IRC logs for #openrisc Thursday, 2012-02-23

-!- Netsplit *.net <-> *.split quits: jonibo_02:22
-!- Netsplit over, joins: jonibo_02:28
jonibostekern: thanks... will give it another g o10:03
jonibostekern: many thanks, that did the trick!10:07
stekernok, cool, wonder what the problem was before though10:13
jonibohmmm... Image Type:   Unknown Architecture10:17
jonibohas the arch type ID changed since things went upstream?10:17
joniboyeah, I don't know what was wrong before... it was a clean checkout from your tree which you said you could build fine10:18
joniboi didn't build the u-boot on the board so it may be from an older u-boot...10:18
stekernjonibo: ah, yeah, it has changed10:22
jonibook10:22
stekernit's a bit of a pain, but it was necessary10:23
jonibosure, no worries, I understand10:23
stekernsince it is an increasing number10:23
joniboas a quick and dirty, can I change it back to the old number in my tree?10:23
joniboif yes, where?10:23
jonibook, found it10:24
jonibowhat was it before?10:24
stekernhttp://git.openrisc.net/cgit.cgi/stefan/u-boot/diff/include/image.h?id=73b553d99a9df8b2cc28c54e3473d8b2c3a115a610:24
jonibothx10:24
jonibo92!10:24
jonibosurprising10:24
joniboi would have guessed 2010:25
stekernyes, don't ask, can't remember why I picked that ;)10:25
stekernI guess I just took a large enough so I wouldn't have to change if new archs got merged before or10:25
jonibointeresting... now I get the undefined reference to 'ethoc_initialize' error again...10:26
stekernhow are you building things?10:26
joniboand changing  the ARCH number back to 21 resolves that...???10:26
joniboCROSS_COMPILE=or32-aac-elf- make -j1210:26
jonibonothing special10:27
stekernsounds strang10:27
stekerne10:27
joniboand now I set it to 92 again and this time it works10:27
jonibosomething doesn't seem to be happening in the right order10:28
jonibooh well...10:28
stekerndoes a make distclean before you make change things?10:28
jonibowill try10:28
joniboworks after a distclean10:29
stekerni usually do make distclean && make board_config, if I have changed anything in the header files10:29
joniboyeah, that's a good idea10:29
jonibobut it was broken right from a clean checkout10:29
stekernyeah, that was strange10:30
jonibowhat load address should i use?10:30
jonibo0, right?10:30
stekernyes, 010:31
stekernand 0x100 for execution10:31
joniboentry point 0x10010:31
joniboright, thx10:31
stekernyeah, the e is probably an abbrevation for entry point, not execution :P10:32
joniboit's almost working now, in any case10:32
jonibojust had the ep wrong10:32
stekerntools/mkimage -n 'Linux for OpenRISC' -A or1k -O linux -T kernel -C none -a 0 -e 0x100 -d /path/to/vmlinux.bin uImage10:32
stekernthat's the command line I'm using10:32
jonibocool, that's what I've got now10:33
joniboyay, it boots!10:34
jonibomany thanks!10:34
jonibowill integrate the uImage build into the kernel10:34
stekernnice one, I've been meaning to myself, but never got around to it10:35
jonibowhat's the status of hte device tree bits10:35
stekernthey are solid :)10:35
jonibodoes u-boot handle that?10:35
jonibowhere does it find the device tree?10:35
stekernfrom wherever you like10:35
jonibook10:36
joniboi'll read up10:36
stekernyou just load it into memory before you boot linux10:36
joniboyup10:36
stekernas long as there is driver support for the device to read in for it should "just work"10:36
jonibogood good10:36
stekerndriver support in u-boot that is, but you probably got that10:37
joniboI saw that it's possible build the device tree right into the u-boot image, too10:37
jonibowas thinking that might be the smartway to do it10:37
stekernyeah, you can do that too10:37
jonibothen again, there are disadvantages to that, too10:37
joniboanyway, nice work10:37
stekernand once you have the device tree in memory, u-boot knows how to modify it, so you can do some on the fly board identification in u-boot and adjust accordingly10:38
joniboaha, why should u-boot do the board identification?  the OS could do that itself?10:39
stekernjust as an example10:39
jonibou-boot should set up stuff that it knows about but that OS can't figure out for itself10:39
jonibolike memory size, for example, if u-boot knows about that10:39
jonibobut I understand what you're saying...10:39
stekernyeah sure, board identification was probably a bad example10:40
jonibo:)10:40
stekernchangin the kernel command line in the device tree on the fly is perhaps a better one10:41
joniboyeah, that's a relevant example ;)10:41
jonibo(couldn't come up with any good example, myself, either)10:41
juliusbjonibo: nice to see you're getting some time to play with OR stuff again :)11:58
joniboyeah, for sure, it's been a while12:49
pgavinonce I've built and run or1ksim on vmlinux, is there any way to interact with it?20:19
olofkInteract with or1ksim?20:30
pgavinum20:48
pgavinyeah, like play with a shell in the simulated linux20:48
pgavinjust to make sure it's working20:48
pgavinI mean, it looks like it works :)20:48
olofkpgavin: Look up the uart section in the or1ksim config. You can either make it start an xterm, or connect via a socket21:10
pgavinok21:11
pgavinthanks :)21:11

Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!