From: Phillip Lougher <phillip@squashfs.org.uk>
To: Xiaoming Ni <nixiaoming@huawei.com>, linux-kernel@vger.kernel.org
Cc: wangle6@huawei.com, yi.zhang@huawei.com, wangbing6@huawei.com,
zhongjubin@huawei.com, chenjianguo3@huawei.com
Subject: Re: [PATCH v4 1/2] squashfs: add the mount parameter theads=<single|multi|percpu>
Date: Wed, 28 Sep 2022 03:20:56 +0100 [thread overview]
Message-ID: <b3018490-b951-01f0-2b86-84049d9470a0@squashfs.org.uk> (raw)
In-Reply-To: <20220916083604.33408-2-nixiaoming@huawei.com>
On 16/09/2022 09:36, Xiaoming Ni wrote:
> 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.
>
> Signed-off-by: Xiaoming Ni <nixiaoming@huawei.com>
>
> +#ifdef CONFIG_SQUASHFS_DECOMP_SINGLE
> + opts->thread_ops = &squashfs_decompressor_single;
> +#elif CONFIG_SQUASHFS_DECOMP_MULTI
> + opts->thread_ops = &squashfs_decompressor_multi;
> +#elif CONFIG_SQUASHFS_DECOMP_MULTI_PERCPU
In my previous review I asked you to fix the above two #elif
lines. You have done so in this patch series, but, only in the
second patch which seems perverse.
The reason why this isn't a good approach. If you *only* apply this
patch, with the following Squashfs configuration options
phillip@phoenix:/external/stripe/linux$ grep SQUASHFS .config
CONFIG_SQUASHFS=y
# CONFIG_SQUASHFS_FILE_CACHE is not set
CONFIG_SQUASHFS_FILE_DIRECT=y
CONFIG_SQUASHFS_DECOMP_MULTI_PERCPU=y
# CONFIG_SQUASHFS_CHOICE_DECOMP_BY_MOUNT is not set
# CONFIG_SQUASHFS_COMPILE_DECOMP_SINGLE is not set
# CONFIG_SQUASHFS_COMPILE_DECOMP_MULTI is not set
CONFIG_SQUASHFS_COMPILE_DECOMP_MULTI_PERCPU=y
CONFIG_SQUASHFS_XATTR=y
CONFIG_SQUASHFS_ZLIB=y
CONFIG_SQUASHFS_LZ4=y
CONFIG_SQUASHFS_LZO=y
CONFIG_SQUASHFS_XZ=y
CONFIG_SQUASHFS_ZSTD=y
# CONFIG_SQUASHFS_4K_DEVBLK_SIZE is not set
# CONFIG_SQUASHFS_EMBEDDED is not set
CONFIG_SQUASHFS_FRAGMENT_CACHE_SIZE=3
You will get the following build warning
phillip@phoenix:/external/stripe/linux$ make bzImage
SYNC include/config/auto.conf.cmd
CALL scripts/checksyscalls.sh
CALL scripts/atomic/check-atomics.sh
DESCEND objtool
CHK include/generated/compile.h
UPD kernel/config_data
GZIP kernel/config_data.gz
CC kernel/configs.o
AR kernel/built-in.a
CC fs/squashfs/block.o
CC fs/squashfs/cache.o
CC fs/squashfs/dir.o
CC fs/squashfs/export.o
CC fs/squashfs/file.o
CC fs/squashfs/fragment.o
CC fs/squashfs/id.o
CC fs/squashfs/inode.o
CC fs/squashfs/namei.o
CC fs/squashfs/super.o
fs/squashfs/super.c: In function ‘squashfs_init_fs_context’:
fs/squashfs/super.c:492:7: warning: "CONFIG_SQUASHFS_DECOMP_MULTI" is
not defined, evaluates to 0 [-Wundef]
492 | #elif CONFIG_SQUASHFS_DECOMP_MULTI
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
This strictly breaks the git bisectability rule. Every patch should
be compilable and not break the build or produce warnings. If not
it makes git bisect difficult to use to find regressions.
This can be avoided by fixing the issue in *this* patch. So
please do so.
Thanks
Phillip
> + opts->thread_ops = &squashfs_decompressor_percpu;
> +#endif
> fc->fs_private = opts;
> fc->ops = &squashfs_context_ops;
> return 0;
> @@ -478,7 +526,7 @@ static void squashfs_put_super(struct super_block *sb)
> squashfs_cache_delete(sbi->block_cache);
> squashfs_cache_delete(sbi->fragment_cache);
> squashfs_cache_delete(sbi->read_page);
> - squashfs_decompressor_destroy(sbi);
> + sbi->thread_ops->destroy(sbi);
> kfree(sbi->id_table);
> kfree(sbi->fragment_index);
> kfree(sbi->meta_index);
next prev parent reply other threads:[~2022-09-28 2:21 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-08-15 3:10 [PATCH 0/2] squashfs: Add the mount parameter "threads=" Xiaoming Ni
2022-08-15 3:10 ` [PATCH 1/2] squashfs: add the mount parameter theads=<single|multi|percpu> Xiaoming Ni
2022-08-15 11:19 ` kernel test robot
2022-08-15 3:11 ` [PATCH 2/2] squashfs: Allows users to configure the number of decompression threads Xiaoming Ni
2022-08-16 1:00 ` [PATCH v2 0/2] squashfs: Add the mount parameter "threads=" Xiaoming Ni
2022-08-16 1:00 ` [PATCH v2 1/2] squashfs: add the mount parameter theads=<single|multi|percpu> Xiaoming Ni
2022-08-16 1:00 ` [PATCH v2 2/2] squashfs: Allows users to configure the number of decompression threads Xiaoming Ni
2022-08-26 6:19 ` ping //Re: [PATCH v2 0/2] squashfs: Add the mount parameter "threads=" Xiaoming Ni
2022-08-28 23:18 ` Phillip Lougher
2022-08-30 13:38 ` Xiaoming Ni
2022-08-30 18:08 ` Phillip Lougher
2022-08-30 18:34 ` Phillip Lougher
2022-08-31 1:09 ` Xiaoming Ni
2022-09-02 9:48 ` [PATCH v3 " Xiaoming Ni
2022-09-02 9:48 ` [PATCH v3 1/2] squashfs: add the mount parameter theads=<single|multi|percpu> Xiaoming Ni
2022-09-09 15:44 ` Phillip Lougher
2022-09-13 2:46 ` Xiaoming Ni
2022-09-09 15:50 ` Phillip Lougher
2022-09-13 2:47 ` Xiaoming Ni
2022-09-02 9:48 ` [PATCH v3 2/2] squashfs: Allows users to configure the number of decompression threads Xiaoming Ni
2022-09-09 15:26 ` [PATCH v3 0/2] squashfs: Add the mount parameter "threads=" Phillip Lougher
2022-09-16 8:36 ` [PATCH v4 " Xiaoming Ni
2022-09-16 8:36 ` [PATCH v4 1/2] squashfs: add the mount parameter theads=<single|multi|percpu> Xiaoming Ni
2022-09-28 2:20 ` Phillip Lougher [this message]
2022-09-28 3:06 ` Xiaoming Ni
2022-09-16 8:36 ` [PATCH v4 2/2] squashfs: Allows users to configure the number of decompression threads Xiaoming Ni
2022-09-27 1:05 ` ping //Re: [PATCH v4 0/2] squashfs: Add the mount parameter "threads=" Xiaoming Ni
2022-09-30 9:14 ` [PATCH v5 " Xiaoming Ni
2022-09-30 9:14 ` [PATCH v5 1/2] squashfs: add the mount parameter theads=<single|multi|percpu> Xiaoming Ni
2022-10-17 21:58 ` Re " Phillip Lougher
2022-09-30 9:14 ` [PATCH v5 2/2] squashfs: Allows users to configure the number of decompression threads Xiaoming Ni
2022-10-17 0:57 ` ping// Re: [PATCH v5 0/2] squashfs: Add the mount parameter "threads=" Xiaoming Ni
2022-10-17 22:59 ` Phillip Lougher
2022-10-18 6:24 ` Xiaoming Ni
2022-10-19 3:09 ` [PATCH v6 " Xiaoming Ni
2022-10-19 3:09 ` [PATCH v6 1/2] squashfs: add the mount parameter theads=<single|multi|percpu> Xiaoming Ni
2022-10-19 3:09 ` [PATCH v6 2/2] squashfs: Allows users to configure the number of decompression threads Xiaoming Ni
2022-10-27 22:44 ` [PATCH v6 0/2] squashfs: Add the mount parameter "threads=" Phillip Lougher
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=b3018490-b951-01f0-2b86-84049d9470a0@squashfs.org.uk \
--to=phillip@squashfs.org.uk \
--cc=chenjianguo3@huawei.com \
--cc=linux-kernel@vger.kernel.org \
--cc=nixiaoming@huawei.com \
--cc=wangbing6@huawei.com \
--cc=wangle6@huawei.com \
--cc=yi.zhang@huawei.com \
--cc=zhongjubin@huawei.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®