linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jan Blunck <jblunck@suse.de>
To: Balbir Singh <balbir@in.ibm.com>
Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	akpm@osdl.org, viro@zeniv.linux.org.uk, dgc@sgi.com,
	neilb@suse.de
Subject: Re: [PATCH 2/5] vfs: d_genocide() doesnt add dentries to unused list
Date: Mon, 19 Jun 2006 11:22:49 +0200	[thread overview]
Message-ID: <20060619092249.GB6824@hasse.suse.de> (raw)
In-Reply-To: <4495AABE.6090007@in.ibm.com>

On Mon, Jun 19, Balbir Singh wrote:

> > 			this_parent = dentry;
> > 			goto repeat;
> > 		}
> >-		atomic_dec(&dentry->d_count);
> >+		if (!list_empty(&dentry->d_lru)) {
> >+			dentry_stat.nr_unused--;
> >+			list_del_init(&dentry->d_lru);
> >+		}
> >+		if (atomic_dec_and_test(&dentry->d_count)) {
> >+			list_add(&dentry->d_lru, dentry_unused.prev);
> >+			dentry_stat.nr_unused++;
> >+		}
> 
> We could have dentries on the LRU list with non-zero d_count. If
> we have a dentry on the LRU list with a count of 1, then the code
> will remove it from LRU list and then add it back subsequently.
> 

So you think this is better?

   if (atomic_dec_and_test(&dentry->d_count)) {
      if (!list_empty(&dentry_d_lru))
         list_move_tail(&dentry->d_lru, dentry_unused);
   } else
      if (!list_empty(&dentry->d_lru)) {
         dentry_stat.nr_unused--;
         list_del_init(&dentry->d_lru);
      }


> I think the condition below should be an else if
> 

No. We always lower the reference count in d_genocide.

> 
> d_genocide() now almost looks like select_parent(). I think we can share a 
> lot
> of code between the two.
> 

Hmm, interesting idea. This would save the dentry-tree walking code in
have_submounts too. Maybe something like this:

+static int select_parent_walker(struct dentry * dentry, int * found)
+{
+       if (!list_empty(&dentry->d_lru)) {
+               dentry_stat.nr_unused--;
+               list_del_init(&dentry->d_lru);
+       }
+
+       /*
+        * move only zero ref count dentries to the end
+        * of the unused list for prune_dcache
+        */
+       if (!atomic_read(&dentry->d_count)) {
+               list_add(&dentry->d_lru, dentry_unused.prev);
+               dentry_stat.nr_unused++;
+               *found++;
+       }
+
+       /*
+        * We can return to the caller if we have found some (this
+        * ensures forward progress). We'll be coming back to find
+        * the rest.
+        */
+       if (*found && need_resched())
+               return -1;
+
+       return 0;
+}
+
+typedef int (*walker_t)(struct dentry * dentry, int * return);
+
+static int dentry_tree_walk(struct dentry * parent, walker_t walker)
+{
+       struct dentry *this_parent = parent;
+       struct list_head *next;
+       int ret = 0;
+
+       spin_lock(&dcache_lock);
+repeat:
+       next = this_parent->d_subdirs.next;
+resume:
+       while (next != &this_parent->d_subdirs) {
+               struct list_head *tmp = next;
+               struct dentry *dentry = list_entry(tmp, struct dentry,
+                                                  d_u.d_child);
+               next = tmp->next;
+
+               if (walker(dentry, &ret))
+                       goto out;
+
+               /*
+                * Descend a level if the d_subdirs list is non-empty.
+                */
+               if (!list_empty(&dentry->d_subdirs)) {
+                       this_parent = dentry;
+                       goto repeat;
+               }
+       }
+       /*
+        * All done at this level ... ascend and resume the search.
+        */
+       if (this_parent != parent) {
+               next = this_parent->d_u.d_child.next;
+               this_parent = this_parent->d_parent;
+               goto resume;
+       }
+out:
+       spin_unlock(&dcache_lock);
+       return ret;
+}

  reply	other threads:[~2006-06-19  9:22 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-06-16 10:43 [PATCH 0/5] vfs: per-superblock unused dentries list (3rd version) jblunck
2006-06-16 10:43 ` [PATCH 1/5] vfs: remove whitespace noise from fs/dcache.c jblunck
2006-06-16 10:43 ` [PATCH 2/5] vfs: d_genocide() doesnt add dentries to unused list jblunck
2006-06-18 19:34   ` Balbir Singh
2006-06-19  9:22     ` Jan Blunck [this message]
2006-06-19 10:38       ` Balbir Singh
2006-06-16 10:43 ` [PATCH 3/5] vfs: remove shrink_dcache_anon() jblunck
2006-06-16 10:43 ` [PATCH 4/5] vfs: per superblock dentry stats jblunck
2006-06-16 10:43 ` [PATCH 5/5] vfs: per superblock dentry unused list jblunck
  -- strict thread matches above, loose matches on Subject: below --
2006-06-01  9:51 [patch 0/5] [PATCH,RFC] vfs: per-superblock unused dentries list (2nd version) jblunck
2006-06-01  9:51 ` [patch 2/5] vfs: d_genocide() doesnt add dentries to unused list jblunck

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=20060619092249.GB6824@hasse.suse.de \
    --to=jblunck@suse.de \
    --cc=akpm@osdl.org \
    --cc=balbir@in.ibm.com \
    --cc=dgc@sgi.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=neilb@suse.de \
    --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;
as well as URLs for NNTP newsgroup(s).