On Wed, Feb 13, 2019 at 12:11 PM David Ahern <dsah...@gmail.com> wrote: > > On 2/13/19 12:57 PM, Peter Oskolkov wrote: > > Thanks, David! I was not able to reproduce the leak, but based on your > > suggestion and similar code elsewhere I made a change in v11 to explicitly > > release a dst with error. > > ok. Did you run the test with a debug kernel - checking refcount, use > after free, etc?
In my tests I was always getting ERR_PTR for unroutable packets, not a full rt/dst with an error flag set. But I checked several similar route lookups, and they all release bad dsts, so I did not feel it was worth it to investigate further.
