check with dpkg -L) [17:45] it puts it in /lib/pkgconfig/libdrm.pc [17:46] oh, that's broken [17:46] does it put anything else in /lib ? [17:46] symlink it to /usr/lib/pkgconfig/libdrm.pc [17:47] /lib/libdrm.a, /lib/libdrm.so [17:47] that looks like a very broken package [17:48] hi outlaw45, ilmari, I am also trying to get psb to work, but first attempting to get vesa working [17:50] what's the mode switch to start installation in vesa mode? [17:50] i tried xdriver=vesa, -vesa.. no good. [17:52] vesa is loaded if with an empty xorg.conf on my system [17:53] outlaw45, i think I might to isntall 915 module, as vesa doesn't seem to find proper modelines [17:53] outlaw45, so you didn't have to put any install boot switches? thanks. [17:53] no nothing.. [17:54] seriously, poulsbo is the worst driver problem I had since my difficulty with Trident drivers years ago. [17:55] * ilmari made damn sure it didn't have poulsbo when he got his new laptop [18:02] outlaw45: there is a bug in our scripts since we changed the libs to /lib to follow libdrm [18:03] it also copies lidrm.pc to /lib/pkgconfig instead of the proper place - /usr/lib/pkgconfig [18:03] yeah found it :p [18:03] just compiled a debug package [18:03] checking if it works at the moment [18:03] sorry about that, we found it yesterday but I haven't yet had the time to do any work on the driver since [18:04] if it works, upload your changes... Don't forget to do a "fakeroot debian/rules clean" before, just to make sure. [18:04] brb [18:05] hi all [18:06] hu lucazade! [18:06] hi [18:09] any news? any good news?! [18:11] I haven't had the time to touch anything since yesterday [18:11] me too [18:11] outlaw45 has just built a debug package for the xorg psb driver, hopefully that will help tracing the segfault [18:12] should i install it to gather some log? [18:14] Let's see if it works for outlaw [18:14] ok [18:14] awoodland was also working on the backtrace [18:15] we really need debug packages for all our psb packages [18:15] brb [18:16] (EE) Failed to load module "psb" (module does not exist, 0) [18:16] but maybe my setup is screwed.. [18:16] gonna eat.. [18:17] but I added the following to debian/rules [18:17] install-stamp: [18:17] dh install [18:17] dh_testdir [18:17] dh_testroot [18:17] dh_installchangelogs [18:17] dh_installdocs [18:17] dh_installexamples [18:17] dh_install --sourcedir=debian/tmp #-Xlibdrm.la --list-missing [18:17] dh_link [18:17] dh_strip -pxserver-xorg-video-psb --dbg-package=xserver-xorg-video-psb-dbg [18:17] dh_strip [18:17] dh_compress [18:17] dh_fixperms [18:17] dh_makeshlibs [18:17] dh_installdeb [18:17] dh_shlibdeps [18:17] dh_gencontrol [18:17] dh_md5sums [18:17] dh_builddeb [18:17] grep PCI_CHIP src/psb_driver.h | sed -e 's/^.*0x/8086/g' > $(CURDIR)/debian/xserver-xorg-video-psb/usr/share/xserver-xorg/pci/psb.ids [18:17] touch install-stamp [18:17] replaced the existing install-stamp [18:17] maybe it is the sourcedir [18:17] and added a debug package to debian/control [18:18] upload your changes to rules and control, I'll try later to build it [18:21] committed [18:33] <|Bernardo|> hi === |Bernardo| is now known as Bernardo [18:40] quite more changes than I expected in rev 27... [18:40] i believe it's generated [18:40] didn't make that much changes [18:42] but the changes I made to debian/rules don't work [18:42] installed the deb from svn and X loads [18:42] I'll check [18:43] did you change anything under libdrm-poulsbo? [18:43] no .. [18:46] well, you uploaded a lot of stuff... :) libdrm-poulsbo configure, xf86drm.c [18:47] xors-xserver-video-psb had also quite a few changed c and h files [18:48] looks like you didn't revert some of the patches [18:49] I'll try to clean things up if you don't mind, dropping most of your changes and keeping only the rules and control, is that ok with you? [18:50] reverting at the moment [18:50] wait a minute [18:52] I've reverted to r 26 (svn up -r 26) then checked out rules and control in rev 27 (checkout -r 27 rules) [18:52] ok.. [18:53] cleaning didn't work here anyway :p [18:54] now, how do I tell it to forget rev 27? :) [18:58] ok, seems like I have to tell it to do a "svn merge -c 27 file" for all the wrongly changed files [19:00] sorry but gotta go [19:00] check in when I'm able [19:00] cya [19:00] I'll try to revert the erroneous changes [19:01] :) [19:01] that's one of the problems with svn [19:02] you need to check what you are uploading before you do a commit, with svn status [19:02] ah good to know [19:03] going to dinner and watchin soccer match :) [19:03] let's see if this work [19:04] what game is on today? [19:04] roma-inter [19:04] italian cup [19:04] open or on sky? [19:04] open i believe [19:05] maybe I'll see it, depends on the wife... :) [19:05] heheh... always forza roma! [19:07] bbl === Bernardo is now known as Bernardo|away [19:16] hm, bugs where 'nomodesetting' boot option is needed for the live-cd, against what should those be reported? linux or xorg or ...? [19:33] ah, I tried multiple distros on my dreaded poulsbo netbook.. only successful was Mandriva's latest beta + force xvesa.. [19:33] various versions of ubuntu fails to get vesa xorg because it seems it can't access vesa bios modeline settings.. === Bernardo|away is now known as Bernardo [19:57] I have karmic working well on my 1101HA [19:57] thanks to lucazade [20:01] ok, rev30 should have rolled back all extranous commits of rev27 [20:02] we'll see if it builds the debug packages [20:07] it builds empty packages... So anyone trying to build xserver-xorg-video-psb should either use rev 26 or fix rev 30 === Bernardo is now known as Bernardo|away === BUGabundo is now known as BUGabundo_Chuck [21:39] hi all [21:39] my new laptop has a bit of an unstable picture on the VGA port [21:40] kind of a pixel or two of wobble [21:40] it gets faster and slower with the refresh rate, but I can't get rid of it. Happens at all resolutions from 800x600 to 2048x1152 [21:43] if i do a restart, when it comes back up there is a lot of ram usage (going to xorg). if i shut down and then start up, ram use is much lower [21:49] the laptop I have is a Samsung R530 with an Intel chipset using the i915 driver === BUGabundo_Chuck is now known as BUGabundo_House === BUGabundo_House is now known as BUGabundo_DrHous === BUGabundo_DrHous is now known as BUGabundoDrHouse === BUGabundoDrHouse is now known as BUGabundo_HIMYM [23:21] hello, i am having problems with Intel Corporation 82852/855GM Integrated Graphics Device (rev 02). try to find solution in lauchpad but could't. ca anyone help [23:22] i belive there is a bug with intel 855. does anyone knows how to solve? === BUGabundo_HIMYM is now known as BUGabundo_NCISLA [23:37] apw: Next-day pong.