| --- Log opened Mon Jan 16 00:00:16 2017 | ||
| olofk_web | promach: Yes, I saw your question. Been offlne a bit due to new internet connection at home | 02:47 |
|---|---|---|
| olofk_web | How do you install FuseSoC? Just running "pip install fusesoc" should work | 02:47 |
| olofk_web | I'm a bit suprised also that it looks like it tries to install fusesoc 1.5. It should use fusesoc 1.6 now | 02:48 |
| promach | Traceback (most recent call last): | 02:50 |
| promach | File "/usr/bin/pip", line 7, in <module> | 02:50 |
| promach | from pip import main | 02:50 |
| promach | ModuleNotFoundError: No module named 'pip' | 02:50 |
| promach | this is also why I tried the virtualenv method | 02:50 |
| olofk_web | stekern: Looking a bit more at the qsys stuff, I see that it's a bit awkward. If I understand it correctly, the qsys file is transformed into a qip and a .v file (plus tons of report files). Is that correct? | 02:51 |
| olofk_web | promarch: Which version of Python do you use, and which os/distro? | 02:51 |
| promach | arch linux | 02:53 |
| olofk_web | Maybe you need to install pip manually first | 02:54 |
| promach | Python 3.6.0 | 02:54 |
| olofk_web | Hey, why did my other me timeout? | 02:55 |
| promach | Be cautious if you're using a Python install that's managed by your operating system or another package manager. get-pip.py does not coordinate with those tools, and may leave your system in an inconsistent state. | 02:55 |
| promach | https://pip.pypa.io/en/stable/installing/ | 02:55 |
| olofk_web | There seems to be a python-pip package for arch | 02:55 |
| promach | looking for conflicting packages... | 02:56 |
| promach | Packages (1) python-pip-9.0.1-2 | 02:56 |
| promach | Total Installed Size: 6.22 MiB | 02:56 |
| promach | :: Proceed with installation? [Y/n] | 02:56 |
| promach | (1/1) checking keys in keyring [###########################] 100% | 02:56 |
| promach | (1/1) checking package integrity [###########################] 100% | 02:56 |
| promach | (1/1) loading package files [###########################] 100% | 02:56 |
| promach | (1/1) checking for file conflicts [###########################] 100% | 02:56 |
| promach | error: failed to commit transaction (conflicting files) | 02:56 |
| promach | python-pip: /usr/bin/pip exists in filesystem | 02:56 |
| promach | python-pip: /usr/bin/pip3 exists in filesystem | 02:56 |
| olofk_web | ??? | 02:56 |
| promach | Errors occurred, no packages were upgraded. | 02:56 |
| promach | but | 02:56 |
| promach | error: No package owns /usr/bin/pip | 02:56 |
| olofk_web | Did you run the get-pip.py script? | 02:57 |
| olofk_web | From the link you posted | 02:57 |
| promach | which means pip is installed not throught python-pip AUR package | 02:57 |
| promach | no | 02:57 |
| olofk_web | Hm.. strange | 02:57 |
| promach | I ran sudo pacman -Qo /usr/bin/pip | 02:57 |
| promach | to check which package installed /usr/bin/pip to my system | 02:58 |
| olofk_web | You could of course try to just move the conflicting files temporarily to see if that helps | 02:58 |
| promach | I see | 02:58 |
| olofk_web | Maybe some other program on your computer installed pip with the get-pip script at some point | 02:58 |
| promach | means I should do it the proper way by using package management system | 02:59 |
| olofk_web | Yeah, I would recommend installing pip through pacman | 02:59 |
| olofk_web | And then install fusesoc through pip | 02:59 |
| promach | done installing python-pip through pacman | 02:59 |
| olofk_web | Python package managment is a mess | 03:00 |
| olofk_web | Cool | 03:00 |
| olofk_web | The moment of truth :) | 03:00 |
| promach | done installing fusesoc using pip | 03:00 |
| olofk_web | Great | 03:01 |
| olofk_web | Now create an empty directory, which will be your workspace directory | 03:01 |
| olofk_web | cd into that and run 'fusesoc init' | 03:01 |
| olofk_web | I also see that the instructions on the OpenRISC tools install page is a bit outdated | 03:03 |
| promach | http://paste2.org/UAYgzDhs | 03:03 |
| promach | I am now trying to follow https://github.com/openrisc/mor1kx/wiki/Setting-up-a-mor1kx-development-and-verification-environment | 03:03 |
| promach | Loading 10086 words | 03:03 |
| promach | 0 : Illegal Wishbone B3 cycle type (xxx) | 03:03 |
| promach | Hello world, from an OpenRISC system! | 03:03 |
| olofk_web | Great! | 03:03 |
| olofk_web | So it's working | 03:03 |
| promach | Illegal cycle ? | 03:04 |
| olofk_web | You can ignore that one. I should fix the warning | 03:04 |
| promach | just one more question: | 03:05 |
| olofk_web | It's a simulation-only check that looks at the cti field of a wishbone transaction. Unfortunately that is usually not defined at time 0. I should just remove the warning I guess | 03:05 |
| olofk_web | Sure | 03:05 |
| promach | what is the purpose of ::adv_debug_sys:0 | 03:05 |
| promach | I saw this early today as well | 03:06 |
| olofk_web | It's a core that is used to control an OpenRISC (or other CPU) during debugging | 03:06 |
| olofk_web | It hooks up to a special interface of the CPU and can be used to start and stop the CPU, read out registers and set breakpoints | 03:06 |
| olofk_web | So usually, when you connect a JTAG debugger to a running system it goes through (JTAG interface) -> adv_debug_sys -> CPU | 03:07 |
| olofk_web | It can also directly read and write memory | 03:07 |
| olofk_web | Very useful when you want to upload an application over JTAG | 03:08 |
| olofk_web | I've been using it also recently to provide an UART over JTAG. That's very handy too | 03:09 |
| promach | uart16550 ? | 03:11 |
| promach | uart over jtag ? | 03:11 |
| olofk_web | Yes. adv_debug_sys supports multiplexing the ordinary JTAG data stream with uart data, and by using OpenOCD on the PC, we can have both jtag and uart over the same physical link | 03:14 |
| olofk_web | These things are much easier to describe with a drawing :) | 03:14 |
| promach | how difficult it is to port openRISC to Parallella (Zynq7010) which is not officially supported by fusesoc ? | 03:17 |
| shorne | pushed changes to next https://github.com/openrisc/linux/commits/for-next | 05:23 |
| promach | olofk: how difficult it is to port openRISC to Parallella (Zynq7010) which is not officially supported by fusesoc ? | 06:21 |
| stekern | olofk: tbh, I have no recollection about the details of the qsys stuff | 07:57 |
| stekern | but I remember it was a bit awkward, yes | 07:57 |
| olofk_web | promach: Hard to say. The problematic parts are usually the device-dependent stuff, such as memory interface, debugging and clocking | 09:07 |
| olofk_web | For memory interfaces you can probably reuse som existing wb to axi4 bridge, as the memory is exposed through axi4 on the PS side of the Zynq | 09:08 |
| olofk_web | For debugging, I don't know. Maybe it's just to instantiate a xilinx_bscan | 09:08 |
| olofk_web | Clocking shouldn't be a problem really | 09:08 |
| --- Log closed Tue Jan 17 00:00:17 2017 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!