**** BEGIN LOGGING AT Tue Nov 02 02:59:57 2010 Nov 02 07:23:02 amitk, http://netbook-remix.archive.canonical.com/updates/pool/public/l/linux-fsl-imx51/ has a couple different trees for imx51 stuff. I think "araneo" is the PC-Z1 and "sendai" the PC-T1, but I'm not 100% sure. Nov 02 08:28:56 morning Nov 02 10:16:16 does anyone here have any experience setting up the TI SDP 4430 (development board)? Nov 02 10:26:22 alf_: you may try on #linux-omap channel as well which may have more SDP4430 users Nov 02 10:27:05 sebjan_: thanks Nov 02 11:00:35 someone remember where I find panda kernel package? Nov 02 11:08:28 hrw: it's linux-ti-omap4 source package. Is this what you look for? Nov 02 11:09:40 was looking for binaries rather but found them already Nov 02 11:10:08 ubuntu/arm concentrates on 'install on sd' way when I have different needs now Nov 02 11:10:45 512MB sd as /boot + 10GB partition on usb harddrive as / + 300GB /home on same hdd Nov 02 11:16:19 hrw: right now, easiest may be to first install on SD, and then move the ext3 partition to your USB storage (while keeping the bootloaders and kernel on the SD card). Nov 02 11:44:49 ~curse cdimages.ubuntu.com mirrors for lack of armel Nov 02 11:54:37 you... you don't make your own images? Nov 02 11:54:40 for _shame_ Nov 02 11:55:44 cwillu: I would rather prefer to grab uimage+uinitrd and boot into installer Nov 02 11:56:04 similar to x86(-64) and debian/arm Nov 02 11:56:44 but instead I fetched 512MB image which unpacks to 2GB image which dds to 4GB card and will be copied to 10GB partition after boot Nov 02 11:57:02 btrfs root, eh? Nov 02 11:57:08 no, ext4 Nov 02 11:58:09 ext4 snapshotted into btrfs? Nov 02 11:59:12 no, no no Nov 02 11:59:30 no btrfs for me until it will get 3 kernel releases without experimental sign Nov 02 12:03:23 suit yourself Nov 02 12:04:02 I prefer my filesystem developers to be overly cautious with that tag than under cautious, like some ext related filesystems I know :p Nov 02 12:04:09 er, unlike Nov 02 12:07:20 ok, pandaboard booted Nov 02 12:52:53 * hrw upgrades pandaboard to natty Nov 02 12:58:14 btw - why only 665MB ram on pandaboard? Nov 02 13:21:56 o... people arrive Nov 02 13:34:25 hrw, I don't remember the details, but I think there were prior conversations about some boot parameter that helped decide how much memory to allocate for dedicated buffers and how much for system RAM. Check the documentation carefully. Nov 02 13:35:26 persia: do you know how the 3PA works? I wonder if we can upload natty source package and test how it builds. Seems our 3PA are set to Maverick only... Nov 02 13:36:15 berco, I don't know much about them. I think that you can upload to any release, and it will do the right thing (just use changelog). You can probably get a more authoritative answer in #launchpad. Nov 02 13:36:47 persia: ok. I'll give it a try first and see how it goes. Nov 02 13:37:51 persia: I thought there was a checkbox or something in the private ppa (3pa) to select build distro and arch? Nov 02 13:39:54 GrueMaster, Could be. I'm unsure. I know that one has to set the target in the changelog for an upload, but don't know if one has to pre-enable first. Nov 02 13:40:09 ah. Nov 02 13:40:59 GrueMaster: that might a be at the admin level... I browsed all the links on launchpad and didn't see such option I could set. However the sources.list only displays maverick in the tech details link Nov 02 13:41:23 persia: will check Nov 02 13:42:14 ok. I personally haven't done anything with ppa building, but I thought I had seen something like that. Nov 02 13:42:40 berco, the "technical details" link is likely to be based on either stock text, or what is there today, rather than being a good guide for what can be there. Nov 02 13:42:59 :) Nov 02 14:10:25 why the hell does the UDF keynote from Mark contains copyrighted material from MontyPython? Nov 02 14:10:27 http://www.youtube.com/watch?v=YUAzicy_01o Nov 02 14:11:18 nice :D Nov 02 14:12:22 hey armin76! got a log for me? Nov 02 14:18:08 (not available at least in germany) Nov 02 14:21:00 * persia didn't think copyright blocked use of small excerpts in other contexts, but isn't qualified to have an actual opinion (but encourages others to either be qualified or seek counsel) Nov 02 14:25:59 hmm, the dead distro sketch :-D Nov 02 14:28:53 on the other hand, the intro sounds like a song form a "little big planet", but maybe I'm also not qualified Nov 02 14:46:27 btw - is there any work on getting normal ubuntu installer working on armel targets? Nov 02 14:50:17 hi prpplague Nov 02 14:50:24 hrw, Last time I played with it, it worked fine. Are you having issues? Nov 02 14:50:50 hrw: hey Nov 02 14:52:01 persia: I got a feeling that the only way to install ubuntu on omap3/4 is to use preinstalled images or rootstock Nov 02 14:53:00 persia: installing by booting uimage+uinitrd would be lovely Nov 02 14:53:01 hrw: That is currently correct. I haven't tested the d-i installer on omap4 yet. Nov 02 14:53:09 hrw, That's because some folk are convinced that omap3/4 can only boot from the SD card, and haven't read the docs that explain how to boot from USB. Nov 02 14:54:16 Since the installer has issues when trying to install to the media from which the installer runs, unless using the HD installer target, and folk felt that using the HD installer target would be slow and annoying because of file copying, they invented preinstalls. Nov 02 14:54:29 persia: I am fine with having /boot on sd Nov 02 14:55:03 That won't work. Nov 02 14:55:35 why? Nov 02 14:55:58 uboot reads kernel+initrd from sd and boots to rootfs on usb-hdd Nov 02 14:57:09 Right, but it won't work with the way the installer works, because it wants to format /boot and *install* a kernel to it. Nov 02 14:57:40 should work when booted kernel+initrd from sd Nov 02 14:57:48 So, if you happen to have enough internal flash, and want to boot from that, it works fine. If you set up a system manually, it works just fine. If you set up an HD-style install from SD, it works just fine. Nov 02 14:57:59 It just doesn't happen to have any way to make the SD /boot Nov 02 14:58:04 as there should be nothing needed on sd during boot Nov 02 14:58:26 The problem is that the boot media needs to be vfat. Nov 02 14:59:06 And it is a broken vfat implementation at that. Nov 02 14:59:33 which can be generated anyway Nov 02 14:59:41 (at least that is where MLO and u-boot need to be). Nov 02 15:00:43 It can be easily generated, and even setup to mount as /boot, but I believe it is against security policies. Nov 02 15:00:52 I'd have to ask our security guru. Nov 02 15:01:55 I do know that updating isn't as easy as just copying new files over, due to the vfat bug issue in silicon. Nov 02 15:02:39 The partition needs to be wiped and rewritten when updating some of the files. Nov 02 15:04:31 ok, so something to try later Nov 02 15:05:06 will have to dig my straight serial cable too Nov 02 15:05:46 I believe we have plans for getting the debian installer to fully support omap4 this cycle, which will make it easier to do alternative installations (netboot installs, etc). Nov 02 15:06:42 good Nov 02 15:12:21 hi robclark Nov 02 15:12:34 hi hrw Nov 02 15:31:04 GrueMaster: where I can find files needed for installation on pandaboard? Nov 02 15:31:29 What do you mean? ppa files? Nov 02 15:31:45 sorry Nov 02 15:32:15 I meant: where to find (or how to generate) installer kernel/initrd Nov 02 15:33:17 Not sure. what are you trying to do? Nov 02 15:34:02 play with other (then preinstall sd) methods of installing ubuntu on pandaboard Nov 02 15:34:15 Ah. Nov 02 15:35:08 That I don't know. I think that they need to be generated by the d-i build process. Not my area of expertise. Nov 02 15:35:24 ok Nov 02 15:36:39 ogra, rsavetti, and ncommander would know more, but I think they are in transit to Boston for Linux Plumbers Conference. Nov 02 15:36:47 ok Nov 02 15:37:17 hrw, http://ports.ubuntu.com/ubuntu-ports/dists/maverick/main/installer-armel/current/images/omap/netboot/omap/ has some omap3 images. Nov 02 15:37:28 Getting omap4 images probably means fiddling with the image build tools. Nov 02 15:37:59 persia: afaik, d-i doesn't currently support omap4. Nov 02 15:38:04 Which is needed. Nov 02 15:38:40 GrueMaster, which part? Are there bugs? I suspect that at least large chunks of d-i *do* support omap4, or oem-config wouldn't work (as it uses d-i as a backend) Nov 02 15:38:49 anyway - it booted from sd to my harddisk Nov 02 15:39:22 oem-install doesn't do any hw-config. It only sets up locale, keyboard, timezone, and user. Nov 02 15:39:52 (i.e. no installation steps) Nov 02 15:40:17 hw installation, that is. Nov 02 15:41:29 GrueMaster, Lots of stuff oem-config does aren't expressed in the UI, but yeah, it's a subset. Nov 02 15:42:16 Well, the parts that we need for installing on new hw aren't supported yet. That is the key. Nov 02 15:51:59 rsalveti: Are you around? Nov 02 15:52:38 Does anyone have a known good MLO and u-boot.bin for the 6 layer? Nov 02 15:56:34 I think the 6 layer was abandoned. Nov 02 15:57:06 But you might be able to pull an earlier version from the archive. Nov 02 16:03:00 But I only have 6 layer and I need to test something Nov 02 16:03:32 You probably have to dig up old stuff out of launchpad and hack an image. Nov 02 16:06:28 lag, if you can pull a package from the archive that is from between Beta and RC, it should work. Otherwise, pull the RC1 image and run dist-upgrade on it. Nov 02 16:07:09 Can't guarantee kernel support though (you know how those kernel guys are). :P Nov 02 16:07:09 I'll do that - cheers Nov 02 16:08:39 What do you need to test? Nov 02 16:10:08 Linaro Nov 02 16:10:28 Ewww. I'm getting page allocation failures in the most recent kernel. Nov 02 16:10:52 I'm unsure whether they have the correct bootloaders Nov 02 16:11:22 They should. I believe that is what we have been using since pre-RC. Nov 02 16:11:42 I could be wrong, though. Nov 02 16:12:39 afair, the changes dropping 6L support were done during the TI-sprint. Nov 02 16:15:43 GrueMaster: Where are the histories? Nov 02 16:15:54 I can only fine the latest builds Nov 02 16:16:08 lag: I can point you to the MLO commit that led to stop supporting 6-layers if it can help. Nov 02 16:18:30 lag: I don't know. I don't see them on ports, which is strange. I have a copy on my mirror though. Nov 02 16:20:49 I don't even really know which we are using, but I would assume it is pool/main/x/x-loader-omap4. Nov 02 16:20:54 There's no history in the archives. Image history is limited to a day or two for selected flavours. LP has richer history. Nov 02 16:21:08 Ah. Nov 02 16:21:33 * GrueMaster needs to do some house-cleaning of private mirror. Nov 02 18:59:15 someone remembers how to get whole 1GB ram on panda? Nov 02 18:59:59 use an updated u-boot/x-loader? Nov 02 19:00:53 armin76: I thought that maverick release had proper ones Nov 02 19:01:14 dunno, blame ogra :P Nov 02 19:02:08 will, when he will be back Nov 02 19:03:57 btw: I am going thought kde components now as they did not built on arm Nov 02 19:04:16 how to request rebuild attempt for armel? Nov 02 19:54:34 hrw|gone, I'd suggest talking with the folk in #kubuntu-devel: a fair number of them can press the button, and they are also tracking KDE-FTBFS-on-armel issues quite closely Nov 02 20:16:39 ok Nov 02 20:17:00 persia: I am testing fixes now. will talk there tomorrow Nov 02 23:03:26 what is good linux serial logger program? Nov 02 23:03:52 I have a gadget device /dev/ttyACM0 and want to see what's coming in Nov 02 23:04:03 and I hate minicom Nov 02 23:04:59 I would be interested in knowing that aswell, the only other I've tried is the gtkterminal thing. But minicom is far better than that Nov 02 23:05:27 * persia remembers plars packaging something recently and hunts Nov 02 23:05:49 Have either of you tried moserial? Nov 02 23:05:59 I also know a few folk like to use screen for that. Nov 02 23:06:30 the problem is that the device should log kernel boot messages Nov 02 23:06:55 and it is only created a short time before logging starts Nov 02 23:07:04 and then its over again Nov 02 23:07:23 need something which can listen to a non existing device Nov 02 23:07:28 Hmm. That gets tricky. Nov 02 23:07:30 and start printing when it comes up Nov 02 23:08:18 If you just want a log, you might be able to do something with while(1) polling for device existence and then dumping output. Nov 02 23:08:38 tail --retry might also help. Nov 02 23:08:48 and how to set the serial parameters? Nov 02 23:08:48 But either solution is subject to race conditions. Nov 02 23:16:49 mmh, the "while poll loop" and cat /dev/ttyACM0 thing does not work :-( Nov 02 23:17:42 That's probably the result of incorrect default serial parameters (and I don't know how to set those: I've not played much with serial ports in a long time) Nov 02 23:29:55 google says to start minicom and configure it there, then quit and use e.g. cat Nov 02 23:30:24 found a program called "cu" from the lrzsz package Nov 02 23:31:38 heh. `cu` is old-school indeed. Nov 02 23:37:55 I used to always use cu, much better than minicom, these days I typically just use screen Nov 02 23:38:08 cu is nice for logging though, in conjunction with script Nov 02 23:38:38 So while until it exists, then cu to grab the data? Nov 02 23:38:39 looks like ttylog should do the job better Nov 02 23:39:03 while loop and ttylog Nov 02 23:39:13 but I still don't get any output Nov 02 23:39:21 maybe nothing is written to the device Nov 02 23:39:26 And you're sure some exists? Nov 02 23:40:52 yeah, the device is created Nov 02 23:41:03 but I don't know if the kernel on the other side sends something Nov 02 23:41:30 I put console=ttyGS0,115200n8 to the cmdline Nov 02 23:41:50 What kernel are you using? I'm not sure the Ubuntu kernels handle gadget-as-serial-console-on-boot very well yet. Nov 02 23:43:07 it's not an ubuntu kernel, but a "future" ubuntu laptop :-) Nov 02 23:43:21 kernel is from android.git.kernel.org Nov 02 23:43:56 kernel has the gadget is compiled in Nov 02 23:45:47 Which laptop? Nov 02 23:46:58 ac100 Nov 02 23:47:07 heh Nov 02 23:47:12 cool Nov 02 23:47:56 Of course, it would come out 3 weeks after I bought a new Acer Aspire One. Nov 02 23:49:42 Yeah. The ac100 is *really* close. Just needs a more recent kernel (although there were some folk at UDS talking about trying to upload a .29) Nov 02 23:51:12 well, 2.6.35 is at least booting to the point it initialized the gadget ... Nov 02 23:51:37 Excellent! Nov 02 23:52:00 Is that Android, or mainline+patches? Nov 02 23:52:17 the latter Nov 02 23:52:29 Oh, lovely. With Ubuntu sauce also? Nov 02 23:52:36 no, http://android.git.kernel.org/?p=kernel/tegra.git;a=summary Nov 02 23:52:48 + a proper machine description Nov 02 23:53:23 but it seems that these patches will get merged sooner or later Nov 03 01:22:00 good evening guys Nov 03 01:22:08 hey mjeamiguel Nov 03 01:22:15 hi persia Nov 03 01:22:27 i just have a noob question Nov 03 01:22:33 which one :) Nov 03 01:23:12 ok, I downloaded this file dhrystone v2.2 Nov 03 01:23:27 it runs when sh is invoked Nov 03 01:23:35 then produces 3 files Nov 03 01:23:49 I am testing it on my x86 machine Nov 03 01:24:16 now my question is, do i need edit the file if i decide to run it on ARM? Nov 03 01:24:23 like cc options edit? Nov 03 01:24:40 i can pastebin the drystone now..gimme a sec Nov 03 01:25:56 here is the file http://paste.ubuntu.com/524783/ Nov 03 01:28:04 Oh, my. What an amusing way to write C code. I7d sugget just trying it. Nov 03 01:28:22 If it doesn't work, then it might be an issue, but there's a good chance it will work. Nov 03 01:28:33 but yeah Nov 03 01:28:37 looking into the code Nov 03 01:28:51 the first few lines are to compile the dhrystone right? Nov 03 01:29:17 but without optimizing it for arm, would it be possible to get proper results? Nov 03 01:29:32 Yeah, it's a shell script *AND* C code, so the shell gets escaped when the C is compiled, and the C gets ignored when the shell is run. Nov 03 01:29:39 it worked on my x86 laptop (dual core) Nov 03 01:29:40 And the shell compiles the C. Nov 03 01:29:45 yes :) Nov 03 01:30:47 is this script good enough to benchmark a beagleboard? Nov 03 01:30:53 or I need to modify it? Nov 03 01:31:19 Depends what you're trying to achieve. Nov 03 01:31:37 If you believe that code accurately reflects some workload you intend to run, then use it unmodified. Nov 03 01:31:45 If you believe it doesn't, it makes for a poor benchmark. Nov 03 01:32:22 oohh ok..any suggestions for armv5 benchmark scripts? Nov 03 01:32:35 Why do you want a benchmark? Nov 03 01:32:43 (and Ubuntu is ARMv7a these days) Nov 03 01:33:03 yeah im sorry bout that, I wanted to try it both Nov 03 01:33:27 its a school project and at the same time, my curiosity Nov 03 01:33:55 OK. Let's start at the beginning :) What are you trying to accomplish? Nov 03 01:35:03 cool :) I wanted to compare the architecture (5 and 7) through bm Nov 03 01:35:12 Ah, OK. Nov 03 01:35:18 yes sir :) Nov 03 01:35:44 Looking to compare different hardware running the same code, or different code opimisation on the same hardware? Nov 03 01:35:44 I havent started yet..i just want to know where Im starting and heading eventually Nov 03 01:37:16 2 different hardware, same codes, speed of execution for compiling ..etc Nov 03 01:37:33 for now, just contrasting both Nov 03 01:37:39 Are both hardware solutions ARMv7a compatible? Nov 03 01:37:57 yes its beagleboardXM Nov 03 01:38:18 my teacher told me that 1 board is already running ..but uses armv5 Nov 03 01:38:37 Hrm. OK. Nov 03 01:38:42 more like optimized for armv5 Nov 03 01:38:44 yes Nov 03 01:38:57 but yeah Nov 03 01:39:00 if i can only Nov 03 01:39:10 get a script for benchmarking Nov 03 01:39:22 i just needed data comparisson.. Nov 03 01:39:24 So there's some stickiness if you're running an ARMv5 kernel and make certain sorts of syscalls from ARMv7a+Thumb code, but I think that's mostly the Thumb stuff. Nov 03 01:39:42 yeah..im more like into arm codes Nov 03 01:39:57 You can probably use your dhrystone script if you just want an artificial benchmark, but be aware it may not be meaningful for any real workload. Nov 03 01:40:18 so no need to edit the script for now right? Nov 03 01:40:26 You will want to modify the top of it to force different ISA compliance. Nov 03 01:40:47 But I'm not convinced you can safely run ARMv5 code in Ubuntu: I'd probably use Debian for such a comparison. Nov 03 01:40:48 will look into that :) Nov 03 01:41:03 ill try that and fedora Nov 03 01:41:16 I wouldn't send you that far away :) Nov 03 01:41:18 its just a little confusing for me, considering its my first time with a project Nov 03 01:41:27 hahahah :) thanks Nov 03 01:41:40 one last question Nov 03 01:42:09 Understood. Folk here are likely to be willing to help with some stuff, but nobody will be willing to actually solve some of the basic issues for the project (but they may be willing to help with issues you discover in the base environment whilst working on your project) Nov 03 01:42:14 No limit :) Nov 03 01:42:45 i also wanted to compile a program lets say written for x86 Nov 03 01:42:58 i jsut needed to change the gcc options right? Nov 03 01:43:04 in the make script? Nov 03 01:44:10 Best effort is made to have sane defaults for gcc so you don't have to do that in the common case. Nov 03 01:44:36 If you're working on something involving comparisons of various compiler flags, the best effort defaults won't necessarily help :) Nov 03 01:45:05 lol..im kinda lost there..can you please explain further thanks persia :) Nov 03 01:45:34 OK, so for each architecture, we ship the compiler with default settings that tend to just work. Nov 03 01:45:41 yes Nov 03 01:46:36 So one can generally get nearly best results for "int main(int argc, char* argv[]){printf("Hello World\n");}" or slightly more complex code. Nov 03 01:47:09 However, if you're working on an academic project to compare different compiler flags, it won't help much, because you need to select specific flags, rather than using the defaults. Nov 03 01:48:30 i guess im not so close to that, thanks for explaining :) Nov 03 01:48:45 any good resource for armv7 readings? Nov 03 01:49:00 i already checked the mailing lists Nov 03 01:51:57 I think you'll have to wait for someone else. I just run the software, and try to fix bugs that aren't that ISA-specific. The few things I know about the ISA changes are mostly hearsay or based on other folks solutions to other, similar bugs. Nov 03 01:52:41 persia: I really really apreciate you helping me tonight..thanks so much! Nov 03 01:54:21 I'm to present about my project soon, and will let my fellow students how this channel and you helped me out a lot Nov 03 01:54:29 thank you :) Nov 03 01:56:22 Thanks for helping advertise us. Nov 03 02:02:27 you guys deserve :) bye! **** ENDING LOGGING AT Wed Nov 03 02:59:57 2010