Handle reset and a few other fixes #14
Conversation
For USB_REQ_SET_CONFIGURATION/USB_REQ_SET_INTERFACE transfers, we should complete the transfer via usb_raw_ep0_read only after we spawned the endpoint threads.
When auto_detach_kernel_driver is set to 1, the kernel will attach a kernel driver when an interface is released. We don't want that, set auto_detach_kernel_driver to 0.
Add handling of the USB_RAW_EVENT_RESET Raw Gadget event. This allows to continue proxying the device if the host decided to reset it, which can happen during normal device operation. On a reset event, stop all endpoint threads and reset the proxied device to effectively restart proxying.
|
Hi @xairy , Thanks for the PR! and sorry that I am about to go on a business trip abroad, so I can only review after I am back |
| @@ -122,7 +122,7 @@ int connect_device(int vendor_id, int product_id) { | |||
| return result; | |||
| } | |||
|
|
|||
| result = libusb_set_auto_detach_kernel_driver(dev_handle, 1); | |||
| result = libusb_set_auto_detach_kernel_driver(dev_handle, 0); | |||
There was a problem hiding this comment.
May I know the reason to set the value to 0? does it fix any issues? Thanks!
There was a problem hiding this comment.
Yes, it prevents the kernel from attaching a driver to the interface when the proxy calls release_interface during a device reset or an altsetting change. I think this was the original intention of the libusb_set_auto_detach_kernel_driver call, but it used a wrong argument by mistake.
|
Let's bring the conversation to here if you don't mind, I noticed that I won't get any notifictation if you have new comments there thanks! |
|
Looks like there is no further discussion, so I will merge this PR soon, and also I will open a new issue to make sure we will have a fix for the issue discussed here |
Sorry for not responding. I was meaning to get back to this, but I've been busy with other things. Thank you for merging! |
No worry, I know people get very busy from time to time, even I am not available all the time maintaining this repo, and thanks for the contributions! we can continue the discussion in #20 if you have any thought. thanks again! |
Add handling of the
USB_RAW_EVENT_RESETRaw Gadget event and fix a few other issues.See the commit descriptions for details.
Fixes #9.