A system with very low autofs timeouts for many mounts of v4.0 exports, while attempting to mount with the non-specific vers=4 option would run out of memory every few days. Each mount was doing the v4.2, v4.1, v4.0 negotiation and constantly setting up and tearing down the backchannel service for positive minorversions. We found a probable cause: wake_up_process() isn't guarunteed to execute the threadfn; kthread_stop() can race in and prevent it. Benjamin Coddington (1): SUNRPC: Fix a server shutdown leak net/sunrpc/svc.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) -- 2.31.1
