| *** tpb has joined #tomu | 00:00 | |
| *** rohitksingh has joined #tomu | 01:47 | |
| *** im-tomu has left #tomu | 02:33 | |
| *** im-tomu has joined #tomu | 02:33 | |
| *** CarlFK has quit IRC | 02:44 | |
| *** rohitksingh has quit IRC | 02:47 | |
| *** CarlFK has joined #tomu | 02:52 | |
| *** Vercas has quit IRC | 02:55 | |
| *** migy has quit IRC | 02:57 | |
| *** Smelly_Cologne has joined #tomu | 03:12 | |
| *** Vercas has joined #tomu | 03:13 | |
| Smelly_Cologne | Hello, I received my fomu today but when I plug it into my USB port, no LED turns on. Further, when I try to run the hello world program in https://github.com/stef/nb-fomu-hw, I get "dfu-util: No DFU capable USB device available" | 03:14 |
|---|---|---|
| Smelly_Cologne | I suspect that the device isn't working, I just wanted to make sure I wasn't missing anything | 03:15 |
| *** migy has joined #tomu | 03:17 | |
| Smelly_Cologne | Also, I should add that the fomu is very hot to the touch after plugging it into the usb port | 03:20 |
| xobs | That does sound like a defective Fomu :( | 03:20 |
| xobs | You should contact Crowd Supply to get a new one. | 03:21 |
| Smelly_Cologne | alright, thanks. Just wanted to make sure before I initiated the return | 03:21 |
| *** Smelly_Cologne has quit IRC | 03:25 | |
| kevans91 | 2x fomu arrived here for some FreeBSD testing today, and I was both prepared and unprepared for how tiny these would be | 03:31 |
| xobs | kevans91: they are surprisingly small. | 03:38 |
| xobs | I started working on building wishbone-tool on CircleCI so we could have freebsd builds, but the deploy just isn't working. I gave up after a day :( | 03:38 |
| kevans91 | I'm trying to get our CI folks to poke Cirrus about their missing functionality | 03:39 |
| xobs | It's the weirdest thing. I suspect curl is behaving differently on freebsd, since the exact same command works locally. | 03:40 |
| kevans91 | It's hard to advocate for it when they don't support cutting a release when a tag is pushed, that seems like a thing more projects would rely on | 03:40 |
| kevans91 | Ideally you wouldn't need to do anything special there and curl can go away | 03:40 |
| kevans91 | (I would think- I guess I didn't look at travis stuff to quickly, but I thought it was preemptively creating the release for you) | 03:41 |
| kevans91 | too thoroughly, even | 03:41 |
| xobs | I think so. I'm still not sure how it works. I know about 5% of the time there's a problem with the release and it mysteriously fails. I think there's an ordering dependency, and there's about a 5 second window during which it can fail if the first two buidls finish at the same time. But it doesn't happen often. | 03:42 |
| kevans91 | yeah, the impression I got was that Travis creates the release implicitly somewhere so you don't need to, hence why there can be no release when Cirrus finishes | 03:44 |
| kevans91 | and that Cirrus should really be coordinating that somehow | 03:45 |
| xobs | RIght, but the weird issue with Cirrus is that even if a release exists, it fails with an auth error. Even if the token is good. | 03:45 |
| kevans91 | oh, odd | 03:46 |
| xobs | I've printed out the curl command that's failed, copied it to my local machine, and had it succeed when I put in the correct token. | 03:46 |
| kevans91 | was it an auth error, or a cert error? The CI environment may not have ca_root_nss installed | 03:46 |
| xobs | I've also printed out a fake token (rot13 so it actually prints), and it's valid as well. | 03:46 |
| kevans91 | we don't bundle a root bundle in base in any release yet | 03:47 |
| kevans91 | I was actually a little surprised that the CI environment had curl, even... we don't install that into the GCE images we upload, as far as I can tlel | 03:48 |
| *** xkapastel has joined #tomu | 03:49 | |
| xobs | Here's an example of the output that fails. The script worked when I did it locally: https://cirrus-ci.com/task/6521247622234112 | 03:50 |
| tpb | Title: Cirrus CI (at cirrus-ci.com) | 03:50 |
| xobs | And an example of me printing out the curl, which worked when I ran it locally with the correct secret: https://cirrus-ci.com/task/5208712226799616 | 03:51 |
| tpb | Title: Cirrus CI (at cirrus-ci.com) | 03:51 |
| kevans91 | hmm | 04:12 |
| kevans91 | I guess I can't just use a personal access token for testing | 04:13 |
| kevans91 | ah, no, just need to inject it via the Cirrus interface | 04:14 |
| *** gio has quit IRC | 04:34 | |
| *** gio has joined #tomu | 04:47 | |
| kevans91 | some light testing in a different repo doesn't reveal anything particularly bogus, but I'm issuing a standard GET to fetch my repos w/ authorization (and content-type, just to match) --headers | 05:07 |
| kevans91 | I'll keep digging, though- | 05:07 |
| *** tpb has joined #tomu | 05:11 | |
| *** CarlFK has quit IRC | 05:30 | |
| *** CarlFK has joined #tomu | 05:46 | |
| *** Yatekii has quit IRC | 06:40 | |
| *** whatnick_ has quit IRC | 07:27 | |
| *** whatnick has joined #tomu | 07:32 | |
| *** futarisIRCcloud has quit IRC | 07:41 | |
| xobs | Let me know if you manage to get it working with your repo! | 07:50 |
| *** whatnick has quit IRC | 08:02 | |
| *** whatnick has joined #tomu | 08:02 | |
| *** Yatekii has joined #tomu | 08:42 | |
| gurke_ | is there something like https://knielsen.github.io/ice40_viewer/ice40_viewer.html for the UP5K? | 08:50 |
| tpb | Title: ICE40 layout viewer (at knielsen.github.io) | 08:50 |
| tnt | gurke_: you can use nextpnr's gui | 08:53 |
| gurke_ | yes, i was looking for sth. online | 08:58 |
| gurke_ | but thx | 08:58 |
| *** xkapastel has quit IRC | 09:09 | |
| *** rohitksingh has joined #tomu | 09:14 | |
| *** im-tomu has left #tomu | 09:17 | |
| *** im-tomu has joined #tomu | 09:17 | |
| *** st-gourichon-f has joined #tomu | 09:25 | |
| *** gourichon has quit IRC | 09:30 | |
| st-gourichon-f | Hello everyone. I work with acathla. We use a FOMU hacker. Success with "dfu-util -D micropython-fomu.dfu". Success with riscv-blink from https://github.com/im-tomu/fomu-workshop/ . | 09:32 |
| tpb | Title: GitHub - im-tomu/fomu-workshop: Support files for participating in a Fomu workshop (at github.com) | 09:32 |
| st-gourichon-f | We are considering synthetizing a RISC-V with extensions P and V for efficient computation of neural networks. | 09:33 |
| st-gourichon-f | So far we start with playing with the softcore included with the FOMU Hacker bootloader. | 09:33 |
| *** TheJJ has quit IRC | 09:34 | |
| st-gourichon-f | We are trying to pseudo-naively compile some code that uses floats. That fails with two variants of RISC-V GCC toolchain. | 09:35 |
| st-gourichon-f | I'm somehow savvy with such compiler errors. For example, I have written a framework, based on SDCC, to easily code in modern C (SDCC mostly supports C11) targetting the Amstrad CPC (yes, that 8-bit computers). | 09:36 |
| st-gourichon-f | That included writing a ctr0, considering what sections make sense in the target OS, etc. | 09:36 |
| xobs | The softcore only implements RV32I. Others have tried adding the M extension to get multiply and divide. | 09:47 |
| xobs | It would be a challenge to fit G (floats and doubles) in there. Floating point is not exactly efficient. | 09:48 |
| st-gourichon-f | I was wondering if the compiler would provide soft implementations. | 09:49 |
| st-gourichon-f | In the case the CPU would not have "hard" floats. | 09:49 |
| xobs | It will if you compile with -march=rv32i | 09:50 |
| st-gourichon-f | I think the makefile does, let me check. | 09:50 |
| st-gourichon-f | ADD_CFLAGS := -I$(BASE_DIR)/include -D__vexriscv__ -march=rv32i -mabi=ilp32 -DSUPERBOT_FOMU_HACKER=1 | 09:50 |
| st-gourichon-f | From https://github.com/im-tomu/fomu-workshop/blob/master/riscv-blink/Makefile#L35 | 09:51 |
| tpb | Title: fomu-workshop/Makefile at master · im-tomu/fomu-workshop · GitHub (at github.com) | 09:51 |
| xobs | What do you mean when you say it "fails"? | 09:51 |
| st-gourichon-f | (Fetching the details.) I tried two compilers. One precompiled and one I compiled with default settings. | 09:53 |
| st-gourichon-f | Preparing a gist snippet. Any preference regarding gist hosting? | 09:53 |
| xobs | Anything works | 09:55 |
| *** TheJJ has joined #tomu | 10:01 | |
| st-gourichon-f | https://gist.github.com/fidergo-stephane-gourichon/870a9f9ba9ee24c91f1cb80cef7b6929 | 10:01 |
| tpb | Title: Floats on RISC-V · GitHub (at gist.github.com) | 10:01 |
| st-gourichon-f | Notice: none of the toolchain was installed globally. I activate one or the other (not both) by changing the PATH of the shell that runs make. | 10:01 |
| xobs | Probably need to add `-lm` to the compiler. | 10:02 |
| xobs | To the linker flags. | 10:02 |
| st-gourichon-f | Yes, I did this. | 10:04 |
| st-gourichon-f | And confirm -lm appears on ld command line. | 10:05 |
| st-gourichon-f | Updated gist. https://gist.github.com/fidergo-stephane-gourichon/870a9f9ba9ee24c91f1cb80cef7b6929 | 10:06 |
| tpb | Title: Floats on RISC-V · GitHub (at gist.github.com) | 10:06 |
| st-gourichon-f | To be precise: I added -lm before asking here. | 10:07 |
| st-gourichon-f | Mmh, could it be some subtlety like position of "-lm" in command line may cause it to be ignored? | 10:08 |
| xobs | It could be. gcc does seem to care about the order of flags. Sometimes it likes it at the start of the command line, sometimes at the end. | 10:09 |
| xobs | The errors you're seeing in the toolchain you compiled yourself appear to be due to the fact that you don't have rv32i libraries. Did you build a multilib toolchain? | 10:11 |
| st-gourichon-f | I just compiled with default : cd build ; ../src/configure --prefix=.../installtree ; make install | 10:13 |
| st-gourichon-f | Which means the "newlib" option, not the "linux" one. | 10:14 |
| xobs | I suspect that builds an rv64g toolchain only, which won't work with Fomu. Which is why your toolchain is producing errors such as `ld: libc.a(lib_a-memcpy.o): file class ELFCLASS64 incompatible with ELFCLASS32` | 10:15 |
| st-gourichon-f | Makes sense. | 10:15 |
| st-gourichon-f | Ok, will try with --with-arch=rv32i --with-abi=ilp32 | 10:16 |
| st-gourichon-f | Compiling. | 10:17 |
| st-gourichon-f | In the meantime, tried to add -lm at many places and ... it appears to work? | 10:20 |
| xobs | That would make sense based on how I've seen -lm behave. | 10:20 |
| st-gourichon-f | I mean: the precompiled toolchain appears to be able to compile a .elf now. | 10:20 |
| st-gourichon-f | Same for me. Never understood very well a simple rule telling where to put it. | 10:21 |
| *** xkapastel has joined #tomu | 10:21 | |
| st-gourichon-f | Okay now both toolchains can compile. | 10:37 |
| st-gourichon-f | And the code runs on our FOMU. It shows a color, runs a neural network step, shows another color, loops. | 10:39 |
| st-gourichon-f | One thing: when compiling the toolchain with configure --with-arch=rv32i --with-abi=ilp32 the generated toolchain binaries are all named riscv32-unknown-elf-* instead of riscv64-unknown-elf-*. | 10:41 |
| st-gourichon-f | (They are locally 64bit binaries.) | 10:41 |
| st-gourichon-f | So I have to compile with make QUIET= CROSS_COMPILE=riscv32-unknown-elf- | 10:42 |
| st-gourichon-f | So I have to compile our code with make QUIET= CROSS_COMPILE=riscv32-unknown-elf- | 10:42 |
| xobs | Yep! That's why that flag is there. | 10:43 |
| daurnimator | Many thanks to xobs; got zig code running on my FOMU :) | 10:44 |
| st-gourichon-f | Thanks for putting it in! | 10:44 |
| st-gourichon-f | zig? Cool! We considered it. | 10:44 |
| st-gourichon-f | So, more thanks to xobs! | 10:46 |
| st-gourichon-f | daurnimator, can you share something about having zig work on the fomu? | 10:47 |
| *** im-tomu has left #tomu | 10:47 | |
| *** im-tomu has joined #tomu | 10:47 | |
| daurnimator | st-gourichon-f: sure? you mean to try yourself right now; or in terms of writing a howto | 10:48 |
| xobs | Good news all around! | 10:51 |
| daurnimator | to upload a new executable I seem to have to unplug/replug it | 10:52 |
| daurnimator | how can I.... not? | 10:52 |
| st-gourichon-f | So far, daurnimator if it's easy to share something please do. Not trying right now. | 10:53 |
| st-gourichon-f | daurnimator, xobs I also wish I didn't have to unplug/replug. | 10:53 |
| st-gourichon-f | There may be Linux-level tricks. | 10:53 |
| daurnimator | st-gourichon-f: yes its easy to share. however I'm still working on things. if you want to get up and going in the next 10 mins I can tell you now; otherwise wait a week :) | 10:54 |
| *** im-tomu has left #tomu | 10:54 | |
| *** im-tomu has joined #tomu | 10:54 | |
| *** futarisIRCcloud has joined #tomu | 11:16 | |
| *** rohitksingh has quit IRC | 11:31 | |
| xobs | st-gourichon-f: daurnimator one trick you can try, especially if you're very careful and on Linux, is setting NO_USB_RESET. | 11:33 |
| daurnimator | That setting is.... a register? | 11:34 |
| daurnimator | and what happens if I'm not careful? reflash bootloader? resend FPGA bitcode? magic smoke? | 11:35 |
| xobs | If you have `0xb469075a` in the first 256 bytes (64 words), then the following memory address is a configuration value. You can set that to 0x00000020, and the bootloader will simply forget to reset USB before jumping to your program. | 11:35 |
| xobs | You can then use the debugger and bridge. | 11:35 |
| xobs | We used to do that in the workshop, but it turns out that doesn't work so well on some operating systems, so we ended up just getting USB working entirely. | 11:36 |
| xobs | You can see how we did that at https://github.com/im-tomu/fomu-workshop/blob/master/riscv-blink/src/crt0-vexriscv.S#L9-L10 | 11:36 |
| tpb | Title: fomu-workshop/crt0-vexriscv.S at master · im-tomu/fomu-workshop · GitHub (at github.com) | 11:36 |
| *** im-tomu has left #tomu | 11:37 | |
| *** im-tomu has joined #tomu | 11:37 | |
| xobs | By "very careful" I mean "try not to request any USB descriptors from the device". For example, if you do "sudo lsusb", then Linux will try to request several descriptors from the device, which will then time out, and Linux will attempt to reset it. Then you won't have any access to the device until you physically unplug it and plug it back in. | 11:37 |
| xobs | If you set that flag, you should be able to use wishbone-tool to debug and generally access the bridge. For example, you can hit https://rm.fomu.im/reboot.html#reboot-ctrl with `wishbone-tool 0xe0006000 0xac`. | 11:40 |
| tpb | Title: REBOOT Fomu Bootloader documentation (at rm.fomu.im) | 11:40 |
| xobs | st-gourichon-f: if you're using C, you can just steal the USB code for Fomu from https://github.com/im-tomu/fomu-workshop/tree/master/riscv-blink/src/ -- it runs entirely inside the ISR, so once it's set up (by calling usb_connect() during setup, and calling usb_isr() in the interrupt handler when USB_INTERRUPT is set), you don't have to worry about it. | 11:42 |
| tpb | Title: fomu-workshop/riscv-blink/src at master · im-tomu/fomu-workshop · GitHub (at github.com) | 11:42 |
| daurnimator | oh cooll.... segfaults in obj-copy >.< | 11:47 |
| st-gourichon-f | xobs, thanks for the tips. I saw USB_NO_RESET in the crt0, no I know how it could be used. | 11:57 |
| st-gourichon-f | So far I'm doing even more than "stealing the USB code from riscv-blink", I even put my toy code inside the project. ;-) | 11:57 |
| xobs | st-gourichon-f: You also can link in tinyusb, if you want a real USB stack. | 11:58 |
| nrossi | xobs: Managed to get fomu bitstreams and software (including micropython) building in OE https://github.com/nathanrossi/meta-hdl#fomu :), foboot too but I only have the one fomu so haven't tested it | 11:58 |
| tpb | Title: GitHub - nathanrossi/meta-hdl: HDL tools layer for OpenEmbedded (at github.com) | 11:58 |
| st-gourichon-f | nrossi, thanks for sharing. | 11:58 |
| st-gourichon-f | daurnimator, lunch time, the 10 minutes have elapsed, I guess I will wait for a week for zig on FOMU. | 11:59 |
| daurnimator | xobs: I tried the 0xb469075a magic... I get: http://sprunge.us/jGZprq | 12:02 |
| xobs | daurnimator: yes, can you use wishbone-tool? e.g. `wishbone-tool 0` to read the ROM entrypoint? | 12:03 |
| daurnimator | xobs: yes I get 7a80106f | 12:05 |
| xobs | Then it's working. Yay! | 12:05 |
| xobs | You can run `wishbone-tool -s gdb`, then you ought to be able to attach gdb to your Fomu. | 12:05 |
| xobs | To reset it, run `wishbone-tool 0xe0006000 0xac` | 12:06 |
| daurnimator | cool :) | 12:06 |
| daurnimator | works :D | 12:06 |
| daurnimator | What is the license of the content on https://rm.fomu.im/index.html ? | 12:07 |
| tpb | Title: Documentation for Fomu Bootloader Fomu Bootloader documentation (at rm.fomu.im) | 12:07 |
| xobs | I'm not sure it can be licensed, since it's facts. But in general, Foboot is released under Apache-2.0. | 12:08 |
| daurnimator | xobs: e.g. I'm wondering if I can copy your verbage from Messible into the doc-comments for my project (which I would prefer to be MIT or GPL) | 12:10 |
| xobs | The Fomu reference manual is generated automatically from Foboot using lxsocdoc (https://github.com/litex-hub/lxsocdoc) so I'd imagine it would take on the original license, which is Apache-2.0, which is compatible with MIT, and should be able to be converted to GPL. | 12:12 |
| tpb | Title: GitHub - litex-hub/lxsocdoc: Document System-on-Chip for LiteX (at github.com) | 12:12 |
| daurnimator | xobs: oh.... those are generated? what is this .svd thing? should i be writing a zig output for it? | 12:46 |
| daurnimator | also. I'm writing to the MESSIBLE bus: how do I read from it with wishbone-tool? | 12:48 |
| xobs | svd is an ARM standard that specifies all the registers in a microcontroller. You can use it to generate C header files, for example. | 12:48 |
| xobs | Or, for example, we run it through `svd2rust` to generate https://docs.rs/fomu-pac/0.0.3/fomu_pac/ | 12:49 |
| daurnimator | sounds like we need an svd2zig | 12:49 |
| xobs | For the messible... that's still somewhat a work in progress. The idea I had was that you could get output when using gdb, so the gdb server in `wishbone-tool` supports writing to gdb any time it sees an output. | 12:50 |
| xobs | Though now that I've got a better idea of how it works, that seems less likely to be useful. Maybe I should add another server to wishbone-tool that just reads the messible. | 12:50 |
| xobs | We already have a `terminal` server that acts as a client to a loopback UART. mithro requested that for debugging things remotely. But Fomu has no uart currently. | 12:51 |
| daurnimator | st-gourichon-f: https://github.com/im-tomu/fomu-workshop/pull/140 | 12:51 |
| tpb | Title: Add zig example by daurnimator · Pull Request #140 · im-tomu/fomu-workshop · GitHub (at github.com) | 12:51 |
| daurnimator | xobs: I was wanting to do some printf-style debugging: was hoping to just shove bytes out messible and read them with wishbone somehow | 12:52 |
| xobs | That's how it's supposed to work. Try running `wishbone-tool -s gdb`, then in another window `riscv64-unknown-elf-gdb -ex 'target remote localhost:1234' -ex 'continue' zig.elf` and see if that doesn't work for you. | 12:53 |
| daurnimator | I don't have a riscv gdb installed | 12:58 |
| *** futarisIRCcloud has quit IRC | 13:46 | |
| *** whatnick has quit IRC | 13:59 | |
| *** whatnick has joined #tomu | 14:04 | |
| *** im-tomu has left #tomu | 14:10 | |
| *** im-tomu has joined #tomu | 14:10 | |
| * kevans91 spots daurnimator and wonders if it some cool lua project is coming to fomu | 14:14 | |
| daurnimator | kevans91: nope; zig | 14:15 |
| kevans91 | huh, this looks cool | 14:17 |
| *** im-tomu has left #tomu | 14:27 | |
| *** im-tomu has joined #tomu | 14:27 | |
| daurnimator | I almost have the blinking done.... I think | 14:29 |
| daurnimator | trouble is its breathing in red instead of blinking in multi colour :P | 14:29 |
| daurnimator | probably got an offset wrong somewhere | 14:29 |
| * daurnimator needs a working debugger... | 14:31 | |
| xobs | daurnimator: For circuitpython, I disable the breathe effect and treat it just like a 24-bit color value: https://github.com/xobs/circuitpython/blob/fomu/ports/litex/common-hal/neopixel_write/__init__.c | 14:35 |
| tpb | Title: circuitpython/__init__.c at fomu · xobs/circuitpython · GitHub (at github.com) | 14:35 |
| * xobs thinks he should add messible support to wishbone-tool. | 14:35 | |
| xobs | daurnimator: I just added messible support to wishbone-tool. | 14:45 |
| xobs | Can you check v0.6.1 and see if it does what you're looking for? | 14:45 |
| xobs | Run "wishbone-tool -s messible" | 14:47 |
| daurnimator | xobs: o.o you did a release already? | 14:47 |
| xobs | building right now | 14:50 |
| * daurnimator is building for arch | 14:51 | |
| xobs | The binaries should work anywhere that's not libmusl. | 14:52 |
| xobs | Maybe even there. | 14:52 |
| daurnimator | I mean in the arch repositories :P | 14:53 |
| xobs | https://github.com/litex-hub/wishbone-utils/releases/tag/v0.6.1 | 14:54 |
| tpb | Title: Release v0.6.1 · litex-hub/wishbone-utils · GitHub (at github.com) | 14:54 |
| *** im-tomu has left #tomu | 14:56 | |
| *** im-tomu has joined #tomu | 14:56 | |
| daurnimator | xobs: https://www.archlinux.org/packages/community/x86_64/wishbone-utils/ | 14:58 |
| tpb | Title: Arch Linux - wishbone-utils 0.6.1-1 (x86_64) (at www.archlinux.org) | 14:58 |
| *** whatnick has quit IRC | 14:59 | |
| daurnimator | xobs: so I run `wishbone-tool -s messible`.. and then run dfu-util to run my program: and it should all just come out? | 15:01 |
| daurnimator | because...... its not | 15:01 |
| *** whatnick has joined #tomu | 15:02 | |
| *** im-tomu has left #tomu | 15:03 | |
| *** im-tomu has joined #tomu | 15:03 | |
| xobs | It should. Assuming I got wishbone-tool right. I don't have any fomus handy to test it out :-/ | 15:04 |
| xobs | (They're in my bag and not convenient to access.) | 15:04 |
| daurnimator | pushed my WIP to my draft PR https://github.com/im-tomu/fomu-workshop/pull/140 will have a look again later in the week | 15:05 |
| tpb | Title: Add zig example by daurnimator · Pull Request #140 · im-tomu/fomu-workshop · GitHub (at github.com) | 15:05 |
| xobs | You ought to just be able to write to 0xe0008000 in your program. | 15:05 |
| xobs | You can use `wishbone-tool 0xe0008008` to see if there's data. if it's not 0, then you're writing. | 15:05 |
| xobs | You can manually read data by doing `wishbone-tool 0xe0008004`. | 15:05 |
| *** im-tomu has left #tomu | 15:13 | |
| *** im-tomu has joined #tomu | 15:14 | |
| daurnimator | <3 https://github.com/ziglang/zig/pull/4269 | 15:15 |
| tpb | Title: compiler-rt: Port __mulsi3 builtin by LemonBoy · Pull Request #4269 · ziglang/zig · GitHub (at github.com) | 15:15 |
| kevans91 | Hmm | 15:18 |
| kevans91 | I think something about our usb stack disagrees with fomu | 15:18 |
| kevans91 | I guess `./wishbone-tool 0` is fine, `./wishbone-tool 0xe0006000 0xac` gives me a nice fomu-reset-loop, and attempting to flash micropython gives me a serial port that causes a reset (from the looks of it) when I attempt to connect | 15:21 |
| kevans91 | serial port's from our standard cdc class driver, so it might be doing something hinky | 15:22 |
| xobs | That serial port isn't the best. | 15:23 |
| xobs | It's totally a hack. | 15:23 |
| xobs | Try the one from circuitpython. | 15:23 |
| xobs | Or the tinyusb demo code. | 15:23 |
| *** emeb has joined #tomu | 15:26 | |
| kevans91 | Sweet, yeah, circuitpython works well | 15:28 |
| kevans91 | so we can generally interact with it and wishbone-tool seems to work as expected (based on the above primitive testing), assuming continuous reset is expected | 15:29 |
| xobs | It is, and I'm not sure how to "fix" it. | 15:31 |
| xobs | The problem is that you're resetting the device, so it doesn't have time to respond, and wishbone-tool thinks the packet got corrupted so it requques it. | 15:32 |
| xobs | Ad infinitum. | 15:32 |
| kevans91 | Seems like it'd be difficult to get "improved" behavior without breaking some potentially legitimate use-case | 15:33 |
| xobs | Yeah. | 15:34 |
| xobs | Since that just happens to be a special case. Nothing else behaves that way. | 15:34 |
| *** im-tomu has left #tomu | 15:45 | |
| st-gourichon-f | daurnimator, implementing messible in wishbone is just... great! | 15:45 |
| *** im-tomu has joined #tomu | 15:45 | |
| st-gourichon-f | It's whay I needed to log FOMU activity to host. | 15:45 |
| st-gourichon-f | s/whay/what/ | 15:45 |
| *** xkapastel has quit IRC | 15:51 | |
| st-gourichon-f | I have a problem with my FOMU hacker : wishbone-tool fails all operations. | 15:57 |
| st-gourichon-f | I can run the micropython, compile riscv-blink code, push it with dfu-util, it works. | 15:58 |
| st-gourichon-f | But wishbone-tool never works. | 15:58 |
| st-gourichon-f | INFO [wishbone_tool::bridge::usb] opened USB device device 103 on bus 003 | 15:58 |
| st-gourichon-f | INFO [wishbone_tool::bridge::usb] waiting for target device | 15:58 |
| st-gourichon-f | For example with this command "wishbone-tool 0x10000000" from https://workshop.fomu.im/en/latest/migen.html#wishbone-bus | 15:59 |
| st-gourichon-f | Whatever the current state of the device, wishbone never worked for me. | 16:08 |
| *** im-tomu has left #tomu | 16:16 | |
| *** im-tomu has joined #tomu | 16:16 | |
| mithro | st-gourichon-f: What operating system? | 16:28 |
| st-gourichon-f | Ubuntu 19.10. Linux 5.3.0-24-generic. x86_64 | 16:44 |
| *** rohitksingh has joined #tomu | 17:01 | |
| st-gourichon-f | Minor observation on https://github.com/im-tomu/fomu-workshop/blob/master/riscv-blink/ld/regions.ld#L3 | 17:12 |
| tpb | Title: fomu-workshop/regions.ld at master · im-tomu/fomu-workshop · GitHub (at github.com) | 17:12 |
| st-gourichon-f | 1MBit is wrong. It's 1MByte | 17:13 |
| st-gourichon-f | Not a life changer, but I can make a PR. | 17:13 |
| st-gourichon-f | First line: 0x20000 is 1Mbit or 128kbytes. Second line: 0x100000 is not 1Mbit but 8Mbit or 1Mbyte. | 17:14 |
| st-gourichon-f | Would you prefer to remove the comment altogether, for fear it becomes wrong again? Or providing correct values? | 17:16 |
| *** im-tomu has left #tomu | 17:18 | |
| *** im-tomu has joined #tomu | 17:18 | |
| *** rohitksingh has quit IRC | 17:33 | |
| *** im-tomu has left #tomu | 17:33 | |
| *** im-tomu has joined #tomu | 17:34 | |
| *** im-tomu has left #tomu | 18:14 | |
| *** im-tomu has joined #tomu | 18:14 | |
| st-gourichon-f | Could the failure of wishbone-tool be related to the fact that it has an old bootloader? | 18:28 |
| st-gourichon-f | See what dfu-util -l reports: Found DFU: [1209:5bf0] ver=0101, devnum=109, cfg=1, intf=0, path="3-5.1", alt=0, name="Fomu DFU Bootloader v1.7.3-1-g82cb20c", serial="UNKNOWN" | 18:29 |
| st-gourichon-f | This refers to this commit from 2019-06-08: https://github.com/im-tomu/foboot/commit/82cb20c3bccda9316a0a053a5d4c701d3bdb7ed9 | 18:30 |
| tpb | Title: v1.7.3: move location of fourth offset · im-tomu/foboot@82cb20c · GitHub (at github.com) | 18:30 |
| st-gourichon-f | Could the failure of wishbone-tool be related to the failure of updating to foboot 2.0.3? | 18:31 |
| CarlFK | neat - I have Fomu Hacker running DFU Bootloader v1.9.1 | 18:35 |
| CarlFK | want me to test something? | 18:35 |
| st-gourichon-f | Mmh | 18:52 |
| st-gourichon-f | CarlFK, can you test "wishbone-tool 0x10000000" ? | 18:52 |
| CarlFK | INFO [wishbone_tool::bridge::usb] opened USB device device 034 on bus 003 | 19:04 |
| CarlFK | Value at 10000000: 00000005 | 19:04 |
| *** im-tomu has left #tomu | 19:07 | |
| *** im-tomu has joined #tomu | 19:07 | |
| st-gourichon-f | Thanks. My FOMU hacker cannot do that. CarlFK what was the bootloader version when you received it? | 19:08 |
| CarlFK | no clue - they are left over from a hackfest in various states, and I made a jig to program some, but I'm surprised this one had 1.x | 19:28 |
| st-gourichon-f | Thanks CarlFK. | 19:28 |
| *** rohitksingh has joined #tomu | 19:34 | |
| *** rohitksingh has quit IRC | 19:49 | |
| *** rohitksingh has joined #tomu | 19:54 | |
| *** rohitksingh has quit IRC | 20:00 | |
| ovf | it was just brought to my attention that for some time now, there has been a wishbone-utils package in archlinux [community], thus no need for the wishbone-tool aur package which can now be retired. | 20:02 |
| kevans91 | xobs: this is going to be another stupid question, but the github token you use to upload assets- is that a "Deploy Key", or can/do you do this with personal access token or some other mechanism? | 20:03 |
| *** whatnick has quit IRC | 20:04 | |
| *** rohitksingh has joined #tomu | 20:11 | |
| st-gourichon-f | sla | 20:13 |
| *** whatnick_ has joined #tomu | 20:38 | |
| *** rohitksingh has quit IRC | 20:51 | |
| *** rohitksingh has joined #tomu | 22:26 | |
| *** rohitksingh has quit IRC | 22:50 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!