| --- Log opened Mon Sep 29 00:00:15 2014 | ||
| olofk | stekern: Wait until you see my off_shoring and out_sourcer cores :) | 06:52 |
|---|---|---|
| * olofk is not awesome :( | 08:03 | |
| olofk | No... wait. Turns out I am awesome after all | 08:07 |
| olofk | Hmm... looks like icarus don't like regs wider thatn 2^18 bits | 08:12 |
| stekern | why do you have 2^18 bit regs? | 09:01 |
| olofk | That's a valid question. I'm stuffing a lot of stimuli into my FIFO BFM | 09:19 |
| olofk | There's no problem though. I just need to split it into chunks first | 09:19 |
| stekern | this makes me upset, vhdl with flattened 2d-arrays in the port decleration | 11:25 |
| olofk | stekern: Agree. I guess the reason for doing that is that people don't know how to export a type that can be used in the parent module | 11:36 |
| wallento | hey guys, can anyone help me with our gcc problem? | 12:13 |
| stekern | wallento: the one you posted to the ml about? | 12:14 |
| wallento | yes | 12:14 |
| wallento | I have no clue | 12:14 |
| wallento | can this be a port problem | 12:14 |
| wallento | it is something with the destructor | 12:14 |
| wallento | and moving stuff | 12:14 |
| wallento | it seems | 12:14 |
| stekern | I just remembered about that today, I did some investigation and I tried to create a smaller test case from the code that fails. But I couldn't. | 12:15 |
| wallento | hm, any clue how I can track it down? I unfortunately have no idea about the GCC internals | 12:16 |
| stekern | you could try to go back to 4.8.1 and see if it's present there too | 12:16 |
| stekern | in general getting a smaller test case would help | 12:17 |
| wallento | yes, maybe we can reconstruct what actually happens, thanks for the moment | 12:18 |
| stekern | this was what I tried: http://pastie.org/9604253 | 12:20 |
| stekern | but I couldn't get that to fail | 12:20 |
| stekern | ...It *does* fail when I try to compile the real fiasco code thoough | 12:21 |
| wallento | I got it to fail now | 13:04 |
| wallento | the following switch creates the problem: -freg-struct-return | 13:04 |
| wallento | this is the code: http://pastie.org/9604359 | 13:05 |
| wallento | this is my command line: or1k-linux-uclibc-g++ -c -o test.o test.cpp -I../src/lib/libk -std=c++0x -freg-struct-return | 13:05 |
| wallento | in mybuild | 13:06 |
| wallento | The SPARC port had the same issue: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=57845 | 13:13 |
| wallento | so we can just remove the flag and it should work | 13:13 |
| wallento | stekern: thanks nevertheless! | 13:13 |
| stekern | wallento: ok, good that you found a workaround. I can take another look at fixing the actual problem with the extra info you provided | 13:44 |
| olofk | Yes! I'm awesome again | 18:20 |
| stekern | olofk: hmm, I thought you already had toggled back to awesome | 18:32 |
| stekern | http://juliusbaxter.net/openrisc-irc/latest.log.html#t08:07 | 18:33 |
| olofk | Probably a glitch somewhere | 18:35 |
| stekern | yeah, we *are* in different timezones, so maybe something got lost in the domain crossing | 18:36 |
| olofk | :) | 18:36 |
| stekern | what's the reason behind these fluctations in your self-esteem btw? | 18:50 |
| -!- Netsplit *.net <-> *.split quits: bentley`` | 19:50 | |
| --- Log closed Tue Sep 30 00:00:16 2014 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!