We should support restarting the interrupt transfer after it's been stopped in the past.
This fixes the breakage of the polling transfer after the "insert card => SCardConnect => SCardDisconnect => remove card" scenario, which led the application to actively poll the reader every 400 ms instead of a 10-minute transfer, which in turn led to excessively big number of inflight chrome.usb transfers and various breakages in production.
We should support restarting the interrupt transfer after it's been stopped in the past.
This fixes the breakage of the polling transfer after the "insert card => SCardConnect => SCardDisconnect => remove card" scenario, which led the application to actively poll the reader every 400 ms instead of a 10-minute transfer, which in turn led to excessively big number of inflight chrome.usb transfers and various breakages in production.