| --- Log opened Sun Jul 27 00:00:37 2014 | ||
| stekern | ah, CONFIG_FIXED_PHY | 05:39 |
|---|---|---|
| stekern | blueCmd: we do have redzone | 05:42 |
| stekern | but we also leave it enabled for the kernel | 05:43 |
| stekern | ...and of course there are no device tree bindings for that... | 06:08 |
| stekern | olofk: lgtm, wanna do a real commit and pull req for that? | 06:17 |
| stekern | minor nit, add a space after // | 06:18 |
| olofk | stekern: Will do | 07:15 |
| olofk | Fantastic forum thread :) http://opencores.org/forum,Cores,0,5410,0 | 07:19 |
| stekern | yeah, it'd be a lost to humanity if the opencores forums didn't exist... | 08:56 |
| -!- olofk_ is now known as olofk | 08:58 | |
| olofk | Just discovered synflow. One day I will have to try all of those high-level HDL efforts | 09:13 |
| olofk | I think all of these new languages that use verilog as an intermediate language would benefit from my idea of creating a small subset of verilog for generated code | 09:16 |
| blueCmd | stekern: oh, I never seen it being generated I think, good to know | 10:33 |
| blueCmd | stekern: have you tried building gcc with enable-multilib ? | 10:59 |
| stekern | yes, it fails iirc | 11:37 |
| stekern | blueCmd: what do you mean, generated? | 12:07 |
| stekern | here is where it's accounted for in the kernel: http://lxr.free-electrons.com/source/arch/openrisc/kernel/head.S#L176 | 12:09 |
| blueCmd | stekern: "generated" as in I never seen accesses that are negative except when I cock up | 12:18 |
| blueCmd | from gcc that is | 12:19 |
| stekern | yeah, they'd only happen in leaf functions | 12:25 |
| stekern | either I'm stupid or this mac->phy connection system in Linux is just a big f*cking mess | 12:27 |
| stekern | could be both too | 12:28 |
| stekern | bah, I give up, I would have already written a fake phy in the amount of time I've wasted on this | 12:54 |
| stekern | and it won't work out of the box anyway, so I'd need to carry patches for it.. | 12:54 |
| blueCmd | stekern: "Warning: Int on line 2 pending: ignored | 15:07 |
| blueCmd | these lock-ups are weird | 15:07 |
| blueCmd | can I dump state in or1ksim somehow while it's running? | 15:08 |
| blueCmd | I guess there would be some fancy command for the "(sim)" prompt | 15:08 |
| blueCmd | I always manage to crash or1ksim when using that though | 15:08 |
| blueCmd | http://0d9324fa90eeee95.paste.se/ | 15:09 |
| blueCmd | weird, I mean that's supposed to be in usermode so clearly things are alive. running "trace" and just watching the PC it jumps between being inside and outside of 0xc0000000 | 15:13 |
| stekern | what are you doing when you get the lock-up? | 15:21 |
| stekern | and that warning comes from or1ksim | 15:21 |
| stekern | ...if you didn't know | 15:21 |
| olofk | Hi! I just invented a new language. It allows you to do assignments from variables that don't exist and you won't get any warnings | 16:25 |
| olofk | It's like the best feature ever | 16:25 |
| blueCmd | stekern: yes, I guessed it did | 16:28 |
| blueCmd | stekern: at that time I tried opening 8 SSH sessions to the simulator at the same time | 16:28 |
| blueCmd | stekern: http://openrisc.debian.net/tmp/rtems-or1k.patch FYI | 16:38 |
| stekern | olofk: I've got a name for it, let's call it verilog! | 16:52 |
| stekern | blueCmd: 8 ssh sessions, ambitous ;) | 16:56 |
| stekern | I don't even have ssh in my test-setup, so I can't try | 16:56 |
| blueCmd | stekern: do you run "make check" on binutils before submitting patches? | 16:58 |
| blueCmd | I always get failures so I don't know if I do it correctly :P | 16:58 |
| olofk | blueCmd: Try make check > /dev/null | 17:26 |
| olofk | Making some progress with my Wishbone stream writer. It seems to read from a FIFO and generate wb transactions fine now | 17:27 |
| olofk | And I got a pair of handy FIFO BFM's in the process | 17:29 |
| olofk | So just give me a FIFO and I will verify the shit out of it | 17:30 |
| blueCmd | olofk: are you dumping wishbone traffic? | 18:04 |
| olofk | blueCmd: No, but writing a monitor is on my TODO list :) | 18:12 |
| olofk | I'm writing a core that takes a data stream from a FIFO interface and writes to a memory area over wishbone | 18:12 |
| blueCmd | ah ok | 18:15 |
| poke53281 | blueCmd: My "Hello world" filesystem works. A tmpfs-type filesystem written in Javascript | 18:35 |
| blueCmd | poke53281: well, as I told you - everything works for me, I just get random lockups after a while | 18:55 |
| olofk | Just read this : http://turbogrill.blogspot.nl/2014/06/doom-on-zpu.html Has anyone tried to run DooM on OpenRISC? | 19:08 |
| stekern | I compiled SDL doom, but never got it to run. | 19:11 |
| poke53281 | nope | 19:12 |
| poke53281 | but sooner or later I will try it again | 19:13 |
| stekern | yeah, we need some new 90's game to show off at orconf | 19:13 |
| stekern | ;) | 19:13 |
| stekern | we can't break that tradition | 19:14 |
| stekern | at orconf2012 we had 2d bump map | 19:14 |
| stekern | at orconf2013, day of the tentacle | 19:15 |
| poke53281 | hmm, I have my c64 emulator somewhere, which runs in the terminal. We can try some nice text adventures. | 19:15 |
| poke53281 | dosbox might be possible to compile | 19:16 |
| poke53281 | But I guess, you didn't show Day of the Tentacle with sound. | 19:21 |
| stekern | no | 19:21 |
| olofk | But the ac97 core works ok on the Atlys though, right? | 19:22 |
| stekern | I tried running it with sound, but it didn't really work | 19:22 |
| olofk | Otherwise I would talk to the maintainer about that | 19:22 |
| stekern | olofk: yes | 19:22 |
| stekern | I wanna do that OPL2 clone though | 19:22 |
| olofk | That requires a full synth, right? | 19:23 |
| stekern | mmm, what would the alternative be, a half synth? | 19:24 |
| olofk | :) | 19:24 |
| blueCmd | one should make an open-source z10 on an FPGA | 19:37 |
| blueCmd | I forgot how long time it takes to compile gcc inside qemu-user | 19:41 |
| stekern | z10? | 19:45 |
| blueCmd | stekern: s390-ish, it's the IBMs Mainframe stuff | 19:48 |
| blueCmd | pretty sure a lot of companies only keep them around because they have ancient programs on them | 19:49 |
| stekern | ah | 19:58 |
| -!- Netsplit *.net <-> *.split quits: bentley`, simoncook, heroux, knz, mboehnert | 20:11 | |
| -!- heroux_ is now known as heroux | 20:11 | |
| --- Log closed Mon Jul 28 00:00:39 2014 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!