Last call for nfs-utils patches
Subsystems (2)
Steve Dickson
Reviewed-by Greg BanksI'm planning on cutting a nfs-utils-1.1.2 release since there has been a number commits since the last 1.1.1 release.
The attached list are all of the commits that have been made to the git://linux-nfs.org/nfs-utils git tree.
(Note: to get a working copy of this tree use:
git clone git://linux-nfs.org/nfs-utils nfs-utils)
So if you don't see a commit you have posted to either the nfs or nfsv4 mailing list it probably means I missed it so you might want to bring it to my attention. (Please Note, the patch authors in some of my earlier commit are incorrect. This was simply due to my lack of git knowledge. Once it was pointed it out to me, the problem was corrected.)
Also if you were holding off posting some (hopefully) non-intrusive patches now it the time. There is a very good chance I will not cut the release until this weekend.
tia,
steved.
commit f24884196dac4cffc3ae79cf87cf51d892cda46c
Author: Neil Brown <[email protected]> Date: Tue Mar 4 09:39:31 2008 -0500
Stop the kernel export table from being flushed on MIPS machines.
commit d15fa02fe38f6f1c9183540810ac7bfeac0c7c17
Author: Drew Middlesworth <[email protected]> Date: Tue Mar 4 09:27:57 2008 -0500
Increase the number of concurrent krb5 mounts by increasing the size of the poll array
commit bcd0fcaf0966c546da5043be700587f73174ae25
Author: NeilBrown <[email protected]> Date: Tue Feb 26 13:57:39 2008 -0500
If validateascii is passed a string containing only non-zero 7bit values, then the loop with exit with i == len, and the following test will access beyond the end of the array.
So add an extra test to fix this.
Found by Marcus Meissner <[email protected]>.
commit acf95d32a44fd8357c24e8a04ec53fc6900bfc58
Author: Peng Haitao <[email protected]> Date: Tue Feb 26 13:52:18 2008 -0500
Fixed typo in the nfsstat command line arugments.
commit 46d439b17f22216ce8f9257a982c6ade5d1c5931
Author: Vince Busam <[email protected]> Date: Tue Feb 26 13:04:52 2008 -0500
Stop gssd from ignoring the machine credential cache defined by the -d flag
commit 9dd9b68c4c44f0d9102eb85ee2fa36a8b7f638e3
Author: Harshula Jayasuriya <[email protected]> Date: Tue Feb 12 16:13:25 2008 -0500
In mountd, if get_exportlist() (utils/mountd/mountd.c) returns NULL it should not be considered a failure. It just means that there are no exports on the system.
The practical problem with the current code is that a showmount -e results in a syslog message from mountd that looks like:
rpc.mountd: export request from 10.250.100.2 failed.
commit 7a817c45eaeb6aa93fdb5ca4d81c363b4e4218f0
Author: Frank Filz <[email protected]> Date: Sat Feb 9 07:48:48 2008 -0500
Cleaned up some typos that were found in the various places in the mountd code.
commit b5651e42cedc2b91aba87ea30520884d7f43b36f
Author: Steve Dickson <[email protected]> Date: Thu Jan 24 17:02:08 2008 -0500
Added in relatime mount option so mount.nfs stays compatible with the mount command in util-linux-ng
commit 2d77e3a27b7b211f303fcc3a813a044838dad970
Author: Steve Dickson <[email protected]> Date: Sat Jan 19 07:59:26 2008 -0500
Fix bug when both crossmnt and fsid are set.
When exported a filesystems with option inherited (by the crossmnt option) from a higherlevel filesystem, ignore filesystem specific options like FSID and explicit UUID.
commit 399cc3dad0cdfe8dca4a0ae9035a0554d7223eb2
Author: Steve Dickson <[email protected]> Date: Thu Jan 17 11:25:37 2008 -0500
Stop segfaults on amd64 during warnings messages by creating a second va_list in xlog_backend() and then use that va_list to print messages on stderr.
commit 72ada109b86622db8009e774ebc96a7a92094e8a
Author: Steve Dickson <[email protected]> Date: Wed Jan 16 12:01:13 2008 -0500
The kernel doesn't support the underlying parts needed for changing the MNT program number.
commit 72bc295da78bafff92762b8afbb298ac16d1ff43
Author: Steve Dickson <[email protected]> Date: Wed Jan 16 12:00:32 2008 -0500
The kernel hasn't supported the underlying parts needed for changing the NFS program number for a very long time.
commit abb00ff7cff3c6540978d9b463dee3d4fdda15f4
Author: Steve Dickson <[email protected]> Date: Wed Jan 16 11:53:41 2008 -0500
Address an inconsistency: the mount.nfs command uses the glibc routines to manipulate /etc/mtab (setmntent) but, everything else in nfs-utils uses a local private version (nfs_setmntent). The local version does some extra mangling of the mtab entries.
We should check what util-linux does these days to be sure, but for now, let's make the mount.nfs command use the nfs_ variants of setmntent().
commit fa087f2dece4b1d322d951e9353e8ab93208194d
Author: Steve Dickson <[email protected]> Date: Wed Jan 16 11:50:41 2008 -0500
add_mtab() calls unlock_mtab() twice in one of its error exit paths.
commit 0f3172012919d36e2a7e5bf46c388f8c028fc93e
Author: Steve Dickson <[email protected]> Date: Wed Jan 16 11:48:23 2008 -0500
When mount.nfs reports that statd isn't running, it suggests using the "-o nolocks" option, which doesn't exist. It should say "-o nolock".
commit ec637de16210c1c6fcb3a0df34d7889592f577dc
Author: Steve Dickson <[email protected]> Date: Wed Jan 9 13:13:31 2008 -0500
Automatically set 'nohide' on referral exports.
commit 5c4f17a97b1260e9c564ddd9887afaf50a26de9d
Author: Steve Dickson <[email protected]> Date: Wed Jan 9 11:51:41 2008 -0500
Redress some nits in the description of the timeo option in the nfs(5) man page.
commit 7c6a4ae0aff5c2cd5fde2909e8af886ccf0bd27d
Author: Steve Dickson <[email protected]> Date: Mon Jan 7 10:39:37 2008 -0500
Fixed typo in rpc.mountd's man page
commit 46704243eb10718c722607cc7f66703e3eb3ac9c
Author: Steve Dickson <[email protected]> Date: Fri Jan 4 10:26:21 2008 -0500
Get rid of the "-i" option for mount.nfs[4] and always use the text- based mount(2) system call for kernel version 2.6.23 and later.
commit ca43d9d8c7bdbd067fb1fb4fa9d6e055f4d34ce5
Author: Steve Dickson <[email protected]> Date: Fri Jan 4 09:27:35 2008 -0500
Incorporated Chuck Lever's and Don Domingo's changes to the nfs(5) manual page.
commit bfdc26916d8dabd319240a3ce5198b6189040be7
Author: Steve Dickson <[email protected]> Date: Sat Nov 3 10:40:36 2007 -0400
When mountd gets a request to export a mountpoint which is not explicitly exported, but is below an export point that is flagged as "crossmnt", it passes the wrong path name to the kernel for the
"filehandle -> directory"
mapping.
This can badly confuse the NFS client, and is certainly wrong. So use the correct path names.
commit 812a43c1e4e055ee684ab95b28b41f0c16cc52fd
Author: Steve Dickson <[email protected]> Date: Sat Nov 3 10:36:36 2007 -0400
When following a list of mount versions to probe -
e.g. probe_mnt1_first or probe_mnt3_first - probe_both will first probe the appropriate NFS version and then, if that succeeds, probe the actual mount version. However instead of probing the target mount version, it probes the "most appropriate" mount version for the given NFS version. This results in it probing:
NFSv2, MOUNTv1
twice rather than
NFSv2, MOUNTv1
NFSv2, MOUNTv2
as would be more correct.
This patch removes the "choose most correct" step and just use the current mouint version for the probe_vers array.
commit d4b159620b26dce8f83d41358d618ccc82fa82d4
Author: Steve Dickson <[email protected]> Date: Sat Nov 3 09:35:05 2007 -0400
This means that if mountd is run with "--no-nfs-version 3",
It will first probe for NFS version 3, which will succeed (assuming the kernel supported NFSv3), then it will check the matching mountd version (3) and probe_port on discovering that isn't supported will try other versions, find "1" is supported will succeed.
This leaves up using mount version 1 for an NFSv3 mount, which doesn't work and leads to a SIGSEGV
There is no case where trying other versions is needed the request one is not supported, so simply remove that code.
commit efbda485ff5b3d5916971e1af656d2c21307af30
Author: Steve Dickson <[email protected]> Date: Sat Nov 3 09:14:57 2007 -0400
The wording in the exportfs man page can be a bit confusing, implying that "exportfs -u :/foo" will unexport /foo from all hosts, which it won't.
This patch cleans that wording up and adds a couple examples on how to unexport directories
commit b1e64b99a97b27d850d7a615b291aa9aeb2972ab
Author: Steve Dickson <[email protected]> Date: Fri Oct 26 07:57:10 2007 -0400
Added -S/--since to the nfsstat(1) manpage
Author: David Richter <[email protected]>
_______________________________________________
NFSv4 mailing list
[email protected]
http://linux-nfs.org/cgi-bin/mailman/listinfo/nfsv4
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 3 lines · Steve Dickson
So if you don't see a commit you have posted to either the nfs or nfsv4 mailing list it probably means I missed it so you might want to bring it to my attention. '
I've posted a few patches (some of them as trivial as spelling fixes in man pages) three times with no response, so I gave up.
FWIW, Debian's current patch set (including the aforemented patches) is at
http://people.debian.org/~sesse/nfs-utils-patches/
I believe they should all, except 02-document-debian-init-scripts.patch, be applicable for upstream. (Some of them are already applied.)
/* Steinar */
Signature · 1 line
Homepage: http://www.sesse.net/
Quoted 18 lines · Steve Dickson
I'm planning on cutting a nfs-utils-1.1.2 release since there has been a number commits since the last 1.1.1 release.
The attached list are all of the commits that have been made to the
git://linux-nfs.org/nfs-utilsgit tree.(Note: to get a working copy of this tree use:
git clonegit://linux-nfs.org/nfs-utilsnfs-utils)So if you don't see a commit you have posted to either the nfs or nfsv4 mailing list it probably means I missed it so you might want to bring it to my attention. (Please Note, the patch authors in some of my earlier commit are incorrect. This was simply due to my lack of git knowledge. Once it was pointed it out to me, the problem was corrected.)
Quoted 3 lines
Also if you were holding off posting some (hopefully) non-intrusive patches now it the time. There is a very good chance I will not cut the release until this weekend.
What about the changes which make rpc.gssd spam the syslog less, ie. lowering the priority of two printerr's? It did the trick wonderfully, no more syslog overflows when user creds have expired.
t
A message that is not in the archive (sent off-list or not yet received): alpine.OSF.0.99.0803060931390.100983-r6D5HDKEOC1C46yoE1Qrng@public.gmane.org
Quoted 20 lines · Timo Aaltonen
Quoted 18 lines · Steve Dickson
I'm planning on cutting a nfs-utils-1.1.2 release since there has been a number commits since the last 1.1.1 release.
The attached list are all of the commits that have been made to the
git://linux-nfs.org/nfs-utilsgit tree.(Note: to get a working copy of this tree use:
git clonegit://linux-nfs.org/nfs-utilsnfs-utils)So if you don't see a commit you have posted to either the nfs or nfsv4 mailing list it probably means I missed it so you might want to bring it to my attention. (Please Note, the patch authors in some of my earlier commit are incorrect. This was simply due to my lack of git knowledge. Once it was pointed it out to me, the problem was corrected.)
Quoted 7 lines
Quoted 3 lines
Also if you were holding off posting some (hopefully) non-intrusive patches now it the time. There is a very good chance I will not cut the release until this weekend.
What about the changes which make rpc.gssd spam the syslog less, ie. lowering the priority of two printerr's? It did the trick wonderfully, no more syslog overflows when user creds have expired.
Which patches are you referring to? It appears I missed them...
steved.
Quoted 5 lines · Steve Dickson
Quoted 3 lines · Timo Aaltonen
What about the changes which make rpc.gssd spam the syslog less, ie. lowering the priority of two printerr's? It did the trick wonderfully, no more syslog overflows when user creds have expired.
Which patches are you referring to? It appears I missed them...
I didn't actually post anything yet, it was discussed recently though. But here it goes:
utils/gssd/gss_util.c 2008-02-27 17:23:40.000000000 +0200 renamed from utils/gssd/gss_util.c 2007-10-19 06:07:28.000000000 +0300 +1 -1
@@ -168,7 +168,7 @@ min = min_gss_buf.value; } - printerr(0, "ERROR: GSS-API: error in %s(): %s - %s\n", + printerr(1, "ERROR: GSS-API: error in %s(): %s - %s\n", m, maj, min); if (maj_gss_buf.length != 0)
utils/gssd/gssd_proc.c 2008-02-27 17:23:10.000000000 +0200 renamed from utils/gssd/gssd_proc.c 2007-10-19 06:07:28.000000000 +0300 +1 -1
@@ -464,7 +464,7 @@ if (write(k5_fd, buf, p - buf) < p - buf) goto out_err; return 0; out_err: - printerr(0, "Failed to write error downcall!\n"); + printerr(1, "Failed to write error downcall!\n"); return -1; }
Quoted 18 lines · Steve Dickson
I'm planning on cutting a nfs-utils-1.1.2 release since there has been a number commits since the last 1.1.1 release.
The attached list are all of the commits that have been made to the
git://linux-nfs.org/nfs-utilsgit tree.(Note: to get a working copy of this tree use:
git clonegit://linux-nfs.org/nfs-utilsnfs-utils)So if you don't see a commit you have posted to either the nfs or nfsv4 mailing list it probably means I missed it so you might want to bring it to my attention. (Please Note, the patch authors in some of my earlier commit are incorrect. This was simply due to my lack of git knowledge. Once it was pointed it out to me, the problem was corrected.)
Quoted 3 lines
Also if you were holding off posting some (hopefully) non-intrusive patches now it the time. There is a very good chance I will not cut the release until this weekend.
Here's another patch (by Niall Sheridan) that needs approval:
- - -
rpc.gssd FD_ALLOC_BLOCK too low
The max number of sec=krb5 mounts that rpc.gssd will handle is hardcoded at 32. From utils/gssd/gssd.h:
#define FD_ALLOC_BLOCK 32
This causes problems and errors like:
RPC: AUTH_GSS upcall timed out.
when the number of kerberized nfs mounts approaches the hard limit.
Since linux supports a maximum of 256 nfs mounts, perhaps this number can be upped to 200 or so.
http://launchpadlibrarian.net/11961434/nfs-utils-fd-alloc-block.debdiff
- - -
what say you?
t
Quoted 38 lines · Timo Aaltonen
Quoted 18 lines · Steve Dickson
I'm planning on cutting a nfs-utils-1.1.2 release since there has been a number commits since the last 1.1.1 release.
The attached list are all of the commits that have been made to the
git://linux-nfs.org/nfs-utilsgit tree.(Note: to get a working copy of this tree use:
git clonegit://linux-nfs.org/nfs-utilsnfs-utils)So if you don't see a commit you have posted to either the nfs or nfsv4 mailing list it probably means I missed it so you might want to bring it to my attention. (Please Note, the patch authors in some of my earlier commit are incorrect. This was simply due to my lack of git knowledge. Once it was pointed it out to me, the problem was corrected.)
Quoted 3 lines
Also if you were holding off posting some (hopefully) non-intrusive patches now it the time. There is a very good chance I will not cut the release until this weekend.
Here's another patch (by Niall Sheridan) that needs approval:
- - -
rpc.gssd
FD_ALLOC_BLOCKtoo lowThe max number of sec=krb5 mounts that rpc.gssd will handle is hardcoded at 32. From
utils/gssd/gssd.h:
#defineFD_ALLOC_BLOCK32This causes problems and errors like:
RPC:AUTH_GSSupcall timed out.
when the number of kerberized nfs mounts approaches the hard limit.
Quoted 8 lines
Since linux supports a maximum of 256 nfs mounts, perhaps this number can be upped to 200 or so.
http://launchpadlibrarian.net/11961434/nfs-utils-fd-alloc-block.debdiff
- - -
what say you?
Oops, should've checked the git first.. this is fixed by commit d15fa02fe38f6f1c9183540810ac7bfeac0c7c17
:)
t