the last component is a symlink to something that isn't a directory.
We introduce a new namei flag, TRAILINGSLASH, which is set by lookup()
if the last component is followed by a slash. The trailing slash is
then stripped, as before. If the final component is a symlink,
lookup() will return to namei(), which will expand the symlink and
call lookup() with the new path. When all symlinks have been
resolved, lookup() checks if the TRAILINGSLASH flag is set, and if it
is, and the vnode it ended up with is not a directory, it returns
ENOTDIR.
PR: kern/21768
Submitted by: Eygene Ryabinkin <rea-fbsd@codelabs.ru>
MFC after: 3 weeks
I don't want people to override the mutex when allocating a TTY. It has
to be there, to keep drivers like syscons happy. So I'm creating a
tty_alloc_mutex() which can be used in those cases. tty_alloc_mutex()
should eventually be removed.
The advantage of this approach, is that we can just remove a function,
without breaking the regular API in the future.
Introduce for this operation the reverse NO_ADAPTIVE_SX option.
The flag SX_ADAPTIVESPIN to be passed to sx_init_flags(9) gets suppressed
and the new flag, offering the reversed logic, SX_NOADAPTIVE is added.
Additively implements adaptive spininning for sx held in shared mode.
The spinning limit can be handled through sysctls in order to be tuned
while the code doesn't reach the release, after which time they should
be dropped probabilly.
This change has made been necessary by recent benchmarks where it does
improve concurrency of workloads in presence of high contention
(ie. ZFS).
KPI breakage is documented by __FreeBSD_version bumping, manpage and
UPDATING updates.
Requested by: jeff, kmacy
Reviewed by: jeff
Tested by: pho
Add support for kernel fault injection using KFAIL_POINT_* macros and
fail_point_* infrastructure. Add example fail point in vfs_bio.c to
simulate VM buf pressure.
Approved by: dfr (mentor)
by creating a child jail, which is visible to that jail and to any
parent jails. Child jails may be restricted more than their parents,
but never less. Jail names reflect this hierarchy, being MIB-style
dot-separated strings.
Every thread now points to a jail, the default being prison0, which
contains information about the physical system. Prison0's root
directory is the same as rootvnode; its hostname is the same as the
global hostname, and its securelevel replaces the global securelevel.
Note that the variable "securelevel" has actually gone away, which
should not cause any problems for code that properly uses
securelevel_gt() and securelevel_ge().
Some jail-related permissions that were kept in global variables and
set via sysctls are now per-jail settings. The sysctls still exist for
backward compatibility, used only by the now-deprecated jail(2) system
call.
Approved by: bz (mentor)
get a quick snapshot of the kernel's symbol table including the symbols
from any loaded modules (the symbols are all merged into one symbol
table). Unlike like other implementations, this ksyms driver maps
memory in the process memory space to store the snapshot at the time
/dev/ksyms is opened. It also checks to see if the process has already
a snapshot open and won't allow it to open /dev/ksyms it again until it
closes first. This prevents kernel and process memory from being
exhausted. Note that /dev/ksyms is used by the lockstat(1) command.
Reviewed by: gallatin kib (freebsd-arch)
Approved by: gnn (mentor)
adds probes for mutexes, reader/writer and shared/exclusive locks to
gather contention statistics and other locking information for
dtrace scripts, the lockstat(1M) command and other potential
consumers.
Reviewed by: attilio jhb jb
Approved by: gnn (mentor)
sleep waiting for conditions when the lock may be granted.
To prevent lf_setlock() from accessing possibly freed memory, add reference
counting to the struct lockf_entry. Bump refcount around the sleep.
Make lf_free_lock() return non-zero when structure was freed, and use
this after the sleep to return EINTR to the caller. The error code might
need a clarification, but we cannot return success to usermode, since
the lock is not owned anymore.
Reviewed by: dfr
Tested by: pho
MFC after: 1 month
is acquired. In the lf_purgelocks(), assert that vnode is doomed and set
*statep to NULL before clearing ls_pending list. Otherwise, we allow for
the thread executing lf_advlockasync() to put new pending entry after
state->ls_lock is dropped in lf_purgelocks().
Reviewed by: dfr
Tested by: pho
MFC after: 1 month
In the original MPSAFE TTY code, I changed the behaviour by returning
EBUSY. I thought this made more sense, because it's basically a race to
see who gets the TTY first.
It turns out this is not a good change, because it also causes EBUSY to
be returned when another process is closing the TTY. This can happen
during startup, when /etc/rc (or one of its children) is still busy
draining its data and /sbin/init is attempting to open the TTY to spawn
a getty.
Reported by: bz
Tested by: bz
to optionally have overlapping unit numbers if attached in different
vnets.
At this stage if_loop is the only clonable ifnet class that has been
extended to allow for such overlapping allocation of unit numbers, i.e.
in each vnet it is possible to have a lo0 interface. Other clonable ifnet
classes remain to operate with traditional semantics, i.e. each instance
of a clonable ifnet will be assigned a globally unique unit number,
regardless in which vnet such an ifnet becomes instantiated.
While here, garbage collect unused _lo_list field in struct vnet_net,
as well as improve indentation for #defines in sys/net/vnet.h.
The layout of struct vnet_net has changed, therefore bump
__FreeBSD_version.
This change has no functional impact on nooptions VIMAGE kernel builds.
Reviewed by: bz, brooks
Approved by: julian (mentor)
for reassigning ifnets from one vnet to another.
if_vmove() works by calling a restricted subset of actions normally
executed by if_detach() on an ifnet in the current vnet, and then
switches to the target vnet and executes an appropriate subset of
if_attach() actions there.
if_attach() and if_detach() have become wrapper functions around
if_attach_internal() and if_detach_internal(), where the later
variants have an additional argument, a flag indicating whether a
full attach or detach sequence is to be executed, or only a
restricted subset suitable for moving an ifnet from one vnet to
another. Hence, if_vmove() will not call if_detach() and if_attach()
directly, but will call the if_detach_internal() and
if_attach_internal() variants instead, with the vmove flag set.
While here, staticize ifnet_setbyindex() since it is not referenced
from outside of sys/net/if.c.
Also rename ifccnt field in struct vimage to ifcnt, and do some minor
whitespace garbage collection where appropriate.
This change should have no functional impact on nooptions VIMAGE kernel
builds.
Reviewed by: bz, rwatson, brooks?
Approved by: julian (mentor)
When enabled all TTY input queue buffers are zeroed when flushing or
closing the TTY. Because TTY input queues are also used to store filled
in passwords, this may be an interesting switch to enable for security
minded people.
the behaviour alredy present with the further malloc() call in
devctl_notify().
This fixes a bug in the CAM layer where the camisr handler finished to
call camperiphfree() (and subsequently destroy_dev() resulting in a new
dev notify) while the xpt lock is held.
PR: kern/130330
Tested by: Riccardo Torrini <riccardo dot torrini at esaote dot com>
shared uses of a resource are recorded on a sub-list hanging off
a main resource object on a main resource list;
without this change a shared resource (e.g. irq) is reported only
once by devinfo -r/-u;
with this change the resource is reported for each driver that
allocates it (which is even more than what vmstat -i -a reports).
Approved by: jhb (mentor)
affinity for the interrupt thread, and requesting that underlying
hardware direct interrupts to the CPU. For software interrupt
threads, implement a no-op interrupt event binder that returns
success, so that the interrupt management code will just set the
ithread's affinity and succeed.
Reviewed by: jhb
MFC after: 1 week
These sysctls don't need any form of locking. At least cp_times is used
by powerd very often, which means I get 50% less calls to non-MPSAFE
sysctls on my system. The other 50% is consumed by dev.cpu.0.freq, but
this seems to need Giant for Newbus.
Provide a more descriptive comment.
Eliminate dead code. The page cannot possibly have PG_ZERO set.
Eliminate unnecessary blank lines.
Reviewed by: tegge
This makes siginfo output look a lot better when pressing it the first
time when in sh(1), for example:
$ load: 0.00 cmd: sh 1945 [ttyin] 3.94r 0.00u 0.00s 0% 1960k
load: 0.00 cmd: sh 1945 [ttyin] 4.19r 0.00u 0.00s 0% 1960k
will now become:
$
load: 0.00 cmd: sh 1945 [ttyin] 3.94r 0.00u 0.00s 0% 1960k
load: 0.00 cmd: sh 1945 [ttyin] 4.19r 0.00u 0.00s 0% 1960k
- Only pick up PROC_LOCK once, which means we can drop the PGRP_LOCK
right after picking up PROC_LOCK for the first time.
- Print the process real time, making it consistent with tools like
time(1).
- Use `p' and `td' to reference the process/thread we are going to
print. Only use pick-variables inside the loops. We already did this
for the threads, but not the processes.
that expect that oldlen is filled with required buffer length even when
supplied buffer is too short and returned error is ENOMEM.
Redo the fix for kern.proc.filedesc, by reverting the req->oldidx when
remaining buffer space is too short for the current kinfo_file structure.
Also, only ignore ENOMEM. We have to convert ENOMEM to no error condition
to keep existing interface for the sysctl, though.
Reported by: ed, Florian Smeets <flo kasimir com>
Tested by: pho
sysctl requests to avoid wiring too much user memory. Only grab this
lock if the user's old buffer is larger than a page as a tradeoff to
allow more concurrency for common small requests.
- Just use a shared lock on the sysctl tree for user sysctl requests now.
MFC after: 1 week
error due to copyout failure or short buffer.
The later breaks the usermode iterators of the sysctl results that pack
arbitrary number of variable-sized structures. Iterator expects that
kernel filled exactly oldlen bytes, and tries to interpret half-filled
or garbage structure at the end of the buffer. In particular,
kinfo_getfile(3) segfaulted.
Reported and tested by: pho
MFC after: 3 weeks
fget_unlocked().
- Save old file descriptor tables created on expansion until
the entire descriptor table is freed so that pointers may be
followed without regard for expanders.
- Mark the file zone as NOFREE so we may attempt to reference
potentially freed files.
- Convert several fget_locked() users to fget_unlocked(). This
requires us to manage reference counts explicitly but reduces
locking overhead in the common case.
following changes:
Rename vfs_page_set_valid() to vfs_page_set_validclean() to reflect
what this function actually does. Suggested by: tegge
Introduce a new version of vfs_page_set_valid() that does no more than
what the function's name implies. Specifically, it does not update
the page's dirty mask, and thus it does not require the page queues
lock to be held.
Update two of the three callers to the old vfs_page_set_valid() to
call vfs_page_set_validclean() instead because they actually require
the page's dirty mask to be cleared.
Introduce vm_page_set_valid().
Reviewed by: tegge
the VFS. Now all the VFS_* functions and relating parts don't want the
context as long as it always refers to curthread.
In some points, in particular when dealing with VOPs and functions living
in the same namespace (eg. vflush) which still need to be converted,
pass curthread explicitly in order to retain the old behaviour.
Such loose ends will be fixed ASAP.
While here fix a bug: now, UFS_EXTATTR can be compiled alone without the
UFS_EXTATTR_AUTOSTART option.
VFS KPI is heavilly changed by this commit so thirdy parts modules needs
to be recompiled. Bump __FreeBSD_version in order to signal such
situation.
suffer from the race condition that motivated revision 1.94. Consequently,
the work-around that was implemented by revision 1.94 is no longer needed.
Moreover, reverting this work-around eliminates the need for
vfs_busy_pages() to acquire the page queues lock when preparing a buffer
for read.
Reviewed by: tegge
Credential might need to hang around longer than its parent and be used
outside of mnt_explock scope controlling netcred lifetime. Use separate
reference-counted ucred allocated separately instead.
While there, extend mnt_explock coverage in vfs_stdexpcheck and clean-up
some unused declarations in new NFS code.
Reported by: John Hickey
PR: kern/133439
Reviewed by: dfr, kib
virtualized instances of hostname and domainname, as well as a new top-level
virtualization struct vimage, which holds pointers to struct vnet and struct
vprocg. Struct vprocg is likely to become replaced in the near future with
a new jail management API import.
As a consequence of this change, change struct ucred to point to a struct
vimage, instead of directly pointing to a vnet.
Merge vnet / vimage / ucred refcounting infrastructure from p4 / vimage
branch.
Permit kldload / kldunload operations to be executed only from the default
vimage context.
This change should have no functional impact on nooptions VIMAGE kernel
builds.
Reviewed by: bz
Approved by: julian (mentor)
OSD-based jail extensions. This allows the Linux MIB to accessed via
jail_set and jail_get, and serves as a demonstration of adding jail support
to a module.
Reviewed by: dchagin, kib
Approved by: bz (mentor)
It turns out if we called cfmakeraw() on a TTY with only a rint handler
in place, it could inject data into the TTY, even though it should be
redirected. Always take a look at the hooks before looking at the
termios flags.
previously always pointing to the default vnet context, to a
dynamically changing thread-local one. The currvnet context
should be set on entry to networking code via CURVNET_SET() macros,
and reverted to previous state via CURVNET_RESTORE(). Recursions
on curvnet are permitted, though strongly discuouraged.
This change should have no functional impact on nooptions VIMAGE
kernel builds, where CURVNET_* macros expand to whitespace.
The curthread->td_vnet (aka curvnet) variable's purpose is to be an
indicator of the vnet context in which the current network-related
operation takes place, in case we cannot deduce the current vnet
context from any other source, such as by looking at mbuf's
m->m_pkthdr.rcvif->if_vnet, sockets's so->so_vnet etc. Moreover, so
far curvnet has turned out to be an invaluable consistency checking
aid: it helps to catch cases when sockets, ifnets or any other
vnet-aware structures may have leaked from one vnet to another.
The exact placement of the CURVNET_SET() / CURVNET_RESTORE() macros
was a result of an empirical iterative process, whith an aim to
reduce recursions on CURVNET_SET() to a minimum, while still reducing
the scope of CURVNET_SET() to networking only operations - the
alternative would be calling CURVNET_SET() on each system call entry.
In general, curvnet has to be set in three typicall cases: when
processing socket-related requests from userspace or from within the
kernel; when processing inbound traffic flowing from device drivers
to upper layers of the networking stack, and when executing
timer-driven networking functions.
This change also introduces a DDB subcommand to show the list of all
vnet instances.
Approved by: julian (mentor)
active network stack instance. Turning on options VIMAGE at compile
time yields the following changes relative to default kernel build:
1) V_ accessor macros for virtualized variables resolve to structure
fields via base pointers, instead of being resolved as fields in global
structs or plain global variables. As an example, V_ifnet becomes:
options VIMAGE: ((struct vnet_net *) vnet_net)->_ifnet
default build: vnet_net_0._ifnet
options VIMAGE_GLOBALS: ifnet
2) INIT_VNET_* macros will declare and set up base pointers to be used
by V_ accessor macros, instead of resolving to whitespace:
INIT_VNET_NET(ifp->if_vnet); becomes
struct vnet_net *vnet_net = (ifp->if_vnet)->mod_data[VNET_MOD_NET];
3) Memory for vnet modules registered via vnet_mod_register() is now
allocated at run time in sys/kern/kern_vimage.c, instead of per vnet
module structs being declared as globals. If required, vnet modules
can now request the framework to provide them with allocated bzeroed
memory by filling in the vmi_size field in their vmi_modinfo structures.
4) structs socket, ifnet, inpcbinfo, tcpcb and syncache_head are
extended to hold a pointer to the parent vnet. options VIMAGE builds
will fill in those fields as required.
5) curvnet is introduced as a new global variable in options VIMAGE
builds, always pointing to the default and only struct vnet.
6) struct sysctl_oid has been extended with additional two fields to
store major and minor virtualization module identifiers, oid_v_subs and
oid_v_mod. SYSCTL_V_* family of macros will fill in those fields
accordingly, and store the offset in the appropriate vnet container
struct in oid_arg1.
In sysctl handlers dealing with virtualized sysctls, the
SYSCTL_RESOLVE_V_ARG1() macro will compute the address of the target
variable and make it available in arg1 variable for further processing.
Unused fields in structs vnet_inet, vnet_inet6 and vnet_ipfw have
been deleted.
Reviewed by: bz, rwatson
Approved by: julian (mentor)
interface as nmount(2). Three new system calls are added:
* jail_set, to create jails and change the parameters of existing jails.
This replaces jail(2).
* jail_get, to read the parameters of existing jails. This replaces the
security.jail.list sysctl.
* jail_remove to kill off a jail's processes and remove the jail.
Most jail parameters may now be changed after creation, and jails may be
set to exist without any attached processes. The current jail(2) system
call still exists, though it is now a stub to jail_set(2).
Approved by: bz (mentor)
import from p4 bms_netdev. Summary of changes:
* Connect netinet6/in6_mcast.c to build.
The legacy KAME KPIs are mostly preserved.
* Eliminate now dead code from ip6_output.c.
Don't do mbuf bingo, we are not going to do RFC 2292 style
CMSG tricks for multicast options as they are not required
by any current IPv6 normative reference.
* Refactor transports (UDP, raw_ip6) to do own mcast filtering.
SCTP, TCP unaffected by this change.
* Add ip6_msource, in6_msource structs to in6_var.h.
* Hookup mld_ifinfo state to in6_ifextra, allocate from
domifattach path.
* Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced.
Kernel consumers which need this should use in6m_lookup().
* Refactor IPv6 socket group memberships to use a vector (like IPv4).
* Update ifmcstat(8) for IPv6 SSM.
* Add witness lock order for IN6_MULTI_LOCK.
* Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths.
* Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup.
* Update carp(4) for new IPv6 SSM KPIs.
* Virtualize ip6_mrouter socket.
Changes mostly localized to IPv6 MROUTING.
* Don't do a local group lookup in MROUTING.
* Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge().
* Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode.
* Bump __FreeBSD_version to 800084.
* Update UPDATING.
NOTE WELL:
* This code hasn't been tested against real MLDv2 queriers
(yet), although the on-wire protocol has been verified in Wireshark.
* There are a few unresolved issues in the socket layer APIs to
do with scope ID propagation.
* There is a LOR present in ip6_output()'s use of
in6_setscope() which needs to be resolved. See comments in mld6.c.
This is believed to be benign and can't be avoided for the moment
without re-introducing an indirect netisr.
This work was mostly derived from the IGMPv3 implementation, and
has been sponsored by a third party.
and it only optimized out an ipi or mwait in very few cases.
- Skip the adaptive idle code when running on SMT or HTT cores. This
just wastes cpu time that could be used on a busy thread on the same
core.
- Rename CG_FLAG_THREAD to CG_FLAG_SMT to be more descriptive. Re-use
CG_FLAG_THREAD to mean SMT or HTT.
Sponsored by: Nokia
root cpuset of that jail.
Processes inside the jail will still be able to change child sets.
A superuser outside of a jail will still be able to change the jail cpuset
and thus limit the number of cpus available to the jail.
Problem reported by: 000.fbsd@quip.cz (Miroslav Lachman)
PR: kern/134050
Reviewed by: jeff
MFC after: 3 weeks
X-MFC: backout r191596
first introduced @ r190909 with a vnet module deregistration
service.
kldunloadable modules, which are currently using vnet_mod_register()
to attach their per-vnet initialization routines to the vnet
initialization framework, should call vnet_mod_deregister() before
acknowledging MOD_UNLOAD requests in their mod_event handlers. Such
changes to the existing code base will follow in subsequent commits.
vnet_mod_deregister() does not check whether departing vnet modules
are registered as prerequisites for another module(s), so it should
be used with care. Currently I'm only aware of vnet modules which
are leafs on module dependency graphs that are kldunloadable.
This change also introduces per-vnet module destructor handler, which
calls vnet's module cleanup function, which (if required) has to be
registered in vnet module's vnet_modinfo_t structure .vmi_idetach
field. Once options VIMAGE becomes operational, the framework will
take care that module's cleanup function become invoked for each
active vnet instance, and that the memory allocated for each instance
gets freed. Currently calls to destructor handlers must always
succeed.
This allows users to increase the maximum amount of pseudo-terminals
without changing any source code. Users must increase UT_LINESIZE before
attempting to increase kern.pts_maxdev.
The main problem is that sbappendrecord_locked() relies on sbcompress()
to set sb_mbtail. This will not happen if sbappendrecord_locked() is
called with mbuf chain made of exactly one mbuf (i.e. m0->m_next == NULL).
In this case sbcompress() will be called with m == NULL and will do
nothing. I'm not entirely sure if m == NULL is a valid argument for
sbcompress(), and, it rather pointless to call it like that, but keep
calling it so it can do SBLASTMBUFCHK().
The problem is triggered by the SOCKBUF_DEBUG kernel option that
enables SBLASTRECORDCHK() and SBLASTMBUFCHK() checks.
PR: kern/126742
Investigated by: pluknet < pluknet -at- gmail -dot- com >
No response from: freebsd-current@, freebsd-bluetooth@
MFC after: 3 days
or ignored SIGCHLD, unconditionally wake up the parent instead of doing
this only when the child is a last child.
This brings us in line with other U**xes that support SA_NOCLDWAIT. If
the parent called waitpid(childpid), then exit of the child should wake
up the parent immediately instead of forcing it to wait for all children
to exit.
Reported by: Alan Ferrency <alan pair com>
Submitted by: Jilles Tjoelker <jilles stack nl>
PR: 108390
MFC after: 2 weeks
In the good old days it was possible to have dev_t's that referred to
nonexistent devices. In these cases devtoname() automatically generated
names. This is no longer possible, so remove this dead code.
Discussed with: kib
Remove the `udev' variable, which has a different type than the original
function argument and si_drv0. The `udev' name is also misleading,
because it is not the number returned by dev2udev(). Rename this
argument to `unit'. It is the same number as returned by dev2unit().
We should still access si_drv0 using dev2unit(). Also change the
KASSERT() to really print the udev instead of the unit number. I suspect
it's still useful to print the unit number, especially for devices that
use clone lists, so keep the unit number in the panic string.
not populated in parent directory if negative entry was being
created, yet entry itself was added to the nc_neg list. It was
possible for parent vnode to get discarded later, leaving negative
entry pointing to now unused memory block.
Reported by: dho
Revewed by: kib
dependency tracking and ordering enforcement.
With this change, per-vnet initialization functions introduced with
r190787 are no longer directly called from traditional initialization
functions (which cc in most cases inlined to pre-r190787 code), but are
instead registered via the vnet framework first, and are invoked only
after all prerequisite modules have been initialized. In the long run,
this framework should allow us to both initialize and dismantle
multiple vnet instances in a correct order.
The problem this change aims to solve is how to replay the
initialization sequence of various network stack components, which
have been traditionally triggered via different mechanisms (SYSINIT,
protosw). Note that this initialization sequence was and still can be
subtly different depending on whether certain pieces of code have been
statically compiled into the kernel, loaded as modules by boot
loader, or kldloaded at run time.
The approach is simple - we record the initialization sequence
established by the traditional mechanisms whenever vnet_mod_register()
is called for a particular vnet module. The vnet_mod_register_multi()
variant allows a single initializer function to be registered multiple
times but with different arguments - currently this is only used in
kern/uipc_domain.c by net_add_domain() with different struct domain *
as arguments, which allows for protosw-registered initialization
routines to be invoked in a correct order by the new vnet
initialization framework.
For the purpose of identifying vnet modules, each vnet module has to
have a unique ID, which is statically assigned in sys/vimage.h.
Dynamic assignment of vnet module IDs is not supported yet.
A vnet module may specify a single prerequisite module at registration
time by filling in the vmi_dependson field of its vnet_modinfo struct
with the ID of the module it depends on. Unless specified otherwise,
all vnet modules depend on VNET_MOD_NET (container for ifnet list head,
rt_tables etc.), which thus has to and will always be initialized
first. The framework will panic if it detects any unresolved
dependencies before completing system initialization. Detection of
unresolved dependencies for vnet modules registered after boot
(kldloaded modules) is not provided.
Note that the fact that each module can specify only a single
prerequisite may become problematic in the long run. In particular,
INET6 depends on INET being already instantiated, due to TCP / UDP
structures residing in INET container. IPSEC also depends on INET,
which will in turn additionally complicate making INET6-only kernel
configs a reality.
The entire registration framework can be compiled out by turning on the
VIMAGE_GLOBALS kernel config option.
Reviewed by: bz
Approved by: julian (mentor)
the removal of NQNFS, but was left in in case it was required for NFSv4.
Since our new NFSv4 client and server can't use it for their
requirements, GC the old mechanism, as well as other unused lease-
related code and interfaces.
Due to its impact on kernel programming and binary interfaces, this
change should not be MFC'd.
Proposed by: jeff
Reviewed by: jeff
Discussed with: rmacklem, zach loafman @ isilon
Check the condition and return ENOENT then.
In nfs_lookup(), respect ENOENT return from cache_lookup() when it is caused
by dvp reclaim.
Reported and tested by: pho
