Re: [PATCH v7 00/14] KVM: mm: fd-based approach for supporting KVM guest private memory
From: Ackerley Tng
Date: Fri Apr 14 2023 - 18:39:03 EST
Sean Christopherson <seanjc@xxxxxxxxxx> writes:
On Thu, Apr 13, 2023, Christian Brauner wrote:
On Thu, Aug 18, 2022 at 04:24:21PM +0300, Kirill A . Shutemov wrote:
> On Wed, Aug 17, 2022 at 10:40:12PM -0700, Hugh Dickins wrote:
> > Here's what I would prefer, and imagine much easier for you to
maintain;
> > but I'm no system designer, and may be misunderstanding throughout.
> >
> > QEMU gets fd from opening /dev/kvm_something, uses ioctls (or perhaps
> > the fallocate syscall interface itself) to allocate and free the
memory,
> > ioctl for initializing some of it too. KVM in control of whether
that
> > fd can be read or written or mmap'ed or whatever, no need to prevent
it
> > in shmem.c, no need for flags, seals, notifications to and fro
because
> > KVM is already in control and knows the history. If shmem actually
has
> > value, call into it underneath - somewhat like SysV SHM, and
/dev/zero
> > mmap, and i915/gem make use of it underneath. If shmem has nothing
to
> > add, just allocate and free kernel memory directly, recorded in your
> > own xarray.
>
> I guess shim layer on top of shmem *can* work. I don't see immediately
why
> it would not. But I'm not sure it is right direction. We risk creating
yet
> another parallel VM with own rules/locking/accounting that opaque to
> core-mm.
Sorry for necrobumping this thread but I've been reviewing the
No worries, I'm just stoked someone who actually knows what they're doing
is
chiming in :-)
+1, thanks Christian!
memfd_restricted() extension that Ackerley is currently working on. I
was pointed to this thread as this is what the extension is building
on but I'll reply to both threads here.
From a glance at v10, memfd_restricted() is currently implemented as an
in-kernel stacking filesystem. A call to memfd_restricted() creates a
new restricted memfd file and a new unlinked tmpfs file and stashes the
tmpfs file into the memfd file's private data member. It then uses the
tmpfs file's f_ops and i_ops to perform the relevant file and inode
operations. So it has the same callstack as a general stacking
filesystem like overlayfs in some cases:
memfd_restricted->getattr()
-> tmpfs->getattr()
...
Since you're effectively acting like a stacking filesystem you should
really use the device number of your memfd restricted filesystem. IOW,
sm like:
stat->dev = memfd_restricted_dentry->d_sb->s_dev;
But then you run into trouble if you want to go forward with Ackerley's
extension that allows to explicitly pass in tmpfs fds to
memfd_restricted(). Afaict, two tmpfs instances might allocate the same
inode number. So now the inode and device number pair isn't unique
anymore.
So you might best be served by allocating and reporting your own inode
numbers as well.
But if you want to preserve the inode number and device number of the
relevant tmpfs instance but still report memfd restricted as your
filesystem type
Unless I missed something along the way, reporting memfd_restricted as a
distinct
filesystem is very much a non-goal. AFAIK it's purely a side effect of
the
proposed implementation.
then I think it's reasonable to ask whether a stacking implementation
really
makes sense here.
If you extend memfd_restricted() or even consider extending it in the
future to take tmpfs file descriptors as arguments to identify the tmpfs
instance in which to allocate the underlying tmpfs file for the new
restricted memfd file you should really consider a tmpfs based
implementation.
Because at that point it just feels like a pointless wrapper to get
custom f_ops and i_ops. Plus it's wasteful because you allocate dentries
and inodes that you don't really care about at all.
Just off the top of my hat you might be better served:
* by a new ioctl() on tmpfs instances that
yield regular tmpfs file descriptors with restricted f_ops and i_ops.
That's not that different from btrfs subvolumes which effectively are
directories but are created through an ioctl().
I think this is more or less what we want to do, except via a dedicated
syscall
instead of an ioctl() so that the primary interface isn't strictly tied
to tmpfs,
e.g. so that it can be extended to other backing types in the future.
* by a mount option to tmpfs that makes it act
in this restricted manner then you don't need an ioctl() and can get
away with regular open calls. Such a tmpfs instance would only create
regular, restricted memfds.
I'd prefer to not go this route, becuase IIUC, it would require
relatively invasive
changes to shmem code, and IIUC would require similar changes to other
support
backings in the future, e.g. hugetlbfs? And as above, I don't think any
of the
potential use cases need restrictedmem to be a uniquely identifiable
mount.
FWIW, I'm starting to look at extending restrictedmem to hugetlbfs and
the separation that the current implementation has is very helpful. Also
helps that hugetlbfs and tmpfs are structured similarly, I guess.
One of the goals (hopefully not a pipe dream) is to design restrictmem in
such a
way that extending it to support other backing types isn't terribly
difficult.
In case it's not obvious, most of us working on this stuff aren't
filesystems
experts, and many of us aren't mm experts either. The more we (KVM folks
for the
most part) can leverage existing code to do the heavy lifting, the better.
After giving myself a bit of a crash course in file systems, would
something like
the below have any chance of (a) working, (b) getting merged, and (c)
being
maintainable?
The idea is similar to a stacking filesystem, but instead of stacking,
restrictedmem
hijacks a f_ops and a_ops to create a lightweight shim around tmpfs.
There are
undoubtedly issues and edge cases, I'm just looking for a quick "yes,
this might
be doable" or a "no, that's absolutely bonkers, don't try it".
Not an FS expert by any means, but I did think of approaching it this
way as well!
"Hijacking" perhaps gives this approach a bit of a negative
connotation. I thought this is pretty close to subclassing (as in Object
Oriented Programming). When some methods (e.g. fallocate) are called,
restrictedmem does some work, and calls the same method in the
superclass.
The existing restrictedmem code is a more like instantiating an shmem
object and keeping that object as a field within the restrictedmem
object.
Some (maybe small) issues I can think of now:
(1)
One difficulty with this approach is that other functions may make
assumptions about private_data being of a certain type, or functions may
use private_data.
I checked and IIUC neither shmem nor hugetlbfs use the private_data
field in the inode's i_mapping (also file's f_mapping).
But there's fs/buffer.c which uses private_data, although those
functions seem to be used by FSes like ext4 and fat, not memory-backed
FSes.
We can probably fix this if any backing filesystems of restrictedmem,
like tmpfs and future ones use private_data.
Could the solution here be to store private_data of the superclass
instance in restrictedmem, and then override every method in the
superclass that uses private_data to first restore private_data before
making the superclass call? Perhaps we can take private_lock to change
private_data.
(2)
Perhaps there are other slightly hidden cases that might need cleaning up.
For example, one of the patches in this series amends the
shmem_mapping() function from
return mapping->a_ops == &shmem_aops;
to
return mapping->host->i_sb->s_magic == TMPFS_MAGIC;
The former/original is more accurate since it checks a property of the
mapping itself instead of checking a property of the mapping's host's
superblock.
The impact of changing this guard is more obvious if we now override
a_ops but keep the mapping's host's superblock's s_magic.
Specifically for this example, maybe we should handle restrictedmem in
the caller (me_pagecache_clean()) specially, in addition to shmem.
Thanks!
struct restrictedmem {
struct rw_semaphore lock;
struct file *file;
const struct file_operations *backing_f_ops;
const struct address_space_operations *backing_a_ops;
struct xarray bindings;
bool exclusive;
};
static int restrictedmem_release(struct inode *inode, struct file *file)
{
struct restrictedmem *rm = inode->i_mapping->private_data;
xa_destroy(&rm->bindings);
kfree(rm);
WARN_ON_ONCE(rm->backing_f_ops->release);
return 0;
}
static long restrictedmem_punch_hole(struct restrictedmem *rm, int mode,
loff_t offset, loff_t len)
{
struct restrictedmem_notifier *notifier;
unsigned long index;
pgoff_t start, end;
int ret;
if (!PAGE_ALIGNED(offset) || !PAGE_ALIGNED(len))
return -EINVAL;
start = offset >> PAGE_SHIFT;
end = (offset + len) >> PAGE_SHIFT;
/*
* Bindings must be stable across invalidation to ensure the start+end
* are balanced.
*/
down_read(&rm->lock);
xa_for_each_range(&rm->bindings, index, notifier, start, end - 1)
notifier->ops->invalidate_start(notifier, start, end);
ret = rm->backing_f_ops->fallocate(rm->file, mode, offset, len);
xa_for_each_range(&rm->bindings, index, notifier, start, end - 1)
notifier->ops->invalidate_end(notifier, start, end);
up_read(&rm->lock);
return ret;
}
static long restrictedmem_fallocate(struct file *file, int mode,
loff_t offset, loff_t len)
{
struct restrictedmem *rm = file->f_mapping->private_data;
if (mode & FALLOC_FL_PUNCH_HOLE)
return restrictedmem_punch_hole(rm, mode, offset, len);
return rm->backing_f_ops->fallocate(file, mode, offset, len);
}
static int restrictedmem_migrate_folio(struct address_space *mapping,
struct folio *dst, struct folio *src,
enum migrate_mode)
{
WARN_ON_ONCE(1);
return -EINVAL;
}
static int restrictedmem_error_page(struct address_space *mapping,
struct page *page)
{
struct restrictedmem *rm = mapping->private_data;
struct restrictedmem_notifier *notifier;
unsigned long index;
pgoff_t start, end;
start = page->index;
end = start + thp_nr_pages(page);
down_read(&rm->lock);
xa_for_each_range(&rm->bindings, index, notifier, start, end - 1)
notifier->ops->error(notifier, start, end);
up_read(&rm->lock);
return rm->backing_a_ops->error_remove_page(mapping, page);
}
When I was thinking of this I was stuck on handling error_remove_page,
because it was looking up the superblock to iterate over the inodes to
find the right mapping. Glad to see that the solution is simply to use
the given mapping from the arguments!
static const struct file_operations restrictedmem_fops = {
.release = restrictedmem_release,
.fallocate = restrictedmem_fallocate,
};
static const struct address_space_operations restrictedmem_aops = {
.dirty_folio = noop_dirty_folio,
#ifdef CONFIG_MIGRATION
.migrate_folio = restrictedmem_migrate_folio,
#endif
.error_remove_page = restrictedmem_error_page,
};
static int restrictedmem_file_create(struct file *file)
{
struct address_space *mapping = file->f_mapping;
struct restrictedmem *rm;
rm = kzalloc(sizeof(*rm), GFP_KERNEL);
if (!rm)
return -ENOMEM;
rm->backing_f_ops = file->f_op;
rm->backing_a_ops = mapping->a_ops;
rm->file = file;
We don't really need to do this, since rm->file is already the same as
file, we could just pass the file itself when it's needed
init_rwsem(&rm->lock);
xa_init(&rm->bindings);
file->f_flags |= O_LARGEFILE;
file->f_op = &restrictedmem_fops;
mapping->a_ops = &restrictedmem_aops;
I think we probably have to override inode_operations as well, because
otherwise other methods would become available to a restrictedmem file
(like link, unlink, mkdir, tmpfile). Or maybe that's a feature instead
of a bug.
mapping_set_unevictable(mapping);
mapping_set_unmovable(mapping);
mapping_set_gfp_mask(mapping,
mapping_gfp_mask(mapping) & ~__GFP_MOVABLE);
return 0;
}
static int restrictedmem_create(struct vfsmount *mount)
{
struct file *file;
int fd, err;
fd = get_unused_fd_flags(0);
if (fd < 0)
return fd;
file = shmem_file_setup_with_mnt(mount, "memfd:restrictedmem", 0,
VM_NORESERVE);
if (IS_ERR(file)) {
err = PTR_ERR(file);
goto err_fd;
}
if (WARN_ON_ONCE(file->private_data)) {
err = -EEXIST;
goto err_fd;
}
Did you intend this as a check that the backing filesystem isn't using
the private_data field in the mapping?
I think you meant file->f_mapping->private_data.
On this note, we will probably have to fix things whenever any backing
filesystems need the private_data field.
file->f_mode |= FMODE_LSEEK | FMODE_PREAD | FMODE_PWRITE;
file->f_flags |= O_LARGEFILE;
err = restrictedmem_file_create(file);
if (err) {
fput(file);
goto err_fd;
}
fd_install(fd, file);
return fd;
err_fd:
put_unused_fd(fd);
return err;
}
SYSCALL_DEFINE2(memfd_restricted, unsigned int, flags, int, mount_fd)
{
struct vfsmount *mnt;
struct path *path;
struct fd f;
int ret;
if (flags)
return -EINVAL;
f = fdget_raw(mount_fd);
if (!f.file)
return -EBADF;
ret = -EINVAL;
path = &f.file->f_path;
if (path->dentry != path->mnt->mnt_root)
goto out;
/* Disallow bind-mounts that aren't bind-mounts of the whole filesystem.
*/
mnt = path->mnt;
if (mnt->mnt_root != mnt->mnt_sb->s_root)
goto out;
/*
* The filesystem must be mounted no-execute, executing from guest
* private memory in the host is nonsensical and unsafe.
*/
if (!(mnt->mnt_sb->s_iflags & SB_I_NOEXEC))
goto out;
/* Currently only TMPFS is supported as underlying storage. */
if (mnt->mnt_sb->s_magic != TMPFS_MAGIC)
goto out;
ret = mnt_want_write(mnt);
if (ret)
goto out;
ret = restrictedmem_create(mnt);
if (mnt)
mnt_drop_write(mnt);
out:
if (f.file)
fdput(f);
return ret;
}