nfsd: don't allow nfsd threads to be signalled.
authorNeilBrown <neilb@suse.de>
Tue, 18 Jul 2023 06:38:08 +0000 (16:38 +1000)
committerChuck Lever <chuck.lever@oracle.com>
Tue, 29 Aug 2023 21:45:22 +0000 (17:45 -0400)
commit3903902401451b1cd9d797a8c79769eb26ac7fe5
tree2baa4e07ddff8293bb1650fd2415bb7eb6f09365
parent8db14cad28ae8ec3fde0fef18e969782bca204d1
nfsd: don't allow nfsd threads to be signalled.

The original implementation of nfsd used signals to stop threads during
shutdown.
In Linux 2.3.46pre5 nfsd gained the ability to shutdown threads
internally it if was asked to run "0" threads.  After this user-space
transitioned to using "rpc.nfsd 0" to stop nfsd and sending signals to
threads was no longer an important part of the API.

In commit 3ebdbe5203a8 ("SUNRPC: discard svo_setup and rename
svc_set_num_threads_sync()") (v5.17-rc1~75^2~41) we finally removed the
use of signals for stopping threads, using kthread_stop() instead.

This patch makes the "obvious" next step and removes the ability to
signal nfsd threads - or any svc threads.  nfsd stops allowing signals
and we don't check for their delivery any more.

This will allow for some simplification in later patches.

A change worth noting is in nfsd4_ssc_setup_dul().  There was previously
a signal_pending() check which would only succeed when the thread was
being shut down.  It should really have tested kthread_should_stop() as
well.  Now it just does the latter, not the former.

Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
fs/nfs/callback.c
fs/nfsd/nfs4proc.c
fs/nfsd/nfssvc.c
net/sunrpc/svc_xprt.c