summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Remove tools/getiversionNeil Brown2007-03-164-63/+2
| | | | | We don't need it (any more). Also add AM_PROG_CC_C_O as we seem to need it..
* Use AC_SYS_LARGEFILE to enable largefile supportNeil Brown2007-03-161-1/+4
| | | | | | rather than explicit -D_FILE_OFFSET_BITS=64 Taken from Redhat -5 sources.
* Look in lib64 for gss libs aswell.Neil Brown2007-03-161-0/+2
| | | | | | On distros that support 32 and 64bit userspace, we need to look a bit more broadly for libraries (taken from Redhat sources for RHEL5 - thanks).
* Allow 'make install' to be run as non-root.Neil Brown2007-03-161-1/+1
| | | | | Providing write access to the DESTDIR is ok, we shouldn't fail the install just because 'chown' fails.
* Use lstat rather than stat in gssd_find_existing_krb5_ccacheNeil Brown2007-03-161-1/+1
| | | | | As this is a file in /tmp, a symlink could take us anywhere... If it was a NFS filesystem with a dead server, we could block for a long time..
* Further tidyup of nfs_umount.Neil Brown2007-03-161-65/+16
| | | | | | | | | | | | | | - remove non-used arguments from del_mtab. - Don't try to pass a "host:/path" string to umount. It used to be possible to umount("/dev/whatever"). It has never been possible to umount("host:/path"). - Don't try to read /proc/mounts first. Some mount options (mount_vers) are only stored in /etc/mtab, not in /proc/mounts. So we have to prefer /etc/mtab as getmntXbackwards do. - Only every call one of getmnt{dir,dev}backwards, depending on whether 'spec' looks like a path name or a host:/path. - Don't call _nfsumount unless we have a host:/path, or del_mtab unless we have a path name.
* Make mtab_head staticNeil Brown2007-03-162-2/+1
| | | | It is only used in one place.
* Rename add_mtab2() to del_mtab()Wei Yongjun2007-03-161-3/+3
| | | | | | | add_mtab2 is used to remove an entry from mtab, rename it to del_mtab. Signed-off-by: Wei Yongjun <yjwei@nanjing-fnst.com> Signed-off-by: Neil Brown <neilb@suse.de>
* Use correct UMNT protocol to do umountWei Yongjun2007-03-161-25/+29
| | | | | | | | | | | | | This patch do following things: 1. When unmounting, we should try /proc/mounts first. Only then in /etc/mtab if it wasn't found. 2. Reject name passed to us which is not starts with a '/' or not contains a ':' 3. Fix the method to get UMNT protocol Signed-off-by: Wei Yongjun <yjwei@nanjing-fnst.com> Signed-off-by: Neil Brown <neilb@suse.de>
* Fix nfs4 umount to not used umount procedureWei Yongjun2007-03-161-27/+13
| | | | | | | | | | | This fix umount.nfs4's BUG, patch cleanup some useless code. 1. Combine nfs3_umount() and nfs2_umount() to nfs_umount() 2. If it is a nfs4 remote file system or unkown mount version, do not probe remote mount port and not send UMNT request wich is not necessary Signed-off-by: Wei Yongjun <yjwei@nanjing-fnst.com> Signed-off-by: Neil Brown <neilb@suse.de>
* Use correct UMNT version to do umountWei Yongjun2007-03-161-3/+10
| | | | | | | | | This patch lets umount to use correct UMNT version to do umount. In latest kernel, the version format is "vers=3" etc., and in old kernel the version format is "v3","v4" etc. Signed-off-by: Wei Yongjun <yjwei@nanjing-fnst.com> Signed-off-by: Neil Brown <neilb@suse.de>
* Don't complain if two exports of different types match the same IP address.Neil Brown2007-03-161-2/+4
| | | | | | | As nfs-utils as an ordering of client types (hostname, netmask, wildcard, netgroup), it is expected that sometimes the one IP will match two or more of these and the first will over-ride. So don't both complaining when that happens.
* Install mount.nfs* in /sbin instead of /usr/sbinNeil Brown2007-03-161-3/+7
| | | | | | This patch is now hard-coded in the Makefile.am and cannot be changed by configure. But as it needs to match what util-linux does, that is probably a good thing.
* Fix memory leak in mountd.Steinar H. Gunderson2007-03-161-0/+1
| | | | Signed-off-by: Neil Brown <neilb@suse.de>
* Automatically start statd when mounting an nfs filesystem.Neil Brown2007-03-164-7/+40
| | | | | | | | | | | If statd is not running (/var/run/rpc.statd.pid) when an nfs filesystem is mounted (v2 or v3, with remote locking enabled), and if /usr/sbin/start-statd (or other program specified at config time) is present, then run that program to start statd. This means that statd does not need to be running "just in case". It only needs to be started at boot time if the nfs server is started.
* Make --enable-secure-statd the default.Neil Brown2007-03-153-7/+39
| | | | | | | | | | | | | | i.e. you now need --disable-secure-statd if you want any client other than lockd to talk to statd. Also relax the RESTRICTED_STATD checks so that a recent kernel with /proc/sys/fs/nfs/nsm_use_hostnames set can still talk to statd. Finally, restrict access to simulate_crash so that only privileged processes on localhost can call it. Having it accessible by the whole world is probably not much more than a minor inconvenience, but it really should be kept closed.
* Remove uid-mapping text from exports.manNeil Brown2007-03-151-99/+0
| | | | | | It was already commented out, and it will never be wanted. Signed-off-by: Neil Brown <neilb@suse.de>
* Remove "maptype" supportJ. Bruce Fields2007-03-155-60/+0
| | | | | | | | It appears that this is used only by unfsd, and is obscure enough that we should be able to just rip it out with no special precautions. Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu> Signed-off-by: Neil Brown <neilb@suse.de>
* If -o sec= is not specified for mount, allow either AUTH_SYS or AUTH_NONE.Neil Brown2007-03-151-2/+10
| | | | | | | | | Solaris servers, when asked to share a filesystem with an anon-uid, will report the only available authentication style as AUTH_NONE in the reply from mountd (even though they actually accept AUTH_SYS and simply ignore the credentiuals). So if no sec= is specified we should really accept anything that can easily be handled. ie. AUTH_SYS or AUTH_NONE.
* Use very long timeout for information in the filehandle->path cache.Simon Peter2007-03-131-1/+9
| | | | | | | | | | | | | /* The fsid -> path lookup can be quite expensive as it * potentially stats and reads lots of devices, and some of those * might have spun-down. The Answer is not likely to * change underneath us, and an 'exportfs -f' can always * remove this from the kernel, so use a really log * timeout. Maybe this should be configurable on the command * line. */ Signed-off-by: Neil Brown <neilb@suse.de>
* nfs-utils patch for mount-reserved-portTalpey, Thomas2007-03-131-18/+18
| | | | | | | | Only create a mount-time reserved port socket for kernels which require it (pre-2.1.32/nfs_mount_version 1). Signed-off-by: Tom Talpey <tmt@netapp.com> Signed-off-by: Neil Brown <neilb@suse.de>
* Restore use of un-connected socket for UDP mount requests.Neil Brown2007-03-133-29/+37
| | | | | | | | | | | | | | | | | | | | | | When connecting to an NFSv4 server we need to find out IP address as it would be seen by the server, to register an address for callbacks. This is most easily done by connecting the socket to the servers address and then getting the address of our endpoint. However with a connected UDP socket, replies that come from a different IP address - as can happen with non-Linux multi-homed servers - will be rejected. So if we connected our UDP socket, we need to be sure to disconnect it before using it. This patch adds an option to get_socket to say if we want it connected or not and, in the case where we do, we disconnect a UDP socket after the connection information has been used. Also clean up the error handling in clnt_ping which was getting clumsy.
* Add support for "mount -o sec=none"Neil Brown2007-03-131-1/+3
| | | | For completeness... and who knows, someone might want it.
* Remove unused variable from add_mtabNeil Brown2007-03-131-1/+0
|
* Merge branch 'branch-1-0'Neil Brown2007-02-278-3/+263
|\
| * Fix silly bug with tcp-wrappers checkNeil Brown2007-02-271-1/+1
| | | | | | | | | | | | A stray '$'. Signed-off-by: Neil Brown <neilb@suse.de>
| * Fix silly bug with gid lookupNeil Brown2007-02-271-1/+1
| | | | | | | | | | | | | | Wrong pointer test meant mountd would alway do gid lookups instead of only if asked to with '-g'. Signed-off-by: Neil Brown <neilb@suse.de>
| * Extend the exportfs/mountd interface to pass fslocations info into the kernelFred Isaman2007-02-275-8/+34
| | | | | | | | | | | | | | | | Fix up a few issues with the fsloc code. Signed-off-by: Fred Isaman <iisaman@citi.umich.edu> Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> Signed-off-by: Neil Brown <neilb@suse.de>
| * Fix reference error when writing fsloc data to cacheKevin Coffman2007-02-271-1/+1
| | | | | | | | | | | | | | | | Use the correct pointer when writing fslocations data to the cache. Also write the fsloc stuff before the uuid stuff so userland code will work with or without the uuid kernel patches. Signed-off-by: Neil Brown <neilb@suse.de>
| * Add missing fsloc.[ch] files - oops.Neil Brown2007-02-262-0/+234
| |
* | Remove rpc.lockdNeil Brown2007-02-229-165/+1
| | | | | | | | | | | | This was only needed for kernels 2.2.14 through 2.2.17. These have long since been superceded, so remove some dead weight.
* | Change default from subtree_check to no_subtree_checkNeil Brown2007-02-221-3/+3
|/ | | | | | | subtree_check causes more problems than it is worth, and it isn't worth much in the first place.. Signed-off-by: Neil Brown <neilb@suse.de>
* Update version to 1.0.11Neil Brown2007-02-221-1/+1
|
* Extend the exportfs interface to pass fslocations info into the kernel.Fred Isaman2007-02-227-3/+98
| | | | | | | | | | Extend exportfs interface to pass fslocations info into the kernel, using syntax modelled after AIX. Adds "refer=" and "replicas=" options to /etc/exports to enable use of the kernel fslocation code. Signed-off-by: Fred Isaman <iisaman@citi.umich.edu> Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> Signed-off-by: Neil Brown <neilb@suse.de>
* Fix showmount bugs.Neil Brown2007-02-221-3/+6
| | | | | | | When doing a nonblocked connect, we need to select for 'write', not 'read'. Also, when a tcp socket has been connected, we should use clnttcp_create to make a tcp client, not clntudp_bufcreate !!
* Support group-id looks for kernels that ask for them.Neil Brown2007-02-124-4/+78
| | | | | | | With "-g" mountd will listen for uid -> gidlist requests from the kernel and provide the required mapping. This is specific to AUTH_USER (aka AUTH_SYS) and is designed to overcome the 16-gid limit in the AUTH_UNIX protocol.
* Use UUIDs to identify filesystems if kernel supports it.Neil Brown2007-02-128-35/+216
| | | | | | | | | | | This introduces a new dependancy on libblkid. If a filesystem being exported has a UUID that libblkid can extract, then that is passed to the kernel for use in identifying the filesystem in filehandles. This means that 'fsid=' is no longer needed to work around the problem of device numbers changing. fsid= is still needed for fielsystems that have no device, and can now be given 16byute uuid instead of just a 32bit one.
* Correct spelling errorNeil Brown2007-02-121-2/+2
| | | | writting -> writing
* Add option to svcgssd to enable libnfsidmap debugging.Kevin Coffman2007-02-092-4/+20
| | | | | | | | | | Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> As suggested by Glenn Machin <GMachin@sandia.gov>. Allow svcgssd to turn on libnfsidmap debugging. This uses a new command-line parameter so that it can be enabled independently from other debugging. Signed-off-by: Neil Brown <neilb@suse.de>
* Remove duplicated code.Kevin Coffman2007-02-091-17/+20
| | | | | | | | Signed-off-by: Fred Isaman <iisaman@citi.umich.edu> Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> Remove duplicated code. Signed-off-by: Neil Brown <neilb@suse.de>
* Share handling of lucid_sec_context for Heimdal and MITKevin Coffman2007-02-095-170/+234
| | | | | | | | | | | | Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> The 0.8 release of Heimdal has (will have) support for the lucid context. The handling of lucid_sec_context can be shared between builds with MIT or Heimdal Kerberos. Split out the lucid_sec_context code from context_mit.c and make a new common file, context_lucid.c. Signed-off-by: Neil Brown <neilb@suse.de>
* Use owner rather than filename format in choosing cred cache filesKevin Coffman2007-02-091-56/+50
| | | | | | | | | | | Signed-off-by: Glenn Machin <gmachin@sandia.gov> Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> Some installations use different name formats for their credentials caches. Instead of checking that the uid is part of the name, just make sure that uid is the owner of the file. This is a modification of the original patch from Glenn. Signed-off-by: Neil Brown <neilb@suse.de>
* Various minor manpage fixes.Kevin Coffman2007-02-091-3/+3
| | | | | | | | | | | | | Signed-off-by: Steinar H. Gunderson <sgunderson@bigfoot.com> Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> This mostly takes care of the difference between - and \-; in man pages, the former is hyphen (which indicates, among others, that a line might be split at that point), while the latter is a dash. For options, the latter is correct. There's also one minor grammatical fix. Signed-off-by: Neil Brown <neilb@suse.de>
* Use the gssglue version of gssapi.h for svcgssd_mech2file.cKevin Coffman2007-02-091-5/+1
| | | | | | | Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> Remove Kerberos implementation dependency from svcgssd_mech2file.c Signed-off-by: Neil Brown <neilb@suse.de>
* Stop using storage after freeKevin Coffman2007-02-091-2/+5
| | | | | | | | | | | | | | | | | | | | | | | | | Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> The previous patch seems to expose a use after free bug in dirscancb. At least, I could reliably reproduce a segfault by doing a bunch of mounts and then unmounting them all. The code uses the following list macro: TAILQ_FOREACH(ic, icq, ic_next) { ...to iterate over all of the ic entries and clean up any that no longer have a corresponding directory in rpc_pipefs. This macro unrolls into: for(ic=icq->tqh_first; ic != NULL; ic=ic->ic_next.tqe_next) { ...but within this loop, we can free ic, and then the for loop can trip over that when it tries to do the iteration. The attached patch works around this by not using the TAILQ_FOREACH macro and saving off the tqe_next pointer prior to the free. Again, this was tested on a patched 1.0.6, but the 1.0.10 code is very similar, and I think the problem exists there as well. Signed-off-by: Neil Brown <neilb@suse.de>
* Fix memory leak in idmapd.Kevin Coffman2007-02-091-3/+8
| | | | | | | | | | | | | | | | Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> There is a pretty nasty memory leak in idmapd in dirscancb(). Some of our customers have reported that idmapd can eat gigabytes of memory on machines with a large number of mounts and unmounts and a long uptime. That function uses scandir(), which malloc's an array of strings, but dirscancb() never frees the strings or the array. The following patch should correct this, but I've not yet tested it on 1.0.10 (only on the RHEL4 1.0.6 version). Still, the code is very similar and I'm fairly certain the problem exists in both versions. Signed-off-by: Neil Brown <neilb@suse.de>
* Treat GSSAPI error codes as unsigned.Kevin Coffman2007-02-093-2/+14
| | | | | | | | | | Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> GSSAPI error codes (major and minor) are defined as unsigned values. However, we treat them as signed while passing them down to the kernel where conversion fails if they include the minus sign. Convert them as unsigned. Signed-off-by: Neil Brown <neilb@suse.de>
* Add AM_MAINTAINER_MODE to configure.inKevin Coffman2007-02-091-0/+1
| | | | | | | | | | | | | | | | | | | Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> As requested by "Steinar H. Gunderson" <sgunderson@bigfoot.com>, add AM_MAINTAINER_MODE to configure.in. See the description of this macro below: `AM_MAINTAINER_MODE' disables the so called "rebuild rules" bys default. If you have `AM_MAINTAINER_MODE' in `configure.ac', and run `./configure && make', then `make' will *never* attempt to rebuild `configure', `Makefile.in's, Lex or Yacc outputs, etc. I.e., this disables build rules for files which are usually distributed and that users should normally not have to update. If you run `./configure --enable-maintainer-mode', then these rebuild rules will be active. Signed-off-by: Neil Brown <neilb@suse.de>
* Touch up some of the autotools filesKevin Coffman2007-02-093-2/+12
| | | | | | | | | | | | | | | | Signed-off-by: Mike Frysinger <vapier@gentoo.org> Signed-off-by: Kevin Coffman <kwc@citi.umich.edu> this patch touches up the autotool code a bit in nfs-utils: - run autogen.sh with -e so if something fails, it'll abort properly - set ACLOCAL_AMFLAGS so that when running autoreconf or when autotools re-runs itself, the m4 files are found properly - make sure we include bsdsignals.m4 in the final tarball - add some cross-compiling fallback logic to bsdsignals.m4 so that when cross-compiling nfs-utils, the configure is a bit more nice than simply: checking for BSD signal semantics... configure: error: cannot run test program while cross compiling Signed-off-by: Neil Brown <neilb@suse.de>
* Make UDP sockets not blockingNeil Brown2007-02-051-4/+15
| | | | | | | | | This is needs if mountd is running multithreaded else multiple threads will be blocked on a UDP port with nothing to read and so won't be able to serve up-calls from the kernel. Thanks to "Murali Krishna V" <vm.krishna@gmail.com> for highlighting the problem.
generated by cgit (git 2.34.1) at 2025-09-25 10:33:37 +0000