| --- Log opened Wed Apr 27 00:00:40 2016 | ||
| _franck_ | shorne: yeah, you should use this repo: https://github.com/openrisc/or1k-src | 00:27 |
|---|---|---|
| _franck_ | and compile with --enable-sim --enable-or1ksim (AFAIR --enable-or1ksim is not mandatory since or1ksim is the default simulator when sim is enabled) | 00:31 |
| _franck_ | mithro: you can try ##fpga | 00:31 |
| _franck_ | or here :) | 00:32 |
| shorne | _franck_: the lastest gdb work we have is now going into binutils-gdb as far as I know. (that is based off of upstream gdb rather than cvs) I am going to try merging the sim patchs from or1k-src to binutils-gdb | 07:00 |
| shorne | ok all sim patches applied (had to skip patches to (/newlib, /libgloss, /cpu) most look like they are upstream | 08:08 |
| shorne | nevermind about /cpu, it looks like they are already applied | 08:12 |
| olofk | shorne: It's really great that you're doing this. It has been annoying that we're probably ready for upstreaming, but that no one really knows | 08:52 |
| shorne | yeah, and our main source is based on cvs while upstream moved to git | 08:53 |
| shorne | progressing with build | 08:53 |
| shorne | it seems there are some API changes, so updating sim for that | 08:53 |
| olofk | shorne: Is it? I thought our binutils-gdb tracked upstream | 08:54 |
| shorne | olofk: yes, binutils-gdb does I mean or1k-src doesnt, which is annoying for preparing any patches | 08:55 |
| shorne | so this should be the last step to get binutils-gdb ready for upstreaming | 08:55 |
| olofk | I thought we had deprecated or1k-src | 08:56 |
| olofk | But maybe this is all the reason why everything is so confusing :) | 08:56 |
| shorne | olofk: thats right, but the patches in or1k-src/sim never made it out of there. So I have extracted them and applied to binutils-gdb | 08:57 |
| olofk | ahh... ok, so our binutils-gdb doesn't include sim and gdb. Got it now | 08:57 |
| olofk | Perfect! | 08:57 |
| shorne | issue is they are 2 years old, and binutils-gdb is modern | 08:57 |
| olofk | Damn bitrot :( | 08:57 |
| shorne | just reading up on some of the api changes | 08:57 |
| shorne | making progress, convering gdb/sim/or1ksim to nrun | 10:41 |
| shorne | then had issue with WITH_PROFILE definition, needed to autoreconf | 10:42 |
| shorne | now looks like I missed some stuff in nrun | 10:42 |
| shorne | FYI, these are the changes I pulled in, not including my patch work in progress https://github.com/stffrdhrn/binutils-gdb/commits/or1k-sim | 10:44 |
| shorne | it builds | 10:57 |
| shorne | looks like testsuites are working | 11:06 |
| shorne | ill push a commit | 11:07 |
| olofk | shorne: I've never managed to understand how to run the test suite. Is it hard? Do you have any instructions? | 11:10 |
| olofk | Might be good material for the tutorials repo | 11:10 |
| shorne | In my build directory I ran this | 11:16 |
| shorne | make check RUNTESTFLAGS="--target_board=or1k-elf-sim" | 11:16 |
| shorne | yeah, I was thinking the same | 11:16 |
| shorne | but.. the --target_board need to be in a path relative to your site.exp or include | 11:17 |
| shorne | in my site.exp I have put 'lappend boards_dir /home/shorne/work/openrisc/or1k-src/boards' | 11:18 |
| shorne | in 'or1k-src/boards' there is a or1k-elf-sim definition that says "use target sim" when we run the | 11:18 |
| shorne | testsuite | 11:18 |
| shorne | Ill put together some better details | 11:20 |
| shorne | the test summary results are very similar to _franck_ 's results from the original commit | 11:43 |
| shorne | https://gist.github.com/stffrdhrn/ace3aa69fd908768a97028ace2d93e69 | 12:05 |
| _franck_ | shorne: great job. We are now ready to get the failure counter down a bit | 15:13 |
| olofk | shorne: Yeah, great job. I'll try to look through the fails when I get some time. Probably a mix of hard and easy bugs | 15:27 |
| olofk | Found a SAT solver to use for FuseSoC dependency tracking. Managed to massage it enough so that it should be possible to shoehorn it into FuseSoC | 16:55 |
| olofk | The codebase for that one is probably way larger than all of FuseSoC :) | 16:56 |
| olofk | Yep. 2.4 times as many lines of Python in the SAT solver :) | 16:58 |
| wallento | shorne: thanks for the work on the sim part! | 18:17 |
| wallento | The item came pretty near to the top of my TODO list for 2016 already ;) | 18:17 |
| wallento | SMDwrk: yes, we have a jenkins instance running and I tried to do as much as possible in travis | 18:18 |
| wallento | actually, most of the tests can be done with travis | 18:19 |
| wallento | please email me to discuss how to progress and give you access to our jenkins instance: [email protected] | 18:19 |
| shorne | wallento: no problem, I said I was going to rebuild my toolchain and run the testsuite, I didnt know I had to merge the sim stuff! | 18:34 |
| shorne | but I figured I would give it a try | 18:34 |
| --- Log closed Thu Apr 28 00:00:41 2016 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!