| 00:55.39 | *** join/#nslu2-linux jacques (n=jacques@nslu2-linux/jacques) |
| 01:26.51 | hillct | the feed seems to be operating properly with respect to the openpbx rev2 packages |
| 01:27.08 | hillct | now anyway |
| 01:27.25 | hillct | duno if that was the result of some manual change or an automated rebuild |
| 01:27.35 | hillct | thanks |
| 01:30.33 | hillct | I guess sox never made it into the feed - just sox-native |
| 01:41.03 | *** join/#nslu2-linux RobNC (n=NSLU2-IR@user-0c8ha9e.cable.mindspring.com) |
| 01:52.03 | RobNC | Any APEX experts here - please let me know. Problem with APEX and FatSlug - memory detected fine but access extremely slow |
| 01:53.47 | rwhitby | hey RobNC |
| 01:54.12 | rwhitby | can you pastebin where apex detects the memory? |
| 01:56.42 | RobNC | hey rwhitby. yeah it detects fine. Will paste URL gimme sec |
| 01:56.55 | RobNC | actually I updated the wiki with the output |
| 01:57.00 | RobNC | gimme sec to find it |
| 01:57.58 | RobNC | http://www.nslu2-linux.org/wiki/Debian/FatSlug - see about 1/3 down the page. |
| 02:02.05 | rwhitby | ok, that looks fine. what's the deal with slow access then? |
| 02:03.33 | RobNC | I dunno, doesn't make sense. When I mean slow, I mean like at least 10-20x slower access. I couldn't do the D-I with 256MB. I reverted the env back, and then installed D-I. Now that I have it installed, I'm rsync'ing to my linux server. I could try now changing the apex env options and boot now, but it bothered me that the whole installer seemed to work significantly slower. Maybe because it wasn't in low-mem mode? |
| 02:04.24 | rwhitby | so just DI was slower, or is normal stuff slower too? |
| 02:04.35 | rwhitby | i.e. can you do any benchmarks ? |
| 02:05.17 | RobNC | it seemed that everything else was slower too, like partitioning the HDD. Do you know of any benchmarking programs I can run in 32MB mode, and again in 256MB mode? I wonder if something dumb like the IXP420 cache is being disabled and never re-enabled or something else |
| 02:06.19 | RobNC | I saw that the new APEX disables the cache (i think I read) so that sdram can be detected properly |
| 02:07.28 | RobNC | yeah they called it "cache lock-down", so that SDRAM can be properly detected. |
| 02:10.03 | jacques | yeah I was just about to say L1 cache |
| 02:10.22 | rwhitby | I've sent an email to Marc Singer - hopefully he sees it and is able to join us here. |
| 02:12.11 | RobNC | ok, I have emailed him bunches of time in the past but maybe he has gmail set to go to spam bait. I am wondering if the cache is re-enabled after SDRAM config changed. |
| 02:13.48 | RobNC | if I had to make a guess, I'd say it performed like PCs in the past - with the old HX chipset. You remember - go above 256MB or something, and then the thing acts very slow due to high mem not being cached and windows running out of high mem first. |
| 02:14.51 | jacques | sounds plausible |
| 02:15.11 | jacques | does setting mem=64M in kernel command line help? |
| 02:17.30 | RobNC | I can try that if necessary. I was using apex to get me into D-I RC2 so not sure if I could access kernel command line, unless it's through the "kernelsrc" parameter |
| 02:18.02 | rwhitby | hmm - wonder if this the dmabounce thingo? |
| 02:18.48 | jacques | ;-) |
| 02:18.58 | rwhitby | RobNC: it would be helpful if you could set yourself up to build from our kernel svn repo |
| 02:19.22 | rwhitby | then we can use that kernel (which does accept cmdline args from apex) to do some testing. |
| 02:19.28 | RobNC | yeah I will... but work projects are driving me crazy - get home at 9PM and stuff. not much brain juice left :-) |
| 02:20.01 | RobNC | I'm more a h/w engineer but I have done linux kernel compiles and program compiles, so it shouldn't be too hard |
| 02:20.11 | rwhitby | ok, we can feed you kernels anyway. |
| 02:21.35 | RobNC | ya, and best of all, I have JTAG *BigGrin* |
| 02:21.35 | rwhitby | can you get the latest SlugOS/LE alpha from http://www.hentges.net/tmp/snapshots/SlugOS/ ? |
| 02:21.35 | RobNC | ok, is that gonna affect debian-install ? |
| 02:21.36 | rwhitby | yes |
| 02:21.47 | rwhitby | it's going to replace the firmware in flash, but not affect anything on your disk. |
| 02:22.05 | rwhitby | if you're in userland now, you can backup your firmware in flash so you can upslug2 it back later |
| 02:22.30 | RobNC | ah ok gotcha, so instead of running 2.6.18-4-ixp4xx kernel |
| 02:22.48 | jacques | we have a tool that backs up flash? cool |
| 02:22.54 | rwhitby | instead of running Debian altogether. |
| 02:23.01 | rwhitby | jacques: yeah, it's called 'cat' |
| 02:23.14 | rwhitby | cat /dev/mtdblock* > backup.bin |
| 02:23.31 | rwhitby | backup.bin is immediately useable by upslug2. |
| 02:23.45 | RobNC | I have my previous D-I RC1 backed up (with functioning Redboot). But once you've used JTAG and get familiar, it's not too "scarey" anymore |
| 02:23.58 | rwhitby | none of this will touch redboot |
| 02:24.03 | rwhitby | so no jtag required |
| 02:24.43 | rwhitby | are you in http://www.hentges.net/tmp/snapshots/SlugOS/accepted/4.3-Beta/20070310/LittleEndian/ yet? |
| 02:24.51 | jacques | familiar? that's a handheld distro |
| 02:25.12 | rwhitby | I think RobNC meant get familiar with using jtag |
| 02:25.14 | RobNC | gotcha, that's a plus obviously. If you guys want, I can run this now. |
| 02:25.31 | rwhitby | sure, I have 30 mins free. |
| 02:26.03 | jacques | ah heh |
| 02:26.11 | RobNC | ok will try now |
| 02:27.07 | rwhitby | you need to read the licese linked off http://www.hentges.net/tmp/snapshots/SlugOS/ and type the password |
| 02:27.14 | rwhitby | s/licese/license/ |
| 02:27.31 | rwhitby | then get the file http://www.hentges.net/tmp/snapshots/SlugOS/accepted/4.3-Beta/20070310/LittleEndian/slugosle-4.3-beta-nslu2.bin |
| 02:27.41 | rwhitby | unpack that with slugimage |
| 02:28.12 | rwhitby | I will email you a new apex to use |
| 02:30.01 | RobNC | ok sorry info overload |
| 02:30.12 | RobNC | ok getting that one |
| 02:30.45 | rwhitby | let me know when you've flashed the new image built from the apex I just sent you and the other files from that slugosle 8MB image. |
| 02:32.16 | RobNC | rwhitby you mean slugimage it too right |
| 02:32.22 | rwhitby | yep |
| 02:33.29 | rwhitby | unpack the file you get from hentges.net, pack it back up with the apex I sent you 5 minutes ago |
| 02:33.29 | rwhitby | oh, I need to give you the right slugimage command for that |
| 02:33.30 | *** join/#nslu2-linux ByronT (n=byron-po@nslu2-linux/ByronT) |
| 02:33.39 | RobNC | ok gotcha... my email etc is on PC I am on now, but PC with serial port, etc. is another machine ( |
| 02:34.51 | RobNC | ok backing up flash now |
| 02:35.07 | RobNC | gonna do the packing in a sec |
| 02:35.11 | rwhitby | I'm prototyping the steps here now |
| 02:36.53 | rwhitby | RobNC: you need "-r Flashdisk:Flashdisk" for the slugimage |
| 02:37.07 | rwhitby | (the ramdisk.gz is dummy for slugosle) |
| 02:38.13 | RobNC | ok backing up flash via sftp right now |
| 02:38.56 | rwhitby | slugimage -u -i slugosle-4.3-beta-nslu2.bin |
| 02:38.57 | RobNC | do I need to get the ramdisk image from that site |
| 02:39.06 | rwhitby | no. just the 8MB .bin file |
| 02:39.10 | RobNC | I meant zimage |
| 02:39.13 | RobNC | ok |
| 02:40.11 | RobNC | ok gonna do slugimage commad now |
| 02:41.15 | rwhitby | slugimage -p -o slugos-newapex.bin -b RedBoot -s SysConf -L |
| 02:41.15 | rwhitby | apex-slugos-nslu2-arm-1.4.17.bin -k vmlinuz -r Flashdisk:Flashdisk -t Trailer |
| 02:42.13 | RobNC | don't I slugimage -u first right |
| 02:42.26 | rwhitby | yep - I pasted that above |
| 02:43.14 | RobNC | yup I see that.. I need to use a real IRC client |
| 02:44.10 | *** join/#nslu2-linux mithro (n=tim@lester.mithis.com) |
| 02:46.01 | rwhitby | RobNC: actually, you need to download the http://www.hentges.net/tmp/snapshots/SlugOS/accepted/4.3-Beta/20070310/LittleEndian/zImage-ixp4xxle file and use that instead of the vmlinuz you unpacked. |
| 02:46.47 | rwhitby | (cause the vmlinuz in the .bin file has the arm-kernel-shim which hardcodes the cmdline) |
| 02:46.53 | RobNC | ok thought there'd be a need for that - gonna do that now |
| 02:48.17 | RobNC | grrr firefox keeps putting in the .htm extension. |
| 02:48.23 | RobNC | in that file (with no extension) |
| 02:49.31 | rwhitby | make sure you get a file that's about 1MB |
| 02:49.52 | RobNC | thing I'm worried about is txt instead of bin mode for xfer. |
| 02:50.34 | RobNC | 1040272 bytes |
| 02:50.55 | rwhitby | md5sum is 79b5b305f7b247b77df57f409a02e13b |
| 02:51.04 | rwhitby | length is correct |
| 02:51.19 | rwhitby | I'm flashing that now |
| 02:51.34 | RobNC | ok checking md5sum now too |
| 02:51.45 | RobNC | yep its ok |
| 02:52.11 | RobNC | that replaces vmlinuz right |
| 02:52.16 | rwhitby | yep |
| 02:52.34 | RobNC | k |
| 02:52.35 | rwhitby | ok, that booted |
| 02:52.57 | rwhitby | flash that image, and make sure it boots normally |
| 02:53.22 | rwhitby | then reboot and interrupt apex |
| 02:53.32 | RobNC | ok gonna try now on mine - gotta update via tftp (b/c using work laptop w/o network access to my linux box) |
| 02:55.20 | rwhitby | ok, I was able to change the apex cmdline and set mem=16M and boot with 16MB |
| 02:55.33 | RobNC | what parameter do you change for that? |
| 02:55.51 | rwhitby | add "mem=XXM" to the end of the cmdline environment variable |
| 02:56.17 | rwhitby | setenv cmdline "root=/dev/mtdblock4 rootfstype=jffs2 console=ttyS0,115200 |
| 02:56.18 | rwhitby | mem=16M" |
| 02:56.47 | rwhitby | jacques: what should we use in slugos for a simple mem access benchmark? |
| 02:57.23 | RobNC | gotcha - brb |
| 02:59.13 | RobNC | ok my cmdline for apex is: cmdline *= console=ttyS0,115200 rtc-x1205.probe=0,0x6f noirqdebug |
| 02:59.55 | RobNC | that is unmodified (from original apex from D-I RC2 - or was it from rwhitby?) |
| 03:00.20 | rwhitby | that would be from DI |
| 03:00.26 | rwhitby | change it to what I have above |
| 03:00.36 | rwhitby | the rtc stuff is done automatically by the slugos kernel |
| 03:01.26 | RobNC | ok gotcha. brb gonna do the flash upgrade, then do the apex env mod |
| 03:01.58 | jacques | rwhitby, hdparm |
| 03:02.04 | jacques | hdparm -T |
| 03:02.52 | rwhitby | jacques: what about without a disk attached? |
| 03:03.13 | RobNC | does it matter about disk attached? |
| 03:03.24 | rwhitby | we want to measure memory only, not disk. |
| 03:03.58 | jacques | -T does measure memory |
| 03:04.15 | jacques | tho hdparm might want a device to look at |
| 03:04.22 | RobNC | ok good catch - measures device cache. |
| 03:05.27 | mwester | Create a ramdisk and dd several MB of data in and out of it? |
| 03:05.50 | mwester | <PROTECTED> |
| 03:06.56 | RobNC | 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 |
| 03:07.26 | rwhitby | go for 256MB first |
| 03:07.28 | jacques | why 16MB and not 32? |
| 03:07.44 | RobNC | also, do I need to run the parameters up front, like "sdram-init" and "memscan -u 0+256m" |
| 03:08.36 | rwhitby | jacques: that was so I could see if it was recognised or not |
| 03:09.25 | mwester | root@grouch:~$ time dd if=/dev/zero of=/media/ram/x bs=32768 count=970 |
| 03:09.25 | mwester | 31784960 bytes (32 MB) copied, 1.57448 seconds, 20.2 MB/s |
| 03:09.35 | rwhitby | mwester: good idea |
| 03:09.35 | RobNC | think I should use a comma before mem=256M ? |
| 03:09.47 | rwhitby | no |
| 03:09.54 | rwhitby | space before it |
| 03:10.02 | RobNC | ah ok sorry gotcha - makes no sense to do so |
| 03:10.05 | rwhitby | the comma that is there is part of the console attribute |
| 03:11.24 | rwhitby | ok, I gotta go do a server outage. back in about 30 mins |
| 03:11.25 | RobNC | yeah I saw that (lack of sleep) |
| 03:12.58 | RobNC | seems like it is stuck at ".....done, booting the kernel." |
| 03:13.08 | RobNC | lemme try 16M to ensure I have the params right |
| 03:13.40 | rwhitby | RobNC: pastebin your slugimage commands to build the image you flashed |
| 03:13.59 | RobNC | k |
| 03:16.03 | rwhitby | mwester: that count is too big for 32MB |
| 03:16.31 | mwester | That fit in my dsmg600's default ramdisk. |
| 03:16.35 | mwester | But only barely. |
| 03:16.53 | mwester | Ah -- the dsm has 64MB ram --- sorry! |
| 03:17.09 | jacques | heh |
| 03:17.14 | jacques | as does the lude |
| 03:17.43 | mwester | Hmmm... yes, and rwhitby has a lude, probably running the same slugos that I'm running. Would be interesting to compare numbers. |
| 03:19.06 | jacques | I need to take the serial adaptor off the slug I took to SCALE and put it back on the JTAG'd one |
| 03:20.48 | RobNC | http://pastebin.ca/414585 |
| 03:22.33 | RobNC | 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? |
| 03:22.39 | mwester | jacques: I bought one of these; it worked: http://cgi.ebay.com/USB-Data-Cable-for-LG-5200-3200-1200-CD_W0QQitemZ160099123895QQcategoryZ50327QQssPageNameZWDVWQQrdZ1QQcmdZViewItem |
| 03:23.36 | jacques | is the picture correct? where are the electronics? |
| 03:24.26 | rwhitby | mwester: I'm testing this on a slug |
| 03:25.21 | mwester | 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 |
| 03:25.22 | RobNC | rwhitby don't read webpge just yet - I made mistake - didnt yet subst zimage |
| 03:25.54 | rwhitby | root@nslu2:/media/ram$ time dd if=/dev/zero of=/media/ram/x bs=32768 count=470 |
| 03:25.54 | rwhitby | 470+0 records in |
| 03:25.54 | rwhitby | 470+0 records out |
| 03:25.54 | rwhitby | real 0m 0.72s |
| 03:25.54 | rwhitby | user 0m 0.00s |
| 03:25.54 | rwhitby | sys 0m 0.72s |
| 03:26.19 | mwester | rwhitby: I just tested on a slug as well. Moving 400 * 32KB took 0.62 real seconds (0.59 sys). |
| 03:26.31 | mwester | Let me try 470... |
| 03:26.52 | mwester | 0.72, 0.00, 0.72 |
| 03:27.29 | rwhitby | ok, that matches then :-) |
| 03:27.58 | rwhitby | anyone know how to access a Dell DRAC II from RHEL4 ? |
| 03:28.05 | mwester | 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. |
| 03:28.50 | rwhitby | mwester: and you get a ringtones CD with it too! ;-) |
| 03:29.34 | mwester | :) Yep! In Taiwanese, too (I think). |
| 03:29.48 | mwester | (it also has the windows driver for the serial chip) |
| 03:33.03 | jacques | hmm, bubble-chip |
| 03:33.22 | jacques | which linux driver it use? |
| 03:33.36 | mwester | ark3116 |
| 03:33.45 | jacques | thanks |
| 03:34.21 | jacques | I need more USB ports on my laptop |
| 03:34.56 | jacques | hmm, maybe I can plug my 7-port hub into the hub in my monitor |
| 03:35.10 | ByronT | 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 |
| 03:35.32 | jacques | my t42 only has 2 :-( |
| 03:36.06 | ByronT | same here (t42p..) |
| 03:37.00 | mwester | 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. |
| 03:37.23 | jacques | I now have usb speakers and monitor hub plugged into laptop |
| 03:37.47 | jacques | usb keyboard, mouse and 7-port hub plugged into monitor |
| 03:38.12 | *** part/#nslu2-linux pierrelux (n=pierre-l@144-125.sh.cgocable.ca) |
| 03:38.20 | rwhitby | RobNC: how goes the battle? |
| 03:38.23 | jacques | every port in the hub is filled, tho not all in use right now |
| 03:39.05 | RobNC | rwhitby: documenting steps - realized I missed some things in trying to scroll up/down the java IRC window |
| 03:39.26 | RobNC | reflashing. Will get you a paste URL soon |
| 03:40.21 | RobNC | sweet it is booting debian. didn't see the scrollby to see if it erased the 256M option and thus more mem |
| 03:40.54 | RobNC | weird, eth broke. can't get addres |
| 03:41.00 | RobNC | got brokenslug msg |
| 03:41.14 | RobNC | (D-I does dhcp and it works for me all the time) |
| 03:42.11 | RobNC | ah gotta change APEX env as scrolling back - the 256M option is gone |
| 03:43.01 | jacques | U-Boot 1.1.4 (Jul 1 2006 - 02:22:45) |
| 03:43.01 | jacques | <PROTECTED> |
| 03:43.01 | jacques | CPU: Intel IXP425 at 266 MHz |
| 03:43.02 | jacques | Board: NSLU2 - aka slug |
| 03:43.02 | jacques | DRAM: 32 MB |
| 03:43.02 | jacques | Flash: 8 MB |
| 03:43.06 | jacques | ... |
| 03:43.12 | jacques | now gotta get jtag-tools back up |
| 03:44.18 | RobNC | seems to still hang on mem=256M parameter |
| 03:44.28 | RobNC | gonna try smaller size to confirm it works on mine |
| 03:45.19 | rwhitby | RobNC: it's booting slugos, not debian :-) |
| 03:45.36 | RobNC | ah ok gotcha so I should disconnect USBHDDs |
| 03:45.42 | rwhitby | doesn't matter |
| 03:45.55 | rwhitby | it's booting and running entirely from flash |
| 03:46.08 | RobNC | ok about to try 16M to confirm the parameter is being read correctly |
| 03:46.21 | rwhitby | ok |
| 03:46.22 | RobNC | you sure I don't need the sdram-init program? |
| 03:46.32 | rwhitby | not for 16, but probably for > 32 |
| 03:46.47 | RobNC | ok will try that next in the boot |
| 03:47.48 | RobNC | 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 :-) |
| 03:48.38 | RobNC | seems to boot into brokenslug prompt |
| 03:48.46 | RobNC | with mem=16M option |
| 03:49.31 | *** join/#nslu2-linux DrZimmerman (n=theo@84-74-134-243.dclient.hispeed.ch) |
| 03:50.32 | RobNC | Do I only change "startup" Apex env var? |
| 03:50.41 | RobNC | i.e., before the "copy -s ..." |
| 03:50.59 | rwhitby | you could just run the sdram init commands manually, then "boot" |
| 03:51.20 | RobNC | ok try that just for kicks |
| 03:51.25 | rwhitby | RobNC: dhcp is probably broken on that image |
| 03:51.49 | RobNC | dunno if memscan is needed |
| 03:51.55 | rwhitby | you can log in with "root" and "opeNSLUg" and "turnup init" to set the network params. |
| 03:52.25 | jacques | ok, jtag seems to work still |
| 03:52.30 | RobNC | yeah I forgot the l/p with openslug - been long time since running that |
| 03:53.15 | rwhitby | jacques: how big is uboot for the nslu2? |
| 03:53.45 | RobNC | doing manual "sdram-init" then "memscan -u 0+256m", then "boot" gets NSLU2 freeze saying "Booting kernel at 0x00008000" |
| 03:54.12 | RobNC | but it had left in the param "mem=16M" in tail of startup (ATAG_CMDLINE) |
| 03:56.43 | jacques | rwhitby, 256k |
| 03:57.11 | jacques | including config area and npe microcode |
| 03:58.04 | jacques | damn, that might mean config area is 128kB... |
| 03:58.19 | jacques | which means if it were moved external to u-boot... |
| 03:59.44 | RobNC | oh rwhitby got some interesting info |
| 04:00.14 | rwhitby | ... |
| 04:00.24 | RobNC | tried mem=256M no go. Tried mem=128M no go. Tried 64M it works. Now, it actually shows that if I login using openslug |
| 04:00.39 | RobNC | however, sdram-init I have not done. |
| 04:01.04 | RobNC | I have tried doing sdram-init, then 256M it hangs. sdram-init, then 128M it hangs. |
| 04:01.25 | RobNC | I know i have 256MiB b/c it is detected :-) |
| 04:02.04 | rwhitby | did you see 256MB when running debian? |
| 04:02.07 | RobNC | (and there is no fuzzy math for 4 @ 32Mx16 != 256MB) |
| 04:02.12 | jacques | detected by the bootloader or the kernel? |
| 04:02.30 | rwhitby | jacques: kernel just displays what the ATAG says, right? |
| 04:02.38 | RobNC | can't remember. about debian. I think debian saw it, using the prev slugimage and ur apex bootloader |
| 04:02.53 | jacques | rwhitby, I'm not sure |
| 04:03.17 | jacques | real test would be (I suppose) detection and memtest in bootloader |
| 04:03.38 | rwhitby | jacques: we're using apex here, and it's detecting 256MB |
| 04:03.43 | RobNC | jacques:I am unsure. hacking with mem=256M or mem=128M or mem-64M option with apex kernel option |
| 04:04.05 | RobNC | yes exactly that is what Apex1.4.17 is supposed to do |
| 04:04.30 | RobNC | yes rwhitby apex detects 256MiB correctly (using 1.4.17 only) |
| 04:04.32 | jacques | does apex have a builtin memory test? |
| 04:04.50 | jacques | to try at least writing to all memory (other than what bootloader is using) and reading it back? |
| 04:05.38 | RobNC | jacques dunno if apex has mem test. I was asking that yesterday actually |
| 04:06.38 | rwhitby | <PROTECTED> |
| 04:06.46 | rwhitby | CONFIG_CMD_MEMLIMIT |
| 04:06.52 | rwhitby | CONFIG_CMD_MEMSCAN |
| 04:06.52 | RobNC | ah that sounds like a compile-time option |
| 04:07.09 | rwhitby | yep, I can quickly do a new version with those enabled. |
| 04:07.18 | RobNC | sure if you don't mind and have the time |
| 04:07.35 | rwhitby | config CMD_MEMLIMIT |
| 04:07.36 | rwhitby | <PROTECTED> |
| 04:07.36 | rwhitby | <PROTECTED> |
| 04:07.36 | rwhitby | <PROTECTED> |
| 04:07.36 | rwhitby | <PROTECTED> |
| 04:07.36 | rwhitby | <PROTECTED> |
| 04:07.37 | RobNC | should I use d-i rc2 or openslug |
| 04:07.40 | rwhitby | <PROTECTED> |
| 04:07.42 | rwhitby | config CMD_MEMSCAN |
| 04:07.44 | rwhitby | <PROTECTED> |
| 04:07.46 | rwhitby | <PROTECTED> |
| 04:07.48 | rwhitby | <PROTECTED> |
| 04:07.50 | rwhitby | <PROTECTED> |
| 04:07.52 | rwhitby | <PROTECTED> |
| 04:07.54 | rwhitby | same image, just replace apex |
| 04:07.58 | rwhitby | or just tftp apex only and flash apex only |
| 04:08.23 | RobNC | true yeah |
| 04:09.56 | RobNC | is that 006000-007FFF? |
| 04:12.17 | *** join/#nslu2-linux nSiclair73de (n=Sinclair@p5B203160.dip0.t-ipconnect.de) |
| 04:13.41 | CIA-5 | 03rwhitby * r806 10kernel/trunk/patches/apex/ (enable-commands.patch series): Enabled memory commands |
| 04:16.03 | rwhitby | RobNC: 0x60000 to 0x7ffff, right? (you missed a digit and would have overwritten redboot) |
| 04:16.27 | rwhitby | 0x50060000 to 0x5007ffff in fact |
| 04:16.45 | rwhitby | RobNC: new apex with memlimit and memscan commands sent to you |
| 04:18.15 | RobNC | yes I see that now - obviously... looking at wiki it appears someone missed a digit. |
| 04:18.31 | rwhitby | which wiki page? |
| 04:19.22 | RobNC | apex over xmodem I think |
| 04:20.23 | RobNC | maybe I mis-read... http://www.nslu2-linux.org/wiki/HowTo/ReflashUsingApexAndXmodem |
| 04:21.17 | rwhitby | no, they've for four zeros there. |
| 04:21.23 | rwhitby | s/for/got/ |
| 04:21.28 | RobNC | ya u are right must have been somewhere else |
| 04:21.48 | rwhitby | can you check your browser history? leaving off a zero is fatal. |
| 04:23.41 | jacques | (unless you have jtag) |
| 04:24.31 | RobNC | ok |
| 04:26.13 | mwester | leaving off a zero when paying one's taxes is often nearly fatal. |
| 04:29.29 | RobNC | ok I looked and looked and can't find the bad page. I'll have to go back to that in a few. |
| 04:30.03 | rwhitby | ok |
| 04:34.35 | RobNC | apex is so small. any need to fill to 7ffff? |
| 04:34.41 | rwhitby | no |
| 04:38.25 | RobNC | trying to not use slugimage |
| 04:39.05 | RobNC | "load -r -v -b 0x01000000 -h TFTPSERVER filename" |
| 04:39.48 | RobNC | "fis write -f 0x50060000 -b 0x01000000 -l 0x0100acc0" |
| 04:40.00 | RobNC | right? says invalid flash address (?) |
| 04:40.40 | RobNC | says "valid range is 0x50000000 - 0x507fffff" but my range is inclusive heh |
| 04:42.15 | RobNC | guess I could always redo the whole 8MB but trying not to kill the flash |
| 04:42.47 | mwester | 0x50060000 + 0x0100acc0 = 0x5106acc0, which is > 0x507fffff |
| 04:43.28 | RobNC | whoops that is weird, how could it be that large |
| 04:43.52 | mwester | did you specify the end address instead of the size? |
| 04:44.06 | RobNC | size I think, that is what -l does right |
| 04:44.19 | mwester | Yes -l is the size, not the end address |
| 04:44.49 | RobNC | ah yep u got it I need to kill the 0x0100 prefix |
| 04:44.54 | mwester | :) |
| 04:45.47 | RobNC | lack of sleep - prolly shouldn't be mucking with slug :-) |
| 04:46.25 | RobNC | ok gonna try memscan if it exists in apex |
| 04:49.58 | jacques | well I have the latest svn u-boot to the same point I had in june |
| 04:50.27 | rwhitby | jacques: what is that point? |
| 04:51.13 | jacques | NPE doesn't work at all |
| 04:51.24 | RobNC | weird... I tried this and same thing both times. next gonna try slugimage to be sure. |
| 04:51.29 | jacques | maybe I need to byte-swap the microcode or something |
| 04:51.58 | jacques | flash works. I can update u-boot with u-boot |
| 04:51.59 | RobNC | I get this type of text: "$T0a0f:00020654;0d:0003dda4;#aa" |
| 04:52.56 | rwhitby | 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 |
| 04:53.21 | rwhitby | (I take all the little things that slugimage does for granted sometimes) |
| 04:53.56 | RobNC | ah ok whew! |
| 04:54.20 | rwhitby | that text you saw is the redboot built-in gdb server |
| 04:54.32 | rwhitby | bbiab (lunch) |
| 04:54.48 | RobNC | ah ok cool... |
| 05:08.42 | rwhitby | back |
| 05:09.17 | mwester | 14 minutes for lunch? That's fast. |
| 05:09.17 | RobNC | still here (not sure for how long b/c 1am here) |
| 05:09.25 | RobNC | yeah true! |
| 05:09.34 | jacques | yah really |
| 05:09.40 | rwhitby | mwester: brought it back to the desk - late lunch due to the scheduled server outage to install the DRACII card |
| 05:09.54 | mwester | Oh ok. |
| 05:10.53 | RobNC | rwhitby I don't see cmd memscan or memlimit in apex u built |
| 05:11.20 | rwhitby | hmm - neither do I :-( |
| 05:11.25 | RobNC | this is after slugimage rebuild |
| 05:12.18 | rwhitby | damn - somehow the patch didn't take |
| 05:12.31 | RobNC | sdram-init still shows 2 banks of 2 512Mib chips |
| 05:13.22 | RobNC | maybe I should check solder job? I doubt it, unless maybe I bumped it somehow (but doubtful) |
| 05:13.25 | rwhitby | looks like I stuffed up. sorry. |
| 05:13.42 | rwhitby | if it's late there, then perhaps another day. |
| 05:14.11 | rwhitby | (tomorrow same time will be fine for me) |
| 05:14.45 | RobNC | I'll look at it under microscope again tomorrow. I did this mod in December. |
| 05:14.46 | rwhitby | I'll make a new apex, and actually test the commands are there before sending it this time |
| 05:15.00 | RobNC | ok no problem, thanks so much for your help. |
| 05:15.04 | rwhitby | did tigerdirect have a slug? |
| 05:15.35 | rwhitby | hmm - maybe wait till we've got this one sorted out before creating another one? |
| 05:15.44 | RobNC | 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). |
| 05:16.28 | RobNC | 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 |
| 05:16.51 | rwhitby | we suspect this is a >64MB kernel problem |
| 05:17.36 | rwhitby | perhaps flash debian back on it, and verify that you're seeing 256MB in debian |
| 05:17.40 | RobNC | 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 |
| 05:18.16 | rwhitby | there's a dmabounce kernel boundary at 64MB too |
| 05:18.36 | RobNC | hmm ok I'll revert the image and see. but I may just keep 64M until speed problem figured out. dmabounce ? |
| 05:18.45 | rwhitby | (although I must admit I don't really understand what the dmabounce thing is all about) |
| 05:19.07 | rwhitby | jacques or mwester might know more about the dmabounce problem |
| 05:19.30 | RobNC | ok, might be something for another evening :-) |
| 05:19.33 | rwhitby | RobNC: if Linux can see 256MB in Debian, then we can start from there. |
| 05:19.50 | rwhitby | (i.e. it might be a simple kernel setting or something) |
| 05:20.40 | RobNC | 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? |
| 05:21.15 | RobNC | I meant is it better to just have apex as 1st stage boot loader |
| 05:21.38 | RobNC | but then serial console is a must for env changes |
| 05:22.01 | RobNC | (b/c no support for microcode for embedded ethernet) |
| 05:23.23 | jacques | dmabounce should onlt affect dma |
| 05:23.48 | jacques | the other slowness RobNC mentioned really sounds like disabled cache |
| 05:24.22 | RobNC | so not memory access. I was using serial console at the time, that doesn't use DMA right? |
| 05:25.53 | jacques | correct |
| 05:26.12 | jacques | usb for example would use dma |
| 05:26.39 | jacques | well, usb-storage |
| 05:27.55 | RobNC | 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. |
| 05:28.54 | *** join/#nslu2-linux jelle-zzzz (n=jelle@ip54518e8c.direct-adsl.nl) |
| 05:29.56 | rwhitby | RobNC: seems that I changed all the apex configs apart from the one that mattered :-( |
| 05:33.49 | RobNC | ah go figure!! rofl |
| 05:35.26 | RobNC | 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 |
| 05:39.05 | rwhitby | one sec |
| 05:40.03 | RobNC | np glad you found the problem! :-) |
| 05:42.58 | CIA-5 | 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. |
| 05:43.31 | rwhitby | Bitten by pattern rules and a lack of .PRECIOUS for the intermediate target. |
| 05:44.19 | rwhitby | Makefile stuff |
| 05:44.24 | rwhitby | ok, new apex sent. |
| 05:45.25 | rwhitby | 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. |
| 05:45.54 | RobNC | Ah okay that makes sense. Thanks for explanation. |
| 05:45.54 | rwhitby | jelle: uploaded the picture, and linked the page in |
| 05:47.43 | jacques | we loooves our .PRECIOUS we does |
| 05:48.04 | RobNC | heh... doing slugimage again with new apex |
| 05:48.08 | rwhitby | I was waiting for someone to say that. |
| 05:48.15 | jacques | I know. |
| 05:49.33 | jacques | I bet Christian Hohnstaedt could get NPE working in u-boot in about 5 minutes |
| 05:50.05 | rwhitby | jacques: challenge him by saying that his open source driver will never work in something like u-boot ... |
| 05:51.30 | jacques | heck I would be willing to offer a bounty (just not a very large one ;-) |
| 05:52.08 | jacques | I think my problems are partially caused by the PHY the nslu2 uses |
| 05:52.36 | jacques | and maybe the fact that I didn't know I was supposed to be initializing NPE-B |
| 05:52.57 | jacques | and maybe the firmware wasn't right in some way |
| 05:53.07 | rwhitby | 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 ... |
| 05:54.46 | rwhitby | apex> memscan -u 0x0+32M |
| 05:54.47 | rwhitby | <PROTECTED> |
| 05:54.47 | rwhitby | apex> memscan -u 0x0+64M |
| 05:54.47 | rwhitby | <PROTECTED> |
| 05:54.54 | RobNC | there we go it is there |
| 05:55.32 | rwhitby | 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. |
| 05:55.47 | RobNC | it returns awfully quickly. |
| 05:56.00 | rwhitby | I don't believe it's a test, just a scan for aliases. |
| 05:56.28 | RobNC | scan for aliases - you mean like stuck bits? |
| 05:56.59 | rwhitby | no, for a physical memory appearing multiple places in the memory map |
| 05:57.00 | RobNC | re: redboot this is true; console-less boot loader is quite tricky |
| 05:57.31 | rwhitby | openmoko neo u-boot has cdc-acm over usb on boot |
| 05:57.35 | jacques | u-boot has some usb drivers, but inexplicably no nec driver |
| 05:57.53 | RobNC | oh for me, I got memscan -u 0x0+32M -> shows 0x0 0x02000000 |
| 05:57.57 | jacques | what's cdc-acm? |
| 05:58.10 | RobNC | memscan -u 0x0+64M => shows 0x0 0x04000000 |
| 05:58.11 | rwhitby | serial over usb |
| 05:58.30 | RobNC | memscan -u 0x0+128M => shows 0x0 0x08000000 |
| 05:58.55 | RobNC | memscan -u 0x0+256M => shows 0x0 0x10000000 (256MiB) |
| 05:59.08 | rwhitby | RobNC: remove the mem= from the command line, then do the 256MB memscan -u, then boot. |
| 05:59.39 | RobNC | right now, since I used your apex w/o parameter, there is no mem= command |
| 05:59.41 | rwhitby | (Apex should pass the most recent memscan results to the kernel via the ATAGs) |
| 06:00.34 | RobNC | ok |
| 06:01.11 | RobNC | seems to hang "booting kernel at 0x00008000" |
| 06:02.12 | rwhitby | redo, with memscan at 32MB and 64MB |
| 06:03.27 | RobNC | ok will do |
| 06:05.19 | RobNC | without sdram-init, the memscan -u 0x0+(any) shows 32M for 32,64,128,256 |
| 06:07.02 | RobNC | after sdram-init, memscan -u 0x0+64M shows 64M |
| 06:08.04 | jelle | rwhitby: great, thanks. |
| 06:08.23 | jelle | any mistakes on the page as far as you know? |
| 06:08.39 | rwhitby | don't know enough about it - you're the expert it seems :-) |
| 06:09.11 | jelle | :-) |
| 06:11.28 | RobNC | ok, thanks for your help Rod, gonna zzz now |
| 06:30.14 | rwhitby | RobNC: did it boot with sdram-init and memscan > 64M ? |
| 06:37.13 | jacques | I too was wondering |
| 07:37.40 | *** join/#nslu2-linux eiffel_ (n=eiffel@p5B0535DA.dip0.t-ipconnect.de) |
| 10:15.13 | *** join/#nslu2-linux jelle_ (n=jelle@ip54518e8c.direct-adsl.nl) |
| 10:15.40 | jelle_ | shouldn't my /tmp directory be empty after reboot? |
| 10:17.32 | jelle_ | because, well, it isn't... quite some files stay. (OpenSlug) |
| 10:31.25 | rwhitby | jelle_: slugos 4.3 alpha, running from flash, /tmp is not preserved. |
| 10:31.39 | rwhitby | (it's in tmpfs) |
| 10:34.57 | jelle_ | good... oh, we are on 4.3 already? I'm still playing around with OpenSlug 3.1... should I upgrade? |
| 10:35.38 | rwhitby | only if you want to alpha test 4.3, and are willing to reinstall from scratch if something goes wrong. |
| 10:36.32 | jelle_ | I have some spare disks, so I could do that |
| 11:04.04 | *** join/#nslu2-linux kolla (n=kolla@firda.uninett.no) |
| 12:12.18 | *** join/#nslu2-linux mithro (n=tim@ppp246-117.static.internode.on.net) |
| 12:14.34 | *** join/#nslu2-linux kolla (n=kolla@firda.uninett.no) |
| 12:29.54 | *** join/#nslu2-linux bullet (n=bullet@160.214.62.81.cust.bluewin.ch) |
| 12:41.14 | rwhitby | jelle: done |
| 12:41.32 | rwhitby | night all |
| 12:47.32 | *** join/#nslu2-linux jelle (n=jelle@ip54518e8c.direct-adsl.nl) [NETSPLIT VICTIM] |
| 12:47.32 | *** join/#nslu2-linux oleo (n=leon@www.lecad.uni-lj.si) [NETSPLIT VICTIM] |
| 12:47.32 | *** join/#nslu2-linux DaKa (n=David@nslu2-linux/daka) [NETSPLIT VICTIM] |
| 12:47.32 | *** join/#nslu2-linux joshin (n=joshin@unaffiliated/joshin) [NETSPLIT VICTIM] |
| 12:47.32 | *** join/#nslu2-linux callum (n=callum@morpheus.notthesame.co.uk) [NETSPLIT VICTIM] |
| 12:47.32 | *** join/#nslu2-linux dwery (n=dwery@nslu2-linux/dwery) |
| 12:47.34 | CIA-5 | 03oleo 07uclibc-unstable * r5844 10optware/sources/buildroot/buildroot.config: unstable buildroot: enable Thread library debugging |
| 13:17.10 | *** join/#nslu2-linux GPSFan (n=kenm@12.10.255.248) |
| 14:42.54 | CIA-5 | 03oleo 07uclibc-unstable * r5845 10optware/make/uclibc-opt.mk: unstable uclibc-opt: reintroduce stack smashing protection and pthead debug libraries |
| 15:07.49 | *** join/#nslu2-linux ByronT_ (i=byron-po@nslu2-linux/ByronT) |
| 15:27.46 | *** join/#nslu2-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
| 16:03.30 | CIA-5 | 03oleo 07uclibc-unstable * r5846 10optware/make/uclibc-opt.mk: unstable uclibc-opt: fix GCC version check for libssp - this was false all time long |
| 16:17.41 | *** join/#nslu2-linux AdamBaker (n=aab@userch028.dsl.pipex.com) |
| 16:25.44 | *** join/#nslu2-linux hillct (n=H@client200-5.dsl.intrex.net) |
| 16:39.01 | *** join/#nslu2-linux kolla (i=kolla@firda.kolla.no) |
| 16:46.10 | *** join/#nslu2-linux Cheef-Daniel (n=chatzill@p5b17acdf.dip0.t-ipconnect.de) |
| 16:55.12 | *** join/#nslu2-linux kami22 (n=kami@pD955C83F.dip.t-dialin.net) |
| 16:55.28 | *** part/#nslu2-linux kami22 (n=kami@pD955C83F.dip.t-dialin.net) |
| 17:21.42 | *** join/#nslu2-linux HopsNBarley (n=hops@nslu2-linux/HopsNBarley) |
| 17:46.00 | *** join/#nslu2-linux anywhere (n=kasper@cpe.atm2-0-107328.0x50a485ae.abnxx9.customer.tele.dk) |
| 17:46.16 | *** part/#nslu2-linux anywhere (n=kasper@cpe.atm2-0-107328.0x50a485ae.abnxx9.customer.tele.dk) |
| 18:23.11 | *** join/#nslu2-linux drif (n=drif@a88-114-66-217.elisa-laajakaista.fi) |
| 18:53.14 | *** join/#nslu2-linux caplink811-log (n=caplink8@dslb-088-073-024-171.pools.arcor-ip.net) |
| 19:13.12 | *** join/#nslu2-linux drath_ (i=vmaster@p5B07FF89.dip.t-dialin.net) |
| 19:49.15 | *** part/#nslu2-linux hillct (n=H@client200-5.dsl.intrex.net) |
| 19:59.02 | *** join/#nslu2-linux hillct (n=H@client200-5.dsl.intrex.net) |
| 20:18.42 | *** join/#nslu2-linux HopsNBarley (n=hops@nslu2-linux/HopsNBarley) |
| 20:24.06 | rwhitby | morning |
| 20:37.56 | CIA-5 | 03oleo * r5847 10optware/trunk/make/esniper.mk: esniper: 2.15.0->2.16.0 |
| 20:46.07 | CIA-5 | 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 |
| 21:09.54 | *** join/#nslu2-linux theblue_ (n=theblue@pool-71-163-245-188.washdc.fios.verizon.net) |
| 21:36.11 | theblue_ | Anyone know of a 5V regulator that accepts 12V in and can supply more than 1A? |
| 21:53.13 | CoreDump|home | hi |
| 21:53.47 | CoreDump|home | theblue: heh 7805-something ;) |
| 21:54.01 | theblue | No, that limits out at 1A. |
| 21:54.11 | theblue | I'm trying to run a slug with USB and Ethernet. |
| 21:55.02 | kolla | hack PoE support :) |
| 21:55.20 | theblue | That'll let me use 12V? |
| 21:55.47 | kolla | no, where do you get 12V from? |
| 21:56.16 | theblue | A gel cell battery. |
| 21:56.21 | theblue | ! |
| 21:56.24 | theblue | I could use a lantern battery! |
| 21:56.28 | theblue | Along with the gel cell! |
| 21:56.47 | kolla | come to think of it... PoE is typically 48V |
| 21:58.29 | theblue | CoreDump|home: Don't think so, the specs say it can't go over 1A. |
| 21:59.10 | CoreDump|home | theblue: there are quite a few different 7805 out there. |
| 21:59.22 | theblue | CoreDump|home: Well, I'll see... |
| 21:59.29 | theblue | Would the slug take 6V very well? |
| 21:59.44 | CoreDump|home | theblue: I don't have any docs about ICs here or I'd have a look, sorry |
| 21:59.53 | theblue | Ok, thanks anyway. |
| 21:59.57 | CoreDump|home | that I do not know |
| 22:00.45 | CoreDump|home | theblue: if everything else fails you could abuse one of these nice 3-to-30VDC converters and hard-wire it to ~5V |
| 22:01.10 | theblue | Aren't those huge? |
| 22:01.14 | theblue | I'm kind of on a space budget. |
| 22:01.17 | CoreDump|home | not really |
| 22:01.29 | theblue | Link please? |
| 22:01.33 | CoreDump|home | but the need a resistor and a cap |
| 22:01.36 | CoreDump|home | hmmm |
| 22:03.18 | CoreDump|home | LM317 is what you want IIRC. |
| 22:04.00 | CoreDump|home | its data-sheet shows cirquit examples for adjustable voltage output |
| 22:04.40 | theblue | Ooo, thanks! |
| 22:04.45 | CoreDump|home | np |
| 22:05.16 | CoreDump|home | theblue: if you can wait for tomorrow around this time, I can have a look for a "native" 5V reg w/ >1A at work |
| 22:05.24 | theblue | CoreDump|home: Absolutely! |
| 22:05.31 | CoreDump|home | ok =) |
| 22:07.31 | theblue | european? |
| 22:07.51 | CoreDump|home | german |
| 22:08.20 | CoreDump|home | http://rafb.net/p/lgaxDm14.html |
| 22:08.22 | theblue | I was guessing British, but that's close. |
| 22:08.27 | CoreDump|home | hehe |
| 22:08.31 | theblue | I know, I do too. |
| 22:09.18 | CoreDump|home | theblue: your politicians suck, but I guess you know that already? =D |
| 22:10.03 | theblue | of course, i've freed my mind. |
| 22:10.13 | theblue | Yay! my TI samples came today! |
| 22:10.19 | kolla | germany has had its share of sucky politicians :) |
| 22:10.25 | theblue | Agreed. |
| 22:10.33 | theblue | I was about to say "As bad as Bush?" |
| 22:10.34 | CoreDump|home | kolla: hehe |
| 22:10.41 | theblue | And scarily enough, I'd think you'd say yes. |
| 22:12.11 | CoreDump|home | anybode good w/ svn here? =\ |
| 22:12.21 | theblue | Sorry. |
| 22:12.50 | kolla | maybe, what's the problem? |
| 22:13.18 | theblue | Woah. |
| 22:13.22 | theblue | TI's got this weird new packing crap. |
| 22:13.36 | CoreDump|home | kolla: http://rafb.net/p/lgaxDm14.html |
| 22:13.38 | theblue | It's like a sheet of kraft paper, with slits cut in it, and the whole thing's pushed at an angle. |
| 22:13.45 | theblue | "Geami", they call it. |
| 22:14.30 | *** join/#nslu2-linux AdamBaker (n=aab@userch028.dsl.pipex.com) |
| 22:15.49 | theblue | Hmm. |
| 22:15.49 | theblue | These are 1.5A 7805s. |
| 22:15.50 | theblue | Guess I was wrong. |
| 22:15.54 | kolla | CoreDump|home: yes, seen that... db4.2? |
| 22:16.22 | CoreDump|home | kolla: dunno actually. "svnadmin recover" thankfully restored the DB |
| 22:16.41 | kolla | yes, that should do it |
| 22:16.54 | kolla | I was thinking of running db4.2_recover |
| 22:17.21 | theblue | Crap! |
| 22:17.23 | kolla | havent seen that error in a while, I hoped to find the recovery command in my bash.history, but it wasnt there :) |
| 22:17.36 | CoreDump|home | kolla: hehe |
| 22:20.36 | rwhitby | CoreDump|home: use fsfs fstype for the svn repo |
| 22:20.57 | CoreDump|home | hmm? |
| 22:21.11 | CoreDump|home | ah, yeah |
| 22:21.13 | rwhitby | CoreDump|home: svnadmin recover the repo, then dump it, then recreate as --fstype fsfs, then load it. |
| 22:21.26 | rwhitby | berkelydb fstype for svn repos sucks badly |
| 22:21.31 | CoreDump|home | rwhitby: I will do it the next time it blows up :\ |
| 22:21.45 | rwhitby | CoreDump|home: I'd do it now, so it never blows up again. |
| 22:22.32 | rwhitby | I had to do that for all the nslu2-linux svn repos, cause it would blow up when I was asleep. |
| 22:23.13 | CoreDump|home | rwhitby: hmmmmm |
| 22:23.25 | CoreDump|home | got an URL for the process by chance? |
| 22:23.35 | rwhitby | CoreDump|home: svnadmin recover the repo, then dump it, then recreate as --fstype fsfs, then load it. |
| 22:23.45 | rwhitby | all svnadmin commands |
| 22:26.55 | CoreDump|home | omg, SVN truly rulez |
| 22:28.50 | *** join/#nslu2-linux pierrelux (n=pierre-l@144-125.sh.cgocable.ca) |
| 22:28.53 | CoreDump|home | rwhitby: thanks! |
| 22:32.47 | kolla | ah yes, we moved to fsfs as well some time ago |
| 22:32.49 | kolla | :) |
| 22:33.18 | kolla | maybe that's why it was so long since I'd done the recovery thing :) |
| 22:33.28 | CoreDump|home | heh |
| 22:35.01 | theblue | Is it possible to wire several 7805s in parallel to get more current capacity? |
| 22:36.10 | CoreDump|home | theblue: not really possible in a safe manner |
| 22:36.18 | theblue | CoreDump|home: How unsafe is it? |
| 22:36.28 | theblue | And do I need any components other than the 7805, or can I just hook it up? |
| 22:36.58 | CoreDump|home | overloaded ICs can cause fire, that is however, unlikely =) |
| 22:37.16 | theblue | These have a temperature-limiting system. |
| 22:37.22 | theblue | Though I'm going to install heatsinks, of course. |
| 22:37.27 | CoreDump|home | theblue: and the 7805 can be operated pretty much standalone IIRC. Might need some cooling |
| 22:37.34 | theblue | But if these get too hot, they shut down, from what I read. |
| 22:37.51 | CoreDump|home | theblue: and hence overloading the _other_ 7805 |
| 22:38.01 | theblue | ? |
| 22:38.15 | theblue | So you mean if they're unbalanced, it's gonna be bad., |
| 22:38.16 | CoreDump|home | lets say you use two 7805 for ~2A |
| 22:38.30 | CoreDump|home | one goes down -> no goes up in smoke =) |
| 22:38.36 | CoreDump|home | s/no/no 2 |
| 22:38.48 | theblue | ? |
| 22:39.20 | CoreDump|home | if you try to pull 2A through a 1A IC, it's going to blow |
| 22:39.24 | theblue | True. |
| 22:39.31 | theblue | So what if I have 4 or 5, as failovers? |
| 22:39.52 | theblue | Since I've got 10 of these ICs. |
| 22:39.53 | CoreDump|home | it would be a _very_ ugly hack but could probably made to work |
| 22:40.01 | theblue | with no extra hardware? |
| 22:40.07 | theblue | If so, that's amazing. |
| 22:40.11 | CoreDump|home | I think so |
| 22:40.14 | theblue | Wow. |
| 22:40.26 | theblue | Should keep them all relatively cool, as well. |
| 22:40.32 | theblue | Since you've got each one handling 500mA. |
| 22:40.47 | theblue | And these things max out at 1.5A. |
| 22:41.05 | theblue | Assuming a 2A load, of course. |
| 22:41.16 | theblue | If one goes out, they each do 750mA, if 2 go out, each one goes 1A. |
| 22:41.23 | CoreDump|home | ugly, ugly hack :\ I would not do that, really. A 2A reg is dead cheap |
| 22:41.37 | theblue | The only way it could go wrong is if every single one except one fails. |
| 22:41.45 | theblue | CoreDump|home: You shold see the rest of the project I'm doing! |
| 22:41.49 | theblue | The whole thing is an ugly hack. |
| 22:44.52 | kolla | ofcourse :) |
| 22:45.30 | CoreDump|home | =) |
| 22:47.04 | CoreDump|home | theblue: the 7805 datasheet shows example cirquits for >1A |
| 22:49.03 | CoreDump|home | it even lists all parts w/ ID and a formula for the resistors |
| 22:56.38 | pierrelux | theblue: http://www.aqra.ca/IMG/bmp/regulation.bmp and http://www.aqra.ca/IMG/bmp/regulationbrd.bmp |
| 22:56.54 | pierrelux | eagle file here http://www.aqra.ca/IMG/zip/Regulation.zip |
