On 14 Jul 2023, at 9:57, Olga Kornievskaia wrote: > > I guess I don't understand what's the problem with the current > approach which is simple. No problem - I'm just bad at code review. I was trying to get my head around why the o_arg->fh couldn't be used, and then wondering why you were hitting that warning. Then I worried that the opendata was already freed at the point the getattr was trying to use it again, and because its a codepath I can't exercise easily my normal approach of just throwing in debug statements didn't work - but I think the opendata is still there because _nfs4_do_open has it pinned. Ben
