On Thu, Sep 24, 2020 at 05:08:03PM +0200, Sedat Dilek wrote:
> On Thu, Sep 24, 2020 at 2:58 PM Matthew Wilcox (Oracle)
> <willy@xxxxxxxxxxxxx> wrote:
> >
> > For filesystems with block size < page size, we need to set all the
> > per-block uptodate bits if the page was already uptodate at the time
> > we create the per-block metadata. This can happen if the page is
> > invalidated (eg by a write to drop_caches) but ultimately not removed
> > from the page cache.
> >
> > This is a data corruption issue as page writeback skips blocks which
> > are marked !uptodate.
> >
> > Fixes: 9dc55f1389f9 ("iomap: add support for sub-pagesize buffered I/O without buffer heads")
>
> This commit is also in Linux v5.9-rc6+ but does not cleanly apply.
> Against which Git tree is this patch?
> When Linux v5.9-rc6+ is affected, do you have a backport?
This applies to v5.8. I'll happily backport this for any other kernel
versions.
>From 51f85a97ccdd7071e5f95b2ac4e41c12bf4d4176 Mon Sep 17 00:00:00 2001
From: "Matthew Wilcox (Oracle)" <willy@xxxxxxxxxxxxx>
Date: Thu, 24 Sep 2020 08:44:56 -0400
Subject: [PATCH] iomap: Set all uptodate bits for an Uptodate page
For filesystems with block size < page size, we need to set all the
per-block uptodate bits if the page was already uptodate at the time
we create the per-block metadata. This can happen if the page is
invalidated (eg by a write to drop_caches) but ultimately not removed
from the page cache.
This is a data corruption issue as page writeback skips blocks which
are marked !uptodate.
Fixes: 9dc55f1389f9 ("iomap: add support for sub-pagesize buffered I/O without buffer heads")
Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
Reported-by: Qian Cai <cai@xxxxxxxxxx>
Cc: Brian Foster <bfoster@xxxxxxxxxx>
---
fs/iomap/buffered-io.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index bcfc288dba3f..810f7dae11d9 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -53,7 +53,10 @@ iomap_page_create(struct inode *inode, struct page *page)
atomic_set(&iop->read_count, 0);
atomic_set(&iop->write_count, 0);
spin_lock_init(&iop->uptodate_lock);
- bitmap_zero(iop->uptodate, PAGE_SIZE / SECTOR_SIZE);
+ if (PageUptodate(page))
+ bitmap_fill(iop->uptodate, PAGE_SIZE / SECTOR_SIZE);
+ else
+ bitmap_zero(iop->uptodate, PAGE_SIZE / SECTOR_SIZE);
/*
* migrate_page_move_mapping() assumes that pages with private data have
--
2.28.0
