Commit Graph

2620 Commits

Author SHA1 Message Date
Rick Macklem
a0c2c3691c Change the new NFS server so that it uses vfs.nfsd naming
for its sysctls instead of vfs.newnfs. This separates the
names from the ones used by the client.
2011-05-08 01:01:27 +00:00
Rick Macklem
1dcad8ec9a Set the initial value of maxfilesize to OFF_MAX in the
new NFS client. It will then be reduced to whatever the
server says it can support. There might be an argument
that this could be one block larger, but since NFS is
a byte granular system, I chose not to do that.

Suggested by:	Matt Dillon
Tested by:	Daniel Braniss (earlier version)
MFC after:	2 weeks
2011-05-06 17:51:00 +00:00
Alexander Motin
08aadbe3b4 Increase NFS_TICKINTVL value from 10 to 500. Now that callout does useful
things only once per second, so other 99 calls per second were useless and
just don't allow idle system to sleep properly.

Reviewed by:	rmacklem
2011-05-06 13:11:50 +00:00
Rick Macklem
78e4b1f838 Change the new NFS server so that it returns 0 when the f_bavail
or f_ffree fields of "struct statfs" are negative, since the
values that go on the wire are unsigned and will appear to be
very large positive values otherwise. This makes the handling
of a negative f_bavail compatible with the old/regular NFS server.

MFC after:	2 weeks
2011-05-06 01:29:14 +00:00
Rick Macklem
f96712c2e6 Fix the new NFS client so that it handles the 64bit fields
that are now in "struct statfs" for NFSv3 and NFSv4. Since
the ffiles value is uint64_t on the wire, I clip the value
to INT64_MAX to avoid setting f_ffree negative.

Tested by:	kib
MFC after:	2 weeks
2011-05-05 00:11:09 +00:00
Rick Macklem
5a816b92a3 Add a comment noting that the NFS code assumes that the
values of error numbers in sys/errno.h will be the same
as the ones specified by the NFS RFCs and that the code
needs to be fixed if error numbers are changed in sys/errno.h.

Suggested by:	Peter Jeremy
MFC after:	2 weeks
2011-05-04 22:02:33 +00:00
Rick Macklem
2e3b981a4d Add kernel support for NFSSVC_ZEROCLTSTATS and NFSSVC_ZEROSRVSTATS
so that they can be used by nfsstat(1) to implement the "-z" option
for the new NFS subsystem.

MFC after:	2 weeks
2011-05-04 13:36:18 +00:00
Rick Macklem
2b08b570cb Revert r221306, since NFSSVC_ZEROSTATS zero'd both client and
server stats, when separate modifiers for NFSSVC_GETSTATS for
each of client and server stats is what it required by nfsstat(1).
2011-05-04 13:30:38 +00:00
Ruslan Ermilov
e2f2b37089 Implemented a mount option "nocto" that disables cache coherency
checking at open time.  It may improve performance for read-only
NFS mounts.  Use deliberately.

MFC after:	1 week
Reviewed by:	rmacklem, jhb (earlier version)
2011-05-04 13:27:45 +00:00
Ruslan Ermilov
55cde634cf In ncl_printf(), call vprintf() instead of printf().
MFC after:	3 days
2011-05-04 11:22:52 +00:00
Rick Macklem
b2946fadcd Add the kernel support needed to zero out the nfsstats
structure for the new NFS subsystem. This will be used
by nfsstats.c to implement the "-z" option.

MFC after:	2 weeks
2011-05-01 22:19:52 +00:00
Konstantin Belousov
4417ac326a Clarify the comment.
MFC after:	1 week
2011-04-30 13:49:03 +00:00
Rick Macklem
8b713a2f8a The build was broken by r221190 for 64bit arches like amd64.
This patch fixes it.

MFC after:	2 weeks
2011-04-29 12:30:15 +00:00
Rick Macklem
61c827204b Fix the new NFS client so that it handles the "nfs_args" value
in mnt_optnew. This is needed so that the old mount(2) syscall
works and that is needed so that amd(8) works. The code was
basically just cribbed from sys/nfsclient/nfs_vfsops.c with minor
changes. This patch is mainly to fix the new NFS client so that
amd(8) works with it. Thanks go to Craig Rodrigues for helping with
this.

Tested by:	Craig Rodrigues (for amd)
MFC after:	2 weeks
2011-04-28 23:21:50 +00:00
John Baldwin
7d74606889 Update a comment since ext2fs does not use SU.
Reviewed by:	kib
2011-04-28 20:25:15 +00:00
John Baldwin
466a71d75e The b_dep field of buffers is always empty for ext2fs, it is only used
for SU in FFS.

Reported by:	kib
2011-04-28 17:36:26 +00:00
John Baldwin
9e880b876d Sync with several changes in UFS/FFS:
- 77115: Implement support for O_DIRECT.
- 98425: Fix a performance issue introduced in 70131 that was causing
  reads before writes even when writing full blocks.
- 98658: Rename the BALLOC flags from B_* to BA_* to avoid confusion with
  the struct buf B_ flags.
- 100344: Merge the BA_ and IO_ flags so so that they may both be used in
  the same flags word. This merger is possible by assigning the IO_ flags
  to the low sixteen bits and the BA_ flags the high sixteen bits.
- 105422: Fix a file-rewrite performance case.
- 129545: Implement IO_INVAL in VOP_WRITE() by marking the buffer as
  "no cache".
- Readd the DOINGASYNC() macro and use it to control asynchronous writes.
  Change i-node updates to honor DOINGASYNC() instead of always being
  synchronous.
- Use a PRIV_VFS_RETAINSUGID check instead of checking cr_uid against 0
  directly when deciding whether or not to clear suid and sgid bits.

Submitted by:	Pedro F. Giffuni  giffunip at yahoo
2011-04-28 14:27:17 +00:00
Rick Macklem
afea74655f Fix module names and dependencies so the NFS clients will
load correctly as modules after r221124.
2011-04-27 20:42:30 +00:00
John Baldwin
bbfe24fbf2 Use a private EXT2_ROOTINO constant instead of redefining ROOTINO.
Submitted by:	Pedro F. Giffuni  giffunip at yahoo
2011-04-27 18:25:35 +00:00
John Baldwin
4d2ede6798 Various style fixes including using uint*_t instead of u_int*_t.
Submitted by:	Pedro F. Giffuni  giffunip at yahoo
2011-04-27 18:15:34 +00:00
Rick Macklem
4309e17add This patch changes head so that the default NFS client is now the new
NFS client (which I guess is no longer experimental). The fstype "newnfs"
is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
Although mounts via fstype "nfs" will usually work without userland
changes, an updated mount_nfs(8) binary is needed for kernels built with
"options NFSCL" but not "options NFSCLIENT". Updated mount_nfs(8) and
mount(8) binaries are needed to do mounts for fstype "oldnfs".
The GENERIC kernel configs have been changed to use options
NFSCL and NFSD (the new client and server) instead of NFSCLIENT and NFSSERVER.
For kernels being used on diskless NFS root systems, "options NFSCL"
must be in the kernel config.
Discussed on freebsd-fs@.
2011-04-27 17:51:51 +00:00
Rick Macklem
541cb7a358 Fix a kernel linking problem introduced by r221032, r221040
when building kernels that don't have "options NFS_ROOT"
specified. I plan on moving the functions that use these
data structures into the shared code in sys/nfs/nfs_diskless.c
in a future commit. At that time, these definitions will no
longer be needed in nfs_vfsops.c and nfs_clvfsops.c.

MFC after:	2 weeks
2011-04-26 13:50:11 +00:00
Rick Macklem
8954032f0d Modify the experimental (newnfs) NFS client so that it uses the
same diskless NFS root code as the regular client, which
was moved to sys/nfs by r221032. This fixes the newnfs
client so that it can do an NFSv3 diskless root file system.

MFC after:	2 weeks
2011-04-25 23:12:18 +00:00
Rick Macklem
151c163e4d Fix the experimental NFS client so that it does not bogusly
set the f_flags field of "struct statfs". This had the interesting
effect of making the NFSv4 mounts "disappear" after r221014,
since NFSMNT_NFSV4 and MNT_IGNORE became the same bit.

MFC after:	2 weeks
2011-04-25 14:51:08 +00:00
Rick Macklem
385edc8e71 Modify the experimental NFS client so that it uses the same
"struct nfs_args" as the regular NFS client. This is needed
so that the old mount(2) syscall will work and it makes
sharing of the diskless NFS root code easier. Eary in the
porting exercise I introduced a new revision of nfs_args, but
didn't actually need it, thanks to nmount(2). I re-introduced the
NFSMNT_KERB flag, since it does essentially the same thing and
the old one would not have been used because it never worked.
I also added a few new NFSMNT_xxx flags to sys/nfsclient/nfs_args.h
that are used by the experimental NFS client.

MFC after:	2 weeks
2011-04-25 13:09:32 +00:00
Rick Macklem
24e2bcc006 Remove the nm_mtx mutex locking from the test for
nm_maxfilesize. This value rarely, if ever, changes
and the nm_mtx mutex is locked/unlocked earlier in
the function, which should be sufficient to avoid
getting a stale cached value for it. There is a
discussion w.r.t. what these tests should be, but
I've left them basically the same as the regular
NFS client for now.

Suggested by:	pjd
MFC after:	2 weeks
2011-04-21 19:56:06 +00:00
Rick Macklem
920ae5d96a Revert r220906, since the vp isn't always locked when
nfscl_request() is called. It will need a more involved
patch.
2011-04-21 12:38:12 +00:00
Rick Macklem
69bcf84509 Add a check for VI_DOOMED at the beginning of nfscl_request()
so that it won't try and use vp->v_mount to do an RPC during
a forced dismount. There needs to be at least one more kernel
commit, plus a change to the umount(8) command before forced
dismounts will work for the experimental NFS client.

MFC after:	2 weeks
2011-04-20 23:25:18 +00:00
Rick Macklem
b29b9bcbfb Modify the offset + size checks for read and write in the
experimental NFS client to take care of overflows for the calls
above the buffer cache layer in a manner similar to r220876.
Thanks go to dillon at apollo.backplane.com for providing the
snippet of code that does this.

MFC after:	2 weeks
2011-04-20 01:15:22 +00:00
Rick Macklem
b1297f142f Modify the offset + size checks for read and write in the
experimental NFS client to take care of overflows. Thanks
go to dillon at apollo.backplane.com for providing the
snippet of code that does this.

MFC after:	2 weeks
2011-04-20 00:21:51 +00:00
Rick Macklem
58c969c8de Fix up handling of the nfsmount structure in read and write
within the experimental NFS client. Mostly add mutex locking
and use the same rsize, wsize during the operation by keeping
a local copy of it. This is another change that brings it
closer to the regular NFS client.

MFC after:	2 weeks
2011-04-19 01:09:51 +00:00
Rick Macklem
a8bafa5d3b Revert r220761 since, as kib@ pointed out, the case of
adding the check to nfsrpc_close() isn't useful. Also,
the check in nfscl_getcl() must be more involved, since
it needs to check before and after the acquisition of
the refcnt on nfsc_lock, while the mutex that protects
the client state data is held.
2011-04-18 23:35:16 +00:00
Rick Macklem
bc62b5cf6a Add a vput() to nfs_lookitup() in the experimental NFS client
for a case that will probably never happen. It can only
happen if a server were to successfully lookup a file, but not
return attributes for that file. Although technically allowed
by the NFSv3 RFC, I doubt any server would ever do this.
However, if it did, the client would have not vput()'d the
new vnode when it needed to do so.

MFC after:	2 weeks
2011-04-18 01:02:43 +00:00
Rick Macklem
ab42af2708 Add vput() calls in two places in the experimental NFS client
that would be needed if, in the future, nfscl_loadattrcache()
were to return an error. Currently nfscl_loadattrcache()
never returns an error, so these cases never currently happen.

MFC after:	2 weeks
2011-04-18 00:41:23 +00:00
Rick Macklem
78d8a60009 Change the mutex locking for several locations in the
experimental NFS client's vnode op functions to make
them compatible with the regular NFS client. I'll admit
I'm not sure that the mutex locks around the assignments
are needed, but the regular client has them, so I added them.
Also, add handling of the case of partial attributes in
setattr to be compatible with the regular client.

MFC after:	2 weeks
2011-04-17 23:56:57 +00:00
Rick Macklem
be8b35eda7 Add checks for MNTK_UNMOUNTF at the beginning of three
functions, so that threads don't get stuck in them during
a forced dismount. nfs_sync/VFS_SYNC() needs this, since it is
called by dounmount() before VFS_UNMOUNT(). The nfscl_nget()
case makes sure that a thread doing an VOP_OPEN() or
VOP_ADVLOCK() call doesn't get blocked before attempting
the RPC. Attempting RPCs don't block, since they all
fail once a forced dismount is in progress.
The third one at the beginning of nfsrpc_close()
is done so threads don't get blocked while doing VOP_INACTIVE()
as the vnodes are cleared out.
With these three changes plus a change to the umount(1)
command so that it doesn't do "sync()" for the forced case
seem to make forced dismounts work for the experimental NFS
client.

MFC after:	2 weeks
2011-04-17 23:04:03 +00:00
Rick Macklem
ebd9ef339f Get rid of the "nfscl: consider increasing kern.ipc.maxsockbuf"
message that was generated when doing experimental NFS client
mounts. I put that message in because the krpc would hang with
the default size for mounts that used large rsize/wsize values.
Since the bug that caused these hangs was fixed by r213756,
I think the message is no longer needed.

MFC after:	2 weeks
2011-04-17 20:01:32 +00:00
Rick Macklem
0a9f005dff Fix up some of the sysctls for the experimental NFS client so
that they use the same names as the regular client. Also add
string descriptions for them.

MFC after:	2 weeks
2011-04-17 18:56:17 +00:00
Rick Macklem
8e82d541da Change some defaults in the experimental NFS client to be the
same as the regular NFS client for NFSv3. The main one is making
use of a reserved port# the default. Also, set the retry limit
for TCP the same and fix the code so that it doesn't disable
readdirplus for NFSv4.

MFC after:	2 weeks
2011-04-17 14:10:12 +00:00
Rick Macklem
f5613c1d97 Fix readdirplus in the experimental NFS client so that it
skips over ".." to avoid a LOR race with nfs_lookup(). This
fix is analagous to r138256 in the regular NFS client.

MFC after:	2 weeks
2011-04-17 02:44:51 +00:00
Rick Macklem
4b3a38ecdf Add a lktype flags argument to nfscl_nget() and ncl_nget() in the
experimental NFS client so that its nfs_lookup() function can use
cn_lkflags in a manner analagous to the regular NFS client.

MFC after:	2 weeks
2011-04-16 23:20:21 +00:00
Rick Macklem
f8a2f6b03a Add mutex locking on the nfs node in ncl_inactive() for the
experimental NFS client.

MFC after:	2 weeks
2011-04-16 22:15:59 +00:00
Rick Macklem
7b8c319be4 Change the experimental NFS client so that it creates nfsiod
threads in the same manner as the regular NFS client after
r214026 was committed. This resolves the lors fixed by r214026
and its predecessors for the regular client.

Reviewed by:	jhb
MFC after:	2 weeks
2011-04-15 23:07:48 +00:00
Rick Macklem
a09001a82b Fix the experimental NFSv4 server so that it uses VOP_PATHCONF()
to determine if a file system supports NFSv4 ACLs. Since
VOP_PATHCONF() must be called with a locked vnode, the function
is called before nfsvno_fillattr() and the result is passed in
as an extra argument.

MFC after:	2 weeks
2011-04-14 23:46:15 +00:00
Rick Macklem
07c0c166e4 Modify the experimental NFSv4 server so that it handles
crossing of server mount points properly. The functions
nfsvno_fillattr() and nfsv4_fillattr() were modified to
take the extra arguments that are the mount point, a flag
to indicate that it is a file system root and the mounted
on fileno. The mount point argument needs to be busy when
nfsvno_fillattr() is called, since the vp argument is not
locked.

Reviewed by:	kib
MFC after:	2 weeks
2011-04-14 21:49:52 +00:00
Rick Macklem
149ce1025c Add VOP_PATHCONF() support to the experimental NFS client
so that it can, along with other things, report whether or
not NFS4 ACLs are supported.

MFC after:	2 weeks
2011-04-13 22:37:28 +00:00
Rick Macklem
3707cf8962 Fix the experimental NFSv4 client so that it recognizes server
mount point crossings correctly. It was testing the wrong flag.
Also, try harder to make sure that the fsid is different than
the one assigned to the client mount point, by hashing the
server's fsid (just to create a different value deterministically)
when it is the same.

MFC after:	2 weeks
2011-04-13 22:16:52 +00:00
Rick Macklem
f659876f01 Vrele ni_startdir in the experimental NFS server for the case
of NFSv2 getting an error return from VOP_MKNOD(). Without this
patch, the server file system remains busy after an NFSv2
VOP_MKNOD() fails.

MFC after:	2 weeks
2011-04-11 20:54:30 +00:00
Rick Macklem
806e2e4bb6 Add some cleanup code to the module unload operation for
the experimental NFS server, so that it doesn't leak memory
when unloaded. However, unloading the NFSv4 server is not
recommended, since all NFSv4 state will be lost by the unload
and clients will have to recover the state after a server
reload/restart as if the server crashed/rebooted.

MFC after:	2 weeks
2011-04-10 20:43:07 +00:00
Rick Macklem
8d2f180ea4 Add a VOP_UNLOCK() for the directory, when that is not what
VOP_LOOKUP() returned. This fixes a bug in the experimental
NFS server for the case where VFS_VGET() fails returning EOPNOTSUPP
in the ReaddirPlus RPC, forcing the use of VOP_LOOKUP() instead.

MFC after:	2 weeks
2011-04-09 23:55:27 +00:00