> I suspect the correct solution is to mark attrs untrusted in > mdc_unreachable().
That doesn't help the entry being renamed though, which is actually the biggest concern. Frank > On Wed, Oct 12, 2016 at 6:29 PM, Frank Filz <ffilz...@mindspring.com> > wrote: > > Dan, > > > > In cache_inode_rename, we did a lookup of both the old name and the > > new name, and refreshed attributes on both objects. > > > > In mdcache, I'm not sure we invalidate the attributes of > > mdc_lookup_dst (if any, the object that is unlinked as a result of rename > onto it). > > > > What this results in is that a rename of a file causes us to not > > reflect the updated ctime to the client. > > > > I think we need to do an mdc_try_get_cached on old name, and if that > > results in a cached object, to invalidate it's attributes. > > > > Similarly, if mdc_lookup_dst is non-NULL, we need to invalidate it's > > attributes (if it had multiple hard links, a rename over it should > > change it's ctime also). If we just evict it from cache, then that works > > also. > > > > What do you think? > > > > Anyone else have thoughts? > > > > Thanks > > > > Frank > > > > > > --- > > This email has been checked for viruses by Avast antivirus software. > > https://www.avast.com/antivirus > > --- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot _______________________________________________ Nfs-ganesha-devel mailing list Nfs-ganesha-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nfs-ganesha-devel
