From: Ritesh Harjani <riteshh@linux.ibm.com>
To: Shiyang Ruan <ruansy.fnst@fujitsu.com>,
linux-kernel@vger.kernel.org, linux-xfs@vger.kernel.org,
linux-nvdimm@lists.01.org, linux-fsdevel@vger.kernel.org
Cc: darrick.wong@oracle.com, dan.j.williams@intel.com,
willy@infradead.org, jack@suse.cz, viro@zeniv.linux.org.uk,
linux-btrfs@vger.kernel.org, ocfs2-devel@oss.oracle.com,
david@fromorbit.com, hch@lst.de, rgoldwyn@suse.de
Subject: Re: [PATCH v3 04/10] fsdax: Introduce dax_iomap_cow_copy()
Date: Tue, 23 Mar 2021 21:38:59 +0530 [thread overview]
Message-ID: <985f720c-0cf0-5ada-4df4-3405d5969b8d@linux.ibm.com> (raw)
In-Reply-To: <20210319015237.993880-5-ruansy.fnst@fujitsu.com>
On 3/19/21 7:22 AM, Shiyang Ruan wrote:
> In the case where the iomap is a write operation and iomap is not equal
> to srcmap after iomap_begin, we consider it is a CoW operation.
>
> The destance extent which iomap indicated is new allocated extent.
> So, it is needed to copy the data from srcmap to new allocated extent.
> In theory, it is better to copy the head and tail ranges which is
> outside of the non-aligned area instead of copying the whole aligned
> range. But in dax page fault, it will always be an aligned range. So,
> we have to copy the whole range in this case.
>
> Signed-off-by: Shiyang Ruan <ruansy.fnst@fujitsu.com>
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> ---
> fs/dax.c | 71 ++++++++++++++++++++++++++++++++++++++++++++++++++++----
> 1 file changed, 66 insertions(+), 5 deletions(-)
>
> diff --git a/fs/dax.c b/fs/dax.c
> index a70e6aa285bb..181aad97136a 100644
> --- a/fs/dax.c
> +++ b/fs/dax.c
> @@ -1037,6 +1037,51 @@ static int dax_iomap_direct_access(struct iomap *iomap, loff_t pos, size_t size,
> return rc;
> }
>
> +/*
> + * Copy the head and tail part of the pages not included in the write but
> + * required for CoW, because pos/pos+length are not page aligned. But in dax
> + * page fault case, the range is page aligned, we need to copy the whole range
> + * of data. Use copy_edge to distinguish these cases.
> + */
Is this version better? Feel free to update/change.
dax_iomap_cow_copy(): This can be called from two places.
Either during DAX write fault, to copy the length size data to daddr.
Or, while doing normal DAX write operation, dax_iomap_actor() might call
this to do the copy of either start or end unaligned address. In this
case the rest of the copy of aligned ranges is taken care by
dax_iomap_actor() itself.
Also, note DAX fault will always result in aligned pos and pos + length.
* @pos: address to do copy from.
* @length: size of copy operation.
* @align_size: aligned w.r.t align_size (either PMD_SIZE or PAGE_SIZE)
* @srcmap: iomap srcmap
* @daddr: destination address to copy to.
> +static int dax_iomap_cow_copy(loff_t pos, loff_t length, size_t align_size,
> + struct iomap *srcmap, void *daddr, bool copy_edge)
do we need bool copy_edge here?
We can detect non-align case directly if head_off != pos or pd_end !=
end no?
> +{
> + loff_t head_off = pos & (align_size - 1);
> + size_t size = ALIGN(head_off + length, align_size);
> + loff_t end = pos + length;
> + loff_t pg_end = round_up(end, align_size);
> + void *saddr = 0;
> + int ret = 0;
> +
> + ret = dax_iomap_direct_access(srcmap, pos, size, &saddr, NULL);
> + if (ret)
> + return ret;
> +
> + if (!copy_edge)
> + return copy_mc_to_kernel(daddr, saddr, length);
> +
> + /* Copy the head part of the range. Note: we pass offset as length. */
> + if (head_off) {
> + if (saddr)
> + ret = copy_mc_to_kernel(daddr, saddr, head_off);
> + else
> + memset(daddr, 0, head_off);
> + }
> + /* Copy the tail part of the range */
> + if (end < pg_end) {
> + loff_t tail_off = head_off + length;
> + loff_t tail_len = pg_end - end;
> +
> + if (saddr)
> + ret = copy_mc_to_kernel(daddr + tail_off,
> + saddr + tail_off, tail_len);
> + else
> + memset(daddr + tail_off, 0, tail_len);
> + }
Can you pls help me understand in which case where saddr is 0 and we
will fall back to memset API ?
I was thinking shouldn't such restrictions be coded inside
copy_mc_to_kernel() function in general?
-ritesh
next prev parent reply other threads:[~2021-03-23 16:11 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-19 1:52 [PATCH v3 00/10] fsdax,xfs: Add reflink&dedupe support for fsdax Shiyang Ruan
2021-03-19 1:52 ` [PATCH v3 01/10] fsdax: Factor helpers to simplify dax fault code Shiyang Ruan
2021-03-23 15:33 ` Ritesh Harjani
2021-03-19 1:52 ` [PATCH v3 02/10] fsdax: Factor helper: dax_fault_actor() Shiyang Ruan
2021-03-23 15:48 ` Ritesh Harjani
2021-03-31 3:57 ` ruansy.fnst
2021-04-02 7:47 ` Christoph Hellwig
2021-03-19 1:52 ` [PATCH v3 03/10] fsdax: Output address in dax_iomap_pfn() and rename it Shiyang Ruan
2021-03-23 15:54 ` Ritesh Harjani
2021-03-19 1:52 ` [PATCH v3 04/10] fsdax: Introduce dax_iomap_cow_copy() Shiyang Ruan
2021-03-23 16:08 ` Ritesh Harjani [this message]
2021-03-19 1:52 ` [PATCH v3 05/10] fsdax: Replace mmap entry in case of CoW Shiyang Ruan
2021-04-01 6:39 ` Ritesh Harjani
2021-04-01 7:03 ` ruansy.fnst
2021-03-19 1:52 ` [PATCH v3 06/10] fsdax: Add dax_iomap_cow_copy() for dax_iomap_zero Shiyang Ruan
2021-04-01 6:45 ` Ritesh Harjani
2021-04-01 7:00 ` ruansy.fnst
2021-03-19 1:52 ` [PATCH v3 07/10] iomap: Introduce iomap_apply2() for operations on two files Shiyang Ruan
2021-04-01 7:12 ` Ritesh Harjani
2021-03-19 1:52 ` [PATCH v3 08/10] fsdax: Dedup file range to use a compare function Shiyang Ruan
2021-04-01 11:11 ` Ritesh Harjani
2021-04-08 3:21 ` ruansy.fnst
2021-03-19 1:52 ` [PATCH v3 09/10] fs/xfs: Handle CoW for fsdax write() path Shiyang Ruan
2021-03-19 1:52 ` [PATCH v3 10/10] fs/xfs: Add dedupe support for fsdax Shiyang Ruan
2021-03-23 15:27 ` [PATCH v3 00/10] fsdax,xfs: Add reflink&dedupe " Ritesh Harjani
2021-04-02 7:49 ` Christoph Hellwig
2021-04-02 8:18 ` ruansy.fnst
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=985f720c-0cf0-5ada-4df4-3405d5969b8d@linux.ibm.com \
--to=riteshh@linux.ibm.com \
--cc=dan.j.williams@intel.com \
--cc=darrick.wong@oracle.com \
--cc=david@fromorbit.com \
--cc=hch@lst.de \
--cc=jack@suse.cz \
--cc=linux-btrfs@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-nvdimm@lists.01.org \
--cc=linux-xfs@vger.kernel.org \
--cc=ocfs2-devel@oss.oracle.com \
--cc=rgoldwyn@suse.de \
--cc=ruansy.fnst@fujitsu.com \
--cc=viro@zeniv.linux.org.uk \
--cc=willy@infradead.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).