irclog2html for blob on 2001.09.28

00:33:48sammymorning ...
04:07:39sammyfeel a little confuse ...
04:09:35sammybefore I flash the kernel and ramdisk to flash , what things I must setup ?
04:10:24sammythen the blob can know where my kernel and ramdisk are ...
04:12:20sammyRuss ? BZFlag ?
04:13:07sammywhat's the point hint ?
05:08:44sammy...
05:13:38BZFlagsammy: there are some platform defines in blob for those locations and the mtd setting in the kernel too.
05:14:23BZFlagthe mtd layer will only be used to rewrite the kernel/rootfs/bootloader, so that might not be importatnt.
05:15:14BZFlagthe kernel gets loaded into ram by the bootloader and so does an initrd (if any). The location of an initrd in ram needs to be passed to the kernel.
05:15:54BZFlagon the tux (and many others) mtd is setup the same as blob for simplicity.
05:16:55BZFlagif you do root on an mtd device ie: jffs2, then you need to setup blob and mtd for that partition and be sure to pass the root device number to the kernel.
05:17:13BZFlagthis is generally the preferred procedure.
05:17:32sammybut the case in LART , there is a choos when blob type help , what's flash use to do ?
05:22:11BZFlag"download" copies over the serial line. "boot" boots whats in RAM. "flash" saves what is is ram to flash so it can get reloaded from flash to ram after a reboot.
05:22:24BZFlagthe contents of ram are lost on reboot.
05:26:14sammyAlright . is suppose that I compiler the new kernel image , I want update my kernel in LART ...
05:26:57sammyfirst I must boot from minicom , then ?
05:27:33sammywhat's next ?
05:28:20sammystop in blob ?
05:30:24BZFlagright. then "download kernel" and send the kernel with uuencode in another window (do not exit minicom, start uuencode at a different shell prompt on the same machine)
05:30:47BZFlagthen "flash kernel" to put it in flash (if you want) and "boot" to test it.
05:31:13BZFlagyou might want to test it a few times before flashing, but I normally just flash it every time.
05:32:23BZFlagthen the same procedure for a "ramdisk". If it is a jffs2 image, then you must flash it. if it's a compressed minix or ext2 image then you can run it from ram without flashing.
05:33:48sammynot clear in "send the kernel with uuencode in another window"...
05:36:36BZFlagread http://TuxScreen.net/wiki/view//InstallingLinux
05:36:53BZFlagyou are running minicom in an Xterm or similar?
05:37:08sammyyes.
05:37:14sammyXterm
05:37:41BZFlagthen leave that open, and open another xterm. in the new xterm us where you run uuencode after switching baud rates in minicom.
05:38:00BZFlags/us/is/
05:38:45sammywhen I type download kernel it show..
05:38:52sammyblob> download kernel
05:38:53sammySwitching to 115200 baud
05:38:54sammyYou have 60 seconds to switch your terminal emulator to the same speed and
05:38:55sammystart downloading. After that blob will switch back to 9600 baud.
05:38:56sammy.......ëK...ò
05:39:39sammyso now my speed is 9600 , need to change to 115200, now
05:40:08sammyright ?
05:40:33sammyand leave it alone , and turn another xterm(next)
05:41:53BZFlagyes, change speed first. in minicom that's ctrl-a, p, i.
05:42:39BZFlagbut you only have 60 seconds to do that and start the upload. else you have to switch back to 9600 and type the download command again.
05:43:00BZFlagbbiab
06:08:38sammyTim , I want to order the Tuxscreen...
06:40:22BZFlagcool.
06:40:35BZFlaghttp://TuxScreen.net/wiki/view//TimRiker
06:44:47sammyif I download the kernel , I must also download the ramdisk too ?
06:57:04BZFlagor have a working ramdisk in flash.
06:58:41sammyif don't do that , then kernel will not find the ramdisk , and can't run it right ?
06:59:49sammybecause , I only download the kernel so when I reboot , it only run to ..
06:59:52sammyINIT: Id "2" respawning too fast: disabled for 5 minutes
06:59:53sammyINIT: no more processes left in this runlevel
07:02:55BZFlagif you go INIT then it did find a root filesystem.
07:03:10BZFlagif you got INIT messages then it did find a root filesystem.
07:03:51BZFlagin this case it looks like your kernel worked, and your root filesystem is a bit messed up.
07:04:20BZFlaggrep ^2 /etc/inittab  to find the line that is failing.
07:06:16sammykmod: failed to exec /sbin/modprobe -s -k char-major-4, errno = 2
07:07:06sammyI download the radmdisk , it show the massage more 30 time's ...
07:08:21BZFlaggot /sbin/modprobe on your ramdisk?
07:08:56sammythen make another ramdisk and download it agail right ?
07:09:14BZFlaghmm... char-major-4 is the tty.
07:09:29BZFlaggot serial console and virtual terminal compiled in?
07:10:28BZFlagmight want to turn off kernel modules support and compile everything static until you at least boot.
07:11:29sammysorry,  but it's still can boot ...
07:12:32sammyjust no idea and less knowledge about why this massage show...
07:15:19BZFlagtry adding "init=/bin/sh" to the kernel command line.
07:21:04sammyis try to find out where to add this command ?
07:23:38BZFlagdo a "boot init=/bin/sh" and that should work. if you have other command line options compiled in, you might want to add those as well.
08:00:19seletzGood morning
08:03:20erikmmorning
08:03:32erikmgets coffee
08:05:04sammymorning erikm
08:06:16sammyBZFlag : I type it , but it show can't fine the bash boot .
08:08:34BZFlagsammy you did init=/bin/sh ?
08:08:51BZFlagthat should be ash or something, not bash.
08:08:57seletzcaffeine refill
08:08:59BZFlagwell, probably anyway.
08:09:22BZFlagack.. if you guys are waking up, I should really get to bed.
08:15:15erikmcalculates
08:15:26erikmyes, BZFlag, it's 2:20 in the morning
08:15:37erikmsame time a I finally got to sleep :)
08:17:32sammyyap, you should need some sleep ...(suggest, if you tired)
08:18:01ed__it's only 5:17 here
08:18:14erikmslaps ed__ into bed
08:18:35ed__ach! das macht spass!
08:19:01sammyed__  morning or evering ?
08:19:10ed__am
08:19:29sammybecause here also 5:17...
08:19:38ed__hooray
08:19:42sammybut is PM..
08:19:48ed__where are you?
08:19:59sammyTaiwan
08:20:06sammyand you ?
08:20:07ed__ahhh
08:20:16ed__US, georgia, atlanta
08:29:02seletzerikm: Hello, got your mail about volatile, barrier() and my goto DONE :^)
08:29:49seletzerikm: my gotos are not for the compile, they are for me. IMHO its easyser to
08:29:54sammyis flash his ramdisk again ...
08:30:18seletzerikm: read when a func has _only_ one exit point.
08:30:23erikmseletz: I disagree
08:30:35erikmseletz: goto is considered dangerous
08:30:55ed__sometimes its ok :P
08:31:00seletzerikm: ahhhh. dijkstra (how do you spell this name)
08:31:01erikmseletz: it's usually bad coding style, and in your cases it definitvely was
08:31:45seletzerikm: _this_ case of goto is _not_ meant in his doc.
08:31:55erikmseletz: the only way goto is considered good is if you want an ordered way to clean up things
08:32:16erikmseletz: like in module initialisation functions in a linux kernel module
08:32:22seletzerikm: mein reden. aerm. yes. thats exactlay what i try to do.
08:33:00seletzerikm: keep only one exit point. after DONE: one can clean up everything neatly.
08:33:13seletzerikm: code complexity goes way down.
08:33:15erikmseletz: Peek() and Poke() didn't clean up anything
08:33:26seletzerikm: yes. agreed.
08:33:38erikmseletz: ret = 2; goto DONE; is evil. you just want return 2
08:33:42seletzerikm: thats because i use macros in vim :^)
08:34:19seletzerikm: no flame wars meant, but _why_ is it evil?
08:35:06erikmseletz: oops, sorry, SIGMANAGEMENT, I have an urgent meeting
08:35:31seletzerikm_meeting: tzhere you go :^D
08:40:15seletzcan someone here explain me the workings of the barrier() macro?
09:26:19seletzcoding
09:41:15sammyerikm: you are back...
09:41:23erikmseletz: the barrier macro is an empty inline assembly statement which fools gcc to think that it makes memory references
09:42:14erikmseletz: but we don't specify what memory reference, so the only thing the compiler can do is to commit all outstanding memory references
09:42:51sammyI use your kernel and ramdisk , but when the LART boot , it's run this about  more then 30 time's then stop and show login : what's the program in this error massage ?
09:42:53sammykmod: failed to exec /sbin/modprobe -s -k char-major-4, errno = 2
09:42:54sammykmod: failed to exec /sbin/modprobe -s -k char-major-4, errno = 2
09:42:55sammyItsy Linux (experimental)
09:42:56sammyKernel 2.0.30 on sa1100/brutus
09:42:57sammylogin:
09:43:26erikmignore the kmod errors
09:44:00sammywhat's that ?
09:44:41erikmit tries to open /dev/tty0 which is not there
09:46:25sammythis is serial important ...as you say before , if I don't need to know how the  kernel work , if I get the error massage how can I know what's problem with that ?
09:47:26sammyif the kernel is stop at place like this ...
09:47:28sammyINIT: Id "1" respawning too fast: disabled for 5 minutes
09:47:29sammykmod: failed to exec /sbin/modprobe -s -k char-major-204, errno = 2
09:47:30sammykmod: failed to exec /sbin/modprobe -s -k char-major-204, errno = 2
09:47:31sammyINIT: Id "2" respawning too fast: disabled for 5 minutes
09:47:32sammyINIT: no more processes left in this runlevel
09:48:13erikmyou are using an incompatible ramdisk-kernel compbination
09:48:24erikmchar-204 is /dev/ttySA0
09:49:26erikmmy ramdisk-new-tty with linux-2.4.9-ac9-rmk1-np1 is a known good combination
09:57:21sammywhat is a secret of success to make my own ramdisk to match the kernel ?
09:57:44erikmeasy: use the correct serial ports
09:58:19sammyblurred ...@_@
10:06:00sammyreload again ...
10:14:32sammynow flash it again...
10:16:38sammyerror....
10:16:49sammyEXT2-fs error (device ramdisk(1,0)): ext2_write_inode: unable to read inode block - inode=1161, block=8202
10:16:50sammyattempt to access beyond end of device
10:16:51sammy01:00: rw=0, want=8204, limit=8192
10:16:52sammyEXT2-fs error (device ramdisk(1,0)): ext2_write_inode: unable to read inode block - inode=1172, block=8203
10:16:59sammyStop ....Why ?
10:19:42erikmbecause of ext2 errors
10:19:56erikmso your ramdisk image got somehow corrupted
10:52:52sammyyou know . that I only change your source and nothing add , but just rebuild again ...so strange
11:02:11seletzhi all
11:02:32seletzcoding
12:31:28prpplaguemorning all
12:32:19seletzhi
12:32:45seletzcoding
13:19:27sammy_night all...
13:19:38seletzgoodnight
13:49:57seletzwell, i'm off too. memory test from Charles Cazabon now runs, we'll see if i
13:50:13seletzhave a mem fault on monday :)
13:50:27erikmok, bye
13:50:36prpplaguehave a good weekend
13:50:54prpplagueoops, alittle slow this morning
13:50:59erikmhehe
13:51:20erikmjust filled the bug and feature request page on the blob sourceforge page
13:52:15prpplagueerikm: should create a seperate patch for blob as secondary bootloader, or should i be work on integrating the code into the standard base?
13:52:24prpplagues/should/should i
13:52:49erikmI think it should be nice to have in the standard code base
13:53:09erikmthough I think it is only a differend first stage loader, right?
13:53:33erikms/differend/different/
13:54:10prpplagueerikm: ya that and a couple changes in the default DEFINES
13:54:45erikmwould be nice enough to have
13:55:17prpplagueerikm: ok then, thats where i'll head
14:45:22erikmhmm, a lart without blob-2.0.4, and it was even lying on my desk.... shame
14:47:35erikmreflashes
14:55:12erikmah, much better
14:58:54erikmhi Russ
14:59:16prpplaguewelcome back Russ
15:18:21prpplaguewell my tuxscreen just arrived...
15:18:46prpplagueit makes my cisco phone look antique
15:18:59erikmhehe
16:06:52prpplagueis away: lunch
17:29:35prpplagueis back (gone 01:22:42)
17:58:48erikmRuss: you there?
18:03:34erikmRuss: anyway, I added a couple of bugs and patches I know about on the sourceforge project pages
18:04:03erikmRuss: is it OK if I assign the parameter block support to you?
18:04:14erikmRuss|werk: or to Russ|werk? :)
18:10:35Russ|werkyah
18:10:54Russ|werkhave you tried 2.4.9-ac10-rmk2-(np1)
18:11:21erikmnot yet
18:11:30erikm-ENOTIME last week
18:11:46Russ|werkmount gets stuck on the 'D' state for me in checkroot.sh
18:12:02Russ|werkdon't know if its teh kernel, or something I did to it
18:12:14Russ|werkhaven't seen anything in linux-arm-kernel
18:13:30erikmno, me neither
18:19:07erikmRuss|werk: could you sysrq-t the system and decode the address to see where it is stuck?
18:20:09erikmRuss|werk: and post results to linux-arm-kernel?
18:20:27erikmRuss|werk: or directly to rmk
18:22:44erikmhi tim
18:23:21Russ|werkerikm: which address would I look at?
18:24:28erikmthe stuck mount
18:25:08erikmRuss|werk: it should be hanging somewhere in kernel, and it's nice to know where exactly
18:25:36Russ|werkthen I can look that up in my symbols?
18:25:56erikmyes
18:37:04erikmgoes home
18:37:08erikmbye
18:37:14prpplaguelater
19:27:48prpplaguehowdy everyone
19:28:08prpplagueBZFlag: got my tuxscreen today!
19:44:14BZFlagprpplague: sweet. the hacking begins. ;-)
23:53:16prpplaguelater all

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.