svn commit: r248196 - head/sys/nfs

[Gleb Smirnoff]

On Wed, Mar 13, 2013 at 04:37:36PM +0100, Andre Oppermann wrote:
A> > A> > Let me repeat: there is a lot of code, that does handmade allocation of
A> > A> > an mbuf chain of an arbitrary length using mbufs and common clusters. This
A> > A> > code can be cut and use m_getm2(), if we can restrict it to avoid page size
A> > A> > clusters. I don't have time to dig deeper in the code and analyze and test
A> > A> > whether it can support page sized clusters in chains.
A> > A>
A> > A> m_getm2() can be used in any such case and doesn't have to avoid PAGE_SIZE
A> > A> clusters.  PAGE_SIZE jumbo clusters are fine.  Larger than PAGE_SIZE is not.
A> > A>
A> > A> When the code is able to work with mbuf chains the exact size of each cluster
A> > A> isn't important anymore.  We can use the optimal size.
A> >
A> > Do you volunteer to fix if anything is broken? If not, I'd prefer to
A> > be conservative and make m_getm() capable to produce chains without
A> > clusters bigger than MCLBYTES.
A> 
A> There is nothing to fix.  It works just fine and in fact is in use for all
A> socket i/o (write) for about 6 years since r163916 and 163916.  :-)
A> 
A> Yes, if there is anything broken with PAGE_SIZE mbuf chains returned by m_getm2()
A> I'm volunteering to fix it.

I caught your tongue :) Be prepared to study netipsec code :)

-- 
Totus tuus, Glebius.