IRC logs for #openrisc Tuesday, 2015-02-24

--- Log opened Tue Feb 24 00:00:54 2015
poke53281Today I was shoicked, when I received a pull request from stekern.08:15
poke53281for Javascript08:16
poke53281stekern on the dark side of the force.08:22
poke53281What was the reason for the patch anyhow. rtems? Do you know, how to use the example?08:23
olofkpoke53281: Is it a timer patch? I know that he needed some extra features for his lk kernel as well08:39
poke53281olofk: Yes, a timer patch.08:43
poke53281lk kernel?08:44
olofkhttps://github.com/travisg/lk08:45
olofkLooks like it's used as a recovery/bootloader kernel for many Android devices08:46
-!- Netsplit *.net <-> *.split quits: _franck_09:05
poke53281Hmm, Ok09:06
stekernpoke53281: http://oompa.chokladfabriken.org/openrisc/jor1k/demos/lk.html09:18
poke53281TODO very urgent: Add terminal history ...09:37
poke53281Again no clue, what to do09:38
poke53281But at least, I can calculate fibonacci numbers.09:39
poke53281But again great job stekern.09:49
olofkpoke53281: Fibonacci series is like the Hello world or blinking LED for mathematicians and language designers :)09:49
poke53281Well, this was the last option in the help command. The first commands I can't see, because I don't have a scrollback buffer.09:50
poke53281Wonder, what the "hello world" equivalent is for physicists? Dropping an apple?09:51
olofklol09:53
olofkYEAH! Gravity works!09:53
juliusbI love this jor1k instance proliferation10:13
olofkYes, it's a great way to show off new stuff10:20
olofkIs anyone using FuseSoC with python 2.7 btw? I'm getting some strange errors10:20
olofkSeems to work now after a reinstall10:22
olofkAt least I didn't have to clean out the registry and disable anti virus programs :)10:23
olofkIs opencores svn down?10:24
stekernpoke53281: it's not so much you can do there, it's just some test applications built in there12:53
stekernbut with the timer patch applied all of the tests works as well in jor1k as in or1ksim12:54
olofkI like the synergies here. jor1k was great to test the toolchain and find bugs/missing features in the RTL. Now we port new OSes and find missing features in jor1k13:03
poke53281olofk: Well, you know that the missing features are obvious for me. I just didn't expect to ever run another operating than lLinux with this emulator.13:23
stekernmmm, no really major things though. Linux doesn't use that mode, I guess poke53281 omitted the support for that13:24
poke53281So I implemented exactly the things necessary to run Linux.13:24
poke53281And Linux uses only one timer mode.13:24
poke53281exactly stekern.13:24
poke53281I also don't support add with carry flag.13:25
poke53281because gcc doesn't use it.13:25
poke53281The flag thingy is more or less broken Did someone ever tested the proper setting of flags in or1ksim?13:26
stekernflag thingy?13:27
poke53281Yes, the flags in the supervisor register.13:27
poke53281I remember that when I read the or1ksim code a long time ago, that there were some comments because of the flags.13:28
stekernah, no probably not...13:28
poke53281That the documentation is wrong or that there was a question in the comments, if the flags are correctly implemented.13:29
poke53281add and sub with carry is kind of important for 64 Bit values.13:30
poke53281on 32 Bit machines13:30
poke53281I have never looked at our compiled code. Probably done with l.sf....13:31
stekernI really want to try out 'bandvigs' fpu stuff13:31
stekernI wonder how complete they are at this point13:31
poke53281I am curious as well.13:32
stekernright now I'm reading code I've written 1.5 years ago and I'm trying to understand what it does..13:35
poke53281You should accustom yourself to write your current alcohol level in the comments when you code :)13:36
poke53281But I have such problems too. But not with programs, which I have written from scratch. Only with programs, were I had to follow other design principles and variables and functions.13:38
stekernI'm pretty sure it was fairly low when I wrote it, but I'm not so sure about my current state13:39
stekernfwiw, this is what I'm looking at: https://github.com/skristiansson/linux/commit/f8dbfc43d7bdeea64810e683f4eec7adb5f3aeda13:39
poke53281Looks like part of your hardware TLB lookup13:41
stekernit is, and there was some nice cleanups in the layout of the Linux PTE done for the sw lookup as weel13:42
stekernand I'd like to apply those to lk npw13:42
stekerns/npw/now13:42
poke53281I remembered, that I asked you to use a supervisor register to activate  this feature :)13:44
olofkLooks like I will have to change all references from opencores svn to github in orpsoc-cores. It's very shaky nowadays13:45
poke53281And you asked about the discrepancies between the implementation in the Linux kernel and in the specification.13:46
stekernhmm, I think that's slightly unrelated14:02
stekernthe discussion was whether a 0 value in the PTBP field would be regarded as hw tlb refill disabled14:04
stekernolofk: yeah, I had some troubles fetching the simple_spi core when I needed to check the orpsoc-cores changes vs orpsocv214:55
olofkLoneTech: Do you know anything about SVN problems on OpenCores?18:47
olofkAsked this before, but can we send in kernel patches without a kernel maintainer?19:10
olofkI'm releasing FuseSoC 1.2 now btw19:42
olofkOk done. Just need to compile a FTP client to upload the tar ball19:54
olofkok done. Bring out the cocaine!20:08
Me1234Where can be working linux on de0_nano downloaded? kernel.org?21:05
--- Log closed Wed Feb 25 00:00:56 2015

Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!