From: Huawei Xie <huawei.xie@intel.com>
To: dev@dpdk.org
Cc: dprovan@bivio.net
Subject: [dpdk-dev] [PATCH v6 1/2] mbuf: provide rte_pktmbuf_alloc_bulk API
Date: Wed, 27 Jan 2016 01:03:34 +0800 [thread overview]
Message-ID: <1453827815-56384-2-git-send-email-huawei.xie@intel.com> (raw)
In-Reply-To: <1453827815-56384-1-git-send-email-huawei.xie@intel.com>
v6 changes:
reflect the changes in release notes and library version map file
revise our duff's code style a bit to make it more readable
v5 changes:
add comment about duff's device and our variant implementation
v3 changes:
move while after case 0
add context about duff's device and why we use while loop in the commit
message
v2 changes:
unroll the loop a bit to help the performance
rte_pktmbuf_alloc_bulk allocates a bulk of packet mbufs.
There is related thread about this bulk API.
http://dpdk.org/dev/patchwork/patch/4718/
Thanks to Konstantin's loop unrolling.
Attached the wiki page about duff's device. It explains the performance
optimization through loop unwinding, and also the most dramatic use of
case label fall-through.
https://en.wikipedia.org/wiki/Duff%27s_device
In our implementation, we use while() loop rather than do{} while() loop
because we could not assume count is strictly positive. Using while()
loop saves one line of check if count is zero.
Signed-off-by: Gerald Rogers <gerald.rogers@intel.com>
Signed-off-by: Huawei Xie <huawei.xie@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
---
doc/guides/rel_notes/release_2_3.rst | 3 ++
lib/librte_mbuf/rte_mbuf.h | 55 ++++++++++++++++++++++++++++++++++++
lib/librte_mbuf/rte_mbuf_version.map | 7 +++++
3 files changed, 65 insertions(+)
diff --git a/doc/guides/rel_notes/release_2_3.rst b/doc/guides/rel_notes/release_2_3.rst
index 99de186..a52cba3 100644
--- a/doc/guides/rel_notes/release_2_3.rst
+++ b/doc/guides/rel_notes/release_2_3.rst
@@ -4,6 +4,9 @@ DPDK Release 2.3
New Features
------------
+* **Enable bulk allocation of mbufs. **
+ A new function ``rte_pktmbuf_alloc_bulk()`` has been added to allow the user
+ to allocate a bulk of mbufs.
Resolved Issues
---------------
diff --git a/lib/librte_mbuf/rte_mbuf.h b/lib/librte_mbuf/rte_mbuf.h
index f234ac9..b2ed479 100644
--- a/lib/librte_mbuf/rte_mbuf.h
+++ b/lib/librte_mbuf/rte_mbuf.h
@@ -1336,6 +1336,61 @@ static inline struct rte_mbuf *rte_pktmbuf_alloc(struct rte_mempool *mp)
}
/**
+ * Allocate a bulk of mbufs, initialize refcnt and reset the fields to default
+ * values.
+ *
+ * @param pool
+ * The mempool from which mbufs are allocated.
+ * @param mbufs
+ * Array of pointers to mbufs
+ * @param count
+ * Array size
+ * @return
+ * - 0: Success
+ */
+static inline int rte_pktmbuf_alloc_bulk(struct rte_mempool *pool,
+ struct rte_mbuf **mbufs, unsigned count)
+{
+ unsigned idx = 0;
+ int rc;
+
+ rc = rte_mempool_get_bulk(pool, (void **)mbufs, count);
+ if (unlikely(rc))
+ return rc;
+
+ /* To understand duff's device on loop unwinding optimization, see
+ * https://en.wikipedia.org/wiki/Duff's_device.
+ * Here while() loop is used rather than do() while{} to avoid extra
+ * check if count is zero.
+ */
+ switch (count % 4) {
+ case 0:
+ while (idx != count) {
+ RTE_MBUF_ASSERT(rte_mbuf_refcnt_read(mbufs[idx]) == 0);
+ rte_mbuf_refcnt_set(mbufs[idx], 1);
+ rte_pktmbuf_reset(mbufs[idx]);
+ idx++;
+ case 3:
+ RTE_MBUF_ASSERT(rte_mbuf_refcnt_read(mbufs[idx]) == 0);
+ rte_mbuf_refcnt_set(mbufs[idx], 1);
+ rte_pktmbuf_reset(mbufs[idx]);
+ idx++;
+ case 2:
+ RTE_MBUF_ASSERT(rte_mbuf_refcnt_read(mbufs[idx]) == 0);
+ rte_mbuf_refcnt_set(mbufs[idx], 1);
+ rte_pktmbuf_reset(mbufs[idx]);
+ idx++;
+ case 1:
+ RTE_MBUF_ASSERT(rte_mbuf_refcnt_read(mbufs[idx]) == 0);
+ rte_mbuf_refcnt_set(mbufs[idx], 1);
+ rte_pktmbuf_reset(mbufs[idx]);
+ idx++;
+ }
+ }
+ return 0;
+}
+
+/**
* Attach packet mbuf to another packet mbuf.
*
* After attachment we refer the mbuf we attached as 'indirect',
diff --git a/lib/librte_mbuf/rte_mbuf_version.map b/lib/librte_mbuf/rte_mbuf_version.map
index e10f6bd..257c65a 100644
--- a/lib/librte_mbuf/rte_mbuf_version.map
+++ b/lib/librte_mbuf/rte_mbuf_version.map
@@ -18,3 +18,10 @@ DPDK_2.1 {
rte_pktmbuf_pool_create;
} DPDK_2.0;
+
+DPDK_2.3 {
+ global:
+
+ rte_pktmbuf_alloc_bulk;
+
+} DPDK_2.1;
--
1.8.1.4
next prev parent reply other threads:[~2016-01-27 8:57 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-13 23:35 [dpdk-dev] [PATCH 0/2] provide rte_pktmbuf_alloc_bulk API and call it in vhost dequeue Huawei Xie
2015-12-13 23:35 ` [dpdk-dev] [PATCH 1/2] mbuf: provide rte_pktmbuf_alloc_bulk API Huawei Xie
2015-12-13 23:35 ` [dpdk-dev] [PATCH 2/2] vhost: call rte_pktmbuf_alloc_bulk in vhost dequeue Huawei Xie
2015-12-14 1:14 ` [dpdk-dev] [PATCH v2 0/2] provide rte_pktmbuf_alloc_bulk API and call it " Huawei Xie
2015-12-14 1:14 ` [dpdk-dev] [PATCH v2 1/2] mbuf: provide rte_pktmbuf_alloc_bulk API Huawei Xie
2015-12-17 6:41 ` Yuanhan Liu
2015-12-17 15:42 ` Ananyev, Konstantin
2015-12-18 2:17 ` Yuanhan Liu
2015-12-18 5:01 ` Stephen Hemminger
2015-12-18 5:21 ` Yuanhan Liu
2015-12-18 7:10 ` Xie, Huawei
2015-12-18 10:44 ` Ananyev, Konstantin
2015-12-18 17:32 ` Stephen Hemminger
2015-12-18 19:27 ` Wiles, Keith
2015-12-21 15:21 ` Xie, Huawei
2015-12-21 17:20 ` Wiles, Keith
2015-12-21 21:30 ` Thomas Monjalon
2015-12-22 1:58 ` Xie, Huawei
2015-12-21 22:34 ` Don Provan
2015-12-21 12:25 ` Xie, Huawei
2015-12-14 1:14 ` [dpdk-dev] [PATCH v2 2/2] vhost: call rte_pktmbuf_alloc_bulk in vhost dequeue Huawei Xie
2015-12-17 6:41 ` Yuanhan Liu
2015-12-22 16:17 ` [dpdk-dev] [PATCH v3 0/2] provide rte_pktmbuf_alloc_bulk API and call it " Huawei Xie
2015-12-22 16:17 ` [dpdk-dev] [PATCH v3 1/2] mbuf: provide rte_pktmbuf_alloc_bulk API Huawei Xie
2015-12-23 18:37 ` Stephen Hemminger
2015-12-23 18:49 ` Ananyev, Konstantin
2015-12-24 1:33 ` Xie, Huawei
2015-12-22 16:17 ` [dpdk-dev] [PATCH v3 2/2] vhost: call rte_pktmbuf_alloc_bulk in vhost dequeue Huawei Xie
2015-12-23 11:22 ` linhaifeng
2015-12-23 11:39 ` Xie, Huawei
2015-12-22 23:05 ` [dpdk-dev] [PATCH v4 0/2] provide rte_pktmbuf_alloc_bulk API and call it " Huawei Xie
2015-12-22 23:05 ` [dpdk-dev] [PATCH v4 1/2] mbuf: provide rte_pktmbuf_alloc_bulk API Huawei Xie
2015-12-22 23:05 ` [dpdk-dev] [PATCH v4 2/2] vhost: call rte_pktmbuf_alloc_bulk in vhost dequeue Huawei Xie
2015-12-27 16:38 ` [dpdk-dev] [PATCH v5 0/2] provide rte_pktmbuf_alloc_bulk API and call it " Huawei Xie
2015-12-27 16:38 ` [dpdk-dev] [PATCH v5 1/2] mbuf: provide rte_pktmbuf_alloc_bulk API Huawei Xie
2015-12-27 16:38 ` [dpdk-dev] [PATCH v5 2/2] vhost: call rte_pktmbuf_alloc_bulk in vhost dequeue Huawei Xie
2016-01-26 17:03 ` [dpdk-dev] [PATCH v6 0/2] provide rte_pktmbuf_alloc_bulk API and call it " Huawei Xie
2016-01-26 17:03 ` Huawei Xie [this message]
2016-01-27 13:56 ` [dpdk-dev] [PATCH v6 1/2] mbuf: provide rte_pktmbuf_alloc_bulk API Panu Matilainen
2016-02-03 17:23 ` Olivier MATZ
2016-02-22 14:49 ` Xie, Huawei
2016-02-23 5:35 ` Xie, Huawei
2016-02-24 12:11 ` Panu Matilainen
2016-02-24 13:23 ` Ananyev, Konstantin
2016-02-26 7:39 ` Xie, Huawei
2016-02-26 8:45 ` Olivier MATZ
2016-02-29 10:51 ` Panu Matilainen
2016-02-29 16:14 ` Thomas Monjalon
2016-02-26 8:55 ` Olivier MATZ
2016-02-26 9:07 ` Xie, Huawei
2016-02-26 9:18 ` Olivier MATZ
2016-01-26 17:03 ` [dpdk-dev] [PATCH v6 2/2] vhost: call rte_pktmbuf_alloc_bulk in vhost dequeue Huawei Xie
2016-02-28 12:44 ` [dpdk-dev] [PATCH v7] mbuf: provide rte_pktmbuf_alloc_bulk API Huawei Xie
2016-02-29 16:27 ` Thomas Monjalon
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1453827815-56384-2-git-send-email-huawei.xie@intel.com \
--to=huawei.xie@intel.com \
--cc=dev@dpdk.org \
--cc=dprovan@bivio.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).