irclog2html for blob on 2002.02.27

01:26:29seletzhas to sleep now
01:26:33seletzbye
01:38:52Russever though of "begin 775 blob" being interpreted as "download blob"?
01:38:59Russs/though/thought
01:42:50Sammyhello
01:43:02Russhey sammy
01:43:24Sammyhowdy Russ :)
01:58:32BZFlagRuss: I agree, yes.
01:58:44BZFlagRuss: and if we add zmodem, even more so. ;-)
02:00:37RussBZFlag: I've been looking at the 32M sodimm situation, I think I can "fix" it
02:00:56RussI need to figure out the decompressor first though
02:01:16BZFlagthe kernel decompression needs to know about blocks, then have the kernel handle more regions and poof, right?
02:25:35Russwell, there are 4 stages
02:25:41Russthe boot loader
02:25:44Russthe decompressor
02:25:50Russthe initial 4M mapping
02:25:55Russand the full kernel MMU mapping
02:27:38Russthe bootloader and full mapping are no problem
02:27:48RussI think I can "fix" the initial 4M mapping
02:48:26BZFlagthe initial 4M mapping is mmuful?
02:52:08Russmmuful?
03:01:16BZFlagbootloader has the mmu off, at what point does it get turned on? in the 4M mapping right?
03:01:55BZFlagso the decompressor does not have the MMU on, but needs to decompress into the right locations so that when the 4M gets mapped it will be one chunk.
03:04:47Russright
03:05:05Russthe kernel currently assumes that 4M of contigous memory is available at TEXTADDR
03:09:15BZFlagwould it help to load a vmlinix image on jffs2? then it's blob's decompressor. ;-)
03:09:42Russblob doesn't know about memory "breaks" right now either
03:09:53BZFlagah, but you know the code better. ;-)
03:20:37BZFlaganyone got a blob system with an initrd handy?
03:20:55BZFlagI changed the code to pass the flash location of the ramdisk to the kernel.
03:21:06BZFlagunless it's downloaded.
03:21:16Russgrab one from erikm
03:21:29Russ(the lart pages)
03:22:11BZFlagI'm fairly confident in the patch, I think I'll just commit it and let folks test it. easy to revert.
03:25:47BZFlagcommited.
03:29:46Russshouldn't you make it not load the ramdisk on boot now?
03:32:02BZFlagif you set NOT_ENOUGH_RAM then it does. ;-)
03:32:14BZFlagor doesn't actually. ;-)
03:42:40Russshouldn't it do that in every case though?
03:42:56Russthe copy would then occur with data cache on, instead of within blob
03:46:54BZFlagagreed. If someone reports initrd from flash working, then I'll change that part too.
08:44:05mmattenhi
10:19:56seletzmorning
15:24:14erikmseletz: when answering trivial questions on LAK just point to the webpage with the answer
15:38:39prpplaguemorning BZFlag
15:38:45prpplagueBZFlag: hows life at lineo?
15:41:05prpplaguedid any of you get that stupid email from "debbie" at idg.com ?
15:42:17erikmno
15:42:32erikmmaybe it was already filtered by spamfilter, though
15:43:42prpplagueso genius named debbie at idg.com sent of an email "LinuxWorld Highlights" with over 600 ppl on the "to:" line and included a ms-word 2000 doc with all the "highlights"
15:45:00prpplagueapparently idg's email server is getting flamed pretty hard
15:45:02erikmprpplague: check the mail headers if it is really from idg.com and send a complaint to abuse@idg.com
17:48:31BZFlagerikm: you think the initrd handling messed up the kernel boot?
17:48:53BZFlagthe kernel chousl not access that memory until after most all kernel init is done.
17:49:03BZFlagmuch kernel text should show up first.
17:50:54erikmBZFlag: rmk told me that it probably wouldn't work
17:51:16BZFlaghmm... I did not see why it would not, but I'm testing now.
17:53:56BZFlagI changed the rest over, so if you have a board setup you can test, that would be nice. ;-)
18:01:14PlugsIs Tim Riker around?
18:03:24BZFlagyes.
18:04:03BZFlagwhy does it say "resume writing at 0x0003fffc" instead of 0x00040000 ?
18:04:11BZFlagis Tim Riker
18:04:33PlugsHow do I request RT_Netlink gets configured into the zaurus kernel
18:04:49erikm_dinnerBZFlag: because that makes the test a bit easier. doesn't matter, doesn't harm the flash
18:05:09erikm_dinnernow really goes to dinner
18:05:29BZFlagjust curious. the first one is ok, so ew should not touch the preceeding partition.
18:05:39BZFlagPlugs: in #Zaurus, k?
18:07:02PlugsI'll go there, thanks
18:22:03BZFlaghmm... first initrd boot worked fine from flash.
18:22:14BZFlagsecond hangs after kernel decompression.
18:22:22BZFlaghard reboot does not fix it.
18:22:41BZFlagsending the kernel over again....
18:25:08BZFlagwhat the crap? Now I can't boot.
18:28:17BZFlaghmm.... perhaps I did not reboot after sending over the initrd so it used the one in memory?
18:28:25BZFlagchecks
18:41:16BZFlaghumpf. /me reverts most systems to ram. =(
18:43:02erikmBZFlag: it *looks* like the kernel crashed somewhere early
18:43:19erikmBZFlag: but because the console driver isn't yet initialised you won't see the output
18:43:49erikmBZFlag: the output gets stored in the printk() buffer, and is actually printed as soon as the console driver is initialised
18:44:13BZFlagconsole dump happens after initrd load?
18:44:25BZFlagI though it happened when the console got setup.
18:44:32BZFlagwhich shouldbe before initrd.
18:45:12BZFlag"boot noinitrd" still shows no output.
18:45:23erikmno, but ATAG params are parsed before the console system is initialised
18:47:42BZFlagparsing should be ok, no?
18:47:55erikmyes, but I don't know what happens next
18:54:22BZFlagtrying a kernel without initrd support.
19:02:36BZFlaghmm.. yep, turn off initrd support, and the atag entry does not kill the kernel. (of course the initrd address should never be accessed either)
19:03:11erikmhmm, IIRC it was an issue with ramdisk
19:03:26erikmrmk mumbled something about solving in 2.5
19:03:34erikmoh, btw, we want to go to OLS
19:03:56BZFlagOLS?
19:04:03erikmlxrbot: ols?
19:04:04lxrbotols is the Ottawa Linux Symposium, see http://www.linuxsymposium.org
19:04:05BZFlagis acronym fried.
19:04:16BZFlagI'd like to go too.
19:04:46erikmOLS is one of the two technical linux conferences (the other one is in sydney)
19:05:28erikmnicolas pitre will be coming, I plan to go, dwmw2 as well, rmk also, jeff sutherland, deepak saxena
19:05:58BZFlagnice
19:06:38erikmcalls jdb
19:12:54BZFlaghmm... from rd.c free_initrd_mem(initrd_start, initrd_end);
19:55:41prpplagueerikm: can i pester you for some urls or reference material on interfacing lcd to SA-11xx ?
20:02:08BZFlagAwayassabet or tuxscreen schmatics?
20:02:41erikmprpplague: abraham vd merwe also has more information
20:03:40BZFlaghttp://developer.intel.com/design/edk/product/strongarm_edk.htm or http://TuxScreen.net/download/schematics/
20:27:18prpplaguei wish i had time to go to OLS
20:29:51prpplagueerikm: is merwe someone on the lart project?
20:30:34erikmno, though he has a LART
20:31:04erikmprpplague: (and his last name is "vd merwe", where the vd is short for "van der")
20:32:31erikmprpplague: (quite a common construction for dutch last names, it means "of", c.f. "von" in german)
20:33:23erikmprpplague: (abraham is from south africa, but south african is just like 16th century dutch)
20:34:49prpplagueerikm: ahh
20:35:17prpplagueerikm: you we yankees aren't educated in such things, lol
20:35:28prpplagues/you/you know
20:35:39prpplaguelol, or how to type,lol
20:36:11erikmprpplague: yeah, I noticed. which *is* a strange thing, if you think about it. the US is an immigration country
20:36:42erikmprpplague: for example: the average US form can't live with more than two first names
20:37:12erikmprpplague: and the idea that your first name can be different from your first given name is completely unknown
20:37:12prpplaguelol, ya that is true
20:39:05prpplaguewell, i'm totaly frustrated, i'm having trouble pulling together and ARM based system
20:39:30prpplaguesince i live in the middle of no-where there isn't much emedded buddies around to work with
20:40:56erikmheh, but do you think there are many around over here?
20:41:34prpplaguewell, atleast your in an academic setting which is more conducive to learn this type of stuff
20:42:18prpplagueerikm: i feel like i'm walking in mud up to my knees learning this stuff, every step is a big effort
20:45:00erikmI do live in a 100.000 people city, but there aren't many embedded people over here
20:45:13erikmbut at least jdb is cluefull
20:46:03erikmand we also learned quite a lot from discussions with a professor from physics who build his own supercomputers for n-body simulations :)
20:47:07prpplagueya jdb seems very intelligent
20:48:43Russprpplague: besides online help, I worked completely alone
20:49:18erikmRuss: yeah, me usually as well
20:49:29erikmprpplague: mailing lists help a lot
20:50:01erikmprpplague: just lurk on a couple of lists and save the interesting posts for later
20:50:32prpplagueya thats what i mainly do as well as read and study other ppls work
20:51:27Russprpplague: btw, if you want the latest pcb (1.99c) lemme know
20:51:32Russaway
20:51:38erikmRuss: you fixed it?
20:51:57Russfixed what?
20:52:19erikmRuss: you told me some time ago that it had some kind of bug
20:52:21RussI have have various arm related gschem parts
20:52:32Russerikm: harry eaton has been making some releases
20:52:47Russthe bugs that were in the 1.7.2 version are pretty simple to fix
20:53:06erikmchecks his current version
20:53:28erikmdoh, old. 1.7.1ALPHA
20:54:10prpplagueRuss: thanks, when i get this lcd stuff under my belt, i'll give you a /msg
20:54:14erikmI think I should become debian maintainer and take over te PCB stuff :)
20:55:05erikm"resistance is futile, you will be LARTed"
20:55:18prpplaguehaha
20:55:47prpplagueas long as the "LART" drone looks like "7 of 9"
20:56:08erikmheh :)
20:56:36erikmfor quite some people 7 of 9 is the single reason to watch star trek
21:05:04prpplagueerikm: she could assimilate me any time
21:05:34erikmheh :)
21:09:12prpplagueerikm: i've been trying to get in contact with these guys http://www.earthlcd.com/marmalade.htm
21:11:30erikmprpplague:  hmm, they need a bootloader, don't you think? ;)
21:11:43prpplagueerikm: yes i do agree
21:12:13prpplagueerikm: someone said they were going to use some stuff from windriver
21:14:24erikmooh, they have some other cool gadgets
21:16:02prpplagueerikm: ya they have some nice stuff, but they can be a hassel to order from sometimes

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with infobot logs, split per channel and by date, etc.