Date: Tue, 08 Jul 2014 16:19:52 +0200 From: Hans Petter Selasky <hps@selasky.org> To: pyunyh@gmail.com Cc: freebsd-net@freebsd.org, freebsd-current@FreeBSD.org Subject: Re: [RFC] Allow m_dup() to use JUMBO clusters Message-ID: <53BBFE08.1080804@selasky.org> In-Reply-To: <53BB7433.2010306@selasky.org> References: <53BA5657.8010309@selasky.org> <20140708021439.GA3965@michelle.fasterthan.com> <53BB7433.2010306@selasky.org>
next in thread | previous in thread | raw e-mail | index | archive | help
This is a multi-part message in MIME format.
--------------060305060509090406020506
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
>
> Hi,
>
> Would it be better if my patch used the PAGE_SIZE clusters instead of
> the 16K ones? Then it should not be affected by memory defragmentation.
> Thanks for shedding some light into this area?
>
> --HPS
>
Hi,
Updated patch attached.
--HPS
--------------060305060509090406020506
Content-Type: text/x-diff;
name="uipc_mbuf.c.diff"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename="uipc_mbuf.c.diff"
=== sys/kern/uipc_mbuf.c
==================================================================
--- sys/kern/uipc_mbuf.c (revision 268358)
+++ sys/kern/uipc_mbuf.c (local)
@@ -917,7 +917,15 @@
struct mbuf *n;
/* Get the next new mbuf */
- if (remain >= MINCLSIZE) {
+ if (remain >= MJUMPAGESIZE) {
+ /*
+ * By allocating a bigger mbuf, we get fewer
+ * scatter gather entries for the hardware to
+ * process:
+ */
+ n = m_getjcl(how, m->m_type, 0, MJUMPAGESIZE);
+ nsize = MJUMPAGESIZE;
+ } else if (remain >= MINCLSIZE) {
n = m_getcl(how, m->m_type, 0);
nsize = MCLBYTES;
} else {
--------------060305060509090406020506--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?53BBFE08.1080804>