=== jkridner_ is now known as jkridner [02:17] Are beagles armv7? [02:18] (f17 arm was released recently and I was idly wondering what their minimum hw target was.) [02:23] twb: Beagles are v7. [02:23] twb: Their armhf port is v7-vfp-d16, same as ours. [02:23] twb: However, theirs is binary incompatible, because they didn't sort out the cross-distro liker mess in time. :/ [02:24] Hehe [02:24] Fortunately I can be idly curious because I don't have to babysit FC boxes anymore :-) === jkridner_ is now known as jkridner === zyga-afk is now known as zyga [09:43] ogra_, ping [10:02] ogra_, pinga ring [11:18] ronoc, yo, sorry, had an early lunch [11:18] no worries [11:19] ogra_, timR is out now [11:19] ogra_, later when he comes back I'll do some introductions if you don't mind [11:19] fine, yeah [11:19] ogra_, thanks [11:53] ogra_, ronoc, hi, just noticed this [11:53] hey tim_ [11:54] hey [11:54] ogra_, this is my mate tim_ i was speaking to you about [11:54] gumstix audio issues [11:54] Hello ogra_! [11:56] tim_, do you have your notifications turned on [11:56] not sure, don't ususally use IRC [11:59] tim_, when I write tim_ the icon on your panel should go blue or something [11:59] what client are you using === ronoc is now known as ronoc-lunch [12:11] tim_, hey [12:12] ogra_ hi! [12:12] so you got sound issues i heard [12:13] yep should we switch to direct message? [12:13] nah, just keep it here [12:13] ok [12:13] I'm just about to run out I'm afraid [12:13] I'll explain quickly [12:14] I'm using 12.04 on gumstix boards [12:14] yep. connor said so [12:14] -n :) [12:14] there is a driver called snd-soc-overo for the built in audio [12:14] It works in openembedded I beleive [12:14] I'm modprobling it and it is loading [12:14] right, do you see anything with "dmesg|tail" if you load it ? [12:15] but no usable devices [12:15] no I don't think so, I'll check that [12:15] and is there anything in /proc /asound ? [12:15] Is it ok if I check that and get back to you in about 2 hours? just have to do something 1st [12:16] sadly we dont have amny gumstix users so we dont really get much feedback for that SoC [12:16] you are in charge of soc audio? [12:16] i have meetings the whole afternoon so i wont be available much [12:16] tim, are you using armhf or armel? [12:16] anytime in the next few days would be great [12:16] armhf [12:16] no, i only care for images usually and for the arm port in general [12:16] but that often enough includes userspace audio fixes [12:17] OK.. will report back in a bit, and thanks [12:17] cool, good luck :) === doko_ is now known as doko === zyga is now known as zyga-food === zyga-food is now known as zyga [14:36] ogra_, dmesg just says - [ 392.393524] overo SoC init [14:37] ogra_, cat /proc/asound/cards says --- no soundcards --- [14:38] hmm, then your driver probably either needs additional modules loaded ... or whats more likely (and happens often on arm) is that the driver only works properly if compiled in [14:39] there are already a bunch of modules loaded [14:39] snd_soc_core 111406 2 snd_soc_overo,snd_soc_twl4030 [14:40] is snd_soc_twl4030 supposed to load I wonder [14:40] twl4030 is the chip that's being used [14:40] and once I have loaded this once, it autoloads on reboot [14:41] btw uname -a : Linux overo 3.2.0-23-omap #36-Ubuntu Tue Apr 10 20:24:21 UTC 2012 armv7l armv7l armv7l GNU/Linux === prpplague^2 is now known as prpplague [15:03] tim_: In my experience with the beagleXM and beagleboard (essentially the same SOC), the audio only worked when the drivers were built into the kernel. You should try rebuilding the kernel with the twl4030 & overo drivers compiled in. [15:04] (I "used" to work on the audio stuff for these boards) [15:09] Thanks! [15:11] I believe these may actually be built into the kernel already? There don't seem to be any .ko files for them.. === zyga is now known as zyga-afk [15:23] Hmmm. [15:24] Try "fgrep TWL4030 /boot/config-*" [15:24] See if they are. [15:49] GrueMaster ok - this checks the kernel build flags? - CONFIG_SND_SOC_TWL4030=m [15:50] and CONFIG_SND_OMAP_SOC_OVERO=m [15:50] Yes. And according to that, the driver is built as a module. [15:50] I guess this means they aren't [15:50] Need to rebuild with those set to Y. [15:51] Great! ok, this is going to take a bit of research.. [15:51] do you know if it needs both of them? [15:52] tim_, which gumstix have you got? (I have an IronStorm) [15:52] I think if you set one in the kernel config menu, it will flag the other, but I am not sure. I would just enable both. [15:52] I have airstorm and tide [15:52] I believe they are all compatible with the same kernels? [15:52] I am a noob ;) [15:53] See http://bugs.launchpad.net/ubuntu/+source/linux/+bug/925094 [15:53] Ubuntu bug 925094 in linux "No audio on omap (beagleXM) system" [Medium,Confirmed] [15:53] (I knew I had filed one - with the fix). [15:54] OK this seems to be a similar issue alright [15:55] At the end there is a post that says its a dependency issue? [15:57] Yea, I just saw that. Can't test it though (not at home, and beaglexm is packed away since I have moved on). [15:58] tim, did you use a preinstalled ubuntu image, I think I tried it earlier and just sat there [15:58] Uncompressing Linux... done, booting the kernel [15:59] I used a preinstalled image yes [15:59] hmm, ok [15:59] cvanvliet: That sounds like a desktop image. It requires video & keyboard. You won't see much else on the serial port. [15:59] I think the current one does that for me also [16:00] cvanvliet, you want a server image for serial output [16:00] GrueMaster, it is set up with DVI, keyboard etc [16:00] aha, I just tried loading snd-soc-omap and snd-soc-omap-mcbsp and the device has appeared! [16:01] without rebuilding the kernel ;-) [16:01] congrats ! [16:01] now to check if there is sound [16:01] cool. You can add the dependency to /etc/modprobe configs. [16:04] ogra, I actually want a desktop, but I can try the server image later [16:04] I actually want 12.04 armel, tbh [16:04] you cant install the desktop image if your screen doesnt work [16:04] and there are no armel images for 12.04 [16:05] I know, I have been trying to make one [16:06] ogra, I am confused, there is a screen attached [16:06] right, but the dirver your kernel ships was only tested on beagleXM [16:07] ok [16:07] its a matter of luck if it works on different omap boards [16:07] thought that may be a the case [16:07] nice to have it confirmed [16:07] and when it worked for tim, I had hope [16:08] desktop images run the installer on the display ... so without working kernel driver you wont be able to install them [16:08] mine is a newer model, so every reason it may not work [16:08] thats why i said use the server image ... it is a minimal ubuntu install that runs completely on serial and also offers to install the desktop at the end if you want to [16:08] ahh ok, [16:08] ahh thanks [16:09] I can try that [16:09] although I need armel, having an armhf may help me figure out the armel [16:09] why do you want armel ? [16:10] it will be lots slower [16:10] (beyond the fact that nobody actually cares if it even works) [16:10] I need the SGX drivers [16:11] ah [16:11] this is for a business of mine [16:11] and we will use opengl for the graphics [16:11] you could ask over in #beagle if there are any people capable of re-rolling the binary driver at TI [16:12] thats essentially what was done for the omap4 hf driver [16:12] one of the guys asked someone here the other day, [16:12] and the answer was dunno :( [16:13] lovely [16:13] ndec, I think it was [16:13] yeah, he isnt working in the omap3 area i think [16:13] I just tried to test the audio.. no audio [16:13] hmm? [16:13] yeah it is quite hard, I feel like i will be using a below par system [16:13] ndec, omap3 SGX armhf [16:13] I am aware that the mixer sometimes is preconfigured for no output [16:13] if I don't use armhf [16:14] i don't do that ... [16:14] i don't do OMAP3 SGX, neither armel, nor armhf [16:14] * ogra_ thought so :) [16:14] its a very complex mixer [16:14] tim_, yeah, thats the problem with bug 925094 [16:14] ndec, sorry if I got that wrong, I thought it was you who responded [16:14] Launchpad bug 925094 in linux "No audio on omap (beagleXM) system" [Medium,Confirmed] https://launchpad.net/bugs/925094 [16:15] if it was easy we would just have a UCM config already :) [16:16] so, accordingly with what tim_ is seeing, i will have audio issues as well , (i need to record sound) [16:16] likely, though for him the driver already works ... should just be a mixer issue now [16:17] if alsamixer actually shows the device and its rulers you are about 90% done [16:17] ogra , ok [16:28] hmm, I have turned up every single bar in the mixer [16:29] alsaplayer thinks its playing a wav [16:29] but.. no audio [17:17] tim_, did you also unmute everything ? [17:18] iirc, the only way I got audio working on the beagle was to rebuild the kernel AND tweak the volume settings. Hopefully you will get it working without rebuilding, but... [17:21] I think I unmuted everything- if OO means on and MM means mute [17:21] Gruemaster, did you also see a device but it didnt work until you rebuilt the kernel? [17:22] Yes. [17:23] OK thats seems pretty clearcut.. [17:25] It seems like the device doesn't power properly unless the driver initiallizes it during kernel init. I don't know enough about the inner workings of the SOC code, but I would guess that the kernel initalization routine powers on devices that need it during boot, but not after. [17:26] On a PC, that's ok as there isn't as much device level power management, but on an SOC that is designed for low power cunsumption.... [17:27] It's like the SOC needs hot plug support for on-die devices. [17:28] Of course, my assumptions could be wrong (it has happened before, once or twice). [17:29] GrueMaster, could the "disabling unused clocks" play into that at all? maybe the SoC is disabling the sound device clocks if it's not configured in kernel init [17:37] That might do it. I'm not a hw designer though. Maybe prpplague would know? [17:37] who summons the plague? [17:38] prpplague: question on the beagleboard/XM/overo Audio. See scrollback. [17:39] yea it is possible that a clock is disabled [17:39] that is something that would take some debugging [17:40] Is there a way to enable it from user land? [17:40] one of the common issues the clock is disabled by mistake when another device enables their clocks [17:40] yea you can use devmem2 to enable most clocks, but without knowing what is enable/disabled, you'd be throwing darts in the dark [17:41] * GrueMaster has done that with...interesting results. [17:41] GrueMaster: i'd bring this to the attention of mdp or tartarus over on #beagle, they'd be able to get to the bottom of it fairly quickly [17:41] Using real darts, of course. [17:42] tim_, bradfa, cvanvliet, there you go. I have done what I can to help from here. [17:43] * cvanvliet passes GrueMaster a beer [17:43] haha thanks a million.. [19:01] lilstevie, no news on the tf101g front? === prp^2 is now known as prpplague === XorA is now known as XorA|gone === gmpff_ is now known as gmpff