irclog2html for tuxscreen on 2001.10.27

irclog2html for tuxscreen on 2001.10.27 00:00:15 SpaceCoaster scanline: the MTD driver I posted won't support JFFS2 yet. 00:00:45 scanline SpaceCoaster: I heard. Something related to the 256-byte pages? 00:02:18 SpaceCoaster Yeah, sortof, that got resolved but I discovered that JFFS2 makes small reads on the flash and I the hardware only supports page reads so...I have to do buffering, page overlap, urgh 00:02:35 scanline yech 00:02:47 scanline maybe there's a FS better suited to serial flash? 00:04:22 Lenolium Jeez, but jffs is supposed to be for flash devices. :) 00:04:54 SpaceCoaster maybe JFFS2 is pretty neat but i wasn't expecting it to do such itsy bitsy reads, I have a problem with corruption of byte 256, which is weird because I transfer 264 bytes! 00:05:33 scanline ha. that is weird 00:05:36 SpaceCoaster Lenolium: yeah but most have wider buses than 1 bit! 00:05:49 SpaceCoaster scanline: yeah I strongly suspect duff code :-( 00:05:59 scanline yup 00:06:32 BZFlag ah, moorman found it. ok, cvs commits are mailing now. 00:06:54 SpaceCoaster Anyway it is close, what would we store in /flash when it works? 00:08:51 BZFlag I'm happy with a block device. I doubt I'd put jffs2 on it. 00:08:59 SpaceCoaster scanline: I guess tuxphone could store/restore state for /flash? 00:09:07 BZFlag probably just compressed tarballs and such. 00:09:17 scanline SpaceCoaster: What state? 00:09:37 SpaceCoaster scanline: my volume setting etc? 00:09:48 BZFlag we could put the kernel up there. ;-) 00:09:51 scanline i suppose 00:09:54 scanline haha 00:10:14 SpaceCoaster BZFlag: go for it, get that kernel into 512K! 00:10:22 scanline SpaceCoaster: Maybe we store the user's .files in the flash? 00:10:35 BZFlag it should be easy if the kernel uses modules. 00:10:43 scanline then we could have a .tuxphone, and whatever else needs to store preferences 00:11:22 SpaceCoaster BZFlag: I saw something that said there was quite a runtime hit for modules on ARM, minimum size of 32k per module 00:11:25 BZFlag the kernel is large now, cause the mtd partition is large. So I configure everything statically. 00:11:48 BZFlag SpaceCoaster: wow, I had not heard that. 00:13:25 SpaceCoaster BZFlag: oops arm2 and arm3, arm6 and up 4k 00:13:43 BZFlag 4k per module? 00:13:58 SpaceCoaster However, arm6 and up machines allow modules to take multiples of 4k, and 00:14:04 SpaceCoaster as such Acorn RiscPCs and other architectures using these processors can 00:14:09 SpaceCoaster make good use of modularisation. 00:15:14 BZFlag cool. could work then. just blast the kernel up there and teach blob to load it. save up to 512k or raw flash for jffs2 to use. 00:15:26 BZFlag s/or/of 00:15:28 SpaceCoaster BZFlag: probably 2k per module (1/2 a page). Not bad. 00:15:47 SpaceCoaster BZFlag: probably easier to put the modules in the serial flash? 00:16:05 BZFlag no, cause they need to be accessed at run time. 00:16:22 BZFlag the bootloader loads the kernel, and then nothing touches it. 00:16:44 BZFlag so put it out there, and free up the main flash. 00:17:18 SpaceCoaster I see, plenty of room in the bootloader area for fancy code to do that. 00:17:32 BZFlag we are talking about 13% of flash here. ;-) 00:17:54 SpaceCoaster Sorry which is 13%? 00:17:54 BZFlag SpaceCoaster: jffs2 likes equal size erase blocks, so that means 128k. 00:18:24 BZFlag that means the 128k at the front (32 + 16 + 26 + 64) is all available for blob 00:18:39 BZFlag 512k is 12.5% of 4M (main flash) 00:19:33 SpaceCoaster BZFlag: is their any magic I should know about for JFFS2 does it require a partition map? 00:19:37 BZFlag er s/26/16/ on the math line... 00:20:10 andersee BZFlag: perhaps we could get the kernel and blob into the serial flash.... 00:20:21 BZFlag SpaceCoaster: no partition required. it just uses all of a mtd device. the partioning (if any) is used by the mtd driver 00:20:42 BZFlag blob can't be in serial flash. execution begins at 0x0 00:21:11 BZFlag which is the first flash block (32k) 00:21:42 andersee hmm. 00:21:43 SpaceCoaster Ok, so when I hack things to allow individual bytes to be read I should be cooking. How do I put JFFS2 onto the serial flash, dd onto the mtd device? 00:21:44 BZFlag serial flash in on in the cpu address space 00:22:29 BZFlag SpaceCoaster: yes, or just erase all the flash and mount it as jffs2. if it is all 0xff then jffs2 will "format" it. 00:23:00 SpaceCoaster BZFlag: wow, that is what I wanted, the docs are thin for MTD and JFFS2! 00:23:19 BZFlag SpaceCoaster: agreed. 00:23:40 BZFlag course, andersee has mtd commit access, so send him doc patches. ;-) 00:23:52 BZFlag ducks 00:24:31 SpaceCoaster tho' David (mr MTD) has answered questions very quickly 00:24:47 andersee is away: I'm busy 00:25:14 andersee is back (gone 00:00:28) 00:25:21 SpaceCoaster BZFlag: you scared him 00:26:48 SpaceCoaster BZFlag: Ok i feel happier about, jffs2, just got to stop the flash driver from returning the odd byte of junk 00:28:53 SpaceCoaster BZFlag: I just realised, I take issue with your 12.5%, that's from what we started with, I want it calculated from what's left :-) 00:30:34 BZFlag SpaceCoaster: ahh, ok.. 00:31:19 11.1111111111111 00:31:19 BZFlag ibot 512 * 1024 * 100 / ( 4 * 1024 * 1024 + 512 * 1024 ) 00:31:35 BZFlag SpaceCoaster: how's 11% then? 00:32:53 SpaceCoaster BZFlag: you are drive a hard bargain. 00:32:57 SpaceCoaster s/are// 00:34:02 GPSFan BZFlag: SpaceCoaster: back from dinner, sorry for the screwup, I should have re-read the data sheet 4 instead of 3 times. ;>)) Any thoughts on the 1M fragmentation when using the mdcnfg params that let 32M SODIMM work? 00:34:11 SpaceCoaster BZFlag: who worked out the SA-1100 memory register settings for 8M? 00:34:32 BZFlag SpaceCoaster: Russ did on trial and error mostly. 00:35:33 BZFlag GPSFan: still thinking on that. I can get the kernel smaller than 1M compressed, but uncompresses would be tough. 00:35:46 BZFlag I might email rmk on that one. 00:36:16 SpaceCoaster BZFlag: can we load it at a different address, i.e. in the 16MB chunk? 00:36:33 GPSFan BZFlag: did you look at the sboot dis-assembly/ it looks like sboot setsup the MMU before booting Inverno. 00:36:57 GPSFan uhhh Inferno ;>) 00:37:05 BZFlag SpaceCoaster: well, it's not relocatable, so we want to do it in a way that allows for the different configs, no? 00:37:35 BZFlag GPSFan: yes, I expect Inferno does that. rmk is the one that thinks bootloaders should not do that. 00:38:04 SpaceCoaster BZFlag: so our bootloader doesn't setup the MMU? 00:39:13 GPSFan BZFlag: could it be a 2 stage boot, where the OS does fully setup the MMU for normal operation, but blob sets it just so as to be able to load & start up the kernel? 00:49:38 BZFlag SpaceCoaster: afaik, no linux arm loaders setup the mmu. 00:50:43 SpaceCoaster hmmm, to boldy go... 01:27:09 BZFlag home 01:33:16 andersee And X windows gurus? Whats the best place to start when trying to build (from source) a tiny X server? 01:33:48 SpaceCoaster andersee: sorry, I just used Xfbdev from familiar 01:34:34 andersee I guess I need to see how familiar builds theirs then... 01:35:13 SpaceCoaster It works fine, needs a stack of libraries tho' 01:35:41 andersee Ahh, found the HOWTO -- http://www.handhelds.org/minihowto/building-x.html 01:37:13 SpaceCoaster I remember when building X took all day and needed all the disks in the building 01:37:57 andersee With luck, this will go a bit faster... I'd like to see if I can get X into the flash on this thing... 01:39:31 andersee I wonder why XFree86 has never dropped the 86 01:39:40 SpaceCoaster The instructions look deceptively simple, tell me how it goes! 01:39:51 SpaceCoaster Maybe it's the year? 01:51:39 andersee I'm checking out the source -- its awfully big 01:52:37 SpaceCoaster andersee: trying to get a kernel into 1M, found out that virtual terminals are not an option :-( 01:53:10 andersee yeah, you sortof need those. 01:53:27 andersee Why 1M? 01:54:26 andersee What you want to try is disabling optional (PCMCIA) device drivers and filesystems 01:57:43 SpaceCoaster I think I've got it, just undefined ref to proc_ucb1200, do you from ucb1200_adc_init, any ideas? 02:09:13 SpaceCoaster hmm, the ucb1200_adc.c won't compile unless CONFIG_PROC_FS is defined, it uses seems to only create /proc entries. 02:09:40 SpaceCoaster -rwxrwxr-x 1 derek wheel 347772 Oct 26 23:11 tuxscreen-kernel 02:09:59 SpaceCoaster tiny tiny kernel! 02:13:57 SpaceCoaster it boots! 02:14:10 andersee very cool 02:14:35 andersee And that has the normal set of builtin drivers (fb, ucb1200, etc?) 02:15:02 andersee That one could easily fit in the serial flash. :-) 02:15:45 SpaceCoaster yes, no networking, pcmcia and /proc, doesn't boot on the revised blob memory mapping though:-( 02:26:57 andersee Ouch. 02:27:25 SpaceCoaster just six ....... away from Uncompressing Linux 02:27:27 andersee Tried adding back basic networking and /proc? 02:27:37 andersee just six? 02:27:53 andersee ahh 02:27:58 andersee uploading 02:29:05 SpaceCoaster correct, it needs to decompress into 1M to fit into the hokey memory map that we currently have with the 16M sodimms 02:29:43 AlHaz 32meg you mean? 02:30:00 SpaceCoaster oops yes. they map as 16+16, sorry 02:30:20 AlHaz *nod* 02:30:29 AlHaz so, lemme guess 02:30:40 AlHaz we have three banks of 16, then, except the first one is swiss cheese 02:30:44 AlHaz ? 02:31:13 AlHaz the way sa1100 addresses memory, I mean 02:31:29 SpaceCoaster yes, the first one is 16x1 on 2m (or maybe 4m boundaries) 02:31:42 AlHaz yech 02:34:28 SpaceCoaster how low can we go..no jffs2, no flash, no nothing ... 247512 Oct 26 23:36 tuxscreen-kernel 02:35:13 BZFlag won't get anywhere without mtd and jffs2. ;-) 02:35:42 SpaceCoaster just a proof of concept, will it boot in 1M 02:35:45 BZFlag but most of the rest can be modules. 02:35:57 BZFlag ahh, trying out memory. good plan. 02:36:05 BZFlag how small is vmlinux ? 02:36:15 SpaceCoaster 247512 02:36:27 BZFlag that's the gzippped version, no? 02:36:45 BZFlag linux/vmlinux ? 02:36:47 SpaceCoaster understand 692870 02:37:08 BZFlag ah. might run in 1M 02:37:35 SpaceCoaster the blob with the new mappings won't download it I have to switch blobs :-( 02:38:35 SpaceCoaster the kernel downloads and flashes quick! 02:38:50 SpaceCoaster Kernel panic: I have no root and I want to scream :-) 02:39:11 nelson thekernel: what are you using to render the tiger data into maps? 02:39:57 BZFlag SpaceCoaster: anyone have a dump of the blob output for a 40M config? 02:40:22 SpaceCoaster BZFlag: I got to Uncompressing Linux.................... done, booting the kernel. 02:40:31 SpaceCoaster but it doesn't run 02:41:03 BZFlag what does blob show for memory detection? 02:41:12 SpaceCoaster 1 second... 02:42:14 BZFlag waits 02:43:03 SpaceCoaster oops don't have the 32mb in, another moment please 02:43:57 BZFlag waits again. ;-) 02:43:57 thekernel was last seen on #tuxscreen 12 hours, 56 minutes and 29 seconds ago, saying: didn't I say I was leaving? bbl [Fri Oct 26 14:47:28 2001] 02:43:57 nelson ibot: seen thekernel? 02:45:11 BZFlag SpaceCoaster: and do you have the memory register values too? 02:45:28 BZFlag is writing the linux-arm-kernel list to ask. 02:46:32 AlHaz Should it also be possible to solder another 8 megs onto the main board? 02:46:36 AlHaz shaggy mentioned that was done to some 02:47:06 BZFlag AlHaz: I don't see why not. should work up to 32M on board, I would think. 02:47:39 AlHaz *nod* 02:47:51 andersee So -- with a theoretically mexed out board -- 32M on board + X in the 72pin SODIMM? 02:47:58 andersee err. maxed 02:48:10 BZFlag yep, 32 + 32 should work. 02:48:26 BZFlag if we reroute a12 then 64 + 64 might even work. 02:48:33 GPSFan BZFlag: the memory chip locations on the Tux are set up to support larger chips, problem is surface mount work isn't fun without the right tools. 02:48:51 BZFlag GPSFan: how much larger? 02:49:05 BZFlag you think 64 on board would work? 02:49:23 GPSFan BZFlag: haven't looked at that, the same chips that are on the 32M should work. 02:49:48 BZFlag GPSFan: yeah, I assumed as much given that the sodimm works. 02:50:05 BZFlag so that's the 32 + 32 that should be possible. 02:50:30 BZFlag we would not have the kernel boot issue either, as the pages would be larger ;-) 02:50:57 BZFlag wonders if SpaceCoaster is coming back anytime soon. 02:51:16 AlHaz I thought i saw more pads for chips on the board, I was mistaken . . 02:51:20 GPSFan BZFlag: address lines A10-A21 are routed corresponding to 12 bits 02:51:48 BZFlag routed to what? 02:52:16 BZFlag a0-a11 are routed to the sodimm right? I assumes that the high bits were set to known states. 02:53:06 BZFlag anyone else have the blob patches to see the 32M? I'll take a patch even and try it here. 02:53:07 GPSFan to no connects on the current chips which correspond to higher addresses on larger chips, and yes there are pins on the board between the notches of the chips. 02:53:46 BZFlag ah, so in theory much larger flash chips could be installed, yes? 02:54:10 GPSFan larger DRAM for sure and prpbably larger flash as well. 02:54:12 4096 02:54:12 BZFlag ibot 2 ** 12 02:54:38 BZFlag ah, yes I meant ram... 02:54:57 GPSFan read section 10.3.1 of the SA1100 manual about DRAM addressing. 02:55:00 4194304 02:55:00 BZFlag ibot 2 ** 22 02:55:17 GPSFan esp. the "special note" 02:55:40 BZFlag don't have it in front of me, summary? 02:56:44 GPSFan the SA1100 was designed to support up to 12 address bits, but 13bits can be supported as well. 02:56:59 GPSFan in a 13 x 8 arrangement. 02:58:30 GPSFan SPECIAL NOTE: At RAS time, all address pins, A(25:0), are driven with the internal address that corresponds to the 02:58:30 GPSFan pin of the same number. This means that a DRAM with 13 bits of row address may be accommodated simply by 02:58:30 GPSFan hooking up the 13th row address line of the DRAM to SA-1100 address pin A22 (MDCNFG:DRAC is a don’t care). 02:58:30 GPSFan However, the column address in this case will be limited to a maximum of 8 bits. 02:58:31 GPSFan In general, DRAMs that utilize less than 8 column address bits can be used, but there will be holes in the memory map 02:58:34 GPSFan due to no physical memory being addressed by the still significant internal address bit IA9. 02:59:21 GPSFan BZFlag: that was a cut&paste from the spec, I realized I couldn't summerize it properly. 03:00:04 BZFlag just fond and read it. 03:00:32 GPSFan BTW, the 13th bit, A22 is NOT connected on our DRAM. 03:00:55 BZFlag with the 32M sodimm we are setup for 12x10 in all four banks? and 2 map to the sodimm? 03:01:22 GPSFan correct, and the onboard 8 fragments into 8 1M chunks. 03:01:49 BZFlag ok. 03:02:22 GPSFan the 2 bits that used to be in the CAS address are now in the RAS address. 03:03:14 GPSFan and fall off the end of the world because our chips expect only 10 bits of RAS not 12 03:03:56 BZFlag right. but they do get mapped later, just fragmented. 03:04:28 BZFlag so all 40M is available, just that we can't boot the kernel. ;-) 03:04:45 GPSFan correct as well. 03:05:20 andersee GPSFan: ever looked at the memory mapping on a NetWinder? 03:05:25 GPSFan because the kernel needs to run to set up the MMU, and 1M isn't enough. 03:05:32 GPSFan Never have. 03:05:40 BZFlag andersee: how's it setup? 03:05:58 andersee bummer. I'm wanting to hack mine to get a 256 Meg SODIMM to work 03:06:06 BZFlag hmm... the netwinder uses the kernel to boot the kernel, yes? does it setup the mmu or not? 03:06:54 andersee The NetWinder takes 144 pin SODIMMs, but apparently it has them wired oddly, since the SODIMM has a couple of wires on it to make it talk... 03:07:30 GPSFan ins't the NetWinder an SA110? 03:07:43 BZFlag GPSFan: yes sa110 03:07:46 andersee BZFlag: I'm not certain. It does use a majorly hacked up linux kernel for the boot loader (which can read ext2) 03:07:50 GPSFan s/ins't/isn't 03:07:50 andersee GPSFan: yup 03:08:17 GPSFan does it have an on-board memory controller, or is that in its support chips? 03:08:23 SpaceCoaster GPSFan: hi 03:08:33 GPSFan SpaceCoaster: howdy.. 03:08:47 BZFlag SpaceCoaster: welcome back. rebooting your tux kicked your dsl line? 03:09:18 SpaceCoaster BZFlag: yes, i turned the tux over, disconnected the handset, phone off hook disconnects DSL! 03:09:20 andersee GPSFan: I'm not sure. I can pull the RAM from it and show you pics of the mod they made so it could address it all... I just bought a 256 Meg SODIMM in the hope of replicating the hack... 03:09:56 BZFlag SpaceCoaster: ahh! ;-) 03:09:59 GPSFan andersee: thanks I'd like to see the pix. 03:10:21 BZFlag SpaceCoaster: you should put the DSL modem on the other side of the tux anyway. ;-) 03:10:40 SpaceCoaster latest kernel figures 220764/621676, loads but doesn't run, maybe refresh ras/cas 03:10:54 BZFlag oh, wait... I take the phone off the hook all the time and my dsl says up... ;-) 03:11:21 BZFlag SpaceCoaster: got that boot text for me? (now that I sent the email off) 03:11:25 andersee I tried installing the 256, but the NetWinder can only see part of it -- I think it was probably writing to all 4 banks 03:11:34 SpaceCoaster BZFlag: at the moment the tux runs into through the extension line on router, is the other better. 03:12:00 SpaceCoaster yeah, I pasted it earlier 03:12:06 SpaceCoaster 0x00100000 @ 0xC0000000 (1 MB) 03:12:11 SpaceCoaster 0x00100000 @ 0xC0400000 (1 MB) 03:12:15 SpaceCoaster 0x00100000 @ 0xC0800000 (1 MB) 03:12:15 andersee GPSFan: gonna be here for a while? I can go disassembe my Netwinder, take picks, then get back online (my netwinder is my external router) 03:12:23 SpaceCoaster 0x00100000 @ 0xC0C00000 (1 MB) 03:12:25 SpaceCoaster 0x00100000 @ 0xC8000000 (1 MB) 03:12:29 BZFlag I run [outside]--[cisco675]--[tux and other phones in house] 03:12:29 SpaceCoaster 0x00100000 @ 0xC8400000 (1 MB) 03:12:34 SpaceCoaster 0x00100000 @ 0xC8800000 (1 MB) 03:12:37 SpaceCoaster 0x00100000 @ 0xC8C00000 (1 MB) 03:12:42 SpaceCoaster 0x01000000 @ 0xD0000000 (16 MB) 03:12:45 SpaceCoaster 0x01000000 @ 0xD8000000 (16 MB) 03:13:07 GPSFan andersee: don't rush, I'm off to Zzzzzzzz in 10-15 min's.. 03:13:13 SpaceCoaster I have a cisco 678, ok 03:13:47 andersee GPSFan: I'll go do it now anyways. 03:13:51 SpaceCoaster GPSFan: I think we have a kernel in 1m with mdcnfg:.long0x00c8993f/* test 1 */ 03:14:07 andersee Back in a bit.... 03:14:10 SpaceCoaster It might be the refresh which stops it from running? 03:14:44 SpaceCoaster BZFlag: is that the info you wanted, going to be soon.... 03:15:42 GPSFan andersee: I found an old Intel block diagram, it looks like the SA110 has its own memory controller. 03:16:26 GPSFan andersee: I'll wait for you to get them up... 03:16:51 BZFlag SpaceCoaster: yes. added to: http://TuxScreen.net/wiki/view/Main/SODIMM 03:18:14 andersee shutting down my netwinder now.... back in a few... 03:18:33 BZFlag cya 03:18:51 BZFlag quick, ping flood him 03:19:05 BZFlag grins 03:21:06 GPSFan SpaceCoaster: the refresh interval in the test 1 mdcncfg is different from the original blob, but uses the original blob waveform settings. 03:23:36 GPSFan the blob 8 + 8 interval was 19a clocks, Inferno 8 + 8 was 193, and Inferno 8 + 32 was 64. 03:25:52 SpaceCoaster GPSFan: ok 03:29:14 SpaceCoaster goodnight all 03:29:25 GPSFan SpaceCoaster: 19a and 193 are not that far appart, but the doc's give a big formula to calculate it, mabe changing to 0x0326993f would help 03:29:35 GPSFan woops... 03:37:25 andersee GPSFan: http://codepoet.org/ram/ 03:37:51 andersee GPSFan: The pics are pretty lousy -- my camera it seems doesn't like closeups.... 03:38:15 GPSFan need closeup lenses... 03:39:43 andersee As you may or may not be able to see, they hacked the SODIMM by adding wires to 1/2 of the chips. 03:40:22 GPSFan I can almost see, did you writ down what the part #s were on the chips? 03:40:56 andersee Presumably, it is only able to address 1/2 the ram directly, so they added wires to presumably do bank switching... 03:41:36 andersee I didn't but I can easily get the numbers -- without powering it down even.... 03:41:48 andersee brb with the part numbers 03:42:27 GPSFan good, I can then look up the # and see what type of chips they are, btw what pins are the wires connected to? 04:00:17 andersee GPSFan: k, I'm back 04:00:43 GPSFan hi, been looking at netwinder.org... 04:00:46 andersee Lemme annotate a couple of the pics for you. That will be easiest. 04:00:52 GPSFan k 04:00:57 andersee Find anything? 04:03:51 GPSFan 2 things, first the blockdiagram i mentioned before was abit too high level, I found the SA110 tech ref manual, which has NO DRAM memory controller. ie. it is external, second the netwinder.org people have had success up to 128M. 04:04:56 GPSFan it lloks like you get 1/2 the SODIMM memory and 256M dimms give you 128 (without hacking) 04:05:07 GPSFan s/lloks/looks 04:17:30 GPSFan andersee: still alive?? 04:29:33 andersee yup 04:29:40 andersee just finised annotating my pics 04:29:47 andersee GPSFan: http://codepoet.org/ram/ 04:30:24 GPSFan that's great! 04:32:25 andersee The 256 Meg SODIMM I plan on hacking is identically configured (8 chips with 27 pins each), but has chips labeled: 04:32:30 andersee SAMSUNG 133 04:32:52 GPSFan have you put it in and gotten 128M? 04:33:06 andersee K4S560832C-TC75 04:33:41 andersee Not this one. I've tried some others and they were only partly recognized (some banks missing) 04:34:52 andersee Anyways, this is how the ram on the NetWinder was hacked to make it work 04:35:18 GPSFan hmmm the parts are Hyundai 64Mbit sync DRAM chips so the memory controller on the netwinder should be able to handle up to 256M. 04:35:32 andersee Apparently (before going out of business) the rebel.com folks used to sell hacked up 128s. 04:35:34 andersee Right. 04:36:27 andersee The firmware prints a message on boot (via serial) mentioning how it is checking for 256 Meg in 4 banks. Which is why I bought the 256M SODIMM. :) 04:37:29 GPSFan what size is the memory in the normal configuration? 04:37:51 andersee 64Meg (what I have). Here is what I know (from one of the formar netwinder guys): 04:37:56 andersee Okay I have the full scoop now. What they did is basically double the 04:37:56 andersee number of banks on a module, while halving the width of the bank. Most 04:37:56 andersee SODIMMs have a single chip select (CS) line that runs to all of the 04:37:56 andersee individual chips. We cut this trace in half and run the second half out 04:37:57 andersee to the next CS line (there can be up to four CS lines). Our memory 04:37:57 andersee modules are made by Shikatronics and have a 0k resistor that can be 04:37:59 andersee moved to route the CS signal either the way we need it, or the regular 04:38:01 andersee way. 04:38:03 andersee So, what you observed makes perfect sense. Only half of the memory is 04:38:05 andersee visible if you don't make the change to the CS line. In fact what 04:38:07 andersee happens is that the same data gets written to _two_ places in your 04:38:09 andersee memory. If you inserted say a 128MB module then the lower 64M and upper 04:38:11 andersee 64M would in fact be mirrors of each other. If you cut the CS trace and 04:38:13 andersee route it to the SODIMM connector, then NeTTrom will recognize it as the 04:38:15 andersee proper number of banks and you'll get the full capacity. 04:39:13 andersee But, lacking the needed hardware knowhow... my project has languished... 04:39:42 GPSFan wow, that sounds like a bug in the netwinder design, that never got fixed. 04:41:48 GPSFan I think that if you are to succeed with your 256M dimm that it must work as a 128 without being modified. then you could research the chips, and solder away. 04:42:45 andersee Apparently they can address 4 banks... 04:43:13 andersee So I should be able to make it work with all 256M 04:43:52 andersee I know that the guy I was emailing with has a 256M machine... 04:44:01 GPSFan I think that it is possible if it works as a 128 without hacking. surface mount rework is not for the faint of heart, but it really isn't that hard if you have a good stereo zoom microscope and a steady hand. 04:44:03 andersee So I _know_ it can be done 04:44:24 GPSFan what type of dimm does he have. 04:44:50 andersee I wish I knew the details. Unfortunately, he wasn;t a hardware guy either 04:44:59 andersee And since the company has gone belly up.... 04:45:33 GPSFan so he is no longer able to provide info on his 256M configuration? 04:46:18 andersee I can try to contact him again. Perhaps I can still reach him... 04:46:55 GPSFan if so and if he will read off the chip #s on his dimm there is a bette chance of success. 04:47:10 GPSFan I'm off to Zzzzzzz land lots to do tomorrow.. 04:49:44 andersee I just placed a few of my emails on my site 04:49:56 andersee That is basically the extent of what I know 04:50:09 andersee GPSFan: http://codepoet.org/ram/emails.txt 04:50:42 GPSFan link doesn't work, something about permissions... 04:50:44 andersee just fixed the perms 04:50:47 andersee sorry 04:55:00 GPSFan ok, I can see how they are doing it, again the first step is does it work as a 128. If so, and if the orginization of the dimm is the same, it should be possible. 04:55:30 andersee if it doesn't work? 04:55:49 andersee I think the last time I tried a 256, it registered as a 64 04:56:00 GPSFan then you have a nice dimm for a notebook, shouldn't hurt it. 04:56:17 as_ Nite all. 04:56:31 andersee Why would that kill things? 04:56:51 andersee i.e. make it not feasible? 04:56:57 GPSFan It might even be possible to make that work. 04:58:34 andersee I'll give it a try... 04:58:47 andersee There is one guy who tried and had success.... 04:58:49 andersee http://www.netwinder.org/cgi-bin/faq.cgi?req=show&file=faq01.007.htp 04:59:00 andersee got a 256 to work as a 128... 04:59:16 GPSFan if it registers as a 64 then it is something different from what netwinder did, that's all there is always an element of risk, such as on the Tux, who could have figured that getting the 32 module to work would result in the onboard memory becomming fragmented. 04:59:55 andersee indeed 05:00:03 andersee pretty wierd stuff 05:00:09 GPSFan yep, 05:00:21 GPSFan time for bed goodnight all.... 05:00:32 andersee night 05:00:34 andersee thanks! 05:00:41 GPSFan sure, anytime... 05:43:28 Lenolium *plong* 07:12:51 andersee BZFlag: still around? 07:13:02 BZFlag yep. sup? 07:13:20 BZFlag no luck hacking your sodimm yet, huh? 07:13:30 andersee What do the A12 and A13 lines do? 07:13:59 andersee I've started putting together a webpage on the subject -- http://codepoet.org/ram/ 07:14:02 BZFlag what do they do? they are address lines. 07:14:14 BZFlag A0 is least sig byte and so on. 07:14:19 BZFlag er bit. 07:14:30 andersee Visit that webpage I just posted and see the side 2 pic 07:14:32 8192 07:14:32 BZFlag ibot: 2 ** 13 07:14:56 65536 07:14:56 BZFlag ibot: 2 ** 13 * 8 07:14:56 andersee I see. 07:15:24 BZFlag so A0-A12 will get you 64k? hmm. that does not seem right. 07:15:28 andersee That pic shows most of what I've figured out so far... 07:16:25 BZFlag wann bring that sodimm down and get a real picture taken? 07:16:39 andersee Maybe.... 07:16:42 4294967296 07:16:42 BZFlag ibot: 2 ** 32 07:16:59 262144 07:16:59 BZFlag ibot: 2 ** 13 * 32 07:17:33 BZFlag ah! yes. addressing on the arm bus is for 32 bit values, right? 07:17:39 andersee right 07:17:46 andersee A SODIMM is 64 bit wide 07:18:05 BZFlag that implies 07:18:09 524288 07:18:09 BZFlag ibot: 2 ** 13 * 64 07:18:35 BZFlag for A0-A12 07:18:55 BZFlag but your box only deals with 32 bits, right? 07:19:01 andersee right 07:19:28 BZFlag so you need to toggle the high bit with a bank select so that it treats that as an address line. 07:20:01 BZFlag or actually the reverse, use the cpu bank select as an address line. 07:22:30 andersee I just updated my page to reflect that comment. 07:24:54 BZFlag sounds logical. I think the 144 pin sodimm is just 2 copies of every pin, right? 07:24:55 andersee Presumably then, I could take any other SODIMM of similar configuration and with the same bit of wire hack, connect the Chip Select lines to A12 and A13, and thereby make things work.... 07:25:06 BZFlag the 72 pin does not have that duplication. 07:25:42 BZFlag andersee: should work. the trick is finding sodimms that have visible reroutable tracings. 07:26:32 andersee I wish I had better pictures.... 07:28:03 BZFlag has a better camera... 07:28:06 andersee If you read over the emails I have linked, it implies thay they cut half the chip select lines by placing a resistor somewhere 07:28:35 BZFlag let's presume there are multiple chips on board. 07:28:50 andersee presumably, the CS lines that are then wired to A12 and A13 07:28:54 BZFlag address 0 (ie A0-A10) are all set to 0 07:29:33 BZFlag the sodimm would have half the chips mapped to the upper and half to the lower bits on the 64bit value returned. 07:30:03 BZFlag now, the arm will never read the upper bits, so... 07:30:22 BZFlag we remove the chips select from those and only return the lower bits. 07:31:02 BZFlag no reattach the chip sellects for those upper bit chips to A13 and you should work. 07:31:45 BZFlag I would think the lowers need to be selected when a12 is off. 07:31:47 BZFlag hmmm 07:31:59 andersee which seems to be what is going on with the NetWinder RAM in the pics 07:32:38 andersee Read the emails I link to .. the one starting with "Okay I have the full scoop now" 07:32:57 BZFlag but you need an inverter no? or is there a chip select that is on normally that will be off when the other is on. 07:33:46 BZFlag ahh, that must be it. there is an existing bank select that will be enabled. use that one for the lower chips only and then attach the a12 to the upper bit chip selects. 07:33:47 andersee It sounds like they have resistors placed on all the CS lines. And then they tweek the resistor as needed. 07:34:39 andersee And then A13 does what? 07:34:52 andersee It is connect to the CS on another bank... 07:35:06 andersee So they have 3 CS lines to twiddle 07:37:06 BZFlag hmm... if you have 2 CS lines, you would need to wire up 1/4 of the chips to each logical state. that's more complicated. 07:37:25 BZFlag unless the sodimm is already doing 1 CS line and you are just adding the other. 07:37:58 BZFlag ditto if the sodimm does 2 and you want to make it do 3 07:38:02 andersee I assume the SODIMM provides one. This hack seems to add 2 more. One on the front, one on the back 07:38:35 BZFlag if you have 64 bits and are creating 32, you only have one more address line to play with. 07:39:34 BZFlag hmm. me ponders fragmentation. 07:40:35 BZFlag no even then, if you are adding more than one CS as address line I can't think of how to do it without adding gates. 07:41:04 andersee But I assume they added those on the motheboard? 07:41:49 andersee Or perhaps they are assering CS on the two other banks concurrently using 2 lines (for some goofy reason)? 07:43:10 BZFlag no clue. trace the wires and label them based on the sodimm doc. that would help. 07:43:45 andersee did that. Thats what I did on those pics.... 07:44:39 andersee Side 1 -- 2 of the 4 chips have their CS lines wired to A12 07:44:52 andersee Side 2 -- 2 of the 4 chips have their CS lines wired to A13 07:45:10 BZFlag I mean label the pin names for the sodimm (from docs) and the ram chips (from docs) 07:45:27 andersee thats what I did. 07:45:40 andersee pin 19 on the ram chips is CS -- chip select. 07:45:50 andersee pin 70 on the SODIMM is A12 07:45:53 BZFlag what's on the exiting CS pins(s)? 07:45:57 andersee pin 72 on the SODIMM is A13 07:46:07 andersee exiting? 07:46:16 BZFlag pin 19 is the only CS? 07:46:23 BZFlag existing, sorry. 07:46:33 andersee lemme bring up the pdf (I have links) 07:47:18 andersee yup. pin 19 is the one and only CS line on the chips 07:48:23 andersee Presumably (if the pic would show it) I would also find that the CS traces for the rewired CS lines are cut.... 07:49:36 andersee I think the idea is this.... 07:49:38 131072 07:49:38 BZFlag ibot: 2 ** 12 * 32 07:49:58 andersee Right now, when it does a write to memory, it writes to 2 spots due to aliasing. 07:50:14 BZFlag ok 07:50:19 andersee By asserting CS, it suppressed one side or the other.... 07:50:29 andersee i kant tipe 07:51:09 BZFlag so if CS is not asserted you think that 1/2 the chips are still addressable? 07:51:22 andersee Yup. 07:51:38 andersee But then it asserts CS on those -- they disapper 07:52:11 andersee And it releases CS on one (or both) of the other banks and they show up instead 07:52:21 BZFlag I thought if CS was off, nothing on the chip would be addressable. 07:52:29 andersee I don't know why they are using 2 addesss lines though 07:52:33 andersee exactly 07:52:58 andersee But they CUT the original CS lines to half the chips.... 07:53:00 BZFlag so disconnect 1/2 the chips and wire thier CS to an address line. 07:53:11 andersee right 07:53:12 BZFlag why do two address lines? 07:53:23 andersee exactly my question. dunno why. 07:53:28 andersee Seems kindof silly to me. 07:53:32 andersee bad planning? 07:53:33 BZFlag me too 07:53:52 BZFlag you're sure that's how it's wired now? 07:53:57 andersee Perhaps to make it easier to mod the SODIMMs.... 07:54:26 andersee Since they solder a wire right to the connector's pad (the one that fits in the socket) 07:54:43 andersee I guess soldering 2 wires to a single pad would be painful.... 07:54:52 BZFlag so they get 1 128M bank and 2 64M banks? 07:55:01 andersee I guess so 07:55:20 andersee Or they may always assert A12 and A13 together... 07:55:41 BZFlag they are address lines.... 07:55:48 andersee so. 07:56:03 andersee They isn't how they are being used.... 07:56:16 andersee Not if they are wired to the CS lines 07:56:42 BZFlag that's 0x3000 07:57:01 BZFlag ah! 07:57:06 andersee I assume they are doing something sneaky in hardware to fake out the addressing. 07:57:11 BZFlag I see, the CS in on the main board. 07:57:26 BZFlag yes, they could be tied together there. 07:58:41 andersee They have some bit of hardware watch the magic address bit, and then they just use the A12/A12 CS lines to bank switch as needed. 07:58:55 andersee I think.... 08:00:16 BZFlag you could check resistance on the main board between a12/a13 with a low voltage ohm meter. ;-) or follow the tracings on the board. 08:00:50 andersee suppose I could try that 08:01:04 andersee I need a new battery for my multimeter.... 08:01:37 andersee Well, I'm off to bed I think 08:01:44 andersee thanks for your help! 08:02:16 BZFlag cya, one more phone to unflash tonight... 08:05:24 andersee night 08:07:16 BZFlag nite all 08:43:15 Tangent wakes up.... 14:30:28 thekernel Hey, quick question about the boards inside the unit 14:32:41 thekernel nevermind, got it off 15:40:47 Lenolium *pling* 15:51:29 thekernel Has anyone done any casemods to their tuxscreen 16:11:57 mallum thekernel: would like too, I guessing you could get it pretty small 16:13:26 thekernel Well, I'm looking at 9.5x7.5x1.75 in 16:14:13 thekernel I'm looking to endup with a tablet that I can use with my GPS 16:15:22 thekernel how is the keypad accessed? 16:25:48 mallum yeah Im think tablet too, may try and build something out of some plexiglass 16:58:56 BZFlag is the 40M still visible with a 11x configuration? anyone have a 16M sodimm? 17:57:57 SteveD does anyone have an "original inferno flash" to compare my md5sum: 31fc77a29bf6795ae8824c95e17276ac ? 17:58:12 Tangent SteveD : It just so happens that I do.. 17:58:28 Tangent SteveD : But my Windoze based md5sum will give different results 17:59:06 SteveD tangent: i'm copying it off twice, but its not a quick process @ 9600bps 18:00:22 BZFlag 71a4a94ffee8252b5a283964ba3aa95d is what I get. 18:00:46 BZFlag I think that the image changes a bit. something must be read/write on it. 18:00:59 BZFlag there are newer inferno images up on the site though. 18:01:05 Tangent 8a22faf068b358e204f9f4016a9bee2e inferno.current 18:01:14 SteveD well, can't imagine I'll want to go back, anyway... 18:01:15 Tangent 33a9eb3f22332ea938a60a6a8a358716 newinferno.img 18:01:36 Tangent SteveD : If you do... one of my Tuxs has the original inferno still in tact anyhow 18:03:03 Lenolium *plong* 18:10:14 BZFlag hmm.. actually I think the unit serial number is somewhere in the data. 18:10:39 BZFlag under inferno the ##314159265359** trick shows the serial number. 19:06:34 SteveD did anyone succeed to "c D!block F!all" from inferno under windows? on my system tux puts up copy bar, pc shows flash: line and then drops back to ';' prompt 19:14:57 SteveD reboots to linux... 20:17:08 SpaceCoaster andersee: did you manage to download X 21:03:52 Lenolium *pling* 21:04:06 BZFlag *plung* 21:04:35 Lenolium Hehe. :) 21:37:07 SteveD ; bind '#'t /dev 21:37:12 SteveD bind: no free devices 21:37:17 SteveD can anyone help? 21:40:58 GPSFan SteveD: the user needs read/write permission for the serial ports. 21:41:43 GPSFan SteveD: there are details on the wiki. 21:45:09 SteveD gpsfan: thought i'd done that - checking... 21:46:11 SteveD gpsfan chmod g+rw /dev/ttyS* /dev/cua* - that's better 21:48:33 GPSFan SteveD: I had the same problem set as you when trying to upgrade my phone, 1st rdp under win98 would not send data to the phone, it would receive just fine, but I could never download anything to the phone. I switched to Linux, and experienced the no free devices problem, I finally got it to work, after creating the proper user as per the Inferno under Linux install instructions. 21:48:57 torquil hi 21:49:07 torquil so whats the current progress of the tuxscreen 21:49:09 torquil i was confused 21:49:15 scanline hi 21:49:20 SteveD ok - about to copy blob over - cross your fingers... 21:49:23 torquil like can i get on the internet with it yet? :) 21:49:50 scanline torquil: we have linux, several GUIs, ethernet, DSP support... 21:49:57 torquil scanline: cool 21:50:10 scanline torquil: Unless you run X, we don't have a working web browser yet 21:50:15 torquil does it have IDE or? 21:50:36 torquil once you have armlinux loaded... can you compile basically anything for the arm? 21:50:36 scanline It only has 3.3v PCMCIA 21:51:29 scanline well, there are memory limitations 21:52:34 torquil so its not really a usable standalone computer yet? 21:53:06 scanline well, what do you want to use it for? 21:53:24 torquil irc and general console linux stuff 21:53:36 scanline it should run most console stuff 21:53:47 torquil you gotta use a PCMCIA HD? 21:54:08 scanline It's a linux box with 16MB of ram and 4MB of flash for jffs2 21:54:24 torquil oh ok 21:54:25 torquil now i see 21:54:26 scanline with a PCMCIA HD you could run anything that would usually run in 16MB of ram 21:54:27 torquil thanks 21:54:40 torquil i only wish i had more than $33 left on my CC :) 21:54:56 SteveD hey - I'm LARTed. :-)) 21:55:22 GPSFan SteveD: congrat's !! 21:56:33 SteveD scanline: you can upgrade ram to max 40M with the right dimm from crucial, right? 21:56:41 SteveD (hope so - since I just ordered one) 21:56:49 GPSFan BZFlag: you around? 21:57:06 scanline SteveD: I dunno, I haven't been following the ram upgrade work 22:01:31 GPSFan SteveD: check the wiki for the latest on the 32M SODIMM issue http://tuxscreen.net/wiki/view//SODIMM 22:02:16 torquil i dunno 22:02:22 torquil the tuxscreen would be great 22:02:31 torquil but i get like 1 phone call a month :) 22:11:06 bhima is there any way to flash a tuxscreen via a PCMCIA card with stuff on it? or over ethernet? 22:11:14 AlHaz not yet 22:11:35 BZFlag GPSFan: here, sup? 22:12:03 BZFlag SteveD: 40M does not yet work in Linux. 22:12:29 BZFlag the 32M fragments the memory such that the kernel will not load/boot. 22:12:38 BZFlag could be fixed with a custom kernel. 22:18:16 GPSFan BZFlag: hi, just wanted to ask your opinion of the Lucent 5V PCMCIA mod. 22:19:29 BZFlag have not tried it yet. of course it would be better if it auto switched instead of hard wiring it, but at least with hard wiring it will work for those with 5v nics. 22:20:15 GPSFan bz, take a look at the 2 sketches, and D2. notice that it connected differently in each sketch. 22:21:05 GPSFan BZFlag: if you wire it up like the schematic on top, both D2 and T1 will fry. 22:21:11 AlHaz yeah, Kurt mentioned something about that. There's also a transistor that's PNP one place, and NPN another place 22:21:55 GPSFan both my sketches have it as a PNP which is correct for tha application. 22:22:12 GPSFan it provides a non-inverting buffer for reset. 22:23:25 GPSFan I guess the arrow in my bottom sketch printed out more like a pnp than npn, I can see where it could be taken for npn... 22:24:34 AlHaz GPSFan: Kurt's got one phone successfully modified already, we will prolly see a howto from him soon 22:25:08 GPSFan AlHaz: great, i guess Kurt is a hardware guy? 22:29:19 AlHaz GPSFan: Yup, EE by trade 22:29:37 AlHaz well, these days he's capering for SLOC . . . . . I think the company he was with went belly-up 22:30:07 GPSFan SLOC?? 22:32:07 AlHaz Salt Lake Olympic Committee 22:32:13 AlHaz his wife works for them 22:32:29 AlHaz so he has a part time job dressing up in a big foam rubber & nylon suit 22:32:39 AlHaz mascots, yaknow 22:33:02 AlHaz kurt's done some SMT soldering for me. He's got a real nice setup in his basement. microscope & everything. 22:33:14 GPSFan Ah... I guess that's comming up soon. Is there much EE work in SLC? 22:33:26 AlHaz Usually 22:33:37 BZFlag AlHaz: have Kurt update the wiki. ;-) 22:33:43 GPSFan really? all I knew about was Iomega. 22:33:44 AlHaz unisys is there, the former Mhz (now usrobotics? 3com? something) is there, etc 22:33:57 AlHaz DEC's got a facility there, as does National 22:34:06 bhima drools. mmm, SMT rework stations are fun. 22:34:10 AlHaz guess DEC -> Compaq -> HPaq? 22:34:27 AlHaz BZFlag: Me & duck keep bugging him to get more involved, he's got four phones after all 22:34:45 GPSFan I was wondering about your DEC reference...;>) 22:35:51 GPSFan with a SMT rework station there are lots of cool mods you could do to the Tux right on the board. 22:37:00 AlHaz I had some boards that needed a serial buffer chip added to 'em 22:37:09 bhima GPS: mmm? what were you thinking of doing? 22:37:22 AlHaz it woulda been a grueling three hour ordeal for me & my silly weller iron 22:37:31 AlHaz took Kurt about 40 seconds 22:37:45 BZFlag heh, get em to build a serial port for your i-opener... and mine too. ;-) 22:37:59 GPSFan a good stereo zoom & a pace is hard to beat. 22:38:15 bhima I saw a big box with hundreds of iOpener mice at a computer show around here a few months back. 22:38:51 GPSFan bhima: those are nice USB mice did you get some? 22:39:07 bhima gps: I have enough mice, I didn't buy any. 22:40:22 bhima I should have bought an iOpener, but I didn't have any money at the time they were around. 22:40:30 GPSFan bhima: oh, I can never seem to have enough mice... What you can do with a SMT station is change the DRAM to bigger parts, do the PCMCIA mod in the I/O board where it belongs, add a real software reset... 22:41:15 GPSFan iOpeners are fun, I've converted 2, and have 2 spares. 22:41:40 bhima GPS: ooh. I should replace the DRAM chips. I already did that on an iPaq, so now I have one with 128MB of DRAM. 22:42:33 GPSFan Midori Linux works great in the iO, and the 800 x 600 screen is sice too. you have a 128M iPaq? cool. 22:42:36 AlHaz i bought my iopener at a thrift store for $20 about four months ago 22:42:48 AlHaz and found the power supply for it today 22:42:51 AlHaz for $2 22:42:53 AlHaz same store . . . 22:42:55 bhima alhaz: damn. that's cool. 22:43:22 GPSFan all those iOpeners that were sold to "real" customers are now door stops.. 22:43:27 AlHaz yep 22:43:36 AlHaz well, some of 'em might have converted to earthlink 22:43:41 AlHaz but I'm not sure even that would work now 22:44:03 AlHaz I'd heard that IO's could work with 256 meg sodimms, so i went and bought one 22:44:06 GPSFan they do make nice win98 and/or Linux boxes though. 22:44:21 AlHaz IO with badflash bios in it detects it as 128 megs, and linux barfs all over itself 22:44:29 AlHaz :/ 22:45:27 GPSFan well mine with 32M may be a bit slow, but it does what I need. I run it as a GPS nav system. 22:45:37 AlHaz Anybody need a 144 pin pc100 256 meg sodimm? 22:45:39 AlHaz Cheap. 22:45:55 GPSFan how much? 22:45:57 bhima alhaz: what brand? how cheap? 22:46:15 AlHaz bhima: unknown brand . . . I could see what make the chips are 22:46:27 GPSFan that's the important part. 22:46:34 AlHaz I'll ship it for $25 22:46:36 AlHaz cost me $30 22:46:40 GPSFan sold\ 22:47:35 AlHaz GPSFan: You're in luck, the chips are Micron 22:47:42 GPSFan bueno!! 22:48:28 bhima I just put some Micron-made 256MB registered ECC DIMMs in my Athlon. 22:48:30 bhima those are _huge_ chips. 22:50:26 bhima hey, where is that mysterious parallel port on the tuxscreen? 22:50:38 bhima I've seen it referenced, but not in much detail. 22:51:06 AlHaz J602, iirc 22:51:33 bhima what can you do with it? 22:51:42 AlHaz I don't think it's really a parallel port 22:52:07 Tangent as_ : Morning 22:53:19 as_ Evnin Tangent. 22:53:26 AlHaz but i don't see any reason why you couldn't adapt a parallel port device to connect to it 22:53:30 AlHaz but how would i know? 22:54:05 as_ To connect to what? 22:54:14 AlHaz as_: j602 22:54:24 AlHaz unpopulated 36 pin header 22:55:25 as_ Yikes, Ive no idea what that is ;] 22:55:42 AlHaz it's hooked up to the sa1100 through buffer chips 22:55:47 GPSFan j602 has lots of usefull stuff, you could all sorts of things 22:55:48 AlHaz but i don't know what it's hooked up to 22:56:21 GPSFan look here... http://tuxscreen.net/download/schematics/j602.pdf 22:56:44 GPSFan off to dinner... 22:58:20 bhima thansk :) 23:07:07 Tangent decides that it's Tuxscreen time 23:08:29 bhima how is the serial port wired? can you use a standard cpu->device cable? 23:09:15 Tangent bhima : Regualar PC Crossover cable 23:09:24 Tangent erm.. null modem 23:09:31 BZFlag only pins 2,3 and 7 required. 23:09:52 bhima null modem and crossover are equivalent enough :) 23:15:03 Tangent: excuse me? 23:15:03 Tangent ibot buildroot 23:18:24 Tangent could someone enlighten me to the CVSROOT I need to checkout buildroot? 23:34:16 mallum Tangent: see the sourceforge project site, under cvs 23:34:25 Tangent sourceforge is down right now 23:34:36 Tangent for maintenance 23:36:32 Tangent sadly it's gone from my bash history now too 23:38:03 mallum cvs.tuxscreen.sourceforge.net:/cvsroot/tuxscreen ? 23:38:42 Tangent I'll give it a try.. thanks 23:41:21 OK, Tangent. 23:41:21 Tangent ibot Tuxscreen CVS is CVSROOT=:pserver:anonymous@cvs.tuxscreen.sourceforge.net:/cvsroot/tuxscreen 23:42:03 Tangent makes world 23:47:33 tuxscreen cvs is mailto:CVSROOT=:pserver:anonymous@cvs.tuxscreen.sourceforge.net:/cvsroot/tuxscreen 23:47:33 andersee ibot tuxscreen cvs? 23:47:47 andersee heh 23:48:02 Tangent Hrm... 23:48:08 Tangent: I forgot tuxscreen cvs 23:48:08 Tangent ibot forget tuxscreen cvs 23:48:16 OK, Tangent. 23:48:16 Tangent ibot Tuxscreen CVS is <reply> CVSROOT=:pserver:anonymous@cvs.tuxscreen.sourceforge.net:/cvsroot/tuxscreen 23:48:24 mailto:CVSROOT=:pserver:anonymous@cvs.tuxscreen.sourceforge.net:/cvsroot/tuxscreen 23:48:24 Tangent ibot tuxscreen cvs? 23:48:30 Tangent Duh Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with infobot logs, split per channel and by date, etc.