Re: [syzbot] [usb?] INFO: task hung in i2c_tiny_usb_disconnect

[weipeng]

On Tue, 13 Jan 2026 10:49:16, Oliver Neukum wrote:

> For once what prevents the driver from being unloaded
> with a work item queued after disconnect() has run?
> How do you make sure that the disconnected flag is ever read?

Hi,

Did you say that we can't make sure the usb_read() and usb_write()
is prevented after the dev->disconnected is set to true?
It is indeed a problem. I will send a new patch soon.

Thanks for review!