mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Jia Zhu <zhujia.zj@bytedance.com>
To: Jingbo Xu <jefflexu@linux.alibaba.com>,
	xiang@kernel.org, chao@kernel.org, linux-erofs@lists.ozlabs.org,
	houtao1@huawei.com
Cc: huyue2@coolpad.com, linux-kernel@vger.kernel.org,
	zhujia.zj@bytedance.com
Subject: Re: [PATCH 4/4] erofs: simplify the name collision checking in share domain mode
Date: Thu, 9 Feb 2023 10:43:28 +0800	[thread overview]
Message-ID: <ce11c1b5-b83e-5c5d-5e03-ea3278059efb@bytedance.com> (raw)
In-Reply-To: <489d47915bcc71ef3c3ec7b8a437ec6e09c4c3db.1675840368.git.jefflexu@linux.alibaba.com>



在 2023/2/8 15:17, Jingbo Xu 写道:
> When share domain is enabled, data blobs can be shared among filesystem
> instances in the same domain for data deduplication, while bootstrap
> blobs are always dedicated to the corresponding filesystem instance, and
> have no need to be shared.
> 
> In the initial implementation of share domain (commit 7d41963759fe
> ("erofs: Support sharing cookies in the same domain")), bootstrap blobs
> are also in the share domain, and thus can be referenced by the
> following filesystem instances.  In this case, mounting twice with the
> same fsid and domain_id will trigger warning in sysfs.  Commit
> 27f2a2dcc626 ("erofs: check the uniqueness of fsid in shared domain in
> advance") fixes this by introducing the name collision checking.
> 
> This patch attempts to fix the above issue in another simpler way.
> Since the bootstrap blobs have no need to be shared, move them out of
> the share domain, so that one bootstrap blob can not be referenced by
> other filesystem instances.  Attempt to mount twice with the same fsid
> and domain_id will fail with info of duplicate cookies, which is
> consistent with the behavior in non-share-domain mode.
> 
> Signed-off-by: Jingbo Xu <jefflexu@linux.alibaba.com>

Reviewed-by: Jia Zhu <zhujia.zj@bytedance.com>

> ---
>   fs/erofs/fscache.c  | 36 ++++++++----------------------------
>   fs/erofs/internal.h |  4 ++--
>   fs/erofs/super.c    |  2 +-
>   3 files changed, 11 insertions(+), 31 deletions(-)
> 
> diff --git a/fs/erofs/fscache.c b/fs/erofs/fscache.c
> index 8353a5fe8c71..8da6e05e9d23 100644
> --- a/fs/erofs/fscache.c
> +++ b/fs/erofs/fscache.c
> @@ -505,25 +505,18 @@ struct erofs_fscache *erofs_fscache_domain_init_cookie(struct super_block *sb,
>   
>   static
>   struct erofs_fscache *erofs_domain_register_cookie(struct super_block *sb,
> -						   char *name,
> -						   unsigned int flags)
> +						   char *name)
>   {
>   	struct erofs_fscache *ctx;
>   	struct erofs_domain *domain = EROFS_SB(sb)->domain;
>   
>   	mutex_lock(&erofs_domain_cookies_lock);
>   	list_for_each_entry(ctx, &erofs_domain_cookies_list, node) {
> -		if (ctx->domain != domain || strcmp(ctx->name, name))
> -			continue;
> -		if (!(flags & EROFS_REG_COOKIE_NEED_NOEXIST)) {
> +		if (ctx->domain == domain && !strcmp(ctx->name, name)) {
>   			refcount_inc(&ctx->ref);
> -		} else {
> -			erofs_err(sb, "%s already exists in domain %s", name,
> -				  domain->domain_id);
> -			ctx = ERR_PTR(-EEXIST);
> +			mutex_unlock(&erofs_domain_cookies_lock);
> +			return ctx;
>   		}
> -		mutex_unlock(&erofs_domain_cookies_lock);
> -		return ctx;
>   	}
>   	ctx = erofs_fscache_domain_init_cookie(sb, name);
>   	mutex_unlock(&erofs_domain_cookies_lock);
> @@ -531,11 +524,10 @@ struct erofs_fscache *erofs_domain_register_cookie(struct super_block *sb,
>   }
>   
>   struct erofs_fscache *erofs_fscache_register_cookie(struct super_block *sb,
> -						    char *name,
> -						    unsigned int flags)
> +						    char *name)
>   {
>   	if (EROFS_SB(sb)->domain_id)
> -		return erofs_domain_register_cookie(sb, name, flags);
> +		return erofs_domain_register_cookie(sb, name);
>   	return erofs_fscache_acquire_cookie(sb, sb, name);
>   }
>   
> @@ -564,7 +556,6 @@ int erofs_fscache_register_fs(struct super_block *sb)
>   	int ret;
>   	struct erofs_sb_info *sbi = EROFS_SB(sb);
>   	struct erofs_fscache *fscache;
> -	unsigned int flags = 0;
>   
>   	if (sbi->domain_id)
>   		ret = erofs_fscache_register_domain(sb);
> @@ -573,19 +564,8 @@ int erofs_fscache_register_fs(struct super_block *sb)
>   	if (ret)
>   		return ret;
>   
> -	/*
> -	 * When shared domain is enabled, using NEED_NOEXIST to guarantee
> -	 * the primary data blob (aka fsid) is unique in the shared domain.
> -	 *
> -	 * For non-shared-domain case, fscache_acquire_volume() invoked by
> -	 * erofs_fscache_register_volume() has already guaranteed
> -	 * the uniqueness of primary data blob.
> -	 *
> -	 * Acquired domain/volume will be relinquished in kill_sb() on error.
> -	 */
> -	if (sbi->domain_id)
> -		flags |= EROFS_REG_COOKIE_NEED_NOEXIST;
> -	fscache = erofs_fscache_register_cookie(sb, sbi->fsid, flags);
> +	/* acquired domain/volume will be relinquished in kill_sb() on error */
> +	fscache = erofs_fscache_acquire_cookie(sb, sb, sbi->fsid);
>   	if (IS_ERR(fscache))
>   		return PTR_ERR(fscache);
>   
> diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
> index 125e4aa8d295..fded736bc3d5 100644
> --- a/fs/erofs/internal.h
> +++ b/fs/erofs/internal.h
> @@ -570,7 +570,7 @@ int erofs_fscache_register_fs(struct super_block *sb);
>   void erofs_fscache_unregister_fs(struct super_block *sb);
>   
>   struct erofs_fscache *erofs_fscache_register_cookie(struct super_block *sb,
> -					char *name, unsigned int flags);
> +						    char *name);
>   void erofs_fscache_unregister_cookie(struct erofs_fscache *fscache);
>   #else
>   static inline int erofs_fscache_register_fs(struct super_block *sb)
> @@ -581,7 +581,7 @@ static inline void erofs_fscache_unregister_fs(struct super_block *sb) {}
>   
>   static inline
>   struct erofs_fscache *erofs_fscache_register_cookie(struct super_block *sb,
> -					char *name, unsigned int flags)
> +						    char *name)
>   {
>   	return ERR_PTR(-EOPNOTSUPP);
>   }
> diff --git a/fs/erofs/super.c b/fs/erofs/super.c
> index 19b1ae79cec4..8706ca34f26a 100644
> --- a/fs/erofs/super.c
> +++ b/fs/erofs/super.c
> @@ -244,7 +244,7 @@ static int erofs_init_device(struct erofs_buf *buf, struct super_block *sb,
>   	}
>   
>   	if (erofs_is_fscache_mode(sb)) {
> -		fscache = erofs_fscache_register_cookie(sb, dif->path, 0);
> +		fscache = erofs_fscache_register_cookie(sb, dif->path);
>   		if (IS_ERR(fscache))
>   			return PTR_ERR(fscache);
>   		dif->fscache = fscache;

  reply	other threads:[~2023-02-09  2:48 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-02-08  7:15 [PATCH 0/4] erofs: cleanup for fscache " Jingbo Xu
2023-02-08  7:16 ` [PATCH 1/4] erofs: remove unused device mapping in meta routine Jingbo Xu
2023-02-08  7:21   ` [External] " Jia Zhu
2023-02-08  7:16 ` [PATCH 2/4] erofs: maintain cookies of share domain in self-contained list Jingbo Xu
2023-02-08  8:16   ` [External] " Jia Zhu
2023-02-08  7:16 ` [PATCH 3/4] erofs: unify anonymous inodes for blob Jingbo Xu
2023-02-08  9:01   ` [External] " Jia Zhu
2023-02-08  7:17 ` [PATCH 4/4] erofs: simplify the name collision checking in share domain mode Jingbo Xu
2023-02-09  2:43   ` Jia Zhu [this message]
2023-02-09  5:01   ` Jingbo Xu
2023-02-08  7:32 ` [PATCH] erofs: relinquish volume with mutex held Jingbo Xu
2023-02-08  9:09   ` [External] " Jia Zhu

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=ce11c1b5-b83e-5c5d-5e03-ea3278059efb@bytedance.com \
    --to=zhujia.zj@bytedance.com \
    --cc=chao@kernel.org \
    --cc=houtao1@huawei.com \
    --cc=huyue2@coolpad.com \
    --cc=jefflexu@linux.alibaba.com \
    --cc=linux-erofs@lists.ozlabs.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=xiang@kernel.org \
    /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®