| --- Log opened Sat Sep 28 00:00:16 2013 | ||
| stekern | poke53281: I noticed you are not using the or1k-linux-uclibc target, is there any particular reason for that? | 03:57 |
|---|---|---|
| stekern | ysionneau: good to hear! | 03:58 |
| hansfbaier | stekern: got my sockit today | 04:02 |
| hansfbaier | stekern: Do you have your work with the sockit in a github repo somewhere? | 04:09 |
| stekern | hansfbaier: yes, here: https://github.com/skristiansson/orpsoc-cores | 04:21 |
| stekern | you also might interested in that I'm in the process of uploading prebuilt or1k-elf toolchains to opencores.org ftp | 04:21 |
| hansfbaier | stekern: RPX, the carrier which handles FedEx in Indonesia has the slogan "One stop logistics" on their logo. I think that is very adequate. The way they treat their customers, they surely only want to stop by once. | 04:22 |
| hansfbaier | stekern: sounds good. As you mentioned, juliusb had the same issue with libgui? Has anyone a solution yet? I guess getting to the bottom of this would need debugging the auto-hell scripts ... | 04:23 |
| stekern | juliusb: ^ did you ever manage to compile or1k-elf-gdb? | 04:25 |
| stekern | and did you have that libgui problem (I recall that you did, but I might be wrong) | 04:25 |
| Powermaniac | Hey | 05:02 |
| stekern | hansfbaier: ftp://openrisc.opencores.org/toolchain/gcc-or1k-elf-4.8.1-x86_64.tar.bz2 | 05:08 |
| hansfbaier | stekern: Wow, thanks! | 05:19 |
| Powermaniac | So how is everyone? | 05:39 |
| hansfbaier | stekern: anonymous ftp doesn seem to work (ftp) says login failure | 05:56 |
| hansfbaier | s/doesn/doesn\'t/ | 05:57 |
| stekern | hansfbaier: try with ocuser/ocuser | 06:11 |
| hansfbaier | stekern: works, thanks | 06:24 |
| stekern | I thought they had enabled anon access when it worked here, but I guess it remembered an earlier login | 06:35 |
| stekern | when i clicked the link in the webbrowser i mean | 06:35 |
| stekern | i'll change the links | 06:36 |
| Powermaniac | knz: are you there? | 06:40 |
| hansfbaier | stekern: The dip switches on that page, aren't they the other way round? http://www.rocketboards.org/foswiki/Documentation/ArrowSoCKITEvaluationBoardLinuxGettingStarted | 06:47 |
| poke53281 | stekern: No particular reason. | 06:48 |
| hansfbaier | stekern: seems to be conflicting with the PDF. Do you have a rev c or b board? | 06:48 |
| poke53281 | stekern: This was the official target when I started developing with the toolchain. When I realized that you changed it, it was already too late. I decided not to recompile everything. | 06:49 |
| poke53281 | But it seems to work. | 06:49 |
| poke53281 | stekern: http://simulationcorner.net/toolchain-downloads/ | 06:52 |
| stekern | poke53281: yeah, the differences are minor iirc, uclibc define _UCLIBC_ (or something like that) and insert the right path to ld ( | 06:52 |
| poke53281 | Here is the content of my downloads folder of the toolchain. | 06:52 |
| stekern | so you don't need the symlink) | 06:53 |
| poke53281 | Ahh, Ok | 06:53 |
| stekern | poke53281: thanks! that will save me some trouble | 06:53 |
| Powermaniac | Minecraft Redstone as a circuit explanation doesn't exactly work if the person explaining decides to leave information out... | 06:54 |
| stekern | hansfbaier: not sure, I can check when I'm close to the board | 06:55 |
| hansfbaier | stekern: never mind | 06:55 |
| hansfbaier | I already got it | 06:55 |
| hansfbaier | stekern: working | 06:55 |
| stekern | I have left the dip-switches the way as was instructed in the workshop where I got it | 06:56 |
| hansfbaier | stekern: confusing the web site the user manual all seemed to be conflicting, but the getting started guide worked finaly | 06:56 |
| hansfbaier | finally | 06:56 |
| poke53281 | I used Linux from Scratch to write those scripts. But there are a lot of cross compiling tricks included. Patching with sed and copying of files basically. | 06:58 |
| poke53281 | I think that on the long term porting buildroot http://buildroot.uclibc.org/ would be better in the future. | 06:59 |
| poke53281 | or Gentoo :) | 07:00 |
| stekern | or yocto | 07:02 |
| poke53281 | it looks more advanced then buildroot | 07:11 |
| poke53281 | It looks very similar | 07:16 |
| hansfbaier | poke53281: what do you talk about? | 07:22 |
| poke53281 | find an easy way to cross-compile a whole distribution. | 07:25 |
| poke53281 | at the moment I am talking about yocto. The web page is clear but not the source. Looks more like chaos. | 07:28 |
| stekern | poke53281: altera use it for their bsp packs for socfpga, that's how I stumbled upon it. I didn't end up using it though | 07:36 |
| stekern | installed ubuntu instead, but saw some potential in it for building openrisc bsps | 07:37 |
| olof | Yeah! I'm all for porting gentoo to OpenRISC. The problem however is that even the base system is actually a bit too large | 07:44 |
| olof | But it would be awesome once it's running. Great tools to work with | 07:44 |
| poke53281 | at the moment http://buildroot.uclibc.org/ looks like the more obvious choice for me. | 07:45 |
| poke53281 | But Gentoo would be cool. | 07:45 |
| poke53281 | And why too large? For what? | 07:45 |
| olof | I think that @system needs at least 50 packages, perl included. I'm used to just having busybox+kernel+one or two extra programs | 07:49 |
| olof | But on the other hand, I'm talking to the guy who runs X on OpenRISC :) | 07:49 |
| poke53281 | But in an image of 10MB (compressed) | 07:52 |
| poke53281 | You are right. Gentoo is the opposite of lightweight. | 07:52 |
| hansfbaier | stekern: ah the toolchain is 64 bit :] | 07:53 |
| hansfbaier | stekern: 32 available too | 07:53 |
| hansfbaier | ? | 07:53 |
| poke53281 | But add support for SD-Card or an hard drive and it should be fine. I wouldn't try it on a de0 board | 07:55 |
| hansfbaier | stekern: ah it is | 07:55 |
| hansfbaier | great | 07:55 |
| hansfbaier | never mind | 07:55 |
| olof | NFS is usually how we do it | 07:55 |
| poke53281 | This is fine too. | 07:56 |
| poke53281 | Ok, have to leave for now. | 07:57 |
| olof | Have fun with your sockit. Mine haven't left the box yet since the workshop | 07:58 |
| stekern | hansfbaier: yes, just remove the _64 in the link | 07:59 |
| stekern | there's even an armhf version there too ;) | 08:00 |
| Powermaniac | stekern: Qucik question, I can't find out what the 6-pin conenctor is on the Digilent Atlys in the center-ish of the board, do you know what it is? | 08:04 |
| stekern | Powermaniac: can't remember, check the user manual? | 08:09 |
| Powermaniac | stkern: Already had a look there =\ | 08:09 |
| stekern | check the schematic? | 08:12 |
| Powermaniac | stekern: Check it, haven't entirely understood it but shall try again! | 08:19 |
| Powermaniac | checked* | 08:19 |
| Powermaniac | Weird the schematic says it is conencted to VCC3V3 which I believe is the power. Also GND which is ground obviously, but then 3 connections to a PIC32 | 08:22 |
| Powermaniac | It's on page 6 directly in the center of the schematics on that page if you want to have a look: http://www.digilentinc.com/Data/Products/ATLYS/Atlys_C2_sch.pdf | 08:27 |
| stekern | the pic32 is handling the usb-hid, but it can also load an configuration frm an usb memory | 08:30 |
| stekern | what are tje sognaös called? | 08:31 |
| stekern | the signals | 08:31 |
| Powermaniac | As in the things named R310/200, R311/200, R313/200? | 08:32 |
| Powermaniac | I'm hoping it is another Pmod but a 6 pin Pmod | 08:32 |
| stekern | no, those are resistors | 08:32 |
| Powermaniac | Then I can connect an ethernet port to the 12 pin Pmod on the front and a 6pin PS/2 controller on this Pmod | 08:32 |
| stekern | it's not | 08:33 |
| Powermaniac | Yeah I thought they had to be resistors, resistor sign and R310 just sounds like a resistor | 08:33 |
| Powermaniac | Do you mean where it says PIC32-MCLR? | 08:33 |
| stekern | that much I know | 08:33 |
| Powermaniac | And PIC32-PGD2 | 08:33 |
| Powermaniac | And PIC32-PGC2? | 08:33 |
| stekern | why would you want to connect an ethernet breakoutboard when you already have ethernet on the board??? | 08:34 |
| Powermaniac | 2x ethernet 1 for HDD 1 for wifi adapter (because I'm lazy and can't be bothered sourcing a cheap open source router | 08:34 |
| Powermaniac | Might be forced to do that anyway | 08:35 |
| Powermaniac | Did you connect the PS/2 controller to the 12 pin Pmod? | 08:35 |
| stekern | that makes no sense.. | 08:36 |
| Powermaniac | which part, or both? | 08:36 |
| Powermaniac | xD | 08:36 |
| stekern | just pick any off-the-shelf switch | 08:37 |
| Powermaniac | What exactly do you mean by Switch? | 08:37 |
| stekern | google "ethernet switch" | 08:40 |
| Powermaniac | Yeah just did, I see what you mean now | 08:40 |
| stekern | ok, so the header you were wondering about is probably a programming header for the pic32 | 08:44 |
| Powermaniac | Ahh okay thanks | 08:44 |
| hansfbaier | stekern: the toolchain is good: It blinketh | 09:03 |
| Powermaniac | hansfbaier: You got it working? | 09:03 |
| Powermaniac | If so yay! | 09:03 |
| hansfbaier | Powermaniac: the downloaded | 09:03 |
| hansfbaier | one | 09:03 |
| hansfbaier | Powermaniac: I know what the singularity will do..... guess...... | 09:04 |
| Powermaniac | So, so many possibilities | 09:04 |
| Powermaniac | Where to start | 09:04 |
| hansfbaier | Powermaniac: I think it will blink LEDs | 09:04 |
| hansfbaier | haha | 09:04 |
| Powermaniac | LOL | 09:04 |
| Powermaniac | I personally don't think it will go all terminator like, as well why would it need to... | 09:07 |
| Powermaniac | (Most often thought of outcome of the singularity thought by 'common' people) | 09:07 |
| Powermaniac | I also think it will depend a lot on how it is programmed | 09:08 |
| Powermaniac | And what it's goals are at first anyway | 09:08 |
| Powermaniac | Did you know they are actually working on how to make robots/AI friendly? | 09:08 |
| hansfbaier | Powermaniac: It will have the possibility to choose its own goals. | 09:09 |
| Powermaniac | Well yes...But I don't think it will at first exactly... | 09:10 |
| Powermaniac | Although it all depends really how it comes about | 09:10 |
| stekern | hansfbaier: just like hal 9000 did | 09:26 |
| stekern | poke53281: minor feedback on the build-script, it fails if you don't manually create a uclibc .config | 12:54 |
| stekern | it doesn't bother me much, just thought I'd let you now anyway ;) | 12:54 |
| Powermaniac | What do you guys think of Steam? | 14:49 |
| Powermaniac | And well Valve | 14:49 |
| poke53281 | stekern: I think you will find more errors. | 16:33 |
| poke53281 | But you can start making a list. | 16:36 |
| stekern | I already have a second item on it =P | 16:43 |
| stekern | uclibc doesn't build for some reason | 16:43 |
| poke53281 | ??? Ok, I will try. But I compiled it the last time 2 days ago. Added locale support to try some stuff | 16:50 |
| stekern | maybe it got confused by my copying around configs, I'm restarting now | 16:57 |
| stekern | and if it doesn't build I'll take a closer look | 16:58 |
| stekern | I mean, I have built uclibc a trillion times on this machine ;) | 16:58 |
| stekern | so it should work | 16:58 |
| poke53281 | But the first time with my toolchain in head. | 17:01 |
| stekern | true | 17:02 |
| stekern | ah, I think I know what I'm doing wrong | 17:28 |
| stekern | you've got to manually export the path and sysroot etc, right? | 17:29 |
| stekern | you should make 'make env' suggest the new toolchain first in PATH, not last btw | 17:30 |
| poke53281 | another problem. Remove the lines: export ac_cv_func_malloc_0_nonnull=yes from X.make | 17:41 |
| poke53281 | they are global and can interfere the configure process of other compilations. | 17:42 |
| poke53281 | stekern: if you want to compile by yourself. You should execute the lines given by "make env" | 17:43 |
| stekern | hmm, but does the script set PATH when it does the toolchain build? | 17:54 |
| poke53281 | Yes | 17:54 |
| poke53281 | seventh line in Makefile | 17:54 |
| stekern | ah, right, so that's the problem | 17:56 |
| poke53281 | PATH := $(TOOLCHAIN_DIR)/bin:$(PATH) | 17:56 |
| poke53281 | Now it should be correct | 17:56 |
| stekern | exactly | 17:57 |
| stekern | it's been picking up some old or1k-linux toolchain I happen to have in my path | 17:57 |
| poke53281 | Unfortunately I have realized that just changing the TOOLCHAIN_NAME will not work. I have used or1k-linux some times without using this variable. | 17:57 |
| poke53281 | Ok | 17:57 |
| stekern | I noticed that too, I changed it to -uclibc in all toolchain related stuff | 17:58 |
| stekern | (but it was failing before that, so that's unrelated) | 17:58 |
| poke53281 | I am fixing this right now | 18:00 |
| poke53281 | stekern: http://www.simulationcorner.net/opencore4/or1k-toolchain.tar.bz2 | 18:49 |
| poke53281 | Corrected a lot | 18:49 |
| poke53281 | You were right. You have to export all this variables by your own via "make env" Now it should work without doing it. | 18:50 |
| poke53281 | And I changed all the target names. | 18:50 |
| olof | Hmmm... looks like I'm disconnecting quite a lot. | 19:08 |
| -!- olof is now known as olofk | 19:08 | |
| stekern | poke53281: great, thanks! | 19:08 |
| poke53281 | between your message and my message I have made additional three correction. This will never end.... | 19:09 |
| stekern | heh ;) | 19:10 |
| stekern | I can confirm that the problems I had were related to that at least | 19:14 |
| stekern | now the whole toolchain built without problems | 19:15 |
| poke53281 | but be careful. If you don't use my uclibc config file maybe X will not compile. I remember that I had to alter it multiple times to include all features. | 19:22 |
| stekern | I'm using it now | 19:28 |
| stekern | haha, looks like I lack wget-fu... I think I downloaded your whole website's content, not just what's in the download folder... | 19:30 |
| poke53281 | Yes,you should not download the parent directory | 19:31 |
| poke53281 | you should stop after 1 recursion. | 19:32 |
| stekern | yeah, I know... but I always forget... | 19:44 |
| stekern | I usually notice when I see stuff that's unrelated, but it seems like everything else was so minor (in filesize) anyway | 19:45 |
| stekern | hmm, when building xpre it picks up the x86_64 machines libpng | 20:53 |
| poke53281 | maybe pkg-config not correctly configured | 21:13 |
| poke53281 | the environment variables for this are set by "make env" | 21:14 |
| poke53281 | and don't forget the patch for the .la files They are very very important | 21:14 |
| poke53281 | sed -i~ -e "s;libdir='/usr;libdir='$(SYSROOT)/usr;" $(SYSROOT)/usr/lib/*.la | 21:15 |
| poke53281 | this should be executed after every compilation of libraries | 21:15 |
| poke53281 | This problem was one of the main issues I had during writing these scripts | 21:17 |
| poke53281 | And it took forever to find solutions. | 21:17 |
| poke53281 | stekern: I am wondering why in xpre? in xpre you are compiling libpng. Why it should pick the x86_64 machines libpng? | 21:20 |
| poke53281 | maybe during compiling pixman? | 21:21 |
| stekern | during freetype | 21:22 |
| stekern | I have the PKG_ env vars as 'make env' told me | 21:23 |
| poke53281 | "make freetype" <- no problem for me | 21:25 |
| poke53281 | what is the error message? | 21:26 |
| poke53281 | During linking or before. | 21:26 |
| stekern | http://pastie.org/8363239 | 21:31 |
| poke53281 | Oh yeah. libtool my friend. Lost some hair because of this. | 21:31 |
| poke53281 | The error is in .la file in $SYSROOT/usr/lib/ | 21:32 |
| poke53281 | grep " /usr/lib/libpng" *.la | 21:32 |
| juliusb | stekern: nice work on the precompiled tool chains | 21:33 |
| juliusb | actually, just reading through today's log, did you get gdb compiling stekern ? | 21:35 |
| juliusb | I never did iirc | 21:35 |
| stekern | poke53281: nope, no such match | 21:36 |
| juliusb | and libgui problems don't ring too many bells | 21:36 |
| poke53281 | grep "libpng" *.la | 21:36 |
| stekern | poke53281: yeah, tried that too | 21:36 |
| poke53281 | then the problem happens already in src/freetype | 21:37 |
| stekern | juliusb: yes, gdb is included in those precompiled | 21:37 |
| juliusb | oh wicked | 21:38 |
| poke53281 | find src/freetype/ -name "*.la" | 21:38 |
| juliusb | I recall _franck_ was doing some hacking on it a while back? | 21:38 |
| stekern | poke53281: nope | 21:40 |
| stekern | I recall building freetype before though | 21:40 |
| poke53281 | Ahh, did you execute "cd src/$@; sed -i~ -e "s;-L/usr/lib ;;" builds/unix/unix-cc.mk" | 21:41 |
| poke53281 | as written in my script? | 21:41 |
| olofk | ORPSoC 3.1 is released now. Would be grateful if someone could try to install it to check if I messed something up | 21:43 |
| juliusb | stekern: sorry, can I be annoying and ask that you put a little more detail with those prebuilt versions. Something like which source they came from and what revision they were at? (or maybe just "HEAD of these repos on 28/09/2013"?) | 21:43 |
| juliusb | ... unless there is something in those archives already? | 21:44 |
| juliusb | olofk: nice one! | 21:44 |
| juliusb | I'll be the guinea pig | 21:44 |
| olofk | juliusb: It's on the opencores FTP | 21:44 |
| juliusb | oh, I can't github it? | 21:45 |
| stekern | poke53281: umm... no | 21:45 |
| juliusb | https://github.com/openrisc/orpsoc/archive/3.1.zip should work too right? | 21:45 |
| olofk | juliusb: Yes, you can. The only difference is that the one on the OpenCores FTP went through make dist, so you don't have to run autoreconf | 21:46 |
| juliusb | ah cool OK | 21:46 |
| olofk | So I think that one will be slightly easier to distribute at the workshop | 21:46 |
| juliusb | makes sense | 21:47 |
| poke53281 | libtool executes with the argument "-L/usr/lib/x86_64-linux-gnu" . It's written somewhere in the config files in freetype. | 21:48 |
| poke53281 | Ahh, I do the patching after configure | 21:49 |
| poke53281 | stekern: I think you have to later the sed patch that it matches -L/usr/lib/x86_64-linux-gnu | 21:49 |
| juliusb | olofk: but just out of interest, can you let me know how I bootstrap it as from the orpsoc git repo? So if I have a fresh clone, what do I do? (I'll document this, too) | 21:49 |
| stekern | poke53281: yeah, but configure get that from libpng-config, should that honour pkg-config flags somehow? | 21:50 |
| olofk | juliusb: For a system installation run autoreconf -i && ./configure && make && make install | 21:50 |
| olofk | But if you are using the version from git and don't want to install it, you can just execute orpsoc from bin/orpsoc.in directly | 21:50 |
| stekern | juliusb: I install it anyway | 21:51 |
| stekern | olofk said that was cool | 21:51 |
| poke53281 | stekern: No. libpng-config --libdir gives me /usr/lib | 21:52 |
| juliusb | :) | 21:52 |
| stekern | poke53281: ok, so our systems differ a bit there, mine gives: | 21:53 |
| stekern | $ libpng-config --libdir | 21:53 |
| stekern | /usr/lib/x86_64-linux-gnu | 21:53 |
| poke53281 | stekern: configure output: "checking for libpng... -L/usr/lib -lpng16" | 21:53 |
| olofk | I actually think it's pretty cool. It makes it feel just like a real program :) | 21:53 |
| poke53281 | So my patch is wrong | 21:53 |
| stekern | olofk: I agree, and it's just handy to "have it around" wherever | 21:54 |
| juliusb | http://pastie.org/8363283 :( | 21:54 |
| olofk | Sorry, autoreconf -i | 21:54 |
| juliusb | oh you wrote that | 21:54 |
| juliusb | yep, all my bad | 21:55 |
| olofk | juliusb: There's an INSTALL file in orpsoc with some basic instructions on how to get started, but it probably needs some extra explaining | 21:55 |
| olofk | ...and you can skip the part where it tells you to mail your credit card number to me. I already got that | 21:55 |
| stekern | juliusb: the prebuilt versions are from the top of our 4.8.1 branch in github | 21:55 |
| juliusb | cool, I'll flesh that stuff out once I get a grip | 21:55 |
| poke53281 | Stekern: Now I am a little bit worried about my other patches. Especially of that one patching the .la files | 21:56 |
| juliusb | olofk: I'm still waiting for the flexlm USB dongle so I can use orpsocv3, though | 21:56 |
| juliusb | you said that was the deal | 21:56 |
| juliusb | you were laughing maniacally and saying how rich you were going to be come once you shipped orpsocv3 | 21:57 |
| juliusb | as you stroked the white cat on your lap | 21:57 |
| olofk | lol | 21:57 |
| olofk | That wasn't a white cat, btw | 21:57 |
| juliusb | oh... now that I think about it, maybe it was a pile of white powder? | 21:58 |
| juliusb | olofk: do you mind if I update instructions on the orpsoc wiki page? | 21:58 |
| stekern | juliusb: olofk have switched to using floppy-disk dongles now though | 21:58 |
| rfajardo | hehe, I will assure myself of not trying orpsocv3 :P | 21:58 |
| stekern | poke53281: hmm, the .la's I looked at looked fine, I think. let me poke around a bit more | 21:59 |
| juliusb | 5.25 inch or 3.5inch? | 21:59 |
| juliusb | i'll need to go commit a crime at the local computer museum if it's the former | 22:00 |
| juliusb | actually... there is a good computer museum here in Cambridge now apparently | 22:00 |
| stekern | juliusb: not sure, I haven't received mine for the orpsocv3 EE fault tolerant version yet | 22:00 |
| juliusb | http://www.computinghistory.org.uk/pages/490/Computers-By-Manufacturer/ | 22:01 |
| juliusb | ok I should stop joking and and start orpsocing | 22:01 |
| rfajardo | go for it :) | 22:02 |
| juliusb | rfajardo: hi, btw! | 22:02 |
| rfajardo | hello there | 22:02 |
| stekern | I almost got framebuffer working with the opencores vga/lcd core from the arm side now | 22:03 |
| rfajardo | I have just pay github openrisc a visit and have browsed through some orpsoc stuff | 22:03 |
| juliusb | rfajardo: sorry to hear you can't make it to ORCONF | 22:03 |
| olofk | juliusb: Updating the instructions would be great | 22:04 |
| olofk | rfajardo: Hey. Haven't heard from you for some time. Are you coming to orconf? | 22:04 |
| rfajardo | Well, I'm sorry too. A lot of things and plans going on this year. I really wish for some more nice open hw/sw work next year. | 22:04 |
| rfajardo | olofk, I'm sorry I won't. I would really like to. But I will content myself with the promise that I will not miss the next one. Hopefuly I will have some more input from my side too. It's probably good to do something every 3 or 4 years :P. | 22:06 |
| poke53281 | stekern: I can set LIBPNG_CFLAGS and LIBPNG_LDFLAGS directly. This should avoid this problem. | 22:07 |
| juliusb | rfajardo: no worries man, we'll drink your beer this year then | 22:09 |
| rfajardo | hehe | 22:10 |
| rfajardo | I think that actually made me really sad | 22:10 |
| olofk | juliusb: Don't do that. I was planning to put rohypnol in rfajardo's beer | 22:10 |
| juliusb | even better, it'll be a cheap night for me | 22:12 |
| rfajardo | I will miss some fun days | 22:13 |
| olofk | rfajardo: Are you still in Germany, btw? Given the number of German contributors lately, it might make sense to have the next conference there | 22:13 |
| rfajardo | where do you have a wiki about the new orpsocv3? | 22:13 |
| juliusb | rfajardo: http://opencores.org/or1k/ORPSoCv3 | 22:13 |
| juliusb | but it looks fairly out of date... | 22:13 |
| poke53281 | stekern: http://pastie.org/8363324 | 22:13 |
| rfajardo | olofk, yes I am. I have seen it too. I guess wallento has been mainly active, isnt it? | 22:13 |
| juliusb | ... I think? | 22:13 |
| olofk | Of course, we are also still waiting for _franck_ to take us to the french riviera | 22:14 |
| juliusb | I'm really hoping we'll have a large number of contributors pop up in Tuscany or the Caribbean over the next 8 months... | 22:15 |
| olofk | rfajardo: Yes, wallento will present some multicore stuff. We will (hopefully) record all the talks, so you can still watch that even if you're not coming | 22:15 |
| olofk | juliusb: Amen that | 22:15 |
| juliusb | plan is to _definitely_ record all of the talks. simoncook will hopefully do the honours | 22:15 |
| rfajardo | olofk, awesome. That's good! | 22:15 |
| poke53281 | stekern: This should do it | 22:15 |
| juliusb | olofk: what's the deal with bin/orpsoc and bin/orpsoc.in? | 22:17 |
| rfajardo | you're certainly posting the link to the list. Or do you have another communication path besides the irc? | 22:17 |
| olofk | Oh well. Time to get off skynet now. Please let me know if there are any troubles with the new orpsoc release. I want it in good shape for the conference | 22:17 |
| juliusb | just some sys.path stuff? | 22:17 |
| stekern | poke53281: I'm testing it out right now | 22:19 |
| stekern | poke53281: those -config scripts are what have gave me the most headaches when crosscompiling | 22:19 |
| stekern | aren't there some sane way to handle that? | 22:20 |
| olofk | juliusb: I can explain more later, but it's basically for allowing orpsoc to be run without being installed | 22:20 |
| juliusb | cool, so I should be running bin/orpsoc.in instead of bin/orpsoc if I hvaen't install it? | 22:20 |
| olofk | The orpsoc binary needs to know where to find the orpsoc python package | 22:20 |
| olofk | If you don't install it, you don't need to run autoreconf and configure at all | 22:21 |
| juliusb | ah right | 22:21 |
| juliusb | makes sense | 22:21 |
| stekern | but if you don't install it, it's not cool | 22:21 |
| juliusb | :( | 22:21 |
| olofk | And I guess you don't want to be one of the uncool kids, right+ | 22:21 |
| olofk | s/+/? | 22:21 |
| juliusb | ok... noob question I could probably find in the logs, but quickly, how do I get a system? list-systems gives me nada | 22:21 |
| juliusb | always was.... | 22:21 |
| poke53281 | stekern: good question. I am not sure. Sandbox like in gentoo maybe. | 22:22 |
| juliusb | :) | 22:22 |
| olofk | juliusb: Did you get the download link in the registration e-mail? | 22:22 |
| stekern | juliusb: I think you are experience the same confusion I had initially | 22:22 |
| stekern | you should have in a directory: orpsoc-cores and build-orpsoc | 22:23 |
| stekern | and in build-orpsoc you should have a orpsoc.conf with the following content | 22:23 |
| poke53281 | but stekern: freetype seems to use the non-standard way for configure libpng. Today the pkg-config script is the standard. | 22:23 |
| stekern | [main] | 22:24 |
| stekern | cores_root =../orpsoc-cores/cores | 22:24 |
| stekern | systems_root =../orpsoc-cores/systems | 22:24 |
| juliusb | my dir looks like so: http://pastie.org/8363335 | 22:24 |
| juliusb | so, sans orpsoc-cores and build-orpsoc but I do have a orpsoc.conf | 22:24 |
| stekern | ok, but you shouldn't be in orpsoc when building (olofk also told me that *that* was uncool) | 22:25 |
| juliusb | oh right... | 22:26 |
| olofk | I'm building from the orpsoc dir all the time | 22:26 |
| olofk | Which is very uncool | 22:26 |
| juliusb | but which one. orpsoc or orpsoc/orpsoc? | 22:26 |
| juliusb | there's 2 orpsoc dirs afaict | 22:26 |
| olofk | But IIRC there is a bug that prevents you from building inside of the orpsoc-cores dir | 22:27 |
| stekern | I think clearest is to just install orpsoc to the system and then forget about the orpsoc dir | 22:27 |
| stekern | it | 22:27 |
| juliusb | ok i like that advice | 22:27 |
| stekern | orpsoc is just a program | 22:27 |
| * juliusb installs orpsocv3 | 22:27 | |
| stekern | the actual SoC stuff is in orpsoc-cores | 22:27 |
| * olofk now has control over juliusb's web cam | 22:27 | |
| stekern | "just" a program (to not offend olofk) | 22:28 |
| stekern | ;) | 22:28 |
| juliusb | haha | 22:28 |
| juliusb | ummm, I don't have an orpsoc-cores dir though. I make one? | 22:28 |
| olofk | juliusb: Clone it from github | 22:28 |
| olofk | openrisc/orpsoc-cores | 22:28 |
| juliusb | ohhhh. separate repo. of course! | 22:29 |
| poke53281 | stekern: Sometimes I have patches like "sed -i~ -e "s; /usr/lib/libdirect.la; $(SYSROOT)/usr/lib/libdirect.la;" $(SYSROOT)/usr/lib/*.la" | 22:29 |
| * juliusb slaps his forehead | 22:29 | |
| poke53281 | stekern: So I am not sure these are working on your system. | 22:29 |
| * olofk watches juliusb slap his forehead | 22:29 | |
| poke53281 | stekern: But it should be obvious how to alter these patches for your system. | 22:30 |
| juliusb | alright, I see things in list-systems | 22:30 |
| olofk | juliusb: There actually isn't a single OpenRISC related line of code in orpsoc. I wanted to have it generic enough to be used for other purposes too | 22:30 |
| olofk | juliusb: Great! | 22:30 |
| juliusb | ah cool. makes sense | 22:31 |
| olofk | Try to run orpsoc sim wb_intercon | 22:31 |
| juliusb | !!! | 22:31 |
| juliusb | wicket | 22:31 |
| juliusb | wicked, even | 22:31 |
| juliusb | (no I'm not watching cricket) | 22:31 |
| juliusb | so what if I don't like how abstracted I am from what's going on. for all I know you could just be printing strings out from orpsoc and not really doing anything.... | 22:32 |
| * juliusb raises an eyebrow suspiciously | 22:32 | |
| juliusb | how do I enable like "verbose" mode or something, so I see verything it's doing? | 22:33 |
| olofk | If you have an elf file with an or1200 testcase from orpsocv2 you can now run orpsoc sim or1200-generic --or1k-elf-load=/path/to/<testcase>.elf | 22:33 |
| juliusb | ah, or1200-specific tests are so orpsocv2 | 22:33 |
| juliusb | you want or1k-generic tests like mor1kx-dev-env :P | 22:33 |
| juliusb | i'll fix that for you | 22:34 |
| olofk | juliusb: It's not very verbose yet :( Franck was asking for the same thing | 22:34 |
| juliusb | what's the default system? | 22:34 |
| olofk | juliusb: I would say it's or1200-generic at the moment, but that will be changed in the future | 22:35 |
| juliusb | actually, I'll poke around, you were going to go, we can talk tomorrow | 22:35 |
| stekern | poke53281: the freetype change worked as intending at least, thanks for that! | 22:35 |
| juliusb | ok, cool, good to know, thanks! | 22:36 |
| olofk | Oh right. I forgot about that :) | 22:36 |
| olofk | Good night | 22:36 |
| juliusb | l8rz | 22:36 |
| * _franck_ just found out we can halt, download an elf, set npc and resume with openocd... :) | 22:36 | |
| stekern | and yeah, the sed changes should be pretty straight forward | 22:36 |
| poke53281 | stekern: Your are almost done. Only 30-40 libraries left. ;) | 22:36 |
| juliusb | _franck_: nice. with an orpsoc (will omit v3 from now on) built image? | 22:36 |
| _franck_ | of course :) | 22:37 |
| juliusb | that's very good news :) | 22:37 |
| _franck_ | I mean with the openocd telnet console. We don't need gdb for this kind of operation | 22:37 |
| juliusb | that is quite cool | 22:38 |
| juliusb | does it work with gdb, though? | 22:38 |
| juliusb | or rather, does gdb work with it? | 22:38 |
| _franck_ | yes it is | 22:38 |
| stekern | juliusb: we were speaking about lifting out the mor1kx-dev-env tests to a "orpsoc-tests" repo or something like that | 22:38 |
| juliusb | yes, that's what was suggested a long time ago (well, taking tests from or1ksim and orpsocv2 and doing that) | 22:39 |
| _franck_ | juliusb: FIY, our openrisc openocd port is upstream now | 22:39 |
| stekern | since there aren't any tests in orpsoc (nor orpsoc-cores) themselves | 22:39 |
| juliusb | now that orpsoc has started the trend of keeping things in their own repo, an or1k-tests is a good idea | 22:39 |
| stekern | that's probably a better name, yes =) | 22:40 |
| rfajardo | Nice talking to you for a bit again. I will make sure I set up a system with which I can work with again and try some stuff :). I see you soon! | 22:40 |
| juliusb | _franck_: now you _really_ deserve a beer for that :) | 22:40 |
| juliusb | rfajardo: you too man, catch you soon! | 22:40 |
| juliusb | _franck_: when was it accepted? | 22:42 |
| _franck_ | some days ago | 22:42 |
| juliusb | oh excellent. where is their repo? | 22:42 |
| juliusb | i see they do releaes to sourceforge but can't find their working git repo | 22:42 |
| _franck_ | http://repo.or.cz/w/openocd.git | 22:43 |
| juliusb | ah but this is where the action went down: http://openocd.zylin.com/#/c/1547/ | 22:44 |
| _franck_ | I dropped the "classic" debug unit support | 22:45 |
| juliusb | yes, that's fair enough. | 22:45 |
| stekern | _franck_: great news! | 22:45 |
| _franck_ | byte / half word read/write was broken | 22:45 |
| _franck_ | and I didn't want to fix it | 22:45 |
| juliusb | oh, that may involve RTL fixes, too? | 22:45 |
| _franck_ | I think so | 22:45 |
| stekern | now I'm even closer to fb working on arm, I've got half of the columns there... | 22:46 |
| stekern | wonder where the other half have went | 22:47 |
| _franck_ | I should update our openocd wiki one day.... | 22:49 |
| juliusb | _franck_: yes, that is certainly a good thing to do. even with just a quick note now to indicate where to get it and the status of what is upstream | 22:50 |
| juliusb | well, is what's on the github openrisc/openOCD still good? | 22:50 |
| _franck_ | yes it's working | 22:51 |
| juliusb | but should we tell people to maybe pull from the openOCD tree now? | 22:51 |
| _franck_ | and, afair, it's up to date | 22:51 |
| _franck_ | if they use the mohor debug unit they need to stick to the github version | 22:52 |
| juliusb | ah right OK | 22:52 |
| _franck_ | one day someone should add the mohor debug unit to openocd (not me :)) | 22:53 |
| _franck_ | I need to sleep, good night | 22:56 |
| stekern | I'm considering doing a dummy/emulated "ps/2" core which just have a register that I can write from the arm side, and then have a mouse/keyboard server running on the arm | 23:24 |
| stekern | but is there some software like that already available? everything seems immensely overkill for my needs | 23:24 |
| --- Log closed Sun Sep 29 00:00:17 2013 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!