| --- Log opened Thu Mar 24 00:00:49 2016 | ||
| _franck__ | http://opencores.org/forum,Cores,0,5762 | 03:45 |
|---|---|---|
| olofk | _franck_: I saw that. Alex Bradbury tweeted about it yesterday. | 03:46 |
| _franck__ | ok, sorry I'm not tweet aware :) | 03:50 |
| olofk | haha | 03:50 |
| stekern | twitter is dangerous, I try to avoid it | 03:54 |
| stekern | you can get stuck in it, in the same way as you can get stuck on youtube | 03:54 |
| olofk | Mmmm :( | 03:56 |
| wallento | hesham: yes, thats the plan after the current updates are upstream | 04:40 |
| wallento | I have already adopted it accordingly and will put it into github.com/openrisc/tutorials | 04:41 |
| shorne | stekern: I think openRISC has helped to cure my youtube addiction | 08:04 |
| stekern | shorne: heh | 08:04 |
| stekern | oh, and I didn't get around to pull out Linux patches last weekend :( | 08:05 |
| shorne | so... from the response from Jonas I spent the past day and a half looking into adding a memcpy builtin to the openrisc gcc backent | 08:06 |
| stekern | let's see if the easter holidays let me spend some cycles on it | 08:06 |
| shorne | backend... I think I figured it all out, but I dont really want to touch that | 08:06 |
| shorne | I see you added some things to gcc before, any thoughts? | 08:07 |
| stekern | not without going back looking at it | 08:07 |
| shorne | the whole .md (lisp like) matchers and code generators are a bit.. much | 08:08 |
| shorne | re: patches, thanks Im just working on the todo list. looking at blockmem cleanups now | 08:09 |
| stekern | .md, yeah... | 08:10 |
| shorne | stekern: do you plan on pulling out the smp patches? | 08:10 |
| stekern | llvm is nicer in that sense | 08:10 |
| stekern | I thought I'd only pull out the really critical once | 08:10 |
| shorne | alright | 08:11 |
| stekern | the ones that fix broken things in the upstream kernel | 08:11 |
| shorne | yeah with llvm you can really easily see the boundary between front end and backend. With GCC its there too but not so well documented or easy to figure out on your own. there are actually ton's of docs but I dont really find what I need there | 08:12 |
| wallento | by the way, whats the state of llvm? | 10:36 |
| wallento | I am ready to build the next crosscompiler :) | 10:36 |
| olofk | Sometimes I wonder why I log in to all these channels, when it's essentially the same people in all of them :) | 13:42 |
| wallento | yeah, thats the same like when facebook, google, whoever buys a social media startup and people say they acquired them for the extra x00 million users | 13:49 |
| wallento | I am always thinking that those are like 99% the same people probably.. | 13:50 |
| wallento | olofk: I played around with fusesoc a bit | 13:52 |
| wallento | I am thinking of a librecores.org integration and have some thoughts how this best works out | 13:52 |
| wallento | should I put them into a Github issue? | 13:53 |
| wallento | the ideas are essentially around online "cores_root" (extend the mechanisms for cacheable web api calls) | 13:54 |
| wallento | they would allow for dependencies like librecores:olofk/spi | 13:54 |
| wallento | and the librecores api resolves this to a core-file URL | 13:54 |
| wallento | to download | 13:54 |
| wallento | probably something to merge with your dependency ideas | 13:55 |
| olofk | wallento: This sounds interesting. It's basically the same idea as Gentoo overlays | 15:08 |
| olofk | One thing to investigate would be to see if it would make sense to use the full IP-XACT vlnv for this | 15:09 |
| olofk | Where in your example, librecores would be vendor, olofk would be library and spi would be component | 15:09 |
| olofk | The general problem with overlays is of course dependencies between them | 15:10 |
| olofk | Gentoo recently improved this area | 15:10 |
| olofk | Which is why I like to look at how Gentoo solve things. They have already been through a lot of these issues. No need to fall into the same traps :) | 15:11 |
| olofk | wallento: Were you also thinking about scraping the .core files for metadata and present on librecores? Because that was my first idea of librecores integration | 15:12 |
| olofk | And with the ambition to move more stuff from .core files into IP-XACT files (now that FuseSoC can parse basic stuff from IP-XACT files), we could basically present decoded IP-XACT info on librecores, with optional FuseSoC integration | 15:14 |
| wallento | olofk: yes, it should probably point to a file URL or so for the beginning and have some versioning. I will put my thoughts in a google doc and share | 17:21 |
| --- Log closed Fri Mar 25 00:00:39 2016 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!