Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
Merge tag 'iomap-6.6-fixes-4' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Pull iomap fixes from Darrick Wong:
- Handle a race between writing and shrinking block devices by
returning EIO
- Fix a typo in a comment
* tag 'iomap-6.6-fixes-4' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
iomap: Spelling s/preceeding/preceding/g
iomap: add a workaround for racy i_size updates on block devices
+11
-2
+10
-1
fs/buffer.c
+10
-1
fs/buffer.c
···
2058
2058
fallthrough;
2059
2059
case IOMAP_MAPPED:
2060
2060
if ((iomap->flags & IOMAP_F_NEW) ||
2061
-
offset >= i_size_read(inode))
2061
+
offset >= i_size_read(inode)) {
2062
+
/*
2063
+
* This can happen if truncating the block device races
2064
+
* with the check in the caller as i_size updates on
2065
+
* block devices aren't synchronized by i_rwsem for
2066
+
* block devices.
2067
+
*/
2068
+
if (S_ISBLK(inode->i_mode))
2069
+
return -EIO;
2062
2070
set_buffer_new(bh);
2071
+
}
2063
2072
bh->b_blocknr = (iomap->addr + offset - iomap->offset) >>
2064
2073
inode->i_blkbits;
2065
2074
set_buffer_mapped(bh);
+1
-1
fs/iomap/buffered-io.c
+1
-1
fs/iomap/buffered-io.c
···
1049
1049
1050
1050
/*
1051
1051
* Scan the data range passed to us for dirty page cache folios. If we find a
1052
-
* dirty folio, punch out the preceeding range and update the offset from which
1052
+
* dirty folio, punch out the preceding range and update the offset from which
1053
1053
* the next punch will start from.
1054
1054
*
1055
1055
* We can punch out storage reservations under clean pages because they either