#rust-embedded on 2024-09-15 — irc logs at libera.irclog.whitequark.org

2022-02-07 19:20 ChanServ changed the topic of #rust-embedded to: Welcome to the Rust Embedded IRC channel! Bridged to #rust-embedded:matrix.org and logged at https://libera.irclog.whitequark.org/rust-embedded, code of conduct at https://www.rust-lang.org/conduct.html

00:00 <i509vcb[m]> (The C110x really is going to be a pain to work with)

00:01 <i509vcb[m]> Although the dev board for it was quite cheap ($7) so why not lol

00:03 <i509vcb[m]> (The L2228 has 32k ram and 128k flash, far more spacious)

00:05 <i509vcb[m]> [probe-rs/src/flashing/flash_algorithm.rs:318:13] ram_for_data = 432

00:05 <i509vcb[m]> [probe-rs/src/flashing/flash_algorithm.rs:318:13] stack_size = 512

00:06 <i509vcb[m]> hmm stripping debug symbols got me under

00:07 <i509vcb[m]> ah right that seems to break defmt

00:19 <JamesMunns[m]> debug symbols aren't flashed to the chip

00:19 <JamesMunns[m]> they only live in the elf

00:24 <jduck[m]> Still release mode optimizations help a ton with code size

00:24 <i509vcb[m]> I am currently using:... (full message at <https://catircservices.org/_matrix/media/v3/download/catircservices.org/oPZxWsEXVuyJVUHPCicCyigc>)

00:24 <i509vcb[m]> Setting the buffer size any smaller doesn't change anything

00:24 <jduck[m]> Debug mode maintains thin abstractions that bloat code

00:24 <i509vcb[m]> * I am currently using:... (full message at <https://catircservices.org/_matrix/media/v3/download/catircservices.org/xThaqHFCXzGuMeaZbyzAaBGx>)

00:25 <jduck[m]> When I had this problem, I added strip option to a toml iirc

00:25 <JamesMunns[m]> there's a difference between debug mode and debug symbols

00:25 <jduck[m]> How are you flashing/programming the device?

00:26 <JamesMunns[m]> they are unrelated

00:26 <JamesMunns[m]> and debug symbols are not flashed to the device.

00:26 <JamesMunns[m]> and do not affect code size or performance at all.

00:26 <i509vcb[m]> jduck[m]: probe-rs, built from source because the target isn't upstream yet

00:26 <i509vcb[m]> although there seems to be an overflow bug in probe-rs as well that I found

00:27 <jduck[m]> Def report

00:27 M9names[m] has joined #rust-embedded

00:27 <M9names[m]> You don't need to build from source to use a new target, you can use the chip-definition-file flag (in case you suspect a bug in probe-rs git)

00:28 <jduck[m]> What is your final code size? Is it just too big for this environment?

00:30 <i509vcb[m]> yeah chip-definition-path gives a different error

00:30 <i509vcb[m]> Failed to configure a stack for the flash algorithm.

00:31 <JamesMunns[m]> probe rs flash algorithms work by loading code into ram on the device, and using that to program the flash

00:32 <JamesMunns[m]> it's possible the algorithm isn't correctly configured, or is too large for the 1k of ram you have

00:32 <JamesMunns[m]> or, overflows the stack while running

00:32 <jduck[m]> For 1k of RAM you will need XIP or something. Just my guess

00:33 <jduck[m]> 1k is very small

00:36 <jduck[m]> Do you have external RAM available?

00:37 <i509vcb[m]> no external ram, this is literally a sub $1 microcontroller lol

00:37 <i509vcb[m]> that's unless the flash algorithm from TI is broken?

00:38 <JamesMunns[m]> Not sure! I've not had the pleasure of debugging flash algos before :D

00:39 <i509vcb[m]> This is what target-gen popped out: https://gist.github.com/i509VCB/fe4ea129bb56a42304eea8f6214e0fae

00:40 <jduck[m]> It's possible TI never tried these with 1k of RAM

00:41 <i509vcb[m]> Well the C110x the only ram size is 1K

00:43 <i509vcb[m]> (there is a 8k version vs the 16k version I have for the sadist in us as well)

00:43 <i509vcb[m]> * a 8k flash version vs

00:43 <jduck[m]> I think that's a very challenging mark to meet for Rust. You could try to reduce your own code size to basically nothing by removing everything and just putting one inline assembly instruction. Maybe you can manage a 1k stack manually lol

00:44 <jduck[m]> But still probe-rs might not work

00:45 <jduck[m]> Rust prefers to put objects on the stack and use RAII, so it naturally has bigger stack frames.

00:52 <i509vcb[m]> I assume there is a way to make rtt work under openocd just to see if the problem might be in probe-rs?

00:52 <i509vcb[m]> There is a patch set to make these work in openocd

01:11 <i509vcb[m]> From the git build of probe-rs I just hardcoded FLASH_ALGO_STACK_SIZE and it works lmao

01:12 <i509vcb[m]> The stuff generated from the pack file doesn't seem to contain a stack_size

01:12 <i509vcb[m]> * The stuff generated from the pack file doesn't seem to contain a stack_size variable and therefore uses the default

01:13 <i509vcb[m]> * From the git build of probe-rs I just hardcoded FLASH_ALGO_STACK_SIZE to 256 and it works lmao

01:13 <i509vcb[m]> Does cargo bloat give a binary size that's accurate enough?

01:14 * i509vcb[m] sent a code block: https://catircservices.org/_matrix/media/v3/download/catircservices.org/zbXzKJHNMglGJnpOGYZucAFV

01:16 <i509vcb[m]> iirc there was a tool which lets you analyze memory usage better? I think cargo bloat just does some stuff with symbols?

01:20 <JamesMunns[m]> cargo bloat, cargo size, or cargo nm, will all show you the static memory usage accurately

01:21 <JamesMunns[m]> cargo-call-stack is used for determinining runtime stack size

01:21 <JamesMunns[m]> or rather, it does static analysis to determine the worst-case runtime stack usage

01:38 <i509vcb[m]> So 144 bytes?

01:38 * i509vcb[m] uploaded an image: (71KiB) < https://catircservices.org/_matrix/media/v3/download/matrix.org/aPcvrwIEUwxxjxfyuqDTEQMT/image.png >

01:38 <i509vcb[m]> Although some flash does get used

01:38 <i509vcb[m]> * 144 bytes of ram?

01:50 sugoi has joined #rust-embedded

01:55 sugoi has quit [Ping timeout: 246 seconds]

01:59 starblue has quit [Ping timeout: 252 seconds]

02:01 starblue has joined #rust-embedded

02:04 <M9names[m]> <i509vcb[m]> "From the git build of probe-rs I..." <- you can set it in the yaml these days, FYI.

02:04 <M9names[m]> https://github.com/9names/lpc-iap/blob/e42a116e75e9dafc2619a199149634ecc12b3f12/algos/lpc81x/template.yaml#L36

02:05 <i509vcb[m]> M9names[m]: > <@9names:matrix.org> you can set it in the yaml these days, FYI.

02:05 <i509vcb[m]> > https://github.com/9names/lpc-iap/blob/e42a116e75e9dafc2619a199149634ecc12b3f12/algos/lpc81x/template.yaml#L36

02:05 <i509vcb[m]> yeah the yaml needs it, but still a bug that it underflowed

02:06 <M9names[m]> but yeah it's really rough building flash algos with so little space. i basically had to evict all error handling to get mine to work (can't remember if this was a 1K or 2K RAM part)

02:07 sugoi has joined #rust-embedded

02:07 <M9names[m]> https://github.com/9names/lpc-iap/blob/e42a116e75e9dafc2619a199149634ecc12b3f12/algos/lpc81x/src/main.rs#L28

02:07 <M9names[m]> haha guess i commented it

02:08 <M9names[m]> > // 1KB RAM is not enough room for error handling. this algo ignores all failures.

02:17 sugoi has quit [Ping timeout: 260 seconds]

02:30 sugoi has joined #rust-embedded

02:35 sugoi has quit [Ping timeout: 260 seconds]

02:48 sugoi has joined #rust-embedded

04:09 sugoi has quit [Ping timeout: 252 seconds]

04:10 cr1901 has quit [Read error: Connection reset by peer]

04:21 cr1901 has joined #rust-embedded

04:21 sugoi has joined #rust-embedded

04:28 sugoi has quit [Ping timeout: 252 seconds]

04:40 sugoi has joined #rust-embedded

04:45 sugoi has quit [Ping timeout: 252 seconds]

04:52 sugoi has joined #rust-embedded

04:56 sugoi has quit [Ping timeout: 252 seconds]

05:08 sugoi has joined #rust-embedded

05:12 sugoi has quit [Ping timeout: 246 seconds]

05:25 sugoi has joined #rust-embedded

05:29 sugoi has quit [Ping timeout: 252 seconds]

05:42 sugoi has joined #rust-embedded

05:47 sugoi has quit [Ping timeout: 264 seconds]

05:53 sugoi has joined #rust-embedded

05:57 sugoi has quit [Ping timeout: 252 seconds]

06:09 sugoi has joined #rust-embedded

06:14 sugoi has quit [Ping timeout: 265 seconds]

06:26 sugoi has joined #rust-embedded

06:31 sugoi has quit [Ping timeout: 276 seconds]

06:44 sugoi has joined #rust-embedded

06:49 sugoi has quit [Ping timeout: 260 seconds]

07:04 juliand[m] has quit [Quit: Idle timeout reached: 172800s]

07:12 sugoi has joined #rust-embedded

07:18 sugoi has quit [Ping timeout: 252 seconds]

07:31 sugoi has joined #rust-embedded

07:36 sugoi has quit [Ping timeout: 260 seconds]

07:48 sugoi has joined #rust-embedded

07:52 sugoi has quit [Ping timeout: 248 seconds]

07:57 sugoi has joined #rust-embedded

08:01 sugoi has quit [Ping timeout: 265 seconds]

08:13 sugoi has joined #rust-embedded

08:19 sugoi has quit [Ping timeout: 246 seconds]

08:30 sugoi has joined #rust-embedded

08:35 sugoi has quit [Ping timeout: 252 seconds]

08:49 sugoi has joined #rust-embedded

08:53 sugoi has quit [Ping timeout: 260 seconds]

09:13 sugoi has joined #rust-embedded

09:14 starblue has quit [Ping timeout: 260 seconds]

09:17 sugoi has quit [Ping timeout: 260 seconds]

09:31 M0xada3[m] has quit [Quit: Idle timeout reached: 172800s]

09:32 sugoi has joined #rust-embedded

09:37 sugoi has quit [Ping timeout: 265 seconds]

09:51 sugoi has joined #rust-embedded

09:55 sugoi has quit [Ping timeout: 246 seconds]

10:09 sugoi has joined #rust-embedded

10:16 sugoi has quit [Ping timeout: 252 seconds]

10:28 sugoi has joined #rust-embedded

10:34 sugoi has quit [Ping timeout: 276 seconds]

10:47 sugoi has joined #rust-embedded

10:52 sugoi has quit [Ping timeout: 260 seconds]

11:07 sugoi has joined #rust-embedded

11:11 sugoi has quit [Ping timeout: 252 seconds]

11:25 sugoi has joined #rust-embedded

11:30 sugoi has quit [Ping timeout: 260 seconds]

11:35 starblue has joined #rust-embedded

11:43 sugoi has joined #rust-embedded

11:48 sugoi has quit [Ping timeout: 265 seconds]

12:02 sugoi has joined #rust-embedded

12:08 sugoi has quit [Ping timeout: 252 seconds]

12:21 sugoi has joined #rust-embedded

12:25 sugoi has quit [Ping timeout: 255 seconds]

12:30 starblue has quit [Ping timeout: 260 seconds]

12:39 sugoi has joined #rust-embedded

12:44 sugoi has quit [Ping timeout: 260 seconds]

12:51 sugoi has joined #rust-embedded

12:55 sugoi has quit [Ping timeout: 252 seconds]

12:58 RobertJrdens[m] has quit [Quit: Idle timeout reached: 172800s]

13:06 sugoi has joined #rust-embedded

13:10 ello has quit [Ping timeout: 252 seconds]

13:10 sugoi has quit [Ping timeout: 244 seconds]

13:16 ello has joined #rust-embedded

13:26 sugoi has joined #rust-embedded

13:26 justacec[m] has quit [Quit: Idle timeout reached: 172800s]

13:30 sugoi has quit [Ping timeout: 246 seconds]

13:30 MurrayToddWillia has quit [Quit: Idle timeout reached: 172800s]

13:44 sugoi has joined #rust-embedded

13:48 sugoi has quit [Ping timeout: 252 seconds]

14:01 sugoi has joined #rust-embedded

14:06 sugoi has quit [Ping timeout: 264 seconds]

14:20 sugoi has joined #rust-embedded

14:25 sugoi has quit [Ping timeout: 260 seconds]

14:38 sugoi has joined #rust-embedded

14:42 sugoi has quit [Ping timeout: 265 seconds]

14:50 sugoi has joined #rust-embedded

14:56 sugoi has quit [Ping timeout: 252 seconds]

15:09 sugoi has joined #rust-embedded

15:13 sugoi has quit [Ping timeout: 246 seconds]

15:26 sugoi has joined #rust-embedded

15:31 sugoi has quit [Ping timeout: 252 seconds]

15:45 sugoi has joined #rust-embedded

15:49 sugoi has quit [Ping timeout: 260 seconds]

16:07 sugoi has joined #rust-embedded

16:11 sugoi has quit [Ping timeout: 260 seconds]

16:23 sugoi has joined #rust-embedded

16:30 sugoi has quit [Ping timeout: 248 seconds]

16:42 sugoi has joined #rust-embedded

16:48 sugoi has quit [Ping timeout: 260 seconds]

16:52 sugoi has joined #rust-embedded

17:02 sugoi has quit [Ping timeout: 276 seconds]

17:06 <thejpster[m]1> The Neotron BMC runs RTIC in 32K Flash and 4K RAM and does quite a bit.

17:06 <thejpster[m]1> I don’t think I’d enjoy programming a 2K part.

17:53 sugoi has joined #rust-embedded

17:57 sugoi has quit [Ping timeout: 252 seconds]

18:12 sugoi has joined #rust-embedded

19:20 starblue has joined #rust-embedded

20:02 sugoi has quit [Ping timeout: 260 seconds]

20:11 sugoi has joined #rust-embedded

20:16 sugoi has quit [Ping timeout: 252 seconds]

20:51 aliceryhl[m] has joined #rust-embedded

20:51 <aliceryhl[m]> <ivmarkov> "> <@aliceryhl:matrix.org> So..." <- Std has a bunch of other dependencies (e.g. hashbrown), whereas core/alloc don't have extra deps.

20:51 <aliceryhl[m]> aliceryhl[m]: (Sorry for the late reply, I am unused to matrix.)

20:53 sugoi has joined #rust-embedded

21:16 AtleoS has quit [Ping timeout: 252 seconds]

21:17 AtleoS has joined #rust-embedded

21:25 sugoi has quit [Ping timeout: 265 seconds]

21:55 sugoi has joined #rust-embedded

22:47 sugoi has quit [Remote host closed the connection]

22:47 sugoi has joined #rust-embedded

23:53 sugoi has quit [Ping timeout: 265 seconds]

23:58 sugoi has joined #rust-embedded