1 commit 684cf4a5e0e84a1367690d7ecf4882cbdf4f3228
2 Author: Prem Karat <prem.karat@linux.vnet.ibm.com>
3 Date: Thu Jun 30 07:29:20 2011 -0400
5 mount.nfs: Fix for the bug in v1.2.4 that breaks mount.nfs
7 commit 30ebf047 failed to include these changes that breaks mount.nfs.
8 mount.nfs will continue to work fine with these changes
10 Signed-off-by: Steve Dickson <steved@redhat.com>
12 commit 057d2add27d1e05fed3ae7206ee043b6c1fda45a
13 Author: NeilBrown <neilb@suse.de>
14 Date: Tue Jun 28 13:24:33 2011 -0400
16 Do not compile unnecessary files when the libmount code is enable
18 Signed-off-by: Steve Dickson <steved@redhat.com>
20 commit 30ebf04700654deddbf5f57d95e84ec69cea8610
21 Author: Prem Karat <prem.karat@linux.vnet.ibm.com>
22 Date: Tue Jun 28 11:53:40 2011 -0400
24 mount.nfs: Don't hard code source and destination
26 Currently souce and destination parameters should be passed as first and
27 second paramter while using mount.nfs. This patch allows them to be passed
28 anywhere while mounting.
30 Current functionality is
31 mount.nfs source destn -o <options>
32 This patch will allow to do this
33 mount.nfs -o <options> source destn
35 mount.nfs -o <options> source -o <options> destn
37 Signed-off-by: Prem Karat <prem.karat@linux.vnet.ibm.com>
38 Signed-off-by: Steve Dickson <steved@redhat.com>
40 commit b3e190c4adfc9ec47567c968bd000d282d07b05e
41 Author: NeilBrown <neilb@suse.de>
42 Date: Tue Jun 28 11:36:31 2011 -0400
44 mount: improve signal management when locking mtab
46 As mount.nfs can run setuid it must be careful about how the user can
47 interact with in. In particular it needs to ensure it does not
48 respond badly to any signals that the user might be able to generate.
50 This is particularly an issue while updating /etc/mtab (when that is
51 not linked to /proc/mounts). If the user can generate a signal which
52 kills mount.nfs while /etc/mtab is locked, then it will leave the file
53 locked, and could possibly corrupt mtab (particularly if 'ulimit 1'
54 was previously issued).
56 Currently lock_mtab does set some handlers for signals, but not
57 enough. It arranges for every signal up to (but not including)
58 SIGCHLD to cause mount.nfs to unlock mdadm promptly exit ... even if
59 the default behaviour would be to ignore the signal. SIGALRM is
60 handled specially, and signals after SIGCHLD are left with their
61 default behaviour. This includes for example SIGXFSZ which can be
62 generated by the user running "ulimit 1".
64 So: change this so that some signals are left unchanged, SIGALRM is
65 handled as required, and all signals that the user can generate are
68 The remainder still cause mount.nfs to print a message, unlock mtab, and
71 Signed-off-by: NeilBrown <neilb@suse.de>
72 Signed-off-by: Steve Dickson <steved@redhat.com>
74 commit a99b7846e2abec5e26ab6b764b921d79559e0a0f
75 Author: J. Bruce Fields <bfields@redhat.com>
76 Date: Mon Jun 27 12:31:07 2011 -0400
78 mountd: move fsidtype-specific code to helpers
80 Now we can move these big switch statements into helper functions.
82 Signed-off-by: J. Bruce Fields <bfields@redhat.com>
83 Signed-off-by: Steve Dickson <steved@redhat.com>
85 commit e6559fd0b7b63f5d152d33d598dc74d78df30ecb
86 Author: J. Bruce Fields <bfields@redhat.com>
87 Date: Mon Jun 27 12:30:36 2011 -0400
89 mountd: gather fsid information into one struct
91 A large part of nfsd_fh() is concerned with extracting
92 fsid-type-specific information from the fsid, then matching that
93 information with information from the export list and the filesystem.
95 Moving all that information into one struct will allow some further
98 Signed-off-by: J. Bruce Fields <bfields@redhat.com>
99 Signed-off-by: Steve Dickson <steved@redhat.com>
101 commit 13a0a61d037f2cc09e7997a96ce5822b9317883b
102 Author: J. Bruce Fields <bfields@redhat.com>
103 Date: Mon Jun 27 12:29:51 2011 -0400
105 mountd: prefer explicit subexports over crossmnt parents
107 If a parent is exported with crossmnt, and if a child is also explicitly
108 exported, then both exports could potentially produce matches in this
109 loop; that isn't a bug.
111 Instead of warning and ignoring the second match we find, we should
112 instead prefer whichever export is deeper in the tree, so that
113 children's options can override those of their parents.
115 Reported-by: Olga Kornievskaia <aglo@citi.umich.edu>
116 Signed-off-by: J. Bruce Fields <bfields@redhat.com>
117 Signed-off-by: Steve Dickson <steved@redhat.com>
119 commit f8d26c1db9a260597828685c7f62e1b29e78285f
120 Author: Jeff Layton <jlayton@redhat.com>
121 Date: Wed Jun 22 15:52:55 2011 -0400
123 manpage: add section on character class matches to exports(5)
125 Signed-off-by: Jeff Layton <jlayton@redhat.com>
126 Signed-off-by: Steve Dickson <steved@redhat.com>
128 commit cb6676aea5bcfcbeaf868e53177eff51f4efe9a8
129 Author: James Pearson <james-p@moving-picture.com>
130 Date: Wed Jun 22 15:51:47 2011 -0400
132 nfs-utils: remove possibly false statement from exports.man
134 A very minor change suggested by J. Bruce Fields <bfields@fieldses.org>
135 to remove the statement that exporting to a single host or IP address is
136 the "most common format" - as it probably isn't.
138 Signed-off-by: James Pearson <james-p@moving-picture.com>
139 Signed-off-by: Steve Dickson <steved@redhat.com>
141 commit 7d71ff8e3825a7f8c2d5c9b5b9344e95e7aa1392
142 Author: Jeff Layton <jlayton@redhat.com>
143 Date: Wed Jun 22 15:51:02 2011 -0400
145 manpage: add info about IPv6 configuration to exports(5)
147 The parts of the exports(5) manpage that discuss IP addressing neglect
148 IPv6 configuration. Update to include info on how to export to IPv6
149 subnets and addresses, and add a line demonstrating that to the EXAMPLE
152 Signed-off-by: Jeff Layton <jlayton@redhat.com>
153 Signed-off-by: Steve Dickson <steved@redhat.com>
155 commit 545ea1098089d6396d5a70111ec231c4de967faa
156 Author: Benny Halevy <benny@tonian.com>
157 Date: Wed Jun 22 15:44:17 2011 -0400
159 nfsstat: reorder nfs4 stats for 2.6.39
161 Signed-off-by: Benny Halevy <benny@tonian.com>
162 Signed-off-by: Steve Dickson <steved@redhat.com>
164 commit c7ce7a4674ad446bee4dd3baf90155ce6b216816
165 Author: Mi Jinlong <mijinlong@cn.fujitsu.com>
166 Date: Wed Jun 22 15:41:27 2011 -0400
168 libexport.a: fix using bad index for loop at cltsetup()
170 In cltsetup(), when checking the address, use clp's naddr for index,
171 instead of cltarg's naddr, which it's always zero there.
173 Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com>
174 Signed-off-by: Steve Dickson <steved@redhat.com>
176 commit b50ad13298b3e9519a9bdecb8c146c9ecf39cef8
177 Author: Jeff Layton <jlayton@redhat.com>
178 Date: Wed Jun 22 14:51:38 2011 -0400
180 nfs: fix host_reliable_addrinfo
182 According to Neil Brown:
184 The point of the word 'reliable' is to check that the name we get
185 really does belong to the host in question - ie that both the
186 forward and reverse maps agree.
188 But the new code doesn't do that check at all. Rather it simply
189 maps the address to a name, then discards the address and maps the
190 name back to a list of addresses and uses that list of addresses as
191 "where the request came from" for permission checking.
193 This bug is exploitable via the following scenario and could allow an
194 attacker access to data that they shouldn't be able to access.
196 Suppose you export a filesystem to some subnet or FQDN and also to a
197 wildcard or netgroup, and I know the details of this (maybe
198 showmount -e tells me) Suppose further that I can get IP packets to
201 Then I create a reverse mapping for my ipaddress to a domain that I
202 own, say "black.hat.org", and a forward mapping from that domain to
203 my IP address, and one of your IP addresses.
205 Then I try to mount your filesystem. The IP address gets correctly
206 mapped to "black.hat.org" and then mapped to both my IP address and
209 Then you search through all of your exports and find that one of the
210 addresses: yours - is allowed to access the filesystem.
212 So you create an export based on the addrinfo you have which allows
213 my IP address the same access as your IP address.
215 Fix this by instead using the forward lookup of the hostname just to
216 verify that the original address is in the list. Then do a numeric
217 lookup using the address and stick the hostname in the ai_canonname.
219 Reviewed-by: NeilBrown <neilb@suse.de>
220 Signed-off-by: Jeff Layton <jlayton@redhat.com>
221 Signed-off-by: Steve Dickson <steved@redhat.com>
223 commit 7235a2164aabfd8dba1f7e1577047bda45053db0
224 Author: James Pearson <james-p@moving-picture.com>
225 Date: Tue Jun 7 16:25:13 2011 -0400
227 exports: Clearly Defining Exports Priorities
229 Added some verbiage to the exports(5) man page
230 that clearly explains the precedence around
231 how exports will work with regard to netgroups.
233 Signed-off-by: Steve Dickson <steved@redhat.com>
235 commit a36f2437caffb23c68fdc0900544d59198bd52b6
236 Author: Neil Brown <neilb@suse.de>
237 Date: Tue Jun 7 13:18:55 2011 -0400
239 Document "-n" for svcgssd
241 The svcgssd man page doesn't mention the "-n" flag.
243 Signed-off-by: NeilBrown <neilb@suse.de>
244 Signed-off-by: Steve Dickson <steved@redhat.com>
246 commit 40aed2c3fb5164195a9975ae7f15ebd4b992839d
247 Author: Pavel Shilovsky <piastry@etersoft.ru>
248 Date: Tue Jun 7 13:18:13 2011 -0400
250 mountd: Fix missing varialble assignment in auth_unix_gid
252 When we get into auth_unix_gid at the second time, groups_len
253 is not 0 and ngroups variable leave as 0. Then we use ngroups
254 in getgrouplist that fails in this case. This patch fixes it.
256 Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
257 Signed-off-by: Steve Dickson <steved@redhat.com>
259 commit 8935933dedcd820c2fb3dddff8b79fd5841dc217
260 Author: Benny Halevy <bhalevy@panasas.com>
261 Date: Mon May 23 08:37:17 2011 -0400
263 nfsstat: reorder nfs4 stats for 2.6.38 and up
265 match order in 2.6.38, 2.6.39 (-rc3) and development tree
266 while at it, get rid of obsolete ds_write and ds_commit
268 Signed-off-by: Benny Halevy <bhalevy@panasas.com>
269 Signed-off-by: Steve Dickson <steved@redhat.com>
271 commit 27dcd8a775024160e741cce53e4a402eaac3501d
272 Author: NeilBrown <neilb@suse.de>
273 Date: Mon May 23 08:23:51 2011 -0400
275 supress socket error when address family is not supported
277 From: Suresh Jayaraman <sjayaraman@suse.de>
279 It was observed that when ipv6 module was not loaded and cannot be auto-loaded,
280 when starting NFS server, the following error occurs:
281 "rpc.nfsd: unable to create inet6 TCP socket: errno 97 (Address
282 family not supported by protocol)"
284 This is obviously a true message, but does not represent an "error" when ipv6
285 is not enabled. Rather, it is an expected condition. As such, it can be
286 confusing / misleading / distracting to display it in this scenario.
288 This patch instead of throwing error when a socket call fails with
289 EAFNOSUPPORT, makes it as a NOTICE.
291 Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
292 Signed-off-by: Neil Brown <neilb@suse.de>
293 Signed-off-by: Steve Dickson <steved@redhat.com>
295 commit 7a802337bfc92d0b30fe94dbd0fa231990a26161
296 Author: NeilBrown <neilb@suse.de>
297 Date: Mon May 23 08:19:57 2011 -0400
299 Remove risk of nfs_addmntent corrupting mtab
301 nfs_addmntent is used to append directly to /etc/mtab.
302 If the write partially fail, e.g. due to RLIMIT_FSIZE,
303 truncate back to original size and return an error.
305 See also https://bugzilla.redhat.com/show_bug.cgi?id=697975
306 (CVE-2011-1749) CVE-2011-1749 nfs-utils: mount.nfs fails to anticipate RLIMIT_FSIZE
308 Signed-off-by: NeilBrown <neilb@suse.de>
309 Signed-off-by: Steve Dickson <steved@redhat.com>
311 commit 56f537535190d034039570bafd9a0de71b79b8f1
312 Author: Ben Myers <bpm@sgi.com>
313 Date: Mon May 23 08:07:00 2011 -0400
315 exportfs: getexportent interprets -test-client- as default options
317 With commit 1374c3861abdc66f3a1410e26cc85f86760b51dd Neil added a
318 -test-client- export to test the exportability of filesystems when exportfs
319 is run. When using the old cache controls (i.e. /proc/fs/nfsd is not
320 mounted) exportfs will read /proc/fs/nfs/exports to process existing
321 exports and find these test client entries. The dash at the beginning of
322 -test-client- will be cause getexportent to look for default options in the
323 rest of the string, which test-client- will not match:
325 exportfs: /proc/fs/nfs/exports:1: unknown keyword "test-client-(rw"
327 This patch resolves that problem (as Steve suggested) by not processing any
328 default options if we are reading the list of existing exports from the
329 kernel. Default options are converted to individual exports by exportfs so
330 the kernel won't have any regardless.
332 Signed-off-by: Ben Myers <bpm@sgi.com>
333 Signed-off-by: Steve Dickson <steved@redhat.com>
335 commit c8e802c036a3f0fcd4481dae8b3ec09fb71f4118
336 Author: Jim Rees <rees@umich.edu>
337 Date: Wed May 18 12:42:02 2011 -0400
339 Removed compilation warnings from mountd/cache.c
341 Commit 5604b35a6 introduced a number of missing initializer
342 warnings that were missed. This patch removes those warnings.
344 Signed-off-by: Steve Dickson <steved@redhat.com>
346 commit 68f7938c85258a8c54b13169dcdeae61cc1bf286
347 Author: Steve Dickson <steved@redhat.com>
348 Date: Tue Apr 26 13:32:35 2011 -0400
350 nfsstat: Output headings mislabled
352 The badclnt and badauth headers were reversed
353 when the server side rpc stats (-s -o rpc) were
356 Signed-off-by: Steve Dickson <steved@redhat.com>
358 commit 2e4a7e9b50a641dc8640a5dd911ee4a2f0b2c073
359 Author: Steve Dickson <steved@redhat.com>
360 Date: Tue Apr 19 12:31:30 2011 -0400
362 rpc.svcgssd: Segmentation fault on error
364 Commit 544ed73d introduced a regression that caused
365 rpc.svcgssd to seg fault on "Wrong principal in request"
366 errors in gss_accept_sec_context()
368 Signed-off-by: Steve Dickson <steved@redhat.com>
370 commit 5604b35a61e22930873ffc4e9971002f578e7978
371 Author: Sean Finney <sean.finney@sonyericsson.com>
372 Date: Tue Apr 19 11:04:35 2011 -0400
374 nfs-utils: Increase the stdio file buffer size for procfs files
376 Previously, when writing to /proc/net/rpc/*/channel, if a cache line
377 were larger than the default buffer size (likely 1024 bytes), mountd
378 and svcgssd would split writes into a number of buffer-sized writes.
379 Each of these writes would get an EINVAL error back from the kernel
380 procfs handle (it expects line-oriented input and does not account for
381 multiple/split writes), and no cache update would occur.
383 When such behavior occurs, NFS clients depending on mountd to finish
384 the cache operation would block/hang, or receive EPERM, depending on
385 the context of the operation. This is likely to happen if a user is a
386 member of a large (~100-200) number of groups.
388 Instead, every fopen() on the procfs files in question is followed by
389 a call to setvbuf(), using a per-file dedicated buffer of
390 RPC_CHAN_BUF_SIZE length.
392 Really, mountd should not be using stdio-style buffered file operations
393 on files in /proc to begin with. A better solution would be to use
394 internally managed buffers and calls to write() instead of these stdio
395 calls, but that would be a more extensive change; so this is proposed
396 as a quick and not-so-dirty fix in the meantime.
398 Signed-off-by: Sean Finney <sean.finney@sonyericsson.com>
399 Signed-off-by: Steve Dickson <steved@redhat.com>
401 commit 9274e94db85bac04e170414cb8e0f4be271cde90
402 Author: Sean Finney <sean.finney@sonyericsson.com>
403 Date: Tue Apr 19 11:05:47 2011 -0400
405 mountd: Use a dynamic buffer for storing lists of gid's
407 Previously, in auth_unix_gid, group lists were stored in an array of
408 hard-coded length 100, and in the situation that the group lists for a
409 particular call were too large, the array was swapped with a dynamically
410 allocated/freed buffer. For environments where users are commonly in
411 a large number of groups, this isn't an ideal approach.
413 Instead, use malloc/realloc to grow the list on an as-needed basis.
415 Signed-off-by: Sean Finney <sean.finney@sonyericsson.com>
416 Signed-off-by: Steve Dickson <steved@redhat.com>
418 commit a99269230a0e77e7bed4fa31c9547f0d61c7f206
419 Author: Karel Zak <kzak@redhat.com>
420 Date: Wed Apr 6 12:39:21 2011 -0400
422 mount: add --enable-libmount-mount
424 This patch allows to link mount.nfs with libmount from util-linux >=
425 v2.19. The new libmount based code is enabled by CONFIG_LIBMOUNT and
426 is stored in mount_libmount.c. The old code is not affected by this
429 The libmount does not have officially stable API yet, so the
430 --enable-libmount-mount is marked as experimental in the configure
433 The ./configure option is the same as we use in util-linux to enable
434 support for libmount in mount(8).
436 The addr= (and some other options necessary for remount/umount) are
437 stored to /etc/mtab or to /dev/.mount/utab. The utab file is *private*
438 libmount file. It's possible that some mount options (for example
439 user=) will be moved to kernel, so the utab will not be necessary.
443 * supports systems without and with regular /etc/mtab
444 * does not store VFS and FS mount options in userspace
445 * manages user= option and evaluate permissions
446 * parses VFS mount options and generate MS_* flags
447 * parses /etc/{fstab,mtab}, /proc/mounts or /proc/self/mountinfo
448 * long-term goal is to use the same code in all mount.<type> helpers
452 LIBMOUNT_DEBUG=0xffff mount.nfs foo:/path /path
454 to debug the library.
456 On systems with util-linux v2.19 the findmnt(8) command uses libmount
457 to list all/selected mount points:
460 $ findmnt --mtab /path
462 the --mtab appends userspace mount options (e.g. user=) to the output.
464 CC: Chuck Lever <chuck.lever@oracle.com>
465 Signed-off-by: Karel Zak <kzak@redhat.com>
466 Signed-off-by: Steve Dickson <steved@redhat.com>
468 commit c01e5ca6179b8f5b041605d9bbd75a0f76812d54
469 Author: Karel Zak <kzak@redhat.com>
470 Date: Wed Apr 6 11:36:40 2011 -0400
472 mount: move generic functions to utils.c and network.c
474 Move generic code that could be shared between standard mount.nfs and
475 libmount version to utils.c and network.c.
477 CC: Chuck Lever <chuck.lever@oracle.com>
478 Signed-off-by: Karel Zak <kzak@redhat.com>
479 Signed-off-by: Steve Dickson <steved@redhat.com>
481 commit d6c1b35c6b40243bfd6fba2591c9f8f2653078c0
482 Author: Kevin Coffman <kwc@citi.umich.edu>
483 Date: Wed Apr 6 11:25:03 2011 -0400
485 nfs-utils: Add support to svcgssd to limit the negotiated enctypes
487 Recent versions of Kerberos libraries negotiate and use
488 an "acceptor subkey". This negotiation does not consider
489 that a service may have limited the encryption keys in its
490 keytab. A patch (http://src.mit.edu/fisheye/changelog/krb5/?cs=24603)
491 has been added to the MIT Kerberos code to allow an application
492 to indicate that it wants to limit the encryption types negotiated.
493 (This functionality has been available on the client/initiator
494 side for a while. The new patch adds this support to the
495 server/acceptor side.)
497 This patch adds support to read a recently added nfsd
498 proc file to determine the encryption types supported by
499 the kernel and calls the function to limit encryption
500 types negotiated for the acceptor subkey.
502 Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
503 Signed-off-by: Steve Dickson <steved@redhat.com>
505 commit 73840ef610accf4cf667427bc64805377c0d8394
506 Author: Chuck Lever <chuck.lever@oracle.com>
507 Date: Wed Apr 6 10:53:57 2011 -0400
509 exports: add a configurable time-to-live for the kernel cache entries
511 From: Trond Myklebust <Trond.Myklebust@netapp.com>
513 The fedfs ldap server will specify a ttl for its entries.
515 Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
517 This is a refactoring change only. There should be no change in
520 Original patch had updates to utils/mountd/junctions.c, which no
521 longer exists. These are not included here.
523 Create a macro for the default cache TTL, which is used in several
524 places besides the export cache.
528 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
529 Signed-off-by: Steve Dickson <steved@redhat.com>
531 commit 38e4c685410885a6d464ddd44eff4fd5e7f8459f
532 Author: Chuck Lever <chuck.lever@oracle.com>
533 Date: Wed Apr 6 10:49:52 2011 -0400
535 statd: Remove vestigial "-w" option from man page synopsis
537 The synopsis of rpc.statd in its man page lists "-w" as a valid
538 option. There is currently no support in the source code for a "-w"
541 BugLink: https://bugzilla.linux-nfs.org/show_bug.cgi?id=199
543 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
544 Signed-off-by: Steve Dickson <steved@redhat.com>
546 commit b57cd77c13831051ad974ae027d96cd88a8d0c59
547 Author: Chuck Lever <chuck.lever@oracle.com>
548 Date: Wed Apr 6 10:48:38 2011 -0400
550 mount.nfs: Don't leak socket in nfs_ca_sockname()
552 Ensure the test socket is always closed before nfs_ca_sockname()
553 returns. Otherwise it's orphaned.
555 BugLink: https://bugzilla.linux-nfs.org/show_bug.cgi?id=197
557 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
558 Signed-off-by: Steve Dickson <steved@redhat.com>
560 commit 502eef09d8050ffb87d394397c2780e1ef042d68
561 Author: Steve Dickson <steved@redhat.com>
562 Date: Wed Apr 6 10:46:06 2011 -0400
564 Removed a warning from v4root.c
566 v4root.c:176:9: warning: variable 'ret' set but not used
568 Signed-off-by: Steve Dickson <steved@redhat.com>
570 commit c2fa189a0467c25666f014cf9ff2576a9f54d682
571 Author: Steve Dickson <steved@redhat.com>
572 Date: Wed Apr 6 10:39:10 2011 -0400
574 Removed a warning from exportfs.c
576 exportfs.c:280:29: warning: 'exp' may be used uninitialized in this function
578 Signed-off-by: Steve Dickson <steved@redhat.com>
580 commit b98ae9df8c4904289c9390288325058b24caa423
581 Author: Steve Dickson <steved@redhat.com>
582 Date: Wed Apr 6 10:36:30 2011 -0400
584 Removed a warning from conffile.c
586 conffile.c:258:19: warning: 'j' may be used uninitialized in this function
588 Signed-off-by: Steve Dickson <steved@redhat.com>
590 commit 012e1a4bf2a002e8cd4d5be3478bfa20a91cbfed
591 Author: Masatake YAMATO <yamato@redhat.com>
592 Date: Mon Mar 7 08:36:19 2011 -0500
594 Update man pages for /etc/exports.d
596 Man page updates for /etc/exports.d.
598 Signed-off-by: Masatake YAMATO <yamato@redhat.com>
599 Signed-off-by: Steve Dickson <steved@redhat.com>
601 commit c7427b57e2be8ef0d57ad0618d4590c062b130f5
602 Author: Masatake YAMATO <yamato@redhat.com>
603 Date: Mon Mar 7 08:18:51 2011 -0500
605 Read /etc/exports.d/*.export as extra export files
607 This patch adding a capability to read /etc/exports.d/*.exports as
608 extra export files to exportfs.
610 If one wants to add or remove an export entry in a script, currently
611 one may have to use sed or something tool for adding or removing the
612 line for the entry in /etc/exports file.
614 With the patch, adding and removing an entry from a script is much
616 cat<<EOF... or mv can be used for adding. rm can be used for removing.
618 Signed-off-by: Masatake YAMATO <yamato@redhat.com>
619 Signed-off-by: Steve Dickson <steved@redhat.com>
621 commit edb9b7f2ab9806afb9af31eabeb505fe454c51df
622 Author: Steve Dickson <steved@redhat.com>
623 Date: Sat Mar 5 16:17:01 2011 -0500
625 Cleaned up a warning in rpcdispatch.c
627 rpcdispatch.c:40:20: warning: comparison between signed and unsigned
630 Signed-off-by: Steve Dickson <steved@redhat.com>
632 commit 930323817b61877d61fb8ef57229013daa2e6091
633 Author: Steve Dickson <steved@redhat.com>
634 Date: Sat Mar 5 16:13:01 2011 -0500
636 mount: Remove MOUNT_CONFIG warnings
638 The following changes are needed to remove compile warnings when
639 MOUNT_CONFIG is not defined
641 Signed-off-by: Steve Dickson <steved@redhat.com>
643 commit 3ef3dc8f1e87ba7a6eaa3c2a6965aff6c80ba414
644 Author: Chuck Lever <chuck.lever@oracle.com>
645 Date: Thu Mar 3 17:26:33 2011 -0500
647 mount: Recognize zero as a valid value for the port= option
649 While zero is not a valid IP port number, zero does represent a valid
650 value for "port=". It means "query rpcbind to discover the actual
651 non-zero port number to use". So the parsing functions that handle
652 "port=" should not flag zero as an invalid value.
654 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
655 Signed-off-by: Steve Dickson <steved@redhat.com>
657 commit b3a4dbfb61dad59829f5191d727267b2ea45937a
658 Author: Mi Jinlong <mijinlong@cn.fujitsu.com>
659 Date: Wed Feb 9 11:29:42 2011 -0500
661 Gssd: modify wrong err message at handle_gssd_upcall
663 Modify wrong err message at handle_gssd_upcall when
664 sscanf encryption types fail.
666 Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com>
667 Signed-off-by: Steve Dickson <steved@redhat.com>
669 commit 45e4597bd570ed40221f51887cde7d7f096f55e7
670 Author: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
671 Date: Wed Feb 9 11:27:19 2011 -0500
673 Support AD style kerberos automatically in rpc.gss
675 An Active Directory KDC will only grant a TGT for UPNs, getting
676 a TGT for SPNs is not possible:
678 $ kinit -k host/ib5@ADS.ORCORP.CA
679 kinit: Client not found in Kerberos database while getting initial
682 The correct thing to do for machine credentials is to get a TGT
683 for the computer UPN <HOSTNAME>$@REALM:
686 12/22/10 11:43:47 12/22/10 21:43:47 krbtgt/ADS.ORCORP.CA@ADS.ORCORP.CA
688 Samba automatically creates /etc/krb5.keytab entry for the computer UPN,
689 this patch makes gssd_refresh_krb5_machine_credential prefer it above
690 the SPNs if it is present.
692 The net result is that nfs client works automatically out of the box
693 if samba has been used to setup kerberos via 'net ads join' 'net ads
696 Tested using Windows Server 2003 R2 as the AD server.
698 Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
699 Signed-off-by: Steve Dickson <steved@redhat.com>
701 commit 730f6986f86873513fa021a450eb55ccd0f2fbff
702 Author: Steve Dickson <steved@redhat.com>
703 Date: Wed Jan 26 07:49:19 2011 -0500
705 Fixed segfault in rpc.mountd
707 A unallocated piece of memory, instead of a NULL point, was being
708 used to initialize a ->next point in the mount link list which
709 caused a segfault after a few remote accesses via the showmount
712 Signed-off-by: Steve Dickson <steved@redhat.com>
714 commit 544ed73d5ab27c1390833d5cf93b9585c151667d
715 Author: Steve Dickson <steved@redhat.com>
716 Date: Fri Jan 14 10:12:28 2011 -0500
718 Improve debugging in svcgssd
720 Added in gss_display_error() which translates the GSS error into the
721 actual GSS macro name. Currently only the translation of these errors
722 are logged. Since those translations are buried deep in the kerberos
723 library code, having the actual GSS macro name makes it easier to
726 Moved the nfs4_init_name_mapping() call into main() so if debug is
727 enabled the DNS name and realms will be logged during start up.
729 Signed-off-by: Steve Dickson <steved@redhat.com>
731 commit 57be18b9ab08148a1cc9d5af588119885720be8b
732 Author: Mi Jinlong <mijinlong@cn.fujitsu.com>
733 Date: Tue Jan 4 11:16:45 2011 -0500
735 libnsm.a: modify return value to false from 0 at nsm_drop_privileges()
737 At nsm_drop_privileges(), for improving readability, unify
740 Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com>
741 Signed-off-by: Steve Dickson <steved@redhat.com>
743 commit 5c498280fd9353ded3ea169841079bdae23418e2
744 Author: Chuck Lever <chuck.lever@oracle.com>
745 Date: Mon Dec 13 14:50:45 2010 -0500
747 libnsm.a: sm-notify sometimes ignores monitored hosts
749 Monitored host information is stored in files under /var/lib/nfs.
750 When visiting entries in the monitored hosts directory, libnsm.a
751 examines the value of dirent.d_type to determine if an entry is a
754 According to readdir(3), the d_type field is not supported by all
755 file system types. My root file system happens to be one where d_type
756 isn't supported. Typical installations that use an ext-derived root
757 file system are not exposed to this issue, but those who use xfs, for
760 On such file systems, not only are remote peers not notified of
761 reboots, but the NSM state number is never incremented. A statd warm
762 restart would not re-monitor any hosts that were monitored before
765 When writing support/nsm/file.c, I copied the use of d_type from the
766 original statd code, so this has likely been an issue for some time.
768 Replace the use of d_type in support/nsm/file.c with a call to
769 lstat(2). It's extra code, but is guaranteed to work on all file
772 Note there is a usage of d_type in gssd. I'll let gssd and rpcpipefs
773 experts decide whether that's worth changing.
777 https://bugzilla.linux-nfs.org/show_bug.cgi?id=193
779 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
780 Signed-off-by: Steve Dickson <steved@redhat.com>
782 commit 089df7c754d9ebab0a7b804f396626ac95fee2e6
783 Author: Chuck Lever <chuck.lever@oracle.com>
784 Date: Mon Dec 13 14:47:42 2010 -0500
786 libnsm.a: Replace __attribute_noinline__
788 Replace the __attribute_noinline__ form with
790 __attribute__((__noinline__)).
792 Even though the compiler didn't complain about __attribute_malloc__,
793 also replace those in order to maintain consistent style throughout the
798 https://bugzilla.linux-nfs.org/show_bug.cgi?id=194
800 Reported-by: "Gabor Z. Papp" <gzp@papp.hu>
801 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
802 Signed-off-by: Steve Dickson <steved@redhat.com>
804 commit 7869a76207d3f4b3bd4ab57b4a7a8807ac2ff0c6
805 Author: Chuck Lever <chuck.lever@oracle.com>
806 Date: Mon Dec 13 14:36:15 2010 -0500
808 sm-notify: Make use of AI_NUMERICSERV conditional
810 Gabor Papp reports nfs-utils-1.2.3 doesn't build on his system that
813 make[3]: Entering directory
814 `/home/gzp/src/nfs-utils-1.2.3/utils/statd'
815 gcc -DHAVE_CONFIG_H -I. -I../../support/include -D_GNU_SOURCE -Wall
816 -Wextra -Wstrict-prototypes -pipe -g -O2 -MT sm-notify.o -MD
817 -MP -MF .deps/sm-notify.Tpo -c -o sm-notify.o sm-notify.c
818 sm-notify.c: In function 'smn_bind_address':
819 sm-notify.c:247: error: 'AI_NUMERICSERV' undeclared (first use in this
821 sm-notify.c:247: error: (Each undeclared identifier is reported only
823 sm-notify.c:247: error: for each function it appears in.)
824 make[3]: *** [sm-notify.o] Error 1
826 According to the getaddrinfo(3) man page, AI_NUMERICSERV is available
827 only since glibc 2.3.4. getaddrinfo(3) seems to convert strings
828 containing a number to the right port value without the use of
829 AI_NUMERICSERV, so I think we can survive on older glibc's without it.
830 It will allow admins to specify service names as well as port numbers
833 There are uses of AI_NUMERICSERV in gssd and in nfs_svc_create(). The
834 one in nfs_svc_create() is behind HAVE_LIBTIRPC, and the other is a
835 issue only for those who want to deploy Kerberos -- likely in both
836 cases, a more modern glibc will be present. I'm going to leave those
841 https://bugzilla.linux-nfs.org/show_bug.cgi?id=195
843 Reported-by: "Gabor Z. Papp" <gzp@papp.hu>
844 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
845 Signed-off-by: Steve Dickson <steved@redhat.com>
847 commit e8dbaddc8465dcd07b53f8e80a537703dd0248ca
848 Author: Sid Moore <learnmost@gmail.com>
849 Date: Fri Dec 3 09:19:06 2010 -0500
851 rpc.mountd: Checking RPC Procedure ID before process it
853 Signed-off-by: Steve Dickson <steved@redhat.com>
855 commit 3c6973c595d62dc6452967d50ae8abe69f9f8bad
856 Author: Mi Jinlong <mijinlong@cn.fujitsu.com>
857 Date: Mon Nov 29 10:59:10 2010 -0500
859 libnfs.a: fix a bug when parse section's arg
861 When parsing section's arg at configure file, the pointer
862 should stop when fetch ']', and give the warning message.
864 Signed-off-by: Steve Dickson <steved@redhat.com>
866 commit 86f7be64cafd17d4a3f164603484eaedb4757431
867 Author: Harshula Jayasuriya <harshula@redhat.com>
868 Date: Mon Nov 22 11:22:31 2010 -0500
870 nfs-utils: nfsstat: has_stats() does not function correctly for NFSv4 client stats
872 The NFSv4 client procs/ops in "struct rpc_procinfo nfs4_procedures" is
873 used to generate the NFS client stats interface:
874 ------------------------------------------------------------
877 proc2 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
878 proc3 22 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 1 0
879 proc4 42 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
882 ------------------------------------------------------------
883 Note, for proc4, the number 42. That is the number of stats that follow
884 on the same line. Currently nfsstat's has_stats() relies on this number
885 to be equal to CLTPROC4_SZ. Unfortunately this is not the case. I have
886 changed has_stats() not to rely on these two values being equal. This
887 should also allow nfsstat to work with different kernel versions that
888 expose a different number of NFS client ops.
891 * Stop print_clnt_list() printing server stats!
892 * Describe the option -3 and -4 completely in the nfsstat manpage.
894 Signed-off-by: Harshula Jayasuriya <harshula@redhat.com>
895 Signed-off-by: Steve Dickson <steved@redhat.com>
897 commit 0868dcccb9a3bf3d022a32ff31311fe371484e77
898 Author: Steve Dickson <steved@redhat.com>
899 Date: Sat Nov 20 15:01:21 2010 -0500
901 Enable nfsidmap to compile
903 Only enable the compilation of nfsidmap when libnfsidmap support it.
905 Signed-off-by: Steve Dickson <steved@redhat.com>
907 commit 6f07548141e710767d425e119d9823691293771d
908 Author: Bryan Schumaker <bjschuma@netapp.com>
909 Date: Fri Nov 19 12:01:10 2010 -0500
911 Add the new nfsidmap program
913 This patch adds the nfsidmap program to nfs-utils. This program is
914 called by the nfs idmapper through request-keys to map between
915 uid / user name and gid / group name.
917 Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
918 Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
919 Signed-off-by: Steve Dickson <steved@redhat.com>
921 commit 409487978593de13ae36be0ee56d8111ad6b3319
922 Author: Steve Dickson <steved@redhat.com>
923 Date: Mon Nov 22 11:33:37 2010 -0500
925 Removed a couple warnings from utils/mount/stropts.c
927 stropts.c:740:6: warning: 'ret' may be used uninitialized in this function
928 stropts.c:653:6: warning: 'ret' may be used uninitialized in this function
930 Signed-off-by: Steve Dickson <steved@redhat.com>
932 commit f4968a724c1d4162a8e2b9f6a19c460cc56c95f7
933 Author: Chuck Lever <chuck.lever@oracle.com>
934 Date: Fri Oct 29 12:56:21 2010 -0400
936 nfs(5): Document remount behavior
938 It appears that, for a long while, NFS "remount" mounts have
939 completely wiped the existing mount options in /etc/mtab for a given
940 mount point. This is a problem for umount.nfs, since it reads its
941 options out of /etc/mtab to find out how to do the unmount.
943 The mount(8) command provides the NFS mount subcommand with the mount
944 options to perform the remount. There are four cases to consider:
946 1. Both the device and mount directory are specified on the
947 command line, and the target mount point is in /etc/fstab
949 2. Only one of the device and mount directory is specified on
950 the command line, and the target mount point is in
953 3. Both the device and mount directory are specified on the
954 command line, and the target mount point is not in /etc/fstab
956 4. Only one of the device and mount directory is specified on
957 the command line, and the target mount point is not in
960 Currently only case 4 works correctly. In that case, mount(8)
961 provides the correct set of mount options to the mount.nfs
962 subcommand and it can update /etc/mtab correctly.
964 Cases 1 and 3 replace all mount options in /etc/mtab with the options
965 provided on the command line during a remount. Case 2 replaces the
966 mount options in /etc/mtab with a mix of options from /etc/fstab and
969 Cases 1 and 3 are historical behavior. Basically this is a formal
970 interface to allow administrators to replace the mount options in
971 /etc/mtab completely, instead of merging in new ones. The present
972 patch documents that behavior in nfs(5), and provides best practice
973 for remounting NFS mount points.
975 There are near-term plans to address case 2 by fixing mount(8)
976 (provided by utils-linux-ng in most distributions).
978 This is a partial fix for:
980 https://bugzilla.linux-nfs.org/show_bug.cgi?id=188
982 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
983 Signed-off-by: Steve Dickson <steved@redhat.com>
985 commit 6d1a82b005994f759f2c847c0354413a24643da5
986 Author: Chuck Lever <chuck.lever@oracle.com>
987 Date: Thu Oct 28 13:15:22 2010 -0400
989 nfs(5): Grammar and style fixes
991 Clean up grammar and style issues introduced by recent updates. Also,
992 I'm not certain inappropriate options are always ignored.
994 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
995 Signed-off-by: Steve Dickson <steved@redhat.com>
997 commit ab2cdb859f738a25e2567a2ec674cfa78a0a175d
998 Author: Chuck Lever <chuck.lever@oracle.com>
999 Date: Thu Oct 28 13:13:19 2010 -0400
1001 mount.nfs: mnt_freq and mnt_pass are always zero
1005 No need to pass constant zeros to add_mtab() from its only call site.
1006 Ensure that initialization of a struct mntent is consistent in both
1007 places that it is done.
1009 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1010 Signed-off-by: Steve Dickson <steved@redhat.com>
1012 commit bc4a0c42570d5620cc1bb32428e16b9c9b5f3863
1013 Author: Chuck Lever <chuck.lever@oracle.com>
1014 Date: Thu Oct 28 13:10:48 2010 -0400
1016 mount.nfs: Fix memory leak in nfs_sys_mount()
1018 This appears to have been left behind by last year's adjustments to
1019 how the extra_opts string is constructed.
1021 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1022 Signed-off-by: Steve Dickson <steved@redhat.com>
1024 commit 1f237ac72e6f563908b350e11fd2bb866c003028
1025 Author: Chuck Lever <chuck.lever@oracle.com>
1026 Date: Thu Oct 28 13:09:38 2010 -0400
1028 mount: Fix compiler warning in nfs_parse_retry_option()
1030 stropts.c: In function nfs_parse_retry_option:
1031 stropts.c:131: warning: conversion to unsigned int from long int may
1034 Make it more clear what the second argument is for, and flag the
1035 switch fallthrough case.
1037 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1038 Signed-off-by: Steve Dickson <steved@redhat.com>
1040 commit 1ea2c3be33f2eb4630c5cdb78edf2bb670b294ab
1041 Author: Chuck Lever <chuck.lever@oracle.com>
1042 Date: Thu Oct 28 12:12:12 2010 -0400
1044 nfs-utils: Remove all uses of AI_ADDRCONFIG
1046 It was reported that, if only "lo" is up,
1048 mount.nfs 127.0.0.1:/export /mount
1050 fails with "Name or service not known".
1052 "man 3 getaddrinfo" says this:
1054 If hints.ai_flags includes the AI_ADDRCONFIG flag, then IPv4
1055 addresses are returned in the list pointed to by res only if the
1056 local system has at least one IPv4 address configured, and IPv6
1057 addresses are only returned if the local system has at least
1058 one IPv6 address configured.
1060 The man page oversimplifies here. A review of glibc shows that
1061 getaddrinfo(3) explicitly ignores loopback addresses when deciding
1062 whether an IPv4 or IPv6 address is configured.
1064 This behavior around loopback is a problem not just for mount.nfs,
1065 but also for RPC daemons that have to start up before a system's
1066 networking is fully configured and started. Given the history of
1067 other problems with AI_ADDRCONFIG and the unpredictable behavior it
1068 introduces, let's just remove it everywhere in nfs-utils.
1072 https://bugzilla.linux-nfs.org/show_bug.cgi?id=191
1074 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1075 Signed-off-by: Steve Dickson <steved@redhat.com>
1077 commit f8e315543b7f1db7f37a4bfe8ede3020cef62868
1078 Author: Jeff Layton <jlayton@redhat.com>
1079 Date: Thu Oct 28 09:18:33 2010 -0400
1081 nfs-utils: fix default value for --enable-tirpc
1083 We need $enable_tirpc to be a tristate. 'yes' means that someone
1084 explicitly requested building with tirpc. 'no' means that it was
1085 explicitly disabled. Anything else means that no one specified a value.
1087 Fix it by setting the value to a blank string so that the default is
1090 Reported-by: Chuck Lever <chuck.lever@oracle.com>
1091 Signed-off-by: Jeff Layton <jlayton@redhat.com>
1092 Signed-off-by: Steve Dickson <steved@redhat.com>
1094 commit c62d756402509ca5d07c1fd4d2e5a9d78dc4171b
1095 Author: Steve Dickson <steved@redhat.com>
1096 Date: Tue Oct 19 15:54:35 2010 -0400
1098 Updated rpc.mountd man page
1100 Updated the rpc.mountd man page to no longer reference
1101 v3 as the "newer" version and also mentioned v4 as
1102 a supported version.
1104 Signed-off-by: Steve Dickson <steved@redhat.com>
1106 commit 79e9079e9af4e5c2aa1d77815df1147b26876eb8
1107 Author: Steve Dickson <steved@redhat.com>
1108 Date: Tue Oct 19 15:54:04 2010 -0400
1110 Cleared up the sync option in exportfs man page
1112 Signed-off-by: Steve Dickson <steved@redhat.com>
1114 commit 6f228ea26be06572de245aed5496aaa122cca5a8
1115 Author: Steve Dickson <steved@redhat.com>
1116 Date: Fri Oct 15 17:20:28 2010 -0400
1118 Removed duplicate entries in export man page
1120 The man page's paragraphs about "refer=" and "replicas="
1123 Signed-off-by: Steve Dickson <steved@redhat.com>
1125 commit 849b7072a04975bb5da09245fbcacb0cb754a909
1126 Author: Chuck Lever <chuck.lever@oracle.com>
1127 Date: Thu Oct 14 10:33:25 2010 -0400
1129 mountd: Clear mountd registrations at start up
1131 Clear stale MNT registrations before mountd tries to create fresh
1132 listeners, to ensure that mountd starts. This is also what statd does.
1134 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1135 Signed-off-by: Steve Dickson <steved@redhat.com>
1137 commit 93dcf64cc4a9e67f693aea35c8193428015f4a30
1138 Author: Chuck Lever <chuck.lever@oracle.com>
1139 Date: Wed Oct 13 13:57:52 2010 -0400
1141 behavior as file systems that use the monolithic /sbin/mount command.
1142 See the MS_NOMTAB macro in utils-linux-ng/mount/mount.c.
1144 Note that mount(8) has MS_USERS and MS_USER in the "nomtab" category
1145 as well, but mount.nfs needs to record those values so that unmounting
1146 a user-mounted NFS file system can work.
1148 While we're here, fix some white space damage in fix_opts_string().
1150 This is a partial fix for:
1152 https://bugzilla.linux-nfs.org/show_bug.cgi?id=188
1154 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1155 Signed-off-by: Steve Dickson <steved@redhat.com>
1157 commit dc08c702a6c7f824f317af561f491635ee898a71
1158 Author: Chuck Lever <chuck.lever@oracle.com>
1159 Date: Wed Oct 13 13:55:10 2010 -0400
1161 umount.nfs: Distinguish between nfs4 and nfs mounts
1163 Neil Brown reports that umount.nfs is still confused by "-t nfs -o
1166 /etc/mtab can be confused. /proc/mounts is authoritative on the
1167 fstype of a mount. Have umount.nfs consult it to determine which
1168 mechanism to use for unmounting. The code to read /proc/mounts was
1169 lifted from the nfsstat command.
1171 The code introduced by this patch may look like belt-n-suspenders, but
1172 we have two use cases to consider:
1174 1. Old kernels don't support the "vers=4" mount option, so
1175 umount.nfs must look for the "nfs4" fstype
1176 2. Upcoming kernels may eliminate support the "nfs4" fstype, so
1177 umount.nfs must look for the "vers=4" mount option
1179 Thus this logic checks for "nfs4" first then looks for the NFS version
1182 Note that we could handle unmounting entirely in the kernel, but that
1183 won't help older kernels that have this issue.
1186 https://bugzilla.linux-nfs.org/show_bug.cgi?id=189
1188 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1189 Signed-off-by: Steve Dickson <steved@redhat.com>
1191 commit 17962b82afb68ca8e6b0d3f432d36c6c7c4980ea
1192 Author: Chuck Lever <chuck.lever@oracle.com>
1193 Date: Wed Oct 13 13:01:51 2010 -0400
1195 mount.nfs: mountproto does not support RDMA
1197 Clean up. Our client does not support the MNT protocol on RDMA.
1199 nfs_mount_protocol() isn't invoked for RDMA mounts (they are shunted
1200 off before nfs_options2pmap() is invoked). But in case it ever is,
1201 it should return the expected response.
1203 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1204 Signed-off-by: Steve Dickson <steved@redhat.com>
1206 commit 73c61fa5cd114fa6eae0e095724ed63aa66a4a6b
1207 Author: NeilBrown <neilb@suse.de>
1208 Date: Wed Oct 13 12:08:41 2010 -0400
1212 client.c: In function 'init_netmask6':
1213 client.c:181:1: warning: no return statement in function returning
1216 and Suse' build system complained
1218 I: Program returns random data in a function
1219 E: nfs-utils no-return-in-nonvoid-function client.c:181
1221 when I built without --enable-ipv6
1223 Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
1224 Signed-off-by: NeilBrown <neilb@suse.de>
1225 Signed-off-by: Steve Dickson <steved@redhat.com>
1227 commit 7e90281b88c05b01c61152b54a0cf2faec45b09c
1228 Author: Chuck Lever <chuck.lever@oracle.com>
1229 Date: Wed Oct 13 12:02:32 2010 -0400
1231 mount.nfs: Eliminate compiler warnings in utils/mount/network.c
1235 network.c: In function get_socket:
1236 network.c:431: warning: dereferencing type-punned pointer might break
1237 strict-aliasing rules
1239 network.c: In function probe_bothports:
1240 network.c:759: warning: dereferencing type-punned pointer might break
1241 strict-aliasing rules
1242 network.c:762: warning: dereferencing type-punned pointer might break
1243 strict-aliasing rules
1245 network.c: In function nfs_probe_statd:
1246 network.c:775: warning: dereferencing type-punned pointer might break
1247 strict-aliasing rules
1249 network.c: In function nfs_call_umount:
1250 network.c:904: warning: dereferencing type-punned pointer might break
1251 strict-aliasing rules
1253 network.c: In function nfs_ca_sockname:
1254 network.c:1106: warning: dereferencing type-punned pointer might break
1255 strict-aliasing rules
1256 network.c:1112: warning: dereferencing type-punned pointer might break
1257 strict-aliasing rules
1259 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1260 Signed-off-by: Steve Dickson <steved@redhat.com>
1262 commit 57385cf87790c0cbdfddfccdde66bd2c8da45923
1263 Author: Chuck Lever <chuck.lever@oracle.com>
1264 Date: Wed Oct 13 11:59:30 2010 -0400
1266 mount.nfs: Eliminate compiler warning in utils/mount/parse_opt.c
1268 parse_opt.c: In function po_rightmost:
1269 parse_opt.c:517: warning: conversion to int from unsigned int may
1270 change the sign of the result
1272 "i" contains the function's result value, so it should be defined as
1273 the same type as the function's return type.
1275 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1276 Signed-off-by: Steve Dickson <steved@redhat.com>
1278 commit 1ee10ef034cbca86da4df271ac4097a948e7ab59
1279 Author: Chuck Lever <chuck.lever@oracle.com>
1280 Date: Wed Oct 13 11:58:27 2010 -0400
1282 mount.nfs: Eliminate compiler warning in utils/mount/nfsumount.c
1286 nfsumount.c:374: warning: ISO C forbids omitting the middle term of
1289 This is also probably harmless, but let's make the code unambiguous.
1291 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1292 Signed-off-by: Steve Dickson <steved@redhat.com>
1294 commit e9c97e4f7075e563d7a442ca8298ac56bafba0d5
1295 Author: Chuck Lever <chuck.lever@oracle.com>
1296 Date: Wed Oct 13 11:56:58 2010 -0400
1298 mount.nfs: Eliminate compiler warning in utils/mount/nfsumount.c
1302 nfsumount.c:265: warning: no previous prototype for nfsumount
1304 It's also a good idea if the compiler can ensure that the prototype
1305 in nfsmount.h matches the actual function defined in nfsumount.c.
1307 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1308 Signed-off-by: Steve Dickson <steved@redhat.com>
1310 commit e2b6d9cbaf20df26dd371a715fce3ae158f37126
1311 Author: Chuck Lever <chuck.lever@oracle.com>
1312 Date: Wed Oct 13 11:54:49 2010 -0400
1314 mount.nfs: Eliminate compiler warnings in utils/mount/mount.c
1318 mount.c: In function parse_opt:
1319 mount.c:354: warning: conversion to size_t from int may change the
1321 mount.c:354: warning: conversion to int from size_t may change the
1323 mount.c:359: warning: conversion to size_t from int may change the
1325 mount.c:359: warning: conversion to int from size_t may change the
1327 mount.c: In function parse_opts:
1328 mount.c:374: warning: conversion to int from size_t may change the
1330 mount.c:377: warning: conversion to size_t from int may change the
1333 Character string lengths are usually size_t anyway. We can easily
1334 avoid the implicit type cast here.
1336 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1337 Signed-off-by: Steve Dickson <steved@redhat.com>
1339 commit 013e8ec9ffb9f28f97e58299719023faf846a029
1340 Author: Chuck Lever <chuck.lever@oracle.com>
1341 Date: Wed Oct 13 11:53:44 2010 -0400
1343 mount.nfs: Eliminate compiler warning in utils/mount/mount.c
1347 mount.c: At top level:
1348 mount.c:324: warning: no previous prototype for ?mount_usage?
1350 mount_usage() has no callers outside of utils/mount/mount.c and no
1351 prototype is provided in a header file. Make it static.
1353 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1354 Signed-off-by: Steve Dickson <steved@redhat.com>
1356 commit 47480475c99335c1203e81662f815b62573c19e8
1357 Author: Chuck Lever <chuck.lever@oracle.com>
1358 Date: Wed Oct 13 11:50:57 2010 -0400
1360 mount.nfs: Eliminate compiler warnings in utils/mount/version.h
1364 In file included from mount.c:50:
1365 version.h: In function linux_version_code:
1366 version.h:48: warning: conversion to unsigned int from int may
1367 change the sign of the result
1368 version.h:48: warning: conversion to unsigned int from int may
1369 change the sign of the result
1370 version.h:48: warning: conversion to unsigned int from int may
1371 change the sign of the result
1373 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1374 Signed-off-by: Steve Dickson <steved@redhat.com>
1376 commit 00885013dccbe00f5cc4e19223cf18e85a8e616a
1377 Author: Chuck Lever <chuck.lever@oracle.com>
1378 Date: Wed Oct 13 11:44:23 2010 -0400
1380 mount.nfs: Eliminate compiler warning in utils/mount/mount.c
1384 In file included from mount.c:41:
1385 mount_config.h:35: warning: no previous prototype for mount_config_opts
1387 Functions defined in include files are usually declared as "static
1388 inline," eliminating the need for a forward declaration.
1390 While I was there, I also fixed the macro that prevents including
1391 mount_config.h multiple times, and fixed some white space damage.
1393 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1394 Signed-off-by: Steve Dickson <steved@redhat.com>
1396 commit ffa577350b03ddd421d455a8cd4cff86e3718264
1397 Author: Chuck Lever <chuck.lever@oracle.com>
1398 Date: Wed Oct 13 11:38:22 2010 -0400
1400 mount.nfs: Eliminate compiler warnings
1404 fstab.c: In function ?lock_mtab?:
1405 fstab.c:385: warning: declaration of ?errsv? shadows a previous local
1406 fstab.c:367: warning: shadowed declaration is here
1407 fstab.c:407: warning: declaration of ?errsv? shadows a previous local
1408 fstab.c:367: warning: shadowed declaration is here
1409 fstab.c:417: warning: declaration of ?tries? shadows a previous local
1410 fstab.c:325: warning: shadowed declaration is here
1411 fstab.c:422: warning: declaration of ?errsv? shadows a previous local
1412 fstab.c:367: warning: shadowed declaration is here
1414 These are probably harmless. Reusing a variable name, however, is a
1415 little confusing to follow when reading the code.
1417 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1418 Signed-off-by: Steve Dickson <steved@redhat.com>
1420 commit 5fe118b838254023d83424c5010ae73a91ec267d
1421 Author: Trond Myklebust <Trond.Myklebust@netapp.com>
1422 Date: Wed Oct 13 11:27:21 2010 -0400
1424 export: Ensure that we free struct exportent->e_uuid
1426 Currently, the exportent->e_uuid is initialised in
1427 support/nfs/exports.c:parseopts(), but it is never freed.
1429 Also ensure that exportent->e_uuid is duplicated correctly in
1432 Adjusted to account for the new export_free() helper.
1434 Also, e_uuid points to memory that is always allocated with strdup(3),
1435 not with xstrdup(). Thus it must be freed via free(3) and not via
1438 Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
1439 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1440 Signed-off-by: Steve Dickson <steved@redhat.com>
1442 commit 656028f9925f5817c5a37565d27159973db84ec3
1443 Author: Chuck Lever <chuck.lever@oracle.com>
1444 Date: Wed Oct 13 11:22:07 2010 -0400
1446 libnfs.a: Allow multiple RPC listeners to share listener port number
1448 Normally, when "-p" is not specified on the mountd command line, the
1449 TI-RPC library chooses random port numbers for each listener. If a
1450 port number _is_ specified on the command line, all the listeners
1451 will get the same port number, so SO_REUSEADDR needs to be set on
1454 Thus we can't let TI-RPC create the listener sockets for us in this
1455 case; we must create them ourselves and then set SO_REUSEADDR (and
1456 other socket options) by hand.
1458 Different versions of the same RPC program have to share the same
1459 listener and SVCXPRT, so we have to cache xprts we create, and re-use
1460 them when additional requests for registration come from the
1463 Though it doesn't look like it, this fix was "copied" from the legacy
1464 rpc_init() function. It's more complicated for TI-RPC, of course,
1465 since a TI-RPC application can set up listeners with a nearly
1466 arbitrary number of address families and socket types, not just the
1467 two listeners that legacy RPC applications can set up (one for AF_INET
1468 UDP and one for AF_INET TCP).
1471 https://bugzilla.linux-nfs.org/show_bug.cgi?id=190
1473 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1474 Signed-off-by: Steve Dickson <steved@redhat.com>
1476 commit 1296be71ebae4c0d7da61cc1077d97562d3bc549
1477 Author: Steve Dickson <steved@redhat.com>
1478 Date: Wed Oct 13 10:17:58 2010 -0400
1480 nfs-utils: Fixed typo in NFS man page
1482 Chuck pointed out there was a grammar typo in addition to the spelling
1483 typo. Here is a revised version of the patch.
1485 Signed-off-by: Jim Rees <rees@umich.edu>
1486 Signed-off-by: Steve Dickson <steved@redhat.com>
1488 commit 9afbdbad4436df6f7a5a501c6e4db04c3a5bbb08
1489 Author: Steve Dickson <steved@redhat.com>
1490 Date: Wed Oct 13 10:15:12 2010 -0400
1492 Fix style nits in atomicio.c
1494 Signed-off-by: Jim Rees <rees@umich.edu>
1495 Signed-off-by: Steve Dickson <steved@redhat.com>
1497 commit c117b7a1f29db65d139824ba5bab2a58bf5609e2
1498 Author: Steve Dickson <steved@redhat.com>
1499 Date: Wed Oct 13 10:09:53 2010 -0400
1501 nfs-utils: Move common code into support
1503 There are several source files and headers present in the ./utils/idmapd
1504 directory which are also usable in a doimapd daemon. Because of this we
1505 move that support into the support directory such that it can be shared by
1508 Signed-off-by: Jim Rees <rees@umich.edu>
1509 Signed-off-by: Steve Dickson <steved@redhat.com>
1511 commit 8c217b9623c8304608196aeb2f7360abfdf987c8
1512 Author: Suresh Jayaraman <sjayaraman@suse.de>
1513 Date: Wed Sep 29 07:14:14 2010 -0400
1515 The kernel 2.6.37 has a add new mount option: local_lock.
1516 Document the new option in the nfs(5) man page.
1518 Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
1519 Signed-off-by: Steve Dickson <steved@redhat.com>
1521 commit 554df6bc8456c6971c1c597f70d3b9131e4e5e11
1522 Author: Steve Dickson <steved@redhat.com>
1523 Date: Tue Sep 28 08:24:16 2010 -0400
1525 Revert "nfs-iostat.py: don't wait for an extra interval when given a count"
1527 This reverts commit 837796686ad8f9178c7b6855ada728a53ae511e3.