[dpdk-dev] [PATCH 1/3] net/mlx5: fix flow director conversion
Flow director must provide the same spec and mask to be sure to be
validated.
Fixes: 4c3e9bcdd52e ("net/mlx5: support flow director")
Cc: sta...@dpdk.org
Signed-off-by: Nelio Laranjeiro
Acked-by: Yongseok Koh
---
drivers/net/mlx5/mlx5_flow.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 13b6483ba..c56ef54cf 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -2656,10 +2656,12 @@ priv_fdir_filter_convert(struct priv *priv,
attributes->items[1] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_IPV4,
.spec = &attributes->l3,
+ .mask = &attributes->l3,
};
attributes->items[2] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_UDP,
.spec = &attributes->l4,
+ .mask = &attributes->l4,
};
break;
case RTE_ETH_FLOW_NONFRAG_IPV4_TCP:
@@ -2677,10 +2679,12 @@ priv_fdir_filter_convert(struct priv *priv,
attributes->items[1] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_IPV4,
.spec = &attributes->l3,
+ .mask = &attributes->l3,
};
attributes->items[2] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_TCP,
.spec = &attributes->l4,
+ .mask = &attributes->l4,
};
break;
case RTE_ETH_FLOW_NONFRAG_IPV4_OTHER:
@@ -2694,6 +2698,7 @@ priv_fdir_filter_convert(struct priv *priv,
attributes->items[1] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_IPV4,
.spec = &attributes->l3,
+ .mask = &attributes->l3,
};
break;
case RTE_ETH_FLOW_NONFRAG_IPV6_UDP:
@@ -2714,10 +2719,12 @@ priv_fdir_filter_convert(struct priv *priv,
attributes->items[1] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_IPV6,
.spec = &attributes->l3,
+ .mask = &attributes->l3,
};
attributes->items[2] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_UDP,
.spec = &attributes->l4,
+ .mask = &attributes->l4,
};
break;
case RTE_ETH_FLOW_NONFRAG_IPV6_TCP:
@@ -2738,10 +2745,12 @@ priv_fdir_filter_convert(struct priv *priv,
attributes->items[1] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_IPV6,
.spec = &attributes->l3,
+ .mask = &attributes->l3,
};
attributes->items[2] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_TCP,
.spec = &attributes->l4,
+ .mask = &attributes->l4,
};
break;
case RTE_ETH_FLOW_NONFRAG_IPV6_OTHER:
@@ -2758,6 +2767,7 @@ priv_fdir_filter_convert(struct priv *priv,
attributes->items[1] = (struct rte_flow_item){
.type = RTE_FLOW_ITEM_TYPE_IPV6,
.spec = &attributes->l3,
+ .mask = &attributes->l3,
};
break;
default:
--
2.11.0
[dpdk-dev] [PATCH 2/3] net/mlx5: fix flow item validation
Two masks were compared instead of verifying the spec was included in the
supported mask.
Fixes: 2097d0d1e2cc ("net/mlx5: support basic flow items and actions")
Cc: sta...@dpdk.org
Signed-off-by: Nelio Laranjeiro
Acked-by: Yongseok Koh
---
drivers/net/mlx5/mlx5_flow.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index c56ef54cf..01e290484 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -529,7 +529,7 @@ mlx5_flow_item_validate(const struct rte_flow_item *item,
}
if (item->mask) {
unsigned int i;
- const uint8_t *spec = item->mask;
+ const uint8_t *spec = item->spec;
for (i = 0; i < size; ++i)
if ((spec[i] | mask[i]) != mask[i])
--
2.11.0
[dpdk-dev] [PATCH 3/3] net/mlx5: support IPv4 time-to-live filter
Signed-off-by: Nelio Laranjeiro
Acked-by: Yongseok Koh
---
drivers/net/mlx5/mlx5_flow.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 01e290484..516f60939 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -353,6 +353,7 @@ static const struct mlx5_flow_items mlx5_flow_items[] = {
.dst_addr = -1,
.type_of_service = -1,
.next_proto_id = -1,
+ .time_to_live = -1,
},
},
.default_mask = &rte_flow_item_ipv4_mask,
--
2.11.0
[dpdk-dev] [PATCH] app/testpmd: do not enable Rx offloads by default
Removed the hardcoded preconfigured Rx offload configuration from
testpmd.
Testers who wish to use these offloads will now have to explicitly
write them in the command-line when running testpmd.
Motivation:
Some PMDs such at the mlx4 may not implement all the offloads.
After the offload API rework assuming no offload is enabled by default,
commit ce17eddefc20 ("ethdev: introduce Rx queue offloads API")
commit cba7f53b717d ("ethdev: introduce Tx queue offloads API") trying
to enable a not supported offload is clearly an error which will cause
configuration failing.
Considering that testpmd is an application to test the PMD, it should
not fail on a configuration which was not explicitly requested.
The behavior of this test application is then turned to an opt-in
model.
Signed-off-by: Moti Haimovsky
---
app/test-pmd/testpmd.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index 5dc8cca..a082352 100644
--- a/app/test-pmd/testpmd.c
+++ b/app/test-pmd/testpmd.c
@@ -305,9 +305,7 @@ struct fwd_engine * fwd_engines[] = {
*/
struct rte_eth_rxmode rx_mode = {
.max_rx_pkt_len = ETHER_MAX_LEN, /**< Default maximum frame length. */
- .offloads = (DEV_RX_OFFLOAD_VLAN_FILTER |
-DEV_RX_OFFLOAD_VLAN_STRIP |
-DEV_RX_OFFLOAD_CRC_STRIP),
+ .offloads = 0,
.ignore_offload_bitfield = 1,
};
--
1.8.3.1
Re: [dpdk-dev] [PATCH 1/2] net/mrvl: switch to the new Rx offload API
Hi Ferruh,
On Mon, Jan 22, 2018 at 05:53:45PM +, Ferruh Yigit wrote:
> On 1/22/2018 12:04 PM, Tomasz Duszynski wrote:
> > Since the old Rx offload API is now depracated
> > update the driver to use the latest one.
> >
> > Signed-off-by: Tomasz Duszynski
>
> <...>
>
> > @@ -1308,6 +1313,42 @@ mrvl_fill_bpool(struct mrvl_rxq *rxq, int num)
> > }
> >
> > /**
> > + * Check whether requested rx queue offloads match port offloads.
> > + *
> > + * @param
> > + * dev Pointer to the device.
> > + * @param
> > + * requested Bitmap of the requested offloads.
> > + *
> > + * @return
> > + * 1 if requested offloads are okay, 0 otherwise.
> > + */
> > +static int
> > +mrvl_rx_queue_offloads_okay(struct rte_eth_dev *dev, uint64_t requested)
> > +{
> > + uint64_t mandatory = dev->data->dev_conf.rxmode.offloads;
> > + uint64_t supported = MRVL_RX_OFFLOADS;
> > + uint64_t unsupported = requested & ~supported;
> > + uint64_t missing = (requested & mandatory) ^ mandatory;
>
> Isn't this same as:
> missing = mandatory & ~requested;
>
> Since "unsupported" use same logic, it can be easier to understand this way.
Fair enough. I'll prepare v2 then.
Thanks for catching this.
>
> Or just putting following comment may be useful enough:
> "mandatory subset of requested subset of supported", assuming it is correct :)
>
> <...>
--
- Tomasz Duszyński
Re: [dpdk-dev] [PATCH] net/mlx5: use PCI BDF as the port name
Hi Yuanhan, On Mon, Jan 22, 2018 at 05:30:06PM +0800, Yuanhan Liu wrote: > It is suggested to use PCI BDF to identify a port for port addition > in OVS-DPDK. While mlx5 has its own naming style: name it by ib dev > name. This breaks the typical OVS DPDK use case and brings more puzzle > to the end users. > > To fix it, this patch changes it to use PCI BDF as the name, too. > Also, a postfix " port %u" is added, just in case their might be more > than 1 port assoicated with a PCI device. Seems only ixgbe, sfc, szedata2 use it whereas ark, bnxt, mlx4, mlx5, mrvl, softnic don't. Why not addressing all drivers at once? > Signed-off-by: Yuanhan Liu Acked-by: Nelio Laranjeiro -- Nélio Laranjeiro 6WIND
[dpdk-dev] [PATCH v2 1/2] net/mrvl: switch to the new Rx offload API
Since the old Rx offload API is now depracated
update the driver to use the latest one.
Signed-off-by: Tomasz Duszynski
---
drivers/net/mrvl/mrvl_ethdev.c | 73 ++
1 file changed, 59 insertions(+), 14 deletions(-)
diff --git a/drivers/net/mrvl/mrvl_ethdev.c b/drivers/net/mrvl/mrvl_ethdev.c
index 4294c56..c313bda 100644
--- a/drivers/net/mrvl/mrvl_ethdev.c
+++ b/drivers/net/mrvl/mrvl_ethdev.c
@@ -94,6 +94,13 @@
/* Memory size (in bytes) for MUSDK dma buffers */
#define MRVL_MUSDK_DMA_MEMSIZE 41943040
+/** Port Rx offload capabilities */
+#define MRVL_RX_OFFLOADS (DEV_RX_OFFLOAD_VLAN_FILTER | \
+ DEV_RX_OFFLOAD_JUMBO_FRAME | \
+ DEV_RX_OFFLOAD_CRC_STRIP | \
+ DEV_RX_OFFLOAD_CHECKSUM)
+
+
static const char * const valid_args[] = {
MRVL_IFACE_NAME_ARG,
MRVL_CFG_ARG,
@@ -302,13 +309,13 @@ mrvl_dev_configure(struct rte_eth_dev *dev)
return -EINVAL;
}
- if (!dev->data->dev_conf.rxmode.hw_strip_crc) {
+ if (!(dev->data->dev_conf.rxmode.offloads & DEV_RX_OFFLOAD_CRC_STRIP)) {
RTE_LOG(INFO, PMD,
"L2 CRC stripping is always enabled in hw\n");
- dev->data->dev_conf.rxmode.hw_strip_crc = 1;
+ dev->data->dev_conf.rxmode.offloads |= DEV_RX_OFFLOAD_CRC_STRIP;
}
- if (dev->data->dev_conf.rxmode.hw_vlan_strip) {
+ if (dev->data->dev_conf.rxmode.offloads & DEV_RX_OFFLOAD_VLAN_STRIP) {
RTE_LOG(INFO, PMD, "VLAN stripping not supported\n");
return -EINVAL;
}
@@ -318,17 +325,17 @@ mrvl_dev_configure(struct rte_eth_dev *dev)
return -EINVAL;
}
- if (dev->data->dev_conf.rxmode.enable_scatter) {
+ if (dev->data->dev_conf.rxmode.offloads & DEV_RX_OFFLOAD_SCATTER) {
RTE_LOG(INFO, PMD, "RX Scatter/Gather not supported\n");
return -EINVAL;
}
- if (dev->data->dev_conf.rxmode.enable_lro) {
+ if (dev->data->dev_conf.rxmode.offloads & DEV_RX_OFFLOAD_TCP_LRO) {
RTE_LOG(INFO, PMD, "LRO not supported\n");
return -EINVAL;
}
- if (dev->data->dev_conf.rxmode.jumbo_frame)
+ if (dev->data->dev_conf.rxmode.offloads & DEV_RX_OFFLOAD_JUMBO_FRAME)
dev->data->mtu = dev->data->dev_conf.rxmode.max_rx_pkt_len -
ETHER_HDR_LEN - ETHER_CRC_LEN;
@@ -1124,11 +1131,8 @@ mrvl_dev_infos_get(struct rte_eth_dev *dev __rte_unused,
info->tx_desc_lim.nb_min = MRVL_PP2_TXD_MIN;
info->tx_desc_lim.nb_align = MRVL_PP2_TXD_ALIGN;
- info->rx_offload_capa = DEV_RX_OFFLOAD_JUMBO_FRAME |
- DEV_RX_OFFLOAD_VLAN_FILTER |
- DEV_RX_OFFLOAD_IPV4_CKSUM |
- DEV_RX_OFFLOAD_UDP_CKSUM |
- DEV_RX_OFFLOAD_TCP_CKSUM;
+ info->rx_offload_capa = MRVL_RX_OFFLOADS;
+ info->rx_queue_offload_capa = MRVL_RX_OFFLOADS;
info->tx_offload_capa = DEV_TX_OFFLOAD_IPV4_CKSUM |
DEV_TX_OFFLOAD_UDP_CKSUM |
@@ -1140,6 +1144,7 @@ mrvl_dev_infos_get(struct rte_eth_dev *dev __rte_unused,
/* By default packets are dropped if no descriptors are available */
info->default_rxconf.rx_drop_en = 1;
+ info->default_rxconf.offloads = DEV_RX_OFFLOAD_CRC_STRIP;
info->max_rx_pktlen = MRVL_PKT_SIZE_MAX;
}
@@ -1308,6 +1313,42 @@ mrvl_fill_bpool(struct mrvl_rxq *rxq, int num)
}
/**
+ * Check whether requested rx queue offloads match port offloads.
+ *
+ * @param
+ * dev Pointer to the device.
+ * @param
+ * requested Bitmap of the requested offloads.
+ *
+ * @return
+ * 1 if requested offloads are okay, 0 otherwise.
+ */
+static int
+mrvl_rx_queue_offloads_okay(struct rte_eth_dev *dev, uint64_t requested)
+{
+ uint64_t mandatory = dev->data->dev_conf.rxmode.offloads;
+ uint64_t supported = MRVL_RX_OFFLOADS;
+ uint64_t unsupported = requested & ~supported;
+ uint64_t missing = mandatory & ~requested;
+
+ if (unsupported) {
+ RTE_LOG(ERR, PMD, "Some Rx offloads are not supported. "
+ "Requested 0x%" PRIx64 " supported 0x%" PRIx64 ".\n",
+ requested, supported);
+ return 0;
+ }
+
+ if (missing) {
+ RTE_LOG(ERR, PMD, "Some Rx offloads are missing. "
+ "Requested 0x%" PRIx64 " missing 0x%" PRIx64 ".\n",
+ requested, missing);
+ return 0;
+ }
+
+ return 1;
+}
+
+/**
* DPDK callback to configure the receive queue.
*
* @param dev
@@ -1319,7 +1360,7 @@ mrvl_fill_bpool(struct mrvl_rxq *rxq, int num)
* @param socket
* NUMA socket on which memory must be allocated.
* @param conf
- * Thresholds paramet
[dpdk-dev] [PATCH v2 0/2] net/mrvl: switch to the new Rx/Tx offloads API
This patch series replaces the old Rx/Tx offload API with the new API. v2: * Follow the same logic for calculating both unsupported and missing flags. Tomasz Duszynski (2): net/mrvl: switch to the new Rx offload API net/mrvl: switch to the new Tx offload API drivers/net/mrvl/mrvl_ethdev.c | 125 ++--- 1 file changed, 106 insertions(+), 19 deletions(-) -- 2.7.4
[dpdk-dev] [PATCH v2 2/2] net/mrvl: switch to the new Tx offload API
Since the old Tx offload API was depracated
update the driver to use the latest one.
Signed-off-by: Tomasz Duszynski
---
drivers/net/mrvl/mrvl_ethdev.c | 52 ++
1 file changed, 47 insertions(+), 5 deletions(-)
diff --git a/drivers/net/mrvl/mrvl_ethdev.c b/drivers/net/mrvl/mrvl_ethdev.c
index c313bda..af8f215 100644
--- a/drivers/net/mrvl/mrvl_ethdev.c
+++ b/drivers/net/mrvl/mrvl_ethdev.c
@@ -100,6 +100,10 @@
DEV_RX_OFFLOAD_CRC_STRIP | \
DEV_RX_OFFLOAD_CHECKSUM)
+/** Port Tx offloads capabilities */
+#define MRVL_TX_OFFLOADS (DEV_TX_OFFLOAD_IPV4_CKSUM | \
+ DEV_TX_OFFLOAD_UDP_CKSUM | \
+ DEV_TX_OFFLOAD_TCP_CKSUM)
static const char * const valid_args[] = {
MRVL_IFACE_NAME_ARG,
@@ -1134,9 +1138,8 @@ mrvl_dev_infos_get(struct rte_eth_dev *dev __rte_unused,
info->rx_offload_capa = MRVL_RX_OFFLOADS;
info->rx_queue_offload_capa = MRVL_RX_OFFLOADS;
- info->tx_offload_capa = DEV_TX_OFFLOAD_IPV4_CKSUM |
- DEV_TX_OFFLOAD_UDP_CKSUM |
- DEV_TX_OFFLOAD_TCP_CKSUM;
+ info->tx_offload_capa = MRVL_TX_OFFLOADS;
+ info->tx_queue_offload_capa = MRVL_TX_OFFLOADS;
info->flow_type_rss_offloads = ETH_RSS_IPV4 |
ETH_RSS_NONFRAG_IPV4_TCP |
@@ -1477,6 +1480,42 @@ mrvl_rx_queue_release(void *rxq)
}
/**
+ * Check whether requested tx queue offloads match port offloads.
+ *
+ * @param
+ * dev Pointer to the device.
+ * @param
+ * requested Bitmap of the requested offloads.
+ *
+ * @return
+ * 1 if requested offloads are okay, 0 otherwise.
+ */
+static int
+mrvl_tx_queue_offloads_okay(struct rte_eth_dev *dev, uint64_t requested)
+{
+ uint64_t mandatory = dev->data->dev_conf.txmode.offloads;
+ uint64_t supported = MRVL_TX_OFFLOADS;
+ uint64_t unsupported = requested & ~supported;
+ uint64_t missing = mandatory & ~requested;
+
+ if (unsupported) {
+ RTE_LOG(ERR, PMD, "Some Rx offloads are not supported. "
+ "Requested 0x%" PRIx64 " supported 0x%" PRIx64 ".\n",
+ requested, supported);
+ return 0;
+ }
+
+ if (missing) {
+ RTE_LOG(ERR, PMD, "Some Rx offloads are missing. "
+ "Requested 0x%" PRIx64 " missing 0x%" PRIx64 ".\n",
+ requested, missing);
+ return 0;
+ }
+
+ return 1;
+}
+
+/**
* DPDK callback to configure the transmit queue.
*
* @param dev
@@ -1488,7 +1527,7 @@ mrvl_rx_queue_release(void *rxq)
* @param socket
* NUMA socket on which memory must be allocated.
* @param conf
- * Thresholds parameters (unused).
+ * Thresholds parameters.
*
* @return
* 0 on success, negative error value otherwise.
@@ -1496,11 +1535,14 @@ mrvl_rx_queue_release(void *rxq)
static int
mrvl_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc,
unsigned int socket,
- const struct rte_eth_txconf *conf __rte_unused)
+ const struct rte_eth_txconf *conf)
{
struct mrvl_priv *priv = dev->data->dev_private;
struct mrvl_txq *txq;
+ if (!mrvl_tx_queue_offloads_okay(dev, conf->offloads))
+ return -ENOTSUP;
+
if (dev->data->tx_queues[idx]) {
rte_free(dev->data->tx_queues[idx]);
dev->data->tx_queues[idx] = NULL;
--
2.7.4
[dpdk-dev] [RFC v2, 1/2] cryptodev: add support to set session private data
Update rte_crypto_op to indicate private data offset.
The application may want to store private data along with the
rte_cryptodev that is transparent to the rte_cryptodev layer.
For e.g., If an eventdev based application is submitting a
rte_cryptodev_sym_session operation and wants to indicate event
information required to construct a new event that will be
enqueued to eventdev after completion of the rte_cryptodev_sym_session
operation. This patch provides a mechanism for the application
to associate this information with the rte_cryptodev_sym_session session.
The application can set the private data using
rte_cryptodev_sym_session_set_private_data() and retrieve it using
rte_cryptodev_sym_session_get_private_data().
Signed-off-by: Abhinandan Gujjar
Signed-off-by: Nikhil Rao
---
Notes:
V2:
1. Removed enum rte_crypto_op_private_data_type
2. Corrected formatting
lib/librte_cryptodev/rte_crypto.h| 8 ++--
lib/librte_cryptodev/rte_cryptodev.h | 32
2 files changed, 38 insertions(+), 2 deletions(-)
diff --git a/lib/librte_cryptodev/rte_crypto.h
b/lib/librte_cryptodev/rte_crypto.h
index 95cf861..14c87c8 100644
--- a/lib/librte_cryptodev/rte_crypto.h
+++ b/lib/librte_cryptodev/rte_crypto.h
@@ -84,8 +84,12 @@ struct rte_crypto_op {
*/
uint8_t sess_type;
/**< operation session type */
-
- uint8_t reserved[5];
+ uint16_t private_data_offset;
+ /**< Offset to indicate start of private data (if any). The private
+* data may be used by the application to store information which
+* should remain untouched in the library/driver
+*/
+ uint8_t reserved[3];
/**< Reserved bytes to fill 64 bits for future additions */
struct rte_mempool *mempool;
/**< crypto operation mempool which operation is allocated from */
diff --git a/lib/librte_cryptodev/rte_cryptodev.h
b/lib/librte_cryptodev/rte_cryptodev.h
index c8fa689..2f4affe 100644
--- a/lib/librte_cryptodev/rte_cryptodev.h
+++ b/lib/librte_cryptodev/rte_cryptodev.h
@@ -1037,6 +1037,38 @@ struct rte_cryptodev_sym_session *
*/
const char *rte_cryptodev_driver_name_get(uint8_t driver_id);
+/**
+ * Set private data for a session.
+ *
+ * @param sessSession pointer allocated by
+ * *rte_cryptodev_sym_session_create*.
+ * @param dataPointer to the private data.
+ * @param sizeSize of the private data.
+ *
+ * @return
+ * - On success, zero.
+ * - On failure, a negative value.
+ */
+int
+rte_cryptodev_sym_session_set_private_data(
+ struct rte_cryptodev_sym_session *sess,
+ void *data,
+ uint16_t size);
+
+/**
+ * Get private data of a session.
+ *
+ * @param sessSession pointer allocated by
+ * *rte_cryptodev_sym_session_create*.
+ *
+ * @return
+ * - On success return pointer to private data.
+ * - On failure returns NULL.
+ */
+void *
+rte_cryptodev_sym_session_get_private_data(
+ const struct rte_cryptodev_sym_session *sess);
+
#ifdef __cplusplus
}
#endif
--
1.9.1
Re: [dpdk-dev] [PATCH v3 7/7] app/testpmd: adjust ethdev port ownership
Hi Konstantin
From: Ananyev, Konstantin, Monday, January 22, 2018 10:49 PM
> Hi Matan,
>
> > -Original Message-
> > From: Matan Azrad [mailto:ma...@mellanox.com]
> > Sent: Monday, January 22, 2018 1:23 PM
> > To: Ananyev, Konstantin ; Gaëtan Rivet
> >
> > Cc: Thomas Monjalon ; Wu, Jingjing
> > ; dev@dpdk.org; Neil Horman
> > ; Richardson, Bruce
> >
> > Subject: RE: [PATCH v3 7/7] app/testpmd: adjust ethdev port ownership
> >
> >
> > Hi
> > From: Ananyev, Konstantin [mailto:konstantin.anan...@intel.com]
> > > Hi lads,
> > >
> > > >
> > > > Hi Matan,
> > > >
> > > > On Fri, Jan 19, 2018 at 01:35:10PM +, Matan Azrad wrote:
> > > > > Hi Konstantin
> > > > >
> > > > > From: Ananyev, Konstantin, Friday, January 19, 2018 3:09 PM
> > > > > > > -Original Message-
> > > > > > > From: Matan Azrad [mailto:ma...@mellanox.com]
> > > > > > > Sent: Friday, January 19, 2018 12:52 PM
> > > > > > > To: Ananyev, Konstantin ;
> > > > > > > Thomas Monjalon ; Gaetan Rivet
> > > > > > ;
> > > > > > > Wu, Jingjing
> > > > > > > Cc: dev@dpdk.org; Neil Horman ;
> > > > > > > Richardson, Bruce
> > > > > > > Subject: RE: [PATCH v3 7/7] app/testpmd: adjust ethdev port
> > > > > > > ownership
> > > > > > >
> > > > > > > Hi Konstantin
> > > > > > >
> > > > > > > From: Ananyev, Konstantin, Friday, January 19, 2018 2:38 PM
> > > > > > > > To: Matan Azrad ; Thomas Monjalon
> > > > > > > > ; Gaetan Rivet
> > > ;
> > > > > > Wu,
> > > > > > > > Jingjing
> > > > > > > > Cc: dev@dpdk.org; Neil Horman ;
> > > > > > > > Richardson, Bruce
> > > > > > > > Subject: RE: [PATCH v3 7/7] app/testpmd: adjust ethdev
> > > > > > > > port ownership
> > > > > > > >
> > > > > > > > Hi Matan,
> > > > > > > >
> > > > > > > > > -Original Message-
> > > > > > > > > From: Matan Azrad [mailto:ma...@mellanox.com]
> > > > > > > > > Sent: Thursday, January 18, 2018 4:35 PM
> > > > > > > > > To: Thomas Monjalon ; Gaetan Rivet
> > > > > > > > > ; Wu, Jingjing
> > > > > > > > >
> > > > > > > > > Cc: dev@dpdk.org; Neil Horman ;
> > > > > > Richardson,
> > > > > > > > > Bruce ; Ananyev, Konstantin
> > > > > > > > >
> > > > > > > > > Subject: [PATCH v3 7/7] app/testpmd: adjust ethdev port
> > > > > > > > > ownership
> > > > > > > > >
> > > > > > > > > Testpmd should not use ethdev ports which are managed by
> > > > > > > > > other DPDK entities.
> > > > > > > > >
> > > > > > > > > Set Testpmd ownership to each port which is not used by
> > > > > > > > > other entity and prevent any usage of ethdev ports which
> > > > > > > > > are not owned by
> > > > > > Testpmd.
> > > > > > > > >
> > > > > > > > > Signed-off-by: Matan Azrad
> > > > > > > > > ---
> > > > > > > > > app/test-pmd/cmdline.c | 89 +++--
>
> > > ---
> > > > > >
> > > > > > > > -
> > > > > > > > > app/test-pmd/cmdline_flow.c | 2 +-
> > > > > > > > > app/test-pmd/config.c | 37 ++-
> > > > > > > > > app/test-pmd/parameters.c | 4 +-
> > > > > > > > > app/test-pmd/testpmd.c | 63
>
> > >
> > > > > > > > > app/test-pmd/testpmd.h | 3 ++
> > > > > > > > > 6 files changed, 103 insertions(+), 95 deletions(-)
> > > > > > > > >
> > > > > > > > > diff --git a/app/test-pmd/cmdline.c
> > > > > > > > > b/app/test-pmd/cmdline.c index
> > > > > > > > > 31919ba..6199c64 100644
> > > > > > > > > --- a/app/test-pmd/cmdline.c
> > > > > > > > > +++ b/app/test-pmd/cmdline.c
> > > > > > > > > @@ -1394,7 +1394,7 @@ struct cmd_config_speed_all {
> > > > > > > > > &link_speed) < 0)
> > > > > > > > > return;
> > > > > > > > >
> > > > > > > > > - RTE_ETH_FOREACH_DEV(pid) {
> > > > > > > > > + RTE_ETH_FOREACH_DEV_OWNED_BY(pid,
> my_owner.id) {
> > > > > > > >
> > > > > > > > Why do we need all these changes?
> > > > > > > > As I understand you changed definition of
> > > > > > > > RTE_ETH_FOREACH_DEV(), so no testpmd should work ok
> > > > > > > > default
> > > (no_owner case).
> > > > > > > > Am I missing something here?
> > > > > > >
> > > > > > > Now, After Gaetan suggestion RTE_ETH_FOREACH_DEV(pid) will
> > > > > > > iterate
> > > > > > over all valid and ownerless ports.
> > > > > >
> > > > > > Yes.
> > > > > >
> > > > > > > Here Testpmd wants to iterate over its owned ports.
> > > > > >
> > > > > > Why? Why it can't just iterate over all valid and ownerless ports?
> > > > > > As I understand it would be enough to fix current problems and
> > > > > > would allow us to avoid any changes in testmpd (which I think
> > > > > > is a good
> > > thing).
> > > > >
> > > > > Yes, I understand that this big change is very daunted, But I
> > > > > think the current a lot of bugs in testpmd(regarding port
> > > > > ownership) even more
> > > > daunted.
> > > > >
> > > > > Look,
> > > > > Testpmd initiates some of its internal databases depends on
> > > > > specific port iteration, In some time someone may take ownership
> > > > > of Testpmd
> > > por
[dpdk-dev] [RFC v2, 2/2] security: add support to set session private data
The application may want to store private data along with the
rte_security that is transparent to the rte_security layer.
For e.g., If an eventdev based application is submitting a
rte_security_session operation and wants to indicate event information
required to construct a new event that will be enqueued to eventdev
after completion of the rte_security operation. This patch provides a
mechanism for the application to associate this information with the
rte_security session. The application can set the private data using
rte_security_session_set_private_data() and retrieve it using
rte_security_session_get_private_data()
Signed-off-by: Abhinandan Gujjar
Signed-off-by: Nikhil Rao
---
lib/librte_security/rte_security.h | 29 +
1 file changed, 29 insertions(+)
diff --git a/lib/librte_security/rte_security.h
b/lib/librte_security/rte_security.h
index d7362f3..1e8d835 100644
--- a/lib/librte_security/rte_security.h
+++ b/lib/librte_security/rte_security.h
@@ -559,6 +559,35 @@ struct rte_security_capability_idx {
rte_security_capability_get(struct rte_security_ctx *instance,
struct rte_security_capability_idx *idx);
+/**
+ * Set private data for a security session.
+ *
+ * @param sesssecurity session
+ * @param datapointer to the private data.
+ * @param sizesize of the private data.
+ *
+ * @return
+ * - On success, zero.
+ * - On failure, a negative value.
+ */
+int
+rte_security_session_set_private_data(struct rte_security_session *sess,
+ void *data,
+ uint16_t size);
+
+/**
+ * Get private data of a security session.
+ *
+ * @param sesssecurity session
+ *
+ * @return
+ * - On success return pointer to private data.
+ * - On failure returns NULL.
+ */
+void *
+rte_security_session_get_private_data(
+ const struct rte_security_session *session);
+
#ifdef __cplusplus
}
#endif
--
1.9.1
Re: [dpdk-dev] [PATCH] build: make compat a universal dependency
On Mon, Jan 22, 2018 at 05:43:06PM +, Luca Boccassi wrote: > On Mon, 2018-01-22 at 15:42 +, Bruce Richardson wrote: > > By making "compat" lib (which consists of a header only) a dependency > > of > > the EAL, we make the header file available to all other libs, drivers > > and > > apps, and thereby make it less work to do ABI versioning. > > > > Signed-off-by: Bruce Richardson > > --- > > drivers/net/bonding/meson.build| 2 +- > > lib/librte_distributor/meson.build | 2 +- > > lib/librte_eal/meson.build | 1 + > > lib/librte_ether/meson.build | 2 +- > > lib/librte_hash/meson.build| 2 +- > > lib/librte_lpm/meson.build | 1 - > > 6 files changed, 5 insertions(+), 5 deletions(-) > > Acked-by: Luca Boccassi > > How's the Meson patchset looking for 18.02? What's on the TODO list? > Since it's going in as experimental, I think the requirements for completeness are not too strict. I'm not aware of any blocking gaps at this stage apart from the release note update which has the patch already submitted. I plan to submit the pull request very soon. For 18.05, the main objective I think is to complete the build, especially all drivers, improve the autotests, [e.g. Harry has some ideas of splitting out the performance tests into a benchmarking target], get the docs building [I see Kevin sent out a patch for that already], and a few cleanups. I'm hoping having the build merged in as experiemental will help encourage maintainers to port their components over, if not already done, and it will certainly help with maintenance - the amount of files added, renamed or moved in a release astounded me! /Bruce
[dpdk-dev] [PATCH] net/virtio-user: fix segfault as features change
Since commit 59fe5e17d93 ("vhost: propagate set features handling error"),
vhost does not allow to set different features without reset.
The virito-user driver fails to reset the device in below commit.
To fix, we send the reset message as stopping the device.
Fixes: c12a26ee209e ("net/virtio-user: fix not properly reset device")
Reported-by: Lei Yao
Reported-by: Tiwei Bie
Signed-off-by: Jianfeng Tan
---
drivers/net/virtio/virtio_user/virtio_user_dev.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c
b/drivers/net/virtio/virtio_user/virtio_user_dev.c
index 7a70c18..3b5f737 100644
--- a/drivers/net/virtio/virtio_user/virtio_user_dev.c
+++ b/drivers/net/virtio/virtio_user/virtio_user_dev.c
@@ -142,6 +142,11 @@ int virtio_user_stop_device(struct virtio_user_dev *dev)
for (i = 0; i < dev->max_queue_pairs; ++i)
dev->ops->enable_qp(dev, i, 0);
+ if (dev->ops->send_request(dev, VHOST_USER_RESET_OWNER, NULL) < 0) {
+ PMD_DRV_LOG(INFO, "Failed to reset the device\n");
+ return -1;
+ }
+
return 0;
}
--
2.7.4
Re: [dpdk-dev] [PATCH] net/virtio-user: fix segfault as features change
Correct yuanhan's email :-)
> -Original Message-
> From: Tan, Jianfeng
> Sent: Tuesday, January 23, 2018 5:47 PM
> To: dev@dpdk.org
> Cc: yuanhan@linux.intel.com; maxime.coque...@redhat.com; Tan,
> Jianfeng
> Subject: [PATCH] net/virtio-user: fix segfault as features change
>
> Since commit 59fe5e17d93 ("vhost: propagate set features handling error"),
> vhost does not allow to set different features without reset.
>
> The virito-user driver fails to reset the device in below commit.
>
> To fix, we send the reset message as stopping the device.
>
> Fixes: c12a26ee209e ("net/virtio-user: fix not properly reset device")
>
> Reported-by: Lei Yao
> Reported-by: Tiwei Bie
> Signed-off-by: Jianfeng Tan
> ---
> drivers/net/virtio/virtio_user/virtio_user_dev.c | 5 +
> 1 file changed, 5 insertions(+)
>
> diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c
> b/drivers/net/virtio/virtio_user/virtio_user_dev.c
> index 7a70c18..3b5f737 100644
> --- a/drivers/net/virtio/virtio_user/virtio_user_dev.c
> +++ b/drivers/net/virtio/virtio_user/virtio_user_dev.c
> @@ -142,6 +142,11 @@ int virtio_user_stop_device(struct virtio_user_dev
> *dev)
> for (i = 0; i < dev->max_queue_pairs; ++i)
> dev->ops->enable_qp(dev, i, 0);
>
> + if (dev->ops->send_request(dev, VHOST_USER_RESET_OWNER,
> NULL) < 0) {
> + PMD_DRV_LOG(INFO, "Failed to reset the device\n");
> + return -1;
> + }
> +
> return 0;
> }
>
> --
> 2.7.4
[dpdk-dev] [PATCH] lib: fix wrong cast
The wrong casts don't cause actual error, but they should conform to C
standard.
Fixes: c261d1431bd8 ("security: introduce security API and framework")
Fixes: b3bbd9e5f265 ("cryptodev: support device independent sessions")
Cc: sta...@dpdk.org
Cc: declan.dohe...@intel.com
Cc: akhil.go...@nxp.com
Signed-off-by: Zhiyong Yang
---
lib/librte_cryptodev/rte_cryptodev.c | 2 +-
lib/librte_security/rte_security.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/librte_cryptodev/rte_cryptodev.c
b/lib/librte_cryptodev/rte_cryptodev.c
index 7726d15f4..8745b6b02 100644
--- a/lib/librte_cryptodev/rte_cryptodev.c
+++ b/lib/librte_cryptodev/rte_cryptodev.c
@@ -1094,7 +1094,7 @@ rte_cryptodev_sym_session_create(struct rte_mempool *mp)
struct rte_cryptodev_sym_session *sess;
/* Allocate a session structure from the session pool */
- if (rte_mempool_get(mp, (void *)&sess)) {
+ if (rte_mempool_get(mp, (void **)&sess)) {
CDEV_LOG_ERR("couldn't get object from session mempool");
return NULL;
}
diff --git a/lib/librte_security/rte_security.c
b/lib/librte_security/rte_security.c
index 6461dba05..d98f46dea 100644
--- a/lib/librte_security/rte_security.c
+++ b/lib/librte_security/rte_security.c
@@ -49,7 +49,7 @@ rte_security_session_create(struct rte_security_ctx *instance,
RTE_FUNC_PTR_OR_ERR_RET(*instance->ops->session_create, NULL);
- if (rte_mempool_get(mp, (void *)&sess))
+ if (rte_mempool_get(mp, (void **)&sess))
return NULL;
if (instance->ops->session_create(instance->device, conf, sess, mp)) {
--
2.13.3
Re: [dpdk-dev] [PATCH] net/mlx5: remmap UAR address for multiple process
Hi Nelio,
> -Original Message-
> From: Nélio Laranjeiro [mailto:nelio.laranje...@6wind.com]
> Sent: Monday, January 22, 2018 10:53 PM
> To: Xueming(Steven) Li
> Cc: Shahaf Shuler ; dev@dpdk.org
> Subject: Re: [PATCH] net/mlx5: remmap UAR address for multiple process
>
> Hi Xueming,
>
> On Fri, Jan 19, 2018 at 11:08:54PM +0800, Xueming Li wrote:
> > UAR(doorbell) is hw resources that have to be same address between
> > primary and secondary process, failed to mmap UAR will make TX packets
> > invisible to HW.
> > Today, UAR address returned from verbs api is mixed in heap and loaded
> > library address space, prone to be occupied in secondary process.
> > This patch reserves a dedicate UAR address space, both primary and
> > secondary process re-mmap UAR pages into this space.
> > Below is a brief picture of dpdk app address space allocation:
> > Before This patch
> > -- --
> > [stack] [stack]
> > [.so, uar, heap][.so, heap]
> > [(empty)] [(empty)]
> > [hugepage] [hugepage]
> > [? others] [? others]
> > [(empty)] [(empty)]
> > [uar]
> > [(empty)]
> > To minimize conflicts, UAR address space comes after hugepage space
> > with an offset to skip potential usage from other drivers.
>
> Seems it is not the case when the memory is contiguous, according to what
> I see in my testpmd /proc//maps:
>
> PMD: mlx5.c:523: mlx5_uar_init_primary(): Reserved UAR address space:
> 0x0x7f4da580
>
> And the fist huge page is at address 0x7f4fa580, new UAR space is
> before and not after.
>
> With this patch I still have the situation described as "before".
>
Your observation is correct, system is allocating address in a high-to-low
manner like stack. UAR address range 0x0x7f4da580 - 0x0x7f4ea580,
4GB size, With another 4G offset, hugepage range start is 0x7f4fa580.
> > Once UAR space reserved successfully, UAR pages are re-mmapped into
> > new area to keep UAR address aligned between primary and secondary
> process.
> >
> > Signed-off-by: Xueming Li
> > ---
> > drivers/net/mlx5/mlx5.c | 107
>
> > drivers/net/mlx5/mlx5.h | 1 +
> > drivers/net/mlx5/mlx5_defs.h| 10
> > drivers/net/mlx5/mlx5_rxtx.h| 3 +-
> > drivers/net/mlx5/mlx5_trigger.c | 7 ++-
> > drivers/net/mlx5/mlx5_txq.c | 51 +--
> > 6 files changed, 163 insertions(+), 16 deletions(-)
> >
> > diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c index
> > fc2d59fee..1539ef608 100644
> > --- a/drivers/net/mlx5/mlx5.c
> > +++ b/drivers/net/mlx5/mlx5.c
> > @@ -39,6 +39,7 @@
> > #include
> > #include
> > #include
> > +#include
> >
> > /* Verbs header. */
> > /* ISO C doesn't support unnamed structs/unions, disabling -pedantic.
> > */ @@ -56,6 +57,7 @@ #include #include
> > #include
> > +#include
> > #include
> >
> > #include "mlx5.h"
> > @@ -466,6 +468,101 @@ mlx5_args(struct mlx5_dev_config *config, struct
> > rte_devargs *devargs)
> >
> > static struct rte_pci_driver mlx5_driver;
> >
> > +/*
> > + * Reserved UAR address space for TXQ UAR(hw doorbell) mapping,
> > +process
> > + * local resource used by both primary and secondary to avoid
> > +duplicate
> > + * reservation.
> > + * The space has to be available on both primary and secondary
> > +process,
> > + * TXQ UAR maps to this area using fixed mmap w/o double check.
> > + */
> > +static void *uar_base;
> > +
> > +/**
> > + * Reserve UAR address space for primary process
> > + *
> > + * @param[in] priv
> > + * Pointer to private structure.
> > + *
> > + * @return
> > + * 0 on success, negative errno value on failure.
> > + */
> > +static int
> > +mlx5_uar_init_primary(struct priv *priv) {
> > + void *addr = (void *)0;
> > + int i;
> > + const struct rte_mem_config *mcfg;
> > +
> > + if (uar_base) { /* UAR address space mapped */
> > + priv->uar_base = uar_base;
> > + return 0;
> > + }
> > + /* find out lower bound of hugepage segments */
> > + mcfg = rte_eal_get_configuration()->mem_config;
> > + for (i = 0; i < RTE_MAX_MEMSEG && mcfg->memseg[i].addr; i++) {
> > + if (addr)
> > + addr = RTE_MIN(addr, mcfg->memseg[i].addr);
> > + else
> > + addr = mcfg->memseg[i].addr;
>
> This if/else is useless as addr is already initialised with the smallest
> possible value.
That's my original code :-) and I always get addr zero then.
Addr here is the lower bound of hugepage, we don't want addr to keep zero.
>
> > + }
> > + /* offset down UAR area */
> > + addr = RTE_PTR_SUB(addr, MLX5_UAR_OFFSET + MLX5_UAR_SIZE);
>
> Seems the error is here, the loops get the address of the memseg with the
> smallest address and then it subtract the UAR size, addr
[dpdk-dev] [PATCH v2] net/virtio-user: fix segfault as features change
Since commit 59fe5e17d930 ("vhost: propagate set features handling error"),
vhost does not allow to set different features without reset.
The virito-user driver fails to reset the device in below commit.
To fix, we send the reset message as stopping the device.
Fixes: c12a26ee209e ("net/virtio-user: fix not properly reset device")
Reported-by: Lei Yao
Reported-by: Tiwei Bie
Signed-off-by: Jianfeng Tan
---
v2: Correct commit reference number format.
drivers/net/virtio/virtio_user/virtio_user_dev.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c
b/drivers/net/virtio/virtio_user/virtio_user_dev.c
index 7a70c18..3b5f737 100644
--- a/drivers/net/virtio/virtio_user/virtio_user_dev.c
+++ b/drivers/net/virtio/virtio_user/virtio_user_dev.c
@@ -142,6 +142,11 @@ int virtio_user_stop_device(struct virtio_user_dev *dev)
for (i = 0; i < dev->max_queue_pairs; ++i)
dev->ops->enable_qp(dev, i, 0);
+ if (dev->ops->send_request(dev, VHOST_USER_RESET_OWNER, NULL) < 0) {
+ PMD_DRV_LOG(INFO, "Failed to reset the device\n");
+ return -1;
+ }
+
return 0;
}
--
2.7.4
[dpdk-dev] [RFC v1 0/1] lib/cryptodev: add support of asymmetric crypto
This an RFC API specification to add support for asymmetric crypto in DPDK cryptodev library. It is derivative of earlier submitted RFC v2 patch series by Umesh Kartha (umesh.kar...@caviumnetworks.com): http://dpdk.org/dev/patchwork/patch/24245/ http://dpdk.org/dev/patchwork/patch/24246/ http://dpdk.org/dev/patchwork/patch/24247/ and *inclusive of review comments on RFC v2* with few minor changes to asymmetric xform structures. Current list of proposed asymmetric algorithm is as per RFC v2. This spec primarily addresses last open review comment regarding queue pair management among different kind of crypto devices. Since, crypto devices can be of different capabilities, such as: - can perform symmetric only operations, - can perform asymmetric only operations, - can do both but may have different ways to manage queue pairs, such as, may have dedicated queues for these or may use all qp to input any kind of operation. Thus, it become key design question how to enable PMDs to do crypto operations with such different capabilities and queue arrangements. So, current specification focuses on defining device capabilities and PMD operations so as to encompass all kind of anticipated crypto devices and intend to invite discussions and feedback on same. Current version only complete with RSA algorithm and associated params. Others algorithm capability and params will be added subsequently. Shally Verma (1): lib/cryptodev: add support of asymmetric crypto lib/librte_cryptodev/Makefile | 1 + lib/librte_cryptodev/rte_crypto.h | 39 +- lib/librte_cryptodev/rte_crypto_asym.h | 949 + lib/librte_cryptodev/rte_cryptodev.c | 287 lib/librte_cryptodev/rte_cryptodev.h | 358 +- lib/librte_cryptodev/rte_cryptodev_pmd.c | 19 +- lib/librte_cryptodev/rte_cryptodev_pmd.h | 154 +++- lib/librte_cryptodev/rte_cryptodev_version.map | 21 + 8 files changed, 1811 insertions(+), 17 deletions(-) create mode 100644 lib/librte_cryptodev/rte_crypto_asym.h -- 1.9.1
Re: [dpdk-dev] [PATCH] build: make compat a universal dependency
O Mon, Jan 22, 2018 at 05:43:06PM +, Luca Boccassi wrote: > On Mon, 2018-01-22 at 15:42 +, Bruce Richardson wrote: > > By making "compat" lib (which consists of a header only) a dependency > > of > > the EAL, we make the header file available to all other libs, drivers > > and > > apps, and thereby make it less work to do ABI versioning. > > > > Signed-off-by: Bruce Richardson > > --- > > drivers/net/bonding/meson.build| 2 +- > > lib/librte_distributor/meson.build | 2 +- > > lib/librte_eal/meson.build | 1 + > > lib/librte_ether/meson.build | 2 +- > > lib/librte_hash/meson.build| 2 +- > > lib/librte_lpm/meson.build | 1 - > > 6 files changed, 5 insertions(+), 5 deletions(-) > > Acked-by: Luca Boccassi > Applied to dpdk-next-build /Bruce
[dpdk-dev] [RFC v1 1/1] lib/cryptodev: add support of asymmetric crypto
From: Shally Verma Add support for asymmetric crypto operations in DPDK lib cryptodev Key feature include: - Only session based asymmetric crypto operations - new get and set APIs for symmetric and asymmetric session private data and other informations - APIs to create, configure and attch queue pair to asymmetric sessions - new capabilities in struct device_info to indicate -- number of dedicated queue pairs available for symmetric and asymmetric operations, if any -- number of asymmetric sessions possible per qp Proposed asymmetric cryptographic operations are: - rsa - dsa - deffie-hellman key pair generation and shared key computation - ecdeffie-hellman - fundamental elliptic curve operations - elliptic curve DSA - modular exponentiation and inversion This patch primarily defines PMD operations and device capabilities to perform asymmetric crypto ops on queue pairs and intend to invite feedbacks on current proposal so as to ensure it encompass all kind of crypto devices with different capabilities and queue pair management. List of TBDs: - Currently, patch only updated for RSA xform and associated params. Other algoritms to be added in subsequent versions. - per-service stats update Signed-off-by: Shally Verma --- It is derivative of RFC v2 asymmetric crypto patch series initiated by Umesh Kartha(mailto:umesh.kar...@caviumnetworks.com): http://dpdk.org/dev/patchwork/patch/24245/ http://dpdk.org/dev/patchwork/patch/24246/ http://dpdk.org/dev/patchwork/patch/24247/ And inclusive of all review comments given on RFC v2. ( See complete discussion thread here: http://dev.dpdk.narkive.com/yqTFFLHw/dpdk-dev-rfc-specifications-for-asymmetric-crypto-algorithms#post12) Some of the RFCv2 Review comments pending for closure: > " [Fiona] The count fn isn't used at all for sym - probably no need to add > for asym better instead to remove the sym fn." It is still present in dpdk-next-crypto for sym, so what has been decision on it? >"[Fiona] if each qp can handle only a specific service, i.e. a subset off the >capabilities Indicated by the device capability list, there's a need for a new API to query the capability of a qp." Current proposal doesn???t distinguish between device capability and qp capability. It rather leave such differences handling internal to PMDs. Thus no capability or API added for qp in current version. It is subject to revisit based on review feedback on current proposal. - Sessionless Support. Current proposal only support Session-based because: 1. All one-time setup i.e. algos and associated params, such as, public-private keys or modulus length can be done in control path using session-init API 2. it???s an easier way to dedicate qp to do specific service (using queue_pair_attach()) which cannot be case in sessionless 3. Couldn???t find any significant advantage going sessionless way. Also existing most of PMDs are session-based. It could be added in subsequent versions, if requirement is identified, based on review comment on this RFC. Summary --- This section provides an overview of key feature enabled in current specification. It comprise of key design challenges as have been identified on RFCv2 and summary description of new interfaces and definitions added to handle same. Description --- This API set assumes that the max_nb_queue_pairs on a device can be allocated to any mix of sym or asym. Some devices may have a fixed max per service. Thus, rte_cryptodev_info is updated with max_sym_nb_queues and max_asym_nb_queues with rule: max_nb_queue_pair = max_nb_sym_qp + max_nb_asym_qp. If device has no restrictions on qp to be used per service, such PMDs can leave max_nb_sym_qp = max_nb_asym_qp = 0. In such case, application can setup any of the service upto limit defined by max_nb_queue_pair. Here, max_nb_sym_qp and max_nb_asym_qp, if non-zero, just define limit on qp which are available for each service and *are not* ids to be used during qp setup and enqueue i.e. if device supports both symmetric and asymmetric with n qp, then any of them can be configured for symmetric or asymmetric subject to limit defined by max_nb_sym_qp and max_nb_asym_qp. For example, if device has 6 queues and 5 for symmetric and 1 for asymmetric that imply application can setup only 1 out of any 6 qp for asymmetric and rest for symmetric. Additionally, application can dedicate qp to perform specific service via optional queue_pair_attach_sym/asym_session() API. Except the ones mentioned above, specification doesn't define any restrictions on enqueue/dequeue API for different crypto services. Application can enqueue both symmetric and asymmetric operations to same device and qp if device info indicate support of both symmetric and asymmetric and qp is not dedicated to any. However in practice, it could result in head-of-line blocking due to the asym ops t
Re: [dpdk-dev] [PATCH] keepalive: fix keepalive state alignment
On 22/01/2018 18:20, Andriy Berestovskyy wrote: [..] On Fri, Jan 19, 2018 at 6:31 PM, Van Haaren, Harry wrote: These changes do reduce false-sharing however is there actually a performance benefit? A lot of cache space will be taken up if each core requires its own cache line, which will reduce performance again.. it's a tradeoff. [..] 2. The original code (prior e70a61ad50ab "keepalive: export states") had each element aligned to the cache line, not the whole array. Aligning each flag element was the original intention, so I see no issue in restoring it. The monitoring core only reads the entries within state_flags for which the corresponding active_core is set, so ultimately the trade-off in cache line usage is one made by the application when it decides which cores need monitoring. ..Remy
Re: [dpdk-dev] [PATCH] doc: add i40e update
> -Original Message- > From: Xing, Beilei > Sent: Thursday, January 18, 2018 2:24 AM > To: Mcnamara, John > Cc: dev@dpdk.org; sta...@dpdk.org > Subject: [PATCH] doc: add i40e update > > Update release note to declare MPLSoUDP/MPLSoGRE/ GTP-U/GTP- > C/PPPoE/PPPoL2TP steering support in i40e driver. > > Signed-off-by: Beilei Xing > Cc: sta...@dpdk.org Acked-by: John McNamara
[dpdk-dev] [PATCH] net/i40e: add device args to disable global configuration
DPDK i40e PMD will modify some global registers during initialization
and post initialization, there'll be impact during use of 700 series
Ethernet Adapter with both Linux kernel and DPDK PMD.
This patch is to add device parameters to disable global configuration.
Signed-off-by: Beilei Xing
---
This patch is based on 16.11.4 LTS version.
drivers/net/i40e/i40e_ethdev.c | 208 -
drivers/net/i40e/i40e_ethdev.h | 2 +
2 files changed, 165 insertions(+), 45 deletions(-)
diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
index 0835c2d..0df5cba 100644
--- a/drivers/net/i40e/i40e_ethdev.c
+++ b/drivers/net/i40e/i40e_ethdev.c
@@ -933,6 +933,67 @@ config_floating_veb(struct rte_eth_dev *dev)
#define I40E_L2_TAGS_S_TAG_SHIFT 1
#define I40E_L2_TAGS_S_TAG_MASK I40E_MASK(0x1, I40E_L2_TAGS_S_TAG_SHIFT)
+#define ETH_I40E_DISABLE_GLOBAL_CFG"disable-global-cfg"
+RTE_PMD_REGISTER_PARAM_STRING(net_i40e,
+ ETH_I40E_DISABLE_GLOBAL_CFG "=0|1");
+
+static int
+i40e_parse_global_cfg_handler(__rte_unused const char *key,
+ const char *value,
+ void *opaque)
+{
+ struct i40e_pf *pf;
+ unsigned long dis_global_cfg;
+ char *end;
+
+ pf = (struct i40e_pf *)opaque;
+
+ errno = 0;
+ dis_global_cfg = strtoul(value, &end, 10);
+ if (errno != 0 || end == value || *end != 0) {
+ PMD_DRV_LOG(WARNING, "Wrong global configuration");
+ return -(EINVAL);
+ }
+
+ if (dis_global_cfg == 1 || dis_global_cfg == 0)
+ pf->dis_global_cfg = (bool)dis_global_cfg;
+ else
+ PMD_DRV_LOG(WARNING, "%s must be 1 or 0,",
+ "enable global configuration by default."
+ ETH_I40E_DISABLE_GLOBAL_CFG);
+ return 0;
+}
+
+static int
+i40e_disable_global_cfg(struct rte_eth_dev *dev)
+{
+ struct i40e_pf *pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+ struct rte_pci_device *pci_dev = dev->pci_dev;
+ static const char *valid_keys[] = {
+ ETH_I40E_DISABLE_GLOBAL_CFG, NULL};
+ struct rte_kvargs *kvlist;
+
+ /* Enable global configuration by default */
+ pf->dis_global_cfg = false;
+
+ if (!pci_dev->device.devargs)
+ return 0;
+
+ kvlist = rte_kvargs_parse(pci_dev->device.devargs->args, valid_keys);
+ if (!kvlist)
+ return -EINVAL;
+
+ if (rte_kvargs_count(kvlist, ETH_I40E_DISABLE_GLOBAL_CFG) > 1)
+ PMD_DRV_LOG(WARNING, "More than one argument \"%s\" and only "
+ "the first invalid or last valid one is used !",
+ ETH_I40E_DISABLE_GLOBAL_CFG);
+
+ rte_kvargs_process(kvlist, ETH_I40E_DISABLE_GLOBAL_CFG,
+ i40e_parse_global_cfg_handler, pf);
+ rte_kvargs_free(kvlist);
+ return 0;
+}
+
static int
eth_i40e_dev_init(struct rte_eth_dev *dev)
{
@@ -982,6 +1043,9 @@ eth_i40e_dev_init(struct rte_eth_dev *dev)
hw->bus.func = pci_dev->addr.function;
hw->adapter_stopped = 0;
+ /* Check if need to disable global registers configuration */
+ i40e_disable_global_cfg(dev);
+
/* Make sure all is clean before doing PF reset */
i40e_clear_hw(hw);
@@ -1008,7 +1072,8 @@ eth_i40e_dev_init(struct rte_eth_dev *dev)
* software. It should be removed once issues are fixed
* in NVM.
*/
- i40e_GLQF_reg_init(hw);
+ if (!pf->dis_global_cfg)
+ i40e_GLQF_reg_init(hw);
/* Initialize the input set for filters (hash and fd) to default value
*/
i40e_filter_input_set_init(pf);
@@ -1104,11 +1169,14 @@ eth_i40e_dev_init(struct rte_eth_dev *dev)
i40e_set_fc(hw, &aq_fail, TRUE);
/* Set the global registers with default ether type value */
- ret = i40e_vlan_tpid_set(dev, ETH_VLAN_TYPE_OUTER, ETHER_TYPE_VLAN);
- if (ret != I40E_SUCCESS) {
- PMD_INIT_LOG(ERR, "Failed to set the default outer "
-"VLAN ether type");
- goto err_setup_pf_switch;
+ if (!pf->dis_global_cfg) {
+ ret = i40e_vlan_tpid_set(dev, ETH_VLAN_TYPE_OUTER,
+ETHER_TYPE_VLAN);
+ if (ret != I40E_SUCCESS) {
+ PMD_INIT_LOG(ERR, "Failed to set the default outer "
+"VLAN ether type");
+ goto err_setup_pf_switch;
+ }
}
/* PF setup, which includes VSI setup */
@@ -2743,11 +2811,17 @@ i40e_vlan_tpid_set(struct rte_eth_dev *dev,
uint16_t tpid)
{
struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ struct i40e_pf *pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private);
uint64_t
Re: [dpdk-dev] [PATCH v5] doc: add queue region feature info to release notes
> -Original Message- > From: Zhao1, Wei > Sent: Monday, January 22, 2018 5:19 AM > To: dev@dpdk.org > Cc: Mcnamara, John ; tho...@monjalon.net; > sta...@dpdk.org; Zhao1, Wei > Subject: [PATCH v5] doc: add queue region feature info to release notes > > This patch add inforation about i40e queue region realted to release > notes, it has been missed before in v17.11 release notes. This feature has > been implemented in v17.11. > > ... > +Queue region configuration > +~~~ The underline for a section header should match the length of the text and should be followed by a blank line, line in my suggested changes. See also: http://dpdk.org/doc/guides/contributing/documentation.html#section-headers However, this isn't an error so: Acked-by: John McNamara
Re: [dpdk-dev] [PATCH] lib: fix wrong cast
Acked-by: Akhil Goyal
[dpdk-dev] OVS+DPDK: deadlock and race condtion, which leads OVS deadlock or crash
Hi Yuanha and all,
Thanks for review. I change the title and let more people involved.
The issues can be reproduced easily with OVS.
The steps are below:
1) virsh start vm
2) ovs-vsctl add port
3) delete the port via ovs-vsctl del-port.
4) shutdown VM via vrish destroy.
I run a script to repeatly do such job for 2 VM, and it needs about 1 hour to
reproduce it.
However, I found there are more issues (I can reproduce it):
1) see below code
static void
vhost_user_read_cb(int connfd, void *dat, int *remove)
{
struct vhost_user_connection *conn = dat;
struct vhost_user_socket *vsocket = conn->vsocket;
int ret;
ret = vhost_user_msg_handler(conn->vid, connfd);
if (ret < 0) {
close(connfd);
*remove = 1;
vhost_destroy_device(conn->vid);
pthread_mutex_lock(&vsocket->conn_mutex);
TAILQ_REMOVE(&vsocket->conn_list, conn, next);
pthread_mutex_unlock(&vsocket->conn_mutex);
free(conn);
-> [Zhike Wang] we can see that the
vsocket does not exist in conn_list or reconn_list at this short time slot. So
in this time, if rte_vhost_driver_unregister() is called, vsocket is freed. So
below vhost_user_create_client() may use raw pointer, and lead to crash.
if (vsocket->reconnect)
vhost_user_create_client(vsocket);
}
}
2) regarding the above issue, I think we need one mutex to lock conn_list and
reconn_list, so the vsocket can in either conn_list or reconn_list, not both or
none. So I use vhost_user.mutex()
diff --git a/dpdk-16.11.3/lib/librte_vhost/fd_man.c
b/dpdk-16.11.3/lib/librte_vhost/fd_man.c
@@ -233,6 +234,7 @@
poll(pfdset->rwfds, numfds, 1000 /* millisecs */);
need_shrink = 0;
+pthread_mutex_lock(&vhost_user.mutex);
for (i = 0; i < numfds; i++) {
pthread_mutex_lock(&pfdset->fd_mutex);
@@ -264,7 +266,10 @@
rcb(fd, dat, &remove1);
if (wcb && pfd->revents & (POLLOUT | FDPOLLERR))
wcb(fd, dat, &remove2);
+
+ pthread_mutex_lock(&pfdset->fd_mutex);
pfdentry->busy = 0;
+ pthread_mutex_unlock(&pfdset->fd_mutex);
/*
* fdset_del needs to check busy flag.
* We don't allow fdset_del to be called in callback
@@ -285,5 +290,6 @@
if (need_shrink)
fdset_shrink(pfdset);
+pthread_mutex_unlock(&vhost_user.mutex);
@@ -390,6 +395,8 @@ struct vhost_user_reconnect_list {
struct vhost_user_reconnect *reconn, *next;
while (1) {
+
+ pthread_mutex_lock(&vhost_user.mutex);
pthread_mutex_lock(&reconn_list.mutex);
/*
@@ -417,11 +424,18 @@ struct vhost_user_reconnect_list {
"%s: connected\n", reconn->vsocket->path);
vhost_user_add_connection(reconn->fd, reconn->vsocket);
remove_fd:
pthread_mutex_unlock(&reconn_list.mutex);
+ pthread_mutex_unlock(&vhost_user.mutex);
sleep(1);
}
Then I met a deadlock (another thread enters rte_vhost_driver_unregister, and
waiting for vhost_user.mutex.)
(gdb) bt
#0 0x7febfaf8ebcd in poll () from /lib64/libc.so.6
#1 0x7fec029b6836 in poll (__timeout=, __nfds=2,
__fds=0x7febe80008c0) at /usr/include/bits/poll2.h:46
#2 time_poll (pollfds=pollfds@entry=0x7febe80008c0, n_pollfds=2,
handles=handles@entry=0x0, timeout_when=4098990819,
elapsed=elapsed@entry=0x7febf7450290) at lib/timeval.c:305
#3 0x7fec0299afea in poll_block () at lib/poll-loop.c:364
#4 0x7fec0297da25 in ovsrcu_synchronize () at lib/ovs-rcu.c:231
#5 0x7fec029e9477 in destroy_device (vid=) at
lib/netdev-dpdk.c:2651
#6 0x7fec0195db04 in vhost_destroy_device (vid=) at
/lib/librte_vhost/vhost.c:277
#7 0x7fec0195cbdc in vhost_user_read_cb (connfd=,
dat=0x7febc80008c0, remove=0x7febf7451420) at /lib/librte_vhost/socket.c:269
#8 0x7fec0195c200 in fdset_event_dispatch
(pfdset=pfdset@entry=0x7fec01b70180 ) at
/lib/librte_vhost/fd_man.c:266
#9 0x7fec0195d647 in rte_vhost_driver_session_start () at
/lib/librte_vhost/socket.c:714
#10 0x7fec029e661b in start_vhost_loop (dummy=) at
lib/netdev-dpdk.c:2736
#11 0x7fec0297f796 in ovsthread_wrapper (aux_=) at
lib/ovs-thread.c:348
#12 0x7febfb9b6dc5 in start_thread () from /lib64/libpthread.so.0
#13 0x7febfaf9921d in clone () from /lib64/libc.so.6
(gdb) n
So now I believe it is hard to fix the issues insides the modules, and I would
like to present identified issues, and want to hear your proposal or fix.
Br,
Zhike Wang
-Original Message-
From:
Re: [dpdk-dev] [PATCH] keepalive: fix keepalive state alignment
> From: Horton, Remy > Sent: Tuesday, January 23, 2018 10:17 AM > To: Andriy Berestovskyy ; Van Haaren, Harry > > Cc: dev@dpdk.org > Subject: Re: [dpdk-dev] [PATCH] keepalive: fix keepalive state alignment > > > On 22/01/2018 18:20, Andriy Berestovskyy wrote: > [..] > > On Fri, Jan 19, 2018 at 6:31 PM, Van Haaren, Harry > > wrote: > >> These changes do reduce false-sharing however is there actually a > >> performance benefit? A lot of cache space will be taken up if each > >> core requires its own cache line, which will reduce performance > >> again.. it's a tradeoff. > [..] > > 2. The original code (prior e70a61ad50ab "keepalive: export states") > > had each element aligned to the cache line, not the whole array. > > Aligning each flag element was the original intention, so I see no issue > in restoring it. The monitoring core only reads the entries within > state_flags for which the corresponding active_core is set, so > ultimately the trade-off in cache line usage is one made by the > application when it decides which cores need monitoring. > > ..Remy No objection here - just making sure the change was intentional and the effects are considered. @Remy as you're the Keepalive maintainer I'll leave Acking to you :) -Harry
Re: [dpdk-dev] [PATCH] doc: add release note entry for meson build
> -Original Message- > From: Richardson, Bruce > Sent: Monday, January 22, 2018 3:14 PM > To: Mcnamara, John ; Kovacevic, Marko > > Cc: bl...@debian.org; dev@dpdk.org; tho...@monjalon.net; Richardson, Bruce > > Subject: [PATCH] doc: add release note entry for meson build > Acked-by: John McNamara
Re: [dpdk-dev] [[PATCH v5] 5/5] doc: Add ABI __experimental tag documentation
> -Original Message- > From: Neil Horman [mailto:nhor...@tuxdriver.com] > Sent: Monday, January 22, 2018 1:48 AM > To: dev@dpdk.org > Cc: Neil Horman ; Thomas Monjalon > ; Mcnamara, John ; > Richardson, Bruce > Subject: [[PATCH v5] 5/5] doc: Add ABI __experimental tag documentation > > Document the need to add the __experimental tag to appropriate functions > > Signed-off-by: Neil Horman > CC: Thomas Monjalon > CC: "Mcnamara, John" > CC: Bruce Richardson > ... > +Note that marking an API as experimental is a multi step process. To > +mark an API as experimental, the symbols which are desired to be > +exported must be placed in an EXPERIMENTAL version block in the > +corresponding libraries' version map script. Secondly, the > +corresponding definitions of those exported functions, and their > +forward declarations (in the development header files), must be marked > +with the __rte_experimental tag (see rte_compat.h). The DPDK build > +makefiles perform a check to ensure that the map file and the C code > +reflect the same list of symbols. This check can be circumvented by > defining ALLOW_EXPERIMENTAL_API during compilation in the corresponding > library Makefile. > + > +In addition to tagging the code with __rte_experimental, the doxygen > +markup must also contain the EXPERIMENTAL string, and the MAINTAINER > +file should note that the library contains EXPERIMENTAL APIs. > + > ABI versions, once released, are available until such time as their > deprecation has been noted in the Release Notes for at least one major > release cycle. For example consider the case where the ABI for DPDK 2.0 > has been > -- > 2.14.3 Thanks for the update, and this work in general. The rendered docs would probably look a better better with __rte_experimental and ALLOW_EXPERIMENTAL_API is fixed width backticks ``var`` but that is only a "nice to have" so no need for a respin. Acked-by: John McNamara
Re: [dpdk-dev] [PATCH v2] doc: add deprecation notice for memory hotplug changes
> -Original Message- > From: Burakov, Anatoly > Sent: Thursday, January 18, 2018 10:32 AM > To: dev@dpdk.org > Cc: Neil Horman ; Mcnamara, John > ; Kovacevic, Marko > Subject: [PATCH v2] doc: add deprecation notice for memory hotplug changes > > Due to coming changes outlined in memory hotplug RFC, there will be > several API/ABI changes. > > Signed-off-by: Anatoly Burakov Acked-by: John McNamara
Re: [dpdk-dev] [PATCH] doc: add ABI change notice for numa_node_count in eal
> -Original Message- > From: Burakov, Anatoly > Sent: Tuesday, January 16, 2018 5:54 PM > To: dev@dpdk.org > Cc: Neil Horman ; Mcnamara, John > ; Kovacevic, Marko > Subject: [PATCH] doc: add ABI change notice for numa_node_count in eal > > There will be a new function added in v18.05 that will return number of > detected sockets, which will change the ABI. > > Signed-off-by: Anatoly Burakov > --- > doc/guides/rel_notes/deprecation.rst | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/doc/guides/rel_notes/deprecation.rst > b/doc/guides/rel_notes/deprecation.rst > index 13e8543..9662150 100644 > --- a/doc/guides/rel_notes/deprecation.rst > +++ b/doc/guides/rel_notes/deprecation.rst > @@ -8,6 +8,8 @@ API and ABI deprecation notices are to be posted here. > Deprecation Notices > --- > > +* eal: new ``numa_node_count`` member will be added to ``rte_config`` > +structure in v18.05. > * eal: several API and ABI changes are planned for ``rte_devargs`` in > v18.02. In general it is best to leave a blank line between the bullet points. But that doesn't affect the rendering so: Acked-by: John McNamara
Re: [dpdk-dev] About : Enable optional dequeue zero copy for vHost User
Hi, For the meantime this feature is proposed as ‘experimental’ for OVS DPDK. Unless you are transmitting to a NIC, you don’t need to set the n_txq_desc. My testing has been only with a DPDK driver in the guest. Have you tried that option? Thanks, Ciara From: liyang07 [mailto:liyan...@corp.netease.com] Sent: Wednesday, January 17, 2018 10:41 AM To: Loftus, Ciara Cc: dev@dpdk.org Subject: About : Enable optional dequeue zero copy for vHost User Hi Ciara, I am tesing the function of "vHost dequeue zero copy" for vm2vm on a host, and I have some problems: 1. The networking is OK before run iperf, I can ping successed from vm1 to vm2，but after run iperf, the networking between vm1 and vm2 is down;(I think n_txq_desc cause the problem) 2. I know the limitation about n_txq_desc, but I cannot set the n_txq_desc for dpdkport while the vms on the same host, because there is no dpdkports work fow my testing; Thus, how can I resolve it, thanks.
Re: [dpdk-dev] [PATCH] doc: add release note entry for meson build
On Tue, Jan 23, 2018 at 10:28:42AM +, Mcnamara, John wrote: > > > > -Original Message- > > From: Richardson, Bruce > > Sent: Monday, January 22, 2018 3:14 PM > > To: Mcnamara, John ; Kovacevic, Marko > > > > Cc: bl...@debian.org; dev@dpdk.org; tho...@monjalon.net; Richardson, Bruce > > > > Subject: [PATCH] doc: add release note entry for meson build > > > > > Acked-by: John McNamara > Applied to dpdk-next-build /Bruce
Re: [dpdk-dev] About : Enable optional dequeue zero copy for vHost User
Hi Which version of dpdk you are using? I have some fixes for dequeue zero copy and now in 18.02-rc1, you can try 18.02-r1. Cheers JJ > -Original Message- > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Loftus, Ciara > Sent: Tuesday, January 23, 2018 6:46 PM > To: liyang07 > Cc: dev@dpdk.org; d...@openvswitch.org > Subject: Re: [dpdk-dev] About : Enable optional dequeue zero copy for vHost > User > > Hi, > > For the meantime this feature is proposed as ‘experimental’ for OVS DPDK. > Unless you are transmitting to a NIC, you don’t need to set the n_txq_desc. > My testing has been only with a DPDK driver in the guest. Have you tried that > option? > > Thanks, > Ciara > > From: liyang07 [mailto:liyan...@corp.netease.com] > Sent: Wednesday, January 17, 2018 10:41 AM > To: Loftus, Ciara > Cc: dev@dpdk.org > Subject: About : Enable optional dequeue zero copy for vHost User > > > Hi Ciara, > > I am tesing the function of "vHost dequeue zero copy" for vm2vm on a > host, and I have some problems: > > 1. The networking is OK before run iperf, I can ping successed from vm1 > to vm2，but after run iperf, the networking between vm1 and vm2 is down;(I > think n_txq_desc cause the problem) > > 2. I know the limitation about n_txq_desc, but I cannot set the > n_txq_desc for dpdkport while the vms on the same host, because there is > no dpdkports work fow my testing; > > > > Thus, how can I resolve it, thanks. > > >
Re: [dpdk-dev] [PATCH v2 0/3] Fixes for QAT PMD
> -Original Message- > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Tomasz Jozwiak > Sent: Monday, January 22, 2018 4:28 PM > To: Trahe, Fiona ; Jain, Deepak K > ; Griffin, John > Cc: dev@dpdk.org; Jozwiak, TomaszX > Subject: [dpdk-dev] [PATCH v2 0/3] Fixes for QAT PMD > > v2: > - Previous patch has been split into 3 as below: > > Tomasz Jozwiak (3): > crypto/qat: fix out-of-bounds compiler error > crypto/qat: fix typo in error message > crypto/qat: fix parameter type > > drivers/crypto/qat/qat_crypto.c | 26 +++--- > 1 file changed, 15 insertions(+), 11 deletions(-) > > -- > 2.7.4 Applied to dpdk-next-crypto. Thanks, Pablo
Re: [dpdk-dev] [RFC v3 1/1] lib: add compressdev API
Hi Fiona
> -Original Message-
> From: Trahe, Fiona [mailto:fiona.tr...@intel.com]
> Sent: 19 January 2018 17:30
> To: Verma, Shally ; dev@dpdk.org;
> akhil.go...@nxp.com
> Cc: Challa, Mahipal ; Athreya, Narayana
> Prasad ; De Lara Guarch, Pablo
> ; Gupta, Ashish
> ; Sahu, Sunila ;
> Jain, Deepak K ; Hemant Agrawal
> ; Roy Pledge ; Youri
> Querry ; Ahmed Mansour
> ; Trahe, Fiona
> Subject: RE: [RFC v3 1/1] lib: add compressdev API
>
> Hi Shally,
>
> > -Original Message-
> > From: Verma, Shally [mailto:shally.ve...@cavium.com]
> > Sent: Thursday, January 18, 2018 12:54 PM
> > To: Trahe, Fiona ; dev@dpdk.org
> > Cc: Challa, Mahipal ; Athreya, Narayana
> Prasad
> > ; De Lara Guarch, Pablo
> ;
> > Gupta, Ashish ; Sahu, Sunila
> ; Jain, Deepak K
> > ; Hemant Agrawal
> ; Roy Pledge
> > ; Youri Querry ;
> Ahmed Mansour
> >
> > Subject: RE: [RFC v3 1/1] lib: add compressdev API
> >
> > Hi Fiona
> >
> > While revisiting this, we identified few questions and additions. Please see
> them inline.
> >
> >
> > > -Original Message-
> > > From: Trahe, Fiona [mailto:fiona.tr...@intel.com]
> > > Sent: 15 December 2017 23:19
> > > To: dev@dpdk.org; Verma, Shally
> > > Cc: Challa, Mahipal ; Athreya, Narayana
> > > Prasad ;
> > > pablo.de.lara.gua...@intel.com; fiona.tr...@intel.com
> > > Subject: [RFC v3 1/1] lib: add compressdev API
> > >
> > > Signed-off-by: Trahe, Fiona
> > > ---
> >
> > //snip
> >
> > > +
> > > +int
> > > +rte_compressdev_queue_pair_setup(uint8_t dev_id, uint16_t
> > > queue_pair_id,
> > > + uint32_t max_inflight_ops, int socket_id)
> >
> > [Shally] Is max_inflights_ops different from nb_streams_per_qp in struct
> rte_compressdev_info?
> > I assume they both carry same purpose. If yes, then it will be better to use
> single naming convention to
> > avoid confusion.
> [Fiona] No, I think they have different purposes.
> max_inflight_ops should be used to configure the qp with the number of ops
> the application expects to be able to submit to the qp before it needs to poll
> for a response. It can be configured differently for each qp. In the QAT case
> it
> dictates the depth of the qp created, it may have different implications on
> other PMDs.
> nb_sessions_per_qp and nb_streams_per_qp are limitations the devices
> reports and are same for all qps on the device. QAT doesn't have those
> limitations and so would report 0, however I assumed they may be necessary
> for other devices.
> This assumption is based on the patch submitted by NXP to cryptodev in Feb
> 2017
> http://dpdk.org/ml/archives/dev/2017-March/060740.html
> I also assume these are not necessarily the max number of sessions in ops on
> the qp at a given time, but the total number attached, i.e. if the device has
> this limitation then sessions must be attached to qps, and presumably
> reserve some resources. Being attached doesn't imply there is an op on the
> qp at that time using that session. So it's not to relating to the inflight op
> count, but to the number of sessions attached/detached to the qp.
> Including Akhil on the To list, maybe NXP can confirm if these params are
> needed.
[Shally] Ok. Then let's wait for NXP to confirm on this requirement as
currently spec doesn't have any API to attach
queue_pair_to_specific_session_or_stream as cryptodev.
But then how application could know limit on max_inflight_ops supported on a
qp? As it can pass any random number during qp_setup().
Do you believe we need to add a capability field in dev_info to indicate limit
on max_inflight_ops?
Thanks
Shally
>
>
> > Also, is it optional API? Like Is this a valid use case?:
> > dev_configure() --> dev_start() --> qp_start() --> enqueue/dequeue() -->
> qp_stop() --> dev_stop() -->
> > dev_close()?
> [Fiona] I don't think it should be optional as some PMDs need to allocate
> resources based on the setup data passed in on this API.
>
> > //snip
> >
> > > +
> > > +#define RTE_COMPRESSDEV_PMD_NAME_ARG
> > > ("name")
> > > +#define RTE_COMPRESSDEV_PMD_MAX_NB_QP_ARG
> > > ("max_nb_queue_pairs")
> > > +#define RTE_COMPRESSDEV_PMD_SOCKET_ID_ARG
> ("socket_id")
> > > +
> >
> > [Shally] Need to define argument macro for max_nb_session_per_qp and
> max_nb_streams_per_qp as
> > well
> [Fiona] ok
>
> > > +
> > > +static const char * const compressdev_pmd_valid_params[] = {
> > > + RTE_COMPRESSDEV_PMD_NAME_ARG,
> > > + RTE_COMPRESSDEV_PMD_MAX_NB_QP_ARG,
> > > + RTE_COMPRESSDEV_PMD_SOCKET_ID_ARG
> > > +};
> >
> > [Shally] Likewise, array need to be updated with other mentioned two
> arguments
> Fiona] ok
>
>
> > > +
> > > +/**
> > > + * @internal
> > > + * Initialisation parameters for comp devices
> > > + */
> > > +struct rte_compressdev_pmd_init_params {
> > > + char name[RTE_COMPRESSDEV_NAME_MAX_LEN];
> > > + size_t private_data_size;
> > > + int socket_id;
> > > + unsigned int max_nb_queue_pairs;
> >
> > [Shally] And this also need to be updated with max_nb_sessions_per_qp
> and max_streams_per_qp
>
[dpdk-dev] [PATCH] app/test-crypto-perf: fix index
Fixes: 27c2e7471961 ("app/crypto-perf: support IMIX")
A bug caused index out-of-range error. This simple patch is to fix
this.
Signed-off-by: Fan Zhang
---
app/test-crypto-perf/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/app/test-crypto-perf/main.c b/app/test-crypto-perf/main.c
index 83a9d7b73..6d850af03 100644
--- a/app/test-crypto-perf/main.c
+++ b/app/test-crypto-perf/main.c
@@ -585,7 +585,7 @@ main(int argc, char **argv)
cdev_id = enabled_cdevs[i];
rte_eal_remote_launch(cperf_testmap[opts.test].runner,
- ctx[cdev_id], lcore_id);
+ ctx[i], lcore_id);
i++;
}
i = 0;
--
2.13.6
Re: [dpdk-dev] [PATCH v3 3/3] crypto/qat: add null point check and fix mem leak
Hi Yong,
> -Original Message-
> From: Yong Wang [mailto:wang.yon...@zte.com.cn]
> Sent: Tuesday, January 23, 2018 2:45 AM
> To: Xing, Beilei ; Lu, Wenzhuo ;
> Trahe, Fiona
>
> Cc: dev@dpdk.org; Yong Wang
> Subject: [PATCH v3 3/3] crypto/qat: add null point check and fix mem leak
>
> There are several func calls to rte_zmalloc() which don't do null
> point check on the return value. And before return, the memory is not
> freed. Fix it by adding null point check and rte_free().
>
> Signed-off-by: Yong Wang
> ---
> v3:
> * Rebase on master and modify again.
> v2:
> * Fix code style warning.
> ---
> drivers/crypto/qat/qat_adf/qat_algs_build_desc.c | 10 ++
> drivers/crypto/qat/qat_qp.c | 8 +++-
> 2 files changed, 17 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/crypto/qat/qat_adf/qat_algs_build_desc.c
> b/drivers/crypto/qat/qat_adf/qat_algs_build_desc.c
> index db6c9a3..26f854c 100644
> --- a/drivers/crypto/qat/qat_adf/qat_algs_build_desc.c
> +++ b/drivers/crypto/qat/qat_adf/qat_algs_build_desc.c
> @@ -359,6 +359,11 @@ static int qat_alg_do_precomputes(enum
> icp_qat_hw_auth_algo hash_alg,
>
> in = rte_zmalloc("working mem for key",
> ICP_QAT_HW_AES_XCBC_MAC_STATE2_SZ, 16);
> + if (in == NULL) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory");
> + return -ENOMEM;
> + }
> +
> rte_memcpy(in, qat_aes_xcbc_key_seed,
> ICP_QAT_HW_AES_XCBC_MAC_STATE2_SZ);
> for (x = 0; x < HASH_XCBC_PRECOMP_KEY_NUM; x++) {
> @@ -389,6 +394,11 @@ static int qat_alg_do_precomputes(enum
> icp_qat_hw_auth_algo hash_alg,
> ICP_QAT_HW_GALOIS_E_CTR0_SZ);
> in = rte_zmalloc("working mem for key",
> ICP_QAT_HW_GALOIS_H_SZ, 16);
> + if (in == NULL) {
> + PMD_DRV_LOG(ERR, "Failed to alloc memory");
> + return -ENOMEM;
> + }
> +
> memset(in, 0, ICP_QAT_HW_GALOIS_H_SZ);
> if (AES_set_encrypt_key(auth_key, auth_keylen << 3,
> &enc_key) != 0) {
> diff --git a/drivers/crypto/qat/qat_qp.c b/drivers/crypto/qat/qat_qp.c
> index 0941a58..812dce9 100644
> --- a/drivers/crypto/qat/qat_qp.c
> +++ b/drivers/crypto/qat/qat_qp.c
> @@ -151,6 +151,11 @@ int qat_crypto_sym_qp_setup(struct rte_cryptodev *dev,
> uint16_t
> queue_pair_id,
> qp->op_cookies = rte_zmalloc("qat PMD op cookie pointer",
> qp_conf->nb_descriptors * sizeof(*qp->op_cookies),
> RTE_CACHE_LINE_SIZE);
> + if (qp->op_cookies == NULL) {
> + PMD_DRV_LOG(ERR, "Failed to alloc mem for cookie");
> + rte_free(qp);
> + return -ENOMEM;
> + }
>
> qp->mmap_bar_addr = pci_dev->mem_resource[0].addr;
> qp->inflights16 = 0;
> @@ -192,7 +197,7 @@ int qat_crypto_sym_qp_setup(struct rte_cryptodev *dev,
> uint16_t queue_pair_id,
> for (i = 0; i < qp->nb_descriptors; i++) {
> if (rte_mempool_get(qp->op_cookie_pool, &qp->op_cookies[i])) {
> PMD_DRV_LOG(ERR, "QAT PMD Cannot get op_cookie");
> - return -EFAULT;
> + goto create_err;
> }
>
> struct qat_crypto_op_cookie *sql_cookie =
> @@ -217,6 +222,7 @@ int qat_crypto_sym_qp_setup(struct rte_cryptodev *dev,
> uint16_t queue_pair_id,
> return 0;
>
> create_err:
[Fiona] Thanks for this - it was on my backlog to fix :)
Can you add following too, to cover errors that happen after the
rte_mempool_create
if (qp->op_cookie_pool)
rte_mempool_free(qp->op_cookie_pool);
> + rte_free(qp->op_cookies);
> rte_free(qp);
> return -EFAULT;
> }
> --
> 1.8.3.1
Re: [dpdk-dev] [PATCH v7 08/15] net/ixgbe: use rte_eth_linkstatus functions
On Mon, Jan 22, 2018 at 03:54:51PM -0800, Stephen Hemminger wrote:
> Use the new helper functions from eth_dev for
> handling atomic link_info update.
>
> Signed-off-by: Stephen Hemminger
> Signed-off-by: Ferruh Yigit
>
> ---
> drivers/net/ixgbe/ixgbe_ethdev.c | 96
> +++-
> 1 file changed, 17 insertions(+), 79 deletions(-)
>
> diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c
> b/drivers/net/ixgbe/ixgbe_ethdev.c
> index 58217680c29c..7b74cbe92c0d 100644
> --- a/drivers/net/ixgbe/ixgbe_ethdev.c
> +++ b/drivers/net/ixgbe/ixgbe_ethdev.c
> @@ -20,7 +20,6 @@
> #include
> #include
> #include
> -#include
> #include
> #include
> #include
> @@ -787,58 +786,6 @@ static const struct rte_ixgbe_xstats_name_off
> rte_ixgbevf_stats_strings[] = {
> #define IXGBEVF_NB_XSTATS (sizeof(rte_ixgbevf_stats_strings) / \
> sizeof(rte_ixgbevf_stats_strings[0]))
>
Getting some build errors here:
CC igb_ethdev.o
/home/nhorman/git/dpdk/drivers/net/e1000/igb_ethdev.c: In function
‘eth_igb_stop’:
/home/nhorman/git/dpdk/drivers/net/e1000/igb_ethdev.c:1510:2: error: implicit
declaration of function ‘rte_eth_linkstatus_set’; did you mean
‘rte_eth_xstats_get’? [-Werror=implicit-function-declaration]
rte_eth_linkstatus_set(dev, &link);
^~
rte_eth_xstats_get
/home/nhorman/git/dpdk/drivers/net/e1000/igb_ethdev.c:1510:2: error: nested
extern declaration of ‘rte_eth_linkstatus_set’ [-Werror=nested-externs]
/home/nhorman/git/dpdk/drivers/net/e1000/igb_ethdev.c: In function
‘eth_igb_interrupt_action’:
/home/nhorman/git/dpdk/drivers/net/e1000/igb_ethdev.c:2831:3: error: implicit
declaration of function ‘rte_eth_linkstatus_get’; did you mean
‘rte_eth_xstats_get’? [-Werror=implicit-function-declaration]
rte_eth_linkstatus_get(dev, &link);
[dpdk-dev] [PATCH v3 2/3] doc: convert license headers to SPDX tags
Signed-off-by: Ferruh Yigit --- v3: Intel ones sharing copyright --- doc/guides/linux_gsg/linux_drivers.rst | 28 +--- 1 file changed, 1 insertion(+), 27 deletions(-) diff --git a/doc/guides/linux_gsg/linux_drivers.rst b/doc/guides/linux_gsg/linux_drivers.rst index 08f7c9baa..14381eed8 100644 --- a/doc/guides/linux_gsg/linux_drivers.rst +++ b/doc/guides/linux_gsg/linux_drivers.rst @@ -1,34 +1,8 @@ -.. BSD LICENSE +.. SPDX-License-Identifier: BSD-3-Clause Copyright(c) 2010-2015 Intel Corporation. Copyright(c) 2017 Mellanox Corporation. All rights reserved. -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions -are met: - -* Redistributions of source code must retain the above copyright -notice, this list of conditions and the following disclaimer. -* Redistributions in binary form must reproduce the above copyright -notice, this list of conditions and the following disclaimer in -the documentation and/or other materials provided with the -distribution. -* Neither the name of Intel Corporation nor the names of its -contributors may be used to endorse or promote products derived -from this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS -"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR -A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT -OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, -SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT -LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, -DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY -THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT -(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE -OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. - .. _linux_gsg_linux_drivers: Linux Drivers -- 2.14.3
[dpdk-dev] [PATCH v3 3/3] doc: convert license headers to SPDX tags
Signed-off-by: Ferruh Yigit --- v3: Non-Intel updates, I am sharing these since I already made the patch, please feel free to ignore if has legal implications --- doc/guides/cryptodevs/mrvl.rst | 28 +--- doc/guides/howto/pvp_reference_benchmark.rst | 29 + doc/guides/howto/rte_flow.rst| 29 + doc/guides/nics/ark.rst | 33 +++- doc/guides/nics/avp.rst | 28 +--- doc/guides/nics/bnx2x.rst| 28 +--- doc/guides/nics/bnxt.rst | 28 +--- doc/guides/nics/cxgbe.rst| 28 +--- doc/guides/nics/ena.rst | 33 +++- doc/guides/nics/enic.rst | 27 +-- doc/guides/nics/fail_safe.rst| 28 +--- doc/guides/nics/mlx4.rst | 28 +--- doc/guides/nics/mlx5.rst | 28 +--- doc/guides/nics/mrvl.rst | 28 +--- doc/guides/nics/nfp.rst | 28 +--- doc/guides/nics/overview.rst | 28 +--- doc/guides/nics/sfc_efx.rst | 23 + doc/guides/nics/szedata2.rst | 28 +--- doc/guides/nics/vhost.rst| 28 +--- doc/guides/prog_guide/port_hotplug_framework.rst | 28 +--- doc/guides/prog_guide/rte_flow.rst | 28 +--- doc/guides/sample_app_ug/flow_filtering.rst | 29 + doc/guides/tools/devbind.rst | 29 + doc/guides/tools/index.rst | 28 +--- doc/guides/tools/pmdinfo.rst | 29 + 25 files changed, 29 insertions(+), 680 deletions(-) diff --git a/doc/guides/cryptodevs/mrvl.rst b/doc/guides/cryptodevs/mrvl.rst index 6a0b08c58..dff0f9533 100644 --- a/doc/guides/cryptodevs/mrvl.rst +++ b/doc/guides/cryptodevs/mrvl.rst @@ -1,34 +1,8 @@ -.. BSD LICENSE +.. SPDX-License-Identifier: BSD-3-Clause Copyright(c) 2017 Marvell International Ltd. Copyright(c) 2017 Semihalf. All rights reserved. -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions -are met: - - * Redistributions of source code must retain the above copyright -notice, this list of conditions and the following disclaimer. - * Redistributions in binary form must reproduce the above copyright -notice, this list of conditions and the following disclaimer in -the documentation and/or other materials provided with the -distribution. - * Neither the name of the copyright holder nor the names of its -contributors may be used to endorse or promote products derived -from this software without specific prior written permission. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS -"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR -A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT -OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, -SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT -LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, -DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY -THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT -(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE -OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. - MRVL Crypto Poll Mode Driver diff --git a/doc/guides/howto/pvp_reference_benchmark.rst b/doc/guides/howto/pvp_reference_benchmark.rst index 228b4a257..3e7a34fb8 100644 --- a/doc/guides/howto/pvp_reference_benchmark.rst +++ b/doc/guides/howto/pvp_reference_benchmark.rst @@ -1,34 +1,7 @@ -.. BSD LICENSE +.. SPDX-License-Identifier: BSD-3-Clause Copyright(c) 2016 Red Hat, Inc. All rights reserved. All rights reserved. -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions -are met: - -* Redistributions of source code must retain the above copyright -notice, this list of conditions and the following disclaimer. -* Redistributions in binary form must reproduce the above copyright -notice, this list of conditions and the following disclaimer in -the documentatio
[dpdk-dev] [PATCH] examples/ipsec_secgw: fix security session
Fixes: 3da37f682173 ("examples/ipsec_secgw: create session mempools for
ethdevs")
Some NICs do not have the rte_security context, this patch fixes the segment
fault
caused by this.
Signed-off-by: Fan Zhang
---
examples/ipsec-secgw/ipsec-secgw.c | 10 --
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/examples/ipsec-secgw/ipsec-secgw.c
b/examples/ipsec-secgw/ipsec-secgw.c
index 3a28fcceb..c67f79db9 100644
--- a/examples/ipsec-secgw/ipsec-secgw.c
+++ b/examples/ipsec-secgw/ipsec-secgw.c
@@ -1384,10 +1384,16 @@ cryptodevs_init(void)
max_sess_sz = sess_sz;
}
for (port_id = 0; port_id < rte_eth_dev_count(); port_id++) {
+ void *sec_ctx;
+
if ((enabled_port_mask & (1 << port_id)) == 0)
continue;
- sess_sz = rte_security_session_get_size(
- rte_eth_dev_get_sec_ctx(port_id));
+
+ sec_ctx = rte_eth_dev_get_sec_ctx(port_id);
+ if (sec_ctx == NULL)
+ continue;
+
+ sess_sz = rte_security_session_get_size(sec_ctx);
if (sess_sz > max_sess_sz)
max_sess_sz = sess_sz;
}
--
2.13.6
Re: [dpdk-dev] [PATCH] doc: document the new devargs syntax
On Thu, Jan 18, 2018 at 10:46:23AM +0100, Gaëtan Rivet wrote: > On Thu, Jan 18, 2018 at 09:46:29AM +0100, Thomas Monjalon wrote: > > 18/01/2018 08:35, Yuanhan Liu: > > > On Wed, Jan 17, 2018 at 12:34:08PM +, Ferruh Yigit wrote: > > > > So does it make sense to separate them logically? Perhaps as "device > > > > identifier" > > > > and "device args". > > > > > > Then I think it returns back to the old issue: how could we identify a > > > port when the bus id (say BDF for PCI bus) is not enough for identifying > > > a port? Such case could happen when a single NIC has 2 ports sharing > > > the same BDF. It could also happen with the VF representors that will > > > be introduced shortly. > > > > Yes, the device matching syntax must include bus category, class category > > and driver category. So any device can be identified in future. > > > > But I think Ferruh is talking about separating device matching > > (which is described in this proposal) and device settings > > (which are usually mixed in -w and --vdev options). > > I agree there are different things and may be separate. > > They could share the same syntax (bus/class/driver) but be separate > > with a semicolon: > > matching;settings > > Can you give an example? Let's take port addition in OVS-DPDK as an example. It happens in 2 steps: - port lookup (if port is already probed) - dev attachment (if lookup fails) And also let's assume we need probe a ConnectX-3 port. Note that for ConnectX-3, there are 2 ports sharing the same PCI addr. Thus, PCI BDF is not enough. And let's assume we use another extra property "port". If the proposal described in this patch is being used, the devarg would look like following: bus=pci,id=04:00.0/class=eth,port=0/driver=mlx4,mlx4_arg_A=val,... Then "bus=pci,id=04:00.0/class=eth,port=0" will be used for lookup, It means we are looking for a port with PCI BDF == 04:00.0 AND port == 0 (the first port of the 2 ports). Note that in my proposal the driver category is not intended for lookup. If any properties needed be looked in the driver category, they would probably need be elevated to the class category. If port not found, then the whole string will be used for dev attachment. It means we are attaching a port with PCI BDF == 04.00.0 AND port == 0 (the 2nd port will not be attached). And here is how the devargs would look like if "matching;settings" is being used: bus=pci,id=04:00.0/class=eth,port=0;bus=pci,id=04:00.0/class=eth,port=0/driver=mlx4,mlx4_arg_A=val,... The part before ";" will be used for lookup and the later part will be used for attachment. It should work. It just looks redundant. -yliu
Re: [dpdk-dev] [PATCH v3 7/7] app/testpmd: adjust ethdev port ownership
Hi all,
On Tue, Jan 23, 2018 at 08:54:27AM +, Matan Azrad wrote:
> > > > > > > > > > Subject: [PATCH v3 7/7] app/testpmd: adjust ethdev port
> > > > > > > > > > ownership
> > > > > > > > > >
> > > > > > > > > > Testpmd should not use ethdev ports which are managed by
> > > > > > > > > > other DPDK entities.
> > > > > > > > > >
> > > > > > > > > > Set Testpmd ownership to each port which is not used by
> > > > > > > > > > other entity and prevent any usage of ethdev ports which
> > > > > > > > > > are not owned by
> > > > > > > Testpmd.
> > > > > > > > > >
> > > > > > > > > > Signed-off-by: Matan Azrad
> > > > > > > > > > ---
> > > > > > > > > > app/test-pmd/cmdline.c | 89 +++--
> >
> > > > ---
> > > > > > >
> > > > > > > > > -
> > > > > > > > > > app/test-pmd/cmdline_flow.c | 2 +-
> > > > > > > > > > app/test-pmd/config.c | 37 ++-
> > > > > > > > > > app/test-pmd/parameters.c | 4 +-
> > > > > > > > > > app/test-pmd/testpmd.c | 63
> >
> > > >
> > > > > > > > > > app/test-pmd/testpmd.h | 3 ++
> > > > > > > > > > 6 files changed, 103 insertions(+), 95 deletions(-)
> > > > > > > > > >
> > > > > > > > > > diff --git a/app/test-pmd/cmdline.c
> > > > > > > > > > b/app/test-pmd/cmdline.c index
> > > > > > > > > > 31919ba..6199c64 100644
> > > > > > > > > > --- a/app/test-pmd/cmdline.c
> > > > > > > > > > +++ b/app/test-pmd/cmdline.c
> > > > > > > > > > @@ -1394,7 +1394,7 @@ struct cmd_config_speed_all {
> > > > > > > > > > &link_speed) < 0)
> > > > > > > > > > return;
> > > > > > > > > >
> > > > > > > > > > - RTE_ETH_FOREACH_DEV(pid) {
> > > > > > > > > > + RTE_ETH_FOREACH_DEV_OWNED_BY(pid,
> > my_owner.id) {
> > > > > > > > >
> > > > > > > > > Why do we need all these changes?
> > > > > > > > > As I understand you changed definition of
> > > > > > > > > RTE_ETH_FOREACH_DEV(), so no testpmd should work ok
> > > > > > > > > default
> > > > (no_owner case).
> > > > > > > > > Am I missing something here?
> > > > > > > >
> > > > > > > > Now, After Gaetan suggestion RTE_ETH_FOREACH_DEV(pid) will
> > > > > > > > iterate
> > > > > > > over all valid and ownerless ports.
> > > > > > >
> > > > > > > Yes.
> > > > > > >
> > > > > > > > Here Testpmd wants to iterate over its owned ports.
> > > > > > >
> > > > > > > Why? Why it can't just iterate over all valid and ownerless ports?
> > > > > > > As I understand it would be enough to fix current problems and
> > > > > > > would allow us to avoid any changes in testmpd (which I think
> > > > > > > is a good
> > > > thing).
> > > > > >
> > > > > > Yes, I understand that this big change is very daunted, But I
> > > > > > think the current a lot of bugs in testpmd(regarding port
> > > > > > ownership) even more
> > > > > daunted.
> > > > > >
> > > > > > Look,
> > > > > > Testpmd initiates some of its internal databases depends on
> > > > > > specific port iteration, In some time someone may take ownership
> > > > > > of Testpmd ports and testpmd will continue to touch them.
> > > >
> > > > But if someone will take the ownership (assign new owner_id) that
> > > > port will not appear in RTE_ETH_FOREACH_DEV() any more.
> > > >
> > >
> > > Yes, but testpmd sometimes depends on previous iteration using internal
> > > database.
> > > So it uses internal database that was updated by old iteration.
> >
> > That sounds like just a bug in testpmd that need to be fixed, no?
>
> If Testpmd already took ownership for these ports(like I did), it is ok.
>
Have you tested using the default iterator (NO_OWNER)?
It worked until now with the bare minimal device tagging using
DEV_DEFERRED. Testpmd did not seem to mind having to skip this port.
I'm sure there were places where this was overlooked, but overall, I'd
think everything should be fixable using only the NO_OWNER iteration.
Can you point to a specific scenario (command line, chain of event) that
would lead to a problem?
> > Any particular places where outdated device info is used?
>
> For example, look for the stream management in testpmd(I think I saw it
> there).
>
The stream management is certainly shaky, but it happens after the EAL
initial port creation, and is not able to update itself for new
hotplugged ports (unless something changed).
> > > > > If I look back on the fail-safe, its sole purpose is to have
> > > > > seamless hotplug with existing applications.
> > > > >
> > > > > Port ownership is a genericization of some functions introduced by
> > > > > the fail-safe, that could structure DPDK further. It should allow
> > > > > applications to have a seamless integration with subsystems using
> > > > > port ownership. Without this, port ownership cannot be used.
> > > > >
> > > > > Testpmd should be fixed, but follow the most common design
> > > > > patterns of DPDK applications. Going with port ownership seems
> > > > > like a paradigm shift.
> > > > >
> >
Re: [dpdk-dev] [PATCH v3 1/3] doc: convert license headers to SPDX tags
On Tue, Jan 23, 2018 at 12:29:16PM +, Ferruh Yigit wrote: > Signed-off-by: Ferruh Yigit > --- > v3: update files only has Intel copyright > --- Acked-by: Bruce Richardson
Re: [dpdk-dev] [PATCH v3 2/3] doc: convert license headers to SPDX tags
On Tue, Jan 23, 2018 at 12:29:17PM +, Ferruh Yigit wrote: > Signed-off-by: Ferruh Yigit > --- > v3: Intel ones sharing copyright > --- > doc/guides/linux_gsg/linux_drivers.rst | 28 +--- > 1 file changed, 1 insertion(+), 27 deletions(-) > >From Intel side: Acked-by: Bruce Richardson
[dpdk-dev] [PATCH 2/7 v2] bus/dpaa: allocate qman portals in thread safe manner
Fixes: 9d32ef0f5d61 ("bus/dpaa: support creating dynamic HW portal")
Cc: sta...@dpdk.org
Signed-off-by: Nipun Gupta
---
drivers/bus/dpaa/base/qbman/qman.c | 8 +++-
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/drivers/bus/dpaa/base/qbman/qman.c
b/drivers/bus/dpaa/base/qbman/qman.c
index e7fdf03..4d8bdae 100644
--- a/drivers/bus/dpaa/base/qbman/qman.c
+++ b/drivers/bus/dpaa/base/qbman/qman.c
@@ -625,7 +625,7 @@ struct qman_portal *qman_create_portal(
#define MAX_GLOBAL_PORTALS 8
static struct qman_portal global_portals[MAX_GLOBAL_PORTALS];
-static int global_portals_used[MAX_GLOBAL_PORTALS];
+rte_atomic16_t global_portals_used[MAX_GLOBAL_PORTALS];
static struct qman_portal *
qman_alloc_global_portal(void)
@@ -633,10 +633,8 @@ struct qman_portal *qman_create_portal(
unsigned int i;
for (i = 0; i < MAX_GLOBAL_PORTALS; i++) {
- if (global_portals_used[i] == 0) {
- global_portals_used[i] = 1;
+ if (rte_atomic16_test_and_set(&global_portals_used[i]))
return &global_portals[i];
- }
}
pr_err("No portal available (%x)\n", MAX_GLOBAL_PORTALS);
@@ -650,7 +648,7 @@ struct qman_portal *qman_create_portal(
for (i = 0; i < MAX_GLOBAL_PORTALS; i++) {
if (&global_portals[i] == portal) {
- global_portals_used[i] = 0;
+ rte_atomic16_clear(&global_portals_used[i]);
return 0;
}
}
--
1.9.1
[dpdk-dev] [PATCH 1/7 v2] bus/dpaa: check flag in qman multi enqueue
A caller may/may not pass the flags in qman enqueue multi API.
This patch adds a check on that flag and only accesses it if passed
by the caller.
Fixes: 43797e7b4774 ("bus/dpaa: support event dequeue and consumption")
Cc: sta...@dpdk.org
Signed-off-by: Nipun Gupta
---
drivers/bus/dpaa/base/qbman/qman.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/bus/dpaa/base/qbman/qman.c
b/drivers/bus/dpaa/base/qbman/qman.c
index 609bc76..e7fdf03 100644
--- a/drivers/bus/dpaa/base/qbman/qman.c
+++ b/drivers/bus/dpaa/base/qbman/qman.c
@@ -2198,7 +2198,7 @@ int qman_enqueue_multi(struct qman_fq *fq,
eq->fd.addr = cpu_to_be40(fd->addr);
eq->fd.status = cpu_to_be32(fd->status);
eq->fd.opaque = cpu_to_be32(fd->opaque);
- if (flags[i] & QMAN_ENQUEUE_FLAG_DCA) {
+ if (flags && (flags[i] & QMAN_ENQUEUE_FLAG_DCA)) {
eq->dca = QM_EQCR_DCA_ENABLE |
((flags[i] >> 8) & QM_EQCR_DCA_IDXMASK);
}
--
1.9.1
[dpdk-dev] [PATCH 0/7 v2] dpaa: fixes and performance improvement changes
Patch 1-4 - Fixes some of the issues in the DPAA bus Patch 5-7 - Performance enhancement changes on DPAA platform Hemant Agrawal (2): mempool/dpaa: fix the phy to virt optimization net/dpaa: use phy to virt optimizations Nipun Gupta (4): bus/dpaa: check flag in qman multi enqueue bus/dpaa: allocate qman portals in thread safe manner bus/dpaa: check portal presence in the caller API net/dpaa: further push mode optimizations Shreyansh Jain (1): bus/dpaa: fix port order shuffling Changes in v2: Fix the checkpatch warnings drivers/bus/dpaa/base/qbman/qman.c| 99 + drivers/bus/dpaa/dpaa_bus.c | 78 ++-- drivers/bus/dpaa/include/fsl_qman.h | 10 +++ drivers/bus/dpaa/rte_bus_dpaa_version.map | 1 + drivers/bus/dpaa/rte_dpaa_bus.h | 2 + drivers/mempool/dpaa/dpaa_mempool.c | 33 + drivers/mempool/dpaa/dpaa_mempool.h | 4 +- drivers/net/dpaa/dpaa_ethdev.c| 19 +++-- drivers/net/dpaa/dpaa_rxtx.c | 115 -- drivers/net/dpaa/dpaa_rxtx.h | 9 ++- 10 files changed, 247 insertions(+), 123 deletions(-) -- 1.9.1
[dpdk-dev] [PATCH 6/7 v2] bus/dpaa: check portal presence in the caller API
In the I/O path we were calling rte_dpaa_portal_init which
internally checks if a portal is affined to the core.
But this lead to calling of that non-static API in every call.
Instead check the portal affinity in the caller itself for
performance reasons
Signed-off-by: Nipun Gupta
---
drivers/bus/dpaa/dpaa_bus.c | 26 ++
drivers/bus/dpaa/rte_bus_dpaa_version.map | 1 +
drivers/bus/dpaa/rte_dpaa_bus.h | 2 ++
drivers/mempool/dpaa/dpaa_mempool.c | 24 ++--
drivers/net/dpaa/dpaa_ethdev.c| 10 ++
drivers/net/dpaa/dpaa_rxtx.c | 20
6 files changed, 41 insertions(+), 42 deletions(-)
diff --git a/drivers/bus/dpaa/dpaa_bus.c b/drivers/bus/dpaa/dpaa_bus.c
index ef2df48..5039067 100644
--- a/drivers/bus/dpaa/dpaa_bus.c
+++ b/drivers/bus/dpaa/dpaa_bus.c
@@ -54,7 +54,7 @@
unsigned int dpaa_svr_family;
-RTE_DEFINE_PER_LCORE(bool, _dpaa_io);
+RTE_DEFINE_PER_LCORE(bool, dpaa_io);
RTE_DEFINE_PER_LCORE(struct dpaa_portal_dqrr, held_bufs);
static int
@@ -230,9 +230,7 @@
}
}
-/** XXX move this function into a separate file */
-static int
-_dpaa_portal_init(void *arg)
+int rte_dpaa_portal_init(void *arg)
{
cpu_set_t cpuset;
pthread_t id;
@@ -303,25 +301,13 @@
return ret;
}
- RTE_PER_LCORE(_dpaa_io) = true;
+ RTE_PER_LCORE(dpaa_io) = true;
DPAA_BUS_LOG(DEBUG, "QMAN thread initialized");
return 0;
}
-/*
- * rte_dpaa_portal_init - Wrapper over _dpaa_portal_init with thread level
check
- * XXX Complete this
- */
-int rte_dpaa_portal_init(void *arg)
-{
- if (unlikely(!RTE_PER_LCORE(_dpaa_io)))
- return _dpaa_portal_init(arg);
-
- return 0;
-}
-
int
rte_dpaa_portal_fq_init(void *arg, struct qman_fq *fq)
{
@@ -329,8 +315,8 @@ int rte_dpaa_portal_init(void *arg)
u32 sdqcr;
struct qman_portal *qp;
- if (unlikely(!RTE_PER_LCORE(_dpaa_io)))
- _dpaa_portal_init(arg);
+ if (unlikely(!RTE_PER_LCORE(dpaa_io)))
+ rte_dpaa_portal_init(arg);
/* Initialise qman specific portals */
qp = fsl_qman_portal_create();
@@ -368,7 +354,7 @@ int rte_dpaa_portal_fq_close(struct qman_fq *fq)
rte_free(dpaa_io_portal);
dpaa_io_portal = NULL;
- RTE_PER_LCORE(_dpaa_io) = false;
+ RTE_PER_LCORE(dpaa_io) = false;
}
#define DPAA_DEV_PATH1 "/sys/devices/platform/soc/soc:fsl,dpaa"
diff --git a/drivers/bus/dpaa/rte_bus_dpaa_version.map
b/drivers/bus/dpaa/rte_bus_dpaa_version.map
index 925cf91..8d90285 100644
--- a/drivers/bus/dpaa/rte_bus_dpaa_version.map
+++ b/drivers/bus/dpaa/rte_bus_dpaa_version.map
@@ -70,6 +70,7 @@ DPDK_18.02 {
dpaa_logtype_eventdev;
dpaa_svr_family;
+ per_lcore_dpaa_io;
per_lcore_held_bufs;
qm_channel_pool1;
qman_alloc_cgrid_range;
diff --git a/drivers/bus/dpaa/rte_dpaa_bus.h b/drivers/bus/dpaa/rte_dpaa_bus.h
index 6fa0c3d..0352abd 100644
--- a/drivers/bus/dpaa/rte_dpaa_bus.h
+++ b/drivers/bus/dpaa/rte_dpaa_bus.h
@@ -31,6 +31,8 @@
extern unsigned int dpaa_svr_family;
+extern RTE_DEFINE_PER_LCORE(bool, dpaa_io);
+
struct rte_dpaa_device;
struct rte_dpaa_driver;
diff --git a/drivers/mempool/dpaa/dpaa_mempool.c
b/drivers/mempool/dpaa/dpaa_mempool.c
index fe22519..eb5b8f9 100644
--- a/drivers/mempool/dpaa/dpaa_mempool.c
+++ b/drivers/mempool/dpaa/dpaa_mempool.c
@@ -139,11 +139,13 @@
DPAA_MEMPOOL_DPDEBUG("Request to free %d buffers in bpid = %d",
n, bp_info->bpid);
- ret = rte_dpaa_portal_init((void *)0);
- if (ret) {
- DPAA_MEMPOOL_ERR("rte_dpaa_portal_init failed with ret: %d",
-ret);
- return 0;
+ if (unlikely(!RTE_PER_LCORE(dpaa_io))) {
+ ret = rte_dpaa_portal_init((void *)0);
+ if (ret) {
+ DPAA_MEMPOOL_ERR("rte_dpaa_portal_init failed with ret:
%d",
+ret);
+ return 0;
+ }
}
while (i < n) {
@@ -193,11 +195,13 @@
return -1;
}
- ret = rte_dpaa_portal_init((void *)0);
- if (ret) {
- DPAA_MEMPOOL_ERR("rte_dpaa_portal_init failed with ret: %d",
-ret);
- return -1;
+ if (unlikely(!RTE_PER_LCORE(dpaa_io))) {
+ ret = rte_dpaa_portal_init((void *)0);
+ if (ret) {
+ DPAA_MEMPOOL_ERR("rte_dpaa_portal_init failed with ret:
%d",
+ret);
+ return -1;
+ }
}
while (n < count) {
diff --git a/drivers/net/dpaa/dpaa_ethdev.c b/drivers/net/dpaa/dpaa_ethdev.c
index bf5eb96..b60ed3b 100644
--- a/drivers/net/dpaa/dpaa_ethdev.c
+++ b/drivers/net
[dpdk-dev] [PATCH 4/7 v2] bus/dpaa: fix port order shuffling
From: Shreyansh Jain
While scanning for devices, the order in which devices appear is
different as compared to MAC sequence.
This can cause confusion for users and automated scripts.
This patch create a sorted list of devices.
Fixes: 919eeaccb2ba ("bus/dpaa: introduce NXP DPAA bus driver skeleton")
Cc: sta...@dpdk.org
Signed-off-by: Shreyansh Jain
---
drivers/bus/dpaa/dpaa_bus.c | 52 +++--
1 file changed, 50 insertions(+), 2 deletions(-)
diff --git a/drivers/bus/dpaa/dpaa_bus.c b/drivers/bus/dpaa/dpaa_bus.c
index ba33566..ef2df48 100644
--- a/drivers/bus/dpaa/dpaa_bus.c
+++ b/drivers/bus/dpaa/dpaa_bus.c
@@ -57,10 +57,58 @@
RTE_DEFINE_PER_LCORE(bool, _dpaa_io);
RTE_DEFINE_PER_LCORE(struct dpaa_portal_dqrr, held_bufs);
+static int
+compare_dpaa_devices(struct rte_dpaa_device *dev1,
+struct rte_dpaa_device *dev2)
+{
+ int comp = 0;
+
+ /* Segragating ETH from SEC devices */
+ if (dev1->device_type > dev2->device_type)
+ comp = 1;
+ else if (dev1->device_type < dev2->device_type)
+ comp = -1;
+ else
+ comp = 0;
+
+ if ((comp != 0) || (dev1->device_type != FSL_DPAA_ETH))
+ return comp;
+
+ if (dev1->id.fman_id > dev2->id.fman_id) {
+ comp = 1;
+ } else if (dev1->id.fman_id < dev2->id.fman_id) {
+ comp = -1;
+ } else {
+ /* FMAN ids match, check for mac_id */
+ if (dev1->id.mac_id > dev2->id.mac_id)
+ comp = 1;
+ else if (dev1->id.mac_id < dev2->id.mac_id)
+ comp = -1;
+ else
+ comp = 0;
+ }
+
+ return comp;
+}
+
static inline void
-dpaa_add_to_device_list(struct rte_dpaa_device *dev)
+dpaa_add_to_device_list(struct rte_dpaa_device *newdev)
{
- TAILQ_INSERT_TAIL(&rte_dpaa_bus.device_list, dev, next);
+ int comp, inserted = 0;
+ struct rte_dpaa_device *dev = NULL;
+ struct rte_dpaa_device *tdev = NULL;
+
+ TAILQ_FOREACH_SAFE(dev, &rte_dpaa_bus.device_list, next, tdev) {
+ comp = compare_dpaa_devices(newdev, dev);
+ if (comp < 0) {
+ TAILQ_INSERT_BEFORE(dev, newdev, next);
+ inserted = 1;
+ break;
+ }
+ }
+
+ if (!inserted)
+ TAILQ_INSERT_TAIL(&rte_dpaa_bus.device_list, newdev, next);
}
static inline void
--
1.9.1
[dpdk-dev] [PATCH 5/7 v2] net/dpaa: use phy to virt optimizations
From: Hemant Agrawal
Use the optimized routine for phy to virt conversion,
when the mempool is allocated from physical contiguous memory.
Signed-off-by: Hemant Agrawal
---
drivers/net/dpaa/dpaa_rxtx.c | 14 +-
1 file changed, 9 insertions(+), 5 deletions(-)
diff --git a/drivers/net/dpaa/dpaa_rxtx.c b/drivers/net/dpaa/dpaa_rxtx.c
index ab23352..b889d03 100644
--- a/drivers/net/dpaa/dpaa_rxtx.c
+++ b/drivers/net/dpaa/dpaa_rxtx.c
@@ -309,7 +309,7 @@ struct rte_mbuf *
DPAA_DP_LOG(DEBUG, "Received an SG frame");
- vaddr = rte_dpaa_mem_ptov(qm_fd_addr(fd));
+ vaddr = DPAA_MEMPOOL_PTOV(bp_info, qm_fd_addr(fd));
if (!vaddr) {
DPAA_PMD_ERR("unable to convert physical address");
return NULL;
@@ -318,7 +318,7 @@ struct rte_mbuf *
sg_temp = &sgt[i++];
hw_sg_to_cpu(sg_temp);
temp = (struct rte_mbuf *)((char *)vaddr - bp_info->meta_data_size);
- sg_vaddr = rte_dpaa_mem_ptov(qm_sg_entry_get64(sg_temp));
+ sg_vaddr = DPAA_MEMPOOL_PTOV(bp_info, qm_sg_entry_get64(sg_temp));
first_seg = (struct rte_mbuf *)((char *)sg_vaddr -
bp_info->meta_data_size);
@@ -334,7 +334,8 @@ struct rte_mbuf *
while (i < DPAA_SGT_MAX_ENTRIES) {
sg_temp = &sgt[i++];
hw_sg_to_cpu(sg_temp);
- sg_vaddr = rte_dpaa_mem_ptov(qm_sg_entry_get64(sg_temp));
+ sg_vaddr = DPAA_MEMPOOL_PTOV(bp_info,
+qm_sg_entry_get64(sg_temp));
cur_seg = (struct rte_mbuf *)((char *)sg_vaddr -
bp_info->meta_data_size);
cur_seg->data_off = sg_temp->offset;
@@ -361,7 +362,7 @@ struct rte_mbuf *
{
struct rte_mbuf *mbuf;
struct dpaa_bp_info *bp_info = DPAA_BPID_TO_POOL_INFO(fd->bpid);
- void *ptr = rte_dpaa_mem_ptov(qm_fd_addr(fd));
+ void *ptr;
uint8_t format =
(fd->opaque & DPAA_FD_FORMAT_MASK) >> DPAA_FD_FORMAT_SHIFT;
uint16_t offset;
@@ -372,6 +373,8 @@ struct rte_mbuf *
if (unlikely(format == qm_fd_sg))
return dpaa_eth_sg_to_mbuf(fd, ifid);
+ ptr = DPAA_MEMPOOL_PTOV(bp_info, qm_fd_addr(fd));
+
rte_prefetch0((void *)((uint8_t *)ptr + DEFAULT_RX_ICEOF));
offset = (fd->opaque & DPAA_FD_OFFSET_MASK) >> DPAA_FD_OFFSET_SHIFT;
@@ -537,7 +540,8 @@ static void *dpaa_get_pktbuf(struct dpaa_bp_info *bp_info)
DPAA_DP_LOG(DEBUG, "got buffer 0x%lx from pool %d",
(uint64_t)bufs.addr, bufs.bpid);
- buf = (uint64_t)rte_dpaa_mem_ptov(bufs.addr) - bp_info->meta_data_size;
+ buf = (uint64_t)DPAA_MEMPOOL_PTOV(bp_info, bufs.addr)
+ - bp_info->meta_data_size;
if (!buf)
goto out;
--
1.9.1
[dpdk-dev] [PATCH 7/7 v2] net/dpaa: further push mode optimizations
This patch supports batch processing of multiple packets
in the Rx side
Signed-off-by: Nipun Gupta
Signed-off-by: Hemant Agrawal
---
drivers/bus/dpaa/base/qbman/qman.c | 89 ++---
drivers/bus/dpaa/include/fsl_qman.h | 10 +
drivers/net/dpaa/dpaa_ethdev.c | 9 +++-
drivers/net/dpaa/dpaa_rxtx.c| 81 +
drivers/net/dpaa/dpaa_rxtx.h| 9 ++--
5 files changed, 137 insertions(+), 61 deletions(-)
diff --git a/drivers/bus/dpaa/base/qbman/qman.c
b/drivers/bus/dpaa/base/qbman/qman.c
index 4d8bdae..2b97671 100644
--- a/drivers/bus/dpaa/base/qbman/qman.c
+++ b/drivers/bus/dpaa/base/qbman/qman.c
@@ -1055,64 +1055,63 @@ unsigned int qman_portal_poll_rx(unsigned int
poll_limit,
void **bufs,
struct qman_portal *p)
{
- const struct qm_dqrr_entry *dq;
- struct qman_fq *fq;
- enum qman_cb_dqrr_result res;
- unsigned int limit = 0;
-#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
- struct qm_dqrr_entry *shadow;
-#endif
- unsigned int rx_number = 0;
+ struct qm_portal *portal = &p->p;
+ register struct qm_dqrr *dqrr = &portal->dqrr;
+ struct qm_dqrr_entry *dq[QM_DQRR_SIZE], *shadow[QM_DQRR_SIZE];
+ struct qman_fq *fq[QM_DQRR_SIZE];
+ unsigned int limit = 0, rx_number = 0;
+ uint32_t consume = 0;
do {
qm_dqrr_pvb_update(&p->p);
- dq = qm_dqrr_current(&p->p);
- if (unlikely(!dq))
+ if (!dqrr->fill)
break;
+
+ dq[rx_number] = dqrr->cursor;
+ dqrr->cursor = DQRR_CARRYCLEAR(dqrr->cursor + 1);
+ /* Prefetch the next DQRR entry */
+ rte_prefetch0(dqrr->cursor);
+
#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
- /* If running on an LE system the fields of the
-* dequeue entry must be swapper. Because the
-* QMan HW will ignore writes the DQRR entry is
-* copied and the index stored within the copy
-*/
- shadow = &p->shadow_dqrr[DQRR_PTR2IDX(dq)];
- *shadow = *dq;
- dq = shadow;
- shadow->fqid = be32_to_cpu(shadow->fqid);
- shadow->contextB = be32_to_cpu(shadow->contextB);
- shadow->seqnum = be16_to_cpu(shadow->seqnum);
- hw_fd_to_cpu(&shadow->fd);
+ /* If running on an LE system the fields of the
+* dequeue entry must be swapper. Because the
+* QMan HW will ignore writes the DQRR entry is
+* copied and the index stored within the copy
+*/
+ shadow[rx_number] =
+ &p->shadow_dqrr[DQRR_PTR2IDX(dq[rx_number])];
+ shadow[rx_number]->fd.opaque_addr =
+ dq[rx_number]->fd.opaque_addr;
+ shadow[rx_number]->fd.addr =
+ be40_to_cpu(dq[rx_number]->fd.addr);
+ shadow[rx_number]->fd.opaque =
+ be32_to_cpu(dq[rx_number]->fd.opaque);
+#else
+ shadow = dq;
#endif
/* SDQCR: context_b points to the FQ */
#ifdef CONFIG_FSL_QMAN_FQ_LOOKUP
- fq = get_fq_table_entry(dq->contextB);
+ fq[rx_number] = qman_fq_lookup_table[be32_to_cpu(
+ dq[rx_number]->contextB)];
#else
- fq = (void *)(uintptr_t)dq->contextB;
+ fq[rx_number] = (void *)(uintptr_t)be32_to_cpu(dq->contextB);
#endif
- /* Now let the callback do its stuff */
- res = fq->cb.dqrr_dpdk_cb(NULL, p, fq, dq, &bufs[rx_number]);
+ fq[rx_number]->cb.dqrr_prepare(shadow[rx_number],
+&bufs[rx_number]);
+
+ consume |= (1 << (31 - DQRR_PTR2IDX(shadow[rx_number])));
rx_number++;
- /* Interpret 'dq' from a driver perspective. */
- /*
-* Parking isn't possible unless HELDACTIVE was set. NB,
-* FORCEELIGIBLE implies HELDACTIVE, so we only need to
-* check for HELDACTIVE to cover both.
-*/
- DPAA_ASSERT((dq->stat & QM_DQRR_STAT_FQ_HELDACTIVE) ||
- (res != qman_cb_dqrr_park));
- qm_dqrr_cdc_consume_1ptr(&p->p, dq, res == qman_cb_dqrr_park);
- /* Move forward */
- qm_dqrr_next(&p->p);
- /*
-* Entry processed and consumed, increment our counter. The
-* callback can request that we exit after consuming the
-* entry, and we also exit if we reach our processing limit,
-* so loop back only if neither of these conditions is met.
-*/
- } while (likely(++limit < poll_limit));
[dpdk-dev] [PATCH 3/7 v2] mempool/dpaa: fix the phy to virt optimization
From: Hemant Agrawal
Fixes: 83a4f267f2e3 ("mempool/dpaa: optimize phy to virt conversion")
Cc: sta...@dpdk.org
Signed-off-by: Hemant Agrawal
---
drivers/mempool/dpaa/dpaa_mempool.c | 9 -
drivers/mempool/dpaa/dpaa_mempool.h | 4 ++--
2 files changed, 6 insertions(+), 7 deletions(-)
diff --git a/drivers/mempool/dpaa/dpaa_mempool.c
b/drivers/mempool/dpaa/dpaa_mempool.c
index ddc4e47..fe22519 100644
--- a/drivers/mempool/dpaa/dpaa_mempool.c
+++ b/drivers/mempool/dpaa/dpaa_mempool.c
@@ -150,8 +150,8 @@
uint64_t phy = rte_mempool_virt2iova(obj_table[i]);
if (unlikely(!bp_info->ptov_off)) {
- /* buffers are not from multiple memzones */
- if (!(bp_info->flags & DPAA_MPOOL_MULTI_MEMZONE)) {
+ /* buffers are from single mem segment */
+ if (bp_info->flags & DPAA_MPOOL_SINGLE_SEGMENT) {
bp_info->ptov_off
= (uint64_t)obj_table[i] - phy;
rte_dpaa_bpid_info[bp_info->bpid].ptov_off
@@ -282,9 +282,8 @@
len, total_elt_sz * mp->size);
/* Detect pool area has sufficient space for elements in this memzone */
- if (len < total_elt_sz * mp->size)
- /* Else, Memory will be allocated from multiple memzones */
- bp_info->flags |= DPAA_MPOOL_MULTI_MEMZONE;
+ if (len >= total_elt_sz * mp->size)
+ bp_info->flags |= DPAA_MPOOL_SINGLE_SEGMENT;
return 0;
}
diff --git a/drivers/mempool/dpaa/dpaa_mempool.h
b/drivers/mempool/dpaa/dpaa_mempool.h
index 02aa513..9435dd2 100644
--- a/drivers/mempool/dpaa/dpaa_mempool.h
+++ b/drivers/mempool/dpaa/dpaa_mempool.h
@@ -28,8 +28,8 @@
/* Maximum release/acquire from BMAN */
#define DPAA_MBUF_MAX_ACQ_REL 8
-/* Buffers are allocated from multiple memzones i.e. non phys contiguous */
-#define DPAA_MPOOL_MULTI_MEMZONE 0x01
+/* Buffers are allocated from single mem segment i.e. phys contiguous */
+#define DPAA_MPOOL_SINGLE_SEGMENT 0x01
struct dpaa_bp_info {
struct rte_mempool *mp;
--
1.9.1
Re: [dpdk-dev] [PATCH v3 3/3] doc: convert license headers to SPDX tags
This patch will be good if you only add SPDX to it and NOT remove the original license text. i.e. only do following: > -.. BSD LICENSE > +.. SPDX-License-Identifier: BSD-3-Clause Around RC2 timeframe, I intend to do that. All the remaining but valid license files, we will add SPDX and NOT remove the license text. Regards, Hemant
[dpdk-dev] [PATCH] event/sw: fix memory leak if self test fails
If the self test is run and a test fails, the allocated
memory would was not freed. The passing case already freed
this memory correctly.
Fixes: 5e6eb5ccd788 ("event/sw: make test standalone")
Coverity ID: 257044
Cc: pbhagavat...@caviumnetworks.com
Signed-off-by: Harry van Haaren
---
---
drivers/event/sw/sw_evdev_selftest.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/event/sw/sw_evdev_selftest.c
b/drivers/event/sw/sw_evdev_selftest.c
index 8c2eef5..dbfa29b 100644
--- a/drivers/event/sw/sw_evdev_selftest.c
+++ b/drivers/event/sw/sw_evdev_selftest.c
@@ -3236,6 +3236,7 @@ test_sw_eventdev(void)
printf("SW Eventdev Selftest Successful.\n");
return 0;
test_fail:
+ free(t);
printf("SW Eventdev Selftest Failed.\n");
return -1;
}
--
2.7.4
[dpdk-dev] [RFC v2 02/17] mempool: add op to calculate memory size to be allocated
Size of memory chunk required to populate mempool objects depends
on how objects are stored in the memory. Different mempool drivers
may have different requirements and a new operation allows to
calculate memory size in accordance with driver requirements and
advertise requirements on minimum memory chunk size and alignment
in a generic way.
Suggested-by: Olivier Matz
Signed-off-by: Andrew Rybchenko
---
lib/librte_mempool/rte_mempool.c | 95 ++
lib/librte_mempool/rte_mempool.h | 63 +++-
lib/librte_mempool/rte_mempool_ops.c | 18 ++
lib/librte_mempool/rte_mempool_version.map | 8 +++
4 files changed, 159 insertions(+), 25 deletions(-)
diff --git a/lib/librte_mempool/rte_mempool.c b/lib/librte_mempool/rte_mempool.c
index e783b9a..1f54f95 100644
--- a/lib/librte_mempool/rte_mempool.c
+++ b/lib/librte_mempool/rte_mempool.c
@@ -233,13 +233,14 @@ rte_mempool_calc_obj_size(uint32_t elt_size, uint32_t
flags,
return sz->total_size;
}
-
/*
- * Calculate maximum amount of memory required to store given number of
objects.
+ * Internal function to calculate required memory chunk size shared
+ * by default implementation of the corresponding callback and
+ * deprecated external function.
*/
-size_t
-rte_mempool_xmem_size(uint32_t elt_num, size_t total_elt_sz, uint32_t pg_shift,
- unsigned int flags)
+static size_t
+rte_mempool_xmem_size_int(uint32_t elt_num, size_t total_elt_sz,
+ uint32_t pg_shift, unsigned int flags)
{
size_t obj_per_page, pg_num, pg_sz;
unsigned int mask;
@@ -264,6 +265,49 @@ rte_mempool_xmem_size(uint32_t elt_num, size_t
total_elt_sz, uint32_t pg_shift,
return pg_num << pg_shift;
}
+ssize_t
+rte_mempool_calc_mem_size_def(const struct rte_mempool *mp,
+ uint32_t obj_num, uint32_t pg_shift,
+ size_t *min_chunk_size,
+ __rte_unused size_t *align)
+{
+ unsigned int mp_flags;
+ int ret;
+ size_t total_elt_sz;
+ size_t mem_size;
+
+ /* Get mempool capabilities */
+ mp_flags = 0;
+ ret = rte_mempool_ops_get_capabilities(mp, &mp_flags);
+ if ((ret < 0) && (ret != -ENOTSUP))
+ return ret;
+
+ total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;
+
+ mem_size = rte_mempool_xmem_size_int(obj_num, total_elt_sz, pg_shift,
+mp->flags | mp_flags);
+
+ if (mp_flags & MEMPOOL_F_CAPA_PHYS_CONTIG)
+ *min_chunk_size = mem_size;
+ else
+ *min_chunk_size = RTE_MAX((size_t)1 << pg_shift, total_elt_sz);
+
+ /* No extra align requirements by default */
+
+ return mem_size;
+}
+
+/*
+ * Calculate maximum amount of memory required to store given number of
objects.
+ */
+size_t
+rte_mempool_xmem_size(uint32_t elt_num, size_t total_elt_sz, uint32_t pg_shift,
+ unsigned int flags)
+{
+ return rte_mempool_xmem_size_int(elt_num, total_elt_sz, pg_shift,
+flags);
+}
+
/*
* Calculate how much memory would be actually required with the
* given memory footprint to store required number of elements.
@@ -570,25 +614,16 @@ rte_mempool_populate_default(struct rte_mempool *mp)
unsigned int mz_flags = RTE_MEMZONE_1GB|RTE_MEMZONE_SIZE_HINT_ONLY;
char mz_name[RTE_MEMZONE_NAMESIZE];
const struct rte_memzone *mz;
- size_t size, total_elt_sz, align, pg_sz, pg_shift;
+ ssize_t mem_size;
+ size_t align, pg_sz, pg_shift;
rte_iova_t iova;
unsigned mz_id, n;
- unsigned int mp_flags;
int ret;
/* mempool must not be populated */
if (mp->nb_mem_chunks != 0)
return -EEXIST;
- /* Get mempool capabilities */
- mp_flags = 0;
- ret = rte_mempool_ops_get_capabilities(mp, &mp_flags);
- if ((ret < 0) && (ret != -ENOTSUP))
- return ret;
-
- /* update mempool capabilities */
- mp->flags |= mp_flags;
-
if (rte_eal_has_hugepages()) {
pg_shift = 0; /* not needed, zone is physically contiguous */
pg_sz = 0;
@@ -599,10 +634,15 @@ rte_mempool_populate_default(struct rte_mempool *mp)
align = pg_sz;
}
- total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;
for (mz_id = 0, n = mp->size; n > 0; mz_id++, n -= ret) {
- size = rte_mempool_xmem_size(n, total_elt_sz, pg_shift,
- mp->flags);
+ size_t min_chunk_size;
+
+ mem_size = rte_mempool_ops_calc_mem_size(mp, n, pg_shift,
+ &min_chunk_size, &align);
+ if (mem_size < 0) {
+ ret = mem_size;
+ goto fail;
+
[dpdk-dev] [RFC v2 10/17] mempool: remove callback to register memory area
The callback is not required any more since there is a new callback
to populate objects using provided memory area which provides
the same information.
Signed-off-by: Andrew Rybchenko
---
lib/librte_mempool/rte_mempool.c | 5 -
lib/librte_mempool/rte_mempool.h | 31 --
lib/librte_mempool/rte_mempool_ops.c | 14 --
lib/librte_mempool/rte_mempool_version.map | 1 -
4 files changed, 51 deletions(-)
diff --git a/lib/librte_mempool/rte_mempool.c b/lib/librte_mempool/rte_mempool.c
index 32b3f94..fc9c95a 100644
--- a/lib/librte_mempool/rte_mempool.c
+++ b/lib/librte_mempool/rte_mempool.c
@@ -416,11 +416,6 @@ rte_mempool_populate_iova(struct rte_mempool *mp, char
*vaddr,
mp->flags |= MEMPOOL_F_POOL_CREATED;
}
- /* Notify memory area to mempool */
- ret = rte_mempool_ops_register_memory_area(mp, vaddr, iova, len);
- if (ret != -ENOTSUP && ret < 0)
- return ret;
-
/* mempool is already populated */
if (mp->populated_size >= mp->size)
return -ENOSPC;
diff --git a/lib/librte_mempool/rte_mempool.h b/lib/librte_mempool/rte_mempool.h
index e95b1a7..6a0039d 100644
--- a/lib/librte_mempool/rte_mempool.h
+++ b/lib/librte_mempool/rte_mempool.h
@@ -399,12 +399,6 @@ typedef int (*rte_mempool_dequeue_t)(struct rte_mempool
*mp,
typedef unsigned (*rte_mempool_get_count)(const struct rte_mempool *mp);
/**
- * Notify new memory area to mempool.
- */
-typedef int (*rte_mempool_ops_register_memory_area_t)
-(const struct rte_mempool *mp, char *vaddr, rte_iova_t iova, size_t len);
-
-/**
* Calculate memory size required to store specified number of objects.
*
* Note that if object size is bigger then page size, then it assumes
@@ -499,10 +493,6 @@ struct rte_mempool_ops {
rte_mempool_dequeue_t dequeue; /**< Dequeue an object. */
rte_mempool_get_count get_count; /**< Get qty of available objs. */
/**
-* Notify new memory area to mempool
-*/
- rte_mempool_ops_register_memory_area_t register_memory_area;
- /**
* Optional callback to calculate memory size required to
* store specified number of objects.
*/
@@ -624,27 +614,6 @@ unsigned
rte_mempool_ops_get_count(const struct rte_mempool *mp);
/**
- * @internal wrapper for mempool_ops register_memory_area callback.
- * API to notify the mempool handler when a new memory area is added to pool.
- *
- * @param mp
- * Pointer to the memory pool.
- * @param vaddr
- * Pointer to the buffer virtual address.
- * @param iova
- * Pointer to the buffer IO address.
- * @param len
- * Pool size.
- * @return
- * - 0: Success;
- * - -ENOTSUP - doesn't support register_memory_area ops (valid error case).
- * - Otherwise, rte_mempool_populate_phys fails thus pool create fails.
- */
-int
-rte_mempool_ops_register_memory_area(const struct rte_mempool *mp,
- char *vaddr, rte_iova_t iova, size_t len);
-
-/**
* @internal wrapper for mempool_ops calc_mem_size callback.
* API to calculate size of memory required to store specified number of
* object.
diff --git a/lib/librte_mempool/rte_mempool_ops.c
b/lib/librte_mempool/rte_mempool_ops.c
index 5ab643b..37b0802 100644
--- a/lib/librte_mempool/rte_mempool_ops.c
+++ b/lib/librte_mempool/rte_mempool_ops.c
@@ -86,7 +86,6 @@ rte_mempool_register_ops(const struct rte_mempool_ops *h)
ops->enqueue = h->enqueue;
ops->dequeue = h->dequeue;
ops->get_count = h->get_count;
- ops->register_memory_area = h->register_memory_area;
ops->calc_mem_size = h->calc_mem_size;
ops->populate = h->populate;
@@ -128,19 +127,6 @@ rte_mempool_ops_get_count(const struct rte_mempool *mp)
}
/* wrapper to notify new memory area to external mempool */
-int
-rte_mempool_ops_register_memory_area(const struct rte_mempool *mp, char *vaddr,
- rte_iova_t iova, size_t len)
-{
- struct rte_mempool_ops *ops;
-
- ops = rte_mempool_get_ops(mp->ops_index);
-
- RTE_FUNC_PTR_OR_ERR_RET(ops->register_memory_area, -ENOTSUP);
- return ops->register_memory_area(mp, vaddr, iova, len);
-}
-
-/* wrapper to notify new memory area to external mempool */
ssize_t
rte_mempool_ops_calc_mem_size(const struct rte_mempool *mp,
uint32_t obj_num, uint32_t pg_shift,
diff --git a/lib/librte_mempool/rte_mempool_version.map
b/lib/librte_mempool/rte_mempool_version.map
index ab30b16..4f7e2b2 100644
--- a/lib/librte_mempool/rte_mempool_version.map
+++ b/lib/librte_mempool/rte_mempool_version.map
@@ -45,7 +45,6 @@ DPDK_16.07 {
DPDK_17.11 {
global:
- rte_mempool_ops_register_memory_area;
rte_mempool_populate_iova;
rte_mempool_populate_iova_tab;
--
2.7.4
[dpdk-dev] [RFC v2 11/17] mempool: ensure the mempool is initialized before populating
From: "Artem V. Andreev"
Callback to calculate required memory area size may require mempool
driver data to be already allocated and initialized.
Signed-off-by: Artem V. Andreev
Signed-off-by: Andrew Rybchenko
---
lib/librte_mempool/rte_mempool.c | 29 ++---
1 file changed, 22 insertions(+), 7 deletions(-)
diff --git a/lib/librte_mempool/rte_mempool.c b/lib/librte_mempool/rte_mempool.c
index fc9c95a..cbb4dd5 100644
--- a/lib/librte_mempool/rte_mempool.c
+++ b/lib/librte_mempool/rte_mempool.c
@@ -370,6 +370,21 @@ rte_mempool_free_memchunks(struct rte_mempool *mp)
}
}
+static int
+mempool_maybe_initialize(struct rte_mempool *mp)
+{
+ int ret;
+
+ /* create the internal ring if not already done */
+ if ((mp->flags & MEMPOOL_F_POOL_CREATED) == 0) {
+ ret = rte_mempool_ops_alloc(mp);
+ if (ret != 0)
+ return ret;
+ mp->flags |= MEMPOOL_F_POOL_CREATED;
+ }
+ return 0;
+}
+
int
rte_mempool_populate_one_by_one(struct rte_mempool *mp, unsigned int max_objs,
void *vaddr, rte_iova_t iova, size_t len,
@@ -408,13 +423,9 @@ rte_mempool_populate_iova(struct rte_mempool *mp, char
*vaddr,
struct rte_mempool_memhdr *memhdr;
int ret;
- /* create the internal ring if not already done */
- if ((mp->flags & MEMPOOL_F_POOL_CREATED) == 0) {
- ret = rte_mempool_ops_alloc(mp);
- if (ret != 0)
- return ret;
- mp->flags |= MEMPOOL_F_POOL_CREATED;
- }
+ ret = mempool_maybe_initialize(mp);
+ if (ret != 0)
+ return ret;
/* mempool is already populated */
if (mp->populated_size >= mp->size)
@@ -587,6 +598,10 @@ rte_mempool_populate_default(struct rte_mempool *mp)
unsigned mz_id, n;
int ret;
+ ret = mempool_maybe_initialize(mp);
+ if (ret != 0)
+ return ret;
+
/* mempool must not be populated */
if (mp->nb_mem_chunks != 0)
return -EEXIST;
--
2.7.4
[dpdk-dev] [RFC v2 01/17] mempool: fix phys contig check if populate default skipped
There is not specified dependency between rte_mempool_populate_default()
and rte_mempool_populate_iova(). So, the second should not rely on the
fact that the first adds capability flags to the mempool flags.
Fixes: 65cf769f5e6a ("mempool: detect physical contiguous objects")
Cc: sta...@dpdk.org
Signed-off-by: Andrew Rybchenko
---
lib/librte_mempool/rte_mempool.c | 11 +--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/lib/librte_mempool/rte_mempool.c b/lib/librte_mempool/rte_mempool.c
index 6d17022..e783b9a 100644
--- a/lib/librte_mempool/rte_mempool.c
+++ b/lib/librte_mempool/rte_mempool.c
@@ -362,6 +362,7 @@ rte_mempool_populate_iova(struct rte_mempool *mp, char
*vaddr,
void *opaque)
{
unsigned total_elt_sz;
+ unsigned int mp_cap_flags;
unsigned i = 0;
size_t off;
struct rte_mempool_memhdr *memhdr;
@@ -386,8 +387,14 @@ rte_mempool_populate_iova(struct rte_mempool *mp, char
*vaddr,
total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;
+ /* Get mempool capabilities */
+ mp_cap_flags = 0;
+ ret = rte_mempool_ops_get_capabilities(mp, &mp_cap_flags);
+ if ((ret < 0) && (ret != -ENOTSUP))
+ return ret;
+
/* Detect pool area has sufficient space for elements */
- if (mp->flags & MEMPOOL_F_CAPA_PHYS_CONTIG) {
+ if (mp_cap_flags & MEMPOOL_F_CAPA_PHYS_CONTIG) {
if (len < total_elt_sz * mp->size) {
RTE_LOG(ERR, MEMPOOL,
"pool area %" PRIx64 " not enough\n",
@@ -407,7 +414,7 @@ rte_mempool_populate_iova(struct rte_mempool *mp, char
*vaddr,
memhdr->free_cb = free_cb;
memhdr->opaque = opaque;
- if (mp->flags & MEMPOOL_F_CAPA_BLK_ALIGNED_OBJECTS)
+ if (mp_cap_flags & MEMPOOL_F_CAPA_BLK_ALIGNED_OBJECTS)
/* align object start address to a multiple of total_elt_sz */
off = total_elt_sz - ((uintptr_t)vaddr % total_elt_sz);
else if (mp->flags & MEMPOOL_F_NO_CACHE_ALIGN)
--
2.7.4
[dpdk-dev] [RFC v2 04/17] mempool: add op to populate objects using provided memory
The callback allows to customize how objects are stored in the
memory chunk. Default implementation of the callback which simply
puts objects one by one is available.
Suggested-by: Olivier Matz
Signed-off-by: Andrew Rybchenko
---
lib/librte_mempool/rte_mempool.c | 44 +++-
lib/librte_mempool/rte_mempool.h | 83 ++
lib/librte_mempool/rte_mempool_ops.c | 18 +++
lib/librte_mempool/rte_mempool_version.map | 1 +
4 files changed, 133 insertions(+), 13 deletions(-)
diff --git a/lib/librte_mempool/rte_mempool.c b/lib/librte_mempool/rte_mempool.c
index 1f54f95..c5003a9 100644
--- a/lib/librte_mempool/rte_mempool.c
+++ b/lib/librte_mempool/rte_mempool.c
@@ -145,9 +145,6 @@ mempool_add_elem(struct rte_mempool *mp, void *obj,
rte_iova_t iova)
tlr = __mempool_get_trailer(obj);
tlr->cookie = RTE_MEMPOOL_TRAILER_COOKIE;
#endif
-
- /* enqueue in ring */
- rte_mempool_ops_enqueue_bulk(mp, &obj, 1);
}
/* call obj_cb() for each mempool element */
@@ -396,6 +393,30 @@ rte_mempool_free_memchunks(struct rte_mempool *mp)
}
}
+int
+rte_mempool_populate_one_by_one(struct rte_mempool *mp, unsigned int max_objs,
+ void *vaddr, rte_iova_t iova, size_t len,
+ rte_mempool_populate_obj_cb_t *obj_cb)
+{
+ size_t total_elt_sz;
+ size_t off;
+ unsigned int i;
+ void *obj;
+
+ total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;
+
+ for (off = 0, i = 0; off + total_elt_sz <= len && i < max_objs; i++) {
+ off += mp->header_size;
+ obj = (char *)vaddr + off;
+ obj_cb(mp, obj,
+ (iova == RTE_BAD_IOVA) ? RTE_BAD_IOVA : (iova + off));
+ rte_mempool_ops_enqueue_bulk(mp, &obj, 1);
+ off += mp->elt_size + mp->trailer_size;
+ }
+
+ return i;
+}
+
/* Add objects in the pool, using a physically contiguous memory
* zone. Return the number of objects added, or a negative value
* on error.
@@ -466,16 +487,13 @@ rte_mempool_populate_iova(struct rte_mempool *mp, char
*vaddr,
else
off = RTE_PTR_ALIGN_CEIL(vaddr, RTE_CACHE_LINE_SIZE) - vaddr;
- while (off + total_elt_sz <= len && mp->populated_size < mp->size) {
- off += mp->header_size;
- if (iova == RTE_BAD_IOVA)
- mempool_add_elem(mp, (char *)vaddr + off,
- RTE_BAD_IOVA);
- else
- mempool_add_elem(mp, (char *)vaddr + off, iova + off);
- off += mp->elt_size + mp->trailer_size;
- i++;
- }
+ if (off > len)
+ return -EINVAL;
+
+ i = rte_mempool_ops_populate(mp, mp->size - mp->populated_size,
+ (char *)vaddr + off,
+ (iova == RTE_BAD_IOVA) ? RTE_BAD_IOVA : (iova + off),
+ len - off, mempool_add_elem);
/* not enough room to store one object */
if (i == 0)
diff --git a/lib/librte_mempool/rte_mempool.h b/lib/librte_mempool/rte_mempool.h
index be8a371..f6ffab9 100644
--- a/lib/librte_mempool/rte_mempool.h
+++ b/lib/librte_mempool/rte_mempool.h
@@ -461,6 +461,59 @@ ssize_t rte_mempool_calc_mem_size_def(const struct
rte_mempool *mp,
uint32_t obj_num, uint32_t pg_shift,
size_t *min_chunk_size, size_t *align);
+/**
+ * Function to be called for each populated object.
+ *
+ * @param mp
+ * A pointer to the mempool structure.
+ * @param vaddr
+ * Object virtual address.
+ * @param iova
+ * Input/output virtual addresss of the object or #RTE_BAD_IOVA.
+ */
+typedef void (rte_mempool_populate_obj_cb_t)(struct rte_mempool *mp,
+ void *vaddr, rte_iova_t iova);
+
+/**
+ * Populate memory pool objects using provided memory chunk.
+ *
+ * Populated objects should be enqueued to the pool, e.g. using
+ * rte_mempool_ops_enqueue_bulk().
+ *
+ * If the given IO address is unknown (iova = RTE_BAD_IOVA),
+ * the chunk doesn't need to be physically contiguous (only virtually),
+ * and allocated objects may span two pages.
+ *
+ * @param mp
+ * A pointer to the mempool structure.
+ * @param max_objs
+ * Maximum number of objects to be populated.
+ * @param vaddr
+ * The virtual address of memory that should be used to store objects.
+ * @param iova
+ * The IO address
+ * @param len
+ * The length of memory in bytes.
+ * @param obj_cb
+ * Callback function to be executed for each populated object.
+ * @return
+ * The number of objects added on success.
+ * On error, no objects are populated and a negative errno is returned.
+ */
+typedef int (*rte_mempool_populate_t)(struct rte_mempool *mp,
+ unsigned int max_objs,
+ void *vaddr, rte_iova_t iova, size_t len,
+ rte_mempool_populate_obj_cb_t *obj_cb);
+
+/**
+ * Defau
[dpdk-dev] [RFC v2 07/17] mempool: deprecate xmem functions
Suggested-by: Olivier Matz Signed-off-by: Andrew Rybchenko --- lib/librte_mempool/rte_mempool.h | 3 +++ 1 file changed, 3 insertions(+) diff --git a/lib/librte_mempool/rte_mempool.h b/lib/librte_mempool/rte_mempool.h index 697d618..e95b1a7 100644 --- a/lib/librte_mempool/rte_mempool.h +++ b/lib/librte_mempool/rte_mempool.h @@ -916,6 +916,7 @@ rte_mempool_create(const char *name, unsigned n, unsigned elt_size, * The pointer to the new allocated mempool, on success. NULL on error * with rte_errno set appropriately. See rte_mempool_create() for details. */ +__rte_deprecated struct rte_mempool * rte_mempool_xmem_create(const char *name, unsigned n, unsigned elt_size, unsigned cache_size, unsigned private_data_size, @@ -1678,6 +1679,7 @@ uint32_t rte_mempool_calc_obj_size(uint32_t elt_size, uint32_t flags, * @return * Required memory size aligned at page boundary. */ +__rte_deprecated size_t rte_mempool_xmem_size(uint32_t elt_num, size_t total_elt_sz, uint32_t pg_shift, unsigned int flags); @@ -1709,6 +1711,7 @@ size_t rte_mempool_xmem_size(uint32_t elt_num, size_t total_elt_sz, * buffer is too small, return a negative value whose absolute value * is the actual number of elements that can be stored in that buffer. */ +__rte_deprecated ssize_t rte_mempool_xmem_usage(void *vaddr, uint32_t elt_num, size_t total_elt_sz, const rte_iova_t iova[], uint32_t pg_num, uint32_t pg_shift, unsigned int flags); -- 2.7.4
[dpdk-dev] [RFC v2 05/17] mempool/octeontx: implement callback to populate objects
Custom callback is required to fullfil requirement to align
object virtual address to total object size.
Signed-off-by: Andrew Rybchenko
---
drivers/mempool/octeontx/rte_mempool_octeontx.c | 28 +
1 file changed, 28 insertions(+)
diff --git a/drivers/mempool/octeontx/rte_mempool_octeontx.c
b/drivers/mempool/octeontx/rte_mempool_octeontx.c
index 4ec5efe..6563e80 100644
--- a/drivers/mempool/octeontx/rte_mempool_octeontx.c
+++ b/drivers/mempool/octeontx/rte_mempool_octeontx.c
@@ -174,6 +174,33 @@ octeontx_fpavf_register_memory_area(const struct
rte_mempool *mp,
return octeontx_fpavf_pool_set_range(pool_bar, len, vaddr, gpool);
}
+static int
+octeontx_fpavf_populate(struct rte_mempool *mp, unsigned int max_objs,
+ void *vaddr, rte_iova_t iova, size_t len,
+ rte_mempool_populate_obj_cb_t *obj_cb)
+{
+ size_t total_elt_sz;
+ size_t off;
+
+ if (iova == RTE_BAD_IOVA)
+ return -EINVAL;
+
+ total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;
+
+ /* align object start address to a multiple of total_elt_sz */
+ off = total_elt_sz - ((uintptr_t)vaddr % total_elt_sz);
+
+ if (len < off)
+ return -EINVAL;
+
+ vaddr = (char *)vaddr + off;
+ iova += off;
+ len -= off;
+
+ return rte_mempool_populate_one_by_one(mp, max_objs, vaddr, iova, len,
+ obj_cb);
+}
+
static struct rte_mempool_ops octeontx_fpavf_ops = {
.name = "octeontx_fpavf",
.alloc = octeontx_fpavf_alloc,
@@ -184,6 +211,7 @@ static struct rte_mempool_ops octeontx_fpavf_ops = {
.get_capabilities = octeontx_fpavf_get_capabilities,
.register_memory_area = octeontx_fpavf_register_memory_area,
.calc_mem_size = octeontx_fpavf_calc_mem_size,
+ .populate = octeontx_fpavf_populate,
};
MEMPOOL_REGISTER_OPS(octeontx_fpavf_ops);
--
2.7.4
[dpdk-dev] [RFC v2 03/17] mempool/octeontx: add callback to calculate memory size
The driver requires one and only one physically contiguous
memory chunk for all objects.
Signed-off-by: Andrew Rybchenko
---
drivers/mempool/octeontx/rte_mempool_octeontx.c | 25 +
1 file changed, 25 insertions(+)
diff --git a/drivers/mempool/octeontx/rte_mempool_octeontx.c
b/drivers/mempool/octeontx/rte_mempool_octeontx.c
index d143d05..4ec5efe 100644
--- a/drivers/mempool/octeontx/rte_mempool_octeontx.c
+++ b/drivers/mempool/octeontx/rte_mempool_octeontx.c
@@ -136,6 +136,30 @@ octeontx_fpavf_get_capabilities(const struct rte_mempool
*mp,
return 0;
}
+static ssize_t
+octeontx_fpavf_calc_mem_size(const struct rte_mempool *mp,
+uint32_t obj_num, uint32_t pg_shift,
+size_t *min_chunk_size, size_t *align)
+{
+ ssize_t mem_size;
+
+ /*
+* Simply need space for one more object to be able to
+* fullfil alignment requirements.
+*/
+ mem_size = rte_mempool_calc_mem_size_def(mp, obj_num + 1, pg_shift,
+min_chunk_size, align);
+ if (mem_size >= 0) {
+ /*
+* The whole memory area containing the objects must be
+* physically contiguous.
+*/
+ *min_chunk_size = mem_size;
+ }
+
+ return mem_size;
+}
+
static int
octeontx_fpavf_register_memory_area(const struct rte_mempool *mp,
char *vaddr, rte_iova_t paddr, size_t len)
@@ -159,6 +183,7 @@ static struct rte_mempool_ops octeontx_fpavf_ops = {
.get_count = octeontx_fpavf_get_count,
.get_capabilities = octeontx_fpavf_get_capabilities,
.register_memory_area = octeontx_fpavf_register_memory_area,
+ .calc_mem_size = octeontx_fpavf_calc_mem_size,
};
MEMPOOL_REGISTER_OPS(octeontx_fpavf_ops);
--
2.7.4
[dpdk-dev] [RFC v2 09/17] mempool/dpaa: convert to use populate driver op
Populate mempool driver callback is executed a bit later than
register memory area, provides the same information and will
substitute the later since it gives more flexibility and in addition
to notification about memory area allows to customize how mempool
objects are stored in memory.
Signed-off-by: Andrew Rybchenko
---
drivers/mempool/dpaa/dpaa_mempool.c | 13 +++--
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/drivers/mempool/dpaa/dpaa_mempool.c
b/drivers/mempool/dpaa/dpaa_mempool.c
index ddc4e47..a179804 100644
--- a/drivers/mempool/dpaa/dpaa_mempool.c
+++ b/drivers/mempool/dpaa/dpaa_mempool.c
@@ -260,10 +260,9 @@ dpaa_mbuf_get_count(const struct rte_mempool *mp)
}
static int
-dpaa_register_memory_area(const struct rte_mempool *mp,
- char *vaddr __rte_unused,
- rte_iova_t paddr __rte_unused,
- size_t len)
+dpaa_populate(const struct rte_mempool *mp, unsigned int max_objs,
+ char *vaddr, rte_iova_t paddr, size_t len,
+ rte_mempool_populate_obj_cb_t *obj_cb)
{
struct dpaa_bp_info *bp_info;
unsigned int total_elt_sz;
@@ -286,7 +285,9 @@ dpaa_register_memory_area(const struct rte_mempool *mp,
/* Else, Memory will be allocated from multiple memzones */
bp_info->flags |= DPAA_MPOOL_MULTI_MEMZONE;
- return 0;
+ return rte_mempool_populate_one_by_one(mp, max_objs, vaddr, paddr, len,
+ obj_cb);
+
}
struct rte_mempool_ops dpaa_mpool_ops = {
@@ -296,7 +297,7 @@ struct rte_mempool_ops dpaa_mpool_ops = {
.enqueue = dpaa_mbuf_free_bulk,
.dequeue = dpaa_mbuf_alloc_bulk,
.get_count = dpaa_mbuf_get_count,
- .register_memory_area = dpaa_register_memory_area,
+ .populate = dpaa_populate,
};
MEMPOOL_REGISTER_OPS(dpaa_mpool_ops);
--
2.7.4
[dpdk-dev] [RFC v2 00/17] mempool: add bucket mempool driver
The patch series starts from generic enhancements suggested by Olivier. Basically it adds driver callbacks to calculate required memory size and to populate objects using provided memory area. It allows to remove so-called capability flags used before to tell generic code how to allocate and slice allocated memory into mempool objects. Clean up which removes get_capabilities and register_memory_area is not strictly required, but I think right thing to do. Existing mempool drivers are updated. I've kept rte_mempool_populate_iova_tab() intact since it seems to be not directly related XMEM API functions. The patch series adds bucket mempool driver which allows to allocate (both physically and virtually) contiguous blocks of objects and adds mempool API to do it. It is still capable to provide separate objects, but it is definitely more heavy-weight than ring/stack drivers. The driver will be used by the future Solarflare driver enhancements which allow to utilize physical contiguous blocks in the NIC hardware/firmware. The target usecase is dequeue in blocks and enqueue separate objects back (which are collected in buckets to be dequeued). So, the memory pool with bucket driver is created by an application and provided to networking PMD receive queue. The choice of bucket driver is done using rte_eth_dev_pool_ops_supported(). A PMD that relies upon contiguous block allocation should report the bucket driver as the only supported and preferred one. Introduction of the contiguous block dequeue operation is proven by performance measurements using autotest with minor enhancements: - in the original test bulks are powers of two, which is unacceptable for us, so they are changed to multiple of contig_block_size; - the test code is duplicated to support plain dequeue and dequeue_contig_blocks; - all the extra test variations (with/without cache etc) are eliminated; - a fake read from the dequeued buffer is added (in both cases) to simulate mbufs access. start performance test for bucket (without cache) mempool_autotest cache= 0 cores= 1 n_get_bulk= 15 n_put_bulk= 1 n_keep= 30 Srate_persec= 111935488 mempool_autotest cache= 0 cores= 1 n_get_bulk= 15 n_put_bulk= 1 n_keep= 60 Srate_persec= 115290931 mempool_autotest cache= 0 cores= 1 n_get_bulk= 15 n_put_bulk= 15 n_keep= 30 Srate_persec= 353055539 mempool_autotest cache= 0 cores= 1 n_get_bulk= 15 n_put_bulk= 15 n_keep= 60 Srate_persec= 353330790 mempool_autotest cache= 0 cores= 2 n_get_bulk= 15 n_put_bulk= 1 n_keep= 30 Srate_persec= 224657407 mempool_autotest cache= 0 cores= 2 n_get_bulk= 15 n_put_bulk= 1 n_keep= 60 Srate_persec= 230411468 mempool_autotest cache= 0 cores= 2 n_get_bulk= 15 n_put_bulk= 15 n_keep= 30 Srate_persec= 706700902 mempool_autotest cache= 0 cores= 2 n_get_bulk= 15 n_put_bulk= 15 n_keep= 60 Srate_persec= 703673139 mempool_autotest cache= 0 cores= 4 n_get_bulk= 15 n_put_bulk= 1 n_keep= 30 Srate_persec= 425236887 mempool_autotest cache= 0 cores= 4 n_get_bulk= 15 n_put_bulk= 1 n_keep= 60 Srate_persec= 437295512 mempool_autotest cache= 0 cores= 4 n_get_bulk= 15 n_put_bulk= 15 n_keep= 30 Srate_persec= 1343409356 mempool_autotest cache= 0 cores= 4 n_get_bulk= 15 n_put_bulk= 15 n_keep= 60 Srate_persec= 1336567397 start performance test for bucket (without cache + contiguous dequeue) mempool_autotest cache= 0 cores= 1 n_get_bulk= 15 n_put_bulk= 1 n_keep= 30 Crate_persec= 122945536 mempool_autotest cache= 0 cores= 1 n_get_bulk= 15 n_put_bulk= 1 n_keep= 60 Crate_persec= 126458265 mempool_autotest cache= 0 cores= 1 n_get_bulk= 15 n_put_bulk= 15 n_keep= 30 Crate_persec= 374262988 mempool_autotest cache= 0 cores= 1 n_get_bulk= 15 n_put_bulk= 15 n_keep= 60 Crate_persec= 377316966 mempool_autotest cache= 0 cores= 2 n_get_bulk= 15 n_put_bulk= 1 n_keep= 30 Crate_persec= 244842496 mempool_autotest cache= 0 cores= 2 n_get_bulk= 15 n_put_bulk= 1 n_keep= 60 Crate_persec= 251618917 mempool_autotest cache= 0 cores= 2 n_get_bulk= 15 n_put_bulk= 15 n_keep= 30 Crate_persec= 751226060 mempool_autotest cache= 0 cores= 2 n_get_bulk= 15 n_put_bulk= 15 n_keep= 60 Crate_persec= 756233010 mempool_autotest cache= 0 cores= 4 n_get_bulk= 15 n_put_bulk= 1 n_keep= 30 Crate_persec= 462068120 mempool_autotest cache= 0 cores= 4 n_get_bulk= 15 n_put_bulk= 1 n_keep= 60 Crate_persec= 476997221 mempool_autotest cache= 0 cores= 4 n_get_bulk= 15 n_put_bulk= 15 n_keep= 30 Crate_persec= 1432171313 mempool_autotest cache= 0 cores= 4 n_get_bulk= 15 n_put_bulk= 15 n_keep= 60 Crate_persec= 1438829771 The number of objects in the contiguous block is a function of bucket memory size (.config option) and total element size. In the future additional API with possibility to pass parameters on mempool allocation may be added. It breaks ABI since changes rte_mempool_ops. Also it removes rt
[dpdk-dev] [RFC v2 06/17] mempool: remove callback to get capabilities
The callback was introduced to let generic code to know octeontx
mempool driver requirements to use single physically contiguous
memory chunk to store all objects and align object address to
total object size. Now these requirements are met using a new
callbacks to calculate required memory chunk size and to populate
objects using provided memory chunk.
These capability flags are not used anywhere else.
Restricting capabilities to flags is not generic and likely to
be insufficient to describe mempool driver features. If required
in the future, API which returns structured information may be
added.
Signed-off-by: Andrew Rybchenko
---
drivers/mempool/octeontx/rte_mempool_octeontx.c | 11 -
lib/librte_mempool/rte_mempool.c| 56 +++--
lib/librte_mempool/rte_mempool.h| 44 ---
lib/librte_mempool/rte_mempool_ops.c| 14 ---
lib/librte_mempool/rte_mempool_version.map | 1 -
5 files changed, 5 insertions(+), 121 deletions(-)
diff --git a/drivers/mempool/octeontx/rte_mempool_octeontx.c
b/drivers/mempool/octeontx/rte_mempool_octeontx.c
index 6563e80..36cc23b 100644
--- a/drivers/mempool/octeontx/rte_mempool_octeontx.c
+++ b/drivers/mempool/octeontx/rte_mempool_octeontx.c
@@ -126,16 +126,6 @@ octeontx_fpavf_get_count(const struct rte_mempool *mp)
return octeontx_fpa_bufpool_free_count(pool);
}
-static int
-octeontx_fpavf_get_capabilities(const struct rte_mempool *mp,
- unsigned int *flags)
-{
- RTE_SET_USED(mp);
- *flags |= (MEMPOOL_F_CAPA_PHYS_CONTIG |
- MEMPOOL_F_CAPA_BLK_ALIGNED_OBJECTS);
- return 0;
-}
-
static ssize_t
octeontx_fpavf_calc_mem_size(const struct rte_mempool *mp,
uint32_t obj_num, uint32_t pg_shift,
@@ -208,7 +198,6 @@ static struct rte_mempool_ops octeontx_fpavf_ops = {
.enqueue = octeontx_fpavf_enqueue,
.dequeue = octeontx_fpavf_dequeue,
.get_count = octeontx_fpavf_get_count,
- .get_capabilities = octeontx_fpavf_get_capabilities,
.register_memory_area = octeontx_fpavf_register_memory_area,
.calc_mem_size = octeontx_fpavf_calc_mem_size,
.populate = octeontx_fpavf_populate,
diff --git a/lib/librte_mempool/rte_mempool.c b/lib/librte_mempool/rte_mempool.c
index c5003a9..32b3f94 100644
--- a/lib/librte_mempool/rte_mempool.c
+++ b/lib/librte_mempool/rte_mempool.c
@@ -237,15 +237,9 @@ rte_mempool_calc_obj_size(uint32_t elt_size, uint32_t
flags,
*/
static size_t
rte_mempool_xmem_size_int(uint32_t elt_num, size_t total_elt_sz,
- uint32_t pg_shift, unsigned int flags)
+ uint32_t pg_shift, __rte_unused unsigned int flags)
{
size_t obj_per_page, pg_num, pg_sz;
- unsigned int mask;
-
- mask = MEMPOOL_F_CAPA_BLK_ALIGNED_OBJECTS | MEMPOOL_F_CAPA_PHYS_CONTIG;
- if ((flags & mask) == mask)
- /* alignment need one additional object */
- elt_num += 1;
if (total_elt_sz == 0)
return 0;
@@ -268,26 +262,15 @@ rte_mempool_calc_mem_size_def(const struct rte_mempool
*mp,
size_t *min_chunk_size,
__rte_unused size_t *align)
{
- unsigned int mp_flags;
- int ret;
size_t total_elt_sz;
size_t mem_size;
- /* Get mempool capabilities */
- mp_flags = 0;
- ret = rte_mempool_ops_get_capabilities(mp, &mp_flags);
- if ((ret < 0) && (ret != -ENOTSUP))
- return ret;
-
total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;
mem_size = rte_mempool_xmem_size_int(obj_num, total_elt_sz, pg_shift,
-mp->flags | mp_flags);
+mp->flags);
- if (mp_flags & MEMPOOL_F_CAPA_PHYS_CONTIG)
- *min_chunk_size = mem_size;
- else
- *min_chunk_size = RTE_MAX((size_t)1 << pg_shift, total_elt_sz);
+ *min_chunk_size = RTE_MAX((size_t)1 << pg_shift, total_elt_sz);
/* No extra align requirements by default */
@@ -312,18 +295,12 @@ rte_mempool_xmem_size(uint32_t elt_num, size_t
total_elt_sz, uint32_t pg_shift,
ssize_t
rte_mempool_xmem_usage(__rte_unused void *vaddr, uint32_t elt_num,
size_t total_elt_sz, const rte_iova_t iova[], uint32_t pg_num,
- uint32_t pg_shift, unsigned int flags)
+ uint32_t pg_shift, __rte_unused unsigned int flags)
{
uint32_t elt_cnt = 0;
rte_iova_t start, end;
uint32_t iova_idx;
size_t pg_sz = (size_t)1 << pg_shift;
- unsigned int mask;
-
- mask = MEMPOOL_F_CAPA_BLK_ALIGNED_OBJECTS | MEMPOOL_F_CAPA_PHYS_CONTIG;
- if ((flags & mask) == mask)
- /* alignment need one additional object */
- elt_num += 1;
/* if iova is NULL, assume contiguous m
[dpdk-dev] [RFC v2 13/17] mempool: support flushing the default cache of the mempool
From: "Artem V. Andreev"
Mempool get/put API cares about cache itself, but sometimes it is
required to flush the cache explicitly.
The function is moved in the file since it now requires
rte_mempool_default_cache().
Signed-off-by: Artem V. Andreev
Signed-off-by: Andrew Rybchenko
---
lib/librte_mempool/rte_mempool.h | 36
1 file changed, 20 insertions(+), 16 deletions(-)
diff --git a/lib/librte_mempool/rte_mempool.h b/lib/librte_mempool/rte_mempool.h
index 6a0039d..16d95ae 100644
--- a/lib/librte_mempool/rte_mempool.h
+++ b/lib/librte_mempool/rte_mempool.h
@@ -1148,22 +1148,6 @@ void
rte_mempool_cache_free(struct rte_mempool_cache *cache);
/**
- * Flush a user-owned mempool cache to the specified mempool.
- *
- * @param cache
- * A pointer to the mempool cache.
- * @param mp
- * A pointer to the mempool.
- */
-static __rte_always_inline void
-rte_mempool_cache_flush(struct rte_mempool_cache *cache,
- struct rte_mempool *mp)
-{
- rte_mempool_ops_enqueue_bulk(mp, cache->objs, cache->len);
- cache->len = 0;
-}
-
-/**
* Get a pointer to the per-lcore default mempool cache.
*
* @param mp
@@ -1186,6 +1170,26 @@ rte_mempool_default_cache(struct rte_mempool *mp,
unsigned lcore_id)
}
/**
+ * Flush a user-owned mempool cache to the specified mempool.
+ *
+ * @param cache
+ * A pointer to the mempool cache.
+ * @param mp
+ * A pointer to the mempool.
+ */
+static __rte_always_inline void
+rte_mempool_cache_flush(struct rte_mempool_cache *cache,
+ struct rte_mempool *mp)
+{
+ if (cache == NULL)
+ cache = rte_mempool_default_cache(mp, rte_lcore_id());
+ if (cache == NULL || cache->len == 0)
+ return;
+ rte_mempool_ops_enqueue_bulk(mp, cache->objs, cache->len);
+ cache->len = 0;
+}
+
+/**
* @internal Put several objects back in the mempool; used internally.
* @param mp
* A pointer to the mempool structure.
--
2.7.4
[dpdk-dev] [RFC v2 14/17] mempool: implement abstract mempool info API
From: "Artem V. Andreev"
Primarily, it is intended as a way for the mempool driver to provide
additional information on how it lays up objects inside the mempool.
Signed-off-by: Artem V. Andreev
Signed-off-by: Andrew Rybchenko
---
lib/librte_mempool/rte_mempool.h | 31 +++
lib/librte_mempool/rte_mempool_ops.c | 15 +++
2 files changed, 46 insertions(+)
diff --git a/lib/librte_mempool/rte_mempool.h b/lib/librte_mempool/rte_mempool.h
index 16d95ae..75630e6 100644
--- a/lib/librte_mempool/rte_mempool.h
+++ b/lib/librte_mempool/rte_mempool.h
@@ -218,6 +218,11 @@ struct rte_mempool_memhdr {
void *opaque;/**< Argument passed to the free callback */
};
+/*
+ * Additional information about the mempool
+ */
+struct rte_mempool_info;
+
/**
* The RTE mempool structure.
*/
@@ -484,6 +489,13 @@ int rte_mempool_populate_one_by_one(struct rte_mempool *mp,
void *vaddr, rte_iova_t iova, size_t len,
rte_mempool_populate_obj_cb_t *obj_cb);
+/**
+ * Get some additional information about a mempool.
+ */
+typedef int (*rte_mempool_get_info_t)(const struct rte_mempool *mp,
+ struct rte_mempool_info *info);
+
+
/** Structure defining mempool operations structure */
struct rte_mempool_ops {
char name[RTE_MEMPOOL_OPS_NAMESIZE]; /**< Name of mempool ops struct. */
@@ -502,6 +514,10 @@ struct rte_mempool_ops {
* provided memory chunk.
*/
rte_mempool_populate_t populate;
+ /**
+* Get mempool info
+*/
+ rte_mempool_get_info_t get_info;
} __rte_cache_aligned;
#define RTE_MEMPOOL_MAX_OPS_IDX 16 /**< Max registered ops structs */
@@ -662,6 +678,21 @@ int rte_mempool_ops_populate(struct rte_mempool *mp,
unsigned int max_objs,
rte_mempool_populate_obj_cb_t *obj_cb);
/**
+ * @internal wrapper for mempool_ops get_info callback.
+ *
+ * @param mp [in]
+ * Pointer to the memory pool.
+ * @param info [out]
+ * Pointer to the rte_mempool_info structure
+ * @return
+ * - 0: Success; The mempool driver supports retrieving supplementary
+ *mempool information
+ * - -ENOTSUP - doesn't support get_info ops (valid case).
+ */
+int rte_mempool_ops_get_info(const struct rte_mempool *mp,
+struct rte_mempool_info *info);
+
+/**
* @internal wrapper for mempool_ops free callback.
*
* @param mp
diff --git a/lib/librte_mempool/rte_mempool_ops.c
b/lib/librte_mempool/rte_mempool_ops.c
index 37b0802..949ab43 100644
--- a/lib/librte_mempool/rte_mempool_ops.c
+++ b/lib/librte_mempool/rte_mempool_ops.c
@@ -88,6 +88,7 @@ rte_mempool_register_ops(const struct rte_mempool_ops *h)
ops->get_count = h->get_count;
ops->calc_mem_size = h->calc_mem_size;
ops->populate = h->populate;
+ ops->get_info = h->get_info;
rte_spinlock_unlock(&rte_mempool_ops_table.sl);
@@ -160,6 +161,20 @@ rte_mempool_ops_populate(struct rte_mempool *mp, unsigned
int max_objs,
return ops->populate(mp, max_objs, vaddr, iova, len, obj_cb);
}
+/* wrapper to get additional mempool info */
+int
+rte_mempool_ops_get_info(const struct rte_mempool *mp,
+struct rte_mempool_info *info)
+{
+ struct rte_mempool_ops *ops;
+
+ ops = rte_mempool_get_ops(mp->ops_index);
+
+ RTE_FUNC_PTR_OR_ERR_RET(ops->get_info, -ENOTSUP);
+ return ops->get_info(mp, info);
+}
+
+
/* sets mempool ops previously registered by rte_mempool_register_ops. */
int
rte_mempool_set_ops_byname(struct rte_mempool *mp, const char *name,
--
2.7.4
[dpdk-dev] [RFC v2 08/17] mempool/octeontx: prepare to remove register memory area op
Callback to populate pool objects has all required information and
executed a bit later than register memory area callback.
Signed-off-by: Andrew Rybchenko
---
drivers/mempool/octeontx/rte_mempool_octeontx.c | 25 ++---
1 file changed, 10 insertions(+), 15 deletions(-)
diff --git a/drivers/mempool/octeontx/rte_mempool_octeontx.c
b/drivers/mempool/octeontx/rte_mempool_octeontx.c
index 36cc23b..8700bfb 100644
--- a/drivers/mempool/octeontx/rte_mempool_octeontx.c
+++ b/drivers/mempool/octeontx/rte_mempool_octeontx.c
@@ -151,26 +151,15 @@ octeontx_fpavf_calc_mem_size(const struct rte_mempool *mp,
}
static int
-octeontx_fpavf_register_memory_area(const struct rte_mempool *mp,
- char *vaddr, rte_iova_t paddr, size_t len)
-{
- RTE_SET_USED(paddr);
- uint8_t gpool;
- uintptr_t pool_bar;
-
- gpool = octeontx_fpa_bufpool_gpool(mp->pool_id);
- pool_bar = mp->pool_id & ~(uint64_t)FPA_GPOOL_MASK;
-
- return octeontx_fpavf_pool_set_range(pool_bar, len, vaddr, gpool);
-}
-
-static int
octeontx_fpavf_populate(struct rte_mempool *mp, unsigned int max_objs,
void *vaddr, rte_iova_t iova, size_t len,
rte_mempool_populate_obj_cb_t *obj_cb)
{
size_t total_elt_sz;
size_t off;
+ uint8_t gpool;
+ uintptr_t pool_bar;
+ int ret;
if (iova == RTE_BAD_IOVA)
return -EINVAL;
@@ -187,6 +176,13 @@ octeontx_fpavf_populate(struct rte_mempool *mp, unsigned
int max_objs,
iova += off;
len -= off;
+ gpool = octeontx_fpa_bufpool_gpool(mp->pool_id);
+ pool_bar = mp->pool_id & ~(uint64_t)FPA_GPOOL_MASK;
+
+ ret = octeontx_fpavf_pool_set_range(pool_bar, len, vaddr, gpool);
+ if (ret < 0)
+ return ret;
+
return rte_mempool_populate_one_by_one(mp, max_objs, vaddr, iova, len,
obj_cb);
}
@@ -198,7 +194,6 @@ static struct rte_mempool_ops octeontx_fpavf_ops = {
.enqueue = octeontx_fpavf_enqueue,
.dequeue = octeontx_fpavf_dequeue,
.get_count = octeontx_fpavf_get_count,
- .register_memory_area = octeontx_fpavf_register_memory_area,
.calc_mem_size = octeontx_fpavf_calc_mem_size,
.populate = octeontx_fpavf_populate,
};
--
2.7.4
[dpdk-dev] [RFC v2 17/17] mempool/bucket: do not allow one lcore to grab all buckets
From: "Artem V. Andreev"
Signed-off-by: Artem V. Andreev
Signed-off-by: Andrew Rybchenko
---
drivers/mempool/bucket/rte_mempool_bucket.c | 13 +
1 file changed, 13 insertions(+)
diff --git a/drivers/mempool/bucket/rte_mempool_bucket.c
b/drivers/mempool/bucket/rte_mempool_bucket.c
index 03fccf1..d1e7c27 100644
--- a/drivers/mempool/bucket/rte_mempool_bucket.c
+++ b/drivers/mempool/bucket/rte_mempool_bucket.c
@@ -42,6 +42,7 @@ struct bucket_data {
unsigned int header_size;
unsigned int total_elt_size;
unsigned int obj_per_bucket;
+ unsigned int bucket_stack_thresh;
uintptr_t bucket_page_mask;
struct rte_ring *shared_bucket_ring;
struct bucket_stack *buckets[RTE_MAX_LCORE];
@@ -139,6 +140,7 @@ bucket_enqueue(struct rte_mempool *mp, void * const
*obj_table,
unsigned int n)
{
struct bucket_data *bd = mp->pool_data;
+ struct bucket_stack *local_stack = bd->buckets[rte_lcore_id()];
unsigned int i;
int rc = 0;
@@ -146,6 +148,15 @@ bucket_enqueue(struct rte_mempool *mp, void * const
*obj_table,
rc = bucket_enqueue_single(bd, obj_table[i]);
RTE_ASSERT(rc == 0);
}
+ if (local_stack->top > bd->bucket_stack_thresh) {
+ rte_ring_enqueue_bulk(bd->shared_bucket_ring,
+ &local_stack->objects
+ [bd->bucket_stack_thresh],
+ local_stack->top -
+ bd->bucket_stack_thresh,
+ NULL);
+ local_stack->top = bd->bucket_stack_thresh;
+ }
return rc;
}
@@ -408,6 +419,8 @@ bucket_alloc(struct rte_mempool *mp)
bd->obj_per_bucket = (bd->bucket_mem_size - bucket_header_size) /
bd->total_elt_size;
bd->bucket_page_mask = ~(rte_align64pow2(bd->bucket_mem_size) - 1);
+ /* eventually this should be a tunable parameter */
+ bd->bucket_stack_thresh = (mp->size / bd->obj_per_bucket) * 4 / 3;
if (mp->flags & MEMPOOL_F_SP_PUT)
rg_flags |= RING_F_SP_ENQ;
--
2.7.4
[dpdk-dev] [RFC v2 15/17] mempool: support block dequeue operation
From: "Artem V. Andreev"
If mempool manager supports object blocks (physically and virtual
contiguous set of objects), it is sufficient to get the first
object only and the function allows to avoid filling in of
information about each block member.
Signed-off-by: Artem V. Andreev
Signed-off-by: Andrew Rybchenko
---
lib/librte_mempool/rte_mempool.h | 125 ++-
lib/librte_mempool/rte_mempool_ops.c | 1 +
2 files changed, 125 insertions(+), 1 deletion(-)
diff --git a/lib/librte_mempool/rte_mempool.h b/lib/librte_mempool/rte_mempool.h
index 75630e6..fa216db 100644
--- a/lib/librte_mempool/rte_mempool.h
+++ b/lib/librte_mempool/rte_mempool.h
@@ -221,7 +221,10 @@ struct rte_mempool_memhdr {
/*
* Additional information about the mempool
*/
-struct rte_mempool_info;
+struct rte_mempool_info {
+ /** Number of objects in the contiguous block */
+ unsigned int contig_block_size;
+};
/**
* The RTE mempool structure.
@@ -399,6 +402,12 @@ typedef int (*rte_mempool_dequeue_t)(struct rte_mempool
*mp,
void **obj_table, unsigned int n);
/**
+ * Dequeue a number of contiquous object blocks from the external pool.
+ */
+typedef int (*rte_mempool_dequeue_contig_blocks_t)(struct rte_mempool *mp,
+void **first_obj_table, unsigned int n);
+
+/**
* Return the number of available objects in the external pool.
*/
typedef unsigned (*rte_mempool_get_count)(const struct rte_mempool *mp);
@@ -518,6 +527,10 @@ struct rte_mempool_ops {
* Get mempool info
*/
rte_mempool_get_info_t get_info;
+ /**
+* Dequeue a number of contiguous object blocks.
+*/
+ rte_mempool_dequeue_contig_blocks_t dequeue_contig_blocks;
} __rte_cache_aligned;
#define RTE_MEMPOOL_MAX_OPS_IDX 16 /**< Max registered ops structs */
@@ -596,6 +609,30 @@ rte_mempool_ops_dequeue_bulk(struct rte_mempool *mp,
}
/**
+ * @internal Wrapper for mempool_ops dequeue_contig_blocks callback.
+ *
+ * @param mp
+ * Pointer to the memory pool.
+ * @param first_obj_table
+ * Pointer to a table of void * pointers (first objects).
+ * @param n
+ * Number of blocks to get.
+ * @return
+ * - 0: Success; got n objects.
+ * - <0: Error; code of dequeue function.
+ */
+static inline int
+rte_mempool_ops_dequeue_contig_blocks(struct rte_mempool *mp,
+ void **first_obj_table, unsigned int n)
+{
+ struct rte_mempool_ops *ops;
+
+ ops = rte_mempool_get_ops(mp->ops_index);
+ RTE_ASSERT(ops->dequeue_contig_blocks != NULL);
+ return ops->dequeue_contig_blocks(mp, first_obj_table, n);
+}
+
+/**
* @internal wrapper for mempool_ops enqueue callback.
*
* @param mp
@@ -1500,6 +1537,92 @@ rte_mempool_get(struct rte_mempool *mp, void **obj_p)
}
/**
+ * @internal Get contiguous blocks of objects from the pool. Used internally.
+ * @param mp
+ * A pointer to the mempool structure.
+ * @param first_obj_table
+ * A pointer to a pointer to the first object in each block.
+ * @param n
+ * A number of blocks to get.
+ * @return
+ * - >0: Success
+ * - <0: Error
+ */
+static __rte_always_inline int
+__mempool_generic_get_contig_blocks(struct rte_mempool *mp,
+ void **first_obj_table, unsigned int n)
+{
+ int ret;
+#ifdef RTE_LIBRTE_MEMPOOL_DEBUG
+ struct rte_mempool_info info;
+ rte_mempool_ops_get_info(mp, &info);
+#endif
+
+ ret = rte_mempool_ops_dequeue_contig_blocks(mp, first_obj_table, n);
+ if (ret < 0)
+ __MEMPOOL_STAT_ADD(mp, get_fail,
+ n * info.contig_block_size);
+ else
+ __MEMPOOL_STAT_ADD(mp, get_success,
+ n * info.contig_block_size);
+
+ return ret;
+}
+
+/**
+ * Get a contiguous blocks of objects from the mempool.
+ *
+ * If cache is enabled, consider to flush it first, to reuse objects
+ * as soon as possible.
+ *
+ * The application should check that the driver supports the operation
+ * by calling rte_mempool_ops_get_info() and checking that `contig_block_size`
+ * is not zero.
+ *
+ * @param mp
+ * A pointer to the mempool structure.
+ * @param first_obj_table
+ * A pointer to a pointer to the first object in each block.
+ * @param n
+ * The number of blocks to get from mempool.
+ * @return
+ * - >0: the size of the block
+ * - -ENOBUFS: Not enough entries in the mempool; no object is retrieved.
+ * - -EOPNOTSUPP: The mempool driver does not support block dequeue
+ */
+static __rte_always_inline int
+rte_mempool_get_contig_blocks(struct rte_mempool *mp,
+ void **first_obj_table, unsigned int n)
+{
+ int ret;
+
+ ret = __mempool_generic_get_contig_blocks(mp, first_obj_table, n);
+#ifdef RTE_LIBRTE_MEMPOOL_DEBUG
+ if (ret == 0) {
+ struct rte_mempool_info info;
+ const size_t total_elt_sz =
+
[dpdk-dev] [RFC v2 12/17] mempool/bucket: implement bucket mempool manager
From: "Artem V. Andreev"
The manager provides a way to allocate physically and virtually
contiguous set of objects.
Signed-off-by: Artem V. Andreev
Signed-off-by: Andrew Rybchenko
---
MAINTAINERS| 9 +
config/common_base | 2 +
drivers/mempool/Makefile | 1 +
drivers/mempool/bucket/Makefile| 27 +
drivers/mempool/bucket/rte_mempool_bucket.c| 561 +
.../mempool/bucket/rte_mempool_bucket_version.map | 4 +
mk/rte.app.mk | 1 +
7 files changed, 605 insertions(+)
create mode 100644 drivers/mempool/bucket/Makefile
create mode 100644 drivers/mempool/bucket/rte_mempool_bucket.c
create mode 100644 drivers/mempool/bucket/rte_mempool_bucket_version.map
diff --git a/MAINTAINERS b/MAINTAINERS
index 5788ea0..9df2cf5 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -303,6 +303,15 @@ F: test/test/test_event_eth_rx_adapter.c
F: doc/guides/prog_guide/event_ethernet_rx_adapter.rst
+Memory Pool Drivers
+---
+
+Bucket memory pool
+M: Artem V. Andreev
+M: Andrew Rybchenko
+F: drivers/mempool/bucket/
+
+
Bus Drivers
---
diff --git a/config/common_base b/config/common_base
index 170a389..4fe42f6 100644
--- a/config/common_base
+++ b/config/common_base
@@ -622,6 +622,8 @@ CONFIG_RTE_LIBRTE_MEMPOOL_DEBUG=n
#
# Compile Mempool drivers
#
+CONFIG_RTE_DRIVER_MEMPOOL_BUCKET=y
+CONFIG_RTE_DRIVER_MEMPOOL_BUCKET_SIZE_KB=64
CONFIG_RTE_DRIVER_MEMPOOL_RING=y
CONFIG_RTE_DRIVER_MEMPOOL_STACK=y
diff --git a/drivers/mempool/Makefile b/drivers/mempool/Makefile
index aae2cb1..45fca04 100644
--- a/drivers/mempool/Makefile
+++ b/drivers/mempool/Makefile
@@ -3,6 +3,7 @@
include $(RTE_SDK)/mk/rte.vars.mk
+DIRS-$(CONFIG_RTE_DRIVER_MEMPOOL_BUCKET) += bucket
DIRS-$(CONFIG_RTE_LIBRTE_DPAA_MEMPOOL) += dpaa
DIRS-$(CONFIG_RTE_LIBRTE_DPAA2_MEMPOOL) += dpaa2
DIRS-$(CONFIG_RTE_DRIVER_MEMPOOL_RING) += ring
diff --git a/drivers/mempool/bucket/Makefile b/drivers/mempool/bucket/Makefile
new file mode 100644
index 000..7364916
--- /dev/null
+++ b/drivers/mempool/bucket/Makefile
@@ -0,0 +1,27 @@
+# SPDX-License-Identifier: BSD-3-Clause
+#
+# Copyright (c) 2017-2018 Solarflare Communications Inc.
+# All rights reserved.
+#
+# This software was jointly developed between OKTET Labs (under contract
+# for Solarflare) and Solarflare Communications, Inc.
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+#
+# library name
+#
+LIB = librte_mempool_bucket.a
+
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS)
+
+LDLIBS += -lrte_eal -lrte_mempool -lrte_ring
+
+EXPORT_MAP := rte_mempool_bucket_version.map
+
+LIBABIVER := 1
+
+SRCS-$(CONFIG_RTE_DRIVER_MEMPOOL_BUCKET) += rte_mempool_bucket.c
+
+include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/mempool/bucket/rte_mempool_bucket.c
b/drivers/mempool/bucket/rte_mempool_bucket.c
new file mode 100644
index 000..dc4e1dc
--- /dev/null
+++ b/drivers/mempool/bucket/rte_mempool_bucket.c
@@ -0,0 +1,561 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ *
+ * Copyright (c) 2017-2018 Solarflare Communications Inc.
+ * All rights reserved.
+ *
+ * This software was jointly developed between OKTET Labs (under contract
+ * for Solarflare) and Solarflare Communications, Inc.
+ */
+
+#include
+#include
+#include
+
+#include
+#include
+#include
+#include
+
+/*
+ * The general idea of the bucket mempool driver is as follows.
+ * We keep track of physically contiguous groups (buckets) of objects
+ * of a certain size. Every such a group has a counter that is
+ * incremented every time an object from that group is enqueued.
+ * Until the bucket is full, no objects from it are eligible for allocation.
+ * If a request is made to dequeue a multiply of bucket size, it is
+ * satisfied by returning the whole buckets, instead of separate objects.
+ */
+
+
+struct bucket_header {
+ unsigned int lcore_id;
+ uint8_t fill_cnt;
+};
+
+struct bucket_stack {
+ unsigned int top;
+ unsigned int limit;
+ void *objects[];
+};
+
+struct bucket_data {
+ unsigned int header_size;
+ unsigned int total_elt_size;
+ unsigned int obj_per_bucket;
+ uintptr_t bucket_page_mask;
+ struct rte_ring *shared_bucket_ring;
+ struct bucket_stack *buckets[RTE_MAX_LCORE];
+ /*
+* Multi-producer single-consumer ring to hold objects that are
+* returned to the mempool at a different lcore than initially
+* dequeued
+*/
+ struct rte_ring *adoption_buffer_rings[RTE_MAX_LCORE];
+ struct rte_ring *shared_orphan_ring;
+ struct rte_mempool *pool;
+ unsigned int bucket_mem_size;
+};
+
+static struct bucket_stack *
+bucket_stack_create(const struct rte_mempool *mp, unsigned int n_elts)
+{
+ struct bucket_stack *stack;
+
+ stack = rte_zmalloc_socket("bucket_stack",
+ sizeof(struct buck
[dpdk-dev] [RFC v2 16/17] mempool/bucket: implement block dequeue operation
From: "Artem V. Andreev"
Signed-off-by: Artem V. Andreev
Signed-off-by: Andrew Rybchenko
---
drivers/mempool/bucket/rte_mempool_bucket.c | 52 +
1 file changed, 52 insertions(+)
diff --git a/drivers/mempool/bucket/rte_mempool_bucket.c
b/drivers/mempool/bucket/rte_mempool_bucket.c
index dc4e1dc..03fccf1 100644
--- a/drivers/mempool/bucket/rte_mempool_bucket.c
+++ b/drivers/mempool/bucket/rte_mempool_bucket.c
@@ -294,6 +294,46 @@ bucket_dequeue(struct rte_mempool *mp, void **obj_table,
unsigned int n)
return rc;
}
+static int
+bucket_dequeue_contig_blocks(struct rte_mempool *mp, void **first_obj_table,
+unsigned int n)
+{
+ struct bucket_data *bd = mp->pool_data;
+ const uint32_t header_size = bd->header_size;
+ struct bucket_stack *cur_stack = bd->buckets[rte_lcore_id()];
+ unsigned int n_buckets_from_stack = RTE_MIN(n, cur_stack->top);
+ struct bucket_header *hdr;
+ void **first_objp = first_obj_table;
+
+ bucket_adopt_orphans(bd);
+
+ n -= n_buckets_from_stack;
+ while (n_buckets_from_stack-- > 0) {
+ hdr = bucket_stack_pop_unsafe(cur_stack);
+ *first_objp++ = (uint8_t *)hdr + header_size;
+ }
+ if (n > 0) {
+ if (unlikely(rte_ring_dequeue_bulk(bd->shared_bucket_ring,
+ first_objp, n, NULL) != n)) {
+ /* Return the already dequeued buckets */
+ while (first_objp-- != first_obj_table) {
+ bucket_stack_push(cur_stack,
+ (uint8_t *)*first_objp -
+ header_size);
+ }
+ rte_errno = ENOBUFS;
+ return -rte_errno;
+ }
+ while (n-- > 0) {
+ hdr = (struct bucket_header *)*first_objp;
+ hdr->lcore_id = rte_lcore_id();
+ *first_objp++ = (uint8_t *)hdr + header_size;
+ }
+ }
+
+ return 0;
+}
+
static void
count_underfilled_buckets(struct rte_mempool *mp,
void *opaque,
@@ -546,6 +586,16 @@ bucket_populate(struct rte_mempool *mp, unsigned int
max_objs,
return n_objs;
}
+static int
+bucket_get_info(const struct rte_mempool *mp, struct rte_mempool_info *info)
+{
+ struct bucket_data *bd = mp->pool_data;
+
+ info->contig_block_size = bd->obj_per_bucket;
+ return 0;
+}
+
+
static const struct rte_mempool_ops ops_bucket = {
.name = "bucket",
.alloc = bucket_alloc,
@@ -555,6 +605,8 @@ static const struct rte_mempool_ops ops_bucket = {
.get_count = bucket_get_count,
.calc_mem_size = bucket_calc_mem_size,
.populate = bucket_populate,
+ .get_info = bucket_get_info,
+ .dequeue_contig_blocks = bucket_dequeue_contig_blocks,
};
--
2.7.4
[dpdk-dev] [PATCH] doc: announce API/ABI changes for mempool
An API/ABI changes are planned for 18.05 [1]: * Allow to customize how mempool objects are stored in memory. * Deprecate mempool XMEM API. * Add mempool driver ops to get information from mempool driver and dequeue contiguous blocks of objects if driver supports it. [1] http://dpdk.org/ml/archives/dev/2018-January/088698.html Signed-off-by: Andrew Rybchenko --- doc/guides/rel_notes/deprecation.rst | 17 + 1 file changed, 17 insertions(+) diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst index d59ad59..9db80da 100644 --- a/doc/guides/rel_notes/deprecation.rst +++ b/doc/guides/rel_notes/deprecation.rst @@ -59,3 +59,20 @@ Deprecation Notices be added between the producer and consumer structures. The size of the structure and the offset of the fields will remain the same on platforms with 64B cache line, but will change on other platforms. + +* mempool: several API and ABI changes are planned in v18.05. + The following functions, introduced for Xen, which is not supported + anymore since v17.11, are hard to use, not used anywhere else in DPDK. + Therefore they will be deprecated in v18.05 and removed in v18.08: + + - ``rte_mempool_xmem_create`` + - ``rte_mempool_xmem_size`` + - ``rte_mempool_xmem_usage`` + + The following changes are planned: + + - removal of ``get_capabilities`` mempool ops and related flags. + - substitute ``register_memory_area`` with ``populate`` ops. + - addition of new ops to customize required memory chunk calculation, +customize objects population and allocate contiguous +block of objects if underlying driver supports it. -- 2.7.4
Re: [dpdk-dev] [PATCH] net/mlx5: remmap UAR address for multiple process
Hi Xueming,
My lonely comments are more related to the commit log which should be
re-written to be more accurate to the issue you try to address, even if
this patch does not solves it.
Please see below,
On Tue, Jan 23, 2018 at 09:50:42AM +, Xueming(Steven) Li wrote:
> Hi Nelio,
>
> > -Original Message-
> > From: Nélio Laranjeiro [mailto:nelio.laranje...@6wind.com]
> > Sent: Monday, January 22, 2018 10:53 PM
> > To: Xueming(Steven) Li
> > Cc: Shahaf Shuler ; dev@dpdk.org
> > Subject: Re: [PATCH] net/mlx5: remmap UAR address for multiple process
> >
> > Hi Xueming,
> >
> > On Fri, Jan 19, 2018 at 11:08:54PM +0800, Xueming Li wrote:
> > > UAR(doorbell) is hw resources that have to be same address between
> > > primary and secondary process, failed to mmap UAR will make TX packets
> > > invisible to HW.
> > > Today, UAR address returned from verbs api is mixed in heap and loaded
> > > library address space, prone to be occupied in secondary process.
> > > This patch reserves a dedicate UAR address space, both primary and
> > > secondary process re-mmap UAR pages into this space.
> > > Below is a brief picture of dpdk app address space allocation:
> > > Before This patch
> > > -- --
> > > [stack] [stack]
> > > [.so, uar, heap][.so, heap]
> > > [(empty)] [(empty)]
> > > [hugepage] [hugepage]
> > > [? others] [? others]
> > > [(empty)] [(empty)]
> > > [uar]
> > > [(empty)]
> > > To minimize conflicts, UAR address space comes after hugepage space
> > > with an offset to skip potential usage from other drivers.
> >
> > Seems it is not the case when the memory is contiguous, according to what
> > I see in my testpmd /proc//maps:
> >
> > PMD: mlx5.c:523: mlx5_uar_init_primary(): Reserved UAR address space:
> > 0x0x7f4da580
> >
> > And the fist huge page is at address 0x7f4fa580, new UAR space is
> > before and not after.
> >
> > With this patch I still have the situation described as "before".
> >
>
> Your observation is correct, system is allocating address in a high-to-low
> manner like stack. UAR address range 0x0x7f4da580 - 0x0x7f4ea580,
> 4GB size, With another 4G offset, hugepage range start is 0x7f4fa580.
>From what I understand, remapping the UAR pages to an address before
the huge pages reduce the situation where the secondaries process cannot
start. This patch does not fix the fact it may fail.
Your small display of the memory mapping between before and after seems
not accurate depending on the OS being run, on Linux v4.14 from debian9
S.I.D. I am still on the situation before no matter how many time I
restart the process. For that I'll suggest you to remove it.
> > > Once UAR space reserved successfully, UAR pages are re-mmapped into
> > > new area to keep UAR address aligned between primary and secondary
> > process.
> > >
> > > Signed-off-by: Xueming Li
> > > ---
> > > drivers/net/mlx5/mlx5.c | 107
> >
> > > drivers/net/mlx5/mlx5.h | 1 +
> > > drivers/net/mlx5/mlx5_defs.h| 10
> > > drivers/net/mlx5/mlx5_rxtx.h| 3 +-
> > > drivers/net/mlx5/mlx5_trigger.c | 7 ++-
> > > drivers/net/mlx5/mlx5_txq.c | 51 +--
> > > 6 files changed, 163 insertions(+), 16 deletions(-)
> > >
> > > diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c index
> > > fc2d59fee..1539ef608 100644
> > > --- a/drivers/net/mlx5/mlx5.c
> > > +++ b/drivers/net/mlx5/mlx5.c
> > > @@ -39,6 +39,7 @@
> > > #include
> > > #include
> > > #include
> > > +#include
> > >
> > > /* Verbs header. */
> > > /* ISO C doesn't support unnamed structs/unions, disabling -pedantic.
> > > */ @@ -56,6 +57,7 @@ #include #include
> > > #include
> > > +#include
> > > #include
> > >
> > > #include "mlx5.h"
> > > @@ -466,6 +468,101 @@ mlx5_args(struct mlx5_dev_config *config, struct
> > > rte_devargs *devargs)
> > >
> > > static struct rte_pci_driver mlx5_driver;
> > >
> > > +/*
> > > + * Reserved UAR address space for TXQ UAR(hw doorbell) mapping,
> > > +process
> > > + * local resource used by both primary and secondary to avoid
> > > +duplicate
> > > + * reservation.
> > > + * The space has to be available on both primary and secondary
> > > +process,
> > > + * TXQ UAR maps to this area using fixed mmap w/o double check.
> > > + */
> > > +static void *uar_base;
> > > +
> > > +/**
> > > + * Reserve UAR address space for primary process
> > > + *
> > > + * @param[in] priv
> > > + * Pointer to private structure.
> > > + *
> > > + * @return
> > > + * 0 on success, negative errno value on failure.
> > > + */
> > > +static int
> > > +mlx5_uar_init_primary(struct priv *priv) {
> > > + void *addr = (void *)0;
> > > + int i;
> > > + const struct rte_mem_config *mcfg;
> > > +
> > > + if (uar_base) { /
[dpdk-dev] [PATCH] doc: replace license text with SPDX tags on SVG images
Signed-off-by: Bruce Richardson --- doc/guides/howto/img/vf_daemon_overview.svg| 36 ++--- .../prog_guide/img/architecture-overview.svg | 35 ++-- doc/guides/prog_guide/img/bond-mode-0.svg | 36 ++--- doc/guides/prog_guide/img/bond-mode-1.svg | 36 ++--- doc/guides/prog_guide/img/bond-mode-2.svg | 36 ++--- doc/guides/prog_guide/img/bond-mode-3.svg | 36 ++--- doc/guides/prog_guide/img/bond-mode-4.svg | 36 ++--- doc/guides/prog_guide/img/bond-mode-5.svg | 36 ++--- doc/guides/prog_guide/img/bond-overview.svg| 36 ++--- doc/guides/prog_guide/img/linuxapp_launch.svg | 35 ++-- doc/guides/prog_guide/img/malloc_heap.svg | 35 ++-- doc/guides/prog_guide/img/mbuf1.svg| 37 ++ doc/guides/prog_guide/img/mbuf2.svg| 37 ++ doc/guides/prog_guide/img/memory-management.svg| 35 ++-- doc/guides/prog_guide/img/memory-management2.svg | 35 ++-- doc/guides/prog_guide/img/mempool.svg | 35 ++-- doc/guides/prog_guide/img/multi_process_memory.svg | 35 ++-- doc/guides/prog_guide/img/ring-dequeue1.svg| 35 ++-- doc/guides/prog_guide/img/ring-dequeue2.svg| 35 ++-- doc/guides/prog_guide/img/ring-dequeue3.svg| 35 ++-- doc/guides/prog_guide/img/ring-enqueue1.svg| 35 ++-- doc/guides/prog_guide/img/ring-enqueue2.svg| 35 ++-- doc/guides/prog_guide/img/ring-enqueue3.svg| 35 ++-- doc/guides/prog_guide/img/ring-modulo1.svg | 36 ++--- doc/guides/prog_guide/img/ring-modulo2.svg | 35 ++-- doc/guides/prog_guide/img/ring-mp-enqueue1.svg | 35 ++-- doc/guides/prog_guide/img/ring-mp-enqueue2.svg | 35 ++-- doc/guides/prog_guide/img/ring-mp-enqueue3.svg | 35 ++-- doc/guides/prog_guide/img/ring-mp-enqueue4.svg | 35 ++-- doc/guides/prog_guide/img/ring-mp-enqueue5.svg | 35 ++-- doc/guides/prog_guide/img/ring1.svg| 35 ++-- doc/guides/sample_app_ug/img/dist_app.svg | 36 ++--- doc/guides/sample_app_ug/img/dist_perf.svg | 36 ++--- .../sample_app_ug/img/exception_path_example.svg | 36 ++--- .../sample_app_ug/img/l2_fwd_benchmark_setup.svg | 35 ++-- .../sample_app_ug/img/vm_power_mgr_highlevel.svg | 36 ++--- .../img/vm_power_mgr_vm_request_seq.svg| 36 ++--- doc/guides/sample_app_ug/img/vmdq_dcb_example.svg | 35 ++-- 38 files changed, 76 insertions(+), 1272 deletions(-) diff --git a/doc/guides/howto/img/vf_daemon_overview.svg b/doc/guides/howto/img/vf_daemon_overview.svg index d4a472340..6a81f2fb8 100644 --- a/doc/guides/howto/img/vf_daemon_overview.svg +++ b/doc/guides/howto/img/vf_daemon_overview.svg @@ -1,38 +1,6 @@ - + + diff --git a/doc/guides/prog_guide/img/architecture-overview.svg b/doc/guides/prog_guide/img/architecture-overview.svg index c0f85bf26..cd8efaae9 100644 --- a/doc/guides/prog_guide/img/architecture-overview.svg +++ b/doc/guides/prog_guide/img/architecture-overview.svg @@ -1,39 +1,8 @@ - + + http://purl.org/dc/elements/1.1/"; diff --git a/doc/guides/prog_guide/img/bond-mode-0.svg b/doc/guides/prog_guide/img/bond-mode-0.svg index e9742c779..850e4d3b6 100644 --- a/doc/guides/prog_guide/img/bond-mode-0.svg +++ b/doc/guides/prog_guide/img/bond-mode-0.svg @@ -1,38 +1,6 @@ - + + - + + - + + - + + - + + - + + - + + http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd";> http://www.w3.org/2000/svg"; xmlns:xlink="http://www.w3.org/1999/xlink"; xmlns:ev="http://www.w3.org/2001/xml-events"; diff --git a/doc/guides/prog_guide/img/linuxapp_launch.svg b/doc/guides/prog_guide/img/linuxapp_launch.svg index 0ac450df8..af685897d 100644 --- a/doc/guides/prog_guide/img/linuxapp_launch.svg +++ b/doc/guides/prog_guide/img/linuxapp_launch.svg @@ -1,39 +1,8 @@ - + + http://purl.org/dc/elements/1.1/"; diff --git a/doc/guides/prog_guide/img/malloc_heap.svg b/doc/guides/prog_guide/img/malloc_heap.svg index d6bcc8489..14e50088c 100644 --- a/doc/guides/prog_guide/img/malloc_heap.svg +++ b/doc/guides/prog_guide/img/malloc_heap.svg @@ -1,39 +1,8 @@ - + + http://purl.org/dc/elements/1.1/"; diff --git a/doc/guides/prog_guide/img/mbuf1.svg b/doc/guides/prog_guide/img/mbuf1.svg index 8750424e3..b80d84216 100644 --- a/doc/guides/prog_guide/img/mbuf1.svg +++ b/doc/guides/prog_guide/img/mbuf1.sv
Re: [dpdk-dev] [PATCH v3 7/7] app/testpmd: adjust ethdev port ownership
Hi Matan,
>
>
> Hi Konstantin
> From: Ananyev, Konstantin, Monday, January 22, 2018 10:49 PM
> > Hi Matan,
> >
> > > -Original Message-
> > > From: Matan Azrad [mailto:ma...@mellanox.com]
> > > Sent: Monday, January 22, 2018 1:23 PM
> > > To: Ananyev, Konstantin ; Gaëtan Rivet
> > >
> > > Cc: Thomas Monjalon ; Wu, Jingjing
> > > ; dev@dpdk.org; Neil Horman
> > > ; Richardson, Bruce
> > >
> > > Subject: RE: [PATCH v3 7/7] app/testpmd: adjust ethdev port ownership
> > >
> > >
> > > Hi
> > > From: Ananyev, Konstantin [mailto:konstantin.anan...@intel.com]
> > > > Hi lads,
> > > >
> > > > >
> > > > > Hi Matan,
> > > > >
> > > > > On Fri, Jan 19, 2018 at 01:35:10PM +, Matan Azrad wrote:
> > > > > > Hi Konstantin
> > > > > >
> > > > > > From: Ananyev, Konstantin, Friday, January 19, 2018 3:09 PM
> > > > > > > > -Original Message-
> > > > > > > > From: Matan Azrad [mailto:ma...@mellanox.com]
> > > > > > > > Sent: Friday, January 19, 2018 12:52 PM
> > > > > > > > To: Ananyev, Konstantin ;
> > > > > > > > Thomas Monjalon ; Gaetan Rivet
> > > > > > > ;
> > > > > > > > Wu, Jingjing
> > > > > > > > Cc: dev@dpdk.org; Neil Horman ;
> > > > > > > > Richardson, Bruce
> > > > > > > > Subject: RE: [PATCH v3 7/7] app/testpmd: adjust ethdev port
> > > > > > > > ownership
> > > > > > > >
> > > > > > > > Hi Konstantin
> > > > > > > >
> > > > > > > > From: Ananyev, Konstantin, Friday, January 19, 2018 2:38 PM
> > > > > > > > > To: Matan Azrad ; Thomas Monjalon
> > > > > > > > > ; Gaetan Rivet
> > > > ;
> > > > > > > Wu,
> > > > > > > > > Jingjing
> > > > > > > > > Cc: dev@dpdk.org; Neil Horman ;
> > > > > > > > > Richardson, Bruce
> > > > > > > > > Subject: RE: [PATCH v3 7/7] app/testpmd: adjust ethdev
> > > > > > > > > port ownership
> > > > > > > > >
> > > > > > > > > Hi Matan,
> > > > > > > > >
> > > > > > > > > > -Original Message-
> > > > > > > > > > From: Matan Azrad [mailto:ma...@mellanox.com]
> > > > > > > > > > Sent: Thursday, January 18, 2018 4:35 PM
> > > > > > > > > > To: Thomas Monjalon ; Gaetan Rivet
> > > > > > > > > > ; Wu, Jingjing
> > > > > > > > > >
> > > > > > > > > > Cc: dev@dpdk.org; Neil Horman ;
> > > > > > > Richardson,
> > > > > > > > > > Bruce ; Ananyev, Konstantin
> > > > > > > > > >
> > > > > > > > > > Subject: [PATCH v3 7/7] app/testpmd: adjust ethdev port
> > > > > > > > > > ownership
> > > > > > > > > >
> > > > > > > > > > Testpmd should not use ethdev ports which are managed by
> > > > > > > > > > other DPDK entities.
> > > > > > > > > >
> > > > > > > > > > Set Testpmd ownership to each port which is not used by
> > > > > > > > > > other entity and prevent any usage of ethdev ports which
> > > > > > > > > > are not owned by
> > > > > > > Testpmd.
> > > > > > > > > >
> > > > > > > > > > Signed-off-by: Matan Azrad
> > > > > > > > > > ---
> > > > > > > > > > app/test-pmd/cmdline.c | 89 +++--
> >
> > > > ---
> > > > > > >
> > > > > > > > > -
> > > > > > > > > > app/test-pmd/cmdline_flow.c | 2 +-
> > > > > > > > > > app/test-pmd/config.c | 37 ++-
> > > > > > > > > > app/test-pmd/parameters.c | 4 +-
> > > > > > > > > > app/test-pmd/testpmd.c | 63
> >
> > > >
> > > > > > > > > > app/test-pmd/testpmd.h | 3 ++
> > > > > > > > > > 6 files changed, 103 insertions(+), 95 deletions(-)
> > > > > > > > > >
> > > > > > > > > > diff --git a/app/test-pmd/cmdline.c
> > > > > > > > > > b/app/test-pmd/cmdline.c index
> > > > > > > > > > 31919ba..6199c64 100644
> > > > > > > > > > --- a/app/test-pmd/cmdline.c
> > > > > > > > > > +++ b/app/test-pmd/cmdline.c
> > > > > > > > > > @@ -1394,7 +1394,7 @@ struct cmd_config_speed_all {
> > > > > > > > > > &link_speed) < 0)
> > > > > > > > > > return;
> > > > > > > > > >
> > > > > > > > > > - RTE_ETH_FOREACH_DEV(pid) {
> > > > > > > > > > + RTE_ETH_FOREACH_DEV_OWNED_BY(pid,
> > my_owner.id) {
> > > > > > > > >
> > > > > > > > > Why do we need all these changes?
> > > > > > > > > As I understand you changed definition of
> > > > > > > > > RTE_ETH_FOREACH_DEV(), so no testpmd should work ok
> > > > > > > > > default
> > > > (no_owner case).
> > > > > > > > > Am I missing something here?
> > > > > > > >
> > > > > > > > Now, After Gaetan suggestion RTE_ETH_FOREACH_DEV(pid) will
> > > > > > > > iterate
> > > > > > > over all valid and ownerless ports.
> > > > > > >
> > > > > > > Yes.
> > > > > > >
> > > > > > > > Here Testpmd wants to iterate over its owned ports.
> > > > > > >
> > > > > > > Why? Why it can't just iterate over all valid and ownerless ports?
> > > > > > > As I understand it would be enough to fix current problems and
> > > > > > > would allow us to avoid any changes in testmpd (which I think
> > > > > > > is a good
> > > > thing).
> > > > > >
> > > > > > Yes, I understand that this big change is very daunted, But I
> > > > > > think the current a lot of bugs in testpmd(reg
[dpdk-dev] questions about new offload ethdev api
Hi, I'm currently porting an application to the new ethdev offload api, and I have two questions about it. 1/ inconsistent offload capa for PMDs using the old API? It is stated in struct rte_eth_txmode: /** * Per-port Tx offloads to be set using DEV_TX_OFFLOAD_* flags. * Only offloads set on tx_offload_capa field on rte_eth_dev_info * structure are allowed to be set. */ uint64_t offloads; So, if I want to enable DEV_TX_OFFLOAD_MULTI_SEGS for the whole ethdev, I must check that DEV_TX_OFFLOAD_MULTI_SEGS is advertised in dev_info->tx_offload_capa. In my understanding, many PMDs are still using the old API, and there is a conversion layer in ethdev when doing dev_configure(). But I don't see any similar mechanism for the dev_info. Therefore, DEV_TX_OFFLOAD_MULTI_SEGS is not present in tx_offload_capa, and if I follow the API comment, I'm not allowed to use this feature. Am I missing something or is it a bug? It looks that testpmd does not check the capa before setting the an offload flag. This could be a workaround in my application. 2/ meaning of rxmode.jumbo_frame, rxmode.enable_scatter, rxmode.max_rx_pkt_len While it's not related to the new API, it is probably a good opportunity to clarify the meaning of these flags. I'm not able to find a good documentation about them. Here is my understanding, the configuration only depends on: - the maximum rx frame length - the amount of data available in a mbuf (minus headroom) Flags to set in rxmode (example): +---+++-+ | |mbuf_data_len=1K|mbuf_data_len=2K|mbuf_data_len=16K| +---+++-+ |max_rx_len=1500|enable_scatter || | +---+++-+ |max_rx_len=9000|enable_scatter, |enable_scatter, |jumbo_frame | | |jumbo_frame |jumbo_frame | | +---+++-+ If this table is correct, the flag jumbo_frame would be equivalent to check if max_rx_pkt_len is above a threshold. And enable_scatter could be deduced from the mbuf size of the given rxq (which is a bit harder but maybe doable). Thanks, Olivier
[dpdk-dev] [PATCH] event/opdl: fix license header and SPDX tags
This patch ensures that the OPDL files all contain correct SPDX tags.
The following changes were made to achieve this:
* replace license text with SPDX tag
* correct occurances where SPDX tag was not on first line of file
* ensure license years were correct
Fixes: 4236ce9bf5bf ("event/opdl: add OPDL ring infrastructure library")
Fixes: 3c7f3dcfb099 ("event/opdl: add PMD main body and helper function")
Fixes: d548ef513cd7 ("event/opdl: add unit tests")
Signed-off-by: Bruce Richardson
---
drivers/event/opdl/Makefile| 2 +-
drivers/event/opdl/opdl_evdev.c| 5 ++---
drivers/event/opdl/opdl_evdev.h| 32 ++--
drivers/event/opdl/opdl_evdev_init.c | 5 ++---
drivers/event/opdl/opdl_evdev_xstats.c | 5 ++---
drivers/event/opdl/opdl_log.h | 5 ++---
drivers/event/opdl/opdl_ring.c | 5 ++---
drivers/event/opdl/opdl_ring.h | 5 ++---
drivers/event/opdl/opdl_test.c | 5 ++---
9 files changed, 17 insertions(+), 52 deletions(-)
diff --git a/drivers/event/opdl/Makefile b/drivers/event/opdl/Makefile
index a8aff2ca6..20216bd97 100644
--- a/drivers/event/opdl/Makefile
+++ b/drivers/event/opdl/Makefile
@@ -1,5 +1,5 @@
# SPDX-License-Identifier: BSD-3-Clause
-# Copyright(c) 2010-2014 Intel Corporation
+# Copyright(c) 2010-2017 Intel Corporation
include $(RTE_SDK)/mk/rte.vars.mk
diff --git a/drivers/event/opdl/opdl_evdev.c b/drivers/event/opdl/opdl_evdev.c
index dcbf404a3..518cb3b3f 100644
--- a/drivers/event/opdl/opdl_evdev.c
+++ b/drivers/event/opdl/opdl_evdev.c
@@ -1,6 +1,5 @@
-/*-
- * SPDX-License-Identifier: BSD-3-Clause
- * Copyright(c) 2010-2014 Intel Corporation
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
*/
#include
diff --git a/drivers/event/opdl/opdl_evdev.h b/drivers/event/opdl/opdl_evdev.h
index 7849af195..610b58b35 100644
--- a/drivers/event/opdl/opdl_evdev.h
+++ b/drivers/event/opdl/opdl_evdev.h
@@ -1,33 +1,5 @@
-/*-
- * BSD LICENSE
- *
- * Copyright(c) 2016-2017 Intel Corporation. All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * * Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in
- * the documentation and/or other materials provided with the
- * distribution.
- * * Neither the name of Intel Corporation nor the names of its
- * contributors may be used to endorse or promote products derived
- * from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
*/
#ifndef _OPDL_EVDEV_H_
diff --git a/drivers/event/opdl/opdl_evdev_init.c
b/drivers/event/opdl/opdl_evdev_init.c
index c37d8bc1b..d7304d67c 100644
--- a/drivers/event/opdl/opdl_evdev_init.c
+++ b/drivers/event/opdl/opdl_evdev_init.c
@@ -1,6 +1,5 @@
-/*-
- * SPDX-License-Identifier: BSD-3-Clause
- * Copyright(c) 2010-2014 Intel Corporation
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
*/
#include
diff --git a/drivers/event/opdl/opdl_evdev_xstats.c
b/drivers/event/opdl/opdl_evdev_xstats.c
index 94dfeee3c..bc0c4dbad 100644
--- a/drivers/event/opdl/opdl_evdev_xstats.c
+++ b/drivers/event/opdl/opdl_evdev_xstats.c
@@ -1,6 +1,5 @@
-/*-
- * SPDX-License-Identifier: BSD-3-Clause
- * Copyright(c) 2010-2014 Intel Corporation
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2017 Intel Corporation
*/
#include "opdl_evdev.h"
diff --git a/drivers/event/opdl/opdl_log.h b/drivers/event/opdl/opdl_log.h
index 0e605aa2d..ae5221c1e 100644
--- a/drivers/event/opdl/opdl_log.h
+++ b/drivers/event/opdl/opdl_log.h
@@ -1,6 +1,5 @@
-/*-
- * SPDX-License-Identifier: BSD-3-Clause
- * Copyright(c) 2010-2014 Intel Corporation
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright
Re: [dpdk-dev] [PATCH v2 2/2] net/ixgbe : backplane port MDIO support
Hi Helin,
Thanks for the review.
Yes , the backplane interfaces(x550em) does not have a phy connected
but it still MDIO lines for control. The requirement for us is to be able
to access
phy registers over backplane MDIO . I 've an updated patchset which is
cleaner.
I 'll send that for review.
Thanks,
Shweta
On Wed, Jan 10, 2018 at 3:17 AM, Zhang, Helin wrote:
>
>
> > -Original Message-
> > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Shweta Choudaha
> > Sent: Monday, November 6, 2017 10:25 PM
> > To: dev@dpdk.org
> > Cc: Shweta Choudaha
> > Subject: [dpdk-dev] [PATCH v2 2/2] net/ixgbe : backplane port MDIO
> support
> >
> > From: Shweta Choudaha
> >
> > Initialize MDIO read/write functions for backplan port
> > (IXGBE_DEV_ID_X550EM_A_KR_L) to enable read/write registers via MDIO
> >
> > Signed-off-by: Shweta Choudaha
> > Reviewed-by: Chas Williams
> > Reviewed-by: Luca Boccassi
> > ---
> > drivers/net/ixgbe/base/ixgbe_x550.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/net/ixgbe/base/ixgbe_x550.c
> > b/drivers/net/ixgbe/base/ixgbe_x550.c
> > index 9862391..3f89dc4 100644
> > --- a/drivers/net/ixgbe/base/ixgbe_x550.c
> > +++ b/drivers/net/ixgbe/base/ixgbe_x550.c
> > @@ -2374,6 +2374,7 @@ s32 ixgbe_init_phy_ops_X550em(struct ixgbe_hw
> > *hw)
> > }
> >
> > switch (hw->device_id) {
> > + case IXGBE_DEV_ID_X550EM_A_KR_L:
> Basically this device ID is for a specific SoC platform, there is no
> external PHY for it.
> We prefer to NACK it. I added more experts here to answer more questions
> if you have.
> Note that they are not working on DPDK, and they are experts on ixgbe
> NIC/SW.
> Or we can discuss more if you have any requests to Intel.
>
> Sorry,
> Helin
>
> > case IXGBE_DEV_ID_X550EM_A_1G_T:
> > case IXGBE_DEV_ID_X550EM_A_1G_T_L:
> > phy->ops.read_reg_mdi = ixgbe_read_phy_reg_mdi_22;
> > --
> > 2.1.4
>
>
Re: [dpdk-dev] [PATCH] net/mlx5: remmap UAR address for multiple process
> -Original Message- > From: Nélio Laranjeiro [mailto:nelio.laranje...@6wind.com] > Sent: Tuesday, January 23, 2018 9:31 PM > To: Xueming(Steven) Li > Cc: Shahaf Shuler ; dev@dpdk.org > Subject: Re: [PATCH] net/mlx5: remmap UAR address for multiple process > > Hi Xueming, > > My lonely comments are more related to the commit log which should be re- > written to be more accurate to the issue you try to address, even if this > patch does not solves it. The current situation is that UAR locates inside top area of address space, VERY easy to conflicts in secondary process, this patch address that with less conflicts. Similar to hugepage mapping in secondary, no good solution to resolve it completely, correct? But it always good to add this warning in commit log. > > Please see below, > > On Tue, Jan 23, 2018 at 09:50:42AM +, Xueming(Steven) Li wrote: > > Hi Nelio, > > > > > -Original Message- > > > From: Nélio Laranjeiro [mailto:nelio.laranje...@6wind.com] > > > Sent: Monday, January 22, 2018 10:53 PM > > > To: Xueming(Steven) Li > > > Cc: Shahaf Shuler ; dev@dpdk.org > > > Subject: Re: [PATCH] net/mlx5: remmap UAR address for multiple > > > process > > > > > > Hi Xueming, > > > > > > On Fri, Jan 19, 2018 at 11:08:54PM +0800, Xueming Li wrote: > > > > UAR(doorbell) is hw resources that have to be same address between > > > > primary and secondary process, failed to mmap UAR will make TX > > > > packets invisible to HW. > > > > Today, UAR address returned from verbs api is mixed in heap and > > > > loaded library address space, prone to be occupied in secondary > process. > > > > This patch reserves a dedicate UAR address space, both primary and > > > > secondary process re-mmap UAR pages into this space. > > > > Below is a brief picture of dpdk app address space allocation: > > > > Before This patch > > > > -- -- > > > > [stack] [stack] > > > > [.so, uar, heap][.so, heap] > > > > [(empty)] [(empty)] > > > > [hugepage] [hugepage] > > > > [? others] [? others] > > > > [(empty)] [(empty)] > > > > [uar] > > > > [(empty)] > > > > To minimize conflicts, UAR address space comes after hugepage > > > > space with an offset to skip potential usage from other drivers. > > > > > > Seems it is not the case when the memory is contiguous, according to > > > what I see in my testpmd /proc//maps: > > > > > > PMD: mlx5.c:523: mlx5_uar_init_primary(): Reserved UAR address space: > > > 0x0x7f4da580 > > > > > > And the fist huge page is at address 0x7f4fa580, new UAR space > > > is before and not after. > > > > > > With this patch I still have the situation described as "before". > > > > > > > Your observation is correct, system is allocating address in a > > high-to-low manner like stack. UAR address range 0x0x7f4da580 - > > 0x0x7f4ea580, 4GB size, With another 4G offset, hugepage range start > is 0x7f4fa580. > > From what I understand, remapping the UAR pages to an address before the > huge pages reduce the situation where the secondaries process cannot start. > This patch does not fix the fact it may fail. > > Your small display of the memory mapping between before and after seems > not accurate depending on the OS being run, on Linux v4.14 from debian9 > S.I.D. I am still on the situation before no matter how many time I > restart the process. For that I'll suggest you to remove it. > > > > > Once UAR space reserved successfully, UAR pages are re-mmapped > > > > into new area to keep UAR address aligned between primary and > > > > secondary > > > process. > > > > > > > > Signed-off-by: Xueming Li > > > > --- > > > > drivers/net/mlx5/mlx5.c | 107 > > > > > > > drivers/net/mlx5/mlx5.h | 1 + > > > > drivers/net/mlx5/mlx5_defs.h| 10 > > > > drivers/net/mlx5/mlx5_rxtx.h| 3 +- > > > > drivers/net/mlx5/mlx5_trigger.c | 7 ++- > > > > drivers/net/mlx5/mlx5_txq.c | 51 +-- > > > > 6 files changed, 163 insertions(+), 16 deletions(-) > > > > > > > > diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c > > > > index > > > > fc2d59fee..1539ef608 100644 > > > > --- a/drivers/net/mlx5/mlx5.c > > > > +++ b/drivers/net/mlx5/mlx5.c > > > > @@ -39,6 +39,7 @@ > > > > #include > > > > #include > > > > #include > > > > +#include > > > > > > > > /* Verbs header. */ > > > > /* ISO C doesn't support unnamed structs/unions, disabling - > pedantic. > > > > */ @@ -56,6 +57,7 @@ #include #include > > > > #include > > > > +#include > > > > #include > > > > > > > > #include "mlx5.h" > > > > @@ -466,6 +468,101 @@ mlx5_args(struct mlx5_dev_config *config, > > > > struct rte_devargs *devargs) > > > > > > > > static struct rte
Re: [dpdk-dev] [PATCH] event/opdl: fix license header and SPDX tags
On 23 Jan 14:11, Bruce Richardson wrote:
> This patch ensures that the OPDL files all contain correct SPDX tags.
> The following changes were made to achieve this:
> * replace license text with SPDX tag
> * correct occurances where SPDX tag was not on first line of file
> * ensure license years were correct
>
> Fixes: 4236ce9bf5bf ("event/opdl: add OPDL ring infrastructure library")
> Fixes: 3c7f3dcfb099 ("event/opdl: add PMD main body and helper function")
> Fixes: d548ef513cd7 ("event/opdl: add unit tests")
>
> Signed-off-by: Bruce Richardson
> ---
> drivers/event/opdl/Makefile| 2 +-
> drivers/event/opdl/opdl_evdev.c| 5 ++---
> drivers/event/opdl/opdl_evdev.h| 32 ++--
> drivers/event/opdl/opdl_evdev_init.c | 5 ++---
> drivers/event/opdl/opdl_evdev_xstats.c | 5 ++---
> drivers/event/opdl/opdl_log.h | 5 ++---
> drivers/event/opdl/opdl_ring.c | 5 ++---
> drivers/event/opdl/opdl_ring.h | 5 ++---
> drivers/event/opdl/opdl_test.c | 5 ++---
> 9 files changed, 17 insertions(+), 52 deletions(-)
>
> diff --git a/drivers/event/opdl/Makefile b/drivers/event/opdl/Makefile
> index a8aff2ca6..20216bd97 100644
> --- a/drivers/event/opdl/Makefile
> +++ b/drivers/event/opdl/Makefile
> @@ -1,5 +1,5 @@
> # SPDX-License-Identifier: BSD-3-Clause
> -# Copyright(c) 2010-2014 Intel Corporation
> +# Copyright(c) 2010-2017 Intel Corporation
>
> include $(RTE_SDK)/mk/rte.vars.mk
>
> diff --git a/drivers/event/opdl/opdl_evdev.c b/drivers/event/opdl/opdl_evdev.c
> index dcbf404a3..518cb3b3f 100644
> --- a/drivers/event/opdl/opdl_evdev.c
> +++ b/drivers/event/opdl/opdl_evdev.c
> @@ -1,6 +1,5 @@
> -/*-
> - * SPDX-License-Identifier: BSD-3-Clause
> - * Copyright(c) 2010-2014 Intel Corporation
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(c) 2017 Intel Corporation
> */
>
> #include
> diff --git a/drivers/event/opdl/opdl_evdev.h b/drivers/event/opdl/opdl_evdev.h
> index 7849af195..610b58b35 100644
> --- a/drivers/event/opdl/opdl_evdev.h
> +++ b/drivers/event/opdl/opdl_evdev.h
> @@ -1,33 +1,5 @@
> -/*-
> - * BSD LICENSE
> - *
> - * Copyright(c) 2016-2017 Intel Corporation. All rights reserved.
> - *
> - * Redistribution and use in source and binary forms, with or without
> - * modification, are permitted provided that the following conditions
> - * are met:
> - *
> - * * Redistributions of source code must retain the above copyright
> - * notice, this list of conditions and the following disclaimer.
> - * * Redistributions in binary form must reproduce the above copyright
> - * notice, this list of conditions and the following disclaimer in
> - * the documentation and/or other materials provided with the
> - * distribution.
> - * * Neither the name of Intel Corporation nor the names of its
> - * contributors may be used to endorse or promote products derived
> - * from this software without specific prior written permission.
> - *
> - * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> - * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> - * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
> - * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> - * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
> - * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
> - * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
> - * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
> - * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
> - * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
> - * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(c) 2017 Intel Corporation
> */
>
> #ifndef _OPDL_EVDEV_H_
> diff --git a/drivers/event/opdl/opdl_evdev_init.c
> b/drivers/event/opdl/opdl_evdev_init.c
> index c37d8bc1b..d7304d67c 100644
> --- a/drivers/event/opdl/opdl_evdev_init.c
> +++ b/drivers/event/opdl/opdl_evdev_init.c
> @@ -1,6 +1,5 @@
> -/*-
> - * SPDX-License-Identifier: BSD-3-Clause
> - * Copyright(c) 2010-2014 Intel Corporation
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(c) 2017 Intel Corporation
> */
>
> #include
> diff --git a/drivers/event/opdl/opdl_evdev_xstats.c
> b/drivers/event/opdl/opdl_evdev_xstats.c
> index 94dfeee3c..bc0c4dbad 100644
> --- a/drivers/event/opdl/opdl_evdev_xstats.c
> +++ b/drivers/event/opdl/opdl_evdev_xstats.c
> @@ -1,6 +1,5 @@
> -/*-
> - * SPDX-License-Identifier: BSD-3-Clause
> - * Copyright(c) 2010-2014 Intel Corporation
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(c) 2017 Intel Corporation
> */
>
> #include "opdl_evdev.h"
> diff --git a/drivers/event/opdl/opdl_log.h b/drivers/event/opdl/opdl_lo
Re: [dpdk-dev] [PATCH] drivers/event: fix resource leak in selftest
> From: Pavan Nikhilesh [mailto:pbhagavat...@caviumnetworks.com]
> Sent: Monday, January 22, 2018 5:46 PM
> To: jerin.ja...@caviumnetworks.com; Van Haaren, Harry
>
> Cc: dev@dpdk.org; Pavan Nikhilesh
> Subject: [dpdk-dev] [PATCH] drivers/event: fix resource leak in selftest
>
> Free resources leak in eventdev selftests.
>
> Coverity issue: 257044
> Coverity issue: 257047
> Coverity issue: 257009
> Fixes: 9ef576176db0 ("test/eventdev: add octeontx multi queue and multi
> port")
> Fixes: 3a17ff401f1e ("test/eventdev: add basic SW tests")
> Fixes: 5e6eb5ccd788 ("event/sw: make test standalone")
>
> Signed-off-by: Pavan Nikhilesh
Acked-by: Harry van Haaren
Re: [dpdk-dev] [PATCH v3 7/7] app/testpmd: adjust ethdev port ownership
23/01/2018 14:34, Ananyev, Konstantin: > If that' s the use case, then I think you need to set device ownership at > creation time - > inside dev_allocate(). > Again that would avoid such racing conditions inside testpmd. The devices must be allocated at a low level layer. When a new device appears (hotplug), an ethdev port should be allocated automatically if it passes the whitelist/blacklist policy test. Then we must decide who will manage this device. I suggest notifying the DPDK libs first. So a DPDK lib or PMD like failsafe can have the priority to take the ownership in its notification callback.
[dpdk-dev] [PATCH] examples/vm_power_manager: fix set VF MAC address
Current code only sets mac address of first VF. Fix code so that it
continues through the loop and sets the mac address of each VF.
Fixes: c9a4779135c9 ("examples/vm_power_mgr: set MAC address of VF")
Signed-off-by: David Coyle
---
examples/vm_power_manager/main.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/examples/vm_power_manager/main.c b/examples/vm_power_manager/main.c
index a50984d..a9f5ad1 100644
--- a/examples/vm_power_manager/main.c
+++ b/examples/vm_power_manager/main.c
@@ -290,7 +290,7 @@ main(int argc, char **argv)
for (portid = 0; portid < nb_ports; portid++) {
struct ether_addr eth;
int w, j;
- int ret = -ENOTSUP;
+ int ret;
if ((enabled_port_mask & (1 << portid)) == 0)
continue;
@@ -308,8 +308,7 @@ main(int argc, char **argv)
for (w = 0; w < MAX_VFS; w++) {
eth.addr_bytes[5] = w + 0xf0;
- if (ret == -ENOTSUP)
- ret = rte_pmd_ixgbe_set_vf_mac_addr(portid,
+ ret = rte_pmd_ixgbe_set_vf_mac_addr(portid,
w, ð);
if (ret == -ENOTSUP)
ret = rte_pmd_i40e_set_vf_mac_addr(portid,
--
2.7.5
--
Intel Research and Development Ireland Limited
Registered in Ireland
Registered Office: Collinstown Industrial Park, Leixlip, County Kildare
Registered Number: 308263
This e-mail and any attachments may contain confidential material for the sole
use of the intended recipient(s). Any review or distribution by others is
strictly prohibited. If you are not the intended recipient, please contact the
sender and delete all copies.
[dpdk-dev] [PATCH] config: replace Intel license headers with SPDX tags
Signed-off-by: Bruce Richardson --- config/common_base| 33 ++- config/common_bsdapp | 33 ++- config/common_linuxapp| 33 ++- config/defconfig_i686-native-linuxapp-gcc | 33 ++- config/defconfig_i686-native-linuxapp-icc | 33 ++- config/defconfig_x86_64-native-bsdapp-clang | 33 ++- config/defconfig_x86_64-native-bsdapp-gcc | 33 ++- config/defconfig_x86_64-native-linuxapp-clang | 33 ++- config/defconfig_x86_64-native-linuxapp-gcc | 33 ++- config/defconfig_x86_64-native-linuxapp-icc | 33 ++- config/defconfig_x86_x32-native-linuxapp-gcc | 33 ++- 11 files changed, 22 insertions(+), 341 deletions(-) diff --git a/config/common_base b/config/common_base index 170a38939..e298bf049 100644 --- a/config/common_base +++ b/config/common_base @@ -1,34 +1,5 @@ -# BSD LICENSE -# -# Copyright(c) 2010-2017 Intel Corporation. All rights reserved. -# All rights reserved. -# -# Redistribution and use in source and binary forms, with or without -# modification, are permitted provided that the following conditions -# are met: -# -# * Redistributions of source code must retain the above copyright -# notice, this list of conditions and the following disclaimer. -# * Redistributions in binary form must reproduce the above copyright -# notice, this list of conditions and the following disclaimer in -# the documentation and/or other materials provided with the -# distribution. -# * Neither the name of Intel Corporation nor the names of its -# contributors may be used to endorse or promote products derived -# from this software without specific prior written permission. -# -# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS -# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR -# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT -# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, -# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT -# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, -# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY -# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT -# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE -# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. -# +# SPDX-License-Identifier: BSD-3-Clause +# Copyright(c) 2010-2017 Intel Corporation # # define executive environment diff --git a/config/common_bsdapp b/config/common_bsdapp index 35a365832..3399246ea 100644 --- a/config/common_bsdapp +++ b/config/common_bsdapp @@ -1,34 +1,5 @@ -# BSD LICENSE -# -# Copyright(c) 2010-2016 Intel Corporation. All rights reserved. -# All rights reserved. -# -# Redistribution and use in source and binary forms, with or without -# modification, are permitted provided that the following conditions -# are met: -# -# * Redistributions of source code must retain the above copyright -# notice, this list of conditions and the following disclaimer. -# * Redistributions in binary form must reproduce the above copyright -# notice, this list of conditions and the following disclaimer in -# the documentation and/or other materials provided with the -# distribution. -# * Neither the name of Intel Corporation nor the names of its -# contributors may be used to endorse or promote products derived -# from this software without specific prior written permission. -# -# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS -# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR -# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT -# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, -# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT -# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, -# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY -# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT -# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE -# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. -# +# SPDX-License-Identifier: BSD-3-Clause +# Copyright(c) 2010-2016 Intel Corporation #include "common_base" diff --git a/config/common_linuxapp b/config/common_linuxapp index 425730d18..ff98f2355 100644 --- a
[dpdk-dev] [PATCH v2 10/10] maintainers: claim ownership of rawdev
Signed-off-by: Shreyansh Jain --- MAINTAINERS | 6 ++ 1 file changed, 6 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index 5788ea004..c3a8769a2 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -302,6 +302,12 @@ F: lib/librte_eventdev/*eth_rx_adapter* F: test/test/test_event_eth_rx_adapter.c F: doc/guides/prog_guide/event_ethernet_rx_adapter.rst +Rawdev API - EXPERIMENTAL +M: Shreyansh Jain +M: Hemant Agrawal +F: lib/librte_rawdev/* +F: test/test/test_rawdev.c +F: drivers/rawdev/skeleton/* Bus Drivers --- -- 2.14.1
[dpdk-dev] [PATCH v2 00/10] Introduce generic 'rawdevice' support
Change History ~~ v2: - restructure comments to prefix model - split patches pivoted on APIs introduced in library - moved test into drivers/rawdev/skeleton from test/test - removed unused RTE_MAX_RAWDEVPORTS and RTE_MAX_RAWDEVS - merge patch configurations - checkpatch fixes Rawdevice Support in DPDK - RFC [1] v1: [2] This patchset introduces rawdevices or generic device support in DPDK. Motivation == In terms of device flavor (type) support, DPDK currently has ethernet (lib_ether), cryptodev (libcryptodev), eventdev (libeventdev) and vdev (virtual device) support. For a new type of device, for example an accelerator, there are not many options except either of: 1. create another lib/librte_MySpecialDev, driver/MySpecialDrv and use it through Bus/PMD model. 2. Or, create a vdev and implement necessary custom APIs which are directly exposed from driver layer. However this may still require changes in bus code in DPDK. Either method is unclean (touching lib for specific context) and possibly non-upstreamable (custom APIs). Applications and customers prefers uniform device view and programming model. Scope = The rawdevice implementation is targetted towards various accelerator use cases which cannot be generalized within existing device models. Aim is to provided a generalized structure at the cost of portability guarantee. Specific PMDs may also expose any specific config APIs. Applications built over such devices are special use-cases involving IP blocks. The rawdevice may also connect to other standard devices using adapter or other methods, similar to eventdev adpter for ethernet/crypto devices. Proposed Solution = Defining a very generic super-set of device type and its device operations that can be exposed such that any new/upcoming/experimental device can be layered over it. 'rawdevice' semantic in this patchset represents a device that doesn't have any flavor/type associated with it which is advertised (like net, crypto etc). A *rte_rawdevice* is a raw/generic device without any standard configuration or input/output method assumption. Thus, driver for a new accelerator block, which requires operations for start/stop/enqueue/dequeue, can be quickly strapped over this rawdevice layer. Thereafter, any appropriate bus can scan for it (assuming device is discoverable over the Linux interfaces like sysfs) and match it against registered drivers. Similarly, for a new accelerator or a wireless device, which doesn't fit the eth type, a driver can be registered with a bus (on which its device would be scannable) and use this layer for configuring the device. It can also serve as a staging area for new type of devices till they find some commonality and can be standardized. The outline of this proposed library is same as existing ether/crypto devices. +---+ |Application(s) | +--.+ | | +--'+ | DPDK Framework (APIs) | +--||-|-+ / \ \ (crypto ops) (eth ops) (rawdev ops)++ / \ \ |DrvA| +-'---++`++---'-+ ++ | crypto || ethdev || raw | +--/--++---/-++/+ ++ /\__/\ / ..|DrvB| / \ /\/ ../\ ++ ++ ++++ +++==/=+ ```Bus Probe |DevA| |DevB||DevC| |DevD||DevF| ++ ++++ ++++ | || | | ``|``||``|`|Bus Scan (PCI) | (PCI) (PCI)(PCI) (BusA) * It is assumed above that DrvB is a PCI type driver which registers itself with PCI Bus * Thereafter, when the PCI scan is done, during probe DrvB would match the rawdev DevF ID and take control of device * Applications can then continue using the device through rawdev API interfaces Proposed Interfaces === Following broad API categories are exposed by the rawdevice: 1) Device State Operations (start/stop/reset) 2) Communication Channel setup/teardown (queue) 3) Attribute get/set operations (~ioctls) 4) Enqueue/Dequeue Operations (using opaque buffers) 5) Firmware Operations (Load/unload) Notes: For (1), other than standard start/stop, reset has been added extra. This is for cases where device power cycle has various definitions. Sema
[dpdk-dev] [PATCH v2 05/10] rawdev: support for firmware management
Some generic operations for firmware management can loading, unloading,
starting, stopping and querying firmware of a device.
This patch adds support for such generic operations.
Signed-off-by: Shreyansh Jain
---
lib/librte_rawdev/rte_rawdev.c | 43
lib/librte_rawdev/rte_rawdev.h | 64 +
lib/librte_rawdev/rte_rawdev_pmd.h | 70
lib/librte_rawdev/rte_rawdev_version.map | 4 ++
4 files changed, 181 insertions(+)
diff --git a/lib/librte_rawdev/rte_rawdev.c b/lib/librte_rawdev/rte_rawdev.c
index 3a7800a45..c1001b34b 100644
--- a/lib/librte_rawdev/rte_rawdev.c
+++ b/lib/librte_rawdev/rte_rawdev.c
@@ -320,6 +320,49 @@ rte_rawdev_xstats_reset(uint16_t dev_id,
return (*dev->dev_ops->xstats_reset)(dev, ids, nb_ids);
}
+int
+rte_rawdev_firmware_status_get(uint16_t dev_id, rte_rawdev_obj_t status_info)
+{
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ struct rte_rawdev *dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->firmware_status_get, -ENOTSUP);
+ return (*dev->dev_ops->firmware_status_get)(dev, status_info);
+}
+
+int
+rte_rawdev_firmware_version_get(uint16_t dev_id, rte_rawdev_obj_t version_info)
+{
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ struct rte_rawdev *dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->firmware_version_get, -ENOTSUP);
+ return (*dev->dev_ops->firmware_version_get)(dev, version_info);
+}
+
+int
+rte_rawdev_firmware_load(uint16_t dev_id, rte_rawdev_obj_t firmware_image)
+{
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ struct rte_rawdev *dev = &rte_rawdevs[dev_id];
+
+ if (!firmware_image)
+ return -EINVAL;
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->firmware_load, -ENOTSUP);
+ return (*dev->dev_ops->firmware_load)(dev, firmware_image);
+}
+
+int
+rte_rawdev_firmware_unload(uint16_t dev_id)
+{
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ struct rte_rawdev *dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->firmware_load, -ENOTSUP);
+ return (*dev->dev_ops->firmware_unload)(dev);
+}
+
int
rte_rawdev_start(uint16_t dev_id)
{
diff --git a/lib/librte_rawdev/rte_rawdev.h b/lib/librte_rawdev/rte_rawdev.h
index 221f85856..388a24cf0 100644
--- a/lib/librte_rawdev/rte_rawdev.h
+++ b/lib/librte_rawdev/rte_rawdev.h
@@ -513,6 +513,70 @@ rte_rawdev_xstats_reset(uint16_t dev_id,
const uint32_t ids[],
uint32_t nb_ids);
+/**
+ * Get Firmware status of the device..
+ * Returns a memory allocated by driver/implementation containing status
+ * information block. It is responsibility of caller to release the buffer.
+ *
+ * @param dev_id
+ * Raw device identifier
+ * @param status_info
+ * Pointer to status information area. Caller is responsible for releasing
+ * the memory associated.
+ * @return
+ * 0 for success,
+ * !0 for failure, `status_info` argument state is undefined
+ */
+int
+rte_rawdev_firmware_status_get(uint16_t dev_id,
+ rte_rawdev_obj_t status_info);
+
+/**
+ * Get Firmware version of the device.
+ * Returns a memory allocated by driver/implementation containing version
+ * information block. It is responsibility of caller to release the buffer.
+ *
+ * @param dev_id
+ * Raw device identifier
+ * @param version_info
+ * Pointer to version information area. Caller is responsible for releasing
+ * the memory associated.
+ * @return
+ * 0 for success,
+ * !0 for failure, `version_info` argument state is undefined
+ */
+int
+rte_rawdev_firmware_version_get(uint16_t dev_id,
+ rte_rawdev_obj_t version_info);
+
+/**
+ * Load firmware on the device.
+ * TODO: In future, methods like directly flashing from file too can be
+ * supported.
+ *
+ * @param dev_id
+ * Raw device identifier
+ * @param firmware_image
+ * Pointer to buffer containing image binary data
+ * @return
+ * 0 for successful load
+ * !0 for failure to load the provided image, or image incorrect.
+ */
+int
+rte_rawdev_firmware_load(uint16_t dev_id, rte_rawdev_obj_t firmware_image);
+
+/**
+ * Unload firmware from the device.
+ *
+ * @param dev_id
+ * Raw device identifiers
+ * @return
+ * 0 for successful Unload
+ * !0 for failure in unloading
+ */
+int
+rte_rawdev_firmware_unload(uint16_t dev_id);
+
#ifdef __cplusplus
}
#endif
diff --git a/lib/librte_rawdev/rte_rawdev_pmd.h
b/lib/librte_rawdev/rte_rawdev_pmd.h
index fd94b39c3..a8da39343 100644
--- a/lib/librte_rawdev/rte_rawdev_pmd.h
+++ b/lib/librte_rawdev/rte_rawdev_pmd.h
@@ -414,6 +414,67 @@ typedef int (*rawdev_xstats_get_names_t)(const struct
rte_rawdev *dev,
typedef uint64_t (*rawdev_xstats_get_by_name_t)(const struct rte_rawdev *dev,
const char *
[dpdk-dev] [PATCH v2 02/10] rawdev: add attribute get and set support
A rawdevice can have various attributes. This patch introduce support
for transparently setting attribute value or getting current attribute
state. This is done by allowing an opaque set of key and value to be
passed through rawdev library.
Signed-off-by: Shreyansh Jain
---
lib/librte_rawdev/rte_rawdev.c | 28 +
lib/librte_rawdev/rte_rawdev.h | 43
lib/librte_rawdev/rte_rawdev_pmd.h | 42 +++
lib/librte_rawdev/rte_rawdev_version.map | 2 ++
4 files changed, 115 insertions(+)
diff --git a/lib/librte_rawdev/rte_rawdev.c b/lib/librte_rawdev/rte_rawdev.c
index 4b89bf07d..427c8a664 100644
--- a/lib/librte_rawdev/rte_rawdev.c
+++ b/lib/librte_rawdev/rte_rawdev.c
@@ -175,6 +175,34 @@ rte_rawdev_queue_release(uint16_t dev_id, uint16_t
queue_id)
return (*dev->dev_ops->queue_release)(dev, queue_id);
}
+int
+rte_rawdev_get_attr(uint16_t dev_id,
+ const char *attr_name,
+ uint64_t *attr_value)
+{
+ struct rte_rawdev *dev;
+
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->attr_get, -ENOTSUP);
+ return (*dev->dev_ops->attr_get)(dev, attr_name, attr_value);
+}
+
+int
+rte_rawdev_set_attr(uint16_t dev_id,
+ const char *attr_name,
+ const uint64_t attr_value)
+{
+ struct rte_rawdev *dev;
+
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->attr_set, -ENOTSUP);
+ return (*dev->dev_ops->attr_set)(dev, attr_name, attr_value);
+}
+
int
rte_rawdev_dump(uint16_t dev_id, FILE *f)
{
diff --git a/lib/librte_rawdev/rte_rawdev.h b/lib/librte_rawdev/rte_rawdev.h
index ef66006f9..44a87c01d 100644
--- a/lib/librte_rawdev/rte_rawdev.h
+++ b/lib/librte_rawdev/rte_rawdev.h
@@ -307,6 +307,49 @@ struct rte_rawdev_buf {
int
rte_rawdev_dump(uint16_t dev_id, FILE *f);
+/**
+ * Get an attribute value from implementation.
+ * Attribute is an opaque handle agreed upon between application and PMD.
+ *
+ * Implementations are expected to maintain an array of attribute-value pairs
+ * based on application calls. Memory management for this structure is
+ * shared responsibility of implementation and application.
+ *
+ * @param dev_id
+ * The identifier of the device to configure.
+ * @attr_name
+ * Opaque object representing an attribute in implementation.
+ * @attr_value [out]
+ * Opaque response to the attribute value. In case of error, this remains
+ * untouched. This is double pointer of void type.
+ * @return
+ * 0 for success
+ * !0 Error; attr_value remains untouched in case of error.
+ */
+int
+rte_rawdev_get_attr(uint16_t dev_id,
+ const char *attr_name,
+ uint64_t *attr_value);
+
+/**
+ * Set an attribute value.
+ * Attribute is an opaque handle agreed upon between application and PMD.
+ *
+ * @param dev_id
+ * The identifier of the device to configure.
+ * @attr_name
+ * Opaque object representing an attribute in implementation.
+ * @attr_value
+ * Value of the attribute represented by attr_name
+ * @return
+ * 0 for success
+ * !0 Error
+ */
+int
+rte_rawdev_set_attr(uint16_t dev_id,
+ const char *attr_name,
+ const uint64_t attr_value);
+
#ifdef __cplusplus
}
#endif
diff --git a/lib/librte_rawdev/rte_rawdev_pmd.h
b/lib/librte_rawdev/rte_rawdev_pmd.h
index a5789cf43..0925d5877 100644
--- a/lib/librte_rawdev/rte_rawdev_pmd.h
+++ b/lib/librte_rawdev/rte_rawdev_pmd.h
@@ -262,6 +262,43 @@ typedef int (*rawdev_queue_release_t)(struct rte_rawdev
*dev,
*/
typedef int (*rawdev_dump_t)(struct rte_rawdev *dev, FILE *f);
+/**
+ * Get an attribute value from implementation.
+ * Attribute is an opaque handle agreed upon between application and PMD.
+ *
+ * @param dev
+ * Raw device pointer
+ * @attr_name
+ * Opaque object representing an attribute in implementation.
+ * @attr_value [out]
+ * Opaque response to the attribute value. In case of error, this remains
+ * untouched. This is double pointer of void type.
+ * @return
+ * 0 for success
+ * !0 Error; attr_value remains untouched in case of error.
+ */
+typedef int (*rawdev_get_attr_t)(struct rte_rawdev *dev,
+const char *attr_name,
+uint64_t *attr_value);
+
+/**
+ * Set an attribute value.
+ * Attribute is an opaque handle agreed upon between application and PMD.
+ *
+ * @param dev
+ * Raw device pointer
+ * @attr_name
+ * Opaque object representing an attribute in implementation.
+ * @attr_value
+ * Value of the attribute represented by attr_name
+ * @return
+ * 0 for success
+ * !0 Error
+ */
+typedef int (*rawdev_set_attr_t)(struct rte_rawdev *dev,
+const char *attr_name
[dpdk-dev] [PATCH v2 04/10] rawdev: support for extended stats
Generic rawdev library cannot define a pre-defined set of stats
for devices which are yet to be defined.
This patch introduces the xstats support for rawdev so that any
implementation can create its own statistics.
Signed-off-by: Shreyansh Jain
---
lib/librte_rawdev/rte_rawdev.c | 75 ++
lib/librte_rawdev/rte_rawdev.h | 105 +++
lib/librte_rawdev/rte_rawdev_pmd.h | 72 +
lib/librte_rawdev/rte_rawdev_version.map | 4 ++
4 files changed, 256 insertions(+)
diff --git a/lib/librte_rawdev/rte_rawdev.c b/lib/librte_rawdev/rte_rawdev.c
index 0c5204f51..3a7800a45 100644
--- a/lib/librte_rawdev/rte_rawdev.c
+++ b/lib/librte_rawdev/rte_rawdev.c
@@ -245,6 +245,81 @@ rte_rawdev_dump(uint16_t dev_id, FILE *f)
return (*dev->dev_ops->dump)(dev, f);
}
+static int
+xstats_get_count(uint16_t dev_id)
+{
+ struct rte_rawdev *dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->xstats_get_names, -ENOTSUP);
+ return (*dev->dev_ops->xstats_get_names)(dev, NULL, 0);
+}
+
+int
+rte_rawdev_xstats_names_get(uint16_t dev_id,
+ struct rte_rawdev_xstats_name *xstats_names,
+ unsigned int size)
+{
+ const struct rte_rawdev *dev;
+ int cnt_expected_entries;
+
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -ENODEV);
+
+ cnt_expected_entries = xstats_get_count(dev_id);
+
+ if (xstats_names == NULL || cnt_expected_entries < 0 ||
+ (int)size < cnt_expected_entries || size <= 0)
+ return cnt_expected_entries;
+
+ dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->xstats_get_names, -ENOTSUP);
+ return (*dev->dev_ops->xstats_get_names)(dev, xstats_names, size);
+}
+
+/* retrieve rawdev extended statistics */
+int
+rte_rawdev_xstats_get(uint16_t dev_id,
+ const unsigned int ids[],
+ uint64_t values[],
+ unsigned int n)
+{
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -ENODEV);
+ const struct rte_rawdev *dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->xstats_get, -ENOTSUP);
+ return (*dev->dev_ops->xstats_get)(dev, ids, values, n);
+}
+
+uint64_t
+rte_rawdev_xstats_by_name_get(uint16_t dev_id,
+ const char *name,
+ unsigned int *id)
+{
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, 0);
+ const struct rte_rawdev *dev = &rte_rawdevs[dev_id];
+ unsigned int temp = -1;
+
+ if (id != NULL)
+ *id = (unsigned int)-1;
+ else
+ id = &temp; /* driver never gets a NULL value */
+
+ /* implemented by driver */
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->xstats_get_by_name, -ENOTSUP);
+ return (*dev->dev_ops->xstats_get_by_name)(dev, name, id);
+}
+
+int
+rte_rawdev_xstats_reset(uint16_t dev_id,
+ const uint32_t ids[], uint32_t nb_ids)
+{
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ struct rte_rawdev *dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->xstats_reset, -ENOTSUP);
+ return (*dev->dev_ops->xstats_reset)(dev, ids, nb_ids);
+}
+
int
rte_rawdev_start(uint16_t dev_id)
{
diff --git a/lib/librte_rawdev/rte_rawdev.h b/lib/librte_rawdev/rte_rawdev.h
index 22f055a40..221f85856 100644
--- a/lib/librte_rawdev/rte_rawdev.h
+++ b/lib/librte_rawdev/rte_rawdev.h
@@ -408,6 +408,111 @@ rte_rawdev_dequeue_buffers(uint16_t dev_id,
unsigned int count,
rte_rawdev_obj_t context);
+/** Maximum name length for extended statistics counters */
+#define RTE_RAW_DEV_XSTATS_NAME_SIZE 64
+
+/**
+ * A name-key lookup element for extended statistics.
+ *
+ * This structure is used to map between names and ID numbers
+ * for extended ethdev statistics.
+ */
+struct rte_rawdev_xstats_name {
+ char name[RTE_RAW_DEV_XSTATS_NAME_SIZE];
+};
+
+/**
+ * Retrieve names of extended statistics of a raw device.
+ *
+ * @param dev_id
+ * The identifier of the raw device.
+ * @param[out] xstats_names
+ * Block of memory to insert names into. Must be at least size in capacity.
+ * If set to NULL, function returns required capacity.
+ * @param size
+ * Capacity of xstats_names (number of names).
+ * @return
+ * - positive value lower or equal to size: success. The return value
+ * is the number of entries filled in the stats table.
+ * - positive value higher than size: error, the given statistics table
+ * is too small. The return value corresponds to the size that should
+ * be given to succeed. The entries in the table are not valid and
+ * shall not be used by the caller.
+ * - negative value on error:
+ *-ENODEV for invalid *dev_id*
+ *-ENOTSUP if the device doesn't support this function.
+ */
+int
+rte_rawdev_xstats_
[dpdk-dev] [PATCH v2 03/10] rawdev: add buffer stream IO support
Introduce handlers for raw buffer enqueue and dequeue. A raw buffer
is essentially a void object which is transparently passed via the
library onto the driver.
Using a context field as argument, any arbitrary meta information
can be passed by application to the driver/implementation. This can
be any data on which driver needs to define the operation semantics.
For example, passing along a queue identifier can suggest the driver
the queue context to perform I/O on.
Signed-off-by: Shreyansh Jain
---
lib/librte_rawdev/rte_rawdev.c | 30 +
lib/librte_rawdev/rte_rawdev.h | 58
lib/librte_rawdev/rte_rawdev_pmd.h | 58
lib/librte_rawdev/rte_rawdev_version.map | 2 ++
4 files changed, 148 insertions(+)
diff --git a/lib/librte_rawdev/rte_rawdev.c b/lib/librte_rawdev/rte_rawdev.c
index 427c8a664..0c5204f51 100644
--- a/lib/librte_rawdev/rte_rawdev.c
+++ b/lib/librte_rawdev/rte_rawdev.c
@@ -203,6 +203,36 @@ rte_rawdev_set_attr(uint16_t dev_id,
return (*dev->dev_ops->attr_set)(dev, attr_name, attr_value);
}
+int
+rte_rawdev_enqueue_buffers(uint16_t dev_id,
+ struct rte_rawdev_buf **buffers,
+ unsigned int count,
+ rte_rawdev_obj_t context)
+{
+ struct rte_rawdev *dev;
+
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->enqueue_bufs, -ENOTSUP);
+ return (*dev->dev_ops->enqueue_bufs)(dev, buffers, count, context);
+}
+
+int
+rte_rawdev_dequeue_buffers(uint16_t dev_id,
+ struct rte_rawdev_buf **buffers,
+ unsigned int count,
+ rte_rawdev_obj_t context)
+{
+ struct rte_rawdev *dev;
+
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->dequeue_bufs, -ENOTSUP);
+ return (*dev->dev_ops->dequeue_bufs)(dev, buffers, count, context);
+}
+
int
rte_rawdev_dump(uint16_t dev_id, FILE *f)
{
diff --git a/lib/librte_rawdev/rte_rawdev.h b/lib/librte_rawdev/rte_rawdev.h
index 44a87c01d..22f055a40 100644
--- a/lib/librte_rawdev/rte_rawdev.h
+++ b/lib/librte_rawdev/rte_rawdev.h
@@ -350,6 +350,64 @@ rte_rawdev_set_attr(uint16_t dev_id,
const char *attr_name,
const uint64_t attr_value);
+/**
+ * Enqueue a stream of buffers to the device.
+ *
+ * Rather than specifying a queue, this API passes along an opaque object
+ * to the driver implementation. That object can be a queue or any other
+ * contextual information necessary for the device to enqueue buffers.
+ *
+ * @param dev_id
+ * The identifier of the device to configure.
+ * @param bufs
+ * Collection of buffers for enqueueing
+ * @param count
+ * Count of buffers to enqueue
+ * @param context
+ * Opaque context information.
+ * @return
+ * >=0 for buffers enqueued
+ * !0 for failure.
+ * Whether partial enqueue is failure or success is defined between app
+ * and driver implementation.
+ */
+int
+rte_rawdev_enqueue_buffers(uint16_t dev_id,
+ struct rte_rawdev_buf **buffers,
+ unsigned int count,
+ rte_rawdev_obj_t context);
+
+/**
+ * Dequeue a stream of buffers from the device.
+ *
+ * Rather than specifying a queue, this API passes along an opaque object
+ * to the driver implementation. That object can be a queue or any other
+ * contextual information necessary for the device to dequeue buffers.
+ *
+ * Application should have allocated enough space to store `count` response
+ * buffers.
+ * Releasing buffers dequeued is responsibility of the application.
+ *
+ * @param dev_id
+ * The identifier of the device to configure.
+ * @param bufs
+ * Collection of buffers dequeued
+ * @param count
+ * Max buffers expected to be dequeued
+ * @param context
+ * Opaque context information.
+ * @return
+ * >=0 for buffers dequeued
+ * !0 for failure.
+ * Whether partial enqueue is failure or success is defined between app
+ * and driver implementation.
+ */
+int
+rte_rawdev_dequeue_buffers(uint16_t dev_id,
+ struct rte_rawdev_buf **buffers,
+ unsigned int count,
+ rte_rawdev_obj_t context);
+
#ifdef __cplusplus
}
#endif
diff --git a/lib/librte_rawdev/rte_rawdev_pmd.h
b/lib/librte_rawdev/rte_rawdev_pmd.h
index 0925d5877..300693f41 100644
--- a/lib/librte_rawdev/rte_rawdev_pmd.h
+++ b/lib/librte_rawdev/rte_rawdev_pmd.h
@@ -248,6 +248,58 @@ typedef int (*rawdev_queue_setup_t)(struct rte_rawdev *dev,
typedef int (*rawdev_queue_release_t)(struct rte_rawdev *dev,
uint16_t queue_id);
+/**
+ * Enqueue an array of raw buffers to the device.
+ *
+ * Buffer being u
[dpdk-dev] [PATCH v2 06/10] rawdev: add self test support
Signed-off-by: Shreyansh Jain
---
lib/librte_rawdev/rte_rawdev.c | 9 +
lib/librte_rawdev/rte_rawdev.h | 12
lib/librte_rawdev/rte_rawdev_pmd.h | 11 +++
lib/librte_rawdev/rte_rawdev_version.map | 1 +
4 files changed, 33 insertions(+)
diff --git a/lib/librte_rawdev/rte_rawdev.c b/lib/librte_rawdev/rte_rawdev.c
index c1001b34b..969244b28 100644
--- a/lib/librte_rawdev/rte_rawdev.c
+++ b/lib/librte_rawdev/rte_rawdev.c
@@ -363,6 +363,15 @@ rte_rawdev_firmware_unload(uint16_t dev_id)
return (*dev->dev_ops->firmware_unload)(dev);
}
+int rte_rawdev_selftest(uint16_t dev_id)
+{
+ RTE_RAWDEV_VALID_DEVID_OR_ERR_RET(dev_id, -EINVAL);
+ struct rte_rawdev *dev = &rte_rawdevs[dev_id];
+
+ RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->dev_selftest, -ENOTSUP);
+ return (*dev->dev_ops->dev_selftest)();
+}
+
int
rte_rawdev_start(uint16_t dev_id)
{
diff --git a/lib/librte_rawdev/rte_rawdev.h b/lib/librte_rawdev/rte_rawdev.h
index 388a24cf0..fd3f7d3d2 100644
--- a/lib/librte_rawdev/rte_rawdev.h
+++ b/lib/librte_rawdev/rte_rawdev.h
@@ -577,6 +577,18 @@ rte_rawdev_firmware_load(uint16_t dev_id, rte_rawdev_obj_t
firmware_image);
int
rte_rawdev_firmware_unload(uint16_t dev_id);
+/**
+ * Trigger the rawdev self test.
+ *
+ * @param dev_id
+ * The identifier of the device
+ * @return
+ * - 0: Selftest successful
+ * - -ENOTSUP if the device doesn't support selftest
+ * - other values < 0 on failure.
+ */
+int rte_rawdev_selftest(uint16_t dev_id);
+
#ifdef __cplusplus
}
#endif
diff --git a/lib/librte_rawdev/rte_rawdev_pmd.h
b/lib/librte_rawdev/rte_rawdev_pmd.h
index a8da39343..dd5c26d57 100644
--- a/lib/librte_rawdev/rte_rawdev_pmd.h
+++ b/lib/librte_rawdev/rte_rawdev_pmd.h
@@ -475,6 +475,14 @@ typedef int (*rawdev_firmware_load_t)(struct rte_rawdev
*dev,
*/
typedef int (*rawdev_firmware_unload_t)(struct rte_rawdev *dev);
+/**
+ * Start rawdev selftest
+ *
+ * @return
+ * Return 0 on success
+ */
+typedef int (*rawdev_selftest_t)(void);
+
/** Rawdevice operations function pointer table */
struct rte_rawdev_ops {
/**< Get device info. */
@@ -528,6 +536,9 @@ struct rte_rawdev_ops {
rawdev_firmware_load_t firmware_load;
/**< Unload firmware */
rawdev_firmware_unload_t firmware_unload;
+
+ /**< Device selftest function */
+ rawdev_selftest_t dev_selftest;
};
/**
diff --git a/lib/librte_rawdev/rte_rawdev_version.map
b/lib/librte_rawdev/rte_rawdev_version.map
index 8de9abdc4..af4465e26 100644
--- a/lib/librte_rawdev/rte_rawdev_version.map
+++ b/lib/librte_rawdev/rte_rawdev_version.map
@@ -19,6 +19,7 @@ EXPERIMENTAL {
rte_rawdev_queue_setup;
rte_rawdev_queue_release;
rte_rawdev_reset;
+ rte_rawdev_selftest;
rte_rawdev_set_attr;
rte_rawdev_socket_id;
rte_rawdev_start;
--
2.14.1
[dpdk-dev] [PATCH v2 01/10] rawdev: introduce raw device library support
Each device in DPDK has a type associated with it - ethernet, crypto,
event etc. This patch introduces 'rawdevice' which is a generic
type of device, not currently handled out-of-the-box by DPDK.
A device which can be scanned on an installed bus (pci, fslmc, ...)
or instantiated through devargs, can be interfaced using
standardized APIs just like other standardized devices.
This library introduces an API set which can be plugged on the
northbound side to the application layer, and on the southbound side
to the driver layer.
The APIs of rawdev library exposes some generic operations which can
enable configuration and I/O with the raw devices. Using opaque
data (pointer) as API arguments, library allows a high flexibility
for application and driver implementation.
This patch introduces basic device operations like start, stop, reset,
queue and info support.
Subsequent patches would introduce other operations like buffer
enqueue/dequeue and firmware support.
Signed-off-by: Shreyansh Jain
---
config/common_base | 7 +
lib/Makefile | 3 +
lib/librte_rawdev/Makefile | 27 +++
lib/librte_rawdev/rte_rawdev.c | 360 +++
lib/librte_rawdev/rte_rawdev.h | 314 +++
lib/librte_rawdev/rte_rawdev_pmd.h | 352 ++
lib/librte_rawdev/rte_rawdev_version.map | 21 ++
mk/rte.app.mk| 1 +
8 files changed, 1085 insertions(+)
create mode 100644 lib/librte_rawdev/Makefile
create mode 100644 lib/librte_rawdev/rte_rawdev.c
create mode 100644 lib/librte_rawdev/rte_rawdev.h
create mode 100644 lib/librte_rawdev/rte_rawdev_pmd.h
create mode 100644 lib/librte_rawdev/rte_rawdev_version.map
diff --git a/config/common_base b/config/common_base
index 170a38939..b8bd25b15 100644
--- a/config/common_base
+++ b/config/common_base
@@ -805,6 +805,13 @@ CONFIG_RTE_LIBRTE_VHOST=n
CONFIG_RTE_LIBRTE_VHOST_NUMA=n
CONFIG_RTE_LIBRTE_VHOST_DEBUG=n
+#
+# Compile raw device support
+# EXPERIMENTAL: API may change without prior notice
+#
+CONFIG_RTE_LIBRTE_RAWDEV=y
+CONFIG_RTE_RAWDEV_MAX_DEVS=10
+
#
# Compile vhost PMD
# To compile, CONFIG_RTE_LIBRTE_VHOST should be enabled.
diff --git a/lib/Makefile b/lib/Makefile
index 427f34b00..97080a590 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -102,4 +102,7 @@ endif
DEPDIRS-librte_kni := librte_eal librte_mempool librte_mbuf librte_ether
DEPDIRS-librte_kni += librte_pci
+DIRS-$(CONFIG_RTE_LIBRTE_RAWDEV) += librte_rawdev
+DEPDIRS-librte_rawdev := librte_eal librte_ether
+
include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/lib/librte_rawdev/Makefile b/lib/librte_rawdev/Makefile
new file mode 100644
index 0..addb288d7
--- /dev/null
+++ b/lib/librte_rawdev/Makefile
@@ -0,0 +1,27 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright 2017 NXP
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+# library name
+LIB = librte_rawdev.a
+
+# library version
+LIBABIVER := 1
+
+# build flags
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS)
+LDLIBS += -lrte_eal
+
+# library source files
+SRCS-y += rte_rawdev.c
+
+# export include files
+SYMLINK-y-include += rte_rawdev.h
+SYMLINK-y-include += rte_rawdev_pmd.h
+
+# versioning export map
+EXPORT_MAP := rte_rawdev_version.map
+
+include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/lib/librte_rawdev/rte_rawdev.c b/lib/librte_rawdev/rte_rawdev.c
new file mode 100644
index 0..4b89bf07d
--- /dev/null
+++ b/lib/librte_rawdev/rte_rawdev.c
@@ -0,0 +1,360 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2017 NXP
+ */
+
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+
+#include "rte_rawdev.h"
+#include "rte_rawdev_pmd.h"
+
+/* dynamic log identifier */
+int librawdev_logtype;
+
+struct rte_rawdev rte_rawdevices[RTE_RAWDEV_MAX_DEVS];
+
+struct rte_rawdev *rte_rawdevs = &rte_rawdevices[0];
+
+static struct rte_rawdev_global rawdev_globals = {
+ .nb_devs= 0
+};
+
+struct rte_rawdev_global *rte_rawdev_globals = &rawdev_globals;
+
+/* Raw device, northbound API implementation */
+uint8_t
+rte_rawdev_count(void)
+{
+ return rte_rawdev_globals->nb_devs;
+}
+
+uint16_t
+rte_rawdev_get_dev_id(const char *name)
+{
+ uint16_t i;
+
+ if (!name)
+ return -EINVAL;
+
+ for (i = 0; i < rte_rawdev_globals->nb_devs; i++)
+ if ((strcmp(rte_rawdevices[i].name, name)
+ == 0) &&
+ (rte_rawdevices[i].attached ==
+ RTE_RAWDEV_ATTACHED))
+ return i;
+ return -ENODEV;
+}
+
+int
+rte_rawdev_socket_id(uint16_t dev_id)
+{
+ struct rte_rawdev *dev;
+
+
[dpdk-dev] [PATCH v2 08/10] drivers/raw: support for rawdev testcases
Patch introduces rawdev unit testcase for validation against the
Skeleton rawdev dummy PMD implementation.
Test cases are added along with the skeleton driver implementation.
It can be enabled by using vdev argument to any DPDK binary:
--vdev="rawdev_skeleton,self_test=1"
In case 'self_test=1' is not provided, autotest doesn't execute the
test cases but the vdev is still available for application use.
Signed-off-by: Shreyansh Jain
---
drivers/raw/skeleton_rawdev/Makefile | 1 +
drivers/raw/skeleton_rawdev/skeleton_rawdev.c | 66 +++-
drivers/raw/skeleton_rawdev/skeleton_rawdev_test.c | 431 +
3 files changed, 497 insertions(+), 1 deletion(-)
create mode 100644 drivers/raw/skeleton_rawdev/skeleton_rawdev_test.c
diff --git a/drivers/raw/skeleton_rawdev/Makefile
b/drivers/raw/skeleton_rawdev/Makefile
index 4d9b2f804..d5e34361a 100644
--- a/drivers/raw/skeleton_rawdev/Makefile
+++ b/drivers/raw/skeleton_rawdev/Makefile
@@ -21,5 +21,6 @@ LIBABIVER := 1
# all source are stored in SRCS-y
#
SRCS-$(CONFIG_RTE_LIBRTE_PMD_SKELETON_RAWDEV) += skeleton_rawdev.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_SKELETON_RAWDEV) += skeleton_rawdev_test.c
include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/raw/skeleton_rawdev/skeleton_rawdev.c
b/drivers/raw/skeleton_rawdev/skeleton_rawdev.c
index b38188e11..2dfba0081 100644
--- a/drivers/raw/skeleton_rawdev/skeleton_rawdev.c
+++ b/drivers/raw/skeleton_rawdev/skeleton_rawdev.c
@@ -541,6 +541,8 @@ static const struct rte_rawdev_ops skeleton_rawdev_ops = {
.firmware_version_get = skeleton_rawdev_firmware_version_get,
.firmware_load = skeleton_rawdev_firmware_load,
.firmware_unload = skeleton_rawdev_firmware_unload,
+
+ .dev_selftest = test_rawdev_skeldev,
};
static int
@@ -630,11 +632,62 @@ skeleton_rawdev_destroy(const char *name)
return 0;
}
+static int
+skeldev_get_selftest(const char *key __rte_unused,
+const char *value,
+void *opaque)
+{
+ int *flag = opaque;
+ *flag = atoi(value);
+ return 0;
+}
+
+static int
+skeldev_parse_vdev_args(struct rte_vdev_device *vdev)
+{
+ int selftest = 0;
+ const char *name;
+ const char *params;
+
+ static const char *const args[] = {
+ SKELETON_SELFTEST_ARG,
+ NULL
+ };
+
+ name = rte_vdev_device_name(vdev);
+
+ params = rte_vdev_device_args(vdev);
+ if (params != NULL && params[0] != '\0') {
+ struct rte_kvargs *kvlist = rte_kvargs_parse(params, args);
+
+ if (!kvlist) {
+ SKELETON_PMD_INFO(
+ "Ignoring unsupported params supplied '%s'",
+ name);
+ } else {
+ int ret = rte_kvargs_process(kvlist,
+ SKELETON_SELFTEST_ARG,
+ skeldev_get_selftest, &selftest);
+ if (ret != 0 || (selftest < 0 || selftest > 1)) {
+ SKELETON_PMD_ERR("%s: Error in parsing args",
+name);
+ rte_kvargs_free(kvlist);
+ ret = -1; /* enforce if selftest is invalid */
+ return ret;
+ }
+ }
+
+ rte_kvargs_free(kvlist);
+ }
+
+ return selftest;
+}
+
static int
skeleton_rawdev_probe(struct rte_vdev_device *vdev)
{
const char *name;
- int ret = 0;
+ int selftest = 0, ret = 0;
name = rte_vdev_device_name(vdev);
@@ -647,7 +700,18 @@ skeleton_rawdev_probe(struct rte_vdev_device *vdev)
SKELETON_PMD_INFO("Init %s on NUMA node %d", name, rte_socket_id());
+ selftest = skeldev_parse_vdev_args(vdev);
+ /* In case of invalid argument, selftest != 1; ignore other values */
+
ret = skeleton_rawdev_create(name, vdev, rte_socket_id());
+ if (!ret) {
+ /* In case command line argument for 'selftest' was passed;
+* if invalid arguments were passed, execution continues but
+* without selftest.
+*/
+ if (selftest == 1)
+ test_rawdev_skeldev();
+ }
/* Device instance created; Second instance not posible */
skeldev_init_once = 1;
diff --git a/drivers/raw/skeleton_rawdev/skeleton_rawdev_test.c
b/drivers/raw/skeleton_rawdev/skeleton_rawdev_test.c
new file mode 100644
index 0..5c6abc13c
--- /dev/null
+++ b/drivers/raw/skeleton_rawdev/skeleton_rawdev_test.c
@@ -0,0 +1,431 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2017 NXP
+ */
+
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+
+/* Using relative path as skeleton_rawdev is not part of exported headers */
+
[dpdk-dev] [PATCH v2 07/10] drivers/raw: introduce skeleton rawdev driver
Skeleton rawdevice driver, on the lines of eventdev skeleton, is for
showcasing the rawdev library. This driver implements some of the
operations of the library based on which a test module can be
developed.
Design of skeleton involves a virtual device which is plugged into
VDEV bus on initialization.
Also, enable compilation of rawdev skeleton driver.
Signed-off-by: Shreyansh Jain
---
config/common_base | 1 +
drivers/Makefile | 2 +
drivers/raw/Makefile | 9 +
drivers/raw/skeleton_rawdev/Makefile | 25 +
.../rte_pmd_skeleton_rawdev_version.map| 4 +
drivers/raw/skeleton_rawdev/skeleton_rawdev.c | 690 +
drivers/raw/skeleton_rawdev/skeleton_rawdev.h | 136
mk/rte.app.mk | 4 +
8 files changed, 871 insertions(+)
create mode 100644 drivers/raw/Makefile
create mode 100644 drivers/raw/skeleton_rawdev/Makefile
create mode 100644
drivers/raw/skeleton_rawdev/rte_pmd_skeleton_rawdev_version.map
create mode 100644 drivers/raw/skeleton_rawdev/skeleton_rawdev.c
create mode 100644 drivers/raw/skeleton_rawdev/skeleton_rawdev.h
diff --git a/config/common_base b/config/common_base
index b8bd25b15..b0f6042d3 100644
--- a/config/common_base
+++ b/config/common_base
@@ -811,6 +811,7 @@ CONFIG_RTE_LIBRTE_VHOST_DEBUG=n
#
CONFIG_RTE_LIBRTE_RAWDEV=y
CONFIG_RTE_RAWDEV_MAX_DEVS=10
+CONFIG_RTE_LIBRTE_PMD_SKELETON_RAWDEV=y
#
# Compile vhost PMD
diff --git a/drivers/Makefile b/drivers/Makefile
index e0488aa2b..ee65c87b0 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -14,5 +14,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += crypto
DEPDIRS-crypto := bus mempool
DIRS-$(CONFIG_RTE_LIBRTE_EVENTDEV) += event
DEPDIRS-event := bus mempool net
+DIRS-$(CONFIG_RTE_LIBRTE_RAWDEV) += raw
+DEPDIRS-raw := bus mempool net event
include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/drivers/raw/Makefile b/drivers/raw/Makefile
new file mode 100644
index 0..da7c8b449
--- /dev/null
+++ b/drivers/raw/Makefile
@@ -0,0 +1,9 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright 2017 NXP
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+# DIRS-$() +=
+DIRS-$(CONFIG_RTE_LIBRTE_PMD_SKELETON_RAWDEV) += skeleton_rawdev
+
+include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/drivers/raw/skeleton_rawdev/Makefile
b/drivers/raw/skeleton_rawdev/Makefile
new file mode 100644
index 0..4d9b2f804
--- /dev/null
+++ b/drivers/raw/skeleton_rawdev/Makefile
@@ -0,0 +1,25 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright 2017 NXP
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+#
+# library name
+#
+LIB = librte_pmd_skeleton_rawdev.a
+
+CFLAGS += $(WERROR_FLAGS)
+LDLIBS += -lrte_eal
+LDLIBS += -lrte_rawdev
+LDLIBS += -lrte_bus_vdev
+
+EXPORT_MAP := rte_pmd_skeleton_rawdev_version.map
+
+LIBABIVER := 1
+
+#
+# all source are stored in SRCS-y
+#
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_SKELETON_RAWDEV) += skeleton_rawdev.c
+
+include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/raw/skeleton_rawdev/rte_pmd_skeleton_rawdev_version.map
b/drivers/raw/skeleton_rawdev/rte_pmd_skeleton_rawdev_version.map
new file mode 100644
index 0..179140fb8
--- /dev/null
+++ b/drivers/raw/skeleton_rawdev/rte_pmd_skeleton_rawdev_version.map
@@ -0,0 +1,4 @@
+DPDK_18.02 {
+
+ local: *;
+};
diff --git a/drivers/raw/skeleton_rawdev/skeleton_rawdev.c
b/drivers/raw/skeleton_rawdev/skeleton_rawdev.c
new file mode 100644
index 0..b38188e11
--- /dev/null
+++ b/drivers/raw/skeleton_rawdev/skeleton_rawdev.c
@@ -0,0 +1,690 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2017 NXP
+ */
+
+#include
+#include
+#include
+#include
+#include
+#include
+
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+
+#include
+#include
+
+#include "skeleton_rawdev.h"
+
+/* Dynamic log type identifier */
+int skeleton_pmd_logtype;
+
+/* Count of instances */
+uint16_t skeldev_init_once;
+
+/**< Rawdev Skeleton dummy driver name */
+#define SKELETON_PMD_RAWDEV_NAME rawdev_skeleton
+
+/**< Skeleton rawdev driver object */
+static struct rte_vdev_driver skeleton_pmd_drv;
+
+struct queue_buffers {
+ void *bufs[SKELETON_QUEUE_MAX_DEPTH];
+};
+
+static struct queue_buffers queue_buf[SKELETON_MAX_QUEUES] = {0};
+static void clear_queue_bufs(int queue_id);
+
+static void skeleton_rawdev_info_get(struct rte_rawdev *dev,
+rte_rawdev_obj_t dev_info)
+{
+ struct skeleton_rawdev *skeldev;
+ struct skeleton_rawdev_conf *skeldev_conf;
+
+ SKELETON_PMD_FUNC_TRACE();
+
+ if (!dev_info) {
+ SKELETON_PMD_ERR("Invalid request");
+ return;
+ }
+
+ skeldev = skeleton_rawdev_get_priv(dev);
+
+ skeldev_conf = dev_info;
+
+ skeldev_conf->num_queues = skeldev->num_queues;
+ skeldev_conf->cap
[dpdk-dev] [PATCH v2 09/10] test: enable rawdev skeleton test
Skeleton rawdevice test cases are part of driver layer. This patch
allows test cases to be executed using 'rawdev_autotest' command
in test framework.
Signed-off-by: Shreyansh Jain
---
test/test/Makefile | 4
test/test/test_rawdev.c | 27 +++
2 files changed, 31 insertions(+)
create mode 100644 test/test/test_rawdev.c
diff --git a/test/test/Makefile b/test/test/Makefile
index 5ba5a9ac7..8dfb122bb 100644
--- a/test/test/Makefile
+++ b/test/test/Makefile
@@ -184,6 +184,10 @@ SRCS-y += test_event_ring.c
SRCS-y += test_event_eth_rx_adapter.c
endif
+ifeq ($(CONFIG_RTE_LIBRTE_RAWDEV),y)
+SRCS-y += test_rawdev.c
+endif
+
SRCS-$(CONFIG_RTE_LIBRTE_KVARGS) += test_kvargs.c
CFLAGS += -O3
diff --git a/test/test/test_rawdev.c b/test/test/test_rawdev.c
new file mode 100644
index 0..043a38a13
--- /dev/null
+++ b/test/test/test_rawdev.c
@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2017 NXP
+ */
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+
+#include "test.h"
+
+static int
+test_rawdev_selftest_impl(const char *pmd, const char *opts)
+{
+ rte_vdev_init(pmd, opts);
+ return rte_rawdev_selftest(rte_rawdev_get_dev_id(pmd));
+}
+
+static int
+test_rawdev_selftest_skeleton(void)
+{
+ return test_rawdev_selftest_impl("rawdev_skeleton", "");
+}
+
+REGISTER_TEST_COMMAND(rawdev_autotest, test_rawdev_selftest_skeleton);
--
2.14.1
[dpdk-dev] [PATCH v2] app/test-crypto-perf: fix index
Fixes: 27c2e7471961 ("app/crypto-perf: support IMIX")
A bug caused index out-of-range error. This simple patch is to fix
this.
Signed-off-by: Fan Zhang
---
v2:
Removed unnecessary cdev_id assignment line.
app/test-crypto-perf/main.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/app/test-crypto-perf/main.c b/app/test-crypto-perf/main.c
index 83a9d7b73..dd16fbe24 100644
--- a/app/test-crypto-perf/main.c
+++ b/app/test-crypto-perf/main.c
@@ -582,10 +582,8 @@ main(int argc, char **argv)
if (i == nb_cryptodevs)
break;
- cdev_id = enabled_cdevs[i];
-
rte_eal_remote_launch(cperf_testmap[opts.test].runner,
- ctx[cdev_id], lcore_id);
+ ctx[i], lcore_id);
i++;
}
i = 0;
--
2.13.6
