DPDK patches and discussions
 help / color / mirror / Atom feed
From: Shahaf Shuler <shahafs@mellanox.com>
To: wenzhuo.lu@intel.com, jingjing.wu@intel.com, bernard.iremonger@intel.com
Cc: dev@dpdk.org, rasland@mellanox.com, thomas@monjalon.net,
	ferruh.yigit@intel.com
Subject: [dpdk-dev] [PATCH v2 3/3] app/testpmd: map anonymous memory for eth devices
Date: Thu,  4 Apr 2019 08:14:43 +0300	[thread overview]
Message-ID: <3ea665747b4eba6f682a82e2f592e79642f02c39.1554354506.git.shahafs@mellanox.com> (raw)
Message-ID: <20190404051443.kKmZz7NRZf_lRZGV6AWBZbKlgUB2V4slbFvfFGpgXW8@z> (raw)
In-Reply-To: <cover.1554354506.git.shahafs@mellanox.com>

Mempools can be populated with anonymous memory when using command line
parameter --mp-alloc=anon.

Considering the mempools are going to be used by the net devices,
it is better to DMA map this memory.

This patch add such mapping now that we have the APIs in place[1].

[1] commit c33a675b6276 ("bus: introduce device level DMA memory mapping")

Signed-off-by: Shahaf Shuler <shahafs@mellanox.com>
---
 app/test-pmd/testpmd.c | 65 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 65 insertions(+)

diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index dd449a9859..3ad111a143 100644
--- a/app/test-pmd/testpmd.c
+++ b/app/test-pmd/testpmd.c
@@ -834,6 +834,63 @@ setup_extmem(uint32_t nb_mbufs, uint32_t mbuf_sz, bool huge)
 
 	return 0;
 }
+static void
+dma_unmap_cb(struct rte_mempool *mp __rte_unused, void *opaque __rte_unused,
+	     struct rte_mempool_memhdr *memhdr, unsigned mem_idx __rte_unused)
+{
+	uint16_t pid = 0;
+	int ret;
+
+	RTE_ETH_FOREACH_DEV(pid) {
+		struct rte_eth_dev *dev =
+			&rte_eth_devices[pid];
+
+		ret = rte_dev_dma_unmap(dev->device, memhdr->addr, 0,
+					memhdr->len);
+		if (ret) {
+			TESTPMD_LOG(DEBUG,
+				    "unable to DMA unmap addr 0x%p "
+				    "for device %s\n",
+				    memhdr->addr, dev->data->name);
+		}
+	}
+	ret = rte_extmem_unregister(memhdr->addr, memhdr->len);
+	if (ret) {
+		TESTPMD_LOG(DEBUG,
+			    "unable to un-register addr 0x%p\n", memhdr->addr);
+		return;
+	}
+}
+
+static void
+dma_map_cb(struct rte_mempool *mp __rte_unused, void *opaque __rte_unused,
+	   struct rte_mempool_memhdr *memhdr, unsigned mem_idx __rte_unused)
+{
+	uint16_t pid = 0;
+	size_t page_size = sysconf(_SC_PAGESIZE);
+	int ret;
+
+	ret = rte_extmem_register(memhdr->addr, memhdr->len, NULL, 0,
+				  page_size);
+	if (ret) {
+		TESTPMD_LOG(DEBUG,
+			    "unable to register addr 0x%p\n", memhdr->addr);
+		return;
+	}
+	RTE_ETH_FOREACH_DEV(pid) {
+		struct rte_eth_dev *dev =
+			&rte_eth_devices[pid];
+
+		ret = rte_dev_dma_map(dev->device, memhdr->addr, 0,
+				      memhdr->len);
+		if (ret) {
+			TESTPMD_LOG(DEBUG,
+				    "unable to DMA map addr 0x%p "
+				    "for device %s\n",
+				    memhdr->addr, dev->data->name);
+		}
+	}
+}
 
 /*
  * Configuration initialisation done once at init time.
@@ -879,6 +936,7 @@ mbuf_pool_create(uint16_t mbuf_seg_size, unsigned nb_mbuf,
 			}
 			rte_pktmbuf_pool_init(rte_mp, NULL);
 			rte_mempool_obj_iter(rte_mp, rte_pktmbuf_init, NULL);
+			rte_mempool_mem_iter(rte_mp, dma_map_cb, NULL);
 			break;
 		}
 	case MP_ALLOC_XMEM:
@@ -2410,6 +2468,13 @@ pmd_test_exit(void)
 	if (test_done == 0)
 		stop_packet_forwarding();
 
+	for (i = 0 ; i < RTE_MAX_NUMA_NODES ; i++) {
+		if (mempools[i]) {
+			if (mp_alloc_type == MP_ALLOC_ANON)
+				rte_mempool_mem_iter(mempools[i], dma_unmap_cb,
+						     NULL);
+		}
+	}
 	if (ports != NULL) {
 		no_link_check = 1;
 		RTE_ETH_FOREACH_DEV(pt_id) {
-- 
2.12.0


  parent reply	other threads:[~2019-04-04  5:14 UTC|newest]

Thread overview: 58+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-01 10:34 [dpdk-dev] [PATCH 0/3] DMA map anonymous memory to " Shahaf Shuler
2019-04-01 10:34 ` Shahaf Shuler
2019-04-01 10:34 ` [dpdk-dev] [PATCH 1/3] app/testpmd: fix mempool free on exit Shahaf Shuler
2019-04-01 10:34   ` Shahaf Shuler
2019-04-01 10:34 ` [dpdk-dev] [PATCH 2/3] app/testpmd: support creation of no IOVA contig mempools Shahaf Shuler
2019-04-01 10:34   ` Shahaf Shuler
2019-04-01 13:50   ` Burakov, Anatoly
2019-04-01 13:50     ` Burakov, Anatoly
2019-04-02  7:02     ` Shahaf Shuler
2019-04-02  7:02       ` Shahaf Shuler
2019-04-02 15:15       ` Burakov, Anatoly
2019-04-02 15:15         ` Burakov, Anatoly
2019-04-03  5:47         ` Shahaf Shuler
2019-04-03  5:47           ` Shahaf Shuler
2019-04-01 10:34 ` [dpdk-dev] [PATCH 3/3] app/testpmd: map anonymous memory for eth devices Shahaf Shuler
2019-04-01 10:34   ` Shahaf Shuler
2019-04-01 13:28   ` Burakov, Anatoly
2019-04-01 13:28     ` Burakov, Anatoly
2019-04-02  7:04     ` Shahaf Shuler
2019-04-02  7:04       ` Shahaf Shuler
2019-04-04  5:14 ` [dpdk-dev] [PATCH v2 0/3] DMA map anonymous memory to " Shahaf Shuler
2019-04-04  5:14   ` Shahaf Shuler
2019-04-04  5:14   ` [dpdk-dev] [PATCH v2 1/3] app/testpmd: fix mempool free on exit Shahaf Shuler
2019-04-04  5:14     ` Shahaf Shuler
2019-04-08 13:39     ` Iremonger, Bernard
2019-04-08 13:39       ` Iremonger, Bernard
2019-04-04  5:14   ` [dpdk-dev] [PATCH v2 2/3] app/testpmd: support creation of no IOVA contig mempools Shahaf Shuler
2019-04-04  5:14     ` Shahaf Shuler
2019-04-04  9:36     ` Burakov, Anatoly
2019-04-04  9:36       ` Burakov, Anatoly
2019-04-04  5:14   ` Shahaf Shuler [this message]
2019-04-04  5:14     ` [dpdk-dev] [PATCH v2 3/3] app/testpmd: map anonymous memory for eth devices Shahaf Shuler
2019-04-04  9:38     ` Burakov, Anatoly
2019-04-04  9:38       ` Burakov, Anatoly
2019-04-04 19:34   ` [dpdk-dev] [PATCH v3 0/3] DMA map anonymous memory to " Shahaf Shuler
2019-04-04 19:34     ` Shahaf Shuler
2019-04-04 19:34     ` [dpdk-dev] [PATCH v3 1/3] app/testpmd: fix mempool free on exit Shahaf Shuler
2019-04-04 19:34       ` Shahaf Shuler
2019-04-05 14:41       ` Ferruh Yigit
2019-04-05 14:41         ` Ferruh Yigit
2019-04-04 19:34     ` [dpdk-dev] [PATCH v3 2/3] app/testpmd: support creation of no IOVA contig mempools Shahaf Shuler
2019-04-04 19:34       ` Shahaf Shuler
2019-04-04 19:35     ` [dpdk-dev] [PATCH v3 3/3] app/testpmd: map anonymous memory for eth devices Shahaf Shuler
2019-04-04 19:35       ` Shahaf Shuler
2019-04-05  0:17     ` [dpdk-dev] [PATCH v3 0/3] DMA map anonymous memory to " Ferruh Yigit
2019-04-05  0:17       ` Ferruh Yigit
2019-04-07  5:02     ` [dpdk-dev] [PATCH v4 " Shahaf Shuler
2019-04-07  5:02       ` Shahaf Shuler
2019-04-07  5:02       ` [dpdk-dev] [PATCH v4 1/3] app/testpmd: fix mempool free on exit Shahaf Shuler
2019-04-07  5:02         ` Shahaf Shuler
2019-04-08 14:14         ` Iremonger, Bernard
2019-04-08 14:14           ` Iremonger, Bernard
2019-04-07  5:02       ` [dpdk-dev] [PATCH v4 2/3] app/testpmd: support creation of no IOVA contig mempools Shahaf Shuler
2019-04-07  5:02         ` Shahaf Shuler
2019-04-07  5:02       ` [dpdk-dev] [PATCH v4 3/3] app/testpmd: map anonymous memory for eth devices Shahaf Shuler
2019-04-07  5:02         ` Shahaf Shuler
2019-04-11 14:07       ` [dpdk-dev] [PATCH v4 0/3] DMA map anonymous memory to " Ferruh Yigit
2019-04-11 14:07         ` Ferruh Yigit

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=3ea665747b4eba6f682a82e2f592e79642f02c39.1554354506.git.shahafs@mellanox.com \
    --to=shahafs@mellanox.com \
    --cc=bernard.iremonger@intel.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=jingjing.wu@intel.com \
    --cc=rasland@mellanox.com \
    --cc=thomas@monjalon.net \
    --cc=wenzhuo.lu@intel.com \
    /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).