When trying to review this, I found the use of the fake kernel here really 
confusing. Why are we using it for a normal user land program? Doesn't it 
strike others as odd when you have a userland program use a `kmutex_t`, but 
then also referring to the userland bits like a normal USYNC_THREAD, which is 
definitely a userland mutex thing. It also makes it much less clear which 
mutex_enter/mutex_exit are we using. The real libc one or the renamed one in 
this kind of fake kernel world.

Unrelated to the above confusion, do you have an analysis somewhere about why 
we're marking specific file systems as requiring to be added to or removed from 
the mount in progress tables? It seems like this is probably an important part 
of this change. Can we get some comments as to why the set of file systems in 
question were noted as being safe to not be included in there.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/openzfs/openzfs/pull/536#issuecomment-363607675
------------------------------------------
openzfs-developer
Archives: 
https://openzfs.topicbox.com/groups/developer/discussions/T22334a01fda83bfd-M144b3c6968c5df77ee77feb7
Powered by Topicbox: https://topicbox.com

Reply via email to