[Ocfs2-devel] [PATCH] ocfs2: Don't print error when listing too many xattrs.

Joel Becker Joel.Becker at oracle.com
Tue May 5 00:18:05 PDT 2009


On Tue, May 05, 2009 at 09:40:31AM +0800, Tao Ma wrote:
> Joel Becker wrote:
>> On Mon, May 04, 2009 at 05:18:09AM +0800, Tao Ma wrote:
>>> Currently, when listing xattrs, kernel define XATTR_LIST_MAX as 65536
>>> in include/linux/limits.h, so it can't handle too many xattrs.
>>>
>>> But with ocfs2 xattr tree, we actually have no limit for the number.
>>> And it will pollute the message with something like this when listing.
>>> (27738,0):ocfs2_iterate_xattr_buckets:3158 ERROR: status = -34
>>> (27738,0):ocfs2_xattr_tree_list_index_block:3264 ERROR: status = -34
>>>
>>> So don't print "ERROR" message as this is not an ocfs2 error.
>>
>> 	Hmm, but this will stop iteration of the xattrs right where we
>> hit the large one.  I don't think that's correct.  At the very least, we
>> should go on listing with the following xattrs.
>> 	Now, the xattr with a too-long name - should we skip it, or
>> should we list a truncated name?  That I'm not sure of.
> I think we should stop at where we are, return errors and let the user  
> space handle this.
>
> Just give you an example.
> 1. for((i=0;i<20000;i++))do setfattr -n "user.name$i" -v "aaaaa"  
> /mnt/ocfs2/a;done
> This will create a very long name list.
>
> 2. getfattr /mnt/ocfs2/a
> /mnt/ocfs2/a: Argument list too long
> So you see it says clearly that the name list is too long.
> and the same goes for listxattr(2).
> If the size of the list buffer is too small to hold the result, errno is  
> set to ERANGE.
>
> But if you go ahead and don't return errors. we will get a name list of  
> about ~5000 xattrs and no error is returned. So the user may wonder why  
> I set 20000 xattrs, but only get 5000 xattrs and look at this as a bug.

	I think I mis-read.  I was thinking you'd error if the name of
an xattr was too long, and thus I could never list the xattrs past that
name.  Instead, this is apparently that the entire list grew past an
arbitrary limit.  If the kernel is enforcing this limit, this is
probably the correct way to go.

Joel

-- 

Life's Little Instruction Book #267

	"Lie on your back and look at the stars."

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127



More information about the Ocfs2-devel mailing list