Interrupts don't work

[Disasm]

It was reported that OTG_HS interrupts never fired on F429 disco configured with "HS in FS" mode.

[TomDeRybel]

I'm also having an issue with OTG2_HS not generating any interrupts. As proposed in the #stm32-rs:matrix.org group, here is a bug report with code to reproduce the issue.

The platform is an ST Nucleo-H743ZI2 (chip HAL identifier: stm32h743v). The USB port is on OTG2.

To reproduce the issue, I have made two examples:

• One using blocking (polling) which works
• One using interrupts in RTIC which fails to enumerate (see log output below).
• For the interrupt case, no interrupts seem to ever occur (the RTIC task is never triggered on any of the USB interrupts).

I hope this helps to diagnose the issue. I'm willing to help and run tests in case you do not have this platform.

blinky-nucleo-h743zi2-blocking.zip

blinky-nucleo-h743zi2-interrupt.zip

Output from "tail -f /var/log/syslog" Jun 22 13:40:53 Feynman kernel: [20339.748558] usb 1-5.3.1: new full-speed USB device number 30 using ehci-pci Jun 22 13:41:09 Feynman kernel: [20355.487171] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:41:25 Feynman kernel: [20371.358790] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:41:25 Feynman kernel: [20371.531792] usb 1-5.3.1: new full-speed USB device number 31 using ehci-pci Jun 22 13:41:41 Feynman kernel: [20387.231404] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:41:57 Feynman kernel: [20403.103016] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:41:57 Feynman kernel: [20403.211345] usb 1-5.3-port1: attempt power cycle Jun 22 13:41:57 Feynman kernel: [20403.791146] usb 1-5.3.1: new full-speed USB device number 32 using ehci-pci Jun 22 13:42:09 Feynman kernel: [20415.019728] usb 1-5.3.1: device not accepting address 32, error -110 Jun 22 13:42:09 Feynman kernel: [20415.086755] usb 1-5.3.1: new full-speed USB device number 33 using ehci-pci Jun 22 13:42:20 Feynman kernel: [20426.282532] usb 1-5.3.1: device not accepting address 33, error -110 Jun 22 13:42:20 Feynman kernel: [20426.282695] usb 1-5.3-port1: unable to enumerate USB device Jun 22 13:42:20 Feynman kernel: [20426.691452] usb 1-5.3.1: new full-speed USB device number 34 using ehci-pci Jun 22 13:42:36 Feynman kernel: [20442.013112] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:42:51 Feynman kernel: [20457.884764] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:42:52 Feynman kernel: [20458.058682] usb 1-5.3.1: new full-speed USB device number 35 using ehci-pci Jun 22 13:43:07 Feynman kernel: [20473.758346] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:43:23 Feynman kernel: [20489.629028] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:43:23 Feynman kernel: [20489.737134] usb 1-5.3-port1: attempt power cycle Jun 22 13:43:24 Feynman kernel: [20490.316936] usb 1-5.3.1: new full-speed USB device number 36 using ehci-pci Jun 22 13:43:35 Feynman kernel: [20501.544651] usb 1-5.3.1: device not accepting address 36, error -110 Jun 22 13:43:35 Feynman kernel: [20501.612632] usb 1-5.3.1: new full-speed USB device number 37 using ehci-pci Jun 22 13:43:38 Feynman kernel: [20504.232701] usb 1-5.3.1: device not accepting address 37, error -32 Jun 22 13:43:38 Feynman kernel: [20504.232877] usb 1-5.3-port1: unable to enumerate USB device Jun 22 13:43:47 Feynman kernel: [20513.736402] usb 1-5.3.1: new full-speed USB device number 38 using ehci-pci Jun 22 13:44:03 Feynman kernel: [20529.051003] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:44:18 Feynman kernel: [20544.922585] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:44:19 Feynman kernel: [20545.095611] usb 1-5.3.1: new full-speed USB device number 39 using ehci-pci Jun 22 13:44:34 Feynman kernel: [20560.794299] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:44:50 Feynman kernel: [20576.665904] usb 1-5.3.1: device descriptor read/64, error -110 Jun 22 13:44:50 Feynman kernel: [20576.775137] usb 1-5.3-port1: attempt power cycle Jun 22 13:44:51 Feynman kernel: [20577.354936] usb 1-5.3.1: new full-speed USB device number 40 using ehci-pci Jun 22 13:44:53 Feynman kernel: [20579.710798] usb 1-5.3.1: device not accepting address 40, error -32

[Disasm]

Thank you for the information! Do you know if the same problem happens with OTG1_HS?

[TomDeRybel]

I don't know. The Nucleo board does not have that port connected. But, give me a few hours and I'll see if I can break it out without breaking anything. BRB (got a telco)

[TomDeRybel]

I've connected the OTG1_HS port. Now, these pins also connect to other things on the Nucleo board (such as LD3), so the USB bus is not happy at all.... I'm holding-off on cutting any traces for now.

--> However, the the interesting thing is that the OTG1_HS interrupts do indeed function, while the OTG2_HS interrupts do not.

When I connect OTG2_HS, the debug console does not show any interrupts firing. When I connect OTG1_HS, the debug console shows the interrupts firing.

Here is the updated firmware for your reference:

blinky-nucleo-h743zi2-interrupt_2021-06-23.zip

Here are the various debug outputs:

Output from "tail -f /var/log/syslog" Jun 23 12:04:34 Feynman kernel: [10600.249633] usb 5-1.3.2: new low-speed USB device number 45 using uhci_hcd Jun 23 12:04:34 Feynman kernel: [10600.328572] usb 5-1.3.2: device descriptor read/64, error -71 Jun 23 12:04:34 Feynman kernel: [10600.516591] usb 5-1.3.2: device descriptor read/64, error -71 Jun 23 12:04:35 Feynman kernel: [10600.689599] usb 5-1.3.2: new low-speed USB device number 46 using uhci_hcd Jun 23 12:04:35 Feynman kernel: [10600.768641] usb 5-1.3.2: device descriptor read/64, error -71 Jun 23 12:04:35 Feynman kernel: [10600.956640] usb 5-1.3.2: device descriptor read/64, error -71 Jun 23 12:04:35 Feynman kernel: [10601.062848] usb 5-1.3-port2: attempt power cycle Jun 23 12:04:35 Feynman kernel: [10601.651704] usb 5-1.3.2: new low-speed USB device number 47 using uhci_hcd Jun 23 12:04:36 Feynman kernel: [10602.069584] usb 5-1.3.2: device not accepting address 47, error -71 Jun 23 12:04:36 Feynman kernel: [10602.137580] usb 5-1.3.2: new low-speed USB device number 48 using uhci_hcd Jun 23 12:04:36 Feynman kernel: [10602.557636] usb 5-1.3.2: device not accepting address 48, error -71 Jun 23 12:04:36 Feynman kernel: [10602.558821] usb 5-1.3-port2: unable to enumerate USB device

Output from the debug console: 0 INFO Blinky on Nucleo-H743ZI2 - stm32h7xx-hal + RTIC example └─ blinky_nucleo_h743zi2_interrupt::app::init @ src/main.rs:116 1 INFO Setup PWR... └─ blinky_nucleo_h743zi2_interrupt::app::init @ src/main.rs:119 2 INFO Setup RCC... └─ blinky_nucleo_h743zi2_interrupt::app::init @ src/main.rs:124 3 INFO Setup Timers... └─ blinky_nucleo_h743zi2_interrupt::app::init @ src/main.rs:163 4 INFO Setup GPIO... └─ blinky_nucleo_h743zi2_interrupt::app::init @ src/main.rs:171 5 INFO Setup USB 2... └─ blinky_nucleo_h743zi2_interrupt::app::init @ src/main.rs:184 6 INFO Setup USB 1... └─ blinky_nucleo_h743zi2_interrupt::app::init @ src/main.rs:213 7 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 8 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 9 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 10 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 11 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 12 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 13 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 14 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 15 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 16 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306 17 INFO USB interrupt received. └─ blinky_nucleo_h743zi2_interrupt::app::usb_handler @ src/main.rs:306

[Disasm]

Thank you! I'll try to look into this when I have time. This looks really strange as in your chip these two USB peripherals should be identical.

[TomDeRybel]

Thank you for looking into this! Please let me know if I can do anything else to help. (I suspect that the OTG1_HS identifying as a low-speed device has to do with the parasitic bus loading,)

[hacknus]

As I already described in this issue I noticed that when connecting a stm32f405 to a linux (ubuntu) host, the interrupts dont get triggered but polling in main works fine. When connecting to a windows or macos host, the interrupts get called fine.

[Dirbaio]

can confirm, f429 disco, also interrupts don't work.

I've found it's somehow related to WFE/WFI.

• The stm32f4xxh-hal example works, but fails once you add a WFE here.
• embassy-stm32 (separate implementation, doesn't use this crate)works if you remove the wfe in the executor, fails if you leave it in.

so it seems like some sort of hardware errata... but there's nothing in the errata sheet! :(

[Dirbaio]

found it. you have to set ahb1lpenr.usb_otg_hs_ulpilpen=0. For some reason that defaults to 1, and it kicks in when entering WFE/WFI, sabotaging the phy choice...

https://github.com/embassy-rs/embassy/pull/2805