ibotlog2html for #crosstool-ng

<< Previous 2014-03-15 Next >>

# 00:02:28 blueness joins #crosstool-ng
# 00:02:29 mnt_real_ joins #crosstool-ng
# 00:03:51 ben1066 joins #crosstool-ng
# 00:04:30 ovf_ joins #crosstool-ng
# 00:10:29 ben1066_ quits : Read error: Connection reset by peer
# 00:10:30 ovf quits : Ping timeout: 246 seconds
# 00:10:33 memleak quits : Ping timeout: 246 seconds
# 00:10:38 xrl____ quits : Ping timeout: 245 seconds
# 00:11:33 memleak joins #crosstool-ng
# 00:11:33 memleak quits : Changing host
# 00:11:33 memleak joins #crosstool-ng
# 00:11:59 xrl____ joins #crosstool-ng
# 00:36:17 xrl_____ joins #crosstool-ng
# 00:37:49 xrl____ quits : Ping timeout: 246 seconds
# 00:37:56 xrl_____ is now known as: xrl____
# 00:50:29 Net147 quits : Quit: HydraIRC -> http://www.hydrairc.com <- Would you like to know more?
# 01:33:35 djerome joins #crosstool-ng
# 06:20:03 bhundven quits : Ping timeout: 245 seconds
# 06:20:19 ovf_ quits : Ping timeout: 264 seconds
# 06:20:44 xrl____ quits : Ping timeout: 265 seconds
# 06:28:00 bhundven_ joins #crosstool-ng
# 07:19:44 xrl_____ joins #crosstool-ng
# 08:18:31 ovf_ joins #crosstool-ng
# 09:43:24 Net147 joins #crosstool-ng
# 10:53:59 ovf_ is now known as: ovf
# 11:02:57 smartin_ joins #crosstool-ng
# 11:11:06 y_morin joins #crosstool-ng
# 11:18:12 smartin_ quits : Quit: leaving
# 11:21:22 smartin_ joins #crosstool-ng
# 12:11:57 ius joins #crosstool-ng
# 13:47:15 perr joins #crosstool-ng
# 13:52:59 perr quits : Ping timeout: 240 seconds
# 14:00:08 Net147 quits : Quit: HydraIRC -> http://www.hydrairc.com <- It'll be on slashdot one day...
# 14:08:02 perr joins #crosstool-ng
# 14:17:13 perr quits : Ping timeout: 240 seconds
# 14:29:25 perr joins #crosstool-ng
# 14:38:55 perr quits : Ping timeout: 264 seconds
# 14:52:30 perr joins #crosstool-ng
# 15:02:54 ius So I'm trying to build a toolchain targeting sh4a, gcc 4.8.2 / glibc 2.24
# 15:03:13 ius (ctng hg checkout a couple of days old)
# 15:04:45 ius But it fails to link libc, missing __emutls_get_address. I believe this is due to failing to link -lgcc_s, but im not really sure what's at fault and how to resolve it
# 15:06:01 ius (I tried adding -lgcc_s to the cc_flags/cflags variable for libc, but then it fails to find gcc_s during configure)
# 15:06:01 perr quits : Ping timeout: 240 seconds
# 15:19:10 perr joins #crosstool-ng
# 15:34:38 perr quits : Quit: Leaving
# 17:43:56 sh4rm4 quits : Ping timeout: 252 seconds
# 17:47:52 sh4rm4 joins #crosstool-ng
# 18:06:41 smartin_ quits : Quit: leaving
# 18:30:08 tlwoerner quits : Ping timeout: 264 seconds
# 18:43:30 tlwoerner joins #crosstool-ng
# 20:51:15 daze joins #crosstool-ng
# 20:51:19 daze parts #crosstool-ng
# 20:51:33 oty joins #crosstool-ng
# 20:57:23 oty Hey. I'm trying to get a crosscompiler working for arm-cortex_a8-linux-gnueabi. I did "ct-ng arm-cortex_a8-linux-gnueabi" and then setup few change "kernel 3.4.69", and because I had few problems I added companions tools autoconf and automake. I also changed gliblc to the latest I could find. But I get this error about "support for the tls_model attribute is required"
# 20:57:33 oty And I don't know what to do from that :o
# 20:58:50 oty (I'm not familiar at all with cross compiling, I just get adviced, I obeyed pretty blindly)
# 21:01:17 y_morin oty: care to share your .config?
# 21:04:12 oty sure y_morin. I that normal I have a .config2 too ?
# 21:04:17 oty *Is
# 21:05:11 y_morin oty: Yes, .cofig2 id just an internal file. We just need the .config
# 21:05:30 y_morin .config2 is....
# 21:05:31 oty http://pastebin.com/uvuuhAQR
# 21:05:51 y_morin oty: I'll do a test build here...
# 21:06:02 oty thanks !
# 21:07:16 oty (but well, it took me about 40 minute to get the error :s)
# 21:10:10 y_morin oty: except for the download, a similar toolchain took about 15-20 minutes here.
# 21:11:11 oty yay o/
# 21:14:11 y_morin oty: What is your host system?
# 21:40:31 y_morin oty: Build failed here too. I'll have a look a bit later...
# 21:49:19 bhundven_ is now known as: bhundven
# 21:49:36 bhundven quits : Changing host
# 21:49:37 bhundven joins #crosstool-ng
# 21:49:37 bhundven quits : Changing host
# 21:49:37 bhundven joins #crosstool-ng
# 21:50:03 bhundven y_morin: back in seattle :D
# 21:56:22 y_morin bhundven: Hey! How did it go?
# 21:58:29 bhundven y_morin: chicago was fun, but our startup got some really big wins.
# 21:59:17 bhundven I will have a lot of work to do, in the near term, so ray and I should start hammering out this multilib stuff
# 21:59:28 bhundven before I get too busy with work
# 22:00:01 y_morin Hey, that's good news for you, then! :-)
# 22:02:11 bhundven tlwoerner: in other words, I may need you to jump in with some arm help
# 22:08:29 bhundven y_morin: yes, indeed.
# 22:08:51 bhundven y_morin: but bad news for any projects I am working on, because I'm going to get so busy
# 22:09:30 bhundven and probably heading to Sydney, back to Chicago, and maybe to Hong-Kong to tie the rest of the shoes to get our startup going.
# 22:10:45 y_morin bhundven: Yeah! You get Miles for your holidays travels! :-)
# 22:11:01 bhundven I need to get a card that rewards me for miles
# 22:22:10 oty y_morin: thanks for trying. I'm on archlinux x86_64 Linux 3.11. Using the AUR crosstool-ng-h version (3243.9838aecd6340-1)
# 22:24:08 oty (I don't particulary care about my actual config ; what is important as far as I know is the arm-cortex_a8-linux-gnueabi, and the 3.4.X kernel :p)
# 22:24:53 y_morin oty: I found the problem.
# 22:25:03 oty o/
# 22:25:06 y_morin oty: Lemme do one more test before, though...
# 22:25:31 oty y_morin: I will wait praising your name :o
# 22:27:29 y_morin oty: Oh, don't! ;-) Results in about 15mins...
# 22:28:06 oty xD
# 22:28:24 oty y_morin: well seen x)
# 22:42:10 y_morin oty: OK, your problem is solved: -mcpu=cortex-a8 conflicts with -march=armv7-a
# 22:42:34 y_morin oty: Just remove the "Architecture level" in "Target options"
# 22:42:40 oty o/
# 22:42:55 y_morin oty: But this uncovered another problem: glibc-2.19 needs binutils 2.21 or above.
# 22:43:22 y_morin oty: I'll let you test this further, now! ;-)
# 22:43:31 oty ^^
# 22:44:07 oty binutils 2.21 is to install on my system, not to chose in the menuconfig, right ?
# 22:44:29 y_morin oty: In the menuconfig, yes.
# 22:44:49 oty oki doki !
# 22:45:02 y_morin oty doki? :-)
# 22:45:07 y_morin hides...
# 22:45:10 oty :o
# 22:45:51 y_morin oty: But basically you would not want to mix too disparate components versions: use a gcc/glib/binutils combo that were released at about the same time.
# 22:46:08 y_morin *glibc not glib
# 22:46:31 y_morin oty: Do not try to mix an old gcc or binutils with a newer glibc. Or any other weird combo.
# 22:46:40 oty I'd want to have everything the most recent possible generally.
# 22:47:08 y_morin oty: Then do so: bump binutils, gcc and the linux headers.
# 22:47:19 oty ok ok !
# 22:51:21 oty mhm
# 22:51:30 oty linux headers ?
# 22:52:30 y_morin oty: Linux headers, yes. They are part of the toolchain. In the "Operating System" sub-menu.
# 22:52:41 oty the kernel version ?
# 22:52:57 y_morin oty: Basically, glibc needs to know the API and ABI to talk to the kernel, so it needs the kernel headers.
# 22:52:57 oty x)
# 22:53:23 oty doki !
# 22:53:27 y_morin :-)
# 22:53:27 oty ><
# 22:54:16 oty (I'll re-use that I think, I've no shame stealing, even shameful stuff :p)
# 22:54:19 y_morin oty: But beware the kernel headers! You should ensure that glibc can talk to the kernel running on your target.
# 22:54:30 y_morin oty: You can achive tht in two ways:
# 22:54:51 y_morin oty: 1) set the same version for kenel headers in ct-ng as the kernel you run on your target
# 22:55:10 y_morin oty: or 2) tell glibc what the oldest kernel it should support will be.
# 22:55:22 y_morin oty: Usually, 1) is the preffered solution.
# 22:55:29 oty I was running the 1) o/
# 22:55:41 oty *for
# 22:55:44 y_morin oty: Good, then keep that version as kernel headers.
# 22:56:29 y_morin oty: The drawback is if you upgrade your kernel on the target: glibc won't be able to use the latest APIs of the newer kernel, unles you recompile your toolchain.
# 22:58:20 oty one hour of compilation is fine, as long as I still have the .config somewhere with everything working ^^
# 22:59:30 y_morin oty: The .config is saved in the toolchain.
# 23:00:11 y_morin oty: It is a script that is named: --ct-ng.config
# 23:01:01 tlwoerner bhundven: okay, np. congrats on your startup wins! sounds exciting :-)
# 23:01:01 y_morin oty: So you can run (eg.) arm-unknown-linux-gnueabi-ct-ng.config >.config to retrieve the .config used for that toolchain.
# 23:02:42 oty y_morin: cool !
# 23:06:12 bhundven tlwoerner: :D , but I do want to get the multilib stuff done and get the patches on patchworks so we can get comments
# 23:06:43 bhundven I don't think y_morin will even look at them and comment from my patchqueue
# 23:11:43 bhundven tlwoerner: to go back to the discussion we were having on email, wrt multlib-list; We are trying to provide a solution that allows for simple multilib configuration, but I'd also like to include a CT_CC_GCC_CUSTOM_MULTILIB_LIST, incase you want to do something weird (and noted that it is not recommended.)
# 23:14:04 oty y_morin: "[ERROR] configure: error: forced unwind support is required" is something you heard about ? :o
# 23:14:43 y_morin bhundven: Yes, we want a default, no-config multilib config option that activates the dfault settings for that arch. And an additional, optional config option for the user to specify a non-default (trimmed-down, experimental...) multilib settings.
# 23:14:57 bhundven :)
# 23:15:02 bhundven exactly
# 23:15:14 y_morin oty: Yes, just say 'Y' to force Unwind support in the menuconfig, in the "C library" sub-menu.
# 23:16:11 y_morin (but it should 'y' by default..)
# 23:16:57 bhundven y_morin: heh, I was about to ask about that.
# 23:17:34 bhundven I notice on most of my configs, I run into the problem where unwind should be on but it's not. I think it's more rare now to not force unwind support
# 23:18:29 oty Maybe I unchecked it by error, but I doubt that x)
# 23:20:15 y_morin bhundven, oty: strange: the default is 'y' (I just checked).
# 23:20:37 bhundven o.O
# 23:20:40 bhundven hrm
# 23:20:42 bhundven your right
# 23:20:48 bhundven feels crazy
# 23:20:50 y_morin oty: Ah, but it is explicitly set to off in the arm-cortex_a8 sample.
# 23:20:56 oty o/
# 23:21:01 oty I'm not crazy yet o/
# 23:21:15 y_morin oty 1, bhundven 0. ;-)
# 23:21:20 bhundven haha
# 23:21:37 y_morin Well, time for bed here...
# 23:21:43 bhundven l8r y_morin
# 23:21:49 oty sweet dreams o/
# 23:22:03 y_morin hates when his body's phisiological needs kick in... :-/
# 23:22:06 y_morin Bye!
# 23:22:12 y_morin quits : Quit: Nighty Night!
# 23:22:26 bhundven well, humans are not machines.
# 23:22:57 bhundven and even machines are affected by entropy
# 23:34:13 mingwandroid joins #crosstool-ng
# 23:53:49 oty error: field 'siginfo' has incomplete type
# 23:54:12 oty y_morin my savior is asleep, if any idea, I'll take it x)

Generated by ibotlog2html by Yann E. MORIN