ibotlog2html for #crosstool-ng

<< Previous 2014-07-29 Next >>

# 00:15:54 sh[4]rm4 joins #crosstool-ng
# 00:17:45 sh4rm4 quits : Ping timeout: 256 seconds
# 00:49:07 djerome joins #crosstool-ng
# 02:58:03 djerome quits : Remote host closed the connection
# 03:03:52 rofl__ joins #crosstool-ng
# 03:05:29 sh[4]rm4 quits : Ping timeout: 256 seconds
# 03:24:26 nandub joins #crosstool-ng
# 03:25:00 nandub quits : Client Quit
# 03:45:14 nandub joins #crosstool-ng
# 03:45:22 sh[4]rm4 joins #crosstool-ng
# 03:47:59 rofl__ quits : Ping timeout: 256 seconds
# 04:04:23 memleak joins #crosstool-ng
# 04:04:36 memleak quits : Changing host
# 04:04:36 memleak joins #crosstool-ng
# 04:11:10 perr joins #crosstool-ng
# 04:49:55 perr quits : Ping timeout: 255 seconds
# 04:57:51 perr joins #crosstool-ng
# 05:45:52 diorcety joins #crosstool-ng
# 05:54:17 diorcety quits : Read error: Connection reset by peer
# 06:04:29 ovf quits : Ping timeout: 272 seconds
# 06:13:45 ovf joins #crosstool-ng
# 06:24:07 ovf quits : Ping timeout: 272 seconds
# 06:28:54 ovf joins #crosstool-ng
# 07:21:11 aalv joins #crosstool-ng
# 07:22:24 aalv quits : Client Quit
# 07:22:41 aalv joins #crosstool-ng
# 07:40:05 rofl__ joins #crosstool-ng
# 07:42:01 sh[4]rm4 quits : Ping timeout: 256 seconds
# 08:09:52 memleak quits : Quit: Leaving
# 08:53:07 perr quits : Ping timeout: 245 seconds
# 09:03:15 UukGoblin quits : Read error: Connection reset by peer
# 09:08:58 perr joins #crosstool-ng
# 09:08:58 perr quits : Changing host
# 09:08:58 perr joins #crosstool-ng
# 09:12:19 UukGoblin joins #crosstool-ng
# 09:38:06 sh4rm4 joins #crosstool-ng
# 09:39:53 rofl__ quits : Ping timeout: 256 seconds
# 09:48:47 maxime__ is now known as: maxime
# 09:49:36 maxime quits : Changing host
# 09:49:36 maxime joins #crosstool-ng
# 11:12:15 sh4rm4 quits : Ping timeout: 256 seconds
# 11:55:02 perr quits : Ping timeout: 255 seconds
# 11:59:44 sh4rm4 joins #crosstool-ng
# 12:09:05 sh4rm4 daggs1-work, never heard of prints(). is that a windows extension?
# 12:09:50 sh4rm4 or were you asking whether musl is littered with debug printf statement that can be activated with a magic configure switch ?
# 12:10:12 sh4rm4 if the latter, than no
# 12:10:27 sh4rm4 you have to add the dprintf's yourself
# 12:11:17 sh4rm4 or just compile musl with CFLAGS="-O0 -g3" ./configure...
# 12:11:28 sh4rm4 then you can comfortably singlestep
# 12:11:44 daggs1-work sh4rm4, managed, used fprintf to stdout
# 12:12:02 daggs1-work found out that musl doesn't interprets scripts
# 12:12:14 sh4rm4 scripts?
# 12:12:37 sh4rm4 btw to debug libc you should use dprintf(2, "...") rather then fprintf
# 12:12:44 sh4rm4 the former is threadsafe
# 12:13:05 daggs1-work if I compile busybox linked to musl and try to run a script by calling the file name directly (like in bash) I get Exec format error
# 12:13:42 sh4rm4 with /bin/sh symlinked to musl-linked busybox ?
# 12:13:50 daggs1-work yes
# 12:14:05 sh4rm4 weird
# 12:14:07 daggs1-work happened with uclibc too
# 12:14:16 sh4rm4 on which arch ?
# 12:14:22 daggs1-work x86_64
# 12:14:50 sh4rm4 is it dynamically linked ?
# 12:15:03 daggs1-work looking as musl code, it gets the cmd directly from busybox and pass to the kernel
# 12:15:09 sh4rm4 if so /lib/ld-musl-$ARCH.so.1 must exist
# 12:15:41 daggs1-work created a musl based cross compiler, made sure busybox with be created with that compiler and linked to that lib
# 12:16:04 sh4rm4 so it is dynamically linked ?
# 12:16:13 daggs1-work it does or else the image won't boot as it run busybox on start
# 12:16:15 daggs1-work yes
# 12:16:33 sh4rm4 hmm ? how would a statically linked binary not boot ?
# 12:16:43 daggs1-work statically?
# 12:16:56 sh4rm4 yes, static linked
# 12:17:07 daggs1-work I'm not using a statically linked busybox
# 12:17:21 daggs1-work I'm using a dynamically linked busybox
# 12:17:29 sh4rm4 i.e. using libc.a rather than libc.so to create a free-standing binary without any external dependencies
# 12:18:04 daggs1-work I need to use dynamically linked busybox due to licensing issues
# 12:18:22 perr joins #crosstool-ng
# 12:18:23 perr quits : Changing host
# 12:18:23 perr joins #crosstool-ng
# 12:18:29 daggs1-work also I have limited size in the image
# 12:19:47 sh4rm4 yes but a statically linked busybox is much smaller than a dynamic linked one + libc.so
# 12:20:22 sh4rm4 anyway, your problem is probably related to one of 2 problems:
# 12:20:33 daggs1-work I understand but my boss wants all to be dynamic linked one
# 12:20:40 sh4rm4 1) lack to compile kernel with a recently added flag that allows to run scripts directly
# 12:21:06 sh4rm4 2) bogus linker that uses mappings > PAGE_SIZE
# 12:21:15 daggs1-work well me img kernel is rather new but the cross compiler kernel is old
# 12:21:25 daggs1-work bogus linker that uses mappings?
# 12:21:29 daggs1-work how can that happen?
# 12:21:33 sh4rm4 for 2, i cant remember the exact details and waiting for the details in #musl
# 12:22:04 daggs1-work this happened to me with uclibc, same scenario worked in glibc
# 12:22:16 sh4rm4 daggs1-work, iirc some ppl use an ld option that makes the dynlinker use huge pages to speed up loadingextremely bloated C++ libs
# 12:22:56 sh4rm4 daggs1-work, maybe the glibc scenario used a different kernel version
# 12:23:10 daggs1-work nope, same kernel usage
# 12:23:39 sh4rm4 anyway, the option is CONFIG_BINFMT_SCRIPT=y
# 12:23:49 sh4rm4 if this is missing in your kernel config, its the reason
# 12:23:52 daggs1-work in the kernel?
# 12:23:57 sh4rm4 yup
# 12:24:21 daggs1-work ok, I see, it isn't set, wil enable it.
# 12:24:31 daggs1-work and try, that might be the issue
# 12:25:26 sh4rm4 seems it was introduced in 3.10
# 12:25:41 daggs1-work there goes... it will be extremely bad (for me) if it works, this means I've wasted most of the day :(
# 12:28:45 sh4rm4 otoh it means you dont have to waste another day
# 12:28:46 daggs1-work :facepalm: it works ;(
# 12:28:59 sh4rm4 but you can thank the kernel guys for this surprise
# 12:29:06 daggs1-work I had the patch ready...
# 12:29:11 sh4rm4 i was quite wtf when i encountered this myself
# 12:29:38 daggs1-work when I encountered this first, I wrote a busybox patch that does it
# 12:29:52 sh4rm4 how could that work ?
# 12:33:47 daggs1-work by trying to read the first few bytes and looking for the shebang format, crude but it suited our needs
# 12:34:46 sh4rm4 you mean in the ash code ?
# 12:34:57 daggs1-work hush code rather
# 12:35:02 sh4rm4 ok that could work as long as you dont start another shell
# 12:35:13 sh4rm4 or you dont use a shell script for init
# 12:35:20 sh4rm4 I need to use dynamically linked busybox due to licensing issues
# 12:35:21 daggs1-work worked even we started anothe rshell
# 12:35:33 sh4rm4 ^ are you sure the license issues apply with musl as well ?
# 12:36:04 sh4rm4 as musl's MIT-licensed, you can link it to anything and distribute the binaries under the license of the programs
# 12:36:18 daggs1-work that are my boss asked me to do.
# 12:36:28 sh4rm4 because of GPL fear
# 12:36:30 daggs1-work but afaik busybox isn't mit license
# 12:36:37 sh4rm4 busybox is GPL
# 12:36:53 sh4rm4 and it will stay that whether you link dynamically or statically
# 12:37:19 daggs1-work I know
# 12:37:56 daggs1-work ask said, my boss handles the licensing issues, I implement.
# 12:38:01 sh4rm4 but if for example bb was GPL 2 only, and your libc GPL3, you couldnt distribute the resulting static linked binary
# 12:38:13 sh4rm4 because the licenses are incompatible
# 12:38:23 daggs1-work as we are constantly understaffed and I'm swamp with work, I don't have timeslot to argue with him :)
# 12:38:35 daggs1-work after all he pays my wage
# 12:38:39 sh4rm4 k
# 12:38:55 daggs1-work thanks for the help :)\
# 12:38:56 daggs1-work thanks for the help :)
# 12:39:02 sh4rm4 welcome
# 13:34:09 tlwoerner quits : Quit: Leaving
# 13:50:28 daggs1-work sh4rm4, question, I'm typing ping dagg in the image but I get this: "ping: bad address 'dagg'", in glibc I was missing libresolv and libnss_dns libs, what are the equivalent in musl?
# 13:51:35 perr quits : Quit: Leaving
# 13:53:03 sh4rm4 musl doesnt need those legacy resolver libs to work
# 13:53:07 sh4rm4 and doesnt provide them
# 13:53:18 sh4rm4 *everything* is in libc.so
# 13:53:23 sh4rm4 including the dynlinker
# 13:54:03 daggs1-work sh4rm4, so I guess I'm missing something...
# 13:55:13 sh4rm4 maybe resolv.conf ?
# 13:55:16 sh4rm4 try strace
# 13:55:46 daggs1-work don't have strace on the system
# 13:56:18 daggs1-work got resolv.conf
# 13:57:35 daggs1-work in the img
# 13:59:43 sh4rm4 no debug tools in the image ? so how are you supposed to debug it ?
# 13:59:53 daggs1-work as said, size limitation
# 13:59:58 sh4rm4 http://landley.net/aboriginal/downloads/binaries/extras
# 14:00:01 daggs1-work will try to find strace
# 14:00:03 sh4rm4 get one from here
# 14:00:10 sh4rm4 it's 400 KB statically linked
# 14:01:42 daggs1-work thanks
# 14:01:48 daggs1-work seems handy
# 14:05:01 daggs1-work ok, it tries /etc/hosts, then opens resolv.conf
# 14:06:24 sh4rm4 and then, does it send udp packets to your configured dns server ?
# 14:06:30 daggs1-work well, frankly I don't know, I see it sends a pkg, gets one but no getting the ip
# 14:07:07 sh4rm4 maybe the response is malformed ?
# 14:08:14 sh4rm4 https://github.com/rofl0r/hardcore-utils/blob/master/host.c
# 14:08:21 daggs1-work not sure what to expect, the image is using kernel 3.12.5 but the cross compiler is build with kernel 2.6.22, I had to tweak the headers when I was building busybox, can that be the issue?
# 14:08:24 sh4rm4 you could try if this host program works
# 14:08:41 sh4rm4 oh, yes this could be an issue.
# 14:08:58 sh4rm4 anyway this host program doesnt rely on kernel headers
# 14:09:06 daggs1-work ok, sec
# 14:10:34 daggs1-work testing in a sec
# 14:13:46 daggs1-work sh4rm4, how do I run it? host ?
# 14:14:03 sh4rm4 yup
# 14:15:35 daggs1-work didn't linked it statically, trying again
# 14:18:09 daggs1-work getting Host not found
# 14:19:28 aalv quits : Ping timeout: 250 seconds
# 14:20:16 sh4rm4 huh
# 14:20:52 sh4rm4 can you sniff the packets ? (wireshark,tcpdump...)
# 14:25:08 daggs1-work can I do it from outside the image?
# 14:26:34 sh4rm4 yes
# 14:27:19 sh4rm4 you could run tcpdump on the dns server
# 14:27:23 sh4rm4 if it's inhouse
# 14:27:34 sh4rm4 or on the router/gateway if its outside
# 14:29:10 sh4rm4 alternatively you can use an old hub (predecessor of network switch) because there all packets are sent to all ports
# 14:29:23 RushPL quits : Ping timeout: 264 seconds
# 14:29:43 daggs1-work maybe I can find static tcpdump
# 14:29:49 RushPL joins #crosstool-ng
# 14:29:53 sh4rm4 yeah i'm just trying to build one for you
# 14:30:35 daggs1-work thanks
# 14:33:08 sh4rm4 sigh... pcap is resisting my attempts to build it without usb support
# 14:37:24 aalv joins #crosstool-ng
# 14:39:33 sh4rm4 http://ftp.barfooze.de/pub/sabotage/bin/tcpdump-static-x86_64.xz
# 14:42:10 daggs1-work thanks
# 14:42:23 daggs1-work how do I use it?
# 14:42:39 sh4rm4 there must be an option to log into pcap files
# 14:42:54 daggs1-work ok
# 14:43:16 sh4rm4 those you can inspect comfortably on a box with wireshark on it
# 14:43:28 daggs1-work ok
# 14:45:10 sh4rm4 tcpdump -w dump.pcap -i eth0
# 14:51:40 daggs1-work got the file
# 14:51:58 daggs1-work I need to go, will check later on with wireshark, thanks for the help
# 14:54:10 sh4rm4 daggs1-work, still there ?
# 14:54:18 sh4rm4 you could send me the pcap via email
# 14:54:23 sh4rm4 so i can take a look
# 14:56:35 loide joins #crosstool-ng
# 15:39:07 memleak joins #crosstool-ng
# 15:41:07 feepbot quits : Remote host closed the connection
# 15:41:58 feepbot joins #crosstool-ng
# 17:13:35 aalv quits : Quit: Leaving.
# 17:35:40 sh4rm4 quits : Remote host closed the connection
# 17:38:31 sh4rm4 joins #crosstool-ng
# 17:53:01 sh4rm4 quits : Remote host closed the connection
# 17:58:52 y_morin joins #crosstool-ng
# 18:04:06 sh4rm4 joins #crosstool-ng
# 18:16:38 sh4rm4 quits : Remote host closed the connection
# 18:26:48 sh4rm4 joins #crosstool-ng
# 19:12:32 sh4rm4 quits : Remote host closed the connection
# 19:43:38 mingwandroid joins #crosstool-ng
# 19:45:09 sh4rm4 joins #crosstool-ng
# 19:50:09 sh4rm4 quits : Remote host closed the connection
# 19:55:41 sh4rm4 joins #crosstool-ng
# 19:57:29 maxime__ joins #crosstool-ng
# 19:59:12 maxime quits : Ping timeout: 260 seconds
# 20:05:44 sh4rm4 quits : Remote host closed the connection
# 20:15:38 diorcety joins #crosstool-ng
# 20:21:56 sh4rm4 joins #crosstool-ng
# 20:23:34 sh4rm4 quits : Remote host closed the connection
# 20:32:42 sh4rm4 joins #crosstool-ng
# 20:32:42 bhundven y_morin: I've updated my branch with the thread fix
# 20:32:46 bhundven https://bitbucket.org/bhundven/crosstool-ng/branch/musl
# 20:33:03 bhundven I'm in the middle of relocating to a different place to work
# 20:33:12 bhundven and found there is some problem with saveconfig
# 20:34:00 bhundven besides that, here is my current arm .config http://dpaste.com/1E1CNRK
# 20:34:07 bhundven which failed saveconfig
# 20:34:12 bhundven bbiab
# 20:35:08 sh4rm4 quits : Remote host closed the connection
# 20:41:00 sh4rm4 joins #crosstool-ng
# 20:45:15 sh4rm4 quits : Remote host closed the connection
# 20:54:14 sh4rm4 joins #crosstool-ng
# 20:54:56 y_morin bhundven: OK, I'll have a look. Thanks! :-)
# 21:36:43 loide quits : Ping timeout: 256 seconds
# 22:01:49 Guma joins #crosstool-ng
# 22:17:46 diorcety quits : Quit: Leaving.
# 23:01:51 y_morin quits : Quit: Nighty Night!
# 23:07:51 Guma quits : Quit: Leaving...

Generated by ibotlog2html by Yann E. MORIN