From: Jan Kara <jack@suse.cz>
To: linux-nvdimm@lists.01.org
Cc: Ted Tso <tytso@mit.edu>,
linux-ext4@vger.kernel.org,
Vishal Verma <vishal.l.verma@intel.com>,
Ross Zwisler <ross.zwisler@linux.intel.com>,
Dan Williams <dan.j.williams@intel.com>,
linux-fsdevel@vger.kernel.org, Jan Kara <jack@suse.cz>
Subject: [PATCH 5/7] dax: Remove zeroing from dax_io()
Date: Wed, 11 May 2016 11:58:51 +0200 [thread overview]
Message-ID: <1462960733-29634-6-git-send-email-jack@suse.cz> (raw)
In-Reply-To: <1462960733-29634-1-git-send-email-jack@suse.cz>
All the filesystems are now zeroing blocks themselves for DAX IO to avoid
races between dax_io() and dax_fault(). Remove the zeroing code from
dax_io() and add warning to catch the case when somebody unexpectedly
returns new or unwritten buffer.
Reviewed-by: Ross Zwisler <ross.zwisler@linux.intel.com>
Signed-off-by: Jan Kara <jack@suse.cz>
---
fs/dax.c | 28 ++++++++++------------------
1 file changed, 10 insertions(+), 18 deletions(-)
diff --git a/fs/dax.c b/fs/dax.c
index ccb8bc399d78..7c0036dd1570 100644
--- a/fs/dax.c
+++ b/fs/dax.c
@@ -119,18 +119,6 @@ int dax_clear_sectors(struct block_device *bdev, sector_t _sector, long _size)
}
EXPORT_SYMBOL_GPL(dax_clear_sectors);
-/* the clear_pmem() calls are ordered by a wmb_pmem() in the caller */
-static void dax_new_buf(void __pmem *addr, unsigned size, unsigned first,
- loff_t pos, loff_t end)
-{
- loff_t final = end - pos + first; /* The final byte of the buffer */
-
- if (first > 0)
- clear_pmem(addr, first);
- if (final < size)
- clear_pmem(addr + final, size - final);
-}
-
static bool buffer_written(struct buffer_head *bh)
{
return buffer_mapped(bh) && !buffer_unwritten(bh);
@@ -169,6 +157,9 @@ static ssize_t dax_io(struct inode *inode, struct iov_iter *iter,
struct blk_dax_ctl dax = {
.addr = (void __pmem *) ERR_PTR(-EIO),
};
+ unsigned blkbits = inode->i_blkbits;
+ sector_t file_blks = (i_size_read(inode) + (1 << blkbits) - 1)
+ >> blkbits;
if (rw == READ)
end = min(end, i_size_read(inode));
@@ -176,7 +167,6 @@ static ssize_t dax_io(struct inode *inode, struct iov_iter *iter,
while (pos < end) {
size_t len;
if (pos == max) {
- unsigned blkbits = inode->i_blkbits;
long page = pos >> PAGE_SHIFT;
sector_t block = page << (PAGE_SHIFT - blkbits);
unsigned first = pos - (block << blkbits);
@@ -192,6 +182,13 @@ static ssize_t dax_io(struct inode *inode, struct iov_iter *iter,
bh->b_size = 1 << blkbits;
bh_max = pos - first + bh->b_size;
bdev = bh->b_bdev;
+ /*
+ * We allow uninitialized buffers for writes
+ * beyond EOF as those cannot race with faults
+ */
+ WARN_ON_ONCE(
+ (buffer_new(bh) && block < file_blks) ||
+ (rw == WRITE && buffer_unwritten(bh)));
} else {
unsigned done = bh->b_size -
(bh_max - (pos - first));
@@ -211,11 +208,6 @@ static ssize_t dax_io(struct inode *inode, struct iov_iter *iter,
rc = map_len;
break;
}
- if (buffer_unwritten(bh) || buffer_new(bh)) {
- dax_new_buf(dax.addr, map_len, first,
- pos, end);
- need_wmb = true;
- }
dax.addr += first;
size = map_len - first;
}
--
2.6.6
next prev parent reply other threads:[~2016-05-11 9:59 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-05-11 9:58 [PATCH 0/7 v4] DAX cleanups and fixes Jan Kara
2016-05-11 9:58 ` [PATCH 1/7] DAX: move RADIX_DAX_ definitions to dax.c Jan Kara
2016-05-11 9:58 ` [PATCH 2/7] dax: Remove complete_unwritten argument Jan Kara
2016-05-11 9:58 ` [PATCH 3/7] ext2: Avoid DAX zeroing to corrupt data Jan Kara
2016-05-12 18:45 ` Ross Zwisler
2016-05-16 15:22 ` Jan Kara
2016-05-17 6:52 ` Verma, Vishal L
2016-05-17 7:19 ` Jan Kara
2016-05-17 17:56 ` Verma, Vishal L
2016-05-11 9:58 ` [PATCH 4/7] dax: Remove dead zeroing code from fault handlers Jan Kara
2016-05-11 9:58 ` Jan Kara [this message]
2016-05-11 9:58 ` [PATCH 6/7] dax: Remove pointless writeback from dax_do_io() Jan Kara
2016-05-11 9:58 ` [PATCH 7/7] dax: Remove redundant inode size checks Jan Kara
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=1462960733-29634-6-git-send-email-jack@suse.cz \
--to=jack@suse.cz \
--cc=dan.j.williams@intel.com \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-nvdimm@lists.01.org \
--cc=ross.zwisler@linux.intel.com \
--cc=tytso@mit.edu \
--cc=vishal.l.verma@intel.com \
/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