[linux-cifs-client] Re: fsx-linux failing with latest cifs-2.6 git tree

[Jeff Layton]

On Fri, 21 Nov 2008 19:04:04 -0600
"Steve French" <smfrench at gmail.com> wrote:

> Don't we also have to add the check for
> if (mapping->host->clientCanCacheRead) before we decide to skip doing
> the cifs_readpage_worker ... otherwise we aren't 100% sure of the file
> size on the server (although it is probably right since we just did
> revalidate)
> 
> so something like
> 
>  if (mapping->host->clientCanCacheRead && (page_start >= i_size ||
> (offset == 0 && (pos + len) >= i_size)))
> 

That certainly would be safer. I'll plan to add that in once I'm more
clear on the page flag handling here...

-- 
Jeff Layton <jlayton at redhat.com>