**** BEGIN LOGGING AT Tue Jul 22 02:59:57 2008 Jul 22 08:24:20 bonjour Jul 22 09:15:48 Jay7: pong Jul 22 09:18:17 ant|work321: kernel can't find init when logical size is increased Jul 22 09:18:42 hmm... Jul 22 09:18:43 now compiled with mtdparts in cmdline and with patched sharpsl Jul 22 09:19:18 I suppose the atag_cmdline is not yet supported... Jul 22 09:19:18 Warning: unable to open an initial console Jul 22 09:19:27 then panic No init found Jul 22 09:19:59 with standard logical is more fun :) Jul 22 09:20:19 all booted properly, but it seems like 'black hole' Jul 22 09:20:23 http://www.rpsys.net/openzaurus/patches/export_atags-r2.patch Jul 22 09:20:41 I have copied to nand 3 files by 128Mb every :) Jul 22 09:20:46 and it's ok.. Jul 22 09:22:40 patch is for 2.6.23 Jul 22 09:22:44 I fear linux-rp removes the cmdline part... Jul 22 09:22:45 I build 2.6.24 Jul 22 09:24:17 patch is also for 2.6.24 Jul 22 09:25:01 but not in 2.6.26 Jul 22 09:25:04 ? Jul 22 09:25:58 http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=blob;f=packages/kexec/kexec-tools-1.101/kexec-arm-atags.patch;h=bf6f64003557db2b4fb01412cb37ec9ec18ac8cd;hb=44ea421ed9b56a58d4ff1941befe61d907452d40 Jul 22 09:26:09 seems it is only for kexec Jul 22 09:26:53 I suppose your cmdline is overwritten with wrong params Jul 22 09:26:57 just guessing Jul 22 09:27:02 ok, what is solution? Jul 22 09:27:27 patch above warns me about already existing atags.c Jul 22 09:27:48 so it seems already in kernel Jul 22 09:27:53 it is... Jul 22 09:28:05 "This is less than ideal as changing the commandline, " Jul 22 09:28:14 initrd and other options would be a useful feature." Jul 22 09:28:33 seems unfinished, but the other patch adds cmdline support... Jul 22 09:28:43 seems work in progress Jul 22 09:28:57 this was in 2007: http://projects.linuxtogo.org/pipermail/openembedded-issues/2007-September/005575.html Jul 22 09:30:52 what should be our next move? Jul 22 09:31:01 so, from the beginning: Jul 22 09:31:02 http://www.nabble.com/Passing-mtdparts-to-the-kernel-to2930591.html Jul 22 09:32:24 if that does not work anymore it is because the cmdline get 'rejected' or filtered Jul 22 09:32:46 by the latest "atag" changes Jul 22 09:32:58 BTW can you check /proc/atag ? Jul 22 09:33:12 now only from emergency image Jul 22 09:33:19 * ant|work321 suspects we need some atag hacker (RP or mwester) Jul 22 09:33:48 no, emergency was 2.6.18 I think (no atag) Jul 22 09:34:48 linux-rp-2.6.24/defconfig-akita:CONFIG_ATAGS_PROC=y Jul 22 09:34:48 atag support was added to 2.6.23 around January IIRC Jul 22 09:34:55 exactly Jul 22 09:34:56 so it should be Jul 22 09:36:23 hm.. Jul 22 09:36:32 if you see, the old way to create cmdline was: Jul 22 09:36:34 http://74.125.39.104/search?q=cache:mPsuwyrp1oYJ:www.openembedded.org/repo/org.openembedded.dev/packages/linux/openzaurus-pxa_2.4.18-rmk7-pxa3-embedix20031107.inc+kernel+mtdparts&hl=it&ct=clnk&cd=10&gl=it Jul 22 09:36:59 is all patches stored in $WRKDIR/patches or only that applied? Jul 22 09:37:22 look in linux-rp_2.6.24.bb Jul 22 09:38:27 now here is no such strings Jul 22 09:40:46 no mtdparts for akita at all Jul 22 09:40:59 in all linux/ dir Jul 22 09:42:00 hm.. I can check strings in kernel Jul 22 09:42:17 he he Jul 22 09:42:41 must be somewhere, unless directly inherited from sharpsl.c Jul 22 09:43:10 sharpsl is patched Jul 22 09:43:32 so it should provide right mdtparts Jul 22 09:47:13 hmm...let's try to bother RP or mwester... Jul 22 09:50:17 Jay7: it's strange because at the time I could boot from flash and SD Jul 22 09:50:43 do you have logical != 7Mb? Jul 22 09:50:51 now yes Jul 22 09:51:00 hm.. Jul 22 09:51:15 but will change later today :-) Jul 22 09:51:45 for the moment I'll give up with that udev issue Jul 22 09:51:57 and come back to u-boot / kexecs Jul 22 09:52:01 and what is your partitions string in u-boot? Jul 22 09:52:11 mtdparts=? Jul 22 09:52:33 was like in 3378 Jul 22 09:52:57 http://bugs.openembedded.net/attachment.cgi?id=5189 Jul 22 09:53:03 this is bootlog Jul 22 09:53:09 with 7mb logical Jul 22 09:53:27 you see the skipped tags at the beginning ? Jul 22 09:53:48 normal boot has only one of them Jul 22 09:54:02 it means u-boot is passing more information Jul 22 09:54:16 (or at least I understand it like this) Jul 22 09:54:45 then is "Using static partision definition" Jul 22 09:55:06 I should go away for some time Jul 22 09:55:12 ~30 min Jul 22 09:55:30 bye Jul 22 11:43:49 Jay7: ping Jul 22 13:17:06 ant|work321: pong Jul 22 13:38:13 ant|work321: what is conclusion about current situation with u-boot and 2.6.24? Jul 22 13:57:12 Jay7: this evening I restart the experiments... Jul 22 14:04:45 first one: flash the kernel from angstrom using nandlogical Jul 22 14:05:46 second one: kernel cmdline tests with mtd1 > 7mb Jul 22 14:11:24 ant|work321: now I shold suspend experiments, but i really need rootfs > 64Mb :) Jul 22 14:11:56 looks like when I boot properly with non-standard mtd1 then I get working rootfs.. Jul 22 14:12:32 working ? Jul 22 14:12:49 df is showing some strange things Jul 22 14:13:02 when changed only root / home partitions Jul 22 14:13:22 cat /proc/mtd Jul 22 14:13:23 I have wrote already at morning about this Jul 22 14:13:42 [13:20] all booted properly, but it seems like 'black hole' Jul 22 14:13:48 yes, you wrote 'kernel can't find init) Jul 22 14:13:54 [13:20] I have copied to nand 3 files by 128Mb every :) Jul 22 14:14:23 no, this about standard logical and changed root size Jul 22 14:14:48 ah, ok Jul 22 14:14:54 kernel can't find init when i change logical too Jul 22 14:15:29 I think it's bound.. Jul 22 14:16:15 *concerned Jul 22 14:17:44 btw, kernel detects right mtd partitions Jul 22 14:18:00 0x0-0x00a00000 - system area Jul 22 14:18:15 0x00a00000 - 0x06e00000 - root filesystem Jul 22 14:18:34 0x06e00000 - 0x08000000 - home Jul 22 14:18:52 in both cases Jul 22 14:19:05 except system area sizes Jul 22 14:20:26 hmmm...init is on /dev/mtd2 Jul 22 14:20:51 I'll try to erase the nand before Jul 22 14:22:39 I'll try add sleep to printk() Jul 22 14:22:55 to check unknown atag's messages Jul 22 14:22:59 you know, booting from SD was failing for the same reason (no init) Jul 22 14:23:13 hm.. kernel is 2.6.24? Jul 22 14:23:14 because it needs some sleep Jul 22 14:23:29 before mounting sd Jul 22 14:24:12 most interesting if before no init message was 'unable to open an initial console' Jul 22 14:24:40 because before 'unable to open' was 'Freeing init memory: 80K' Jul 22 14:25:12 kernel freeing init memory after reading init or before? Jul 22 14:25:59 before...possibly udev related see new logs 4118 Jul 22 14:27:56 some fun on akita: spitz keyboard and corgi touchscreen found in dmesg :) Jul 22 14:28:20 recycling... Jul 22 14:48:46 ant|work321: I found a way to check /proc/cmdline Jul 22 14:49:13 I can left logical as is and resize rootfs Jul 22 14:49:27 then boot and check cmdline Jul 22 14:50:45 hardcoded partitioning is known Jul 22 14:50:49 atags left.. Jul 22 15:37:57 ant|work321: stupid question.. Jul 22 15:38:24 I'm looking now into u-boot's akita.h and comparing it with svs57's corgi.h Jul 22 15:38:44 please, ? Jul 22 15:38:45 in akita.h: Jul 22 15:38:48 #define MTDIDS_DEFAULT "nand0=laze-0" Jul 22 15:38:53 #define MTDPARTS_DEFAULT "mtdparts=sharpsl-nand:7168k(Logical),54272k(root),69632k(home)" Jul 22 15:39:00 in svs57's corgi.h Jul 22 15:39:05 #define MTDIDS_DEFAULT "nand0=laze-0" Jul 22 15:39:10 #define MTDPARTS_DEFAULT "mtdparts=laze-0:10240k(Logical),-(JFFS2)" Jul 22 15:39:26 should MTIDS be equal to first part of MTDPARTS? Jul 22 15:39:44 laze-0 != sharpsl-nand Jul 22 15:40:27 moment Jul 22 15:40:48 standard for akita should be +define MTDPARTS_DEFAULT "mtdparts=sharpsl-nand:7168k(Logical),59392k(root),64512k(home)" Jul 22 15:41:10 btw, +#define :) Jul 22 15:41:20 ok, fixed Jul 22 15:41:24 :-) Jul 22 15:41:44 I'll put a mtn diff of my u-boot layout later Jul 22 15:42:16 ok, but I'll try with 'laze-0' name :) Jul 22 15:43:21 Jay7: I'm looking at +/* Note: fake mtd_id used, no linux mtd map file */ Jul 22 15:43:32 btw, in default corgi.h, names also equals Jul 22 15:43:57 laze is the name of a developer :-) Jul 22 15:44:05 :)) Jul 22 15:44:55 the only map file is sharpsl.c Jul 22 15:45:25 so here it differs from it Jul 22 15:45:57 hey, mwester is on #oe ! Jul 22 15:46:05 let try to fish him too? Jul 22 15:46:41 why not? :) Jul 22 15:54:47 ant|work321: btw, entry point and load address for mkimage is same for corgi and akita? Jul 22 15:56:04 yes, it is copied in ram and executed i think Jul 22 15:56:22 but you're right about the #define MTDIDS_DEFAULT "nand0=laze-0" Jul 22 15:56:45 ant|work321: now I've set both to 'sharpsl-nand' Jul 22 15:57:17 I'll reflash in 5 min :) Jul 22 15:57:22 sorry again, never tried personally on akita/spitz Jul 22 15:58:27 btw, svs57 says at hpc.ru about patched nandlogical Jul 22 15:59:13 patched nandlogical should 'flash with new partitions sizes' Jul 22 15:59:23 what he can mean? :) Jul 22 16:00:32 mhh, you mean pdaxrom nandlogical? Jul 22 16:00:44 for 2.6.1x ? Jul 22 16:00:58 ant|work321: may be.. here is no version Jul 22 16:01:06 or our one for 2.6.2x? Jul 22 16:01:12 it's thread about debian with eabi Jul 22 16:01:36 I suppose we should full-erase / full overwrite mtd1 Jul 22 16:01:45 to cancel any garbage Jul 22 16:02:48 http://www.translate.ru/url/tran_url.asp?lang=ru&url=http%3A%2F%2Fwww.hpc.ru%2Fpda%2Fboard%2Findex.php%3Ft%3D121323%26start%3D240&direction=re&template=General&cp1=NO&cp2=NO&autotranslate=on&psubmit2.x=61&psubmit2.y=2 Jul 22 16:03:00 here is autotranslated version Jul 22 16:03:54 how we can erase mtd1? Jul 22 16:04:01 full nand erase from D+M? Jul 22 16:04:34 no, I don't think it is necessary Jul 22 16:04:52 I'm flashing Jul 22 16:04:52 I'd fill ('format') using nandlogical Jul 22 16:05:02 hm Jul 22 16:05:12 I can do it right now Jul 22 16:05:24 the code is in full_updater.sh Jul 22 16:05:33 oops, full_autoboot.sh Jul 22 16:06:45 offset 0x0, and size? Jul 22 16:07:02 0x700000? Jul 22 16:10:03 sorry, phone call, I'm back Jul 22 16:10:31 nandlogical /dev/mtd1 WRITE 0x0 0x700000 /dev/zero Jul 22 16:10:46 now I have 7Mb logical in u-boot Jul 22 16:11:23 yes, definetely 0x00700000 Jul 22 16:11:36 ok, doing it :) Jul 22 16:13:22 booting ok.. Jul 22 16:13:32 mdelay seems not working :) Jul 22 16:13:47 oh Jul 22 16:13:58 ah, I set mdelay(100) Jul 22 16:14:09 but it should take delay anyway.. Jul 22 16:14:41 Jay: the dmesg and /proc/cmdline will tell... Jul 22 16:15:00 yes Jul 22 16:15:16 Ignoring unrecognised tag 0x0000000 Jul 22 16:15:19 4 times Jul 22 16:15:28 ok, got it too Jul 22 16:15:32 :) Jul 22 16:15:48 Using static partitions definition Jul 22 16:15:53 these must be misinterpreted atags Jul 22 16:16:01 and 3 default mtdblock's Jul 22 16:17:26 do you see mtdparts in the cmdline? Jul 22 16:17:41 no, default cmdline from u-boot Jul 22 16:17:57 without mtdparts Jul 22 16:18:17 console=ttyS0,115200 console=tty1 root=/dev/mtdblock2 rootfstype=jffs2 fbcon=rotate:1 Jul 22 16:18:30 so defaulting to sharpslc (static) Jul 22 16:18:33 yes Jul 22 16:18:55 and I have not patched sharpsl.c, so having default partitioning Jul 22 16:19:17 perhaps this old version of u-boot puts the atag in the old/wrong memory address Jul 22 16:19:18 how we can resolve unrecognised tag's? Jul 22 16:19:33 so these are 'ignored' Jul 22 16:19:54 oh.. patching needed Jul 22 16:20:00 it must be something very Zaurus-specific Jul 22 16:20:17 can I just write mtdparts to default cmdline? :) Jul 22 16:20:19 even RP cannot remember... Jul 22 16:20:28 setenv bootargs .... mtdparts... Jul 22 16:21:21 btw.. Jul 22 16:21:37 when I flash /dev/mtd2 from emergency Jul 22 16:21:50 it works with default partitions Jul 22 16:22:11 then I reboot with other kernel and other partitioning Jul 22 16:22:20 how it can works at all? :) Jul 22 16:22:43 the emergency kernel has hardcoded partitioning too... Jul 22 16:23:01 only 2 Jul 22 16:23:11 ? Jul 22 16:23:59 should check Jul 22 16:24:08 mtd1 and mtd2 are 'different' Jul 22 16:24:21 because of wear leveling feature for mtd1 Jul 22 16:24:30 so I suppose if we extend mtd1 Jul 22 16:24:54 over the boundaries, it could need to be erased Jul 22 16:25:10 while mtd2 can be written with nandwrite Jul 22 16:25:29 now, the doubt is: do nandlogical knows about boudaries? Jul 22 16:25:33 how? Jul 22 16:25:42 yes, but how nandwrite detect new mtd2 boundaries? Jul 22 16:25:49 it needs mtd-utils Jul 22 16:25:52 if we nandwrite from emergency Jul 22 16:26:10 but after normal boot there is new partitioning Jul 22 16:26:11 perhaps we should recompile nandlogical after changing partitioning Jul 22 16:26:34 ah, I see your point Jul 22 16:26:36 grr.. Jul 22 16:26:52 so, we need right emergency :) Jul 22 16:27:16 or nandwrite with directly specified offset/size Jul 22 16:27:54 but if you could write 128 mb file... Jul 22 16:28:13 3x :) Jul 22 16:28:33 hm.. Jul 22 16:28:47 zaurus does'nt go into emergency.. Jul 22 16:29:01 normal boot started Jul 22 16:30:44 I have erased emergency :)) Jul 22 16:30:55 ah, yes Jul 22 16:31:18 zaurus is unkillable thing :) Jul 22 16:31:41 last hope: Fn+D+M Jul 22 16:32:01 :) Jul 22 16:32:13 now just reflash u-boot and emergency from pdaxrom Jul 22 16:32:32 if you have the cmdline in the kernel, ok Jul 22 16:32:48 but really should work straight in u-boot Jul 22 16:34:20 Jay7: you could flash from angstrom if you include nandlogical in rootfs Jul 22 16:35:01 I know, but I testing with just console-image Jul 22 16:35:12 that's why one single kernel could do emergency, menu-boot and kexec a new kernel in mtd2 or SD or NFS... Jul 22 16:35:40 I think I can pack a super console image in the 7mb Jul 22 16:36:13 I'd use the initramfs-bootmenu-image for this 'universal' kernel Jul 22 16:36:29 I have reflashed with default pdaxrom files and replaced u-boot :) Jul 22 16:36:51 btw, here is same unrecognised tag Jul 22 16:37:15 yes, one line is 'usual', 4 is not ;-) Jul 22 16:37:59 and really, here is two mtd parts Jul 22 16:38:12 one 0x0 - 0x00700000 Jul 22 16:38:26 -#define MTDPARTS_DEFAULT "mtdparts=laze-0:7168k(Logical),-(JFFS2)" Jul 22 16:38:32 and second 0x00700000 - 0x08000000 Jul 22 16:38:32 ^^ Jul 22 16:39:24 nandwrite have -s switch Jul 22 16:39:33 'set start address' Jul 22 16:39:35 default is 0 Jul 22 16:40:21 well, it should be relative to mtd2 Jul 22 16:40:33 so we can calculate new address Jul 22 16:40:37 hm.. Jul 22 16:40:39 0 = 7168 offset Jul 22 16:41:30 wait, I remember I'v seen the sources Jul 22 16:46:58 canot find it anymore... Jul 22 16:47:26 it was some pdaxrom updater Jul 22 16:48:36 now I want to specify mtdparts in setenv bootargs Jul 22 16:48:49 then reflash with nandwrite -s Jul 22 16:49:17 no -s :) I have standard logical now Jul 22 16:56:50 Jay7: I'm not a nand expert but I think the 'blocks' are written differently in mtd1 (block wear levelling). Then there is the thing 'badblocks', very complicated... Jul 22 16:57:44 yes, I know Jul 22 16:57:58 I'll use nandwrite on mtd2 Jul 22 16:58:53 but not now Jul 22 17:03:50 now it takes mtdparts from cmdline Jul 22 17:04:41 3 cmdlines partitions found on MTD device sharpsl-nand Jul 22 17:07:53 but there is the same problem Jul 22 17:08:13 i can create files bigger than fs size.. Jul 22 17:08:31 dd if=/dev/zero of=/testfile bs=1M count=200 Jul 22 17:09:03 du -sx / => 221308 :) Jul 22 17:09:51 df / => 102400 17376 85024 17% / Jul 22 17:10:03 ant|work321: how it can be? :) Jul 22 17:10:23 well Jul 22 17:10:26 scary Jul 22 17:11:19 may be initrd shoul be fixed? Jul 22 17:11:43 did youreflash only kernel? Jul 22 17:12:03 initrd too Jul 22 17:12:48 now you have standard mtd1 (7m) and cmdline set in u-boot isn't? Jul 22 17:13:25 which u-boot? pdax? Jul 22 17:13:28 setenv bootargs console=ttyS0,115200 console=tty1 root=/dev/mtdblock2 rootfstype=jffs2 mtdparts=sharpsl-nand:7168k(Logical),102400k(root),-(home) fbcon=rotate:1; Jul 22 17:14:04 self compiled u-boot Jul 22 17:15:07 ok, only root resized Jul 22 17:15:15 in sharpsl.c too? Jul 22 17:15:26 no, sharpsl left as is Jul 22 17:15:44 kernel takes mtdparts and uses them Jul 22 17:15:59 info from sharpsl.c should not be considered then Jul 22 17:16:05 yep.. Jul 22 17:16:19 so it somehow can read u-boot args Jul 22 17:16:43 via cmdline Jul 22 17:16:52 as shown abowe Jul 22 17:19:15 may be change ROOT_FLASH_SIZE in conf? Jul 22 17:19:46 hmm...don't think so Jul 22 17:19:52 and rebuild console-image.. Jul 22 17:20:30 who on #oe can resolve such problem? Jul 22 17:20:48 I was been told this param is only used by opie (or gpe)? Jul 22 17:21:26 Jay7: cat /proc/mtd ? Jul 22 17:22:05 mtd0: 006b0000 00800000 "Boot PROM Filesystem" Jul 22 17:22:07 and what is in /proc/atags? Jul 22 17:22:29 mtd1: 00700000 0002000 Logical Jul 22 17:22:54 mtd2: 06400000 00020000 root Jul 22 17:23:06 oops, 0002000 Jul 22 17:23:29 mtd3: 01500000 002000 home Jul 22 17:23:36 hm.. looking strange Jul 22 17:23:40 yes Jul 22 17:23:45 erasesize is equal Jul 22 17:24:23 atags is same as cmdline with AT????' AT prepended Jul 22 17:24:39 and ?root? appended Jul 22 17:24:46 ?root@ Jul 22 17:25:04 ? is squares Jul 22 17:27:05 can you check what is normal values for /proc/mtd? Jul 22 17:27:15 with default partitioning Jul 22 17:27:26 sure, later Jul 22 17:28:11 ROOT_FLASH_SIZE does not matter as grep shows :) Jul 22 17:28:20 I'm looking in that nand.xls Jul 22 17:28:24 used only in IMAGE_LINGUAS Jul 22 17:28:39 ugly naming btw Jul 22 17:28:54 not clear Jul 22 17:29:54 ok, the output of /proc/mtd is: mtdx offset size Jul 22 17:30:02 offset before Jul 22 17:30:35 dev: size erasesize name Jul 22 17:30:51 header from my /proc/mtd Jul 22 17:31:07 root should have an offset of 0x0070000 Jul 22 17:31:34 hmm.. Jul 22 17:31:44 did you invert the params? Jul 22 17:32:10 I've seen some other syntax for mtdparts Jul 22 17:32:10 no, as is Jul 22 17:32:30 e.g. 123456@65332 Jul 22 17:33:01 I hope u-boot and kernel could have different syntax Jul 22 17:33:09 I fear Jul 22 17:33:13 not hope! Jul 22 17:33:31 :) Jul 22 17:33:38 I need a break... Jul 22 17:34:00 going home now Jul 22 17:34:08 will collect some data for you Jul 22 17:34:11 cu Jul 22 17:35:24 bbl Jul 22 17:35:27 ok, bbl Jul 22 20:37:58 Jay7: re Jul 22 20:38:05 ant_: hi Jul 22 20:38:12 ant_: it works :) Jul 22 20:38:16 :-] Jul 22 20:38:23 hrw save my brain :) Jul 22 20:38:31 yes, compression... Jul 22 20:39:01 so I stop now with experiments up to 4 august Jul 22 20:39:12 I'm revising the u-boot patch, what do you think about the name of mtd? Jul 22 20:39:17 laze-0 ??? Jul 22 20:39:35 s/mtd/nand/ Jul 22 20:39:49 I use sharpsl-nand in both places Jul 22 20:40:06 should we leave sharpsl-nand for compatibilty? Jul 22 20:40:07 ok Jul 22 20:40:19 agrd Jul 22 20:40:29 done Jul 22 20:40:50 so we should resolve ATAG's problem later.. Jul 22 20:41:00 I'll test again the u-boot-zaurus recipe and replace the old attachment Jul 22 20:41:40 ok, I've update later Jul 22 20:41:56 it would be possible to automate the kernel+uImage creation Jul 22 20:42:09 without sharp's headers Jul 22 20:42:37 but yes, we should check the atag again Jul 22 20:42:55 it is somehow obscure to me Jul 22 20:43:04 where it is filtered Jul 22 20:43:45 Jay7: if you have time, please investigate on initramfs-bootmenu-image Jul 22 20:43:56 it is 98% done Jul 22 20:44:05 if someone can test u-boot 1.1.4 on other board Jul 22 20:44:06 with ATAG's Jul 22 20:44:06 and with klibc is very small Jul 22 20:44:31 it would be the base for the new emergency-image Jul 22 20:44:34 ant_: after 4 august only Jul 22 20:44:38 np Jul 22 20:44:52 it is simply this: the script uses 'expr' Jul 22 20:45:03 we going to single kernel :) Jul 22 20:45:04 we don't want perl in initramfs..! Jul 22 20:45:22 I'm thinking how to rewrite this file: Jul 22 20:45:33 sh + dialog? Jul 22 20:45:38 or cdialog Jul 22 20:45:59 http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=blob;f=packages/initrdscripts/files/30-bootmenu.sh;h=c77bd662afa1145bbd2023477816d59fe19f3ffa;hb=HEAD Jul 22 20:46:07 we don't have ncurses Jul 22 20:46:17 nor glibc, nor busybox Jul 22 20:46:37 I've seen other distros are enabling 'awk' Jul 22 20:46:44 in initramfs-images Jul 22 20:47:04