**** BEGIN LOGGING AT Fri Jun 22 02:59:59 2012
Jun 22 03:12:58 <robclark> lilstevie, infinity, from what I understand it is running a sort of customized ubuntu 10.10 already.. but hopefully when they post the kernel the board support isn't too hard to forward port and get newer kernel and 12.04 running on it
Jun 22 03:21:18 <infinity> robclark: Hang a USB drive off the thing, and it would be a perfect devel kit to recommend to people.
Jun 22 03:21:25 <rsalveti> robclark: that's cool
Jun 22 03:21:30 <robclark> yup
Jun 22 03:25:21 <ojn> robclark, do you know if it's a 4430 or 4460? hard to argue with that price for something that has a monitor, etc.
Jun 22 03:25:41 <robclark> I think it is 4430
Jun 22 03:27:16 <heathkid> support for realtek USB wifi?
Jun 22 03:28:41 <robclark> http://h18000.www1.hp.com/products/quickspecs/14358_na/14358_na.HTML#Technical Specifications
Jun 22 03:29:18 <ojn> robclark, excellent, thanks!
Jun 22 03:36:30 <lilstevie> I kinda want one
Jun 22 03:42:39 <lilstevie> I also noticed that it has flash
Jun 22 06:54:35 <cvanvliet> Tinti, I am trying to do a 12.04 image for gumstix (armel), and got errors - Failed to create pty - disabling logging for job
Jun 22 06:55:34 <infinity> cvanvliet: Does your kernel have support for devpts?
Jun 22 06:57:24 <cvanvliet> I only compiled my first kernel  a few days ago, and am not great with kernel foo yet, wil check it
Jun 22 06:59:43 <cvanvliet> (what I did do was enable initramfs, and move init to / and it helped, but ended up in init being /bin/sh?)
Jun 22 07:00:10 <cvanvliet> bbs, thanks infinity
Jun 22 07:07:38 <cvanvliet> quick check -> yes, unix98 and bsd are both enabled
Jun 22 07:13:18 <infinity> Okay, but it sounds like you're doing some odd things with your initramfs...
Jun 22 07:13:26 <infinity> Are you creating it with update-initramfs?
Jun 22 07:13:46 <infinity> I mean, talk of moving init around and whatnot sounds a bit unconventional. :P
Jun 22 07:13:59 <cvanvliet> yes
Jun 22 07:14:37 <cvanvliet> infinity, I am mainly trying to learn, don’t take anything I say as being correct
Jun 22 07:14:42 <infinity> Anyhow, I probably shouldn't get into this tonight, I need to sleep and be up early.
Jun 22 07:14:48 <cvanvliet> thanks
Jun 22 07:15:00 <infinity> But, yeah.  If you think you need to move init around, you're almost certainly doing something wrong elsewhere. ;)
Jun 22 07:15:19 <cvanvliet> I may have just seen something :/, let me try
Jun 22 07:15:27 <cvanvliet> thanks!
Jun 22 13:32:43 <ppisati> ogra_: we are going to stay with this kernel for omap4, sorry
Jun 22 13:32:53 <ppisati> ogra_: tilt-tracking is heavily based on linaro kernels
Jun 22 13:33:09 <ppisati> ogra_: and a rebase gave me all kinds of clashes
Jun 22 13:33:35 <ppisati> ogra_: after alpha2 is out, i'll start building a 3.5 kernel for omap4 (as the email on the ubuntu kernel ml stated)
Jun 22 13:36:03 <ogra_> damned
Jun 22 13:36:05 <ogra_> but ok
Jun 22 13:36:23 <ogra_> i guess then we cant really make A2
Jun 22 13:37:02 <ogra_> (with the display issues and the switch to full live images we wont have a way to install)
Jun 22 13:37:33 <ppisati> ogra_: it works for some monitors actually
Jun 22 13:37:44 <ogra_> i have tried three so far
Jun 22 13:37:48 <ogra_> none of them works
Jun 22 13:38:03 <ppisati> i dont' know what to tell
Jun 22 13:38:10 <ppisati> i've two here and they work
Jun 22 13:38:24 <ogra_> n othing to tell, ndec already said they completely reworked the graphics driver
Jun 22 13:38:48 <ogra_> we just need that code dump ....
Jun 22 13:39:17 <ndec> but we don't have it for 3.5
Jun 22 13:39:25 * ndec checks ubuntu kernel list...
Jun 22 13:39:40 <ogra_> oh, i thought you said it was 3.5
Jun 22 13:39:49 <ppisati> nope
Jun 22 13:39:52 <ppisati> still 3.4 for them
Jun 22 13:39:59 <ogra_> my prob is that omap doesnt boot either
Jun 22 13:40:07 <ppisati> and it's actually 3.4 + a lot of stuff from linaro + ...
Jun 22 13:40:08 <ogra_> and ac100 doesnt build yet
Jun 22 13:40:19 <ndec> we have 'frozen' tilt-3.4 just today. so tilt-tracking will start tracking 3.5 any day now.
Jun 22 13:40:19 <ppisati> a disaster! :)
Jun 22 13:40:21 <ogra_> so we wont have any arm images for A2
Jun 22 13:40:30 <ppisati> wait wait
Jun 22 13:40:35 <ppisati> what's wrong with omap?
Jun 22 13:40:45 <ndec> however we (TI) will not spent a lot of energy on 3.5...
Jun 22 13:40:58 <ndec> so i am not sure which features will work in 3.5
Jun 22 13:40:59 <ogra_> ppisati, given that the QA team only has 5 weeks left to get their automated testing working, it is a desaster, there are no images yet for them to write scripts for
Jun 22 13:41:00 <ndec> yet
Jun 22 13:41:31 <ppisati> ndec: i know i know, we are going to use 3.5 upstream after alpha2 + cherry picks useful stuff from tilt
Jun 22 13:41:43 <ndec> it's likely that the TILT 3.5 won't get many more than what's in mainline already
Jun 22 13:42:01 <ndec> so ppisati what's the problem exactly?
Jun 22 13:42:03 <ppisati> ogra_: what's wrong with omap?
Jun 22 13:42:14 <ogra_> oopses all over the place
Jun 22 13:42:24 <ogra_> we didnt release it for A1 if you remember
Jun 22 13:42:32 <ppisati> ogra_: i've been testing it till an hour ago
Jun 22 13:42:33 <ndec> 3.4?
Jun 22 13:42:40 <ppisati> ndec: 3.5
Jun 22 13:42:43 <ndec> ah...
Jun 22 13:42:52 <ogra_> ndec, oopses -> omap3 ...
Jun 22 13:42:56 <ogra_> not omap4
Jun 22 13:43:02 <ndec> tilt-3.4 should be in quite a good shape on OMAP4 and 5.
Jun 22 13:43:11 <ndec> hum. omap3?
Jun 22 13:43:30 <ogra_> omap4 3.5 simply freaks out with the wlan device (demsg gets massively spammed) and doesnt detect any monitor here for me
Jun 22 13:43:50 <ogra_> err 3.4 that is ...
Jun 22 13:44:23 <ppisati> ogra_: me and ming were tracking down an mmc/fs issue in omap for 3.5
Jun 22 13:44:31 <ogra_> ah
Jun 22 13:44:31 <ppisati> ogra_: but apart from that, kernel works here
Jun 22 13:44:35 <ndec> we use wlan with 3.4-tilt with *precise* these days. and it seems to be working.
Jun 22 13:44:52 <ppisati> it seems it's an sd card problem with a late commit
Jun 22 13:44:52 <ogra_> ppisati, well, it didnt for A1 thats why it was decided to be dropped
Jun 22 13:45:22 <ppisati> ogra_: video wasn't working iirc
Jun 22 13:45:27 <ppisati> ogra_: anyway, let me check
Jun 22 13:45:51 <ogra_> and with the switch to full live images the XM install (with only 512M) wont be any fun anymore ... apart from fully relying on video
Jun 22 13:46:33 <ogra_> infinity, around ?
Jun 22 14:12:53 <ppisati> ogra_: i still believe we should enable serial console output by default
Jun 22 14:16:51 <ogra_> that wont help anymore
Jun 22 14:17:14 <ppisati> why?
Jun 22 14:17:23 <ogra_> we are just in the process to switch to full live images, serial wont gain you anything there
Jun 22 14:18:00 <ppisati> and how do we debug problems during installation?
Jun 22 14:18:04 <ogra_> (upstart doesnt enable serial by default so you wont have a login, quiet and splash are set as default so you wont get much kernel output)
Jun 22 14:18:24 <ogra_> oh, for debugging you just dump an uEnv.txt in place
Jun 22 14:18:37 <ogra_> with the cmdline you want
Jun 22 14:19:31 <ppisati> ogra_: so, i;m trying the daily image installation
Jun 22 14:19:38 <ppisati> ogra_: omap3 on a beagle xm
Jun 22 14:19:41 <ogra_> but that wont get you anywhere with a live image
Jun 22 14:19:52 <ogra_> since the installer doesnt run on serial
Jun 22 14:20:02 <ppisati> ogra_: i modifided boot.scr to have srial output
Jun 22 14:20:11 <ppisati> and it's X failing
Jun 22 14:20:21 <ogra_> right, you can just use uEnv.txt nowadays
Jun 22 14:20:29 <ogra_> no need for fiddling with mkimage etc
Jun 22 14:20:50 <ogra_> lucky you, for me it didnt boot (with A1 at least)
Jun 22 14:21:01 <ppisati> daily from today
Jun 22 14:21:09 <ppisati> so maybe we can fix it
Jun 22 14:21:19 <ogra_> yeah, but there werent any omap related changes, were there ?
Jun 22 14:21:21 <ppisati> eutehr something screwed in the kernel (omapfb&c)
Jun 22 14:21:32 <ppisati> well, we passed from 3.4 to 3.5
Jun 22 14:21:42 <ppisati> so, everything could have happened
Jun 22 14:22:00 <ogra_> ah, well, i'll try an omap image on monday, currently i'm focused to get omap4 live working
Jun 22 14:22:01 <ppisati> ok, let's do that
Jun 22 14:22:22 <ppisati> btw, can you get a new sd card and try an old P release on that panda?
Jun 22 14:22:32 <ogra_> why a new SD card ?
Jun 22 14:22:32 <ppisati> i wonder if video output is ok
Jun 22 14:22:43 <ppisati> or rewrite it, whatever
Jun 22 14:22:51 <ppisati> i mean, don't dump your installation
Jun 22 14:22:54 <ogra_> i teasted all püanda images during precise development
Jun 22 14:22:58 <ogra_> well, all milestones
Jun 22 14:23:09 <ogra_> including the final reelase
Jun 22 14:23:16 <ogra_> there were no issues at all
Jun 22 14:23:36 <ogra_> (and you asked me to test the precise kernel when my A1 image didnt work if you remember)
Jun 22 14:24:08 <ogra_> downgrading to the precise kernel fixes everything here
Jun 22 14:25:35 <ppisati> ah, when was last time you had a working video kon that?
Jun 22 14:25:38 <ppisati> because
Jun 22 14:25:41 <ppisati> http://www.kernelhub.org/?msg=30802&p=2
Jun 22 14:25:50 <ppisati> there was a window of time
Jun 22 14:26:02 <ppisati> where you could destroy your video output
Jun 22 14:37:35 <ppisati> reboot
Jun 22 14:51:25 <cvanvliet> what are the differences between ubuntu and linaro ( particularly omap3 , SGX)
Jun 22 14:53:01 <ogra_> linaro takes snapshots of the ubuntu archive and rolls monthly images from them
Jun 22 14:53:34 <ogra_> on top of that they include their imporvements that will soon land in the ubuntu archive ...
Jun 22 14:53:55 <ogra_> they dont support their images though ... no security fixes updates etc
Jun 22 14:55:11 <ogra_> so you get an image with outdated (or the same) SW from the last ubuntu release but with added linaro sweetness on top (toolchain fixes, other kernels, bootloaders etc)
Jun 22 14:57:13 <cvanvliet> ok, that clears a lot up
Jun 22 14:57:28 <Tinti> hello ogra_ did you wrote rootstock tool?
Jun 22 14:57:49 <ogra_> Tinti, yes, but its dead since over a year now
Jun 22 14:58:07 <cvanvliet> was Tinti the gumstix guy earlier?
Jun 22 14:58:29 <Tinti> I saw, I want to start developing in linaro do you have any advices or tools that I should use?
Jun 22 14:59:00 <ogra_> ask in #linaro ? :)
Jun 22 14:59:04 <Tinti> cvanvliet, no I am Tinti for a long time :)
Jun 22 14:59:17 <cvanvliet> I think I jumped to conclusion
Jun 22 14:59:26 <cvanvliet> sorry long night :/
Jun 22 15:00:05 <Tinti> in fact I am there but I am not getting much things ... in fact I would like to get recommendations from you. probably you have switched from rootstock to something else right?
Jun 22 15:01:00 <ogra_> nope
Jun 22 15:01:26 <ogra_> rootstock was never actually used anywhere in ubuntu, it was just a script combining a lot of hacks
Jun 22 15:01:56 <ogra_> to roll a rootfs you can use live-build ... and i think there are a ton of linaro tools to create images
Jun 22 15:02:26 <ogra_> so best ask in linaro (we dont use these tools either in ubuntu, #linaro is a better place to ask about this)
Jun 22 15:02:35 <Tinti> yes there are, I was thinking in use it
Jun 22 15:02:59 <cvanvliet> are they any tutorials to get 12.04 armel up an running?
Jun 22 15:03:48 <ogra_> well, there are installation howtos for the official images
Jun 22 15:03:50 <Tinti> yeah thanks. I need to compile an environment for iMX51
Jun 22 15:04:07 <Tinti> I would like to do it from scratch in fact
Jun 22 15:04:16 <infinity> ogra_: Around now...
Jun 22 15:04:22 <cvanvliet> ogra, I have tried most of these (they all seem beagle based)?
Jun 22 15:04:48 <ogra_> infinity, so i will switch the crontab to roll live images for omap, omap4 and mx5 today
Jun 22 15:04:59 <ogra_> infinity, what do we do about preinstalled server ?
Jun 22 15:05:22 <infinity> ogra_: We leave it as-is for now.
Jun 22 15:05:22 <Tinti> cvanvliet, I really would like to have this
Jun 22 15:05:23 <ogra_> i dont want to introduce alternate again
Jun 22 15:05:27 <ogra_> ah, cool
Jun 22 15:05:35 <infinity> ogra_: If x86 server switches to live, we'll follow suit.
Jun 22 15:05:49 <ogra_> cvanvliet, right, they are either beagle, panda or freescale mx5 based
Jun 22 15:05:52 <cvanvliet> netboot, server image (armhf), build up form ubuntu core, qemu-debootstrap
Jun 22 15:05:53 <infinity> ogra_: Most sane server users install from d-i netboot anyway, regardless of the platform.
Jun 22 15:06:05 <cvanvliet> ogra I have a gumstix
Jun 22 15:06:05 <ogra_> infinity, they dont switch to live, they just use a squashfs in d-i
Jun 22 15:06:12 <Tinti> the tutorials seems to be old in fact no?
Jun 22 15:06:18 <Tinti> at least on wiki
Jun 22 15:06:29 <ogra_> right, server on omap or omap4 is mostly intresting as minimal dev image
Jun 22 15:06:36 <infinity> ogra_: That's a "live-style" install.
Jun 22 15:06:44 <ogra_> well, its still d-i
Jun 22 15:06:48 <infinity> ogra_: As in, no more installing packages.
Jun 22 15:06:49 <ogra_> :)
Jun 22 15:06:52 <ogra_> right
Jun 22 15:06:55 <infinity> ogra_: ubiquity is d-i.
Jun 22 15:07:06 <ogra_> pfft, yeah
Jun 22 15:07:12 <infinity> I see this as "text ubiquity". :P
Jun 22 15:07:28 <ogra_> k ... so i wont touch server ... thats good
Jun 22 15:07:40 <ogra_> wrt install speed live is pretty horrid btw
Jun 22 15:07:59 <ogra_> especially the partitioner takes like 5min after you clicked anything
Jun 22 15:08:16 <ogra_> my test install took about 90min until it failed in the bootloader install
Jun 22 15:08:21 <infinity> That seems odd.
Jun 22 15:08:28 <infinity> The 5m partitioner thing, I mean.
Jun 22 15:08:36 <infinity> I expect the copy to be slow, not much else.
Jun 22 15:08:40 <ogra_> thoguh i'm trying out the slowest HW combo i can atm
Jun 22 15:08:51 <infinity> Oh, installing to a USB stick?
Jun 22 15:08:53 <cvanvliet> ogra, the live installs ar armhf? I assume
Jun 22 15:09:00 <ogra_> i.e. installing to an SD in a cardreader as target
Jun 22 15:09:09 <ogra_> cvanvliet, everything is armhf
Jun 22 15:09:15 <ogra_> we dropped armel last release
Jun 22 15:09:16 <cvanvliet> yep
Jun 22 15:09:37 <cvanvliet> armel is dropped after precise ?
Jun 22 15:09:47 <ogra_> while the archive is still there it is currently largely unmaintained and we dont roll any images for that arch
Jun 22 15:09:56 <infinity> cvanvliet: We didn't ship armel images for precise either.
Jun 22 15:10:05 <infinity> cvanvliet: Just netboot images (which we still ship).
Jun 22 15:10:07 <cvanvliet> there is a core though
Jun 22 15:10:11 <ogra_> and its likely it will be completely dropped before final release
Jun 22 15:10:30 <cvanvliet> infinity, it wont find my ethernet
Jun 22 15:11:25 <ogra_> file a bug and ask for the NIC driver module to be included in the installer
Jun 22 15:11:34 <ogra_> (file it against linux and debian-installer)
Jun 22 15:12:48 <infinity> We might not have the driver at all, if this is an SoC we don't support.
Jun 22 15:13:01 <ogra_> gumstix ?
Jun 22 15:13:06 <ogra_> should all be in mainline
Jun 22 15:13:07 <infinity> Or a device.
Jun 22 15:13:15 <cvanvliet> ok
Jun 22 15:13:46 <cvanvliet> the base is a Tobi board
Jun 22 15:14:20 <ogra_> never heard of that
Jun 22 15:16:01 <cvanvliet> I have to reconsider somethings now (re support of armel)
Jun 22 15:16:47 <cvanvliet> wish they would just release the drivers for armhf
Jun 22 15:17:06 <ogra_> they will at some point, all distros are switching
Jun 22 15:20:01 <cvanvliet> I feel that way to, but, I am trying to find the best tools for use (this is for my business)
Jun 22 15:21:41 <ogra_> ppisati, hmm, https://lists.ubuntu.com/archives/ubuntu-release/2012-June/001440.html only talks about a 3.4 tilt kernel
Jun 22 15:22:02 <infinity> cvanvliet: Just keep pestering people in #beagle until an sgx drop appears, and let me know? ;)
Jun 22 15:22:21 <cvanvliet> infinity, you will know within seconds
Jun 22 15:22:46 <cvanvliet> I may also try Imagination (the powervr guys)
Jun 22 15:22:55 <infinity> We have PVR on armhf.
Jun 22 15:23:06 <infinity> Err, for omap4, that is.
Jun 22 15:23:10 <ogra_> just not for the chip in the beagle
Jun 22 15:23:12 <infinity> YOu could just switch SoCs. :)
Jun 22 15:23:12 <cvanvliet> they are in the UK, and may have a symapthetic ear for a UK startup :p
Jun 22 15:23:18 <cvanvliet> sorta
Jun 22 15:23:46 <cvanvliet> that was my thinking above in my reconsider
Jun 22 15:23:58 <infinity> I suspect they won't do much for you.  Their licensees are the ones who are responsible for code drops.
Jun 22 15:24:08 <infinity> (In this case, TI)
Jun 22 15:24:30 <cvanvliet> ok
Jun 22 15:24:37 <ppisati> ogra_: it's a 3.4 + all linaro stuff
Jun 22 15:24:55 <ogra_> ah, k
Jun 22 15:25:20 <ppisati> ndec: AFAYK, FB_OMAP + DSS is enough to get decent video support for omap3, right?
Jun 22 15:25:33 <ppisati> ndec: and what about DRM_OMAP?
Jun 22 15:25:45 <ndec> well, i think robclark tested DRM_OMAP on beagle...
Jun 22 15:25:46 <ppisati> ndec: DRM_OMAP and FB_OMAP can't live together
Jun 22 15:25:58 <ndec> if that works, that's what we should be using i think...
Jun 22 15:26:02 <ogra_> ppisati, and very very likely three tons of commandline options you need to set to even see a single pixel :)
Jun 22 15:26:12 <cvanvliet> I may be wrong but omap4 is not industrial grade
Jun 22 15:26:18 <ogra_> if not five tons
Jun 22 15:26:25 <ogra_> (metric tons)
Jun 22 15:26:32 <infinity> cvanvliet: Define "industrial grade".
Jun 22 15:26:35 <ppisati> the problem is that, FB_OMAP i get /dev/fb, right?
Jun 22 15:26:50 <ppisati> while with DRM_OMAP do i need an external video driver? like pvr or what?
Jun 22 15:26:54 <ogra_> well, fb0
Jun 22 15:27:08 <ogra_> but yeh, that should be the difference
Jun 22 15:27:20 <cvanvliet> infinity, that was the answer I got form our supplier, only yesterday, I have to follow that up
Jun 22 15:27:22 <ogra_> (for details ask koen in #beagle ;) )
Jun 22 15:27:42 <robclark> w/ omapdrm you will get a /dev/fb0 as well.. and xf86-video-omap will work without pvr (but without accel)
Jun 22 15:28:05 <infinity> cvanvliet: I mean, what do you mean by "industrial grade".  Are you building life-and-death-critical systems, telco backbones...?
Jun 22 15:28:08 <ppisati> robclark: so, which one is better/the right one? DRM or FB_OMAP?
Jun 22 15:28:25 <ogra_> robclark, on omap3 ?
Jun 22 15:28:26 <infinity> cvanvliet: Or is it just about wanting "reliable parts"?  In which case, omap4 seems to be good enough for Amazon and HP (and many others).
Jun 22 15:28:49 <infinity> ogra_: He claims all the way back to omap2, though not tested. ;)
Jun 22 15:28:58 <ogra_> heh
Jun 22 15:29:43 <robclark> ogra_, it should work on omap3..  admittedly it gets tested more on omap4, but I'll fix bugs if there are any on omap3
Jun 22 15:30:39 <robclark> infinity, some of the o3 based parts (and other various parts) are avail in industrial package for harsh environments
Jun 22 15:30:58 <infinity> robclark: Ahh, kay.
Jun 22 15:31:11 <infinity> robclark: Again, I'm curious if that's actually what cvanvliet needs.
Jun 22 15:31:12 <cvanvliet> infinity, yeah our supplier is generally in that area, and we may be at a later point, but not right now.
Jun 22 15:31:41 <cvanvliet> but, they already haev a product, very similar to what we need
Jun 22 15:32:00 <cvanvliet> so we don’t have the engineering costs
Jun 22 15:32:12 <cvanvliet> and we can tweak it a little
Jun 22 15:32:34 <cvanvliet> to suite our stuff, it is based on omap3 (igep boards)
Jun 22 15:33:47 <cvanvliet> infinity, it is more which is the easiest (and cheapest ) way for a self backed start up :)
Jun 22 15:34:34 <infinity> cvanvliet: Absolutely.  Build-your-own hardware isn't a sane way to startup, unless your business model is, in fact, designing hardware.
Jun 22 15:34:34 <cvanvliet> but our stuff is not life or death, that is for sure
Jun 22 15:35:53 <cvanvliet> so that is why I want 12.04, armel, SGX in omap3
Jun 22 15:36:15 <cvanvliet> we do a lot of 3d display
Jun 22 15:37:24 <infinity> cvanvliet: Check.  Well, we'd all prefer if we could make that be armhf (but until we see an sgx driver, meh), but there's certainly no one stopping you from using 12.04 armel.  It's not "officially supported", but it's not going away for 5y either.
Jun 22 15:38:18 <cvanvliet> i just need to get it going ;)
Jun 22 15:38:52 <cvanvliet> infinity, we make things like hawk-eye for sports
Jun 22 15:39:20 <infinity> cvanvliet: Sounds fun.
Jun 22 15:39:40 <cvanvliet> accelerometers, forceplates,  camera etc
Jun 22 15:39:53 <cvanvliet> starting in golf industry
Jun 22 15:40:30 <infinity> Sounds like an industry where price-point isn't dead critical, and omap4 might be viable.
Jun 22 15:40:36 <cvanvliet> yes
Jun 22 15:40:38 <infinity> (And I can't think insane durability is important either)
Jun 22 15:41:12 <infinity> Not that I'm recommending you start from scratch. :P
Jun 22 15:41:20 <GrueMaster> Depends on if you are hitting the device with a driver.
Jun 22 15:41:28 <cvanvliet> we have enough to do as it is
Jun 22 15:41:30 <infinity> Starting from scratch every 6 months leads to Duke Nukem Forever disease.
Jun 22 15:41:52 <GrueMaster> Most electronics don't take well to being clubbed.
Jun 22 15:41:53 * ogra_ prefers hiting devices with wrenches
Jun 22 15:42:16 <cvanvliet> GrueMaster, Xbee plus arduino can fly 30 metres ;)
Jun 22 15:42:36 <cvanvliet> when gaffa tape breaks on a golf club :p
Jun 22 15:42:50 <GrueMaster> Heh
Jun 22 15:42:52 <ogra_> drivers look so new-rich ... wrenches are more "male" :)
Jun 22 15:43:05 <ppisati> uhm
Jun 22 15:43:13 <ppisati> is it normal for my board to use fbdev?
Jun 22 15:43:20 <ogra_> yes
Jun 22 15:43:25 <ppisati> well
Jun 22 15:43:32 <ogra_> on all arm boards
Jun 22 15:43:36 <ppisati> X says i'm having a 1280x720 display
Jun 22 15:43:43 <ppisati> but to me it looks like 640x480
Jun 22 15:43:47 <ogra_> (until we are installing SGX or whatever)
Jun 22 15:43:56 <ogra_> thats omap3 ?
Jun 22 15:43:59 <ppisati> yes
Jun 22 15:44:09 <ogra_> well, as i said, three tons of cdmline options
Jun 22 15:44:33 <ppisati> i'm using FB_OMAP here
Jun 22 15:44:44 <ppisati> with
Jun 22 15:44:46 <ppisati> vram=12M omapfb.mode=dvi:1280x720MR-16@60
Jun 22 15:44:51 <ppisati> isn'it enough?
Jun 22 15:44:57 <ppisati> and what do we use for our installer?
Jun 22 15:45:06 <ogra_> nothing
Jun 22 15:45:16 <ppisati> so how does it work?
Jun 22 15:45:24 <ogra_> we always asked the kernel team to set a proper default :P
Jun 22 15:45:43 <ppisati> uh?
Jun 22 15:46:03 <ogra_> and with the merge of omap in mainline that default should have been carried over
Jun 22 15:46:20 <ogra_> it likely wasnt though and we need to set a cmdline option again to get a proper mode
Jun 22 15:46:43 <ogra_> i remember in the first images we had omapfb.mode=dvi:1280x720MR-16@60
Jun 22 15:46:56 <ogra_> but i'm not even sure the option is still called like that
Jun 22 15:47:12 <rcn-ee> it is, just not for omapdrm/kms..
Jun 22 15:47:47 <GrueMaster> Note that the above setting is for wide screen.  Looked like crap on my 19" standard LCD monitors.
Jun 22 15:47:50 <ogra_> well, he built with FB_OMAP
Jun 22 15:48:00 <rcn-ee> yeap, so it should work. ;)
Jun 22 15:48:01 <ogra_> not DSS
Jun 22 15:48:09 <GrueMaster> I used 1280x1024.
Jun 22 15:48:16 <ppisati> so, in P we were using FB_OMAP
Jun 22 15:48:24 <ppisati> so, either we passed something in the cmdline
Jun 22 15:48:35 <ppisati> or i don't know
Jun 22 15:49:25 * ppisati trying with a kernel with FB_OMAP
Jun 22 15:49:26 <rcn-ee> once it's up, can you access /proc/cmdline...
Jun 22 15:49:44 <ppisati> rcn-ee: talking about the installer
Jun 22 15:50:33 <rcn-ee> the older d-i? or the live-cd thing..
Jun 22 15:50:51 <ogra_> ogra@nusakan:~/branches/debian-cd$ grep omapfb tools/boot/quantal/post-boot-armhf+omap
Jun 22 15:50:51 <ogra_>     v_opts="vram=12M omapfb.mode=dvi:1280x720MR-16@60"
Jun 22 15:50:54 <ppisati> the one that we use with the preinstalled-omap image
Jun 22 15:51:02 <ogra_> ppisati, you are right, we do set it on omap3
Jun 22 15:51:08 <ppisati> ok, cool
Jun 22 15:51:18 <ogra_> probably needs some adjustments
Jun 22 15:51:57 <GrueMaster> Too bad it can't do autodetect.
Jun 22 15:52:18 <ogra_> i was promised autodetection for ages
Jun 22 15:52:35 <rcn-ee> it was set to that, since the default u-boot value was so small.. http://git.denx.de/?p=u-boot.git;a=blob;f=include/configs/omap3_beagle.h;hb=HEAD#l229
Jun 22 15:52:42 <ogra_> a bitz longe than i was promised unity 3d on arm (which only twook 2 years)
Jun 22 15:52:57 <rcn-ee> actually autodetection works pretty good with v3.3/3.4 (omapdrm/kms) on the beagle. ;)
Jun 22 15:53:24 <ogra_> "pretty good" ?
Jun 22 15:53:34 <ogra_> like every tenth time ? on every second board ?
Jun 22 15:53:40 <ogra_> or is it already reliable
Jun 22 15:53:45 <ppisati> rcn-ee: so, what's the difference between drm and fb_omap?
Jun 22 15:53:50 <rcn-ee> works on my 4-5 monitors.. (but until recently it wasn't stable enough to push to all users)
Jun 22 15:54:21 <rcn-ee> both use omapdss internally. ;)
Jun 22 15:55:02 <rcn-ee> ppisati, you better ask robclark, he wrote alot of the omapdrm/kms stuff.. i just enable/tweak it...
Jun 22 15:56:02 <ppisati> robclark: ^
Jun 22 15:56:35 <robclark> omapfb is using old fbdev driver infrastructure/interface..  omapdrm is using drm/kms..
Jun 22 15:56:48 <robclark> short version, w/ omapdrm you can get xrandr
Jun 22 15:57:13 <robclark> (and dri2, hotplug detection, etc)
Jun 22 15:57:19 <ogra_> well, if its stable i'm all for switching
Jun 22 15:57:56 <ppisati> robclark: any drawback?
Jun 22 15:58:11 <ppisati> btw, just compiled in fb_omap, X is up, but i don't have any output
Jun 22 15:59:38 <robclark> well.. omapdrm is newer... sometimes exposes some issues in omapdss, relies on monitors to have sane EDID (although more recently there is some drm bootargs added to allow reading EDID from firmware file to deal with broken monitors)
Jun 22 16:03:28 <ppisati> omapfb omapfb: no driver for display: dvi
Jun 22 16:03:30 <ppisati> uhm
Jun 22 16:03:55 <rcn-ee> ppisati, this is 3.4 or 3.5?
Jun 22 16:04:03 <ppisati> 3.5
Jun 22 16:04:39 <rcn-ee> ah, haven't figured that one out yet, it was broken in rc3 last tried, sorry..
Jun 22 16:05:17 <ppisati> rcn-ee: are you saying fb_omap is broken somehow?
Jun 22 16:06:36 <rcn-ee> with 3.5-rc3, it oppes'ed on my beagle_xm, and the display didn't come up, didn't dig into to much as i'm mostly on 3.4.x at the moment.. but the stuff for rc4 probally fixed that..
Jun 22 16:43:20 <ppisati> this time fb seems to be ok
Jun 22 16:43:25 <ppisati> fbcvt: 1280x720@60: CVT Name - .921M9-R
Jun 22 16:43:34 <ppisati> but i still get an ugly 640x480
Jun 22 16:43:47 <ppisati> any idea how to debug this?
Jun 22 16:55:29 <GrueMaster> ppisati: What does xdpyinfo thing the resolution is?  And does your monitor display the incoming signal info?
Jun 22 17:18:19 <rsalveti> ogra_: edid autodetection should work on omap3 just fine
Jun 22 17:18:29 <rsalveti> it was supported by upstream since 3.2 I believe
Jun 22 17:20:28 <rsalveti> if you're using omapfb with panel-dvi you should also get edid detection
Jun 22 17:21:05 <ppisati> rsalveti: panel-dvi?
Jun 22 17:21:30 <rsalveti> CONFIG_PANEL_DVI
Jun 22 17:21:40 <rsalveti> drivers/video/omap2/displays/panel-dvi.c
Jun 22 17:21:58 <rsalveti> this is based on an implementation I had for quite a while ago
Jun 22 17:22:10 <ppisati> rsalveti: it seems they killed it in 3.5
Jun 22 17:22:32 <ppisati> ls -la drivers/video/omap2/displays/*dvi* | wc -l
Jun 22 17:22:33 <ppisati> 0
Jun 22 17:22:43 <rsalveti> could be, let me check
Jun 22 17:23:01 <rsalveti> omapdrm should probably be replacing most of the stuff
Jun 22 17:23:18 <ppisati> trying that right now
Jun 22 17:25:07 <ppisati> PANEL_TFP410?
Jun 22 17:26:36 <ppisati> ok, it was renamed
Jun 22 17:33:42 <rsalveti> ppisati: yup
Jun 22 17:34:49 <ppisati> rsalveti: so i assume i need to change cmdline too
Jun 22 17:35:09 <ppisati> rsalveti: from omapfb.mode=dvi:1280x720MR-16@60 to omapfb.mode=tfp410:...
Jun 22 17:36:41 <ppisati> btw, enabling DRM_OMAP didn't give me any /dev/fb*, what's wrong?
Jun 22 17:36:43 <ppisati> robclark: ^
Jun 22 17:38:36 <robclark> ppisati, if using omapdrm, then omapfb.mode won't do anything.. but you also shouldn't need it..
Jun 22 17:38:53 <robclark> but possibly the kernel is not registering the platform device?
Jun 22 17:39:00 <rsalveti> ppisati: you don't need any cmdline for this panel as well
Jun 22 17:39:19 <ppisati> wait
Jun 22 17:39:36 * robclark wonders what kernel we are talking about?
Jun 22 17:39:57 <ppisati> robclark: didn't you say that with DRM_OMAP we would get a /dev/fb, right?
Jun 22 17:40:19 <ppisati> robclark: foerget about the omapfb stuff, i'm testing DRM and FB_OMAP in parallel
Jun 22 17:40:25 <ppisati> 3.5rc3
Jun 22 17:40:35 <robclark> yes, you will..  but only if the driver is loaded (which it won't be if the device is not registered)
Jun 22 17:40:44 <robclark> if pure upstream kernel, there is one patch you need
Jun 22 17:40:55 <ppisati> robclark: it's pure upstream
Jun 22 17:41:02 <robclark> k, hang on
Jun 22 17:41:48 <ppisati> rsalveti: on the other hand, with FB_OMAP we always passed omapfb.etcetc
Jun 22 17:42:03 <rsalveti> ppisati: but it's not needed with this panel, afaik
Jun 22 17:42:12 <ppisati> rsalveti: uhm ok
Jun 22 17:42:14 <robclark> ppisati, http://permalink.gmane.org/gmane.linux.ports.arm.omap/77544
Jun 22 17:43:18 <ppisati> robclark: ack, i'll try
Jun 22 18:55:15 <ppisati> robclark: added your patch on top of 3.5rc3, got /dev/fb0 but still no output
Jun 22 18:55:41 <ppisati> anywa, i'm off for the day
Jun 22 18:56:14 <robclark> ppisati, I guess next step, turning on omapdss.debug=1 and possibly drm.debug=7 in bootargs and then start havign look at the log
Jun 22 19:24:18 <Pokinawa> hello people....
Jun 22 19:25:10 <Pokinawa> anyone here work at ARM? :P
**** ENDING LOGGING AT Sat Jun 23 02:59:58 2012