USB Bulk IN Becomes Busy

RE: USB Bulk IN Becomes Busy

Håkon Alseth — Wed, 23 Oct 2019 14:12:09 GMT

Hi,

This thread is quite old, and related to issues in engineering silicon. Please create a new ticket with your inquiry.

Kind regards,

Håkon

RE: USB Bulk IN Becomes Busy

iceseyes — Wed, 23 Oct 2019 13:26:56 GMT

Hi,

I'm experiencing the same problem with the nRF52840 chips using SDK 15.3. In my main loop I call the nrf_drv_usbd_ep_is_busy() but on linux and os x after the first read the ep return ready and I can read new input from host. On Windows, m_ep_dma_waiting never return to 0 and I can't read any other data. my test is this.

I have noticed that on Linux and on Mac after I sent the reply to host two RESET event are dispatched. But on Windows this does not happen. On Host I use libusb to speak with device.

Can you help me?

RE: USB Bulk IN Becomes Busy

Laurence — Wed, 21 Feb 2018 17:43:18 GMT

Fantastic!

It works. Thank you Hakon.

I can also confirm that the original issue is gone too with this rev of the PDK.

Cheers,

Larry

RE: USB Bulk IN Becomes Busy

Håkon Alseth — Wed, 21 Feb 2018 13:57:41 GMT

Hi Larry,

Unfortunately, SDK v14.2 does not include errata workarounds for rev B yet, but will in SDK v15. I suspect that your firmware hangs because of this errata:

http://infocenter.nordicsemi.com/topic/com.nordic.infocenter.nrf52840.EngB.errata/dita/errata/nRF52840/EngineeringB/latest/anomaly_840_187.html?cp=2_0_1_1_1_21

Could you try implementing this in nrf_drv_usbd.h::nrf_usbd_enable() ?

void nrf_usbd_enable(void)
{
    *(volatile uint32_t *)0x4006EC00 = 0x00009375;
    *(volatile uint32_t *)0x4006ED14 = 0x00000003;
    *(volatile uint32_t *)0x4006EC00 = 0x00009375;

    /* Enable the peripheral */
    NRF_USBD->ENABLE = USBD_ENABLE_ENABLE_Enabled<< USBD_ENABLE_ENABLE_Pos;

    /* Waiting for peripheral to enable, this should take a few �s */
    while (0 == (NRF_USBD->EVENTCAUSE & USBD_EVENTCAUSE_READY_Msk))
    {
        /* Empty loop */
    }
    NRF_USBD->EVENTCAUSE &= ~USBD_EVENTCAUSE_READY_Msk;

    *(volatile uint32_t *)0x4006EC00 = 0x00009375;
    *(volatile uint32_t *)0x4006ED14 = 0x00000000;
    *(volatile uint32_t *)0x4006EC00 = 0x00009375;
}

Best regards,

Håkon

RE: USB Bulk IN Becomes Busy

Laurence — Tue, 20 Feb 2018 22:19:43 GMT

Hi,

I have a new PDK. PCA10056 Version 0.11.0.

Unfortunately USB doesn't seem to run at all.

It can't even enable USB.

In the debugger I can trace it to nrf_usbd_eventcause_clear(), where it seems to hang.

Called from nrf_drv_usbd_enable()

Any ideas?

Do I need a specific version of firmware to be loaded?

Thanks,

Larry

RE: USB Bulk IN Becomes Busy

Håkon Alseth — Wed, 14 Feb 2018 15:19:06 GMT

Hi Larry,

RSM = regional sales manager, and PM=personal message.

Cheers,

Håkon

RE: USB Bulk IN Becomes Busy

Laurence — Wed, 14 Feb 2018 13:46:15 GMT

Hi Hakon,

I'm not clear on the acronyms. What do you mean by PM and RSM?

Thanks,

Larry

RE: USB Bulk IN Becomes Busy

Håkon Alseth — Wed, 14 Feb 2018 10:17:03 GMT

You can contact your regional sales manager to get more information on how to obtain the newest revision. If you do not already have the details of your RSM, send me a PM with your location and I can provide you with this.

Cheers,

Håkon

RE: USB Bulk IN Becomes Busy

Laurence — Tue, 13 Feb 2018 21:01:47 GMT

How can I get a new revision? Is this available now in a development kit?

RE: USB Bulk IN Becomes Busy

Håkon Alseth — Tue, 13 Feb 2018 09:10:43 GMT

Hi,

It is very likely that you are experiencing the side-effects of erratum #104.

As shown in this list, there are quite a few erratas on the USBD for engineering A of the nRF52840.

I'd recommend to get a newer revision of the nRF52840 silicon to avoid the erratum #104.

Best regards,

Håkon

RE: USB Bulk IN Becomes Busy

Laurence — Mon, 12 Feb 2018 14:33:14 GMT

Is it possible that I am experiencing Errata 104 (EPDATA event is not always generated)?