Re: [PATCH] usb: gadget: eem: fix command packet transfer issue

['Greg Kroah-Hartman']

On Thu, Jun 10, 2021 at 02:10:40PM +0800, linyyuan@xxxxxxxxxxxxxx wrote:
> From: Linyu Yuan <linyyuan@xxxxxxxxxxxxxx>
> 
> there is following warning,
> [<ffffff8008905a94>] dwc3_gadget_ep_queue+0x1b4/0x1c8
> [<ffffff800895ec9c>] usb_ep_queue+0x3c/0x120
> [<ffffff80089677a0>] eem_unwrap+0x180/0x330
> [<ffffff80089634f8>] rx_complete+0x70/0x230
> [<ffffff800895edbc>] usb_gadget_giveback_request+0x3c/0xe8
> [<ffffff8008901e7c>] dwc3_gadget_giveback+0xb4/0x190
> [<ffffff8008905254>] dwc3_endpoint_transfer_complete+0x32c/0x410
> [<ffffff80089060fc>] dwc3_bh_work+0x654/0x12e8
> [<ffffff80080c63fc>] process_one_work+0x1d4/0x4a8
> [<ffffff80080c6720>] worker_thread+0x50/0x4a8
> [<ffffff80080cc8e8>] kthread+0xe8/0x100
> [<ffffff8008083980>] ret_from_fork+0x10/0x50
> request ffffffc0716bf200 belongs to 'ep0out'
> 
> when gadget receive a eem command packet from host, it need to response,
> but queue usb request to wrong endpoint.

What is the full warning here?  THe above traceback is a bit odd and
does not show what is happening.

> fix it by queue usb request to eem IN endpoint and allow host read it.

I do not understand how this matches up with your kernel patch, when you
resend this, can you please expand on this and make it more obvious what
you are doing here?

thanks,

greg k-h