| *** tpb has joined #litex | 00:00 | |
| *** roboknight has quit IRC | 00:36 | |
| *** lf has quit IRC | 00:53 | |
| *** lf has joined #litex | 00:53 | |
| *** Degi_ has joined #litex | 01:06 | |
| *** Degi has quit IRC | 01:09 | |
| *** Degi_ is now known as Degi | 01:09 | |
| *** FFY00 has joined #litex | 02:04 | |
| *** Dolu has quit IRC | 02:51 | |
| *** Bertl_oO is now known as Bertl_zZ | 03:00 | |
| *** shenki has joined #litex | 04:53 | |
| *** futarisIRCcloud has joined #litex | 05:22 | |
| futarisIRCcloud | https://shop.lambdaconcept.com/home/46-1-ecpix-5.html | 05:22 |
|---|---|---|
| tpb | Title: ECPIX-5 (at shop.lambdaconcept.com) | 05:22 |
| futarisIRCcloud | https://groupgets.com/campaigns/868-butterstick | 05:23 |
| tpb | Title: ButterStick | GroupGets (at groupgets.com) | 05:23 |
| *** proteusguy has quit IRC | 06:16 | |
| *** proteusguy has joined #litex | 06:36 | |
| *** lindemer has joined #litex | 07:11 | |
| lindemer | Can anyone tell me how to debug LiteX/VexRiscv with GDB on an Arty FPGA? The best instructions I can find are https://github.com/timvideos/litex-buildenv/wiki/Debugging but running "load firmware.elf" from GDB once connected returns communication errors and doesn't boot the image | 07:14 |
| *** lindemer has quit IRC | 07:16 | |
| *** lindemer has joined #litex | 07:16 | |
| *** lf has quit IRC | 07:16 | |
| *** lf has joined #litex | 07:18 | |
| *** kgugala_ has joined #litex | 07:21 | |
| *** kgugala has quit IRC | 07:23 | |
| *** kgugala has joined #litex | 07:23 | |
| *** kgugala_ has quit IRC | 07:26 | |
| *** futarisIRCcloud has quit IRC | 07:31 | |
| *** kgugala has quit IRC | 07:46 | |
| *** kgugala has joined #litex | 07:46 | |
| _florent_ | somlo: nice! | 08:16 |
| *** keesj has quit IRC | 08:45 | |
| *** keesj has joined #litex | 08:52 | |
| *** Bertl_zZ is now known as Bertl | 10:06 | |
| *** Dolu has joined #litex | 10:36 | |
| *** futarisIRCcloud has joined #litex | 13:28 | |
| *** CarlFK has quit IRC | 15:06 | |
| *** lindemer has quit IRC | 15:38 | |
| *** kgugala_ has joined #litex | 16:34 | |
| *** kgugala has quit IRC | 16:38 | |
| *** kgugala_ has quit IRC | 16:47 | |
| *** kgugala has joined #litex | 16:48 | |
| *** kgugala has quit IRC | 16:48 | |
| *** kgugala has joined #litex | 16:48 | |
| *** kgugala_ has joined #litex | 16:56 | |
| *** kgugala has quit IRC | 16:58 | |
| *** kgugala has joined #litex | 16:58 | |
| *** kgugala_ has quit IRC | 17:01 | |
| *** SpaceCoaster has quit IRC | 17:02 | |
| *** SpaceCoaster has joined #litex | 17:03 | |
| *** kgugala_ has joined #litex | 17:43 | |
| *** kgugala has quit IRC | 17:46 | |
| *** kgugala has joined #litex | 17:59 | |
| *** kgugala_ has quit IRC | 18:01 | |
| mithro | somlo: I'm somewhat excited about "linux-on-litex-rocket" | 18:35 |
| mithro | somlo / _florent_: I really think that the two repositories will kind of end up trending towards, linux-on-litex-vexriscv --> linux-32bit-on-litex and linux-on-litex-rocket --> linux-64bit-on-litex direction (I don't think we should rename them however) | 18:36 |
| _florent_ | mithro: I see things differently. There is an history behind linux-on-litex-vexriscv, linked to VexRiscv and to the collaboration with @Dolu @daveshah Antmicro, etc... so this project is not going to support other CPUs | 18:52 |
| _florent_ | mithro: same for linux-on-litex-rocket, that is in fact just a name for the work done by @somlo during the last year and that was already available at http://www.contrib.andrew.cmu.edu/~somlo/BTCP/ | 18:53 |
| tpb | Title: A Trustworthy, Free (Libre), Linux Capable, Self-Hosting 64bit RISC-V Computer (at www.contrib.andrew.cmu.edu) | 18:53 |
| mithro | I would say that BlackParrot for example is more similar to Rocket than VexRISCV | 18:55 |
| *** Bertl is now known as Bertl_oO | 18:55 | |
| leons | _florent_: I'm personally equally interested in VexRiscv (which is awesome) and ibex (because it has more mature than VexRiscv) in the OS I'm working on | 18:55 |
| leons | Does this comment mean that https://github.com/enjoy-digital/litex/issues/695 isn't going to happen? | 18:55 |
| mithro | _florent_: Generally I have found that optionality tends to lead to both options being better than either one by itself | 18:56 |
| mithro | leons: I'm sure that #695 will happen at some point, but it will probably be zephyr / microcontroller targetted rather than Linux | 18:56 |
| leons | _florent_: Or was that only referring to Linux support? | 18:57 |
| mithro | _florent_: But we can keep the linux-on-litex-vexriscv specific repository to VexRISCV | 18:57 |
| leons | mithro: Ah, that makes sense (and conveniently is exactly what I want :) ) | 18:57 |
| mithro | It might get a bit silly when we end up having linux-on-litex-vexirscv, linux-on-litex-rocket, linux-on-litex-blackparrot, linux-on-litex-microwatt, linux-on-litex-XXXX | 18:58 |
| _florent_ | @leons yes that's the way I see thing, linux-on-litex-vexriscv, linux-on-litex-rocket are applications projects for both LiteX/Vexriscv/Rocket | 18:58 |
| mithro | I do wonder if Dolu will end up looking into an VexRISCV which is 64bit capable but more efficient than rocket | 18:59 |
| leons | And there's nothing stopping anyone from putting in the work to make Linux run on LiteX with these CPUs, right? Just not an official project? | 19:00 |
| somlo | mithro, _florent_: with luck (and I'm personally pretty optimistic about that) we can ensure upstream Linux drivers (and support in general) can be done in a way that works out-of-the-box for both 32 and 64 bit CPUs, without too much in the way of special-case hackery | 19:00 |
| mithro | leons: There isn't really any "official" project here -- everything is mostly driven by what people are interested in | 19:01 |
| mithro | somlo: So am I! | 19:01 |
| mithro | Here is my previous attempt ad trying to "map" the "LiteX ecosystem" - https://docs.google.com/drawings/d/15hSX1cHwz2_Lm-CxQxlelGUuVPP4AmEEljyRb1ncjjk/edit | 19:01 |
| tpb | Title: LiteX Ecosystem Diagram - Google Drawings (at docs.google.com) | 19:01 |
| _florent_ | @leons: I would be happy to see https://github.com/enjoy-digital/litex/issues/695 happens, just that I'm trying to limit the scope of linux-on-litex-vexriscv not LiteX | 19:03 |
| _florent_ | somlo: that's indeed the direction we are going I think thanks to your work (BTW linux-on-litex-vexriscv is now also using your litex-rocket-rebase branch) | 19:05 |
| somlo | speaking of LiteX vs. linux-on-litex-vexriscv: one of my highest priorities for linux-on-litex-rocket was to use unmodified upstream LiteX, with no customizations in terms of gateware :) | 19:05 |
| somlo | _florent_: I *thought* I saw something about Antmicro's LiteX soc controller landing upstream yesterday (not 100% sure, haven't had time to check) | 19:07 |
| _florent_ | somlo: re no customization: that's indeed nice, that also the direction I'd like to go with linux-on-litex-vexriscv | 19:07 |
| _florent_ | there are still some customization but I think we could remove it easily | 19:08 |
| somlo | I think that'd be the 8-bit-csr-data-width version that's been winding its way through reviews, so my next priority (next week I hope) is to rebase "litex-rocket-rebase" on top of that work, and to submit upstream the "works with both 8-bit and 32-bit CSRs" portion of that ASAP, before anyone gets too attached to what just got committed :D | 19:08 |
| _florent_ | somlo: indeed: https://twitter.com/stffrdhrn/status/1339688913409957888 | 19:09 |
| _florent_ | sorry I have to go | 19:09 |
| somlo | have a good weekend! | 19:09 |
| somlo | hmmm, apparently the 85k version of the ECPIX-5 is out of stock... :( | 19:29 |
| somlo | https://shop.lambdaconcept.com/home/46-2-ecpix-5.html#/2-ecpix_5_fpga-ecpix_5_85f | 19:29 |
| tpb | Title: ECPIX-5 (at shop.lambdaconcept.com) | 19:29 |
| *** chrisps has joined #litex | 19:31 | |
| daveshah | somlo: I've got one if you want me to test anything | 19:32 |
| daveshah | Almost as hot as new GPUs, it seeme | 19:33 |
| somlo | daveshah: nothing immediately earth-shattering -- I was going to get one for myself as a Christmas present, and putter around with LiteX+Rocket (maybe look into whether the sata port would come in handy) | 19:35 |
| somlo | but I can wait for the next production run for all that, no big deal | 19:36 |
| somlo | maybe the next version will come with a heatsink ;) | 19:37 |
| chrisps | i recently got a vcu1525 cheap despite having no ee experience to use as an accelerator for some projects of mine. did i make a mistake? | 19:39 |
| Finde | chrisps: well I'm jealous at least :) | 19:42 |
| mithro | somlo: Antmicro went with mor1k as a start as Stafford could help with getting it upstream and there is less politics | 20:12 |
| mithro | somlo: LiteX + Rocket is probably easier than LiteX + VexRISCV in Linux because Linux 32bit RISC-V has a lot of haters (and there is increasing moment behind the "drop 32bit in Linux all together" idea). | 20:14 |
| mithro | I would really like to figure out getting CI working which shows that Linux will boot on key hardware in linux-on-litex-vexriscv | 20:16 |
| somlo | mithro: interesting to know, thanks for the context! I will probably collapse the litex-vexriscv-rebase and litex-rocket-rebase branches into a single litex-rebase branch that applies cleanly on the current stuff upstream | 20:18 |
| somlo | then we get to bicker about upstreaming the accessors that handle all corner cases some more, but once that's done, additional device drivers should have a much easier time | 20:19 |
| mithro | somlo: I was hoping that the microwatt people would help with Linux on LiteX upstream work but that didn't seem to pan out | 20:19 |
| somlo | imho, device drivers have no business *knowing* whether the CSR data width is 8 or 32, or whether the underlying cpu is 32 or 64 bits :) | 20:20 |
| somlo | we have to encapsulate that magic somewhere (i.e. in the accessors), in a way that can be cleaned up to everyone's satisfaction without further disturbing the way device drivers are written | 20:21 |
| mithro | somlo: if you ever want / need a Nexys Video or Gensys 2 boards, I would happily send you one as long as you promise to continue to support this stuff on small devices like Arty :-) | 20:21 |
| mithro | Specially if you wanted to help make sure LiteX + Rocket works with Yosys + VPR + Prj X-Ray flows | 20:22 |
| somlo | mithro: thanks, but I don't have an arty :) smallest device I do have is nexys4ddr | 20:22 |
| mithro | I'm happy to send you an Arty too | 20:22 |
| somlo | in the xilinx category, at least | 20:22 |
| somlo | In the Xilinx category I was waiting for the dust to settle on the vpr vs. nextpnr-xilinx vs prj-xray situation, and stick with ecp5 for the turtles-all-the-way-down-FOSS self-hosting ideal | 20:28 |
| somlo | but maybe it's time to re-evaluate that at some point... Personally, my first priority is getting something that can mount and read/write an ext2 filesystem in a not-too-slow way | 20:29 |
| somlo | so I can boot Fedora on it, and run yosys/trellis/nextpnr and/or vpr on the "self-hosting" thing itself | 20:30 |
| acathla | "<somlo> hmmm, apparently the 85k version of the ECPIX-5 is out of stock... :(" | 20:36 |
| acathla | It seems both versions are out of stock. That's a good news, it means the business is good :) | 20:36 |
| daveshah | Or the AMD/NVIDIA trick of not making enough to create an illusion of popularity ;) | 20:40 |
| zyp | I'm not sure whether I'm more excited for the ECPIX-5 or the butterstick | 20:41 |
| zyp | I ordered one of each the other day | 20:41 |
| *** FFY00 has quit IRC | 20:43 | |
| *** FFY00 has joined #litex | 20:44 | |
| acathla | I'm not sure they had a clue about how many the were going to sell, so they probably prefered to produce just a few. | 20:47 |
| acathla | Who's behind lambdaconcept? I'm sure they are hiding here :) | 20:49 |
| _florent_ | acathla: key2 probably know them :) | 21:09 |
| key2 | We really sold out | 21:10 |
| key2 | what pisses me off is companies buying sometimes 5 unit making the batch go fast | 21:11 |
| acathla | Like pasta or toilet paper in France during lockdown :) | 21:18 |
| futarisIRCcloud | Is there another batch in the queue? | 21:20 |
| felix_ | just had a look in my desk drawers and found a digilent atlys board with corresponding i/o breakout board and a numato opsis in there; if someone wants those and actually has the time to do something useful with those, i cold ship you those; preferrably in germany or at least europe due to shipping cost. it's been quite some time that i did something with the boards, so i'm not sure how functional those still | 22:09 |
| felix_ | are | 22:09 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!