IRC logs for #openrisc Friday, 2013-02-15

poke53281blueCmd: same error here. cloned the kernel and got this error message02:41
stekernjuliusb: the ticksyscall test is also problematic, it get stuck in a branch-delay-slot-interrupt loop03:09
joniboblueCmd:  what kernel and what config are you using?  I'm not seeing any compile errors with the branches I named in my email and the defconfig10:20
jonibonot sure what issues you were having with ethoc... i'll try to take a look and see if it works for me10:20
jonibomaster branch ethoc and for-upstream ethoc do differ a bit10:21
blueCmdjonibo: I had problems with ethoc on 3.6.10 mainline, but I haven't tested with your git version.11:41
blueCmdjonibo: I emailed a sort of bug report11:53
LokeI'm having problems compiling c++ code for the openrisc, using the stable toolchain. Please, anyone here knows how to solve it?14:11
LokeI would appreciate any help14:11
blueCmdLoke: what's your problem?14:12
LokeWhen I try to compile c++ code, the linker would give me this error:14:13
Loke /opt/openrisc/lib/gcc/or32-elf/4.5.1-or32-1.0rc4/../../../../or32-elf/bin/ld: error: no memory region specified for loadable section `.gcc_except_table'14:13
blueCmdare you using newlib or uclibc?14:13
Lokethe current stable version14:14
blueCmdsorry, I have no experience with newlib. if this is a blocker for you I would recommend that you use the development toolchain14:14
blueCmd_franck_ uses newlib I think14:14
Lokecorrect me if i'm wrong but, using uclibc would mean simulating linux as well as the code i'm compiling right?14:15
blueCmdyes, uclibc is for a linux system14:16
blueCmdthe new toolchain (or1k-elf as the target is named nowadays) would be where bugs are fixed anyway.14:17
LokeUhm ok, I will try that and see what happens14:18
LokeThank you14:18
blueCmdnp, stick around here on IRC if you can - it's great for getting help on these kind of things14:18
LokeAnyway, the uclic can compile c++ wihtout problems?14:19
blueCmdyeah, and newlib should be fine aswell. I just looked at the regression testing for 4.5.1 and it passes a lot of C++ tests so it should be fine14:19
LokeOh ok14:20
LokeReally appreciate it, this error is driving me crazy..14:20
blueCmdThey often do :-).14:21
Lokehaha I guess so14:21
joniboblueCmd: thanks for the bug report... I'll look into it14:21
blueCmdjonibo: great! I'm probably missing something very simple14:25
blueCmdjeremybennett: do you have any copy of the work you had done to parallelise the regression tests?14:58
jeremybennettblueCmd: I thought it was in the or32 SVN repository. It was a set of scripts that manually broke up the tests.14:59
blueCmdmaybe it is, i will have a look14:59
blueCmdI rarely poke around in the SVN repo :)14:59
jeremybennettIt is now pretty static, so you can just take a copy and use it for reference.15:00
jeremybennettIf it's not there, I'll see if I can find the scripts and zip them up for you.15:00
blueCmdjeremybennett: do we have some automatic builders or anything like that?15:01
blueCmdjeremybennett: I think i found the script15:03
jeremybennettThat's good15:03
jeremybennettIf only I had time, I'd like to write it up as an application note. I talked about it at the GCC cauldron, and there was a lot of interest.15:03
jeremybennettI get the strong impression that every group hacks the test environment to get it to do what they want, but little of that is actually shared.15:04
-!- Kevin` is now known as Kevin`_15:26

Generated by 2.15.2 by Marius Gedminas - find it at!