**** BEGIN LOGGING AT Tue May 12 02:59:57 2009 May 12 06:01:11 heyho... i have a git repo, and in one of the subfolders, there is a Java application. How can I use Eclipse to work on this java application within the git subdirectory? May 12 06:06:26 ough... wrong channel, sorry ^^ May 12 06:30:34 rabeeh: I am using the elf image that I got in the SheevaPlug Development Kit. May 12 06:31:03 I am executing "sheevaplug_init" followed by "load_image " followed by "resume 0x600000" May 12 07:54:11 guerby: http://openplug.org/plugforum/index.php?topic=272.0 May 12 07:55:33 rabeeh, wow the scope must be more expansive than my el cheapo killawatt :) May 12 07:56:14 guerby: :) May 12 07:56:36 guerby: i assume that he measured that on the 220v, before the power supply. May 12 07:57:06 rabeeh, measure seem coherent with mine so yes May 12 07:57:28 i'll ask if he can measure after the power supply on the 5V connector. May 12 09:19:52 tbm: I installed debian on my plug from scratch using debootrap --foreign, but I think that the sid kernel is not setting the correct MTD partitions. any hints? this is what I get in /proc/mtd: May 12 09:19:53 mtd0: 00100000 00020000 "u-boot" May 12 09:19:53 mtd1: 00400000 00020000 "uImage" May 12 09:19:53 mtd2: 1fb00000 00020000 "root" May 12 09:20:14 I am unable to mount the root May 12 09:20:43 Md: dunno, I haven't done anything with flash yet May 12 09:20:53 Md: btw, flash-kernel in sid doesn't support it, but the one in my tar ball does May 12 09:21:21 do you mind mailing me the script? May 12 09:22:14 nevermind, the partitions *are* correct. it's just that mount and vol_id do not recognize jffs2 May 12 09:23:05 tbm: BTW, I think that the sid kernel also lacks crypto acceleration May 12 09:23:38 Md: there is no crypto support yet May 12 09:24:02 also, after mounting the jffs2 file system pdflush spins at 100% CPU the whole time... May 12 09:24:57 and now jffs2_gcd_mtd2 replaced it. which is weird, considering that I am not writing to that fs May 12 09:33:38 hello eFfeM May 12 09:34:40 hi likewise May 12 09:34:49 everything ok? May 12 09:35:14 yes pretty well, and you? May 12 09:35:54 oe/florian and oe/XorA got my other plugs, so they will probably appear here at some time :-) May 12 09:36:02 reasonably busy, using my time to do some delayed cleaning (meaning throwing things away) May 12 09:36:17 ah ok, nice May 12 09:36:38 i have coherence 0.9.2 on it (hi dev!) May 12 09:50:43 eFfeM: \o/ May 12 09:51:25 probably 0.6.2, but maybe under all your gadgets is a time-machine too ;-) May 12 09:53:48 sorry, yes 0.6.2 May 12 09:54:37 too much gadgets and books and saved stuff. that is why my wife forced me to do some cleanup May 12 10:08:06 mine was delivered this morning - to the neighbor as I was away that 10min DHL came, and now my neighbor is away 8) May 12 10:12:58 dev: is your neighbor a Coherence developer too? :) May 12 10:13:34 dev, don't tell what it is, otherwise your neighbour might not give it back :-) May 12 10:15:16 dev, start with reading http://wiki.openplug.org/plugwiki/index.php/New_Plugger_How_To May 12 10:16:17 and maybe find yourself the right cable to connect it to the electricity network, it is a cable like you get with shavers etc (the device side looks like a figure 8 May 12 10:18:55 dev: why was it shipped with DHL and not FedEx? May 12 10:19:15 europe DHL, US FedEx? May 12 10:19:49 rabeeh: no, otherwise as eFfeM said I probably never would get it back :-) May 12 10:20:14 eFfeM: the cable is already on my table May 12 10:20:22 ah cool May 12 10:20:48 * eFfeM expects a report on coherence 0.6.4 on sheeva later today :-) May 12 10:21:13 eFfeM: I need to do the release, was busy with work-work :-( May 12 10:21:28 eFfeM: http://openplug.org/plugforum/index.php?topic=213.0 May 12 10:21:51 or was that an old version? May 12 10:22:42 that is 0.6.2 May 12 10:24:35 rabeeh, that is indeed 0.6.2 the version that comes with ubuntu, not sure if 0.6.4 is already available May 12 10:25:04 btw it is not an "old" version, it is still the current one (as long as 0.6.4 is not there) May 12 10:29:28 rabeeh: 0.6.2 is latest stable, 0.6.4 I hope to release today but it will probably take a few days until packeteers pick it up May 12 10:34:04 * spiyush waiting for response from rabeeh May 12 10:49:33 spiyush: http://www.openplug.org/plugwiki/index.php/Setting_Up_OpenOCD_Under_Linux#Example_1:_Running_Das_U-Boot_from_DRAM May 12 10:49:47 spiyush: i can't tell more than that (not openocd expert) May 12 10:50:26 rabeeh: I have already tried that. It dint work :( May 12 10:50:45 rabeeh: Thanks for ur reply though May 12 10:50:46 spiyush: is your system bricked or what? May 12 10:50:55 it isn't May 12 10:51:09 spiyush: oh. so you are just playing with it :) May 12 10:51:21 yeah May 12 10:51:24 spiyush: can you try other address - jumping to 0x670000 May 12 10:51:51 ok i'll try that May 12 10:53:41 spiyush: are you expecting something on the console too? i can see that when i'm burning a bricked plug with u-boot, i need to power off the plug and then power it back again May 12 10:54:04 spiyush: something like the FTDI chip inside the plug doesn't go back to it's default state as usb serial controller. May 12 10:57:02 does anybody know in which openocd SVN changeset sheevaplug support was introduced? May 12 11:10:05 rabeeh: I tried the other address. Dint work May 12 11:11:09 rabeeh: And I can't poweroff the board, since I want to run the uboot from DRAM May 12 11:11:29 rabeeh: unpacking plug now - thx a million! May 12 11:37:49 Md: version 1520 May 12 14:04:40 rabeeh: hows the status wrt another donation to gentoo? :) /me wants discovery May 12 14:52:00 mwester there? May 12 14:52:21 * mwester jumps awake, and looks about... May 12 14:52:25 netpbm May 12 14:52:42 Ah yes. May 12 14:52:42 a better fix maybe to change in the Makefile.config May 12 14:52:44 STRIPFLAG = -s --strip-program=$(STRIP) May 12 14:53:05 also i have a fix for the QA gnu hash error May 12 14:54:19 Ok, but not sure what that fixes. May 12 14:54:39 * mwester just duplicated the solution used elsewhere in OE. May 12 14:55:26 if stripping is ment to be done there it strips the symbol tables May 12 14:56:15 * mwester goes to check the generated executables... May 12 14:57:06 netpbm: ELF 32-bit MSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.14, stripped May 12 14:57:24 ahha, i thought it may have been stripped elsewhere May 12 14:57:40 i think oe does that automagically? May 12 14:57:47 .debug/netpbm: ELF 32-bit MSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.14, not stripped May 12 14:57:55 :) Sometimes OE "just works"... May 12 14:58:00 That's cool. May 12 14:58:26 ok, i'll chuck in the gnu hash fix which is just a small patch to libs/Makefile May 12 14:58:56 And if you remove the spots where I coded specifically for the NSLU2, I have yet to find anything in SlugOS that _doesn't_ work on the plug. :) May 12 14:59:02 after i check it works :) May 12 14:59:21 Who needs debian on the plug?! May 12 15:00:51 :D May 12 15:11:55 me May 12 15:11:57 :D May 12 15:14:30 use ubuntu :P May 12 15:24:30 who needs a distro? May 12 15:34:56 where can I buy a plug? May 12 15:35:10 I mean this "plug" May 12 15:41:23 not ear plugs? May 12 15:54:38 nope May 12 16:08:02 armin76: sorry again. no updates May 12 16:08:32 timtimred: did you see the wiki sidebar with distros? I couldn't tell how to add OE May 12 16:11:02 OE isnt really a distro... :) May 12 16:11:13 but worthy of a mention there May 12 16:11:17 you need to edit.... May 12 16:12:47 what? what to do? May 12 16:13:09 ok. going home now May 12 16:13:23 timtimred: please drop me a line what to mention. May 12 16:13:49 if you do the same as for the others we can fill in the rest May 12 16:14:02 cya rabeeh :) May 12 21:02:49 timtimred: I added SlugOS and Angstrom May 12 21:03:00 cool that would do May 12 21:03:07 timtimred: should I also add OE category? Does it make sense? May 12 21:03:09 they are both OE based May 12 21:03:22 OE has its own docs, could link to them May 12 21:03:58 if you make a category we can treat it as a page and point to getting started and docs May 12 21:04:09 (but its still not a distribution!) :) May 12 21:04:13 well, one of things is that every distro has some explanation on the wiki; except OE stuff that is very general and pointers to OE web site May 12 21:04:36 yep May 12 21:04:41 that sounds good May 12 21:04:43 rabeeh: will the recovery mechanim you're working on also allow you upgrade u-boot from a usb stick? May 12 21:05:24 that would be useful (so people can upgrade to the version with SD support when it's integrated) May 12 21:05:35 tbm: it will boot Linux, and by then you can do anything (upgrading u-boot etc...) May 12 21:05:53 tbm: does that sound reasonable enough? May 12 21:07:24 tbm: my dream at the end is that any one that has bricked his plug, or willing to change distro will have a tarball that will run part of it on the host; and part on a USB stick and will magically move between Ubuntu, Debian and all other friends May 12 21:07:44 rabeeh: it would be nice if it would detect a u-boot.img file on the usb stick and upgrade u-boot. but it's definitely not essential; as you say, people can use Linux to upgrade u-boot May 12 21:07:57 or tftp a u-boot image and upgrae May 12 21:08:03 tbm: the process must be that simple that any Java or mono developer will be able to burn and embedded system with a single click May 12 21:08:20 tftp will always be supported; but this is really for smart people May 12 21:09:26 rabeeh: you tested the SD patch for u-boot, right? Any opinion how "ready" it is? (and will you make ready for integration? :) May 12 21:09:30 tbm: as for the NAND ECC stuff; nico and lennert are working on a 4bit ECC for both openocd and kernel May 12 21:09:41 rabeeh: it's committed in openocd svn May 12 21:09:53 rabeeh: now i need to finish the decoder May 12 21:10:17 tbm: the FAT32 fix worked without an issue; for the SD cards it worked on 128MB old SD card but not on the 8GB Sandisk card May 12 21:10:37 lennert: right. version 1780? May 12 21:11:22 rabeeh: nico says 1768 works, haven't tested myself May 12 21:11:27 tbm: so potentially when lennert has the decoder up and running, we can also have u-boot env variable access from LK May 12 21:11:54 lennert: i tried 1520 with nico's patch. works perfectly fine May 12 21:12:00 rabeeh: kewl May 12 21:12:55 lennert: maybe the only issue with openocd right now is that when the process quit, it doesn't revert the FTDI chip to it's orginal state as USB serial controller May 12 21:13:02 lennert: i'll report that to nico May 12 21:13:11 right May 12 21:17:33 rabeeh: hopefully it will be possible to disable the u-boot recovery feature, since many people value quick booting... May 12 21:18:04 BTW, with the debian kernel I can set the eth0 MTU. apparently it even supports baby jumbo May 12 21:18:29 Md: it should support jumbo but as it does multi-page allocs for receive it might run out of memory when memory is fragmented May 12 21:19:56 Md: what do you mean? recovery feature is triggered only when you really need it. May 12 21:20:11 Md: only when you write 'recvr' on the prompt? May 12 21:20:41 rabeeh: I read in the forum about it checking at every boot for an USB device, maybe I did not understand how it works then May 12 21:21:08 BTW, how is integration in mainline u-boot progressing? May 12 21:21:56 Md: i'm not familiar with the status of that. I know Prafulla is trying to push something mainstream; but have no clue about status and timelines May 12 21:22:15 s/mainstream/mainline May 12 21:22:22 who is Prafulla? Marvell employee? May 12 21:22:27 yep May 12 21:23:26 Prafulla is another engineer the lennert needs to support :) May 12 21:23:34 like most of us :) May 12 21:23:35 :))) May 12 21:23:40 oh come on :) May 12 21:25:56 any more nice kit coming soon btw? May 12 21:26:19 lennert: your sculpture is almost done. we are at final stages of calculating it's ECC May 12 21:26:40 timtimred: sheevaplug pro? :) May 12 21:26:41 rabeeh: hehe May 12 21:27:18 or sheevaplug - 2 May 12 21:27:56 an updated sheevaplug - what are the changes? May 12 21:27:56 sheevaunplug May 12 21:28:29 sheevaplugless May 12 21:28:59 Sheevaplug Reloaded May 12 21:29:06 timtimred: we will have a sheeva without the USB and without the gigE May 12 21:29:34 and without sd and usb May 12 21:29:35 is the sd driver known to have problem May 12 21:29:37 . May 12 21:29:39 timtimred: we might also get red of some DDR too May 12 21:29:42 when booting on it, I get: May 12 21:29:44 [42949384.150000] mmc0: unhandled error status 0x4000 May 12 21:29:44 [42949384.160000] mmcblk0: error -42 sending read/write command, response 0x900, card status 0xc00 May 12 21:30:30 lennert: i know what do to. we will release a plug with audio and call it 'sheeva unplugged' May 12 21:30:35 cool it just sits there using a low amount of power. May 12 21:30:38 :) May 12 21:31:54 As long as it has the Red Button, it's good. May 12 21:33:29 aurel32: can you email nico@cam.org about that? May 12 21:33:47 I have just found a mini-sd adapter May 12 21:33:56 let's try another sd card before May 12 21:34:04 mini-sd is not that common May 12 21:34:30 aurel32: which kernel are you using? May 12 21:34:43 aurel32: do you have the mvsdio.patch in place? May 12 21:35:04 sheeva unplugged would just need batteries and a wireless nic May 12 21:35:28 rabeeh: 2.6.30-rc5 May 12 21:35:39 rabeeh: where I can get this patch? May 12 21:37:13 aurel32: best to get instructions on the forums. there is a post that explains where to get a uImage with all relevant patches. May 12 21:37:21 aurel32: sources ofcourse included May 12 21:37:26 * rabeeh searches May 12 21:38:31 * aurel32 was hacking his sheevaplug (just received it today) without knowing about the forums May 12 21:38:47 I only looked at the marvell git tree May 12 21:38:53 http://openplug.org/plugforum/index.php?topic=265.0 May 12 21:39:14 thanks May 12 21:39:25 aurel32: also look for the error code on the forums, there are few tips over there (search for 0x4000 ) May 12 21:39:53 maybe there is also a patch for to fix the map of the nand driver May 12 21:40:05 instead of overriding it from the command line May 12 21:40:22 aurel32: if that is still broken, please drop a line May 12 21:40:33 aurel32: no patch for the nand mtdparts May 12 21:41:12 are the mtdparts wrong in mainline? May 12 21:41:28 you have all the mtd broken message May 12 21:41:30 s May 12 21:42:17 i think mainline uses different name for the command line that the LSP. May 12 21:42:29 I am using mtdparts=orion_nand:0x400000@0x100000(uImage),0x1fb00000@0x500000(rootfs) May 12 21:42:32 as i recall it uses orion_mtd, instead of something else May 12 21:42:49 that fixes the problem and I am able to use the flash May 12 21:42:58 to put a kernel/initrd May 12 21:43:27 aurel32: right. we probably need get a patch to permanently fix that May 12 21:43:38 the machine support file should just be corrected then May 12 21:43:54 and now I am trying to boot onto an sdcard to get rid of the usb drive May 12 21:44:20 usb drive should be faster May 12 21:44:30 mechanical drive? May 12 21:44:38 no flash disk May 12 21:44:52 but I want to use the usb port for another thing May 12 21:44:59 and I don't want a hub... May 12 21:45:35 lennert: how would we access the first MByte of NAND flash (for u-boot env variables), if we would add the patch for the fixed mtdparts? May 12 21:46:02 notice that currently it provides 4MB for uImage, but at offset 1MB May 12 21:46:10 rabeeh: as far as i know the partition sizes just aren't correct? May 12 21:46:19 the offset seems wrong May 12 21:46:27 for root May 12 21:46:53 well, originally we had 1MB for U-Boot at offset 0x0, 4MB for uImage at offset 1MB and the rest for the rootfs May 12 21:47:16 i mean, the kernel table should correspond with whatever we ship May 12 21:47:16 i think the change was made so that /dev/mtd0 wouldn't go to u-boot, so people might mistakenly override it May 12 21:47:34 I have just taken the default command line of the original software May 12 21:47:45 just replacing the name May 12 21:48:13 right. so the default start with uImage at offset 1MB and skips the u-boot partition May 12 21:48:28 which is the first 1MB May 12 21:48:31 that is nand_mtd -> orion_nand May 12 21:51:40 ok worse with the new card May 12 21:51:43 [42949380.660000] end_request: I/O error, dev mmcblk0, sector 7146913 May 12 21:51:44 [42949380.660000] mmcblk0: error -110 sending status comand<3>mmcblk0: error -110 sending read/write command, response 0x900, card status 0x900 May 12 21:52:06 at least the first one was working during part of the boot process May 12 21:57:59 tbm: do you think we can include the msvdio patches in the debian kernel? maybe the clock gating one too May 12 21:58:03 I mean in trunk/ May 12 21:59:22 aurel32: one part of the mvsdio patch is included May 12 21:59:28 the other part is still under discussiona faik May 12 21:59:51 ok May 12 22:00:21 just found an sd-card that works \o/ May 12 22:02:50 "brand" name on the ones that won't work? May 12 22:03:42 does not work: lexar (SD 2GB) and kingston (mini-SD 4GB) May 12 22:03:53 work: corsair (mini-SD 2GB) May 12 22:05:48 aurel32: the mdelay() should fix the kingston -110 problem (at least did here) May 12 22:05:58 ok May 12 22:06:11 http://git.marvell.com/?p=orion.git;a=commitdiff;h=e99684bf8862c8c7342e9e0fce2bf1c64a7c83f5 May 12 22:06:13 tomorrow, I'll try to buy a new fast one May 12 22:06:24 I'll avoid kingston ones then May 12 22:06:24 mine wasn't mini-sd tho May 12 22:06:45 well I think micro and mini-sd are slower May 12 22:06:50 but I only have that at home May 12 22:29:53 does anybody have a sheevaplug configuration for uboot-envtools? May 12 22:31:19 or can I just copy the mtd0 entry from /proc/mtd? May 12 22:41:55 apparently not... May 12 23:46:53 why do I need to saveenv (and reload?) after setting mainlineLinux and arcNumber? If I don't do this the kernel does not notice the change May 13 02:10:44 when using these which is that base, and which is the size limit? mtdparts=orion_nand:0x100000@0x00000(u-boot),0x300000@0x100000(uImage),0x1fc00000@0x400000(rootfs) May 13 02:11:06 trying to figure out how to expand and which direction May 13 02:19:11 @0x00000 is base and 0x100000 is the size May 13 02:19:15 cat /proc/mtd May 13 02:25:43 so my uImage butts up against my root, thats what I was afraid of, ty May 13 02:28:07 intersting the size doens't match what I have here if I cat mtd like you mentioned it shows mtd1: 00200000 00020000 "uImage" May 13 02:59:56 Hmm, I changed my password, upgraded and rebooted. Now my password no longer works. Known problem? **** ENDING LOGGING AT Wed May 13 02:59:57 2009