**** BEGIN LOGGING AT Sun May 31 02:59:59 2015
May 31 04:27:25 <L29Ah> Loading ui!
May 31 04:27:25 <L29Ah> zsh: segmentation fault  LANG=en snappy *
May 31 04:27:25 <L29Ah> nice one snappy
May 31 05:22:48 <KotCzarny> :)
May 31 15:18:22 <NIN101> new rescueOS version http://n900.quitesimple.org/rescueOS/rescueOS-1.2/ 
May 31 15:20:37 <kerio> nice
May 31 15:22:18 <kerio> NIN101: are you sure about "rootdelay" without a parameter
May 31 15:22:26 <kerio> shouldn't it be rootwait
May 31 15:23:43 <NIN101> well, it works :-).
May 31 15:24:06 <kerio> and what's the point?
May 31 15:24:55 <bencoh> oh ... yay, thx :-)
May 31 15:25:07 <KotCzarny> :)
May 31 15:26:41 <bencoh> "NEW kernel rebuilt (gnoutchd)" anything added to kernel ?
May 31 15:30:30 <KotCzarny> nin101: feature idea: virtual terminals and switching via ctrl-right/left
May 31 15:31:28 <NIN101> kerio: there was probably a reason it was added back  then (so it boots),  or it is a leftover of something. it can't hurt. 
May 31 15:31:37 <NIN101> bencoh: nothing added.
May 31 15:32:16 <NIN101> KotCzarny: hmm.
May 31 15:32:42 <kerio> NIN101: i would've called the kernel rescueOS-1.2.zImage or something
May 31 15:33:11 <KotCzarny> more like rescueOS-n900-1.2.zImage
May 31 15:33:49 <NIN101> ok, I'll rename it.
May 31 15:38:05 <kerio> it works without rootwait
May 31 15:38:09 <kerio> do you even need the root= parameter
May 31 15:38:34 <KotCzarny> you always need the root= parameter
May 31 15:38:42 <kerio> but
May 31 15:38:43 <KotCzarny> unless it was known at the build time
May 31 15:38:45 <kerio> there's an initrd
May 31 15:38:57 <KotCzarny> initrd != initramfs
May 31 15:39:26 <kerio> hold on why is it needed to explicitly unmount the maemo root?
May 31 15:39:43 <kerio> KotCzarny: yes but nothing should be mounting any root fs here
May 31 15:39:49 <NIN101> who says it's needed?
May 31 15:39:49 <kerio> ...right?
May 31 15:40:10 <NIN101> it should be taken care of on shutdown
May 31 15:40:32 <KotCzarny> kerio, if you dont specify any root= option kernel would just hang because it wouldnt find init
May 31 15:40:51 <kerio> NIN101: i meant automatically
May 31 15:40:55 <KotCzarny> unless you supply cmd line at the build time
May 31 15:40:57 <kerio> as in the kernel would deal with it
May 31 15:42:32 <NIN101> kerio: not sure I follow, why should the kernel unmount things automatically?
May 31 15:42:47 <kerio> why shouldn't it?
May 31 15:43:46 <NIN101> kerio: write your suggestion to lkml.
May 31 15:43:48 <NIN101> :-)
May 31 15:43:51 <kerio> alright
May 31 15:43:56 <kerio> anyway, no need for any magic cmdline apparently
May 31 15:43:59 <kerio> unless it was compiled in
May 31 15:44:17 <kerio> oh rofl
May 31 15:44:22 <KotCzarny> cat /proc/cmdline
May 31 15:44:30 <kerio> poweroff caused a panic
May 31 15:44:38 <kerio> due to missing root
May 31 15:45:47 <NIN101> hmm?
May 31 15:46:04 <KotCzarny> dont mind him, he is using it in the way its not supposed to be
May 31 15:46:14 <NIN101> yeah 
May 31 15:46:48 <kerio> well, the initrd exited
May 31 15:47:03 <KotCzarny> good for funsies and learning, but otherwise not important
May 31 15:47:07 <kerio> so the kernel tried to keep booting
May 31 15:47:37 <KotCzarny> kerio, ever tried kill -9 1 ?
May 31 15:47:49 <kerio> it doesn't seem like a good idea
May 31 15:47:56 <kerio> wait nvm that's totally harmless
May 31 15:47:57 <KotCzarny> its just the same thing
May 31 15:47:59 <L29Ah> yeah it must be the same
May 31 15:48:01 <kerio> init can't be SIGKILLed
May 31 15:48:17 <KotCzarny> everything can be killed via kill -9
May 31 15:48:21 <KotCzarny> unless its a zombie
May 31 15:48:28 <L29Ah> kerio: do this on your current machine then ;p
May 31 15:49:05 <kerio> the linux kernel will not deliver any fatal signal to init, apparently
May 31 15:49:08 <kerio> which includes SIGKILL
May 31 15:49:16 <L29Ah> it delivers SIGSEGV pretty well :3
May 31 15:49:57 <KotCzarny> kerio, just try it
May 31 15:50:14 <kerio> hold on, what *is* ram0
May 31 15:50:15 <KotCzarny> save your work tho
May 31 15:50:28 <KotCzarny> seriously, you are asking newbie questions
May 31 15:50:30 <KotCzarny> :)
May 31 15:50:47 <kerio> wow that's not condescending at all
May 31 15:51:06 <KotCzarny> ram* is a ramdisk block device
May 31 15:51:19 <kerio> ...yes, i figured as much even without reading the documentation
May 31 15:51:21 <KotCzarny> can be created and destroyed dynamically
May 31 15:51:25 <kerio> what is it used for in rescueos?
May 31 15:51:30 <KotCzarny> for initrd
May 31 15:51:57 <KotCzarny> and differs from initramfs in that way, that its seen as a block device, and not a plain fs
May 31 15:52:37 <KotCzarny> so you have to prepare it via mkfs earlier
May 31 15:52:55 <KotCzarny> initramfs is just cpio archive glued to kernel
May 31 15:53:15 <kerio> yeah i was just missing the fact that "root=/dev/ram0" is actually special-cased in the kernel
May 31 15:53:22 <kerio> to mean "mount the initrd as the only root"
May 31 15:53:43 <KotCzarny> you cant have more than one root at a time
May 31 15:53:50 <KotCzarny> you can use pivotroot etc
May 31 15:53:52 <L29Ah> how is it special?
May 31 15:53:58 <KotCzarny> but there is alsways only one /
May 31 15:54:07 <L29Ah> there's chroot ;)
May 31 15:54:27 <KotCzarny> chroot is just an emulation ;)
May 31 15:54:28 <kerio> L29Ah: the initrd, if present, is *always* mounted as / on the early phases of booting
May 31 15:54:34 <L29Ah> KotCzarny: you can't tell
May 31 15:54:47 <KotCzarny> l29ah: cd /proc/1/cwd, done
May 31 15:54:51 <KotCzarny> or something
May 31 15:54:56 <KotCzarny> i can tell
May 31 15:55:04 <kerio> it will then do its own shit and eventually it should mount the actual root and pivot_root it
May 31 15:55:18 <L29Ah> lrwxrwxrwx 1 root root 0 May 31 15:55 /proc/1/cwd -> /
May 31 15:55:18 <kerio> and then exec*() the new init
May 31 15:55:19 <L29Ah> it lies
May 31 15:55:28 <KotCzarny> no, not ls, do cd, then do ls
May 31 15:55:45 <L29Ah> whoa
May 31 15:55:48 <L29Ah> sick shit
May 31 15:55:57 <KotCzarny> yeah, learned it quite a time ago
May 31 15:55:57 <L29Ah> KotCzarny: :*
May 31 15:56:02 <L29Ah> tho i can use PID namespaces
May 31 15:56:06 <KotCzarny> chroot is not safe at all
May 31 15:56:14 <KotCzarny> it's just a convenience
May 31 15:56:29 <kerio> yeah, if you're leaving /proc in the jail
May 31 15:56:43 <KotCzarny> kerio: there are more ways to escape chroot
May 31 15:56:48 <KotCzarny> that is just easiest
May 31 15:56:57 <kerio> they all involve having root tho
May 31 15:57:03 <kerio> or CAP_CHROOT, whatever
May 31 15:57:55 <kerio> the easiest is definetely mkdir("tmp"), chroot("tmp"), chdir("..") x 1024, chroot(".") 
May 31 15:58:22 <KotCzarny> mine was 1 cmd
May 31 15:58:28 <KotCzarny> yours is 1027
May 31 16:00:20 <L29Ah> kerio: didn't work
May 31 20:06:35 <L29Ah> is it possible to paginate uboot output or scroll it back?
May 31 20:06:49 <L29Ah> help and printenv don't fit the screen
May 31 20:31:01 <L29Ah> damn, uboot stopped to see my internal-mmc-hosted kernel after i've plugged in the external one
May 31 20:37:16 <L29Ah> oh no it does see
May 31 20:37:23 <L29Ah> it's just emmc "forgot" the kernel
May 31 20:37:27 <L29Ah> damn
May 31 20:37:39 <L29Ah> i woder how does this happen
May 31 20:38:56 <L29Ah> wtf the directory is said to be modified two minutes ago
May 31 20:53:31 <L29Ah> yay after a reboot it REMEMBERED it back ><
May 31 20:53:58 <L29Ah> i guess i should attach a power kernel to uboot to be able to reformat emmc to ext4
May 31 20:54:15 <L29Ah> oh, i can haz ext3
May 31 21:12:25 <L29Ah> qemu-user sucked so much i had to move gcc job back to n900 >_>
May 31 21:12:36 <L29Ah> i take the phrase about two days back
May 31 21:39:22 <chainsawbike> L29Ah, https://wiki.gentoo.org/wiki/Cross_Container_Support_Project - native cross-compiling gcc toolchain in a qemu-arm chroot
May 31 22:08:07 <bencoh> chainsawbike: does gcc runs in qemu ?
May 31 22:09:05 <bencoh> or does it work like scratchbox (native crosscompiler/tools and binfmt_misc trick to run target binaries in qemu) ?
May 31 22:37:29 <chainsawbike> like scratchbox
**** ENDING LOGGING AT Mon Jun 01 02:59:59 2015