Re: [NFS] [PATCH] NFS server does not update mtime on setattr request

[Peter Staubach <staubach@redhat.com>]

Trond Myklebust wrote:

>On Wed, 2006-06-07 at 11:26 -0400, Peter Staubach wrote:
>  
>
>>J. Bruce Fields wrote:
>>
>>    
>>
>>>On Wed, Jun 07, 2006 at 10:44:50AM -0400, Peter Staubach wrote:
>>> 
>>>
>>>      
>>>
>>>>I saw that wording too and assumed what I think that you assumed.  I
>>>>assumed that that meant that if the new size is equal to the old size,
>>>>then nothing should be changed.  However, that does not seem to be how
>>>>those words are to be interpreted.  They are to be interpreted as "if
>>>>the new length of the file can be successfully set, then the
>>>>mtime/ctime should be changed".
>>>>   
>>>>
>>>>        
>>>>
>>>What's the basis for that interpretation?  The language seems extremely
>>>clear:
>>>
>>>	"On successful completion, if the file size is changed, these
>>>	functions will mark for update the st_ctime and st_mtime fields
>>>	of the file, and if the file is a regular file, the S_ISUID and
>>>	S_ISGID bits of the file mode may be cleared."
>>>
>>>Why are you concerned about this?  Do you have an actual application
>>>that breaks?
>>>
>>>      
>>>
>>Yes, there is a customer who is quite unhappy that the semantics over Linux
>>client NFS are different than those of BSD, Solaris, and local file system
>>access on Linux itself.  The basis for my work is based on a bugzilla from
>>this customer.
>>
>>My interpretation is based on looking at the local behavior on Linux, which
>>changes mtime/ctime even if the file size does not change, and SunOS, which
>>changes mtime/ctime even if the file size does not change and is very
>>heavily SUSv3 compliant.
>>
>>In this case, "changed" does not mean "made different".  It simply means
>>that the file size is set to the new value.
>>
>>I would have chosen different words or a different interpretation too,
>>but all of the evidence suggests that the semantics are as I stated.
>>    
>>
>
>We've already fixed this to be SuSv3 compliant for both create and
>truncate. Your "safe" suggestion would break truncate again. That is why
>it is being vetoed.
>

You are speaking of the client side changes?  I don't have a problem with
not taking them.

I am curious about how this would break truncate?

    Thanx...

       ps