=== persia_ is now known as persia [07:28] anyone there? [07:29] Snookie: Lots of people here, but all are less likely to respond without a question or statement related to the channel topic :) [07:30] = P [07:30] I had a question regarding embedded systems and linux, hoping someone could help me out.... I'm working on a kernel for my board, I'm a little confused, I have a NAND flash I'm going to use for a file system, do I need to map the physical address inside the kernel menuconfig? [07:34] Heh. That's a lower-level question than I usually see here. I have a couple systems that boot off NAND, but the NAND is presented as a filesystem device (PATA and usb-storage, respectively), so there's no memory mapping. [07:35] persia: so you're using just a usb stick that has nand flash? [07:36] Snookie: No. In one case, it's a IDE adaptor that has some NAND on-board, and in the other, the manufacturer used the USB bus for the NAND connection. [07:37] Also, I'll admit to not knowing enough about the deep kernel to answer your question: my familiarity is mostly userspace. [07:37] persia: ah okay, not really the same thing, do you have to have NFTL compiled in the kernel or is this something completely different? [07:39] NFTL is the mtd device stuff? I'm really not sure. [07:40] yeah, nand flash translation layer [07:40] I'd guess that you need to either have NFTL compiled-in, or available as a module in the initramfs. That said, you'll need your bootloader to be able to get to it in either case. [07:40] yeah [07:40] = ? [07:40] = ( [07:41] this stuff is like black magic, seems like people just understand it without being taught, I've read books and can't find answers [07:41] but I appreciate your help [07:41] linux nand stuff is available at: http://www.linux-mtd.infradead.org/ [07:42] there is also the irc channel #mtd [07:42] Snookie: Best advice I can give you is 1) make sure you can access the NAND in your bootloader, 2) Boot off something else and fiddle with kernels & modules until you have something that allows you to partition/access the NAND, 3) Put the working solution on the NAND, and try loading from the boot loader. [07:43] awesome thank you persia [07:43] suihkulokki: I'll give it a shot, thanks [08:54] hi [08:56] Anybody knows how UME saves system sound (Such as login sound, logout sound) setting ?? [08:58] I'd like to change login sound but can't find where it is defined. I know login sound is saved in /usr/share/sounds/login.wav. But I'd like to change the config [08:59] jerryfan: I've not specifically investigated for Ubuntu Mobile, but for Ubuntu Desktop, that name was awkward to modify last I investigated. [09:00] You may have the best luck with replacing the file, although something could have happened in the last 8 months about which I'm unaware. [09:00] I think ume system sound setting might be same as desktop [09:03] persia, thanks I will try that. But i'd like to change it through a config file (if it exists) ultimately. [09:04] jerryfan: If you find a way to do that, please share. Further, if you can find a way to do it without disturbing package updates, etc. it would be even better. There's a few groups that have a similar desire. [09:04] persia, sure thing [09:18] hi all [10:35] Ah missed jerrygan [10:35] *fan [10:35] persia: In the future, we might get the sound theme specs [10:36] persia: Lennart has been pushing for new FD specs copying the icon theme stuff -- for sounds [10:37] lool: That's excellent news. I know TheMuso was chasing some hacks for hardy, and took a bit of a look, but it seemed the infrastructure wasn't quite there. I've read some of lennart's stuff, and it seems very different than the current model: nice for us, but I wonder how long it will take the Desktops to adapt. [10:42] * lool goes for lunch & [14:03] jerryfan, did you see my response? [14:04] pmcgowan, I am reading my post and try to recall what bug is it. [16:12] jerryfan: I see you pinged me about sound card configuration over the WE, but I came too late to the computer; did you clear your questions there? [16:12] lool, No. Still can't find where is system sound config is. Such as login.wav and logout.wav [16:15] jerryfan: I have no idea how login and logout is implemented in your builds; on Ubuntu it's handled by GDM and on UME we don't have them [16:15] jerryfan: Other sounds, I would have thought would be in GConf, but I don't find them [16:15] did you just see my last lines? [16:16] lool, We'd like to change login.wav and logout.wav. But we can't find where it is defined. I think there is some kind of conf files in the system for this purpose. Do u know where it is? [16:16] lool, for ur previous question, yea we did found a way to switch sound card manually [16:16] jerryfan2: http://paste.ubuntu.com/23990/ [16:17] jerryfan2: I'm asking seb128 (doing Ubuntu desktop) about the other sounds [16:18] lool, are u sure system notification sound is handled by GDM?? [16:19] jerryfan2: I'm pretty sure for a regular Ubuntu desktop for login [16:19] jerryfan2: Not for UME of course, we don't use gdm [16:19] SoundOnLoginFile=/usr/share/sounds/question.wav [16:20] in /etc/gdm/gdm.conf [16:20] lool, let me check [16:20] jerryfan2: So perhaps your flash movie sends an event to play the various sounds [16:20] lool, I think i have seen /home/ume/.gdm before [16:21] jerryfan2: I think I saw some work to bridge sound events over m-b-f recently [16:21] lool, what is m-b-f? [16:21] mobile-basic-flash [16:21] lool, there is no .gdm in /home/ume [16:21] The engine displaying the home screen [16:21] jerryfan2: As I said, UME doesn't use GDM [16:22] jerryfan2: I was just comparing regular Ubuntu versus UME [16:22] lool, but if u killall moblin-setting-daemon and then start again, u will heard login sound again [16:23] lool, so it doesn't look like login.wav is handled by m-b-f becasue on our system, we actually see flash 1st and then hear sound 20seconds later [16:24] jerryfan2: you have mbf source? [16:24] jerryfan2: It is in GConf as I thought [16:24] To enable/disable them [16:24] and list is in .gnome2 [16:25] jerryfan, gdm has never shouldnt and will never write to homedirs, its a system service, usually running as the gdm user even [16:26] have a look at the edubuntu-artwork and xubunt-settings packages how to set up cusatom gdm.conf setups, we have a special hack for that in ubuntus gdm [16:26] (look for gdm-cdd.conf) [16:27] lool, I'll look into mbf src later. [16:27] lool, But I checked on my Ubuntu 8.04 desktop and can't find any sound setting in gconf. [16:28] jerryfan2: The list should be in .gnome2/sound/events/gnome-2.soundlist [16:29] jerryfan2: See libsounds/sound-properties.c in control-center's source [16:29] lool, is " .gnome2/sound/events/gnome-2.soundlist" a dir path or gconf path? [16:29] It's a regular file [16:29] The gconf part is only to switch sounds on or off [16:30] /etc/sound/events/gnome-2.soundlist would be the system equivalent btw [16:30] lool, Hmmmm........I did see sound.lst on my desktop, but not on ume [16:30] ogra_cmpc: right [16:30] I am trying to find a toolchain and embedded linux for development on a http://www.vortex86sx.com/default.htm. I am new to embedded linux and it looks like most of the budding distros have died out. Could I use Ubuntu for this board, if so, where to start reading :) [16:30] not easy to change though ... apart from a dpkg-divert or something [16:30] Sciri: what hostname do i use for sledhammer from mobilepartners? [16:30] jerryfan2: only created in your home if you diverge from the defaults [16:31] ogra_cmpc: don't divert conffiles! [16:31] Sciri: sorry, wrong channel.. [16:31] lool, oh, i didnt know it was one :) [16:31] ogra_cmpc, lool : I found /etc/sound/events/gnome-2.soundlist on ume. [16:32] that issue will likely come up more often, we should add a proper events.d or so [16:32] ogra_cmpc, lool , I will try to mod soundlist in etc and see does it work. Thank you guys [16:32] so OEM adjustments can go there [16:32] lool, merci ! [16:33] lool, Merci mille fois de nous avoir aidés [16:33] Eh you speak French! [16:33] Je me sens moins seul ! [16:33] lool, Frech Canadian [16:33] jerryfan, if its really a conffile, note that you will get questions from update-manager if there are security fixes to libgnome2-common if you modify it [16:34] lool, just joking. Canadian. [16:34] Erf calorifères and all [16:34] jerryfan2: You can drop a file in /home/ume/.gnome2/ etc. too [16:35] It's mostly how it was handled until now for other settings I'd say [16:35] we shoud get that into proper global settings in intrepid, really [16:35] lool, so if I add sound.list in /home/ume, then we will only change notification for ume and not globally right? [16:36] (thats noting you can use if you actually have users (i.e. on netbooks) a generic override will be better for such stuff) [16:36] jerryfan2: Correct [16:36] but thats surely a bunch of packages to fix ... [16:37] lool and ogra_cmpc , then I think that would be our solution. Great. [16:37] ogra_cmpc: I would guess it's easier to have a one shot default config in /home/ume rather than fork libgnome2 just for this [16:37] not fork [16:37] add support for override configs should be a simple hack [16:37] Or patch to read another location [16:37] It's a fork in the end at this point [16:38] and surely something upstream would at least think about for making it esaier for OEMs and derivatives to adjust [16:39] imho its a bug that it uses hardcoded paths with no way to cleanly override [16:39] It's being fixed by using a sound theme spec [16:39] Just like for icons [16:39] I think OEM would like to adopt some kind of theming idea like cell phone, to apply sounding and mbf look. [16:39] right [16:39] xactly, so lets make it easy for them [16:42] lool, one more question. I have seen documention on hildon official page saying something about app state save. Why this feature is not implemented in UME? We are now suffering big time on thermal issue. [16:44] jerryfan2: Good question [16:45] jerryfan2: I am not quite sure what type of app state save you mean though [16:45] I don't think you mean the Androir low memory thing [16:46] *Android [16:46] lool, in Nokia N810, maemo launcher automatically check for memory usage and close unused apps on the fly after saving its state. And next time user open up the deleted app, maemo will restore it to previous state. [16:47] jerryfan2: Ok; I've read some things about state saving on memory constrained devices in Android too [16:48] And I also knew of a project which was about saving any process under Linux and try to restore it later [16:48] lool, sounds like application suspend. [16:48] jerryfan2: This would definitely be something interesting for UME apps, but it requires important efforts to properly save/restore apps [16:49] It's not trivial for things like your IM status [16:49] We'd be happy to merge such work though [16:53] lool, btw, had a boring weekend again after some heavy ltsp testing so i needed some distraction ... http://people.ubuntu.com/~ogra/MightyMailApp/ [16:53] ;) [16:53] have u guys tried Intel c compiler? Does it really help? I tried few times but notice no improvement at all [16:55] ogra_cmpc: haha [16:55] ogra_cmpc: you're crazy :) [16:55] just posessed somtimes :) [16:55] it looks real cool [16:55] another booring sunday with libgmail and it will actually act as gmail client [16:55] jerryfan2: It might help for some heavy maths, and I know it helps in some particular cases but I wouldn't expect a huge gain in the average case [16:56] ogra_cmpc: Nice idea [16:56] ogra_cmpc: If you're bored next weekend, come around and help me paint my walls [16:56] yeah, i looked at libgmail before going to bed yesterday, seems fairly trivial [16:57] * lool did some Debian unpiling this WE [16:57] * ogra_cmpc has a 200sqm house with rotten walls himself :) [16:58] and a garden that starts looking like a jungle ... ot very german :) (they use to mow their lawn with nail scissors here)) [16:58] *not [17:00] But python app runs real bad on UME, especially the old media center [17:05] jerryfan2: What about elisa? [17:06] Ah too late [17:07] oyu should really make it default in intrepid if you are so intrested in feedback ;) switching back is cheap if it doesnt work good enough [17:07] True [17:10] pmcgowan: Tobin here. Heard you are having problems finding the psb headers. [17:11] does ubuntu mid work on any i386? How about a SOC that emulates a 486 with no Math Coprocessor? [17:13] Can you point me to a website to get started. I have a Hardy box here I could use to put the toolchain, etc. [17:14] the default images are all built for the lpia architecture which is compiled with 686 instructions ... you will need to build an image with the -386 kernel flavour and i386 packges to support 486 [17:15] and i doubt the performance of that will be great ... [17:15] lol..I hear ya [17:15] i got the board for free...so here we go [17:15] but theoretically its all possible :) [17:16] what do I have to add to heron to get started with this...I see bits and pieces here and there on wikis/webs/ but I don't seem to find a clear-cut process to follow... [17:21] Hmm https://wiki.ubuntu.com/MobileAndEmbedded/InstallingUMEWhichIsFor686OnMy486 ? [17:21] :-P [17:22] see what I mean...LOL [17:22] :P [17:23] the meetings pages are not very useful either... ;) [17:25] Try the released kvm image perhaps? [17:39] GrueMaster, hey, I think Calvin is figuring out why they went missing from the libdrm package [17:40] They never were part of that package, that I know of (and I build them). They are part of the kernel module package. [17:41] libdrm requires them as well. [17:41] I think they were in libdrm until shortly before PV [17:41] BTW most drm headers are duplicated between kernel and libdrm, so it seemed quite normal to have them duplicated for psb as well [17:41] Only if they were repackaged by your team. [17:42] GrueMaster: I mean they were in libdrm in your releases [17:42] I'll double check that. [17:43] (If you have an /usr/include/drm, I'm sure you see what I mean with the other drivers doing this already) [17:45] I see the others in the libdrm-dev, but not the psb specific headers. I went back to 2.0.0.32L.0010. [17:45] I went back there too [17:46] I'm checking 2.0.0.3 now :) [17:46] GrueMaster: Ah hold on, these were in xf86-psb [17:51] GrueMaster: It was around 14 it seems [17:51] GrueMaster: Check https://launchpad.net/~ubuntu-mobile/+archive?field.name_filter=xserver-xorg-video-psb&field.status_filter=any [17:51] GrueMaster: Version 0.13.0ubuntu1~804um1 [17:51] Bryce had to copy the headers as a patch in -psb because the headers were dropped from -psb [17:51] Later on, I suggested him to move the patch to libdrm where other similar headers were living [17:53] Well, they didn't ship that way from here. I just checked all of my ubuntu builds from 0010 through 0019. They are only in the psb-kmd deb packages I built, which should be what is up on arms. [17:57] GrueMaster, lool so what is the correct place for these? [17:58] I would suggest that since they come with the kernel modules, they should be part of your linux-headers-lum package. [17:59] GrueMaster: That's what I thought as well but I discussed this with amitk and it turned out other drms are in libdrm [17:59] GrueMaster: I wanted to do exactly what you proposed initially [18:03] GrueMaster: So would it be possible to copy them in libdrm like the other drm headers? [18:04] It's not up to me, but I'll check. This whole mess needs to be cleaned up so it can go upstream anyways. [18:06] pmcgowan: on a different note, I'm having issues with your 20080618 image booting on a ODM device. No X, unless I load the kernel modules manually. Any suggestions? [18:13] GrueMaster, no idea, this is the target hardware? [18:13] yes [18:13] GrueMaster, what BIOS version? [18:13] Trying it on a CB as well. [18:13] Not sure. Let me check. [18:13] CB will need some tweaks in X [18:13] to disable TS driver I think [18:15] dmidecode shows 37AT02 [18:16] sounds pretty recent, let me check [18:16] 37AT02 sounds like our machine. [18:18] jerryfan, yes [18:19] GrueMaster, we are running the same bios version [18:19] ok. [18:19] we have no such problem actually. is it 100% failure rate? [18:19] yes [18:19] same on cb so far. [18:20] we have seen this problem before only if image-creator is set incorrectly [18:22] GrueMaster: The 20080618 image will not work on a CB without disabling the touchscreen driver in xorg.conf. It should be fine on the target HW [18:22] GrueMaster: What version of image creator do you have? [18:23] I don't know. Tip, I guess. I have a cron job that pulls it daily. [18:23] but I'm not using it on your image. [18:25] I have commented out the respawn command from /etc/event.d/session. That at least gives me a prompt. When it boots, it dumps down to a login, and I can run startx -- -br -config xorg-cronbeach.conf and the system will come up. [18:26] GrueMaster: So I checked where the headers were removed [18:27] GrueMaster: It was in moblin's libdrm [18:27] GrueMaster: libdrm for UME came from moblin.org's git tree for a while [18:27] GrueMaster: Where they had the same fix in place as the one we use now: shipping headers in libdrm copied from the kernel module [18:28] GrueMaster: When moblin.org's libdrm stopped doing so, we readded the patch ourselves [18:28] GrueMaster: Anyway, I think there's no doubt that other drivers are shipping the headers in libdrm, so it would be nice to get it in there upstream [18:29] GrueMaster: is there anything in xorg.log when x fails to come up? [18:29] actually, no. [18:29] which is odd. [18:30] GrueMaster: If there is no error message, it must be working. :) [18:30] none that I have seen. Not in Xorg.o.log or dmesg. [18:31] I'll try reimaging just for grins. [18:32] check ur default setting in image creator [18:32] if it is boot=1024, then u are wrong [18:32] is there any framebuffer kernel driver that could come in your way ? the speeded up bootprocess could cause it to load exactly when x is about to start up [18:32] shoud be 768 and 512 [18:32] andy told me that [18:34] no boot= options in grub that I see. And I am not using image creator. [18:35] there are boot swap and fat32. boot=768 swap=768 [18:36] nono. i am talking about default in image-creator. [18:37] u made wrong settign in default, ur build might fail during X start up [18:37] jerryfan, I think he is just installing the stock image, not making a new one [18:37] exactly. [18:37] I haven't even downloaded the project file yet. [18:38] Just the usb-install image. [18:38] GrueMaster, then I have no idea then ......... :( [18:42] but i can assure u that 037+0618 are good [18:43] ok. I'll mess with it some more here. [18:44] When is the PV image for this system? [18:44] any minute now :-) [18:44] cool. I'll keep my eyes open. [18:48] jerryfan: =.= not the same things. [18:48] GrueMaster: since you can boot in console mode. [18:49] can you use 'df' to check, is disk full ? [18:49] will do. [19:03] ok, reimaging it seems to have fixed the issue. Sorry for the long delay, had power supply issues as well. [19:12] GrueMaster: I know when disk full, it can't enter X forever (100%). [19:14] I'm thinking that maybe the partition reformatting didn't clear everything the first time through. I discovered that I didn't have power from the supply until I moved it to a different outlet, and I'm thinking that may have effected the ssd erase procedure. [19:15] It's running fine now, on a different outlet and reimaged with no other changes. [19:29] what is the platform you guys are running on? [19:30] a moblin or your desktop in some virtual machine [21:28] hi [21:28] need small help [21:29] i made a live RW usb with Image-Creator, booted in htc-shift, it asks for login [21:30] tryed different logins, all doesnt work [21:32] hello [21:33] userid is ume. no password [21:33] oh thaaaaaaaaanks [21:33] dude [21:33] thanks [21:36] actually the htc-shift is with my dumb boss, he doesnt let me sleep and he has something else he cant pass from now [21:36] he says it prompts for some command now, what that would be ? [21:38] can i get a page where i can read some guide for "after making" usb.. [21:38] i'll take a shower, brb in 10 minuts === Zic_ is now known as Zic