Re: nfsd: supports read buffer from multiples pages

["J. Bruce Fields" <bfields@fieldses.org>]

On Mon, Feb 01, 2016 at 01:38:05PM -0500, bfields wrote:
> On Sun, Jan 31, 2016 at 08:50:10PM +0800, Kinglong Mee wrote:
> > ltp fsync02 will cause nfs sending LAYOUTCOMMIT with length
> > larger than two pages. nfsd returns NFSERR_BAD_XDR right now.
> 
> This is with the xfs block layout?
> 
> Christoph, do we know anything about average or worst-case sizes for
> that layout update field?
> 
> > This patch lets nfsd supports read buffer from multiples pages.
> 
> Hm.  We'll end up kmalloc()ing the passed-in field length:
> 
> 		p = argp->tmpp = kmalloc(nbytes, GFP_KERNEL);
> 
> We still do still have that (avail + argp->pagelen) limit, so we're not
> going to pass arbitrarily large nbytes straight from the network to
> kmalloc.  But we do try to avoid depending on higher-order allocations.

(Which it looks like we were allowing before, possibly by accident.  But
I doubt they were actually happening in practice, so that's not evidence
that we don't need to worry about allocations greater than a page.)

--b.