From: "Darrick J. Wong" <djwong@kernel.org>
To: Christoph Hellwig <hch@lst.de>
Cc: Dan Williams <dan.j.williams@intel.com>,
Mike Snitzer <snitzer@redhat.com>,
Ira Weiny <ira.weiny@intel.com>,
dm-devel@redhat.com, linux-xfs@vger.kernel.org,
nvdimm@lists.linux.dev, linux-s390@vger.kernel.org,
linux-fsdevel@vger.kernel.org, linux-erofs@lists.ozlabs.org,
linux-ext4@vger.kernel.org,
virtualization@lists.linux-foundation.org
Subject: Re: [PATCH 26/29] fsdax: shift partition offset handling into the file systems
Date: Tue, 23 Nov 2021 15:11:46 -0800 [thread overview]
Message-ID: <20211123231146.GT266024@magnolia> (raw)
In-Reply-To: <20211109083309.584081-27-hch@lst.de>
On Tue, Nov 09, 2021 at 09:33:06AM +0100, Christoph Hellwig wrote:
> Remove the last user of ->bdev in dax.c by requiring the file system to
> pass in an address that already includes the DAX offset. As part of the
> only set ->bdev or ->daxdev when actually required in the ->iomap_begin
As part of the ... ?
"...impending disentanglement of block_device and dax_device"?
Which I assume is why we make filesystems know about partition offsets
now?
> methods.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
> fs/dax.c | 6 +-----
> fs/erofs/data.c | 11 ++++++++--
> fs/erofs/internal.h | 1 +
> fs/ext2/inode.c | 8 +++++--
> fs/ext4/inode.c | 16 +++++++++-----
> fs/xfs/libxfs/xfs_bmap.c | 4 ++--
> fs/xfs/xfs_aops.c | 2 +-
> fs/xfs/xfs_iomap.c | 45 +++++++++++++++++++++++++---------------
> fs/xfs/xfs_iomap.h | 5 +++--
> fs/xfs/xfs_pnfs.c | 2 +-
> 10 files changed, 63 insertions(+), 37 deletions(-)
>
> diff --git a/fs/dax.c b/fs/dax.c
> index 0bd6cdcbacfc4..2c13c681edf09 100644
> --- a/fs/dax.c
> +++ b/fs/dax.c
> @@ -711,11 +711,7 @@ int dax_invalidate_mapping_entry_sync(struct address_space *mapping,
>
> static pgoff_t dax_iomap_pgoff(const struct iomap *iomap, loff_t pos)
> {
> - phys_addr_t paddr = iomap->addr + (pos & PAGE_MASK) - iomap->offset;
> -
> - if (iomap->bdev)
> - paddr += (get_start_sect(iomap->bdev) << SECTOR_SHIFT);
> - return PHYS_PFN(paddr);
> + return PHYS_PFN(iomap->addr + (pos & PAGE_MASK) - iomap->offset);
> }
>
> static int copy_cow_page_dax(struct vm_fault *vmf, const struct iomap_iter *iter)
<skip to the xfs part, the ext* parts look ok and I didn't look at erofs>
> diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
> index 4dccd4d90622d..74198dd82b035 100644
> --- a/fs/xfs/libxfs/xfs_bmap.c
> +++ b/fs/xfs/libxfs/xfs_bmap.c
> @@ -4551,7 +4551,7 @@ xfs_bmapi_convert_delalloc(
> * the extent. Just return the real extent at this offset.
> */
> if (!isnullstartblock(bma.got.br_startblock)) {
> - xfs_bmbt_to_iomap(ip, iomap, &bma.got, flags);
> + xfs_bmbt_to_iomap(ip, iomap, &bma.got, 0, flags);
> *seq = READ_ONCE(ifp->if_seq);
> goto out_trans_cancel;
> }
> @@ -4598,7 +4598,7 @@ xfs_bmapi_convert_delalloc(
> XFS_STATS_INC(mp, xs_xstrat_quick);
>
> ASSERT(!isnullstartblock(bma.got.br_startblock));
> - xfs_bmbt_to_iomap(ip, iomap, &bma.got, flags);
> + xfs_bmbt_to_iomap(ip, iomap, &bma.got, 0, flags);
> *seq = READ_ONCE(ifp->if_seq);
>
> if (whichfork == XFS_COW_FORK)
> diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
> index c8c15c3c31471..6ac3449a68ba0 100644
> --- a/fs/xfs/xfs_aops.c
> +++ b/fs/xfs/xfs_aops.c
> @@ -359,7 +359,7 @@ xfs_map_blocks(
> isnullstartblock(imap.br_startblock))
> goto allocate_blocks;
>
> - xfs_bmbt_to_iomap(ip, &wpc->iomap, &imap, 0);
> + xfs_bmbt_to_iomap(ip, &wpc->iomap, &imap, 0, 0);
> trace_xfs_map_blocks_found(ip, offset, count, whichfork, &imap);
> return 0;
> allocate_blocks:
> diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
> index 704292c6ce0c7..74dbf1fd99d39 100644
> --- a/fs/xfs/xfs_iomap.c
> +++ b/fs/xfs/xfs_iomap.c
> @@ -54,7 +54,8 @@ xfs_bmbt_to_iomap(
> struct xfs_inode *ip,
> struct iomap *iomap,
> struct xfs_bmbt_irec *imap,
> - u16 flags)
> + unsigned int flags,
> + u16 iomap_flags)
The argument names confused me. Do @flags contains IOMAP_$FOO flags,
whereas @iomap_flags contains IOMAP_F_$FOO flags? Can these be changed
to "unsigned int iomap_flags" and "u16 iomap_f_flags" to make the flags
domain more obvious, please?
I'd also take "u16 mapping_flags" for the last parameter.
--D
> {
> struct xfs_mount *mp = ip->i_mount;
> struct xfs_buftarg *target = xfs_inode_buftarg(ip);
> @@ -71,16 +72,22 @@ xfs_bmbt_to_iomap(
> iomap->type = IOMAP_DELALLOC;
> } else {
> iomap->addr = BBTOB(xfs_fsb_to_db(ip, imap->br_startblock));
> + if (flags & IOMAP_DAX)
> + iomap->addr += target->bt_dax_part_off;
> +
> if (imap->br_state == XFS_EXT_UNWRITTEN)
> iomap->type = IOMAP_UNWRITTEN;
> else
> iomap->type = IOMAP_MAPPED;
> +
> }
> iomap->offset = XFS_FSB_TO_B(mp, imap->br_startoff);
> iomap->length = XFS_FSB_TO_B(mp, imap->br_blockcount);
> - iomap->bdev = target->bt_bdev;
> - iomap->dax_dev = target->bt_daxdev;
> - iomap->flags = flags;
> + if (flags & IOMAP_DAX)
> + iomap->dax_dev = target->bt_daxdev;
> + else
> + iomap->bdev = target->bt_bdev;
> + iomap->flags = iomap_flags;
>
> if (xfs_ipincount(ip) &&
> (ip->i_itemp->ili_fsync_fields & ~XFS_ILOG_TIMESTAMP))
> @@ -801,7 +808,7 @@ xfs_direct_write_iomap_begin(
>
> xfs_iunlock(ip, lockmode);
> trace_xfs_iomap_found(ip, offset, length, XFS_DATA_FORK, &imap);
> - return xfs_bmbt_to_iomap(ip, iomap, &imap, iomap_flags);
> + return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, iomap_flags);
>
> allocate_blocks:
> error = -EAGAIN;
> @@ -832,18 +839,19 @@ xfs_direct_write_iomap_begin(
> return error;
>
> trace_xfs_iomap_alloc(ip, offset, length, XFS_DATA_FORK, &imap);
> - return xfs_bmbt_to_iomap(ip, iomap, &imap, iomap_flags | IOMAP_F_NEW);
> + return xfs_bmbt_to_iomap(ip, iomap, &imap, flags,
> + iomap_flags | IOMAP_F_NEW);
>
> out_found_cow:
> xfs_iunlock(ip, lockmode);
> length = XFS_FSB_TO_B(mp, cmap.br_startoff + cmap.br_blockcount);
> trace_xfs_iomap_found(ip, offset, length - offset, XFS_COW_FORK, &cmap);
> if (imap.br_startblock != HOLESTARTBLOCK) {
> - error = xfs_bmbt_to_iomap(ip, srcmap, &imap, 0);
> + error = xfs_bmbt_to_iomap(ip, srcmap, &imap, flags, 0);
> if (error)
> return error;
> }
> - return xfs_bmbt_to_iomap(ip, iomap, &cmap, IOMAP_F_SHARED);
> + return xfs_bmbt_to_iomap(ip, iomap, &cmap, flags, IOMAP_F_SHARED);
>
> out_unlock:
> if (lockmode)
> @@ -1053,23 +1061,24 @@ xfs_buffered_write_iomap_begin(
> */
> xfs_iunlock(ip, XFS_ILOCK_EXCL);
> trace_xfs_iomap_alloc(ip, offset, count, allocfork, &imap);
> - return xfs_bmbt_to_iomap(ip, iomap, &imap, IOMAP_F_NEW);
> + return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, IOMAP_F_NEW);
>
> found_imap:
> xfs_iunlock(ip, XFS_ILOCK_EXCL);
> - return xfs_bmbt_to_iomap(ip, iomap, &imap, 0);
> + return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, 0);
>
> found_cow:
> xfs_iunlock(ip, XFS_ILOCK_EXCL);
> if (imap.br_startoff <= offset_fsb) {
> - error = xfs_bmbt_to_iomap(ip, srcmap, &imap, 0);
> + error = xfs_bmbt_to_iomap(ip, srcmap, &imap, flags, 0);
> if (error)
> return error;
> - return xfs_bmbt_to_iomap(ip, iomap, &cmap, IOMAP_F_SHARED);
> + return xfs_bmbt_to_iomap(ip, iomap, &cmap, flags,
> + IOMAP_F_SHARED);
> }
>
> xfs_trim_extent(&cmap, offset_fsb, imap.br_startoff - offset_fsb);
> - return xfs_bmbt_to_iomap(ip, iomap, &cmap, 0);
> + return xfs_bmbt_to_iomap(ip, iomap, &cmap, flags, 0);
>
> out_unlock:
> xfs_iunlock(ip, XFS_ILOCK_EXCL);
> @@ -1178,7 +1187,8 @@ xfs_read_iomap_begin(
> if (error)
> return error;
> trace_xfs_iomap_found(ip, offset, length, XFS_DATA_FORK, &imap);
> - return xfs_bmbt_to_iomap(ip, iomap, &imap, shared ? IOMAP_F_SHARED : 0);
> + return xfs_bmbt_to_iomap(ip, iomap, &imap, flags,
> + shared ? IOMAP_F_SHARED : 0);
> }
>
> const struct iomap_ops xfs_read_iomap_ops = {
> @@ -1237,7 +1247,8 @@ xfs_seek_iomap_begin(
> if (data_fsb < cow_fsb + cmap.br_blockcount)
> end_fsb = min(end_fsb, data_fsb);
> xfs_trim_extent(&cmap, offset_fsb, end_fsb);
> - error = xfs_bmbt_to_iomap(ip, iomap, &cmap, IOMAP_F_SHARED);
> + error = xfs_bmbt_to_iomap(ip, iomap, &cmap, flags,
> + IOMAP_F_SHARED);
> /*
> * This is a COW extent, so we must probe the page cache
> * because there could be dirty page cache being backed
> @@ -1259,7 +1270,7 @@ xfs_seek_iomap_begin(
> imap.br_state = XFS_EXT_NORM;
> done:
> xfs_trim_extent(&imap, offset_fsb, end_fsb);
> - error = xfs_bmbt_to_iomap(ip, iomap, &imap, 0);
> + error = xfs_bmbt_to_iomap(ip, iomap, &imap, flags, 0);
> out_unlock:
> xfs_iunlock(ip, lockmode);
> return error;
> @@ -1306,7 +1317,7 @@ xfs_xattr_iomap_begin(
> if (error)
> return error;
> ASSERT(nimaps);
> - return xfs_bmbt_to_iomap(ip, iomap, &imap, 0);
> + return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, 0);
> }
>
> const struct iomap_ops xfs_xattr_iomap_ops = {
> diff --git a/fs/xfs/xfs_iomap.h b/fs/xfs/xfs_iomap.h
> index 5648262a71736..fe7a625361d95 100644
> --- a/fs/xfs/xfs_iomap.h
> +++ b/fs/xfs/xfs_iomap.h
> @@ -18,8 +18,9 @@ int xfs_iomap_write_unwritten(struct xfs_inode *, xfs_off_t, xfs_off_t, bool);
> xfs_fileoff_t xfs_iomap_eof_align_last_fsb(struct xfs_inode *ip,
> xfs_fileoff_t end_fsb);
>
> -int xfs_bmbt_to_iomap(struct xfs_inode *, struct iomap *,
> - struct xfs_bmbt_irec *, u16);
> +int xfs_bmbt_to_iomap(struct xfs_inode *ip, struct iomap *iomap,
> + struct xfs_bmbt_irec *imap, unsigned int flags,
> + u16 iomap_flags);
>
> int xfs_zero_range(struct xfs_inode *ip, loff_t pos, loff_t len,
> bool *did_zero);
> diff --git a/fs/xfs/xfs_pnfs.c b/fs/xfs/xfs_pnfs.c
> index e188e1cf97cc5..d6334abbc0b3e 100644
> --- a/fs/xfs/xfs_pnfs.c
> +++ b/fs/xfs/xfs_pnfs.c
> @@ -173,7 +173,7 @@ xfs_fs_map_blocks(
> }
> xfs_iunlock(ip, XFS_IOLOCK_EXCL);
>
> - error = xfs_bmbt_to_iomap(ip, iomap, &imap, 0);
> + error = xfs_bmbt_to_iomap(ip, iomap, &imap, 0, 0);
> *device_generation = mp->m_generation;
> return error;
> out_unlock:
> --
> 2.30.2
>
next prev parent reply other threads:[~2021-11-23 23:11 UTC|newest]
Thread overview: 102+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-09 8:32 decouple DAX from block devices Christoph Hellwig
2021-11-09 8:32 ` [PATCH 01/29] nvdimm/pmem: move dax_attribute_group from dax to pmem Christoph Hellwig
2021-11-17 17:44 ` Dan Williams
2021-11-19 6:56 ` Christoph Hellwig
2021-11-19 17:21 ` Dan Williams
2021-11-09 8:32 ` [PATCH 02/29] dm: make the DAX support dependend on CONFIG_FS_DAX Christoph Hellwig
2021-11-17 17:23 ` Dan Williams
2021-11-19 6:54 ` Christoph Hellwig
2021-11-23 2:54 ` Dan Williams
2021-11-23 5:56 ` Christoph Hellwig
2021-11-09 8:32 ` [PATCH 03/29] dax: remove CONFIG_DAX_DRIVER Christoph Hellwig
2021-11-17 17:43 ` Dan Williams
2021-11-23 2:54 ` Dan Williams
2021-11-09 8:32 ` [PATCH 04/29] dax: simplify the dax_device <-> gendisk association Christoph Hellwig
2021-11-23 3:33 ` Dan Williams
2021-11-23 5:57 ` Christoph Hellwig
2021-11-23 19:33 ` Dan Williams
2021-11-09 8:32 ` [PATCH 05/29] dax: remove the pgmap sanity checks in generic_fsdax_supported Christoph Hellwig
2021-11-23 3:35 ` Dan Williams
2021-11-09 8:32 ` [PATCH 06/29] dax: move the partition alignment check into fs_dax_get_by_bdev Christoph Hellwig
2021-11-23 3:41 ` Dan Williams
2021-11-23 22:25 ` Darrick J. Wong
2021-11-24 6:36 ` Christoph Hellwig
2021-11-09 8:32 ` [PATCH 07/29] xfs: factor out a xfs_setup_dax_always helper Christoph Hellwig
2021-11-17 17:31 ` Darrick J. Wong
2021-11-23 3:51 ` Dan Williams
2021-11-09 8:32 ` [PATCH 08/29] dax: remove dax_capable Christoph Hellwig
2021-11-23 3:58 ` Dan Williams
2021-11-23 22:31 ` Darrick J. Wong
2021-11-24 6:37 ` Christoph Hellwig
2021-11-09 8:32 ` [PATCH 09/29] dm-linear: add a linear_dax_pgoff helper Christoph Hellwig
2021-11-23 3:59 ` Dan Williams
2021-11-09 8:32 ` [PATCH 10/29] dm-log-writes: add a log_writes_dax_pgoff helper Christoph Hellwig
2021-11-23 4:02 ` Dan Williams
2021-11-09 8:32 ` [PATCH 11/29] dm-stripe: add a stripe_dax_pgoff helper Christoph Hellwig
2021-11-23 4:16 ` Dan Williams
2021-11-09 8:32 ` [PATCH 12/29] fsdax: remove a pointless __force cast in copy_cow_page_dax Christoph Hellwig
2021-11-23 4:17 ` Dan Williams
2021-11-23 22:33 ` Darrick J. Wong
2021-11-09 8:32 ` [PATCH 13/29] fsdax: use a saner calling convention for copy_cow_page_dax Christoph Hellwig
2021-11-23 19:35 ` Dan Williams
2021-11-23 22:33 ` Darrick J. Wong
2021-11-09 8:32 ` [PATCH 14/29] fsdax: simplify the pgoff calculation Christoph Hellwig
2021-11-23 19:44 ` Dan Williams
2021-11-23 22:36 ` Darrick J. Wong
2021-11-24 6:39 ` Christoph Hellwig
2021-11-09 8:32 ` [PATCH 15/29] xfs: add xfs_zero_range and xfs_truncate_page helpers Christoph Hellwig
2021-11-23 21:15 ` Dan Williams
2021-11-23 22:37 ` Darrick J. Wong
2021-11-09 8:32 ` [PATCH 16/29] fsdax: simplify the offset check in dax_iomap_zero Christoph Hellwig
2021-11-23 21:16 ` Dan Williams
2021-11-23 22:40 ` Darrick J. Wong
2021-11-09 8:32 ` [PATCH 17/29] fsdax: factor out a dax_memzero helper Christoph Hellwig
2021-11-23 21:22 ` Dan Williams
2021-11-23 22:44 ` Darrick J. Wong
2021-11-24 6:47 ` Christoph Hellwig
2021-11-09 8:32 ` [PATCH 18/29] fsdax: decouple zeroing from the iomap buffered I/O code Christoph Hellwig
2021-11-23 21:46 ` Dan Williams
2021-11-24 6:50 ` Christoph Hellwig
2021-11-23 22:53 ` Darrick J. Wong
2021-11-24 6:52 ` Christoph Hellwig
2021-11-09 8:32 ` [PATCH 19/29] ext2: cleanup the dax handling in ext2_fill_super Christoph Hellwig
2021-11-23 21:48 ` Dan Williams
2021-11-23 22:54 ` Darrick J. Wong
2021-11-09 8:33 ` [PATCH 20/29] ext4: cleanup the dax handling in ext4_fill_super Christoph Hellwig
2021-11-23 21:49 ` Dan Williams
2021-11-23 22:54 ` Darrick J. Wong
2021-11-24 6:53 ` Christoph Hellwig
2021-11-09 8:33 ` [PATCH 21/29] xfs: move dax device handling into xfs_{alloc,free}_buftarg Christoph Hellwig
2021-11-23 22:55 ` Darrick J. Wong
2021-11-24 2:40 ` Dan Williams
2021-11-24 3:05 ` Darrick J. Wong
2021-11-24 6:55 ` Christoph Hellwig
2021-11-09 8:33 ` [PATCH 22/29] iomap: add a IOMAP_DAX flag Christoph Hellwig
2021-11-23 23:00 ` Darrick J. Wong
2021-11-24 2:47 ` Dan Williams
2021-11-24 7:10 ` Christoph Hellwig
2021-11-09 8:33 ` [PATCH 23/29] xfs: use IOMAP_DAX to check for DAX mappings Christoph Hellwig
2021-11-23 23:01 ` Darrick J. Wong
2021-11-24 7:14 ` Christoph Hellwig
2021-11-24 2:49 ` Dan Williams
2021-11-09 8:33 ` [PATCH 24/29] xfs: use xfs_direct_write_iomap_ops for DAX zeroing Christoph Hellwig
2021-11-23 23:02 ` Darrick J. Wong
2021-11-24 2:52 ` Dan Williams
2021-11-09 8:33 ` [PATCH 25/29] dax: return the partition offset from fs_dax_get_by_bdev Christoph Hellwig
2021-11-24 2:56 ` Dan Williams
2021-11-24 6:59 ` Christoph Hellwig
2021-11-09 8:33 ` [PATCH 26/29] fsdax: shift partition offset handling into the file systems Christoph Hellwig
2021-11-23 23:11 ` Darrick J. Wong [this message]
2021-11-24 3:21 ` Dan Williams
2021-11-24 3:44 ` Gao Xiang
2021-11-09 8:33 ` [PATCH 27/29] dax: fix up some of the block device related ifdefs Christoph Hellwig
2021-11-23 23:13 ` Darrick J. Wong
2021-11-24 3:47 ` Dan Williams
2021-11-09 8:33 ` [PATCH 28/29] iomap: build the block based code conditionally Christoph Hellwig
2021-11-23 23:13 ` Darrick J. Wong
2021-11-24 3:51 ` Dan Williams
2021-11-09 8:33 ` [PATCH 29/29] fsdax: don't require CONFIG_BLOCK Christoph Hellwig
2021-11-23 23:13 ` Darrick J. Wong
2021-11-24 3:52 ` Dan Williams
2021-11-29 10:21 decouple DAX from block devices v2 Christoph Hellwig
2021-11-29 10:22 ` [PATCH 26/29] fsdax: shift partition offset handling into the file systems Christoph Hellwig
2021-11-30 19:06 ` Darrick J. Wong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20211123231146.GT266024@magnolia \
--to=djwong@kernel.org \
--cc=dan.j.williams@intel.com \
--cc=dm-devel@redhat.com \
--cc=hch@lst.de \
--cc=ira.weiny@intel.com \
--cc=linux-erofs@lists.ozlabs.org \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-s390@vger.kernel.org \
--cc=linux-xfs@vger.kernel.org \
--cc=nvdimm@lists.linux.dev \
--cc=snitzer@redhat.com \
--cc=virtualization@lists.linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).