USB MSC example in interrupt mode

[ebray]

I'm experimenting with the USB MSC example (SDK 15.3) on my nRF52840 DK. It seems to work well in the polling mode as delivered. However, when I try to run in interrupt mode it no longer functions properly. The first time I connect the USB cable, Windows 10 recognizes the 3 drives as expected. However, after disconnecting the USB cable, all subsequent attempts to reconnect it fail to bring up the drives. Windows doesn't even see the USB device. During this process the board is constantly powered through the debug USB port. The only way to get the MSC drives working again is to restart the micro.

I assumed the only steps needed to get this working would be to comment would be the ones suggested in this post:

https://devzone.nordicsemi.com/f/nordic-q-a/34933/usb-cdc-acm-example-is-not-working-in-interrupt-mode

All you should have to do to make event handling interrupt driven in SDK v15 is to set APP_USBD_CONFIG_EVENT_QUEUE_ENABLE to 0 in sdk_config.h, and comment out the call to app_usbd_event_queue_process() in main.c

Any assistance getting this resolved would be greatly appreciated!

[Edvin]

Hello,

I see that the thread that you link to died. I have tested, and I see the same behavior as you, but I am not certain that the interrupt handler handles all the events correctly, as it does when you use the event queue. I have asked our SDK team if they can take a look at what you would need to do to use this mode. I will let you know as soon as I know something.

Best regards,

Edvin

[ebray]

Odd. The link still works fine for me. If you search for the text below in the forum you can find the post that way.

usb-cdc-acm-example-is-not-working-in-interrupt-mode

Thank you for testing and verifying the behavior on the MSC example. Since it works properly for the CDC example I'm assuming it's either a bug or a configuration issue as I wouldn't think there should be any difference in the USBD user events generated.

[ebray]

It appears that the real issue is that when the USB is disconnected the fatfs_init call never returns. It seems to get hung up waiting for the QSPI flash to become not busy but that never occurs.

[Edvin]

ebray said:

Odd. The link still works fine for me. If you search for the text below in the forum you can find the post that way.

 Sorry. I meant that they didn't find the cause of the issue. 

 

ebray said:

It seems to get hung up waiting for the QPSI to become not busy but that never occurs.

 Have you tested the example without the QSPI part? Or you can try to turn off and on the power register on the QSPI like described in the workaround here.

[ebray]

No, because the QSPI flash is the only "drive" that is configured to work with FatFS. Also, it works fine in polling mode.