Linux-mtd Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Yufen Yu <yuyufen@huawei.com>
To: <dwmw2@infradead.org>
Cc: <viro@zeniv.linux.org.uk>, <linux-mtd@lists.infradead.org>,
	<linux-kernel@vger.kernel.org>, Yufen Yu <yuyufen@huawei.com>
Subject: [PATCH] jffs2: remove fd from the f->dents list immediately.
Date: Fri, 16 Mar 2018 19:05:22 +0800	[thread overview]
Message-ID: <20180316110522.1120-1-yuyufen@huawei.com> (raw)

commit 15953580e79b ("[JFFS2] Improve getdents vs. f_pos handling on NOR flash.")
is introduced to resolve 'rm -r', which cannot remove all files:
    http://lists.infradead.org/pipermail/linux-mtd/2007-October/019658.html

However, it can cause the following issues:

1. 'deletion' dirents is alway in the f->dents list, wasting memory
    resource. For example:
        There is a file named 'file1'. Then we rename it:
        mv file1 file2;
        mv file2 file3;
        ...
        mv file99999 file1000000

        When CONFIG_JFFS2_SUMMARY is not set, file1~file1000000
        always in the f->dents list.

2. Since the list become longer and longer, more CPU time is used
    to traverse it.

After reverting the commit, we test 'rm -r', which can remove all
files, and all seems OK!

Signed-off-by: Yufen Yu <yuyufen@huawei.com>
---
 fs/jffs2/write.c | 33 +++++++++++++++++----------------
 1 file changed, 17 insertions(+), 16 deletions(-)

diff --git a/fs/jffs2/write.c b/fs/jffs2/write.c
index cda9a361368e..1deed35beb50 100644
--- a/fs/jffs2/write.c
+++ b/fs/jffs2/write.c
@@ -598,31 +598,32 @@ int jffs2_do_unlink(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f,
 		jffs2_add_fd_to_list(c, fd, &dir_f->dents);
 		mutex_unlock(&dir_f->sem);
 	} else {
+		struct jffs2_full_dirent **prev = &dir_f->dents;
 		uint32_t nhash = full_name_hash(NULL, name, namelen);
 
-		fd = dir_f->dents;
 		/* We don't actually want to reserve any space, but we do
 		   want to be holding the alloc_sem when we write to flash */
 		mutex_lock(&c->alloc_sem);
 		mutex_lock(&dir_f->sem);
 
-		for (fd = dir_f->dents; fd; fd = fd->next) {
-			if (fd->nhash == nhash &&
-			    !memcmp(fd->name, name, namelen) &&
-			    !fd->name[namelen]) {
-
-				jffs2_dbg(1, "Marking old dirent node (ino #%u) @%08x obsolete\n",
-					  fd->ino, ref_offset(fd->raw));
-				jffs2_mark_node_obsolete(c, fd->raw);
-				/* We don't want to remove it from the list immediately,
-				   because that screws up getdents()/seek() semantics even
-				   more than they're screwed already. Turn it into a
-				   node-less deletion dirent instead -- a placeholder */
-				fd->raw = NULL;
-				fd->ino = 0;
-				break;
+		while ((*prev) && (*prev)->nhash <= nhash) {
+			if ((*prev)->nhash == nhash &&
+				!memcmp((*prev)->name, name, namelen) &&
+				!(*prev)->name[namelen]) {
+
+					struct jffs2_full_dirent *this = *prev;
+
+					jffs2_dbg(1, "Marking old dirent node (ino #%u) @%08x obsolete\n",
+							this->ino, ref_offset(this->raw));
+					*prev = this->next;
+					jffs2_mark_node_obsolete(c, this->raw);
+					jffs2_free_full_dirent(this);
+
+					break;
 			}
+			prev = &((*prev)->next);
 		}
+
 		mutex_unlock(&dir_f->sem);
 	}
 
-- 
2.13.6

             reply	other threads:[~2018-03-16 10:56 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-16 11:05 Yufen Yu [this message]
2018-03-16 12:39 ` [PATCH] jffs2: remove fd from the f->dents list immediately Joakim Tjernlund
2018-03-16 12:52   ` David Woodhouse
2018-03-19  8:27     ` yuyufen

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=20180316110522.1120-1-yuyufen@huawei.com \
    --to=yuyufen@huawei.com \
    --cc=dwmw2@infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mtd@lists.infradead.org \
    --cc=viro@zeniv.linux.org.uk \
    /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