Description of problem:
After the `nfsrahead` introduced via Bug 1946283 the automatic testcase gets to be failed sometimes as the result read_ahead_kb value doesn't match what the config file set.

But it's weird that when I check that by hand it turns to work as expected. Finally I find this problem can be reliablely reproduced in arch s390x. Please help check it.

Version-Release number of selected component (if applicable):
nfs-utils-2.3.3-57.el8

How reproducible:
always in arch s390x

Steps to Reproduce:
1. Configure nfsrahead in /etc/nfs.conf
2. Mount and check read_ahead_kb value
3.

Actual results:
[root@ibm-z-509 ~]# cat /etc/exports
/export_test *(rw,no_root_squash)
[root@ibm-z-509 ~]# cat /etc/nfs.conf
[nfsrahead]
nfs=15000
nfs4=16000 <<<<<<<<<<<<<<
default=128
[root@ibm-z-509 ~]# systemctl restart nfs-server
[root@ibm-z-509 ~]# mount localhost:/export_test/ /mnt_test/ -o vers=3
[root@ibm-z-509 ~]# cat /proc/self/mountinfo | grep test
334 95 0:45 / /mnt_test rw,relatime shared:178 - nfs localhost:/export_test/ rw,vers=3,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp6,timeo=600,retrans=2,sec=sys,mountaddr=::1,mountvers=3,mountport=20048,mountproto=udp6,local_lock=none,addr=::1
[root@ibm-z-509 ~]# cat /sys/class/bdi/0\:45/read_ahead_kb
15000
[root@ibm-z-509 ~]# umount /mnt_test/
[root@ibm-z-509 ~]# mount localhost:/export_test/ /mnt_test/ -o vers=4
[root@ibm-z-509 ~]# cat /proc/self/mountinfo | grep test
336 95 0:46 / /mnt_test rw,relatime shared:178 - nfs4 localhost:/export_test rw,vers=4.2,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp6,timeo=600,retrans=2,sec=sys,clientaddr=::1,local_lock=none,addr=::1
[root@ibm-z-509 ~]# cat /sys/class/bdi/0\:46/read_ahead_kb
128 <<<<<<<<<<<<<
[root@ibm-z-509 ~]# umount /mnt_test/
[root@ibm-z-509 ~]# rpm -q kernel nfs-utils
kernel-4.18.0-425.el8.s390x
nfs-utils-2.3.3-57.el8.s390x
[root@ibm-z-509 ~]#

Expected results:

Additional info: