1 commit 057d2add27d1e05fed3ae7206ee043b6c1fda45a
2 Author: NeilBrown <neilb@suse.de>
3 Date: Tue Jun 28 13:24:33 2011 -0400
5 Do not compile unnecessary files when the libmount code is enable
7 Signed-off-by: Steve Dickson <steved@redhat.com>
9 commit 30ebf04700654deddbf5f57d95e84ec69cea8610
10 Author: Prem Karat <prem.karat@linux.vnet.ibm.com>
11 Date: Tue Jun 28 11:53:40 2011 -0400
13 mount.nfs: Don't hard code source and destination
15 Currently souce and destination parameters should be passed as first and
16 second paramter while using mount.nfs. This patch allows them to be passed
17 anywhere while mounting.
19 Current functionality is
20 mount.nfs source destn -o <options>
21 This patch will allow to do this
22 mount.nfs -o <options> source destn
24 mount.nfs -o <options> source -o <options> destn
26 Signed-off-by: Prem Karat <prem.karat@linux.vnet.ibm.com>
27 Signed-off-by: Steve Dickson <steved@redhat.com>
29 commit b3e190c4adfc9ec47567c968bd000d282d07b05e
30 Author: NeilBrown <neilb@suse.de>
31 Date: Tue Jun 28 11:36:31 2011 -0400
33 mount: improve signal management when locking mtab
35 As mount.nfs can run setuid it must be careful about how the user can
36 interact with in. In particular it needs to ensure it does not
37 respond badly to any signals that the user might be able to generate.
39 This is particularly an issue while updating /etc/mtab (when that is
40 not linked to /proc/mounts). If the user can generate a signal which
41 kills mount.nfs while /etc/mtab is locked, then it will leave the file
42 locked, and could possibly corrupt mtab (particularly if 'ulimit 1'
43 was previously issued).
45 Currently lock_mtab does set some handlers for signals, but not
46 enough. It arranges for every signal up to (but not including)
47 SIGCHLD to cause mount.nfs to unlock mdadm promptly exit ... even if
48 the default behaviour would be to ignore the signal. SIGALRM is
49 handled specially, and signals after SIGCHLD are left with their
50 default behaviour. This includes for example SIGXFSZ which can be
51 generated by the user running "ulimit 1".
53 So: change this so that some signals are left unchanged, SIGALRM is
54 handled as required, and all signals that the user can generate are
57 The remainder still cause mount.nfs to print a message, unlock mtab, and
60 Signed-off-by: NeilBrown <neilb@suse.de>
61 Signed-off-by: Steve Dickson <steved@redhat.com>
63 commit a99b7846e2abec5e26ab6b764b921d79559e0a0f
64 Author: J. Bruce Fields <bfields@redhat.com>
65 Date: Mon Jun 27 12:31:07 2011 -0400
67 mountd: move fsidtype-specific code to helpers
69 Now we can move these big switch statements into helper functions.
71 Signed-off-by: J. Bruce Fields <bfields@redhat.com>
72 Signed-off-by: Steve Dickson <steved@redhat.com>
74 commit e6559fd0b7b63f5d152d33d598dc74d78df30ecb
75 Author: J. Bruce Fields <bfields@redhat.com>
76 Date: Mon Jun 27 12:30:36 2011 -0400
78 mountd: gather fsid information into one struct
80 A large part of nfsd_fh() is concerned with extracting
81 fsid-type-specific information from the fsid, then matching that
82 information with information from the export list and the filesystem.
84 Moving all that information into one struct will allow some further
87 Signed-off-by: J. Bruce Fields <bfields@redhat.com>
88 Signed-off-by: Steve Dickson <steved@redhat.com>
90 commit 13a0a61d037f2cc09e7997a96ce5822b9317883b
91 Author: J. Bruce Fields <bfields@redhat.com>
92 Date: Mon Jun 27 12:29:51 2011 -0400
94 mountd: prefer explicit subexports over crossmnt parents
96 If a parent is exported with crossmnt, and if a child is also explicitly
97 exported, then both exports could potentially produce matches in this
98 loop; that isn't a bug.
100 Instead of warning and ignoring the second match we find, we should
101 instead prefer whichever export is deeper in the tree, so that
102 children's options can override those of their parents.
104 Reported-by: Olga Kornievskaia <aglo@citi.umich.edu>
105 Signed-off-by: J. Bruce Fields <bfields@redhat.com>
106 Signed-off-by: Steve Dickson <steved@redhat.com>
108 commit f8d26c1db9a260597828685c7f62e1b29e78285f
109 Author: Jeff Layton <jlayton@redhat.com>
110 Date: Wed Jun 22 15:52:55 2011 -0400
112 manpage: add section on character class matches to exports(5)
114 Signed-off-by: Jeff Layton <jlayton@redhat.com>
115 Signed-off-by: Steve Dickson <steved@redhat.com>
117 commit cb6676aea5bcfcbeaf868e53177eff51f4efe9a8
118 Author: James Pearson <james-p@moving-picture.com>
119 Date: Wed Jun 22 15:51:47 2011 -0400
121 nfs-utils: remove possibly false statement from exports.man
123 A very minor change suggested by J. Bruce Fields <bfields@fieldses.org>
124 to remove the statement that exporting to a single host or IP address is
125 the "most common format" - as it probably isn't.
127 Signed-off-by: James Pearson <james-p@moving-picture.com>
128 Signed-off-by: Steve Dickson <steved@redhat.com>
130 commit 7d71ff8e3825a7f8c2d5c9b5b9344e95e7aa1392
131 Author: Jeff Layton <jlayton@redhat.com>
132 Date: Wed Jun 22 15:51:02 2011 -0400
134 manpage: add info about IPv6 configuration to exports(5)
136 The parts of the exports(5) manpage that discuss IP addressing neglect
137 IPv6 configuration. Update to include info on how to export to IPv6
138 subnets and addresses, and add a line demonstrating that to the EXAMPLE
141 Signed-off-by: Jeff Layton <jlayton@redhat.com>
142 Signed-off-by: Steve Dickson <steved@redhat.com>
144 commit 545ea1098089d6396d5a70111ec231c4de967faa
145 Author: Benny Halevy <benny@tonian.com>
146 Date: Wed Jun 22 15:44:17 2011 -0400
148 nfsstat: reorder nfs4 stats for 2.6.39
150 Signed-off-by: Benny Halevy <benny@tonian.com>
151 Signed-off-by: Steve Dickson <steved@redhat.com>
153 commit c7ce7a4674ad446bee4dd3baf90155ce6b216816
154 Author: Mi Jinlong <mijinlong@cn.fujitsu.com>
155 Date: Wed Jun 22 15:41:27 2011 -0400
157 libexport.a: fix using bad index for loop at cltsetup()
159 In cltsetup(), when checking the address, use clp's naddr for index,
160 instead of cltarg's naddr, which it's always zero there.
162 Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com>
163 Signed-off-by: Steve Dickson <steved@redhat.com>
165 commit b50ad13298b3e9519a9bdecb8c146c9ecf39cef8
166 Author: Jeff Layton <jlayton@redhat.com>
167 Date: Wed Jun 22 14:51:38 2011 -0400
169 nfs: fix host_reliable_addrinfo
171 According to Neil Brown:
173 The point of the word 'reliable' is to check that the name we get
174 really does belong to the host in question - ie that both the
175 forward and reverse maps agree.
177 But the new code doesn't do that check at all. Rather it simply
178 maps the address to a name, then discards the address and maps the
179 name back to a list of addresses and uses that list of addresses as
180 "where the request came from" for permission checking.
182 This bug is exploitable via the following scenario and could allow an
183 attacker access to data that they shouldn't be able to access.
185 Suppose you export a filesystem to some subnet or FQDN and also to a
186 wildcard or netgroup, and I know the details of this (maybe
187 showmount -e tells me) Suppose further that I can get IP packets to
190 Then I create a reverse mapping for my ipaddress to a domain that I
191 own, say "black.hat.org", and a forward mapping from that domain to
192 my IP address, and one of your IP addresses.
194 Then I try to mount your filesystem. The IP address gets correctly
195 mapped to "black.hat.org" and then mapped to both my IP address and
198 Then you search through all of your exports and find that one of the
199 addresses: yours - is allowed to access the filesystem.
201 So you create an export based on the addrinfo you have which allows
202 my IP address the same access as your IP address.
204 Fix this by instead using the forward lookup of the hostname just to
205 verify that the original address is in the list. Then do a numeric
206 lookup using the address and stick the hostname in the ai_canonname.
208 Reviewed-by: NeilBrown <neilb@suse.de>
209 Signed-off-by: Jeff Layton <jlayton@redhat.com>
210 Signed-off-by: Steve Dickson <steved@redhat.com>
212 commit 7235a2164aabfd8dba1f7e1577047bda45053db0
213 Author: James Pearson <james-p@moving-picture.com>
214 Date: Tue Jun 7 16:25:13 2011 -0400
216 exports: Clearly Defining Exports Priorities
218 Added some verbiage to the exports(5) man page
219 that clearly explains the precedence around
220 how exports will work with regard to netgroups.
222 Signed-off-by: Steve Dickson <steved@redhat.com>
224 commit a36f2437caffb23c68fdc0900544d59198bd52b6
225 Author: Neil Brown <neilb@suse.de>
226 Date: Tue Jun 7 13:18:55 2011 -0400
228 Document "-n" for svcgssd
230 The svcgssd man page doesn't mention the "-n" flag.
232 Signed-off-by: NeilBrown <neilb@suse.de>
233 Signed-off-by: Steve Dickson <steved@redhat.com>
235 commit 40aed2c3fb5164195a9975ae7f15ebd4b992839d
236 Author: Pavel Shilovsky <piastry@etersoft.ru>
237 Date: Tue Jun 7 13:18:13 2011 -0400
239 mountd: Fix missing varialble assignment in auth_unix_gid
241 When we get into auth_unix_gid at the second time, groups_len
242 is not 0 and ngroups variable leave as 0. Then we use ngroups
243 in getgrouplist that fails in this case. This patch fixes it.
245 Signed-off-by: Pavel Shilovsky <piastry@etersoft.ru>
246 Signed-off-by: Steve Dickson <steved@redhat.com>
248 commit 8935933dedcd820c2fb3dddff8b79fd5841dc217
249 Author: Benny Halevy <bhalevy@panasas.com>
250 Date: Mon May 23 08:37:17 2011 -0400
252 nfsstat: reorder nfs4 stats for 2.6.38 and up
254 match order in 2.6.38, 2.6.39 (-rc3) and development tree
255 while at it, get rid of obsolete ds_write and ds_commit
257 Signed-off-by: Benny Halevy <bhalevy@panasas.com>
258 Signed-off-by: Steve Dickson <steved@redhat.com>
260 commit 27dcd8a775024160e741cce53e4a402eaac3501d
261 Author: NeilBrown <neilb@suse.de>
262 Date: Mon May 23 08:23:51 2011 -0400
264 supress socket error when address family is not supported
266 From: Suresh Jayaraman <sjayaraman@suse.de>
268 It was observed that when ipv6 module was not loaded and cannot be auto-loaded,
269 when starting NFS server, the following error occurs:
270 "rpc.nfsd: unable to create inet6 TCP socket: errno 97 (Address
271 family not supported by protocol)"
273 This is obviously a true message, but does not represent an "error" when ipv6
274 is not enabled. Rather, it is an expected condition. As such, it can be
275 confusing / misleading / distracting to display it in this scenario.
277 This patch instead of throwing error when a socket call fails with
278 EAFNOSUPPORT, makes it as a NOTICE.
280 Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
281 Signed-off-by: Neil Brown <neilb@suse.de>
282 Signed-off-by: Steve Dickson <steved@redhat.com>
284 commit 7a802337bfc92d0b30fe94dbd0fa231990a26161
285 Author: NeilBrown <neilb@suse.de>
286 Date: Mon May 23 08:19:57 2011 -0400
288 Remove risk of nfs_addmntent corrupting mtab
290 nfs_addmntent is used to append directly to /etc/mtab.
291 If the write partially fail, e.g. due to RLIMIT_FSIZE,
292 truncate back to original size and return an error.
294 See also https://bugzilla.redhat.com/show_bug.cgi?id=697975
295 (CVE-2011-1749) CVE-2011-1749 nfs-utils: mount.nfs fails to anticipate RLIMIT_FSIZE
297 Signed-off-by: NeilBrown <neilb@suse.de>
298 Signed-off-by: Steve Dickson <steved@redhat.com>
300 commit 56f537535190d034039570bafd9a0de71b79b8f1
301 Author: Ben Myers <bpm@sgi.com>
302 Date: Mon May 23 08:07:00 2011 -0400
304 exportfs: getexportent interprets -test-client- as default options
306 With commit 1374c3861abdc66f3a1410e26cc85f86760b51dd Neil added a
307 -test-client- export to test the exportability of filesystems when exportfs
308 is run. When using the old cache controls (i.e. /proc/fs/nfsd is not
309 mounted) exportfs will read /proc/fs/nfs/exports to process existing
310 exports and find these test client entries. The dash at the beginning of
311 -test-client- will be cause getexportent to look for default options in the
312 rest of the string, which test-client- will not match:
314 exportfs: /proc/fs/nfs/exports:1: unknown keyword "test-client-(rw"
316 This patch resolves that problem (as Steve suggested) by not processing any
317 default options if we are reading the list of existing exports from the
318 kernel. Default options are converted to individual exports by exportfs so
319 the kernel won't have any regardless.
321 Signed-off-by: Ben Myers <bpm@sgi.com>
322 Signed-off-by: Steve Dickson <steved@redhat.com>
324 commit c8e802c036a3f0fcd4481dae8b3ec09fb71f4118
325 Author: Jim Rees <rees@umich.edu>
326 Date: Wed May 18 12:42:02 2011 -0400
328 Removed compilation warnings from mountd/cache.c
330 Commit 5604b35a6 introduced a number of missing initializer
331 warnings that were missed. This patch removes those warnings.
333 Signed-off-by: Steve Dickson <steved@redhat.com>
335 commit 68f7938c85258a8c54b13169dcdeae61cc1bf286
336 Author: Steve Dickson <steved@redhat.com>
337 Date: Tue Apr 26 13:32:35 2011 -0400
339 nfsstat: Output headings mislabled
341 The badclnt and badauth headers were reversed
342 when the server side rpc stats (-s -o rpc) were
345 Signed-off-by: Steve Dickson <steved@redhat.com>
347 commit 2e4a7e9b50a641dc8640a5dd911ee4a2f0b2c073
348 Author: Steve Dickson <steved@redhat.com>
349 Date: Tue Apr 19 12:31:30 2011 -0400
351 rpc.svcgssd: Segmentation fault on error
353 Commit 544ed73d introduced a regression that caused
354 rpc.svcgssd to seg fault on "Wrong principal in request"
355 errors in gss_accept_sec_context()
357 Signed-off-by: Steve Dickson <steved@redhat.com>
359 commit 5604b35a61e22930873ffc4e9971002f578e7978
360 Author: Sean Finney <sean.finney@sonyericsson.com>
361 Date: Tue Apr 19 11:04:35 2011 -0400
363 nfs-utils: Increase the stdio file buffer size for procfs files
365 Previously, when writing to /proc/net/rpc/*/channel, if a cache line
366 were larger than the default buffer size (likely 1024 bytes), mountd
367 and svcgssd would split writes into a number of buffer-sized writes.
368 Each of these writes would get an EINVAL error back from the kernel
369 procfs handle (it expects line-oriented input and does not account for
370 multiple/split writes), and no cache update would occur.
372 When such behavior occurs, NFS clients depending on mountd to finish
373 the cache operation would block/hang, or receive EPERM, depending on
374 the context of the operation. This is likely to happen if a user is a
375 member of a large (~100-200) number of groups.
377 Instead, every fopen() on the procfs files in question is followed by
378 a call to setvbuf(), using a per-file dedicated buffer of
379 RPC_CHAN_BUF_SIZE length.
381 Really, mountd should not be using stdio-style buffered file operations
382 on files in /proc to begin with. A better solution would be to use
383 internally managed buffers and calls to write() instead of these stdio
384 calls, but that would be a more extensive change; so this is proposed
385 as a quick and not-so-dirty fix in the meantime.
387 Signed-off-by: Sean Finney <sean.finney@sonyericsson.com>
388 Signed-off-by: Steve Dickson <steved@redhat.com>
390 commit 9274e94db85bac04e170414cb8e0f4be271cde90
391 Author: Sean Finney <sean.finney@sonyericsson.com>
392 Date: Tue Apr 19 11:05:47 2011 -0400
394 mountd: Use a dynamic buffer for storing lists of gid's
396 Previously, in auth_unix_gid, group lists were stored in an array of
397 hard-coded length 100, and in the situation that the group lists for a
398 particular call were too large, the array was swapped with a dynamically
399 allocated/freed buffer. For environments where users are commonly in
400 a large number of groups, this isn't an ideal approach.
402 Instead, use malloc/realloc to grow the list on an as-needed basis.
404 Signed-off-by: Sean Finney <sean.finney@sonyericsson.com>
405 Signed-off-by: Steve Dickson <steved@redhat.com>
407 commit a99269230a0e77e7bed4fa31c9547f0d61c7f206
408 Author: Karel Zak <kzak@redhat.com>
409 Date: Wed Apr 6 12:39:21 2011 -0400
411 mount: add --enable-libmount-mount
413 This patch allows to link mount.nfs with libmount from util-linux >=
414 v2.19. The new libmount based code is enabled by CONFIG_LIBMOUNT and
415 is stored in mount_libmount.c. The old code is not affected by this
418 The libmount does not have officially stable API yet, so the
419 --enable-libmount-mount is marked as experimental in the configure
422 The ./configure option is the same as we use in util-linux to enable
423 support for libmount in mount(8).
425 The addr= (and some other options necessary for remount/umount) are
426 stored to /etc/mtab or to /dev/.mount/utab. The utab file is *private*
427 libmount file. It's possible that some mount options (for example
428 user=) will be moved to kernel, so the utab will not be necessary.
432 * supports systems without and with regular /etc/mtab
433 * does not store VFS and FS mount options in userspace
434 * manages user= option and evaluate permissions
435 * parses VFS mount options and generate MS_* flags
436 * parses /etc/{fstab,mtab}, /proc/mounts or /proc/self/mountinfo
437 * long-term goal is to use the same code in all mount.<type> helpers
441 LIBMOUNT_DEBUG=0xffff mount.nfs foo:/path /path
443 to debug the library.
445 On systems with util-linux v2.19 the findmnt(8) command uses libmount
446 to list all/selected mount points:
449 $ findmnt --mtab /path
451 the --mtab appends userspace mount options (e.g. user=) to the output.
453 CC: Chuck Lever <chuck.lever@oracle.com>
454 Signed-off-by: Karel Zak <kzak@redhat.com>
455 Signed-off-by: Steve Dickson <steved@redhat.com>
457 commit c01e5ca6179b8f5b041605d9bbd75a0f76812d54
458 Author: Karel Zak <kzak@redhat.com>
459 Date: Wed Apr 6 11:36:40 2011 -0400
461 mount: move generic functions to utils.c and network.c
463 Move generic code that could be shared between standard mount.nfs and
464 libmount version to utils.c and network.c.
466 CC: Chuck Lever <chuck.lever@oracle.com>
467 Signed-off-by: Karel Zak <kzak@redhat.com>
468 Signed-off-by: Steve Dickson <steved@redhat.com>
470 commit d6c1b35c6b40243bfd6fba2591c9f8f2653078c0
471 Author: Kevin Coffman <kwc@citi.umich.edu>
472 Date: Wed Apr 6 11:25:03 2011 -0400
474 nfs-utils: Add support to svcgssd to limit the negotiated enctypes
476 Recent versions of Kerberos libraries negotiate and use
477 an "acceptor subkey". This negotiation does not consider
478 that a service may have limited the encryption keys in its
479 keytab. A patch (http://src.mit.edu/fisheye/changelog/krb5/?cs=24603)
480 has been added to the MIT Kerberos code to allow an application
481 to indicate that it wants to limit the encryption types negotiated.
482 (This functionality has been available on the client/initiator
483 side for a while. The new patch adds this support to the
484 server/acceptor side.)
486 This patch adds support to read a recently added nfsd
487 proc file to determine the encryption types supported by
488 the kernel and calls the function to limit encryption
489 types negotiated for the acceptor subkey.
491 Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
492 Signed-off-by: Steve Dickson <steved@redhat.com>
494 commit 73840ef610accf4cf667427bc64805377c0d8394
495 Author: Chuck Lever <chuck.lever@oracle.com>
496 Date: Wed Apr 6 10:53:57 2011 -0400
498 exports: add a configurable time-to-live for the kernel cache entries
500 From: Trond Myklebust <Trond.Myklebust@netapp.com>
502 The fedfs ldap server will specify a ttl for its entries.
504 Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
506 This is a refactoring change only. There should be no change in
509 Original patch had updates to utils/mountd/junctions.c, which no
510 longer exists. These are not included here.
512 Create a macro for the default cache TTL, which is used in several
513 places besides the export cache.
517 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
518 Signed-off-by: Steve Dickson <steved@redhat.com>
520 commit 38e4c685410885a6d464ddd44eff4fd5e7f8459f
521 Author: Chuck Lever <chuck.lever@oracle.com>
522 Date: Wed Apr 6 10:49:52 2011 -0400
524 statd: Remove vestigial "-w" option from man page synopsis
526 The synopsis of rpc.statd in its man page lists "-w" as a valid
527 option. There is currently no support in the source code for a "-w"
530 BugLink: https://bugzilla.linux-nfs.org/show_bug.cgi?id=199
532 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
533 Signed-off-by: Steve Dickson <steved@redhat.com>
535 commit b57cd77c13831051ad974ae027d96cd88a8d0c59
536 Author: Chuck Lever <chuck.lever@oracle.com>
537 Date: Wed Apr 6 10:48:38 2011 -0400
539 mount.nfs: Don't leak socket in nfs_ca_sockname()
541 Ensure the test socket is always closed before nfs_ca_sockname()
542 returns. Otherwise it's orphaned.
544 BugLink: https://bugzilla.linux-nfs.org/show_bug.cgi?id=197
546 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
547 Signed-off-by: Steve Dickson <steved@redhat.com>
549 commit 502eef09d8050ffb87d394397c2780e1ef042d68
550 Author: Steve Dickson <steved@redhat.com>
551 Date: Wed Apr 6 10:46:06 2011 -0400
553 Removed a warning from v4root.c
555 v4root.c:176:9: warning: variable 'ret' set but not used
557 Signed-off-by: Steve Dickson <steved@redhat.com>
559 commit c2fa189a0467c25666f014cf9ff2576a9f54d682
560 Author: Steve Dickson <steved@redhat.com>
561 Date: Wed Apr 6 10:39:10 2011 -0400
563 Removed a warning from exportfs.c
565 exportfs.c:280:29: warning: 'exp' may be used uninitialized in this function
567 Signed-off-by: Steve Dickson <steved@redhat.com>
569 commit b98ae9df8c4904289c9390288325058b24caa423
570 Author: Steve Dickson <steved@redhat.com>
571 Date: Wed Apr 6 10:36:30 2011 -0400
573 Removed a warning from conffile.c
575 conffile.c:258:19: warning: 'j' may be used uninitialized in this function
577 Signed-off-by: Steve Dickson <steved@redhat.com>
579 commit 012e1a4bf2a002e8cd4d5be3478bfa20a91cbfed
580 Author: Masatake YAMATO <yamato@redhat.com>
581 Date: Mon Mar 7 08:36:19 2011 -0500
583 Update man pages for /etc/exports.d
585 Man page updates for /etc/exports.d.
587 Signed-off-by: Masatake YAMATO <yamato@redhat.com>
588 Signed-off-by: Steve Dickson <steved@redhat.com>
590 commit c7427b57e2be8ef0d57ad0618d4590c062b130f5
591 Author: Masatake YAMATO <yamato@redhat.com>
592 Date: Mon Mar 7 08:18:51 2011 -0500
594 Read /etc/exports.d/*.export as extra export files
596 This patch adding a capability to read /etc/exports.d/*.exports as
597 extra export files to exportfs.
599 If one wants to add or remove an export entry in a script, currently
600 one may have to use sed or something tool for adding or removing the
601 line for the entry in /etc/exports file.
603 With the patch, adding and removing an entry from a script is much
605 cat<<EOF... or mv can be used for adding. rm can be used for removing.
607 Signed-off-by: Masatake YAMATO <yamato@redhat.com>
608 Signed-off-by: Steve Dickson <steved@redhat.com>
610 commit edb9b7f2ab9806afb9af31eabeb505fe454c51df
611 Author: Steve Dickson <steved@redhat.com>
612 Date: Sat Mar 5 16:17:01 2011 -0500
614 Cleaned up a warning in rpcdispatch.c
616 rpcdispatch.c:40:20: warning: comparison between signed and unsigned
619 Signed-off-by: Steve Dickson <steved@redhat.com>
621 commit 930323817b61877d61fb8ef57229013daa2e6091
622 Author: Steve Dickson <steved@redhat.com>
623 Date: Sat Mar 5 16:13:01 2011 -0500
625 mount: Remove MOUNT_CONFIG warnings
627 The following changes are needed to remove compile warnings when
628 MOUNT_CONFIG is not defined
630 Signed-off-by: Steve Dickson <steved@redhat.com>
632 commit 3ef3dc8f1e87ba7a6eaa3c2a6965aff6c80ba414
633 Author: Chuck Lever <chuck.lever@oracle.com>
634 Date: Thu Mar 3 17:26:33 2011 -0500
636 mount: Recognize zero as a valid value for the port= option
638 While zero is not a valid IP port number, zero does represent a valid
639 value for "port=". It means "query rpcbind to discover the actual
640 non-zero port number to use". So the parsing functions that handle
641 "port=" should not flag zero as an invalid value.
643 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
644 Signed-off-by: Steve Dickson <steved@redhat.com>
646 commit b3a4dbfb61dad59829f5191d727267b2ea45937a
647 Author: Mi Jinlong <mijinlong@cn.fujitsu.com>
648 Date: Wed Feb 9 11:29:42 2011 -0500
650 Gssd: modify wrong err message at handle_gssd_upcall
652 Modify wrong err message at handle_gssd_upcall when
653 sscanf encryption types fail.
655 Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com>
656 Signed-off-by: Steve Dickson <steved@redhat.com>
658 commit 45e4597bd570ed40221f51887cde7d7f096f55e7
659 Author: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
660 Date: Wed Feb 9 11:27:19 2011 -0500
662 Support AD style kerberos automatically in rpc.gss
664 An Active Directory KDC will only grant a TGT for UPNs, getting
665 a TGT for SPNs is not possible:
667 $ kinit -k host/ib5@ADS.ORCORP.CA
668 kinit: Client not found in Kerberos database while getting initial
671 The correct thing to do for machine credentials is to get a TGT
672 for the computer UPN <HOSTNAME>$@REALM:
675 12/22/10 11:43:47 12/22/10 21:43:47 krbtgt/ADS.ORCORP.CA@ADS.ORCORP.CA
677 Samba automatically creates /etc/krb5.keytab entry for the computer UPN,
678 this patch makes gssd_refresh_krb5_machine_credential prefer it above
679 the SPNs if it is present.
681 The net result is that nfs client works automatically out of the box
682 if samba has been used to setup kerberos via 'net ads join' 'net ads
685 Tested using Windows Server 2003 R2 as the AD server.
687 Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
688 Signed-off-by: Steve Dickson <steved@redhat.com>
690 commit 730f6986f86873513fa021a450eb55ccd0f2fbff
691 Author: Steve Dickson <steved@redhat.com>
692 Date: Wed Jan 26 07:49:19 2011 -0500
694 Fixed segfault in rpc.mountd
696 A unallocated piece of memory, instead of a NULL point, was being
697 used to initialize a ->next point in the mount link list which
698 caused a segfault after a few remote accesses via the showmount
701 Signed-off-by: Steve Dickson <steved@redhat.com>
703 commit 544ed73d5ab27c1390833d5cf93b9585c151667d
704 Author: Steve Dickson <steved@redhat.com>
705 Date: Fri Jan 14 10:12:28 2011 -0500
707 Improve debugging in svcgssd
709 Added in gss_display_error() which translates the GSS error into the
710 actual GSS macro name. Currently only the translation of these errors
711 are logged. Since those translations are buried deep in the kerberos
712 library code, having the actual GSS macro name makes it easier to
715 Moved the nfs4_init_name_mapping() call into main() so if debug is
716 enabled the DNS name and realms will be logged during start up.
718 Signed-off-by: Steve Dickson <steved@redhat.com>
720 commit 57be18b9ab08148a1cc9d5af588119885720be8b
721 Author: Mi Jinlong <mijinlong@cn.fujitsu.com>
722 Date: Tue Jan 4 11:16:45 2011 -0500
724 libnsm.a: modify return value to false from 0 at nsm_drop_privileges()
726 At nsm_drop_privileges(), for improving readability, unify
729 Signed-off-by: Mi Jinlong <mijinlong@cn.fujitsu.com>
730 Signed-off-by: Steve Dickson <steved@redhat.com>
732 commit 5c498280fd9353ded3ea169841079bdae23418e2
733 Author: Chuck Lever <chuck.lever@oracle.com>
734 Date: Mon Dec 13 14:50:45 2010 -0500
736 libnsm.a: sm-notify sometimes ignores monitored hosts
738 Monitored host information is stored in files under /var/lib/nfs.
739 When visiting entries in the monitored hosts directory, libnsm.a
740 examines the value of dirent.d_type to determine if an entry is a
743 According to readdir(3), the d_type field is not supported by all
744 file system types. My root file system happens to be one where d_type
745 isn't supported. Typical installations that use an ext-derived root
746 file system are not exposed to this issue, but those who use xfs, for
749 On such file systems, not only are remote peers not notified of
750 reboots, but the NSM state number is never incremented. A statd warm
751 restart would not re-monitor any hosts that were monitored before
754 When writing support/nsm/file.c, I copied the use of d_type from the
755 original statd code, so this has likely been an issue for some time.
757 Replace the use of d_type in support/nsm/file.c with a call to
758 lstat(2). It's extra code, but is guaranteed to work on all file
761 Note there is a usage of d_type in gssd. I'll let gssd and rpcpipefs
762 experts decide whether that's worth changing.
766 https://bugzilla.linux-nfs.org/show_bug.cgi?id=193
768 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
769 Signed-off-by: Steve Dickson <steved@redhat.com>
771 commit 089df7c754d9ebab0a7b804f396626ac95fee2e6
772 Author: Chuck Lever <chuck.lever@oracle.com>
773 Date: Mon Dec 13 14:47:42 2010 -0500
775 libnsm.a: Replace __attribute_noinline__
777 Replace the __attribute_noinline__ form with
779 __attribute__((__noinline__)).
781 Even though the compiler didn't complain about __attribute_malloc__,
782 also replace those in order to maintain consistent style throughout the
787 https://bugzilla.linux-nfs.org/show_bug.cgi?id=194
789 Reported-by: "Gabor Z. Papp" <gzp@papp.hu>
790 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
791 Signed-off-by: Steve Dickson <steved@redhat.com>
793 commit 7869a76207d3f4b3bd4ab57b4a7a8807ac2ff0c6
794 Author: Chuck Lever <chuck.lever@oracle.com>
795 Date: Mon Dec 13 14:36:15 2010 -0500
797 sm-notify: Make use of AI_NUMERICSERV conditional
799 Gabor Papp reports nfs-utils-1.2.3 doesn't build on his system that
802 make[3]: Entering directory
803 `/home/gzp/src/nfs-utils-1.2.3/utils/statd'
804 gcc -DHAVE_CONFIG_H -I. -I../../support/include -D_GNU_SOURCE -Wall
805 -Wextra -Wstrict-prototypes -pipe -g -O2 -MT sm-notify.o -MD
806 -MP -MF .deps/sm-notify.Tpo -c -o sm-notify.o sm-notify.c
807 sm-notify.c: In function 'smn_bind_address':
808 sm-notify.c:247: error: 'AI_NUMERICSERV' undeclared (first use in this
810 sm-notify.c:247: error: (Each undeclared identifier is reported only
812 sm-notify.c:247: error: for each function it appears in.)
813 make[3]: *** [sm-notify.o] Error 1
815 According to the getaddrinfo(3) man page, AI_NUMERICSERV is available
816 only since glibc 2.3.4. getaddrinfo(3) seems to convert strings
817 containing a number to the right port value without the use of
818 AI_NUMERICSERV, so I think we can survive on older glibc's without it.
819 It will allow admins to specify service names as well as port numbers
822 There are uses of AI_NUMERICSERV in gssd and in nfs_svc_create(). The
823 one in nfs_svc_create() is behind HAVE_LIBTIRPC, and the other is a
824 issue only for those who want to deploy Kerberos -- likely in both
825 cases, a more modern glibc will be present. I'm going to leave those
830 https://bugzilla.linux-nfs.org/show_bug.cgi?id=195
832 Reported-by: "Gabor Z. Papp" <gzp@papp.hu>
833 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
834 Signed-off-by: Steve Dickson <steved@redhat.com>
836 commit e8dbaddc8465dcd07b53f8e80a537703dd0248ca
837 Author: Sid Moore <learnmost@gmail.com>
838 Date: Fri Dec 3 09:19:06 2010 -0500
840 rpc.mountd: Checking RPC Procedure ID before process it
842 Signed-off-by: Steve Dickson <steved@redhat.com>
844 commit 3c6973c595d62dc6452967d50ae8abe69f9f8bad
845 Author: Mi Jinlong <mijinlong@cn.fujitsu.com>
846 Date: Mon Nov 29 10:59:10 2010 -0500
848 libnfs.a: fix a bug when parse section's arg
850 When parsing section's arg at configure file, the pointer
851 should stop when fetch ']', and give the warning message.
853 Signed-off-by: Steve Dickson <steved@redhat.com>
855 commit 86f7be64cafd17d4a3f164603484eaedb4757431
856 Author: Harshula Jayasuriya <harshula@redhat.com>
857 Date: Mon Nov 22 11:22:31 2010 -0500
859 nfs-utils: nfsstat: has_stats() does not function correctly for NFSv4 client stats
861 The NFSv4 client procs/ops in "struct rpc_procinfo nfs4_procedures" is
862 used to generate the NFS client stats interface:
863 ------------------------------------------------------------
866 proc2 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
867 proc3 22 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 1 0
868 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
871 ------------------------------------------------------------
872 Note, for proc4, the number 42. That is the number of stats that follow
873 on the same line. Currently nfsstat's has_stats() relies on this number
874 to be equal to CLTPROC4_SZ. Unfortunately this is not the case. I have
875 changed has_stats() not to rely on these two values being equal. This
876 should also allow nfsstat to work with different kernel versions that
877 expose a different number of NFS client ops.
880 * Stop print_clnt_list() printing server stats!
881 * Describe the option -3 and -4 completely in the nfsstat manpage.
883 Signed-off-by: Harshula Jayasuriya <harshula@redhat.com>
884 Signed-off-by: Steve Dickson <steved@redhat.com>
886 commit 0868dcccb9a3bf3d022a32ff31311fe371484e77
887 Author: Steve Dickson <steved@redhat.com>
888 Date: Sat Nov 20 15:01:21 2010 -0500
890 Enable nfsidmap to compile
892 Only enable the compilation of nfsidmap when libnfsidmap support it.
894 Signed-off-by: Steve Dickson <steved@redhat.com>
896 commit 6f07548141e710767d425e119d9823691293771d
897 Author: Bryan Schumaker <bjschuma@netapp.com>
898 Date: Fri Nov 19 12:01:10 2010 -0500
900 Add the new nfsidmap program
902 This patch adds the nfsidmap program to nfs-utils. This program is
903 called by the nfs idmapper through request-keys to map between
904 uid / user name and gid / group name.
906 Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
907 Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
908 Signed-off-by: Steve Dickson <steved@redhat.com>
910 commit 409487978593de13ae36be0ee56d8111ad6b3319
911 Author: Steve Dickson <steved@redhat.com>
912 Date: Mon Nov 22 11:33:37 2010 -0500
914 Removed a couple warnings from utils/mount/stropts.c
916 stropts.c:740:6: warning: 'ret' may be used uninitialized in this function
917 stropts.c:653:6: warning: 'ret' may be used uninitialized in this function
919 Signed-off-by: Steve Dickson <steved@redhat.com>
921 commit f4968a724c1d4162a8e2b9f6a19c460cc56c95f7
922 Author: Chuck Lever <chuck.lever@oracle.com>
923 Date: Fri Oct 29 12:56:21 2010 -0400
925 nfs(5): Document remount behavior
927 It appears that, for a long while, NFS "remount" mounts have
928 completely wiped the existing mount options in /etc/mtab for a given
929 mount point. This is a problem for umount.nfs, since it reads its
930 options out of /etc/mtab to find out how to do the unmount.
932 The mount(8) command provides the NFS mount subcommand with the mount
933 options to perform the remount. There are four cases to consider:
935 1. Both the device and mount directory are specified on the
936 command line, and the target mount point is in /etc/fstab
938 2. Only one of the device and mount directory is specified on
939 the command line, and the target mount point is in
942 3. Both the device and mount directory are specified on the
943 command line, and the target mount point is not in /etc/fstab
945 4. Only one of the device and mount directory is specified on
946 the command line, and the target mount point is not in
949 Currently only case 4 works correctly. In that case, mount(8)
950 provides the correct set of mount options to the mount.nfs
951 subcommand and it can update /etc/mtab correctly.
953 Cases 1 and 3 replace all mount options in /etc/mtab with the options
954 provided on the command line during a remount. Case 2 replaces the
955 mount options in /etc/mtab with a mix of options from /etc/fstab and
958 Cases 1 and 3 are historical behavior. Basically this is a formal
959 interface to allow administrators to replace the mount options in
960 /etc/mtab completely, instead of merging in new ones. The present
961 patch documents that behavior in nfs(5), and provides best practice
962 for remounting NFS mount points.
964 There are near-term plans to address case 2 by fixing mount(8)
965 (provided by utils-linux-ng in most distributions).
967 This is a partial fix for:
969 https://bugzilla.linux-nfs.org/show_bug.cgi?id=188
971 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
972 Signed-off-by: Steve Dickson <steved@redhat.com>
974 commit 6d1a82b005994f759f2c847c0354413a24643da5
975 Author: Chuck Lever <chuck.lever@oracle.com>
976 Date: Thu Oct 28 13:15:22 2010 -0400
978 nfs(5): Grammar and style fixes
980 Clean up grammar and style issues introduced by recent updates. Also,
981 I'm not certain inappropriate options are always ignored.
983 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
984 Signed-off-by: Steve Dickson <steved@redhat.com>
986 commit ab2cdb859f738a25e2567a2ec674cfa78a0a175d
987 Author: Chuck Lever <chuck.lever@oracle.com>
988 Date: Thu Oct 28 13:13:19 2010 -0400
990 mount.nfs: mnt_freq and mnt_pass are always zero
994 No need to pass constant zeros to add_mtab() from its only call site.
995 Ensure that initialization of a struct mntent is consistent in both
996 places that it is done.
998 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
999 Signed-off-by: Steve Dickson <steved@redhat.com>
1001 commit bc4a0c42570d5620cc1bb32428e16b9c9b5f3863
1002 Author: Chuck Lever <chuck.lever@oracle.com>
1003 Date: Thu Oct 28 13:10:48 2010 -0400
1005 mount.nfs: Fix memory leak in nfs_sys_mount()
1007 This appears to have been left behind by last year's adjustments to
1008 how the extra_opts string is constructed.
1010 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1011 Signed-off-by: Steve Dickson <steved@redhat.com>
1013 commit 1f237ac72e6f563908b350e11fd2bb866c003028
1014 Author: Chuck Lever <chuck.lever@oracle.com>
1015 Date: Thu Oct 28 13:09:38 2010 -0400
1017 mount: Fix compiler warning in nfs_parse_retry_option()
1019 stropts.c: In function nfs_parse_retry_option:
1020 stropts.c:131: warning: conversion to unsigned int from long int may
1023 Make it more clear what the second argument is for, and flag the
1024 switch fallthrough case.
1026 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1027 Signed-off-by: Steve Dickson <steved@redhat.com>
1029 commit 1ea2c3be33f2eb4630c5cdb78edf2bb670b294ab
1030 Author: Chuck Lever <chuck.lever@oracle.com>
1031 Date: Thu Oct 28 12:12:12 2010 -0400
1033 nfs-utils: Remove all uses of AI_ADDRCONFIG
1035 It was reported that, if only "lo" is up,
1037 mount.nfs 127.0.0.1:/export /mount
1039 fails with "Name or service not known".
1041 "man 3 getaddrinfo" says this:
1043 If hints.ai_flags includes the AI_ADDRCONFIG flag, then IPv4
1044 addresses are returned in the list pointed to by res only if the
1045 local system has at least one IPv4 address configured, and IPv6
1046 addresses are only returned if the local system has at least
1047 one IPv6 address configured.
1049 The man page oversimplifies here. A review of glibc shows that
1050 getaddrinfo(3) explicitly ignores loopback addresses when deciding
1051 whether an IPv4 or IPv6 address is configured.
1053 This behavior around loopback is a problem not just for mount.nfs,
1054 but also for RPC daemons that have to start up before a system's
1055 networking is fully configured and started. Given the history of
1056 other problems with AI_ADDRCONFIG and the unpredictable behavior it
1057 introduces, let's just remove it everywhere in nfs-utils.
1061 https://bugzilla.linux-nfs.org/show_bug.cgi?id=191
1063 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1064 Signed-off-by: Steve Dickson <steved@redhat.com>
1066 commit f8e315543b7f1db7f37a4bfe8ede3020cef62868
1067 Author: Jeff Layton <jlayton@redhat.com>
1068 Date: Thu Oct 28 09:18:33 2010 -0400
1070 nfs-utils: fix default value for --enable-tirpc
1072 We need $enable_tirpc to be a tristate. 'yes' means that someone
1073 explicitly requested building with tirpc. 'no' means that it was
1074 explicitly disabled. Anything else means that no one specified a value.
1076 Fix it by setting the value to a blank string so that the default is
1079 Reported-by: Chuck Lever <chuck.lever@oracle.com>
1080 Signed-off-by: Jeff Layton <jlayton@redhat.com>
1081 Signed-off-by: Steve Dickson <steved@redhat.com>
1083 commit c62d756402509ca5d07c1fd4d2e5a9d78dc4171b
1084 Author: Steve Dickson <steved@redhat.com>
1085 Date: Tue Oct 19 15:54:35 2010 -0400
1087 Updated rpc.mountd man page
1089 Updated the rpc.mountd man page to no longer reference
1090 v3 as the "newer" version and also mentioned v4 as
1091 a supported version.
1093 Signed-off-by: Steve Dickson <steved@redhat.com>
1095 commit 79e9079e9af4e5c2aa1d77815df1147b26876eb8
1096 Author: Steve Dickson <steved@redhat.com>
1097 Date: Tue Oct 19 15:54:04 2010 -0400
1099 Cleared up the sync option in exportfs man page
1101 Signed-off-by: Steve Dickson <steved@redhat.com>
1103 commit 6f228ea26be06572de245aed5496aaa122cca5a8
1104 Author: Steve Dickson <steved@redhat.com>
1105 Date: Fri Oct 15 17:20:28 2010 -0400
1107 Removed duplicate entries in export man page
1109 The man page's paragraphs about "refer=" and "replicas="
1112 Signed-off-by: Steve Dickson <steved@redhat.com>
1114 commit 849b7072a04975bb5da09245fbcacb0cb754a909
1115 Author: Chuck Lever <chuck.lever@oracle.com>
1116 Date: Thu Oct 14 10:33:25 2010 -0400
1118 mountd: Clear mountd registrations at start up
1120 Clear stale MNT registrations before mountd tries to create fresh
1121 listeners, to ensure that mountd starts. This is also what statd does.
1123 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1124 Signed-off-by: Steve Dickson <steved@redhat.com>
1126 commit 93dcf64cc4a9e67f693aea35c8193428015f4a30
1127 Author: Chuck Lever <chuck.lever@oracle.com>
1128 Date: Wed Oct 13 13:57:52 2010 -0400
1130 behavior as file systems that use the monolithic /sbin/mount command.
1131 See the MS_NOMTAB macro in utils-linux-ng/mount/mount.c.
1133 Note that mount(8) has MS_USERS and MS_USER in the "nomtab" category
1134 as well, but mount.nfs needs to record those values so that unmounting
1135 a user-mounted NFS file system can work.
1137 While we're here, fix some white space damage in fix_opts_string().
1139 This is a partial fix for:
1141 https://bugzilla.linux-nfs.org/show_bug.cgi?id=188
1143 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1144 Signed-off-by: Steve Dickson <steved@redhat.com>
1146 commit dc08c702a6c7f824f317af561f491635ee898a71
1147 Author: Chuck Lever <chuck.lever@oracle.com>
1148 Date: Wed Oct 13 13:55:10 2010 -0400
1150 umount.nfs: Distinguish between nfs4 and nfs mounts
1152 Neil Brown reports that umount.nfs is still confused by "-t nfs -o
1155 /etc/mtab can be confused. /proc/mounts is authoritative on the
1156 fstype of a mount. Have umount.nfs consult it to determine which
1157 mechanism to use for unmounting. The code to read /proc/mounts was
1158 lifted from the nfsstat command.
1160 The code introduced by this patch may look like belt-n-suspenders, but
1161 we have two use cases to consider:
1163 1. Old kernels don't support the "vers=4" mount option, so
1164 umount.nfs must look for the "nfs4" fstype
1165 2. Upcoming kernels may eliminate support the "nfs4" fstype, so
1166 umount.nfs must look for the "vers=4" mount option
1168 Thus this logic checks for "nfs4" first then looks for the NFS version
1171 Note that we could handle unmounting entirely in the kernel, but that
1172 won't help older kernels that have this issue.
1175 https://bugzilla.linux-nfs.org/show_bug.cgi?id=189
1177 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1178 Signed-off-by: Steve Dickson <steved@redhat.com>
1180 commit 17962b82afb68ca8e6b0d3f432d36c6c7c4980ea
1181 Author: Chuck Lever <chuck.lever@oracle.com>
1182 Date: Wed Oct 13 13:01:51 2010 -0400
1184 mount.nfs: mountproto does not support RDMA
1186 Clean up. Our client does not support the MNT protocol on RDMA.
1188 nfs_mount_protocol() isn't invoked for RDMA mounts (they are shunted
1189 off before nfs_options2pmap() is invoked). But in case it ever is,
1190 it should return the expected response.
1192 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1193 Signed-off-by: Steve Dickson <steved@redhat.com>
1195 commit 73c61fa5cd114fa6eae0e095724ed63aa66a4a6b
1196 Author: NeilBrown <neilb@suse.de>
1197 Date: Wed Oct 13 12:08:41 2010 -0400
1201 client.c: In function 'init_netmask6':
1202 client.c:181:1: warning: no return statement in function returning
1205 and Suse' build system complained
1207 I: Program returns random data in a function
1208 E: nfs-utils no-return-in-nonvoid-function client.c:181
1210 when I built without --enable-ipv6
1212 Reviewed-by: Chuck Lever <chuck.lever@oracle.com>
1213 Signed-off-by: NeilBrown <neilb@suse.de>
1214 Signed-off-by: Steve Dickson <steved@redhat.com>
1216 commit 7e90281b88c05b01c61152b54a0cf2faec45b09c
1217 Author: Chuck Lever <chuck.lever@oracle.com>
1218 Date: Wed Oct 13 12:02:32 2010 -0400
1220 mount.nfs: Eliminate compiler warnings in utils/mount/network.c
1224 network.c: In function get_socket:
1225 network.c:431: warning: dereferencing type-punned pointer might break
1226 strict-aliasing rules
1228 network.c: In function probe_bothports:
1229 network.c:759: warning: dereferencing type-punned pointer might break
1230 strict-aliasing rules
1231 network.c:762: warning: dereferencing type-punned pointer might break
1232 strict-aliasing rules
1234 network.c: In function nfs_probe_statd:
1235 network.c:775: warning: dereferencing type-punned pointer might break
1236 strict-aliasing rules
1238 network.c: In function nfs_call_umount:
1239 network.c:904: warning: dereferencing type-punned pointer might break
1240 strict-aliasing rules
1242 network.c: In function nfs_ca_sockname:
1243 network.c:1106: warning: dereferencing type-punned pointer might break
1244 strict-aliasing rules
1245 network.c:1112: warning: dereferencing type-punned pointer might break
1246 strict-aliasing rules
1248 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1249 Signed-off-by: Steve Dickson <steved@redhat.com>
1251 commit 57385cf87790c0cbdfddfccdde66bd2c8da45923
1252 Author: Chuck Lever <chuck.lever@oracle.com>
1253 Date: Wed Oct 13 11:59:30 2010 -0400
1255 mount.nfs: Eliminate compiler warning in utils/mount/parse_opt.c
1257 parse_opt.c: In function po_rightmost:
1258 parse_opt.c:517: warning: conversion to int from unsigned int may
1259 change the sign of the result
1261 "i" contains the function's result value, so it should be defined as
1262 the same type as the function's return type.
1264 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1265 Signed-off-by: Steve Dickson <steved@redhat.com>
1267 commit 1ee10ef034cbca86da4df271ac4097a948e7ab59
1268 Author: Chuck Lever <chuck.lever@oracle.com>
1269 Date: Wed Oct 13 11:58:27 2010 -0400
1271 mount.nfs: Eliminate compiler warning in utils/mount/nfsumount.c
1275 nfsumount.c:374: warning: ISO C forbids omitting the middle term of
1278 This is also probably harmless, but let's make the code unambiguous.
1280 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1281 Signed-off-by: Steve Dickson <steved@redhat.com>
1283 commit e9c97e4f7075e563d7a442ca8298ac56bafba0d5
1284 Author: Chuck Lever <chuck.lever@oracle.com>
1285 Date: Wed Oct 13 11:56:58 2010 -0400
1287 mount.nfs: Eliminate compiler warning in utils/mount/nfsumount.c
1291 nfsumount.c:265: warning: no previous prototype for nfsumount
1293 It's also a good idea if the compiler can ensure that the prototype
1294 in nfsmount.h matches the actual function defined in nfsumount.c.
1296 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1297 Signed-off-by: Steve Dickson <steved@redhat.com>
1299 commit e2b6d9cbaf20df26dd371a715fce3ae158f37126
1300 Author: Chuck Lever <chuck.lever@oracle.com>
1301 Date: Wed Oct 13 11:54:49 2010 -0400
1303 mount.nfs: Eliminate compiler warnings in utils/mount/mount.c
1307 mount.c: In function parse_opt:
1308 mount.c:354: warning: conversion to size_t from int may change the
1310 mount.c:354: warning: conversion to int from size_t may change the
1312 mount.c:359: warning: conversion to size_t from int may change the
1314 mount.c:359: warning: conversion to int from size_t may change the
1316 mount.c: In function parse_opts:
1317 mount.c:374: warning: conversion to int from size_t may change the
1319 mount.c:377: warning: conversion to size_t from int may change the
1322 Character string lengths are usually size_t anyway. We can easily
1323 avoid the implicit type cast here.
1325 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1326 Signed-off-by: Steve Dickson <steved@redhat.com>
1328 commit 013e8ec9ffb9f28f97e58299719023faf846a029
1329 Author: Chuck Lever <chuck.lever@oracle.com>
1330 Date: Wed Oct 13 11:53:44 2010 -0400
1332 mount.nfs: Eliminate compiler warning in utils/mount/mount.c
1336 mount.c: At top level:
1337 mount.c:324: warning: no previous prototype for ?mount_usage?
1339 mount_usage() has no callers outside of utils/mount/mount.c and no
1340 prototype is provided in a header file. Make it static.
1342 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1343 Signed-off-by: Steve Dickson <steved@redhat.com>
1345 commit 47480475c99335c1203e81662f815b62573c19e8
1346 Author: Chuck Lever <chuck.lever@oracle.com>
1347 Date: Wed Oct 13 11:50:57 2010 -0400
1349 mount.nfs: Eliminate compiler warnings in utils/mount/version.h
1353 In file included from mount.c:50:
1354 version.h: In function linux_version_code:
1355 version.h:48: warning: conversion to unsigned int from int may
1356 change the sign of the result
1357 version.h:48: warning: conversion to unsigned int from int may
1358 change the sign of the result
1359 version.h:48: warning: conversion to unsigned int from int may
1360 change the sign of the result
1362 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1363 Signed-off-by: Steve Dickson <steved@redhat.com>
1365 commit 00885013dccbe00f5cc4e19223cf18e85a8e616a
1366 Author: Chuck Lever <chuck.lever@oracle.com>
1367 Date: Wed Oct 13 11:44:23 2010 -0400
1369 mount.nfs: Eliminate compiler warning in utils/mount/mount.c
1373 In file included from mount.c:41:
1374 mount_config.h:35: warning: no previous prototype for mount_config_opts
1376 Functions defined in include files are usually declared as "static
1377 inline," eliminating the need for a forward declaration.
1379 While I was there, I also fixed the macro that prevents including
1380 mount_config.h multiple times, and fixed some white space damage.
1382 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1383 Signed-off-by: Steve Dickson <steved@redhat.com>
1385 commit ffa577350b03ddd421d455a8cd4cff86e3718264
1386 Author: Chuck Lever <chuck.lever@oracle.com>
1387 Date: Wed Oct 13 11:38:22 2010 -0400
1389 mount.nfs: Eliminate compiler warnings
1393 fstab.c: In function ?lock_mtab?:
1394 fstab.c:385: warning: declaration of ?errsv? shadows a previous local
1395 fstab.c:367: warning: shadowed declaration is here
1396 fstab.c:407: warning: declaration of ?errsv? shadows a previous local
1397 fstab.c:367: warning: shadowed declaration is here
1398 fstab.c:417: warning: declaration of ?tries? shadows a previous local
1399 fstab.c:325: warning: shadowed declaration is here
1400 fstab.c:422: warning: declaration of ?errsv? shadows a previous local
1401 fstab.c:367: warning: shadowed declaration is here
1403 These are probably harmless. Reusing a variable name, however, is a
1404 little confusing to follow when reading the code.
1406 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1407 Signed-off-by: Steve Dickson <steved@redhat.com>
1409 commit 5fe118b838254023d83424c5010ae73a91ec267d
1410 Author: Trond Myklebust <Trond.Myklebust@netapp.com>
1411 Date: Wed Oct 13 11:27:21 2010 -0400
1413 export: Ensure that we free struct exportent->e_uuid
1415 Currently, the exportent->e_uuid is initialised in
1416 support/nfs/exports.c:parseopts(), but it is never freed.
1418 Also ensure that exportent->e_uuid is duplicated correctly in
1421 Adjusted to account for the new export_free() helper.
1423 Also, e_uuid points to memory that is always allocated with strdup(3),
1424 not with xstrdup(). Thus it must be freed via free(3) and not via
1427 Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
1428 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1429 Signed-off-by: Steve Dickson <steved@redhat.com>
1431 commit 656028f9925f5817c5a37565d27159973db84ec3
1432 Author: Chuck Lever <chuck.lever@oracle.com>
1433 Date: Wed Oct 13 11:22:07 2010 -0400
1435 libnfs.a: Allow multiple RPC listeners to share listener port number
1437 Normally, when "-p" is not specified on the mountd command line, the
1438 TI-RPC library chooses random port numbers for each listener. If a
1439 port number _is_ specified on the command line, all the listeners
1440 will get the same port number, so SO_REUSEADDR needs to be set on
1443 Thus we can't let TI-RPC create the listener sockets for us in this
1444 case; we must create them ourselves and then set SO_REUSEADDR (and
1445 other socket options) by hand.
1447 Different versions of the same RPC program have to share the same
1448 listener and SVCXPRT, so we have to cache xprts we create, and re-use
1449 them when additional requests for registration come from the
1452 Though it doesn't look like it, this fix was "copied" from the legacy
1453 rpc_init() function. It's more complicated for TI-RPC, of course,
1454 since a TI-RPC application can set up listeners with a nearly
1455 arbitrary number of address families and socket types, not just the
1456 two listeners that legacy RPC applications can set up (one for AF_INET
1457 UDP and one for AF_INET TCP).
1460 https://bugzilla.linux-nfs.org/show_bug.cgi?id=190
1462 Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
1463 Signed-off-by: Steve Dickson <steved@redhat.com>
1465 commit 1296be71ebae4c0d7da61cc1077d97562d3bc549
1466 Author: Steve Dickson <steved@redhat.com>
1467 Date: Wed Oct 13 10:17:58 2010 -0400
1469 nfs-utils: Fixed typo in NFS man page
1471 Chuck pointed out there was a grammar typo in addition to the spelling
1472 typo. Here is a revised version of the patch.
1474 Signed-off-by: Jim Rees <rees@umich.edu>
1475 Signed-off-by: Steve Dickson <steved@redhat.com>
1477 commit 9afbdbad4436df6f7a5a501c6e4db04c3a5bbb08
1478 Author: Steve Dickson <steved@redhat.com>
1479 Date: Wed Oct 13 10:15:12 2010 -0400
1481 Fix style nits in atomicio.c
1483 Signed-off-by: Jim Rees <rees@umich.edu>
1484 Signed-off-by: Steve Dickson <steved@redhat.com>
1486 commit c117b7a1f29db65d139824ba5bab2a58bf5609e2
1487 Author: Steve Dickson <steved@redhat.com>
1488 Date: Wed Oct 13 10:09:53 2010 -0400
1490 nfs-utils: Move common code into support
1492 There are several source files and headers present in the ./utils/idmapd
1493 directory which are also usable in a doimapd daemon. Because of this we
1494 move that support into the support directory such that it can be shared by
1497 Signed-off-by: Jim Rees <rees@umich.edu>
1498 Signed-off-by: Steve Dickson <steved@redhat.com>
1500 commit 8c217b9623c8304608196aeb2f7360abfdf987c8
1501 Author: Suresh Jayaraman <sjayaraman@suse.de>
1502 Date: Wed Sep 29 07:14:14 2010 -0400
1504 The kernel 2.6.37 has a add new mount option: local_lock.
1505 Document the new option in the nfs(5) man page.
1507 Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
1508 Signed-off-by: Steve Dickson <steved@redhat.com>
1510 commit 554df6bc8456c6971c1c597f70d3b9131e4e5e11
1511 Author: Steve Dickson <steved@redhat.com>
1512 Date: Tue Sep 28 08:24:16 2010 -0400
1514 Revert "nfs-iostat.py: don't wait for an extra interval when given a count"
1516 This reverts commit 837796686ad8f9178c7b6855ada728a53ae511e3.