> Hint: readahead via the page cache is the way to go...
That's not the problem: knfsd has always done readahead via the page
cache.
The problem is that the generic_file_read() stuff caches info about
the status of readaheads in the 'struct file', which is a volatile
object in NFS (it is released after the NFS request has been serviced).
To work around that, knfsd copies that readahead status data from the
struct file into a private cache, and trots it out again upon the next
READ call to reference that particular inode.
Cheers,
Trond
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/