[dpdk-dev] [PATCH] i40e: support xen domain0
i40e was failing to run in XEN domain0, as the physical memory for adminq DMA should be allocated and translated in a different way for XEN domain0. So rte_memzone_reserve_bounded() should be used for DMA memory allocation, and rte_mem_phy2mch() should be used for DMA memory address translation to support running i40e PMD in XEN domain0. Signed-off-by: Helin Zhang --- lib/librte_pmd_i40e/i40e_ethdev.c | 9 + 1 file changed, 9 insertions(+) diff --git a/lib/librte_pmd_i40e/i40e_ethdev.c b/lib/librte_pmd_i40e/i40e_ethdev.c index 9ed31b5..7a823a0 100644 --- a/lib/librte_pmd_i40e/i40e_ethdev.c +++ b/lib/librte_pmd_i40e/i40e_ethdev.c @@ -1515,14 +1515,23 @@ i40e_allocate_dma_mem_d(__attribute__((unused)) struct i40e_hw *hw, id++; snprintf(z_name, sizeof(z_name), "i40e_dma_%"PRIu64, id); +#ifdef RTE_LIBRTE_XEN_DOM0 + mz = rte_memzone_reserve_bounded(z_name, size, 0, 0, alignment, + RTE_PGSIZE_2M); +#else mz = rte_memzone_reserve_aligned(z_name, size, 0, 0, alignment); +#endif if (!mz) return I40E_ERR_NO_MEMORY; mem->id = id; mem->size = size; mem->va = mz->addr; +#ifdef RTE_LIBRTE_XEN_DOM0 + mem->pa = rte_mem_phy2mch(mz->memseg_id, mz->phys_addr); +#else mem->pa = mz->phys_addr; +#endif return I40E_SUCCESS; } -- 1.8.1.4
[dpdk-dev] [PATCH] i40e: support xen domain0
> i40e was failing to run in XEN domain0, as the physical > memory for adminq DMA should be allocated and translated > in a different way for XEN domain0. So > rte_memzone_reserve_bounded() should be used for DMA > memory allocation, and rte_mem_phy2mch() should be used > for DMA memory address translation to support running > i40e PMD in XEN domain0. > > Signed-off-by: Helin Zhang Tested-by: Zhaochen Zhan This patch updates i40e to support xen domain0, testpmd can works OK with it. It is ready to be integrated to dpdk.org.
[dpdk-dev] [PATCH 0/4] RX/TX queue start/stop enhancement
From: "Chen Jing D(Mark)" This patch mainly includes 2 changes. One is in testpmd to add command testing specific RX/TX queue start/stop in Port X. Another change is in i40e, which implemented rx/tx_queue_start/stop in both PF and VF driver. In the meanwhile, support field start_rx_per_q in i40e_dev_rx_queue_setup and start_tx_per_q in i40e_dev_tx_queue_setup. Chen Jing D(Mark) (4): testpmd: add command to start/stop specfic queue i40e: PF Add support for per-queue start/stop i40e: PF driver to support RX/TX config paramter i40e: VF driver to support per-queue RX/TX start/stop app/test-pmd/cmdline.c | 95 ++ app/test-pmd/config.c|6 +- app/test-pmd/testpmd.c | 12 ++ app/test-pmd/testpmd.h |4 + lib/librte_pmd_i40e/i40e_ethdev.c| 50 +--- lib/librte_pmd_i40e/i40e_ethdev_vf.c | 232 ++ lib/librte_pmd_i40e/i40e_rxtx.c | 133 +-- lib/librte_pmd_i40e/i40e_rxtx.h |9 ++ 8 files changed, 453 insertions(+), 88 deletions(-) -- 1.7.7.6
[dpdk-dev] [PATCH 1/4] testpmd: add command to start/stop specfic queue
From: "Chen Jing D(Mark)"
rte_ether library provide function pointer to start/stop specific
RX/TX queue, NIC driver also implemented functions. This change
adds command in testpmd to start/stop specific RX/TX queue.
Signed-off-by: Chen Jing D(Mark)
Reviewed-by: Konstantin Ananyev
Reviewed-by: Changchun Ouyang
Reviewed-by: Huawei Xie
---
app/test-pmd/cmdline.c | 95
app/test-pmd/config.c |6 ++--
app/test-pmd/testpmd.c | 12 ++
app/test-pmd/testpmd.h |4 ++
4 files changed, 114 insertions(+), 3 deletions(-)
diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index 345be11..ddf5def 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -565,6 +565,10 @@ static void cmd_help_long_parsed(void *parsed_result,
" tx rs bit threshold.\n\n"
"port config mtu X value\n"
"Set the MTU of port X to a given value\n\n"
+
+ "port (port_id) (rxq|txq) (queue_id) (start|stop)\n"
+ "Start/stop a rx/tx queue of port X. Only take
effect"
+ " when port X is started\n"
);
}
@@ -1425,6 +1429,96 @@ cmdline_parse_inst_t cmd_config_rss_hash_key = {
},
};
+/* *** configure port rxq/txq start/stop *** */
+struct cmd_config_rxtx_queue {
+ cmdline_fixed_string_t port;
+ uint8_t portid;
+ cmdline_fixed_string_t rxtxq;
+ uint16_t qid;
+ cmdline_fixed_string_t opname;
+};
+
+static void
+cmd_config_rxtx_queue_parsed(void *parsed_result,
+ __attribute__((unused)) struct cmdline *cl,
+ __attribute__((unused)) void *data)
+{
+ struct cmd_config_rxtx_queue *res = parsed_result;
+ uint8_t isrx;
+ uint8_t isstart;
+
+ if (test_done == 0) {
+ printf("Please stop forwarding first\n");
+ return;
+ }
+
+ if (port_id_is_invalid(res->portid))
+ return;
+
+ if (port_is_started(res->portid) != 1) {
+ printf("Please start port %u first\n", res->portid);
+ return;
+ }
+
+ if (!strcmp(res->rxtxq, "rxq"))
+ isrx = 1;
+ else if (!strcmp(res->rxtxq, "txq"))
+ isrx = 0;
+ else {
+ printf("Unknown parameter\n");
+ return;
+ }
+
+ if (isrx && rx_queue_id_is_invalid(res->qid))
+ return;
+ else if (!isrx && tx_queue_id_is_invalid(res->qid))
+ return;
+
+ if (!strcmp(res->opname, "start"))
+ isstart = 1;
+ else if (!strcmp(res->opname, "stop"))
+ isstart = 0;
+ else {
+ printf("Unknown parameter\n");
+ return;
+ }
+
+ if (isstart && isrx)
+ rte_eth_dev_rx_queue_start(res->portid, res->qid);
+ else if (!isstart && isrx)
+ rte_eth_dev_rx_queue_stop(res->portid, res->qid);
+ else if (isstart && !isrx)
+ rte_eth_dev_tx_queue_start(res->portid, res->qid);
+ else
+ rte_eth_dev_tx_queue_stop(res->portid, res->qid);
+}
+
+cmdline_parse_token_string_t cmd_config_rxtx_queue_port =
+ TOKEN_STRING_INITIALIZER(struct cmd_config_rxtx_queue, port, "port");
+cmdline_parse_token_num_t cmd_config_rxtx_queue_portid =
+ TOKEN_NUM_INITIALIZER(struct cmd_config_rxtx_queue, portid, UINT8);
+cmdline_parse_token_string_t cmd_config_rxtx_queue_rxtxq =
+ TOKEN_STRING_INITIALIZER(struct cmd_config_rxtx_queue, rxtxq,
"rxq#txq");
+cmdline_parse_token_num_t cmd_config_rxtx_queue_qid =
+ TOKEN_NUM_INITIALIZER(struct cmd_config_rxtx_queue, qid, UINT16);
+cmdline_parse_token_string_t cmd_config_rxtx_queue_opname =
+ TOKEN_STRING_INITIALIZER(struct cmd_config_rxtx_queue, opname,
+ "start#stop");
+
+cmdline_parse_inst_t cmd_config_rxtx_queue = {
+ .f = cmd_config_rxtx_queue_parsed,
+ .data = NULL,
+ .help_str = "port X rxq|txq ID start|stop",
+ .tokens = {
+ (void *)&cmd_config_speed_all_port,
+ (void *)&cmd_config_rxtx_queue_portid,
+ (void *)&cmd_config_rxtx_queue_rxtxq,
+ (void *)&cmd_config_rxtx_queue_qid,
+ (void *)&cmd_config_rxtx_queue_opname,
+ NULL,
+ },
+};
+
/* *** Configure RSS RETA *** */
struct cmd_config_rss_reta {
cmdline_fixed_string_t port;
@@ -7393,6 +7487,7 @@ cmdline_parse_ctx_t main_ctx[] = {
(cmdline_parse_inst_t *)&cmd_config_max_pkt_len,
(cmdline_parse_inst_t *)&cmd_config_rx_mode_flag,
(cmdline_parse_inst_t *)&cmd_config_rss,
+ (cmdline_parse_inst_t *)&cmd_config_rxtx_queue,
(cmdline_parse_inst_t *)&cmd_config_rss_reta,
(cmdline_parse_inst_t *)&cmd_showport_reta,
(cmdline_parse_inst_t *)&cm
[dpdk-dev] [PATCH 2/4] i40e: PF Add support for per-queue start/stop
From: "Chen Jing D(Mark)"
I40E driver add function pointer to start/stop specific RX/TX queue.
Signed-off-by: Chen Jing D(Mark)
Reviewed-by: Konstantin Ananyev
Reviewed-by: Changchun Ouyang
Reviewed-by: Huawei Xie
---
lib/librte_pmd_i40e/i40e_ethdev.c |4 +
lib/librte_pmd_i40e/i40e_rxtx.c | 112 +
lib/librte_pmd_i40e/i40e_rxtx.h |4 +
3 files changed, 120 insertions(+), 0 deletions(-)
diff --git a/lib/librte_pmd_i40e/i40e_ethdev.c
b/lib/librte_pmd_i40e/i40e_ethdev.c
index 9ed31b5..81a1deb 100644
--- a/lib/librte_pmd_i40e/i40e_ethdev.c
+++ b/lib/librte_pmd_i40e/i40e_ethdev.c
@@ -232,6 +232,10 @@ static struct eth_dev_ops i40e_eth_dev_ops = {
.vlan_offload_set = i40e_vlan_offload_set,
.vlan_strip_queue_set = i40e_vlan_strip_queue_set,
.vlan_pvid_set= i40e_vlan_pvid_set,
+ .rx_queue_start = i40e_dev_rx_queue_start,
+ .rx_queue_stop= i40e_dev_rx_queue_stop,
+ .tx_queue_start = i40e_dev_tx_queue_start,
+ .tx_queue_stop= i40e_dev_tx_queue_stop,
.rx_queue_setup = i40e_dev_rx_queue_setup,
.rx_queue_release = i40e_dev_rx_queue_release,
.rx_queue_count = i40e_dev_rx_queue_count,
diff --git a/lib/librte_pmd_i40e/i40e_rxtx.c b/lib/librte_pmd_i40e/i40e_rxtx.c
index 83b9462..323c004 100644
--- a/lib/librte_pmd_i40e/i40e_rxtx.c
+++ b/lib/librte_pmd_i40e/i40e_rxtx.c
@@ -1429,6 +1429,118 @@ i40e_xmit_pkts_simple(void *tx_queue,
}
int
+i40e_dev_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id)
+{
+ struct i40e_vsi *vsi = I40E_DEV_PRIVATE_TO_VSI(dev->data->dev_private);
+ struct i40e_rx_queue *rxq;
+ int err = -1;
+ struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
+ uint16_t q_base = vsi->base_queue;
+
+ PMD_INIT_FUNC_TRACE();
+
+ if (rx_queue_id < dev->data->nb_rx_queues) {
+ rxq = dev->data->rx_queues[rx_queue_id];
+
+ err = i40e_alloc_rx_queue_mbufs(rxq);
+ if (err) {
+ PMD_DRV_LOG(ERR, "Failed to allocate RX queue mbuf\n");
+ return err;
+ }
+
+ rte_wmb();
+
+ /* Init the RX tail regieter. */
+ I40E_PCI_REG_WRITE(rxq->qrx_tail, rxq->nb_rx_desc - 1);
+
+ err = i40e_switch_rx_queue(hw, rx_queue_id + q_base, TRUE);
+
+ if (err) {
+ PMD_DRV_LOG(ERR, "Failed to switch RX queue %u on\n",
+ rx_queue_id);
+
+ i40e_rx_queue_release_mbufs(rxq);
+ i40e_reset_rx_queue(rxq);
+ }
+ }
+
+ return err;
+}
+
+int
+i40e_dev_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id)
+{
+ struct i40e_vsi *vsi = I40E_DEV_PRIVATE_TO_VSI(dev->data->dev_private);
+ struct i40e_rx_queue *rxq;
+ int err;
+ struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
+ uint16_t q_base = vsi->base_queue;
+
+ if (rx_queue_id < dev->data->nb_rx_queues) {
+ rxq = dev->data->rx_queues[rx_queue_id];
+
+ err = i40e_switch_rx_queue(hw, rx_queue_id + q_base, FALSE);
+
+ if (err) {
+ PMD_DRV_LOG(ERR, "Failed to switch RX queue %u off\n",
+ rx_queue_id);
+ return err;
+ }
+ i40e_rx_queue_release_mbufs(rxq);
+ i40e_reset_rx_queue(rxq);
+ }
+
+ return 0;
+}
+
+int
+i40e_dev_tx_queue_start(struct rte_eth_dev *dev, uint16_t tx_queue_id)
+{
+ struct i40e_vsi *vsi = I40E_DEV_PRIVATE_TO_VSI(dev->data->dev_private);
+ int err = -1;
+ struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
+ uint16_t q_base = vsi->base_queue;
+
+ PMD_INIT_FUNC_TRACE();
+
+ if (tx_queue_id < dev->data->nb_tx_queues) {
+ err = i40e_switch_tx_queue(hw, tx_queue_id + q_base, TRUE);
+ if (err)
+ PMD_DRV_LOG(ERR, "Failed to switch TX queue %u on\n",
+ tx_queue_id);
+ }
+
+ return err;
+}
+
+int
+i40e_dev_tx_queue_stop(struct rte_eth_dev *dev, uint16_t tx_queue_id)
+{
+ struct i40e_vsi *vsi = I40E_DEV_PRIVATE_TO_VSI(dev->data->dev_private);
+ struct i40e_tx_queue *txq;
+ int err;
+ struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
+ uint16_t q_base = vsi->base_queue;
+
+ if (tx_queue_id < dev->data->nb_tx_queues) {
+ txq = dev->data->tx_queues[tx_queue_id];
+
+ err = i40e_switch_tx_queue(hw, tx_queue_id + q_base, FALSE);
+
+ if (err) {
+ PMD_DRV_LOG(ERR, "Failed to switch TX queue %u of\n",
+ tx_queue_id);
+ return err;
+ }
+
+
[dpdk-dev] [PATCH 3/4] i40e: PF driver to support RX/TX config paramter
From: "Chen Jing D(Mark)"
PF driver to support field start_rx_per_q in i40e_dev_rx_queue_setup
and start_tx_per_q in i40e_dev_tx_queue_setup. In the meanwhile,
Change dev_start/stop function to call per-queue RX/TX function.
Signed-off-by: Chen Jing D(Mark)
Reviewed-by: Konstantin Ananyev
Reviewed-by: Changchun Ouyang
Reviewed-by: Huawei Xie
---
lib/librte_pmd_i40e/i40e_ethdev.c | 46 +---
lib/librte_pmd_i40e/i40e_rxtx.c | 21 +---
lib/librte_pmd_i40e/i40e_rxtx.h |5
3 files changed, 39 insertions(+), 33 deletions(-)
diff --git a/lib/librte_pmd_i40e/i40e_ethdev.c
b/lib/librte_pmd_i40e/i40e_ethdev.c
index 81a1deb..d5dd6e3 100644
--- a/lib/librte_pmd_i40e/i40e_ethdev.c
+++ b/lib/librte_pmd_i40e/i40e_ethdev.c
@@ -690,7 +690,6 @@ i40e_dev_start(struct rte_eth_dev *dev)
err_up:
i40e_vsi_switch_queues(vsi, FALSE);
- i40e_dev_clear_queues(dev);
return ret;
}
@@ -704,9 +703,6 @@ i40e_dev_stop(struct rte_eth_dev *dev)
/* Disable all queues */
i40e_vsi_switch_queues(vsi, FALSE);
- /* Clear all queues and release memory */
- i40e_dev_clear_queues(dev);
-
/* un-map queues with interrupt registers */
i40e_vsi_disable_queues_intr(vsi);
i40e_vsi_queues_unbind_intr(vsi);
@@ -2845,17 +2841,23 @@ static int
i40e_vsi_switch_tx_queues(struct i40e_vsi *vsi, bool on)
{
struct rte_eth_dev_data *dev_data = I40E_VSI_TO_DEV_DATA(vsi);
- struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
struct i40e_tx_queue *txq;
- uint16_t i, pf_q;
+ struct rte_eth_dev *dev = I40E_VSI_TO_ETH_DEV(vsi);
+ uint16_t i;
int ret;
- pf_q = vsi->base_queue;
- for (i = 0; i < dev_data->nb_tx_queues; i++, pf_q++) {
+ for (i = 0; i < dev_data->nb_tx_queues; i++) {
txq = dev_data->tx_queues[i];
- if (!txq->q_set)
- continue; /* Queue not configured */
- ret = i40e_switch_tx_queue(hw, pf_q, on);
+ /**
+* Don't operate the queue in 2 cases. One is queue is never
configured,
+* another is it's start operation and start_tx_per_q is set.
+**/
+ if (!txq->q_set || (on && txq->start_tx_per_q))
+ continue;
+ if (on)
+ ret = i40e_dev_tx_queue_start(dev, i);
+ else
+ ret = i40e_dev_tx_queue_stop(dev, i);
if ( ret != I40E_SUCCESS)
return ret;
}
@@ -2919,18 +2921,24 @@ static int
i40e_vsi_switch_rx_queues(struct i40e_vsi *vsi, bool on)
{
struct rte_eth_dev_data *dev_data = I40E_VSI_TO_DEV_DATA(vsi);
- struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
struct i40e_rx_queue *rxq;
- uint16_t i, pf_q;
+ struct rte_eth_dev *dev = I40E_VSI_TO_ETH_DEV(vsi);
+ uint16_t i;
int ret;
- pf_q = vsi->base_queue;
- for (i = 0; i < dev_data->nb_rx_queues; i++, pf_q++) {
+ for (i = 0; i < dev_data->nb_rx_queues; i++) {
rxq = dev_data->rx_queues[i];
- if (!rxq->q_set)
- continue; /* Queue not configured */
- ret = i40e_switch_rx_queue(hw, pf_q, on);
- if ( ret != I40E_SUCCESS)
+ /**
+* Don't operate the queue in 2 cases. One is queue is never
configured,
+* another is it's start operation and start_rx_per_q is set.
+**/
+ if (!rxq->q_set || (on && rxq->start_rx_per_q))
+ continue;
+ if (on)
+ ret = i40e_dev_rx_queue_start(dev, i);
+ else
+ ret = i40e_dev_rx_queue_stop(dev, i);
+ if (ret != I40E_SUCCESS)
return ret;
}
diff --git a/lib/librte_pmd_i40e/i40e_rxtx.c b/lib/librte_pmd_i40e/i40e_rxtx.c
index 323c004..f153844 100644
--- a/lib/librte_pmd_i40e/i40e_rxtx.c
+++ b/lib/librte_pmd_i40e/i40e_rxtx.c
@@ -88,9 +88,6 @@ i40e_ring_dma_zone_reserve(struct rte_eth_dev *dev,
uint16_t queue_id,
uint32_t ring_size,
int socket_id);
-static void i40e_reset_rx_queue(struct i40e_rx_queue *rxq);
-static void i40e_reset_tx_queue(struct i40e_tx_queue *txq);
-static void i40e_tx_queue_release_mbufs(struct i40e_tx_queue *txq);
static uint16_t i40e_xmit_pkts_simple(void *tx_queue,
struct rte_mbuf **tx_pkts,
uint16_t nb_pkts);
@@ -1594,6 +1591,7 @@ i40e_dev_rx_queue_setup(struct rte_eth_dev *dev,
0 : ETHER_CRC_LEN);
rxq->drop_en = rx_conf->rx_drop_en;
rxq->vsi = vsi;
+ rxq->start_rx_per_q = rx_conf->start_rx_per_q;
/* Allocate the m
[dpdk-dev] [PATCH 4/4] i40e: VF driver to support per-queue RX/TX start/stop
From: "Chen Jing D(Mark)"
VF driver to add per-queue RX/TX start/stop function. Support
field start_rx_per_q in rx_queue_setup and start_tx_per_q in
tx_queue_setup. In the meanwhile, Change dev_start/stop function
to call per-queue RX/TX function.
Signed-off-by: Chen Jing D(Mark)
Reviewed-by: Konstantin Ananyev
Reviewed-by: Changchun Ouyang
Reviewed-by: Huawei Xie
---
lib/librte_pmd_i40e/i40e_ethdev_vf.c | 232 ++
1 files changed, 180 insertions(+), 52 deletions(-)
diff --git a/lib/librte_pmd_i40e/i40e_ethdev_vf.c
b/lib/librte_pmd_i40e/i40e_ethdev_vf.c
index 2726bfb..a9a5633 100644
--- a/lib/librte_pmd_i40e/i40e_ethdev_vf.c
+++ b/lib/librte_pmd_i40e/i40e_ethdev_vf.c
@@ -125,6 +125,10 @@ static void i40evf_dev_allmulticast_disable(struct
rte_eth_dev *dev);
static int i40evf_get_link_status(struct rte_eth_dev *dev,
struct rte_eth_link *link);
static int i40evf_init_vlan(struct rte_eth_dev *dev);
+static int i40evf_dev_rx_queue_start(struct rte_eth_dev *, uint16_t);
+static int i40evf_dev_rx_queue_stop(struct rte_eth_dev *, uint16_t);
+static int i40evf_dev_tx_queue_start(struct rte_eth_dev *, uint16_t);
+static int i40evf_dev_tx_queue_stop(struct rte_eth_dev *, uint16_t);
static struct eth_dev_ops i40evf_eth_dev_ops = {
.dev_configure= i40evf_dev_configure,
.dev_start= i40evf_dev_start,
@@ -140,6 +144,10 @@ static struct eth_dev_ops i40evf_eth_dev_ops = {
.vlan_filter_set = i40evf_vlan_filter_set,
.vlan_offload_set = i40evf_vlan_offload_set,
.vlan_pvid_set= i40evf_vlan_pvid_set,
+ .rx_queue_start = i40evf_dev_rx_queue_start,
+ .rx_queue_stop= i40evf_dev_rx_queue_stop,
+ .tx_queue_start = i40evf_dev_tx_queue_start,
+ .tx_queue_stop= i40evf_dev_tx_queue_stop,
.rx_queue_setup = i40e_dev_rx_queue_setup,
.rx_queue_release = i40e_dev_rx_queue_release,
.tx_queue_setup = i40e_dev_tx_queue_setup,
@@ -628,65 +636,94 @@ i40evf_config_irq_map(struct rte_eth_dev *dev)
}
static int
-i40evf_enable_queues(struct rte_eth_dev *dev)
+i40evf_switch_queue(struct rte_eth_dev *dev, bool isrx, uint16_t qid,
+ bool on)
{
struct i40e_vf *vf = I40EVF_DEV_PRIVATE_TO_VF(dev->data->dev_private);
struct i40e_virtchnl_queue_select queue_select;
- int err, i;
+ int err;
struct vf_cmd_info args;
-
+ memset(&queue_select, 0, sizeof(queue_select));
queue_select.vsi_id = vf->vsi_res->vsi_id;
- queue_select.rx_queues = 0;
- /* Enable configured RX queues */
- for (i = 0; i < dev->data->nb_rx_queues; i++)
- queue_select.rx_queues |= 1 << i;
-
- /* Enable configured TX queues */
- queue_select.tx_queues = 0;
- for (i = 0; i < dev->data->nb_tx_queues; i++)
- queue_select.tx_queues |= 1 << i;
+ if (isrx)
+ queue_select.rx_queues |= 1 << qid;
+ else
+ queue_select.tx_queues |= 1 << qid;
- args.ops = I40E_VIRTCHNL_OP_ENABLE_QUEUES;
+ if (on)
+ args.ops = I40E_VIRTCHNL_OP_ENABLE_QUEUES;
+ else
+ args.ops = I40E_VIRTCHNL_OP_DISABLE_QUEUES;
args.in_args = (u8 *)&queue_select;
args.in_args_size = sizeof(queue_select);
args.out_buffer = cmd_result_buffer;
args.out_size = I40E_AQ_BUF_SZ;
err = i40evf_execute_vf_cmd(dev, &args);
if (err)
- PMD_DRV_LOG(ERR, "fail to execute command OP_ENABLE_QUEUES\n");
+ PMD_DRV_LOG(ERR, "fail to switch %s %u %s\n", isrx ? "RX" :
"TX",
+ qid, on ? "on" : "off");
return err;
}
static int
-i40evf_disable_queues(struct rte_eth_dev *dev)
+i40evf_start_queues(struct rte_eth_dev *dev)
{
- struct i40e_virtchnl_queue_select queue_select;
- int err, i;
- struct vf_cmd_info args;
+ struct rte_eth_dev_data *dev_data = dev->data;
+ int i;
+ struct i40e_rx_queue *rxq;
+ struct i40e_tx_queue *txq;
- /* Enable configured RX queues */
- queue_select.rx_queues = 0;
- for (i = 0; i < dev->data->nb_rx_queues; i++)
- queue_select.rx_queues |= 1 << i;
+ for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ rxq = dev_data->rx_queues[i];
+ if (rxq->start_rx_per_q)
+ continue;
+ if (i40evf_dev_rx_queue_start(dev, i) != 0) {
+ PMD_DRV_LOG(ERR, "Fail to start queue %u\n",
+ i);
+ return -1;
+ }
+ }
- /* Enable configured TX queues */
- queue_select.tx_queues = 0;
- for (i = 0; i < dev->data->nb_tx_queues; i++)
- queue_select.tx_queues |= 1 << i;
+ for (i = 0; i < dev->data->nb_tx_queues; i++) {
[dpdk-dev] [PATCH 2/4] i40e: PF Add support for per-queue start/stop
I will verified this next week. :)
-Original Message-
From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Chen Jing D(Mark)
Sent: Thursday, August 14, 2014 3:35 PM
To: dev at dpdk.org
Subject: [dpdk-dev] [PATCH 2/4] i40e: PF Add support for per-queue start/stop
From: "Chen Jing D(Mark)"
I40E driver add function pointer to start/stop specific RX/TX queue.
Signed-off-by: Chen Jing D(Mark)
Reviewed-by: Konstantin Ananyev
Reviewed-by: Changchun Ouyang
Reviewed-by: Huawei Xie
---
lib/librte_pmd_i40e/i40e_ethdev.c |4 +
lib/librte_pmd_i40e/i40e_rxtx.c | 112 +
lib/librte_pmd_i40e/i40e_rxtx.h |4 +
3 files changed, 120 insertions(+), 0 deletions(-)
diff --git a/lib/librte_pmd_i40e/i40e_ethdev.c
b/lib/librte_pmd_i40e/i40e_ethdev.c
index 9ed31b5..81a1deb 100644
--- a/lib/librte_pmd_i40e/i40e_ethdev.c
+++ b/lib/librte_pmd_i40e/i40e_ethdev.c
@@ -232,6 +232,10 @@ static struct eth_dev_ops i40e_eth_dev_ops = {
.vlan_offload_set = i40e_vlan_offload_set,
.vlan_strip_queue_set = i40e_vlan_strip_queue_set,
.vlan_pvid_set= i40e_vlan_pvid_set,
+ .rx_queue_start = i40e_dev_rx_queue_start,
+ .rx_queue_stop= i40e_dev_rx_queue_stop,
+ .tx_queue_start = i40e_dev_tx_queue_start,
+ .tx_queue_stop= i40e_dev_tx_queue_stop,
.rx_queue_setup = i40e_dev_rx_queue_setup,
.rx_queue_release = i40e_dev_rx_queue_release,
.rx_queue_count = i40e_dev_rx_queue_count,
diff --git a/lib/librte_pmd_i40e/i40e_rxtx.c b/lib/librte_pmd_i40e/i40e_rxtx.c
index 83b9462..323c004 100644
--- a/lib/librte_pmd_i40e/i40e_rxtx.c
+++ b/lib/librte_pmd_i40e/i40e_rxtx.c
@@ -1429,6 +1429,118 @@ i40e_xmit_pkts_simple(void *tx_queue,
}
int
+i40e_dev_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id)
+{
+ struct i40e_vsi *vsi = I40E_DEV_PRIVATE_TO_VSI(dev->data->dev_private);
+ struct i40e_rx_queue *rxq;
+ int err = -1;
+ struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
+ uint16_t q_base = vsi->base_queue;
+
+ PMD_INIT_FUNC_TRACE();
+
+ if (rx_queue_id < dev->data->nb_rx_queues) {
+ rxq = dev->data->rx_queues[rx_queue_id];
+
+ err = i40e_alloc_rx_queue_mbufs(rxq);
+ if (err) {
+ PMD_DRV_LOG(ERR, "Failed to allocate RX queue mbuf\n");
+ return err;
+ }
+
+ rte_wmb();
+
+ /* Init the RX tail regieter. */
+ I40E_PCI_REG_WRITE(rxq->qrx_tail, rxq->nb_rx_desc - 1);
+
+ err = i40e_switch_rx_queue(hw, rx_queue_id + q_base, TRUE);
+
+ if (err) {
+ PMD_DRV_LOG(ERR, "Failed to switch RX queue %u on\n",
+ rx_queue_id);
+
+ i40e_rx_queue_release_mbufs(rxq);
+ i40e_reset_rx_queue(rxq);
+ }
+ }
+
+ return err;
+}
+
+int
+i40e_dev_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id)
+{
+ struct i40e_vsi *vsi = I40E_DEV_PRIVATE_TO_VSI(dev->data->dev_private);
+ struct i40e_rx_queue *rxq;
+ int err;
+ struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
+ uint16_t q_base = vsi->base_queue;
+
+ if (rx_queue_id < dev->data->nb_rx_queues) {
+ rxq = dev->data->rx_queues[rx_queue_id];
+
+ err = i40e_switch_rx_queue(hw, rx_queue_id + q_base, FALSE);
+
+ if (err) {
+ PMD_DRV_LOG(ERR, "Failed to switch RX queue %u off\n",
+ rx_queue_id);
+ return err;
+ }
+ i40e_rx_queue_release_mbufs(rxq);
+ i40e_reset_rx_queue(rxq);
+ }
+
+ return 0;
+}
+
+int
+i40e_dev_tx_queue_start(struct rte_eth_dev *dev, uint16_t tx_queue_id)
+{
+ struct i40e_vsi *vsi = I40E_DEV_PRIVATE_TO_VSI(dev->data->dev_private);
+ int err = -1;
+ struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
+ uint16_t q_base = vsi->base_queue;
+
+ PMD_INIT_FUNC_TRACE();
+
+ if (tx_queue_id < dev->data->nb_tx_queues) {
+ err = i40e_switch_tx_queue(hw, tx_queue_id + q_base, TRUE);
+ if (err)
+ PMD_DRV_LOG(ERR, "Failed to switch TX queue %u on\n",
+ tx_queue_id);
+ }
+
+ return err;
+}
+
+int
+i40e_dev_tx_queue_stop(struct rte_eth_dev *dev, uint16_t tx_queue_id)
+{
+ struct i40e_vsi *vsi = I40E_DEV_PRIVATE_TO_VSI(dev->data->dev_private);
+ struct i40e_tx_queue *txq;
+ int err;
+ struct i40e_hw *hw = I40E_VSI_TO_HW(vsi);
+ uint16_t q_base = vsi->base_queue;
+
+ if (tx_queue_id < dev->data->nb_tx_queues) {
+ txq = dev->data->tx_queues[tx_queue_id];
+
+ err = i40e_sw
[dpdk-dev] [PATCH v3] virtio: Support mergeable buffer in virtio pmd
v3 change:
- Investigate the comments from Huawei and fix one potential issue of wrong
offset to
the number of descriptor in buffer; also fix other tiny comments.
v2 change:
- Resolve conflicts with the tip code;
- And resolve 2 issues:
-- fix mbuf leak when discard an uncompleted packet.
-- refine pkt.data to point to actual payload data start point.
v1 change:
- This patch supports mergeable buffer feature in DPDK based virtio PMD, which
can
receive jumbo frame with larger size, like 3K, 4K or even 9K.
Signed-off-by: Changchun Ouyang
Acked-by: Huawei Xie
---
lib/librte_pmd_virtio/virtio_ethdev.c | 20 +--
lib/librte_pmd_virtio/virtio_ethdev.h | 3 +
lib/librte_pmd_virtio/virtio_rxtx.c | 221 +-
3 files changed, 207 insertions(+), 37 deletions(-)
diff --git a/lib/librte_pmd_virtio/virtio_ethdev.c
b/lib/librte_pmd_virtio/virtio_ethdev.c
index b9f5529..535d798 100644
--- a/lib/librte_pmd_virtio/virtio_ethdev.c
+++ b/lib/librte_pmd_virtio/virtio_ethdev.c
@@ -337,7 +337,7 @@ int virtio_dev_queue_setup(struct rte_eth_dev *dev,
snprintf(vq_name, sizeof(vq_name), "port%d_tvq%d_hdrzone",
dev->data->port_id, queue_idx);
vq->virtio_net_hdr_mz = rte_memzone_reserve_aligned(vq_name,
- vq_size * sizeof(struct virtio_net_hdr),
+ vq_size * hw->vtnet_hdr_size,
socket_id, 0, CACHE_LINE_SIZE);
if (vq->virtio_net_hdr_mz == NULL) {
rte_free(vq);
@@ -346,7 +346,7 @@ int virtio_dev_queue_setup(struct rte_eth_dev *dev,
vq->virtio_net_hdr_mem =
vq->virtio_net_hdr_mz->phys_addr;
memset(vq->virtio_net_hdr_mz->addr, 0,
- vq_size * sizeof(struct virtio_net_hdr));
+ vq_size * hw->vtnet_hdr_size);
} else if (queue_type == VTNET_CQ) {
/* Allocate a page for control vq command, data and status */
snprintf(vq_name, sizeof(vq_name), "port%d_cvq_hdrzone",
@@ -571,9 +571,6 @@ virtio_negotiate_features(struct virtio_hw *hw)
mask |= VIRTIO_NET_F_GUEST_TSO4 | VIRTIO_NET_F_GUEST_TSO6 |
VIRTIO_NET_F_GUEST_ECN;
mask |= VTNET_LRO_FEATURES;
- /* rx_mbuf should not be in multiple merged segments */
- mask |= VIRTIO_NET_F_MRG_RXBUF;
-
/* not negotiating INDIRECT descriptor table support */
mask |= VIRTIO_RING_F_INDIRECT_DESC;
@@ -746,7 +743,6 @@ eth_virtio_dev_init(__rte_unused struct eth_driver *eth_drv,
}
eth_dev->dev_ops = &virtio_eth_dev_ops;
- eth_dev->rx_pkt_burst = &virtio_recv_pkts;
eth_dev->tx_pkt_burst = &virtio_xmit_pkts;
if (rte_eal_process_type() == RTE_PROC_SECONDARY)
@@ -801,10 +797,13 @@ eth_virtio_dev_init(__rte_unused struct eth_driver
*eth_drv,
virtio_negotiate_features(hw);
/* Setting up rx_header size for the device */
- if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF))
+ if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
+ eth_dev->rx_pkt_burst = &virtio_recv_mergeable_pkts;
hw->vtnet_hdr_size = sizeof(struct virtio_net_hdr_mrg_rxbuf);
- else
+ } else {
+ eth_dev->rx_pkt_burst = &virtio_recv_pkts;
hw->vtnet_hdr_size = sizeof(struct virtio_net_hdr);
+ }
/* Allocate memory for storing MAC addresses */
eth_dev->data->mac_addrs = rte_zmalloc("virtio", ETHER_ADDR_LEN, 0);
@@ -1009,7 +1008,7 @@ static void virtio_dev_free_mbufs(struct rte_eth_dev *dev)
while ((buf = (struct rte_mbuf *)virtqueue_detatch_unused(
dev->data->rx_queues[i])) != NULL) {
- rte_pktmbuf_free_seg(buf);
+ rte_pktmbuf_free(buf);
mbuf_num++;
}
@@ -1028,7 +1027,8 @@ static void virtio_dev_free_mbufs(struct rte_eth_dev *dev)
mbuf_num = 0;
while ((buf = (struct rte_mbuf *)virtqueue_detatch_unused(
dev->data->tx_queues[i])) != NULL) {
- rte_pktmbuf_free_seg(buf);
+ rte_pktmbuf_free(buf);
+
mbuf_num++;
}
diff --git a/lib/librte_pmd_virtio/virtio_ethdev.h
b/lib/librte_pmd_virtio/virtio_ethdev.h
index 858e644..d2e1eed 100644
--- a/lib/librte_pmd_virtio/virtio_ethdev.h
+++ b/lib/librte_pmd_virtio/virtio_ethdev.h
@@ -104,6 +104,9 @@ int virtio_dev_tx_queue_setup(struct rte_eth_dev *dev,
uint16_t tx_queue_id,
uint16_t virtio_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts);
+uint16_t virtio_recv_mergeable_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
+ uint16_t nb_pkts);
+
uint16_t virtio_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
[dpdk-dev] Signature filter for virtual function
Hello, I'm trying to configure the flow director of VF 82599 port using rte_eth_dev_fdir_add_signature_filter (). The function returns with ENOTSUP. Doe's flow director supported for VF ? What can be the reason for ENOTSUP ? Thanks, Ilan B
[dpdk-dev] 1.7.0 release failed to compile on linux
Hi, I got the following error when I tried to compile 1.7.0 release tarball on a Linux box (Ubuntu 12.04/kernel=3.13.0-32-generic, gcc=4.6.3): In file included from /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_osdep.h:41:0, from /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_hw.h:31, from /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_api.h:31, from /home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.c:38: /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3853:1: error: conflicting types for ?skb_set_hash? /usr/src/linux-headers-3.13.0-32-generic/include/linux/skbuff.h:740:1: note: previous definition of ?skb_set_hash? was here make[8]: *** [/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.o] Error 1 Apparently the version check around 'skb_set_hash' in kcompat.h failed for some reason. To work around it, I temporarily commented out that 'skb_set_hash' definition code and rte_kni complied fine. But it failed again in librte_pmd_ixgbe: = In file included from /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:41:0: /usr/lib/gcc/x86_64-linux-gnu/4.6/include/nmmintrin.h:31:3: error: #error "SSE4.2 instruction set not enabled" /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c: In function ?ixgbe_recv_pkts_vec?: /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3: error: implicit declaration of function ?_mm_shuffle_epi8? [-Werror=implicit-function-declaration] /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3: error: nested extern declaration of ?_mm_shuffle_epi8? [-Werror=nested-externs] /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:11: error: incompatible types when assigning to type ?__m128i? from type ?int? /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:298:11: error: incompatible types when assigning to type ?__m128i? from type ?int? /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:313:11: error: incompatible types when assigning to type ?__m128i? from type ?int? /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:314:11: error: incompatible types when assigning to type ?__m128i? from type ?int? /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3: error: implicit declaration of function ?_mm_popcnt_u64? [-Werror=implicit-function-declaration] /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3: error: nested extern declaration of ?_mm_popcnt_u64? [-Werror=nested-externs] cc1: all warnings being treated as errors make[3]: *** [ixgbe_rxtx_vec.o] Error 1 == My CPU is the older Intel C2D E6550 which does not have sse4.2/sse4.1 support. Is there a way to work around this? -Q
[dpdk-dev] 1.7.0 release failed to compile on linux
For the SKB function change I had to change the ifdef in the kcompat.h file
around line 3848 (I think) to:
/* Changed the next line to use (3,13,8) instead of (3,14,0) KeithW */
#if ( LINUX_VERSION_CODE < KERNEL_VERSION(3,13,8) )
#if (!(RHEL_RELEASE_CODE && RHEL_RELEASE_CODE >= RHEL_RELEASE_VERSION(7,0)))
#ifdef NETIF_F_RXHASH
#define PKT_HASH_TYPE_L3 0
static inline void
skb_set_hash(struct sk_buff *skb, __u32 hash, __always_unused int type)
{
skb->rxhash = hash;
}
#endif /* NETIF_F_RXHASH */
#endif /* < RHEL7 */
#endif /* < 3.14.0 */
Keith Wiles, Principal Technologist with CTO office, Wind River mobile
972-213-5533
On Aug 14, 2014, at 4:44 PM, Qiaobing Xie mailto:qiaobing.xie at gmail.com>> wrote:
Hi,
I got the following error when I tried to compile 1.7.0 release tarball on a
Linux box (Ubuntu 12.04/kernel=3.13.0-32-generic, gcc=4.6.3):
In file included from
/home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_osdep.h:41:0,
from
/home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_hw.h:31,
from
/home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_api.h:31,
from
/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.c:38:
/home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3853:1:
error: conflicting types for ?skb_set_hash?
/usr/src/linux-headers-3.13.0-32-generic/include/linux/skbuff.h:740:1: note:
previous definition of ?skb_set_hash? was here
make[8]: ***
[/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.o]
Error 1
Apparently the version check around 'skb_set_hash' in kcompat.h failed for some
reason. To work around it, I temporarily commented out that 'skb_set_hash'
definition code and rte_kni complied fine. But it failed again in
librte_pmd_ixgbe:
=
In file included from
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:41:0:
/usr/lib/gcc/x86_64-linux-gnu/4.6/include/nmmintrin.h:31:3: error: #error
"SSE4.2 instruction set not enabled"
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c: In function
?ixgbe_recv_pkts_vec?:
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3: error:
implicit declaration of function ?_mm_shuffle_epi8?
[-Werror=implicit-function-declaration]
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3: error:
nested extern declaration of ?_mm_shuffle_epi8? [-Werror=nested-externs]
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:11: error:
incompatible types when assigning to type ?__m128i? from type ?int?
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:298:11: error:
incompatible types when assigning to type ?__m128i? from type ?int?
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:313:11: error:
incompatible types when assigning to type ?__m128i? from type ?int?
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:314:11: error:
incompatible types when assigning to type ?__m128i? from type ?int?
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3: error:
implicit declaration of function ?_mm_popcnt_u64?
[-Werror=implicit-function-declaration]
/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3: error:
nested extern declaration of ?_mm_popcnt_u64? [-Werror=nested-externs]
cc1: all warnings being treated as errors
make[3]: *** [ixgbe_rxtx_vec.o] Error 1
==
My CPU is the older Intel C2D E6550 which does not have sse4.2/sse4.1 support.
Is there a way to work around this?
-Q
[dpdk-dev] 1.7.0 release failed to compile on linux
On 08/14/2014 05:44 PM, Qiaobing Xie wrote: > Hi, > > I got the following error when I tried to compile 1.7.0 release > tarball on a Linux box (Ubuntu 12.04/kernel=3.13.0-32-generic, > gcc=4.6.3): > > > In file included from > /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_osdep.h:41:0, > from > /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_hw.h:31, > from > /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_api.h:31, > from > /home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.c:38: > /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3853:1: > > error: conflicting types for ?skb_set_hash? > /usr/src/linux-headers-3.13.0-32-generic/include/linux/skbuff.h:740:1: > note: previous definition of ?skb_set_hash? was here > make[8]: *** > [/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.o] > Error 1 > > > Apparently the version check around 'skb_set_hash' in kcompat.h failed > for some reason. To work around it, I temporarily commented out that > 'skb_set_hash' definition code and rte_kni complied fine. But it > failed again in librte_pmd_ixgbe: > > = > In file included from > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:41:0: > /usr/lib/gcc/x86_64-linux-gnu/4.6/include/nmmintrin.h:31:3: error: > #error "SSE4.2 instruction set not enabled" From the error, nmmintrin.h contains built-in CPU calls that depend on the CPU having SSE4.2 extensions. You need to hunt down the call that is being executed and see which file is including nmmintrin.h. For example, I have encountered similar error on a CPU that does not have SSE4.1 and 4.2 extension and it was trying to call the built-in HW crc hashing call. instead I switched with rte_jhash.h, the alternative SW implementation of the hashing call. Looking below, nmmintrin.h is being included by the lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c, from the looks of it. This means that your defconfig (did you use the default one?) has CONFIG_RTE_IXGBE_INC_VECTOR value set to "y". try setting it to "n" and see where the build would stop next looking for another SSE4.x based builtin call. My ultimate recommendation is to update you CPU to a newer one or perform the build/run on a separate newer CPU. Aws\ > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c: In > function ?ixgbe_recv_pkts_vec?: > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3: > error: implicit declaration of function ?_mm_shuffle_epi8? > [-Werror=implicit-function-declaration] > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3: > error: nested extern declaration of ?_mm_shuffle_epi8? > [-Werror=nested-externs] > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:11: > error: incompatible types when assigning to type ?__m128i? from type > ?int? > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:298:11: > error: incompatible types when assigning to type ?__m128i? from type > ?int? > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:313:11: > error: incompatible types when assigning to type ?__m128i? from type > ?int? > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:314:11: > error: incompatible types when assigning to type ?__m128i? from type > ?int? > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3: > error: implicit declaration of function ?_mm_popcnt_u64? > [-Werror=implicit-function-declaration] > /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3: > error: nested extern declaration of ?_mm_popcnt_u64? > [-Werror=nested-externs] > cc1: all warnings being treated as errors > make[3]: *** [ixgbe_rxtx_vec.o] Error 1 > == > > My CPU is the older Intel C2D E6550 which does not have sse4.2/sse4.1 > support. Is there a way to work around this? > > -Q > > > > >
[dpdk-dev] 1.7.0 release failed to compile on linux
On 08/14/2014 05:55 PM, Wiles, Roger Keith wrote:
> For the SKB function change I had to change the ifdef in the kcompat.h file
> around line 3848 (I think) to:
>
> /* Changed the next line to use (3,13,8) instead of (3,14,0) KeithW */
> #if ( LINUX_VERSION_CODE < KERNEL_VERSION(3,13,8) )
> #if (!(RHEL_RELEASE_CODE && RHEL_RELEASE_CODE >= RHEL_RELEASE_VERSION(7,0)))
> #ifdef NETIF_F_RXHASH
> #define PKT_HASH_TYPE_L3 0
> static inline void
> skb_set_hash(struct sk_buff *skb, __u32 hash, __always_unused int type)
> {
> skb->rxhash = hash;
> }
> #endif /* NETIF_F_RXHASH */
> #endif /* < RHEL7 */
> #endif /* < 3.14.0 */
>
>
> Keith Wiles, Principal Technologist with CTO office, Wind River mobile
> 972-213-5533
>
> On Aug 14, 2014, at 4:44 PM, Qiaobing Xie gmail.com> wrote:
>
> Hi,
>
> I got the following error when I tried to compile 1.7.0 release tarball on a
> Linux box (Ubuntu 12.04/kernel=3.13.0-32-generic, gcc=4.6.3):
>
>
> In file included from
> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_osdep.h:41:0,
> from
> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_hw.h:31,
> from
> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_api.h:31,
> from
> /home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.c:38:
> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3853:1:
> error: conflicting types for ?skb_set_hash?
> /usr/src/linux-headers-3.13.0-32-generic/include/linux/skbuff.h:740:1: note:
> previous definition of ?skb_set_hash? was here
Last time I checked, I overcame this error by not using the tarball but
by simply cloning the dpdk git repo and checking out the v1.7.0-rc4 tag.
For example:
git clone git://dpdk.org/dpdk
git checkout v1.7.0-rc4
git checkout -b dpdk1.7.0-rc4 < the branch name is just an example,
you could decide on any other name as you see fit.
My machine info:
3.11.0-20 kernel, ubuntu 12.04, gcc 4.6.3
From the v1.7.0-rc4 tag, viewing
lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3848, you could see:
#if ( LINUX_VERSION_CODE < KERNEL_VERSION(3,14,0) )
#if (!(RHEL_RELEASE_CODE && RHEL_RELEASE_CODE >= RHEL_RELEASE_VERSION(7,0)))
#ifdef NETIF_F_RXHASH
#define PKT_HASH_TYPE_L3 0
static inline void
skb_set_hash(struct sk_buff *skb, __u32 hash, __always_unused int type)
{
skb->rxhash = hash;
}
#endif /* NETIF_F_RXHASH */
#endif /* < RHEL7 */
#endif /* < 3.14.0 */
> make[8]: ***
> [/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.o]
> Error 1
>
>
> Apparently the version check around 'skb_set_hash' in kcompat.h failed for
> some reason. To work around it, I temporarily commented out that
> 'skb_set_hash' definition code and rte_kni complied fine. But it failed again
> in librte_pmd_ixgbe:
>
> =
> In file included from
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:41:0:
> /usr/lib/gcc/x86_64-linux-gnu/4.6/include/nmmintrin.h:31:3: error: #error
> "SSE4.2 instruction set not enabled"
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c: In function
> ?ixgbe_recv_pkts_vec?:
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3: error:
> implicit declaration of function ?_mm_shuffle_epi8?
> [-Werror=implicit-function-declaration]
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3: error:
> nested extern declaration of ?_mm_shuffle_epi8? [-Werror=nested-externs]
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:11: error:
> incompatible types when assigning to type ?__m128i? from type ?int?
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:298:11: error:
> incompatible types when assigning to type ?__m128i? from type ?int?
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:313:11: error:
> incompatible types when assigning to type ?__m128i? from type ?int?
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:314:11: error:
> incompatible types when assigning to type ?__m128i? from type ?int?
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3: error:
> implicit declaration of function ?_mm_popcnt_u64?
> [-Werror=implicit-function-declaration]
> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3: error:
> nested extern declaration of ?_mm_popcnt_u64? [-Werror=nested-externs]
> cc1: all warnings being treated as errors
> make[3]: *** [ixgbe_rxtx_vec.o] Error 1
> ==
>
> My CPU is the older Intel C2D E6550 which does not have sse4.2/sse4.1
> support. Is there a way to work around this?
>
> -Q
>
>
>
>
>
>
[dpdk-dev] 1.7.0 release failed to compile on linux
Aws,
Thanks, I will move to v1.7.0-rc4 and get it a try...
-Q
On 8/14/14 5:24 PM, Aws Ismail wrote:
>
>
> On 08/14/2014 05:55 PM, Wiles, Roger Keith wrote:
>> For the SKB function change I had to change the ifdef in the
>> kcompat.h file around line 3848 (I think) to:
>>
>> /* Changed the next line to use (3,13,8) instead of (3,14,0)
>> KeithW */
>> #if ( LINUX_VERSION_CODE < KERNEL_VERSION(3,13,8) )
>> #if (!(RHEL_RELEASE_CODE && RHEL_RELEASE_CODE >=
>> RHEL_RELEASE_VERSION(7,0)))
>> #ifdef NETIF_F_RXHASH
>> #define PKT_HASH_TYPE_L3 0
>> static inline void
>> skb_set_hash(struct sk_buff *skb, __u32 hash, __always_unused int type)
>> {
>> skb->rxhash = hash;
>> }
>> #endif /* NETIF_F_RXHASH */
>> #endif /* < RHEL7 */
>> #endif /* < 3.14.0 */
>>
>>
>> Keith Wiles, Principal Technologist with CTO office, Wind River
>> mobile 972-213-5533
>>
>> On Aug 14, 2014, at 4:44 PM, Qiaobing Xie
>> mailto:qiaobing.xie at gmail.com>> wrote:
>>
>> Hi,
>>
>> I got the following error when I tried to compile 1.7.0 release
>> tarball on a Linux box (Ubuntu 12.04/kernel=3.13.0-32-generic,
>> gcc=4.6.3):
>>
>>
>> In file included from
>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_osdep.h:41:0,
>> from
>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_hw.h:31,
>> from
>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_api.h:31,
>> from
>> /home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.c:38:
>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3853:1:
>>
>> error: conflicting types for ?skb_set_hash?
>> /usr/src/linux-headers-3.13.0-32-generic/include/linux/skbuff.h:740:1: note:
>> previous definition of ?skb_set_hash? was here
> Last time I checked, I overcame this error by not using the tarball
> but by simply cloning the dpdk git repo and checking out the
> v1.7.0-rc4 tag. For example:
>
> git clone git://dpdk.org/dpdk
> git checkout v1.7.0-rc4
> git checkout -b dpdk1.7.0-rc4 < the branch name is just an
> example, you could decide on any other name as you see fit.
>
> My machine info:
>
> 3.11.0-20 kernel, ubuntu 12.04, gcc 4.6.3
>
> From the v1.7.0-rc4 tag, viewing
> lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3848, you could see:
>
>
> #if ( LINUX_VERSION_CODE < KERNEL_VERSION(3,14,0) )
> #if (!(RHEL_RELEASE_CODE && RHEL_RELEASE_CODE >=
> RHEL_RELEASE_VERSION(7,0)))
> #ifdef NETIF_F_RXHASH
> #define PKT_HASH_TYPE_L3 0
> static inline void
> skb_set_hash(struct sk_buff *skb, __u32 hash, __always_unused int type)
> {
> skb->rxhash = hash;
> }
> #endif /* NETIF_F_RXHASH */
> #endif /* < RHEL7 */
> #endif /* < 3.14.0 */
>
>
>> make[8]: ***
>> [/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.o]
>>
>> Error 1
>>
>>
>> Apparently the version check around 'skb_set_hash' in kcompat.h
>> failed for some reason. To work around it, I temporarily commented
>> out that 'skb_set_hash' definition code and rte_kni complied fine.
>> But it failed again in librte_pmd_ixgbe:
>>
>> =
>> In file included from
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:41:0:
>> /usr/lib/gcc/x86_64-linux-gnu/4.6/include/nmmintrin.h:31:3: error:
>> #error "SSE4.2 instruction set not enabled"
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c: In
>> function ?ixgbe_recv_pkts_vec?:
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3:
>> error: implicit declaration of function ?_mm_shuffle_epi8?
>> [-Werror=implicit-function-declaration]
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3:
>> error: nested extern declaration of ?_mm_shuffle_epi8?
>> [-Werror=nested-externs]
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:11:
>> error: incompatible types when assigning to type ?__m128i? from type
>> ?int?
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:298:11:
>> error: incompatible types when assigning to type ?__m128i? from type
>> ?int?
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:313:11:
>> error: incompatible types when assigning to type ?__m128i? from type
>> ?int?
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:314:11:
>> error: incompatible types when assigning to type ?__m128i? from type
>> ?int?
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3:
>> error: implicit declaration of function ?_mm_popcnt_u64?
>> [-Werror=implicit-function-declaration]
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3:
>> error: nested extern declaration of ?_mm_popcnt_u64?
>> [-Werror=nested-externs]
>> cc1: all warnings being treated as errors
>> make[3]: *** [ixgbe_rxtx_vec.o] Error 1
>> ==
>>
>> My CPU is the older Intel C2D E6550 which does not have sse4.2/sse4.1
>> support. Is there a way to work around this?
>>
>> -Q
>>
>>
>>
>>
>>
>>
>
[dpdk-dev] 1.7.0 release failed to compile on linux
OK, I made the following 3 changes and the v1.7.0 release tarball got
built cleanly on my older machine (Ubuntu 12.04/kernel
3.13.0-32-generic, Intel C2D E6550 without SSE4.x, gcc 4.6.3):
1) kcompat.h change version check from 3.14.0 -> 3.13.8, as Roger suggested.
2) disable RTE_IXGBE_INC_VECTOR:
-CONFIG_RTE_IXGBE_INC_VECTOR=y
+CONFIG_RTE_IXGBE_INC_VECTOR=n
3) disable CONFIG_RTE_LIBRTE_ACL:
-CONFIG_RTE_LIBRTE_ACL=y
+CONFIG_RTE_LIBRTE_ACL=n
I haven't tried to test run anything yet... What would be the
implication without LIBRTE_ACL and IXGBE_INC_VECTOR built?
-Q
On 8/14/14 5:24 PM, Aws Ismail wrote:
>
>
> On 08/14/2014 05:55 PM, Wiles, Roger Keith wrote:
>> For the SKB function change I had to change the ifdef in the
>> kcompat.h file around line 3848 (I think) to:
>>
>> /* Changed the next line to use (3,13,8) instead of (3,14,0)
>> KeithW */
>> #if ( LINUX_VERSION_CODE < KERNEL_VERSION(3,13,8) )
>> #if (!(RHEL_RELEASE_CODE && RHEL_RELEASE_CODE >=
>> RHEL_RELEASE_VERSION(7,0)))
>> #ifdef NETIF_F_RXHASH
>> #define PKT_HASH_TYPE_L3 0
>> static inline void
>> skb_set_hash(struct sk_buff *skb, __u32 hash, __always_unused int type)
>> {
>> skb->rxhash = hash;
>> }
>> #endif /* NETIF_F_RXHASH */
>> #endif /* < RHEL7 */
>> #endif /* < 3.14.0 */
>>
>>
>> Keith Wiles, Principal Technologist with CTO office, Wind River
>> mobile 972-213-5533
>>
>> On Aug 14, 2014, at 4:44 PM, Qiaobing Xie
>> mailto:qiaobing.xie at gmail.com>> wrote:
>>
>> Hi,
>>
>> I got the following error when I tried to compile 1.7.0 release
>> tarball on a Linux box (Ubuntu 12.04/kernel=3.13.0-32-generic,
>> gcc=4.6.3):
>>
>>
>> In file included from
>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_osdep.h:41:0,
>> from
>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_hw.h:31,
>> from
>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_api.h:31,
>> from
>> /home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.c:38:
>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3853:1:
>>
>> error: conflicting types for ?skb_set_hash?
>> /usr/src/linux-headers-3.13.0-32-generic/include/linux/skbuff.h:740:1: note:
>> previous definition of ?skb_set_hash? was here
> Last time I checked, I overcame this error by not using the tarball
> but by simply cloning the dpdk git repo and checking out the
> v1.7.0-rc4 tag. For example:
>
> git clone git://dpdk.org/dpdk
> git checkout v1.7.0-rc4
> git checkout -b dpdk1.7.0-rc4 < the branch name is just an
> example, you could decide on any other name as you see fit.
>
> My machine info:
>
> 3.11.0-20 kernel, ubuntu 12.04, gcc 4.6.3
>
> From the v1.7.0-rc4 tag, viewing
> lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3848, you could see:
>
>
> #if ( LINUX_VERSION_CODE < KERNEL_VERSION(3,14,0) )
> #if (!(RHEL_RELEASE_CODE && RHEL_RELEASE_CODE >=
> RHEL_RELEASE_VERSION(7,0)))
> #ifdef NETIF_F_RXHASH
> #define PKT_HASH_TYPE_L3 0
> static inline void
> skb_set_hash(struct sk_buff *skb, __u32 hash, __always_unused int type)
> {
> skb->rxhash = hash;
> }
> #endif /* NETIF_F_RXHASH */
> #endif /* < RHEL7 */
> #endif /* < 3.14.0 */
>
>
>> make[8]: ***
>> [/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.o]
>>
>> Error 1
>>
>>
>> Apparently the version check around 'skb_set_hash' in kcompat.h
>> failed for some reason. To work around it, I temporarily commented
>> out that 'skb_set_hash' definition code and rte_kni complied fine.
>> But it failed again in librte_pmd_ixgbe:
>>
>> =
>> In file included from
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:41:0:
>> /usr/lib/gcc/x86_64-linux-gnu/4.6/include/nmmintrin.h:31:3: error:
>> #error "SSE4.2 instruction set not enabled"
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c: In
>> function ?ixgbe_recv_pkts_vec?:
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3:
>> error: implicit declaration of function ?_mm_shuffle_epi8?
>> [-Werror=implicit-function-declaration]
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:3:
>> error: nested extern declaration of ?_mm_shuffle_epi8?
>> [-Werror=nested-externs]
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:297:11:
>> error: incompatible types when assigning to type ?__m128i? from type
>> ?int?
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:298:11:
>> error: incompatible types when assigning to type ?__m128i? from type
>> ?int?
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:313:11:
>> error: incompatible types when assigning to type ?__m128i? from type
>> ?int?
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:314:11:
>> error: incompatible types when assigning to type ?__m128i? from type
>> ?int?
>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:341:3:
>> error: impli
[dpdk-dev] 1.7.0 release failed to compile on linux
On Thu, Aug 14, 2014 at 06:15:24PM -0400, Aws Ismail wrote: > > On 08/14/2014 05:44 PM, Qiaobing Xie wrote: > >Hi, > > > >I got the following error when I tried to compile 1.7.0 release tarball on > >a Linux box (Ubuntu 12.04/kernel=3.13.0-32-generic, gcc=4.6.3): > > > > > >In file included from > >/home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_osdep.h:41:0, > >from > >/home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_hw.h:31, > >from > >/home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_api.h:31, > >from > >/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.c:38: > >/home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3853:1: > >error: conflicting types for ?skb_set_hash? > >/usr/src/linux-headers-3.13.0-32-generic/include/linux/skbuff.h:740:1: > >note: previous definition of ?skb_set_hash? was here > >make[8]: *** > >[/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.o] > >Error 1 > > > > > >Apparently the version check around 'skb_set_hash' in kcompat.h failed for > >some reason. To work around it, I temporarily commented out that > >'skb_set_hash' definition code and rte_kni complied fine. But it failed > >again in librte_pmd_ixgbe: > > > >= > >In file included from > >/home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:41:0: > >/usr/lib/gcc/x86_64-linux-gnu/4.6/include/nmmintrin.h:31:3: error: #error > >"SSE4.2 instruction set not enabled" > From the error, nmmintrin.h contains built-in CPU calls that depend on the > CPU having SSE4.2 extensions. You need to hunt down the call that is being > executed and see which file is including nmmintrin.h. For example, I have > encountered similar error on a CPU that does not have SSE4.1 and 4.2 > extension and it was trying to call the built-in HW crc hashing call. > instead I switched with rte_jhash.h, the alternative SW implementation of > the hashing call. > > Looking below, nmmintrin.h is being included by the > lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c, from the looks of it. This means that > your defconfig (did you use the default one?) has > CONFIG_RTE_IXGBE_INC_VECTOR value set to "y". try setting it to "n" and see > where the build would stop next looking for another SSE4.x based builtin > call. > This is fixed at the head of the dpdk git tree. The vector rx code has been converted to use builtin intrinsics that can work without sse support in the cpu. The ACL library has a simmilar problem, but we are working at enabling it for all cpus, with runtime updates to faster paths when the executing cpu supports it. > My ultimate recommendation is to update you CPU to a newer one or perform > the build/run on a separate newer CPU. > Building on a newer cpu won't help if you intend to run it on an older cpu. Neil
[dpdk-dev] 1.7.0 release failed to compile on linux
Nail, Thanks for the info... I got v1.7.0 tarball compile after configuring out both IXGBE_INC_VECTOR and ACL. I will take a look at the head. -Q On 8/14/14 7:16 PM, Neil Horman wrote: > On Thu, Aug 14, 2014 at 06:15:24PM -0400, Aws Ismail wrote: >> On 08/14/2014 05:44 PM, Qiaobing Xie wrote: >>> Hi, >>> >>> I got the following error when I tried to compile 1.7.0 release tarball on >>> a Linux box (Ubuntu 12.04/kernel=3.13.0-32-generic, gcc=4.6.3): >>> >>> >>> In file included from >>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_osdep.h:41:0, >>> from >>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_hw.h:31, >> >from >> >/home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/e1000_api.h:31, >> >from >> >/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.c:38: >>> /home/qxie/dpdk-1.7.0/lib/librte_eal/linuxapp/kni/ethtool/igb/kcompat.h:3853:1: >>> error: conflicting types for ?skb_set_hash? >>> /usr/src/linux-headers-3.13.0-32-generic/include/linux/skbuff.h:740:1: >>> note: previous definition of ?skb_set_hash? was here >>> make[8]: *** >>> [/home/qxie/dpdk-1.7.0/build/build/lib/librte_eal/linuxapp/kni/e1000_82575.o] >>> Error 1 >>> >>> >>> Apparently the version check around 'skb_set_hash' in kcompat.h failed for >>> some reason. To work around it, I temporarily commented out that >>> 'skb_set_hash' definition code and rte_kni complied fine. But it failed >>> again in librte_pmd_ixgbe: >>> >>> = >>> In file included from >>> /home/qxie/dpdk-1.7.0/lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c:41:0: >>> /usr/lib/gcc/x86_64-linux-gnu/4.6/include/nmmintrin.h:31:3: error: #error >>> "SSE4.2 instruction set not enabled" >> From the error, nmmintrin.h contains built-in CPU calls that depend on the >> CPU having SSE4.2 extensions. You need to hunt down the call that is being >> executed and see which file is including nmmintrin.h. For example, I have >> encountered similar error on a CPU that does not have SSE4.1 and 4.2 >> extension and it was trying to call the built-in HW crc hashing call. >> instead I switched with rte_jhash.h, the alternative SW implementation of >> the hashing call. >> >> Looking below, nmmintrin.h is being included by the >> lib/librte_pmd_ixgbe/ixgbe_rxtx_vec.c, from the looks of it. This means that >> your defconfig (did you use the default one?) has >> CONFIG_RTE_IXGBE_INC_VECTOR value set to "y". try setting it to "n" and see >> where the build would stop next looking for another SSE4.x based builtin >> call. >> > This is fixed at the head of the dpdk git tree. The vector rx code has been > converted to use builtin intrinsics that can work without sse support in the > cpu. The ACL library has a simmilar problem, but we are working at enabling > it > for all cpus, with runtime updates to faster paths when the executing cpu > supports it. > >> My ultimate recommendation is to update you CPU to a newer one or perform >> the build/run on a separate newer CPU. >> > Building on a newer cpu won't help if you intend to run it on an older cpu. > > Neil >
