01:20 <somlo_> sajattack[m]: instructions for opensbi are relatively simple: `make CROSS_COMPILE=riscv64-unknown-linux-gnu- PLATFORM=generic FW_PAYLOAD_PATH=~/linux/arc/riscv/boot/Image FW_FDT_PATH=~/litex_dts/nexys4ddr_fpu.dtb`

01:21 <somlo_> requires the uart patch applied on top of upstream opensbi at https://github.com/litex-hub/opensbi/tree/litex-rocket-rebase/

01:21 <somlo_> needs the compiled .dtb rather than using directly the .dts file shipped with the linux-on-litex-rocket project

01:22 <somlo_> produces a single `fw_payload.bin` blob that can be loaded as `boot.bin` same as when using bbl

01:23 <somlo_> I'll post proper instructions on linux-on-litex-rocket once the patch makes it into upstream opensbi

01:23 <somlo_> you can of course do the `boot.json` thing and ship each piece as a separate file, I just find it more convenient to package it all up into a single boot.bin :)

01:29 <sajattack[m]> cool thanks

01:30 <sajattack[m]> I'm experiencing a weird behavior where litepcie_util info, litepcie_util scratch_test and litepcie_util dma_test all work, but I can't get a uart. Could it be a timing fail or did I somehow screw up my crossover?

01:42 <sajattack[m]> ummm?... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/5aa94d99a05f30e6a60868cb7627af0aab42e0d3)

01:48 <sajattack[m]> look guiz I made windows for riscv :P

01:50 <tpw_rules> EFI applications are almost exactly PE .EXE files

01:51 <tpw_rules> (and EFI filesystems are almost exactly FAT32)

01:54 <sajattack[m]> yeah weird

02:00 <cr1901> Oh God, I just realized all my parts/joins are showing up as I test a new IRC client

02:00 <cr1901> oops

02:07 <sajattack[m]> what did I do to my crossover???? https://hatebin.com/fnjeevkqys

02:07 <tpb> Title: hatebin (at hatebin.com)

02:08 <sajattack[m]> I did --uart-name=crossover

02:29 <sajattack[m]> is it crossover+bridge now?

05:38 <sajattack[m]> somlo: it looks like crossover uart is broken on full variants

05:38 <sajattack[m]> tried full and full4d

05:39 <sajattack[m]> gonna try linux4 next to confirm

05:40 <sajattack[m]> also for some reason DMA is slower on the full variant

05:56 <sajattack[m]> nevermind, linux4 isn't working either

05:56 <sajattack[m]> so maybe it's something else

06:00 <sajattack[m]> I shouldn't have updated litex :P

06:11 <sajattack[m]> yep it's the toolchain

06:47 <sajattack[m]> even a fresh clone doesn't work

06:47 <sajattack[m]> I'm out of ideas for tonight

21:37 <SMB784> Hey everyone, I finally was able to integrate my custom bitstream for the SQRL Acorn CLE-215 with the LiteX litepcie bitstream. I used this to integrate the SQRL Acorn with a thunderbolt 3 enclosure, so now these boards can be accessed via PCIe without actually needing a PCIe board (just a USB-C Thunderbolt connection common on most computers). This has full x4 speed over the PCIe connection, and I have provided the verilog files so you can add your o