Re: [PATCH v9 02/10] dax: Introduce holder for dax_device

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Jan 5, 2022 at 2:47 PM Darrick J. Wong <djwong@xxxxxxxxxx> wrote:
>
> On Wed, Jan 05, 2022 at 11:20:12AM -0800, Dan Williams wrote:
> > On Wed, Jan 5, 2022 at 10:56 AM Darrick J. Wong <djwong@xxxxxxxxxx> wrote:
> > >
> > > On Wed, Jan 05, 2022 at 10:23:08AM -0800, Dan Williams wrote:
> > > > On Wed, Jan 5, 2022 at 10:12 AM Darrick J. Wong <djwong@xxxxxxxxxx> wrote:
> > > > >
> > > > > On Sun, Dec 26, 2021 at 10:34:31PM +0800, Shiyang Ruan wrote:
> > > > > > To easily track filesystem from a pmem device, we introduce a holder for
> > > > > > dax_device structure, and also its operation.  This holder is used to
> > > > > > remember who is using this dax_device:
> > > > > >  - When it is the backend of a filesystem, the holder will be the
> > > > > >    instance of this filesystem.
> > > > > >  - When this pmem device is one of the targets in a mapped device, the
> > > > > >    holder will be this mapped device.  In this case, the mapped device
> > > > > >    has its own dax_device and it will follow the first rule.  So that we
> > > > > >    can finally track to the filesystem we needed.
> > > > > >
> > > > > > The holder and holder_ops will be set when filesystem is being mounted,
> > > > > > or an target device is being activated.
> > > > > >
> > > > > > Signed-off-by: Shiyang Ruan <ruansy.fnst@xxxxxxxxxxx>
> > > > > > ---
> > > > > >  drivers/dax/super.c | 62 +++++++++++++++++++++++++++++++++++++++++++++
> > > > > >  include/linux/dax.h | 29 +++++++++++++++++++++
> > > > > >  2 files changed, 91 insertions(+)
> > > > > >
> > > > > > diff --git a/drivers/dax/super.c b/drivers/dax/super.c
> > > > > > index c46f56e33d40..94c51f2ee133 100644
> > > > > > --- a/drivers/dax/super.c
> > > > > > +++ b/drivers/dax/super.c
> > > > > > @@ -20,15 +20,20 @@
> > > > > >   * @inode: core vfs
> > > > > >   * @cdev: optional character interface for "device dax"
> > > > > >   * @private: dax driver private data
> > > > > > + * @holder_data: holder of a dax_device: could be filesystem or mapped device
> > > > > >   * @flags: state and boolean properties
> > > > > > + * @ops: operations for dax_device
> > > > > > + * @holder_ops: operations for the inner holder
> > > > > >   */
> > > > > >  struct dax_device {
> > > > > >       struct inode inode;
> > > > > >       struct cdev cdev;
> > > > > >       void *private;
> > > > > >       struct percpu_rw_semaphore rwsem;
> > > > > > +     void *holder_data;
> > > > > >       unsigned long flags;
> > > > > >       const struct dax_operations *ops;
> > > > > > +     const struct dax_holder_operations *holder_ops;
> > > > > >  };
> > > > > >
> > > > > >  static dev_t dax_devt;
> > > > > > @@ -192,6 +197,29 @@ int dax_zero_page_range(struct dax_device *dax_dev, pgoff_t pgoff,
> > > > > >  }
> > > > > >  EXPORT_SYMBOL_GPL(dax_zero_page_range);
> > > > > >
> > > > > > +int dax_holder_notify_failure(struct dax_device *dax_dev, u64 off,
> > > > > > +                           u64 len, int mf_flags)
> > > > > > +{
> > > > > > +     int rc;
> > > > > > +
> > > > > > +     dax_read_lock(dax_dev);
> > > > > > +     if (!dax_alive(dax_dev)) {
> > > > > > +             rc = -ENXIO;
> > > > > > +             goto out;
> > > > > > +     }
> > > > > > +
> > > > > > +     if (!dax_dev->holder_ops) {
> > > > > > +             rc = -EOPNOTSUPP;
> > > > > > +             goto out;
> > > > > > +     }
> > > > > > +
> > > > > > +     rc = dax_dev->holder_ops->notify_failure(dax_dev, off, len, mf_flags);
> > > > > > +out:
> > > > > > +     dax_read_unlock(dax_dev);
> > > > > > +     return rc;
> > > > > > +}
> > > > > > +EXPORT_SYMBOL_GPL(dax_holder_notify_failure);
> > > > > > +
> > > > > >  #ifdef CONFIG_ARCH_HAS_PMEM_API
> > > > > >  void arch_wb_cache_pmem(void *addr, size_t size);
> > > > > >  void dax_flush(struct dax_device *dax_dev, void *addr, size_t size)
> > > > > > @@ -254,6 +282,10 @@ void kill_dax(struct dax_device *dax_dev)
> > > > > >               return;
> > > > > >       dax_write_lock(dax_dev);
> > > > > >       clear_bit(DAXDEV_ALIVE, &dax_dev->flags);
> > > > > > +
> > > > > > +     /* clear holder data */
> > > > > > +     dax_dev->holder_ops = NULL;
> > > > > > +     dax_dev->holder_data = NULL;
> > > > > >       dax_write_unlock(dax_dev);
> > > > > >  }
> > > > > >  EXPORT_SYMBOL_GPL(kill_dax);
> > > > > > @@ -401,6 +433,36 @@ void put_dax(struct dax_device *dax_dev)
> > > > > >  }
> > > > > >  EXPORT_SYMBOL_GPL(put_dax);
> > > > > >
> > > > > > +void dax_register_holder(struct dax_device *dax_dev, void *holder,
> > > > > > +             const struct dax_holder_operations *ops)
> > > > > > +{
> > > > > > +     if (!dax_alive(dax_dev))
> > > > > > +             return;
> > > > > > +
> > > > > > +     dax_dev->holder_data = holder;
> > > > > > +     dax_dev->holder_ops = ops;
> > > > >
> > > > > Shouldn't this return an error code if the dax device is dead or if
> > > > > someone already registered a holder?  I'm pretty sure XFS should not
> > > > > bind to a dax device if someone else already registered for it...
> > > >
> > > > Agree, yes.
> > > >
> > > > >
> > > > > ...unless you want to use a notifier chain for failure events so that
> > > > > there can be multiple consumers of dax failure events?
> > > >
> > > > No, I would hope not. It should be 1:1 holders to dax-devices. Similar
> > > > ownership semantics like bd_prepare_to_claim().
> > >
> > > Does each partition on a pmem device still have its own dax_device?
> >
> > No, it never did...
> >
> > Just as before, each dax-device is still associated with a gendisk /
> > whole-block_device. The recent change is that instead of needing that
> > partition-block_device plumbed to convert a relative block number to
> > its absolute whole-block_device offset the filesystem now handles that
> > at iomap_begin() time. See:
> >
> >                 if (mapping_flags & IOMAP_DAX)
> >                         iomap->addr += target->bt_dax_part_off;
> >
> > ...in xfs_bmbt_to_iomap() (in -next). I.e. bdev_dax_pgoff() is gone
> > with the lead-in reworks.
>
> OH.  Crap, Dan's right:
>
> XFS (pmem0p1): ddev dax = 0xffff88800304ed00 bdev = 0xffff8880480d6180
> XFS (pmem0p2): ddev dax = 0xffff88800304ed00 bdev = 0xffff8880480d4380
>
> Unless you're planning to remove partition support too, this patch needs
> to be reworked so that multiple filesystems in separate partitions can
> each call dax_register_holder to receive memory_failure notifications
> for their partition.

Oh, crap indeed. I think this gets back to the slow tip-toeing away
from dax + partition support. While FSDAX can continue to support
"legacy/experimental" operation on partitions of DAX capable block
devices, I think failure notification + reflink support is where we
draw the line and say "DAX on partitions was a mistake, it's too late
to undo that mistake, but going forward for new FSDAX features it
requires switching from block-device partitions to pmem-namespaces if
you need sub-division support and new DAX features."

> /methinks this sharing is all a little scary...

Yes, I think we should just fail the holder registration and
DAX+reflink unless the FS being mounted on a whole device. I know Ted
and others had reservations about moving filesystems to be mounted on
dax-devices directly, but hopefully the whole-block_device requirement
is a workable middle ground?



[Index of Archives]     [Linux Ext4 Filesystem]     [Union Filesystem]     [Filesystem Testing]     [Ceph Users]     [Ecryptfs]     [NTFS 3]     [AutoFS]     [Kernel Newbies]     [Share Photos]     [Security]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux Cachefs]     [Reiser Filesystem]     [Linux RAID]     [NTFS 3]     [Samba]     [Device Mapper]     [CEPH Development]

  Powered by Linux