| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
| |
Signed-off-by: Yang Bo <rslovers@yandex.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
EBUSY can happen when mounting a filesystem that
is already mounted or when the context= are
different when using the -o sharecache
Signed-off-by: Jianhong Yin <yin-jianhong@163.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
When using nfsconf tool to set values, allow it to uncomment a setting
it finds in the relevant section if it would otherwise have had to create
a new one.
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If you run systemctl isolate multi-user.target
then all services not needed by multi-user.target are stopped.
This currently includes rpc-statd, which is a problem.
Systemd doesn't know that rpc-statd.service is needed by
multi-user.target, or more accurately nfs-client.target, because
sometimes it isn't.
It is started dynamically by mount.nfs when a v3 or v2 filesystem
is mounted (not when v4 is mounted) - so systemd doesn't see the connection.
mount units default to IgnoreOnIsolate=yes, as mounts are expected to be
left mounted when you run "systemctl isolate ....". As rpc-statd.service
is started by a mount, and needed for some mounts, it makes sense for it
to declare IgnoreOnIsolate=yes too.
With this declaration in place rpc.statd remains running when
systemctl isolate multi-user.target is run.
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
CONFIG_JUNCTION was defined for use in Makefiles, but that does not
provide a -DCONFIG_JUNCTION on the compiler command line. Add logic
to configure.ac to get the compiler flags right.
Now that mountd junction support is getting built, a few last minute
porting bugs popped out. Fix those up.
Reported-by: Yongcheng Yang <yoyang@redhat.com>
BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=1543126
Fixes: 79978ed34973 ("mountd: Solder in support for NFS basic ... ")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Modernize the libxml2 script to use PKG_PROG_PKG_CONFIG instead of
an ad hoc macro. This automates the population of AM_CPPFLAGS with
"-I/usr/include/libxml2" when it is needed.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
Replace an unmatched single quote to help syntax coloring editors.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
| |
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
To make it easier to turn values on
and off, start with the section headings
not being commented out.
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
Allow the used of the gssprox-mech(8) through a the
boolean variable in the [gssd] section of nfs.conf
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
EBUSY errors can be caused by using the -o sharecache
flag and different context= values. So make sure
the filesytem is mounted before ignoring the EBUSY error.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1629705
Found-by: Jianhong.Yin <yin-jianhong@163.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
This ancient script has not been used
in years, if used at all.
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Remove last use of dnotify in nfs-utils by bringing idmapd upto
date with (required) inotify use by gssd and blkmapd.
Acked-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Alan Swanson <reiver@improbability.net>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Add nfs.conf entries for various cmdline options
These options did not have a corresponding nfs.conf entry which is
currently forcing use of cmdline flags
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Setting "vers4=n/off" (disabling the major NFS version 4) in nfs.conf
doesn't work now.
Refer to rpc.nfsd(8), "vers3/vers4" in the [nfsd] section of the
/etc/nfs.conf file is to enable or disable a major NFS version 3 or 4
(which are normally enabled by default).
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1625032
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
| |
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
TCP wrappers is a pretty old and obsolete technology. The
support has be terminated in current distro release, so
only require the library to exist when asked for via the
--with-tcp-wrappers
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
If nsm_insert_monitored_host() fails while saving the record to
stable storage, we can't just assume the entry was new. Existing
records must be removed from the list before being freed.
Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Frank Sorenson <sorenson@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since both 'rpc.nfsd -V4' and 'rpc.nfsd -V4.0' alter bit 0 in 'minorversset'
and 'minorvers', it's not possible to know whether the user explicitly
requested version 4.0 or not. Add another variable 'force4dot0' to track
when the user explicitly requested version 4.0 so that we can write '4.0'
to /proc/fs/nfsd/versions on kernel versions 4.11 and higher. On earlier
kernels, we'll still write '4' as in the past.
Signed-off-by: Scott Mayhew <smayhew@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
The nfs.conf file has a 'vers4.0' parameter, but nfsd doesn't
currently check it.
Signed-off-by: Scott Mayhew <smayhew@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
| |
Signed-off-by: Yongcheng Yang <yoyang@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
utils/gssd_proc.c uses SYS_setresuid and SYS_setresgid in
change_identity when it should use SYS_setresuid32 and
SYS_setresgid32 instead. This causes it to truncate
UIDs/GIDs > 65536.
Fixes: https://bugs.launchpad.net/ubuntu/+source/nfs-utils/+bug/1779962
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1595927
Tested-by: James Ettle <theholyettlz@googlemail.com>
Tested-by: Sree <Sree@gmail.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
| |
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
This reverts commit 98118f5e382cdeb1acf25370768d5dfdc254ba5d.
Added back the 'm' character which cause memory to be
allocated for the string convertions.
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
This reverts commit acf95d32a44fd8357c24e8a04ec53fc6900bfc58
Signed-off-by: Yongcheng Yang <yongcheng.yang@gmail.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
| |
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
Signed-off-by: Frank Sorenson <sorenson@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
Fix some typos in the nfs.conf related code
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
Use the new conf_write() function to add setting and
unsetting of config file values to the cli tool
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
Adds a function to nfsconf handling to write a single config
entry, creating the file and section headers as required.
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
Use of xlog_err is immediately fatal, switch to using
xlog(L_ERROR, ...) instead so that the error handling and
cleanup mechanisms can operate properly
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
When comparing two config section headers it incorrectly
handled matching section names where only one had a
subsection value
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
The test for empty lines didn't ignore whitespace properly
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
getopt_long() returns an int, not char, which
matters on non-x86_64 archs since a char is
signed on x86_64 arch but unsigned on other
archs.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1551903
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
Free allocated memory for path before return.
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
We're changing the kernel to allow gss requests from high ports even
when "secure" is set.
If the change gets backported to distro kernels, the kernel version may
be an imperfect predictor of the behavior, but I think it's the best we
can do.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
Document the use of IPv4/IPv6_ANY address as values to the clientaddr
as a way to tell the server that the client does not want delegations.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
If the NFS client administrator supplies the clientaddr mount
option, it should be either a special value of either IPv4/IPv6
any address or one of the machine's network addresses. Otherwise,
warn the administrator about the use of an arbitrary value for
the clientaddr value.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
umich_ldap.c:1128:3: warning: ‘strncat’ specified bound 128
equals destination size [-Wstringop-overflow=]
umich_ldap.c:1130:3: warning: ‘strncat’ specified bound 128
equals destination size [-Wstringop-overflow=]
idmapd.c:313:2: warning: ‘strncat’ specified bound 4096
equals destination size [-Wstringop-overflow=]
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
client.c:486:3: warning: ‘strncpy’ destination unchanged after
copying no bytes [-Wstringop-truncation]
file.c:99:2: warning: ‘strncpy’ specified bound 4096
equals destination size [-Wstringop-truncation]
v4root.c:95:2: warning: ‘strncpy’ specified bound 1025
equals destination size [-Wstringop-truncation]
sm-notify.c:572:3: warning: ‘strncpy’ specified bound 1025
equals destination size [-Wstringop-truncation]
nfs4mount.c:221:3: warning: ‘strncpy’ specified bound 1024
equals destination size [-Wstringop-truncation]
nfsmount.c:831:2: warning: ‘strncpy’ specified bound 256
equals destination size [-Wstringop-truncation]
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
gssd.c:291:7: error: ISO C does not support the 'm' scanf flag
[-Werror=format=]
gssd.c:291:7: error: format '%s' expects argument of type 'char *', but
argument 4 has type 'char **' [-Werror=format=]
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
idmapd.c: In function 'getfield':
idmapd.c:936:4: error: format '%o' expects argument of type 'unsigned
int *', but argument 3 has type 'int *' [-Werror=format=]
if ((n = sscanf(bp, "\\%03o", &val)) != 1)
nfsidmap.c: In function 'key_invalidate':
nfsidmap.c:322:3: error: format '%x' expects argument of type 'unsigned
int *', but argument 3 has type 'key_serial_t *' [-Werror=format=]
sscanf(buf, "%x", &key);
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
nfsstat.c: In function 'mounts':
nfsstat.c:1016:3: error: ISO C does not support the '%m' gnu_printf
format [-Werror=format=]
fprintf(stderr, "Warning: %s: %m\n", name);
^
nfsstat.c: In function 'get_stats':
nfsstat.c:1093:6: error: ISO C does not support the '%m' gnu_printf
format [-Werror=format=]
label, file);
statd.c: In function 'set_nlm_port':
statd.c:235:5: error: ISO C does not support the '%m' gnu_printf format
[-Werror=format=]
name_p, type);
^
statd.c:238:3: error: ISO C does not support the '%m' gnu_printf format
[-Werror=format=]
fprintf(stderr, "%s: failed to open %s: %m\n", name_p, pathbuf);
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
xtab.c: In function 'xtab_read':
xtab.c:60:4: error: overflow in implicit constant conversion
[-Werror=overflow]
exp->m_xtabent = 1;
^
xtab.c:61:4: error: overflow in implicit constant conversion
[-Werror=overflow]
exp->m_mayexport = 1;
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
| |
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
Only compile and use the internal rpcgen
when required via the --with-rpcgen flag
It is strongly suggested to use the rpcgen from
the new created rpcsvc-proto package be used.
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The callback client used by NFSv4.0 servers to send CB calls has to
use a source principal that is the same as the target principal
that the client used to establish the forward channel.
In multi-homed server set-ups, the domain part of the principal's
hostname may not be the same as server's DNS domain. So gssd can
no longer assume that. The kernel can scrape that domain off the
forward channel's principal and pass that up to gssd.
This patch adds a new parameter, "srchost", to the kernel upcall.
When the kernel presents this new parameter in an upcall, gssd will
use it, along with the "service" parameter, to construct the service
principal for the keytab lookup.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
| |
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
| |
Provide a man page for nfsconf
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
| |
Uses the nfsconftool cli to test that the configuration
library functions are operating as expected
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|
|
|
|
|
|
|
|
|
| |
This tool uses the conffile facilities to allow commandline
querying of configuration settings and to dump the current
config for diagnosis and testing
Signed-off-by: Justin Mitchell <jumitche@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
|