On Mo, 2019-03-25 at 10:21 +0000, Kento.A.Kobayashi@xxxxxxxx wrote: > Hi, > > > > > And the return value(-ENODEV) will be returned to error handler. > > > > uas_eh_device_reset_handler->usb_reset_device -> usb_reset_and_verify_device (return -ENODEV) Then I wrote that commit message that we think even if we ignore "ENODEV" in post reset to avoid hang issue but the error will also be reported to error handler. > > > > #If ignore an error and the error will not be reported then it is not good. > > > > > > Well, what do we do now? Are you saying that the state model SCSI is using is wrong? > > No, it is not wrong. > > But I think we may want to prevend hang by sleep peocess with continuing to take a lock at blk_execute_rq() from >usb_unbind_and_rebind_marded_interfaces(). > > Could you please you can apply this patch for mainline or not? > If not, could you please tell me why you can't allow to apply this patch? Sorry, I thought this was clear. Your patch is making the assumption that the reset is triggered by the SCSI layer. You cannot make that assumption, as there is an ioctl for resetting a USB device. In case we are getting an error during the reset (our endpoints vanish), the device driver must report that to the USB layer, so the driver will always be disconnected. We cannot drop errors. Regards Oliver
