a vnode. Not bumped into with asserts in the main tree because we
run the NFS server with Giant by default. Discovered by inspection.
Complete annotations of Giant acquisition/release to note that it's
only because of VFS that we acquire Giant in most places in the NFS
server.
subsystem lock to avoid tripping over an assertion regarding whether
the lock is held or not. This is likely to be the cause of a panic
tripped over by Andrea Campi.
mechanism so that early processing on mbufs can be performed before
a context switch to the NFS server threads. Because of this, if
the socket code is running without Giant, the NFS server also needs
to be able to run the upcall code without relying on the presence on
Giant. This change modifies the NFS server to run using a "giant
code lock" covering operation of the whole subsystem. Work is in
progress to move to data-based locking as part of the NFSv4 server
changes.
Introduce an NFS server subsystem lock, 'nfsd_mtx', and a set of
macros to operate on the lock:
NFSD_LOCK_ASSERT() Assert nfsd_mtx owned by current thread
NFSD_UNLOCK_ASSERT() Assert nfsd_mtx not owned by current thread
NFSD_LOCK_DONTCARE() Advisory: this function doesn't care
NFSD_LOCK() Lock nfsd_mtx
NFSD_UNLOCK() Unlock nfsd_mtx
Constify a number of global variables/structures in the NFS server
code, as they are not modified and contain constants only:
nfsrvv2_procid nfsrv_nfsv3_procid nonidempotent
nfsv2_repstat nfsv2_type nfsrv_nfsv3_procid
nfsrvv2_procid nfsrv_v2errmap nfsv3err_null
nfsv3err_getattr nfsv3err_setattr nfsv3err_lookup
nfsv3err_access nfsv3err_readlink nfsv3err_read
nfsv3err_write nfsv3err_create nfsv3err_mkdir
nfsv3err_symlink nfsv3err_mknod nfsv3err_remove
nfsv3err_rmdir nfsv3err_rename nfsv3err_link
nfsv3err_readdir nfsv3err_readdirplus nfsv3err_fsstat
nfsv3err_fsinfo nfsv3err_pathconf nfsv3err_commit
nfsrv_v3errmap
There are additional structures that should be constified but due
to their being passed into general purpose functions without const
arguments, I have not yet converted.
In general, acquire nfsd_mtx when accessing any of the global NFS
structures, including struct nfssvc_sock, struct nfsd, struct
nfsrv_descript.
Release nfsd_mtx whenever calling into VFS, and acquire Giant for
calls into VFS. Giant is not required for any part of the
operation of the NFS server with the exception of calls into VFS.
Giant will never by acquired in the upcall code path. However, it
may operate entirely covered by Giant, or not. If debug.mpsafenet
is set to 0, the system calls will acquire Giant across all
operations, and the upcall will assert Giant. As such, by default,
this enables locking and allows us to test assertions, but should not
cause any substantial new amount of code to be run without Giant.
Bugs should manifest in the form of lock assertion failures for now.
This approach is similar (but not identical) to modifications to the
BSD/OS NFS server code snapshot provided by BSDi as part of their
SMPng snapshot. The strategy is almost the same (single lock over
the NFS server), but differs in the following ways:
- Our NFS client and server code bases don't overlap, which means
both fewer bugs and easier locking (thanks Peter!). Also means
NFSD_*() as opposed to NFS_*().
- We make broad use of assertions, whereas the BSD/OS code does not.
- Made slightly different choices about how to handle macros building
packets but operating with side effects.
- We acquire Giant only when entering VFS from the NFS server daemon
threads.
- Serious bugs in BSD/OS implementation corrected -- the snapshot we
received was clearly a work in progress.
Based on ideas from: BSDi SMPng Snapshot
Reviewed by: rick@snowhite.cis.uoguelph.ca
Extensive testing by: kris
FreeBSD, we can have a negative available space value, but the
corresponding fields in the NFS protocol are unsigned. So
trnucate the value to 0 if it's negative, so that the client
doesn't receive absurdly high values.
Tested by: cognet
short read operations at the end of a file to not have the "eof"
flag set as they should. The problem is that the requested read
count was compared against the rounded-up reply data length instead
of the actual reply data length. This bug appears to have been
introduced in revision 1.78 (June 1999). It causes first-time reads
of certain file sizes (e.g 4094 bytes) to fail with EIO on a RedHat
9.0 NFSv3 client.
MFC after: 1 week
when serving up more than about 32 active files. For details see
section 6.3 (pg 111) of Daniel Ellard and Margo Seltzer, ``NFS
Tricks and Benchmarking Traps'' in the Proceedings of the Usenix
2003 Freenix Track, June 9-14, 2003 pg 101-114.
Obtained from: Daniel Ellard <ellard@eecs.harvard.edu>
Sponsored by: DARPA & NAI Labs.
is not pretty, but it fixes the code so that it no longer violates the
vnode locking rules in the VFS API and doesn't trip any of the locking
assertions enabled by the DEBUG_VFS_LOCKS kernel configuration option.
There is one report that this patch fixed a "locking against myself"
panic on an NFS server that was tripped by a diskless client.
Approved by: re (scottl)
- Add a parameter to vm_pageout_flush() that tells vm_pageout_flush()
whether its caller has locked the vm_object. (This is a temporary
measure to bootstrap vm_object locking.)
- Remove the buftimelock mutex and acquire the buf's interlock to protect
these fields instead.
- Hold the vnode interlock while locking bufs on the clean/dirty queues.
This reduces some cases from one BUF_LOCK with a LK_NOWAIT and another
BUF_LOCK with a LK_TIMEFAIL to a single lock.
Reviewed by: arch, mckusick
32k read and write operations on datagram sockets when in fact we
reject requests larger than 16k. It must be the case that virtually
all clients use data sizes of 16k or less for UDP transport (FreeBSD's
client defaults to 8k and never exceeds 16k), as this bug has been
present ever since NFSv3 support was added.
Reported by: Senthil <lihtnes78@netscape.net>
Reviewed by: dillon
Approved by: re
MFC-after: 1 week
nfsm_srvpathsiz. This macro plucks a length out of an rpc request and
verifies that its size does not exceed NFS_MAXPATHLEN. If it does
it generates an ENAMETOOLONG response.
- Use this macro, and the existing nfsm_srvnamsiz macro in two places
where we deal with paths passed in by the client.
This fixes a linux interoperability bug. Linux was sending oversized path
components which would cause us to ignore the request all together. This
causes linux to hang indefinitly while it waits for a response. This
could still happen in other cases where we error out with EBADRPC.
Sponsored by: Isilon Systems, Inc.
Reviewed by: alfred, fabbri@isilon.com, neal@isilon.com
- v_vflag is protected by the vnode lock and is used when synchronization
with VOP calls is needed.
- v_iflag is protected by interlock and is used for dealing with vnode
management issues. These flags include X/O LOCK, FREE, DOOMED, etc.
- All accesses to v_iflag and v_vflag have either been locked or marked with
mp_fixme's.
- Many ASSERT_VOP_LOCKED calls have been added where the locking was not
clear.
- Many functions in vfs_subr.c were restructured to provide for stronger
locking.
Idea stolen from: BSD/OS
methodology similar to the vm_map_entry splay and the VM splay that Alan
Cox is working on. Extensive testing has appeared to have shown no
increase in overhead.
Disadvantages
Dirties more cache lines during lookups.
Not as fast as a hash table lookup (but still N log N and optimal
when there is locality of reference).
Advantages
vnode->v_dirtyblkhd is now perfectly sorted, making fsync/sync/filesystem
syncer operate more efficiently.
I get to rip out all the old hacks (some of which were mine) that tried
to keep the v_dirtyblkhd tailq sorted.
The per-vnode splay tree should be easier to lock / SMPng pushdown on
vnodes will be easier.
This commit along with another that Alan is working on for the VM page
global hash table will allow me to implement ranged fsync(), optimize
server-side nfs commit rpcs, and implement partial syncs by the
filesystem syncer (aka filesystem syncer would detect that someone is
trying to get the vnode lock, remembers its place, and skip to the
next vnode).
Note that the buffer cache splay is somewhat more complex then other splays
due to special handling of background bitmap writes (multiple buffers with
the same lblkno in the same vnode), and B_INVAL discontinuities between the
old hash table and the existence of the buffer on the v_cleanblkhd list.
Suggested by: alc
nfsrv_readdir and nfsrv_readdirplus can return. A client request
containing an over-large `count' field could trigger the "Bad nfs
svc reply" panic in nfs_syscalls.c.
Spotted while trying to reproduce kern/37304, which turned out to
be fixed in FreeBSD a long time ago.
MFC after: 1 week
general cleanup of the API. The entire API now consists of two functions
similar to the pre-KSE API. The suser() function takes a thread pointer
as its only argument. The td_ucred member of this thread must be valid
so the only valid thread pointers are curthread and a few kernel threads
such as thread0. The suser_cred() function takes a pointer to a struct
ucred as its first argument and an integer flag as its second argument.
The flag is currently only used for the PRISON_ROOT flag.
Discussed on: smp@
locking flags when acquiring a vnode. The immediate purpose is
to allow polling lock requests (LK_NOWAIT) needed by soft updates
to avoid deadlock when enlisting other processes to help with
the background cleanup. For the future it will allow the use of
shared locks for read access to vnodes. This change touches a
lot of files as it affects most filesystems within the system.
It has been well tested on FFS, loopback, and CD-ROM filesystems.
only lightly on the others, so if you find a problem there, please
let me (mckusick@mckusick.com) know.
server side. This can lead to a system deadlock.
Reviewed by: iedowse
Tested by: Alexey G Misurenko <mag@caravan.ru>, iedowse
Bug found with help by: Alexey G Misurenko <mag@caravan.ru>
MFC at: earliest convenience
that va_mode of the supplied attributes is filled in with a valid
file mode (i.e not VNOVAL, and only ALLPERM bits set). However,
some NFS server op functions didn't guarantee this for all possible
request messages:
If a V3 client chose not include to a mode specification, we could
end up creating an ffs inode with mode 0177777, requiring a manual
fsck on the next reboot. Fix this by setting va_mode to 0 before
calling the VOP if a mode hasn't been supplied by the client.
In nfsrv_symlink(), S_IFMT bits supplied by a V2 client could end
up in the va_mode passed to VOP_SYMLINK with similar effects. We
now use the macro nfstov_mode() to correctly mask the bits.
semantics of the nfsm_reply() macro were changed so that the caller
has to explicitly handle the V2 error case, whereas before,
nfsm_reply() did a `goto nfsmout' then. A few server ops (setattr,
readlink, create, mkdir) weren't updated to match, so errors in the
V2 case could cause protocol hangs and leaked mbufs.
Correct some comments that describe the old nfsm_reply behaviour.
[older, harmless nit] Remove the unnecessary `nfsmreply0' label in
nfsrv_create(), since for its users, the main `ereply' label does
the same thing.
temporary storage. In the old NFS code it wasn't at all clear if
the value of `tl' was used across or after macro calls, but I'm
fairly confident that the convention was to keep its use local.
Each ex-macro function now uses a local version of this variable,
so all of the double-indirection goes away.
The only exception to the `local use' rule for `tl' is nfsm_clget(),
which is left unchanged by this commit.
Reviewed by: peter
out nd.ni_vp to prevent the resource cleanup code at the end of
nfsrv_symlink from trying to vrele it. This fixes a "vrele: negative
ref cnt" panic that can occur when a symlink is attempted on an NFS
filesystem with no free space. Found locally, but the symptoms
correspond to those in the PR referenced below.
PR: kern/26878
MFC after: 3 days
to do it explicitly in nfsrv_noop so that the reply gets sent back
to the client. This fixes the generation of a selection of RPC
error replies (RPC_PROGMISMATCH, RPC_PROGUNAVAIL, RPC_PROCUNAVAIL
etc.) that are used by some clients to detect support for optional
protocols and features.
Reviewed by: peter
Reported by: Thomas Quinot <quinot@inf.enst.fr>
PR: kern/31479
next to equivalent m_len adjustments. Move the nfsm_subs.h macros
into groups depending on which phase they are used in, since that
affects the error recovery requirements. Collect some of the common error
checking into a single macro as preparation for unwinding some more.
Have nfs_rephead return a value instead of secretly modifying args.
Remove some unused function arguments that were being passed around.
Clarify nfsm_reply()'s error handling (I hope).
Note ALL MODULES MUST BE RECOMPILED
make the kernel aware that there are smaller units of scheduling than the
process. (but only allow one thread per process at this time).
This is functionally equivalent to teh previousl -current except
that there is a thread associated with each process.
Sorry john! (your next MFC will be a doosie!)
Reviewed by: peter@freebsd.org, dillon@freebsd.org
X-MFC after: ha ha ha ha
of returning an error code to the caller, NFS server op routines
must themselves build an error reply and return 0 to the caller.
This is achieved by replacing the erroneous return statements with
code that jumps forward to the op function's reply code. We need
to be careful to ensure that the 'struct mount' pointer is NULL
though, so that the final vn_finished_write() call becomes a no-op.
Reviewed by: mckusick, dillon
This is because calls with M_WAIT (now M_TRYWAIT) may not wait
forever when nothing is available for allocation, and may end up
returning NULL. Hopefully we now communicate more of the right thing
to developers and make it very clear that it's necessary to check whether
calls with M_(TRY)WAIT also resulted in a failed allocation.
M_TRYWAIT basically means "try harder, block if necessary, but don't
necessarily wait forever." The time spent blocking is tunable with
the kern.ipc.mbuf_wait sysctl.
M_WAIT is now deprecated but still defined for the next little while.
* Fix a typo in a comment in mbuf.h
* Fix some code that was actually passing the mbuf subsystem's M_WAIT to
malloc(). Made it pass M_WAITOK instead. If we were ever to redefine the
value of the M_WAIT flag, this could have became a big problem.
the gating of system calls that cause modifications to the underlying
filesystem. The gating can be enabled by any filesystem that needs
to consistently suspend operations by adding the vop_stdgetwritemount
to their set of vnops. Once gating is enabled, the function
vfs_write_suspend stops all new write operations to a filesystem,
allows any filesystem modifying system calls already in progress
to complete, then sync's the filesystem to disk and returns. The
function vfs_write_resume allows the suspended write operations to
begin again. Gating is not added by default for all filesystems as
for SMP systems it adds two extra locks to such critical kernel
paths as the write system call. Thus, gating should only be added
as needed.
Details on the use and current status of snapshots in FFS can be
found in /sys/ufs/ffs/README.snapshot so for brevity and timelyness
is not included here. Unless and until you create a snapshot file,
these changes should have no effect on your system (famous last words).
