DPDK patches and discussions
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH] rte_mbuf: scattered pktmbufs freeing optimization
@ 2015-02-26 23:15 vadim.suraev
  2015-02-27  0:49 ` Stephen Hemminger
  2015-02-27 11:17 ` Ananyev, Konstantin
  0 siblings, 2 replies; 10+ messages in thread
From: vadim.suraev @ 2015-02-26 23:15 UTC (permalink / raw)
  To: dev

From: "vadim.suraev@gmail.com" <vadim.suraev@gmail.com>

new function - rte_pktmbuf_free_bulk makes freeing long
scattered (chained) pktmbufs belonging to the same pool 
more optimal using rte_mempool_put_bulk rather than calling
rte_mempool_put for each segment. 
Inlike rte_pktmbuf_free, which calls rte_pktmbuf_free_seg,
this function calls __rte_pktmbuf_prefree_seg. If non-NULL
returned, the pointer is placed in an array. When array is
filled or when the last segment is processed, rte_mempool_put_bulk 
is called. In case of multiple producers, performs 3 times better.


Signed-off-by: vadim.suraev@gmail.com <vadim.suraev@gmail.com>
---
 lib/librte_mbuf/rte_mbuf.h |   55 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 55 insertions(+)

diff --git a/lib/librte_mbuf/rte_mbuf.h b/lib/librte_mbuf/rte_mbuf.h
index 17ba791..1d6f848 100644
--- a/lib/librte_mbuf/rte_mbuf.h
+++ b/lib/librte_mbuf/rte_mbuf.h
@@ -824,6 +824,61 @@ static inline void rte_pktmbuf_free(struct rte_mbuf *m)
 	}
 }
 
+/* This macro defines the size of max bulk of mbufs to free for rte_pktmbuf_free_bulk */
+#define MAX_MBUF_FREE_SIZE 32
+
+/* If RTE_LIBRTE_MBUF_DEBUG is enabled, checks if all mbufs must belong to the same mempool */
+#ifdef RTE_LIBRTE_MBUF_DEBUG
+
+#define RTE_MBUF_MEMPOOL_CHECK1(m) struct rte_mempool *first_buffers_mempool = (m) ? (m)->pool : NULL
+
+#define RTE_MBUF_MEMPOOL_CHECK2(m) RTE_MBUF_ASSERT(first_buffers_mempool == (m)->pool)
+
+#else
+
+#define RTE_MBUF_MEMPOOL_CHECK1(m)
+
+#define RTE_MBUF_MEMPOOL_CHECK2(m)
+
+#endif
+
+/**
+ * Free chained (scattered) mbuf into its original mempool.
+ *
+ * All the mbufs in the chain must belong to the same mempool.
+ *
+ * @param head
+ *   The head of mbufs to be freed chain
+ */
+
+static inline void __attribute__((always_inline))
+rte_pktmbuf_free_bulk(struct rte_mbuf *head)
+{
+    void *mbufs[MAX_MBUF_FREE_SIZE];
+    unsigned mbufs_count = 0;
+    struct rte_mbuf *next;
+
+    RTE_MBUF_MEMPOOL_CHECK1(head);
+
+    while(head) {
+        next = head->next;
+        head->next = NULL;
+        if(__rte_pktmbuf_prefree_seg(head)) {
+            RTE_MBUF_ASSERT(rte_mbuf_refcnt_read(head) == 0);
+            RTE_MBUF_MEMPOOL_CHECK2(head);
+            mbufs[mbufs_count++] = head;
+        }
+        head = next;
+        if(mbufs_count == MAX_MBUF_FREE_SIZE) {
+            rte_mempool_put_bulk(((struct rte_mbuf *)mbufs[0])->pool,mbufs,mbufs_count);
+            mbufs_count = 0;
+        }
+    }
+    if(mbufs_count > 0) {
+        rte_mempool_put_bulk(((struct rte_mbuf *)mbufs[0])->pool,mbufs,mbufs_count);
+    }
+}
+
 /**
  * Creates a "clone" of the given packet mbuf.
  *
-- 
1.7.9.5

^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2015-03-09  8:38 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-02-26 23:15 [dpdk-dev] [PATCH] rte_mbuf: scattered pktmbufs freeing optimization vadim.suraev
2015-02-27  0:49 ` Stephen Hemminger
2015-02-27 11:17 ` Ananyev, Konstantin
2015-02-27 12:18   ` Vadim Suraev
     [not found]     ` <2601191342CEEE43887BDE71AB977258213F2E3D@irsmsx105.ger.corp.intel.com>
2015-02-27 13:10       ` Ananyev, Konstantin
2015-02-27 13:20     ` Olivier MATZ
2015-02-27 17:09       ` Vadim Suraev
2015-03-04  8:54         ` Olivier MATZ
2015-03-06 23:24           ` Vadim Suraev
2015-03-09  8:38             ` Olivier MATZ

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).