05:01 <re_irc> <ignormies> Back again with another weird SPI issue. If I try to ".write" more than 13 bytes (exactly) in a single call/packet, the SCK signal stops working properly.

05:01 <re_irc> First image shows a normal, working signal, with a packet of 13 bytes of "0b10101010" (alternating 1s and 0s) and the second image shows what happens if I add just a single more byte. (top graph is MOSI, bottom is SCK). SCK hiccups every (exactly) 24 periods, preventing the data getting sent.

05:02 <re_irc> This issue goes away if I use "opt-level = 3", but with anything less this is present. Anyone know what's up with this? How can I investigate the true cause of this? GrantM11235 mentioned above that some non-inlined code was fixed here in the next version of "stm32f1xx_hal" but I tried using the git master version of the crate and the issue persisted (unless I crank optimizations all the way up).

05:02 <re_irc> loop {

05:02 <re_irc> ... long message truncated: https://psion.agg.io/_matrix/media/r0/download/psion.agg.io/2653ed2fde614d58131938c138710d8dc8b7bfa2e9bba0a28e30a7a64e0fb97b (15 lines)

05:02 <re_irc> <ignormies> https://psion.agg.io/_matrix/media/r0/download/matrix.org/ztVLZkVQcHAQNjASextbOOKJ/Screenshot%20from%202022-09-08%2021-51-58.png

05:02 <re_irc> <ignormies> https://psion.agg.io/_matrix/media/r0/download/matrix.org/rUPOClaIeozopAOGfVDZQvxP/Screenshot%20from%202022-09-08%2021-53-07.png

05:24 <re_irc> <ignormies> Actually, this occurs with "opt-level = 3" with more than exactly 28 bytes in the packet, so there's more to this going on

05:24 <re_irc> <ignormies> +even

07:44 <re_irc> <almindor> ignormies: I had something similar happen if I misconfigured my MISO line (I didn't even use it in most cases but I did have the pin used in the SPI)

14:26 <re_irc> <ignormies> almindor: Can you elaborate what you mean by "misconfigured"? I'm currently using "NoMiso".

14:44 <re_irc> <newam> ignormies: I had a similar issue in the stm32wlxx-hal. Still need to track it down. I think it is a compiler bug, older compiler versions don't show it.

15:10 <re_irc> <newam> ignormies: do you have a GitHub issue for this? I want to subscribe if there is 🙂

15:10 <re_irc> Now that I know it's not just a me issue I should dig into the codegen and find out what is going on.

15:35 <re_irc> <ignormies> newam: I can definitely open one this evening 👍. Should it just be on the HAL repo?

15:37 <re_irc> <newam> Yeah, that makes the most sense for now I think.

18:14 <re_irc> <almindor> ignormies: I had the actual physical pin connected to the device (I didn't use it to read anything tho), but I didn't configure the PIN on the MCU (I just left it out and use a SCK, MOSI, CS combo). This seems to have caused some issues IIRC, but it was some time ago I might've misremembered