linux.git/fs/squashfs/super.c, branch for-next Linux kernel source tree squashfs: cache partial compressed blocks 2023-06-10T00:44:14+00:00 Vincent Whitchurch vincent.whitchurch@axis.com 2023-05-17T14:18:19+00:00 e994f5b677ee016a2535d9df826315122da1ae65 Before commit 93e72b3c612adcaca1 ("squashfs: migrate from ll_rw_block usage to BIO"), compressed blocks read by squashfs were cached in the page cache, but that is not the case after that commit. That has lead to squashfs having to re-read a lot of sectors from disk/flash. For example, the first sectors of every metadata block need to be read twice from the disk. Once partially to read the length, and a second time to read the block itself. Also, in linear reads of large files, the last sectors of one data block are re-read from disk when reading the next data block, since the compressed blocks are of variable sizes and not aligned to device blocks. This extra I/O results in a degrade in read performance of, for example, ~16% in one scenario on my ARM platform using squashfs with dm-verity and NAND. Since the decompressed data is cached in the page cache or squashfs' internal metadata and fragment caches, caching _all_ compressed pages would lead to a lot of double caching and is undesirable. But make the code cache any disk blocks which were only partially requested, since these are the ones likely to include data which is needed by other file system blocks. This restores read performance in my test scenario. The compressed block caching is only applied when the disk block size is equal to the page size, to avoid having to deal with caching sub-page reads. [akpm@linux-foundation.org: fs/squashfs/block.c needs linux/pagemap.h] [vincent.whitchurch@axis.com: fix page update race] Link: https://lkml.kernel.org/r/20230526-squashfs-cache-fixup-v1-1-d54a7fa23e7b@axis.com [vincent.whitchurch@axis.com: fix page indices] Link: https://lkml.kernel.org/r/20230526-squashfs-cache-fixup-v1-2-d54a7fa23e7b@axis.com [akpm@linux-foundation.org: fix layout, per hch] Link: https://lkml.kernel.org/r/20230510-squashfs-cache-v4-1-3bd394e1ee71@axis.com Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> 
Before commit 93e72b3c612adcaca1 ("squashfs: migrate from ll_rw_block
usage to BIO"), compressed blocks read by squashfs were cached in the page
cache, but that is not the case after that commit.  That has lead to
squashfs having to re-read a lot of sectors from disk/flash.

For example, the first sectors of every metadata block need to be read
twice from the disk.  Once partially to read the length, and a second time
to read the block itself.  Also, in linear reads of large files, the last
sectors of one data block are re-read from disk when reading the next data
block, since the compressed blocks are of variable sizes and not aligned
to device blocks.  This extra I/O results in a degrade in read performance
of, for example, ~16% in one scenario on my ARM platform using squashfs
with dm-verity and NAND.

Since the decompressed data is cached in the page cache or squashfs'
internal metadata and fragment caches, caching _all_ compressed pages
would lead to a lot of double caching and is undesirable.  But make the
code cache any disk blocks which were only partially requested, since
these are the ones likely to include data which is needed by other file
system blocks.  This restores read performance in my test scenario.

The compressed block caching is only applied when the disk block size is
equal to the page size, to avoid having to deal with caching sub-page
reads.

[akpm@linux-foundation.org: fs/squashfs/block.c needs linux/pagemap.h]
[vincent.whitchurch@axis.com: fix page update race]
  Link: https://lkml.kernel.org/r/20230526-squashfs-cache-fixup-v1-1-d54a7fa23e7b@axis.com
[vincent.whitchurch@axis.com: fix page indices]
  Link: https://lkml.kernel.org/r/20230526-squashfs-cache-fixup-v1-2-d54a7fa23e7b@axis.com
[akpm@linux-foundation.org: fix layout, per hch]
Link: https://lkml.kernel.org/r/20230510-squashfs-cache-v4-1-3bd394e1ee71@axis.com
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Merge tag 'fs.idmapped.squashfs.v6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/idmapping 2022-12-13T04:24:51+00:00 Linus Torvalds torvalds@linux-foundation.org 2022-12-13T04:24:51+00:00 6e8948a0630f9ee46cf03dbf65949c1f4b6f6dd2 Pull squashfs update from Seth Forshee: "This is a simple patch to enable idmapped mounts for squashfs. All functionality squashfs needs to support idmapped mounts is already implemented in generic VFS code, so all that is needed is to set FS_ALLOW_IDMAP in fs_flags" * tag 'fs.idmapped.squashfs.v6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/idmapping: squashfs: enable idmapped mounts 
Pull squashfs update from Seth Forshee:
 "This is a simple patch to enable idmapped mounts for squashfs.

  All functionality squashfs needs to support idmapped mounts is already
  implemented in generic VFS code, so all that is needed is to set
  FS_ALLOW_IDMAP in fs_flags"

* tag 'fs.idmapped.squashfs.v6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/idmapping:
  squashfs: enable idmapped mounts
squashfs: fix null-ptr-deref in squashfs_fill_super 2022-11-18T21:55:09+00:00 Baokun Li libaokun1@huawei.com 2022-11-01T07:33:43+00:00 c7e8d3279c984e41165a7b510759bd1771ac3941 When squashfs_read_table() returns an error or `sb->s_magic != SQUASHFS_MAGIC`, enters the error branch and calls msblk->thread_ops->destroy(msblk) to destroy msblk. However, msblk->thread_ops has not been initialized. Therefore, the following problem is triggered: ================================================================== BUG: KASAN: null-ptr-deref in squashfs_fill_super+0xe7a/0x13b0 Read of size 8 at addr 0000000000000008 by task swapper/0/1 CPU: 0 PID: 1 Comm: swapper/0 Not tainted 6.1.0-rc3-next-20221031 #367 Call Trace: <TASK> dump_stack_lvl+0x73/0x9f print_report+0x743/0x759 kasan_report+0xc0/0x120 __asan_load8+0xd3/0x140 squashfs_fill_super+0xe7a/0x13b0 get_tree_bdev+0x27b/0x450 squashfs_get_tree+0x19/0x30 vfs_get_tree+0x49/0x150 path_mount+0xaae/0x1350 init_mount+0xad/0x100 do_mount_root+0xbc/0x1d0 mount_block_root+0x173/0x316 mount_root+0x223/0x236 prepare_namespace+0x1eb/0x237 kernel_init_freeable+0x528/0x576 kernel_init+0x29/0x250 ret_from_fork+0x1f/0x30 </TASK> ================================================================== To solve this issue, msblk->thread_ops is initialized immediately after msblk is assigned a value. Link: https://lkml.kernel.org/r/20221101073343.3961562-1-libaokun1@huawei.com Fixes: b0645770d3c7 ("squashfs: add the mount parameter theads=<single|multi|percpu>") Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Xiaoming Ni <nixiaoming@huawei.com> Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk> Cc: Yu Kuai <yukuai3@huawei.com> Cc: Zhang Yi <yi.zhang@huawei.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> 
When squashfs_read_table() returns an error or `sb->s_magic !=
SQUASHFS_MAGIC`, enters the error branch and calls
msblk->thread_ops->destroy(msblk) to destroy msblk.  However,
msblk->thread_ops has not been initialized.  Therefore, the following
problem is triggered:

==================================================================
BUG: KASAN: null-ptr-deref in squashfs_fill_super+0xe7a/0x13b0
Read of size 8 at addr 0000000000000008 by task swapper/0/1

CPU: 0 PID: 1 Comm: swapper/0 Not tainted 6.1.0-rc3-next-20221031 #367
Call Trace:
 <TASK>
 dump_stack_lvl+0x73/0x9f
 print_report+0x743/0x759
 kasan_report+0xc0/0x120
 __asan_load8+0xd3/0x140
 squashfs_fill_super+0xe7a/0x13b0
 get_tree_bdev+0x27b/0x450
 squashfs_get_tree+0x19/0x30
 vfs_get_tree+0x49/0x150
 path_mount+0xaae/0x1350
 init_mount+0xad/0x100
 do_mount_root+0xbc/0x1d0
 mount_block_root+0x173/0x316
 mount_root+0x223/0x236
 prepare_namespace+0x1eb/0x237
 kernel_init_freeable+0x528/0x576
 kernel_init+0x29/0x250
 ret_from_fork+0x1f/0x30
 </TASK>
==================================================================

To solve this issue, msblk->thread_ops is initialized immediately after
msblk is assigned a value.

Link: https://lkml.kernel.org/r/20221101073343.3961562-1-libaokun1@huawei.com
Fixes: b0645770d3c7 ("squashfs: add the mount parameter theads=<single|multi|percpu>")
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Xiaoming Ni <nixiaoming@huawei.com>
Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk>
Cc: Yu Kuai <yukuai3@huawei.com>
Cc: Zhang Yi <yi.zhang@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
squashfs: allows users to configure the number of decompression threads 2022-11-18T21:55:08+00:00 Xiaoming Ni nixiaoming@huawei.com 2022-10-19T03:09:30+00:00 fb40fe04f9df23114782d5edd1c5d017ae9d0ca8 The maximum number of threads in the decompressor_multi.c file is fixed and cannot be adjusted according to user needs. Therefore, the mount parameter needs to be added to allow users to configure the number of threads as required. The upper limit is num_online_cpus() * 2. Link: https://lkml.kernel.org/r/20221019030930.130456-3-nixiaoming@huawei.com Signed-off-by: Xiaoming Ni <nixiaoming@huawei.com> Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk> Cc: Jianguo Chen <chenjianguo3@huawei.com> Cc: Jubin Zhong <zhongjubin@huawei.com> Cc: Zhang Yi <yi.zhang@huawei.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> 
The maximum number of threads in the decompressor_multi.c file is fixed
and cannot be adjusted according to user needs.  Therefore, the mount
parameter needs to be added to allow users to configure the number of
threads as required.  The upper limit is num_online_cpus() * 2.

Link: https://lkml.kernel.org/r/20221019030930.130456-3-nixiaoming@huawei.com
Signed-off-by: Xiaoming Ni <nixiaoming@huawei.com>
Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk>
Cc: Jianguo Chen <chenjianguo3@huawei.com>
Cc: Jubin Zhong <zhongjubin@huawei.com>
Cc: Zhang Yi <yi.zhang@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
squashfs: add the mount parameter theads=<single|multi|percpu> 2022-11-18T21:55:08+00:00 Xiaoming Ni nixiaoming@huawei.com 2022-10-19T03:09:29+00:00 80f784098ff44e086f68f0e8c98b6c6da8702ec4 Patch series 'squashfs: Add the mount parameter "threads="'. Currently, Squashfs supports multiple decompressor parallel modes. However, this mode can be configured only during kernel building and does not support flexible selection during runtime. In the current patch set, the mount parameter "threads=" is added to allow users to select the parallel decompressor mode and configure the number of decompressors when mounting a file system. "threads=<single|multi|percpu|1|2|3|...>" The upper limit is num_online_cpus() * 2. This patch (of 2): Squashfs supports three decompression concurrency modes: Single-thread mode: concurrent reads are blocked and the memory overhead is small. Multi-thread mode/percpu mode: reduces concurrent read blocking but increases memory overhead. The corresponding schema must be fixed at compile time. During mounting, the concurrent decompression mode cannot be adjusted based on file read blocking. The mount parameter theads=<single|multi|percpu> is added to select the concurrent decompression mode of a single SquashFS file system image. Link: https://lkml.kernel.org/r/20221019030930.130456-1-nixiaoming@huawei.com Link: https://lkml.kernel.org/r/20221019030930.130456-2-nixiaoming@huawei.com Signed-off-by: Xiaoming Ni <nixiaoming@huawei.com> Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk> Cc: Jianguo Chen <chenjianguo3@huawei.com> Cc: Jubin Zhong <zhongjubin@huawei.com> Cc: Zhang Yi <yi.zhang@huawei.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> 
Patch series 'squashfs: Add the mount parameter "threads="'.

Currently, Squashfs supports multiple decompressor parallel modes. 
However, this mode can be configured only during kernel building and does
not support flexible selection during runtime.

In the current patch set, the mount parameter "threads=" is added to allow
users to select the parallel decompressor mode and configure the number of
decompressors when mounting a file system.

"threads=<single|multi|percpu|1|2|3|...>"
The upper limit is num_online_cpus() * 2.


This patch (of 2):

Squashfs supports three decompression concurrency modes:
	Single-thread mode: concurrent reads are blocked and the memory
		overhead is small.
	Multi-thread mode/percpu mode: reduces concurrent read blocking but
		increases memory overhead.

The corresponding schema must be fixed at compile time. During mounting,
the concurrent decompression mode cannot be adjusted based on file read
blocking.

The mount parameter theads=<single|multi|percpu> is added to select
the concurrent decompression mode of a single SquashFS file system
image.

Link: https://lkml.kernel.org/r/20221019030930.130456-1-nixiaoming@huawei.com
Link: https://lkml.kernel.org/r/20221019030930.130456-2-nixiaoming@huawei.com
Signed-off-by: Xiaoming Ni <nixiaoming@huawei.com>
Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk>
Cc: Jianguo Chen <chenjianguo3@huawei.com>
Cc: Jubin Zhong <zhongjubin@huawei.com>
Cc: Zhang Yi <yi.zhang@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
squashfs: enable idmapped mounts 2022-11-07T09:24:22+00:00 Michael Weiß michael.weiss@aisec.fraunhofer.de 2022-10-24T19:15:52+00:00 42da66ac7bcb19181385e851094ceedfe7c81984 For squashfs all needed functionality for idmapped mounts is already implemented by the generic handlers in the VFS. Thus, it is sufficient to just enable the corresponding FS_ALLOW_IDMAP flag to support idmapped mounts. We use this for unprivileged (user namespaced) containers based on squashfs images as rootfs in GyroidOS. A simple test using the mount-idmapped tool executed as user with uid=1000 looks as follows: $ mkdir test $ echo "test" > test/test_file $ mksquashfs test/ fs.img $ sudo mkdir /mnt/test $ sudo mkdir /mnt/mapped $ sudo mount fs.img -o loop /mnt/test/ $ sudo ./mount-idmapped --map-mount b:1000:2000:1 /mnt/test/ /mnt/mapped/ $ mount | tail -n2 fs.img on /mnt/test type squashfs (ro,relatime,errors=continue) fs.img on /mnt/mapped type squashfs (ro,relatime,idmapped,errors=continue) $ ls -lan /mnt/test/ total 5 drwxr-xr-x 2 1000 1000 32 Okt 24 13:36 . drwxr-xr-x 6 0 0 4096 Okt 24 13:38 .. -rw-r--r-- 1 1000 1000 5 Okt 24 13:36 test_file $ ls -lan /mnt/mapped/ total 5 drwxr-xr-x 2 2000 2000 32 Okt 24 13:36 . drwxr-xr-x 6 0 0 4096 Okt 24 13:38 .. -rw-r--r-- 1 2000 2000 5 Okt 24 13:36 test_file Signed-off-by: Michael Weiß <michael.weiss@aisec.fraunhofer.de> Reviewed-by: Christian Brauner <brauner@kernel.org> Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk> Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org> 
For squashfs all needed functionality for idmapped mounts is already
implemented by the generic handlers in the VFS. Thus, it is sufficient
to just enable the corresponding FS_ALLOW_IDMAP flag to support
idmapped mounts.

We use this for unprivileged (user namespaced) containers based on
squashfs images as rootfs in GyroidOS.

A simple test using the mount-idmapped tool executed as user with
uid=1000 looks as follows:

$ mkdir test
$ echo "test" > test/test_file
$ mksquashfs test/ fs.img
$ sudo mkdir /mnt/test
$ sudo mkdir /mnt/mapped
$ sudo mount fs.img -o loop /mnt/test/
$ sudo ./mount-idmapped --map-mount b:1000:2000:1 /mnt/test/ /mnt/mapped/

$ mount | tail -n2
fs.img on /mnt/test type squashfs (ro,relatime,errors=continue)
fs.img on /mnt/mapped type squashfs (ro,relatime,idmapped,errors=continue)

$ ls -lan /mnt/test/
total 5
drwxr-xr-x 2 1000 1000   32 Okt 24 13:36 .
drwxr-xr-x 6    0    0 4096 Okt 24 13:38 ..
-rw-r--r-- 1 1000 1000    5 Okt 24 13:36 test_file

$ ls -lan /mnt/mapped/
total 5
drwxr-xr-x 2 2000 2000   32 Okt 24 13:36 .
drwxr-xr-x 6    0    0 4096 Okt 24 13:38 ..
-rw-r--r-- 1 2000 2000    5 Okt 24 13:36 test_file

Signed-off-by: Michael Weiß <michael.weiss@aisec.fraunhofer.de>
Reviewed-by: Christian Brauner <brauner@kernel.org>
Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk>
Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org>
Revert "squashfs: provide backing_dev_info in order to disable read-ahead" 2022-07-30T01:12:34+00:00 Hsin-Yi Wang hsinyi@chromium.org 2022-06-17T08:38:09+00:00 0c12185728d602c27cd12a845249e7f37197f71f Patch series "Implement readahead for squashfs", v7. Commit 9eec1d897139("squashfs: provide backing_dev_info in order to disable read-ahead") mitigates the performance drop issue for squashfs by closing readahead for it. This series implements readahead callback for squashfs. This patch (of 4): This reverts 9eec1d897139e5 ("squashfs: provide backing_dev_info in order to disable read-ahead"). Revert closing the readahead to squashfs since the readahead callback for squashfs is implemented. Link: https://lkml.kernel.org/r/20220617083810.337573-1-hsinyi@chromium.org Link: https://lkml.kernel.org/r/20220617083810.337573-2-hsinyi@chromium.org Signed-off-by: Hsin-Yi Wang <hsinyi@chromium.org> Suggested-by: Xiongwei Song <Xiongwei.Song@windriver.com> Cc: Phillip Lougher <phillip@squashfs.org.uk> Cc: Matthew Wilcox <willy@infradead.org> Cc: Marek Szyprowski <m.szyprowski@samsung.com> Cc: Zheng Liang <zhengliang6@huawei.com> Cc: Zhang Yi <yi.zhang@huawei.com> Cc: Hou Tao <houtao1@huawei.com> Cc: Miao Xie <miaoxie@huawei.com> Cc: kernel test robot <lkp@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> 
Patch series "Implement readahead for squashfs", v7.

Commit 9eec1d897139("squashfs: provide backing_dev_info in order to
disable read-ahead") mitigates the performance drop issue for squashfs by
closing readahead for it.

This series implements readahead callback for squashfs.


This patch (of 4):

This reverts 9eec1d897139e5 ("squashfs: provide backing_dev_info in order
to disable read-ahead").

Revert closing the readahead to squashfs since the readahead callback for
squashfs is implemented.

Link: https://lkml.kernel.org/r/20220617083810.337573-1-hsinyi@chromium.org
Link: https://lkml.kernel.org/r/20220617083810.337573-2-hsinyi@chromium.org
Signed-off-by: Hsin-Yi Wang <hsinyi@chromium.org>
Suggested-by: Xiongwei Song <Xiongwei.Song@windriver.com>
Cc: Phillip Lougher <phillip@squashfs.org.uk>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Zheng Liang <zhengliang6@huawei.com>
Cc: Zhang Yi <yi.zhang@huawei.com>
Cc: Hou Tao <houtao1@huawei.com>
Cc: Miao Xie <miaoxie@huawei.com>

Cc: kernel test robot <lkp@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
squashfs: Convert squashfs to read_folio 2022-05-09T20:21:46+00:00 Matthew Wilcox (Oracle) willy@infradead.org 2022-04-29T15:12:16+00:00 124cfc154f6cfb530bfb36e7728406c56ebf37ad This is a "weak" conversion which converts straight back to using pages. A full conversion should be performed at some point, hopefully by someone familiar with the filesystem. Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> 
This is a "weak" conversion which converts straight back to using pages.
A full conversion should be performed at some point, hopefully by
someone familiar with the filesystem.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
fs: allocate inode by using alloc_inode_sb() 2022-03-22T22:57:03+00:00 Muchun Song songmuchun@bytedance.com 2022-03-22T21:41:03+00:00 fd60b28842df833477c42da6a6d63d0d114a5fcc The inode allocation is supposed to use alloc_inode_sb(), so convert kmem_cache_alloc() of all filesystems to alloc_inode_sb(). Link: https://lkml.kernel.org/r/20220228122126.37293-5-songmuchun@bytedance.com Signed-off-by: Muchun Song <songmuchun@bytedance.com> Acked-by: Theodore Ts'o <tytso@mit.edu> [ext4] Acked-by: Roman Gushchin <roman.gushchin@linux.dev> Cc: Alex Shi <alexs@kernel.org> Cc: Anna Schumaker <Anna.Schumaker@Netapp.com> Cc: Chao Yu <chao@kernel.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Fam Zheng <fam.zheng@bytedance.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Kari Argillander <kari.argillander@gmail.com> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Qi Zheng <zhengqi.arch@bytedance.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Wei Yang <richard.weiyang@gmail.com> Cc: Xiongchun Duan <duanxiongchun@bytedance.com> Cc: Yang Shi <shy828301@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 
The inode allocation is supposed to use alloc_inode_sb(), so convert
kmem_cache_alloc() of all filesystems to alloc_inode_sb().

Link: https://lkml.kernel.org/r/20220228122126.37293-5-songmuchun@bytedance.com
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Acked-by: Theodore Ts'o <tytso@mit.edu>		[ext4]
Acked-by: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Alex Shi <alexs@kernel.org>
Cc: Anna Schumaker <Anna.Schumaker@Netapp.com>
Cc: Chao Yu <chao@kernel.org>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Fam Zheng <fam.zheng@bytedance.com>
Cc: Jaegeuk Kim <jaegeuk@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Kari Argillander <kari.argillander@gmail.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Qi Zheng <zhengqi.arch@bytedance.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Trond Myklebust <trond.myklebust@hammerspace.com>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Xiongchun Duan <duanxiongchun@bytedance.com>
Cc: Yang Shi <shy828301@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
squashfs: provide backing_dev_info in order to disable read-ahead 2022-01-15T14:30:24+00:00 Zheng Liang zhengliang6@huawei.com 2022-01-14T22:03:31+00:00 9eec1d897139e5de287af5d559a02b811b844d82 Commit c1f6925e1091 ("mm: put readahead pages in cache earlier") causes the read performance of squashfs to deteriorate.Through testing, we find that the performance will be back by closing the readahead of squashfs. So we want to learn the way of ubifs, provides backing_dev_info and disable read-ahead We tested the following data by fio. squashfs image blocksize=128K test command: fio --name basic --bs=? --filename="/mnt/test_file" --rw=? --iodepth=1 --ioengine=psync --runtime=200 --time_based turn on squashfs readahead in 5.10 kernel bs(k) read/randread MB/s 4 randread 271 128 randread 231 1024 randread 246 4 read 310 128 read 245 1024 read 247 turn off squashfs readahead in 5.10 kernel bs(k) read/randread MB/s 4 randread 293 128 randread 330 1024 randread 363 4 read 338 128 read 360 1024 read 365 turn on squashfs readahead and revert the commit c1f6925e1091("mm: put readahead pages in cache earlier") in 5.10 kernel bs(k) read/randread MB/s 4 randread 289 128 randread 306 1024 randread 335 4 read 337 128 read 336 1024 read 338 Link: https://lkml.kernel.org/r/20211116113141.1391026-1-zhengliang6@huawei.com Signed-off-by: Zheng Liang <zhengliang6@huawei.com> Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk> Cc: Zhang Yi <yi.zhang@huawei.com> Cc: Hou Tao <houtao1@huawei.com> Cc: Miao Xie <miaoxie@huawei.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 
Commit c1f6925e1091 ("mm: put readahead pages in cache earlier") causes
the read performance of squashfs to deteriorate.Through testing, we find
that the performance will be back by closing the readahead of squashfs.

So we want to learn the way of ubifs, provides backing_dev_info and
disable read-ahead

We tested the following data by fio.
squashfs image blocksize=128K
test command:

  fio --name basic --bs=? --filename="/mnt/test_file" --rw=? --iodepth=1 --ioengine=psync --runtime=200 --time_based

  turn on squashfs readahead in 5.10 kernel
  bs(k)      read/randread           MB/s
  4            randread              271
  128          randread              231
  1024         randread              246
  4            read                  310
  128          read                  245
  1024         read                  247

  turn off squashfs readahead in 5.10 kernel
  bs(k)      read/randread           MB/s
  4            randread              293
  128          randread              330
  1024         randread              363
  4            read                  338
  128          read                  360
  1024         read                  365

  turn on squashfs readahead and revert the
  commit c1f6925e1091("mm: put readahead
  pages in cache earlier") in 5.10 kernel
  bs(k)      read/randread           MB/s
  4           randread               289
  128         randread               306
  1024        randread               335
  4           read                   337
  128         read                   336
  1024        read                   338

Link: https://lkml.kernel.org/r/20211116113141.1391026-1-zhengliang6@huawei.com
Signed-off-by: Zheng Liang <zhengliang6@huawei.com>
Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk>
Cc: Zhang Yi <yi.zhang@huawei.com>
Cc: Hou Tao <houtao1@huawei.com>
Cc: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>