**** BEGIN LOGGING AT Thu Mar 29 02:59:56 2007 Mar 29 03:00:20 that would be from DI Mar 29 03:00:26 change it to what I have above Mar 29 03:00:36 the rtc stuff is done automatically by the slugos kernel Mar 29 03:01:26 ok gotcha. brb gonna do the flash upgrade, then do the apex env mod Mar 29 03:01:58 rwhitby, hdparm Mar 29 03:02:04 hdparm -T Mar 29 03:02:52 jacques: what about without a disk attached? Mar 29 03:03:13 does it matter about disk attached? Mar 29 03:03:24 we want to measure memory only, not disk. Mar 29 03:03:58 -T does measure memory Mar 29 03:04:15 tho hdparm might want a device to look at Mar 29 03:04:22 ok good catch - measures device cache. Mar 29 03:05:27 Create a ramdisk and dd several MB of data in and out of it? Mar 29 03:05:50 /media/ram should be such a device, no? Mar 29 03:06:56 rwhitby I see that cmdline is different now so I just have to add the mem=16MB. Is that needed now - or should I just go for 256MB Mar 29 03:07:26 go for 256MB first Mar 29 03:07:28 why 16MB and not 32? Mar 29 03:07:44 also, do I need to run the parameters up front, like "sdram-init" and "memscan -u 0+256m" Mar 29 03:08:36 jacques: that was so I could see if it was recognised or not Mar 29 03:09:13 root@grouch:~$ time dd if=/dev/zero of=/media/ram/x bs=32768 count=970 Mar 29 03:09:24 31784960 bytes (32 MB) copied, 1.57448 seconds, 20.2 MB/s Mar 29 03:09:27 mwester: good idea Mar 29 03:09:35 think I should use a comma before mem=256M ? Mar 29 03:09:47 no Mar 29 03:09:54 space before it Mar 29 03:10:02 ah ok sorry gotcha - makes no sense to do so Mar 29 03:10:05 the comma that is there is part of the console attribute Mar 29 03:11:24 ok, I gotta go do a server outage. back in about 30 mins Mar 29 03:11:25 yeah I saw that (lack of sleep) Mar 29 03:12:58 seems like it is stuck at ".....done, booting the kernel." Mar 29 03:13:08 lemme try 16M to ensure I have the params right Mar 29 03:13:40 RobNC: pastebin your slugimage commands to build the image you flashed Mar 29 03:13:59 k Mar 29 03:16:03 mwester: that count is too big for 32MB Mar 29 03:16:31 That fit in my dsmg600's default ramdisk. Mar 29 03:16:35 But only barely. Mar 29 03:16:53 Ah -- the dsm has 64MB ram --- sorry! Mar 29 03:17:09 heh Mar 29 03:17:13 as does the lude Mar 29 03:17:43 Hmmm... yes, and rwhitby has a lude, probably running the same slugos that I'm running. Would be interesting to compare numbers. Mar 29 03:18:19 * mwester goes to see if he has a bootable slugos NSLU2 about... Mar 29 03:19:06 I need to take the serial adaptor off the slug I took to SCALE and put it back on the JTAG'd one Mar 29 03:20:48 http://pastebin.ca/414585 Mar 29 03:22:33 that's a good idea u guys use header and serial adapter. I got MAX3202 and flipped it, glued it, wired, decoupling caps, and soldered cable permanently to SLUG. Is there a cheap (~10) cable that can be hacked to use RX/TX/GND for serial port? Mar 29 03:22:39 jacques: I bought one of these; it worked: http://cgi.ebay.com/USB-Data-Cable-for-LG-5200-3200-1200-CD_W0QQitemZ160099123895QQcategoryZ50327QQssPageNameZWDVWQQrdZ1QQcmdZViewItem Mar 29 03:23:36 is the picture correct? where are the electronics? Mar 29 03:24:26 mwester: I'm testing this on a slug Mar 29 03:25:20 The picture is very close to the actual item. Electronics on mine were in the USB connector. Here's a link to the actual item I got: http://avr.auctionant.de/ark3116_linux_driver/pictures-inside.html Mar 29 03:25:22 rwhitby don't read webpge just yet - I made mistake - didnt yet subst zimage Mar 29 03:25:53 root@nslu2:/media/ram$ time dd if=/dev/zero of=/media/ram/x bs=32768 count=470 Mar 29 03:25:54 470+0 records in Mar 29 03:25:54 470+0 records out Mar 29 03:25:54 real 0m 0.72s Mar 29 03:25:54 user 0m 0.00s Mar 29 03:25:54 sys 0m 0.72s Mar 29 03:26:19 rwhitby: I just tested on a slug as well. Moving 400 * 32KB took 0.62 real seconds (0.59 sys). Mar 29 03:26:31 Let me try 470... Mar 29 03:26:52 0.72, 0.00, 0.72 Mar 29 03:27:29 ok, that matches then :-) Mar 29 03:27:58 anyone know how to access a Dell DRAC II from RHEL4 ? Mar 29 03:28:05 jacques: for the price, you can't beat it. It took a while to figure out the wires, but if you get one and it's identical, I'll be happy to share the color codes on mine. Mar 29 03:28:50 mwester: and you get a ringtones CD with it too! ;-) Mar 29 03:29:34 :) Yep! In Taiwanese, too (I think). Mar 29 03:29:48 (it also has the windows driver for the serial chip) Mar 29 03:33:03 hmm, bubble-chip Mar 29 03:33:22 which linux driver it use? Mar 29 03:33:36 ark3116 Mar 29 03:33:44 thanks Mar 29 03:34:21 I need more USB ports on my laptop Mar 29 03:34:56 hmm, maybe I can plug my 7-port hub into the hub in my monitor Mar 29 03:35:10 jacques, I had to replace my father's 6 year old Dell on Saturday... I was a bit jealous that the Toshiba that I got him had 4 USB ports Mar 29 03:35:32 my t42 only has 2 :-( Mar 29 03:36:06 same here (t42p..) Mar 29 03:37:00 I was happy that my IBM T42 (work computer) has two as well -- until I discovered that it has no native serial port or mouse port. There goes the USB ports, just to bring it up to what the Dell it replaced had built-in. Mar 29 03:37:23 I now have usb speakers and monitor hub plugged into laptop Mar 29 03:37:42 usb keyboard, mouse and 7-port hub plugged into monitor Mar 29 03:38:20 RobNC: how goes the battle? Mar 29 03:38:23 every port in the hub is filled, tho not all in use right now Mar 29 03:39:05 rwhitby: documenting steps - realized I missed some things in trying to scroll up/down the java IRC window Mar 29 03:39:26 reflashing. Will get you a paste URL soon Mar 29 03:40:21 sweet it is booting debian. didn't see the scrollby to see if it erased the 256M option and thus more mem Mar 29 03:40:54 weird, eth broke. can't get addres Mar 29 03:41:00 got brokenslug msg Mar 29 03:41:14 (D-I does dhcp and it works for me all the time) Mar 29 03:42:11 ah gotta change APEX env as scrolling back - the 256M option is gone Mar 29 03:43:01 U-Boot 1.1.4 (Jul 1 2006 - 02:22:45) Mar 29 03:43:01 Mar 29 03:43:01 CPU: Intel IXP425 at 266 MHz Mar 29 03:43:01 Board: NSLU2 - aka slug Mar 29 03:43:01 DRAM: 32 MB Mar 29 03:43:02 Flash: 8 MB Mar 29 03:43:06 ... Mar 29 03:43:12 now gotta get jtag-tools back up Mar 29 03:44:18 seems to still hang on mem=256M parameter Mar 29 03:44:28 gonna try smaller size to confirm it works on mine Mar 29 03:45:19 RobNC: it's booting slugos, not debian :-) Mar 29 03:45:36 ah ok gotcha so I should disconnect USBHDDs Mar 29 03:45:42 doesn't matter Mar 29 03:45:55 it's booting and running entirely from flash Mar 29 03:46:08 ok about to try 16M to confirm the parameter is being read correctly Mar 29 03:46:21 ok Mar 29 03:46:22 you sure I don't need the sdram-init program? Mar 29 03:46:32 not for 16, but probably for > 32 Mar 29 03:46:47 ok will try that next in the boot Mar 29 03:47:48 for some reason it sees eth0 link up but no discover response. weird. I am using that router now with this IRC so I know the router works :-) Mar 29 03:48:38 seems to boot into brokenslug prompt Mar 29 03:48:46 with mem=16M option Mar 29 03:50:32 Do I only change "startup" Apex env var? Mar 29 03:50:41 i.e., before the "copy -s ..." Mar 29 03:50:59 you could just run the sdram init commands manually, then "boot" Mar 29 03:51:20 ok try that just for kicks Mar 29 03:51:25 RobNC: dhcp is probably broken on that image Mar 29 03:51:49 dunno if memscan is needed Mar 29 03:51:55 you can log in with "root" and "opeNSLUg" and "turnup init" to set the network params. Mar 29 03:52:24 ok, jtag seems to work still Mar 29 03:52:30 yeah I forgot the l/p with openslug - been long time since running that Mar 29 03:53:15 jacques: how big is uboot for the nslu2? Mar 29 03:53:45 doing manual "sdram-init" then "memscan -u 0+256m", then "boot" gets NSLU2 freeze saying "Booting kernel at 0x00008000" Mar 29 03:54:12 but it had left in the param "mem=16M" in tail of startup (ATAG_CMDLINE) Mar 29 03:56:43 rwhitby, 256k Mar 29 03:57:11 including config area and npe microcode Mar 29 03:58:04 damn, that might mean config area is 128kB... Mar 29 03:58:19 which means if it were moved external to u-boot... Mar 29 03:59:44 oh rwhitby got some interesting info Mar 29 04:00:14 ... Mar 29 04:00:24 tried mem=256M no go. Tried mem=128M no go. Tried 64M it works. Now, it actually shows that if I login using openslug Mar 29 04:00:39 however, sdram-init I have not done. Mar 29 04:01:04 I have tried doing sdram-init, then 256M it hangs. sdram-init, then 128M it hangs. Mar 29 04:01:25 I know i have 256MiB b/c it is detected :-) Mar 29 04:02:04 did you see 256MB when running debian? Mar 29 04:02:06 (and there is no fuzzy math for 4 @ 32Mx16 != 256MB) Mar 29 04:02:12 detected by the bootloader or the kernel? Mar 29 04:02:30 jacques: kernel just displays what the ATAG says, right? Mar 29 04:02:38 can't remember. about debian. I think debian saw it, using the prev slugimage and ur apex bootloader Mar 29 04:02:53 rwhitby, I'm not sure Mar 29 04:03:17 real test would be (I suppose) detection and memtest in bootloader Mar 29 04:03:38 jacques: we're using apex here, and it's detecting 256MB Mar 29 04:03:43 jacques:I am unsure. hacking with mem=256M or mem=128M or mem-64M option with apex kernel option Mar 29 04:04:05 yes exactly that is what Apex1.4.17 is supposed to do Mar 29 04:04:30 yes rwhitby apex detects 256MiB correctly (using 1.4.17 only) Mar 29 04:04:32 does apex have a builtin memory test? Mar 29 04:04:50 to try at least writing to all memory (other than what bootloader is using) and reading it back? Mar 29 04:05:38 jacques dunno if apex has mem test. I was asking that yesterday actually Mar 29 04:06:38 CONFIG_BOOTSTRAP_MEMTEST Mar 29 04:06:45 CONFIG_CMD_MEMLIMIT Mar 29 04:06:52 CONFIG_CMD_MEMSCAN Mar 29 04:06:52 ah that sounds like a compile-time option Mar 29 04:07:09 yep, I can quickly do a new version with those enabled. Mar 29 04:07:18 sure if you don't mind and have the time Mar 29 04:07:35 config CMD_MEMLIMIT Mar 29 04:07:35 bool "Define Memlimit Command" Mar 29 04:07:35 default n Mar 29 04:07:35 help Mar 29 04:07:35 This command constrains amount of RAM available to Linux Mar 29 04:07:36 kernel. It is useful for testing how the kernel Mar 29 04:07:36 should I use d-i rc2 or openslug Mar 29 04:07:40 responds to a limited memory resource. Mar 29 04:07:42 config CMD_MEMSCAN Mar 29 04:07:44 bool "Define Memscan Command" Mar 29 04:07:46 default n Mar 29 04:07:48 help Mar 29 04:07:50 This command scans memory as APEX does when initializing. Mar 29 04:07:52 It is used primarily for debugging. Mar 29 04:07:54 same image, just replace apex Mar 29 04:07:58 or just tftp apex only and flash apex only Mar 29 04:08:22 true yeah Mar 29 04:09:56 is that 006000-007FFF? Mar 29 04:13:41 03rwhitby * r806 10kernel/trunk/patches/apex/ (enable-commands.patch series): Enabled memory commands Mar 29 04:16:03 RobNC: 0x60000 to 0x7ffff, right? (you missed a digit and would have overwritten redboot) Mar 29 04:16:27 0x50060000 to 0x5007ffff in fact Mar 29 04:16:45 RobNC: new apex with memlimit and memscan commands sent to you Mar 29 04:18:15 yes I see that now - obviously... looking at wiki it appears someone missed a digit. Mar 29 04:18:31 which wiki page? Mar 29 04:19:22 apex over xmodem I think Mar 29 04:20:23 maybe I mis-read... http://www.nslu2-linux.org/wiki/HowTo/ReflashUsingApexAndXmodem Mar 29 04:21:17 no, they've for four zeros there. Mar 29 04:21:23 s/for/got/ Mar 29 04:21:24 rwhitby meant: no, they've got four zeros there. Mar 29 04:21:28 ya u are right must have been somewhere else Mar 29 04:21:48 can you check your browser history? leaving off a zero is fatal. Mar 29 04:23:41 (unless you have jtag) Mar 29 04:24:31 ok Mar 29 04:26:12 leaving off a zero when paying one's taxes is often nearly fatal. Mar 29 04:26:50 * mwester goes back to installing "TurboTax" and gathering his paperwork... Mar 29 04:29:29 ok I looked and looked and can't find the bad page. I'll have to go back to that in a few. Mar 29 04:30:03 ok Mar 29 04:34:34 apex is so small. any need to fill to 7ffff? Mar 29 04:34:41 no Mar 29 04:38:25 trying to not use slugimage Mar 29 04:39:05 "load -r -v -b 0x01000000 -h TFTPSERVER filename" Mar 29 04:39:48 "fis write -f 0x50060000 -b 0x01000000 -l 0x0100acc0" Mar 29 04:40:00 right? says invalid flash address (?) Mar 29 04:40:40 says "valid range is 0x50000000 - 0x507fffff" but my range is inclusive heh Mar 29 04:42:15 guess I could always redo the whole 8MB but trying not to kill the flash Mar 29 04:42:47 0x50060000 + 0x0100acc0 = 0x5106acc0, which is > 0x507fffff Mar 29 04:43:28 whoops that is weird, how could it be that large Mar 29 04:43:52 did you specify the end address instead of the size? Mar 29 04:44:06 size I think, that is what -l does right Mar 29 04:44:19 Yes -l is the size, not the end address Mar 29 04:44:49 ah yep u got it I need to kill the 0x0100 prefix Mar 29 04:44:54 :) Mar 29 04:45:47 lack of sleep - prolly shouldn't be mucking with slug :-) Mar 29 04:46:25 ok gonna try memscan if it exists in apex Mar 29 04:49:58 well I have the latest svn u-boot to the same point I had in june Mar 29 04:50:27 jacques: what is that point? Mar 29 04:51:12 NPE doesn't work at all Mar 29 04:51:24 weird... I tried this and same thing both times. next gonna try slugimage to be sure. Mar 29 04:51:29 maybe I need to byte-swap the microcode or something Mar 29 04:51:58 flash works. I can update u-boot with u-boot Mar 29 04:51:59 I get this type of text: "$T0a0f:00020654;0d:0003dda4;#aa" Mar 29 04:52:56 RobNC: oh, yeah, you'll need to use slugimage, cause you probably won't have the 16 byte header on the front of that apex Mar 29 04:53:20 (I take all the little things that slugimage does for granted sometimes) Mar 29 04:53:56 ah ok whew! Mar 29 04:54:20 that text you saw is the redboot built-in gdb server Mar 29 04:54:32 bbiab (lunch) Mar 29 04:54:48 ah ok cool... Mar 29 05:08:42 back Mar 29 05:09:17 14 minutes for lunch? That's fast. Mar 29 05:09:17 still here (not sure for how long b/c 1am here) Mar 29 05:09:25 yeah true! Mar 29 05:09:34 yah really Mar 29 05:09:40 mwester: brought it back to the desk - late lunch due to the scheduled server outage to install the DRACII card Mar 29 05:09:54 Oh ok. Mar 29 05:10:53 rwhitby I don't see cmd memscan or memlimit in apex u built Mar 29 05:11:20 hmm - neither do I :-( Mar 29 05:11:25 this is after slugimage rebuild Mar 29 05:12:18 damn - somehow the patch didn't take Mar 29 05:12:31 sdram-init still shows 2 banks of 2 512Mib chips Mar 29 05:13:22 maybe I should check solder job? I doubt it, unless maybe I bumped it somehow (but doubtful) Mar 29 05:13:24 looks like I stuffed up. sorry. Mar 29 05:13:42 if it's late there, then perhaps another day. Mar 29 05:14:11 (tomorrow same time will be fine for me) Mar 29 05:14:45 I'll look at it under microscope again tomorrow. I did this mod in December. Mar 29 05:14:46 I'll make a new apex, and actually test the commands are there before sending it this time Mar 29 05:15:00 ok no problem, thanks so much for your help. Mar 29 05:15:03 did tigerdirect have a slug? Mar 29 05:15:35 hmm - maybe wait till we've got this one sorted out before creating another one? Mar 29 05:15:44 I would think if I had a solder bad job, the slug would never have lasted (since I ran rsync every 2 hours and it frequently used all RAM and even about 20M of swap). Mar 29 05:16:28 I didn't get chance to go to tigerdirect b/c vendors on-site debugging serious problem (work-related) - and I was at work until 8:45PM Mar 29 05:16:51 we suspect this is a >64MB kernel problem Mar 29 05:17:36 perhaps flash debian back on it, and verify that you're seeing 256MB in debian Mar 29 05:17:40 yeah I think so. Even other wiki notes said that no changes were necessary to get redboot to detect 64MB. However, the apex > 1.4.16 implies that it works w/o mucking with redboot Mar 29 05:18:16 there's a dmabounce kernel boundary at 64MB too Mar 29 05:18:35 hmm ok I'll revert the image and see. but I may just keep 64M until speed problem figured out. dmabounce ? Mar 29 05:18:44 (although I must admit I don't really understand what the dmabounce thing is all about) Mar 29 05:19:07 jacques or mwester might know more about the dmabounce problem Mar 29 05:19:30 ok, might be something for another evening :-) Mar 29 05:19:33 RobNC: if Linux can see 256MB in Debian, then we can start from there. Mar 29 05:19:49 (i.e. it might be a simple kernel setting or something) Mar 29 05:20:40 yeah that would be the best thing... but I'm wondering if apex can actually work in lieu of redboot. I.e., is it better to just have redboot for boot loader as 1st stage boot loader for fatslug? Mar 29 05:21:15 I meant is it better to just have apex as 1st stage boot loader Mar 29 05:21:38 but then serial console is a must for env changes Mar 29 05:22:01 (b/c no support for microcode for embedded ethernet) Mar 29 05:23:23 dmabounce should onlt affect dma Mar 29 05:23:48 the other slowness RobNC mentioned really sounds like disabled cache Mar 29 05:24:22 so not memory access. I was using serial console at the time, that doesn't use DMA right? Mar 29 05:25:53 correct Mar 29 05:26:12 usb for example would use dma Mar 29 05:26:39 well, usb-storage Mar 29 05:27:55 hmm... well, the slowness was while installing D-I beta 2, to USB2 storage devices. I let the thing run and download, and 6 hours later (went to sleep) it was still stuck. I could ping device, but dunno why it was so slow. There may have been both the cache issue and the dmabounce issue. Mar 29 05:29:55 RobNC: seems that I changed all the apex configs apart from the one that mattered :-( Mar 29 05:33:49 ah go figure!! rofl Mar 29 05:35:23 I was gonna resume the rsync and restore d-i but if you want me to try another slugimage, I think I can make time for one more test before turning in Mar 29 05:39:05 one sec Mar 29 05:40:03 np glad you found the problem! :-) Mar 29 05:42:58 03rwhitby * r807 10kernel/trunk/ (Makefile patches/apex/enable-commands.patch): Added .PRECIOUS target to the Makefile to save the apex configs, and fixed the mem commands for nslu2 and nas100d arm slugos apex configs. Mar 29 05:43:31 Bitten by pattern rules and a lack of .PRECIOUS for the intermediate target. Mar 29 05:44:07 * RobNC says that last statement just went right over his head Mar 29 05:44:19 Makefile stuff Mar 29 05:44:24 ok, new apex sent. Mar 29 05:45:24 if you have a chained pattern rule, which involves an intermediate dependency which is a real file, and which is only created by the pattern rule chain, then by default it is deleted after the final target is built. .PRECIOUS stops that. That's why I didn't see the config files to patch them. Mar 29 05:45:54 Ah okay that makes sense. Thanks for explanation. Mar 29 05:45:54 jelle: uploaded the picture, and linked the page in Mar 29 05:47:43 we loooves our .PRECIOUS we does Mar 29 05:48:04 heh... doing slugimage again with new apex Mar 29 05:48:08 I was waiting for someone to say that. Mar 29 05:48:15 I know. Mar 29 05:49:33 I bet Christian Hohnstaedt could get NPE working in u-boot in about 5 minutes Mar 29 05:50:05 jacques: challenge him by saying that his open source driver will never work in something like u-boot ... Mar 29 05:51:30 heck I would be willing to offer a bounty (just not a very large one ;-) Mar 29 05:52:08 I think my problems are partially caused by the PHY the nslu2 uses Mar 29 05:52:36 and maybe the fact that I didn't know I was supposed to be initializing NPE-B Mar 29 05:52:57 and maybe the firmware wasn't right in some way Mar 29 05:53:07 jacques: nslu2-linux could offer a bounty if we thought that having u-boot for ixp4xx was something that was very useful - e.g. if it had usb serial console or something ... Mar 29 05:53:52 * RobNC uses uboot at work and it has tftp support Mar 29 05:54:46 apex> memscan -u 0x0+32M Mar 29 05:54:47 0x0 0x02000000 (32 MiB) Mar 29 05:54:47 apex> memscan -u 0x0+64M Mar 29 05:54:47 0x0 0x02000000 (32 MiB) Mar 29 05:54:54 there we go it is there Mar 29 05:55:32 RobNC: RedBoot has tftp support - we need some new killer feature that allows users to get into the bootloader console without requiring *any* hardware mods. Mar 29 05:55:47 it returns awfully quickly. Mar 29 05:56:00 I don't believe it's a test, just a scan for aliases. Mar 29 05:56:28 scan for aliases - you mean like stuck bits? Mar 29 05:56:59 no, for a physical memory appearing multiple places in the memory map Mar 29 05:57:00 re: redboot this is true; console-less boot loader is quite tricky Mar 29 05:57:31 openmoko neo u-boot has cdc-acm over usb on boot Mar 29 05:57:35 u-boot has some usb drivers, but inexplicably no nec driver Mar 29 05:57:53 oh for me, I got memscan -u 0x0+32M -> shows 0x0 0x02000000 Mar 29 05:57:57 what's cdc-acm? Mar 29 05:58:10 memscan -u 0x0+64M => shows 0x0 0x04000000 Mar 29 05:58:11 serial over usb Mar 29 05:58:30 memscan -u 0x0+128M => shows 0x0 0x08000000 Mar 29 05:58:55 memscan -u 0x0+256M => shows 0x0 0x10000000 (256MiB) Mar 29 05:59:08 RobNC: remove the mem= from the command line, then do the 256MB memscan -u, then boot. Mar 29 05:59:39 right now, since I used your apex w/o parameter, there is no mem= command Mar 29 05:59:41 (Apex should pass the most recent memscan results to the kernel via the ATAGs) Mar 29 06:00:34 ok Mar 29 06:01:11 seems to hang "booting kernel at 0x00008000" Mar 29 06:02:12 redo, with memscan at 32MB and 64MB Mar 29 06:03:27 ok will do Mar 29 06:05:19 without sdram-init, the memscan -u 0x0+(any) shows 32M for 32,64,128,256 Mar 29 06:07:02 after sdram-init, memscan -u 0x0+64M shows 64M Mar 29 06:08:04 rwhitby: great, thanks. Mar 29 06:08:23 any mistakes on the page as far as you know? Mar 29 06:08:38 don't know enough about it - you're the expert it seems :-) Mar 29 06:09:11 :-) Mar 29 06:11:28 ok, thanks for your help Rod, gonna zzz now Mar 29 06:30:14 RobNC: did it boot with sdram-init and memscan > 64M ? Mar 29 06:37:13 I too was wondering Mar 29 10:15:40 shouldn't my /tmp directory be empty after reboot? Mar 29 10:17:15 because, well, it isn't... quite some files stay. (OpenSlug) Mar 29 10:31:25 jelle_: slugos 4.3 alpha, running from flash, /tmp is not preserved. Mar 29 10:31:38 (it's in tmpfs) Mar 29 10:34:57 good... oh, we are on 4.3 already? I'm still playing around with OpenSlug 3.1... should I upgrade? Mar 29 10:35:38 only if you want to alpha test 4.3, and are willing to reinstall from scratch if something goes wrong. Mar 29 10:36:32 I have some spare disks, so I could do that Mar 29 12:41:14 jelle: done Mar 29 12:41:32 night all Mar 29 12:47:25 03oleo 07uclibc-unstable * r5844 10optware/sources/buildroot/buildroot.config: unstable buildroot: enable Thread library debugging Mar 29 14:42:53 03oleo 07uclibc-unstable * r5845 10optware/make/uclibc-opt.mk: unstable uclibc-opt: reintroduce stack smashing protection and pthead debug libraries Mar 29 16:03:30 03oleo 07uclibc-unstable * r5846 10optware/make/uclibc-opt.mk: unstable uclibc-opt: fix GCC version check for libssp - this was false all time long Mar 29 20:24:06 morning Mar 29 20:37:56 03oleo * r5847 10optware/trunk/make/esniper.mk: esniper: 2.15.0->2.16.0 Mar 29 20:46:07 03bzhou * r5848 10optware/trunk/ (make/cups.mk sources/cups/printers.conf): cups: change from "DeviceURI usb:/dev/lp0" to "DeviceURI parallel:/dev/lp0" in printers.conf Mar 29 21:36:11 Anyone know of a 5V regulator that accepts 12V in and can supply more than 1A? Mar 29 21:53:13 hi Mar 29 21:53:47 theblue: heh 7805-something ;) Mar 29 21:54:01 No, that limits out at 1A. Mar 29 21:54:10 I'm trying to run a slug with USB and Ethernet. Mar 29 21:55:02 hack PoE support :) Mar 29 21:55:20 That'll let me use 12V? Mar 29 21:55:47 no, where do you get 12V from? Mar 29 21:56:16 A gel cell battery. Mar 29 21:56:20 ! Mar 29 21:56:24 I could use a lantern battery! Mar 29 21:56:28 Along with the gel cell! Mar 29 21:56:38 * CoreDump|home is pretty certain he used 7805-something w/ >1A before... Mar 29 21:56:47 come to think of it... PoE is typically 48V Mar 29 21:58:29 CoreDump|home: Don't think so, the specs say it can't go over 1A. Mar 29 21:59:10 theblue: there are quite a few different 7805 out there. Mar 29 21:59:21 CoreDump|home: Well, I'll see... Mar 29 21:59:29 Would the slug take 6V very well? Mar 29 21:59:44 theblue: I don't have any docs about ICs here or I'd have a look, sorry Mar 29 21:59:53 Ok, thanks anyway. Mar 29 21:59:57 that I do not know Mar 29 22:00:45 theblue: if everything else fails you could abuse one of these nice 3-to-30VDC converters and hard-wire it to ~5V Mar 29 22:01:10 Aren't those huge? Mar 29 22:01:14 I'm kind of on a space budget. Mar 29 22:01:17 not really Mar 29 22:01:29 Link please? Mar 29 22:01:33 but the need a resistor and a cap Mar 29 22:01:36 hmmm Mar 29 22:01:47 * CoreDump|home s technical english sucks Mar 29 22:03:18 LM317 is what you want IIRC. Mar 29 22:04:00 its data-sheet shows cirquit examples for adjustable voltage output Mar 29 22:04:40 Ooo, thanks! Mar 29 22:04:45 np Mar 29 22:05:16 theblue: if you can wait for tomorrow around this time, I can have a look for a "native" 5V reg w/ >1A at work Mar 29 22:05:24 CoreDump|home: Absolutely! Mar 29 22:05:31 ok =) Mar 29 22:06:39 * CoreDump|home s SVN repo just went tits up :( Mar 29 22:07:31 european? Mar 29 22:07:51 german Mar 29 22:08:20 http://rafb.net/p/lgaxDm14.html Mar 29 22:08:22 I was guessing British, but that's close. Mar 29 22:08:24 * CoreDump|home curses Mar 29 22:08:27 hehe Mar 29 22:08:31 I know, I do too. Mar 29 22:08:35 * theblue is American, sadly. Mar 29 22:09:17 theblue: your politicians suck, but I guess you know that already? =D Mar 29 22:10:03 of course, i've freed my mind. Mar 29 22:10:13 Yay! my TI samples came today! Mar 29 22:10:19 germany has had its share of sucky politicians :) Mar 29 22:10:25 Agreed. Mar 29 22:10:33 I was about to say "As bad as Bush?" Mar 29 22:10:34 kolla: hehe Mar 29 22:10:41 And scarily enough, I'd think you'd say yes. Mar 29 22:12:11 anybode good w/ svn here? =\ Mar 29 22:12:21 Sorry. Mar 29 22:12:50 maybe, what's the problem? Mar 29 22:13:17 Woah. Mar 29 22:13:22 TI's got this weird new packing crap. Mar 29 22:13:36 kolla: http://rafb.net/p/lgaxDm14.html Mar 29 22:13:38 It's like a sheet of kraft paper, with slits cut in it, and the whole thing's pushed at an angle. Mar 29 22:13:45 "Geami", they call it. Mar 29 22:15:39 Hmm. Mar 29 22:15:44 These are 1.5A 7805s. Mar 29 22:15:46 Guess I was wrong. Mar 29 22:15:54 CoreDump|home: yes, seen that... db4.2? Mar 29 22:16:22 kolla: dunno actually. "svnadmin recover" thankfully restored the DB Mar 29 22:16:41 yes, that should do it Mar 29 22:16:54 I was thinking of running db4.2_recover Mar 29 22:17:20 Crap! Mar 29 22:17:23 havent seen that error in a while, I hoped to find the recovery command in my bash.history, but it wasnt there :) Mar 29 22:17:36 kolla: hehe Mar 29 22:20:36 CoreDump|home: use fsfs fstype for the svn repo Mar 29 22:20:57 hmm? Mar 29 22:21:10 ah, yeah Mar 29 22:21:13