On Thu, May 13, 2021 at 12:21:52PM +0800, Chris Chiu wrote: > On Wed, May 12, 2021 at 11:04 PM Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote: > > > > For testing purposes, set status to 0 and jump to a new goto label in > > the "else" branch. In other words, treat it as if the suspend really > > had worked and go back to the successful pathway. Try this out and see > > if it fixes the problem. > > > > If it does then the reset-resume isn't needed. If it doesn't, post your > > patch again, and mention in the patch description that testing shows the > > reset-resume really is necessary. > > > > Alan Stern > > Thanks for the suggestion. I revised the patch to create a new goto > label in the "else" branch and clear the status to zero if the port is > really suspended. It fixed the problem in my 100 time suspend/resume > test. I will send v3 patch w/ the modification. Great! This is a much better solution. Alan Stern
