The current mainline has copies propagated to *all* nodes, then
tears down the copies we made for nodes that do not contain
counterparts of the desired mountpoint. That sets the right
propagation graph for the copies (at teardown time we move
the slaves of removed node to a surviving peer or directly
to master), but we end up paying a fairly steep price in
useless allocations. It's fairly easy to create a situation
where N calls of mount(2) create exactly N bindings, with
O(N^2) vfsmounts allocated and freed in process.
Fortunately, it is possible to avoid those allocations/freeings.
The trick is to create copies in the right order and find which
one would've eventually become a master with the current algorithm.
It turns out to be possible in O(nodes getting propagation) time
and with no extra allocations at all.
One part is that we need to make sure that eventual master will be
created before its slaves, so we need to walk the propagation
tree in a different order - by peer groups. And iterate through
the peers before dealing with the next group.
Another thing is finding the (earlier) copy that will be a master
of one we are about to create; to do that we are (temporary) marking
the masters of mountpoints we are attaching the copies to.
Either we are in a peer of the last mountpoint we'd dealt with,
or we have the following situation: we are attaching to mountpoint M,
the last copy S_0 had been attached to M_0 and there are sequences
S_0...S_n, M_0...M_n such that S_{i+1} is a master of S_{i},
S_{i} mounted on M{i} and we need to create a slave of the first S_{k}
such that M is getting propagation from M_{k}. It means that the master
of M_{k} will be among the sequence of masters of M. On the
other hand, the nearest marked node in that sequence will either
be the master of M_{k} or the master of M_{k-1} (the latter -
in the case if M_{k-1} is a slave of something M gets propagation
from, but in a wrong peer group).
So we go through the sequence of masters of M until we find
a marked one (P). Let N be the one before it. Then we go through
the sequence of masters of S_0 until we find one (say, S) mounted
on a node D that has P as master and check if D is a peer of N.
If it is, S will be the master of new copy, if not - the master of S
will be.
That's it for the hard part; the rest is fairly simple. Iterator
is in next_group(), handling of one prospective mountpoint is
propagate_one().
It seems to survive all tests and gives a noticably better performance
than the current mainline for setups that are seriously using shared
subtrees.
Change-Id: I45648e8a405544f768c5956711bdbdf509e2705a
Cc: stable@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This starts to add private data associated directly
to mount points. The intent is to give filesystems
a sense of where they have come from, as a means of
letting a filesystem take different actions based on
this information.
Change-Id: Ie769d7b3bb2f5972afe05c1bf16cf88c91647ab2
Signed-off-by: Daniel Rosenberg <drosen@google.com>
commit 9566d6742852c527bf5af38af5cbb878dad75705 upstream.
While invesgiating the issue where in "mount --bind -oremount,ro ..."
would result in later "mount --bind -oremount,rw" succeeding even if
the mount started off locked I realized that there are several
additional mount flags that should be locked and are not.
In particular MNT_NOSUID, MNT_NODEV, MNT_NOEXEC, and the atime
flags in addition to MNT_READONLY should all be locked. These
flags are all per superblock, can all be changed with MS_BIND,
and should not be changable if set by a more privileged user.
The following additions to the current logic are added in this patch.
- nosuid may not be clearable by a less privileged user.
- nodev may not be clearable by a less privielged user.
- noexec may not be clearable by a less privileged user.
- atime flags may not be changeable by a less privileged user.
The logic with atime is that always setting atime on access is a
global policy and backup software and auditing software could break if
atime bits are not updated (when they are configured to be updated),
and serious performance degradation could result (DOS attack) if atime
updates happen when they have been explicitly disabled. Therefore an
unprivileged user should not be able to mess with the atime bits set
by a more privileged user.
The additional restrictions are implemented with the addition of
MNT_LOCK_NOSUID, MNT_LOCK_NODEV, MNT_LOCK_NOEXEC, and MNT_LOCK_ATIME
mnt flags.
Taken together these changes and the fixes for MNT_LOCK_READONLY
should make it safe for an unprivileged user to create a user
namespace and to call "mount --bind -o remount,... ..." without
the danger of mount flags being changed maliciously.
Acked-by: Serge E. Hallyn <serge.hallyn@ubuntu.com>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a6138db815df5ee542d848318e5dae681590fccd upstream.
Kenton Varda <kenton@sandstorm.io> discovered that by remounting a
read-only bind mount read-only in a user namespace the
MNT_LOCK_READONLY bit would be cleared, allowing an unprivileged user
to the remount a read-only mount read-write.
Correct this by replacing the mask of mount flags to preserve
with a mask of mount flags that may be changed, and preserve
all others. This ensures that any future bugs with this mask and
remount will fail in an easy to detect way where new mount flags
simply won't change.
Acked-by: Serge E. Hallyn <serge.hallyn@ubuntu.com>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When a read-only bind mount is copied from mount namespace in a higher
privileged user namespace to a mount namespace in a lesser privileged
user namespace, it should not be possible to remove the the read-only
restriction.
Add a MNT_LOCK_READONLY mount flag to indicate that a mount must
remain read-only.
CC: stable@vger.kernel.org
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
This allows us to move duplicated code in <asm/atomic.h>
(atomic_inc_not_zero() for now) to <linux/atomic.h>
Signed-off-by: Arun Sharma <asharma@fb.com>
Reviewed-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: David Miller <davem@davemloft.net>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Instead of splitting refcount between (per-cpu) mnt_count
and (SMP-only) mnt_longrefs, make all references contribute
to mnt_count again and keep track of how many are longterm
ones.
Accounting rules for longterm count:
* 1 for each fs_struct.root.mnt
* 1 for each fs_struct.pwd.mnt
* 1 for having non-NULL ->mnt_ns
* decrement to 0 happens only under vfsmount lock exclusive
That allows nice common case for mntput() - since we can't drop the
final reference until after mnt_longterm has reached 0 due to the rules
above, mntput() can grab vfsmount lock shared and check mnt_longterm.
If it turns out to be non-zero (which is the common case), we know
that this is not the final mntput() and can just blindly decrement
percpu mnt_count. Otherwise we grab vfsmount lock exclusive and
do usual decrement-and-check of percpu mnt_count.
For fs_struct.c we have mnt_make_longterm() and mnt_make_shortterm();
namespace.c uses the latter in places where we don't already hold
vfsmount lock exclusive and opencodes a few remaining spots where
we need to manipulate mnt_longterm.
Note that we mostly revert the code outside of fs/namespace.c back
to what we used to have; in particular, normal code doesn't need
to care about two kinds of references, etc. And we get to keep
the optimization Nick's variant had bought us...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Unexport do_add_mount() and make ->d_automount() return the vfsmount to be
added rather than calling do_add_mount() itself. follow_automount() will then
do the addition.
This slightly complicates things as ->d_automount() normally wants to add the
new vfsmount to an expiration list and start an expiration timer. The problem
with that is that the vfsmount will be deleted if it has a refcount of 1 and
the timer will not repeat if the expiration list is empty.
To this end, we require the vfsmount to be returned from d_automount() with a
refcount of (at least) 2. One of these refs will be dropped unconditionally.
In addition, follow_automount() must get a 3rd ref around the call to
do_add_mount() lest it eat a ref and return an error, leaving the mount we
have open to being expired as we would otherwise have only 1 ref on it.
d_automount() should also add the the vfsmount to the expiration list (by
calling mnt_set_expiry()) and start the expiration timer before returning, if
this mechanism is to be used. The vfsmount will be unlinked from the
expiration list by follow_automount() if do_add_mount() fails.
This patch also fixes the call to do_add_mount() for AFS to propagate the mount
flags from the parent vfsmount.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
The problem that this patch aims to fix is vfsmount refcounting scalability.
We need to take a reference on the vfsmount for every successful path lookup,
which often go to the same mount point.
The fundamental difficulty is that a "simple" reference count can never be made
scalable, because any time a reference is dropped, we must check whether that
was the last reference. To do that requires communication with all other CPUs
that may have taken a reference count.
We can make refcounts more scalable in a couple of ways, involving keeping
distributed counters, and checking for the global-zero condition less
frequently.
- check the global sum once every interval (this will delay zero detection
for some interval, so it's probably a showstopper for vfsmounts).
- keep a local count and only taking the global sum when local reaches 0 (this
is difficult for vfsmounts, because we can't hold preempt off for the life of
a reference, so a counter would need to be per-thread or tied strongly to a
particular CPU which requires more locking).
- keep a local difference of increments and decrements, which allows us to sum
the total difference and hence find the refcount when summing all CPUs. Then,
keep a single integer "long" refcount for slow and long lasting references,
and only take the global sum of local counters when the long refcount is 0.
This last scheme is what I implemented here. Attached mounts and process root
and working directory references are "long" references, and everything else is
a short reference.
This allows scalable vfsmount references during path walking over mounted
subtrees and unattached (lazy umounted) mounts with processes still running
in them.
This results in one fewer atomic op in the fastpath: mntget is now just a
per-CPU inc, rather than an atomic inc; and mntput just requires a spinlock
and non-atomic decrement in the common case. However code is otherwise bigger
and heavier, so single threaded performance is basically a wash.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Commit d0adde574b added MNT_STRICTATIME
but it isn't actually used (MS_STRICTATIME clears MNT_RELATIME and
MNT_NOATIME rather than setting any mount flag).
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This patch adds the list and mask fields needed to support vfsmount marks.
These are the same fields fsnotify needs on an inode. They are not used,
just declared and we note where the cleanup hook should be (the function is
not yet defined)
Signed-off-by: Andreas Gruenbacher <agruen@suse.de>
Signed-off-by: Eric Paris <eparis@redhat.com>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6: (52 commits)
init: Open /dev/console from rootfs
mqueue: fix typo "failues" -> "failures"
mqueue: only set error codes if they are really necessary
mqueue: simplify do_open() error handling
mqueue: apply mathematics distributivity on mq_bytes calculation
mqueue: remove unneeded info->messages initialization
mqueue: fix mq_open() file descriptor leak on user-space processes
fix race in d_splice_alias()
set S_DEAD on unlink() and non-directory rename() victims
vfs: add NOFOLLOW flag to umount(2)
get rid of ->mnt_parent in tomoyo/realpath
hppfs can use existing proc_mnt, no need for do_kern_mount() in there
Mirror MS_KERNMOUNT in ->mnt_flags
get rid of useless vfsmount_lock use in put_mnt_ns()
Take vfsmount_lock to fs/internal.h
get rid of insanity with namespace roots in tomoyo
take check for new events in namespace (guts of mounts_poll()) to namespace.c
Don't mess with generic_permission() under ->d_lock in hpfs
sanitize const/signedness for udf
nilfs: sanitize const/signedness in dealing with ->d_name.name
...
Fix up fairly trivial (famous last words...) conflicts in
drivers/infiniband/core/uverbs_main.c and security/tomoyo/realpath.c
The handling of mount flags in set_mnt_shared() got a little tangled
up during previous cleanups, with the following problems:
* MNT_PNODE_MASK is defined as a literal constant when it should be a
bitwise xor of other MNT_* flags
* set_mnt_shared() clears and then sets MNT_SHARED (part of MNT_PNODE_MASK)
* MNT_PNODE_MASK could use a comment in mount.h
* MNT_PNODE_MASK is a terrible name, change to MNT_SHARED_MASK
This patch fixes these problems.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Add __percpu sparse annotations to fs.
These annotations are to make sparse consider percpu variables to be
in a different address space and warn if accessed without going
through percpu accessors. This patch doesn't affect normal builds.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Alex Elder <aelder@sgi.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
This patch speeds up lmbench lat_mmap test by about another 2% after the
first patch.
Before:
avg = 462.286
std = 5.46106
After:
avg = 453.12
std = 9.58257
(50 runs of each, stddev gives a reasonable confidence)
It does this by introducing mnt_clone_write, which avoids some heavyweight
operations of mnt_want_write if called on a vfsmount which we know already
has a write count; and mnt_want_write_file, which can call mnt_clone_write
if the file is open for write.
After these two patches, mnt_want_write and mnt_drop_write go from 7% on
the profile down to 1.3% (including mnt_clone_write).
[AV: mnt_want_write_file() should take file alone and derive mnt from it;
not only all callers have that form, but that's the only mnt about which
we know that it's already held for write if file is opened for write]
Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This patch speeds up lmbench lat_mmap test by about 8%. lat_mmap is set up
basically to mmap a 64MB file on tmpfs, fault in its pages, then unmap it.
A microbenchmark yes, but it exercises some important paths in the mm.
Before:
avg = 501.9
std = 14.7773
After:
avg = 462.286
std = 5.46106
(50 runs of each, stddev gives a reasonable confidence, but there is quite
a bit of variation there still)
It does this by removing the complex per-cpu locking and counter-cache and
replaces it with a percpu counter in struct vfsmount. This makes the code
much simpler, and avoids spinlocks (although the msync is still pretty
costly, unfortunately). It results in about 900 bytes smaller code too. It
does increase the size of a vfsmount, however.
It should also give a speedup on large systems if CPUs are frequently operating
on different mounts (because the existing scheme has to operate on an atomic in
the struct vfsmount when switching between mounts). But I'm most interested in
the single threaded path performance for the moment.
[AV: minor cleanup]
Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Add support for explicitly requesting full atime updates. This makes it
possible for kernels to default to relatime but still allow userspace to
override it.
Signed-off-by: Matthew Garrett <mjg@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Remove a CVS keyword that wasn't updated for a long time from a comment.
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
- use kstrdup() instead of kmalloc() + memcpy()
- return NULL if allocating ->mnt_devname failed
- mnt_devname should be const
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Remove the "#ifdef __KERNEL__" tests from unexported header files in
linux/include whose entire contents are wrapped in that preprocessor
test.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add a unique ID to each peer group using the IDR infrastructure. The
identifiers are reused after the peer group dissolves.
The IDR structures are protected by holding namepspace_sem for write
while allocating or deallocating IDs.
IDs are allocated when a previously unshared vfsmount becomes the
first member of a peer group. When a new member is added to an
existing group, the ID is copied from one of the old members.
IDs are freed when the last member of a peer group is unshared.
Setting the MNT_SHARED flag on members of a subtree is done as a
separate step, after all the IDs have been allocated. This way an
allocation failure can be cleaned up easilty, without affecting the
propagation state.
Based on design sketch by Al Viro.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Add a unique ID to each vfsmount using the IDR infrastructure. The
identifiers are reused after the vfsmount is freed.
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Originally from: Herbert Poetzl <herbert@13thfloor.at>
This is the core of the read-only bind mount patch set.
Note that this does _not_ add a "ro" option directly to the bind mount
operation. If you require such a mount, you must first do the bind, then
follow it up with a 'mount -o remount,ro' operation:
If you wish to have a r/o bind mount of /foo on bar:
mount --bind /foo /bar
mount -o remount,ro /bar
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This is the real meat of the entire series. It actually
implements the tracking of the number of writers to a mount.
However, it causes scalability problems because there can be
hundreds of cpus doing open()/close() on files on the same mnt at
the same time. Even an atomic_t in the mnt has massive scalaing
problems because the cacheline gets so terribly contended.
This uses a statically-allocated percpu variable. All want/drop
operations are local to a cpu as long that cpu operates on the same
mount, and there are no writer count imbalances. Writer count
imbalances happen when a write is taken on one cpu, and released
on another, like when an open/close pair is performed on two
Upon a remount,ro request, all of the data from the percpu
variables is collected (expensive, but very rare) and we determine
if there are any outstanding writers to the mount.
I've written a little benchmark to sit in a loop for a couple of
seconds in several cpus in parallel doing open/write/close loops.
http://sr71.net/~dave/linux/openbench.c
The code in here is a a worst-possible case for this patch. It
does opens on a _pair_ of files in two different mounts in parallel.
This should cause my code to lose its "operate on the same mount"
optimization completely. This worst-case scenario causes a 3%
degredation in the benchmark.
I could probably get rid of even this 3%, but it would be more
complex than what I have here, and I think this is getting into
acceptable territory. In practice, I expect writing more than 3
bytes to a file, as well as disk I/O to mask any effects that this
has.
(To get rid of that 3%, we could have an #defined number of mounts
in the percpu variable. So, instead of a CPU getting operate only
on percpu data when it accesses only one mount, it could stay on
percpu data when it only accesses N or fewer mounts.)
[AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This patch adds two function mnt_want_write() and mnt_drop_write(). These are
used like a lock pair around and fs operations that might cause a write to the
filesystem.
Before these can become useful, we must first cover each place in the VFS
where writes are performed with a want/drop pair. When that is complete, we
can actually introduce code that will safely check the counts before allowing
r/w<->r/o transitions to occur.
Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
... and take it out of ->umount_begin() instances. Call with all locks
already taken (by do_umount()) and leave calling release_mounts() to
caller (it will do release_mounts() anyway, so we can just put into
the same list).
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
make propagate_mount_busy() exclude references from the vfsmounts
that had been isolated by umount_tree() and are just waiting for
release_mounts() to dispose of their ->mnt_parent/->mnt_mountpoint.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Fix the misspellings of "propogate", "writting" and (oh, the shame
:-) "kenrel" in the source tree.
Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
I noticed cache misses in touch_atime() that can be avoided if we keep
mnt_count & mnt_expiry_mark in a different cache line than mnt_flags
(mostly read)
mnt_count & mnt_expiry_mark are modified each time a file is opened/closed
in a file system.
touch_atime() is called each time a file is read, and generally needs to
read mnt_flags.
Other fields of struct vfsmount are mostly read so I chose to move
mnt_count & mnt_expiry_mark at the end of struct vfsmount. And adding a
comment so that nobody tries to re-arrange fields to fill the holes :)
On 64bits platforms, the new offsetof(mnt_count) is 0xC0
On 32bits platforms, it is 0x60, so I didnot add a
____cacheline_aligned_in_smp because it would have a too big impact on the
size of this object (in particular if CONFIG_X86_L1_CACHE_SHIFT=7)
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add "relatime" (relative atime) support. Relative atime only updates the
atime if the previous atime is older than the mtime or ctime. Like
noatime, but useful for applications like mutt that need to know when a
file has been read since it was last modified.
A corresponding patch against mount(8) is available at
http://userweb.kernel.org/~akpm/mount-relative-atime.txt
Signed-off-by: Valerie Henson <val_henson@linux.intel.com>
Cc: Mark Fasheh <mark.fasheh@oracle.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Karel Zak <kzak@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Rename 'struct namespace' to 'struct mnt_namespace' to avoid confusion with
other namespaces being developped for the containers : pid, uts, ipc, etc.
'namespace' variables and attributes are also renamed to 'mnt_ns'
Signed-off-by: Kirill Korotaev <dev@sw.ru>
Signed-off-by: Cedric Le Goater <clg@fr.ibm.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Herbert Poetzl <herbert@13thfloor.at>
Cc: Sukadev Bhattiprolu <sukadev@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Al Viro
Daniel Rosenberg
Eric W. Biederman
Arun Sharma
David Howells
Nick Piggin
Miklos Szeredi
Andreas Gruenbacher
Linus Torvalds
Valerie Aurora
Tejun Heo
npiggin@suse.de
Matthew Garrett
Adrian Bunk
Li Zefan
Robert P. J. Day
Dave Hansen
Robert P. J. Day
Eric Dumazet
Valerie Henson
Kirill Korotaev