* [PATCH 0/1] nilfs2: fix data loss with mmap()
@ 2014-09-18 14:56 Ryusuke Konishi
2014-09-18 14:56 ` [PATCH] " Ryusuke Konishi
0 siblings, 1 reply; 4+ messages in thread
From: Ryusuke Konishi @ 2014-09-18 14:56 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-nilfs, LKML, Ryusuke Konishi, Andreas Rohner
Andrew,
Please apply the following bugfix and send it to upstream.
It fixes a regression of nilfs2's mmap which was brought by the commit
136e8770cd5d1fe3 "nilfs2: fix issue of nilfs_set_page_dirty() for page
at EOF boundary". Andreas Rohner recently found that the commit was
not perfect and causing reproducible silent data loss issue.
Thanks in advance,
Ryusuke Konishi
--
Andreas Rohner (1):
nilfs2: fix data loss with mmap()
fs/nilfs2/inode.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH] nilfs2: fix data loss with mmap()
2014-09-18 14:56 [PATCH 0/1] nilfs2: fix data loss with mmap() Ryusuke Konishi
@ 2014-09-18 14:56 ` Ryusuke Konishi
2014-09-18 19:17 ` Andrew Morton
0 siblings, 1 reply; 4+ messages in thread
From: Ryusuke Konishi @ 2014-09-18 14:56 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-nilfs, LKML, Ryusuke Konishi, Andreas Rohner
From: Andreas Rohner <andreas.rohner@gmx.net>
This bug leads to reproducible silent data loss, despite the use of
msync(), sync() and a clean unmount of the file system. It is easily
reproducible with the following script:
----------------[BEGIN SCRIPT]--------------------
mkfs.nilfs2 -f /dev/sdb
mount /dev/sdb /mnt
dd if=/dev/zero bs=1M count=30 of=/mnt/testfile
umount /mnt
mount /dev/sdb /mnt
CHECKSUM_BEFORE="$(md5sum /mnt/testfile)"
/root/mmaptest/mmaptest /mnt/testfile 30 10 5
sync
CHECKSUM_AFTER="$(md5sum /mnt/testfile)"
umount /mnt
mount /dev/sdb /mnt
CHECKSUM_AFTER_REMOUNT="$(md5sum /mnt/testfile)"
umount /mnt
echo "BEFORE MMAP:\t$CHECKSUM_BEFORE"
echo "AFTER MMAP:\t$CHECKSUM_AFTER"
echo "AFTER REMOUNT:\t$CHECKSUM_AFTER_REMOUNT"
----------------[END SCRIPT]--------------------
The mmaptest tool looks something like this (very simplified, with
error checking removed):
----------------[BEGIN mmaptest]--------------------
data = mmap(NULL, file_size - file_offset, PROT_READ | PROT_WRITE,
MAP_SHARED, fd, file_offset);
for (i = 0; i < write_count; ++i) {
memcpy(data + i * 4096, buf, sizeof(buf));
msync(data, file_size - file_offset, MS_SYNC))
}
----------------[END mmaptest]--------------------
The output of the script looks something like this:
BEFORE MMAP: 281ed1d5ae50e8419f9b978aab16de83 /mnt/testfile
AFTER MMAP: 6604a1c31f10780331a6850371b3a313 /mnt/testfile
AFTER REMOUNT: 281ed1d5ae50e8419f9b978aab16de83 /mnt/testfile
So it is clear, that the changes done using mmap() do not survive a
remount. This can be reproduced a 100% of the time. The problem was
introduced with the following commit:
136e877 nilfs2: fix issue of nilfs_set_page_dirty() for page at EOF
boundary
If the page was read with mpage_readpage() or mpage_readpages() for
example, then it has no buffers attached to it. In that case
page_has_buffers(page) in nilfs_set_page_dirty() will be false.
Therefore nilfs_set_file_dirty() is never called and the pages are never
collected and never written to disk.
This patch fixes the problem by also calling nilfs_set_file_dirty() if
the page has no buffers attached to it.
Signed-off-by: Andreas Rohner <andreas.rohner@gmx.net>
Tested-by: Andreas Rohner <andreas.rohner@gmx.net>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Cc: <stable@vger.kernel.org>
---
fs/nilfs2/inode.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c
index 6252b17..9e3c525 100644
--- a/fs/nilfs2/inode.c
+++ b/fs/nilfs2/inode.c
@@ -219,10 +219,10 @@ static int nilfs_writepage(struct page *page, struct writeback_control *wbc)
static int nilfs_set_page_dirty(struct page *page)
{
+ struct inode *inode = page->mapping->host;
int ret = __set_page_dirty_nobuffers(page);
if (page_has_buffers(page)) {
- struct inode *inode = page->mapping->host;
unsigned nr_dirty = 0;
struct buffer_head *bh, *head;
@@ -245,6 +245,10 @@ static int nilfs_set_page_dirty(struct page *page)
if (nr_dirty)
nilfs_set_file_dirty(inode, nr_dirty);
+ } else if (ret) {
+ unsigned nr_dirty = 1 << (PAGE_SHIFT - inode->i_blkbits);
+
+ nilfs_set_file_dirty(inode, nr_dirty);
}
return ret;
}
--
1.7.9.4
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] nilfs2: fix data loss with mmap()
2014-09-18 14:56 ` [PATCH] " Ryusuke Konishi
@ 2014-09-18 19:17 ` Andrew Morton
2014-09-18 23:41 ` Ryusuke Konishi
0 siblings, 1 reply; 4+ messages in thread
From: Andrew Morton @ 2014-09-18 19:17 UTC (permalink / raw)
To: Ryusuke Konishi; +Cc: linux-nilfs, LKML, Andreas Rohner
On Thu, 18 Sep 2014 23:56:25 +0900 Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> wrote:
> From: Andreas Rohner <andreas.rohner@gmx.net>
>
> This bug leads to reproducible silent data loss, despite the use of
> msync(), sync() and a clean unmount of the file system. It is easily
> reproducible with the following script:
>
> ...
>
> --- a/fs/nilfs2/inode.c
> +++ b/fs/nilfs2/inode.c
> @@ -219,10 +219,10 @@ static int nilfs_writepage(struct page *page, struct writeback_control *wbc)
>
> static int nilfs_set_page_dirty(struct page *page)
> {
> + struct inode *inode = page->mapping->host;
> int ret = __set_page_dirty_nobuffers(page);
>
> if (page_has_buffers(page)) {
> - struct inode *inode = page->mapping->host;
> unsigned nr_dirty = 0;
> struct buffer_head *bh, *head;
>
> @@ -245,6 +245,10 @@ static int nilfs_set_page_dirty(struct page *page)
>
> if (nr_dirty)
> nilfs_set_file_dirty(inode, nr_dirty);
> + } else if (ret) {
> + unsigned nr_dirty = 1 << (PAGE_SHIFT - inode->i_blkbits);
It's quite cosmetic, but it is conventional to use PAGE_CACHE_SHIFT here.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] nilfs2: fix data loss with mmap()
2014-09-18 19:17 ` Andrew Morton
@ 2014-09-18 23:41 ` Ryusuke Konishi
0 siblings, 0 replies; 4+ messages in thread
From: Ryusuke Konishi @ 2014-09-18 23:41 UTC (permalink / raw)
To: Andrew Morton; +Cc: linux-nilfs, LKML, Andreas Rohner
On Thu, 18 Sep 2014 12:17:08 -0700, Andrew Morton wrote:
> On Thu, 18 Sep 2014 23:56:25 +0900 Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> wrote:
>
>> From: Andreas Rohner <andreas.rohner@gmx.net>
>>
>> This bug leads to reproducible silent data loss, despite the use of
>> msync(), sync() and a clean unmount of the file system. It is easily
>> reproducible with the following script:
>>
>> ...
>>
>> --- a/fs/nilfs2/inode.c
>> +++ b/fs/nilfs2/inode.c
>> @@ -219,10 +219,10 @@ static int nilfs_writepage(struct page *page, struct writeback_control *wbc)
>>
>> static int nilfs_set_page_dirty(struct page *page)
>> {
>> + struct inode *inode = page->mapping->host;
>> int ret = __set_page_dirty_nobuffers(page);
>>
>> if (page_has_buffers(page)) {
>> - struct inode *inode = page->mapping->host;
>> unsigned nr_dirty = 0;
>> struct buffer_head *bh, *head;
>>
>> @@ -245,6 +245,10 @@ static int nilfs_set_page_dirty(struct page *page)
>>
>> if (nr_dirty)
>> nilfs_set_file_dirty(inode, nr_dirty);
>> + } else if (ret) {
>> + unsigned nr_dirty = 1 << (PAGE_SHIFT - inode->i_blkbits);
>
> It's quite cosmetic, but it is conventional to use PAGE_CACHE_SHIFT here.
Agreed.
Thanks,
Ryusuke Konishi
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2014-09-18 23:41 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-09-18 14:56 [PATCH 0/1] nilfs2: fix data loss with mmap() Ryusuke Konishi
2014-09-18 14:56 ` [PATCH] " Ryusuke Konishi
2014-09-18 19:17 ` Andrew Morton
2014-09-18 23:41 ` Ryusuke Konishi
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).