[Ocfs2-devel] [PATCH 1/1] OCFS2: add nlink check in ocfs2_inode_revalidate()

[Ocfs2-devel] [PATCH 1/1] OCFS2: add nlink check in ocfs2_inode_revalidate()

[wangang wang]

nlink should be also checked in ocfs2_inode_revalidate().
before setting flag OCFS2_INODE_DELETED ip_flags (between 
unlink and delete vote), the nlink may be 0.

the patch is against 1.4 git. 
1.2 svn should has the same patch with different line number.

Signed-off-by: Wengang wang <wen.gang.wang@oracle.com>
--
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index 591e693..6b5a83e 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -1186,6 +1186,13 @@ int ocfs2_inode_revalidate(struct dentry *dentry)
 		status = -ENOENT;
 		goto bail;
 	}
+
+	if (!inode->i_nlink) {
+		spin_unlock(&OCFS2_I(inode)->ip_lock);
+		mlog(0, "inode deleted!\n");
+		status = -ENOENT;
+		goto bail;
+	}
 	spin_unlock(&OCFS2_I(inode)->ip_lock);
 
 	/* Let ocfs2_inode_lock do the work of updating our struct

[Ocfs2-devel] [PATCH 1/1] OCFS2: add nlink check in ocfs2_inode_revalidate()

[Sunil Mushran]

NAK

There is no voting in ocfs2 1.4.

Secondly, i_nlink=0 does not mean inode is deleted. It simply
means it has no more links. So returning enoent for a valid
inode would be incorrect.

wangang wang wrote:
> nlink should be also checked in ocfs2_inode_revalidate().
> before setting flag OCFS2_INODE_DELETED ip_flags (between 
> unlink and delete vote), the nlink may be 0.
>
> the patch is against 1.4 git. 
> 1.2 svn should has the same patch with different line number.
>
> Signed-off-by: Wengang wang <wen.gang.wang@oracle.com>
> --
> diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
> index 591e693..6b5a83e 100644
> --- a/fs/ocfs2/inode.c
> +++ b/fs/ocfs2/inode.c
> @@ -1186,6 +1186,13 @@ int ocfs2_inode_revalidate(struct dentry *dentry)
>  		status = -ENOENT;
>  		goto bail;
>  	}
> +
> +	if (!inode->i_nlink) {
> +		spin_unlock(&OCFS2_I(inode)->ip_lock);
> +		mlog(0, "inode deleted!\n");
> +		status = -ENOENT;
> +		goto bail;
> +	}
>  	spin_unlock(&OCFS2_I(inode)->ip_lock);
>  
>  	/* Let ocfs2_inode_lock do the work of updating our struct
>
> _______________________________________________
> Ocfs2-devel mailing list
> Ocfs2-devel at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-devel
>   

[Ocfs2-devel] [PATCH 1/1] OCFS2: add nlink check in ocfs2_inode_revalidate()

[wengang wang]

Sunil,

Sunil Mushran wrote:
> NAK
>
> There is no voting in ocfs2 1.4.
>
Agree.
This patch has nothing to do with voting. and it helps nothing on bug 
7029797.
I found these problems when testing with OCFS2/NFS.
> Secondly, i_nlink=0 does not mean inode is deleted. It simply
> means it has no more links. So returning enoent for a valid
> inode would be incorrect.
ha, as my opinion, i_nlink changing from non-zero to zero, the 
inode/file is deleted. maybe it's wrong :)

thanks,
wengang.
>
> wangang wang wrote:
>> nlink should be also checked in ocfs2_inode_revalidate().
>> before setting flag OCFS2_INODE_DELETED ip_flags (between unlink and 
>> delete vote), the nlink may be 0.
>>
>> the patch is against 1.4 git. 1.2 svn should has the same patch with 
>> different line number.
>>
>> Signed-off-by: Wengang wang <wen.gang.wang@oracle.com>
>> -- 
>> diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
>> index 591e693..6b5a83e 100644
>> --- a/fs/ocfs2/inode.c
>> +++ b/fs/ocfs2/inode.c
>> @@ -1186,6 +1186,13 @@ int ocfs2_inode_revalidate(struct dentry *dentry)
>>          status = -ENOENT;
>>          goto bail;
>>      }
>> +
>> +    if (!inode->i_nlink) {
>> +        spin_unlock(&OCFS2_I(inode)->ip_lock);
>> +        mlog(0, "inode deleted!\n");
>> +        status = -ENOENT;
>> +        goto bail;
>> +    }
>>      spin_unlock(&OCFS2_I(inode)->ip_lock);
>>  
>>      /* Let ocfs2_inode_lock do the work of updating our struct
>>
>> _______________________________________________
>> Ocfs2-devel mailing list
>> Ocfs2-devel at oss.oracle.com
>> http://oss.oracle.com/mailman/listinfo/ocfs2-devel