commit 95be3b747365bce3593688ea9e3c824a7dd24b6f
Author: Steve Dickson <steved@redhat.com>
Date: Tue Jan 4 13:23:04 2022 -0500
mount: removed unused lable
Commit f054365d removed the unnecessary bind() call
but did not remove the error lable.
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 252769799eff4c5cd7bac28480385043e781d76e
Author: Sergey V. Lobanov <sergey@lobanov.in>
Date: Tue Jan 4 12:02:04 2022 -0500
tools/rpcgen: fix build on macos arm64 (stat64 issue)
__DARWIN_ONLY_64_BIT_INO_T is true on macos arm64 so struct stat64
and stat64() are not available. This patch defines stat64 as stat if
__DARWIN_ONLY_64_BIT_INO_T is true
Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit ba79e680b50d18125f4d066a45989b17b71289ea
Author: Steve Dickson <steved@redhat.com>
Date: Wed Dec 8 11:09:39 2021 -0500
mount: Remove NFS v2 support from mount.nfs
This patch removes the ability to do NFS v2
mounts. They will now fail with EOPNOTSUPP.
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 522837f03a4a1b77fc4275504a4c932a11a9bc6f
Author: Steve Dickson <steved@redhat.com>
Date: Mon Nov 29 11:08:03 2021 -0500
nfs.man: Remove references to NFS v2 from the man pages
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 2c2c36c59fa1de2ff7fd28917e54700ecb39b730
Author: Steve Dickson <steved@redhat.com>
Date: Tue Nov 23 12:26:40 2021 -0500
nfsd: Remove the ability to enable NFS v2.
Signed-off-by: Steve Dickson <steved@redhat.com>
commit f054365d7f8e69b103edca2fbcd8bfa3b1cd6a14
Author: NeilBrown <neilb@suse.de>
Date: Wed Nov 10 13:57:58 2021 -0500
mount: don't bind a socket needlessly.
When clnt_ping() calls get_socket(), get_socket() will create a socket,
call bind() to choose an unused local port, and then connect to the
given address.
The "bind()" call is unnecessary and problematic.
It is unnecessary as the "connect()" call will bind the socket as
required.
It is problematic as it requires a completely unused port number, rather
than just a port number which isn't currently in use for connecting to
the given remote address.
If all local ports (net.ipv4.ip_local_port_range) are in use, the bind()
will fail. However the connect() call will only fail if all those port
are in use for connecting to the same address.
So remove the unnecessary bind() call.
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 49822c2130a2a6e1d5a9e414f5a0726a461d9b00
Author: NeilBrown <neilb@suse.de>
Date: Mon Oct 25 12:08:46 2021 -0400
Add --disable-sbin-override for when /sbin is a symlink
mount.nfs* umount.nfs* and nfsdcltrack are currently always installed in
/sbin.
Many distros are moving to a "merged /usr" where /sbin and others are
symlinks into /usr/sbin or similar. In these cases it is inelegant to
install in /sbin (i.e. install through a symlink).
So we add "--disable-sbin-override" as a configure option. This causes
the same sbindir to be used for *mount.nfs* and nfsdcltrack as for other
system binaries.
Note that autotools notices if we simply define "sbindir=/sbin"
inside an "if CONFIG_foo" clause, gives a warning, and defeats our
intent.
So instead, we use the @CONFIG_SBIN_OVERRIDE_TRUE@ prefix to find
the new declaration when we don't want it.
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 9d48962e8c6d807064163e2deb7360d062478502
Author: Stig <thatsafunnyname@gmail.com>
Date: Thu Sep 23 11:31:02 2021 -0400
mountstats: division by zero error on new mount when 0==rpcsends
When rpcsends is 0 this is the error seen when mounstats is run on a
NFSv4.2 mount
Fixes: https://bugzilla.linux-nfs.org/show_bug.cgi?id=367
Signed-off-by: Steve Dickson <steved@redhat.com>
commit feb3dfc7127cf1337530ccb06ed90e818b026a07
Author: Steve Dickson <steved@redhat.com>
Date: Wed Sep 22 11:31:56 2021 -0400
mountd: only do NFSv4 logging on supported kernels.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1979816
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 47e6d7667cd7cd82c9033a3176afbdd2341065b0
Author: Steve Dickson <steved@redhat.com>
Date: Wed Sep 22 11:29:34 2021 -0400
Move version.h into a common include directory
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 294454e5170c6531727536cdbcdbe2606ac87b35
Author: Petr Vorel <pvorel@suse.cz>
Date: Tue Sep 21 13:13:38 2021 -0400
install-dep: Use command -v instead of which
`command -v' is shell builtin required by POSIX [1] and supported on all
common shells (bash, zsh, dash, busybox sh, mksh). `which' utility is not
presented on some containers (e.g. Fedora, openSUSE), also going to be
removed from future Debian versions.
Also remove stderr redirection to /dev/null as it's unnecessary when
using 'command': POSIX says "no output shall be written" if the command
isn't found.
[1] https://pubs.opengroup.org/onlinepubs/9699919799/utilities/command.html
[2] https://salsa.debian.org/debian/debianutils/-/commit/3a8dd10b4502f7bae8fc6973c13ce23fc9da7efb
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 28deb4f398363e4e75ea41ff0fe604b11f6ee91a
Author: Olga Kornievskaia <kolga@netapp.com>
Date: Tue Sep 21 13:00:12 2021 -0400
nfs.man: adding new mount option max_connect
When client discovers trunkable servers, instead of dropping newly
created trunkable connections, add this connection to the existing
RPC client.
Signed-off-by: Olga Kornievskaia <kolga@netapp.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 1a661f02b0a44dd1200476f0e684a1a4ef809935
Author: Steve Dickson <steved@redhat.com>
Date: Tue Sep 21 12:53:52 2021 -0400
cacheio.c:216:21: warning: unused variable 'stb' [-Wunused-variable]
Signed-off-by: Steve Dickson <steved@redhat.com>
commit cfe41d6f06af0e7744c1ca30503f93d28aca4d8b
Author: NeilBrown <neilb@suse.de>
Date: Tue Sep 21 12:47:10 2021 -0400
gssd: fix crash in debug message.
A recent cleanup of debug messages added func and tid format specifiers
to a debug message (when full hostname was different), but the func name
and tid were NOT added as arguments.
Consequently there weren't enough args, random bytes of the stack were
interpreted as a pointer, and rpc.gssd crashed (when -v was specified).
Fixes: b538862a5135 ("gssd: Cleaned up debug messages")
Reviewed-by: Petr Vorel <pvorel@suse.cz>
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 9bd7e5814c38d2ac7ea218dbd8fa4dbfdf58f48a
Author: Salvatore Bonaccorso <carnil@debian.org>
Date: Tue Sep 21 12:25:18 2021 -0400
systemd generators: Install depending on location for systemd unit files
While it is possible to configure the systemd unit files directory on
configure time with --with-systemd=unit-dir-path, this path is not taken
into account for installing the systemd unit generators, as they are
installed unconditionally in /usr/lib/systemd/system-generators.
Signed-off-by: Salvatore Bonaccorso <carnil@debian.org>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 191b610a74dc603cc966258813fc6b671e750ef5
Author: Salvatore Bonaccorso <carnil@debian.org>
Date: Tue Sep 21 12:12:34 2021 -0400
systemd/Makefile: Drop exlicit setting of unit_dir
The used variable is named unitdir in configure.ac and
systemd/Makefile.am otherwise but is used in a single place as unit_dir.
The setting has no effect, but if later commits would use a base to the
systemd unit files directory for installing further files this would
void the possibility to explicitly set a systemd unit files directory
via configure with --with-systemd=unit-dir-path.
Fixes: 0fbf91a4fd90 ("Include systemd unit files in "dist" and "install".")
Signed-off-by: Salvatore Bonaccorso <carnil@debian.org>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit e8ff27d6a29550d997372f41249dac3f83668b5b
Author: Jianhong Yin <yin-jianhong@163.com>
Date: Sat Aug 21 13:52:57 2021 -0400
nfs-utils: add install-dep for installing all dependencies
Whenever user want to compile and install from source code, they
have to constantly install dependencies based on error message.
Verified on RHEL-8/Fedora-34/debian-10/openSUSE-15.3
Signed-off-by: Jianhong Yin <yin-jianhong@163.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 5724a33d43430ea993ea41bc9536d185a96bd44b
Author: Alice Mitchell <ajmitchell@redhat.com>
Date: Sat Aug 21 13:46:40 2021 -0400
nfs-utils: Fix mem leak in mountd
leak of mountlist struct and content on error
Signed-off-by: Alice Mitchell <ajmitchell@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit e1ba9f52c965aa632ce1289b16dbaf0f65fc1a21
Author: Alice Mitchell <ajmitchell@redhat.com>
Date: Sat Aug 21 13:44:09 2021 -0400
nfs-utils: Fix mem leaks in krb5_util
query_krb5_ccache: if the ret_realm strdup fails then ret_princname leaks
gssd_get_krb5_machine_cred_list: l was being leaked if the realloc failed
it was also leaked if the strdup of ccname failed
Signed-off-by: Alice Mitchell <ajmitchell@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 8d8f2792eb266a42ecf40ed5473af6a24f0d3bdf
Author: Alice Mitchell <ajmitchell@redhat.com>
Date: Sat Aug 21 13:42:33 2021 -0400
nfs-utils: Fix mem leaks in gssd
ccachedir_copy isnt used properly and is leaking, ccachedir gets modified
by a strtok, altering the original argv or conf parameter which is an
undesirable side-effect
Signed-off-by: Alice Mitchell <ajmitchell@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 647f3796d11ed6122d13331b96d15c1151b59e23
Author: Alice Mitchell <ajmitchell@redhat.com>
Date: Sat Aug 21 13:38:02 2021 -0400
nfs-utils: Fix potential memory leaks in idmap
regex.c: regex_getpwnam() would leak memory if the name was not found.
nss.c: nss_name_to_gid() the conditional frees look like a potential
memory leak, removed the unnecessary conditions.
Signed-off-by: Alice Mitchell <ajmitchell@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 383d787d1b77f165da68495cb0363220b66935a4
Author: Steve Dickson <steved@redhat.com>
Date: Tue Jul 27 21:12:17 2021 -0400
nfsdcltrack: Use uint64_t instead of time_t
With recent commits (4f2a5b64,5a53426c) that fixed
compile errors on x86_64 machines, caused similar
errors on i686 machines.
The variable type that was being used was a time_t,
which changes size between architects, which
caused the compile error.
Changing the variable to uint64_t fixed the issue.
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 364b7fefd6680d592fe14cb2ba838843b1124048
Author: James Hilliard <james.hilliard1@gmail.com>
Date: Mon Jul 26 12:03:09 2021 -0400
systemd: Fix non-default statedir paths.
Make the state directory configurable.
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 5a53426c0f4e84f6644c11e4f01015597d3bb82c
Author: Petr Vorel <petr.vorel@gmail.com>
Date: Sun Jul 25 13:13:40 2021 -0400
nfsdcltrack/nfsdcltrack.c: Fix printf format
nfsdcltrack.c: In function 'cltrack_gracedone':
nfsdcltrack.c:528:47: error: format '%ld' expects argument of type 'long int', but argument 4 has type 'time_t' {aka 'long long int'} [-Werror=format=]
Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 4f2a5b64386600430ec6b71e57b88c5efaa70aed
Author: Petr Vorel <petr.vorel@gmail.com>
Date: Sun Jul 25 13:09:59 2021 -0400
nfsdcltrack/sqlite: Fix printf format
sqlite.c: In function 'sqlite_remove_unreclaimed':
sqlite.c:547:71: error: format '%ld' expects argument of type 'long int', but argument 4 has type 'time_t' {aka 'long long int'} [-Werror=format=]
Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Signed-off-by: Steve Dickson <steved@redhat.com>
commit 4dd8d833c9350d42528ada0fd65aee41b712f41d
Author: Steve Dickson <steved@redhat.com>
Date: Tue Jul 20 17:14:04 2021 -0400
mount.nfs: Fix the sloppy option processing
The new mount API broke how the sloppy option is parsed.
So the option processing needs to be moved up in
the mount.nfs command.
The option needs to be the first option in the string
that is passed into the kernel with the -s mount(8)
and/or the -o sloppy is used.
Commit 92b664ef fixed the process of the -s flag
and this version fixes the -o sloppy processing
as well works when libmount-mount is and is not
enabled plus cleans up the mount options passed
to the kernel.
Reviewed-and-tested-by: Dave Wysochanski <dwysocha@redhat.com>
Signed-off-by: Steve Dickson <steved@redhat.com>