| 00:03:30 | andersee | BZFlag: vodz has a telnetd for busybox at ftp://ftp.simtreas.ru/pub/my/bb |
| 00:04:40 | andersee | BZFlag: specifically in ftp://ftp.simtreas.ru/pub/my/bb/new/ports |
| 00:05:26 | TomW | andersee: after banging make for a while, it appears that you have tamed the pcmcia stuff, again. |
| 00:05:35 | TomW | andersee: works fine now. |
| 00:06:12 | jacques | hi RedBeard |
| 00:06:40 | jacques | BZFlag: WOW i was just thinking about that this morning |
| 00:08:17 | jacques | i have my tuxes on the network now but no way to login in remotely |
| 00:08:20 | TomW | andersee: I have the ftp, telnetd & inetd from uClinux working on the Tux. |
| 00:08:55 | jacques | hmm that .ru site is extremely slow coming up |
| 00:09:00 | jacques | as in not |
| 00:09:23 | jacques | TomW: was it hard to get telnetd working? |
| 00:10:40 | TomW | no, the uClinux sources are already stripped down. You have to remove the vfork() and replace with fork() in the source, then edit the Makefile(s) to get them to behave under Erik's staging dir. |
| 00:10:54 | TomW | jacques: you want them? |
| 00:10:59 | RedBeard | doesn't busybox offer a telnetd? or is it just telnet client? |
| 00:11:08 | jacques | TomW: that would be great |
| 00:11:16 | TomW | telnet AFAIK |
| 00:11:24 | RedBeard | *nod* |
| 00:12:07 | BZFlag | andersee: what's the long term plan on a inetd/telnetd ? |
| 00:12:09 | andersee | TomW: you can use them stock -- vfork works just fine on mmu-full or mmu-less |
| 00:12:21 | andersee | BZFlag: dunno yet |
| 00:12:30 | BZFlag | well, make one up. ;-) |
| 00:12:48 | BZFlag | TomW: can you send in patches for buildroot for the ones you have working? |
| 00:13:42 | TomW | Um, yes. I will have to do some stuff first, get a fresh buildroot image. wait a bit. |
| 00:14:37 | TomW | I had the stuff working in the previous 2.4.16 version of buildroot, am now flashing the newer 2.4.18 with these apps to test them. |
| 00:14:55 | andersee | BZFlag: biggest concern I have with daemons is that busybox has not been carefully audited... |
| 00:15:14 | andersee | BZFlag: So if these is a buffer overflow in telnetd, for example.... |
| 00:15:22 | BZFlag | I hear ya. |
| 00:16:17 | BZFlag | I'd like as small a remote shell as possible that can handle some form of passwdless connect for things that tar over a pipe. |
| 00:16:38 | BZFlag | I do not need full terminal type handling, but enough for bb vi would be nice. |
| 00:17:30 | BZFlag | some folks want ftpd, I do not. that's what pipes and tar/rsync are for. |
| 00:17:35 | BZFlag | ;-) |
| 00:17:51 | andersee | BZFlag: I have a _very_ old password-less rsh-like app I can send you... |
| 00:18:17 | andersee | BZFlag: It would need a little work to update it, but not much |
| 00:20:28 | BZFlag | no I want it password full, but .rhosts or some such handling. |
| 00:20:36 | TomW | Ok, inetd & telnetd are working. BTW you have to edit sources/device_table.txt and increase the number of ptyp(s) to a reasonable number. They are set at only 1 device, I increased it to four. |
| 00:21:07 | BZFlag | TomW: add that to the patch. 16 or so might be a good number. |
| 00:21:38 | andersee | TomW: the apps need to be taught to use /dev/pts |
| 00:22:03 | TomW | ok. As there is no password for root (and I suspect that securetty does not list the ptyp0) you have to adduser name and login as that username, then su. |
| 00:22:05 | BZFlag | heh, yeah that would be better. |
| 00:22:38 | jacques | TomW: I have same pty problem with rxvt |
| 00:22:47 | BZFlag | TomW: we should probably add a shannon/shannon to the build then. |
| 00:22:49 | jacques | 16 is a good number |
| 00:24:22 | BZFlag | or perhaps a more generix default/default user/password? |
| 00:25:37 | BZFlag | hmm... the non-root user can be passwd-less, yes? |
| 00:25:59 | CosmicPenguin | is away: Uh, oh.. the weasel got out again... |
| 00:26:51 | BZFlag | adds default user to cvs |
| 00:27:23 | jacques | maybe i can teach rxvt to use /dev/pts |
| 00:33:55 | BZFlag | jacques: that would be nice. |
| 00:34:21 | BZFlag | what happened to tiny x patches for buildroot? |
| 00:39:22 | TomW | andersee: Erik. buildroot-tux/sources/cvspass is broken for infradead.org. To fix it, take out the port number (2401) then buildroot will login into his cvs okay. |
| 00:40:06 | andersee | TomW: What we really need to do is eliminate cvs. We need to create our own working tarball, and post it somewhere... |
| 00:41:12 | TomW | yes. Also, I have been using his 20020302 snapshot and it seems to work okay. buildroot calls for 20020102, may be a bit dated. |
| 00:42:16 | BZFlag | TomW: port removed. |
| 00:42:51 | TomW | wish that I could offer the bandwidth, but my ISP busted me for running an 800K cable modem on a 150K account (their fault). |
| 00:43:20 | BZFlag | plenty of space on sf, we can just upload it there. |
| 00:44:03 | andersee | BZFlag: make it so... :-) |
| 00:44:09 | BZFlag | TomW: got a sf account? I can add you |
| 00:46:42 | BZFlag | andersee: I just added you and russ as admins on the project. |
| 00:50:37 | andersee | cool, thx |
| 00:51:44 | BZFlag | later. andersee add TomW etc as you like. cya |
| 00:52:23 | kergoth | andersee: any quick ideas offhand on what would cause an arm build of busybox sh to hang on execution? |
| 00:52:34 | scanline | is back (gone 01:36:57) |
| 00:53:21 | andersee | How was sh invoked? |
| 00:53:35 | andersee | Latest busybox.stable? |
| 00:53:45 | kergoth | andersee: either by login (tinylogin) or init=/bin/sh .. busybox 0.60.2 |
| 00:54:03 | kergoth | andersee: built against a build of glibc, not uclibc |
| 00:56:01 | andersee | dunno |
| 00:56:08 | andersee | Which shell? |
| 00:56:16 | andersee | lash? msh? ash? |
| 00:56:19 | kergoth | ash |
| 00:56:38 | kergoth | just thought id see if you had any suggestions. i must be missing something |
| 00:57:06 | andersee | I have an updated version of ash you can try out |
| 00:57:17 | andersee | I've been using it here with good results... |
| 00:57:24 | kergoth | sure, why not |
| 00:57:46 | kergoth | ive got no better ideas. the behavior is odd |
| 00:58:44 | TomW | patches for inetd, telnetd & ftp (client) were just posted to the TuxScreen list. |
| 00:59:00 | TomW | hope I got everything... |
| 01:30:27 | TomW | Well, looks like the patches for indetd & telnetd will be held up for a while. The file came out to 100K zipped and is being held for moderator approval. If anyone wants that patch, I can dcc / email it to you. |
| 01:58:22 | jacques | heh i forgot there was a tuxscreen list |
| 01:58:47 | RedBeard | I'd like to subscribe to it, is it the sourceforge list? |
| 01:59:49 | TomW | <https://lists.sourceforge.net/lists/listinfo/tuxscreen-devel>, |
| 01:59:49 | TomW | <mailto:tuxscreen-devel-request@lists.sourceforge.net?subject=subscribe> |
| 01:59:49 | TomW | List-Id: |
| 01:59:49 | TomW | <tuxscreen-devel.lists.sourceforge.net> |
| 02:00:08 | RedBeard | yeah, I'm there, thanks |
| 02:00:30 | TomW | ok. not much activity there, most developers seem to be over here! |
| 02:00:54 | TomW | BZFlag: you want that inetd & telnetd patch? |
| 02:01:06 | BZFlag | you just mailed it, yes? |
| 02:01:22 | TomW | BZFlag: http://www.openhardware.net/TuxScreen/files/ |
| 02:01:42 | jacques | cool did you guys see the list post about RAM ? http://sourceforge.net/mailarchive/forum.php?thread_id=558853&forum_id=2321 |
| 02:02:34 | BZFlag | TomW: hmm.... |
| 02:02:53 | BZFlag | I was expecting a patch to the Makefile that gets the sources from someplace and builds them |
| 02:02:56 | BZFlag | not all the sources. |
| 02:04:32 | TomW | Sorry. I am not that far along with the Tux development stuff and my server will be disappearing soon. It will be moved to a low reliability, low speed ISDN line soon. :-( |
| 02:06:48 | TomW | jacques: yes, I saw that post on the RAM. It was to be expected that you could replace the lower density DRAM with higher density ones, rewire the control lines and reprogram the DRAM controller to accept them. A bit more work than I would want to do right now. |
| 02:06:53 | RedBeard | TomW: you coupd put patches to the source on my site for the time being, and submit a patch to the make file if you'd like |
| 02:07:33 | RedBeard | s/coupd/could |
| 02:07:50 | TomW | Ok, let me come up with a patch to the Makefile to wget the sources, then give Erik the patch so he can incorporate it into the buildroot in cvs? |
| 02:08:19 | RedBeard | okay |
| 02:10:00 | jacques | wow the blob in cvs also supports xmodem. |
| 02:10:05 | TomW | CMIIAW: cvs listens on port 2401 and has no ports < 1024 that it listens on? |
| 02:11:47 | TomW | if that is the case, I can put the stuff on my cvs here. Once I switch to a residential account, the cvs would still be reachable. My ISP blocks SYN packets sent to ports 0..1023 on the residential accounts. |
| 02:15:18 | jacques | BZFlag: I got the howto done. Just never went further integrating tinyX build into buildroot. |
| 02:49:05 | BZFlag | TomW: where did all the sources come from? |
| 02:49:18 | BZFlag | jacques: well? ;-) |
| 02:49:26 | TomW | BZFlag: uClinux |
| 02:49:47 | BZFlag | are there tarballs with just that? or is it part of the large tarball? |
| 02:50:18 | TomW | BZFlag: I had been using uClinux on some M68K stuff, so the sources were at hand. Also, the apps have been stripped to keep the size small. |
| 02:50:52 | TomW | Just the one tarball contains all files. The target_skeleton/etc/inetd.conf etc. |
| 02:52:08 | TomW | Only three files you might want to check on: Makefile, sources/device_table.txt and target_skeleton/linuxrc to see what changes I did. Most of the tarball is the build/UserApps sources. |
| 02:53:55 | BZFlag | perhaps I'm not being clear.... where is a tarball on the internet that contains telnetd/pathnames.h (for example) that you used? |
| 02:54:15 | BZFlag | I know your patch contains them... where did they come from? url? |
| 02:56:06 | BZFlag | ie: telnetd contains: Copyright (c) 1989 Regents of the University of California |
| 02:56:15 | BZFlag | so surely it came from someplace else? |
| 03:01:37 | BZFlag | hmm.. looks like that's the same telnetd that is in debian sources... /me will diff em |
| 03:03:41 | BZFlag | ftp://ftp.uk.linux.org/pub/linux/Networking/netkit/ looks like the original source. |
| 03:28:11 | jacques | hi Tangent |
| 03:28:30 | Tangent | jacques : Hey |
| 03:29:10 | Tangent | jacques : Actually.. I was just off to bed... |
| 03:29:15 | jacques | Tangent: heh |
| 03:29:22 | Tangent | just finished long multiplayer game session... |
| 03:29:39 | jacques | mmmm |
| 03:29:41 | | i guess game is full? |
| 03:29:41 | jacques | what game? |
| 03:29:41 | Tangent | 3:30AM here now.... got places to go tommorow... |
| 03:29:53 | | jacques: huh? |
| 03:29:53 | jacques | ibot: you ain't got game |
| 03:30:03 | Tangent | jacques : Total Anillihation.... It's about 5 years old |
| 03:30:20 | jacques | Tangent: I wanna get jedi II |
| 03:30:31 | jacques | i actually have TA but have never installed it |
| 03:30:43 | Tangent | jacques : It's an excellent game... well worth a spin |
| 03:30:50 | Tangent | Anyhow... I must go and sleep now.. |
| 03:30:53 | Tangent | catch you soon |
| 03:30:54 | TomW | BZFlag: sorry, bbl, have phone call. |
| 03:30:56 | jacques | later |
| 03:30:58 | TomW | is away: I'm busy |
| 03:39:09 | BZFlag | splits up the makefiles. |
| 03:53:26 | Kergoth | BZFlag: fyi the OZ buildroot will have staging files split from inst files shortly |
| 03:57:15 | BZFlag | Kergoth: nice. ooh capital nick today... |
| 03:58:14 | Kergoth | BZFlag: hehe. gotta mix it up once in a while.. you know.. ;-) |
| 04:44:15 | BZFlag | argh... shared libraries are a Good Thing people. do not link them statically into your code. ;-) |
| 04:44:58 | jacques | ? |
| 04:50:50 | BZFlag | picogui and microwindows both include libjpeg statically |
| 04:51:46 | jacques | heh |
| 04:53:51 | TomW | is back (gone 01:22:52) |
| 04:54:58 | BZFlag | jpeg is now shared, though I'm not sure pgui and mwin will use em that way |
| 04:55:00 | TomW | BZFlag: The uClinux is one large tarball distro. It is an "all in one" type of distro, Kernel + Apps + vendor configs. |
| 04:55:17 | BZFlag | TomW: yeah, shame on them. |
| 04:55:38 | BZFlag | I found the original source, and will reference that with diffs. |
| 04:55:53 | BZFlag | thanx for the working version to refer to. ;-) |
| 04:56:23 | TomW | BZFlag: Well, they are trying to promote it to windoze type of developers. You know Gerg don't you? Lineo: Morton Bay. Greg is a prime issuer of the distro. |
| 04:56:38 | TomW | s/Gerg/Gerg/ |
| 04:57:01 | BZFlag | yep. tuxscreen buildroot does that as well, and you can include sources/dl/* if you want. |
| 04:57:36 | BZFlag | but you don't need to. and we certianly don't need yet another copy in yet another cvs tree. |
| 04:58:13 | BZFlag | make the world on buildroot, then nuke build/ and you have a tarball ready tree. |
| 04:58:43 | TomW | BZFlag: I'll make it an option similar to the microwindows. If you want it pulled in, set the UserApps var to "yes" |
| 04:58:48 | BZFlag | I plan to add a DL target at somepoint so that you can type "make download" and just get the sources/dl populated. |
| 04:59:15 | BZFlag | TomW: I'm not planning on adding the sources to cvs. just a makefile. |
| 04:59:16 | RedBeard | BZFlag: I like that idea, it would allow for tweeking before compiling |
| 04:59:31 | BZFlag | RedBeard: agreed. |
| 04:59:49 | BZFlag | TomW: do you want to convert it to dl and patch the original sources? |
| 04:59:58 | TomW | I don't know how useful it would be. It seems to me that a lot of the developers here are pretty experinced in integrating apps into the buildroot. |
| 05:00:05 | BZFlag | ftp://ftp.uk.linux.org/pub/linux/Networking/netkit/netkit-tftp-0.17.tar.gz |
| 05:00:12 | BZFlag | oops... |
| 05:00:24 | BZFlag | ftp://ftp.uk.linux.org/pub/linux/Networking/netkit/netkit-telnet-0.17.tar.gz |
| 05:00:54 | BZFlag | TomW: the idea is to expand the user base to those that are not. |
| 05:01:00 | RedBeard | I'm learning how to integrate the apps |
| 05:01:18 | BZFlag | RedBeard: do a cvs up -PdA and check out make/*.mk |
| 05:01:26 | BZFlag | see if that makes it easier. ;-) |
| 05:01:57 | TomW | Tnx for the url, I can always use source. :-D |
| 05:03:34 | TomW | BZFlag: speaking of cleaning up the source tree: that microwindows snapshot stuff is a real pain |
| 05:03:35 | TomW | ! |
| 05:03:56 | TomW | We need something cleaner, ready to run "out of the box". |
| 05:05:30 | TomW | I puzzled over the config file for a while, finally got that working. Then had to debug the makefile where the linker was failing to find symbols in a later lib in the lib chain, had to re-arrange the order of libs. Then had to remove nxscribble to keep that from bombing the compile. |
| 05:06:03 | TomW | Why don't I put my own snapshot up for that? |
| 05:06:11 | TomW | A fixed one... |
| 05:22:49 | BZFlag | patches to the make and sources would be welcome. |
| 05:23:26 | BZFlag | I think it was SpaceCoaster that added it? Whoever owns it should add an email address to the microwindows.mk now. |
| 05:23:59 | BZFlag | feel fee to put up a snapshot, but we need to put up sources too. |
| 05:24:05 | BZFlag | s/fee/free/ |
| 06:07:38 | RedBeard | BZFlag: I'll do it tomorrow, I'm too sleepy right now (I know, not enough caffeine) ;) |
| 06:07:41 | Kergoth | argh. stupid tinylogin build doesnt like mem |
| 06:09:17 | RedBeard | g'nite |
| 06:16:42 | jacques | how fast is my tux running in MHz? |
| 06:16:46 | jacques | how can I find out? |
| 06:17:10 | jacques | is 127.38 bogomips right for 192MHz ? seems low |
| 06:30:14 | Russ | jacques: do you know what bogomips stands for? |
| 06:30:44 | TomW | bogus mips |
| 06:31:36 | Russ | not quite |
| 06:31:49 | Russ | bogus, meaningless, interpretation of processor speed |
| 06:32:49 | TomW | a bogomip has no real direct correlation to a Dhrystone mip, it is close. It is meant to give a "feeling" for how fast linux is running. |
| 06:33:15 | Russ | TomW: no |
| 06:33:20 | Russ | its meant for delay loops |
| 06:33:41 | Russ | it isn't meant to have any meaning to the user, and kernel revisions change it regularly |
| 06:34:17 | TomW | right, same thing. How many loops do I need to insert here? For that you need some relative measure of the CPU performance. |
| 06:34:37 | TomW | hence: a bogomip (bogus mip). |
| 06:44:38 | jacques | all |
| 06:44:39 | jacques | i |
| 06:44:40 | jacques | want |
| 06:44:41 | jacques | to |
| 06:44:42 | jacques | know |
| 06:44:43 | jacques | is |
| 06:44:47 | jacques | how |
| 06:44:48 | jacques | fast |
| 06:44:49 | jacques | in |
| 06:44:51 | jacques | mhz |
| 06:44:52 | jacques | the |
| 06:44:53 | jacques | proc |
| 06:44:54 | jacques | in |
| 06:44:55 | jacques | my |
| 06:44:57 | jacques | tux |
| 06:44:59 | jacques | is |
| 06:45:01 | jacques | running |
| 06:46:55 | jacques | since we have no /proc/scale or /proc/sys/cpu the only thing left to me is the bogomips value |
| 06:47:18 | jacques | i am not trying to compare performance between different archs |
| 06:47:37 | jacques | i am not even trying to compare drastically different kernel versions |
| 06:48:43 | jacques | in my statement above where I said 127.38 seemed low for 192MHz, I was comparing to what I thought I remembered I was getting on my ipaq |
| 06:49:52 | jacques | but if I knew a different way of getting the proc clock speed on the tux I would never have mentioned bogomips at all. |
| 07:01:43 | BZFlag | #elif defined SHANNON |
| 07:01:43 | BZFlag | cpuspeed: .long 0x09 /* 191.7 MHz */ |
| 07:03:36 | jacques | BZFlag: thanks. do you know of anybody using /proc/scale on the tux? or that is does not work? |
| 07:04:04 | TomW | jacques: From experience, a bogomip is a fair value of the performance. A significant factor in "slowing" the bogos down is background activity such as the LCD controller. My M68EZ328 systems are rated for 2.1 mips, the bogos get up to approx 1.9 mips, when I turn off the LCD controller. With it on, the bogos are only 1.48. |
| 07:04:18 | BZFlag | needs some kernel work, should not be too tough. lcd timing needs to change as cpu timing changes. |
| 07:10:26 | jacques | aah, fascinating |
| 07:10:53 | jacques | i wonder if they dealt with that on the ipaq yet - i know different lcd/display |
| 07:11:33 | jacques | what about DRAM refresh timing? |
| 07:12:06 | TomW | The slowdown would vary according to the amount of data that the LCD DMA has to "push" onto the display. We are using a full vga display, the ipaq is, what, QVGA? |
| 07:12:19 | jacques | I had my ipaq u pto 236MHz for a while :-) it did improve compile performance |
| 07:12:58 | jacques | TomW: that's correct - 320x240 on ipaq - but at 16bit instead of 8bit as on tux |
| 07:13:16 | jacques | so i guess ipaq is half amount of data as tux |
| 07:13:47 | TomW | Refresh timings are not much of an issue, not that you have much control over them anyway. Most modern processors "weave" the refresh into the machine cycles so as to cause minimal disturbance to the CPU. This is one reason for a larger L1 cache. |
| 07:15:20 | TomW | DRAM wait states can cause a slight problem: from 5% .. 15% degradation in system performance. 0 waits are nice, but expensive, usually 1 wait state ram is cost effective. |
| 07:16:11 | TomW | check the wait state settings on the blob, they may be using conservative settings (one size fits all). |
| 07:16:40 | jacques | fascinating info. this is why i love these channels. :-) |
| 07:17:28 | jacques | TomW: would I have to look at the blob source or are those wait state settings changeable from boot prompt? |
| 07:17:55 | TomW | Try hardware design sometime, it can stretch your mind as much as a software system can. All those timing propogations you have to add up, what will heat do the timings, etc. |
| 07:18:53 | jacques | TomW: I figures out a long time ago that I'm not smart enough for hardware design :-) |
| 07:19:09 | jacques | i have a really good friend who does it |
| 07:19:26 | TomW | Depends on the processor. Some processors lock vital registers after they have been set once. Those type of processors need a reset pulse to place them back into "receptive" mode. They are not common, but I would not say "yes" to that question without researching it first. |
| 07:21:15 | TomW | one way to try: ioremap the registers so you can get at them, then start messing with them, if you crash the system, you will know that you definately can change the DRAM timings. :-D |
| 07:21:56 | jacques | heh |
| 07:22:34 | TomW | be careful of the clock settings though, it could be possible to overclock it to the point that you start generating a lot of internal heat. |
| 07:22:54 | jacques | TomW: I remember ppl talking about the cpuscale code in the arm kernel changing those timings as the cpu clock was changed so maybe they are modifiable |
| 07:23:09 | TomW | especially if you crash the system, then go for coffee and forget to shut it off. |
| 07:23:13 | jacques | TomW: have you ever messed with /proc/scale ? you can cat a value into it to change clock speed |
| 07:23:32 | TomW | no. I am new to ARM. |
| 07:23:36 | jacques | it's standard on the ipaq kernels these days |
| 07:24:34 | TomW | sounds odd to have it, the system should be using the idle() loop of the kernel to modify the system clockrate. |
| 07:24:59 | jacques | 9 was 192, 10 was 206, 11 was i forget and 12 was 236 |
| 07:25:18 | jacques | i usually set mine on 11 :-) |
| 07:25:22 | TomW | 12 sounds nice.... |
| 07:25:34 | jacques | 12 was good, some ipaqs would run at it some would not |
| 07:25:46 | jacques | mine would run indefinitely at it AFAICT |
| 07:25:58 | TomW | Well, I do have a dongle from Tim, but not anxious to try it out yet. |
| 07:25:59 | jacques | I definitely saw compile time speedups |
| 07:27:01 | jacques | heh |
| 07:27:16 | TomW | heat & lockups would be your only real issue of concern. Usually, you can get specs on the thermal resistance of the chip, measure the surface temp, then calculate the internal temp of the chip. |
| 07:27:39 | jacques | but what I really want in my tuxes is more RAM |
| 07:27:47 | TomW | yes. |
| 07:27:54 | jacques | 32 would be nice, 64 would be heavenly |
| 07:28:31 | TomW | so would 128 / 256. Doesn't the StrongARM support 4 banks of 64 meg ram? |
| 07:28:52 | jacques | well I know that ppl have made 128MB ipaqs |
| 07:29:01 | jacques | there is a company that will do the conversions |
| 07:29:07 | jacques | since it requires SMT |
| 07:29:44 | jacques | I'm waiting for the 412MHz XScale 128MB ipaqs |
| 07:30:10 | jacques | maybe the company would convert tuxes too, but the cost would probably be prohibitive |
| 07:30:41 | TomW | Depends on the footprint of the RAM chips have on the PCB, you need something higher in density that: has the same number of pins, pinouts really close to the originals, and organized in a way that the SA-1100 can handle (12x10, 14x10, 12x12, etc.). |
| 07:31:28 | jacques | well for the ipaq I know the parts for 128MB became available a few months ago, and AFAIK are pin compatible with what was on there before |
| 07:31:36 | TomW | SMT is not a big deal, it takes patience to work with it. I know of people that just use a 40W Weller iron, solder wick and rosin to mount / unmount chips. |
| 07:31:58 | jacques | it seems like all that was in volved was disassembly, remove chip, solder new chip, reassemble |
| 07:32:27 | jacques | TomW: heh. I have neither the eyes nor the patience for that :-) |
| 07:32:29 | TomW | I have a Zephyrtronics Hot Air system that I use for working with SMT. Cost me about $2k for the equipment. |
| 07:32:50 | jacques | i saw a hot air smt system at fry's for like $500 ? |
| 07:32:51 | TomW | Also have a 20X binocular microscope. :-D |
| 07:33:01 | jacques | yeah i would need that |
| 07:33:38 | jacques | my hands are steady but my near vision has been killed by years of computer use |
| 07:35:07 | TomW | http://zephyrtronics.com/home.html |
| 07:35:35 | TomW | use mozilla, Netscape 4.77 locks up on that site. |
| 07:35:47 | jacques | yeah galeon works fine on it |
| 07:36:05 | jacques | wow I would love to work around ppl who even use equipment like that |
| 07:36:06 | jacques | :-D |
| 07:36:18 | TomW | I like their desoldering solution: low temperature alloy that dilutes the solder. |
| 07:37:21 | TomW | I sold mytoys.com in 2001 for $50K, so I invested a lot into equipment, software, office equipmnet. Also payed myself for a year of exploration into embedding linux. |
| 07:37:58 | jacques | Nice. |
| 07:38:23 | TomW | It was to a German outfit, before the dot-com crash. He just showed up one day in email, asked if I would sell.. |
| 07:38:23 | jacques | i want to get into embedded/pervasive linux |
| 07:38:46 | jacques | that's cool |
| 07:39:10 | TomW | I would say that the work on the Tux would be good experience. |
| 07:39:48 | jacques | yeah I spent a year or so on the handhelds.org ipaq project before i got the tuxes |
| 07:40:07 | TomW | I did a port of uClinux over to two M68K systems I built. It was good experience to get inside the kernel and hack around to make my hardware work. |
| 07:40:10 | jacques | i think ARM is going to get bigger |
| 07:40:26 | jacques | heh that was bad timing |
| 07:40:40 | jacques | not in response to what you said about M68K |
| 07:40:51 | jacques | just I think ARM is going to get bigger than it is |
| 07:40:58 | jacques | not than M68K |
| 07:41:33 | TomW | Don't know. ARM has a really nice nitch right now. Powerfull, easy on batteries, and there are a lot of processors with nice subsystems (VGA, LCD, Ethernet, SDRAM controllers, etc). |
| 07:41:51 | TomW | M68k is going no where. |
| 07:42:10 | jacques | with all the wince machines ARM now, and Palm switching to ARM... |
| 07:42:11 | TomW | s/nitch/niche/ ?? |
| 07:42:18 | jacques | niche is correct |
| 07:42:46 | TomW | I cannot spell, so that is why I write software and not documentation. :-) |
| 07:42:59 | jacques | French, from Old French, from nichier, to nest, from Vulgar Latin * nidicare, from Latin nidus, nest, See sed-. or from Old Italian nicchio, seashell, perhaps from Latin mitulus, mussel. |
| 07:43:26 | TomW | Mmmm... |
| 07:43:51 | TomW | mumbles something about online lookups. |
| 07:44:06 | jacques | heh my company makes IM bots |
| 07:44:10 | jacques | and I just asked one of them |
| 07:44:29 | jacques | but I knew the spelling before I looked it up |
| 07:45:49 | TomW | What version kernel are you running now in the Tux? |
| 07:45:52 | jacques | i also knew it was a french borrowing, but not the latin root |
| 07:46:01 | jacques | 2.4.18-rmk2-tux1 |
| 07:46:36 | TomW | Are you seeing a problem with pcmcia not properly processing card insertions? |
| 07:46:39 | jacques | that's what current (as of yesterday) buildroot-tux created |
| 07:47:18 | jacques | I've only tried my netgear MA411 10/100 network cards, but they seem to be working as normal |
| 07:47:27 | jacques | what symptoms are you seeing? |
| 07:47:35 | TomW | my kernel / software sees the device, but doesn't trigger the /etc/pcmcia/network file. I have to make a manual entry in linuxrc for "ifconfig eth0 192.168.1.201" |
| 07:48:01 | jacques | are you inserting the card after the system is already booted and expecting it to get configured? |
| 07:48:07 | TomW | Once in a *great* while, I can cold boot the system and it will configure the card okay. |
| 07:48:08 | jacques | it's never worked that way on tux |
| 07:48:21 | jacques | I think it's because the drivers are builtin to the kernel |
| 07:48:26 | jacques | instead of being modules |
| 07:48:51 | jacques | I can insert the card in a running tux, and cardctl status and cardctl ident work fine, but the network device is not there |
| 07:48:59 | jacques | ifconfig -a shows no eth0 |
| 07:49:00 | TomW | One system I have, I can insert the CF & ethernet card at anytime and it mostly configures it correctly. But that was a 2.4.16 kernel. |
| 07:49:19 | jacques | TomW: is that on a tux? |
| 07:49:31 | TomW | Try "ifconfig eth0 ipaddress", it will probably come up okay. |
| 07:49:46 | jacques | no, the device is simply not there |
| 07:49:57 | jacques | ifconfig -a should show even unconfigged devices |
| 07:50:11 | jacques | the driver gets initialized at load time |
| 07:50:23 | jacques | when the driver is builtin to the kernel, loadtime is boottine |
| 07:50:37 | TomW | Yes, on a tux. That board that I canabalized from my first unit: http://www.openhardware.net/TuxScreen/StrongARM-perf.html |
| 07:50:45 | jacques | we really ned the jffs2-aware blob and modular kernel |
| 07:51:00 | jacques | TomW: was the driver for the device a module or compiled into the kernel? |
| 07:51:23 | TomW | into the kernel. But the 2.4.16 was the same way, compiled into the kernel. |
| 07:51:31 | TomW | ... I think.. |
| 07:51:32 | jacques | hmmm then I am confused |
| 07:51:55 | jacques | because that's not my understanding of how it works |
| 07:52:16 | TomW | I had a /lib/modules/<kernel-version> directory under 2.4.16, but the kernel was configured to not have any modules. |
| 07:52:35 | jacques | do you still have that system running? |
| 07:52:54 | jacques | I would be interested in what lsmod says after the card is inserted and configured |
| 07:53:05 | TomW | So, you are saying that without a /lib/modules/<kernel-version> that cardctl won't work properly. |
| 07:53:19 | TomW | YES! I will not mess with that dir! It works! |
| 07:53:26 | jacques | my understanding is that the device has to e present when the driver is loaded |
| 07:53:50 | BZFlag | not so. |
| 07:54:08 | jacques | BZFlag: I was hoping someone would jump in |
| 07:54:23 | BZFlag | card drivers will be reused. this happens even with modules if you insert a second card that uses the same driver. |
| 07:54:40 | jacques | BZFlag: then why does cardmgr not configure my network cards when I insert them into a running system? |
| 07:54:46 | jacques | but they config fine on boot |
| 07:54:50 | BZFlag | don't know. works here. |
| 07:54:51 | TomW | Hmmm... you are right. if it is compiled into the kernel, then it would be probed for at boot time. The only way to do otherwise would to have the cardmanager stuff insert the module into the kernel. Then when the module inits, it will probe for the hardware. Not so? |
| 07:54:57 | BZFlag | check system log to see if it tried. |
| 07:55:01 | BZFlag | dmesg etc. |
| 07:55:10 | jacques | it does not try |
| 07:55:21 | jacques | but cardctl ident and status show expcted values |
| 07:55:36 | BZFlag | TomW: no. modprobe can fail cause the module is already there, the driver init will still be called. |
| 07:55:59 | jacques | hmm just nt for me |
| 07:56:08 | TomW | jacques: Right, now do "ifconfig eth0 YourCardIpAddress" and it will start working and show up in the ifconfig list. |
| 07:56:16 | BZFlag | jacques: is cardmgr running? |
| 07:56:28 | jacques | maybe it's a discrepancy with pcmcia userspace / kernel versions |
| 07:56:48 | jacques | d'oh |
| 07:56:56 | BZFlag | jacques: does the cardctl ident exactly match an entry in /etc/pcmcia/config ? |
| 07:56:56 | jacques | maybe that's what those errors were about on boot |
| 07:57:01 | BZFlag | heh |
| 07:57:06 | jacques | :-( |
| 07:57:53 | BZFlag | try commenting out all the include lines in /etc/pcmcia/config.opts |
| 07:58:00 | jacques | hmm I just manually started cardmgr |
| 07:58:01 | TomW | That is it! I ran cardmgr (cardmgr &) and it found my CF device. Then I could mount it to /mnt. |
| 07:58:12 | jacques | it complained a lot but started |
| 07:58:20 | jacques | i wonder why it was not running before |
| 07:58:26 | TomW | So, cardmgr is dying somehow, or was never loaded. |
| 07:58:53 | TomW | jacques: but, your ethernet started up? |
| 07:59:02 | BZFlag | it should run once in /linuxrc and then exit. then init should start it when it starts. |
| 07:59:13 | jacques | cardmgr is not running after boot on either of my tuxes |
| 07:59:19 | jacques | but starts manually on both |
| 07:59:28 | BZFlag | so your card might get started when the "cardmgr -o" runs from /linuxrc |
| 07:59:34 | TomW | Erik is starting cardmgr from busybox? |
| 07:59:46 | jacques | TomW: my ethernet only starts when I boot the tux with the card inserted |
| 07:59:48 | BZFlag | that happens so that pivot_root to someplace can start another version of pcmcia services if you want. |
| 08:00:00 | TomW | Ah, .. |
| 08:00:10 | jacques | what's the -o ? |
| 08:00:11 | BZFlag | TomW: no. I'm starting it from /linuxrc and then it starts again from.... |
| 08:00:23 | TomW | makes sense, we have another Erik condition. :-D |
| 08:00:29 | BZFlag | etc/init.d/S80pcmcia |
| 08:00:51 | jacques | I need to look into why it's not running after boot |
| 08:00:56 | BZFlag | jacques: man cardmgr - tell it to run once, ie: init all cards and exit. |
| 08:00:57 | TomW | Ok, will look into this now that I understand the problem. Thanks Tim. |
| 08:01:14 | BZFlag | np, I did it, so you can blame me. ;-) |
| 08:01:22 | TomW | heh |
| 08:01:54 | TomW | humility == willingness to take the blame for one's own actions. |
| 08:03:23 | jacques | ah so this is expected behavior? |
| 08:04:03 | jacques | hmm, why is S80pcmcia not starting it I wonder |
| 08:05:31 | BZFlag | try commenting out all the include lines in /etc/pcmcia/config.opts |
| 08:06:44 | jacques | ok just did that |
| 08:06:50 | jacques | now to reboot |
| 08:07:38 | jacques | now it's running after boot! |
| 08:07:43 | jacques | :-D |
| 08:08:09 | jacques | thanks BZFlag |
| 08:09:14 | TomW | same here, except ethernet card still needs the ifconfig eth0 <ip> before it inits the card. |
| 08:09:37 | jacques | I have my card config in the /linuxrc |
| 08:09:53 | jacques | so it would not autoconfig even now if I stick it in after boot |
| 08:10:04 | jacques | but there are files in /etc/pcmcia to configure that |
| 08:10:23 | jacques | but before this, the eth0 device was not even there so ifconfig would not work |
| 08:10:57 | jacques | I think you want /etc/pcmcia/network.opts |
| 08:11:21 | jacques | but then pcmcia configuration is not my forte |
| 08:12:02 | jacques | I gotta head home - will be on from there |
| 08:13:22 | BZFlag | setting in /etc/pcmcia/network.opts should work even at boot. just not for dhcp cause udhcpc would not know to exit after getting an address. |
| 08:14:24 | BZFlag | actually, I think hinds added a run once env var to 3.1.33 for me, so I ought to read up on it. |
| 08:16:58 | TomW | Well, using /etc/pcmcia/network.opts from a cold boot works fine. I cannot cold / warm boot then insert the card and have it recognized. At least it is something to have it come up on a cold boot, without the ifconfig entry inside linuxrc. I will dig deeper. |
| 08:22:26 | TomW | BZFlag: looks like the /etc/pcmcia/options.opts includes were the key. AFAIR, I did have erratic operation from this card. After testing, it works most of the time when inserted. Thanks for your help. |
| 08:22:52 | BZFlag | np. |
| 08:23:42 | fontenot | home now |
| 08:24:20 | TomW | fontenot: nothing happened while you were away... ;-) |
| 08:24:49 | fontenot | heh cool I hate to miss stuff :-) |
| 08:24:57 | BZFlag | made the change in buildroot |
| 08:25:16 | fontenot | :-) |
| 08:25:22 | TomW | lots of good info floating around here, should put your irc on "record". |
| 08:25:39 | fontenot | irc logs http://ipaq.secret.org.uk/~timr |
| 08:25:47 | fontenot | from the topic |
| 08:25:53 | | well, irc is "where the men are men, the women are men, the boys are FBI agents, but some of the are men, the women are men, the boys are FBI agents, but some of the men are actually women." |
| 08:25:53 | TomW | ibot: irc |
| 08:26:13 | TomW | LOL |
| 08:26:28 | | fontenot: excuse me? |
| 08:26:28 | fontenot | ibot: you kill me sometimes |
| 08:26:50 | | fontenot: excuse me? |
| 08:26:50 | fontenot | ibot: fontenot |
| 08:26:55 | TomW | poor ibot, victum of a bunch of screwy geeks! |
| 08:26:58 | | fontenot: bugger all, i dunno |
| 08:26:58 | fontenot | ibot: fontenot? |
| 08:27:04 | | jacques is not trying to compare performance between different archs |
| 08:27:04 | fontenot | ibot: jacques? |
| 08:27:11 | | fontenot: I forgot jacques |
| 08:27:11 | fontenot | ibot: forget jacques |
| 08:27:25 | | OK, TomW. |
| 08:27:25 | TomW | ibot: tom is asleep |
| 08:27:29 | fontenot | I am king of the world! |
| 08:27:30 | | i guess tom is asleep |
| 08:27:30 | TomW | ibot: tom |
| 08:27:38 | | fontenot: i haven't a clue |
| 08:27:38 | fontenot | ibot: fontenot? |
| 08:27:40 | fontenot | arg |
| 08:28:01 | | OK, TomW. |
| 08:28:01 | TomW | ibot: fontenot is the king of the World. God save the King! |
| 08:28:06 | | well, fontenot is the king of the World. God save the King! |
| 08:28:06 | TomW | ibot: fontenot |
| 08:28:13 | fontenot | whoo! |
| 08:28:19 | TomW | LOL |
| 08:28:24 | | Since Thu Mar 28 12:18:48 2002, there have been 55 modifications and 34 questions. I have been awake for 20 hours, 9 minutes, 36 seconds this session, and currently reference 4813 factoids. Addressing is in required mode. |
| 08:28:24 | fontenot | ibot: status |
| 08:28:55 | | TomW: excuse me? |
| 08:28:55 | TomW | ibot: microwin |
| 08:29:00 | | it has been said that microwindows is at http://microwindows.org/ |
| 08:29:00 | TomW | ibot: microwindows |
| 08:29:13 | | TomW: huh? |
| 08:29:13 | TomW | ibot: air |
| 08:29:46 | | OK, TomW. |
| 08:29:46 | TomW | ibot: it has been suggested that air is something that some people need. |
| 08:29:52 | | TomW: excuse me? |
| 08:29:52 | TomW | ibot: air |
| 08:29:57 | | TomW: huh? |
| 08:29:57 | TomW | ibot: it |
| 08:30:03 | | i haven't a clue, fontenot |
| 08:30:03 | fontenot | ibot: it? |
| 08:30:13 | TomW | wonder where that went? |
| 08:30:18 | | somebody said it has been suggested that air was something that some people need. |
| 08:30:18 | fontenot | ibot: it has been suggested that air |
| 08:30:24 | fontenot | lol |
| 08:30:33 | | TomW: sorry... |
| 08:30:33 | TomW | ibot: has |
| 08:30:39 | | TomW: excuse me? |
| 08:30:39 | TomW | ibot: been |
| 08:30:40 | fontenot | you gotta think like an ibot |
| 08:30:48 | | TomW: what? |
| 08:30:48 | TomW | ibot: suggested |
| 08:30:59 | fontenot | he parses everything before the is as a token |
| 08:31:10 | fontenot | ibot: is is is is |
| 08:31:21 | fontenot | oops |
| 08:31:35 | TomW | hmmm.. it looks phrase based as well. When you use particular verbs, it extracts a relationship |
| 08:31:49 | fontenot | i think I melted his little brain |
| 08:31:50 | | i already had it that way, TomW. |
| 08:31:50 | TomW | ibot: air: it has been suggested that air is something that some people need. |
| 08:31:59 | | TomW: sorry... |
| 08:31:59 | TomW | ibot: air |
| 08:32:18 | | TomW: I forgot it has been suggested that air |
| 08:32:18 | TomW | ibot: forget it has been suggested that air |
| 08:32:21 | | OK, fontenot. |
| 08:32:21 | fontenot | ibot: air is <reply> it has been suggested that air is something that some people need. |
| 08:32:25 | | it has been suggested that air is something that some people need. |
| 08:32:25 | fontenot | ibot: air |
| 08:32:28 | fontenot | :-D |
| 08:32:40 | BZFlag | fun with ibot? ;-) |
| 08:32:46 | | 54865566756 |
| 08:32:46 | BZFlag | ibot 234234 * 234234 |
| 08:32:51 | fontenot | he's a fun little fellow |
| 08:33:03 | | TomW is nothing but an imp-bladdereddle-headed ass-full of horn-beat Stimpy-drool. |
| 08:33:03 | BZFlag | ibot: insult TomW |
| 08:33:03 | TomW | hey! Cool! where did you come up with that <reply> keyword stuff? |
| 08:33:11 | fontenot | TomW: I read the help |
| 08:33:16 | | TomW: what? |
| 08:33:16 | TomW | ibot: bad ibot. |
| 08:33:17 | fontenot | TomW: ask ibot for help |
| 08:33:24 | TomW | ibot: help me |
| 08:33:33 | fontenot | he should open a private chat with you |
| 08:33:41 | | BZFlag: Provo Municipal Automatic Weather Observing / Reporting System, UT, United States; (KPVU) 40-13N 111-43W; last updated: Mar 29, 2002 03:15 AM EST; Dew Point: 33 F (1 C); Pressure (altimeter): 30.04 in. Hg (1017 hPa); Relative Humidity: 70%; Sky conditions: clear; Temperature: 42 F (6 C); Visibility: 10 mile(s); Wind: Calm |
| 08:33:41 | BZFlag | ibot: weather for kpvu |
| 08:33:48 | | fontenot: The hardware bus needs a new token. |
| 08:33:48 | fontenot | ibot: excuse |
| 08:33:52 | TomW | shit, he gets up close and personal when you ask for help! |
| 08:33:57 | fontenot | right |
| 08:34:06 | fontenot | no need for everyone to see that |
| 08:34:26 | | fontenot is nothing but a clapper-clawed pile of spongy poop. |
| 08:34:26 | TomW | ibot: insult fontenot |
| 08:34:33 | fontenot | :-( |
| 08:34:34 | | fontenot is nothing but an hedge-born thimbleful of beslubbering craptacular carpet droppings. |
| 08:34:34 | TomW | ibot: insult fontenot |
| 08:34:38 | TomW | ahhh! |
| 08:34:58 | | eight ball is probably Yes - definitely |
| 08:34:58 | fontenot | ibot: eight ball |
| 08:35:08 | fontenot | probably definitely? |
| 08:35:22 | | ...but tom is asleep... |
| 08:35:22 | TomW | ibot: Tom is <reply> just another fool scribbling around in kernel space. |
| 08:35:29 | | TomW: I forgot tom |
| 08:35:29 | TomW | ibot: forget tom |
| 08:35:31 | | OK, TomW. |
| 08:35:31 | TomW | ibot: Tom is <reply> just another fool scribbling around in kernel space. |
| 08:35:37 | | just another fool scribbling around in kernel space. |
| 08:35:37 | TomW | ibot: tom |
| 08:35:43 | TomW | cool. |
| 08:35:53 | fontenot | i like ibot |
| 08:36:19 | TomW | it is interesting to play with the thing. Too bad it has the IQ of a dog. |
| 08:36:25 | | okay, fontenot. |
| 08:36:25 | fontenot | ibot: tom is also obsessed with air |
| 08:36:31 | | just another fool scribbling around in kernel space. or obsessed with air |
| 08:36:31 | fontenot | ibot: tom |
| 08:36:38 | fontenot | heh heh |
| 08:36:57 | TomW | Bill Cosby: "Why is there air?" |
| 08:37:05 | BZFlag | fontenot: I ran out of length on eightball so I could not make them all <reply>s |
| 08:37:22 | fontenot | heh |
| 08:37:58 | TomW | is away: coding |
| 08:45:02 | fontenot | heh the only thing bigger than the reiserfs driver is the ext3 driver |
| 09:18:08 | LoneTech | hello |
| 09:18:54 | LoneTech | I've been trying to unlock the flash in the tuxscreen, but it only went down to 3, not 0 |
| 09:19:23 | LoneTech | BZFlag: can you help? |
| 09:22:53 | LoneTech | anyone? |
| 09:23:16 | BZFlag | LoneTech: do it again. |
| 09:23:24 | BZFlag | and again... until you get 0s |
| 09:23:42 | BZFlag | make sure you have a real 12v |
| 09:24:00 | BZFlag | I suspect that low voltage will cause incomplete unlocks. |
| 09:24:05 | LoneTech | my instrument says it's about 12.01-12.03 |
| 09:24:14 | LoneTech | and I've tried something like 60 times |
| 09:24:17 | BZFlag | sounds good. |
| 09:24:23 | BZFlag | oh. |
| 09:24:31 | BZFlag | turn it up a little? |
| 09:24:35 | LoneTech | trying to relock it changes the first one to f, unlocking goes back to 3 |
| 09:24:38 | fontenot | LoneTech: do you have a 12v battery you can try? |
| 09:24:43 | BZFlag | I use 8 AA batteries. |
| 09:24:53 | LoneTech | I can try with my separate stepless supply |
| 09:24:57 | fontenot | i use 1 12V garage door opener battery |
| 09:25:10 | fontenot | tiny little thing about 1/3 AA in length |
| 09:25:39 | BZFlag | I'd say try risking a bit more voltage. say 14v or so. |
| 09:25:40 | fontenot | hmm more like 1/2 |
| 09:25:56 | LoneTech | so far I've tried with the same ATX supply the 3.3 and 5 came from |
| 09:26:14 | LoneTech | but I have a big adjustable supply.. |
| 09:26:20 | fontenot | some supplies can do strange things with no load |
| 09:26:28 | LoneTech | moves laptop to proximity of tuxscreen |
| 09:26:39 | BZFlag | batteries are the safe bet. |
| 09:26:59 | fontenot | yeah |
| 09:27:18 | LoneTech | I don't have many |
| 09:27:34 | LoneTech | hm. perhaps the pack I used for the camera |
| 09:28:15 | LoneTech | 8 AA cells. should do, right? |
| 09:28:25 | BZFlag | that's what I use. |
| 09:28:37 | BZFlag | goes back to boxing phones. |
| 09:28:59 | LoneTech | only it reads 11.26 not 12 |
| 09:29:02 | fontenot | shipping on fridat now? |
| 09:29:11 | LoneTech | may be too spent. I'll try. |
| 09:29:12 | fontenot | LoneTech: please try it anyway |
| 09:29:36 | fontenot | BZFlag: you're not shipping the last 100 are you? :-) |
| 09:30:59 | fontenot | oh there's 300 left |
| 09:31:02 | fontenot | never mind |
| 09:32:13 | LoneTech | it didn't regard that battery as 12V |
| 09:33:01 | fontenot | maybe your connection is not good? |
| 09:33:19 | fontenot | just trying ideas |
| 09:34:58 | LoneTech | I wouldn't expect it to work partway if the connection was bad |
| 09:35:34 | fontenot | so the battery does not work at all? |
| 09:35:39 | LoneTech | I can try exchanging half the cells in the pack |
| 09:35:52 | LoneTech | fontenot: it stayed at waiting for 12v |
| 09:36:02 | fontenot | by bad connection I meany high resistance |
| 09:36:34 | fontenot | but I suspect the switching supply is confusing it |
| 09:36:41 | LoneTech | hm. the cells I got from BZFlag are also not up to 1.5V |
| 09:36:52 | fontenot | I would keep trying to get the battery to work |
| 09:36:57 | fontenot | yes the cells are old |
| 09:36:58 | LoneTech | shall I measure the resistance? |
| 09:37:05 | fontenot | mine died quickly :-) |
| 09:37:14 | LoneTech | or get dressed and go buy two packs of AA cells |
| 09:37:34 | fontenot | LoneTech: how are you making the connection? soldered wires or holding by hand? |
| 09:37:53 | LoneTech | wires are soldered onto the tuxscreen pads.. |
| 09:38:00 | fontenot | ok |
| 09:38:14 | fontenot | that's how I did it but I also know ppl who just help the wires to the pads |
| 09:38:17 | fontenot | held |
| 09:38:34 | LoneTech | the other end, ATX supply: jammed into connector, battery: held by hand, alternate supply: pinched in place (screw grip thingies) |
| 09:39:11 | fontenot | when I used battery I had to use electrical tape to connect wire to battery |
| 09:39:17 | LoneTech | I only have two hands, and wouldn't trust my ability to touch only that small pad among that group anyway |
| 09:39:22 | fontenot | it did not want to work with me just holding wires to battery |
| 09:39:44 | TomW | BZFlag: question on the make system for buildroot. |
| 09:39:53 | fontenot | LoneTech: you have more than one tuxscreen right? |
| 09:40:01 | LoneTech | no, a friend has the other. |
| 09:40:08 | fontenot | ok |
| 09:40:17 | LoneTech | he's building PFI-capable power supplies |
| 09:40:31 | fontenot | nice |
| 09:40:35 | LoneTech | (he's more into hardware, I'm more into software) |
| 09:40:59 | fontenot | TomW: BZFlag is boxing phones |
| 09:41:00 | LoneTech | is it normal that the dsp runs occasionally even in the absence of pfi? |
| 09:41:12 | TomW | ok, later then. make money! |
| 09:41:13 | fontenot | i do not know |
| 09:41:15 | LoneTech | it sometimes clicks or dongs |
| 09:41:24 | fontenot | hmm |
| 09:41:41 | LoneTech | only while running inferno tho. not while in sboot. |
| 09:41:52 | LoneTech | I also haven't managed to get the remote connection thing to work |
| 09:42:07 | | hmmm... BZFlag is TimRiker or a 3D internet multiplayer multiplatform (linux, irix, win32, etc) opensource opengl Battlezone capture the flag game that you must try at http://BZFlag.org/ debian arm linux contains a version that will run on the ipaq at like 6 fps |
| 09:42:07 | TomW | ibot: BZFlag |
| 09:42:09 | LoneTech | which I believe would be helpful when downloading blob |
| 09:42:17 | | BZFlag was last seen on #tuxscreen 13 minutes and 40 seconds ago, saying: goes back to boxing phones. [Fri Mar 29 09:28:37 2002] |
| 09:42:17 | TomW | ibot: seen BZFlag |
| 09:43:02 | LoneTech | ok, I think I'll try with fresh AA cells. then perhaps jam the wires into place. |
| 09:44:26 | LoneTech | btw, I don't use only wires, to do the temporary 12v, I also put a microswitch on the 12v wire |
| 09:45:02 | fontenot | that should be ok |
| 09:45:26 | fontenot | as long as switch is ok |
| 09:45:37 | LoneTech | I opened it up.. simple mechanism |
| 09:45:43 | LoneTech | button has a spring connected to it |
| 09:45:57 | LoneTech | that spring pulls at a metal blade, connected to the "output" |
| 09:46:24 | LoneTech | pressing the button pulls the spring down, causing the blade to jump to the other pad |
| 09:46:33 | LoneTech | very little bouncing, I think |
| 09:46:51 | LoneTech | and I've tried many times having connected it before doing P/u 0 |
| 09:47:44 | LoneTech | spots a bag with cells |
| 09:48:40 | LoneTech | nah, only one at full capacity. |
| 09:48:59 | LoneTech | well, off to the store, I guess |
| 09:49:02 | fontenot | maybe use 9 AA cells? |
| 09:49:06 | fontenot | or 10 |
| 09:49:11 | LoneTech | hm |
| 09:49:12 | fontenot | whatever it takes to get > 12V |
| 09:49:17 | LoneTech | tricky. but I can try. |
| 09:52:38 | LoneTech | tried 10 cells for 14V |
| 09:52:41 | LoneTech | failed |
| 09:52:59 | LoneTech | it recognized it as 12V but didn't go below 3 in unlocking |
| 09:54:13 | LoneTech | I'm off to the store |
| 09:54:19 | fontenot | ok |
| 09:54:54 | TomW | one nice thing about being a Ham radio operator, there is always a 12v supply around. |
| 09:55:13 | TomW | well, 13.8v |
| 09:55:47 | fontenot | :-) |
| 09:56:37 | TomW | While I wasn't looking BZFlag changed the makefile system in buildroot. looks better. |
| 09:56:56 | LoneTech | my alternate supply is marked Model: AL 370/5 stabilized DC power supply 3 - 15 V = 7 A Max. 1' |
| 09:57:26 | TomW | variable supply? |
| 09:57:30 | LoneTech | says at 14-15V it can give 3A continuously or 6A max 1' |
| 09:57:33 | LoneTech | yes. |
| 09:57:45 | LoneTech | alpha+elettronica |
| 09:57:51 | LoneTech | italian |
| 09:58:06 | fontenot | you dont need much current at all for the unlock |
| 09:58:15 | fontenot | i used a tiny A23 batt |
| 09:58:17 | TomW | I came in the middle of this conversation. I am assuming you were talking about unlocking the Flash and the supply wasn't strong enough? |
| 09:58:50 | LoneTech | TomW: well, it only goes to 3, not 0, when I try to unlock. I don't know for certain where the problem is. |
| 09:59:07 | fontenot | he's tried supply and batteries but neither worked |
| 09:59:10 | LoneTech | I'll go to the store and buy 8 fresh AA cells to try |
| 09:59:31 | TomW | isn't there a zener on the +12 volt programming line to the CPU, thought I saw one there.. |
| 09:59:33 | LoneTech | should I measure the resistance over the wires? |
| 09:59:57 | LoneTech | tomw: well, below 12V it just says waiting for 12V. that includes 11.3V. |
| 10:00:53 | LoneTech | well, off to the store.. see you soon |
| 10:01:23 | TomW | yes, you have a digital gate sensing the +12 supply: 74LVC04. That gate is a level shifter, you have to get below 0.8 volts before it will switch. 3 volts is not low enough, it will still think that +12 is still applied. |
| 10:03:21 | TomW | with the voltage divider of 100K & 33K, a 3 volt applied will give 1 volt to the input of the 74LVC04. |
| 10:23:23 | LoneTech | umm.. |
| 10:23:34 | LoneTech | I don't understand that. Sorry. |
| 10:26:36 | LoneTech | fresh cells, 8, giving 12.87V, no better. |
| 10:26:57 | TomW | The designers of the Shannon also put in a protection feature that is used to sense if the programming voltage is being applied. They used a simple digital gate and resistor devider to look for a voltabe being applied. I suspect that you could apply as little as 3 volts to the programming supply, remove it, and it would unlock. |
| 10:27:07 | fontenot | LoneTech: please check your soldering and your wires |
| 10:27:23 | fontenot | LoneTech: you may have a weak solder joint |
| 10:27:31 | LoneTech | guess so |
| 10:27:38 | LoneTech | but why would it always stop at 3 then? |
| 10:27:54 | LoneTech | shall I check the resistance? I am very new to this :/ |
| 10:29:37 | TomW | did you remove the supply when you were checking the flash to see if it was unlocked, or did you simply dial it down to the 3 volt setting? |
| 10:29:57 | LoneTech | I released the switch. the 12V wire has a microswitch on it. |
| 10:29:58 | fontenot | he tried batteries some times too |
| 10:31:35 | TomW | checking data sheets... |
| 10:34:04 | TomW | ok, the 29LV160 needs a 11.5 .. 12.5 volt supply to unprotect the sectors. |
| 10:34:57 | LoneTech | it has gone from f to 3 |
| 10:36:02 | TomW | The +12 is applied to pin 12 (#RESET) of the device to unlock the protected sectors. |
| 10:38:16 | TomW | The +12 must remain applied during the sector unprotect operation that you issue via and inferno command. After the inferno has finished its' operation, then you can remove the +12 supply, not before. |
| 10:39:00 | LoneTech | it finishes.. with "undefined error" or something like that. |
| 10:39:06 | TomW | I think that the Wiki suggested that you could "pulse" the +12v on the unlock supply line, the datasheet says that this is not so. |
| 10:40:06 | TomW | Do a ps -ax, take a look at the process stack of your machine. You will probably find 10..20 of those damned inferno programs on the stack. I have seen that. |
| 10:40:22 | LoneTech | I never got the remote to work |
| 10:40:28 | LoneTech | I am doing P/u from the console |
| 10:40:50 | TomW | OH! You went in via minicom? |
| 10:41:04 | LoneTech | no |
| 10:41:09 | LoneTech | Ctrl+P mode |
| 10:42:35 | TomW | eeehhh... You are on your own there. I could help you with inferno. I have a 2.4 kernel + Mandrake 8.0 system here, you have to do just a wee bit more than install inferno, you have to replace one of the inferno files with a later version to get it to work under a 2.4 kenrel system |
| 10:43:00 | LoneTech | I did that.. only it claims to be an earlier version, not later.. |
| 10:43:05 | LoneTech | and they behave the same. |
| 10:43:18 | fontenot | LoneTech: what error do you get from inferno? |
| 10:43:35 | LoneTech | none, it goes to sleep at <sending reset> or something like that |
| 10:43:36 | fontenot | LoneTech: I had all sorts of problems until I made my serial devices 666 |
| 10:44:05 | LoneTech | it keeps ttyS0 open |
| 10:44:11 | fontenot | also like TomW said, you have to make sure there are not a lot of them running |
| 10:44:14 | TomW | just edit /etc/group and put your username begin the uucp entry, you won't need chmod 666 |
| 10:44:34 | TomW | s/begin/behind/ |
| 10:44:38 | LoneTech | I believe it's serial or dialout. and I have that group. |
| 10:44:46 | LoneTech | Debian sid system |
| 10:44:53 | TomW | hmmm. |
| 10:45:02 | TomW | debian... |
| 10:45:20 | TomW | should still work though. |
| 10:45:44 | TomW | LoneTech: How long you gonna hang out? |
| 10:46:04 | LoneTech | oh, I can hang here about 11 hours more, I think |
| 10:46:15 | TomW | wait! how about I send you a tarball of my inferno? |
| 10:46:20 | TomW | the one that works. |
| 10:46:55 | LoneTech | sure. I'll just reconnect from donkey, brb |
| 10:47:45 | LoneTech | re |
| 10:47:53 | LoneTech | (phone) |
| 10:48:04 | TomW | It is 20M, how fat is your pipe? |
| 10:48:21 | LoneTech | 512kbps |
| 10:48:30 | TomW | ok, here goes. |
| 10:50:28 | LoneTech | thank you |
| 10:52:39 | TomW | put the tarball in your $HOME dir. then: "tar zxf Inferno.tar.gz". It will create a dir named Inferno. then edit the paths inside Inferno/RunInferno.sh |
| 10:53:02 | TomW | Then launch the RunInferno.sh script from a Xterm. |
| 10:53:33 | LoneTech | ok |
| 10:54:16 | TomW | Note: the -g1024x768 inside RunInferno.sh, that tells it how big your graphics screen is. Or something like that. |
| 10:54:25 | LoneTech | ic |
| 10:54:29 | TomW | I use 1024x768 here. |
| 10:54:51 | LoneTech | donkey has 1600x1200 |
| 10:55:22 | TomW | You may have to set it to that, dunno (-g1600x1200) |
| 10:56:24 | TomW | I know what happened to you, you read the Wiki and grabbed the "updated" file from there. For some reason it was the same file as what it was to replace. I don't recall where I got the updated one... |
| 10:56:57 | LoneTech | no, that updated one was older than the one I got in the Linux.tgz thingy |
| 10:57:29 | TomW | right. it was linuxemu to replace the emu file. |
| 10:57:38 | LoneTech | yes |
| 10:57:51 | LoneTech | the linuxemu.gz file contained an emu that claimed to be older |
| 11:00:30 | LoneTech | heh |
| 11:00:35 | TomW | running? |
| 11:00:37 | LoneTech | all I have to change is tom to yann |
| 11:00:54 | TomW | great minds think alike. |
| 11:02:14 | | BZFlag is, like, TimRiker or a 3D internet multiplayer multiplatform (linux, irix, win32, etc) opensource opengl Battlezone capture the flag game that you must try at http://BZFlag.org/ debian arm linux contains a version that will run on the ipaq at like 6 fps |
| 11:02:14 | TomW | ibot: BZFlag |
| 11:03:29 | LoneTech | ; ; bind -a '#t' /dev ; rdp -s115200 -r |
| 11:03:29 | LoneTech | rdp 0.17 (port=/dev/eia0, bps=115200) |
| 11:03:29 | LoneTech | <sending reset> |
| 11:03:59 | fontenot | LoneTech: maybe it is your cable? |
| 11:04:15 | LoneTech | it's the same cable I've used to control the Sun. |
| 11:04:16 | TomW | You have to reset the tux to get it to connect. |
| 11:04:32 | LoneTech | cold boot enough? |
| 11:04:43 | ctr | has debian running on his psion 5 |
| 11:04:49 | LoneTech | after hitting Escape and Ctrl+A, the reset switch on the back doesn't seem to do anything |
| 11:05:41 | TomW | I used a null modem connected to my /dev/ttyS0 and the db9 on the Tux. |
| 11:06:34 | TomW | right. Reset (or cold boot), then hit <esc>, followed by CTRL-A. |
| 11:06:45 | TomW | WAIT one, I have a virgin unit here. |
| 11:09:09 | TomW | okay, that worked. |
| 11:11:48 | TomW | did you get the screen that says "Activate Failsafe mode?" then when you hit ctrl-a it goes blank (this is how it behaves without inferno)? |
| 11:12:12 | LoneTech | babbling friend on the phone.. can't concentrate |
| 11:12:22 | TomW | Kill him... |
| 11:12:41 | TomW | ok, that wasn't nice. sorry. |
| 11:13:00 | LoneTech | done |
| 11:13:28 | LoneTech | yes. |
| 11:13:41 | LoneTech | it is now blank |
| 11:14:04 | TomW | Ok, then while it is in that state, you do the " bind -a '#t' /dev ; rdp -s115200 -r" from inferno. |
| 11:14:58 | LoneTech | rdp 0.17 (port=/dev/eia0, bps=115200) |
| 11:14:59 | LoneTech | <sending reset> |
| 11:15:17 | LoneTech | and nothing. |
| 11:15:40 | LoneTech | does it require handshaking? |
| 11:15:49 | TomW | Aha! Debian is screwing you! What is the /dev nanme of your com1 ? |
| 11:16:26 | TomW | I think you callled it /dev/eia0 ? |
| 11:16:53 | LoneTech | crw-rw---- 1 root dialout 4, 64 Mar 29 12:17 /dev/tts/0 |
| 11:16:53 | LoneTech | lr-xr-xr-x 1 root root 5 Mar 24 19:39 /dev/ttyS0 -> tts/0 |
| 11:16:54 | LoneTech | lrwxrwxrwx 1 yann yann 0 Mar 29 12:19 /proc/17940/exe -> /home/yann/TuxScreen/TomW/Inferno/Linux/386/bin/emu |
| 11:16:54 | LoneTech | lrwx------ 1 yann yann 64 Mar 29 12:19 /proc/17940/fd/3 -> /dev/tts/0 |
| 11:17:05 | LoneTech | inferno calls it eia0. |
| 11:18:15 | TomW | ok, that wasn't it. I thought that was what debian called their ports. see: "strings emu | grep ttyS0". |
| 11:18:33 | LoneTech | as you can see on the proc lines, inferno has it open |
| 11:19:00 | LoneTech | it said /dev/ttyS0 |
| 11:21:07 | LoneTech | what mode does the serial port initially come up in? |
| 11:21:36 | TomW | I am using a full modem port to connect to the tux. The ttyS0 port is a nine pin, I have a nine pin cable running out to a null modem then to a gender changer and it is plugged into the AUX port of the Tux. |
| 11:22:27 | TomW | mode? |
| 11:22:29 | LoneTech | yes! |
| 11:22:36 | LoneTech | flash: id=1/49 (AMD Am29LV160BB), ss=20000 fs=400000 w=2 b=0-20000 f=5 |
| 11:23:02 | TomW | Roger, Houston, we have a go for seperation. |
| 11:23:14 | fontenot | ?? |
| 11:23:17 | fontenot | what was it? |
| 11:23:31 | TomW | operator head space error? |
| 11:23:41 | LoneTech | looks like my inferno was broken.. |
| 11:23:50 | LoneTech | and in trying to get it connected, I had tried ttyS1 too |
| 11:23:57 | LoneTech | that's why yours failed initially |
| 11:24:08 | TomW | operator head space error. |
| 11:24:13 | fontenot | heh |
| 11:24:17 | TomW | lol |
| 11:24:38 | LoneTech | still all at 3 though |
| 11:24:48 | LoneTech | thanks |
| 11:25:16 | fontenot | d'oh |
| 11:26:05 | LoneTech | but now there's a greater distance between the sboot prompt and the switch. it still doesn't unlock. |
| 11:26:08 | TomW | keep up with connecting volts, eraseing, remove volts, and checking til they all go to 0. One unit I had it took 5 tries to completely erase it. |
| 11:26:20 | LoneTech | didn't take 70 though |
| 11:26:49 | TomW | DO NOT STOP WHAT YOU ARE DOING! Once you erase, then put blob on the unit. Do not power it down or remove inferno! |
| 11:27:11 | LoneTech | hm? |
| 11:27:20 | LoneTech | I thought the unlock was reversible |
| 11:27:29 | LoneTech | the critical part ought to be when I download blob |
| 11:28:17 | LoneTech | 00000000-00007fff and 0000c000-0000ffff through 000c0000-000dffff still are at 3 |
| 11:28:21 | LoneTech | the rest are at 0 |
| 11:28:24 | TomW | While the Flash is unlocked it can be easily scrambled. Then you, potentially, cannot get inferno to link up again. Some people have had this happen. |
| 11:28:34 | LoneTech | k |
| 11:28:41 | LoneTech | it's not scrambled so far |
| 11:28:48 | LoneTech | and I have a jtag dongle, just in case. |
| 11:28:48 | TomW | k |
| 11:28:53 | TomW | aha! |
| 11:29:12 | fontenot | so some of them are going to 0 now? |
| 11:29:22 | LoneTech | fontenot: they did that the first try. |
| 11:29:38 | LoneTech | some of them are stubborn at 3 |
| 11:29:46 | TomW | you are not verifying the erasure with the voltage applied, are you? |
| 11:29:53 | LoneTech | I can get the first to f with P/p, no other effects. |
| 11:29:57 | LoneTech | TomW: no. |
| 11:30:00 | TomW | ok. |
| 11:30:13 | LoneTech | although it does say the exact same with the voltage applied |
| 11:30:29 | LoneTech | but I thought this was unprotecting, not erasing |
| 11:32:08 | TomW | "P/u 0" to unlock under volts, then P/m to check without volts. |
| 11:32:42 | LoneTech | yes. that's what I've done all the time, except once I did P/m with 12v out of curiosity. |
| 11:32:53 | LoneTech | (after about 20 unsuccessful tries, iirc) |
| 11:33:26 | TomW | So did I, that didn't hurt anything. |
| 11:33:42 | BZFlag | LoneTech: you running devfs? |
| 11:33:46 | LoneTech | yes |
| 11:34:01 | BZFlag | I think I tried that once, and Inferno did not like it at all. |
| 11:34:12 | LoneTech | well, it's connected now |
| 11:34:14 | BZFlag | no idea why, but reverting to normal /dev fixed it. |
| 11:34:27 | LoneTech | and ttyS0 is a symlink to tts/0, which inferno has open. |
| 11:34:54 | BZFlag | cool. |
| 11:34:57 | LoneTech | yann@donkey:~/TuxScreen/TomW/Inferno$ groups yann |
| 11:34:58 | LoneTech | yann : yann tty disk dialout cdrom floppy tape audio dip video staff |
| 11:35:04 | LoneTech | crw-rw---- 1 root dialout 4, 64 Mar 29 12:36 /dev/tts/0 |
| 11:35:22 | BZFlag | reads the backlog |
| 11:36:35 | LoneTech | btw, the screen went blue and says debug mode when the serial inferno connection started working. |
| 11:38:24 | LoneTech | my laptop has a 640x480 colour touchscreen |
| 11:38:32 | LoneTech | smaller than that on the tuxscreen tho |
| 11:43:40 | BZFlag | LoneTech: is it possible you over heated the connection when soldering? I suspect that one of the two chips may be disconnected now from the 12v feed somehow. |
| 11:47:22 | LoneTech | possible, I suppose |
| 11:47:24 | LoneTech | phone |
| 11:51:18 | BZFlag | gotta sleep. check the board closely, check the reverse side too. I suspect a tracing is broken. |
| 11:51:36 | BZFlag | sorry to leave you, but I must sleep. |
| 12:07:18 | LoneTech | the via hole closest to the 12V pad, should it be connected to the 12V pad? |
| 12:08:07 | TomW | Wait a bit, I am erasing the last of the Tuxes now. will check soon. |
| 12:09:29 | LoneTech | I was careless with the solder (inexperienced) and connected them while soldering. Noticing this when verifying (did not apply power) I disconnected them, but I am not sure if they were connected before I applied too much solder |
| 12:12:10 | TomW | ok. There is a via at 10:00 near the +12v pad, it is connected to the +12v |
| 12:13:54 | TomW | I am looking at the unit with it lying face down and the LED lens is pointed to my chest. |
| 12:14:18 | LoneTech | 10:00? |
| 12:14:31 | LoneTech | same way I look at it |
| 12:14:41 | LoneTech | you mean a bit up/left? |
| 12:15:03 | LoneTech | it's closer than any of the other vias |
| 12:15:16 | TomW | the pad at the 2:00 position is NOT connected to +12v. The two pads look like ears on mickey mouse: 10 o'clock & 2 o'clock position. |
| 12:15:52 | LoneTech | but the 2 o'clock is slightly further away. about halfway between 12v and the next pad. |
| 12:15:54 | LoneTech | okay |
| 12:16:09 | LoneTech | so I should _not_ have severed the connection between the via and 12V pad! |
| 12:16:12 | LoneTech | thank you |
| 12:16:14 | TomW | Yes, the 2:00 is slightly further away. |
| 12:16:24 | LoneTech | I will now attack it with my soldering iron again |
| 12:17:56 | TomW | correct, you need that connected to +12. Notice that the +12 trace runs to your right on the PCB and there is another large pad there? That is where you will connect your +12. Just fix the broken trace for mow, use a strand of wire to do it with, then trim it off (carefully) with your X-acto knife. |
| 12:18:32 | LoneTech | wire? it's such a short distance solder seems easier |
| 12:18:32 | TomW | You don't need a big wire, just a wire strand will do. |
| 12:18:40 | TomW | Solder is okay. |
| 12:19:12 | TomW | Watch out for the trace near that pad, don't get solder on it. |
| 12:19:27 | TomW | the thin trace above the 10:00 via. |
| 12:19:55 | LoneTech | the one that goes to a via some ways to the right of the upper row of pads (4 of them)? |
| 12:20:04 | LoneTech | around the right hole |
| 12:20:15 | LoneTech | I'll be careful. |
| 12:20:33 | LoneTech | I made the mistake of not checking _before_ I soldered the first time |
| 12:20:42 | LoneTech | that's why I didn't know the via was to be 12v |
| 12:21:39 | TomW | I always get nervous when mod-ing equipment, I have to work myself up to it.. Unless it is my own design, then I get brutal with it. :-D as in "don't you dare fail". |
| 12:25:38 | TomW | LoneTech: When you reconnect the inferno, use 115200 not 9600 baud. I just flashed the previous Tux at 9600 and I have a bad / broken blob in that unit. |
| 12:29:11 | LoneTech | OK |
| 12:29:23 | LoneTech | hope the jtag works then |
| 12:29:58 | LoneTech | <contact has been made> |
| 12:30:21 | LoneTech | yay! all 0s! |
| 12:31:07 | TomW | It is always the last thing that you did that was wrong, eh? |
| 12:31:14 | LoneTech | md5sum okay.. |
| 12:31:20 | LoneTech | sure |
| 12:31:28 | LoneTech | -rw-r--r-- 1 yann yann 14054 Mar 28 16:06 /home/yann/Download/tuxscreen-blob |
| 12:31:31 | LoneTech | bea2e6a180ec25c3af7dd4d2966f243a /home/yann/Download/tuxscreen-blob |
| 12:31:37 | LoneTech | that's the same you have? |
| 12:32:18 | TomW | bea2e6a180ec25c3af7dd4d2966f243a tuxscreen-blob |
| 12:32:26 | TomW | yup. |
| 12:32:39 | LoneTech | good. |
| 12:32:49 | TomW | Move that file under TuxScreen/Inferno. |
| 12:33:06 | LoneTech | is it normal that no progress is visible when doing "c F!all D!inferno.backup"? |
| 12:33:30 | TomW | Then you simply do: "c D!tuxscreen-blob F!all" to program the blob. |
| 12:33:54 | LoneTech | c is copy.. D is disk.. F is flash |
| 12:34:00 | LoneTech | at least that's my guess (: |
| 12:34:02 | TomW | Very few people have actaully saved Inferno. I tried to download it for over 30 minutes and I gave up. |
| 12:34:10 | LoneTech | I see |
| 12:34:14 | LoneTech | that big? |
| 12:34:37 | LoneTech | should be possible to fetch within 10minutes |
| 12:34:44 | LoneTech | not important to me, really |
| 12:34:49 | TomW | Others have said the same, left it running for several hours and still no file downloaded. I think the Inferno in the Tux is incomplete. |
| 12:35:01 | TomW | c D!tuxscreen-blob F!all |
| 12:35:28 | LoneTech | it certainly hasn't written anything into the file |
| 12:35:33 | LoneTech | can it be interrupted? |
| 12:35:37 | TomW | heh |
| 12:35:50 | TomW | yes, but you'd better do a 'killall emu' |
| 12:36:02 | TomW | if it crashes on you, inferno that is. |
| 12:36:06 | LoneTech | they all died properly. |
| 12:37:11 | LoneTech | ok.. |
| 12:37:19 | LoneTech | here we go. |
| 12:37:40 | TomW | did you see my site for the TuxScreen. I give a blow-by-blow how to do this? |
| 12:37:52 | LoneTech | flash: id=1/49 (AMD Am29LV160BB), ss=20000 fs=400000 w=2 b=0-20000 f=5 |
| 12:37:52 | LoneTech | Flash write: fa15d4,36e6 -> 0-36e5 |
| 12:37:52 | LoneTech | flash: 0-7fff: erasing/writing... flash: erasing boot area! (0-7fff) |
| 12:37:52 | LoneTech | <38?><terminal mode> |
| 12:37:52 | LoneTech | 8ÂÂÂ8ÂÂÂ8ÂÂÂ8ÂÂÂ8ÂÂÂ8ÂÂÂ8ÂÂÂ8ÂÂ<rdp mode> |
| 12:37:53 | LoneTech | <sending reset> |
| 12:38:34 | TomW | Good, now quit inferno, killall emu, run minicom at no handshakes (hardware / software) at 115200 8 n 1 |
| 12:38:47 | TomW | Then reset the Tux, you should be Larted. |
| 12:39:12 | LoneTech | W.Äç..æc"ækC.Gó.öZï...c..!)rBBB#BA`.á..Äÿÿÿ..kÿÿÿÿÿÿÿÿÿJ.æ.Húÿÿvg. |
| 12:39:12 | TomW | It always dumps crap after flashing blob. |
| 12:39:35 | LoneTech | Consider yourself LARTed! |
| 12:39:43 | LoneTech | YAY! |
| 12:39:45 | TomW | hit returm to stop. |
| 12:39:54 | LoneTech | it's at 115200, not 9600. |
| 12:40:27 | TomW | Yes, it should be, some units (few) do end up at 9600 baud. |
| 12:40:38 | TomW | now: http://www.openhardware.net/TuxScreen/InstallingLinux.html |
| 12:40:39 | LoneTech | it looks happy (: |
| 12:40:49 | LoneTech | I have wanted a lart since I first heard about them |
| 12:40:52 | TomW | got a kernel? |
| 12:40:58 | TomW | got a ramdisk? |
| 12:42:12 | LoneTech | not yet |
| 12:42:22 | LoneTech | but soon |
| 12:42:38 | LoneTech | thanks |
| 12:42:54 | LoneTech | now that I have blob in there, I feel much safer. |
| 12:44:30 | TomW | LoneTech: you will also need the ramdisk image as well. |
| 12:45:09 | TomW | Ok, now you are at the blob prompt, right? |
| 12:45:22 | LoneTech | I'm trying to uuencode the kernel in |
| 12:45:28 | TomW | ok. |
| 12:45:28 | LoneTech | it seems kind of quiet |
| 12:45:46 | TomW | uuencode tuxscreen-kernel foo > /dev/ttyS0 |
| 12:45:52 | LoneTech | why foo? |
| 12:46:08 | TomW | temp file that is there for syntax. |
| 12:46:25 | TomW | it does nothing. just keeps uuencode happy. |
| 12:46:28 | LoneTech | right |
| 12:46:42 | TomW | You will see dots running in blob...................................................... |
| 12:48:58 | TomW | LoneTech: The suspense is killing me! |
| 12:49:06 | TomW | heh |
| 12:49:15 | LoneTech | don't worry |
| 12:49:19 | LoneTech | I'm just messing with minicom |
| 12:49:23 | TomW | heh |
| 12:49:36 | TomW | no handshakes. |
| 12:50:53 | TomW | Run minicom as user: minicom -s, do your port setups, make sure you also delete all the AT krap so it doesn't condition a "modem" when it starts. |
| 12:51:23 | LoneTech | dots |
| 12:51:29 | TomW | cool! |
| 12:51:32 | LoneTech | I use minicom -o |
| 12:51:58 | LoneTech | and since I only use this for serial consoles on the Sun, SGI, and TuxScreen, I've removed the modem crap (and the handshaking too) |
| 12:52:12 | TomW | ok, I just never did a man minicom before. |
| 12:52:20 | LoneTech | -o means no resets |
| 12:52:45 | LoneTech | turns minicom linewrap on |
| 12:52:52 | TomW | heh |
| 12:53:10 | LoneTech | ran out of space for the dots |
| 12:53:14 | LoneTech | Received 901432 (0x000DC138) bytes. |
| 12:53:14 | LoneTech | blob> |
| 12:53:20 | TomW | flash kernel |
| 12:53:33 | LoneTech | Saving kernel to flash .... |
| 12:53:39 | ctr | does the "I have X11 working on my psion5mx" dance |
| 12:53:43 | LoneTech | Saving kernel to flash ....... ....... done |
| 12:53:43 | LoneTech | blob> |
| 12:53:54 | LoneTech | ctr: good, now put picogui there. |
| 12:54:03 | TomW | then power cycle the unit, you must power cycle. if you don't then the ramdisk download / flash will run REAL SLOWLY. |
| 12:54:08 | ctr | i've tried it |
| 12:54:13 | LoneTech | typing reset ok? |
| 12:54:19 | ctr | wasn't that impressed, there are zero applications for picogui :) |
| 12:54:24 | TomW | No, power cycle it. |
| 12:54:24 | LoneTech | or maybe even reblob? |
| 12:54:31 | LoneTech | ctr: hey, we have xchat! |
| 12:54:39 | LoneTech | sure |
| 12:54:41 | TomW | stay away from reblob! |
| 12:54:48 | ctr | uh |
| 12:54:53 | ctr | xchat |
| 12:54:53 | ctr | wow :) |
| 12:54:59 | LoneTech | corrupt jffs! |
| 12:55:22 | LoneTech | more importantly, happy kernel |
| 12:55:28 | TomW | ?? corrupt, how? during download? |
| 12:55:36 | LoneTech | no, it's just not there yet |
| 12:55:51 | LoneTech | jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x002c2c24: 0x0978 in |
| 12:55:51 | LoneTech | stead |
| 12:56:30 | LoneTech | downloading your ramdisk image now |
| 12:56:38 | LoneTech | thank you so much for all your help (: |
| 12:56:53 | TomW | Don't thank me yet, unitl it boots.... |
| 12:57:09 | LoneTech | maybe put a warning that stopping at 3 means that line was severed? |
| 12:57:18 | LoneTech | for the flash unlocking |
| 12:57:44 | LoneTech | btw, it still said undefined error even when it got it right. |
| 12:57:44 | TomW | No, something else may have happened. |
| 12:57:59 | LoneTech | but I know that connection was severed |
| 12:58:04 | LoneTech | so it's one thing to check |
| 12:58:15 | LoneTech | flashing ramdisk now |
| 12:58:18 | TomW | best not to cut the trace.. |
| 12:58:47 | LoneTech | yes. but I didn't see the trace. careless enough not to check beforehand |
| 12:59:14 | LoneTech | and the picture on http://tuxscreen.net/wiki/view/Main/FlashUnlock wasn't clear enough to tell me if there was one there |
| 12:59:54 | TomW | Tough to get these digital cameras to focus at that short a range. |
| 13:00:05 | LoneTech | I know |
| 13:00:09 | LoneTech | Freeing init memory: 72K |
| 13:00:10 | LoneTech | tuxscreen login: |
| 13:00:14 | TomW | root |
| 13:00:20 | LoneTech | Welcome to the Shannon |
| 13:01:05 | TomW | cool! You got it, just download the buildroot-tux from the cvs and do a make. Don't edit anything in it yet, just do the defaults. |
| 13:01:09 | LoneTech | (: |
| 13:02:19 | TomW | Yup, just checked to see if cvs has the last anomoly fixed, it does, so you should be in business shortly. I need sleep. |
| 13:02:44 | LoneTech | cvs.. |
| 13:03:12 | TomW | http://sourceforge.net/cvs/?group_id=20228 |
| 13:03:22 | LoneTech | thank you |
| 13:03:46 | TomW | also keep this url handy: http://sourceforge.net/projects/tuxscreen |
| 13:04:17 | TomW | jacques helped me get started, so this was payback. Enjoy. |
| 13:04:43 | LoneTech | (: |
| 13:15:36 | LoneTech | make[3]: arm-linux-gcc: Command not found |
| 13:16:14 | LoneTech | tpkg-make arm (: |
| 13:16:43 | LoneTech | hello GPSFan |
| 13:32:43 | GPSFan | hi LoneTech: how are you this fine Friday? |
| 13:36:57 | LoneTech | happy that my tuxscreen works and runs linux |
| 13:37:14 | GPSFan | great!,you were having problems? |
| 13:37:37 | LoneTech | yes.. I had mistakenly cut a lead on the motherboard, preventing me from unlocking all the flash. |
| 13:37:56 | scanline | scrolled up and read a whole bunch... |
| 13:38:07 | scanline | LoneTech: I prefer the "chopsticks of doom" approach for unlocking flash ;) |
| 13:38:07 | LoneTech | and tomw sent me a working inferno for linux. before that, the remote sboot control didn't work at all. |
| 13:38:41 | GPSFan | ah, what did you cut, and why? |
| 13:38:58 | LoneTech | there's a via hole very close to the 12V pad |
| 13:39:30 | LoneTech | I applied a bit much solder, thought I'd connected those two by mistake, and cut the trace underneath the solder in the process of disconnecting them. |
| 13:40:04 | GPSFan | I see, well, soldering SMT stuff is still more of an art than science. ;>) |
| 13:40:09 | LoneTech | but that via was to give the 12V to 1/4th of the flash chips or something like that |
| 13:40:15 | GPSFan | Hi scanline: how goes pico? |
| 13:40:37 | scanline | GPSFan: not bad! It's spring break, and i'm having fun with my new dreamcast :) |
| 13:40:54 | scanline | having fun == hacking pgserver and eventually linux |
| 13:41:18 | GPSFan | ooowww! are you & BZ working on geting pico to compile under the new buildroot? |
| 13:41:27 | LoneTech | scanline: but those both came after making your homebuilt CDs boot, right? |
| 13:41:37 | scanline | mmm... nyet |
| 13:42:08 | scanline | LoneTech: yeah, compiling the toolchain and kernel was a pain.. though actually I haven't made a CD with the new kernel yet, i've been sending them over the network |
| 13:42:19 | LoneTech | ic |
| 13:42:23 | scanline | there's a handy dc-tool CD that accepts binaries over ethernet or serial |
| 13:42:37 | LoneTech | good. |
| 13:43:09 | scanline | i think i would have used a couple 100-packs of CD-Rs otherwise trying to get this kernel working ;) |
| 13:43:13 | GPSFan | scanline: dc's are almost as good a value as Tuxscreens these days |
| 13:43:17 | LoneTech | does the buildroot depend on the compiler being called arm-gcc? I think mine will be arm-linux-gcc |
| 13:43:35 | LoneTech | GPSFan: easier to find pc card ethernet, I believe |
| 13:43:43 | scanline | GPSFan: Yeah, the price of DC/tux + network card is about the same |
| 13:43:55 | scanline | DC is cheap, but the network cards are real expensive |
| 13:44:14 | scanline | and i didn't have any SuperH hardware yet :) |
| 13:44:24 | LoneTech | how many other SuperH platforms are there? |
| 13:44:30 | LoneTech | I don't know any |
| 13:44:33 | GPSFan | is the DC limited to TV resolution displays? |
| 13:44:39 | scanline | a lot of older handheld PCs were SuperH |
| 13:45:01 | scanline | GPSFan: nope, i usually use it with a VGA monitor. It usually runs at 640x480, though there's a rarely-used 800x608 mode |
| 13:45:10 | LoneTech | my Palmax qualifies as handheld PC. wince thingies do not. |
| 13:45:32 | scanline | ok, a lot of older WinCE thingies were SuperH. I'm not sure what new hardware uses it, if any |
| 13:45:58 | LoneTech | I have one of the predecessors here.. a H8 based Lego brick. |
| 13:46:38 | scanline | Though so far the only actual porting i've done is dealing with peculiarities of the Dreamcast's VRAM, haven't had to change anything for SuperH |
| 13:46:53 | scanline | i found out all those weird problems i was having earlier was due to a buggy -O3 in gcc 3.0 |
| 13:47:01 | LoneTech | k |
| 13:49:27 | LoneTech | scanline: looks like a few more episodes of excel saga have been subtitled. do you have any 19+? |
| 13:49:42 | scanline | ooh... no, i don't have 'em yet |
| 13:50:02 | scanline | Lain is cool... maybe i should get it on DVD |
| 13:50:18 | LoneTech | I have no anime on dvd |
| 13:50:25 | LoneTech | ghost in the shell on vhs |
| 13:51:40 | LoneTech | dvds I have.. Kull, Nirvana, Ghormenghast (full series), the hitchhiker's guide to the galaxy (BBC TV version), Lexx 1.1-2.02 |
| 13:52:30 | LoneTech | dvd pack 2.02 contains episodes 2.3 and 2.4 |
| 13:52:36 | LoneTech | er |
| 13:52:40 | LoneTech | dvds 2.3 and 2.4.. |
| 13:52:44 | LoneTech | episodes up to 2.08 |
| 13:52:54 | LoneTech | smacks self |
| 13:53:03 | LoneTech | 2.10 |
| 13:53:08 | LoneTech | wrong disc (: |
| 13:54:12 | scanline | whoa.. hitchhiker's guide |
| 13:54:15 | LoneTech | yes. |
| 13:54:20 | scanline | i heard the TV version was pretty bad though |
| 13:54:27 | LoneTech | pretty bad.. not extremely |
| 13:54:34 | LoneTech | it's very much like the radio version |
| 13:54:38 | scanline | haha |
| 13:54:48 | scanline | i have the radio version |
| 13:54:58 | scanline | decides to listen to it after the next epsiode of MacGyver |
| 13:56:03 | scanline | macgyver is so cheesy :) |
| 13:59:58 | LoneTech | scanline: same actors, btw |
| 14:00:19 | scanline | ah, that's good at least |
| 14:00:47 | scanline | grr... only one update per mouse move, so my bug is elsewhere... |
| 14:02:51 | scanline | hey... maybe i just got the buffers backwards |
| 15:29:33 | cpeugin_away | is back (gone 15:03:34) |
| 16:02:16 | LoneTech | (: |
| 16:07:21 | LoneTech | it works (: |
| 16:07:39 | CosmicPenguin | If it works, why are you sad... |
| 16:08:17 | LoneTech | I'm not |
| 16:08:21 | LoneTech | (: is happy |
| 17:13:29 | Keyser[a] | eek. Lonetech, your head is on backwards. |
| 17:15:57 | LoneTech | huh? |
| 17:16:52 | Keyser[a] | (: vs. :) |
| 17:17:14 | LoneTech | I'm entitled to do that! |
| 17:17:36 | LoneTech | http://www.algonet.se/~yann/upsidedown.jpg |
| 17:17:38 | Keyser[a] | i know, it just looks painful. |
| 17:45:24 | RedBeard | LoneTech: dude! my beard is bigger than yours! ;) |
| 17:46:07 | LoneTech | RedBeard: it's an old pic |
| 17:46:31 | RedBeard | LoneTech: I'm going to be trimming mine much shorter, for interviewing. |
| 17:46:42 | LoneTech | bah |
| 17:52:45 | ctr | anyone got any ideas on how to remove jitter from the touch scren adc? |
| 18:49:22 | LoneTech | ctr: there's some filtering code in one of picogui's touchscreen drivers |
| 18:49:29 | LoneTech | feel free to grab that |
| 18:50:03 | ctr | sure |
| 18:50:06 | ctr | i think i wilkl |
| 18:50:09 | ctr | how good is it? |
| 18:50:15 | LoneTech | dunno |
| 18:50:23 | CosmicPenguin | tslib has some filtering in it too |
| 18:50:37 | LoneTech | haven't copied it to use it on the other drivers, and it's not a touchscreen I have |
| 18:50:59 | ctr | i was going to put it in the kernel level drivers for the 5mx touchscreen |
| 18:51:13 | ctr | but abstract it enough so all ts's can use it's logic |
| 18:51:39 | LoneTech | nice |
| 18:51:57 | ctr | because i know all touch screens suffer the problem |
| 18:53:21 | LoneTech | yep. the palmax one I have suffers especially horizontally. |
| 18:53:55 | LoneTech | I made some shared calibration stuff in picogui, but haven't rewritten it to handle filtering yet |
| 20:20:28 | LoneTech | BZFlag: it works.. I had cut a trace. |
| 20:20:43 | LoneTech | the via closest to the 12V pad should be connected to it |
| 20:21:06 | LoneTech | if it's not, the flash chip with the 0x3 part doesn't get the unlocking power |
| 20:27:56 | BZFlag | so you jumped the trace and fixed it? Great! |
| 20:29:18 | Russ | LoneTech: oops |
| 20:30:27 | LoneTech | it works now (: |
| 20:31:13 | LoneTech | I was going to rebuild the software, but haven't got gcc working yet. will soonish. |
| 20:31:25 | LoneTech | familiar territory now anyway (: |
| 22:43:45 | fontenot | CosmicPenguin: you awake? |
| 23:04:46 | CosmicPenguin | fontenot: I am here |
| 23:24:20 | fontenot | CosmicPenguin: heh was in shower |
| 23:25:00 | fontenot | CosmicPenguin: are you going to be around a little later? I have some questions about tslib |
| 23:28:51 | CosmicPenguin | I'm probably not the best person to ask |
| 23:29:45 | Russ|werk | fontenot: ask on the linux-arm list |
| 23:30:13 | CosmicPenguin | I have only glanced it over, not actually beat the thing up |
| 23:31:32 | fontenot | CosmicPenguin: I thought you used it to get mw ts working? |
| 23:31:36 | fontenot | maybe I am misremembering |
| 23:32:02 | fontenot | Russ|werk: the linux arm list would laugh me away |
| 23:32:24 | Russ|werk | fontenot: what's the question? |
| 23:32:37 | fontenot | someone said they wrote a custom version of tslib to get mw working |
| 23:32:41 | Russ|werk | can't be nearly as stupid as some of the stuff I've seen on arm kernel |
| 23:32:51 | fontenot | which implies to me that they understand it |
| 23:33:06 | fontenot | Russ|werk: OK I was playing with tslib on my tux last night - the test programs |
| 23:33:08 | Russ|werk | fontenot: do you have a CVS checkout (or snapshot) of tslib? |
| 23:33:17 | fontenot | Russ|werk: yes as of a couple of days ago |
| 23:33:53 | fontenot | I had to modify the ts_test program because it was not even setup for 8bit displays - would not show the cross at all |
| 23:34:14 | Russ|werk | fontenot: rmk really wants tslib to take off I think |
| 23:34:19 | fontenot | and the ts_calibrate program seems to generate a consistent page fault complete with stack trace |
| 23:34:26 | CosmicPenguin | fontenot: I used some of the filtering ideas, but thats all |
| 23:34:33 | Russ|werk | fontenot: if you talk to him about it, or email arm linux, you'll get answers |
| 23:34:41 | Russ|werk | (re: the 8 bit thing) |
| 23:34:46 | CosmicPenguin | Russ|werk: Hopefully, RMK will start moving toward his input subsystem idea.... |
| 23:35:09 | fontenot | Russ|werk: well the 8bit thing was trivial - there was a case statement and 16bit was there but 8bit was just "break;" |
| 23:36:05 | fontenot | my *main* question is how I am supposed to translate coords to screen coords - right now I am getting about 1/2 res in either direction |
| 23:36:24 | fontenot | and there are magic numbers in the linear filter which I do not understand |
| 23:36:32 | CosmicPenguin | prpplague: you around? |
| 23:36:47 | prpplague | CosmicPenguin: ya what up |
| 23:37:00 | fontenot | ok, maybe I should ask this on the arm list. but I will try to do a search first |
| 23:37:11 | prpplague | CosmicPenguin: been workin like crazy lately |
| 23:37:34 | fontenot | it almost seems like the linear filter is hardwired for 320x240 displays |
| 23:37:58 | Russ|werk | fontenot: I assume you've been doing some skimming through the source then |
| 23:42:03 | fontenot | Russ|werk: yeah |
| 23:42:24 | Russ|werk | fontenot: then whatever you email will probably sound reasonably intelligent |
| 23:44:05 | | CosmicPenguin was last seen on #tuxscreen 7 minutes and 33 seconds ago, saying: prpplague: you around? [Fri Mar 29 23:36:32 2002] |
| 23:44:05 | prpplague | ibot: seen CosmicPenguin |
| 23:44:08 | fontenot | Russ|werk: heh ok you've convinced me :-) |
| 23:44:46 | | You are nothing but a vain stack of reeky waffle-house grits. |
| 23:44:46 | prpplague | ibot: insult prpplague |
| 23:45:31 | CosmicPenguin | prpplague: Do you know if perl bindings are available for sqlite? |
| 23:49:58 | prpplague | CosmicPenguin: i know several were available but i don't know about perl |
| 23:50:43 | CosmicPenguin | prpplague: Ok... just wondering |
| 23:51:30 | prpplague | CosmicPenguin: i check the list of messages, and it looks like a lot of ppl discussed getting it, but no one has posted anything about it actually existing |
| 23:52:17 | prpplague | CosmicPenguin: only thing i see is a php binding |
| 23:53:06 | | prpplague_ is nothing but a fen-sucked bucket of fermented fish heads. |
| 23:53:06 | prpplague | ibot: insult prpplague_ |
| 23:54:07 | | yo momma is nothing but a wretched ass-full of it-fowling snake bait. |
| 23:54:07 | Russ|werk | ibot: insult yo momma |
