On Sun, Jul 26, 2015 at 3:53 AM, Helg <[email protected]> wrote: > > The problem is that fuse doesn't respond properly to this situation and the > kernel then become unstable.
I am quite sure that this is not why your kernel become unstable. What is this instability? can you give me a way reproduce it ? I think you probably have an issue elsewhere (few vnop like vnop_rename() still need some love). What you see is probably a side effect of something going wrong in a vnop... > Also, there are some operations, like the one you mention, that just take a > long time and the user has to be patient. The best solution is to block > indefinitely but handle interrupts gracefully so that if the user loses > patience then the operation can be interrupted and handled gracefully. I > almost had this working at one stage but it wasn't that easy to do and I > opted to focus on other stability improvements. > > It's good to have you back Syl. I submitted this and other patches a while > back but there wasn't anyone to do much with them so I lost motivation. > OK, I will have a look a it. Cheers, -- Sylvestre Gallon
