Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
Merge tag 'erofs-for-7.0-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs
Pull erofs fixes from Gao Xiang:
- Mark I/Os as failed when encountering short reads on file-backed
mounts
- Label GFP_NOIO in the BIO completion when the completion is in the
process context, and directly call into the decompression to avoid
deadlocks
- Improve Kconfig descriptions to better highlight the overall efforts
- Fix .fadvise() for page cache sharing
* tag 'erofs-for-7.0-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs:
erofs: fix .fadvise() for page cache sharing
erofs: update the Kconfig description
erofs: add GFP_NOIO in the bio completion if needed
erofs: set fileio bio failed in short read case
+47
-20
+29
-14
fs/erofs/Kconfig
+29
-14
fs/erofs/Kconfig
···
16
16
select ZLIB_INFLATE if EROFS_FS_ZIP_DEFLATE
17
17
select ZSTD_DECOMPRESS if EROFS_FS_ZIP_ZSTD
18
18
help
19
-
EROFS (Enhanced Read-Only File System) is a lightweight read-only
20
-
file system with modern designs (e.g. no buffer heads, inline
21
-
xattrs/data, chunk-based deduplication, multiple devices, etc.) for
22
-
scenarios which need high-performance read-only solutions, e.g.
23
-
smartphones with Android OS, LiveCDs and high-density hosts with
24
-
numerous containers;
19
+
EROFS (Enhanced Read-Only File System) is a modern, lightweight,
20
+
secure read-only filesystem for various use cases, such as immutable
21
+
system images, container images, application sandboxes, and datasets.
25
22
26
-
It also provides transparent compression and deduplication support to
27
-
improve storage density and maintain relatively high compression
28
-
ratios, and it implements in-place decompression to temporarily reuse
29
-
page cache for compressed data using proper strategies, which is
30
-
quite useful for ensuring guaranteed end-to-end runtime decompression
23
+
EROFS uses a flexible, hierarchical on-disk design so that features
24
+
can be enabled on demand: the core on-disk format is block-aligned in
25
+
order to perform optimally on all kinds of devices, including block
26
+
and memory-backed devices; the format is easy to parse and has zero
27
+
metadata redundancy, unlike generic filesystems, making it ideal for
28
+
filesystem auditing and remote access; inline data, random-access
29
+
friendly directory data, inline/shared extended attributes and
30
+
chunk-based deduplication ensure space efficiency while maintaining
31
+
high performance.
32
+
33
+
Optionally, it supports multiple devices to reference external data,
34
+
enabling data sharing for container images.
35
+
36
+
It also has advanced encoded on-disk layouts, particularly for data
37
+
compression and fine-grained deduplication. It utilizes fixed-size
38
+
output compression to improve storage density while keeping relatively
39
+
high compression ratios. Furthermore, it implements in-place
40
+
decompression to reuse file pages to keep compressed data temporarily
41
+
with proper strategies, which ensures guaranteed end-to-end runtime
31
42
performance under extreme memory pressure without extra cost.
32
43
33
-
See the documentation at <file:Documentation/filesystems/erofs.rst>
34
-
and the web pages at <https://erofs.docs.kernel.org> for more details.
44
+
For more details, see the web pages at <https://erofs.docs.kernel.org>
45
+
and the documentation at <file:Documentation/filesystems/erofs.rst>.
46
+
47
+
To compile EROFS filesystem support as a module, choose M here. The
48
+
module will be called erofs.
35
49
36
50
If unsure, say N.
37
51
···
119
105
depends on EROFS_FS
120
106
default y
121
107
help
122
-
Enable transparent compression support for EROFS file systems.
108
+
Enable EROFS compression layouts so that filesystems containing
109
+
compressed files can be parsed by the kernel.
123
110
124
111
If you don't want to enable compression feature, say N.
125
112
+2
-4
fs/erofs/fileio.c
+2
-4
fs/erofs/fileio.c
···
25
25
container_of(iocb, struct erofs_fileio_rq, iocb);
26
26
struct folio_iter fi;
27
27
28
-
if (ret >= 0 && ret != rq->bio.bi_iter.bi_size) {
29
-
bio_advance(&rq->bio, ret);
30
-
zero_fill_bio(&rq->bio);
31
-
}
28
+
if (ret >= 0 && ret != rq->bio.bi_iter.bi_size)
29
+
ret = -EIO;
32
30
if (!rq->bio.bi_end_io) {
33
31
bio_for_each_folio_all(fi, &rq->bio) {
34
32
DBG_BUGON(folio_test_uptodate(fi.folio));
+3
fs/erofs/zdata.c
+3
fs/erofs/zdata.c
···
1445
1445
int bios)
1446
1446
{
1447
1447
struct erofs_sb_info *const sbi = EROFS_SB(io->sb);
1448
+
int gfp_flag;
1448
1449
1449
1450
/* wake up the caller thread for sync decompression */
1450
1451
if (io->sync) {
···
1478
1477
sbi->sync_decompress = EROFS_SYNC_DECOMPRESS_FORCE_ON;
1479
1478
return;
1480
1479
}
1480
+
gfp_flag = memalloc_noio_save();
1481
1481
z_erofs_decompressqueue_work(&io->u.work);
1482
+
memalloc_noio_restore(gfp_flag);
1482
1483
}
1483
1484
1484
1485
static void z_erofs_fill_bio_vec(struct bio_vec *bvec,