@ppenna hm, strange. I've created the readme and somehow it got removed again!?
anyways, its there again and I can also fork the repo now
Pedro Henrique Penna
@ppenna
now it works!
Philipp Wagner
@imphil
@ppenna I've looked through the report again, looks mostly good! Please open a issue on GitHub for the two remaining issues you mentioned at the end so that can get tracked (and also link to the issue # in your report)
Pedro Henrique Penna
@ppenna
@imphil thanks! I'm working on the pull requests... Once I'm done with that I'll open the issues
Ghost
@ghost~5910ddcfd73408ce4f5dfa87
Hello, sorry I missed the call, yesterday was a long day I was in bed a few hours before the meeting
If it is ok for you, let me know which task I should work on next: dib in the interrupt system bug or update the documentation for the NoC driver
@imphil the blog post has the build instructions for Linux on OpTiMSoC. I thought that the documentation of the driver itself should be on the Linux source tree... but I don't know...
Philipp Wagner
@imphil
@ppenna Thanks for the blog post draft. How about making a bit more of a story what you did, and possibly add a nice screenshot or a screencast of how things are working? Let's get people excited about what you did!
Pedro Henrique Penna
@ppenna
@imphil sure. I was saving that for the final implementation, but I guess that you want to have something closer to the report that I have written, right?
Philipp Wagner
@imphil
@ppenna in a project like that, there's never a "final" version :) Just take the chance of GSoC ending to tell the world what you did and how much better the world is now :)
Pedro Henrique Penna
@ppenna
=D
_
Ghost
@ghost~5910ddcfd73408ce4f5dfa87
Hello. It's a bit late maybe but I put some comments on the Linux pull request and the gsoc gdoc
Olof Kindgren
@olofk
@stffrdhrn The latest fix for mor1kx with Icarus broke modelsim :/
** Error: ../src/mor1kx_5.0-r2/rtl/verilog/mor1kx_cpu_cappuccino.v(1132): External function 'mor1kx_rf_cappuccino.calc_rf_addr_width' may not be used in a constant expression.
** Error: ../src/mor1kx_5.0-r2/rtl/verilog/mor1kx_cpu_cappuccino.v(1132): Parameter value must be constant.
woah... that wasn't my intention :)
MAKE IT BIG
Ghost
@ghost~5910ddcfd73408ce4f5dfa87
Ah
But not quartus or iverilog
Any suggestion? Make the function
Sorry. I leave
I mean. Move the function to utils?
Olof Kindgren
@olofk
That's probably a good idea. Should be pretty tool-agnostic
And just pull in the function with an `include, you mean ?
Ghost
@ghost~5910ddcfd73408ce4f5dfa87
Yes
Philipp Wagner
@imphil
it's too bad icarus verilog is not supporting any recent SV standards, so it's become pretty much unusable on my side
Ghost
@ghost~5910ddcfd73408ce4f5dfa87
thats unfortunate its pretty fast
Created a pr to fix it
openrisc/mor1kx#57
Plus I have an old PR still open
openrisc/mor1kx#56
@olofk what are your tips for managing local development versions of cores with fusesoc? I have my local versions ahead of fusesoc-cores and orpsoc-cores in my PATH
but... sometimes if I don't have the versions just right it takes the non local versions
So I have to be very careful and always do fusesoc core-info and watch which versions are pulled in
Olof Kindgren
@olofk
Yeah, same here
You can keep multiple workspaces with different fusesoc.conf
Ghost
@ghost~5910ddcfd73408ce4f5dfa87
ah, that might be nice
Olof Kindgren
@olofk
Or if you're just working on a specific core, you can pull that one in with fusesoc --cores-root=/path/to/core sim ...