IRC logs for #openrisc Wednesday, 2015-04-29

--- Log opened Wed Apr 29 00:00:27 2015
latifolofk: I am so sory but my reinsttaled toolchain returns an error again, but a different error!! When I try to compile a very easy c++ code (just helloworld) it gives this error07:47
latifundefined reference to std::ios::Inıt:Inıt07:50
olofklatif: Don't have a clue where to start looking this time unfortunately :/07:52
olofkCan you give me the code you're trying to run07:52
latifolofk: this my code...  #include <iostream>   int main() { std::cout << "test\n";  return 0; }08:09
latifolofk: Please try this code on your system... I hope It works on your system..08:10
olofkYep. That works fine here08:28
olofk(with the additions to the toolchain I showed you yesterday08:28
latifAaa..So?? I did something wrong?!! Can you please tell me the additions again..08:38
latiffirst..I wrote _open_r (for _open)...then I added const for definition..(const char).. in newlib/libgloss/or1k/ syscalls.c... Anythng else??08:40
olofklatif: That should do it09:24
olofkhmmm.. did you say that you were using or1k-src btw?09:24
olofkOr are you using the instructions here ?09:24
latifolofk: I am using the instrunctions..
latifbut as you know or1k-src also exist in the instructions on the what do you mean by or1k-src btw??.09:34
latifby the way I did not change anything on is it rightt??09:35
olofkYes, that correct. It turned out no changes to was needed09:40
olofklatif: Ah right. We still point to or1k-src for the uClibc toolchain and for GDB09:41
olofkor1k-src contains binutils, newlib and gdb bundled together, and until last year we used that to build most parts of the toolchain. But it's not updated anymore, so I just wanted to rule out that you were using old code09:42
latifolofk: my _open_r function is defined like: int _open_r(struct _reent *reent,const char *file, int flags, int mode) { reent->_errno = ENOSYS; return -1; }  is it the same with yours??10:18
latifand it is the all we should change right?? By the way can you please try my code (easy c++ code) in simulator.. I wonder if it works correctly??10:20
heshambandvig: Did you get Linux running on your Atlys board?13:09
bandvighesham: I’ve got one. However, I use uClibc, configured with my original configuration file (to overcome some issues I had with compiling uClibc). And I use old version of linux sources downloaded from (with already precompiled busybox and configured file system).13:23
heshambandvig: I don't know if this is can be done on OpenRISC, please tell me. I want to be able to get Linux running (even on or1ksim) and be able to compile some POSIX programs over there, and run it, and do some timing analysis.13:27
heshamThat would need gcc, running on Linux/OpenRISC, and a c library.13:29
heshamIs glibc upstreamed (yet) and working with linux? I saw this e-mail
bandvighesham: In fact, I’m not linux guru. I’m afraid I couldn’t help you. I just followed the blog.13:53
bandvigI have never compiled any application separately from linux kernel. I just compiled kernel+busybox, saw that it works, executed several trivial commands like “ls” and stopped playing with linux.13:53
bandvigI just know, that it is possible to compile a linux application on PC with appropriate tool chain (uClibc or musl based) than upload it into linux file system (by tftp? by uart?) and execute it on target board.13:53
bandvigHowever, I’ve never done that. You should asked somebody (stekern, for example) who have got much more linux experience.13:54
heshambandvig: Me neither, that's why I am asking. Thanks anyway. Hope someone would see question this and reply.13:58
latifolofk: no problem.. now I can compile  std::cout on my system... you are realy awesome thank you so muchh :)16:56
olofklatif: Great to hear that it's working. I'll make sure the patch is applied to newlib21:11
--- Log closed Thu Apr 30 00:00:29 2015

Generated by 2.15.2 by Marius Gedminas - find it at!