2021-04-08 15:41:25

by Jeff Layton

[permalink] [raw]
Subject: Re: [PATCH v6 00/30] Network fs helper library & fscache kiocb API

On Thu, 2021-04-08 at 15:03 +0100, David Howells wrote:
> Here's a set of patches to do two things:
>
> (1) Add a helper library to handle the new VM readahead interface. This
> is intended to be used unconditionally by the filesystem (whether or
> not caching is enabled) and provides a common framework for doing
> caching, transparent huge pages and, in the future, possibly fscrypt
> and read bandwidth maximisation. It also allows the netfs and the
> cache to align, expand and slice up a read request from the VM in
> various ways; the netfs need only provide a function to read a stretch
> of data to the pagecache and the helper takes care of the rest.
>
> (2) Add an alternative fscache/cachfiles I/O API that uses the kiocb
> facility to do async DIO to transfer data to/from the netfs's pages,
> rather than using readpage with wait queue snooping on one side and
> vfs_write() on the other. It also uses less memory, since it doesn't
> do buffered I/O on the backing file.
>
> Note that this uses SEEK_HOLE/SEEK_DATA to locate the data available
> to be read from the cache. Whilst this is an improvement from the
> bmap interface, it still has a problem with regard to a modern
> extent-based filesystem inserting or removing bridging blocks of
> zeros. Fixing that requires a much greater overhaul.
>
> This is a step towards overhauling the fscache API. The change is opt-in
> on the part of the network filesystem. A netfs should not try to mix the
> old and the new API because of conflicting ways of handling pages and the
> PG_fscache page flag and because it would be mixing DIO with buffered I/O.
> Further, the helper library can't be used with the old API.
>
> This does not change any of the fscache cookie handling APIs or the way
> invalidation is done.
>
> In the near term, I intend to deprecate and remove the old I/O API
> (fscache_allocate_page{,s}(), fscache_read_or_alloc_page{,s}(),
> fscache_write_page() and fscache_uncache_page()) and eventually replace
> most of fscache/cachefiles with something simpler and easier to follow.
>
> The patchset contains the following parts:
>
> (1) Some helper patches, including provision of an ITER_XARRAY iov
> iterator and a function to do readahead expansion.
>
> (2) Patches to add the netfs helper library.
>
> (3) A patch to add the fscache/cachefiles kiocb API.
>
> (4) Patches to add support in AFS for this.
>
> Jeff Layton has patches to add support in Ceph for this.
>
> With this, AFS without a cache passes all expected xfstests; with a cache,
> there's an extra failure, but that's also there before these patches.
> Fixing that probably requires a greater overhaul. Ceph also passes the
> expected tests.
>
> The netfs lib and fscache/cachefiles patches can be found also on:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git/log/?h=netfs-lib
>
> The afs patches can be found on:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git/log/?h=afs-netfs-lib
>
>
> Changes
> =======
>
> ver #6:
> Merged in some fixes and added an additional tracepoint[8], including
> fixing the amalgamation of contiguous subrequests that are to be
> written to the cache.
>
> Added/merged some patches from Matthew Wilcox to make
> readahead_expand() appropriately adjust the trigger for the next
> readahead[9]. Also included is a patch to kerneldocify the
> file_ra_state struct.
>
> Altered netfs_write_begin() to use DEFINE_READAHEAD()[10].
>
> Split the afs patches out into their own branch.
>
> ver #5:
> Fixed some review comments from Matthew Wilcox:
>
> - Put a comment into netfs_readahead() to indicate why there's a loop
> that puts, but doesn't unlock, "unconsumed" pages at the end when
> it could just return said pages to the caller to dispose of[6].
> (This is because where those pages are marked consumed).
>
> - Use the page_file_mapping() and page_index() helper functions
> rather than accessing the page struct directly[6].
>
> - Better names for wrangling functions for PG_private_2 and
> PG_fscache wrangling functions[7]. Came up with
> {set,end,wait_for}_page_private_2() and aliased these for fscache.
>
> Moved the taking of/dropping a page ref for the PG_private_2 flag
> into the set and end functions.
>
> ver #4:
> Fixed some review comments from Christoph Hellwig, including dropping
> the export of rw_verify_area()[3] and some minor stuff[4].
>
> Moved the declaration of readahead_expand() to a better location[5].
>
> Rebased to v5.12-rc2 and added a bunch of references into individual
> commits.
>
> Dropped Ceph support - that will go through the maintainer's tree.
>
> Added interface documentation for the netfs helper library.
>
> ver #3:
> Rolled in the bug fixes.
>
> Adjusted the functions that unlock and wait for PG_fscache according
> to Linus's suggestion[1].
>
> Hold a ref on a page when PG_fscache is set as per Linus's
> suggestion[2].
>
> Dropped NFS support and added Ceph support.
>
> ver #2:
> Fixed some bugs and added NFS support.
>
> Link: https://lore.kernel.org/r/[email protected]om/ [1]
> Link: https://lore.kernel.org/r/[email protected]om/ [2]
> Link: https://lore.kernel.org/r/[email protected]/ [3]
> Link: https://lore.kernel.org/r/[email protected]/ [4]
> Link: https://lore.kernel.org/r/[email protected]/ [5]
> Link: https://lore.kernel.org/r/[email protected]/ [6]
> Link: https://lore.kernel.org/r/[email protected]/ [7]
> Link: https://lore.kernel.org/r/[email protected]rg.uk/ [8]
> Link: https://lore.kernel.org/r/[email protected]/ [9]
> Link: https://lore.kernel.org/r/[email protected]/ [10]
>
> References
> ==========
>
> These patches have been published for review before, firstly as part of a
> larger set:
>
> Link: https://lore.kernel.org/r/[email protected]g.uk/
>
> Link: https://lore.kernel.org/r/[email protected]org.uk/
> Link: https://lore.kernel.org/r/[email protected]org.uk/
> Link: https://lore.kernel.org/r/[email protected]rg.uk/
>
> Link: https://lore.kernel.org/r/[email protected]rg.uk/
>
> Then as a cut-down set:
>
> Link: https://lore.kernel.org/r/[email protected]org.uk/ # v1
>
> Link: https://lore.kernel.org/r/[email protected]rg.uk/ # v2
>
> Link: https://lore.kernel.org/r/[email protected]rg.uk/ # v3
>
> Link: https://lore.kernel.org/r/[email protected]g.uk/ # v4
>
> Link: https://lore.kernel.org/r/[email protected]org.uk/ # v5
>
> Proposals/information about the design has been published here:
>
> Link: https://lore.kernel.org/r/[email protected]/
> Link: https://lore.kernel.org/r/[email protected]/
> Link: https://lore.kernel.org/r/[email protected]/
> Link: https://lore.kernel.org/r/[email protected]/
>
> And requests for information:
>
> Link: https://lore.kernel.org/r/[email protected]/
> Link: https://lore.kernel.org/r/[email protected]/
> Link: https://lore.kernel.org/r/[email protected]/
>
> I've posted partial patches to try and help 9p and cifs along:
>
> Link: https://lore.kernel.org/r/[email protected]/
> Link: https://lore.kernel.org/r/[email protected]/
> Link: https://lore.kernel.org/r/[email protected]/
> Link: https://lore.kernel.org/r/[email protected]/
>
> David
> ---
> David Howells (28):
> iov_iter: Add ITER_XARRAY
> mm: Add set/end/wait functions for PG_private_2
> mm: Implement readahead_control pageset expansion
> netfs: Make a netfs helper module
> netfs: Documentation for helper library
> netfs, mm: Move PG_fscache helper funcs to linux/netfs.h
> netfs, mm: Add set/end/wait_on_page_fscache() aliases
> netfs: Provide readahead and readpage netfs helpers
> netfs: Add tracepoints
> netfs: Gather stats
> netfs: Add write_begin helper
> netfs: Define an interface to talk to a cache
> netfs: Add a tracepoint to log failures that would be otherwise unseen
> fscache, cachefiles: Add alternate API to use kiocb for read/write to cache
> afs: Disable use of the fscache I/O routines
> afs: Pass page into dirty region helpers to provide THP size
> afs: Print the operation debug_id when logging an unexpected data version
> afs: Move key to afs_read struct
> afs: Don't truncate iter during data fetch
> afs: Log remote unmarshalling errors
> afs: Set up the iov_iter before calling afs_extract_data()
> afs: Use ITER_XARRAY for writing
> afs: Wait on PG_fscache before modifying/releasing a page
> afs: Extract writeback extension into its own function
> afs: Prepare for use of THPs
> afs: Use the fs operation ops to handle FetchData completion
> afs: Use new netfs lib read helper API
> afs: Use the netfs_write_begin() helper
>
> Matthew Wilcox (Oracle) (2):
> mm/filemap: Pass the file_ra_state in the ractl
> fs: Document file_ra_state
>
>
> Documentation/filesystems/index.rst | 1 +
> Documentation/filesystems/netfs_library.rst | 526 ++++++++
> fs/Kconfig | 1 +
> fs/Makefile | 1 +
> fs/afs/Kconfig | 1 +
> fs/afs/dir.c | 225 ++--
> fs/afs/file.c | 483 ++------
> fs/afs/fs_operation.c | 4 +-
> fs/afs/fsclient.c | 108 +-
> fs/afs/inode.c | 7 +-
> fs/afs/internal.h | 59 +-
> fs/afs/rxrpc.c | 150 +--
> fs/afs/write.c | 657 +++++-----
> fs/afs/yfsclient.c | 82 +-
> fs/cachefiles/Makefile | 1 +
> fs/cachefiles/interface.c | 5 +-
> fs/cachefiles/internal.h | 9 +
> fs/cachefiles/io.c | 420 +++++++
> fs/ext4/verity.c | 2 +-
> fs/f2fs/file.c | 2 +-
> fs/f2fs/verity.c | 2 +-
> fs/fscache/Kconfig | 1 +
> fs/fscache/Makefile | 1 +
> fs/fscache/internal.h | 4 +
> fs/fscache/io.c | 116 ++
> fs/fscache/page.c | 2 +-
> fs/fscache/stats.c | 1 +
> fs/netfs/Kconfig | 23 +
> fs/netfs/Makefile | 5 +
> fs/netfs/internal.h | 97 ++
> fs/netfs/read_helper.c | 1185 +++++++++++++++++++
> fs/netfs/stats.c | 59 +
> include/linux/fs.h | 24 +-
> include/linux/fscache-cache.h | 4 +
> include/linux/fscache.h | 50 +-
> include/linux/netfs.h | 234 ++++
> include/linux/pagemap.h | 41 +-
> include/net/af_rxrpc.h | 2 +-
> include/trace/events/afs.h | 74 +-
> include/trace/events/netfs.h | 261 ++++
> mm/filemap.c | 63 +-
> mm/internal.h | 7 +-
> mm/readahead.c | 97 +-
> net/rxrpc/recvmsg.c | 9 +-
> 44 files changed, 3998 insertions(+), 1108 deletions(-)
> create mode 100644 Documentation/filesystems/netfs_library.rst
> create mode 100644 fs/cachefiles/io.c
> create mode 100644 fs/fscache/io.c
> create mode 100644 fs/netfs/Kconfig
> create mode 100644 fs/netfs/Makefile
> create mode 100644 fs/netfs/internal.h
> create mode 100644 fs/netfs/read_helper.c
> create mode 100644 fs/netfs/stats.c
> create mode 100644 include/linux/netfs.h
> create mode 100644 include/trace/events/netfs.h
>
>

You can add this to 1-16:

Tested-by: Jeff Layton <[email protected]>

Tested with xfstests with fscache enabled on my local machine.