OpenVZ Forum


Home » Mailing lists » Devel » [Q] missing unused dentry in prune_dcache()?
Re: [Q] missing unused dentry in prune_dcache()? [message #7837 is a reply to message #7815] Fri, 27 October 2006 06:32 Go to previous messageGo to previous message
vaverin is currently offline  vaverin
Messages: 708
Registered: September 2005
Senior Member
David Howells wrote:
> Vasily Averin <vvs@sw.ru> wrote:
>
>> I've noticed one more minor issue in your patch: in
>> shrink_dcache_for_umount_subtree() function you decrement
>> dentry_stat.nr_dentry without dcache_lock.
>
> How about the attached patch?
I'm sorry, but your patch is wrong:
you have mixed calculation of 2 variables:
dentry_stat.nr_unused -- were correct, it was decremented under dcache_lock.
dentry_stat.nr_dentry -- were incorrect, it was decremented without dcache_lock.

You should correct dentry_stat.nr_dentry, but instead you broke calculation of
dentry_stat.nr_unused.

I've fixed this issue by following patch.

Thank you,
Vasily Averin
---
VFS: Fix an error in dentry_stat.nr_dentry counting

From: Vasily Averin <vvs@sw.ru>

Fix an error in dentry_stat.nr_dentry counting in
shrink_dcache_for_umount_subtree() in which the count is modified without the
dcache_lock held.

Signed-Off-By: Vasily Averin <vvs@sw.ru>

--- linux-2.6.19-rc3/fs/dcache.c.nrdntr 2006-10-26 15:14:51.000000000 +0400
+++ linux-2.6.19-rc3/fs/dcache.c 2006-10-27 10:24:07.000000000 +0400
@@ -554,6 +554,7 @@ repeat:
static void shrink_dcache_for_umount_subtree(struct dentry *dentry)
{
struct dentry *parent;
+ unsigned detached = 0;

BUG_ON(!IS_ROOT(dentry));

@@ -618,7 +619,7 @@ static void shrink_dcache_for_umount_sub
atomic_dec(&parent->d_count);

list_del(&dentry->d_u.d_child);
- dentry_stat.nr_dentry--; /* For d_free, below */
+ detached++;

inode = dentry->d_inode;
if (inode) {
@@ -635,8 +636,8 @@ static void shrink_dcache_for_umount_sub
/* finished when we fall off the top of the tree,
* otherwise we ascend to the parent and move to the
* next sibling if there is one */
- if (!parent)
- return;
+ if (!parent)
+ goto out;

dentry = parent;

@@ -645,6 +646,11 @@ static void shrink_dcache_for_umount_sub
dentry = list_entry(dentry->d_subdirs.next,
struct dentry, d_u.d_child);
}
+out:
+ /* several dentries were freed, need to correct nr_dentry */
+ spin_lock(&dcache_lock);
+ dentry_stat.nr_dentry -= detached;
+ spin_unlock(&dcache_lock);
}

/*

Report message to a moderator

[Message index]
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: [PATCH]: OOM can panic due to processes stuck in __alloc_pages()
Next Topic: [PATCH] move_task_off_dead_cpu() should be called with disabled ints
Goto Forum:
  

[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Sat Jul 05 09:19:34 GMT 2025

Total time taken to generate the page: 0.03319 seconds
Powered by FUDforum Powered by Virtuozzo Containers