| 01:14.32 | *** join/#edev roxfan (n=dunno@69.1-200-80.adsl-dyn.isp.belgacom.be) |
| 02:14.31 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 02:48.21 | *** join/#edev landley (n=landley@cpe-72-177-114-214.austin.res.rr.com) |
| 02:52.53 | RobotGuy | landley: Hello |
| 02:56.55 | landley | Hi robotguy. |
| 02:57.00 | landley | How goes? |
| 03:07.49 | RobotGuy | Not bad. I have a nice kernel built that leaves me about 20 MB RAM after modules load. |
| 03:08.46 | *** join/#edev _lifeless (n=patrick@89.20.118.236) |
| 03:11.54 | RobotGuy | The way that software is built using buildroot bothers me. It does not make any sense the way it is done for embedded systems. |
| 03:13.13 | RobotGuy | landley: How goes with you? |
| 03:14.11 | landley | Eh, medium. |
| 03:14.35 | landley | I seem to be developing an allergy to cats, which sucks since I have four of them. |
| 03:15.15 | RobotGuy | I have just one cat. Allergies to your pets are not good. |
| 03:16.22 | RobotGuy | If software is built to go into /usr, why are the config files not put in /usr/etc? |
| 03:18.11 | landley | I'm unaware of anything using /usr/etc. |
| 03:18.24 | landley | <PROTECTED> |
| 03:18.35 | RobotGuy | I know that. That is not my point though. |
| 03:18.37 | landley | Once upon a time it was general configuration stuff, a lot of which was dynamic. |
| 03:18.55 | landley | Then people started making read only root filesystems, and started moving the volatile stuff out of /etc. |
| 03:19.05 | landley | It's kind of stuck halfway these days. |
| 03:19.12 | RobotGuy | Buildroot builts software just like it was for a big system, not an embedded system. That does not work. |
| 03:19.36 | landley | Depending on whether or not they're writable. |
| 03:19.46 | landley | So I actually _do_ use /usr/etc. :) |
| 03:20.15 | RobotGuy | The builds still put config files in /etc. |
| 03:20.19 | landley | In general, it's nice to have the entire OS in one directory, for the same reason it's nice to have your home directory not be / |
| 03:20.51 | landley | In practice, /etc, /bin, /sbin, /tmp, /mnt, /var, /usr, /proc, and /sys are hardwired into all sorts of programs as assumptions. |
| 03:21.00 | landley | Oh, and /dev. |
| 03:21.00 | landley | And probably more. |
| 03:21.11 | RobotGuy | Those assumptions do not work for embedded systems. |
| 03:21.29 | landley | MacOS X threw it _all_ out the window. |
| 03:21.41 | landley | Why do those assumptions not work on embedded systems? |
| 03:21.58 | landley | looking for /etc/passwd is something software is just going to _do_. Symlink /etc if you've moved it. |
| 03:22.07 | RobotGuy | Because config files for software added to an embedded system should not go into /etc. |
| 03:22.15 | landley | So symlink it. |
| 03:22.28 | RobotGuy | Symlink what? |
| 03:22.32 | landley | maybe then we can break up /proc. |
| 03:22.39 | landley | Make /etc a symlink to /var/etc or some such. |
| 03:22.48 | landley | <PROTECTED> |
| 03:23.10 | landley | You can also modify every single program you put on the system to use a custom location. |
| 03:23.17 | landley | In which case you can ignore FHS completely. |
| 03:23.25 | landley | But then it's up to you to police it. |
| 03:23.38 | RobotGuy | The same problem exists for using /var as it does for using /etc for config files. That's why I want config files to be in /usr/etc for software I add. |
| 03:24.36 | RobotGuy | My thumb drive mounts at /usr, and has software, but not of the config files that go with the software. |
| 03:25.47 | landley | So symlink /etc to /usr/etc. |
| 03:25.53 | landley | ln -s /usr/etc /etc |
| 03:25.56 | landley | From initramfs. |
| 03:26.31 | RobotGuy | I know how to do that, but I don't know how to do it for the initrd build and it wouldn't work anyway because there would not be any /usr/etc then. |
| 03:39.53 | landley | Sorry, can't help you there... |
| 03:39.59 | landley | Sorry, I mean I don't use buildroot. |
| 03:40.11 | landley | Which is what I believe you're using to build your initrd...? |
| 03:43.59 | RobotGuy | Yes, of course I am using buildroot. What do you use? |
| 03:44.17 | RobotGuy | Do you understand what I am saying? |
| 03:45.30 | *** part/#edev RobotGuy (n=RobotGuy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 03:45.42 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 03:51.45 | landley | RobotGuy: I use my firmware linux thing. |
| 03:51.58 | landley | However, you don't have to modify your buildroot build. |
| 03:52.05 | landley | You just have to modify your packaging step. |
| 03:52.20 | landley | Build it, them move stuff after you build it but before you make the ext2 image for your initrd. |
| 03:52.43 | landley | Or have a script in your initrd move stuff. I'm not sure sure how you mount your /usr currently. |
| 03:56.19 | RobotGuy | I think you are not understanding me for some reason. |
| 03:56.48 | landley | Quite possibly. Could you elaborate? |
| 03:57.01 | RobotGuy | I been trying to. |
| 03:57.15 | landley | You build a filesystem. You package it into an initrd. You boot the initrd. |
| 03:57.49 | landley | Correct? |
| 03:58.04 | landley | When the initrd boots and runs, it mounts /usr from your flash disk. |
| 03:58.09 | RobotGuy | That much is correct, but that is not the problem. |
| 03:59.05 | RobotGuy | The initrd does not load from my thumb drive. |
| 04:00.10 | RobotGuy | It loads from the Hammer's internal flash. |
| 04:01.57 | RobotGuy | The initrd has to be as small as possible to leave as much RAM free as possible for loading programs to run. |
| 04:05.10 | landley | ok |
| 04:05.30 | landley | You don't actually need anything in /etc to run the initrd, that I'm aware of. |
| 04:05.44 | landley | Busybox mount will happily read /proc/mounts if there's no /etc/mtab. |
| 04:05.56 | landley | You may have to say mount -n or some such, I forget. |
| 04:06.17 | landley | (Although I wrote this code, It's been a while since I've played with the guts of busybox and I haven't added mount to toybox yet.) |
| 04:06.52 | landley | So not having an /etc until you mount /usr shouldn't be too bad if you do it early. The dangling symlink doesn't hurt anything. |
| 04:07.03 | landley | <PROTECTED> |
| 04:10.39 | RobotGuy | You are not getting what I am saying. I don't know how to say it any different. |
| 04:11.45 | RobotGuy | <PROTECTED> |
| 04:12.48 | RobotGuy | My thumb drive should have all the config files for software I add to Hammer. This is software that does NOT go into the initrd. |
| 04:14.16 | landley | Ok. |
| 04:14.28 | landley | Sounds like you need a union mount. |
| 04:14.43 | landley | You can also copy /usr/etc to /etc after the mount. |
| 04:14.58 | landley | Assuming that doesn't eat too much ram. |
| 04:15.12 | RobotGuy | I do not want config files in /etc for software I add to Hammer. |
| 04:15.38 | landley | Would copying the contents of the initrd /etc into /usr/etc cause problems? |
| 04:15.47 | landley | By which I mean doing so every boot, overwriting the old files if they're there? |
| 04:16.10 | landley | In which case "mv /etc/* /usr/etc; rm /etc; ln -s /usr/etc /etc" can be part of the bootup scripts. |
| 04:16.14 | RobotGuy | I don't know. If it does work, it is not the correct solution. |
| 04:16.22 | landley | Back to union mounts then. |
| 04:16.37 | RobotGuy | No, it does not have to be back to union mounts then. |
| 04:16.46 | landley | You can try configuring your software packages to put their config files somewhere else, but that's whack-a-mole with every package. |
| 04:17.29 | RobotGuy | It's the correct solution though - the software build should put configs in a proper place. /etc is not that place. |
| 04:18.14 | RobotGuy | If software is built to run from /usr, configs should be in /usr/etc. |
| 04:31.10 | landley | RobotGuy: so if you install the binary in /usr/bin, it should look for its config files in /usr/etc? |
| 04:31.27 | RobotGuy | landley: Yes. |
| 04:31.29 | landley | By that logic, "useradd" and "userdel" should modify /usr/etc/passwd. |
| 04:31.35 | landley | They're in /usr/bin after all. |
| 04:31.56 | RobotGuy | passwd is not a configuration file - it's a system file. |
| 04:32.30 | landley | So dpkg should have the system's installed package database in /usr/etc? |
| 04:32.59 | RobotGuy | I give up. |
| 04:33.00 | *** join/#edev btraynor (n=btraynor@gromit.mixdown.ca) |
| 04:33.18 | landley | The distinction between /bin and /usr/bin has always been deeply arbitrary. |
| 04:33.36 | landley | There have been attempts to retcon justification for it, but the reality is "ken and dennis ran out of space". |
| 04:33.58 | landley | But /etc doesn't take up much space, and lots of utilities read common configuration files. |
| 04:34.08 | landley | Splitting the config files across multiple locations never had any advantage, so they didn't do it. |
| 04:34.15 | RobotGuy | I don't care what Ken and Dennis do. I care about building software for my Hammer and NOT having the config files be put in /etc. |
| 04:34.28 | landley | Feel free to customize it then. |
| 04:34.50 | landley | But installing binaries in /bin vs /usr/bin (or /sbin vs /usr/sbin) has never affected /etc. |
| 04:35.16 | landley | I still think union mounts are the correct thing to do. |
| 04:35.41 | landley | But it's your system. |
| 04:36.18 | RobotGuy | All I am saying is that the config files for the software being built should not go into /etc. The configs should go into the --prefix where the software is built. |
| 04:37.25 | landley | There may be a separate way to specify a prefix for /etc (that's up to whoever wrote the configure), but /etc has traditionally not had a prefix. |
| 04:37.42 | landley | Exactly because most of the files in /etc were common files shared by multiple packages, some of which were in /bin and some of which were in /usr/bin. |
| 04:37.42 | RobotGuy | I am not talking about a prefix for /etc |
| 04:38.10 | landley | You can also put the config files into a user's home directory. That's common too. |
| 04:38.35 | landley | But things like /usr are often a shared read-only nfs mount common across dozens of machines. |
| 04:38.45 | landley | Having /etc not be in there means the configurations can be customized for each machine if necessary. |
| 04:39.00 | landley | If the config was in /usr/etc, every machine would have to have the same config. |
| 04:39.37 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 04:39.40 | RobotGuy | You are talking about big systems, and I am talking about little systems. BIG DIFFERENCE. |
| 04:41.25 | RobotGuy | You are confusing the discussion with things that do not apply. |
| 04:45.38 | landley | No, I'm saying "these packages we're using came from a certain environment". |
| 04:45.58 | landley | Now if you want to argue that busybox and uClibc shouldn't be making these sort of assumptions, sure. |
| 04:46.16 | RobotGuy | It does not matter where these packages came from. They are being built for a different environment. |
| 04:46.22 | landley | Once you get beyond a core dozen packages or so and start adding more stuff, you wind up in traditional unix-land that has 30 years of history not being battery powered. |
| 04:46.49 | landley | They are being built for an environment they were not designed for, written to a set of standards they often have to explicitly violate to do what you want. |
| 04:47.03 | RobotGuy | Just because something has always been done a certain way, neither makes it right nor makes it correct for what needs to be done now. |
| 04:47.04 | landley | The LSB is kind of evil in places. |
| 04:47.15 | jbevren | concur lol |
| 04:47.18 | landley | It was right and correct for the environment it was designed in. |
| 04:47.18 | jbevren | and on that I"m going to bed :) |
| 04:47.32 | landley | You're adapting them to a different environment, and railing at what they currently do in the process. |
| 04:48.03 | RobotGuy | They supposedly have already been adapted for this different environment. |
| 04:48.18 | RobotGuy | Or perhaps I misunderstand what buildroot is all about. |
| 04:48.52 | RobotGuy | I am saying the software package builds put the config files for the software in the wrong place for an embedded system. |
| 04:48.54 | landley | I don't use buildroot, but where it came from was a test harness for uClibc and busybox. |
| 04:49.07 | RobotGuy | What do you use? |
| 04:49.13 | landley | "It builds against uClibc and works with it". That's what buildroot was for. |
| 04:49.16 | landley | Anything else is gravy. |
| 04:49.30 | landley | As I said, I use my own homebrew, http://landley.net/code/firmware |
| 04:49.37 | RobotGuy | There are still packages in buildroot that try building glibc. |
| 04:50.14 | landley | Do they succeed? |
| 04:50.38 | RobotGuy | I never let them continue. |
| 04:51.00 | CP|home | RobotGuy: I think you have mis interpeted what buildroot is |
| 04:51.16 | CP|home | its a build system, nothing more, nothing less |
| 04:51.24 | landley | Buildroot has been _adapted_ into a distribution generator. |
| 04:51.34 | landley | But it evolved out of the need to build a toolchain to link things against uClibc with. |
| 04:51.41 | RobotGuy | CP|home: Buildroot does not seem to be entirely appropriate for building software for embedded systems. |
| 04:51.52 | landley | Buildroot happened because the old gcc wrapper script couldn't fix the design flaw in libgcc_s.so. |
| 04:51.53 | CP|home | Nobody ever claimed that it was |
| 04:52.09 | CP|home | That it builds embedded packages well is because thats what it was originally built for |
| 04:52.34 | landley | Then once it build a toolchain, it was easy to add other packages to build against uClibc as a way of testing the toolchain and current snapshot of uClibc. |
| 04:53.03 | RobotGuy | I guess I'll just have to change buildroot to do things the way I think they should be done then. |
| 04:53.10 | CP|home | And at the same time, it served a purpose for building packages for early embedded apps |
| 04:53.17 | landley | People added lots and lots of packages along the way so uClibc got automatic regression testing against that package when things changed in uClibc. |
| 04:53.19 | CP|home | Such as the beloved Tuxscreen and other such things |
| 04:54.07 | CP|home | The problem is - embedded systems have a tendency to be... special |
| 04:54.20 | CP|home | So inevitably, every generic solution ends up being wrong |
| 04:54.21 | RobotGuy | Or I need to find a build system more appropriate to embedded systems. |
| 04:54.31 | CP|home | so a new solution is invented - and then that gets generic, and ceases being useful |
| 04:54.40 | landley | RobotGuy: as I said, I rolled my own build system, specifically for creating an embedded development environment. |
| 04:54.46 | landley | Because buildroot didn't match what I wanted to do. |
| 04:54.53 | RobotGuy | landley: Hmmm. |
| 04:55.00 | landley | The point of my build system is it's _simple_. |
| 04:55.09 | landley | It gets you to a native build environment for your target, where you can build anything you want. |
| 04:55.13 | landley | But doing that is a separate layer. |
| 04:55.20 | landley | Rigidly isolated from the base layer. |
| 04:56.26 | landley | Basically, I'm automating the hard parts of Linux From Scratch, and adapting it to use uClibc and busybox on non-x86 targets. |
| 04:56.28 | RobotGuy | I've built a few Linux From Scrach systems. |
| 04:56.32 | landley | But where you go from there? Read LFS. :) |
| 04:56.44 | landley | You'll know why the system I build lives in /tools then. |
| 04:56.52 | RobotGuy | I know about LFS and BLFS. I've done them. :) |
| 04:56.54 | landley | What you build _with_ that system is Not My Problem. :) |
| 04:57.09 | landley | (Although I need to check in the distcc acceleration trick so you don't have to leave builds running overnight.) |
| 04:57.48 | landley | I've gotten sidetracked into making it build on an MacOS X host, which requires some environmental bootstrapping. |
| 04:57.51 | landley | 2/3 done... |
| 04:58.52 | RobotGuy | All I want to do is build software for my Hammer (/usr) so configs for that software are where the software is (also /usr). |
| 04:59.18 | landley | I'm still waiting for prpplague to send me broken up kernel patches for hammer so I can push 'em upstream... |
| 04:59.40 | landley | I'm trying to use packages with as little modification as possible. |
| 04:59.57 | landley | Hammer kernel needing extensive patches says to me "it's not ready yet". |
| 05:00.29 | RobotGuy | It just says there needs to be some Hammer specific stuff in the kernel, like there is for other systems. |
| 05:01.26 | landley | The current patch is a half-dozen unrelated things. I sent a longish analysis to prpplague via email. |
| 05:01.36 | RobotGuy | I still haven't figured out how to init the second serial port on Hammer. setserial doesn't seem to do the job. |
| 05:01.39 | landley | He's breaking 'em up and I plan to push the ones I can as soon as I get 'em. |
| 05:01.48 | RobotGuy | What things are unrelated specifically to Hammer? |
| 05:02.53 | landley | msg me your email and I'll forward you the review I sent him... |
| 05:04.02 | landley | Sent. |
| 05:04.16 | RobotGuy | I'm going to start looking at buildroot tomorrow to see how difficult it will be to change the config file path for software builds. |
| 05:04.19 | RobotGuy | Thanks. |
| 05:15.42 | *** join/#edev redblue (i=star@ppp016.109-253-207.mtl.mt.videotron.ca) |
| 05:23.18 | RobotGuy | I need to find which Ubuntu package has libtinfo in it. This does not seem to be a simple thing to find out. |
| 05:33.03 | landley | You have it installed? |
| 05:33.36 | landley | dpkg-query -S /path/to/filename |
| 05:33.57 | RobotGuy | No, I don't have it - that is the problem. |
| 05:34.07 | landley | Ah. |
| 05:34.25 | landley | What is libtinfo? |
| 05:34.26 | RobotGuy | buildroot needs it. |
| 05:34.32 | RobotGuy | No idea what it is. |
| 05:36.24 | landley | It's a curses variant. |
| 05:36.50 | RobotGuy | OK. Do you know what package has it? |
| 05:37.03 | landley | http://209.85.165.104/search?q=cache:SZ7MMPxV0zsJ:lists.pld-linux.org/mailman/pipermail/feedback/2001-February/000030.html+libtinfo&hl=en&ct=clnk&cd=4&gl=us |
| 05:37.16 | landley | You can use libncurses instead. Just substitute and/or symlink. I think. |
| 05:37.46 | landley | If it really wants libtinfo, make a libtinfo symlink in your libpath pointing to libncurses. |
| 05:40.22 | Mirell | Got me a pizza stone |
| 05:40.35 | landley | what's a pizza stone? |
| 05:40.37 | RobotGuy | I guess I won't be building anymore software for Hammer. It wants libtinfo and I do not find where to get that. |
| 05:40.47 | landley | A stone of +5 to pizza summoning? |
| 05:40.52 | RobotGuy | Good Night. This is too frustrating right now. |
| 05:40.58 | landley | RobotGuy: libtinfo is libncurses. |
| 05:40.59 | Mirell | landley: A stone on which to cook pizzas in the oven. |
| 05:41.11 | landley | Just replace each occurrence of libtinfo in the makefile with libncurses. |
| 05:41.19 | RobotGuy | landley: I can't find what package to install for that. :( |
| 05:41.22 | landley | It's a stripped down version of libncurses. |
| 05:41.27 | landley | In theory libncurses 5.2, I think. |
| 05:41.31 | landley | The red hat version, anyway. |
| 05:41.37 | landley | This might be a redhat-ism. |
| 05:41.48 | landley | Google isn't quite being articulate enough tonight... |
| 05:41.54 | RobotGuy | I am running Kubuntu. |
| 05:42.02 | RobotGuy | Good Night. |
| 05:42.02 | landley | RobotGuy: me too. |
| 05:42.05 | landley | Good night. |
| 05:42.11 | landley | Mirell: does it help cook pizzas? |
| 05:42.11 | Mirell | Hmm.. |
| 05:42.15 | Mirell | landley: Yes. |
| 05:42.24 | Mirell | landley: It basically ensures an even temperature on the bottom of the pizza, |
| 05:42.34 | landley | Cool. |
| 05:42.35 | Mirell | landley: And also absorb moisture in order to make it crispier. |
| 05:42.38 | landley | I made soup again. |
| 05:44.52 | landley | I must try this pizza sometime. |
| 05:45.24 | Mirell | I just made pizza tonight with someone who had a pizza stone. |
| 05:45.33 | Mirell | 'twas delicious |
| 05:46.26 | Mirell | I also got a collections notice for an overdue parking ticket for my Mazda. |
| 05:47.10 | landley | Were you previously aware of this ticket? |
| 05:47.20 | landley | If it's already gone to collections, it's unlikely to be time-critical at this point. |
| 05:47.37 | Mirell | Note. |
| 05:47.39 | Mirell | Mazda. |
| 05:47.42 | Mirell | Never owned a Mazda. |
| 05:47.49 | landley | Ah. |
| 05:47.55 | landley | You might want to point this out to them. |
| 05:48.02 | Mirell | Indeed. |
| 05:51.03 | Mirell | Hmm...reading up on them, they appear to be really shady. |
| 05:59.30 | landley | The collections agency? |
| 05:59.43 | landley | Report 'em to the better business bureau and/or whoever regulates collection agencies. |
| 05:59.57 | landley | Entirely possible it's a scam. |
| 05:59.59 | Mirell | Hah... |
| 06:00.02 | Mirell | BBB is worthless. |
| 06:19.41 | Mirell | I need to start going to the gym after work. |
| 06:25.47 | *** join/#edev anxt (n=epoxy@S0106001c101b4132.ed.shawcable.net) |
| 06:40.01 | landley | I need to renew my old gym membership. |
| 06:40.14 | anxt | do irc pushups |
| 06:40.42 | anxt | i am pumped i got some stupid image put together with uclibc |
| 06:40.54 | anxt | managed to compile dhcpcd |
| 06:41.19 | anxt | now ot get libogg to compile |
| 06:42.02 | anxt | i guess i didn't make a c++ |
| 06:42.06 | anxt | oops lol |
| 06:52.07 | Mirell | Mmm. |
| 06:52.36 | anxt | dammit |
| 06:52.47 | anxt | at least i am getting paid for this lol |
| 07:07.32 | davidc__ | Mirell: tsk tsk tsk ;) |
| 07:11.10 | Mirell | davidc__: Already own a license to 10.5 |
| 07:11.18 | Mirell | However, that laptop is dead. |
| 07:11.20 | Mirell | So, eh. |
| 07:26.55 | anxt | 10.5 sucks |
| 07:32.16 | Mirell | Howso. |
| 07:36.56 | *** join/#edev mattepiu (i=59ca918f@gateway/web/ajax/mibbit.com/x-0a51f8ee49580b43) |
| 07:37.59 | mattepiu | new computer, new system, moved from redhat to gentoo, still 3,1 Gigabytes kernel image (but 4 Mb zImage). |
| 08:04.34 | *** join/#edev lyakh (n=lyakh@p57BD282B.dip0.t-ipconnect.de) |
| 08:14.11 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 08:15.36 | Mirell | Wait, what? |
| 08:15.42 | Mirell | 3.1GB kernel image? |
| 08:18.53 | anxt | THAT IS A LARGE KERNEL |
| 08:18.58 | RobotGuy | Mirell: Greets |
| 08:19.05 | RobotGuy | I don't think that is possible. |
| 08:19.16 | mattepiu | is useless Image, qemu starts unpacking and keeps forever |
| 08:19.56 | mattepiu | and is the only result I get from latest buildroot script |
| 08:20.30 | Mirell | 'lo RobotGuy |
| 08:21.06 | RobotGuy | My entire initrd image is 1.6 MB compressed, 6.7 MB uncompressed. |
| 08:21.41 | anxt | my kernel is under 1mb |
| 08:21.57 | anxt | and my rootfs is like 70 mb |
| 08:21.57 | RobotGuy | I have I2C, RTC, wireless, etc stuff in my kernel. |
| 08:22.19 | anxt | and i want to stab people to death |
| 08:22.28 | RobotGuy | anxt: Are you using a thumb drive? |
| 08:22.34 | anxt | no |
| 08:22.40 | anxt | i am using virtualbox |
| 08:22.56 | anxt | i made a thumbdrive image before though |
| 08:22.58 | RobotGuy | Are you using a Hammer? |
| 08:23.01 | anxt | with grub |
| 08:23.10 | anxt | x86 |
| 08:23.16 | RobotGuy | Oh, OK. |
| 08:23.25 | anxt | but i would like to hit something with a hammer |
| 08:23.34 | Mirell | Stop. |
| 08:23.36 | Mirell | Hammer Time. |
| 08:23.38 | Mirell | Doo dee dee doo. |
| 08:23.39 | anxt | target is prolly an arm |
| 08:23.51 | RobotGuy | Hammer is an ARM. |
| 08:24.02 | anxt | i know |
| 08:24.59 | anxt | someone put high in my boots |
| 08:25.01 | RobotGuy | My Hammer will hit your arm if you don't behave. ;) |
| 08:25.17 | anxt | i wasn't high a minuter ago |
| 08:25.17 | RobotGuy | anxt has high boots. |
| 08:25.44 | RobotGuy | Mirell: Do you have a Hammer? |
| 08:27.11 | RobotGuy | Is the second serial port /dev/ttySAC1 ? |
| 08:27.18 | anxt | wonder if g++ will work in new image |
| 08:27.35 | RobotGuy | New image? |
| 08:27.58 | anxt | using buildroot to make .ext2 |
| 08:28.07 | RobotGuy | I make ext2 also. |
| 08:28.54 | *** join/#edev harpal (n=Harpal@124.125.76.8) |
| 08:29.37 | mattepiu | my target is arm customn board |
| 08:30.40 | anxt | i am not as high now |
| 08:30.49 | anxt | what happened there |
| 08:31.40 | RobotGuy | mattepiu: What is the ARM MCU you use? |
| 08:32.06 | anxt | i have an arm board coming to work |
| 08:32.17 | anxt | i havce noclue how to setr up a bootstrap loader |
| 08:37.30 | *** join/#edev steliosk (n=Stelios@85.72.92.83) |
| 08:37.31 | RobotGuy | anxt: Do you know which board you are getting? |
| 08:37.51 | anxt | i think the first one is an arm |
| 08:37.57 | anxt | then an x86 after that |
| 08:37.58 | mattepiu | ARM MCU? arm926 integrator board |
| 08:38.12 | anxt | x86 with cf wounds easy |
| 08:38.14 | RobotGuy | mattepiu: Do you have a URL for that? |
| 08:38.23 | anxt | sounds |
| 08:38.45 | anxt | a guy from work is ordering them |
| 08:38.51 | anxt | without consuklting me |
| 08:39.26 | anxt | so i might hafta stick one in the closet |
| 08:39.32 | anxt | to dl torrents for me |
| 08:39.34 | mattepiu | Robotguy: no, the board is still in development, I´m still trying to build a generic minimal system with buildroot |
| 08:40.14 | anxt | i am gonna do usb alsa on it |
| 08:40.22 | anxt | unless i can find a minipci card |
| 08:40.43 | mattepiu | Robotguy: till now only glibc toolchain described in Building embedded linux system book worked |
| 08:41.43 | RobotGuy | mattepiu: buildroot uses uLibc. |
| 08:42.20 | anxt | i manaed to bbuild a saystem with that |
| 08:42.36 | anxt | and wget dhcpcd |
| 08:42.37 | mattepiu | Robotguy I know, I'd like uclibc worked also, here |
| 08:43.43 | mattepiu | RobotGuy: but I tried buildroot on multiple system and multiple distro, result is the same - 3 Gb Image and 4 Mb zImage... |
| 08:44.05 | anxt | mattepiu: THAT IS WHACKED |
| 08:44.27 | anxt | DID YOU INCLDE KDE? |
| 08:44.32 | anxt | err sorry caps |
| 08:44.38 | anxt | i am pretty high |
| 08:44.47 | RobotGuy | mattepiu: I don't get how you could get such a large kernel image. |
| 08:45.06 | anxt | my x86 kernel is like under 2 mb |
| 08:45.10 | anxt | on my normal box |
| 08:45.28 | RobotGuy | I've never seen a Linux kernel over 2 MB. |
| 08:45.38 | mattepiu | RobotGuy: I suspect is full of zero. My x86 and x86_64 kernels are normally sized |
| 08:45.51 | anxt | mattepiu: hexdump it |
| 08:45.53 | anxt | and grep |
| 08:45.58 | RobotGuy | mattepiu: You must mean the entire initrd image. |
| 08:46.24 | mattepiu | RobotGuy: Image = 3 G , zImage = 4/5 Mb |
| 08:46.31 | anxt | even an initrd of 3 g |
| 08:46.39 | RobotGuy | You shouldn't include any extra software except Busybox in your initrd. |
| 08:47.25 | RobotGuy | That's the only way I can see you getting a large image. |
| 08:47.48 | mattepiu | I have just busybox activated in menuconfig.... can I ask to NOT create anything but the kernel? (binutils-gcc-uclibc of course) |
| 08:48.48 | mattepiu | I think is a bug, but nobody here seems to have got it, and this is strange |
| 08:49.08 | *** join/#edev bbradley (n=bbradley@82-32-35-42.cable.ubr05.azte.blueyonder.co.uk) |
| 08:50.07 | RobotGuy | mattepiu: What exactly do you have in your .config for the initrd? |
| 08:51.20 | mattepiu | machines givin me 3 Gb kernel Image: x86 gentoo , x86_64 red hat, x86_64 4_dual_xeon red hat server |
| 09:09.38 | mattepiu | RobotGuy: I didn't modified any initrd option after unpacking buildroot... should I do something there? |
| 09:11.49 | mattepiu | I mean, I did "make config" and didn't touched any initrd option |
| 09:12.02 | RobotGuy | mattepiu: You need to be sure there is not any additional software selected, except for Busybox, in the Packages section. |
| 09:12.57 | mattepiu | RobotGuy: there's nothing else, I'm sure |
| 09:13.54 | RobotGuy | You need to check to be sure. |
| 09:14.15 | RobotGuy | You shouldn't need to do make config. It should be make menuconfig. |
| 09:20.01 | mattepiu | yep , make menuconfig. Just checked it gets busybox and fakeroot.... could fakeroot be the issue? |
| 09:22.56 | mattepiu | it slows down a lot at the line: GZIP arch/arm/boot/compressed/piggy.gz |
| 09:23.25 | RobotGuy | mattepiu: I don't have anything but Busybox in my initrd image. |
| 09:23.56 | RobotGuy | Well, Busybox and kernel modules. |
| 09:24.03 | mattepiu | RobotGuy: ok thanks, I'll try to remove fakeroot, I suspect is the skeleton option |
| 09:24.35 | RobotGuy | mattepiu: That could be what is causing you problems. |
| 09:29.18 | mattepiu | RobotGuy: just to check, could you give me your .config? (through pastebin or what you want...) |
| 09:34.53 | RobotGuy | mattepiu: Sure, hang on a sec. |
| 09:37.50 | RobotGuy | mattepiu: http://www.pastebin.ca/867300 |
| 09:40.25 | mattepiu | thank you... however it looks like it's the kernel, I got in kernel dir and did http://pastebin.com/m28cc296e make command (no initrd) and still got 3,1 Gb |
| 09:51.59 | RobotGuy | mattepiu: No, it is not the kernel. |
| 09:52.22 | RobotGuy | Kernel is built outside of buildroot at present. |
| 09:53.31 | RobotGuy | I don't really know what all that means, sorry. |
| 09:53.50 | RobotGuy | What do you think is the kernel? |
| 09:57.09 | *** join/#edev emil|k (n=emil@e179238033.adsl.alicedsl.de) |
| 10:02.46 | mattepiu | RobotGuy: I built the kernel without any root initrd (see command above) an got same result... is kernel or gmp/uclibc |
| 10:03.21 | mattepiu | RobotGuy: now trying with your config, modified as less as possible |
| 10:29.39 | harpal | I have made symbolic link to /proc/mounts and /etc/mtab but when I remount my file system as readwrite mode than it shows me warning: cant open /etc/mtab: No such file or directory |
| 10:35.56 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 10:58.22 | *** join/#edev __lifeless (n=patrick@89.20.118.236) |
| 11:19.37 | mattepiu | RobotGuy-Out: buildroot doesn't even starts with that config, you'rre using an old buildroot snapshot and I'm reportng a bug of recent snapshots |
| 13:11.13 | *** join/#edev purl (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
| 13:11.13 | *** topic/#edev is a channel for the discussion of hacking small consumer electronic devices, mostly ARM based; Hammer Development boards now available at http://www.tincantools.com/ |
| 13:15.04 | *** join/#edev agaffney (n=agaffney@gentoo/developer/pdpc.active.agaffney) |
| 13:25.52 | *** join/#edev dbtid (i=qwi0tu5z@cpe-71-72-252-171.columbus.res.rr.com) [NETSPLIT VICTIM] |
| 13:25.52 | *** join/#edev dfarnsworth (n=dfarnswo@65.200.49.156) [NETSPLIT VICTIM] |
| 13:25.54 | *** join/#edev bbradley (n=bbradley@82-32-35-42.cable.ubr05.azte.blueyonder.co.uk) [NETSPLIT VICTIM] |
| 13:25.55 | *** join/#edev Redhatter (n=stuartl@gentoo/developer/redhatter) |
| 13:25.55 | *** join/#edev CP|home (n=nobody@www.crouse-house.com) [NETSPLIT VICTIM] |
| 13:25.55 | *** join/#edev harpal (n=Harpal@124.125.76.8) [NETSPLIT VICTIM] |
| 13:25.55 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) [NETSPLIT VICTIM] |
| 13:25.55 | *** join/#edev wmat (n=btraynor@gromit.mixdown.ca) [NETSPLIT VICTIM] |
| 13:25.55 | *** join/#edev roxfan (n=dunno@69.1-200-80.adsl-dyn.isp.belgacom.be) [NETSPLIT VICTIM] |
| 13:25.55 | *** join/#edev ccherrett (n=chris@S0106000c413c715f.ed.shawcable.net) [NETSPLIT VICTIM] |
| 13:25.55 | *** join/#edev davidc__ (n=davidc__@S01060002b360aacd.vc.shawcable.net) [NETSPLIT VICTIM] |
| 13:25.55 | *** join/#edev MonMotha (n=monmotha@c-69-245-153-193.hsd1.in.comcast.net) [NETSPLIT VICTIM] |
| 13:25.55 | *** join/#edev SpeedEvil (n=mauve@tor/regular/SpeedEvil) [NETSPLIT VICTIM] |
| 13:27.38 | *** join/#edev sjhill (n=sjhill@real.realitydiluted.com) |
| 13:27.38 | *** join/#edev steliosk (n=Stelios@85.72.92.83) |
| 13:28.30 | *** join/#edev cbrake_a1ay (n=cbrake@oh-69-34-21-229.sta.embarqhsd.net) |
| 13:28.30 | *** join/#edev Mirell (n=mirell@quiche.ling.utexas.edu) |
| 13:28.30 | *** join/#edev ds2 (i=noinf@netblock-66-245-251-24.dslextreme.com) |
| 13:28.55 | *** join/#edev jbevren (n=jbevren@dumt-04-044.dsl.netins.net) [NETSPLIT VICTIM] |
| 13:28.55 | *** join/#edev __lifeless (n=patrick@89.20.118.236) |
| 13:28.55 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 13:28.55 | *** join/#edev emil|k (n=emil@e179238033.adsl.alicedsl.de) |
| 13:28.55 | *** join/#edev anxt (n=epoxy@S0106001c101b4132.ed.shawcable.net) |
| 13:28.55 | *** join/#edev chouimat|busy (n=dieu@kde/developer/chouinard) |
| 13:28.55 | *** join/#edev sunrunner20 (n=jamesdoe@pool-96-226-38-56.dllstx.fios.verizon.net) [NETSPLIT VICTIM] |
| 13:28.56 | *** join/#edev huaynac (n=huaynac@61.6.65.198) |
| 13:28.56 | *** join/#edev ccjoe (n=ccjoe@75-169-14-207.slkc.qwest.net) |
| 13:28.56 | *** join/#edev Crofton (n=balister@66.207.66.26) |
| 13:28.56 | *** mode/#edev [+o jbevren] by irc.freenode.net |
| 13:29.01 | *** join/#edev ico2 (n=ico2@modem-3876.cougar.dialup.pol.co.uk) |
| 13:36.32 | *** join/#edev ico2 (n=ico2@modem-3876.cougar.dialup.pol.co.uk) |
| 14:06.38 | *** join/#edev cajaka (n=none@cpe-24-58-193-200.twcny.res.rr.com) |
| 14:09.06 | *** join/#edev prpplague (n=dave@mail.americanmicrosystems.com) |
| 14:10.41 | *** join/#edev RobotGuy-Out (n=RobotGuy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 14:20.38 | *** join/#edev GPSFan (n=kenm@12.10.255.246) |
| 14:20.38 | *** mode/#edev [+o GPSFan] by ChanServ |
| 14:33.02 | *** join/#edev chouimat|work (n=mchouina@209.217.106.98) |
| 14:33.45 | chouimat|work | morning |
| 14:34.48 | prpplague | chouimat|work: greetings |
| 14:36.47 | chouimat|work | hi prpplague |
| 14:38.15 | prpplague | chouimat|work: whats cookin? |
| 14:38.17 | prpplague | wmat: ping |
| 14:38.20 | GPSFan | prpplague: morning.. |
| 14:38.43 | chouimat|work | prpplague: nothing much ... try to understand why stuff work on my unit and not on the QA ones |
| 14:38.47 | prpplague | GPSFan: greetings |
| 14:38.54 | prpplague | GPSFan: whats new? |
| 14:39.09 | prpplague | chouimat|work: oh thats always fun |
| 14:39.43 | prpplague | chouimat|work: i love it when they give the software guys a "pre-production" unit that has all kinds of items that aren't the same |
| 14:39.52 | chouimat|work | prpplague: I think I know the problem ... you know one in the kci (keyboard chair interface) |
| 14:40.12 | prpplague | chouimat|work: i.e. "oh we forgot and change the interrupt from eint11 to eint10 since it was easier to route on the pcb" |
| 14:40.25 | chouimat|work | prpplague: hehe |
| 14:40.26 | GPSFan | prpplague: do you have any userland tools for tweaking the lcdcon registers on the S3C2410? I've got my monochrome LCD hooked up and am trying to setup the LCDCONx reg's and it's a pain to recompile for each tweak. |
| 14:40.31 | wmat | prpplague: pong |
| 14:40.42 | prpplague | GPSFan: use devmem |
| 14:40.48 | prpplague | wmat: get your kit? |
| 14:40.59 | prpplague | GPSFan: devmem works great for doing those tweaks |
| 14:41.13 | prpplague | GPSFan: what display are you using? |
| 14:41.22 | prpplague | GPSFan: one sec |
| 14:41.30 | prpplague | GPSFan: it should be on your rootfs for the hammer |
| 14:41.41 | wmat | prpplague: nothing yet, it's likely stuck in customs |
| 14:41.49 | GPSFan | prpplague: sharp lm32k10 |
| 14:42.26 | GPSFan | prpplague: are there man pages anywhere? |
| 14:42.56 | prpplague | GPSFan: http://www.lartmaker.nl/lartware/port/ |
| 14:43.03 | prpplague | GPSFan: look at the bottom of the page |
| 14:43.13 | prpplague | GPSFan: its so simple, no man page has ever been needed |
| 14:43.16 | *** join/#edev FireEgl (i=FireEgl@4.0.0.0.1.0.0.0.c.d.4.8.0.c.5.0.1.0.0.2.ip6.arpa) |
| 14:43.45 | prpplague | GPSFan: its just a command line util to allow you to read and write to physical memory addresses |
| 14:43.54 | GPSFan | prpplague: thanks that's just what I was looking for... |
| 14:44.07 | prpplague | GPSFan: its included in your hammer rootfs |
| 14:45.05 | GPSFan | LCDCON3 => 0x4d000008 is the address from the manual is that mapped somewhere else or do I use that address? |
| 14:45.55 | prpplague | GPSFan: devmem allows you to use the actually phsical address, so you would use the 0x4d000008 |
| 14:46.03 | GPSFan | ;>)) |
| 14:47.16 | prpplague | GPSFan: got a url for the datasheet on that lcd? |
| 14:48.43 | prpplague | GPSFan: you get that from earthlcd? |
| 14:48.52 | GPSFan | http://www.esskabel.de/Datenblaetter/LCD-Specs/Sharp/LM/LM32K10.pdf |
| 14:49.24 | GPSFan | prpplague: no I had it (and a few like it) left over from a "long-ago" project. |
| 14:49.31 | prpplague | ahh ok |
| 14:50.34 | prpplague | GPSFan: yea thats gonna be pretty easy to get working |
| 14:50.58 | prpplague | GPSFan: look at the lcd config data for the aml-m5900 |
| 14:51.12 | prpplague | GPSFan: only difference is that the m5900 lcd is 160x160 |
| 14:51.19 | prpplague | GPSFan: just need to tweak those values |
| 14:51.29 | GPSFan | I got most of the settings figured out, and checked the waveforms out with a scope, it's just the last bit of tweaking that I needed |
| 14:52.16 | GPSFan | I may have to buffer the VLINE signal, it seems to be loaded down more than the rest by the lcd. |
| 14:54.01 | prpplague | GPSFan: you aren't buffering the signals? i'd have expected you'd need that with a 5v display like that |
| 14:54.10 | prpplague | GPSFan: you are using it with the hammer? |
| 14:54.57 | GPSFan | prpplague: yes with the hammer, I was going to try without a buffer, but after seeing the condition of the VLINE signal I may need to add one. |
| 14:55.17 | prpplague | GPSFan: yea, i'd suspect you'd need one |
| 14:55.29 | prpplague | we generally use a 245 to do it |
| 14:55.49 | GPSFan | yeah, that's all thats needed, 8 signals is enough. |
| 14:55.56 | *** join/#edev emil|k (n=emil@e179238033.adsl.alicedsl.de) |
| 14:55.57 | prpplague | GPSFan: yep |
| 14:56.06 | prpplague | GPSFan: you gonna post this info? |
| 14:56.17 | GPSFan | oh yes, as soon as I'm done, |
| 14:56.35 | prpplague | GPSFan: want some more lcd's to tinker with? |
| 14:57.16 | GPSFan | the lm32k10 is long since out of production, but there are decendents that run of 3.3V. |
| 14:57.35 | prpplague | http://www.allelectronics.com/cgi-bin/item/LCD-101/365/256_X_128_LCD_PANEL_.html |
| 14:57.43 | prpplague | thats the one i need to get working |
| 14:57.47 | prpplague | just haven't had time |
| 14:59.32 | GPSFan | it needs an op-amp bias generator. one quad op-amp should do. |
| 15:00.08 | prpplague | GPSFan: yea, i have a couple of parts picked out that would do it |
| 15:01.41 | GPSFan | wait, it's not a real LCD interface, it has a controller with command structure etc. |
| 15:01.47 | prpplague | GPSFan: yea |
| 15:05.37 | prpplague | GPSFan: SpeedEvil thinks we can tweak the tft fb support to use just 8-bit red data lines to support a full tft display |
| 15:06.43 | prpplague | GPSFan: http://store.earthlcd.com/ACX705AKM-7-ACX705AKM7?category=193 |
| 15:11.14 | prpplague | GPSFan: we use a lcd similiar to this - http://www.sparkfun.com/commerce/product_info.php?products_id=8490 |
| 15:11.31 | prpplague | GPSFan: we can get them cheap, so if we can that working, those would be nice to sale |
| 15:11.54 | SpeedEvil | prpplague: it's almost trivial |
| 15:12.14 | prpplague | SpeedEvil: yea, thats what i'm thinking, just gotta wire it up and test it |
| 15:12.14 | SpeedEvil | prpplague: The hard part isn't the hardware, it's making the software play nice |
| 15:12.28 | prpplague | SpeedEvil: indeed |
| 15:12.42 | SpeedEvil | As a trivial example, RGB24 you just get whatever red you set |
| 15:13.12 | prpplague | indeed |
| 15:13.45 | prpplague | wmat: surprised you haven't received your kit yet |
| 15:13.45 | SpeedEvil | Though you don't do that unless you're silly - memory bandwidth hurts |
| 15:13.47 | prpplague | wmat: do you normally have problems with customs on components/dev parts? |
| 15:14.17 | prpplague | SpeedEvil: yea no way around it perse |
| 15:14.26 | SpeedEvil | prpplague: unfortunately, I've been unable to get nice code working for you - I've had a heating failure, and am currently dismantliing my boiler. |
| 15:14.28 | prpplague | SpeedEvil: since we have to format the data as rgb |
| 15:14.35 | SpeedEvil | prpplague: why? |
| 15:14.37 | prpplague | SpeedEvil: ug |
| 15:14.50 | prpplague | SpeedEvil: thats just the way the controller works |
| 15:15.18 | SpeedEvil | prpplague: You can misuse pallette mode - and patch the software to set the 'wrong' colors. |
| 15:15.42 | prpplague | SpeedEvil: right, we'll parse the red bits to configure it for rgb |
| 15:16.16 | SpeedEvil | App wants FF00FF, you mis-set it as E3XX or something |
| 15:16.21 | SpeedEvil | App wants FF00FF, you mis-set it as E3XXXX or something |
| 15:16.50 | prpplague | yea |
| 15:16.51 | SpeedEvil | (where first three bits are red, second three are green, and last two are blue |
| 15:19.46 | prpplague | yep |
| 15:19.49 | prpplague | thats what i was thinking |
| 15:28.23 | SpeedEvil | Also. Dealextreme.com has PSP LCDs at $30 delivered. |
| 15:28.59 | prpplague | SpeedEvil: yea, sparkfun has them as well, but we can't get them in bulk, plus the size is pushing the limits of the s3c2410 |
| 15:29.23 | prpplague | SpeedEvil: need to keep the size below 320x240 |
| 15:30.01 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 15:30.15 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 15:30.24 | SpeedEvil | Though that has much better memory bandwidth. |
| 15:34.09 | prpplague | SpeedEvil: yea you can run a 640x480, it just takes alot more to operate |
| 15:34.20 | prpplague | SpeedEvil: tomw is running a 640x480 |
| 15:37.09 | SpeedEvil | (It's in the neo) |
| 15:37.31 | SpeedEvil | On my stack of stuff to do is to measure the supply current at different pixel-clocks |
| 15:38.11 | SpeedEvil | But just got the multimeter out to head into the attic now. House <11C is more urgent :) |
| 15:39.19 | prpplague | SpeedEvil: the neo uses a standalone controller |
| 15:41.56 | prpplague | SpeedEvil: you have to get a lcd with a smart lcd driver chipset to really get the advantage of the 640x480 - "Toshiba JBT6K74 TFT LCD Driver Chipset" |
| 15:57.58 | *** join/#edev ico2 (n=ico2@modem-3516.chimpanzee.dialup.pol.co.uk) |
| 16:01.33 | *** join/#edev chouimat (n=mchouina@209.217.106.98) |
| 16:04.07 | SpeedEvil | prpplague: no it doesn't. |
| 16:04.36 | SpeedEvil | The freerunner uses a seperate graphics chip - on a 30Mb/s bus :/ the current version is connected directly to the LCD pins |
| 16:04.54 | prpplague | SpeedEvil: but the lcd has a "smart" driver chipset |
| 16:05.06 | SpeedEvil | The toshiba chip is in fact a _really_ dumb shift-register |
| 16:06.00 | SpeedEvil | I'm fairly certain - I was talking to the lead software architect for the latest version on this about how the new version and the current one compare |
| 16:06.43 | prpplague | SpeedEvil: still there is a pretty big performance hit on going from 320x240 to 640x480 |
| 16:06.57 | SpeedEvil | Oh of course. It's a lot of data a second. |
| 16:07.09 | prpplague | SpeedEvil: i guess there isn't any reason we couldn't do some tests between the 320x240 and the 480x320 psp display |
| 16:07.11 | SpeedEvil | Especially if you've only got a 16 bit wide memory bus like the hammer |
| 16:07.21 | prpplague | SpeedEvil: but the fact remains that we can't get the psp display in bulk |
| 16:08.01 | SpeedEvil | prpplague: dealextreme from my experience on other lines are quite happy to deal in bulk |
| 16:08.20 | SpeedEvil | Though admittedly this was for a lower value item. |
| 16:09.23 | prpplague | SpeedEvil: for 1k qty? |
| 16:14.35 | prpplague | CosmicPenguin: hehe |
| 16:14.46 | prpplague | CosmicPenguin: that happening at work? |
| 16:14.52 | CosmicPenguin | mainly in the community |
| 16:14.55 | prpplague | ahh |
| 16:15.14 | CosmicPenguin | If you configure a TFT panel in the BIOS or in the X configuration, the behavior of the X driver is dramatically different |
| 16:15.51 | CosmicPenguin | and so we have people thinking they have a panel, and then complaining when only a few modes are available or it starts scaling when you switch modes |
| 16:16.56 | CosmicPenguin | And unfortunately, we don't have any way to tell if a panel is phsyically attached |
| 16:21.35 | prpplague | ahh |
| 16:24.34 | *** join/#edev adyer (n=adyer@adsl-69-223-13-78.dsl.chcgil.ameritech.net) |
| 16:24.43 | *** join/#edev _lifeless (n=patrick@89.20.115.112) |
| 16:26.21 | *** join/#edev emil|k (n=emil@f054012199.adsl.alicedsl.de) |
| 16:45.12 | SpeedEvil | prpplague: no - it was a query on 150 toys actually - but they got back to me promptly. It turned out they weren't required, so I never went ahead. |
| 16:46.15 | SpeedEvil | Total value would have been $1K |
| 16:47.03 | SpeedEvil | At $30K, personally I'd want to arrange a 'I pick them up in person' arrangement or something. I have had no problems with DX for multiple $10-$100 orders - but... |
| 17:19.21 | *** join/#edev samuelr (n=samuelr@dsl093-061-094.pit1.dsl.speakeasy.net) |
| 18:01.20 | *** join/#edev ico2 (n=ico2@modem-3516.chimpanzee.dialup.pol.co.uk) |
| 18:12.55 | *** join/#edev redblue (i=star@ppp090.108-253-207.mtl.mt.videotron.ca) |
| 18:17.31 | *** join/#edev chouimat|busy (n=dieu@r2351064.cidc.net) |
| 18:23.08 | prpplague | lordy lordy, i love patent trolls |
| 18:23.10 | prpplague | not |
| 18:23.44 | wmat | prpplague: I received a parcel notice from the post office, so I assume that's the hammer |
| 18:23.53 | prpplague | wmat: ahh ok good |
| 18:23.59 | prpplague | wmat: that mean its cleared customs? |
| 18:24.01 | wmat | prpplague: freakin $39.82 due though |
| 18:24.26 | wmat | stupid customs |
| 18:24.38 | prpplague | yikes |
| 18:26.22 | prpplague | anyone heard of patent issues revolving around implementation of a standard 802.11b/g ? |
| 18:41.31 | prpplague | ho ho hum |
| 18:41.33 | prpplague | http://www.setexasrecord.com/news/205753-texas-company-seeks-judgement-to-stop-potential-patent-suit |
| 18:44.18 | prpplague | i can't submit it to groklaw............ |
| 18:44.21 | prpplague | *cough* |
| 18:59.58 | *** mode/#edev [+o prpplague] by ChanServ |
| 19:19.49 | *** join/#edev _lifeless (n=patrick@89.20.118.121) |
| 19:33.04 | *** join/#edev davidc__ (n=davidc__@S01060002b360aacd.vc.shawcable.net) |
| 19:52.07 | dbtid | under linux 2.6.2x on an at91rm9200, isn't the write to the SPI dataflash supported?? |
| 19:52.14 | dbtid | reading is, but is write (for a block device) |
| 19:52.26 | dbtid | e.g. dd if=/some/file of=/dev/mtdblock/2 |
| 19:52.43 | dbtid | trying to change my u-boot environment on the fly |
| 19:54.12 | RobotGuy | Good Morning |
| 19:54.56 | *** join/#edev davidc__ (n=davidc__@S01060002b360aacd.vc.shawcable.net) |
| 19:58.13 | prpplague | RobotGuy: greetings |
| 19:59.05 | *** join/#edev ico2 (n=ico2@modem-1500.cougar.dialup.pol.co.uk) |
| 20:00.02 | RobotGuy | I built a bunch of software last night, to see what would and would not build. Some stuff can't be downloaded because it is not found. |
| 20:02.17 | RobotGuy | I have a 6.7 MB uncompressed initrd now. |
| 20:02.27 | prpplague | url's or version numbers just need to be updated |
| 20:09.11 | RobotGuy | Robert Sheer has his Hammer Kit now. |
| 20:09.35 | RobotGuy | I still have not gotten the Gadget Ethernet to work, although the Hammer side looks correct now. |
| 20:11.39 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 20:20.22 | *** join/#edev dataghost_ (n=_datagho@ppp-62-216-201-189.dynamic.mnet-online.de) |
| 20:27.39 | RobotGuy | prpplague: I haven't figured out how to initialize the second serial port yet. Is it /dev/ttySAC1 ? |
| 20:27.58 | prpplague | RobotGuy: yea, it should already be initialized |
| 20:28.41 | RobotGuy | prpplague: What do I need on the host side, besides cdc-ether, to have the Gadget Ethernet work? |
| 20:28.58 | prpplague | that should be it |
| 20:29.54 | RobotGuy | prpplague: Hmmmmmm. I believe I have the Hammer side setup right - the route to the host comes up. I can't ping from Hammer though. |
| 20:30.33 | prpplague | RobotGuy: did you bring up the interface on the hammer and assign an ip? |
| 20:30.41 | wmat | prpplague: hammer is on my desk now |
| 20:30.47 | prpplague | wmat: good deal |
| 20:31.02 | dataghost_ | Did someone cross-compile "screen-4.0.3" with uClibc? Any expirence? Got a wired problem :/ |
| 20:31.11 | RobotGuy | prpplague: Yes. Everything looks good on the Hammer side. Default gateway is the host at the other side of the link. |
| 20:31.35 | prpplague | RobotGuy: you have ip_forward set? |
| 20:31.39 | prpplague | RobotGuy: on the host pc? |
| 20:32.58 | *** join/#edev __lifeless (n=patrick@90.150.214.219) |
| 20:32.59 | RobotGuy | prpplague: Not sure. I just switched from Fedora 8 to Kubuntu. |
| 20:33.46 | RobotGuy | prpplague: I can't even ping the host from Hammer though. Is that still affected by ip_forward? |
| 20:35.41 | prpplague | RobotGuy: try this: echo 1 > /proc/sys/net/ipv4/ip_forward |
| 20:35.48 | prpplague | RobotGuy: they try pinging it |
| 20:38.14 | RobotGuy | prpplague: No ping |
| 20:38.50 | prpplague | RobotGuy: paste your ifconfig and netstat -rn on pastebin.ca |
| 20:38.52 | RobotGuy | prpplague: It might not be turned on in the kernel. I am about to configure and build a new 2.6.23.12 kernel. |
| 20:38.57 | prpplague | RobotGuy: from the host side |
| 20:39.26 | prpplague | RobotGuy: why don't you stick with a stock distro while you are doing dev work, that way you don't fight two battles at once |
| 20:39.42 | prpplague | wmat: what do you think of the hammer? |
| 20:40.20 | RobotGuy | prpplague: I am on a stock distro. |
| 20:40.37 | prpplague | RobotGuy: why are you recompiling the kernel then? |
| 20:40.43 | wmat | prpplague: so far, it's well wrapped ;) |
| 20:40.53 | prpplague | hehe |
| 20:41.30 | RobotGuy | prpplague: I know I don't have an interfacce up that I should, but don't know what to setup. http://www.pastebin.ca/867651 |
| 20:41.45 | RobotGuy | prpplague: Kernel for PC. |
| 20:42.41 | prpplague | RobotGuy: uh, how do you expect to ping it with no inferface up? |
| 20:42.56 | prpplague | RobotGuy: do a ifconfig -a |
| 20:43.49 | RobotGuy | prpplague: There is no other interface to configure. |
| 20:44.07 | prpplague | RobotGuy: ok , then you don't have something working |
| 20:44.41 | prpplague | RobotGuy: if you do a cat /proc/bus/usb/devices , is the gadget listed? |
| 20:44.50 | *** part/#edev dataghost_ (n=_datagho@ppp-62-216-201-189.dynamic.mnet-online.de) |
| 20:49.58 | prpplague | eh gads there is no way to check to see which patent in 802.11 covers roaming, you have to thru each patent one by one :( |
| 20:50.11 | *** join/#edev chouimat|work (n=mchouina@209.217.106.98) |
| 20:51.06 | adyer | what, you want have wireless AND roaming? egads, man - the horror! :-) |
| 20:51.48 | adyer | of course you could just use the old "frame of reference" argument - you're still and the base stations are moving |
| 20:53.30 | RobotGuy | prpplague: I don't see anything recognizable as the Gadget Ethernet. |
| 20:53.59 | prpplague | RobotGuy: pastebin.ca your /proc/bus/net/devices |
| 20:55.40 | RobotGuy | prpplague: http://www.pastebin.ca/867676 |
| 20:56.40 | prpplague | RobotGuy: indeed , looks like maybe you still don't have something correct on your hammer config |
| 21:00.24 | prpplague | code looks correct and matches the schematic |
| 21:00.35 | RobotGuy | prpplague: Hmmmm |
| 21:00.39 | prpplague | RobotGuy: pastebin.ca your kernel config for the hammer |
| 21:03.08 | RobotGuy | prpplague: Kernel config - http://www.pastebin.ca/867692 |
| 21:04.56 | prpplague | uh why do you have ipv6 enable? thats a total waste |
| 21:06.00 | prpplague | RobotGuy: do you have devmem in your rootfs? |
| 21:07.16 | RobotGuy | prpplague: Probably not. Didn't see it in the packages. |
| 21:07.33 | prpplague | RobotGuy: its not in the packages |
| 21:07.52 | prpplague | RobotGuy: its going to be hard to debug unless you add that back into your rootfs |
| 21:08.31 | RobotGuy | prpplague: How do I add it if it isn't in packages? |
| 21:08.45 | prpplague | RobotGuy: http://www.lartmaker.nl/lartware/port/devmem2.c |
| 21:08.51 | prpplague | RobotGuy: download and compile that |
| 21:09.04 | prpplague | RobotGuy: arm-linux-uclibc-gcc -o devmem devmem2.c |
| 21:09.30 | prpplague | wmat: you get it unwrapped yet? |
| 21:10.07 | wmat | prpplague: yes, i'm staring at it now :) |
| 21:10.19 | prpplague | wmat: hehe |
| 21:10.36 | prpplague | wmat: i'm curious what the folks there at codesourcery think of it |
| 21:11.54 | *** join/#edev redblue (i=star@ppp166.108-253-207.mtl.mt.videotron.ca) |
| 21:13.59 | prpplague | RobotGuy: if you do a small initrd, you can chroot to your thumb drive with no problems |
| 21:14.00 | prpplague | RobotGuy: per your email on PARTS |
| 21:14.12 | prpplague | RobotGuy: all you need is about a 512k rootfs that mounts the thumbdrive and then does a chroot |
| 21:16.59 | wmat | prpplague: the feedback i got from a few guys was that it looked cool |
| 21:17.12 | wmat | prpplague: i'm the only one with actual hardware, of course |
| 21:17.13 | prpplague | wmat: great |
| 21:17.17 | RobotGuy | prpplague: Ahh, OK, I think I follow that. So I have everything I need on the thumb drive then, like I was creating a huge initrd? |
| 21:17.42 | prpplague | wmat: yep |
| 21:18.22 | prpplague | RobotGuy: you get devmem compiled? |
| 21:18.38 | RobotGuy | prpplague: Not yet. Fixing another problem. |
| 21:18.47 | prpplague | ok |
| 21:19.33 | prpplague | RobotGuy: no need to use kexec like that |
| 21:19.44 | prpplague | RobotGuy: per the email from dave hylands |
| 21:20.07 | prpplague | RobotGuy: just use a 1mb kernel, plus a 512k initrd and then chroot to your external usb storage |
| 21:20.16 | RobotGuy | prpplague: OK, got devmem compiled. |
| 21:20.34 | prpplague | RobotGuy: send that over to your hammer, |
| 21:20.58 | prpplague | i usually like to put devmem in the jffs2 that way its always available while testing initrd's |
| 21:22.45 | RobotGuy | prpplague: In /mnt/flash ? |
| 21:22.53 | prpplague | yea |
| 21:23.41 | prpplague | RobotGuy: you have rx enabled in busybox? |
| 21:24.11 | RobotGuy | prpplague: I have everything enabled that is stock - didn't change anything for Busybox. |
| 21:24.28 | prpplague | RobotGuy: hehe, nothing really "Stock" about busybox |
| 21:24.39 | prpplague | RobotGuy: check to see if you have rx enable |
| 21:24.50 | prpplague | RobotGuy: i.e. issue the busybox command with no args |
| 21:24.57 | prpplague | RobotGuy: it will list all the commands |
| 21:25.19 | prpplague | RobotGuy: i guess since you have your thumb drive, you could always just copy it to /usr/bin |
| 21:25.45 | RobotGuy | prpplague: No rx. |
| 21:25.51 | prpplague | RobotGuy: :( |
| 21:26.05 | prpplague | RobotGuy: just copy devmem to your thumb drive's /usr/bin dir |
| 21:26.21 | prpplague | just make sure the permissions are set 755 |
| 21:26.22 | RobotGuy | prpplague: I installed rz, rx, etc though. |
| 21:26.48 | prpplague | RobotGuy: ahh ok, well if you have rz you can transfer it using that |
| 21:26.53 | RobotGuy | prpplague: :D |
| 21:27.23 | prpplague | RobotGuy: just not used to such big distros |
| 21:28.15 | *** join/#edev wiml (n=wiml@underhill.hhhh.org) |
| 21:28.48 | RobotGuy | prpplague: I am finding things that are broken. ;) |
| 21:29.00 | prpplague | RobotGuy: since most people never use them |
| 21:29.03 | prpplague | hehe |
| 21:29.14 | RobotGuy | prpplague: I am nit most people. ;P |
| 21:29.21 | prpplague | indeed |
| 21:29.41 | RobotGuy | prpplague: That tends to make me a pretty good tester. |
| 21:30.29 | prpplague | indeed |
| 21:31.28 | RobotGuy | prpplague: Speaking of which, I can't build a native toolchain because binutils is blowing up because of the makeinfo thing. |
| 21:32.12 | RobotGuy | prpplague: Do I have to put that patch in another location? |
| 21:32.14 | prpplague | RobotGuy: we fixed the makeinfo "thing" |
| 21:32.26 | RobotGuy | prpplague: For the cross tool chain, yes we did. |
| 21:32.49 | RobotGuy | prpplague: This is for the native toolchain build. |
| 21:33.04 | prpplague | RobotGuy: why build a native toolchain? |
| 21:33.31 | RobotGuy | prpplague: Because it will be cool to show I can build software right on Hammer, and it has a high coolness factor, AND because I can. :D |
| 21:33.48 | prpplague | doesn't make any sense to me |
| 21:33.59 | RobotGuy | prpplague: Doesn't have to make sense to you. :) |
| 21:34.34 | prpplague | indeed |
| 21:35.11 | RobotGuy | prpplague: I know where to turn it on. :) It's called native toolchain |
| 21:35.18 | prpplague | RobotGuy: where? |
| 21:35.56 | RobotGuy | prpplague: native toolchain in the target filesystem, right below gawk. |
| 21:36.18 | *** join/#edev ico2 (n=ico2@modem-1500.cougar.dialup.pol.co.uk) |
| 21:36.30 | prpplague | RobotGuy: in the package options? |
| 21:36.34 | RobotGuy | prpplague: Yes |
| 21:37.00 | prpplague | ahh |
| 21:37.14 | prpplague | RobotGuy: got devmem on your device yet? |
| 21:37.36 | RobotGuy | prpplague: Not quite yet. Got to get rz/sz on my PC first. |
| 21:40.20 | prpplague | interesting, native toolchain downloaded 2.17 binutils instead if using the already downloaded one |
| 21:40.41 | RobotGuy | prpplague: That would be a problem, right? |
| 21:40.49 | prpplague | yea |
| 21:41.07 | RobotGuy | prpplague: I thought that was weird. |
| 21:41.25 | RobotGuy | I knew we had already fixed that problem. |
| 21:46.22 | prpplague | RobotGuy: yea, looks like its building a different set of bintutils |
| 21:46.25 | prpplague | RobotGuy: and gcc |
| 21:46.42 | prpplague | RobotGuy: so you'll need to add that patch to the 2.17 binutils too |
| 21:46.45 | RobotGuy | prpplague: OK, got devmem on Hammer now. |
| 21:46.55 | RobotGuy | prpplague: Could not put it in /mnt/flash |
| 21:47.00 | prpplague | RobotGuy: why? |
| 21:47.13 | RobotGuy | prpplague: I think it's read only. |
| 21:47.23 | prpplague | uh shouldn't be |
| 21:47.28 | prpplague | oh well whatever |
| 21:47.42 | prpplague | do ./devmem 0x56000020 |
| 21:47.51 | RobotGuy | prpplague: JFFS2 doesn't begin on erase boundary, making read only. |
| 21:48.19 | prpplague | RobotGuy: ahh you've got the boundaries messed up |
| 21:48.21 | prpplague | RobotGuy: oh well |
| 21:48.23 | prpplague | do ./devmem 0x56000020 |
| 21:48.29 | RobotGuy | prpplague: Did that. |
| 21:48.36 | prpplague | response? |
| 21:48.49 | RobotGuy | prpplague: I didn't change anything on the jffs2 boundary. |
| 21:49.03 | RobotGuy | Memory mapped at address 0x40006000. |
| 21:49.03 | RobotGuy | Value at address 0x56000020 (0x40006020): 0x1000000 |
| 21:49.50 | prpplague | RobotGuy: are the usb gadget drivers loaded? |
| 21:50.16 | RobotGuy | prpplague: usbnet and cdc-ether are loaded |
| 21:50.42 | prpplague | RobotGuy: and the s3c2410 udc driver? |
| 21:51.24 | adyer | isn't it g_ether on the gadget side? |
| 21:51.42 | RobotGuy | adyer: I don't think so. |
| 21:52.05 | prpplague | adyer: they changed awhile back |
| 21:52.22 | prpplague | RobotGuy: is the udc driver loaded? |
| 21:55.01 | prpplague | adyer: actually you are right |
| 21:55.06 | prpplague | adyer: its listed as g_ether |
| 21:55.12 | prpplague | adyer: i always compile it in |
| 21:55.23 | prpplague | adyer: g_ether.ko |
| 21:55.32 | RobotGuy | prpplague: I believe I compiled that in also. |
| 21:55.44 | RobotGuy | prpplague: What is udc? |
| 21:55.48 | prpplague | RobotGuy: so they aren't as modules |
| 21:55.57 | prpplague | RobotGuy: repost your kernel config on pastebin.ca |
| 21:56.03 | RobotGuy | prpplague: OK |
| 21:58.09 | RobotGuy | prpplague: http://www.pastebin.ca/867760 |
| 21:59.16 | prpplague | RobotGuy: ok that looks correct, can you pastebin your dmesg? |
| 22:02.11 | RobotGuy | prpplague: http://www.pastebin.ca/867774 |
| 22:06.19 | prpplague | RobotGuy: hmm |
| 22:08.05 | prpplague | man have you got something funky going on |
| 22:12.17 | RobotGuy | prpplague: How though? I didn't change anything for the hammer.cfg and only changed hammer.ocd to look for a different initrd file name. |
| 22:12.46 | prpplague | RobotGuy: yea, you have lots of other problems that are conspiring against you |
| 22:13.01 | RobotGuy | prpplague: Like what? Being me? :P |
| 22:13.22 | prpplague | RobotGuy: 1) you have a thousand options turned on that don't need to be |
| 22:13.31 | prpplague | RobotGuy: 2) the jffs isn't on a 256k boundry |
| 22:13.43 | RobotGuy | prpplague: I didn't do anything with jffs. |
| 22:14.03 | prpplague | RobotGuy: you changed the .size values which changes the jffs2 |
| 22:14.11 | RobotGuy | prpplague: I turned ipv6 off in my kernel. Everything else stays. |
| 22:14.25 | RobotGuy | prpplague: Ahhh, the .size is not right.. |
| 22:16.31 | *** join/#edev samuelr (n=samuelr@dsl093-061-094.pit1.dsl.speakeasy.net) |
| 22:16.32 | ds2 | Mmmmm JFFS2 |
| 22:16.46 | prpplague | RobotGuy: how big is your compressed initrd? |
| 22:18.15 | RobotGuy | prpplague: I have a 1.6 MB compressed initrd, < 1 MB kernel. |
| 22:18.31 | RobotGuy | prpplague: It will be smaller without ipv6 |
| 22:19.02 | ds2 | Mmmm IPv6 addressable robots... Nanites }:-) |
| 22:19.08 | RobotGuy | prpplague: All the sizes I have on the wiki are wrong then. |
| 22:19.15 | prpplague | RobotGuy: ok so you are you going for 2mb for the initrd? |
| 22:19.39 | RobotGuy | prpplague: Hang on and I will give you exact info. Got to remake my initrd. |
| 22:20.00 | RobotGuy | prpplague: Well, that would be a good target. |
| 22:20.10 | RobotGuy | prpplague: < 2 MB initrd. |
| 22:20.21 | RobotGuy | prpplague: (compressed) |
| 22:20.21 | prpplague | RobotGuy: the size you need then is 0x340000 |
| 22:21.04 | prpplague | RobotGuy: 256k for bootloader, 1024k for the kernel, and 2048k for the initrd |
| 22:22.09 | prpplague | RobotGuy: and just a note, if you keep using the jtag to reflash your nor flash, you are going to use up all of your erase cycles |
| 22:23.11 | adyer | what flash are you using? most nor is > 10k erases these days |
| 22:27.26 | RobotGuy | prpplague: OK, well I don't want to use up all my flashes. I need to learn another way to do it then. |
| 22:28.49 | prpplague | one sec |
| 22:28.50 | prpplague | phone |
| 22:29.56 | MonMotha | most NAND flash is >10k these days.. |
| 22:30.04 | MonMotha | heck, lots of NOR devices are >100k |
| 22:30.08 | prpplague | RobotGuy: i think you got something misconfigured, mine worked first try |
| 22:31.03 | prpplague | MonMotha: indeed, however using the jtag app to reflash erases all 127 blocks everytime |
| 22:31.19 | RobotGuy | prpplague: What would I have misconfigured? |
| 22:31.40 | adyer | monmotha I was being conservative :-) |
| 22:32.08 | prpplague | RobotGuy: i'm guessing your s3c2410_udc is bailing with that error about debugfs |
| 22:32.15 | adyer | prpplague - doesn't matter the 100k erase spec is per sector |
| 22:32.39 | prpplague | adyer: well every sector is being erased |
| 22:34.39 | adyer | prpplague so? if robotguy doesn't put any writable filesystem he can flash the part 100000 times (insert quote about monkeys, typewriters, and shakespeare here) |
| 22:34.58 | MonMotha | prpplague: still, it would take a LONG time to wear it out |
| 22:35.59 | adyer | also the 100k is a min spec - what really happens is the time to erase a sector lengthens as it gets 'older |
| 22:37.05 | prpplague | i've had several flash devices fail due to erase cycles |
| 22:37.22 | MonMotha | I've had precisely zero flash devices fail due to erase cycles |
| 22:38.20 | adyer | I've had a few intentionally fail under erase testing |
| 22:38.23 | prpplague | MonMotha: hehe, must not be pushing them hard enough, hehe |
| 22:38.32 | MonMotha | prpplague: I pretty much just use them |
| 22:38.49 | MonMotha | I ignore the erase cycle limitations for everything other than machine generated erases |
| 22:38.59 | prpplague | MonMotha: yea, most of the failures were on devices we were pounding with reflashes |
| 22:39.06 | MonMotha | I figure if I'm manually triggering the erase, there's pretty much no chance of me breaking it |
| 22:39.20 | adyer | the failure is really the time to erase > some preset limit in the chip. erase takes longer the more you erase the part |
| 22:39.40 | MonMotha | I'm waiting for MRAM to become viable |
| 22:39.53 | MonMotha | it's out and it works, but it's too expensive |
| 22:40.04 | MonMotha | volume is low, so the manufacturers are keeping it on older processes |
| 22:41.21 | prpplague | RobotGuy: my usb gadget drivers working perfectly fine |
| 22:41.26 | RobotGuy | prpplague: I am confused about s3c2410_udc |
| 22:41.42 | prpplague | RobotGuy: you have it enabled |
| 22:41.48 | prpplague | its in your kernel config |
| 22:41.53 | prpplague | and is showing up in your dmesg |
| 22:42.05 | RobotGuy | prpplague: Also where I need to put the makeinfo patch for the native toolchain. |
| 22:42.14 | RobotGuy | prpplague: OK, if it's enabled, what is wrong? |
| 22:42.36 | prpplague | RobotGuy: just like you put it in the 2.17.50.0.17 dir of binutils, put it also in the 2.17 dir |
| 22:42.54 | prpplague | RobotGuy: you have so many things turned on, it could be just about anything |
| 22:43.07 | prpplague | RobotGuy: my suggestion is you start with the hammer-config |
| 22:43.18 | prpplague | RobotGuy: and just change the gadget driver to the ethernet |
| 22:43.22 | prpplague | RobotGuy: then retest |
| 22:44.04 | prpplague | RobotGuy: just to make sure....... |
| 22:44.26 | prpplague | RobotGuy: you are connecting a cable from the usb slave port on the carrier board to the host port on your pc? right? |
| 22:44.44 | RobotGuy | prpplague: Yes. |
| 22:44.51 | prpplague | just checking |
| 22:45.12 | prpplague | RobotGuy: you wanna try my kernel? |
| 22:46.01 | Mirell | Don't show off your kernel in public. |
| 22:46.06 | Mirell | You might catch something. |
| 22:46.12 | prpplague | hehe\ |
| 22:46.57 | ds2 | now repeat the claims at the extremes of the temperature listed on the datasheet ;) |
| 22:47.32 | MonMotha | ds2: my experience is that there's often quite a factor built in even at the extremes! |
| 22:47.41 | RobotGuy | prpplague: Building new kernel as suggested. |
| 22:48.02 | ds2 | MonMotha: I've manage to mess up a serial flash device by writing to it too much |
| 22:48.11 | ds2 | and it is at root temp |
| 22:49.07 | RobotGuy | prpplague: I use /dev/ttySAC1 for the second serial port, right? |
| 22:49.46 | prpplague | one sec phone |
| 22:50.30 | MonMotha | ds2: well, my tales of little experience apply mostly to manually triggered erases, like "reflashing" the firmware on a devie |
| 22:50.42 | MonMotha | I can certainly see erases initiated via automatic means hiting the limit |
| 22:51.04 | prpplague | RobotGuy: yea |
| 22:51.17 | RobotGuy | prpplague: Is it inited at 115200? |
| 22:51.45 | prpplague | RobotGuy: yea, why? |
| 22:52.13 | prpplague | RobotGuy: ok, looks like something is reseting the gpio that controls the usb slave port |
| 22:52.19 | RobotGuy | prpplague: Because I need to use it. I am going to either hook my bluetooth module or my SSC-32 to that port. |
| 22:52.45 | prpplague | RobotGuy: hehe, you make me laugh, you are all over the place on dev work |
| 22:53.04 | prpplague | RobotGuy: still got devmem on the hammer? |
| 22:53.09 | RobotGuy | prpplague: Yes |
| 22:53.24 | prpplague | RobotGuy: is it booted? |
| 22:53.37 | RobotGuy | prpplague: I don't have any trouble keeping track of multiple threads of dev. :) |
| 22:53.38 | prpplague | ds2: dart board? |
| 22:53.51 | prpplague | RobotGuy: i find it counter productive |
| 22:54.04 | ds2 | prpplague: yeah.. complete with velcro tags of current problems...it'll be used to determine the current problem to be looked at |
| 22:54.04 | prpplague | RobotGuy: is the hammer still booted up? |
| 22:54.12 | prpplague | ds2: hehe |
| 22:54.14 | RobotGuy | prpplague: Hammer is booted and I have console to it. |
| 22:54.20 | ds2 | it'll just as effective as having management around ;) |
| 22:54.30 | prpplague | ds2: no problems with the hammer per se, its just doing the dev people want |
| 22:54.37 | prpplague | RobotGuy: do ./devmem 0x56000020 |
| 22:55.11 | RobotGuy | Memory mapped at address 0x40006000. |
| 22:55.12 | RobotGuy | Value at address 0x56000020 (0x40006020): 0x0 |
| 22:55.22 | RobotGuy | prpplague: Note, I have not reflashed with the new kernel yet. |
| 22:56.03 | prpplague | RobotGuy: ok do this: ./devmem 0x56000020 w 0x40000000 |
| 22:56.53 | RobotGuy | prpplague: 3 lines. pastebin? |
| 22:57.00 | prpplague | RobotGuy: here is fine |
| 22:57.03 | RobotGuy | Memory mapped at address 0x40006000. |
| 22:57.03 | RobotGuy | Value at address 0x56000020 (0x40006020): 0x0 |
| 22:57.03 | RobotGuy | Written 0x40000000; readback 0x40000000 |
| 22:57.16 | ds2 | back then, people did that same thing with 'peek' and 'poke' |
| 22:57.23 | prpplague | RobotGuy: ok do this: ./devmem 0x56000024 |
| 22:57.36 | prpplague | ds2: hehe yea, pretty much the same |
| 22:57.53 | RobotGuy | Memory mapped at address 0x40006000. |
| 22:57.53 | RobotGuy | Value at address 0x56000024 (0x40006024): 0xFFFF |
| 22:58.30 | prpplague | RobotGuy: ok do this: ./devmem 0x56000024 w 0x00007fff |
| 22:59.41 | prpplague | RobotGuy: if your cable is plugged in, you should be able to do a ifconfig -a and see a usb0 |
| 22:59.48 | RobotGuy | prpplague: http://www.pastebin.ca/867858 |
| 23:00.11 | prpplague | RobotGuy: yep, thats a good thing |
| 23:00.25 | prpplague | RobotGuy: one of the drivers you have enabled is overwritting the bit config for that |
| 23:00.35 | prpplague | RobotGuy: is suspect a sloppy bitmask somewhere |
| 23:00.38 | RobotGuy | prpplague: I got some activity on the host also. |
| 23:00.52 | prpplague | RobotGuy: yea, you should be able to do ifconfig usb0 now |
| 23:00.59 | RobotGuy | prpplague: On host? |
| 23:01.04 | prpplague | RobotGuy: on both |
| 23:01.12 | prpplague | RobotGuy: i.e. ifconfig usb0 10.1.1.1 |
| 23:01.16 | prpplague | RobotGuy: then do |
| 23:01.18 | prpplague | RobotGuy: i.e. ifconfig usb0 10.1.1.2 |
| 23:01.23 | prpplague | on the host |
| 23:01.31 | RobotGuy | prpplague: On the host: |
| 23:01.34 | RobotGuy | usb0:avah Link encap:Ethernet HWaddr C6:24:57:2C:FE:F8 |
| 23:01.34 | RobotGuy | <PROTECTED> |
| 23:01.34 | RobotGuy | <PROTECTED> |
| 23:02.03 | prpplague | RobotGuy: you probably got some script running there to assing ips |
| 23:02.14 | prpplague | RobotGuy: there an ip on the hammer? |
| 23:02.16 | RobotGuy | prpplague: It's in inittab |
| 23:02.26 | RobotGuy | prpplague: Yes, have IP on Hammer |
| 23:02.34 | prpplague | should be working then |
| 23:07.05 | *** join/#edev CosmicPe1guin (n=nobody@163.181.251.103) |
| 23:07.20 | RobotGuy | prpplague: Still no ping on host. |
| 23:07.35 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 23:07.50 | prpplague | RobotGuy: are you able to ping from the hammer? |
| 23:10.05 | RobotGuy | prpplague: No |
| 23:11.23 | prpplague | RobotGuy: try the basics |
| 23:11.33 | prpplague | RobotGuy: on the hammer do this: ifconfig usb0 10.1.1.1 |
| 23:11.47 | prpplague | RobotGuy: on the host do this: ifconfig usb0 10.1.1.2 |
| 23:11.55 | prpplague | RobotGuy: then give it a try |
| 23:12.03 | prpplague | RobotGuy: from teh hammer do: ping 10.1.1.2 |
| 23:13.25 | RobotGuy | prpplague: I have ping |
| 23:13.38 | prpplague | RobotGuy: ok that means your scripts are @#$@#$ |
| 23:16.02 | RobotGuy | prpplague: OK. The problem must be with my routes. |
| 23:16.10 | prpplague | yea |
| 23:16.42 | RobotGuy | prpplague: Do I need a host route from hammer to host and the reverse? |
| 23:17.24 | prpplague | RobotGuy: yea, that helps |
| 23:24.19 | RobotGuy | prpplague: I can't seem to get it to work using my static ips. |
| 23:24.53 | prpplague | RobotGuy: thats gonna be routing and setup for you |
| 23:25.21 | prpplague | RobotGuy: you'll have to make sure you don't have some firewalls running that are blocking routing |
| 23:25.52 | prpplague | RobotGuy: probably best if you setup ip masquerading on your pc |
| 23:26.14 | prpplague | RobotGuy: check out the handhelds.org web site for details info on getting the routing correct |
| 23:26.34 | ds2 | <PROTECTED> |
| 23:27.30 | prpplague | RobotGuy: i've added a patch for the udc |
| 23:27.32 | prpplague | RobotGuy: http://www.elinux.org/Hammer_Board_Software_System |