On Fri, Dec 15, 2023 at 02:15:07AM +0000, Roy Luo wrote:
> The KOBJ_CHANGE uevent is sent before gadget unbind is actually
> executed, resulting in inaccurate uevent emitted at incorrect timing
> (the uevent would have USB_UDC_DRIVER variable set while it would
> soon be removed).
> Move the KOBJ_CHANGE uevent to the end of the unbind function so that
> uevent is sent only after the change has been made.
>
> Fixes: 2ccea03a8f7e ("usb: gadget: introduce UDC Class")
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Roy Luo <royluo@xxxxxxxxxx>
> Link: https://lore.kernel.org/r/20231128221756.2591158-1-royluo@xxxxxxxxxx
> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> (cherry picked from commit 73ea73affe8622bdf292de898da869d441da6a9d)
> ---
> drivers/usb/gadget/udc/core.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
Why just a 5.10.y backport? What about 5.15.y as well? You can't
upgrade kernels and have a regression :(
thanks,
greg k-h
