[ENBD] 2.4.32

[Peter T. Breuer]

Anyone want to confirm or deny that 2.4.32 is now stable and correct?

I believe we fixed the bug tracked to somewhere in do_srv_write() (by
regressing the routine), and I moved all clientside locking (per area on
device) through the same lock routines as used by the server, which are
now microsecond resolution instead of second resolution, leading to
faster reaction if one ever has to pause waiting on a userland lock,
while still allowing breakout from the lock wait in who-knows-what
circumstances ...

Ho hum.

Peter