commit 8f54511aefe1455161a6c4406ed8c770139f61e3 Author: Seiichi Ikarashi Date: Sat Jan 31 14:22:55 2026 -0500 statd: fix a typo in a debug message Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit d0ed742a6394ced5c3622a45cdfe4a050aabdaf1 Author: Seiichi Ikarashi Date: Sat Jan 31 14:19:25 2026 -0500 nfsd: fix a typo in man page Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit af4887d32bf7f3fd22ee3d6289bcd32553f52c01 Author: Seiichi Ikarashi Date: Sat Jan 31 14:08:59 2026 -0500 mountd: fix a typo in man page Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit 5e692c5a549c398e0a6ee97451cc496ae42a202d Author: Seiichi Ikarashi Date: Sat Jan 31 14:05:35 2026 -0500 nfs.man: fix a typo in man page Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit b99b6b2971b0665ba65b4bf5fddc60ac801d0558 Author: Seiichi Ikarashi Date: Sat Jan 31 14:01:19 2026 -0500 gssd: fix typos in man page Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit 730cb6c278d35572a25fc05e455a952fab31d510 Author: Seiichi Ikarashi Date: Sat Jan 31 13:39:33 2026 -0500 exports: fix a typo in man page Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit 06045ffb3e8de9d3ebbb96a47b5aae89f97eb8cd Author: Seiichi Ikarashi Date: Sat Jan 31 13:34:39 2026 -0500 rpcctl: fix a typo in man page Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit c4f2523439178a8eb225b3e9cd12f933fc57bce5 Author: Seiichi Ikarashi Date: Sat Jan 31 13:32:52 2026 -0500 nfsiostat: fix typos in man page Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit ab0d78a638208e86c8d09202fffefb04adec021b Author: Seiichi Ikarashi Date: Sat Jan 31 13:31:12 2026 -0500 mountstats: fix a typo in man page Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit a641896b9cde4010ae937b8e918d98a65232d9e0 Author: Seiichi Ikarashi Date: Sat Jan 31 13:27:45 2026 -0500 nfs.conf: fix a typo in man page Signed-off-by: Seiichi Ikarashi Signed-off-by: Steve Dickson commit 00e2e62b89984f2341d2f8ffc8d1eb6df33ce718 Author: Jeff Layton Date: Sat Jan 31 13:04:52 2026 -0500 nfsdctl: add support for min-threads parameter This patch adds proper support to nfsdctl to manage nfsd's new dynamic threading when NFSD_A_SERVER_MIN_THREADS is defined in the header: If the user sets min-threads, and has a new enough kernel, the traditional threads parameters will represent the max number of threads. The min-threads value represents the minimum number of threads to run in each pool. The kernel will start the minimum number of threads at startup time, and the thread count will dynamically fluctuate between the min and max based on load. Allow the "autostart" subcommand to find the "min-threads" parameter in nfs.conf and set it in the netlink call appropriately. If it's not set, then assume that it's 0, which will disable dynamic threading. For the "threads" subcommand, add a new command-line option. If it's not set, then don't send it in the netlink command at all. That allows the kernel's existing setting to remain as-is, unless explicitly changed. Also, update the documentation with info about min-threads. Signed-off-by: Jeff Layton Signed-off-by: Steve Dickson commit 4c2754421f4c67dc75a6e053baf2305d8ad99e82 Author: Dusty Mabe Date: Sat Jan 31 13:00:17 2026 -0500 systemd: drop Wants=network-online.target for rpc-statd-notify Pulling in network-online.target by default is generally discouraged. From [1]: > It is strongly recommended not to make use of this target too liberally: > for example network server software should generally not pull this in > (since server software generally is happy to accept local connections > even before any routable network interface is up). Its primary purpose > is network client software that cannot operate without network. On systems where nfs-client.target is enabled by default via presets (this is the case on Fedora today) then nfs-client.target will pull in rpc-statd-notify.service which will pull in network-online.target. This is the case even if the user hasn't configured any NFS mounts. The man page for sm-notify already mentions: > Notifications are retried if sending fails, the remote does not respond, > the remote's NSM service is not registered, or if there is a DNS failure > which prevents the remote's mon_name from being resolved to an address. So I would think it would be able to cope with a short period of time if it were to start before Networking were fully online. Making this change means if you are in an offline scenario (one example is an install environment) then you won't have to wait for network-online.target to timeout before the system comes up fully. We originally discussed this problem over in [2], [3]. [1] https://systemd.io/NETWORK_ONLINE/ [2] https://github.com/coreos/fedora-coreos-config/pull/3530 [3] https://bugzilla.redhat.com/show_bug.cgi?id=2383585 Signed-off-by: Dusty Mabe Signed-off-by: Steve Dickson commit 03b9c540152b0630427056c5ffe7ea5f4a65cd3d Author: Scott Mayhew Date: Mon Jan 26 08:28:41 2026 -0500 nfsiostat: normalize the mountpoints passed in from the command line If the mountpoint passed in from the command line contains a trailing '/' character, then nfsiostat winds up printing statistics for all mounts instead of printing statistics for the specific mount that was requested. Signed-off-by: Scott Mayhew Signed-off-by: Steve Dickson commit 59e85671e59f7ccbf66dcf72ce7462bf4932cc78 Author: Scott Mayhew Date: Sun Jan 18 14:56:15 2026 -0500 Rename CONFIG_NFSV41 to CONFIG_BLKMAPD and disable by default pNFS block layout is deprecated in favor of pNFS SCSI layout, which doesn't require the use of blkmapd. Since CONFIG_NFSV41 (enabled by default, but can be disabled via --disable-nfsv41) is only used to enable blkmapd, let's rename it to CONFIG_BLKMAPD and change the default to disabled. Distributions that wish to continue to include blkmapd can do so by adding --enable-blkmapd to their configure script invocation. Reviewed-by: Christoph Hellwig Tested-by: Salvatore Bonaccorso Signed-off-by: Scott Mayhew Signed-off-by: Steve Dickson commit 3b7de50fe8f59c64762a015693b705db5d0f15e7 Author: Olga Kornievskaia Date: Sun Jan 18 14:30:45 2026 -0500 nfsdctl: ignore ipv6 listener creation error If ipv6 is not supported by the kernel and during the nfsd start we are trying to create an ipv6 listener, ignore the error (this restores how nfsd handled this failure). Fixes: 8b02f0d5590e ("nfsdctl: cleanup listeners if some failed") Signed-off-by: Olga Kornievskaia Signed-off-by: Steve Dickson commit 0e71be58cdead21b7bc0285fa6afbf1d0eca3049 Author: Ross Burton Date: Tue Nov 18 11:21:30 2025 -0500 locktest: use correct build flags This makefile uses CFLAGS_FOR_BUILD etc but since 2020[1] hasn't used CC_FOR_BUILD as CC. This means in cross-compile environments this uninstalled binary is built for the host machine using build machine flags, which can result in incorrect paths being passed. As this binary hasn't been built with CC_FOR_BUILD for five years, we can assume that this isn't actually used and just remove the _FOR_BUILD flags entirely. Original patch by Khem Raj . [1] nfs-utils 1fee8caa ("locktest: Makefile.am: remove host compiler costraint") Acked-by: Jeff Layton Signed-off-by: Ross Burton Signed-off-by: Steve Dickson commit 077b70fee5d84d07d35b28514e4e1b054b2544bb Author: Peter Schwenke Date: Tue Nov 18 10:45:14 2025 -0500 sm-notify: Do not drop privileges if running as non-root user If sm-notify is not running root and is running as the owner of the state directory, it is not necessary to drop privileges. The patch checks if the current user is the owner of the state directory. In the case of a ha-callout, the ha-callout will, more than likely, be running as rpcuser - or similar. That means the sm-notify program can not be called from the ha-callout. nsm_drop_privileges() is also called by statd. However, that will fail for a number of reasons such as not being able to create the PID file etc. So, this patch should not cause any issues for statd. statd will still need to be started as root. Without this patch capabalities such as cap_setpcap, cap_setgid, cap_net_bind_service=ep need to be set. Clearly, that is an ugly hack and best avoided. Signed-off-by: Peter Schwenke Signed-off-by: Peter Schwenke Signed-off-by: Steve Dickson commit 8600bbb7727df779ba1104c9f0c574b06be116a3 Author: Olga Kornievskaia Date: Tue Nov 18 10:23:27 2025 -0500 gssd: protect kerberos ticket cache access gssd_get_single_krb5_cred() is a function that's will (for when needed) send a TGT request to the KDC and then store it in a credential cache. If multiple threads (eg., parallel mounts) are making an upcall at the same time then getting creds and storing creds need to be serialized due to do kerberos API not being concurrency safe. Fixes: https://issues.redhat.com/browse/RHEL-103627 Signed-off-by: Olga Kornievskaia Signed-off-by: Steve Dickson