| lkcl | anton really struggled to get microwatt into 10 mm^2 | 01:39 |
|---|---|---|
| lkcl | and ls2 (2nd ASIC) will be about 100 mm^2 in 180nm, so half that for 130 nm is estimated 50 mm^2 | 01:41 |
| lkcl | programmerjake, email michiel. but first add each bugreport full URL link into the Schedule A | 10:33 |
| lkcl | i'm overloaded | 10:33 |
| markos | lkcl, what is a good day for you to have a call reg. video SVP64 code? | 12:01 |
| markos | just fixed a long standing bug yesterday so I am a bit better wrt time these days | 12:01 |
| lkcl | nice | 12:16 |
| lkcl | how about monday? | 12:17 |
| lkcl | markos, i've no other meetings monday | 12:21 |
| markos | monday is good for me too | 12:21 |
| lkcl | not too early, if that's ok :) after 12:00 UTC? any time's good | 12:23 |
| markos | could we do it a couple of hours later? | 12:59 |
| lkcl | sure | 13:28 |
| lkcl | threw together the libdram firmware from a hybrid of microwatt bare-metal programs and the gram library | 16:18 |
| lkcl | now compiling it into the VERSA_ECP5 to see if it works | 16:19 |
| lkcl | simulation at least shows the UART working but there's no verilator simulation for the DRAM so i can't test that | 16:19 |
| tplaten | I'm trying to install nmigen-boards which fails. I've posted a mail to the list | 16:22 |
| lkcl | tplaten, under no circumstances utilise the trademark-infringing amaranth | 16:32 |
| lkcl | you need only get the 32 lines of code here and place it into a file | 16:32 |
| lkcl | https://github.com/orangecrab-fpga/orangecrab-examples/blob/main/nmigen/blink.py | 16:32 |
| lkcl | you should already have nmigen-boards installed by the libre-soc dev-env-setup scripts | 16:33 |
| lkcl | i repeat, UNDER NO CIRCUMSTANCES utilise the trademark-infringing amarath | 16:33 |
| lkcl | you can if you want to re-run the dev-env-setup scripts in a separate chroot but if you already have them installed and already have nextpnr-ecp5 installed then you have everything that you need, already | 16:34 |
| tplaten | OK, i'll use that one: https://git.libre-soc.org/git/nmigen-boards.git | 16:50 |
| tplaten | I ran the scripts some month ago | 16:51 |
| lkcl | best to use gitlab | 16:51 |
| lkcl | i updated the dev-env-scripts to point to gitlab | 16:51 |
| lkcl | https://gitlab.com/nmigen/nmigen-boards/-/blob/master/nmigen_boards/orangecrab_r0_2.py | 16:51 |
| lkcl | hmmm that's a 25k ECP5 | 16:52 |
| tplaten | once that is fixed, I can try that test. | 16:52 |
| tplaten | Unfortunately LUNA Now Uses Amaranth HDL, the project looks quite interesting | 16:53 |
| lkcl | yes i know. it's quite disturbing and shocking how the entire HDL community has completely disregarded Trademark Law and basic respect | 16:54 |
| lkcl | a temporary "hack" of orangecrab_r0_2.py you can substitute line 14 | 16:55 |
| lkcl | device = "LFE5U-85F" | 16:55 |
| tplaten | Until half a year ago, I was unaware that nMigen is trademarked. | 16:56 |
| lkcl | it took some considerable time for Sebastien to prepare it. | 17:04 |
| lkcl | we (Libre-SOC) and Toshaan Bharvani are the authorised maintainers of nmigen, under License by M-Labs. | 17:04 |
| lkcl | we're literally the only team protected by - and compliant with - Trademark Law. | 17:05 |
| tplaten | What are the Amaranth developers doing wrong, the software License is the BSD-2-Clause and Trademark law is completely different from Copyright law. | 17:08 |
| lkcl | yes, ok so you know that they are two completely different branches of Law | 17:08 |
| lkcl | and that you have to comply with both | 17:08 |
| lkcl | and, further, that compliance with just one of them does *not* confer rights or remove any obligation *to* comply with the other | 17:09 |
| lkcl | i.e. just because you are complying with the BSD license does *not* automatically eliminate or remove a need to comply with Trademark Law | 17:09 |
| tplaten | I've read https://www.gnu.org/philosophy/not-ipr.en.html, but the Trademark thing is harder to understand | 17:10 |
| lkcl | yes it is, because it's just not so commonly used in FOSS | 17:10 |
| lkcl | but basically it protects "reputation". | 17:10 |
| lkcl | it provides actual legal teeth for your "reputation" | 17:11 |
| lkcl | so you know that if there is no Copyright License, you have to go to the author to ask permission? | 17:11 |
| tplaten | And "reputation" may be damaged my renaming or forking a project | 17:11 |
| lkcl | well, if the fork is a "minor" one that does not change the API, then yes, that's fine | 17:12 |
| tplaten | For copyright I know, and for Trademarks I thing there is some asking for permission too. | 17:12 |
| lkcl | and if the "renaming" is done in such a way that the owner of the Trademark's business is literally destroyed, then yes, that's clearly "damage" | 17:12 |
| lkcl | such as stating "nmigen is DEAD. it HAS been RENAMED to amaranth" | 17:13 |
| lkcl | i mean, that's so blindly obviously completely destroying nmigen and nmigen's reputation that i'm deeply shocked at the total ignorance of the developers of the consequences of their actions | 17:13 |
| lkcl | yes, there is, i was going to say, "default permissions" for Trademarks, as well | 17:14 |
| tplaten | That is clearly a false statement. nmigen is stable software that is still maintained. When changing the API you can do that in a 2.0 branch without renaming. | 17:14 |
| lkcl | you can *use* a Trademark without seeking permission, as long as you *only* use it. | 17:14 |
| lkcl | by default, distribution is not permitted. | 17:15 |
| lkcl | i know! | 17:15 |
| tplaten | That is definively too complex for me. | 17:15 |
| lkcl | well, if you change the API (any changes), it could cause harm to the main Trademark, by making it "incompatible". that would hurt the Trademark's reputation. | 17:16 |
| lkcl | yes, it's complex. we're covered because we have sought - and obtained - a Trademark license as the exclusive maintainers | 17:16 |
| tplaten | I now continue on the orangecrab-examples. | 17:17 |
| lkcl | ok :) | 17:17 |
| lkcl | just take those 32 lines of code, blink.py | 17:17 |
| tplaten | I'll do | 17:17 |
| lkcl | modify orangecrab_r0_2.py in nmigen-boards to say "LFE5U-85F" | 17:18 |
| lkcl | and it should work. | 17:18 |
| lkcl | i should have mine soon | 17:18 |
| lkcl | the parcelforce website is offline at the moment, doh | 17:18 |
| lkcl | oh you will need dfu-util | 17:19 |
| lkcl | git clone git://git.code.sf.net/p/dfu-util/dfu-util | 17:19 |
| lkcl | we really need a new dev-env-setup script which installs fujprog, dfu-util, and all the other installers like openocd | 17:20 |
| tplaten | done: modify orangecrab_r0_2.py in nmigen-boards to say "LFE5U-85F" | 17:22 |
| lkcl | i'll have to create a new board definition which overrides that | 17:26 |
| lkcl | OrangeCrabR0_2_85k_Platform | 17:27 |
| lkcl | or something | 17:27 |
| lkcl | similar to how it's done for the ulx3s | 17:27 |
| tplaten | OrangeCrabR0_2_85k_Platform looks good. | 17:31 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!