* [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils
@ 2024-07-22 17:01 Jeff Layton
2024-07-22 17:01 ` [PATCH nfs-utils v6 1/3] nfsdctl: add the nfsdctl utility " Jeff Layton
` (5 more replies)
0 siblings, 6 replies; 16+ messages in thread
From: Jeff Layton @ 2024-07-22 17:01 UTC (permalink / raw)
To: Steve Dickson
Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey,
Chuck Lever, Lorenzo Bianconi, Jeff Layton
Hi Steve,
Here's an squashed version of the nfsdctl patches, that represents
the latest changes. Let me know if you run into any other problems,
and thanks for helping to test this!
Signed-off-by: Jeff Layton <jlayton@kernel.org>
---
Changes in v6:
- make the default number of threads 16 in autostart
- doc updates
Changes in v5:
- add support for pool-mode setting
- fix up the handling of nfsd_netlink.h in autoconf
- Link to v4: https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org
Changes in v4:
- add ability to specify an array of pool thread counts in nfs.conf
- Link to v3: https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org
Changes in v3:
- split nfsdctl.h so we can include the UAPI header as-is
- squash the patches together that added Lorenzo's version and convert
it to the new interface
- adapt to latest version of netlink interface changes
+ have THREADS_SET/GET report an array of thread counts (one per pool)
+ pass scope in as a string to THREADS_SET instead of using unshare() trick
Changes in v2:
- Adapt to latest kernel netlink interface changes (in particular, send
the leastime and gracetime when they are set in the config).
- More help text for different subcommands
- New nfsdctl(8) manpage
- Patch to make systemd preferentially use nfsdctl instead of rpc.nfsd
- Link to v1: https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org
---
Jeff Layton (3):
nfsdctl: add the nfsdctl utility to nfs-utils
nfsdctl: asciidoc source for the manpage
systemd: use nfsdctl to start and stop the nfs server
configure.ac | 19 +
systemd/nfs-server.service | 4 +-
utils/Makefile.am | 4 +
utils/nfsdctl/Makefile.am | 13 +
utils/nfsdctl/nfsd_netlink.h | 96 +++
utils/nfsdctl/nfsdctl.8 | 304 ++++++++
utils/nfsdctl/nfsdctl.adoc | 158 +++++
utils/nfsdctl/nfsdctl.c | 1570 ++++++++++++++++++++++++++++++++++++++++++
utils/nfsdctl/nfsdctl.h | 93 +++
9 files changed, 2259 insertions(+), 2 deletions(-)
---
base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be
change-id: 20240412-nfsdctl-fa8bd8430cfd
Best regards,
--
Jeff Layton <jlayton@kernel.org>
^ permalink raw reply [flat|nested] 16+ messages in thread* [PATCH nfs-utils v6 1/3] nfsdctl: add the nfsdctl utility to nfs-utils 2024-07-22 17:01 [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Jeff Layton @ 2024-07-22 17:01 ` Jeff Layton 2024-07-22 17:01 ` [PATCH nfs-utils v6 2/3] nfsdctl: asciidoc source for the manpage Jeff Layton ` (4 subsequent siblings) 5 siblings, 0 replies; 16+ messages in thread From: Jeff Layton @ 2024-07-22 17:01 UTC (permalink / raw) To: Steve Dickson Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi, Jeff Layton This tool is based on Lorenzo's original nfsdctl tool [1]. His original tool used getopt_long to indicate the command, but that's somewhat limiting. This converts it to a subcommand-based interface, where each subcommand can take its own options, in the spirit of commands like nmcli or virsh. There are currently 6 different subcommands: pool-mode get/set current pool mode setting listener get/set listener info version get/set supported NFS versions threads get/set nfsd thread settings status get current RPC processing info autostart start server with settings from /etc/nfs.conf Each can take different options, and we can expand this interface later with more commands as necessary. This is based on Lorenzo's original userland tool: https://github.com/LorenzoBianconi/nfsdctl Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org> Signed-off-by: Jeff Layton <jlayton@kernel.org> --- configure.ac | 19 + utils/Makefile.am | 4 + utils/nfsdctl/Makefile.am | 13 + utils/nfsdctl/nfsd_netlink.h | 96 +++ utils/nfsdctl/nfsdctl.8 | 304 ++++++++ utils/nfsdctl/nfsdctl.c | 1570 ++++++++++++++++++++++++++++++++++++++++++ utils/nfsdctl/nfsdctl.h | 93 +++ 7 files changed, 2099 insertions(+) diff --git a/configure.ac b/configure.ac index 58d1728c5bc6..1aebc3a767a7 100644 --- a/configure.ac +++ b/configure.ac @@ -252,6 +252,24 @@ AC_ARG_ENABLE(nfsdcltrack, enable_nfsdcltrack=$enableval, enable_nfsdcltrack="yes") +AC_ARG_ENABLE(nfsdctl, + [AS_HELP_STRING([--disable-nfsdctl],[disable nfsdctl program for controlling nfsd@<:@default=no@:>@])], + enable_nfsdctl=$enableval, + enable_nfsdctl="yes") + AM_CONDITIONAL(CONFIG_NFSDCTL, [test "$enable_nfsdctl" = "yes" ]) + if test "$enable_nfsdctl" = yes; then + PKG_CHECK_MODULES(LIBNL3, libnl-3.0 >= 3.1) + PKG_CHECK_MODULES(LIBNLGENL3, libnl-genl-3.0 >= 3.1) + PKG_CHECK_MODULES(LIBREADLINE, readline) + AC_CHECK_HEADERS(linux/nfsd_netlink.h) + + # ensure we have the pool-mode commands + AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[#include <linux/nfsd_netlink.h>]], + [[int foo = NFSD_CMD_POOL_MODE_GET;]])], + [AC_DEFINE([USE_SYSTEM_NFSD_NETLINK_H], 1, + ["Use system's linux/nfsd_netlink.h"])]) + fi + AC_ARG_ENABLE(nfsv4server, [AS_HELP_STRING([--enable-nfsv4server],[enable support for NFSv4 only server @<:@default=no@:>@])], enable_nfsv4server=$enableval, @@ -739,6 +757,7 @@ AC_CONFIG_FILES([ utils/mountd/Makefile utils/exportd/Makefile utils/nfsd/Makefile + utils/nfsdctl/Makefile utils/nfsref/Makefile utils/nfsstat/Makefile utils/nfsidmap/Makefile diff --git a/utils/Makefile.am b/utils/Makefile.am index ab584190f17e..e5cb81e796bd 100644 --- a/utils/Makefile.am +++ b/utils/Makefile.am @@ -35,6 +35,10 @@ if CONFIG_JUNCTION OPTDIRS += nfsref endif +if CONFIG_NFSDCTL +OPTDIRS += nfsdctl +endif + SUBDIRS = \ exportfs \ mountd \ diff --git a/utils/nfsdctl/Makefile.am b/utils/nfsdctl/Makefile.am new file mode 100644 index 000000000000..89c7ecd6f30b --- /dev/null +++ b/utils/nfsdctl/Makefile.am @@ -0,0 +1,13 @@ +## Process this file with automake to produce Makefile.in + +man8_MANS = nfsdctl.8 +EXTRA_DIST = $(man8_MANS) + +sbin_PROGRAMS = nfsdctl + +noinst_HEADERS = nfsdctl.h +nfsdctl_SOURCES = nfsdctl.c +nfsdctl_CFLAGS = $(LIBNL3_CFLAGS) $(LIBNLGENL3_CFLAGS) $(LIBREADLINE_CFLAGS) +nfsdctl_LDADD = ../../support/nfs/libnfs.la $(LIBNL3_LIBS) $(LIBNLGENL3_LIBS) $(LIBREADLINE_LIBS) + +MAINTAINERCLEANFILES = Makefile.in diff --git a/utils/nfsdctl/nfsd_netlink.h b/utils/nfsdctl/nfsd_netlink.h new file mode 100644 index 000000000000..887cbd12b695 --- /dev/null +++ b/utils/nfsdctl/nfsd_netlink.h @@ -0,0 +1,96 @@ +/* SPDX-License-Identifier: ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) */ +/* Do not edit directly, auto-generated from: */ +/* Documentation/netlink/specs/nfsd.yaml */ +/* YNL-GEN uapi header */ + +#ifndef _UAPI_LINUX_NFSD_NETLINK_H +#define _UAPI_LINUX_NFSD_NETLINK_H + +#define NFSD_FAMILY_NAME "nfsd" +#define NFSD_FAMILY_VERSION 1 + +enum { + NFSD_A_RPC_STATUS_XID = 1, + NFSD_A_RPC_STATUS_FLAGS, + NFSD_A_RPC_STATUS_PROG, + NFSD_A_RPC_STATUS_VERSION, + NFSD_A_RPC_STATUS_PROC, + NFSD_A_RPC_STATUS_SERVICE_TIME, + NFSD_A_RPC_STATUS_PAD, + NFSD_A_RPC_STATUS_SADDR4, + NFSD_A_RPC_STATUS_DADDR4, + NFSD_A_RPC_STATUS_SADDR6, + NFSD_A_RPC_STATUS_DADDR6, + NFSD_A_RPC_STATUS_SPORT, + NFSD_A_RPC_STATUS_DPORT, + NFSD_A_RPC_STATUS_COMPOUND_OPS, + + __NFSD_A_RPC_STATUS_MAX, + NFSD_A_RPC_STATUS_MAX = (__NFSD_A_RPC_STATUS_MAX - 1) +}; + +enum { + NFSD_A_SERVER_THREADS = 1, + NFSD_A_SERVER_GRACETIME, + NFSD_A_SERVER_LEASETIME, + NFSD_A_SERVER_SCOPE, + + __NFSD_A_SERVER_MAX, + NFSD_A_SERVER_MAX = (__NFSD_A_SERVER_MAX - 1) +}; + +enum { + NFSD_A_VERSION_MAJOR = 1, + NFSD_A_VERSION_MINOR, + NFSD_A_VERSION_ENABLED, + + __NFSD_A_VERSION_MAX, + NFSD_A_VERSION_MAX = (__NFSD_A_VERSION_MAX - 1) +}; + +enum { + NFSD_A_SERVER_PROTO_VERSION = 1, + + __NFSD_A_SERVER_PROTO_MAX, + NFSD_A_SERVER_PROTO_MAX = (__NFSD_A_SERVER_PROTO_MAX - 1) +}; + +enum { + NFSD_A_SOCK_ADDR = 1, + NFSD_A_SOCK_TRANSPORT_NAME, + + __NFSD_A_SOCK_MAX, + NFSD_A_SOCK_MAX = (__NFSD_A_SOCK_MAX - 1) +}; + +enum { + NFSD_A_SERVER_SOCK_ADDR = 1, + + __NFSD_A_SERVER_SOCK_MAX, + NFSD_A_SERVER_SOCK_MAX = (__NFSD_A_SERVER_SOCK_MAX - 1) +}; + +enum { + NFSD_A_POOL_MODE_MODE = 1, + NFSD_A_POOL_MODE_NPOOLS, + + __NFSD_A_POOL_MODE_MAX, + NFSD_A_POOL_MODE_MAX = (__NFSD_A_POOL_MODE_MAX - 1) +}; + +enum { + NFSD_CMD_RPC_STATUS_GET = 1, + NFSD_CMD_THREADS_SET, + NFSD_CMD_THREADS_GET, + NFSD_CMD_VERSION_SET, + NFSD_CMD_VERSION_GET, + NFSD_CMD_LISTENER_SET, + NFSD_CMD_LISTENER_GET, + NFSD_CMD_POOL_MODE_SET, + NFSD_CMD_POOL_MODE_GET, + + __NFSD_CMD_MAX, + NFSD_CMD_MAX = (__NFSD_CMD_MAX - 1) +}; + +#endif /* _UAPI_LINUX_NFSD_NETLINK_H */ diff --git a/utils/nfsdctl/nfsdctl.8 b/utils/nfsdctl/nfsdctl.8 new file mode 100644 index 000000000000..dae1863d3238 --- /dev/null +++ b/utils/nfsdctl/nfsdctl.8 @@ -0,0 +1,304 @@ +'\" t +.\" Title: nfsdctl +.\" Author: Jeff Layton +.\" Generator: Asciidoctor 2.0.20 +.\" Date: 2024-07-22 +.\" Manual: \ \& +.\" Source: \ \& +.\" Language: English +.\" +.TH "NFSDCTL" "8" "2024-07-22" "\ \&" "\ \&" +.ie \n(.g .ds Aq \(aq +.el .ds Aq ' +.ss \n[.ss] 0 +.nh +.ad l +.de URL +\fI\\$2\fP <\\$1>\\$3 +.. +.als MTO URL +.if \n[.g] \{\ +. mso www.tmac +. am URL +. ad l +. . +. am MTO +. ad l +. . +. LINKSTYLE blue R < > +.\} +.SH "NAME" +nfsdctl \- control program for the Linux kernel NFS server +.SH "SYNOPSIS" +.sp +\fBnfsdctl\fP [ \fIOPTION\fP ] COMMAND ... +.SH "DESCRIPTION" +.sp +nfsdctl is a control and query program for the in\-kernel NFS server. There are several +subcommands (documented below) that allow the admin to configure or query different +aspects of the NFS server. +.sp +To get information about different subcommand usage, pass the subcommand the +\-\-help parameter. For example: +.sp +.if n .RS 4 +.nf +.fam C +nfsdctl listener \-\-help +.fam +.fi +.if n .RE +.SH "OPTIONS" +.sp +\fB\-d, \-\-debug\fP +.RS 4 +Enable debug logging +.RE +.sp +\fB\-h, \-\-help\fP +.RS 4 +Print program help text +.RE +.sp +\fB\-V, \-\-version\fP +.RS 4 +Print program version +.RE +.SH "SUBCOMMANDS" +.sp +Each subcommand can also accept its own set of options and arguments. The +\-\-help option is standard for all subcommands: +.sp +\fBautostart\fP +.RS 4 +Start the server using the settings in the [nfsd] section of /etc/nfs.conf. +This subcommand takes no arguments. Note that if a "threads=" value is not set in +nfs.conf, 16 server threads will be brought online. +.RE +.sp +\fBlistener\fP +.RS 4 +Get/set the listening sockets for the server. Run this without arguments to +get a list of the sockets on which the server is currently listening. To add +or remove sockets, pass it whitespace\-separated strings in the format: +.sp +.if n .RS 4 +.nf +.fam C +{ +|\- }{ protocol }:{ address }:{ port } +.fam +.fi +.if n .RE +.sp +.if n .RS 4 +.nf +.fam C +The fields are: +.fam +.fi +.if n .RE +.sp +.if n .RS 4 +.nf +.fam C ++ to add a listener, \- to remove one +protocol: protocol name (e.g. tcp, udp, rdma) +address: hostname or address +port: port number or service name +.fam +.fi +.if n .RE +.sp +.if n .RS 4 +.nf +.fam C +All fields are required, except for the address. If address is an empty string, +then the listeners will be opened for INADDR_ANY and IN6ADDR_ANY_INIT for ipv6 +(if enabled). The address can be either a hostname or an IP address. IPv4 +addresses must be in dotted\-quad form. IPv6 addresses should be in standard +colon separated form, and must be enclosed in square brackets. +.fam +.fi +.if n .RE +.RE +.sp +\fBstatus\fP +.RS 4 +Get information about RPCs currently executing in the server. This subcommand +takes no arguments. +.RE +.sp +\fBthreads\fP +.RS 4 +Get/set the number of running nfsd threads in each pool. Pass a list of +integers to change the currently active number of threads. Passing it a +value of 0 will shut down the NFS server. Run this without arguments to +get the current number of running threads in each pool. +.RE +.sp +\fBversion\fP +.RS 4 +Get/set the enabled NFS versions for the server. Run without arguments to +get a list of supported versions and whether they are currently enabled or +disabled. To enable or disable a version, pass it a string in the format: +.sp +.if n .RS 4 +.nf +.fam C +{ +|\- }{ MAJOR }{.{ MINOR }} +.fam +.fi +.if n .RE +.sp +.if n .RS 4 +.nf +.fam C +The fields are: +.fam +.fi +.if n .RE +.sp +.if n .RS 4 +.nf +.fam C ++ to enable a version, \- to disable +MAJOR: the major version integer value +MINOR: the minor version integet value +.fam +.fi +.if n .RE +.sp +.if n .RS 4 +.nf +.fam C +The minorversion field is optional. If not given, it will disable or enable +all minorversions for that major version. +.fam +.fi +.if n .RE +.RE +.sp +\fBpool\-mode\fP +.RS 4 +Get/set the host\(cqs pool mode. This will cause the server to start threads +that are pinned to either the CPU or the NUMA node. This can only be set +when there are no nfsd threads running. +.sp +.if n .RS 4 +.nf +.fam C +The available options are: + global: single large pool + percpu: pool per CPU + pernode: pool per NUMA node + auto: choose a mode based on host configuration +.fam +.fi +.if n .RE +.RE +.SH "EXAMPLES" +.sp +Start the server with the settings in nfs.conf: +.sp +.if n .RS 4 +.nf +.fam C +nfsdctl autostart +.fam +.fi +.if n .RE +.sp +Get a list of current listening sockets: +.sp +.if n .RS 4 +.nf +.fam C +nfsdctl listener +.fam +.fi +.if n .RE +.sp +Show the supported and enabled NFS versions: +.sp +.if n .RS 4 +.nf +.fam C +nfsdctl version +.fam +.fi +.if n .RE +.sp +Add TCP listener on all addresses (both v4 and v6), port 2049: +.sp +.if n .RS 4 +.nf +.fam C +nfsdctl listener +tcp::2049 +.fam +.fi +.if n .RE +.sp +Add RDMA listener on 1.2.3.4 port 20049: +.sp +.if n .RS 4 +.nf +.fam C +nfsdctl listener +rdma:1.2.3.4:20049 +.fam +.fi +.if n .RE +.sp +Add same listener on IPv6 address f00::ba4 port 20050: +.sp +.if n .RS 4 +.nf +.fam C +nfsdctl listener +rdma:[f00::ba4]:20050 +.fam +.fi +.if n .RE +.sp +Enable NFS version 3, disable v4.0: +.sp +.if n .RS 4 +.nf +.fam C +nfsdctl version +3 \-4.0 +.fam +.fi +.if n .RE +.sp +Change the number of running threads in first pool to 256: +.sp +.if n .RS 4 +.nf +.fam C +nfsdctl threads 256 +.fam +.fi +.if n .RE +.sp +Set the pool\-mode to "pernode": +.sp +.if n .RS 4 +.nf +.fam C +nfsctl pool\-mode pernode +.fam +.fi +.if n .RE +.SH "NOTES" +.sp +nfsdctl is intended to supersede rpc.nfsd(8), which controls the nfs server +using the files under /proc/fs/nfsd. nfsdctl instead uses a netlink(7) +interface to achieve the same goals. +.sp +Most of the commands that query the NFS server can be run as an unprivileged +user, but configuring the server typically requires an account with elevated +privileges. +.SH "SEE ALSO" +.sp +nfs.conf(5), rpc.nfsd(8), rpc.mountd(8), exports(5), exportfs(8), nfs.conf(5), rpc.rquotad(8), nfsstat(8), netconfig(5) +.SH "AUTHOR" +.sp +Jeff Layton \ No newline at end of file diff --git a/utils/nfsdctl/nfsdctl.c b/utils/nfsdctl/nfsdctl.c new file mode 100644 index 000000000000..f7c276320085 --- /dev/null +++ b/utils/nfsdctl/nfsdctl.c @@ -0,0 +1,1570 @@ +#define _GNU_SOURCE 1 + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#include <linux/module.h> +#include <linux/version.h> +#include <netlink/genl/genl.h> +#include <errno.h> +#include <stdio.h> +#include <stdlib.h> +#include <stdarg.h> +#include <time.h> +#include <stdbool.h> +#include <arpa/inet.h> +#include <unistd.h> +#include <getopt.h> +#include <string.h> +#include <sched.h> +#include <sys/queue.h> + +#include <netlink/genl/family.h> +#include <netlink/genl/ctrl.h> +#include <netlink/msg.h> +#include <netlink/attr.h> +#include <linux/netlink.h> + +#include <readline/readline.h> +#include <readline/history.h> + +#ifdef USE_SYSTEM_NFSD_NETLINK_H +#include <linux/nfsd_netlink.h> +#else +#include "nfsd_netlink.h" +#endif + +#include "nfsdctl.h" +#include "conffile.h" +#include "xlog.h" + +/* compile note: + * gcc -I/usr/include/libnl3/ -o <prog-name> <prog-name>.c -lnl-3 -lnl-genl-3 + */ + +static int debug_level; + +struct nfs_version { + uint8_t major; + uint8_t minor; + bool enabled; +}; + +/* + * The NFS server should only have around 5 versions or so, so we don't bother + * with memory allocation here, and just use a global array. + */ +#define MAX_NFS_VERSIONS 16 + +struct nfs_version nfsd_versions[MAX_NFS_VERSIONS]; + +/* + * All of the existing netids are short strings (3-4 chars), but let's allow + * for up to 16. + */ +#define MAX_CLASS_NAME_LEN 16 + +struct server_socket { + struct sockaddr_storage ss; + char name[MAX_CLASS_NAME_LEN]; + bool active; +}; + +#define MAX_NFSD_SOCKETS 256 + +int nfsd_socket_count; +struct server_socket nfsd_sockets[MAX_NFSD_SOCKETS]; + +const char *taskname; + +static const struct option help_only_options[] = { + { "help", no_argument, NULL, 'h' }, + { }, +}; + +#define NFSD4_OPS_MAX_LEN sizeof(nfsd4_ops) / sizeof(nfsd4_ops[0]) +static const char *nfsd4_ops[] = { + [OP_ACCESS] = "OP_ACCESS", + [OP_CLOSE] = "OP_CLOSE", + [OP_COMMIT] = "OP_COMMIT", + [OP_CREATE] = "OP_CREATE", + [OP_DELEGRETURN] = "OP_DELEGRETURN", + [OP_GETATTR] = "OP_GETATTR", + [OP_GETFH] = "OP_GETFH", + [OP_LINK] = "OP_LINK", + [OP_LOCK] = "OP_LOCK", + [OP_LOCKT] = "OP_LOCKT", + [OP_LOCKU] = "OP_LOCKU", + [OP_LOOKUP] = "OP_LOOKUP", + [OP_LOOKUPP] = "OP_LOOKUPP", + [OP_NVERIFY] = "OP_NVERIFY", + [OP_OPEN] = "OP_OPEN", + [OP_OPEN_CONFIRM] = "OP_OPEN_CONFIRM", + [OP_OPEN_DOWNGRADE] = "OP_OPEN_DOWNGRADE", + [OP_PUTFH] = "OP_PUTFH", + [OP_PUTPUBFH] = "OP_PUTPUBFH", + [OP_PUTROOTFH] = "OP_PUTROOTFH", + [OP_READ] = "OP_READ", + [OP_READDIR] = "OP_READDIR", + [OP_READLINK] = "OP_READLINK", + [OP_REMOVE] = "OP_REMOVE", + [OP_RENAME] = "OP_RENAME", + [OP_RENEW] = "OP_RENEW", + [OP_RESTOREFH] = "OP_RESTOREFH", + [OP_SAVEFH] = "OP_SAVEFH", + [OP_SECINFO] = "OP_SECINFO", + [OP_SETATTR] = "OP_SETATTR", + [OP_SETCLIENTID] = "OP_SETCLIENTID", + [OP_SETCLIENTID_CONFIRM] = "OP_SETCLIENTID_CONFIRM", + [OP_VERIFY] = "OP_VERIFY", + [OP_WRITE] = "OP_WRITE", + [OP_RELEASE_LOCKOWNER] = "OP_RELEASE_LOCKOWNER", + /* NFSv4.1 operations */ + [OP_EXCHANGE_ID] = "OP_EXCHANGE_ID", + [OP_BACKCHANNEL_CTL] = "OP_BACKCHANNEL_CTL", + [OP_BIND_CONN_TO_SESSION] = "OP_BIND_CONN_TO_SESSION", + [OP_CREATE_SESSION] = "OP_CREATE_SESSION", + [OP_DESTROY_SESSION] = "OP_DESTROY_SESSION", + [OP_SEQUENCE] = "OP_SEQUENCE", + [OP_DESTROY_CLIENTID] = "OP_DESTROY_CLIENTID", + [OP_RECLAIM_COMPLETE] = "OP_RECLAIM_COMPLETE", + [OP_SECINFO_NO_NAME] = "OP_SECINFO_NO_NAME", + [OP_TEST_STATEID] = "OP_TEST_STATEID", + [OP_FREE_STATEID] = "OP_FREE_STATEID", + [OP_GETDEVICEINFO] = "OP_GETDEVICEINFO", + [OP_LAYOUTGET] = "OP_LAYOUTGET", + [OP_LAYOUTCOMMIT] = "OP_LAYOUTCOMMIT", + [OP_LAYOUTRETURN] = "OP_LAYOUTRETURN", + /* NFSv4.2 operations */ + [OP_ALLOCATE] = "OP_ALLOCATE", + [OP_DEALLOCATE] = "OP_DEALLOCATE", + [OP_CLONE] = "OP_CLONE", + [OP_COPY] = "OP_COPY", + [OP_READ_PLUS] = "OP_READ_PLUS", + [OP_SEEK] = "OP_SEEK", + [OP_OFFLOAD_STATUS] = "OP_OFFLOAD_STATUS", + [OP_OFFLOAD_CANCEL] = "OP_OFFLOAD_CANCEL", + [OP_COPY_NOTIFY] = "OP_COPY_NOTIFY", + [OP_GETXATTR] = "OP_GETXATTR", + [OP_SETXATTR] = "OP_SETXATTR", + [OP_LISTXATTRS] = "OP_LISTXATTRS", + [OP_REMOVEXATTR] = "OP_REMOVEXATTR", +}; + +static int error_handler(struct sockaddr_nl *nla, struct nlmsgerr *err, + void *arg) +{ + int *ret = arg; + + *ret = err->error; + return NL_SKIP; +} + +static int finish_handler(struct nl_msg *msg, void *arg) +{ + int *ret = arg; + + *ret = 0; + return NL_SKIP; +} + +static int ack_handler(struct nl_msg *msg, void *arg) +{ + int *ret = arg; + + *ret = 0; + return NL_STOP; +} + +static void parse_rpc_status_get(struct genlmsghdr *gnlh) +{ + struct nlattr *attr; + int rem; + + nla_for_each_attr(attr, genlmsg_attrdata(gnlh, 0), + genlmsg_attrlen(gnlh, 0), rem) { + switch (nla_type(attr)) { + case NFSD_A_RPC_STATUS_XID: + case NFSD_A_RPC_STATUS_FLAGS: + printf(" 0x%08x", nla_get_u32(attr)); + break; + case NFSD_A_RPC_STATUS_PROC: + case NFSD_A_RPC_STATUS_PROG: + printf(" %d", nla_get_u32(attr)); + break; + case NFSD_A_RPC_STATUS_VERSION: + printf(" NFS%d", nla_get_u8(attr)); + break; + case NFSD_A_RPC_STATUS_SERVICE_TIME: + printf(" %ld", nla_get_u64(attr)); + break; + case NFSD_A_RPC_STATUS_DADDR4: + case NFSD_A_RPC_STATUS_SADDR4: { + struct in_addr addr = { + .s_addr = nla_get_u32(attr), + }; + + printf(" %s", inet_ntoa(addr)); + break; + } + case NFSD_A_RPC_STATUS_DPORT: + case NFSD_A_RPC_STATUS_SPORT: + printf(" %hu", nla_get_u16(attr)); + break; + case NFSD_A_RPC_STATUS_COMPOUND_OPS: { + unsigned int op = nla_get_u32(attr); + + if (op < NFSD4_OPS_MAX_LEN) + printf(" %s", nfsd4_ops[op]); + break; + } + default: + break; + } + } + printf("\n"); +} + +static void parse_version_get(struct genlmsghdr *gnlh) +{ + struct nlattr *attr; + int rem, idx = 0; + + /* clear the nfsd_versions array */ + memset(nfsd_versions, '\0', sizeof(*nfsd_versions) * MAX_NFS_VERSIONS); + + nla_for_each_attr(attr, genlmsg_attrdata(gnlh, 0), + genlmsg_attrlen(gnlh, 0), rem) { + struct nlattr *a; + int i; + + nla_for_each_nested(a, attr, i) { + switch (nla_type(a)) { + case NFSD_A_VERSION_MAJOR: + nfsd_versions[idx].major = nla_get_u32(a); + break; + case NFSD_A_VERSION_MINOR: + nfsd_versions[idx].minor = nla_get_u32(a); + break; + case NFSD_A_VERSION_ENABLED: + nfsd_versions[idx].enabled = nla_get_flag(a); + break; + default: + break; + } + } + ++idx; + } +} + +static void parse_listener_get(struct genlmsghdr *gnlh) +{ + struct nlattr *attr; + int rem, idx = 0; + + /* clear the nfsd_sockets array */ + memset(nfsd_sockets, '\0', sizeof(*nfsd_sockets) * MAX_NFSD_SOCKETS); + + nla_for_each_attr(attr, genlmsg_attrdata(gnlh, 0), + genlmsg_attrlen(gnlh, 0), rem) { + struct nlattr *a; + char *res; + int i; + + nla_for_each_nested(a, attr, i) { + switch (nla_type(a)) { + case NFSD_A_SOCK_TRANSPORT_NAME: + res = strncpy(nfsd_sockets[idx].name, nla_data(a), + MAX_CLASS_NAME_LEN); + res[MAX_CLASS_NAME_LEN - 1] = '\0'; // just to be sure + break; + case NFSD_A_SOCK_ADDR: + memcpy(&nfsd_sockets[idx].ss, nla_data(a), + sizeof(nfsd_sockets[idx].ss)); + break; + } + nfsd_sockets[idx].active = true; + } + ++idx; + } + nfsd_socket_count = idx; +} + +static void parse_threads_get(struct genlmsghdr *gnlh) +{ + struct nlattr *attr; + int rem, pools = 0, i = 0; + uint32_t *pool_threads = NULL; + + nla_for_each_attr(attr, genlmsg_attrdata(gnlh, 0), + genlmsg_attrlen(gnlh, 0), rem) + if (nla_type(attr) == NFSD_A_SERVER_THREADS) + ++pools; + + pool_threads = alloca(pools * sizeof(*pool_threads)); + + nla_for_each_attr(attr, genlmsg_attrdata(gnlh, 0), + genlmsg_attrlen(gnlh, 0), rem) { + switch (nla_type(attr)) { + case NFSD_A_SERVER_GRACETIME: + printf("gracetime: %u\n", nla_get_u32(attr)); + break; + case NFSD_A_SERVER_LEASETIME: + printf("leasetime: %u\n", nla_get_u32(attr)); + break; + case NFSD_A_SERVER_SCOPE: + printf("scope: %s\n", (const char *)nla_data(attr)); + break; + case NFSD_A_SERVER_THREADS: + pool_threads[i++] = nla_get_u32(attr); + break; + default: + break; + } + } + + printf("pool-threads:"); + for (i = 0; i < pools; ++i) + printf(" %d", pool_threads[i]); + putchar('\n'); +} + +static void parse_pool_mode_get(struct genlmsghdr *gnlh) +{ + struct nlattr *attr; + int rem; + + nla_for_each_attr(attr, genlmsg_attrdata(gnlh, 0), + genlmsg_attrlen(gnlh, 0), rem) { + switch (nla_type(attr)) { + case NFSD_A_POOL_MODE_MODE: + printf("pool-mode: %s\n", (const char *)nla_data(attr)); + break; + case NFSD_A_POOL_MODE_NPOOLS: + printf("npools: %u\n", nla_get_u32(attr)); + break; + default: + break; + } + } +} + +static int recv_handler(struct nl_msg *msg, void *arg) +{ + struct genlmsghdr *gnlh = nlmsg_data(nlmsg_hdr(msg)); + + switch (gnlh->cmd) { + case NFSD_CMD_RPC_STATUS_GET: + parse_rpc_status_get(gnlh); + break; + case NFSD_CMD_THREADS_GET: + parse_threads_get(gnlh); + break; + case NFSD_CMD_VERSION_GET: + parse_version_get(gnlh); + break; + case NFSD_CMD_LISTENER_GET: + parse_listener_get(gnlh); + break; + case NFSD_CMD_POOL_MODE_GET: + parse_pool_mode_get(gnlh); + break; + default: + break; + } + + return NL_SKIP; +} + +#define BUFFER_SIZE 8192 +static struct nl_sock *netlink_sock_alloc(void) +{ + struct nl_sock *sock; + int ret; + + sock = nl_socket_alloc(); + if (!sock) + return NULL; + + if (genl_connect(sock)) { + fprintf(stderr, "Failed to connect to generic netlink\n"); + nl_socket_free(sock); + return NULL; + } + + nl_socket_set_buffer_size(sock, BUFFER_SIZE, BUFFER_SIZE); + setsockopt(nl_socket_get_fd(sock), SOL_NETLINK, NETLINK_EXT_ACK, + &ret, sizeof(ret)); + + return sock; +} + +static struct nl_msg *netlink_msg_alloc(struct nl_sock *sock) +{ + struct nl_msg *msg; + int id; + + id = genl_ctrl_resolve(sock, NFSD_FAMILY_NAME); + if (id < 0) { + fprintf(stderr, "%s not found\n", NFSD_FAMILY_NAME); + return NULL; + } + + msg = nlmsg_alloc(); + if (!msg) { + fprintf(stderr, "failed to allocate netlink message\n"); + return NULL; + } + + if (!genlmsg_put(msg, 0, 0, id, 0, 0, 0, 0)) { + fprintf(stderr, "failed to allocate netlink message\n"); + nlmsg_free(msg); + return NULL; + } + + return msg; +} + +static void status_usage(void) +{ + printf("Usage: %s status\n", taskname); + printf(" Display RPC jobs currently in flight on the server.\n"); +} + +static int status_func(struct nl_sock *sock, int argc, char ** argv) +{ + struct genlmsghdr *ghdr; + struct nlmsghdr *nlh; + struct nl_msg *msg; + struct nl_cb *cb; + int opt, ret; + + optind = 1; + while ((opt = getopt_long(argc, argv, "h", help_only_options, NULL)) != -1) { + switch (opt) { + case 'h': + status_usage(); + return 0; + } + } + + msg = netlink_msg_alloc(sock); + if (!msg) + return 1; + + nlh = nlmsg_hdr(msg); + nlh->nlmsg_flags |= NLM_F_DUMP; + ghdr = nlmsg_data(nlh); + ghdr->cmd = NFSD_CMD_RPC_STATUS_GET; + + cb = nl_cb_alloc(NL_CB_CUSTOM); + if (!cb) { + fprintf(stderr, "failed to allocate netlink callbacks\n"); + ret = 1; + goto out; + } + + ret = nl_send_auto(sock, msg); + if (ret < 0) + goto out_cb; + + ret = 1; + nl_cb_err(cb, NL_CB_CUSTOM, error_handler, &ret); + nl_cb_set(cb, NL_CB_FINISH, NL_CB_CUSTOM, finish_handler, &ret); + nl_cb_set(cb, NL_CB_ACK, NL_CB_CUSTOM, ack_handler, &ret); + nl_cb_set(cb, NL_CB_VALID, NL_CB_CUSTOM, recv_handler, NULL); + + while (ret > 0) + nl_recvmsgs(sock, cb); + if (ret < 0) { + fprintf(stderr, "Error: %s\n", strerror(-ret)); + ret = 1; + } +out_cb: + nl_cb_put(cb); +out: + nlmsg_free(msg); + return ret; +} + +static int threads_doit(struct nl_sock *sock, int cmd, int grace, int lease, + int pool_count, int *pool_threads, char *scope) +{ + struct genlmsghdr *ghdr; + struct nlmsghdr *nlh; + struct nl_msg *msg; + struct nl_cb *cb; + int ret; + + msg = netlink_msg_alloc(sock); + if (!msg) + return 1; + + nlh = nlmsg_hdr(msg); + if (cmd == NFSD_CMD_THREADS_SET) { + int i; + + if (grace) + nla_put_u32(msg, NFSD_A_SERVER_GRACETIME, grace); + if (lease) + nla_put_u32(msg, NFSD_A_SERVER_LEASETIME, lease); + if (scope) + nla_put_string(msg, NFSD_A_SERVER_SCOPE, scope); + for (i = 0; i < pool_count; ++i) + nla_put_u32(msg, NFSD_A_SERVER_THREADS, pool_threads[i]); + } + ghdr = nlmsg_data(nlh); + ghdr->cmd = cmd; + + cb = nl_cb_alloc(NL_CB_CUSTOM); + if (!cb) { + fprintf(stderr, "failed to allocate netlink callbacks\n"); + ret = 1; + goto out; + } + + ret = nl_send_auto(sock, msg); + if (ret < 0) { + fprintf(stderr, "send failed (%d)!\n", ret); + goto out_cb; + } + + ret = 1; + nl_cb_err(cb, NL_CB_CUSTOM, error_handler, &ret); + nl_cb_set(cb, NL_CB_FINISH, NL_CB_CUSTOM, finish_handler, &ret); + nl_cb_set(cb, NL_CB_ACK, NL_CB_CUSTOM, ack_handler, &ret); + nl_cb_set(cb, NL_CB_VALID, NL_CB_CUSTOM, recv_handler, NULL); + + while (ret > 0) + nl_recvmsgs(sock, cb); + if (ret < 0) { + fprintf(stderr, "Error: %s\n", strerror(-ret)); + ret = 1; + } +out_cb: + nl_cb_put(cb); +out: + nlmsg_free(msg); + return ret; +} + +static void threads_usage(void) +{ + printf("Usage: %s threads [ pool0_count ] [ pool1_count ] ...\n", taskname); + printf(" pool0_count: thread count for pool0, etc...\n"); + printf("Omit any arguments to show current thread counts.\n"); +} + +static int threads_func(struct nl_sock *sock, int argc, char **argv) +{ + uint8_t cmd = NFSD_CMD_THREADS_GET; + int *pool_threads = NULL; + int opt, pools = 0; + + optind = 1; + while ((opt = getopt_long(argc, argv, "h", help_only_options, NULL)) != -1) { + switch (opt) { + case 'h': + threads_usage(); + return 0; + } + } + + if (optind < argc) { + char **targv = &argv[optind]; + int i; + + pools = argc - optind; + pool_threads = alloca(pools * sizeof(*pool_threads)); + cmd = NFSD_CMD_THREADS_SET; + + for (i = 0; i < pools; ++i) { + char *endptr = NULL; + + /* empty string? */ + if (targv[i][0] == '\0') { + fprintf(stderr, "Invalid threads value %s.\n", targv[i]); + return 1; + } + + pool_threads[i] = strtol(targv[i], &endptr, 0); + if (!endptr || *endptr != '\0') { + fprintf(stderr, "Invalid threads value %s.\n", argv[1]); + return 1; + } + } + } + return threads_doit(sock, cmd, 0, 0, pools, pool_threads, NULL); +} + +/* + * Update the nfsd_versions array with the latest info from the kernel + */ +static int fetch_nfsd_versions(struct nl_sock *sock) +{ + struct genlmsghdr *ghdr; + struct nlmsghdr *nlh; + struct nl_msg *msg; + struct nl_cb *cb; + int ret; + + msg = netlink_msg_alloc(sock); + if (!msg) + return 1; + + nlh = nlmsg_hdr(msg); + ghdr = nlmsg_data(nlh); + ghdr->cmd = NFSD_CMD_VERSION_GET; + + cb = nl_cb_alloc(NL_CB_CUSTOM); + if (!cb) { + fprintf(stderr, "failed to allocate netlink callbacks\n"); + ret = 1; + goto out; + } + + ret = nl_send_auto(sock, msg); + if (ret < 0) { + fprintf(stderr, "send failed: %d\n", ret); + goto out_cb; + } + + ret = 1; + nl_cb_err(cb, NL_CB_CUSTOM, error_handler, &ret); + nl_cb_set(cb, NL_CB_FINISH, NL_CB_CUSTOM, finish_handler, &ret); + nl_cb_set(cb, NL_CB_ACK, NL_CB_CUSTOM, ack_handler, &ret); + nl_cb_set(cb, NL_CB_VALID, NL_CB_CUSTOM, recv_handler, NULL); + + while (ret > 0) + nl_recvmsgs(sock, cb); + if (ret < 0) { + fprintf(stderr, "Error: %s\n", strerror(-ret)); + ret = 1; + } +out_cb: + nl_cb_put(cb); +out: + nlmsg_free(msg); + return ret; +} + +static void print_versions_array(void) +{ + int i; + + for (i = 0; i < MAX_NFS_VERSIONS; ++i) { + /* A major of zero indicates the end of the array */ + if (nfsd_versions[i].major == 0) + break; + if (i != 0) + printf(" "); + printf("%c%hhd.%hhd", + nfsd_versions[i].enabled ? '+' : '-', + nfsd_versions[i].major, nfsd_versions[i].minor); + } + putchar('\n'); +} + +static int set_nfsd_versions(struct nl_sock *sock) +{ + struct genlmsghdr *ghdr; + struct nlmsghdr *nlh; + struct nl_msg *msg; + struct nl_cb *cb; + int i, ret; + + msg = netlink_msg_alloc(sock); + if (!msg) + return 1; + + nlh = nlmsg_hdr(msg); + + for (i = 0; i < MAX_NFS_VERSIONS; ++i) { + struct nlattr *a; + + if (nfsd_versions[i].major == 0) + break; + + a = nla_nest_start(msg, NLA_F_NESTED | NFSD_A_SERVER_PROTO_VERSION); + if (!a) { + fprintf(stderr, "Unable to allocate version nest!\n"); + ret = 1; + goto out; + } + + nla_put_u32(msg, NFSD_A_VERSION_MAJOR, nfsd_versions[i].major); + nla_put_u32(msg, NFSD_A_VERSION_MINOR, nfsd_versions[i].minor); + if (nfsd_versions[i].enabled) + nla_put_flag(msg, NFSD_A_VERSION_ENABLED); + nla_nest_end(msg, a); + } + + ghdr = nlmsg_data(nlh); + ghdr->cmd = NFSD_CMD_VERSION_SET; + + cb = nl_cb_alloc(NL_CB_CUSTOM); + if (!cb) { + fprintf(stderr, "Failed to allocate netlink callbacks\n"); + ret = 1; + goto out; + } + + ret = nl_send_auto(sock, msg); + if (ret < 0) { + fprintf(stderr, "Send failed: %d\n", ret); + goto out_cb; + } + + ret = 1; + nl_cb_err(cb, NL_CB_CUSTOM, error_handler, &ret); + nl_cb_set(cb, NL_CB_FINISH, NL_CB_CUSTOM, finish_handler, &ret); + nl_cb_set(cb, NL_CB_ACK, NL_CB_CUSTOM, ack_handler, &ret); + nl_cb_set(cb, NL_CB_VALID, NL_CB_CUSTOM, recv_handler, NULL); + + while (ret > 0) + nl_recvmsgs(sock, cb); + if (ret < 0) { + fprintf(stderr, "Error: %s\n", strerror(-ret)); + ret = 1; + } +out_cb: + nl_cb_put(cb); +out: + nlmsg_free(msg); + return ret; +} + +static int update_nfsd_version(int major, int minor, bool enabled) +{ + int i; + + for (i = 0; i < MAX_NFS_VERSIONS; ++i) { + if (nfsd_versions[i].major == 0) + break; + if (nfsd_versions[i].major == major && nfsd_versions[i].minor == minor) { + nfsd_versions[i].enabled = enabled; + return 0; + } + } + /* the kernel doesn't support this version */ + if (!enabled) + return 0; + fprintf(stderr, "This kernel does not support NFS version %d.%d\n", major, minor); + return -EINVAL; +} + +static void version_usage(void) +{ + printf("Usage: %s version { {+,-}major.minor } ...\n", taskname); + printf(" + to enable a version, - to disable it\n"); + printf(" @major: major version number\n"); + printf(" @minor: minor version number\n"); + printf("Examples:\n"); + printf(" Display currently enabled and disabled versions:\n"); + printf(" version\n"); + printf(" Disable NFSv4.0:\n"); + printf(" version -v4.0\n"); + printf(" Enable v4.1, v4.2, disable v2, v3 and v4.0:\n"); + printf(" version -2 -3 -v4.0 +4.1 +v4.2\n"); +} + +static int version_func(struct nl_sock *sock, int argc, char ** argv) +{ + int ret, i; + + /* help is only valid as first argument after command */ + if (argc > 1 && + (!strcmp(argv[1], "-h") || !strcmp(argv[1], "--help"))) { + version_usage(); + return 0; + } + + ret = fetch_nfsd_versions(sock); + if (ret) + return ret; + + if (argc > 1) { + for (i = 1; i < argc; ++i) { + int ret, major, minor = 0; + char sign = '\0', *str = argv[i]; + bool enabled; + + ret = sscanf(str, "%c%d.%d\n", &sign, &major, &minor); + if (ret < 2) { + fprintf(stderr, "Invalid version string (%d) %s\n", ret, str); + return -EINVAL; + } + + switch(sign) { + case '+': + enabled = true; + break; + case '-': + enabled = false; + break; + default: + fprintf(stderr, "Invalid version string %s\n", str); + return -EINVAL; + } + + ret = update_nfsd_version(major, minor, enabled); + if (ret) + return ret; + } + return set_nfsd_versions(sock); + } + + print_versions_array(); + return 0; +} + +static int fetch_current_listeners(struct nl_sock *sock) +{ + struct genlmsghdr *ghdr; + struct nlmsghdr *nlh; + struct nl_msg *msg; + struct nl_cb *cb; + int ret; + + msg = netlink_msg_alloc(sock); + if (!msg) + return 1; + + nlh = nlmsg_hdr(msg); + ghdr = nlmsg_data(nlh); + ghdr->cmd = NFSD_CMD_LISTENER_GET; + + cb = nl_cb_alloc(NL_CB_CUSTOM); + if (!cb) { + fprintf(stderr, "failed to allocate netlink callbacks\n"); + ret = 1; + goto out; + } + + ret = nl_send_auto(sock, msg); + if (ret < 0) { + fprintf(stderr, "send failed: %d\n", ret); + goto out_cb; + } + + ret = 1; + nl_cb_err(cb, NL_CB_CUSTOM, error_handler, &ret); + nl_cb_set(cb, NL_CB_FINISH, NL_CB_CUSTOM, finish_handler, &ret); + nl_cb_set(cb, NL_CB_ACK, NL_CB_CUSTOM, ack_handler, &ret); + nl_cb_set(cb, NL_CB_VALID, NL_CB_CUSTOM, recv_handler, NULL); + + while (ret > 0) + nl_recvmsgs(sock, cb); + if (ret < 0) { + fprintf(stderr, "Error: %s\n", strerror(-ret)); + ret = 1; + } +out_cb: + nl_cb_put(cb); +out: + nlmsg_free(msg); + return ret; +} + +static void print_listeners(void) +{ + int i; + const char *res; + + for (i = 0; i < MAX_NFSD_SOCKETS; ++i) { + struct server_socket *sock = &nfsd_sockets[i]; + char addr[INET6_ADDRSTRLEN + 1]; + in_port_t port = 0; + + if (*sock->name == '\0') + break; + + if (!sock->active) + continue; + + switch(sock->ss.ss_family) { + case AF_INET: + res = inet_ntop(AF_INET, &((struct sockaddr_in *)(&sock->ss))->sin_addr, + addr, INET6_ADDRSTRLEN); + port = ((struct sockaddr_in *)(&sock->ss))->sin_port; + if (res == NULL) + perror("inet_ntop"); + else + printf("%s:%s:%hu\n", sock->name, addr, ntohs(port)); + break; + case AF_INET6: + res = inet_ntop(AF_INET6, &((struct sockaddr_in6 *)(&sock->ss))->sin6_addr, + addr, INET6_ADDRSTRLEN); + port = ((struct sockaddr_in6 *)(&sock->ss))->sin6_port; + if (res == NULL) + perror("inet_ntop"); + else + printf("%s:[%s]:%hu\n", sock->name, addr, ntohs(port)); + break; + default: + snprintf(addr, INET6_ADDRSTRLEN, "Unknown address family: %d\n", + sock->ss.ss_family); + addr[INET6_ADDRSTRLEN - 1] = '\0'; + } + } +} + +#define BUFLEN (INET6_ADDRSTRLEN + 16) + +/* + * Format is <+/-><netid>:<address>:port + * + * + or -: denotes whether we're adding or removing a socket + * netid: tcp, udp, rdma (something else in the future?( + * address: IPv4 or IPv6 address. IPv6 addr should be in square brackets + * port: decimal port value + */ +static int update_listeners(const char *str) +{ + char buf[INET6_ADDRSTRLEN + 16]; + char sign = *str; + char *netid, *addr, *port, *end; + struct addrinfo *res; + int i, ret; + struct addrinfo hints = { .ai_flags = AI_PASSIVE, + .ai_family = AF_INET, + .ai_socktype = SOCK_STREAM, + .ai_protocol = IPPROTO_TCP }; + + if (sign != '+' && sign != '-') + goto out_inval; + + strcpy(buf, str + 1); + + /* netid is start */ + netid = buf; + + /* find first ':' */ + addr = strchr(buf, ':'); + if (!addr) + goto out_inval; + + if (addr == buf) { + /* empty netid */ + goto out_inval; + } + *addr = '\0'; + ++addr; + + port = strrchr(addr, ':'); + if (!port) + goto out_inval; + if (port == addr) { + /* empty address, give gai a NULL ptr */ + addr = NULL; + } + *port = '\0'; + port++; + + if (*port == '\0') { + /* empty port */ + goto out_inval; + } + + /* IPv6 addrs must be in square brackets */ + if (addr && *addr == '[') { + hints.ai_family = AF_INET6; + ++addr; + end = strchr(addr, ']'); + if (!end) + goto out_inval; + if (end == addr) + addr = NULL; + *end = '\0'; + } + + /* + * If we're looking for wildcard address, look for both + * families. + */ + if (!addr) + hints.ai_family = AF_UNSPEC; + + /* + * Note that we hint for a stream/tcp socket just to limit the number of + * entries that come back. We're only interested in the sockaddrs. + */ + ret = getaddrinfo(addr, port, &hints, &res); + if (ret) { + fprintf(stderr, "getaddrinfo of \"%s\" failed: %s\n", + addr, gai_strerror(ret)); + return -EINVAL; + } + + for ( ; res; res = res->ai_next) { + struct sockaddr_in6 *r6 = (struct sockaddr_in6 *)res->ai_addr; + struct sockaddr_in *r4 = (struct sockaddr_in *)res->ai_addr; + bool found = false; + + for (i = 0; i < MAX_NFSD_SOCKETS; ++i) { + struct server_socket *sock = &nfsd_sockets[i]; + struct sockaddr_in6 *l6 = (struct sockaddr_in6 *)&sock->ss; + struct sockaddr_in *l4 = (struct sockaddr_in *)&sock->ss; + + if (sock->ss.ss_family == AF_UNSPEC) + break; + + if (sock->ss.ss_family != res->ai_addr->sa_family) + continue; + + if (strcmp(sock->name, netid)) + continue; + + switch(sock->ss.ss_family) { + case AF_INET: + if (r4->sin_port != l4->sin_port || + memcmp(&r4->sin_addr, &l4->sin_addr, sizeof(l4->sin_addr))) + continue; + case AF_INET6: + if (r6->sin6_port != l6->sin6_port || + memcmp(&r6->sin6_addr, &l6->sin6_addr, sizeof(l6->sin6_addr))) + continue; + default: + + } + sock->active = (sign == '+'); + found = true; + break; + } + if (!found && sign == '+') { + struct server_socket *sock = &nfsd_sockets[nfsd_socket_count]; + + memcpy(&sock->ss, res->ai_addr, res->ai_addrlen); + strncpy(sock->name, netid, MAX_CLASS_NAME_LEN); + sock->name[MAX_CLASS_NAME_LEN - 1] = '\0'; + sock->active = true; + ++nfsd_socket_count; + } + } + return 0; +out_inval: + fprintf(stderr, "Invalid listener update string: %s", str); + return -EINVAL; +} + +static int set_listeners(struct nl_sock *sock) +{ + struct genlmsghdr *ghdr; + struct nlmsghdr *nlh; + struct nl_msg *msg; + struct nl_cb *cb; + int i, ret; + + msg = netlink_msg_alloc(sock); + if (!msg) + return 1; + + nlh = nlmsg_hdr(msg); + + for (i = 0; i < MAX_NFSD_SOCKETS; ++i) { + struct server_socket *sock = &nfsd_sockets[i]; + struct nlattr *a; + + if (sock->ss.ss_family == 0) + break; + + if (!sock->active) + continue; + + a = nla_nest_start(msg, NLA_F_NESTED | NFSD_A_SERVER_SOCK_ADDR); + if (!a) { + fprintf(stderr, "Unable to allocate listener nest!\n"); + ret = 1; + goto out; + } + + nla_put(msg, NFSD_A_SOCK_ADDR, sizeof(sock->ss), &sock->ss); + nla_put_string(msg, NFSD_A_SOCK_TRANSPORT_NAME, sock->name); + nla_nest_end(msg, a); + } + + ghdr = nlmsg_data(nlh); + ghdr->cmd = NFSD_CMD_LISTENER_SET; + + cb = nl_cb_alloc(NL_CB_CUSTOM); + if (!cb) { + fprintf(stderr, "Failed to allocate netlink callbacks\n"); + ret = 1; + goto out; + } + + ret = nl_send_auto(sock, msg); + if (ret < 0) { + fprintf(stderr, "Send failed: %d\n", ret); + goto out_cb; + } + + ret = 1; + nl_cb_err(cb, NL_CB_CUSTOM, error_handler, &ret); + nl_cb_set(cb, NL_CB_FINISH, NL_CB_CUSTOM, finish_handler, &ret); + nl_cb_set(cb, NL_CB_ACK, NL_CB_CUSTOM, ack_handler, &ret); + nl_cb_set(cb, NL_CB_VALID, NL_CB_CUSTOM, recv_handler, NULL); + + while (ret > 0) + nl_recvmsgs(sock, cb); + if (ret < 0) { + fprintf(stderr, "Error: %s\n", strerror(-ret)); + ret = 1; + } +out_cb: + nl_cb_put(cb); +out: + nlmsg_free(msg); + return ret; +} + +static void listener_usage(void) +{ + printf("Usage: %s listener { {+,-}proto:addr:port } ...\n", taskname); + printf(" + to add a listener, - to remove one\n"); + printf(" @proto: protocol (e.g. tcp, udp, rdma)\n"); + printf(" @addr: hostname or address to listen on (blank string == wildcard addresses)\n"); + printf(" @port: port number or service name to listen on\n\n"); + printf("Examples:\n"); + printf(" Display currently configured listeners:\n"); + printf(" listener\n"); + printf(" Add TCP listener on all addresses (both v4 and v6), port 2049:\n"); + printf(" listener +tcp::2049\n"); + printf(" Add RDMA listener on 1.2.3.4 port 20049:\n"); + printf(" listener +rdma:1.2.3.4:20049\n"); + printf(" Add same listener on IPv6 address f00::ba4 port 20050:\n"); + printf(" listener +rdma:[f00::ba4]:20050\n"); + printf(" Remove UDP listener from nfsserver.example.org, nfs port:\n"); + printf(" listener -udp:nfsserver.example.org:nfs\n\n"); +} + +static int listener_func(struct nl_sock *sock, int argc, char ** argv) +{ + int ret, i; + + /* help is only valid as first argument after command */ + if (argc > 1 && + (!strcmp(argv[1], "-h") || !strcmp(argv[1], "--help"))) { + listener_usage(); + return 0; + } + + ret = fetch_current_listeners(sock); + if (ret) + return ret; + + if (argc > 1) { + for (i = 1; i < argc; ++i) + update_listeners(argv[i]); + return set_listeners(sock); + } + + print_listeners(); + return 0; +} + +static int pool_mode_doit(struct nl_sock *sock, int cmd, const char *pool_mode) +{ + struct genlmsghdr *ghdr; + struct nlmsghdr *nlh; + struct nl_msg *msg; + struct nl_cb *cb; + int ret; + + msg = netlink_msg_alloc(sock); + if (!msg) + return 1; + + nlh = nlmsg_hdr(msg); + if (cmd == NFSD_CMD_POOL_MODE_SET) { + if (pool_mode) + nla_put_string(msg, NFSD_A_POOL_MODE_MODE, pool_mode); + } + ghdr = nlmsg_data(nlh); + ghdr->cmd = cmd; + + cb = nl_cb_alloc(NL_CB_CUSTOM); + if (!cb) { + fprintf(stderr, "failed to allocate netlink callbacks\n"); + ret = 1; + goto out; + } + + ret = nl_send_auto(sock, msg); + if (ret < 0) { + fprintf(stderr, "send failed (%d)!\n", ret); + goto out_cb; + } + + ret = 1; + nl_cb_err(cb, NL_CB_CUSTOM, error_handler, &ret); + nl_cb_set(cb, NL_CB_FINISH, NL_CB_CUSTOM, finish_handler, &ret); + nl_cb_set(cb, NL_CB_ACK, NL_CB_CUSTOM, ack_handler, &ret); + nl_cb_set(cb, NL_CB_VALID, NL_CB_CUSTOM, recv_handler, NULL); + + while (ret > 0) + nl_recvmsgs(sock, cb); + if (ret < 0) { + fprintf(stderr, "Error: %s\n", strerror(-ret)); + ret = 1; + } +out_cb: + nl_cb_put(cb); +out: + nlmsg_free(msg); + return ret; +} + +static void pool_mode_usage(void) +{ + printf("Usage: %s pool-mode { global | auto | percpu | pernode } ...\n", taskname); +} + +static int pool_mode_func(struct nl_sock *sock, int argc, char **argv) +{ + uint8_t cmd = NFSD_CMD_POOL_MODE_GET; + const char *pool_mode = NULL; + int opt; + + optind = 1; + while ((opt = getopt_long(argc, argv, "h", help_only_options, NULL)) != -1) { + switch (opt) { + case 'h': + pool_mode_usage(); + return 0; + } + } + + if (optind < argc) { + char **targv = &argv[optind]; + + cmd = NFSD_CMD_POOL_MODE_SET; + + /* empty string? */ + if (*targv[0] == '\0') { + fprintf(stderr, "Invalid threads value %s.\n", targv[0]); + return 1; + } + pool_mode = targv[0]; + } + return pool_mode_doit(sock, cmd, pool_mode); +} + +#define MAX_LISTENER_LEN (64 * 2 + 16) + +static void +add_listener(const char *netid, const char *addr, const char *port) +{ + char buf[MAX_LISTENER_LEN]; + + if (strchr(addr, ':')) + snprintf(buf, MAX_LISTENER_LEN, "+%s:[%s]:%s", + netid, addr, port); + else + snprintf(buf, MAX_LISTENER_LEN, "+%s:%s:%s", + netid, addr, port); + buf[MAX_LISTENER_LEN - 1] = '\0'; + update_listeners(buf); +} + +static void +read_nfsd_conf(void) +{ + conf_init_file(NFS_CONFFILE); + xlog_set_debug("nfsd"); +} + +static void configure_versions(void) +{ + bool v4 = conf_get_bool("nfsd", "vers4", true); + + update_nfsd_version(2, 0, conf_get_bool("nfsd", "vers2", false)); + update_nfsd_version(3, 0, conf_get_bool("nfsd", "vers3", true)); + update_nfsd_version(4, 0, v4 && conf_get_bool("nfsd", "vers4.0", true)); + update_nfsd_version(4, 1, v4 && conf_get_bool("nfsd", "vers4.1", true)); + update_nfsd_version(4, 2, v4 && conf_get_bool("nfsd", "vers4.2", true)); +} + +static void configure_listeners(void) +{ + char *port, *rdma_port; + bool rdma, udp, tcp; + struct conf_list *hosts; + + udp = conf_get_bool("nfsd", "udp", false); + tcp = conf_get_bool("nfsd", "tcp", true); + port = conf_get_str("nfsd", "port"); + if (!port) + port = "nfs"; + + rdma = conf_get_bool("nfsd", "rdma", false); + if (rdma) { + rdma_port = conf_get_str("nfsd", "rdma-port"); + if (!rdma_port) + rdma_port = "nfsrdma"; + } + + /* backward compatibility - nfs.conf used to set rdma port directly */ + if (!rdma_port) + rdma_port = conf_get_str("nfsd", "rdma"); + + hosts = conf_get_list("nfsd", "host"); + if (hosts && hosts->cnt) { + struct conf_list_node *n; + TAILQ_FOREACH(n, &(hosts->fields), link) { + if (udp) + add_listener("udp", n->field, port); + if (tcp) + add_listener("tcp", n->field, port); + if (rdma) + add_listener("rdma", n->field, rdma_port); + } + } else { + if (udp) + add_listener("udp", "", port); + if (tcp) + add_listener("tcp", "", port); + if (rdma) + add_listener("rdma", "", rdma_port); + } +} + +static void autostart_usage(void) +{ + printf("Usage: %s autostart\n", taskname); + printf(" Start the server with the settings in /etc/nfs.conf.\n"); +} + +/* default number of nfsd threads when not specified in nfs.conf */ +#define DEFAULT_AUTOSTART_THREADS 16 + +static int autostart_func(struct nl_sock *sock, int argc, char ** argv) +{ + int *threads, grace, lease, idx, ret, opt, pools; + struct conf_list *thread_str; + struct conf_list_node *n; + char *scope, *pool_mode; + + optind = 1; + while ((opt = getopt_long(argc, argv, "h", help_only_options, NULL)) != -1) { + switch (opt) { + case 'h': + autostart_usage(); + return 0; + } + } + + read_nfsd_conf(); + + pool_mode = conf_get_str("nfsd", "pool-mode"); + if (pool_mode) { + ret = pool_mode_doit(sock, NFSD_CMD_POOL_MODE_SET, pool_mode); + if (ret) + return ret; + } + + ret = fetch_nfsd_versions(sock); + if (ret) + return ret; + configure_versions(); + ret = set_nfsd_versions(sock); + if (ret) + return ret; + + configure_listeners(); + ret = set_listeners(sock); + if (ret) + return ret; + + grace = conf_get_num("nfsd", "grace-time", 0); + lease = conf_get_num("nfsd", "lease-time", 0); + scope = conf_get_str("nfsd", "scope"); + + thread_str = conf_get_list("nfsd", "threads"); + pools = thread_str ? thread_str->cnt : 1; + + threads = calloc(pools, sizeof(int)); + if (!threads) + return -ENOMEM; + + if (thread_str) { + idx = 0; + TAILQ_FOREACH(n, &(thread_str->fields), link) { + char *endptr = NULL; + + threads[idx++] = strtol(n->field, &endptr, 0); + if (!endptr || *endptr != '\0') { + fprintf(stderr, "Invalid threads value %s.\n", n->field); + ret = -EINVAL; + goto out; + } + } + } else { + threads[0] = DEFAULT_AUTOSTART_THREADS; + } + + ret = threads_doit(sock, NFSD_CMD_THREADS_SET, grace, lease, pools, + threads, scope); +out: + free(threads); + return ret; +} + +enum nfsdctl_commands { + NFSDCTL_STATUS, + NFSDCTL_THREADS, + NFSDCTL_VERSION, + NFSDCTL_LISTENER, + NFSDCTL_AUTOSTART, + NFSDCTL_POOL_MODE, +}; + +static int parse_command(char *str) +{ + if (!strcmp(str, "status")) + return NFSDCTL_STATUS; + if (!strcmp(str, "threads")) + return NFSDCTL_THREADS; + if (!strcmp(str, "version")) + return NFSDCTL_VERSION; + if (!strcmp(str, "listener")) + return NFSDCTL_LISTENER; + if (!strcmp(str, "autostart")) + return NFSDCTL_AUTOSTART; + if (!strcmp(str, "pool-mode")) + return NFSDCTL_POOL_MODE; + return -1; +} + +typedef int (*nfsdctl_func)(struct nl_sock *sock, int argc, char **argv); + +static nfsdctl_func func[] = { + [NFSDCTL_STATUS] = status_func, + [NFSDCTL_THREADS] = threads_func, + [NFSDCTL_VERSION] = version_func, + [NFSDCTL_LISTENER] = listener_func, + [NFSDCTL_AUTOSTART] = autostart_func, + [NFSDCTL_POOL_MODE] = pool_mode_func, +}; + +static void usage(void) +{ + printf("Usage:\n"); + printf("%s [-hv] [COMMAND] [ARGS]\n", taskname); + printf(" options:\n"); + printf(" -h | --help usage info\n"); + printf(" -d | --debug=NUM enable debugging\n"); + printf(" -V | --version print version info\n"); + printf(" commands:\n"); + printf(" pool-mode get/set host pool mode setting\n"); + printf(" listener get/set listener info\n"); + printf(" version get/set supported NFS versions\n"); + printf(" threads get/set nfsd thread settings\n"); + printf(" status get current RPC processing info\n"); + printf(" autostart start server with settings from /etc/nfs.conf\n"); +} + +/* Options given before the command string */ +static const struct option pre_options[] = { + { "help", no_argument, NULL, 'h' }, + { "debug", required_argument, NULL, 'd' }, + { "version", no_argument, NULL, 'V' }, + { }, +}; + +static int run_one_command(struct nl_sock *sock, int argc, char **argv) +{ + int cmd = parse_command(argv[0]); + + if (cmd < 0) { + usage(); + return 1; + } + return func[cmd](sock, argc, argv); +} + +#define MAX_ARGUMENTS 256 + +static int tokenize_string(char *line, int *argc, char **argv) +{ + int idx = 0; + char *arg, *save; + + memset(argv, '\0', sizeof(*argv) * MAX_ARGUMENTS); + + arg = strtok_r(line, " \t", &save); + while(arg) { + argv[idx] = arg; + ++idx; + if (idx >= MAX_ARGUMENTS) + return -E2BIG; + arg = strtok_r(NULL, " \t", &save); + } + *argc = idx; + return 0; +} + +static int run_commandline(struct nl_sock *sock) +{ + char *argv[MAX_ARGUMENTS]; + char *line; + int ret, argc; + + for (;;) { + line = readline("nfsdctl> "); + if (!line || !strcmp(line, "quit")) + break; + if (*line == '\0') + continue; + add_history(line); + ret = tokenize_string(line, &argc, argv); + if (!ret) + ret = run_one_command(sock, argc, argv); + if (ret) + fprintf(stderr, "Error: %s\n", strerror(ret)); + free(line); + } + return 0; +} + +int main(int argc, char **argv) +{ + int opt, ret; + struct nl_sock *sock = netlink_sock_alloc(); + + if (!sock) { + fprintf(stderr, "Unable to allocate netlink socket!"); + return 1; + } + + taskname = argv[0]; + + /* Parse the preliminary options */ + while ((opt = getopt_long(argc, argv, "+hd:V", pre_options, NULL)) != -1) { + switch (opt) { + case 'h': + usage(); + return 0; + case 'd': + debug_level = atoi(optarg); + break; + case 'V': + // FIXME: print_version(); + return 0; + } + } + + if (optind > argc) { + usage(); + return 1; + } + + if (optind == argc) + ret = run_commandline(sock); + else + ret = run_one_command(sock, argc - optind, &argv[optind]); + + nl_socket_free(sock); + return ret; +} diff --git a/utils/nfsdctl/nfsdctl.h b/utils/nfsdctl/nfsdctl.h new file mode 100644 index 000000000000..f0aa3ab862c2 --- /dev/null +++ b/utils/nfsdctl/nfsdctl.h @@ -0,0 +1,93 @@ +/* SPDX-License-Identifier: ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) */ +/* Do not edit directly, auto-generated from: */ +/* Documentation/netlink/specs/nfsd.yaml */ +/* YNL-GEN uapi header */ + +#ifndef _UTILS_NFSDCTL_NFSDCTL_H +#define _UTILS_NFSDCTL_NFSDCTL_H + +enum nfs_opnum4 { + OP_ACCESS = 3, + OP_CLOSE = 4, + OP_COMMIT = 5, + OP_CREATE = 6, + OP_DELEGPURGE = 7, + OP_DELEGRETURN = 8, + OP_GETATTR = 9, + OP_GETFH = 10, + OP_LINK = 11, + OP_LOCK = 12, + OP_LOCKT = 13, + OP_LOCKU = 14, + OP_LOOKUP = 15, + OP_LOOKUPP = 16, + OP_NVERIFY = 17, + OP_OPEN = 18, + OP_OPENATTR = 19, + OP_OPEN_CONFIRM = 20, + OP_OPEN_DOWNGRADE = 21, + OP_PUTFH = 22, + OP_PUTPUBFH = 23, + OP_PUTROOTFH = 24, + OP_READ = 25, + OP_READDIR = 26, + OP_READLINK = 27, + OP_REMOVE = 28, + OP_RENAME = 29, + OP_RENEW = 30, + OP_RESTOREFH = 31, + OP_SAVEFH = 32, + OP_SECINFO = 33, + OP_SETATTR = 34, + OP_SETCLIENTID = 35, + OP_SETCLIENTID_CONFIRM = 36, + OP_VERIFY = 37, + OP_WRITE = 38, + OP_RELEASE_LOCKOWNER = 39, + + /* nfs41 */ + OP_BACKCHANNEL_CTL = 40, + OP_BIND_CONN_TO_SESSION = 41, + OP_EXCHANGE_ID = 42, + OP_CREATE_SESSION = 43, + OP_DESTROY_SESSION = 44, + OP_FREE_STATEID = 45, + OP_GET_DIR_DELEGATION = 46, + OP_GETDEVICEINFO = 47, + OP_GETDEVICELIST = 48, + OP_LAYOUTCOMMIT = 49, + OP_LAYOUTGET = 50, + OP_LAYOUTRETURN = 51, + OP_SECINFO_NO_NAME = 52, + OP_SEQUENCE = 53, + OP_SET_SSV = 54, + OP_TEST_STATEID = 55, + OP_WANT_DELEGATION = 56, + OP_DESTROY_CLIENTID = 57, + OP_RECLAIM_COMPLETE = 58, + + /* nfs42 */ + OP_ALLOCATE = 59, + OP_COPY = 60, + OP_COPY_NOTIFY = 61, + OP_DEALLOCATE = 62, + OP_IO_ADVISE = 63, + OP_LAYOUTERROR = 64, + OP_LAYOUTSTATS = 65, + OP_OFFLOAD_CANCEL = 66, + OP_OFFLOAD_STATUS = 67, + OP_READ_PLUS = 68, + OP_SEEK = 69, + OP_WRITE_SAME = 70, + OP_CLONE = 71, + + /* xattr support (RFC8726) */ + OP_GETXATTR = 72, + OP_SETXATTR = 73, + OP_LISTXATTRS = 74, + OP_REMOVEXATTR = 75, + + OP_ILLEGAL = 10044, +}; + +#endif /* _UTILS_NFSDCTL_NFSDCTL_H */ -- 2.45.2 ^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH nfs-utils v6 2/3] nfsdctl: asciidoc source for the manpage 2024-07-22 17:01 [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Jeff Layton 2024-07-22 17:01 ` [PATCH nfs-utils v6 1/3] nfsdctl: add the nfsdctl utility " Jeff Layton @ 2024-07-22 17:01 ` Jeff Layton 2024-07-22 17:01 ` [PATCH nfs-utils v6 3/3] systemd: use nfsdctl to start and stop the nfs server Jeff Layton ` (3 subsequent siblings) 5 siblings, 0 replies; 16+ messages in thread From: Jeff Layton @ 2024-07-22 17:01 UTC (permalink / raw) To: Steve Dickson Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi, Jeff Layton Convert to manpage with: asciidoctor -b manpage nfsdctl.adoc Signed-off-by: Jeff Layton <jlayton@kernel.org> --- utils/nfsdctl/nfsdctl.adoc | 158 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 158 insertions(+) diff --git a/utils/nfsdctl/nfsdctl.adoc b/utils/nfsdctl/nfsdctl.adoc new file mode 100644 index 000000000000..7119b44db6d2 --- /dev/null +++ b/utils/nfsdctl/nfsdctl.adoc @@ -0,0 +1,158 @@ +// Convert to manpage via: asciidoctor -b manpage nfsdctl.adoc += nfsdctl(8) +Jeff Layton +:doctype: manpage + +== NAME + +nfsdctl - control program for the Linux kernel NFS server + +== SYNOPSIS + +*nfsdctl* [ _OPTION_ ] COMMAND ... + +== DESCRIPTION + +nfsdctl is a control and query program for the in-kernel NFS server. There are several +subcommands (documented below) that allow the admin to configure or query different +aspects of the NFS server. + +To get information about different subcommand usage, pass the subcommand the +--help parameter. For example: + + nfsdctl listener --help + +== OPTIONS + +*-d, --debug*:: + Enable debug logging + +*-h, --help*:: + Print program help text + +*-V, --version*:: + Print program version + +== SUBCOMMANDS + +Each subcommand can also accept its own set of options and arguments. The +--help option is standard for all subcommands: + +*autostart*:: + Start the server using the settings in the [nfsd] section of /etc/nfs.conf. + This subcommand takes no arguments. Note that if a "threads=" value is not set in + nfs.conf, 16 server threads will be brought online. + +*listener*:: + + Get/set the listening sockets for the server. Run this without arguments to + get a list of the sockets on which the server is currently listening. To add + or remove sockets, pass it whitespace-separated strings in the format: + + { +|- }{ protocol }:{ address }:{ port } + + The fields are: + + + to add a listener, - to remove one + protocol: protocol name (e.g. tcp, udp, rdma) + address: hostname or address + port: port number or service name + + All fields are required, except for the address. If address is an empty string, + then the listeners will be opened for INADDR_ANY and IN6ADDR_ANY_INIT for ipv6 + (if enabled). The address can be either a hostname or an IP address. IPv4 + addresses must be in dotted-quad form. IPv6 addresses should be in standard + colon separated form, and must be enclosed in square brackets. + +*status*:: + + Get information about RPCs currently executing in the server. This subcommand + takes no arguments. + +*threads*:: + + Get/set the number of running nfsd threads in each pool. Pass a list of + integers to change the currently active number of threads. Passing it a + value of 0 will shut down the NFS server. Run this without arguments to + get the current number of running threads in each pool. + +*version*:: + + Get/set the enabled NFS versions for the server. Run without arguments to + get a list of supported versions and whether they are currently enabled or + disabled. To enable or disable a version, pass it a string in the format: + + { +|- }{ MAJOR }{.{ MINOR }} + + The fields are: + + + to enable a version, - to disable + MAJOR: the major version integer value + MINOR: the minor version integet value + + The minorversion field is optional. If not given, it will disable or enable + all minorversions for that major version. + +*pool-mode*:: + + Get/set the host's pool mode. This will cause the server to start threads + that are pinned to either the CPU or the NUMA node. This can only be set + when there are no nfsd threads running. + + The available options are: + global: single large pool + percpu: pool per CPU + pernode: pool per NUMA node + auto: choose a mode based on host configuration + +== EXAMPLES + +Start the server with the settings in nfs.conf: + + nfsdctl autostart + +Get a list of current listening sockets: + + nfsdctl listener + +Show the supported and enabled NFS versions: + + nfsdctl version + +Add TCP listener on all addresses (both v4 and v6), port 2049: + + nfsdctl listener +tcp::2049 + +Add RDMA listener on 1.2.3.4 port 20049: + + nfsdctl listener +rdma:1.2.3.4:20049 + +Add same listener on IPv6 address f00::ba4 port 20050: + + nfsdctl listener +rdma:[f00::ba4]:20050 + +Enable NFS version 3, disable v4.0: + + nfsdctl version +3 -4.0 + +Change the number of running threads in first pool to 256: + + nfsdctl threads 256 + +Set the pool-mode to "pernode": + + nfsctl pool-mode pernode + +== NOTES + +nfsdctl is intended to supersede rpc.nfsd(8), which controls the nfs server +using the files under /proc/fs/nfsd. nfsdctl instead uses a netlink(7) +interface to achieve the same goals. + +Most of the commands that query the NFS server can be run as an unprivileged +user, but configuring the server typically requires an account with elevated +privileges. + +== SEE ALSO + +nfs.conf(5), rpc.nfsd(8), rpc.mountd(8), exports(5), exportfs(8), nfs.conf(5), rpc.rquotad(8), nfsstat(8), netconfig(5) -- 2.45.2 ^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH nfs-utils v6 3/3] systemd: use nfsdctl to start and stop the nfs server 2024-07-22 17:01 [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Jeff Layton 2024-07-22 17:01 ` [PATCH nfs-utils v6 1/3] nfsdctl: add the nfsdctl utility " Jeff Layton 2024-07-22 17:01 ` [PATCH nfs-utils v6 2/3] nfsdctl: asciidoc source for the manpage Jeff Layton @ 2024-07-22 17:01 ` Jeff Layton 2024-07-26 19:40 ` [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Steve Dickson ` (2 subsequent siblings) 5 siblings, 0 replies; 16+ messages in thread From: Jeff Layton @ 2024-07-22 17:01 UTC (permalink / raw) To: Steve Dickson Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi, Jeff Layton Attempt to use nfsdctl to start and stop the nfs-server. If that fails for any reason, use rpc.nfsd to do it instead. Signed-off-by: Jeff Layton <jlayton@kernel.org> --- systemd/nfs-server.service | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/systemd/nfs-server.service b/systemd/nfs-server.service index ac17d5286d3b..8793856aa707 100644 --- a/systemd/nfs-server.service +++ b/systemd/nfs-server.service @@ -23,8 +23,8 @@ After=rpc-gssd.service gssproxy.service rpc-svcgssd.service Type=oneshot RemainAfterExit=yes ExecStartPre=-/usr/sbin/exportfs -r -ExecStart=/usr/sbin/rpc.nfsd -ExecStop=/usr/sbin/rpc.nfsd 0 +ExecStart=/bin/sh -c '/usr/sbin/nfsdctl autostart || /usr/sbin/rpc.nfsd' +ExecStop=/bin/sh -c '/usr/sbin/nfsdctl threads 0 || /usr/sbin/rpc.nfsd 0' ExecStopPost=/usr/sbin/exportfs -au ExecStopPost=/usr/sbin/exportfs -f -- 2.45.2 ^ permalink raw reply related [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-07-22 17:01 [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Jeff Layton ` (2 preceding siblings ...) 2024-07-22 17:01 ` [PATCH nfs-utils v6 3/3] systemd: use nfsdctl to start and stop the nfs server Jeff Layton @ 2024-07-26 19:40 ` Steve Dickson 2024-07-26 21:32 ` Jeff Layton 2024-09-30 15:52 ` Jeff Layton 2024-10-17 11:40 ` Steve Dickson 5 siblings, 1 reply; 16+ messages in thread From: Steve Dickson @ 2024-07-26 19:40 UTC (permalink / raw) To: Jeff Layton Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi Hey! On 7/22/24 1:01 PM, Jeff Layton wrote: > Hi Steve, > > Here's an squashed version of the nfsdctl patches, that represents > the latest changes. Let me know if you run into any other problems, > and thanks for helping to test this! > > Signed-off-by: Jeff Layton <jlayton@kernel.org> > --- > Changes in v6: > - make the default number of threads 16 in autostart > - doc updates > > Changes in v5: > - add support for pool-mode setting > - fix up the handling of nfsd_netlink.h in autoconf > - Link to v4: https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org > > Changes in v4: > - add ability to specify an array of pool thread counts in nfs.conf > - Link to v3: https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org > > Changes in v3: > - split nfsdctl.h so we can include the UAPI header as-is > - squash the patches together that added Lorenzo's version and convert > it to the new interface > - adapt to latest version of netlink interface changes > + have THREADS_SET/GET report an array of thread counts (one per pool) > + pass scope in as a string to THREADS_SET instead of using unshare() trick > > Changes in v2: > - Adapt to latest kernel netlink interface changes (in particular, send > the leastime and gracetime when they are set in the config). > - More help text for different subcommands > - New nfsdctl(8) manpage > - Patch to make systemd preferentially use nfsdctl instead of rpc.nfsd > - Link to v1: https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org > > --- > Jeff Layton (3): > nfsdctl: add the nfsdctl utility to nfs-utils > nfsdctl: asciidoc source for the manpage > systemd: use nfsdctl to start and stop the nfs server > > configure.ac | 19 + > systemd/nfs-server.service | 4 +- > utils/Makefile.am | 4 + > utils/nfsdctl/Makefile.am | 13 + > utils/nfsdctl/nfsd_netlink.h | 96 +++ > utils/nfsdctl/nfsdctl.8 | 304 ++++++++ > utils/nfsdctl/nfsdctl.adoc | 158 +++++ > utils/nfsdctl/nfsdctl.c | 1570 ++++++++++++++++++++++++++++++++++++++++++ > utils/nfsdctl/nfsdctl.h | 93 +++ > 9 files changed, 2259 insertions(+), 2 deletions(-) > --- > base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be > change-id: 20240412-nfsdctl-fa8bd8430cfd The patches apply very cleaning and thank you for squashing them down... but... bring up the NFS server with 'nfsdctl autostart' v3 is not being registered with rpcbind which means v3 mount will not work. Just curious are you trying support my idea of deprecating V3 :-) (That's a joke!) steved. ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-07-26 19:40 ` [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Steve Dickson @ 2024-07-26 21:32 ` Jeff Layton 2024-07-26 21:59 ` Steve Dickson 0 siblings, 1 reply; 16+ messages in thread From: Jeff Layton @ 2024-07-26 21:32 UTC (permalink / raw) To: Steve Dickson Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On Fri, 2024-07-26 at 15:40 -0400, Steve Dickson wrote: > Hey! > > On 7/22/24 1:01 PM, Jeff Layton wrote: > > Hi Steve, > > > > Here's an squashed version of the nfsdctl patches, that represents > > the latest changes. Let me know if you run into any other problems, > > and thanks for helping to test this! > > > > Signed-off-by: Jeff Layton <jlayton@kernel.org> > > --- > > Changes in v6: > > - make the default number of threads 16 in autostart > > - doc updates > > > > Changes in v5: > > - add support for pool-mode setting > > - fix up the handling of nfsd_netlink.h in autoconf > > - Link to v4: > > https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org > > > > Changes in v4: > > - add ability to specify an array of pool thread counts in nfs.conf > > - Link to v3: > > https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org > > > > Changes in v3: > > - split nfsdctl.h so we can include the UAPI header as-is > > - squash the patches together that added Lorenzo's version and > > convert > > it to the new interface > > - adapt to latest version of netlink interface changes > > + have THREADS_SET/GET report an array of thread counts (one per > > pool) > > + pass scope in as a string to THREADS_SET instead of using > > unshare() trick > > > > Changes in v2: > > - Adapt to latest kernel netlink interface changes (in particular, > > send > > the leastime and gracetime when they are set in the config). > > - More help text for different subcommands > > - New nfsdctl(8) manpage > > - Patch to make systemd preferentially use nfsdctl instead of > > rpc.nfsd > > - Link to v1: > > https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org > > > > --- > > Jeff Layton (3): > > nfsdctl: add the nfsdctl utility to nfs-utils > > nfsdctl: asciidoc source for the manpage > > systemd: use nfsdctl to start and stop the nfs server > > > > configure.ac | 19 + > > systemd/nfs-server.service | 4 +- > > utils/Makefile.am | 4 + > > utils/nfsdctl/Makefile.am | 13 + > > utils/nfsdctl/nfsd_netlink.h | 96 +++ > > utils/nfsdctl/nfsdctl.8 | 304 ++++++++ > > utils/nfsdctl/nfsdctl.adoc | 158 +++++ > > utils/nfsdctl/nfsdctl.c | 1570 > > ++++++++++++++++++++++++++++++++++++++++++ > > utils/nfsdctl/nfsdctl.h | 93 +++ > > 9 files changed, 2259 insertions(+), 2 deletions(-) > > --- > > base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be > > change-id: 20240412-nfsdctl-fa8bd8430cfd > > The patches apply very cleaning and thank you > for squashing them down... but... bring up the > NFS server with 'nfsdctl autostart' v3 is not > being registered with rpcbind which means > v3 mount will not work. > > Just curious are you trying support my > idea of deprecating V3 :-) (That's a joke!) > > steved. > You do need a patched kernel for this: https://lore.kernel.org/linux-nfs/Zp5j2DW+2BNaIPif@tissot.1015granger.net/T/#e675642639c59b1c0070f4b19cd03b89cff7983ba With a patched kernel, I get this with autostart: [kdevops@kdevops-nfsd ~]$ rpcinfo -p program vers proto port service 100000 4 tcp 111 portmapper 100000 3 tcp 111 portmapper 100000 2 tcp 111 portmapper 100000 4 udp 111 portmapper 100000 3 udp 111 portmapper 100000 2 udp 111 portmapper 100024 1 udp 42104 status 100024 1 tcp 40159 status 100003 3 udp 2049 nfs 100227 3 udp 2049 nfs_acl 100003 3 tcp 2049 nfs 100003 4 tcp 2049 nfs 100227 3 tcp 2049 nfs_acl 100021 1 udp 46387 nlockmgr 100021 3 udp 46387 nlockmgr 100021 4 udp 46387 nlockmgr 100021 1 tcp 36565 nlockmgr 100021 3 tcp 36565 nlockmgr 100021 4 tcp 36565 nlockmgr Are you seeing different results? -- Jeff Layton <jlayton@kernel.org> ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-07-26 21:32 ` Jeff Layton @ 2024-07-26 21:59 ` Steve Dickson 2024-07-26 22:20 ` Jeff Layton 0 siblings, 1 reply; 16+ messages in thread From: Steve Dickson @ 2024-07-26 21:59 UTC (permalink / raw) To: Jeff Layton Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On 7/26/24 5:32 PM, Jeff Layton wrote: > On Fri, 2024-07-26 at 15:40 -0400, Steve Dickson wrote: >> Hey! >> >> On 7/22/24 1:01 PM, Jeff Layton wrote: >>> Hi Steve, >>> >>> Here's an squashed version of the nfsdctl patches, that represents >>> the latest changes. Let me know if you run into any other problems, >>> and thanks for helping to test this! >>> >>> Signed-off-by: Jeff Layton <jlayton@kernel.org> >>> --- >>> Changes in v6: >>> - make the default number of threads 16 in autostart >>> - doc updates >>> >>> Changes in v5: >>> - add support for pool-mode setting >>> - fix up the handling of nfsd_netlink.h in autoconf >>> - Link to v4: >>> https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org >>> >>> Changes in v4: >>> - add ability to specify an array of pool thread counts in nfs.conf >>> - Link to v3: >>> https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org >>> >>> Changes in v3: >>> - split nfsdctl.h so we can include the UAPI header as-is >>> - squash the patches together that added Lorenzo's version and >>> convert >>> it to the new interface >>> - adapt to latest version of netlink interface changes >>> + have THREADS_SET/GET report an array of thread counts (one per >>> pool) >>> + pass scope in as a string to THREADS_SET instead of using >>> unshare() trick >>> >>> Changes in v2: >>> - Adapt to latest kernel netlink interface changes (in particular, >>> send >>> the leastime and gracetime when they are set in the config). >>> - More help text for different subcommands >>> - New nfsdctl(8) manpage >>> - Patch to make systemd preferentially use nfsdctl instead of >>> rpc.nfsd >>> - Link to v1: >>> https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org >>> >>> --- >>> Jeff Layton (3): >>> nfsdctl: add the nfsdctl utility to nfs-utils >>> nfsdctl: asciidoc source for the manpage >>> systemd: use nfsdctl to start and stop the nfs server >>> >>> configure.ac | 19 + >>> systemd/nfs-server.service | 4 +- >>> utils/Makefile.am | 4 + >>> utils/nfsdctl/Makefile.am | 13 + >>> utils/nfsdctl/nfsd_netlink.h | 96 +++ >>> utils/nfsdctl/nfsdctl.8 | 304 ++++++++ >>> utils/nfsdctl/nfsdctl.adoc | 158 +++++ >>> utils/nfsdctl/nfsdctl.c | 1570 >>> ++++++++++++++++++++++++++++++++++++++++++ >>> utils/nfsdctl/nfsdctl.h | 93 +++ >>> 9 files changed, 2259 insertions(+), 2 deletions(-) >>> --- >>> base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be >>> change-id: 20240412-nfsdctl-fa8bd8430cfd >> >> The patches apply very cleaning and thank you >> for squashing them down... but... bring up the >> NFS server with 'nfsdctl autostart' v3 is not >> being registered with rpcbind which means >> v3 mount will not work. >> >> Just curious are you trying support my >> idea of deprecating V3 :-) (That's a joke!) >> >> steved. >> > > You do need a patched kernel for this: > > https://lore.kernel.org/linux-nfs/Zp5j2DW+2BNaIPif@tissot.1015granger.net/T/#e675642639c59b1c0070f4b19cd03b89cff7983ba > > With a patched kernel, I get this with autostart: > > [kdevops@kdevops-nfsd ~]$ rpcinfo -p > program vers proto port service > 100000 4 tcp 111 portmapper > 100000 3 tcp 111 portmapper > 100000 2 tcp 111 portmapper > 100000 4 udp 111 portmapper > 100000 3 udp 111 portmapper > 100000 2 udp 111 portmapper > 100024 1 udp 42104 status > 100024 1 tcp 40159 status > 100003 3 udp 2049 nfs > 100227 3 udp 2049 nfs_acl > 100003 3 tcp 2049 nfs > 100003 4 tcp 2049 nfs > 100227 3 tcp 2049 nfs_acl > 100021 1 udp 46387 nlockmgr > 100021 3 udp 46387 nlockmgr > 100021 4 udp 46387 nlockmgr > 100021 1 tcp 36565 nlockmgr > 100021 3 tcp 36565 nlockmgr > 100021 4 tcp 36565 nlockmgr > > > Are you seeing different results? Yup uname -r 6.11.0-0.rc0.20240724git786c8248dbd3.12.fc41.x86_64 (rawhide) rpcinfo -p program vers proto port service 100000 4 tcp 111 portmapper 100000 3 tcp 111 portmapper 100000 2 tcp 111 portmapper 100000 4 udp 111 portmapper 100000 3 udp 111 portmapper 100000 2 udp 111 portmapper 100005 1 udp 20048 mountd 100005 1 tcp 20048 mountd 100024 1 udp 38596 status 100024 1 tcp 60257 status 100005 2 udp 20048 mountd 100005 2 tcp 20048 mountd 100005 3 udp 20048 mountd 100005 3 tcp 20048 mountd 100021 1 udp 55966 nlockmgr 100021 3 udp 55966 nlockmgr 100021 4 udp 55966 nlockmgr 100021 1 tcp 40995 nlockmgr 100021 3 tcp 40995 nlockmgr 100021 4 tcp 40995 nlockmgr # mount -o v3 fedora:/home/tmp /mnt/tmp mount.nfs: requested NFS version or transport protocol is not supported for /mnt/tmp steved. ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-07-26 21:59 ` Steve Dickson @ 2024-07-26 22:20 ` Jeff Layton 2024-07-27 11:28 ` Steve Dickson 2024-07-27 16:03 ` Chuck Lever III 0 siblings, 2 replies; 16+ messages in thread From: Jeff Layton @ 2024-07-26 22:20 UTC (permalink / raw) To: Steve Dickson Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On Fri, 2024-07-26 at 17:59 -0400, Steve Dickson wrote: > > On 7/26/24 5:32 PM, Jeff Layton wrote: > > On Fri, 2024-07-26 at 15:40 -0400, Steve Dickson wrote: > > > Hey! > > > > > > On 7/22/24 1:01 PM, Jeff Layton wrote: > > > > Hi Steve, > > > > > > > > Here's an squashed version of the nfsdctl patches, that represents > > > > the latest changes. Let me know if you run into any other problems, > > > > and thanks for helping to test this! > > > > > > > > Signed-off-by: Jeff Layton <jlayton@kernel.org> > > > > --- > > > > Changes in v6: > > > > - make the default number of threads 16 in autostart > > > > - doc updates > > > > > > > > Changes in v5: > > > > - add support for pool-mode setting > > > > - fix up the handling of nfsd_netlink.h in autoconf > > > > - Link to v4: > > > > https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org > > > > > > > > Changes in v4: > > > > - add ability to specify an array of pool thread counts in nfs.conf > > > > - Link to v3: > > > > https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org > > > > > > > > Changes in v3: > > > > - split nfsdctl.h so we can include the UAPI header as-is > > > > - squash the patches together that added Lorenzo's version and > > > > convert > > > > it to the new interface > > > > - adapt to latest version of netlink interface changes > > > > + have THREADS_SET/GET report an array of thread counts (one per > > > > pool) > > > > + pass scope in as a string to THREADS_SET instead of using > > > > unshare() trick > > > > > > > > Changes in v2: > > > > - Adapt to latest kernel netlink interface changes (in particular, > > > > send > > > > the leastime and gracetime when they are set in the config). > > > > - More help text for different subcommands > > > > - New nfsdctl(8) manpage > > > > - Patch to make systemd preferentially use nfsdctl instead of > > > > rpc.nfsd > > > > - Link to v1: > > > > https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org > > > > > > > > --- > > > > Jeff Layton (3): > > > > nfsdctl: add the nfsdctl utility to nfs-utils > > > > nfsdctl: asciidoc source for the manpage > > > > systemd: use nfsdctl to start and stop the nfs server > > > > > > > > configure.ac | 19 + > > > > systemd/nfs-server.service | 4 +- > > > > utils/Makefile.am | 4 + > > > > utils/nfsdctl/Makefile.am | 13 + > > > > utils/nfsdctl/nfsd_netlink.h | 96 +++ > > > > utils/nfsdctl/nfsdctl.8 | 304 ++++++++ > > > > utils/nfsdctl/nfsdctl.adoc | 158 +++++ > > > > utils/nfsdctl/nfsdctl.c | 1570 > > > > ++++++++++++++++++++++++++++++++++++++++++ > > > > utils/nfsdctl/nfsdctl.h | 93 +++ > > > > 9 files changed, 2259 insertions(+), 2 deletions(-) > > > > --- > > > > base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be > > > > change-id: 20240412-nfsdctl-fa8bd8430cfd > > > > > > The patches apply very cleaning and thank you > > > for squashing them down... but... bring up the > > > NFS server with 'nfsdctl autostart' v3 is not > > > being registered with rpcbind which means > > > v3 mount will not work. > > > > > > Just curious are you trying support my > > > idea of deprecating V3 :-) (That's a joke!) > > > > > > steved. > > > > > > > You do need a patched kernel for this: > > > > https://lore.kernel.org/linux-nfs/Zp5j2DW+2BNaIPif@tissot.1015granger.net/T/#e675642639c59b1c0070f4b19cd03b89cff7983ba > > > > With a patched kernel, I get this with autostart: > > > > [kdevops@kdevops-nfsd ~]$ rpcinfo -p > > program vers proto port service > > 100000 4 tcp 111 portmapper > > 100000 3 tcp 111 portmapper > > 100000 2 tcp 111 portmapper > > 100000 4 udp 111 portmapper > > 100000 3 udp 111 portmapper > > 100000 2 udp 111 portmapper > > 100024 1 udp 42104 status > > 100024 1 tcp 40159 status > > 100003 3 udp 2049 nfs > > 100227 3 udp 2049 nfs_acl > > 100003 3 tcp 2049 nfs > > 100003 4 tcp 2049 nfs > > 100227 3 tcp 2049 nfs_acl > > 100021 1 udp 46387 nlockmgr > > 100021 3 udp 46387 nlockmgr > > 100021 4 udp 46387 nlockmgr > > 100021 1 tcp 36565 nlockmgr > > 100021 3 tcp 36565 nlockmgr > > 100021 4 tcp 36565 nlockmgr > > > > > > Are you seeing different results? > Yup > uname -r > 6.11.0-0.rc0.20240724git786c8248dbd3.12.fc41.x86_64 (rawhide) > Did you patch that kernel by hand then? AFAICT, that git hash doesn't have the necessary fix. I don't think Chuck has sent a PR to Linus for it just yet. > rpcinfo -p > program vers proto port service > 100000 4 tcp 111 portmapper > 100000 3 tcp 111 portmapper > 100000 2 tcp 111 portmapper > 100000 4 udp 111 portmapper > 100000 3 udp 111 portmapper > 100000 2 udp 111 portmapper > 100005 1 udp 20048 mountd > 100005 1 tcp 20048 mountd > 100024 1 udp 38596 status > 100024 1 tcp 60257 status > 100005 2 udp 20048 mountd > 100005 2 tcp 20048 mountd > 100005 3 udp 20048 mountd > 100005 3 tcp 20048 mountd > 100021 1 udp 55966 nlockmgr > 100021 3 udp 55966 nlockmgr > 100021 4 udp 55966 nlockmgr > 100021 1 tcp 40995 nlockmgr > 100021 3 tcp 40995 nlockmgr > 100021 4 tcp 40995 nlockmgr > > # mount -o v3 fedora:/home/tmp /mnt/tmp > mount.nfs: requested NFS version or transport protocol is not supported > for /mnt/tmp > -- Jeff Layton <jlayton@kernel.org> ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-07-26 22:20 ` Jeff Layton @ 2024-07-27 11:28 ` Steve Dickson 2024-07-27 16:03 ` Chuck Lever III 1 sibling, 0 replies; 16+ messages in thread From: Steve Dickson @ 2024-07-27 11:28 UTC (permalink / raw) To: Jeff Layton Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On 7/26/24 6:20 PM, Jeff Layton wrote: > On Fri, 2024-07-26 at 17:59 -0400, Steve Dickson wrote: >> >> On 7/26/24 5:32 PM, Jeff Layton wrote: >>> On Fri, 2024-07-26 at 15:40 -0400, Steve Dickson wrote: >>>> Hey! >>>> >>>> On 7/22/24 1:01 PM, Jeff Layton wrote: >>>>> Hi Steve, >>>>> >>>>> Here's an squashed version of the nfsdctl patches, that represents >>>>> the latest changes. Let me know if you run into any other problems, >>>>> and thanks for helping to test this! >>>>> >>>>> Signed-off-by: Jeff Layton <jlayton@kernel.org> >>>>> --- >>>>> Changes in v6: >>>>> - make the default number of threads 16 in autostart >>>>> - doc updates >>>>> >>>>> Changes in v5: >>>>> - add support for pool-mode setting >>>>> - fix up the handling of nfsd_netlink.h in autoconf >>>>> - Link to v4: >>>>> https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org >>>>> >>>>> Changes in v4: >>>>> - add ability to specify an array of pool thread counts in nfs.conf >>>>> - Link to v3: >>>>> https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org >>>>> >>>>> Changes in v3: >>>>> - split nfsdctl.h so we can include the UAPI header as-is >>>>> - squash the patches together that added Lorenzo's version and >>>>> convert >>>>> it to the new interface >>>>> - adapt to latest version of netlink interface changes >>>>> + have THREADS_SET/GET report an array of thread counts (one per >>>>> pool) >>>>> + pass scope in as a string to THREADS_SET instead of using >>>>> unshare() trick >>>>> >>>>> Changes in v2: >>>>> - Adapt to latest kernel netlink interface changes (in particular, >>>>> send >>>>> the leastime and gracetime when they are set in the config). >>>>> - More help text for different subcommands >>>>> - New nfsdctl(8) manpage >>>>> - Patch to make systemd preferentially use nfsdctl instead of >>>>> rpc.nfsd >>>>> - Link to v1: >>>>> https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org >>>>> >>>>> --- >>>>> Jeff Layton (3): >>>>> nfsdctl: add the nfsdctl utility to nfs-utils >>>>> nfsdctl: asciidoc source for the manpage >>>>> systemd: use nfsdctl to start and stop the nfs server >>>>> >>>>> configure.ac | 19 + >>>>> systemd/nfs-server.service | 4 +- >>>>> utils/Makefile.am | 4 + >>>>> utils/nfsdctl/Makefile.am | 13 + >>>>> utils/nfsdctl/nfsd_netlink.h | 96 +++ >>>>> utils/nfsdctl/nfsdctl.8 | 304 ++++++++ >>>>> utils/nfsdctl/nfsdctl.adoc | 158 +++++ >>>>> utils/nfsdctl/nfsdctl.c | 1570 >>>>> ++++++++++++++++++++++++++++++++++++++++++ >>>>> utils/nfsdctl/nfsdctl.h | 93 +++ >>>>> 9 files changed, 2259 insertions(+), 2 deletions(-) >>>>> --- >>>>> base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be >>>>> change-id: 20240412-nfsdctl-fa8bd8430cfd >>>> >>>> The patches apply very cleaning and thank you >>>> for squashing them down... but... bring up the >>>> NFS server with 'nfsdctl autostart' v3 is not >>>> being registered with rpcbind which means >>>> v3 mount will not work. >>>> >>>> Just curious are you trying support my >>>> idea of deprecating V3 :-) (That's a joke!) >>>> >>>> steved. >>>> >>> >>> You do need a patched kernel for this: >>> >>> https://lore.kernel.org/linux-nfs/Zp5j2DW+2BNaIPif@tissot.1015granger.net/T/#e675642639c59b1c0070f4b19cd03b89cff7983ba >>> >>> With a patched kernel, I get this with autostart: >>> >>> [kdevops@kdevops-nfsd ~]$ rpcinfo -p >>> program vers proto port service >>> 100000 4 tcp 111 portmapper >>> 100000 3 tcp 111 portmapper >>> 100000 2 tcp 111 portmapper >>> 100000 4 udp 111 portmapper >>> 100000 3 udp 111 portmapper >>> 100000 2 udp 111 portmapper >>> 100024 1 udp 42104 status >>> 100024 1 tcp 40159 status >>> 100003 3 udp 2049 nfs >>> 100227 3 udp 2049 nfs_acl >>> 100003 3 tcp 2049 nfs >>> 100003 4 tcp 2049 nfs >>> 100227 3 tcp 2049 nfs_acl >>> 100021 1 udp 46387 nlockmgr >>> 100021 3 udp 46387 nlockmgr >>> 100021 4 udp 46387 nlockmgr >>> 100021 1 tcp 36565 nlockmgr >>> 100021 3 tcp 36565 nlockmgr >>> 100021 4 tcp 36565 nlockmgr >>> >>> >>> Are you seeing different results? >> Yup >> uname -r >> 6.11.0-0.rc0.20240724git786c8248dbd3.12.fc41.x86_64 (rawhide) >> > > Did you patch that kernel by hand then? AFAICT, that git hash doesn't > have the necessary fix. I don't think Chuck has sent a PR to Linus for > it just yet. No I did not... Due to the fact I can not commit theses patches until the kernel patch land into the distros. steved. ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-07-26 22:20 ` Jeff Layton 2024-07-27 11:28 ` Steve Dickson @ 2024-07-27 16:03 ` Chuck Lever III 1 sibling, 0 replies; 16+ messages in thread From: Chuck Lever III @ 2024-07-27 16:03 UTC (permalink / raw) To: Jeff Layton, Steve Dickson Cc: Linux NFS Mailing List, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Lorenzo Bianconi > On Jul 26, 2024, at 6:20 PM, Jeff Layton <jlayton@kernel.org> wrote: > > On Fri, 2024-07-26 at 17:59 -0400, Steve Dickson wrote: >> >> On 7/26/24 5:32 PM, Jeff Layton wrote: >>> On Fri, 2024-07-26 at 15:40 -0400, Steve Dickson wrote: >>>> Hey! >>>> >>>> On 7/22/24 1:01 PM, Jeff Layton wrote: >>>>> Hi Steve, >>>>> >>>>> Here's an squashed version of the nfsdctl patches, that represents >>>>> the latest changes. Let me know if you run into any other problems, >>>>> and thanks for helping to test this! >>>>> >>>>> Signed-off-by: Jeff Layton <jlayton@kernel.org> >>>>> --- >>>>> Changes in v6: >>>>> - make the default number of threads 16 in autostart >>>>> - doc updates >>>>> >>>>> Changes in v5: >>>>> - add support for pool-mode setting >>>>> - fix up the handling of nfsd_netlink.h in autoconf >>>>> - Link to v4: >>>>> https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org >>>>> >>>>> Changes in v4: >>>>> - add ability to specify an array of pool thread counts in nfs.conf >>>>> - Link to v3: >>>>> https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org >>>>> >>>>> Changes in v3: >>>>> - split nfsdctl.h so we can include the UAPI header as-is >>>>> - squash the patches together that added Lorenzo's version and >>>>> convert >>>>> it to the new interface >>>>> - adapt to latest version of netlink interface changes >>>>> + have THREADS_SET/GET report an array of thread counts (one per >>>>> pool) >>>>> + pass scope in as a string to THREADS_SET instead of using >>>>> unshare() trick >>>>> >>>>> Changes in v2: >>>>> - Adapt to latest kernel netlink interface changes (in particular, >>>>> send >>>>> the leastime and gracetime when they are set in the config). >>>>> - More help text for different subcommands >>>>> - New nfsdctl(8) manpage >>>>> - Patch to make systemd preferentially use nfsdctl instead of >>>>> rpc.nfsd >>>>> - Link to v1: >>>>> https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org >>>>> >>>>> --- >>>>> Jeff Layton (3): >>>>> nfsdctl: add the nfsdctl utility to nfs-utils >>>>> nfsdctl: asciidoc source for the manpage >>>>> systemd: use nfsdctl to start and stop the nfs server >>>>> >>>>> configure.ac | 19 + >>>>> systemd/nfs-server.service | 4 +- >>>>> utils/Makefile.am | 4 + >>>>> utils/nfsdctl/Makefile.am | 13 + >>>>> utils/nfsdctl/nfsd_netlink.h | 96 +++ >>>>> utils/nfsdctl/nfsdctl.8 | 304 ++++++++ >>>>> utils/nfsdctl/nfsdctl.adoc | 158 +++++ >>>>> utils/nfsdctl/nfsdctl.c | 1570 >>>>> ++++++++++++++++++++++++++++++++++++++++++ >>>>> utils/nfsdctl/nfsdctl.h | 93 +++ >>>>> 9 files changed, 2259 insertions(+), 2 deletions(-) >>>>> --- >>>>> base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be >>>>> change-id: 20240412-nfsdctl-fa8bd8430cfd >>>> >>>> The patches apply very cleaning and thank you >>>> for squashing them down... but... bring up the >>>> NFS server with 'nfsdctl autostart' v3 is not >>>> being registered with rpcbind which means >>>> v3 mount will not work. >>>> >>>> Just curious are you trying support my >>>> idea of deprecating V3 :-) (That's a joke!) >>>> >>>> steved. >>>> >>> >>> You do need a patched kernel for this: >>> >>> https://lore.kernel.org/linux-nfs/Zp5j2DW+2BNaIPif@tissot.1015granger.net/T/#e675642639c59b1c0070f4b19cd03b89cff7983ba >>> >>> With a patched kernel, I get this with autostart: >>> >>> [kdevops@kdevops-nfsd ~]$ rpcinfo -p >>> program vers proto port service >>> 100000 4 tcp 111 portmapper >>> 100000 3 tcp 111 portmapper >>> 100000 2 tcp 111 portmapper >>> 100000 4 udp 111 portmapper >>> 100000 3 udp 111 portmapper >>> 100000 2 udp 111 portmapper >>> 100024 1 udp 42104 status >>> 100024 1 tcp 40159 status >>> 100003 3 udp 2049 nfs >>> 100227 3 udp 2049 nfs_acl >>> 100003 3 tcp 2049 nfs >>> 100003 4 tcp 2049 nfs >>> 100227 3 tcp 2049 nfs_acl >>> 100021 1 udp 46387 nlockmgr >>> 100021 3 udp 46387 nlockmgr >>> 100021 4 udp 46387 nlockmgr >>> 100021 1 tcp 36565 nlockmgr >>> 100021 3 tcp 36565 nlockmgr >>> 100021 4 tcp 36565 nlockmgr >>> >>> >>> Are you seeing different results? >> Yup >> uname -r >> 6.11.0-0.rc0.20240724git786c8248dbd3.12.fc41.x86_64 (rawhide) > > Did you patch that kernel by hand then? AFAICT, that git hash doesn't > have the necessary fix. I don't think Chuck has sent a PR to Linus for > it just yet. "nfsd: don't set SVC_SOCK_ANONYMOUS when creating nfsd sockets" is queued up for v6.11-rc. -- Chuck Lever ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-07-22 17:01 [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Jeff Layton ` (3 preceding siblings ...) 2024-07-26 19:40 ` [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Steve Dickson @ 2024-09-30 15:52 ` Jeff Layton 2024-09-30 16:16 ` Steve Dickson 2024-10-17 11:40 ` Steve Dickson 5 siblings, 1 reply; 16+ messages in thread From: Jeff Layton @ 2024-09-30 15:52 UTC (permalink / raw) To: Steve Dickson Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On Mon, 2024-07-22 at 13:01 -0400, Jeff Layton wrote: > Hi Steve, > > Here's an squashed version of the nfsdctl patches, that represents > the latest changes. Let me know if you run into any other problems, > and thanks for helping to test this! > > Signed-off-by: Jeff Layton <jlayton@kernel.org> > --- > Changes in v6: > - make the default number of threads 16 in autostart > - doc updates > > Changes in v5: > - add support for pool-mode setting > - fix up the handling of nfsd_netlink.h in autoconf > - Link to v4: https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org > > Changes in v4: > - add ability to specify an array of pool thread counts in nfs.conf > - Link to v3: https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org > > Changes in v3: > - split nfsdctl.h so we can include the UAPI header as-is > - squash the patches together that added Lorenzo's version and convert > it to the new interface > - adapt to latest version of netlink interface changes > + have THREADS_SET/GET report an array of thread counts (one per pool) > + pass scope in as a string to THREADS_SET instead of using unshare() trick > > Changes in v2: > - Adapt to latest kernel netlink interface changes (in particular, send > the leastime and gracetime when they are set in the config). > - More help text for different subcommands > - New nfsdctl(8) manpage > - Patch to make systemd preferentially use nfsdctl instead of rpc.nfsd > - Link to v1: https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org > > --- > Jeff Layton (3): > nfsdctl: add the nfsdctl utility to nfs-utils > nfsdctl: asciidoc source for the manpage > systemd: use nfsdctl to start and stop the nfs server > > configure.ac | 19 + > systemd/nfs-server.service | 4 +- > utils/Makefile.am | 4 + > utils/nfsdctl/Makefile.am | 13 + > utils/nfsdctl/nfsd_netlink.h | 96 +++ > utils/nfsdctl/nfsdctl.8 | 304 ++++++++ > utils/nfsdctl/nfsdctl.adoc | 158 +++++ > utils/nfsdctl/nfsdctl.c | 1570 ++++++++++++++++++++++++++++++++++++++++++ > utils/nfsdctl/nfsdctl.h | 93 +++ > 9 files changed, 2259 insertions(+), 2 deletions(-) > --- > base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be > change-id: 20240412-nfsdctl-fa8bd8430cfd > > Best regards, Ping? -- Jeff Layton <jlayton@kernel.org> ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-09-30 15:52 ` Jeff Layton @ 2024-09-30 16:16 ` Steve Dickson 2024-10-16 15:02 ` Jeff Layton 0 siblings, 1 reply; 16+ messages in thread From: Steve Dickson @ 2024-09-30 16:16 UTC (permalink / raw) To: Jeff Layton Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On 9/30/24 11:52 AM, Jeff Layton wrote: > On Mon, 2024-07-22 at 13:01 -0400, Jeff Layton wrote: >> Hi Steve, >> >> Here's an squashed version of the nfsdctl patches, that represents >> the latest changes. Let me know if you run into any other problems, >> and thanks for helping to test this! >> >> Signed-off-by: Jeff Layton <jlayton@kernel.org> >> --- >> Changes in v6: >> - make the default number of threads 16 in autostart >> - doc updates >> >> Changes in v5: >> - add support for pool-mode setting >> - fix up the handling of nfsd_netlink.h in autoconf >> - Link to v4: https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org >> >> Changes in v4: >> - add ability to specify an array of pool thread counts in nfs.conf >> - Link to v3: https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org >> >> Changes in v3: >> - split nfsdctl.h so we can include the UAPI header as-is >> - squash the patches together that added Lorenzo's version and convert >> it to the new interface >> - adapt to latest version of netlink interface changes >> + have THREADS_SET/GET report an array of thread counts (one per pool) >> + pass scope in as a string to THREADS_SET instead of using unshare() trick >> >> Changes in v2: >> - Adapt to latest kernel netlink interface changes (in particular, send >> the leastime and gracetime when they are set in the config). >> - More help text for different subcommands >> - New nfsdctl(8) manpage >> - Patch to make systemd preferentially use nfsdctl instead of rpc.nfsd >> - Link to v1: https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org >> >> --- >> Jeff Layton (3): >> nfsdctl: add the nfsdctl utility to nfs-utils >> nfsdctl: asciidoc source for the manpage >> systemd: use nfsdctl to start and stop the nfs server >> >> configure.ac | 19 + >> systemd/nfs-server.service | 4 +- >> utils/Makefile.am | 4 + >> utils/nfsdctl/Makefile.am | 13 + >> utils/nfsdctl/nfsd_netlink.h | 96 +++ >> utils/nfsdctl/nfsdctl.8 | 304 ++++++++ >> utils/nfsdctl/nfsdctl.adoc | 158 +++++ >> utils/nfsdctl/nfsdctl.c | 1570 ++++++++++++++++++++++++++++++++++++++++++ >> utils/nfsdctl/nfsdctl.h | 93 +++ >> 9 files changed, 2259 insertions(+), 2 deletions(-) >> --- >> base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be >> change-id: 20240412-nfsdctl-fa8bd8430cfd >> >> Best regards, > > Ping? I'm keeping my eye on it.... I have the bits on a private branch... I just don't want to break v3... when the distros get to 6.11... I'll make it happen steved. ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-09-30 16:16 ` Steve Dickson @ 2024-10-16 15:02 ` Jeff Layton 2024-10-16 20:10 ` Steve Dickson 0 siblings, 1 reply; 16+ messages in thread From: Jeff Layton @ 2024-10-16 15:02 UTC (permalink / raw) To: Steve Dickson Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On Mon, 2024-09-30 at 12:16 -0400, Steve Dickson wrote: > > On 9/30/24 11:52 AM, Jeff Layton wrote: > > On Mon, 2024-07-22 at 13:01 -0400, Jeff Layton wrote: > > > Hi Steve, > > > > > > Here's an squashed version of the nfsdctl patches, that represents > > > the latest changes. Let me know if you run into any other problems, > > > and thanks for helping to test this! > > > > > > Signed-off-by: Jeff Layton <jlayton@kernel.org> > > > --- > > > Changes in v6: > > > - make the default number of threads 16 in autostart > > > - doc updates > > > > > > Changes in v5: > > > - add support for pool-mode setting > > > - fix up the handling of nfsd_netlink.h in autoconf > > > - Link to v4: https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org > > > > > > Changes in v4: > > > - add ability to specify an array of pool thread counts in nfs.conf > > > - Link to v3: https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org > > > > > > Changes in v3: > > > - split nfsdctl.h so we can include the UAPI header as-is > > > - squash the patches together that added Lorenzo's version and convert > > > it to the new interface > > > - adapt to latest version of netlink interface changes > > > + have THREADS_SET/GET report an array of thread counts (one per pool) > > > + pass scope in as a string to THREADS_SET instead of using unshare() trick > > > > > > Changes in v2: > > > - Adapt to latest kernel netlink interface changes (in particular, send > > > the leastime and gracetime when they are set in the config). > > > - More help text for different subcommands > > > - New nfsdctl(8) manpage > > > - Patch to make systemd preferentially use nfsdctl instead of rpc.nfsd > > > - Link to v1: https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org > > > > > > --- > > > Jeff Layton (3): > > > nfsdctl: add the nfsdctl utility to nfs-utils > > > nfsdctl: asciidoc source for the manpage > > > systemd: use nfsdctl to start and stop the nfs server > > > > > > configure.ac | 19 + > > > systemd/nfs-server.service | 4 +- > > > utils/Makefile.am | 4 + > > > utils/nfsdctl/Makefile.am | 13 + > > > utils/nfsdctl/nfsd_netlink.h | 96 +++ > > > utils/nfsdctl/nfsdctl.8 | 304 ++++++++ > > > utils/nfsdctl/nfsdctl.adoc | 158 +++++ > > > utils/nfsdctl/nfsdctl.c | 1570 ++++++++++++++++++++++++++++++++++++++++++ > > > utils/nfsdctl/nfsdctl.h | 93 +++ > > > 9 files changed, 2259 insertions(+), 2 deletions(-) > > > --- > > > base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be > > > change-id: 20240412-nfsdctl-fa8bd8430cfd > > > > > > Best regards, > > > > Ping? > I'm keeping my eye on it.... I have the bits on > a private branch... I just don't want to break > v3... when the distros get to 6.11... I'll > make it happen > Previously, you said you would merge it once rawhide had a sufficient kernel, which it has had for months. Fedora 40 is now shipping a v6.11 kernel. Can we get this in now? Thanks, -- Jeff Layton <jlayton@kernel.org> ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-10-16 15:02 ` Jeff Layton @ 2024-10-16 20:10 ` Steve Dickson 0 siblings, 0 replies; 16+ messages in thread From: Steve Dickson @ 2024-10-16 20:10 UTC (permalink / raw) To: Jeff Layton Cc: linux-nfs, Neil Brown, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On 10/16/24 11:02 AM, Jeff Layton wrote: > On Mon, 2024-09-30 at 12:16 -0400, Steve Dickson wrote: >> >> On 9/30/24 11:52 AM, Jeff Layton wrote: >>> On Mon, 2024-07-22 at 13:01 -0400, Jeff Layton wrote: >>>> Hi Steve, >>>> >>>> Here's an squashed version of the nfsdctl patches, that represents >>>> the latest changes. Let me know if you run into any other problems, >>>> and thanks for helping to test this! >>>> >>>> Signed-off-by: Jeff Layton <jlayton@kernel.org> >>>> --- >>>> Changes in v6: >>>> - make the default number of threads 16 in autostart >>>> - doc updates >>>> >>>> Changes in v5: >>>> - add support for pool-mode setting >>>> - fix up the handling of nfsd_netlink.h in autoconf >>>> - Link to v4: https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org >>>> >>>> Changes in v4: >>>> - add ability to specify an array of pool thread counts in nfs.conf >>>> - Link to v3: https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org >>>> >>>> Changes in v3: >>>> - split nfsdctl.h so we can include the UAPI header as-is >>>> - squash the patches together that added Lorenzo's version and convert >>>> it to the new interface >>>> - adapt to latest version of netlink interface changes >>>> + have THREADS_SET/GET report an array of thread counts (one per pool) >>>> + pass scope in as a string to THREADS_SET instead of using unshare() trick >>>> >>>> Changes in v2: >>>> - Adapt to latest kernel netlink interface changes (in particular, send >>>> the leastime and gracetime when they are set in the config). >>>> - More help text for different subcommands >>>> - New nfsdctl(8) manpage >>>> - Patch to make systemd preferentially use nfsdctl instead of rpc.nfsd >>>> - Link to v1: https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org >>>> >>>> --- >>>> Jeff Layton (3): >>>> nfsdctl: add the nfsdctl utility to nfs-utils >>>> nfsdctl: asciidoc source for the manpage >>>> systemd: use nfsdctl to start and stop the nfs server >>>> >>>> configure.ac | 19 + >>>> systemd/nfs-server.service | 4 +- >>>> utils/Makefile.am | 4 + >>>> utils/nfsdctl/Makefile.am | 13 + >>>> utils/nfsdctl/nfsd_netlink.h | 96 +++ >>>> utils/nfsdctl/nfsdctl.8 | 304 ++++++++ >>>> utils/nfsdctl/nfsdctl.adoc | 158 +++++ >>>> utils/nfsdctl/nfsdctl.c | 1570 ++++++++++++++++++++++++++++++++++++++++++ >>>> utils/nfsdctl/nfsdctl.h | 93 +++ >>>> 9 files changed, 2259 insertions(+), 2 deletions(-) >>>> --- >>>> base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be >>>> change-id: 20240412-nfsdctl-fa8bd8430cfd >>>> >>>> Best regards, >>> >>> Ping? >> I'm keeping my eye on it.... I have the bits on >> a private branch... I just don't want to break >> v3... when the distros get to 6.11... I'll >> make it happen >> > > Previously, you said you would merge it once rawhide had a sufficient > kernel, which it has had for months. Fedora 40 is now shipping a v6.11 > kernel. Can we get this in now? Very true... I did notice just f40 got to v6.11... so now I can make the release... just in time for Bakeathon. I just hoping the release will not break V3 (which it will) on other distros that are not at v6.11. steved. ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-07-22 17:01 [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Jeff Layton ` (4 preceding siblings ...) 2024-09-30 15:52 ` Jeff Layton @ 2024-10-17 11:40 ` Steve Dickson 2024-10-17 15:22 ` Jeff Layton 5 siblings, 1 reply; 16+ messages in thread From: Steve Dickson @ 2024-10-17 11:40 UTC (permalink / raw) To: Jeff Layton Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On 7/22/24 1:01 PM, Jeff Layton wrote: > Hi Steve, > > Here's an squashed version of the nfsdctl patches, that represents > the latest changes. Let me know if you run into any other problems, > and thanks for helping to test this! > > Signed-off-by: Jeff Layton <jlayton@kernel.org> > --- > Changes in v6: > - make the default number of threads 16 in autostart > - doc updates > > Changes in v5: > - add support for pool-mode setting > - fix up the handling of nfsd_netlink.h in autoconf > - Link to v4: https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org > > Changes in v4: > - add ability to specify an array of pool thread counts in nfs.conf > - Link to v3: https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org > > Changes in v3: > - split nfsdctl.h so we can include the UAPI header as-is > - squash the patches together that added Lorenzo's version and convert > it to the new interface > - adapt to latest version of netlink interface changes > + have THREADS_SET/GET report an array of thread counts (one per pool) > + pass scope in as a string to THREADS_SET instead of using unshare() trick > > Changes in v2: > - Adapt to latest kernel netlink interface changes (in particular, send > the leastime and gracetime when they are set in the config). > - More help text for different subcommands > - New nfsdctl(8) manpage > - Patch to make systemd preferentially use nfsdctl instead of rpc.nfsd > - Link to v1: https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org > > --- > Jeff Layton (3): > nfsdctl: add the nfsdctl utility to nfs-utils > nfsdctl: asciidoc source for the manpage > systemd: use nfsdctl to start and stop the nfs server > > configure.ac | 19 + > systemd/nfs-server.service | 4 +- > utils/Makefile.am | 4 + > utils/nfsdctl/Makefile.am | 13 + > utils/nfsdctl/nfsd_netlink.h | 96 +++ > utils/nfsdctl/nfsdctl.8 | 304 ++++++++ > utils/nfsdctl/nfsdctl.adoc | 158 +++++ > utils/nfsdctl/nfsdctl.c | 1570 ++++++++++++++++++++++++++++++++++++++++++ > utils/nfsdctl/nfsdctl.h | 93 +++ > 9 files changed, 2259 insertions(+), 2 deletions(-) > --- > base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be > change-id: 20240412-nfsdctl-fa8bd8430cfd > > Best regards, Committed... (tag: nfs-utils-2-7-2-rc2) Note: These changes need a v6.11 kernel for V3 to come up properly steved. ^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils 2024-10-17 11:40 ` Steve Dickson @ 2024-10-17 15:22 ` Jeff Layton 0 siblings, 0 replies; 16+ messages in thread From: Jeff Layton @ 2024-10-17 15:22 UTC (permalink / raw) To: Steve Dickson Cc: linux-nfs, Neil Brown, Olga Kornievskaia, Dai Ngo, Tom Talpey, Chuck Lever, Lorenzo Bianconi On Thu, 2024-10-17 at 07:40 -0400, Steve Dickson wrote: > > On 7/22/24 1:01 PM, Jeff Layton wrote: > > Hi Steve, > > > > Here's an squashed version of the nfsdctl patches, that represents > > the latest changes. Let me know if you run into any other problems, > > and thanks for helping to test this! > > > > Signed-off-by: Jeff Layton <jlayton@kernel.org> > > --- > > Changes in v6: > > - make the default number of threads 16 in autostart > > - doc updates > > > > Changes in v5: > > - add support for pool-mode setting > > - fix up the handling of nfsd_netlink.h in autoconf > > - Link to v4: https://lore.kernel.org/r/20240604-nfsdctl-v4-0-a2941f782e4c@kernel.org > > > > Changes in v4: > > - add ability to specify an array of pool thread counts in nfs.conf > > - Link to v3: https://lore.kernel.org/r/20240423-nfsdctl-v3-0-9e68181c846d@kernel.org > > > > Changes in v3: > > - split nfsdctl.h so we can include the UAPI header as-is > > - squash the patches together that added Lorenzo's version and convert > > it to the new interface > > - adapt to latest version of netlink interface changes > > + have THREADS_SET/GET report an array of thread counts (one per pool) > > + pass scope in as a string to THREADS_SET instead of using unshare() trick > > > > Changes in v2: > > - Adapt to latest kernel netlink interface changes (in particular, send > > the leastime and gracetime when they are set in the config). > > - More help text for different subcommands > > - New nfsdctl(8) manpage > > - Patch to make systemd preferentially use nfsdctl instead of rpc.nfsd > > - Link to v1: https://lore.kernel.org/r/20240412-nfsdctl-v1-0-efd6dcebcc04@kernel.org > > > > --- > > Jeff Layton (3): > > nfsdctl: add the nfsdctl utility to nfs-utils > > nfsdctl: asciidoc source for the manpage > > systemd: use nfsdctl to start and stop the nfs server > > > > configure.ac | 19 + > > systemd/nfs-server.service | 4 +- > > utils/Makefile.am | 4 + > > utils/nfsdctl/Makefile.am | 13 + > > utils/nfsdctl/nfsd_netlink.h | 96 +++ > > utils/nfsdctl/nfsdctl.8 | 304 ++++++++ > > utils/nfsdctl/nfsdctl.adoc | 158 +++++ > > utils/nfsdctl/nfsdctl.c | 1570 ++++++++++++++++++++++++++++++++++++++++++ > > utils/nfsdctl/nfsdctl.h | 93 +++ > > 9 files changed, 2259 insertions(+), 2 deletions(-) > > --- > > base-commit: b76dbaa48f7c239accb0c2d1e1d51ddd73f4d6be > > change-id: 20240412-nfsdctl-fa8bd8430cfd > > > > Best regards, > Committed... (tag: nfs-utils-2-7-2-rc2) > > Note: These changes need a v6.11 kernel for V3 to > come up properly > Thanks! To be clear: The "v3 problem" you're referring to was fixed in mainline by: 91da337e5d50 nfsd: don't set SVC_SOCK_ANONYMOUS when creating nfsd sockets That patch also went into v6.10.5 (as 04c2c4d836ff) as well. Earlier kernels don't need it since they lack the relevant netlink interfaces altogether. -- Jeff Layton <jlayton@kernel.org> ^ permalink raw reply [flat|nested] 16+ messages in thread
end of thread, other threads:[~2024-10-17 15:22 UTC | newest] Thread overview: 16+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2024-07-22 17:01 [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Jeff Layton 2024-07-22 17:01 ` [PATCH nfs-utils v6 1/3] nfsdctl: add the nfsdctl utility " Jeff Layton 2024-07-22 17:01 ` [PATCH nfs-utils v6 2/3] nfsdctl: asciidoc source for the manpage Jeff Layton 2024-07-22 17:01 ` [PATCH nfs-utils v6 3/3] systemd: use nfsdctl to start and stop the nfs server Jeff Layton 2024-07-26 19:40 ` [PATCH nfs-utils v6 0/3] nfsdctl: add a new nfsdctl tool to nfs-utils Steve Dickson 2024-07-26 21:32 ` Jeff Layton 2024-07-26 21:59 ` Steve Dickson 2024-07-26 22:20 ` Jeff Layton 2024-07-27 11:28 ` Steve Dickson 2024-07-27 16:03 ` Chuck Lever III 2024-09-30 15:52 ` Jeff Layton 2024-09-30 16:16 ` Steve Dickson 2024-10-16 15:02 ` Jeff Layton 2024-10-16 20:10 ` Steve Dickson 2024-10-17 11:40 ` Steve Dickson 2024-10-17 15:22 ` Jeff Layton
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox