* [dpdk-dev] [PATCH v1 1/1] net/octeontx2: add SDP interface support
@ 2019-10-30 14:08 Subrahmanyam Nilla
2019-11-06 6:45 ` Jerin Jacob
0 siblings, 1 reply; 2+ messages in thread
From: Subrahmanyam Nilla @ 2019-10-30 14:08 UTC (permalink / raw)
To: dev; +Cc: jerinj, shallyv, venkatn, mchalla
System DPI Packet Interface Unit (SDP) is a co-processor
of OCTEON TX2 which provides PCIe endpoint support for a
remote host to DMA packets into and out of the OCTEON TX2 SoC.
SDP interface comes in to live only when it is connected in
EP mode. It exposes input and output queue pairs to remote host
for instruction input and packet output. It can be used as
a communication channel between remote host and OCTEON TX2.
Host machine needs to use corresponding user/kernel mode
driver to communicate with SDP interface on OCTEON TX2 SoC.
SDP interface support is limited to SDP PF device now.
No SDP VF support.
Signed-off-by: Subrahmanyam Nilla <snilla@marvell.com>
Signed-off-by: Venkateshwarlu Nalla <venkatn@marvell.com>
---
doc/guides/nics/octeontx2.rst | 4 ++++
doc/guides/platform/octeontx2.rst | 21 +++++++++++++++++++++
drivers/common/octeontx2/otx2_common.h | 2 ++
drivers/common/octeontx2/otx2_dev.c | 1 +
drivers/common/octeontx2/otx2_dev.h | 4 ++++
drivers/net/octeontx2/otx2_ethdev.c | 31 ++++++++++++++++++++++++++-----
drivers/net/octeontx2/otx2_ethdev.h | 1 +
drivers/net/octeontx2/otx2_ethdev_ops.c | 6 +++++-
drivers/net/octeontx2/otx2_flow_ctrl.c | 6 ++++++
drivers/net/octeontx2/otx2_link.c | 6 +++---
drivers/net/octeontx2/otx2_mac.c | 8 ++++----
drivers/net/octeontx2/otx2_ptp.c | 6 +++---
drivers/net/octeontx2/otx2_tm.c | 18 ++++++++++++------
13 files changed, 92 insertions(+), 22 deletions(-)
diff --git a/doc/guides/nics/octeontx2.rst b/doc/guides/nics/octeontx2.rst
index fc8a130fb..7e901cec1 100644
--- a/doc/guides/nics/octeontx2.rst
+++ b/doc/guides/nics/octeontx2.rst
@@ -200,6 +200,10 @@ Multicast MAC filtering
``net_octeontx2`` pmd supports multicast mac filtering feature only on physical
function devices.
+SDP interface support
+~~~~~~~~~~~~~~~~~~~~~
+OCTEON TX2 SDP interface support is limited to PF device, No VF support.
+
Debugging Options
-----------------
diff --git a/doc/guides/platform/octeontx2.rst b/doc/guides/platform/octeontx2.rst
index 1c7d98f73..15b1641cf 100644
--- a/doc/guides/platform/octeontx2.rst
+++ b/doc/guides/platform/octeontx2.rst
@@ -64,6 +64,8 @@ DPDK subsystem.
+---+-----+--------------------------------------------------------------+
| 8 | DPI | rte_rawdev |
+---+-----+--------------------------------------------------------------+
+ | 9 | SDP | rte_ethdev |
+ +---+-----+--------------------------------------------------------------+
PF0 is called the administrative / admin function (AF) and has exclusive
privileges to provision RVU functional block's LFs to each of the PF/VF.
@@ -102,6 +104,25 @@ Typical application usage models are,
#. Exception path to Linux kernel from DPDK application as SW ``KNI`` replacement.
#. Communication between two different DPDK applications.
+SDP interface
+-------------
+
+System DPI Packet Interface unit(SDP) provides PCIe endpoint support for remote host
+to DMA packets into and out of OCTEON TX2 SoC. SDP interface comes in to live only when
+OCTEON TX2 SoC is connected in PCIe endpoint mode. It can be used to send/receive
+packets to/from remote host machine using input/output queue pairs exposed to it.
+SDP interface receives input packets from remote host from NIX-RX and sends packets
+to remote host using NIX-TX. Remote host machine need to use corresponding driver
+(kernel/user mode) to communicate with SDP interface on OCTEON TX2 SoC. SDP supports
+single PCIe SRIOV physical function(PF) and multiple virtual functions(VF's). Users
+can bind PF or VF to use SDP interface and it will be enumerated as ethdev ports.
+
+The primary use case for SDP is to enable the smart NIC use case. Typical usage models are,
+
+#. Communication channel between remote host and OCTEON TX2 SoC over PCIe.
+#. Transfer packets received from network interface to remote host over PCIe and
+ vice-versa.
+
OCTEON TX2 packet flow
----------------------
diff --git a/drivers/common/octeontx2/otx2_common.h b/drivers/common/octeontx2/otx2_common.h
index 0d0b87ee6..f62c45deb 100644
--- a/drivers/common/octeontx2/otx2_common.h
+++ b/drivers/common/octeontx2/otx2_common.h
@@ -121,6 +121,8 @@ extern int otx2_logtype_dpi;
#define PCI_DEVID_OCTEONTX2_RVU_CPT_VF 0xA0FE
#define PCI_DEVID_OCTEONTX2_RVU_AF_VF 0xA0f8
#define PCI_DEVID_OCTEONTX2_DPI_VF 0xA081
+#define PCI_DEVID_OCTEONTX2_RVU_SDP_PF 0xA0f6
+#define PCI_DEVID_OCTEONTX2_RVU_SDP_VF 0xA0f7
/* Subsystem Device ID */
#define PCI_SUBSYS_DEVID_96XX_95XX 0xB200
diff --git a/drivers/common/octeontx2/otx2_dev.c b/drivers/common/octeontx2/otx2_dev.c
index 910270975..0fc799e4a 100644
--- a/drivers/common/octeontx2/otx2_dev.c
+++ b/drivers/common/octeontx2/otx2_dev.c
@@ -885,6 +885,7 @@ otx2_update_vf_hwcap(struct rte_pci_device *pci_dev, struct otx2_dev *dev)
case PCI_DEVID_OCTEONTX2_RVU_CPT_VF:
case PCI_DEVID_OCTEONTX2_RVU_AF_VF:
case PCI_DEVID_OCTEONTX2_RVU_VF:
+ case PCI_DEVID_OCTEONTX2_RVU_SDP_VF:
dev->hwcap |= OTX2_HWCAP_F_VF;
break;
}
diff --git a/drivers/common/octeontx2/otx2_dev.h b/drivers/common/octeontx2/otx2_dev.h
index 5fd05fa17..7d9839c33 100644
--- a/drivers/common/octeontx2/otx2_dev.h
+++ b/drivers/common/octeontx2/otx2_dev.h
@@ -19,6 +19,10 @@
#define otx2_dev_is_lbk(dev) ((dev->hwcap & OTX2_HWCAP_F_VF) && \
(dev->tx_chan_base < 0x700))
#define otx2_dev_revid(dev) (dev->hwcap & 0xFF)
+#define otx2_dev_is_sdp(dev) (dev->sdp_link)
+
+#define otx2_dev_is_vf_or_sdp(dev) \
+ (otx2_dev_is_vf(dev) || otx2_dev_is_sdp(dev))
#define otx2_dev_is_A0(dev) \
((RVU_PCI_REV_MAJOR(otx2_dev_revid(dev)) == 0x0) && \
diff --git a/drivers/net/octeontx2/otx2_ethdev.c b/drivers/net/octeontx2/otx2_ethdev.c
index 62291c698..b476b5ffb 100644
--- a/drivers/net/octeontx2/otx2_ethdev.c
+++ b/drivers/net/octeontx2/otx2_ethdev.c
@@ -120,7 +120,7 @@ otx2_cgx_rxtx_start(struct otx2_eth_dev *dev)
{
struct otx2_mbox *mbox = dev->mbox;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return 0;
otx2_mbox_alloc_msg_cgx_start_rxtx(mbox);
@@ -133,7 +133,7 @@ otx2_cgx_rxtx_stop(struct otx2_eth_dev *dev)
{
struct otx2_mbox *mbox = dev->mbox;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return 0;
otx2_mbox_alloc_msg_cgx_stop_rxtx(mbox);
@@ -166,7 +166,7 @@ nix_cgx_start_link_event(struct otx2_eth_dev *dev)
{
struct otx2_mbox *mbox = dev->mbox;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return 0;
otx2_mbox_alloc_msg_cgx_start_linkevents(mbox);
@@ -179,7 +179,7 @@ cgx_intlbk_enable(struct otx2_eth_dev *dev, bool en)
{
struct otx2_mbox *mbox = dev->mbox;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return 0;
if (en)
@@ -195,7 +195,7 @@ nix_cgx_stop_link_event(struct otx2_eth_dev *dev)
{
struct otx2_mbox *mbox = dev->mbox;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return 0;
otx2_mbox_alloc_msg_cgx_stop_linkevents(mbox);
@@ -2056,6 +2056,15 @@ otx2_eth_dev_lf_detach(struct otx2_mbox *mbox)
return otx2_mbox_process(mbox);
}
+static bool
+otx2_eth_dev_is_sdp(struct rte_pci_device *pci_dev)
+{
+ if (pci_dev->id.device_id == PCI_DEVID_OCTEONTX2_RVU_SDP_PF ||
+ pci_dev->id.device_id == PCI_DEVID_OCTEONTX2_RVU_SDP_VF)
+ return true;
+ return false;
+}
+
static int
otx2_eth_dev_init(struct rte_eth_dev *eth_dev)
{
@@ -2099,6 +2108,10 @@ otx2_eth_dev_init(struct rte_eth_dev *eth_dev)
goto error;
}
}
+ if (otx2_eth_dev_is_sdp(pci_dev))
+ dev->sdp_link = true;
+ else
+ dev->sdp_link = false;
/* Device generic callbacks */
dev->ops = &otx2_dev_ops;
dev->eth_dev = eth_dev;
@@ -2387,6 +2400,14 @@ static const struct rte_pci_id pci_nix_map[] = {
PCI_DEVID_OCTEONTX2_RVU_AF_VF)
},
{
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_CAVIUM,
+ PCI_DEVID_OCTEONTX2_RVU_SDP_PF)
+ },
+ {
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_CAVIUM,
+ PCI_DEVID_OCTEONTX2_RVU_SDP_VF)
+ },
+ {
.vendor_id = 0,
},
};
diff --git a/drivers/net/octeontx2/otx2_ethdev.h b/drivers/net/octeontx2/otx2_ethdev.h
index 4d9ed4870..b49e309fd 100644
--- a/drivers/net/octeontx2/otx2_ethdev.h
+++ b/drivers/net/octeontx2/otx2_ethdev.h
@@ -321,6 +321,7 @@ struct otx2_eth_dev {
uint64_t clk_delta;
bool mc_tbl_set;
struct otx2_nix_mc_filter_tbl mc_fltr_tbl;
+ bool sdp_link; /* SDP flag */
} __rte_cache_aligned;
struct otx2_eth_txq {
diff --git a/drivers/net/octeontx2/otx2_ethdev_ops.c b/drivers/net/octeontx2/otx2_ethdev_ops.c
index fc0fbd9f2..d1e77c324 100644
--- a/drivers/net/octeontx2/otx2_ethdev_ops.c
+++ b/drivers/net/octeontx2/otx2_ethdev_ops.c
@@ -36,6 +36,8 @@ otx2_nix_mtu_set(struct rte_eth_dev *eth_dev, uint16_t mtu)
req = otx2_mbox_alloc_msg_nix_set_hw_frs(mbox);
req->update_smq = true;
+ if (otx2_dev_is_sdp(dev))
+ req->sdp_link = true;
/* FRS HW config should exclude FCS but include NPC VTAG insert size */
req->maxlen = frame_size - RTE_ETHER_CRC_LEN + NIX_MAX_VTAG_ACT_SIZE;
@@ -46,6 +48,8 @@ otx2_nix_mtu_set(struct rte_eth_dev *eth_dev, uint16_t mtu)
/* Now just update Rx MAXLEN */
req = otx2_mbox_alloc_msg_nix_set_hw_frs(mbox);
req->maxlen = frame_size - RTE_ETHER_CRC_LEN;
+ if (otx2_dev_is_sdp(dev))
+ req->sdp_link = true;
rc = otx2_mbox_process(mbox);
if (rc)
@@ -98,7 +102,7 @@ nix_cgx_promisc_config(struct rte_eth_dev *eth_dev, int en)
struct otx2_eth_dev *dev = otx2_eth_pmd_priv(eth_dev);
struct otx2_mbox *mbox = dev->mbox;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return;
if (en)
diff --git a/drivers/net/octeontx2/otx2_flow_ctrl.c b/drivers/net/octeontx2/otx2_flow_ctrl.c
index 1d00e4687..0e725493b 100644
--- a/drivers/net/octeontx2/otx2_flow_ctrl.c
+++ b/drivers/net/octeontx2/otx2_flow_ctrl.c
@@ -14,6 +14,9 @@ otx2_nix_rxchan_bpid_cfg(struct rte_eth_dev *eth_dev, bool enb)
struct nix_bp_cfg_rsp *rsp;
int rc;
+ if (otx2_dev_is_sdp(dev))
+ return 0;
+
if (enb) {
req = otx2_mbox_alloc_msg_nix_bp_enable(mbox);
req->chan_base = 0;
@@ -189,6 +192,9 @@ otx2_nix_update_flow_ctrl_mode(struct rte_eth_dev *eth_dev)
struct otx2_eth_dev *dev = otx2_eth_pmd_priv(eth_dev);
struct rte_eth_fc_conf fc_conf;
+ if (otx2_dev_is_sdp(dev))
+ return 0;
+
memset(&fc_conf, 0, sizeof(struct rte_eth_fc_conf));
/* Both Rx & Tx flow ctrl get enabled(RTE_FC_FULL) in HW
* by AF driver, update those info in PMD structure.
diff --git a/drivers/net/octeontx2/otx2_link.c b/drivers/net/octeontx2/otx2_link.c
index 725b793d4..f5679b06e 100644
--- a/drivers/net/octeontx2/otx2_link.c
+++ b/drivers/net/octeontx2/otx2_link.c
@@ -93,7 +93,7 @@ otx2_nix_link_update(struct rte_eth_dev *eth_dev, int wait_to_complete)
RTE_SET_USED(wait_to_complete);
- if (otx2_dev_is_lbk(dev))
+ if (otx2_dev_is_lbk(dev) || otx2_dev_is_sdp(dev))
return 0;
otx2_mbox_alloc_msg_cgx_get_linkinfo(mbox);
@@ -129,7 +129,7 @@ otx2_nix_dev_set_link_up(struct rte_eth_dev *eth_dev)
struct otx2_eth_dev *dev = otx2_eth_pmd_priv(eth_dev);
int rc, i;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return -ENOTSUP;
rc = nix_dev_set_link_state(eth_dev, 1);
@@ -150,7 +150,7 @@ otx2_nix_dev_set_link_down(struct rte_eth_dev *eth_dev)
struct otx2_eth_dev *dev = otx2_eth_pmd_priv(eth_dev);
int i;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return -ENOTSUP;
/* Stop tx queues */
diff --git a/drivers/net/octeontx2/otx2_mac.c b/drivers/net/octeontx2/otx2_mac.c
index b4bcc61f8..262d185e5 100644
--- a/drivers/net/octeontx2/otx2_mac.c
+++ b/drivers/net/octeontx2/otx2_mac.c
@@ -15,7 +15,7 @@ otx2_cgx_mac_addr_set(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr)
struct otx2_mbox *mbox = dev->mbox;
int rc;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return -ENOTSUP;
if (otx2_dev_active_vfs(dev))
@@ -38,7 +38,7 @@ otx2_cgx_mac_max_entries_get(struct otx2_eth_dev *dev)
struct otx2_mbox *mbox = dev->mbox;
int rc;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return 0;
otx2_mbox_alloc_msg_cgx_mac_max_entries_get(mbox);
@@ -59,7 +59,7 @@ otx2_nix_mac_addr_add(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr,
struct cgx_mac_addr_add_rsp *rsp;
int rc;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return -ENOTSUP;
if (otx2_dev_active_vfs(dev))
@@ -89,7 +89,7 @@ otx2_nix_mac_addr_del(struct rte_eth_dev *eth_dev, uint32_t index)
struct cgx_mac_addr_del_req *req;
int rc;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return;
req = otx2_mbox_alloc_msg_cgx_mac_addr_del(mbox);
diff --git a/drivers/net/octeontx2/otx2_ptp.c b/drivers/net/octeontx2/otx2_ptp.c
index fd13c2678..4cd1cc8e5 100644
--- a/drivers/net/octeontx2/otx2_ptp.c
+++ b/drivers/net/octeontx2/otx2_ptp.c
@@ -104,7 +104,7 @@ nix_ptp_config(struct rte_eth_dev *eth_dev, int en)
struct otx2_mbox *mbox = dev->mbox;
uint8_t rc = -EINVAL;
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return rc;
if (en) {
@@ -168,7 +168,7 @@ otx2_nix_timesync_enable(struct rte_eth_dev *eth_dev)
}
/* If we are VF, no further action can be taken */
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return -EINVAL;
if (!(dev->rx_offload_flags & NIX_RX_OFFLOAD_PTYPE_F)) {
@@ -222,7 +222,7 @@ otx2_nix_timesync_disable(struct rte_eth_dev *eth_dev)
}
/* If we are VF, nothing else can be done */
- if (otx2_dev_is_vf(dev))
+ if (otx2_dev_is_vf_or_sdp(dev))
return -EINVAL;
dev->rx_offloads &= ~DEV_RX_OFFLOAD_TIMESTAMP;
diff --git a/drivers/net/octeontx2/otx2_tm.c b/drivers/net/octeontx2/otx2_tm.c
index 12903ec1b..ba615cebc 100644
--- a/drivers/net/octeontx2/otx2_tm.c
+++ b/drivers/net/octeontx2/otx2_tm.c
@@ -410,6 +410,12 @@ populate_tm_registers(struct otx2_eth_dev *dev,
*regval++ = shaper2regval(&cir) | 1;
req->num_regs++;
}
+ /* Configure TL4 to send to SDP channel instead of CGX/LBK */
+ if (otx2_dev_is_sdp(dev)) {
+ *reg++ = NIX_AF_TL4X_SDP_LINK_CFG(schq);
+ *regval++ = BIT_ULL(12);
+ req->num_regs++;
+ }
rc = send_tm_reqval(mbox, req);
if (rc)
@@ -465,9 +471,12 @@ populate_tm_registers(struct otx2_eth_dev *dev,
else
*regval++ = (strict_schedul_prio << 24) | rr_quantum;
req->num_regs++;
- *reg++ = NIX_AF_TL3_TL2X_LINKX_CFG(schq, nix_get_link(dev));
- *regval++ = BIT_ULL(12) | nix_get_relchan(dev);
- req->num_regs++;
+ if (!otx2_dev_is_sdp(dev)) {
+ *reg++ = NIX_AF_TL3_TL2X_LINKX_CFG(schq,
+ nix_get_link(dev));
+ *regval++ = BIT_ULL(12) | nix_get_relchan(dev);
+ req->num_regs++;
+ }
if (pir.rate && pir.burst) {
*reg++ = NIX_AF_TL2X_PIR(schq);
*regval++ = shaper2regval(&pir) | 1;
@@ -522,9 +531,6 @@ nix_tm_txsch_reg_config(struct otx2_eth_dev *dev)
uint32_t lvl;
int rc = 0;
- if (nix_get_link(dev) == 13)
- return -EPERM;
-
for (lvl = 0; lvl < (uint32_t)dev->otx2_tm_root_lvl + 1; lvl++) {
TAILQ_FOREACH(tm_node, &dev->node_list, node) {
if (tm_node->hw_lvl_id == lvl) {
--
2.13.6
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [dpdk-dev] [PATCH v1 1/1] net/octeontx2: add SDP interface support
2019-10-30 14:08 [dpdk-dev] [PATCH v1 1/1] net/octeontx2: add SDP interface support Subrahmanyam Nilla
@ 2019-11-06 6:45 ` Jerin Jacob
0 siblings, 0 replies; 2+ messages in thread
From: Jerin Jacob @ 2019-11-06 6:45 UTC (permalink / raw)
To: Subrahmanyam Nilla; +Cc: dpdk-dev, Jerin Jacob, shallyv, venkatn, mchalla
On Wed, Oct 30, 2019 at 7:38 PM Subrahmanyam Nilla <snilla@marvell.com> wrote:
>
> System DPI Packet Interface Unit (SDP) is a co-processor
> of OCTEON TX2 which provides PCIe endpoint support for a
> remote host to DMA packets into and out of the OCTEON TX2 SoC.
> SDP interface comes in to live only when it is connected in
> EP mode. It exposes input and output queue pairs to remote host
> for instruction input and packet output. It can be used as
> a communication channel between remote host and OCTEON TX2.
> Host machine needs to use corresponding user/kernel mode
> driver to communicate with SDP interface on OCTEON TX2 SoC.
>
> SDP interface support is limited to SDP PF device now.
> No SDP VF support.
>
> Signed-off-by: Subrahmanyam Nilla <snilla@marvell.com>
> Signed-off-by: Venkateshwarlu Nalla <venkatn@marvell.com>
This change looks good to me.
Could rebase the patch to dpdk-next-net-mrvl.git.
Applying: net/octeontx2: add SDP interface support
Using index info to reconstruct a base tree...
M doc/guides/nics/octeontx2.rst
M drivers/net/octeontx2/otx2_ethdev.c
M drivers/net/octeontx2/otx2_flow_ctrl.c
M drivers/net/octeontx2/otx2_ptp.c
Falling back to patching base and 3-way merge...
Auto-merging drivers/net/octeontx2/otx2_ptp.c
Auto-merging drivers/net/octeontx2/otx2_flow_ctrl.c
CONFLICT (content): Merge conflict in drivers/net/octeontx2/otx2_flow_ctrl.c
Auto-merging drivers/net/octeontx2/otx2_ethdev.c
Auto-merging doc/guides/nics/octeontx2.rst
error: Failed to merge in the changes.
Patch failed at 0001 net/octeontx2: add SDP interface support
hint: Use 'git am --show-current-patch' to see the failed patch
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".
HEAD is now at b7ce8bb35 net/octeontx2: fix compilation with x86-x32
git am failed /tmp/r//v1-1-1-net-octeontx2-add-SDP-interface-support
> ---
> doc/guides/nics/octeontx2.rst | 4 ++++
> doc/guides/platform/octeontx2.rst | 21 +++++++++++++++++++++
> drivers/common/octeontx2/otx2_common.h | 2 ++
> drivers/common/octeontx2/otx2_dev.c | 1 +
> drivers/common/octeontx2/otx2_dev.h | 4 ++++
> drivers/net/octeontx2/otx2_ethdev.c | 31 ++++++++++++++++++++++++++-----
> drivers/net/octeontx2/otx2_ethdev.h | 1 +
> drivers/net/octeontx2/otx2_ethdev_ops.c | 6 +++++-
> drivers/net/octeontx2/otx2_flow_ctrl.c | 6 ++++++
> drivers/net/octeontx2/otx2_link.c | 6 +++---
> drivers/net/octeontx2/otx2_mac.c | 8 ++++----
> drivers/net/octeontx2/otx2_ptp.c | 6 +++---
> drivers/net/octeontx2/otx2_tm.c | 18 ++++++++++++------
> 13 files changed, 92 insertions(+), 22 deletions(-)
>
> diff --git a/doc/guides/nics/octeontx2.rst b/doc/guides/nics/octeontx2.rst
> index fc8a130fb..7e901cec1 100644
> --- a/doc/guides/nics/octeontx2.rst
> +++ b/doc/guides/nics/octeontx2.rst
> @@ -200,6 +200,10 @@ Multicast MAC filtering
> ``net_octeontx2`` pmd supports multicast mac filtering feature only on physical
> function devices.
>
> +SDP interface support
> +~~~~~~~~~~~~~~~~~~~~~
> +OCTEON TX2 SDP interface support is limited to PF device, No VF support.
> +
> Debugging Options
> -----------------
>
> diff --git a/doc/guides/platform/octeontx2.rst b/doc/guides/platform/octeontx2.rst
> index 1c7d98f73..15b1641cf 100644
> --- a/doc/guides/platform/octeontx2.rst
> +++ b/doc/guides/platform/octeontx2.rst
> @@ -64,6 +64,8 @@ DPDK subsystem.
> +---+-----+--------------------------------------------------------------+
> | 8 | DPI | rte_rawdev |
> +---+-----+--------------------------------------------------------------+
> + | 9 | SDP | rte_ethdev |
> + +---+-----+--------------------------------------------------------------+
>
> PF0 is called the administrative / admin function (AF) and has exclusive
> privileges to provision RVU functional block's LFs to each of the PF/VF.
> @@ -102,6 +104,25 @@ Typical application usage models are,
> #. Exception path to Linux kernel from DPDK application as SW ``KNI`` replacement.
> #. Communication between two different DPDK applications.
>
> +SDP interface
> +-------------
> +
> +System DPI Packet Interface unit(SDP) provides PCIe endpoint support for remote host
> +to DMA packets into and out of OCTEON TX2 SoC. SDP interface comes in to live only when
> +OCTEON TX2 SoC is connected in PCIe endpoint mode. It can be used to send/receive
> +packets to/from remote host machine using input/output queue pairs exposed to it.
> +SDP interface receives input packets from remote host from NIX-RX and sends packets
> +to remote host using NIX-TX. Remote host machine need to use corresponding driver
> +(kernel/user mode) to communicate with SDP interface on OCTEON TX2 SoC. SDP supports
> +single PCIe SRIOV physical function(PF) and multiple virtual functions(VF's). Users
> +can bind PF or VF to use SDP interface and it will be enumerated as ethdev ports.
> +
> +The primary use case for SDP is to enable the smart NIC use case. Typical usage models are,
> +
> +#. Communication channel between remote host and OCTEON TX2 SoC over PCIe.
> +#. Transfer packets received from network interface to remote host over PCIe and
> + vice-versa.
> +
> OCTEON TX2 packet flow
> ----------------------
>
> diff --git a/drivers/common/octeontx2/otx2_common.h b/drivers/common/octeontx2/otx2_common.h
> index 0d0b87ee6..f62c45deb 100644
> --- a/drivers/common/octeontx2/otx2_common.h
> +++ b/drivers/common/octeontx2/otx2_common.h
> @@ -121,6 +121,8 @@ extern int otx2_logtype_dpi;
> #define PCI_DEVID_OCTEONTX2_RVU_CPT_VF 0xA0FE
> #define PCI_DEVID_OCTEONTX2_RVU_AF_VF 0xA0f8
> #define PCI_DEVID_OCTEONTX2_DPI_VF 0xA081
> +#define PCI_DEVID_OCTEONTX2_RVU_SDP_PF 0xA0f6
> +#define PCI_DEVID_OCTEONTX2_RVU_SDP_VF 0xA0f7
>
> /* Subsystem Device ID */
> #define PCI_SUBSYS_DEVID_96XX_95XX 0xB200
> diff --git a/drivers/common/octeontx2/otx2_dev.c b/drivers/common/octeontx2/otx2_dev.c
> index 910270975..0fc799e4a 100644
> --- a/drivers/common/octeontx2/otx2_dev.c
> +++ b/drivers/common/octeontx2/otx2_dev.c
> @@ -885,6 +885,7 @@ otx2_update_vf_hwcap(struct rte_pci_device *pci_dev, struct otx2_dev *dev)
> case PCI_DEVID_OCTEONTX2_RVU_CPT_VF:
> case PCI_DEVID_OCTEONTX2_RVU_AF_VF:
> case PCI_DEVID_OCTEONTX2_RVU_VF:
> + case PCI_DEVID_OCTEONTX2_RVU_SDP_VF:
> dev->hwcap |= OTX2_HWCAP_F_VF;
> break;
> }
> diff --git a/drivers/common/octeontx2/otx2_dev.h b/drivers/common/octeontx2/otx2_dev.h
> index 5fd05fa17..7d9839c33 100644
> --- a/drivers/common/octeontx2/otx2_dev.h
> +++ b/drivers/common/octeontx2/otx2_dev.h
> @@ -19,6 +19,10 @@
> #define otx2_dev_is_lbk(dev) ((dev->hwcap & OTX2_HWCAP_F_VF) && \
> (dev->tx_chan_base < 0x700))
> #define otx2_dev_revid(dev) (dev->hwcap & 0xFF)
> +#define otx2_dev_is_sdp(dev) (dev->sdp_link)
> +
> +#define otx2_dev_is_vf_or_sdp(dev) \
> + (otx2_dev_is_vf(dev) || otx2_dev_is_sdp(dev))
>
> #define otx2_dev_is_A0(dev) \
> ((RVU_PCI_REV_MAJOR(otx2_dev_revid(dev)) == 0x0) && \
> diff --git a/drivers/net/octeontx2/otx2_ethdev.c b/drivers/net/octeontx2/otx2_ethdev.c
> index 62291c698..b476b5ffb 100644
> --- a/drivers/net/octeontx2/otx2_ethdev.c
> +++ b/drivers/net/octeontx2/otx2_ethdev.c
> @@ -120,7 +120,7 @@ otx2_cgx_rxtx_start(struct otx2_eth_dev *dev)
> {
> struct otx2_mbox *mbox = dev->mbox;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return 0;
>
> otx2_mbox_alloc_msg_cgx_start_rxtx(mbox);
> @@ -133,7 +133,7 @@ otx2_cgx_rxtx_stop(struct otx2_eth_dev *dev)
> {
> struct otx2_mbox *mbox = dev->mbox;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return 0;
>
> otx2_mbox_alloc_msg_cgx_stop_rxtx(mbox);
> @@ -166,7 +166,7 @@ nix_cgx_start_link_event(struct otx2_eth_dev *dev)
> {
> struct otx2_mbox *mbox = dev->mbox;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return 0;
>
> otx2_mbox_alloc_msg_cgx_start_linkevents(mbox);
> @@ -179,7 +179,7 @@ cgx_intlbk_enable(struct otx2_eth_dev *dev, bool en)
> {
> struct otx2_mbox *mbox = dev->mbox;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return 0;
>
> if (en)
> @@ -195,7 +195,7 @@ nix_cgx_stop_link_event(struct otx2_eth_dev *dev)
> {
> struct otx2_mbox *mbox = dev->mbox;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return 0;
>
> otx2_mbox_alloc_msg_cgx_stop_linkevents(mbox);
> @@ -2056,6 +2056,15 @@ otx2_eth_dev_lf_detach(struct otx2_mbox *mbox)
> return otx2_mbox_process(mbox);
> }
>
> +static bool
> +otx2_eth_dev_is_sdp(struct rte_pci_device *pci_dev)
> +{
> + if (pci_dev->id.device_id == PCI_DEVID_OCTEONTX2_RVU_SDP_PF ||
> + pci_dev->id.device_id == PCI_DEVID_OCTEONTX2_RVU_SDP_VF)
> + return true;
> + return false;
> +}
> +
> static int
> otx2_eth_dev_init(struct rte_eth_dev *eth_dev)
> {
> @@ -2099,6 +2108,10 @@ otx2_eth_dev_init(struct rte_eth_dev *eth_dev)
> goto error;
> }
> }
> + if (otx2_eth_dev_is_sdp(pci_dev))
> + dev->sdp_link = true;
> + else
> + dev->sdp_link = false;
> /* Device generic callbacks */
> dev->ops = &otx2_dev_ops;
> dev->eth_dev = eth_dev;
> @@ -2387,6 +2400,14 @@ static const struct rte_pci_id pci_nix_map[] = {
> PCI_DEVID_OCTEONTX2_RVU_AF_VF)
> },
> {
> + RTE_PCI_DEVICE(PCI_VENDOR_ID_CAVIUM,
> + PCI_DEVID_OCTEONTX2_RVU_SDP_PF)
> + },
> + {
> + RTE_PCI_DEVICE(PCI_VENDOR_ID_CAVIUM,
> + PCI_DEVID_OCTEONTX2_RVU_SDP_VF)
> + },
> + {
> .vendor_id = 0,
> },
> };
> diff --git a/drivers/net/octeontx2/otx2_ethdev.h b/drivers/net/octeontx2/otx2_ethdev.h
> index 4d9ed4870..b49e309fd 100644
> --- a/drivers/net/octeontx2/otx2_ethdev.h
> +++ b/drivers/net/octeontx2/otx2_ethdev.h
> @@ -321,6 +321,7 @@ struct otx2_eth_dev {
> uint64_t clk_delta;
> bool mc_tbl_set;
> struct otx2_nix_mc_filter_tbl mc_fltr_tbl;
> + bool sdp_link; /* SDP flag */
> } __rte_cache_aligned;
>
> struct otx2_eth_txq {
> diff --git a/drivers/net/octeontx2/otx2_ethdev_ops.c b/drivers/net/octeontx2/otx2_ethdev_ops.c
> index fc0fbd9f2..d1e77c324 100644
> --- a/drivers/net/octeontx2/otx2_ethdev_ops.c
> +++ b/drivers/net/octeontx2/otx2_ethdev_ops.c
> @@ -36,6 +36,8 @@ otx2_nix_mtu_set(struct rte_eth_dev *eth_dev, uint16_t mtu)
>
> req = otx2_mbox_alloc_msg_nix_set_hw_frs(mbox);
> req->update_smq = true;
> + if (otx2_dev_is_sdp(dev))
> + req->sdp_link = true;
> /* FRS HW config should exclude FCS but include NPC VTAG insert size */
> req->maxlen = frame_size - RTE_ETHER_CRC_LEN + NIX_MAX_VTAG_ACT_SIZE;
>
> @@ -46,6 +48,8 @@ otx2_nix_mtu_set(struct rte_eth_dev *eth_dev, uint16_t mtu)
> /* Now just update Rx MAXLEN */
> req = otx2_mbox_alloc_msg_nix_set_hw_frs(mbox);
> req->maxlen = frame_size - RTE_ETHER_CRC_LEN;
> + if (otx2_dev_is_sdp(dev))
> + req->sdp_link = true;
>
> rc = otx2_mbox_process(mbox);
> if (rc)
> @@ -98,7 +102,7 @@ nix_cgx_promisc_config(struct rte_eth_dev *eth_dev, int en)
> struct otx2_eth_dev *dev = otx2_eth_pmd_priv(eth_dev);
> struct otx2_mbox *mbox = dev->mbox;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return;
>
> if (en)
> diff --git a/drivers/net/octeontx2/otx2_flow_ctrl.c b/drivers/net/octeontx2/otx2_flow_ctrl.c
> index 1d00e4687..0e725493b 100644
> --- a/drivers/net/octeontx2/otx2_flow_ctrl.c
> +++ b/drivers/net/octeontx2/otx2_flow_ctrl.c
> @@ -14,6 +14,9 @@ otx2_nix_rxchan_bpid_cfg(struct rte_eth_dev *eth_dev, bool enb)
> struct nix_bp_cfg_rsp *rsp;
> int rc;
>
> + if (otx2_dev_is_sdp(dev))
> + return 0;
> +
> if (enb) {
> req = otx2_mbox_alloc_msg_nix_bp_enable(mbox);
> req->chan_base = 0;
> @@ -189,6 +192,9 @@ otx2_nix_update_flow_ctrl_mode(struct rte_eth_dev *eth_dev)
> struct otx2_eth_dev *dev = otx2_eth_pmd_priv(eth_dev);
> struct rte_eth_fc_conf fc_conf;
>
> + if (otx2_dev_is_sdp(dev))
> + return 0;
> +
> memset(&fc_conf, 0, sizeof(struct rte_eth_fc_conf));
> /* Both Rx & Tx flow ctrl get enabled(RTE_FC_FULL) in HW
> * by AF driver, update those info in PMD structure.
> diff --git a/drivers/net/octeontx2/otx2_link.c b/drivers/net/octeontx2/otx2_link.c
> index 725b793d4..f5679b06e 100644
> --- a/drivers/net/octeontx2/otx2_link.c
> +++ b/drivers/net/octeontx2/otx2_link.c
> @@ -93,7 +93,7 @@ otx2_nix_link_update(struct rte_eth_dev *eth_dev, int wait_to_complete)
>
> RTE_SET_USED(wait_to_complete);
>
> - if (otx2_dev_is_lbk(dev))
> + if (otx2_dev_is_lbk(dev) || otx2_dev_is_sdp(dev))
> return 0;
>
> otx2_mbox_alloc_msg_cgx_get_linkinfo(mbox);
> @@ -129,7 +129,7 @@ otx2_nix_dev_set_link_up(struct rte_eth_dev *eth_dev)
> struct otx2_eth_dev *dev = otx2_eth_pmd_priv(eth_dev);
> int rc, i;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return -ENOTSUP;
>
> rc = nix_dev_set_link_state(eth_dev, 1);
> @@ -150,7 +150,7 @@ otx2_nix_dev_set_link_down(struct rte_eth_dev *eth_dev)
> struct otx2_eth_dev *dev = otx2_eth_pmd_priv(eth_dev);
> int i;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return -ENOTSUP;
>
> /* Stop tx queues */
> diff --git a/drivers/net/octeontx2/otx2_mac.c b/drivers/net/octeontx2/otx2_mac.c
> index b4bcc61f8..262d185e5 100644
> --- a/drivers/net/octeontx2/otx2_mac.c
> +++ b/drivers/net/octeontx2/otx2_mac.c
> @@ -15,7 +15,7 @@ otx2_cgx_mac_addr_set(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr)
> struct otx2_mbox *mbox = dev->mbox;
> int rc;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return -ENOTSUP;
>
> if (otx2_dev_active_vfs(dev))
> @@ -38,7 +38,7 @@ otx2_cgx_mac_max_entries_get(struct otx2_eth_dev *dev)
> struct otx2_mbox *mbox = dev->mbox;
> int rc;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return 0;
>
> otx2_mbox_alloc_msg_cgx_mac_max_entries_get(mbox);
> @@ -59,7 +59,7 @@ otx2_nix_mac_addr_add(struct rte_eth_dev *eth_dev, struct rte_ether_addr *addr,
> struct cgx_mac_addr_add_rsp *rsp;
> int rc;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return -ENOTSUP;
>
> if (otx2_dev_active_vfs(dev))
> @@ -89,7 +89,7 @@ otx2_nix_mac_addr_del(struct rte_eth_dev *eth_dev, uint32_t index)
> struct cgx_mac_addr_del_req *req;
> int rc;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return;
>
> req = otx2_mbox_alloc_msg_cgx_mac_addr_del(mbox);
> diff --git a/drivers/net/octeontx2/otx2_ptp.c b/drivers/net/octeontx2/otx2_ptp.c
> index fd13c2678..4cd1cc8e5 100644
> --- a/drivers/net/octeontx2/otx2_ptp.c
> +++ b/drivers/net/octeontx2/otx2_ptp.c
> @@ -104,7 +104,7 @@ nix_ptp_config(struct rte_eth_dev *eth_dev, int en)
> struct otx2_mbox *mbox = dev->mbox;
> uint8_t rc = -EINVAL;
>
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return rc;
>
> if (en) {
> @@ -168,7 +168,7 @@ otx2_nix_timesync_enable(struct rte_eth_dev *eth_dev)
> }
>
> /* If we are VF, no further action can be taken */
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return -EINVAL;
>
> if (!(dev->rx_offload_flags & NIX_RX_OFFLOAD_PTYPE_F)) {
> @@ -222,7 +222,7 @@ otx2_nix_timesync_disable(struct rte_eth_dev *eth_dev)
> }
>
> /* If we are VF, nothing else can be done */
> - if (otx2_dev_is_vf(dev))
> + if (otx2_dev_is_vf_or_sdp(dev))
> return -EINVAL;
>
> dev->rx_offloads &= ~DEV_RX_OFFLOAD_TIMESTAMP;
> diff --git a/drivers/net/octeontx2/otx2_tm.c b/drivers/net/octeontx2/otx2_tm.c
> index 12903ec1b..ba615cebc 100644
> --- a/drivers/net/octeontx2/otx2_tm.c
> +++ b/drivers/net/octeontx2/otx2_tm.c
> @@ -410,6 +410,12 @@ populate_tm_registers(struct otx2_eth_dev *dev,
> *regval++ = shaper2regval(&cir) | 1;
> req->num_regs++;
> }
> + /* Configure TL4 to send to SDP channel instead of CGX/LBK */
> + if (otx2_dev_is_sdp(dev)) {
> + *reg++ = NIX_AF_TL4X_SDP_LINK_CFG(schq);
> + *regval++ = BIT_ULL(12);
> + req->num_regs++;
> + }
>
> rc = send_tm_reqval(mbox, req);
> if (rc)
> @@ -465,9 +471,12 @@ populate_tm_registers(struct otx2_eth_dev *dev,
> else
> *regval++ = (strict_schedul_prio << 24) | rr_quantum;
> req->num_regs++;
> - *reg++ = NIX_AF_TL3_TL2X_LINKX_CFG(schq, nix_get_link(dev));
> - *regval++ = BIT_ULL(12) | nix_get_relchan(dev);
> - req->num_regs++;
> + if (!otx2_dev_is_sdp(dev)) {
> + *reg++ = NIX_AF_TL3_TL2X_LINKX_CFG(schq,
> + nix_get_link(dev));
> + *regval++ = BIT_ULL(12) | nix_get_relchan(dev);
> + req->num_regs++;
> + }
> if (pir.rate && pir.burst) {
> *reg++ = NIX_AF_TL2X_PIR(schq);
> *regval++ = shaper2regval(&pir) | 1;
> @@ -522,9 +531,6 @@ nix_tm_txsch_reg_config(struct otx2_eth_dev *dev)
> uint32_t lvl;
> int rc = 0;
>
> - if (nix_get_link(dev) == 13)
> - return -EPERM;
> -
> for (lvl = 0; lvl < (uint32_t)dev->otx2_tm_root_lvl + 1; lvl++) {
> TAILQ_FOREACH(tm_node, &dev->node_list, node) {
> if (tm_node->hw_lvl_id == lvl) {
> --
> 2.13.6
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2019-11-06 6:46 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-10-30 14:08 [dpdk-dev] [PATCH v1 1/1] net/octeontx2: add SDP interface support Subrahmanyam Nilla
2019-11-06 6:45 ` Jerin Jacob
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).