* [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
@ 2016-12-15 7:59 Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 01/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) documentation Vishwanathapura, Niranjana
` (6 more replies)
0 siblings, 7 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w
Thanks Jason for the valuable feedback.
Here is the revised HFI VNIC patch series.
ChangeLog:
=========
v1 => v2:
a) Removed hfi_vnic bus, instead make hfi_vnic driver an 'ib client',
as per feedback from Jason Gunthorpe.
b) Interface changes, data structure changes and variable name changes
associated with (a).
c) Add hfi_ibdev abstraction to provide VNIC control operations to
hfi_vnic client.
d) Minor fixes
e) Moved hfi_vnic driver from .../sw/intel/vnic/hfi_vnic to
.../sw/intel/hfi_vnic.
v1: Initial post @ https://www.spinics.net/lists/linux-rdma/msg43158.html
Description:
============
Intel Omni-Path Host Fabric Interface (HFI) Virtual Network Interface
Controller (VNIC) feature supports Ethernet functionality over Omni-Path
fabric by encapsulating the Ethernet packets between HFI nodes.
The patterns of exchanges of Omni-Path encapsulated Ethernet packets
involves one or more virtual Ethernet switches overlaid on the Omni-Path
fabric topology. A subset of HFI nodes on the Omni-Path fabric are
permitted to exchange encapsulated Ethernet packets across a particular
virtual Ethernet switch. The virtual Ethernet switches are logical
abstractions achieved by configuring the HFI nodes on the fabric for
header generation and processing. In the simplest configuration all HFI
nodes across the fabric exchange encapsulated Ethernet packets over a
single virtual Ethernet switch. A virtual Ethernet switch, is effectively
an independent Ethernet network. The configuration is performed by an
Ethernet Manager (EM) which is part of the trusted Fabric Manager (FM)
application. HFI nodes can have multiple VNICs each connected to a
different virtual Ethernet switch. The below diagram presents a case
of two virtual Ethernet switches with two HFI nodes.
+-------------------+
| Subnet/ |
| Ethernet |
| Manager |
+-------------------+
/ /
/ /
/ /
/ /
+-----------------------------+ +------------------------------+
| Virtual Ethernet Switch | | Virtual Ethernet Switch |
| +---------+ +---------+ | | +---------+ +---------+ |
| | VPORT | | VPORT | | | | VPORT | | VPORT | |
+--+---------+----+---------+-+ +-+---------+----+---------+---+
| \ / |
| \ / |
| \/ |
| / \ |
| / \ |
+-----------+------------+ +-----------+------------+
| VNIC | VNIC | | VNIC | VNIC |
+-----------+------------+ +-----------+------------+
| HFI | | HFI |
+------------------------+ +------------------------+
Intel HFI VNIC software design is presented in the below diagram.
HFI VNIC functionality has a HW dependent component and a HW
independent component.
The HW dependent VNIC functionality is part of the HFI1 driver. It
implements the callback functions to do various tasks which includes
adding and removing of VNIC ports, HW resource allocation for VNIC
functionality and actual transmission and reception of encapsulated
Ethernet packets over the fabric. Each VNIC port is addressed by the
HFI port number, and the VNIC port number on that HFI port.
The HFI VNIC module implements the HW independent VNIC functionality.
It consists of two parts. The VNIC Ethernet Management Agent (VEMA)
registers itself with IB core as an IB client and interfaces with the
IB MAD stack. It exchanges the management information with the Ethernet
Manager (EM) and the VNIC netdev. The VNIC netdev part interfaces with
the Linux network stack, thus providing standard Ethernet network
interfaces. It invokes HFI device's VNIC callback functions for HW access.
The VNIC netdev encapsulates the Ethernet packets with an Omni-Path
header before passing them to the HFI1 driver for transmission.
Similarly, it de-encapsulates the received Omni-Path packets before
passing them to the network stack. For each VNIC interface, the
information required for encapsulation is configured by EM via VEMA MAD
interface.
+-------------------+ +----------------------+
| | | Linux |
| IB MAD | | Network |
| | | Stack |
+-------------------+ +----------------------+
| |
| |
+--------------------------------------------+
| |
| HFI VNIC Module |
| (HFI VNIC Netdev and EMA drivers) |
| |
+--------------------------------------------+
|
|
+------------------+
| IB core |
+------------------+
|
|
+--------------------------------------------+
| |
| HFI1 Driver with VNIC support |
| |
+--------------------------------------------+
Vishwanathapura, Niranjana (10):
IB/hfi-vnic: Virtual Network Interface Controller (VNIC) documentation
IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface
IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
IB/hfi-vnic: VNIC Ethernet Management (EM) structure definitions
IB/hfi-vnic: VNIC statistics support
IB/hfi-vnic: VNIC MAC table support
IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) interface
IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) function
IB/hfi1: Virtual Network Interface Controller (VNIC) support
IB/hfi1: VNIC SDMA support
Documentation/infiniband/hfi_vnic.txt | 95 ++
MAINTAINERS | 7 +
drivers/infiniband/Kconfig | 1 +
drivers/infiniband/hw/hfi1/Makefile | 2 +-
drivers/infiniband/hw/hfi1/aspm.h | 13 +-
drivers/infiniband/hw/hfi1/chip.c | 272 +++++-
drivers/infiniband/hw/hfi1/chip.h | 2 +
drivers/infiniband/hw/hfi1/debugfs.c | 6 +-
drivers/infiniband/hw/hfi1/driver.c | 84 +-
drivers/infiniband/hw/hfi1/file_ops.c | 25 +-
drivers/infiniband/hw/hfi1/hfi.h | 52 +-
drivers/infiniband/hw/hfi1/init.c | 41 +-
drivers/infiniband/hw/hfi1/intr.c | 2 +-
drivers/infiniband/hw/hfi1/mad.c | 10 +-
drivers/infiniband/hw/hfi1/pio.c | 17 +
drivers/infiniband/hw/hfi1/pio.h | 6 +
drivers/infiniband/hw/hfi1/qp.c | 24 +-
drivers/infiniband/hw/hfi1/ruc.c | 2 +-
drivers/infiniband/hw/hfi1/sysfs.c | 24 +-
drivers/infiniband/hw/hfi1/user_exp_rcv.c | 6 +-
drivers/infiniband/hw/hfi1/user_pages.c | 3 +-
drivers/infiniband/hw/hfi1/verbs.c | 120 +--
drivers/infiniband/hw/hfi1/verbs.h | 9 +-
drivers/infiniband/hw/hfi1/vnic.h | 173 ++++
drivers/infiniband/hw/hfi1/vnic_main.c | 631 ++++++++++++
drivers/infiniband/hw/hfi1/vnic_sdma.c | 320 ++++++
drivers/infiniband/sw/Makefile | 1 +
drivers/infiniband/sw/intel/hfi_vnic/Kconfig | 8 +
drivers/infiniband/sw/intel/hfi_vnic/Makefile | 7 +
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c | 489 ++++++++++
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h | 510 ++++++++++
.../sw/intel/hfi_vnic/hfi_vnic_ethtool.c | 208 ++++
.../sw/intel/hfi_vnic/hfi_vnic_internal.h | 443 +++++++++
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c | 810 ++++++++++++++++
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_vema.c | 1024 ++++++++++++++++++++
.../sw/intel/hfi_vnic/hfi_vnic_vema_iface.c | 432 +++++++++
include/rdma/opa_hfi.h | 199 ++++
include/rdma/opa_port_info.h | 2 +-
38 files changed, 5891 insertions(+), 189 deletions(-)
create mode 100644 Documentation/infiniband/hfi_vnic.txt
create mode 100644 drivers/infiniband/hw/hfi1/vnic.h
create mode 100644 drivers/infiniband/hw/hfi1/vnic_main.c
create mode 100644 drivers/infiniband/hw/hfi1/vnic_sdma.c
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/Kconfig
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/Makefile
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema.c
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema_iface.c
create mode 100644 include/rdma/opa_hfi.h
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* [RFC v2 01/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) documentation
2016-12-15 7:59 [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
[not found] ` <1481788782-89964-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
` (5 subsequent siblings)
6 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura
Add HFI VNIC design document explaining the VNIC architecture and the
driver design.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
---
Documentation/infiniband/hfi_vnic.txt | 95 +++++++++++++++++++++++++++++++++++
1 file changed, 95 insertions(+)
create mode 100644 Documentation/infiniband/hfi_vnic.txt
diff --git a/Documentation/infiniband/hfi_vnic.txt b/Documentation/infiniband/hfi_vnic.txt
new file mode 100644
index 0000000..1f39d8b
--- /dev/null
+++ b/Documentation/infiniband/hfi_vnic.txt
@@ -0,0 +1,95 @@
+Intel Omni-Path Host Fabric Interface (HFI) Virtual Network Interface
+Controller (VNIC) feature supports Ethernet functionality over Omni-Path
+fabric by encapsulating the Ethernet packets between HFI nodes.
+
+The patterns of exchanges of Omni-Path encapsulated Ethernet packets
+involves one or more virtual Ethernet switches overlaid on the Omni-Path
+fabric topology. A subset of HFI nodes on the Omni-Path fabric are
+permitted to exchange encapsulated Ethernet packets across a particular
+virtual Ethernet switch. The virtual Ethernet switches are logical
+abstractions achieved by configuring the HFI nodes on the fabric for
+header generation and processing. In the simplest configuration all HFI
+nodes across the fabric exchange encapsulated Ethernet packets over a
+single virtual Ethernet switch. A virtual Ethernet switch, is effectively
+an independent Ethernet network. The configuration is performed by an
+Ethernet Manager (EM) which is part of the trusted Fabric Manager (FM)
+application. HFI nodes can have multiple VNICs each connected to a
+different virtual Ethernet switch. The below diagram presents a case
+of two virtual Ethernet switches with two HFI nodes.
+
+ +-------------------+
+ | Subnet/ |
+ | Ethernet |
+ | Manager |
+ +-------------------+
+ / /
+ / /
+ / /
+ / /
++-----------------------------+ +------------------------------+
+| Virtual Ethernet Switch | | Virtual Ethernet Switch |
+| +---------+ +---------+ | | +---------+ +---------+ |
+| | VPORT | | VPORT | | | | VPORT | | VPORT | |
++--+---------+----+---------+-+ +-+---------+----+---------+---+
+ | \ / |
+ | \ / |
+ | \/ |
+ | / \ |
+ | / \ |
+ +-----------+------------+ +-----------+------------+
+ | VNIC | VNIC | | VNIC | VNIC |
+ +-----------+------------+ +-----------+------------+
+ | HFI | | HFI |
+ +------------------------+ +------------------------+
+
+Intel HFI VNIC software design is presented in the below diagram.
+HFI VNIC functionality has a HW dependent component and a HW
+independent component.
+
+The HW dependent VNIC functionality is part of the HFI1 driver. It
+implements the callback functions to do various tasks which includes
+adding and removing of VNIC ports, HW resource allocation for VNIC
+functionality and actual transmission and reception of encapsulated
+Ethernet packets over the fabric. Each VNIC port is addressed by the
+HFI port number, and the VNIC port number on that HFI port.
+
+The HFI VNIC module implements the HW independent VNIC functionality.
+It consists of two parts. The VNIC Ethernet Management Agent (VEMA)
+registers itself with IB core as an IB client and interfaces with the
+IB MAD stack. It exchanges the management information with the Ethernet
+Manager (EM) and the VNIC netdev. The VNIC netdev part interfaces with
+the Linux network stack, thus providing standard Ethernet network
+interfaces. It invokes HFI device's VNIC callback functions for HW access.
+The VNIC netdev encapsulates the Ethernet packets with an Omni-Path
+header before passing them to the HFI1 driver for transmission.
+Similarly, it de-encapsulates the received Omni-Path packets before
+passing them to the network stack. For each VNIC interface, the
+information required for encapsulation is configured by EM via VEMA MAD
+interface.
+
+
+ +-------------------+ +----------------------+
+ | | | Linux |
+ | IB MAD | | Network |
+ | | | Stack |
+ +-------------------+ +----------------------+
+ | |
+ | |
+ +--------------------------------------------+
+ | |
+ | HFI VNIC Module |
+ | (HFI VNIC Netdev and EMA drivers) |
+ | |
+ +--------------------------------------------+
+ |
+ |
+ +------------------+
+ | IB core |
+ +------------------+
+ |
+ |
+ +--------------------------------------------+
+ | |
+ | HFI1 Driver with VNIC support |
+ | |
+ +--------------------------------------------+
--
1.8.3.1
^ permalink raw reply related [flat|nested] 41+ messages in thread
* [RFC v2 02/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface
[not found] ` <1481788782-89964-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
[not found] ` <1481788782-89964-3-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2016-12-15 7:59 ` [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev Vishwanathapura, Niranjana
` (4 subsequent siblings)
5 siblings, 1 reply; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Niranjana Vishwanathapura
Create hfi_ibdev abstraction which hfi1_ibdev will extend.
Define HFI VNIC interface between hardware independent VNIC
functionality and the hardware dependent VNIC functionality.
Add VNIC control operations to add and remove VNIC devices,
to the hfi_ibdev structure.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/hw/hfi1/chip.c | 2 +-
drivers/infiniband/hw/hfi1/driver.c | 10 +-
drivers/infiniband/hw/hfi1/hfi.h | 2 +-
drivers/infiniband/hw/hfi1/init.c | 4 +-
drivers/infiniband/hw/hfi1/intr.c | 2 +-
drivers/infiniband/hw/hfi1/mad.c | 2 +-
drivers/infiniband/hw/hfi1/qp.c | 24 +++--
drivers/infiniband/hw/hfi1/ruc.c | 2 +-
drivers/infiniband/hw/hfi1/sysfs.c | 22 ++--
drivers/infiniband/hw/hfi1/verbs.c | 113 ++++++++++----------
drivers/infiniband/hw/hfi1/verbs.h | 9 +-
include/rdma/opa_hfi.h | 199 ++++++++++++++++++++++++++++++++++++
12 files changed, 298 insertions(+), 93 deletions(-)
create mode 100644 include/rdma/opa_hfi.h
diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c
index 37d8af5..9263984 100644
--- a/drivers/infiniband/hw/hfi1/chip.c
+++ b/drivers/infiniband/hw/hfi1/chip.c
@@ -10452,7 +10452,7 @@ int set_link_state(struct hfi1_pportdata *ppd, u32 state)
sdma_all_running(dd);
/* Signal the IB layer that the port has went active */
- event.device = &dd->verbs_dev.rdi.ibdev;
+ event.device = &dd->verbs_dev.hfidev.rdi.ibdev;
event.element.port_num = ppd->port;
event.event = IB_EVENT_PORT_ACTIVE;
}
diff --git a/drivers/infiniband/hw/hfi1/driver.c b/drivers/infiniband/hw/hfi1/driver.c
index d426116..e219c3b 100644
--- a/drivers/infiniband/hw/hfi1/driver.c
+++ b/drivers/infiniband/hw/hfi1/driver.c
@@ -163,7 +163,8 @@ const char *get_unit_name(int unit)
const char *get_card_name(struct rvt_dev_info *rdi)
{
- struct hfi1_ibdev *ibdev = container_of(rdi, struct hfi1_ibdev, rdi);
+ struct hfi1_ibdev *ibdev = container_of(rdi, struct hfi1_ibdev,
+ hfidev.rdi);
struct hfi1_devdata *dd = container_of(ibdev,
struct hfi1_devdata, verbs_dev);
return get_unit_name(dd->unit);
@@ -171,7 +172,8 @@ const char *get_card_name(struct rvt_dev_info *rdi)
struct pci_dev *get_pci_dev(struct rvt_dev_info *rdi)
{
- struct hfi1_ibdev *ibdev = container_of(rdi, struct hfi1_ibdev, rdi);
+ struct hfi1_ibdev *ibdev = container_of(rdi, struct hfi1_ibdev,
+ hfidev.rdi);
struct hfi1_devdata *dd = container_of(ibdev,
struct hfi1_devdata, verbs_dev);
return dd->pcidev;
@@ -281,7 +283,7 @@ static void rcv_hdrerr(struct hfi1_ctxtdata *rcd, struct hfi1_pportdata *ppd,
int lnh = be16_to_cpu(rhdr->lrh[0]) & 3;
struct hfi1_ibport *ibp = &ppd->ibport_data;
struct hfi1_devdata *dd = ppd->dd;
- struct rvt_dev_info *rdi = &dd->verbs_dev.rdi;
+ struct rvt_dev_info *rdi = &dd->verbs_dev.hfidev.rdi;
if (packet->rhf & (RHF_VCRC_ERR | RHF_ICRC_ERR))
return;
@@ -600,7 +602,7 @@ static void __prescan_rxq(struct hfi1_packet *packet)
struct rvt_qp *qp;
struct ib_header *hdr;
struct ib_other_headers *ohdr;
- struct rvt_dev_info *rdi = &dd->verbs_dev.rdi;
+ struct rvt_dev_info *rdi = &dd->verbs_dev.hfidev.rdi;
u64 rhf = rhf_to_cpu(rhf_addr);
u32 etype = rhf_rcv_type(rhf), qpn, bth1;
int is_ecn = 0;
diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h
index 4163596..1fc5b68 100644
--- a/drivers/infiniband/hw/hfi1/hfi.h
+++ b/drivers/infiniband/hw/hfi1/hfi.h
@@ -1601,7 +1601,7 @@ static inline struct hfi1_pportdata *ppd_from_ibp(struct hfi1_ibport *ibp)
static inline struct hfi1_ibdev *dev_from_rdi(struct rvt_dev_info *rdi)
{
- return container_of(rdi, struct hfi1_ibdev, rdi);
+ return container_of(rdi, struct hfi1_ibdev, hfidev.rdi);
}
static inline struct hfi1_ibport *to_iport(struct ib_device *ibdev, u8 port)
diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c
index 60db615..13f6862 100644
--- a/drivers/infiniband/hw/hfi1/init.c
+++ b/drivers/infiniband/hw/hfi1/init.c
@@ -1020,7 +1020,7 @@ static void __hfi1_free_devdata(struct kobject *kobj)
free_percpu(dd->int_counter);
free_percpu(dd->rcv_limit);
free_percpu(dd->send_schedule);
- rvt_dealloc_device(&dd->verbs_dev.rdi);
+ rvt_dealloc_device(&dd->verbs_dev.hfidev.rdi);
}
static struct kobj_type hfi1_devdata_type = {
@@ -1133,7 +1133,7 @@ struct hfi1_devdata *hfi1_alloc_devdata(struct pci_dev *pdev, size_t extra)
bail:
if (!list_empty(&dd->list))
list_del_init(&dd->list);
- rvt_dealloc_device(&dd->verbs_dev.rdi);
+ rvt_dealloc_device(&dd->verbs_dev.hfidev.rdi);
return ERR_PTR(ret);
}
diff --git a/drivers/infiniband/hw/hfi1/intr.c b/drivers/infiniband/hw/hfi1/intr.c
index 65348d1..b4f94b1 100644
--- a/drivers/infiniband/hw/hfi1/intr.c
+++ b/drivers/infiniband/hw/hfi1/intr.c
@@ -95,7 +95,7 @@ static void signal_ib_event(struct hfi1_pportdata *ppd, enum ib_event_type ev)
*/
if (!(dd->flags & HFI1_INITTED))
return;
- event.device = &dd->verbs_dev.rdi.ibdev;
+ event.device = &dd->verbs_dev.hfidev.rdi.ibdev;
event.element.port_num = ppd->port;
event.event = ev;
ib_dispatch_event(&event);
diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c
index 6e595af..ed8ae22 100644
--- a/drivers/infiniband/hw/hfi1/mad.c
+++ b/drivers/infiniband/hw/hfi1/mad.c
@@ -83,7 +83,7 @@ void hfi1_event_pkey_change(struct hfi1_devdata *dd, u8 port)
struct ib_event event;
event.event = IB_EVENT_PKEY_CHANGE;
- event.device = &dd->verbs_dev.rdi.ibdev;
+ event.device = &dd->verbs_dev.hfidev.rdi.ibdev;
event.element.port_num = port;
ib_dispatch_event(&event);
}
diff --git a/drivers/infiniband/hw/hfi1/qp.c b/drivers/infiniband/hw/hfi1/qp.c
index d752d67..cf0f6ed 100644
--- a/drivers/infiniband/hw/hfi1/qp.c
+++ b/drivers/infiniband/hw/hfi1/qp.c
@@ -659,7 +659,7 @@ struct qp_iter *qp_iter_init(struct hfi1_ibdev *dev)
return NULL;
iter->dev = dev;
- iter->specials = dev->rdi.ibdev.phys_port_cnt * 2;
+ iter->specials = dev->hfidev.rdi.ibdev.phys_port_cnt * 2;
return iter;
}
@@ -682,11 +682,13 @@ int qp_iter_next(struct qp_iter *iter)
*
* n = 0..iter->specials is the special qp indices
*
- * n = iter->specials..dev->rdi.qp_dev->qp_table_size+iter->specials are
+ * n = iter->specials..
+ * dev->hfidev.rdi.qp_dev->qp_table_size+iter->specials are
* the potential hash bucket entries
*
*/
- for (; n < dev->rdi.qp_dev->qp_table_size + iter->specials; n++) {
+ for (; n < dev->hfidev.rdi.qp_dev->qp_table_size + iter->specials;
+ n++) {
if (pqp) {
qp = rcu_dereference(pqp->next);
} else {
@@ -695,7 +697,7 @@ int qp_iter_next(struct qp_iter *iter)
struct hfi1_ibport *ibp;
int pidx;
- pidx = n % dev->rdi.ibdev.phys_port_cnt;
+ pidx = n % dev->hfidev.rdi.ibdev.phys_port_cnt;
ppd = &dd_from_dev(dev)->pport[pidx];
ibp = &ppd->ibport_data;
@@ -705,7 +707,7 @@ int qp_iter_next(struct qp_iter *iter)
qp = rcu_dereference(ibp->rvp.qp[1]);
} else {
qp = rcu_dereference(
- dev->rdi.qp_dev->qp_table[
+ dev->hfidev.rdi.qp_dev->qp_table[
(n - iter->specials)]);
}
}
@@ -836,7 +838,7 @@ unsigned free_all_qps(struct rvt_dev_info *rdi)
{
struct hfi1_ibdev *verbs_dev = container_of(rdi,
struct hfi1_ibdev,
- rdi);
+ hfidev.rdi);
struct hfi1_devdata *dd = container_of(verbs_dev,
struct hfi1_devdata,
verbs_dev);
@@ -922,7 +924,7 @@ u32 mtu_from_qp(struct rvt_dev_info *rdi, struct rvt_qp *qp, u32 pmtu)
u32 mtu;
struct hfi1_ibdev *verbs_dev = container_of(rdi,
struct hfi1_ibdev,
- rdi);
+ hfidev.rdi);
struct hfi1_devdata *dd = container_of(verbs_dev,
struct hfi1_devdata,
verbs_dev);
@@ -945,7 +947,7 @@ int get_pmtu_from_attr(struct rvt_dev_info *rdi, struct rvt_qp *qp,
int mtu, pidx = qp->port_num - 1;
struct hfi1_ibdev *verbs_dev = container_of(rdi,
struct hfi1_ibdev,
- rdi);
+ hfidev.rdi);
struct hfi1_devdata *dd = container_of(verbs_dev,
struct hfi1_devdata,
verbs_dev);
@@ -1004,9 +1006,9 @@ void hfi1_error_port_qps(struct hfi1_ibport *ibp, u8 sl)
rcu_read_lock();
/* Deal only with RC/UC qps that use the given SL. */
- for (n = 0; n < dev->rdi.qp_dev->qp_table_size; n++) {
- for (qp = rcu_dereference(dev->rdi.qp_dev->qp_table[n]); qp;
- qp = rcu_dereference(qp->next)) {
+ for (n = 0; n < dev->hfidev.rdi.qp_dev->qp_table_size; n++) {
+ for (qp = rcu_dereference(dev->hfidev.rdi.qp_dev->qp_table[n]);
+ qp; qp = rcu_dereference(qp->next)) {
if (qp->port_num == ppd->port &&
(qp->ibqp.qp_type == IB_QPT_UC ||
qp->ibqp.qp_type == IB_QPT_RC) &&
diff --git a/drivers/infiniband/hw/hfi1/ruc.c b/drivers/infiniband/hw/hfi1/ruc.c
index 717ed4b15..175d352 100644
--- a/drivers/infiniband/hw/hfi1/ruc.c
+++ b/drivers/infiniband/hw/hfi1/ruc.c
@@ -103,7 +103,7 @@ static int init_sge(struct rvt_qp *qp, struct rvt_rwqe *wqe)
struct rvt_pd *pd;
struct rvt_sge_state *ss;
- rkt = &to_idev(qp->ibqp.device)->rdi.lkey_table;
+ rkt = &to_idev(qp->ibqp.device)->hfidev.rdi.lkey_table;
pd = ibpd_to_rvtpd(qp->ibqp.srq ? qp->ibqp.srq->pd : qp->ibqp.pd);
ss = &qp->r_sge;
ss->sg_list = qp->r_sg_list;
diff --git a/drivers/infiniband/hw/hfi1/sysfs.c b/drivers/infiniband/hw/hfi1/sysfs.c
index edba224..06bff50 100644
--- a/drivers/infiniband/hw/hfi1/sysfs.c
+++ b/drivers/infiniband/hw/hfi1/sysfs.c
@@ -498,7 +498,7 @@ static ssize_t show_rev(struct device *device, struct device_attribute *attr,
char *buf)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
return sprintf(buf, "%x\n", dd_from_dev(dev)->minrev);
}
@@ -507,7 +507,7 @@ static ssize_t show_hfi(struct device *device, struct device_attribute *attr,
char *buf)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
struct hfi1_devdata *dd = dd_from_dev(dev);
int ret;
@@ -522,7 +522,7 @@ static ssize_t show_boardversion(struct device *device,
struct device_attribute *attr, char *buf)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
struct hfi1_devdata *dd = dd_from_dev(dev);
/* The string printed here is already newline-terminated. */
@@ -533,7 +533,7 @@ static ssize_t show_nctxts(struct device *device,
struct device_attribute *attr, char *buf)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
struct hfi1_devdata *dd = dd_from_dev(dev);
/*
@@ -551,7 +551,7 @@ static ssize_t show_nfreectxts(struct device *device,
struct device_attribute *attr, char *buf)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
struct hfi1_devdata *dd = dd_from_dev(dev);
/* Return the number of free user ports (contexts) available. */
@@ -562,7 +562,7 @@ static ssize_t show_serial(struct device *device,
struct device_attribute *attr, char *buf)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
struct hfi1_devdata *dd = dd_from_dev(dev);
return scnprintf(buf, PAGE_SIZE, "%s", dd->serial);
@@ -573,7 +573,7 @@ static ssize_t store_chip_reset(struct device *device,
size_t count)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
struct hfi1_devdata *dd = dd_from_dev(dev);
int ret;
@@ -602,7 +602,7 @@ static ssize_t show_tempsense(struct device *device,
struct device_attribute *attr, char *buf)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
struct hfi1_devdata *dd = dd_from_dev(dev);
struct hfi1_temp temp;
int ret;
@@ -627,7 +627,7 @@ static ssize_t show_sdma_affinity(struct device *device,
struct device_attribute *attr, char *buf)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
struct hfi1_devdata *dd = dd_from_dev(dev);
return hfi1_get_sdma_affinity(dd, buf);
@@ -638,7 +638,7 @@ static ssize_t store_sdma_affinity(struct device *device,
const char *buf, size_t count)
{
struct hfi1_ibdev *dev =
- container_of(device, struct hfi1_ibdev, rdi.ibdev.dev);
+ container_of(device, struct hfi1_ibdev, hfidev.rdi.ibdev.dev);
struct hfi1_devdata *dd = dd_from_dev(dev);
return hfi1_set_sdma_affinity(dd, buf, count);
@@ -852,7 +852,7 @@ static SDE_ATTR(cpu_list, S_IWUSR | S_IRUGO,
*/
int hfi1_verbs_register_sysfs(struct hfi1_devdata *dd)
{
- struct ib_device *dev = &dd->verbs_dev.rdi.ibdev;
+ struct ib_device *dev = &dd->verbs_dev.hfidev.rdi.ibdev;
struct device *class_dev = &dev->dev;
int i, j, ret;
diff --git a/drivers/infiniband/hw/hfi1/verbs.c b/drivers/infiniband/hw/hfi1/verbs.c
index 95ed4d6..9355c4c 100644
--- a/drivers/infiniband/hw/hfi1/verbs.c
+++ b/drivers/infiniband/hw/hfi1/verbs.c
@@ -577,7 +577,7 @@ void hfi1_ib_rcv(struct hfi1_packet *packet)
u32 tlen = packet->tlen;
struct hfi1_pportdata *ppd = rcd->ppd;
struct hfi1_ibport *ibp = &ppd->ibport_data;
- struct rvt_dev_info *rdi = &ppd->dd->verbs_dev.rdi;
+ struct rvt_dev_info *rdi = &ppd->dd->verbs_dev.hfidev.rdi;
opcode_handler packet_handler;
unsigned long flags;
u32 qp_num;
@@ -1301,7 +1301,7 @@ int hfi1_verbs_send(struct rvt_qp *qp, struct hfi1_pkt_state *ps)
*/
static void hfi1_fill_device_attr(struct hfi1_devdata *dd)
{
- struct rvt_dev_info *rdi = &dd->verbs_dev.rdi;
+ struct rvt_dev_info *rdi = &dd->verbs_dev.hfidev.rdi;
u16 ver = dd->dc8051_ver;
memset(&rdi->dparms.props, 0, sizeof(rdi->dparms.props));
@@ -1747,9 +1747,10 @@ static int get_hw_stats(struct ib_device *ibdev, struct rdma_hw_stats *stats,
int hfi1_register_ib_device(struct hfi1_devdata *dd)
{
struct hfi1_ibdev *dev = &dd->verbs_dev;
- struct ib_device *ibdev = &dev->rdi.ibdev;
+ struct ib_device *ibdev = &dev->hfidev.rdi.ibdev;
struct hfi1_pportdata *ppd = dd->pport;
struct hfi1_ibport *ibp = &ppd->ibport_data;
+ struct rvt_dev_info *rdi = &dd->verbs_dev.hfidev.rdi;
unsigned i;
int ret;
size_t lcpysz = IB_DEVICE_NAME_MAX;
@@ -1799,77 +1800,77 @@ int hfi1_register_ib_device(struct hfi1_devdata *dd)
/*
* Fill in rvt info object.
*/
- dd->verbs_dev.rdi.driver_f.port_callback = hfi1_create_port_files;
- dd->verbs_dev.rdi.driver_f.get_card_name = get_card_name;
- dd->verbs_dev.rdi.driver_f.get_pci_dev = get_pci_dev;
- dd->verbs_dev.rdi.driver_f.check_ah = hfi1_check_ah;
- dd->verbs_dev.rdi.driver_f.notify_new_ah = hfi1_notify_new_ah;
- dd->verbs_dev.rdi.driver_f.get_guid_be = hfi1_get_guid_be;
- dd->verbs_dev.rdi.driver_f.query_port_state = query_port;
- dd->verbs_dev.rdi.driver_f.shut_down_port = shut_down_port;
- dd->verbs_dev.rdi.driver_f.cap_mask_chg = hfi1_cap_mask_chg;
+ rdi->driver_f.port_callback = hfi1_create_port_files;
+ rdi->driver_f.get_card_name = get_card_name;
+ rdi->driver_f.get_pci_dev = get_pci_dev;
+ rdi->driver_f.check_ah = hfi1_check_ah;
+ rdi->driver_f.notify_new_ah = hfi1_notify_new_ah;
+ rdi->driver_f.get_guid_be = hfi1_get_guid_be;
+ rdi->driver_f.query_port_state = query_port;
+ rdi->driver_f.shut_down_port = shut_down_port;
+ rdi->driver_f.cap_mask_chg = hfi1_cap_mask_chg;
/*
* Fill in rvt info device attributes.
*/
hfi1_fill_device_attr(dd);
/* queue pair */
- dd->verbs_dev.rdi.dparms.qp_table_size = hfi1_qp_table_size;
- dd->verbs_dev.rdi.dparms.qpn_start = 0;
- dd->verbs_dev.rdi.dparms.qpn_inc = 1;
- dd->verbs_dev.rdi.dparms.qos_shift = dd->qos_shift;
- dd->verbs_dev.rdi.dparms.qpn_res_start = kdeth_qp << 16;
- dd->verbs_dev.rdi.dparms.qpn_res_end =
- dd->verbs_dev.rdi.dparms.qpn_res_start + 65535;
- dd->verbs_dev.rdi.dparms.max_rdma_atomic = HFI1_MAX_RDMA_ATOMIC;
- dd->verbs_dev.rdi.dparms.psn_mask = PSN_MASK;
- dd->verbs_dev.rdi.dparms.psn_shift = PSN_SHIFT;
- dd->verbs_dev.rdi.dparms.psn_modify_mask = PSN_MODIFY_MASK;
- dd->verbs_dev.rdi.dparms.core_cap_flags = RDMA_CORE_PORT_INTEL_OPA;
- dd->verbs_dev.rdi.dparms.max_mad_size = OPA_MGMT_MAD_SIZE;
-
- dd->verbs_dev.rdi.driver_f.qp_priv_alloc = qp_priv_alloc;
- dd->verbs_dev.rdi.driver_f.qp_priv_free = qp_priv_free;
- dd->verbs_dev.rdi.driver_f.free_all_qps = free_all_qps;
- dd->verbs_dev.rdi.driver_f.notify_qp_reset = notify_qp_reset;
- dd->verbs_dev.rdi.driver_f.do_send = hfi1_do_send;
- dd->verbs_dev.rdi.driver_f.schedule_send = hfi1_schedule_send;
- dd->verbs_dev.rdi.driver_f.schedule_send_no_lock = _hfi1_schedule_send;
- dd->verbs_dev.rdi.driver_f.get_pmtu_from_attr = get_pmtu_from_attr;
- dd->verbs_dev.rdi.driver_f.notify_error_qp = notify_error_qp;
- dd->verbs_dev.rdi.driver_f.flush_qp_waiters = flush_qp_waiters;
- dd->verbs_dev.rdi.driver_f.stop_send_queue = stop_send_queue;
- dd->verbs_dev.rdi.driver_f.quiesce_qp = quiesce_qp;
- dd->verbs_dev.rdi.driver_f.notify_error_qp = notify_error_qp;
- dd->verbs_dev.rdi.driver_f.mtu_from_qp = mtu_from_qp;
- dd->verbs_dev.rdi.driver_f.mtu_to_path_mtu = mtu_to_path_mtu;
- dd->verbs_dev.rdi.driver_f.check_modify_qp = hfi1_check_modify_qp;
- dd->verbs_dev.rdi.driver_f.modify_qp = hfi1_modify_qp;
- dd->verbs_dev.rdi.driver_f.check_send_wqe = hfi1_check_send_wqe;
+ rdi->dparms.qp_table_size = hfi1_qp_table_size;
+ rdi->dparms.qpn_start = 0;
+ rdi->dparms.qpn_inc = 1;
+ rdi->dparms.qos_shift = dd->qos_shift;
+ rdi->dparms.qpn_res_start = kdeth_qp << 16;
+ rdi->dparms.qpn_res_end =
+ rdi->dparms.qpn_res_start + 65535;
+ rdi->dparms.max_rdma_atomic = HFI1_MAX_RDMA_ATOMIC;
+ rdi->dparms.psn_mask = PSN_MASK;
+ rdi->dparms.psn_shift = PSN_SHIFT;
+ rdi->dparms.psn_modify_mask = PSN_MODIFY_MASK;
+ rdi->dparms.core_cap_flags = RDMA_CORE_PORT_INTEL_OPA;
+ rdi->dparms.max_mad_size = OPA_MGMT_MAD_SIZE;
+
+ rdi->driver_f.qp_priv_alloc = qp_priv_alloc;
+ rdi->driver_f.qp_priv_free = qp_priv_free;
+ rdi->driver_f.free_all_qps = free_all_qps;
+ rdi->driver_f.notify_qp_reset = notify_qp_reset;
+ rdi->driver_f.do_send = hfi1_do_send;
+ rdi->driver_f.schedule_send = hfi1_schedule_send;
+ rdi->driver_f.schedule_send_no_lock = _hfi1_schedule_send;
+ rdi->driver_f.get_pmtu_from_attr = get_pmtu_from_attr;
+ rdi->driver_f.notify_error_qp = notify_error_qp;
+ rdi->driver_f.flush_qp_waiters = flush_qp_waiters;
+ rdi->driver_f.stop_send_queue = stop_send_queue;
+ rdi->driver_f.quiesce_qp = quiesce_qp;
+ rdi->driver_f.notify_error_qp = notify_error_qp;
+ rdi->driver_f.mtu_from_qp = mtu_from_qp;
+ rdi->driver_f.mtu_to_path_mtu = mtu_to_path_mtu;
+ rdi->driver_f.check_modify_qp = hfi1_check_modify_qp;
+ rdi->driver_f.modify_qp = hfi1_modify_qp;
+ rdi->driver_f.check_send_wqe = hfi1_check_send_wqe;
/* completeion queue */
- snprintf(dd->verbs_dev.rdi.dparms.cq_name,
- sizeof(dd->verbs_dev.rdi.dparms.cq_name),
+ snprintf(rdi->dparms.cq_name,
+ sizeof(rdi->dparms.cq_name),
"hfi1_cq%d", dd->unit);
- dd->verbs_dev.rdi.dparms.node = dd->node;
+ rdi->dparms.node = dd->node;
/* misc settings */
- dd->verbs_dev.rdi.flags = 0; /* Let rdmavt handle it all */
- dd->verbs_dev.rdi.dparms.lkey_table_size = hfi1_lkey_table_size;
- dd->verbs_dev.rdi.dparms.nports = dd->num_pports;
- dd->verbs_dev.rdi.dparms.npkeys = hfi1_get_npkeys(dd);
+ rdi->flags = 0; /* Let rdmavt handle it all */
+ rdi->dparms.lkey_table_size = hfi1_lkey_table_size;
+ rdi->dparms.nports = dd->num_pports;
+ rdi->dparms.npkeys = hfi1_get_npkeys(dd);
/* post send table */
- dd->verbs_dev.rdi.post_parms = hfi1_post_parms;
+ rdi->post_parms = hfi1_post_parms;
ppd = dd->pport;
for (i = 0; i < dd->num_pports; i++, ppd++)
- rvt_init_port(&dd->verbs_dev.rdi,
+ rvt_init_port(rdi,
&ppd->ibport_data.rvp,
i,
ppd->pkeys);
- ret = rvt_register_device(&dd->verbs_dev.rdi);
+ ret = rvt_register_device(rdi);
if (ret)
goto err_verbs_txreq;
@@ -1880,7 +1881,7 @@ int hfi1_register_ib_device(struct hfi1_devdata *dd)
return ret;
err_class:
- rvt_unregister_device(&dd->verbs_dev.rdi);
+ rvt_unregister_device(rdi);
err_verbs_txreq:
verbs_txreq_exit(dev);
dd_dev_err(dd, "cannot register verbs: %d!\n", -ret);
@@ -1893,7 +1894,7 @@ void hfi1_unregister_ib_device(struct hfi1_devdata *dd)
hfi1_verbs_unregister_sysfs(dd);
- rvt_unregister_device(&dd->verbs_dev.rdi);
+ rvt_unregister_device(&dd->verbs_dev.hfidev.rdi);
if (!list_empty(&dev->txwait))
dd_dev_err(dd, "txwait list not empty!\n");
diff --git a/drivers/infiniband/hw/hfi1/verbs.h b/drivers/infiniband/hw/hfi1/verbs.h
index e6b8930..72412ed 100644
--- a/drivers/infiniband/hw/hfi1/verbs.h
+++ b/drivers/infiniband/hw/hfi1/verbs.h
@@ -64,6 +64,7 @@
#include <rdma/rdma_vt.h>
#include <rdma/rdmavt_qp.h>
#include <rdma/rdmavt_cq.h>
+#include <rdma/opa_hfi.h>
struct hfi1_ctxtdata;
struct hfi1_pportdata;
@@ -174,7 +175,7 @@ struct hfi1_ibport {
};
struct hfi1_ibdev {
- struct rvt_dev_info rdi; /* Must be first */
+ struct hfi_ibdev hfidev; /* Must be first */
/* QP numbers are shared by all IB ports */
/* protect txwait list */
@@ -201,10 +202,10 @@ struct hfi1_ibdev {
static inline struct hfi1_ibdev *to_idev(struct ib_device *ibdev)
{
- struct rvt_dev_info *rdi;
+ struct hfi_ibdev *hfidev;
- rdi = container_of(ibdev, struct rvt_dev_info, ibdev);
- return container_of(rdi, struct hfi1_ibdev, rdi);
+ hfidev = to_hfi_ibdev(ibdev);
+ return container_of(hfidev, struct hfi1_ibdev, hfidev);
}
static inline struct rvt_qp *iowait_to_qp(struct iowait *s_iowait)
diff --git a/include/rdma/opa_hfi.h b/include/rdma/opa_hfi.h
new file mode 100644
index 0000000..ef61d65
--- /dev/null
+++ b/include/rdma/opa_hfi.h
@@ -0,0 +1,199 @@
+#ifndef _OPA_HFI_H
+#define _OPA_HFI_H
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains Intel Omni-Path (OPA) Host Fabric Interface (HFI)
+ * specific declarations.
+ */
+
+#include <linux/etherdevice.h>
+#include <linux/module.h>
+#include <rdma/ib_verbs.h>
+#include <rdma/rdma_vt.h>
+
+/* Maximum possible number of VNICs */
+#define HFI_MAX_NUM_VNICS 255
+
+#define HFI_VNIC_MAX_QUEUE 16
+
+#define HFI_VNIC_CAP_SG BIT(0)
+
+enum {
+ /* Packet received on queue 0 */
+ HFI_VNIC_EVT_RX0,
+ /* Tx wakeup notification on queue 0 */
+ HFI_VNIC_EVT_TX0
+ = HFI_VNIC_EVT_RX0 + HFI_VNIC_MAX_QUEUE,
+ HFI_VNIC_NUM_EVTS
+ = HFI_VNIC_EVT_TX0 + HFI_VNIC_MAX_QUEUE,
+};
+
+struct hfi_vnic_port;
+
+typedef void (*hfi_vnic_evt_cb_fn)(struct hfi_vnic_port *vport, u8 evt);
+
+/**
+ * struct hfi_vnic_ops - HFI HW specific VNIC functions
+ * @open: Open the vnic port
+ * @close: Close the vnic port
+ * @put_skb: transmit an skb
+ * @get_skb: receive an skb
+ * @get_read_avail: return number of available to read
+ * @get_write_avail: return whether write space is available or not
+ * @select_queue: select tx queue
+ * @config_notify: enable/disable notification
+ */
+struct hfi_vnic_ops {
+ int (*open)(struct hfi_vnic_port *vport,
+ hfi_vnic_evt_cb_fn cb);
+ void (*close)(struct hfi_vnic_port *vport);
+ int (*put_skb)(struct hfi_vnic_port *vport,
+ u8 q_idx, struct sk_buff *skb);
+ struct sk_buff *(*get_skb)(struct hfi_vnic_port *vport, u8 q_idx);
+ u16 (*get_read_avail)(struct hfi_vnic_port *vport, u8 q_idx);
+ bool (*get_write_avail)(struct hfi_vnic_port *vport, u8 q_idx);
+ u8 (*select_queue)(struct hfi_vnic_port *vport, u8 vl, u8 entropy);
+ void (*config_notify)(struct hfi_vnic_port *vport,
+ u8 evt, bool enable);
+};
+
+/**
+ * struct hfi_vnic_stats - HFI HW specific statistics
+ * @rx_fifo_errors: receive packets dropped due to fifo full
+ * @tx_fifo_errors: transmit packets dropped due to fifo full
+ * @rx_missed_errors: receive packets missed due to no memory
+ * @tx_carrier_errors: packet transmits when STL link is down
+ * @rx_bad_veswid: receive packets with invalid vesw id
+ * @rx_logic_errors: receive packets dropped due to other errors
+ * @tx_logic_errors: transmit packets dropped due to other errors
+ *
+ * This structure holds any statistics information that is
+ * collected by HW specific driver layer.
+ */
+struct hfi_vnic_stats {
+ u64 rx_fifo_errors;
+ u64 tx_fifo_errors;
+ u64 rx_missed_errors;
+ u64 tx_carrier_errors;
+ u64 rx_bad_veswid;
+ u64 rx_logic_errors;
+ u64 tx_logic_errors;
+};
+
+/**
+ * struct hfi_vnic_info - HFI HW specific VNIC information
+ * @cap: capabilities
+ * @num_rx_q: number of receive queues
+ * @num_tx_q: number of transmit queues
+ */
+struct hfi_vnic_info {
+ u32 cap;
+ u8 num_rx_q;
+ u8 num_tx_q;
+};
+
+/**
+ * struct hfi_vnic_port - HFI virtual NIC port
+ * @vesw_id: virtual ethernet switch id
+ * @netdev: pointer to associated netdev
+ * @port_num: hfi port instance
+ * @vport_num: vnic port instance on the hfi port
+ * @ops: hfi vnic operations
+ * @hfi_priv: hfi private data pointer
+ * @hfi_info: hfi information
+ * @hfi_stats: per queue hfi statistics
+ */
+struct hfi_vnic_port {
+ u16 vesw_id;
+ struct net_device *netdev;
+ u8 port_num;
+ u8 vport_num;
+
+ struct hfi_vnic_ops *ops;
+ void *hfi_priv;
+ struct hfi_vnic_info hfi_info;
+ struct hfi_vnic_stats hfi_stats[HFI_VNIC_MAX_QUEUE];
+};
+
+/**
+ * struct hfi_vnic_ctrl_ops - HFI HW specific VNIC control functions
+ * @add_vport: add a vnic port
+ * @rem_vport: remove a vnic port
+ */
+struct hfi_vnic_ctrl_ops {
+ struct hfi_vnic_port *(*add_vport)(struct ib_device *device,
+ u8 port_num, u8 vport_num);
+ void (*rem_vport)(struct hfi_vnic_port *vport);
+};
+
+/**
+ * struct hfi_ibdev - HFI extension of rdmavt device
+ * @rdi: rdmavt device interface
+ * @vnic_ctrl_ops: VNIC control operations
+ */
+struct hfi_ibdev {
+ struct rvt_dev_info rdi; /* Must be first */
+
+ struct hfi_vnic_ctrl_ops vnic_ctrl_ops;
+};
+
+static inline struct hfi_ibdev *to_hfi_ibdev(struct ib_device *ibdev)
+{
+ struct rvt_dev_info *rdi;
+
+ rdi = container_of(ibdev, struct rvt_dev_info, ibdev);
+ return container_of(rdi, struct hfi_ibdev, rdi);
+}
+
+static inline bool is_hfi_ibdev(struct ib_device *ibdev)
+{
+ return !memcmp(ibdev->name, "hfi", 3);
+}
+
+#endif /* _OPA_HFI_H */
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 41+ messages in thread
* [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
[not found] ` <1481788782-89964-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2016-12-15 7:59 ` [RFC v2 02/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface Vishwanathapura, Niranjana
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
[not found] ` <1481788782-89964-4-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2016-12-15 7:59 ` [RFC v2 04/10] IB/hfi-vnic: VNIC Ethernet Management (EM) structure definitions Vishwanathapura, Niranjana
` (3 subsequent siblings)
5 siblings, 1 reply; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Niranjana Vishwanathapura,
Sadanand Warrier, Sudeep Dutt, Tanya K Jajodia,
Andrzej Kacprowski
HFI VNIC netdev function supports Ethernet functionality over Omni-Path
fabric by encapsulating Ethernet packets inside Omni-Path packet header.
It interfaces with the network stack to provide standard Ethernet network
interfaces. It invokes HFI device's VNIC callback functions for HW access.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Sadanand Warrier <sadanand.warrier-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Sudeep Dutt <sudeep.dutt-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Tanya K Jajodia <tanya.k.jajodia-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Andrzej Kacprowski <andrzej.kacprowski-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
MAINTAINERS | 7 +
drivers/infiniband/Kconfig | 1 +
drivers/infiniband/sw/Makefile | 1 +
drivers/infiniband/sw/intel/hfi_vnic/Kconfig | 8 +
drivers/infiniband/sw/intel/hfi_vnic/Makefile | 6 +
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c | 238 ++++++++++++
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h | 62 ++++
.../sw/intel/hfi_vnic/hfi_vnic_ethtool.c | 65 ++++
.../sw/intel/hfi_vnic/hfi_vnic_internal.h | 220 +++++++++++
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c | 409 +++++++++++++++++++++
10 files changed, 1017 insertions(+)
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/Kconfig
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/Makefile
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
diff --git a/MAINTAINERS b/MAINTAINERS
index 2c7a7b6..62db3ea 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -5628,6 +5628,13 @@ F: drivers/block/cciss*
F: include/linux/cciss_ioctl.h
F: include/uapi/linux/cciss_ioctl.h
+HFI-VNIC DRIVER
+M: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
+M: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
+L: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
+S: Supported
+F: drivers/infiniband/sw/intel/hfi_vnic
+
HFI1 DRIVER
M: Mike Marciniszyn <mike.marciniszyn-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
M: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
diff --git a/drivers/infiniband/Kconfig b/drivers/infiniband/Kconfig
index 6709173..900daf3 100644
--- a/drivers/infiniband/Kconfig
+++ b/drivers/infiniband/Kconfig
@@ -85,6 +85,7 @@ source "drivers/infiniband/ulp/srpt/Kconfig"
source "drivers/infiniband/ulp/iser/Kconfig"
source "drivers/infiniband/ulp/isert/Kconfig"
+source "drivers/infiniband/sw/intel/hfi_vnic/Kconfig"
source "drivers/infiniband/sw/rdmavt/Kconfig"
source "drivers/infiniband/sw/rxe/Kconfig"
diff --git a/drivers/infiniband/sw/Makefile b/drivers/infiniband/sw/Makefile
index 8b095b2..2792559 100644
--- a/drivers/infiniband/sw/Makefile
+++ b/drivers/infiniband/sw/Makefile
@@ -1,2 +1,3 @@
obj-$(CONFIG_INFINIBAND_RDMAVT) += rdmavt/
obj-$(CONFIG_RDMA_RXE) += rxe/
+obj-$(CONFIG_HFI_VNIC) += intel/hfi_vnic/
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/Kconfig b/drivers/infiniband/sw/intel/hfi_vnic/Kconfig
new file mode 100644
index 0000000..84d13e7
--- /dev/null
+++ b/drivers/infiniband/sw/intel/hfi_vnic/Kconfig
@@ -0,0 +1,8 @@
+config HFI_VNIC
+ tristate "Intel HFI VNIC support"
+ depends on X86_64 && INFINIBAND
+ ---help---
+ This is HFI Virtual Network Interface Controller (VNIC) driver
+ for Ethernet over HFI feature. It implements the HW independent
+ VNIC functionality. It interfaces with Linux stack for data path
+ and IB MAD for the control path.
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/Makefile b/drivers/infiniband/sw/intel/hfi_vnic/Makefile
new file mode 100644
index 0000000..8e3dca7
--- /dev/null
+++ b/drivers/infiniband/sw/intel/hfi_vnic/Makefile
@@ -0,0 +1,6 @@
+# Makefile - Intel HFI Virtual Network Controller driver
+# Copyright(c) 2016, Intel Corporation.
+#
+obj-$(CONFIG_HFI_VNIC) += hfi_vnic.o
+
+hfi_vnic-y := hfi_vnic_netdev.o hfi_vnic_encap.o hfi_vnic_ethtool.o
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
new file mode 100644
index 0000000..093df67
--- /dev/null
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
@@ -0,0 +1,238 @@
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains HFI VNIC encapsulation/decapsulation function.
+ */
+
+#include <linux/if_ether.h>
+#include <linux/if_vlan.h>
+
+#include "hfi_vnic_internal.h"
+
+/**
+ * union hfi_vnic_bypass_hdr - VNIC bypass header
+ * @slid: source lid
+ * @length: length of packet
+ * @becn: backward explicit congestion notification
+ * @dlid: destination lid
+ * @sc: service class
+ * @fecn: forward explicit congestion notification
+ * @l2: L2 type (2=16B)
+ * @lt: link transfer field
+ * @l4: L4 type
+ * @slid_high: upper 4 bits of source lid
+ * @dlid_high: upper 4 bits of destination lid
+ * @pkey: partition key
+ * @entropy: entropy
+ * @age: packet age
+ * @l4_hdr: L4 header
+ */
+union hfi_vnic_bypass_hdr {
+ struct {
+ struct {
+ uint64_t slid : 20;
+ uint64_t length : 11;
+ uint64_t becn : 1;
+ uint64_t dlid : 20;
+ uint64_t sc : 5;
+ uint64_t rsvd : 3;
+ uint64_t fecn : 1;
+ uint64_t l2 : 2;
+ uint64_t lt : 1;
+ };
+ struct {
+ uint64_t l4 : 8;
+ uint64_t slid_high : 4;
+ uint64_t dlid_high : 4;
+ uint64_t pkey : 16;
+ uint64_t entropy : 16;
+ uint64_t age : 8;
+ uint64_t rsvd1 : 8;
+ };
+ struct {
+ uint32_t rsvd2 : 16;
+ uint32_t l4_hdr : 16;
+ };
+ } __packed;
+ u32 dw[5];
+};
+
+#define HFI_VNIC_SC_MASK 0x1f
+
+/* hfi_vnic_get_dlid - find and return the DLID */
+static uint32_t hfi_vnic_get_dlid(struct hfi_vnic_adapter *adapter,
+ struct sk_buff *skb, u8 def_port)
+{
+ struct __hfi_veswport_info *info = &adapter->info;
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
+ u32 dlid;
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest)) {
+ dlid = info->vesw.u_mcast_dlid;
+ } else {
+ if (is_local_ether_addr(mac_hdr->h_dest)) {
+ dlid = ((uint32_t)mac_hdr->h_dest[5] << 16) |
+ ((uint32_t)mac_hdr->h_dest[4] << 8) |
+ mac_hdr->h_dest[3];
+ if (unlikely(!dlid))
+ v_warn("Null dlid in MAC address\n");
+ } else if (def_port != HFI_VNIC_INVALID_PORT) {
+ dlid = info->vesw.u_ucast_dlid[def_port];
+ }
+ }
+
+ return dlid;
+}
+
+/* hfi_vnic_get_sc - return the service class */
+static u8 hfi_vnic_get_sc(struct __hfi_veswport_info *info,
+ struct sk_buff *skb)
+{
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
+ u16 vlan_tci;
+ u8 sc;
+
+ if (!__vlan_get_tag(skb, &vlan_tci)) {
+ u8 pcp = HFI_VNIC_VLAN_PCP(vlan_tci);
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ sc = info->vport.pcp_to_sc_mc[pcp];
+ else
+ sc = info->vport.pcp_to_sc_uc[pcp];
+ } else {
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ sc = info->vport.non_vlan_sc_mc;
+ else
+ sc = info->vport.non_vlan_sc_uc;
+ }
+
+ return sc & HFI_VNIC_SC_MASK;
+}
+
+/* hfi_vnic_calc_entropy - calculate the packet entropy */
+u8 hfi_vnic_calc_entropy(struct hfi_vnic_adapter *adapter, struct sk_buff *skb)
+{
+ u16 hash16;
+
+ /*
+ * Get flow based 16-bit hash and then XOR the upper and lower bytes
+ * to get the entropy.
+ * __skb_tx_hash limits qcount to 16 bits. Hence, get 15-bit hash.
+ */
+ hash16 = __skb_tx_hash(adapter->netdev, skb, BIT(15));
+ return (u8)((hash16 >> 8) ^ (hash16 & 0xff));
+}
+
+/* hfi_vnic_get_def_port - get default port based on entropy */
+static inline u8 hfi_vnic_get_def_port(struct hfi_vnic_adapter *adapter,
+ u8 entropy)
+{
+ u8 flow_id;
+
+ /* Add the upper and lower 4-bits of entropy to get the flow id */
+ flow_id = ((entropy & 0xf) + (entropy >> 4));
+ return adapter->flow_tbl[flow_id & (HFI_VNIC_FLOW_TBL_SIZE - 1)];
+}
+
+/* Calculate packet length including OPA header, crc and padding */
+static inline int hfi_vnic_wire_length(struct sk_buff *skb)
+{
+ u32 pad_len, hlen = HFI_VNIC_HDR_LEN;
+
+ /* padding for 8 bytes size alignment */
+ pad_len = -(skb->len + hlen + HFI_VNIC_ICRC_TAIL_LEN) & 0x7;
+ pad_len += HFI_VNIC_ICRC_TAIL_LEN;
+
+ return (skb->len + hlen + pad_len) >> 3;
+}
+
+/* hfi_vnic_encap_skb - encapsulate skb (ethernet) packet with OPA header */
+int hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb)
+{
+ struct __hfi_veswport_info *info = &adapter->info;
+ union hfi_vnic_bypass_hdr *hdr;
+ u32 dlid;
+ u8 def_port;
+
+ hdr = (union hfi_vnic_bypass_hdr *)(skb->data - HFI_VNIC_HDR_LEN);
+ memset(hdr, 0, HFI_VNIC_HDR_LEN);
+
+ hdr->entropy = hfi_vnic_calc_entropy(adapter, skb);
+ def_port = hfi_vnic_get_def_port(adapter, hdr->entropy);
+
+ hdr->slid = info->vport.encap_slid;
+ hdr->slid_high = info->vport.encap_slid >> 20;
+
+ dlid = hfi_vnic_get_dlid(adapter, skb, def_port);
+ if (unlikely(!dlid))
+ return -EFAULT;
+
+ hdr->dlid = dlid;
+ hdr->dlid_high = dlid >> 20;
+
+ hdr->length = hfi_vnic_wire_length(skb);
+ hdr->sc = hfi_vnic_get_sc(info, skb);
+
+ hdr->l2 = HFI_VNIC_L2_TYPE;
+ hdr->lt = 1;
+
+ hdr->pkey = info->vesw.pkey;
+
+ hdr->l4 = HFI_VNIC_L4_ETHR;
+ hdr->l4_hdr = info->vesw.vesw_id;
+
+ skb_push(skb, HFI_VNIC_HDR_LEN);
+ return 0;
+}
+
+/* hfi_vnic_decap_skb - strip OPA header from the skb (ethernet) packet */
+int hfi_vnic_decap_skb(struct hfi_vnic_rx_queue *rxq, struct sk_buff *skb)
+{
+ skb_pull(skb, HFI_VNIC_HDR_LEN);
+ return 0;
+}
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
new file mode 100644
index 0000000..6786cce
--- /dev/null
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
@@ -0,0 +1,62 @@
+#ifndef _HFI_VNIC_ENCAP_H
+#define _HFI_VNIC_ENCAP_H
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains all HFI VNIC declaration required for encapsulation
+ * and decapsulation of Ethernet packets
+ */
+
+#define HFI_VESW_MAX_NUM_DEF_PORT 16
+#define HFI_VNIC_MAX_NUM_PCP 8
+
+/* VNIC configured and operational state values */
+#define HFI_VNIC_STATE_DROP_ALL 0x1
+#define HFI_VNIC_STATE_FORWARDING 0x3
+
+#endif /* _HFI_VNIC_ENCAP_H */
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
new file mode 100644
index 0000000..0b4da5e
--- /dev/null
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
@@ -0,0 +1,65 @@
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains HFI VNIC ethtool functions
+ */
+
+#include <linux/ethtool.h>
+
+#include "hfi_vnic_internal.h"
+
+/* ethtool ops */
+static const struct ethtool_ops hfi_vnic_ethtool_ops = {
+ .get_link = ethtool_op_get_link,
+};
+
+/* hfi_vnic_set_ethtool_ops - set ethtool ops */
+void hfi_vnic_set_ethtool_ops(struct net_device *netdev)
+{
+ netdev->ethtool_ops = &hfi_vnic_ethtool_ops;
+}
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
new file mode 100644
index 0000000..30731b4
--- /dev/null
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
@@ -0,0 +1,220 @@
+#ifndef _HFI_VNIC_INTERNAL_H
+#define _HFI_VNIC_INTERNAL_H
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains HFI VNIC driver internal declarations
+ */
+
+#include <linux/bitops.h>
+#include <linux/etherdevice.h>
+#include <linux/hashtable.h>
+#include <linux/sizes.h>
+#include <rdma/opa_hfi.h>
+
+#include "hfi_vnic_encap.h"
+
+/* VNIC uses 16B header format */
+#define HFI_VNIC_L2_TYPE 0x2
+
+/* 16 header bytes + 2 reserved bytes */
+#define HFI_VNIC_L2_HDR_LEN (16 + 2)
+
+#define HFI_VNIC_L4_HDR_LEN 2
+
+#define HFI_VNIC_HDR_LEN (HFI_VNIC_L2_HDR_LEN + \
+ HFI_VNIC_L4_HDR_LEN)
+
+#define HFI_VNIC_L4_ETHR 0x78
+
+#define HFI_VNIC_ICRC_LEN 4
+#define HFI_VNIC_TAIL_LEN 1
+#define HFI_VNIC_ICRC_TAIL_LEN (HFI_VNIC_ICRC_LEN + HFI_VNIC_TAIL_LEN)
+
+#define HFI_VNIC_VLAN_PCP(vlan_tci) \
+ (((vlan_tci) & VLAN_PRIO_MASK) >> VLAN_PRIO_SHIFT)
+
+#define HFI_VNIC_SKB_HEADROOM ALIGN(HFI_VNIC_HDR_LEN, 8)
+
+/* Flow to default port redirection table size */
+#define HFI_VNIC_FLOW_TBL_SIZE 32
+
+/* Invalid port number */
+#define HFI_VNIC_INVALID_PORT 0xff
+
+enum hfi_vnic_flags_t {
+ HFI_VNIC_UP,
+ HFI_VNIC_OPEN,
+};
+
+struct hfi_vnic_adapter;
+
+/**
+ * struct __hfi_vesw_info - HFI vnic virtual switch info
+ */
+struct __hfi_vesw_info {
+ u16 fabric_id;
+ u16 vesw_id;
+
+ u8 rsvd0[6];
+ u16 def_port_mask;
+
+ u8 rsvd1[2];
+ u16 pkey;
+
+ u8 rsvd2[4];
+ u32 u_mcast_dlid;
+ u32 u_ucast_dlid[HFI_VESW_MAX_NUM_DEF_PORT];
+
+ u8 rsvd3[44];
+ u16 eth_mtu[HFI_VNIC_MAX_NUM_PCP];
+ u16 eth_mtu_non_vlan;
+ u8 rsvd4[2];
+} __packed;
+
+/**
+ * struct __hfi_per_veswport_info - HFI vnic per port info
+ */
+struct __hfi_per_veswport_info {
+ u32 port_num;
+
+ u8 eth_link_status;
+ u8 rsvd0[3];
+
+ u8 base_mac_addr[ETH_ALEN];
+ u8 config_state;
+ u8 oper_state;
+
+ u16 max_mac_tbl_ent;
+ u16 max_smac_ent;
+ u32 mac_tbl_digest;
+ u8 rsvd1[4];
+
+ u32 encap_slid;
+
+ u8 pcp_to_sc_uc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_vl_uc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_sc_mc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_vl_mc[HFI_VNIC_MAX_NUM_PCP];
+
+ u8 non_vlan_sc_uc;
+ u8 non_vlan_vl_uc;
+ u8 non_vlan_sc_mc;
+ u8 non_vlan_vl_mc;
+
+ u8 rsvd2[48];
+
+ u16 uc_macs_gen_count;
+ u16 mc_macs_gen_count;
+
+ u8 rsvd3[8];
+} __packed;
+
+/**
+ * struct __hfi_veswport_info - HFI vnic port info
+ */
+struct __hfi_veswport_info {
+ struct __hfi_vesw_info vesw;
+ struct __hfi_per_veswport_info vport;
+};
+
+/**
+ * struct hfi_vnic_rx_queue - HFI VNIC receive queue
+ * @idx: queue index
+ * @adapter: netdev adapter
+ * @napi: netdev napi structure
+ */
+struct hfi_vnic_rx_queue {
+ u8 idx;
+ struct hfi_vnic_adapter *adapter;
+ struct napi_struct napi;
+};
+
+/**
+ * struct hfi_vnic_adapter - HFI VNIC netdev private data structure
+ * @netdev: pointer to associated netdev
+ * @vport: pointer to hfi vnic port
+ * @flags: flags indicating various states
+ * @lock: adapter lock
+ * @rxq: receive queue array
+ * @info: virtual ethernet switch port information
+ * @flow_tbl: flow to default port redirection table
+ */
+struct hfi_vnic_adapter {
+ struct net_device *netdev;
+ struct hfi_vnic_port *vport;
+ unsigned long flags;
+
+ /* Lock used around state updates */
+ struct mutex lock;
+
+ struct hfi_vnic_rx_queue rxq[HFI_VNIC_MAX_QUEUE];
+
+ struct __hfi_veswport_info info;
+
+ u8 flow_tbl[HFI_VNIC_FLOW_TBL_SIZE];
+};
+
+#define v_dbg(format, arg...) \
+ netdev_dbg(adapter->netdev, format, ## arg)
+#define v_err(format, arg...) \
+ netdev_err(adapter->netdev, format, ## arg)
+#define v_info(format, arg...) \
+ netdev_info(adapter->netdev, format, ## arg)
+#define v_warn(format, arg...) \
+ netdev_warn(adapter->netdev, format, ## arg)
+
+struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
+ struct device *parent);
+void hfi_vnic_rem_netdev(struct hfi_vnic_port *vport);
+int hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+int hfi_vnic_decap_skb(struct hfi_vnic_rx_queue *rxq, struct sk_buff *skb);
+u8 hfi_vnic_calc_entropy(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+void hfi_vnic_set_ethtool_ops(struct net_device *netdev);
+
+#endif /* _HFI_VNIC_INTERNAL_H */
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
new file mode 100644
index 0000000..6360d37
--- /dev/null
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
@@ -0,0 +1,409 @@
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains HFI Virtual Network Interface Controller (VNIC) driver
+ */
+
+#include <linux/module.h>
+#include <linux/if_vlan.h>
+
+#include "hfi_vnic_internal.h"
+
+#define HFI_TX_TIMEOUT_MS 1000
+
+#define HFI_VNIC_MIN_ETH_MTU (ETH_ZLEN - ETH_HLEN)
+
+/* hfi_vnic_maybe_stop_tx - stop tx queue if required */
+static void hfi_vnic_maybe_stop_tx(struct hfi_vnic_adapter *adapter, u8 q_idx)
+{
+ struct hfi_vnic_port *vport = adapter->vport;
+
+ netif_stop_subqueue(vport->netdev, q_idx);
+ if (!vport->ops->get_write_avail(vport, q_idx))
+ return;
+
+ netif_start_subqueue(vport->netdev, q_idx);
+}
+
+/* hfi_netdev_start_xmit - transmit function */
+static netdev_tx_t hfi_netdev_start_xmit(struct sk_buff *skb,
+ struct net_device *netdev)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ struct hfi_vnic_port *vport = adapter->vport;
+ u8 q_idx = skb->queue_mapping;
+ bool skip_skb_free = false;
+ int rc = -1;
+
+ v_dbg("xmit: queue %d skb len %d\n", q_idx, skb->len);
+ if (unlikely(adapter->info.vport.oper_state !=
+ HFI_VNIC_STATE_FORWARDING))
+ goto tx_finish;
+
+ /* pad to ensure mininum ethernet packet length */
+ if (unlikely(skb->len < ETH_ZLEN)) {
+ if (skb_padto(skb, ETH_ZLEN)) {
+ skip_skb_free = true;
+ goto tx_finish;
+ }
+ skb_put(skb, ETH_ZLEN - skb->len);
+ }
+
+ rc = hfi_vnic_encap_skb(adapter, skb);
+ if (unlikely(rc))
+ goto tx_finish;
+
+ /* Get reference to skb as hfi driver might release it */
+ skb_get(skb);
+ rc = vport->ops->put_skb(vport, q_idx, skb);
+ /* remove the header */
+ skb_pull(skb, HFI_VNIC_HDR_LEN);
+
+tx_finish:
+ if (unlikely(rc == -EBUSY)) {
+ hfi_vnic_maybe_stop_tx(adapter, q_idx);
+ dev_kfree_skb_any(skb);
+ return NETDEV_TX_BUSY;
+ }
+
+ if (!skip_skb_free)
+ dev_kfree_skb_any(skb);
+ return NETDEV_TX_OK;
+}
+
+/* vnic_handle_rx - handle skb receive */
+static void vnic_handle_rx(struct hfi_vnic_rx_queue *rxq,
+ int *work_done, int work_to_do)
+{
+ struct hfi_vnic_adapter *adapter = rxq->adapter;
+ struct hfi_vnic_port *vport = adapter->vport;
+ struct sk_buff *skb;
+
+ while (1) {
+ if (*work_done >= work_to_do)
+ break;
+
+ skb = vport->ops->get_skb(vport, rxq->idx);
+ if (!skb)
+ break;
+
+ if (hfi_vnic_decap_skb(rxq, skb)) {
+ dev_kfree_skb_any(skb);
+ continue;
+ }
+
+ skb_checksum_none_assert(skb);
+ skb->protocol = eth_type_trans(skb, vport->netdev);
+
+ napi_gro_receive(&rxq->napi, skb);
+ (*work_done)++;
+ }
+}
+
+/* vnic_napi - napi receive polling callback function */
+static int vnic_napi(struct napi_struct *napi, int budget)
+{
+ struct hfi_vnic_rx_queue *rxq = container_of(napi,
+ struct hfi_vnic_rx_queue, napi);
+ struct hfi_vnic_adapter *adapter = rxq->adapter;
+ struct hfi_vnic_port *vport = adapter->vport;
+ u8 evt = rxq->idx + HFI_VNIC_EVT_RX0;
+ int work_done = 0;
+
+ v_dbg("napi %d budget %d\n", rxq->idx, budget);
+ vnic_handle_rx(rxq, &work_done, budget);
+
+ v_dbg("napi %d work_done %d\n", rxq->idx, work_done);
+ if (work_done < budget) {
+ napi_complete(napi);
+ vport->ops->config_notify(vport, evt, true);
+ }
+
+ return work_done;
+}
+
+/* vnic_event_cb - handle events from vnic hfi driver */
+static void vnic_event_cb(struct hfi_vnic_port *vport, u8 evt)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(vport->netdev);
+ struct hfi_vnic_rx_queue *rxq;
+ u8 q_idx;
+
+ v_dbg("received event %d\n", evt);
+ if (evt < vport->hfi_info.num_rx_q) {
+ q_idx = evt;
+ if (unlikely(adapter->info.vport.oper_state !=
+ HFI_VNIC_STATE_FORWARDING))
+ return;
+
+ rxq = &adapter->rxq[q_idx];
+ if (napi_schedule_prep(&rxq->napi)) {
+ v_dbg("napi %d scheduling\n", q_idx);
+ vport->ops->config_notify(vport, evt, false);
+ __napi_schedule(&rxq->napi);
+ }
+ return;
+ }
+ if ((evt >= HFI_VNIC_EVT_TX0) &&
+ (evt < (HFI_VNIC_EVT_TX0 + vport->hfi_info.num_tx_q))) {
+ q_idx = evt - HFI_VNIC_EVT_TX0;
+
+ if (__netif_subqueue_stopped(vport->netdev, q_idx))
+ netif_wake_subqueue(vport->netdev, q_idx);
+
+ return;
+ }
+ v_err("Invalid event\n");
+}
+
+static u16 hfi_vnic_select_queue(struct net_device *netdev, struct sk_buff *skb,
+ void *accel_priv,
+ select_queue_fallback_t fallback)
+{
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ struct __hfi_veswport_info *info = &adapter->info;
+ struct hfi_vnic_port *vport = adapter->vport;
+ u8 vl, entropy;
+
+ if (skb_vlan_tag_present(skb)) {
+ u8 pcp = skb_vlan_tag_get(skb) >> VLAN_PRIO_SHIFT;
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ vl = info->vport.pcp_to_vl_mc[pcp];
+ else
+ vl = info->vport.pcp_to_vl_uc[pcp];
+ } else {
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ vl = info->vport.non_vlan_vl_mc;
+ else
+ vl = info->vport.non_vlan_vl_uc;
+ }
+
+ entropy = hfi_vnic_calc_entropy(adapter, skb);
+ return vport->ops->select_queue(vport, vl, entropy);
+}
+
+/* hfi_netdev_change_mtu - change the MTU */
+static int hfi_netdev_change_mtu(struct net_device *netdev, int new_mtu)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ struct __hfi_veswport_info *info = &adapter->info;
+ u16 min_mtu = HFI_VNIC_MIN_ETH_MTU;
+ u16 max_mtu = max(min_mtu, info->vesw.eth_mtu_non_vlan);
+
+ /* Supported MTUs */
+ if ((new_mtu < min_mtu) || (new_mtu > max_mtu)) {
+ v_err("Unsupported MTU setting\n");
+ return -EINVAL;
+ }
+
+ v_info("changing MTU from %d to %d\n", netdev->mtu, new_mtu);
+ netdev->mtu = new_mtu;
+ return 0;
+}
+
+/* hfi_vnic_up - enable vnic data flow */
+static int hfi_vnic_up(struct hfi_vnic_adapter *adapter)
+{
+ struct hfi_vnic_port *vport = adapter->vport;
+ int i, rc;
+
+ rc = vport->ops->open(vport, vnic_event_cb);
+ if (rc) {
+ v_dbg("hfi_open failed %d\n", rc);
+ return rc;
+ }
+
+ netif_carrier_on(adapter->netdev);
+ netif_tx_start_all_queues(adapter->netdev);
+ for (i = 0; i < vport->hfi_info.num_rx_q; i++)
+ napi_enable(&adapter->rxq[i].napi);
+
+ set_bit(HFI_VNIC_UP, &adapter->flags);
+ return 0;
+}
+
+/* hfi_vnic_down - disable vnic data flow */
+static void hfi_vnic_down(struct hfi_vnic_adapter *adapter)
+{
+ struct hfi_vnic_port *vport = adapter->vport;
+ int i;
+
+ netif_carrier_off(adapter->netdev);
+ netif_tx_disable(adapter->netdev);
+ for (i = 0; i < vport->hfi_info.num_rx_q; i++)
+ napi_disable(&adapter->rxq[i].napi);
+
+ vport->ops->close(vport);
+ clear_bit(HFI_VNIC_UP, &adapter->flags);
+}
+
+/* hfi_vnic_set_mac_addr - change mac address */
+static int hfi_vnic_set_mac_addr(struct net_device *netdev, void *addr)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ struct sockaddr *sa = addr;
+ int rc;
+
+ if (!memcmp(netdev->dev_addr, sa->sa_data, ETH_ALEN))
+ return 0;
+
+ mutex_lock(&adapter->lock);
+ rc = eth_mac_addr(netdev, addr);
+ mutex_unlock(&adapter->lock);
+
+ return rc;
+}
+
+/* hfi_netdev_open - activate network interface */
+static int hfi_netdev_open(struct net_device *netdev)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ int rc;
+
+ mutex_lock(&adapter->lock);
+ rc = hfi_vnic_up(adapter);
+ if (rc)
+ goto open_done;
+
+ set_bit(HFI_VNIC_OPEN, &adapter->flags);
+ v_info("opened\n");
+open_done:
+ mutex_unlock(&adapter->lock);
+ return rc;
+}
+
+/* hfi_netdev_close - disable network interface */
+static int hfi_netdev_close(struct net_device *netdev)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+
+ mutex_lock(&adapter->lock);
+ if (test_bit(HFI_VNIC_UP, &adapter->flags))
+ hfi_vnic_down(adapter);
+
+ clear_bit(HFI_VNIC_OPEN, &adapter->flags);
+ mutex_unlock(&adapter->lock);
+ v_info("closed\n");
+ return 0;
+}
+
+/* netdev ops */
+static const struct net_device_ops hfi_netdev_ops = {
+ .ndo_open = hfi_netdev_open,
+ .ndo_stop = hfi_netdev_close,
+ .ndo_start_xmit = hfi_netdev_start_xmit,
+ .ndo_change_mtu = hfi_netdev_change_mtu,
+ .ndo_select_queue = hfi_vnic_select_queue,
+ .ndo_set_mac_address = hfi_vnic_set_mac_addr,
+};
+
+/* hfi_vnic_add_netdev - create vnic netdev interface */
+struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
+ struct device *parent)
+{
+ struct net_device *netdev;
+ struct hfi_vnic_adapter *adapter;
+ int i, rc;
+
+ netdev = alloc_etherdev_mqs(sizeof(struct hfi_vnic_adapter),
+ vport->hfi_info.num_tx_q,
+ vport->hfi_info.num_rx_q);
+ if (!netdev)
+ return ERR_PTR(-ENOMEM);
+ adapter = netdev_priv(netdev);
+ adapter->netdev = netdev;
+ adapter->vport = vport;
+ vport->netdev = netdev;
+ netdev->features = NETIF_F_HIGHDMA;
+ if (vport->hfi_info.cap & HFI_VNIC_CAP_SG)
+ netdev->features |= NETIF_F_SG;
+ netdev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
+ netdev->hw_features = netdev->features;
+ netdev->vlan_features = netdev->features;
+ netdev->watchdog_timeo = msecs_to_jiffies(HFI_TX_TIMEOUT_MS);
+ netdev->netdev_ops = &hfi_netdev_ops;
+ netdev->hard_header_len += HFI_VNIC_SKB_HEADROOM;
+ mutex_init(&adapter->lock);
+ strcpy(netdev->name, "veth%d");
+
+ SET_NETDEV_DEV(netdev, parent);
+
+ hfi_vnic_set_ethtool_ops(netdev);
+ for (i = 0; i < vport->hfi_info.num_rx_q; i++) {
+ adapter->rxq[i].idx = i;
+ adapter->rxq[i].adapter = adapter;
+ netif_napi_add(netdev, &adapter->rxq[i].napi, vnic_napi, 64);
+ }
+
+ rc = register_netdev(netdev);
+ if (rc)
+ goto netdev_err;
+
+ netif_carrier_off(netdev);
+ v_info("initialized\n");
+
+ return adapter;
+netdev_err:
+ mutex_destroy(&adapter->lock);
+ free_netdev(netdev);
+
+ return ERR_PTR(rc);
+}
+
+/* hfi_vnic_rem_netdev - remove vnic netdev interface */
+void hfi_vnic_rem_netdev(struct hfi_vnic_port *vport)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(vport->netdev);
+
+ v_info("removing\n");
+ unregister_netdev(vport->netdev);
+ mutex_destroy(&adapter->lock);
+ free_netdev(vport->netdev);
+}
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 41+ messages in thread
* [RFC v2 04/10] IB/hfi-vnic: VNIC Ethernet Management (EM) structure definitions
[not found] ` <1481788782-89964-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2016-12-15 7:59 ` [RFC v2 02/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev Vishwanathapura, Niranjana
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 06/10] IB/hfi-vnic: VNIC MAC table support Vishwanathapura, Niranjana
` (2 subsequent siblings)
5 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Niranjana Vishwanathapura,
Sadanand Warrier, Tanya K Jajodia
Define VNIC EM MAD structures and the associated macros. These structures
are used for information exchange between VNIC EM agent (EMA) on the HFI
host and the Ethernet manager. These include the virtual ethernet switch
(vesw) port information, vesw port mac table, summay and error counters,
vesw port interface mac lists and the EMA trap.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Sadanand Warrier <sadanand.warrier-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Tanya K Jajodia <tanya.k.jajodia-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h | 444 +++++++++++++++++++++
.../sw/intel/hfi_vnic/hfi_vnic_internal.h | 33 ++
2 files changed, 477 insertions(+)
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
index 6786cce..a6770ef 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
@@ -52,11 +52,455 @@
* and decapsulation of Ethernet packets
*/
+#include <linux/types.h>
+#include <rdma/ib_mad.h>
+
+/* Maximum number of vnics supported */
+#define HFI_MAX_VPORTS_SUPPORTED 256
+
+/* EMA class version */
+#define HFI_EMA_CLASS_VERSION 0x80
+
+/*
+ * Define the Intel vendor management class for HFI
+ * ETHERNET MANAGEMENT
+ */
+#define HFI_MGMT_CLASS_INTEL_EMA 0x34
+
+/* EM attribute IDs */
+#define HFI_EM_ATTR_CLASS_PORT_INFO 0x0001
+#define HFI_EM_ATTR_VESWPORT_INFO 0x0011
+#define HFI_EM_ATTR_VESWPORT_MAC_ENTRIES 0x0012
+#define HFI_EM_ATTR_IFACE_UCAST_MACS 0x0013
+#define HFI_EM_ATTR_IFACE_MCAST_MACS 0x0014
+#define HFI_EM_ATTR_DELETE_VESW 0x0015
+#define HFI_EM_ATTR_VESWPORT_SUMMARY_COUNTERS 0x0020
+#define HFI_EM_ATTR_VESWPORT_ERROR_COUNTERS 0x0022
+
#define HFI_VESW_MAX_NUM_DEF_PORT 16
#define HFI_VNIC_MAX_NUM_PCP 8
+#define HFI_VNIC_EMA_DATA (OPA_MGMT_MAD_SIZE - IB_MGMT_VENDOR_HDR)
+
+/* Defines for vendor specific notice(trap) attributes */
+#define HFI_INTEL_EMA_NOTICE_TYPE_INFO 0x04
+
+/* INTEL OUI */
+#define INTEL_OUI_1 0x00
+#define INTEL_OUI_2 0x06
+#define INTEL_OUI_3 0x6a
+
+/* Trap opcodes sent from VNIC */
+#define HFI_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE 0x1
+#define HFI_VESWPORT_TRAP_IFACE_MCAST_MAC_CHANGE 0x2
+#define HFI_VESWPORT_TRAP_ETH_LINK_STATUS_CHANGE 0x3
+
/* VNIC configured and operational state values */
#define HFI_VNIC_STATE_DROP_ALL 0x1
#define HFI_VNIC_STATE_FORWARDING 0x3
+/**
+ * struct hfi_vesw_info - HFI vnic switch information
+ * @fabric_id: 10-bit fabric id
+ * @vesw_id: 12-bit virtual ethernet switch id
+ * @def_port_mask: bitmask of default ports
+ * @pkey: partition key
+ * @u_mcast_dlid: unknown multicast dlid
+ * @u_ucast_dlid: array of unknown unicast dlids
+ * @eth_mtu: MTUs for each vlan PCP
+ * @eth_mtu_non_vlan: MTU for non vlan packets
+ */
+struct hfi_vesw_info {
+ __be16 fabric_id;
+ __be16 vesw_id;
+
+ u8 rsvd0[6];
+ __be16 def_port_mask;
+
+ u8 rsvd1[2];
+ __be16 pkey;
+
+ u8 rsvd2[4];
+ __be32 u_mcast_dlid;
+ __be32 u_ucast_dlid[HFI_VESW_MAX_NUM_DEF_PORT];
+
+ u8 rsvd3[44];
+ __be16 eth_mtu[HFI_VNIC_MAX_NUM_PCP];
+ __be16 eth_mtu_non_vlan;
+ u8 rsvd4[2];
+} __packed;
+
+/**
+ * struct hfi_per_veswport_info - HFI vnic per port information
+ * @port_num: port number
+ * @eth_link_status: current ethernet link state
+ * @base_mac_addr: base mac address
+ * @config_state: configured port state
+ * @oper_state: operational port state
+ * @max_mac_tbl_ent: max number of mac table entries
+ * @max_smac_ent: max smac entries in mac table
+ * @mac_tbl_digest: mac table digest
+ * @encap_slid: base slid for the port
+ * @pcp_to_sc_uc: sc by pcp index for unicast ethernet packets
+ * @pcp_to_vl_uc: vl by pcp index for unicast ethernet packets
+ * @pcp_to_sc_mc: sc by pcp index for multicast ethernet packets
+ * @pcp_to_vl_mc: vl by pcp index for multicast ethernet packets
+ * @non_vlan_sc_uc: sc for non-vlan unicast ethernet packets
+ * @non_vlan_vl_uc: vl for non-vlan unicast ethernet packets
+ * @non_vlan_sc_mc: sc for non-vlan multicast ethernet packets
+ * @non_vlan_vl_mc: vl for non-vlan multicast ethernet packets
+ * @uc_macs_gen_count: generation count for unicast macs list
+ * @mc_macs_gen_count: generation count for multicast macs list
+ */
+struct hfi_per_veswport_info {
+ __be32 port_num;
+
+ u8 eth_link_status;
+ u8 rsvd0[3];
+
+ u8 base_mac_addr[ETH_ALEN];
+ u8 config_state;
+ u8 oper_state;
+
+ __be16 max_mac_tbl_ent;
+ __be16 max_smac_ent;
+ __be32 mac_tbl_digest;
+ u8 rsvd1[4];
+
+ __be32 encap_slid;
+
+ u8 pcp_to_sc_uc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_vl_uc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_sc_mc[HFI_VNIC_MAX_NUM_PCP];
+ u8 pcp_to_vl_mc[HFI_VNIC_MAX_NUM_PCP];
+
+ u8 non_vlan_sc_uc;
+ u8 non_vlan_vl_uc;
+ u8 non_vlan_sc_mc;
+ u8 non_vlan_vl_mc;
+
+ u8 rsvd2[48];
+
+ __be16 uc_macs_gen_count;
+ __be16 mc_macs_gen_count;
+
+ u8 rsvd3[8];
+} __packed;
+
+/**
+ * struct hfi_veswport_info - HFI vnic port information
+ * @vesw: HFI vnic switch information
+ * @vport: HFI vnic per port information
+ *
+ * On host, each of the virtual ethernet ports belongs
+ * to a different virtual ethernet switches.
+ */
+struct hfi_veswport_info {
+ struct hfi_vesw_info vesw;
+ struct hfi_per_veswport_info vport;
+};
+
+/**
+ * union __hfi_vnic_dlid_sd - vnic dlid and side data needed.
+ * @sd_is_src_mac: 1 = entry is SMAC, 0 = not SMAC
+ * @dlid: Destination lid corresponding to MAC addr
+ */
+union __hfi_vnic_dlid_sd {
+ struct {
+ u32 sd_reserved : 5;
+ u32 sd_is_src_mac : 1;
+ u32 rsvd0 : 2;
+ u32 dlid : 24;
+ };
+ u32 dw;
+};
+
+/* Same as __hfi_vnic_dlid_sd, but with a big endian attribute */
+union hfi_vnic_dlid_sd {
+ union __hfi_vnic_dlid_sd u;
+ __be32 dw;
+};
+
+/**
+ * struct hfi_veswport_mactable_entry - single entry in the forwarding table
+ * @mac_addr: MAC address
+ * @mac_addr_mask: MAC address bit mask
+ * @dlid_sd: Matching DLID and side data
+ *
+ * On the host each virtual ethernet port will have
+ * a forwarding table. These tables are used to
+ * map a MAC to a LID and other data. For more
+ * details see struct hfi_veswport_mactable_entries.
+ * This is the structure of a single mactable entry
+ */
+struct hfi_veswport_mactable_entry {
+ u8 mac_addr[ETH_ALEN];
+ u8 mac_addr_mask[ETH_ALEN];
+ union hfi_vnic_dlid_sd dlid_sd;
+} __packed;
+
+/**
+ * struct hfi_veswport_mactable - Forwarding table array
+ * @offset: mac table starting offset
+ * @num_entries: Number of entries to get or set
+ * @mac_tbl_digest: mac table digest
+ * @tbl_entries[]: Array of table entries
+ *
+ * The EM sends down this structure in a MAD indicating
+ * the starting offset in the forwarding table that this
+ * entry is to be loaded into and the number of entries
+ * that that this MAD instance contains
+ * The mac_tbl_digest has been added to this MAD structure. It will be set by
+ * the EM and it will be used by the EM to check if there are any
+ * discrepancies with this value and the value
+ * maintained by the EM in the case of VNIC port being deleted or unloaded
+ * A new instantiation of a VNIC will always have a value of zero.
+ * This value is stored as part of the vnic adapter structure and will be
+ * accessed by the GET and SET routines for both the mactable entries and the
+ * veswport info.
+ */
+struct hfi_veswport_mactable {
+ __be16 offset;
+ __be16 num_entries;
+ __be32 mac_tbl_digest;
+ struct hfi_veswport_mactable_entry tbl_entries[0];
+} __packed;
+
+/**
+ * struct hfi_veswport_summary_counters - summary counters
+ * @vp_instance: vport instance on the HFI port
+ * @vesw_id: virtual ethernet switch id
+ * @veswport_num: virtual ethernet switch port number
+ * @tx_errors: transmit errors
+ * @rx_errors: receive errors
+ * @tx_packets: transmit packets
+ * @rx_packets: receive packets
+ * @tx_bytes: transmit bytes
+ * @rx_bytes: receive bytes
+ * @tx_unicast: unicast packets transmitted
+ * @tx_mcastbcast: multicast/broadcast packets transmitted
+ * @tx_untagged: non-vlan packets transmitted
+ * @tx_vlan: vlan packets transmitted
+ * @tx_64_size: transmit packet length is 64 bytes
+ * @tx_65_127: transmit packet length is >=65 and < 127 bytes
+ * @tx_128_255: transmit packet length is >=128 and < 255 bytes
+ * @tx_256_511: transmit packet length is >=256 and < 511 bytes
+ * @tx_512_1023: transmit packet length is >=512 and < 1023 bytes
+ * @tx_1024_1518: transmit packet length is >=1024 and < 1518 bytes
+ * @tx_1519_max: transmit packet length >= 1519 bytes
+ * @rx_unicast: unicast packets received
+ * @rx_mcastbcast: multicast/broadcast packets received
+ * @rx_untagged: non-vlan packets received
+ * @rx_vlan: vlan packets received
+ * @rx_64_size: received packet length is 64 bytes
+ * @rx_65_127: received packet length is >=65 and < 127 bytes
+ * @rx_128_255: received packet length is >=128 and < 255 bytes
+ * @rx_256_511: received packet length is >=256 and < 511 bytes
+ * @rx_512_1023: received packet length is >=512 and < 1023 bytes
+ * @rx_1024_1518: received packet length is >=1024 and < 1518 bytes
+ * @rx_1519_max: received packet length >= 1519 bytes
+ *
+ * All the above are counters of corresponding conditions.
+ */
+struct hfi_veswport_summary_counters {
+ __be16 vp_instance;
+ __be16 vesw_id;
+ __be32 veswport_num;
+
+ __be64 tx_errors;
+ __be64 rx_errors;
+ __be64 tx_packets;
+ __be64 rx_packets;
+ __be64 tx_bytes;
+ __be64 rx_bytes;
+
+ __be64 tx_unicast;
+ __be64 tx_mcastbcast;
+
+ __be64 tx_untagged;
+ __be64 tx_vlan;
+
+ __be64 tx_64_size;
+ __be64 tx_65_127;
+ __be64 tx_128_255;
+ __be64 tx_256_511;
+ __be64 tx_512_1023;
+ __be64 tx_1024_1518;
+ __be64 tx_1519_max;
+
+ __be64 rx_unicast;
+ __be64 rx_mcastbcast;
+
+ __be64 rx_untagged;
+ __be64 rx_vlan;
+
+ __be64 rx_64_size;
+ __be64 rx_65_127;
+ __be64 rx_128_255;
+ __be64 rx_256_511;
+ __be64 rx_512_1023;
+ __be64 rx_1024_1518;
+ __be64 rx_1519_max;
+
+ __be64 reserved[16];
+} __packed;
+
+/**
+ * struct hfi_veswport_error_counters - error counters
+ * @vp_instance: vport instance on the HFI port
+ * @vesw_id: virtual ethernet switch id
+ * @veswport_num: virtual ethernet switch port number
+ * @tx_errors: transmit errors
+ * @rx_errors: receive errors
+ * @tx_smac_filt: smac filter errors
+ * @tx_dlid_zero: transmit packets with invalid dlid
+ * @tx_logic: other transmit errors
+ * @tx_drop_state: packet tansmission in non-forward port state
+ * @rx_bad_veswid: received packet with invalid vesw id
+ * @rx_runt: received ethernet packet with length < 64 bytes
+ * @rx_oversize: received ethernet packet with length > MTU size
+ * @rx_eth_down: received packets when interface is down
+ * @rx_drop_state: received packets in non-forwarding port state
+ * @rx_logic: other receive errors
+ *
+ * All the above are counters of corresponding erorr conditions.
+ */
+struct hfi_veswport_error_counters {
+ __be16 vp_instance;
+ __be16 vesw_id;
+ __be32 veswport_num;
+
+ __be64 tx_errors;
+ __be64 rx_errors;
+
+ __be64 rsvd0;
+ __be64 tx_smac_filt;
+ __be64 rsvd1;
+ __be64 rsvd2;
+ __be64 rsvd3;
+ __be64 tx_dlid_zero;
+ __be64 rsvd4;
+ __be64 tx_logic;
+ __be64 rsvd5;
+ __be64 tx_drop_state;
+
+ __be64 rx_bad_veswid;
+ __be64 rsvd6;
+ __be64 rx_runt;
+ __be64 rx_oversize;
+ __be64 rsvd7;
+ __be64 rx_eth_down;
+ __be64 rx_drop_state;
+ __be64 rx_logic;
+ __be64 rsvd8;
+
+ __be64 rsvd9[16];
+} __packed;
+
+/**
+ * struct hfi_veswport_trap - Trap message sent to EM by VNIC
+ * @fabric_id: 10 bit fabric id
+ * @veswid: 12 bit virtual ethernet switch id
+ * @veswportnum: logical port number on the Virtual switch
+ * @hfiportnum: physical port num (redundant on host)
+ * @veswportindex: switch port index on hfi port 0 based
+ * @opcode: operation
+ * @reserved: 32 bit for alignment
+ *
+ * The VNIC will send trap messages to the Ethernet manager to
+ * inform it about changes to the VNIC config, behaviour etc.
+ * This is the format of the trap payload.
+ */
+struct hfi_veswport_trap {
+ __be16 fabric_id;
+ __be16 veswid;
+ __be32 veswportnum;
+ __be16 hfiportnum;
+ u8 veswportindex;
+ u8 opcode;
+ __be32 reserved;
+} __packed;
+
+/**
+ * struct hfi_vnic_iface_macs_entry - single entry in the mac list
+ * @mac_addr: MAC address
+ */
+struct hfi_vnic_iface_mac_entry {
+ u8 mac_addr[ETH_ALEN];
+};
+
+/**
+ * struct hfi_veswport_iface_macs - Msg to set globally administered MAC
+ * @start_idx: position of first entry (0 based)
+ * @num_macs_in_msg: number of MACs in this message
+ * @tot_macs_in_lst: The total number of MACs the agent has
+ * @gen_count: gen_count to indicate change
+ * @entry: The mac list entry
+ *
+ * Same attribute IDS and attribute modifiers as in locally administered
+ * addresses used to set globally administered addresses
+ */
+struct hfi_veswport_iface_macs {
+ __be16 start_idx;
+ __be16 num_macs_in_msg;
+ __be16 tot_macs_in_lst;
+ __be16 gen_count;
+ struct hfi_vnic_iface_mac_entry entry[0];
+} __packed;
+
+/**
+ * struct hfi_vnic_vema_mad - Generic VEMA MAD
+ * @mad_hdr: Generic MAD header
+ * @rmpp_hdr: RMPP header for vendor specific MADs
+ * @oui: Unique org identifier
+ * @data: MAD data
+ */
+struct hfi_vnic_vema_mad {
+ struct ib_mad_hdr mad_hdr;
+ struct ib_rmpp_hdr rmpp_hdr;
+ u8 reserved;
+ u8 oui[3];
+ u8 data[HFI_VNIC_EMA_DATA];
+};
+
+/**
+ * struct hfi_vnic_notice_attr - Generic Notice MAD
+ * @gen_type: Generic/Specific bit and type of notice
+ * @oui_1: Vendor ID byte 1
+ * @oui_2: Vendor ID byte 2
+ * @oui_3: Vendor ID byte 3
+ * @trap_num: Trap number
+ * @toggle_count: Notice toggle bit and count value
+ * @issuer_lid: Trap issuer's lid
+ * @issuer_gid: Issuer GID (only if Report method)
+ * @raw_data: Trap message body
+ */
+struct hfi_vnic_notice_attr {
+ u8 gen_type;
+ u8 oui_1;
+ u8 oui_2;
+ u8 oui_3;
+ __be16 trap_num;
+ __be16 toggle_count;
+ __be32 issuer_lid;
+ __be32 reserved;
+ u8 issuer_gid[16];
+ u8 raw_data[64];
+} __packed;
+
+/**
+ * struct hfi_vnic_vema_mad_trap - Generic VEMA MAD Trap
+ * @mad_hdr: Generic MAD header
+ * @rmpp_hdr: RMPP header for vendor specific MADs
+ * @oui: Unique org identifier
+ * @notice: Notice structure
+ */
+struct hfi_vnic_vema_mad_trap {
+ struct ib_mad_hdr mad_hdr;
+ struct ib_rmpp_hdr rmpp_hdr;
+ u8 reserved;
+ u8 oui[3];
+ struct hfi_vnic_notice_attr notice;
+};
+
#endif /* _HFI_VNIC_ENCAP_H */
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
index 30731b4..63d6db6 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
@@ -96,6 +96,8 @@ enum hfi_vnic_flags_t {
/**
* struct __hfi_vesw_info - HFI vnic virtual switch info
+ *
+ * Same as hfi_vesw_info without bitwise attribute.
*/
struct __hfi_vesw_info {
u16 fabric_id;
@@ -119,6 +121,8 @@ struct __hfi_vesw_info {
/**
* struct __hfi_per_veswport_info - HFI vnic per port info
+ *
+ * Same as hfi_per_veswport_info without bitwise attribute.
*/
struct __hfi_per_veswport_info {
u32 port_num;
@@ -157,6 +161,8 @@ struct __hfi_per_veswport_info {
/**
* struct __hfi_veswport_info - HFI vnic port info
+ *
+ * Same as hfi_veswport_info without bitwise attribute.
*/
struct __hfi_veswport_info {
struct __hfi_vesw_info vesw;
@@ -164,6 +170,21 @@ struct __hfi_veswport_info {
};
/**
+ * struct __hfi_veswport_trap - HFI vnic trap info
+ *
+ * Same as hfi_veswport_trap without bitwise attribute.
+ */
+struct __hfi_veswport_trap {
+ u16 fabric_id;
+ u16 veswid;
+ u32 veswportnum;
+ u16 hfiportnum;
+ u8 veswportindex;
+ u8 opcode;
+ u32 reserved;
+} __packed;
+
+/**
* struct hfi_vnic_rx_queue - HFI VNIC receive queue
* @idx: queue index
* @adapter: netdev adapter
@@ -209,6 +230,18 @@ struct hfi_vnic_adapter {
#define v_warn(format, arg...) \
netdev_warn(adapter->netdev, format, ## arg)
+/* The maximum allowed entries in the mac table */
+#define HFI_VNIC_MAC_TBL_MAX_ENTRIES 2048
+/* Limit of smac entries in mac table */
+#define HFI_VNIC_MAX_SMAC_LIMIT 256
+
+/* The last octet of the MAC address is used as the key to the hash table */
+#define HFI_VNIC_MAC_HASH_IDX 5
+
+/* The VNIC MAC hash table is of size 2^8 */
+#define HFI_VNIC_MAC_TBL_HASH_BITS 8
+#define HFI_VNIC_MAC_TBL_SIZE BIT(HFI_VNIC_MAC_TBL_HASH_BITS)
+
struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
struct device *parent);
void hfi_vnic_rem_netdev(struct hfi_vnic_port *vport);
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 41+ messages in thread
* [RFC v2 05/10] IB/hfi-vnic: VNIC statistics support
2016-12-15 7:59 [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 01/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) documentation Vishwanathapura, Niranjana
[not found] ` <1481788782-89964-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 08/10] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) function Vishwanathapura, Niranjana
` (3 subsequent siblings)
6 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura
HFI VNIC driver statistics support maintains various counters including
standard netdev counters and the Ethernet manager defined counters.
Add the Ethtool hook to read the counters.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
---
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c | 19 +-
.../sw/intel/hfi_vnic/hfi_vnic_ethtool.c | 131 +++++++++++
.../sw/intel/hfi_vnic/hfi_vnic_internal.h | 84 +++++++
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c | 260 ++++++++++++++++++++-
4 files changed, 486 insertions(+), 8 deletions(-)
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
index 093df67..3fdfb7b 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
@@ -209,8 +209,10 @@ int hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb)
hdr->slid_high = info->vport.encap_slid >> 20;
dlid = hfi_vnic_get_dlid(adapter, skb, def_port);
- if (unlikely(!dlid))
+ if (unlikely(!dlid)) {
+ adapter->q_err_cntrs[skb->queue_mapping].tx_dlid_zero++;
return -EFAULT;
+ }
hdr->dlid = dlid;
hdr->dlid_high = dlid >> 20;
@@ -233,6 +235,19 @@ int hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb)
/* hfi_vnic_decap_skb - strip OPA header from the skb (ethernet) packet */
int hfi_vnic_decap_skb(struct hfi_vnic_rx_queue *rxq, struct sk_buff *skb)
{
+ struct hfi_vnic_adapter *adapter = rxq->adapter;
+ int max_len = adapter->netdev->mtu + VLAN_ETH_HLEN;
+ int rc = -EFAULT;
+
skb_pull(skb, HFI_VNIC_HDR_LEN);
- return 0;
+
+ /* Validate Packet length */
+ if (skb->len > max_len)
+ adapter->q_err_cntrs[rxq->idx].rx_oversize++;
+ else if (skb->len < ETH_ZLEN)
+ adapter->q_err_cntrs[rxq->idx].rx_runt++;
+ else
+ rc = 0;
+
+ return rc;
}
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
index 0b4da5e..9289ab2 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
@@ -53,9 +53,140 @@
#include "hfi_vnic_internal.h"
+enum {NETDEV_STATS, VNIC_STATS};
+
+struct vnic_stats {
+ char stat_string[ETH_GSTRING_LEN];
+ struct {
+ int type;
+ int sizeof_stat;
+ int stat_offset;
+ };
+};
+
+#define VNIC_STAT(m) { VNIC_STATS, \
+ FIELD_SIZEOF(struct hfi_vnic_adapter, m), \
+ offsetof(struct hfi_vnic_adapter, m) }
+#define VNIC_NETDEV_STAT(m) { NETDEV_STATS, \
+ FIELD_SIZEOF(struct net_device, m), \
+ offsetof(struct net_device, m) }
+
+static struct vnic_stats vnic_gstrings_stats[] = {
+ /* NETDEV stats */
+ {"rx_packets", VNIC_NETDEV_STAT(stats.rx_packets)},
+ {"tx_packets", VNIC_NETDEV_STAT(stats.tx_packets)},
+ {"rx_bytes", VNIC_NETDEV_STAT(stats.rx_bytes)},
+ {"tx_bytes", VNIC_NETDEV_STAT(stats.tx_bytes)},
+ {"rx_errors", VNIC_NETDEV_STAT(stats.rx_errors)},
+ {"tx_errors", VNIC_NETDEV_STAT(stats.tx_errors)},
+ {"rx_dropped", VNIC_NETDEV_STAT(stats.rx_dropped)},
+ {"tx_dropped", VNIC_NETDEV_STAT(stats.tx_dropped)},
+
+ {"rx_fifo_errors", VNIC_NETDEV_STAT(stats.rx_fifo_errors)},
+ {"rx_missed_errors", VNIC_NETDEV_STAT(stats.rx_missed_errors)},
+ {"tx_carrier_errors", VNIC_NETDEV_STAT(stats.tx_carrier_errors)},
+ {"tx_fifo_errors", VNIC_NETDEV_STAT(stats.tx_fifo_errors)},
+
+ /* SUMMARY counters */
+ {"tx_unicast", VNIC_STAT(sum_cntrs.tx_grp.unicast)},
+ {"tx_mcastbcast", VNIC_STAT(sum_cntrs.tx_grp.mcastbcast)},
+ {"tx_untagged", VNIC_STAT(sum_cntrs.tx_grp.untagged)},
+ {"tx_vlan", VNIC_STAT(sum_cntrs.tx_grp.vlan)},
+
+ {"tx_64_size", VNIC_STAT(sum_cntrs.tx_grp.xx_64_size)},
+ {"tx_65_127", VNIC_STAT(sum_cntrs.tx_grp.xx_65_127)},
+ {"tx_128_255", VNIC_STAT(sum_cntrs.tx_grp.xx_128_255)},
+ {"tx_256_511", VNIC_STAT(sum_cntrs.tx_grp.xx_256_511)},
+ {"tx_512_1023", VNIC_STAT(sum_cntrs.tx_grp.xx_512_1023)},
+ {"tx_1024_1518", VNIC_STAT(sum_cntrs.tx_grp.xx_1024_1518)},
+ {"tx_1519_max", VNIC_STAT(sum_cntrs.tx_grp.xx_1519_max)},
+
+ {"rx_unicast", VNIC_STAT(sum_cntrs.rx_grp.unicast)},
+ {"rx_mcastbcast", VNIC_STAT(sum_cntrs.rx_grp.mcastbcast)},
+ {"rx_untagged", VNIC_STAT(sum_cntrs.rx_grp.untagged)},
+ {"rx_vlan", VNIC_STAT(sum_cntrs.rx_grp.vlan)},
+
+ {"rx_64_size", VNIC_STAT(sum_cntrs.rx_grp.xx_64_size)},
+ {"rx_65_127", VNIC_STAT(sum_cntrs.rx_grp.xx_65_127)},
+ {"rx_128_255", VNIC_STAT(sum_cntrs.rx_grp.xx_128_255)},
+ {"rx_256_511", VNIC_STAT(sum_cntrs.rx_grp.xx_256_511)},
+ {"rx_512_1023", VNIC_STAT(sum_cntrs.rx_grp.xx_512_1023)},
+ {"rx_1024_1518", VNIC_STAT(sum_cntrs.rx_grp.xx_1024_1518)},
+ {"rx_1519_max", VNIC_STAT(sum_cntrs.rx_grp.xx_1519_max)},
+
+ /* ERROR counters */
+ {"tx_smac_filt", VNIC_STAT(err_cntrs.tx_smac_filt)},
+ {"tx_dlid_zero", VNIC_STAT(err_cntrs.tx_dlid_zero)},
+ {"tx_logic", VNIC_STAT(err_cntrs.tx_logic)},
+ {"tx_drop_state", VNIC_STAT(err_cntrs.tx_drop_state)},
+
+ {"rx_bad_veswid", VNIC_STAT(err_cntrs.rx_bad_veswid)},
+ {"rx_runt", VNIC_STAT(err_cntrs.rx_runt)},
+ {"rx_oversize", VNIC_STAT(err_cntrs.rx_oversize)},
+ {"rx_eth_down", VNIC_STAT(err_cntrs.rx_eth_down)},
+ {"rx_drop_state", VNIC_STAT(err_cntrs.rx_drop_state)},
+ {"rx_logic", VNIC_STAT(err_cntrs.rx_logic)},
+};
+
+#define VNIC_STATS_LEN ARRAY_SIZE(vnic_gstrings_stats)
+
+/* vnic_get_sset_count - get string set count */
+static int vnic_get_sset_count(struct net_device *netdev, int sset)
+{
+ return (sset == ETH_SS_STATS) ? VNIC_STATS_LEN : -EOPNOTSUPP;
+}
+
+/* vnic_get_ethtool_stats - get statistics */
+static void vnic_get_ethtool_stats(struct net_device *netdev,
+ struct ethtool_stats *stats, u64 *data)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ int i;
+ char *p = NULL;
+
+ mutex_lock(&adapter->stats_lock);
+ hfi_vnic_update_stats(netdev);
+ for (i = 0; i < VNIC_STATS_LEN; i++) {
+ switch (vnic_gstrings_stats[i].type) {
+ case NETDEV_STATS:
+ p = (char *)netdev +
+ vnic_gstrings_stats[i].stat_offset;
+ break;
+ case VNIC_STATS:
+ p = (char *)adapter +
+ vnic_gstrings_stats[i].stat_offset;
+ break;
+ default:
+ p = NULL;
+ }
+
+ if (p)
+ data[i] = (vnic_gstrings_stats[i].sizeof_stat ==
+ sizeof(u64)) ? *(u64 *)p : *(u32 *)p;
+ }
+ mutex_unlock(&adapter->stats_lock);
+}
+
+/* vnic_get_strings - get strings */
+static void vnic_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
+{
+ int i;
+
+ if (stringset != ETH_SS_STATS)
+ return;
+
+ for (i = 0; i < VNIC_STATS_LEN; i++)
+ memcpy(data + i * ETH_GSTRING_LEN,
+ vnic_gstrings_stats[i].stat_string,
+ ETH_GSTRING_LEN);
+}
+
/* ethtool ops */
static const struct ethtool_ops hfi_vnic_ethtool_ops = {
.get_link = ethtool_op_get_link,
+ .get_strings = vnic_get_strings,
+ .get_sset_count = vnic_get_sset_count,
+ .get_ethtool_stats = vnic_get_ethtool_stats,
};
/* hfi_vnic_set_ethtool_ops - set ethtool ops */
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
index 63d6db6..af3ff0e 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
@@ -95,6 +95,64 @@ enum hfi_vnic_flags_t {
struct hfi_vnic_adapter;
/**
+ * struct __hfi_vnic_summary_counters - HFI summary counters
+ *
+ * Same as __hfi_veswport_summary_counters without bitwise
+ * attribute and reserved fields.
+ */
+struct __hfi_vnic_summary_counters {
+ u64 tx_errors;
+ u64 rx_errors;
+ u64 tx_packets;
+ u64 rx_packets;
+ u64 tx_bytes;
+ u64 rx_bytes;
+
+ /* Group of histogram statistic counters */
+ struct __hfi_vnic_group_scs {
+ u64 unicast;
+ u64 mcastbcast;
+
+ u64 untagged;
+ u64 vlan;
+
+ u64 xx_64_size;
+ u64 xx_65_127;
+ u64 xx_128_255;
+ u64 xx_256_511;
+ u64 xx_512_1023;
+ u64 xx_1024_1518;
+ u64 xx_1519_max;
+ } tx_grp;
+
+ struct __hfi_vnic_group_scs rx_grp;
+
+} __packed;
+
+/**
+ * struct __hfi_vnic_error_counters - HFI error counters
+ *
+ * Same as hfi_veswport_error_counters without bitwise
+ * attribute and reserved fields.
+ */
+struct __hfi_vnic_error_counters {
+ u64 tx_errors;
+ u64 rx_errors;
+
+ u64 tx_smac_filt;
+ u64 tx_dlid_zero;
+ u64 tx_logic;
+ u64 tx_drop_state;
+
+ u64 rx_bad_veswid;
+ u64 rx_runt;
+ u64 rx_oversize;
+ u64 rx_eth_down;
+ u64 rx_drop_state;
+ u64 rx_logic;
+} __packed;
+
+/**
* struct __hfi_vesw_info - HFI vnic virtual switch info
*
* Same as hfi_vesw_info without bitwise attribute.
@@ -204,7 +262,17 @@ struct hfi_vnic_rx_queue {
* @lock: adapter lock
* @rxq: receive queue array
* @info: virtual ethernet switch port information
+ * @stats_lock: statistics lock
* @flow_tbl: flow to default port redirection table
+ * @q_sum_cntrs: per queue EM summary counters
+ * @q_err_cntrs: per queue EM error counters
+ * @q_rx_logic_errors: per queue rx logic (default) errors
+ * @q_tx_logic_errors: per queue tx logic (default) errors
+ * @q_tx_halt: per queue tx halt counts
+ * @q_tx_restart: per queue tx restart counts
+ * @q_tx_wakeup: per queue tx wakeup counts
+ * @sum_cntrs: Total EM summary counters (from all queues)
+ * @err_cntrs: Total EM error counters (from all queues)
*/
struct hfi_vnic_adapter {
struct net_device *netdev;
@@ -218,7 +286,22 @@ struct hfi_vnic_adapter {
struct __hfi_veswport_info info;
+ /* Lock used to protect access to vnic counters */
+ struct mutex stats_lock;
+
u8 flow_tbl[HFI_VNIC_FLOW_TBL_SIZE];
+
+ struct __hfi_vnic_summary_counters q_sum_cntrs[HFI_VNIC_MAX_QUEUE];
+ struct __hfi_vnic_error_counters q_err_cntrs[HFI_VNIC_MAX_QUEUE];
+ u64 q_rx_logic_errors[HFI_VNIC_MAX_QUEUE];
+ u64 q_tx_logic_errors[HFI_VNIC_MAX_QUEUE];
+
+ u64 q_tx_halt[HFI_VNIC_MAX_QUEUE];
+ u64 q_tx_restart[HFI_VNIC_MAX_QUEUE];
+ u64 q_tx_wakeup[HFI_VNIC_MAX_QUEUE];
+
+ struct __hfi_vnic_summary_counters sum_cntrs;
+ struct __hfi_vnic_error_counters err_cntrs;
};
#define v_dbg(format, arg...) \
@@ -248,6 +331,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
int hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
int hfi_vnic_decap_skb(struct hfi_vnic_rx_queue *rxq, struct sk_buff *skb);
u8 hfi_vnic_calc_entropy(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+void hfi_vnic_update_stats(struct net_device *netdev);
void hfi_vnic_set_ethtool_ops(struct net_device *netdev);
#endif /* _HFI_VNIC_INTERNAL_H */
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
index 6360d37..1626e44 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
@@ -58,6 +58,235 @@
#define HFI_VNIC_MIN_ETH_MTU (ETH_ZLEN - ETH_HLEN)
+#define SUM_GRP_COUNTERS(adpt, summary, x_grp) do { \
+ u64 *src64, *dst64; \
+ for (src64 = &summary->x_grp.unicast, \
+ dst64 = &adpt->sum_cntrs.x_grp.unicast; \
+ dst64 <= &adpt->sum_cntrs.x_grp.xx_1519_max;) { \
+ *dst64++ += *src64++; \
+ } \
+ } while (0)
+
+/* hfi_vnic_update_stats - update statistics */
+void hfi_vnic_update_stats(struct net_device *netdev)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ struct hfi_vnic_port *vport = adapter->vport;
+ struct hfi_vnic_stats h_stats = { 0 };
+ u64 tx_logic_errors = 0;
+ u64 rx_logic_errors = 0;
+ u8 i;
+
+ /* first clear the total counters */
+ memset(&adapter->sum_cntrs, 0, sizeof(adapter->sum_cntrs));
+ memset(&adapter->err_cntrs, 0, sizeof(adapter->err_cntrs));
+
+ /* add tx counters on different queues */
+ for (i = 0; i < vport->hfi_info.num_tx_q; i++) {
+ struct hfi_vnic_stats *hfi_stats = &vport->hfi_stats[i];
+ struct __hfi_vnic_summary_counters *sum_cntrs =
+ &adapter->q_sum_cntrs[i];
+ struct __hfi_vnic_error_counters *err_cntrs =
+ &adapter->q_err_cntrs[i];
+
+ h_stats.tx_fifo_errors += hfi_stats->tx_fifo_errors;
+ h_stats.tx_carrier_errors += hfi_stats->tx_carrier_errors;
+ h_stats.tx_logic_errors += hfi_stats->tx_logic_errors;
+
+ SUM_GRP_COUNTERS(adapter, sum_cntrs, tx_grp);
+ adapter->sum_cntrs.tx_packets += sum_cntrs->tx_packets;
+ adapter->sum_cntrs.tx_bytes += sum_cntrs->tx_bytes;
+
+ adapter->err_cntrs.tx_smac_filt += err_cntrs->tx_smac_filt;
+ adapter->err_cntrs.tx_dlid_zero += err_cntrs->tx_dlid_zero;
+ adapter->err_cntrs.tx_drop_state += err_cntrs->tx_drop_state;
+
+ tx_logic_errors += adapter->q_tx_logic_errors[i];
+ }
+
+ /* add rx counters on different queues */
+ for (i = 0; i < vport->hfi_info.num_rx_q; i++) {
+ struct hfi_vnic_stats *hfi_stats = &vport->hfi_stats[i];
+ struct __hfi_vnic_summary_counters *sum_cntrs =
+ &adapter->q_sum_cntrs[i];
+ struct __hfi_vnic_error_counters *err_cntrs =
+ &adapter->q_err_cntrs[i];
+
+ h_stats.rx_fifo_errors += hfi_stats->rx_fifo_errors;
+ h_stats.rx_missed_errors += hfi_stats->rx_missed_errors;
+ h_stats.rx_bad_veswid += hfi_stats->rx_bad_veswid;
+ h_stats.rx_logic_errors += hfi_stats->rx_logic_errors;
+
+ SUM_GRP_COUNTERS(adapter, sum_cntrs, rx_grp);
+ adapter->sum_cntrs.rx_packets += sum_cntrs->rx_packets;
+ adapter->sum_cntrs.rx_bytes += sum_cntrs->rx_bytes;
+
+ adapter->err_cntrs.rx_drop_state += err_cntrs->rx_drop_state;
+ adapter->err_cntrs.rx_runt += err_cntrs->rx_runt;
+ adapter->err_cntrs.rx_oversize += err_cntrs->rx_oversize;
+
+ rx_logic_errors += adapter->q_rx_logic_errors[i];
+ }
+
+ /* update hfi errors */
+ netdev->stats.rx_fifo_errors = h_stats.rx_fifo_errors;
+ netdev->stats.tx_fifo_errors = h_stats.tx_fifo_errors;
+ netdev->stats.rx_missed_errors = h_stats.rx_missed_errors;
+ netdev->stats.tx_carrier_errors = h_stats.tx_carrier_errors;
+ adapter->err_cntrs.rx_bad_veswid = h_stats.rx_bad_veswid;
+
+ /* update tx counters */
+ netdev->stats.tx_packets = adapter->sum_cntrs.tx_packets;
+ netdev->stats.tx_bytes = adapter->sum_cntrs.tx_bytes;
+
+ adapter->err_cntrs.tx_logic = netdev->stats.tx_carrier_errors +
+ netdev->stats.tx_fifo_errors +
+ h_stats.tx_logic_errors +
+ tx_logic_errors;
+
+ netdev->stats.tx_errors = adapter->err_cntrs.tx_smac_filt +
+ adapter->err_cntrs.tx_dlid_zero +
+ adapter->err_cntrs.tx_drop_state +
+ adapter->err_cntrs.tx_logic;
+
+ netdev->stats.tx_dropped = netdev->stats.tx_errors;
+ adapter->sum_cntrs.tx_errors = netdev->stats.tx_errors;
+ adapter->err_cntrs.tx_errors = netdev->stats.tx_errors;
+
+ /* update rx counters */
+ netdev->stats.rx_packets = adapter->sum_cntrs.rx_packets;
+ netdev->stats.rx_bytes = adapter->sum_cntrs.rx_bytes;
+ netdev->stats.multicast = adapter->sum_cntrs.rx_grp.mcastbcast;
+ netdev->stats.rx_over_errors = adapter->err_cntrs.rx_oversize;
+ netdev->stats.rx_length_errors = adapter->err_cntrs.rx_oversize +
+ adapter->err_cntrs.rx_runt;
+
+ adapter->err_cntrs.rx_logic = netdev->stats.rx_missed_errors +
+ netdev->stats.rx_fifo_errors +
+ h_stats.rx_logic_errors +
+ rx_logic_errors;
+
+ netdev->stats.rx_errors = adapter->err_cntrs.rx_bad_veswid +
+ adapter->err_cntrs.rx_runt +
+ adapter->err_cntrs.rx_oversize +
+ adapter->err_cntrs.rx_eth_down +
+ adapter->err_cntrs.rx_drop_state +
+ adapter->err_cntrs.rx_logic;
+
+ netdev->stats.rx_dropped = netdev->stats.rx_errors;
+ adapter->sum_cntrs.rx_errors = netdev->stats.rx_errors;
+ adapter->err_cntrs.rx_errors = netdev->stats.rx_errors;
+}
+
+/* update_len_counters - update pkt's len histogram counters */
+static inline void update_len_counters(struct __hfi_vnic_group_scs *grp,
+ int len)
+{
+ /* account for 4 byte FCS */
+ if (len >= 1515)
+ grp->xx_1519_max++;
+ else if (len >= 1020)
+ grp->xx_1024_1518++;
+ else if (len >= 508)
+ grp->xx_512_1023++;
+ else if (len >= 252)
+ grp->xx_256_511++;
+ else if (len >= 124)
+ grp->xx_128_255++;
+ else if (len >= 61)
+ grp->xx_65_127++;
+ else
+ grp->xx_64_size++;
+}
+
+/* hfi_vnic_update_tx_counters - update transmit counters */
+static void hfi_vnic_update_tx_counters(struct net_device *netdev, u8 q_idx,
+ struct sk_buff *skb, int err)
+{
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ struct __hfi_vnic_group_scs *grp_cntrs =
+ &adapter->q_sum_cntrs[q_idx].tx_grp;
+ u16 vlan_tci;
+
+ adapter->q_sum_cntrs[q_idx].tx_packets++;
+ adapter->q_sum_cntrs[q_idx].tx_bytes += skb->len + ETH_FCS_LEN;
+
+ update_len_counters(grp_cntrs, skb->len);
+
+ /* rest of the counts are for good packets only */
+ if (err)
+ return;
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ grp_cntrs->mcastbcast++;
+ else
+ grp_cntrs->unicast++;
+
+ if (!__vlan_get_tag(skb, &vlan_tci))
+ grp_cntrs->vlan++;
+ else
+ grp_cntrs->untagged++;
+}
+
+/* hfi_vnic_update_rx_counters - update receive counters */
+static void hfi_vnic_update_rx_counters(struct net_device *netdev, u8 q_idx,
+ struct sk_buff *skb, int err)
+{
+ struct ethhdr *mac_hdr = (struct ethhdr *)skb->data;
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ struct __hfi_vnic_group_scs *grp_cntrs =
+ &adapter->q_sum_cntrs[q_idx].rx_grp;
+ u16 vlan_tci;
+
+ adapter->q_sum_cntrs[q_idx].rx_packets++;
+ adapter->q_sum_cntrs[q_idx].rx_bytes += skb->len + ETH_FCS_LEN;
+
+ update_len_counters(grp_cntrs, skb->len);
+
+ /* rest of the counts are for good packets only */
+ if (err)
+ return;
+
+ if (is_multicast_ether_addr(mac_hdr->h_dest))
+ grp_cntrs->mcastbcast++;
+ else
+ grp_cntrs->unicast++;
+
+ if (!__vlan_get_tag(skb, &vlan_tci))
+ grp_cntrs->vlan++;
+ else
+ grp_cntrs->untagged++;
+}
+
+static struct rtnl_link_stats64 *
+hfi_vnic_get_stats64(struct net_device *netdev,
+ struct rtnl_link_stats64 *stats)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+
+ mutex_lock(&adapter->stats_lock);
+ hfi_vnic_update_stats(netdev);
+
+ stats->rx_packets = netdev->stats.rx_packets;
+ stats->tx_packets = netdev->stats.tx_packets;
+ stats->rx_bytes = netdev->stats.rx_bytes;
+ stats->tx_bytes = netdev->stats.tx_bytes;
+ stats->rx_errors = netdev->stats.rx_errors;
+ stats->tx_errors = netdev->stats.tx_errors;
+ stats->rx_dropped = netdev->stats.rx_dropped;
+ stats->tx_dropped = netdev->stats.tx_dropped;
+ stats->multicast = netdev->stats.multicast;
+ stats->rx_length_errors = netdev->stats.rx_length_errors;
+ stats->rx_over_errors = netdev->stats.rx_over_errors;
+ stats->rx_fifo_errors = netdev->stats.rx_fifo_errors;
+ stats->rx_missed_errors = netdev->stats.rx_missed_errors;
+ stats->tx_carrier_errors = netdev->stats.tx_carrier_errors;
+ stats->tx_fifo_errors = netdev->stats.tx_fifo_errors;
+ mutex_unlock(&adapter->stats_lock);
+ return stats;
+}
+
/* hfi_vnic_maybe_stop_tx - stop tx queue if required */
static void hfi_vnic_maybe_stop_tx(struct hfi_vnic_adapter *adapter, u8 q_idx)
{
@@ -67,6 +296,7 @@ static void hfi_vnic_maybe_stop_tx(struct hfi_vnic_adapter *adapter, u8 q_idx)
if (!vport->ops->get_write_avail(vport, q_idx))
return;
+ adapter->q_tx_restart[q_idx]++;
netif_start_subqueue(vport->netdev, q_idx);
}
@@ -82,12 +312,15 @@ static netdev_tx_t hfi_netdev_start_xmit(struct sk_buff *skb,
v_dbg("xmit: queue %d skb len %d\n", q_idx, skb->len);
if (unlikely(adapter->info.vport.oper_state !=
- HFI_VNIC_STATE_FORWARDING))
+ HFI_VNIC_STATE_FORWARDING)) {
+ adapter->q_err_cntrs[q_idx].tx_drop_state++;
goto tx_finish;
+ }
/* pad to ensure mininum ethernet packet length */
if (unlikely(skb->len < ETH_ZLEN)) {
if (skb_padto(skb, ETH_ZLEN)) {
+ adapter->q_tx_logic_errors[q_idx]++;
skip_skb_free = true;
goto tx_finish;
}
@@ -101,16 +334,19 @@ static netdev_tx_t hfi_netdev_start_xmit(struct sk_buff *skb,
/* Get reference to skb as hfi driver might release it */
skb_get(skb);
rc = vport->ops->put_skb(vport, q_idx, skb);
- /* remove the header */
+ /* remove the header before updating tx counters */
skb_pull(skb, HFI_VNIC_HDR_LEN);
tx_finish:
if (unlikely(rc == -EBUSY)) {
hfi_vnic_maybe_stop_tx(adapter, q_idx);
+ adapter->q_tx_halt[q_idx]++;
dev_kfree_skb_any(skb);
return NETDEV_TX_BUSY;
}
+ /* update tx counters */
+ hfi_vnic_update_tx_counters(netdev, q_idx, skb, rc);
if (!skip_skb_free)
dev_kfree_skb_any(skb);
return NETDEV_TX_OK;
@@ -123,6 +359,7 @@ static void vnic_handle_rx(struct hfi_vnic_rx_queue *rxq,
struct hfi_vnic_adapter *adapter = rxq->adapter;
struct hfi_vnic_port *vport = adapter->vport;
struct sk_buff *skb;
+ int rc;
while (1) {
if (*work_done >= work_to_do)
@@ -132,7 +369,11 @@ static void vnic_handle_rx(struct hfi_vnic_rx_queue *rxq,
if (!skb)
break;
- if (hfi_vnic_decap_skb(rxq, skb)) {
+ rc = hfi_vnic_decap_skb(rxq, skb);
+
+ /* update rx counters */
+ hfi_vnic_update_rx_counters(adapter->netdev, rxq->idx, skb, rc);
+ if (rc) {
dev_kfree_skb_any(skb);
continue;
}
@@ -178,8 +419,10 @@ static void vnic_event_cb(struct hfi_vnic_port *vport, u8 evt)
if (evt < vport->hfi_info.num_rx_q) {
q_idx = evt;
if (unlikely(adapter->info.vport.oper_state !=
- HFI_VNIC_STATE_FORWARDING))
+ HFI_VNIC_STATE_FORWARDING)) {
+ adapter->q_err_cntrs[q_idx].rx_drop_state++;
return;
+ }
rxq = &adapter->rxq[q_idx];
if (napi_schedule_prep(&rxq->napi)) {
@@ -193,9 +436,10 @@ static void vnic_event_cb(struct hfi_vnic_port *vport, u8 evt)
(evt < (HFI_VNIC_EVT_TX0 + vport->hfi_info.num_tx_q))) {
q_idx = evt - HFI_VNIC_EVT_TX0;
- if (__netif_subqueue_stopped(vport->netdev, q_idx))
+ if (__netif_subqueue_stopped(vport->netdev, q_idx)) {
netif_wake_subqueue(vport->netdev, q_idx);
-
+ adapter->q_tx_wakeup[q_idx]++;
+ }
return;
}
v_err("Invalid event\n");
@@ -340,6 +584,7 @@ static int hfi_netdev_close(struct net_device *netdev)
.ndo_stop = hfi_netdev_close,
.ndo_start_xmit = hfi_netdev_start_xmit,
.ndo_change_mtu = hfi_netdev_change_mtu,
+ .ndo_get_stats64 = hfi_vnic_get_stats64,
.ndo_select_queue = hfi_vnic_select_queue,
.ndo_set_mac_address = hfi_vnic_set_mac_addr,
};
@@ -371,6 +616,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
netdev->netdev_ops = &hfi_netdev_ops;
netdev->hard_header_len += HFI_VNIC_SKB_HEADROOM;
mutex_init(&adapter->lock);
+ mutex_init(&adapter->stats_lock);
strcpy(netdev->name, "veth%d");
SET_NETDEV_DEV(netdev, parent);
@@ -392,6 +638,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
return adapter;
netdev_err:
mutex_destroy(&adapter->lock);
+ mutex_destroy(&adapter->stats_lock);
free_netdev(netdev);
return ERR_PTR(rc);
@@ -405,5 +652,6 @@ void hfi_vnic_rem_netdev(struct hfi_vnic_port *vport)
v_info("removing\n");
unregister_netdev(vport->netdev);
mutex_destroy(&adapter->lock);
+ mutex_destroy(&adapter->stats_lock);
free_netdev(vport->netdev);
}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 41+ messages in thread
* [RFC v2 06/10] IB/hfi-vnic: VNIC MAC table support
[not found] ` <1481788782-89964-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
` (2 preceding siblings ...)
2016-12-15 7:59 ` [RFC v2 04/10] IB/hfi-vnic: VNIC Ethernet Management (EM) structure definitions Vishwanathapura, Niranjana
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 07/10] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) interface Vishwanathapura, Niranjana
2016-12-15 9:12 ` [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Leon Romanovsky
5 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Niranjana Vishwanathapura,
Sadanand Warrier
HFI VNIC MAC table contains the MAC address to DLID mappings provided by
the Ethernet manager. During transmission, the MAC table provides the MAC
address to DLID translation. Implement MAC table using simple hash list.
Also provide support to update/query the MAC table by Ethernet manager.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Sadanand Warrier <sadanand.warrier-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c | 236 +++++++++++++++++++++
.../sw/intel/hfi_vnic/hfi_vnic_internal.h | 53 ++++-
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c | 4 +
3 files changed, 292 insertions(+), 1 deletion(-)
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
index 3fdfb7b..e45cff8 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.c
@@ -104,6 +104,238 @@
#define HFI_VNIC_SC_MASK 0x1f
+/*
+ * Using a simple hash table for mac table implementation with the last octet
+ * of mac address as a key.
+ */
+static void hfi_vnic_free_mac_tbl(struct hlist_head *mactbl)
+{
+ struct hfi_vnic_mac_tbl_node *node;
+ struct hlist_node *tmp;
+ int bkt;
+
+ if (!mactbl)
+ return;
+
+ vnic_hash_for_each_safe(mactbl, bkt, tmp, node, hlist) {
+ hash_del(&node->hlist);
+ kfree(node);
+ }
+ kfree(mactbl);
+}
+
+static struct hlist_head *hfi_vnic_alloc_mac_tbl(void)
+{
+ u32 size = sizeof(struct hlist_head) * HFI_VNIC_MAC_TBL_SIZE;
+ struct hlist_head *mactbl;
+
+ mactbl = kzalloc(size, GFP_KERNEL);
+ if (!mactbl)
+ return ERR_PTR(-ENOMEM);
+
+ vnic_hash_init(mactbl);
+ return mactbl;
+}
+
+/* hfi_vnic_release_mac_tbl - empty and free the mac table */
+void hfi_vnic_release_mac_tbl(struct hfi_vnic_adapter *adapter)
+{
+ struct hlist_head *mactbl;
+
+ mutex_lock(&adapter->mactbl_lock);
+ mactbl = rcu_access_pointer(adapter->mactbl);
+ rcu_assign_pointer(adapter->mactbl, NULL);
+ synchronize_rcu();
+ hfi_vnic_free_mac_tbl(mactbl);
+ mutex_unlock(&adapter->mactbl_lock);
+}
+
+/*
+ * hfi_vnic_query_mac_tbl - query the mac table for a section
+ *
+ * This function implements query of specific function of the mac table.
+ * The function also expects the requested range to be valid.
+ */
+void hfi_vnic_query_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_mactable *tbl)
+{
+ struct hfi_vnic_mac_tbl_node *node;
+ struct hlist_head *mactbl;
+ int bkt;
+ u16 loffset, lnum_entries;
+
+ rcu_read_lock();
+ mactbl = rcu_dereference(adapter->mactbl);
+ if (!mactbl)
+ goto get_mac_done;
+
+ loffset = be16_to_cpu(tbl->offset);
+ lnum_entries = be16_to_cpu(tbl->num_entries);
+
+ vnic_hash_for_each(mactbl, bkt, node, hlist) {
+ struct __hfi_vnic_mactable_entry *nentry = &node->entry;
+ struct hfi_veswport_mactable_entry *entry;
+
+ if ((node->index < loffset) ||
+ (node->index >= (loffset + lnum_entries)))
+ continue;
+
+ /* populate entry in the tbl corresponding to the index */
+ entry = &tbl->tbl_entries[node->index - loffset];
+ memcpy(entry->mac_addr, nentry->mac_addr,
+ ARRAY_SIZE(entry->mac_addr));
+ memcpy(entry->mac_addr_mask, nentry->mac_addr_mask,
+ ARRAY_SIZE(entry->mac_addr_mask));
+ entry->dlid_sd.dw = cpu_to_be32(nentry->dlid_sd.dw);
+ }
+ tbl->mac_tbl_digest = cpu_to_be32(adapter->info.vport.mac_tbl_digest);
+get_mac_done:
+ rcu_read_unlock();
+}
+
+/*
+ * hfi_vnic_update_mac_tbl - update mac table section
+ *
+ * This function updates the specified section of the mac table.
+ * The procedure includes following steps.
+ * - Allocate a new mac (hash) table.
+ * - Add the specified entries to the new table.
+ * (except the ones that are requested to be deleted).
+ * - Add all the other entries from the old mac table.
+ * - If there is a failure, free the new table and return.
+ * - Switch to the new table.
+ * - Free the old table and return.
+ *
+ * The function also expects the requested range to be valid.
+ */
+int hfi_vnic_update_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_mactable *tbl)
+{
+ struct hfi_vnic_mac_tbl_node *node, *new_node;
+ struct hlist_head *new_mactbl, *old_mactbl;
+ int i, bkt, rc = 0;
+ u8 key;
+ u16 loffset, lnum_entries;
+
+ mutex_lock(&adapter->mactbl_lock);
+ /* allocate new mac table */
+ new_mactbl = hfi_vnic_alloc_mac_tbl();
+ if (IS_ERR(new_mactbl)) {
+ mutex_unlock(&adapter->mactbl_lock);
+ return PTR_ERR(new_mactbl);
+ }
+
+ loffset = be16_to_cpu(tbl->offset);
+ lnum_entries = be16_to_cpu(tbl->num_entries);
+
+ /* add updated entries to the new mac table */
+ for (i = 0; i < lnum_entries; i++) {
+ struct __hfi_vnic_mactable_entry *nentry;
+ struct hfi_veswport_mactable_entry *entry =
+ &tbl->tbl_entries[i];
+ u8 *mac_addr = entry->mac_addr;
+ u8 empty_mac[ETH_ALEN] = { 0 };
+
+ v_dbg("new mac entry %4d: %02x:%02x:%02x:%02x:%02x:%02x %x\n",
+ loffset + i, mac_addr[0], mac_addr[1], mac_addr[2],
+ mac_addr[3], mac_addr[4], mac_addr[5],
+ entry->dlid_sd.dw);
+
+ /* if the entry is being removed, do not add it */
+ if (!memcmp(mac_addr, empty_mac, ARRAY_SIZE(empty_mac)))
+ continue;
+
+ node = kzalloc(sizeof(*node), GFP_KERNEL);
+ if (!node) {
+ rc = -ENOMEM;
+ goto updt_done;
+ }
+
+ node->index = loffset + i;
+ nentry = &node->entry;
+ memcpy(nentry->mac_addr, entry->mac_addr,
+ ARRAY_SIZE(nentry->mac_addr));
+ memcpy(nentry->mac_addr_mask, entry->mac_addr_mask,
+ ARRAY_SIZE(nentry->mac_addr_mask));
+ nentry->dlid_sd.dw = be32_to_cpu(entry->dlid_sd.dw);
+ key = node->entry.mac_addr[HFI_VNIC_MAC_HASH_IDX];
+ vnic_hash_add(new_mactbl, &node->hlist, key);
+ }
+
+ /* add other entries from current mac table to new mac table */
+ old_mactbl = rcu_access_pointer(adapter->mactbl);
+ if (!old_mactbl)
+ goto switch_tbl;
+
+ vnic_hash_for_each(old_mactbl, bkt, node, hlist) {
+ if ((node->index >= loffset) &&
+ (node->index < (loffset + lnum_entries)))
+ continue;
+
+ new_node = kzalloc(sizeof(*new_node), GFP_KERNEL);
+ if (!new_node) {
+ rc = -ENOMEM;
+ goto updt_done;
+ }
+
+ new_node->index = node->index;
+ memcpy(&new_node->entry, &node->entry, sizeof(node->entry));
+ key = new_node->entry.mac_addr[HFI_VNIC_MAC_HASH_IDX];
+ vnic_hash_add(new_mactbl, &new_node->hlist, key);
+ }
+
+switch_tbl:
+ /* switch to new table */
+ rcu_assign_pointer(adapter->mactbl, new_mactbl);
+ synchronize_rcu();
+
+ adapter->info.vport.mac_tbl_digest = be32_to_cpu(tbl->mac_tbl_digest);
+updt_done:
+ /* upon failure, free the new table; otherwise, free the old table */
+ if (rc)
+ hfi_vnic_free_mac_tbl(new_mactbl);
+ else
+ hfi_vnic_free_mac_tbl(old_mactbl);
+
+ mutex_unlock(&adapter->mactbl_lock);
+ return rc;
+}
+
+/* hfi_vnic_chk_mac_tbl - check mac table for dlid */
+static uint32_t hfi_vnic_chk_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct ethhdr *mac_hdr)
+{
+ struct hfi_vnic_mac_tbl_node *node;
+ struct hlist_head *mactbl;
+ u32 dlid = 0;
+ u8 key;
+
+ rcu_read_lock();
+ mactbl = rcu_dereference(adapter->mactbl);
+ if (!mactbl)
+ goto chk_done;
+
+ key = mac_hdr->h_dest[HFI_VNIC_MAC_HASH_IDX];
+ vnic_hash_for_each_possible(mactbl, node, hlist, key) {
+ struct __hfi_vnic_mactable_entry *entry = &node->entry;
+
+ /* if related to source mac, skip */
+ if (entry->dlid_sd.sd_is_src_mac)
+ continue;
+
+ if (!memcmp(node->entry.mac_addr, mac_hdr->h_dest,
+ ARRAY_SIZE(node->entry.mac_addr))) {
+ /* mac address found */
+ dlid = node->entry.dlid_sd.dlid;
+ break;
+ }
+ }
+
+chk_done:
+ rcu_read_unlock();
+ return dlid;
+}
+
/* hfi_vnic_get_dlid - find and return the DLID */
static uint32_t hfi_vnic_get_dlid(struct hfi_vnic_adapter *adapter,
struct sk_buff *skb, u8 def_port)
@@ -112,6 +344,10 @@ static uint32_t hfi_vnic_get_dlid(struct hfi_vnic_adapter *adapter,
struct ethhdr *mac_hdr = (struct ethhdr *)skb_mac_header(skb);
u32 dlid;
+ dlid = hfi_vnic_chk_mac_tbl(adapter, mac_hdr);
+ if (dlid)
+ return dlid;
+
if (is_multicast_ether_addr(mac_hdr->h_dest)) {
dlid = info->vesw.u_mcast_dlid;
} else {
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
index af3ff0e..6d5c5f8 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
@@ -262,6 +262,8 @@ struct hfi_vnic_rx_queue {
* @lock: adapter lock
* @rxq: receive queue array
* @info: virtual ethernet switch port information
+ * @mactbl: hash table of MAC entries
+ * @mactbl_lock: mac table lock
* @stats_lock: statistics lock
* @flow_tbl: flow to default port redirection table
* @q_sum_cntrs: per queue EM summary counters
@@ -284,7 +286,11 @@ struct hfi_vnic_adapter {
struct hfi_vnic_rx_queue rxq[HFI_VNIC_MAX_QUEUE];
- struct __hfi_veswport_info info;
+ struct __hfi_veswport_info info;
+ struct hlist_head __rcu *mactbl;
+
+ /* Lock used to protect updates to mac table */
+ struct mutex mactbl_lock;
/* Lock used to protect access to vnic counters */
struct mutex stats_lock;
@@ -304,6 +310,25 @@ struct hfi_vnic_adapter {
struct __hfi_vnic_error_counters err_cntrs;
};
+/* Same as hfi_veswport_mactable_entry, but without bitwise attribute */
+struct __hfi_vnic_mactable_entry {
+ u8 mac_addr[ETH_ALEN];
+ u8 mac_addr_mask[ETH_ALEN];
+ union __hfi_vnic_dlid_sd dlid_sd;
+} __packed;
+
+/**
+ * struct hfi_vnic_mac_tbl_node - HFI VNIC mac table node
+ * @hlist: hash list handle
+ * @index: index of entry in the mac table
+ * @entry: entry in the table
+ */
+struct hfi_vnic_mac_tbl_node {
+ struct hlist_node hlist;
+ u16 index;
+ struct __hfi_vnic_mactable_entry entry;
+};
+
#define v_dbg(format, arg...) \
netdev_dbg(adapter->netdev, format, ## arg)
#define v_err(format, arg...) \
@@ -325,12 +350,38 @@ struct hfi_vnic_adapter {
#define HFI_VNIC_MAC_TBL_HASH_BITS 8
#define HFI_VNIC_MAC_TBL_SIZE BIT(HFI_VNIC_MAC_TBL_HASH_BITS)
+/* VNIC HASH MACROS */
+#define vnic_hash_init(hashtable) __hash_init(hashtable, HFI_VNIC_MAC_TBL_SIZE)
+
+#define vnic_hash_add(hashtable, node, key) \
+ hlist_add_head(node, \
+ &hashtable[hash_min(key, ilog2(HFI_VNIC_MAC_TBL_SIZE))])
+
+#define vnic_hash_for_each_safe(name, bkt, tmp, obj, member) \
+ for ((bkt) = 0, obj = NULL; \
+ !obj && (bkt) < HFI_VNIC_MAC_TBL_SIZE; (bkt)++) \
+ hlist_for_each_entry_safe(obj, tmp, &name[bkt], member)
+
+#define vnic_hash_for_each_possible(name, obj, member, key) \
+ hlist_for_each_entry(obj, \
+ &name[hash_min(key, ilog2(HFI_VNIC_MAC_TBL_SIZE))], member)
+
+#define vnic_hash_for_each(name, bkt, obj, member) \
+ for ((bkt) = 0, obj = NULL; \
+ !obj && (bkt) < HFI_VNIC_MAC_TBL_SIZE; (bkt)++) \
+ hlist_for_each_entry(obj, &name[bkt], member)
+
struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
struct device *parent);
void hfi_vnic_rem_netdev(struct hfi_vnic_port *vport);
int hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
int hfi_vnic_decap_skb(struct hfi_vnic_rx_queue *rxq, struct sk_buff *skb);
u8 hfi_vnic_calc_entropy(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+void hfi_vnic_release_mac_tbl(struct hfi_vnic_adapter *adapter);
+void hfi_vnic_query_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_mactable *tbl);
+int hfi_vnic_update_mac_tbl(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_mactable *tbl);
void hfi_vnic_update_stats(struct net_device *netdev);
void hfi_vnic_set_ethtool_ops(struct net_device *netdev);
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
index 1626e44..04edafa 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
@@ -616,6 +616,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
netdev->netdev_ops = &hfi_netdev_ops;
netdev->hard_header_len += HFI_VNIC_SKB_HEADROOM;
mutex_init(&adapter->lock);
+ mutex_init(&adapter->mactbl_lock);
mutex_init(&adapter->stats_lock);
strcpy(netdev->name, "veth%d");
@@ -638,6 +639,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
return adapter;
netdev_err:
mutex_destroy(&adapter->lock);
+ mutex_destroy(&adapter->mactbl_lock);
mutex_destroy(&adapter->stats_lock);
free_netdev(netdev);
@@ -651,7 +653,9 @@ void hfi_vnic_rem_netdev(struct hfi_vnic_port *vport)
v_info("removing\n");
unregister_netdev(vport->netdev);
+ hfi_vnic_release_mac_tbl(adapter);
mutex_destroy(&adapter->lock);
+ mutex_destroy(&adapter->mactbl_lock);
mutex_destroy(&adapter->stats_lock);
free_netdev(vport->netdev);
}
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 41+ messages in thread
* [RFC v2 07/10] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) interface
[not found] ` <1481788782-89964-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
` (3 preceding siblings ...)
2016-12-15 7:59 ` [RFC v2 06/10] IB/hfi-vnic: VNIC MAC table support Vishwanathapura, Niranjana
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
2016-12-15 9:12 ` [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Leon Romanovsky
5 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Niranjana Vishwanathapura,
Sadanand Warrier, Tanya K Jajodia
HFI VNIC EMA interface functions are the management interfaces to the HFI
VNIC netdev. Add support to add and remove VNIC ports. Implement the
required GET/SET management interface functions and processing of new
management information. Add support to send trap notifications upon various
events like interface status change, unicast/multicast mac list update and
mac address change.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Reviewed-by: Ira Weiny <ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Sadanand Warrier <sadanand.warrier-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Tanya K Jajodia <tanya.k.jajodia-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
drivers/infiniband/sw/intel/hfi_vnic/Makefile | 3 +-
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h | 4 +
.../sw/intel/hfi_vnic/hfi_vnic_internal.h | 44 +++
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c | 153 +++++++-
.../sw/intel/hfi_vnic/hfi_vnic_vema_iface.c | 432 +++++++++++++++++++++
5 files changed, 633 insertions(+), 3 deletions(-)
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema_iface.c
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/Makefile b/drivers/infiniband/sw/intel/hfi_vnic/Makefile
index 8e3dca7..a0562af 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/Makefile
+++ b/drivers/infiniband/sw/intel/hfi_vnic/Makefile
@@ -3,4 +3,5 @@
#
obj-$(CONFIG_HFI_VNIC) += hfi_vnic.o
-hfi_vnic-y := hfi_vnic_netdev.o hfi_vnic_encap.o hfi_vnic_ethtool.o
+hfi_vnic-y := hfi_vnic_netdev.o hfi_vnic_encap.o hfi_vnic_ethtool.o \
+ hfi_vnic_vema_iface.o
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
index a6770ef..54e9081 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_encap.h
@@ -99,6 +99,10 @@
#define HFI_VNIC_STATE_DROP_ALL 0x1
#define HFI_VNIC_STATE_FORWARDING 0x3
+/* VNIC Ethernet link status */
+#define HFI_VNIC_ETH_LINK_UP 1
+#define HFI_VNIC_ETH_LINK_DOWN 2
+
/**
* struct hfi_vesw_info - HFI vnic switch information
* @fabric_id: 10-bit fabric id
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
index 6d5c5f8..7723a4e 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
@@ -243,6 +243,16 @@ struct __hfi_veswport_trap {
} __packed;
/**
+ * struct hfi_vnic_ctrl_port - HFI virtual NIC control port
+ * @ibdev: pointer to ib device
+ * @ops: hfi vnic control operations
+ */
+struct hfi_vnic_ctrl_port {
+ struct ib_device *ibdev;
+ struct hfi_vnic_ctrl_ops *ops;
+};
+
+/**
* struct hfi_vnic_rx_queue - HFI VNIC receive queue
* @idx: queue index
* @adapter: netdev adapter
@@ -257,11 +267,15 @@ struct hfi_vnic_rx_queue {
/**
* struct hfi_vnic_adapter - HFI VNIC netdev private data structure
* @netdev: pointer to associated netdev
+ * @cport: pointer to hfi vnic control port
* @vport: pointer to hfi vnic port
* @flags: flags indicating various states
* @lock: adapter lock
* @rxq: receive queue array
* @info: virtual ethernet switch port information
+ * @vema_mac_addr: mac address configured by vema
+ * @umac_hash: unicast maclist hash
+ * @mmac_hash: multicast maclist hash
* @mactbl: hash table of MAC entries
* @mactbl_lock: mac table lock
* @stats_lock: statistics lock
@@ -278,6 +292,7 @@ struct hfi_vnic_rx_queue {
*/
struct hfi_vnic_adapter {
struct net_device *netdev;
+ struct hfi_vnic_ctrl_port *cport;
struct hfi_vnic_port *vport;
unsigned long flags;
@@ -287,6 +302,9 @@ struct hfi_vnic_adapter {
struct hfi_vnic_rx_queue rxq[HFI_VNIC_MAX_QUEUE];
struct __hfi_veswport_info info;
+ u8 vema_mac_addr[ETH_ALEN];
+ u32 umac_hash;
+ u32 mmac_hash;
struct hlist_head __rcu *mactbl;
/* Lock used to protect updates to mac table */
@@ -338,6 +356,11 @@ struct hfi_vnic_mac_tbl_node {
#define v_warn(format, arg...) \
netdev_warn(adapter->netdev, format, ## arg)
+#define c_err(format, arg...) \
+ dev_err(&cport->ibdev->dev, format, ## arg)
+#define c_info(format, arg...) \
+ dev_info(&cport->ibdev->dev, format, ## arg)
+
/* The maximum allowed entries in the mac table */
#define HFI_VNIC_MAC_TBL_MAX_ENTRIES 2048
/* Limit of smac entries in mac table */
@@ -377,12 +400,33 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
int hfi_vnic_encap_skb(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
int hfi_vnic_decap_skb(struct hfi_vnic_rx_queue *rxq, struct sk_buff *skb);
u8 hfi_vnic_calc_entropy(struct hfi_vnic_adapter *adapter, struct sk_buff *skb);
+void hfi_vnic_process_vema_config(struct hfi_vnic_adapter *adapter);
void hfi_vnic_release_mac_tbl(struct hfi_vnic_adapter *adapter);
void hfi_vnic_query_mac_tbl(struct hfi_vnic_adapter *adapter,
struct hfi_veswport_mactable *tbl);
int hfi_vnic_update_mac_tbl(struct hfi_vnic_adapter *adapter,
struct hfi_veswport_mactable *tbl);
+void hfi_vnic_query_ucast_macs(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_iface_macs *macs);
+void hfi_vnic_query_mcast_macs(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_iface_macs *macs);
void hfi_vnic_update_stats(struct net_device *netdev);
+void hfi_vnic_get_summary_counters(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_summary_counters *cntrs);
+void hfi_vnic_get_error_counters(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_error_counters *cntrs);
+void hfi_vnic_get_vesw_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_vesw_info *info);
+void hfi_vnic_set_vesw_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_vesw_info *info);
+void hfi_vnic_get_per_veswport_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_per_veswport_info *info);
+void hfi_vnic_set_per_veswport_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_per_veswport_info *info);
+void hfi_vnic_vema_report_event(struct hfi_vnic_adapter *adapter, u8 event);
+struct hfi_vnic_adapter *hfi_vnic_add_vport(struct hfi_vnic_ctrl_port *cport,
+ u8 port_num, u8 vport_num);
+void hfi_vnic_rem_vport(struct hfi_vnic_adapter *adapter);
void hfi_vnic_set_ethtool_ops(struct net_device *netdev);
#endif /* _HFI_VNIC_INTERNAL_H */
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
index 04edafa..7e58c1c 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_netdev.c
@@ -51,6 +51,7 @@
#include <linux/module.h>
#include <linux/if_vlan.h>
+#include <linux/crc32.h>
#include "hfi_vnic_internal.h"
@@ -528,7 +529,90 @@ static void hfi_vnic_down(struct hfi_vnic_adapter *adapter)
clear_bit(HFI_VNIC_UP, &adapter->flags);
}
-/* hfi_vnic_set_mac_addr - change mac address */
+/* hfi_vnic_process_vema_config - process vema configuration updates */
+void hfi_vnic_process_vema_config(struct hfi_vnic_adapter *adapter)
+{
+ struct __hfi_veswport_info *info = &adapter->info;
+ struct hfi_vnic_port *vport = adapter->vport;
+ u8 port_num[HFI_VESW_MAX_NUM_DEF_PORT] = { 0 };
+ u16 port_mask, mtu_limit = ETH_ZLEN - ETH_HLEN;
+ u8 i, port_count = 0;
+
+ /* If the base_mac_addr is changed, update the interface mac address */
+ if (memcmp(info->vport.base_mac_addr, adapter->vema_mac_addr,
+ ARRAY_SIZE(info->vport.base_mac_addr))) {
+ struct sockaddr saddr;
+
+ memcpy(saddr.sa_data, info->vport.base_mac_addr,
+ ARRAY_SIZE(info->vport.base_mac_addr));
+ mutex_lock(&adapter->lock);
+ eth_mac_addr(adapter->netdev, &saddr);
+ memcpy(adapter->vema_mac_addr,
+ info->vport.base_mac_addr, ETH_ALEN);
+ mutex_unlock(&adapter->lock);
+ }
+
+ /*
+ * If vesw_id is being changed, and if the vnic interface
+ * is up, reset the hfi interface to ensure new vesw_id
+ * is picked by hfi driver
+ */
+ if (vport->vesw_id != info->vesw.vesw_id) {
+ mutex_lock(&adapter->lock);
+ if (test_bit(HFI_VNIC_UP, &adapter->flags))
+ hfi_vnic_down(adapter);
+
+ vport->vesw_id = info->vesw.vesw_id;
+ if (test_bit(HFI_VNIC_OPEN, &adapter->flags))
+ hfi_vnic_up(adapter);
+
+ mutex_unlock(&adapter->lock);
+ }
+
+ /* Handle MTU limit change */
+ mtu_limit = max(info->vesw.eth_mtu_non_vlan, mtu_limit);
+ rtnl_lock();
+ if (adapter->netdev->mtu > mtu_limit)
+ dev_set_mtu(adapter->netdev, mtu_limit);
+ rtnl_unlock();
+
+ /* Update flow to default port redirection table */
+ port_mask = info->vesw.def_port_mask;
+ for (i = 0; i < HFI_VESW_MAX_NUM_DEF_PORT; i++) {
+ if (port_mask & 1)
+ port_num[port_count++] = i;
+ port_mask >>= 1;
+ }
+
+ /*
+ * Build the flow table. Flow table is required when destination LID
+ * is not available. Up to HFI_VNIC_FLOW_TBL_SIZE flows supported.
+ * Each flow need a default port number to get its dlid from the
+ * u_ucast_dlid array.
+ */
+ for (i = 0; i < HFI_VNIC_FLOW_TBL_SIZE; i++)
+ adapter->flow_tbl[i] = port_count ? port_num[i % port_count] :
+ HFI_VNIC_INVALID_PORT;
+
+ /* Operational state can only be DROP_ALL or FORWARDING */
+ if (info->vport.config_state == HFI_VNIC_STATE_FORWARDING)
+ info->vport.oper_state = HFI_VNIC_STATE_FORWARDING;
+ else
+ info->vport.oper_state = HFI_VNIC_STATE_DROP_ALL;
+}
+
+/*
+ * Set the power on default values in adapter's vema interface structure.
+ */
+static inline void hfi_vnic_set_pod_values(struct hfi_vnic_adapter *adapter)
+{
+ adapter->info.vport.max_mac_tbl_ent = HFI_VNIC_MAC_TBL_MAX_ENTRIES;
+ adapter->info.vport.max_smac_ent = HFI_VNIC_MAX_SMAC_LIMIT;
+ adapter->info.vport.config_state = HFI_VNIC_STATE_DROP_ALL;
+ adapter->info.vport.eth_link_status = HFI_VNIC_ETH_LINK_DOWN;
+}
+
+/* hfi_vnic_set_mac_addr - change mac address and send trap */
static int hfi_vnic_set_mac_addr(struct net_device *netdev, void *addr)
{
struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
@@ -541,8 +625,62 @@ static int hfi_vnic_set_mac_addr(struct net_device *netdev, void *addr)
mutex_lock(&adapter->lock);
rc = eth_mac_addr(netdev, addr);
mutex_unlock(&adapter->lock);
+ if (rc)
+ return rc;
- return rc;
+ adapter->info.vport.uc_macs_gen_count++;
+ hfi_vnic_vema_report_event(adapter,
+ HFI_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE);
+ return 0;
+}
+
+/*
+ * hfi_vnic_mac_send_event - post event on possible mac list exchange
+ * Send trap when digest from uc/mc mac list differs from previous run.
+ * Digest is evaluated similar to how cksum does.
+ */
+static void hfi_vnic_mac_send_event(struct net_device *netdev, u8 event)
+{
+ struct hfi_vnic_adapter *adapter = netdev_priv(netdev);
+ struct netdev_hw_addr *ha;
+ struct netdev_hw_addr_list *hw_list;
+ u32 *ref_crc;
+ u32 l, crc = 0;
+
+ switch (event) {
+ case HFI_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE:
+ hw_list = &netdev->uc;
+ adapter->info.vport.uc_macs_gen_count++;
+ ref_crc = &adapter->umac_hash;
+ break;
+ case HFI_VESWPORT_TRAP_IFACE_MCAST_MAC_CHANGE:
+ hw_list = &netdev->mc;
+ adapter->info.vport.mc_macs_gen_count++;
+ ref_crc = &adapter->mmac_hash;
+ break;
+ default:
+ return;
+ }
+ netdev_hw_addr_list_for_each(ha, hw_list) {
+ crc = crc32_le(crc, ha->addr, ETH_ALEN);
+ }
+ l = netdev_hw_addr_list_count(hw_list) * ETH_ALEN;
+ crc = ~crc32_le(crc, (void *)&l, sizeof(l));
+
+ if (crc != *ref_crc) {
+ *ref_crc = crc;
+ hfi_vnic_vema_report_event(adapter, event);
+ }
+}
+
+/* hfi_vnic_set_rx_mode - handle uc/mc mac list change */
+static void hfi_vnic_set_rx_mode(struct net_device *netdev)
+{
+ hfi_vnic_mac_send_event(netdev,
+ HFI_VESWPORT_TRAP_IFACE_UCAST_MAC_CHANGE);
+
+ hfi_vnic_mac_send_event(netdev,
+ HFI_VESWPORT_TRAP_IFACE_MCAST_MAC_CHANGE);
}
/* hfi_netdev_open - activate network interface */
@@ -556,6 +694,10 @@ static int hfi_netdev_open(struct net_device *netdev)
if (rc)
goto open_done;
+ /* Update eth link status and send trap */
+ adapter->info.vport.eth_link_status = HFI_VNIC_ETH_LINK_UP;
+ hfi_vnic_vema_report_event(adapter,
+ HFI_VESWPORT_TRAP_ETH_LINK_STATUS_CHANGE);
set_bit(HFI_VNIC_OPEN, &adapter->flags);
v_info("opened\n");
open_done:
@@ -572,6 +714,10 @@ static int hfi_netdev_close(struct net_device *netdev)
if (test_bit(HFI_VNIC_UP, &adapter->flags))
hfi_vnic_down(adapter);
+ /* Update eth link status and send trap */
+ adapter->info.vport.eth_link_status = HFI_VNIC_ETH_LINK_DOWN;
+ hfi_vnic_vema_report_event(adapter,
+ HFI_VESWPORT_TRAP_ETH_LINK_STATUS_CHANGE);
clear_bit(HFI_VNIC_OPEN, &adapter->flags);
mutex_unlock(&adapter->lock);
v_info("closed\n");
@@ -585,6 +731,7 @@ static int hfi_netdev_close(struct net_device *netdev)
.ndo_start_xmit = hfi_netdev_start_xmit,
.ndo_change_mtu = hfi_netdev_change_mtu,
.ndo_get_stats64 = hfi_vnic_get_stats64,
+ .ndo_set_rx_mode = hfi_vnic_set_rx_mode,
.ndo_select_queue = hfi_vnic_select_queue,
.ndo_set_mac_address = hfi_vnic_set_mac_addr,
};
@@ -629,6 +776,8 @@ struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
netif_napi_add(netdev, &adapter->rxq[i].napi, vnic_napi, 64);
}
+ hfi_vnic_set_pod_values(adapter);
+
rc = register_netdev(netdev);
if (rc)
goto netdev_err;
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema_iface.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema_iface.c
new file mode 100644
index 0000000..f912171f
--- /dev/null
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema_iface.c
@@ -0,0 +1,432 @@
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains HFI VNIC EMA Interface functions.
+ */
+
+#include "hfi_vnic_internal.h"
+
+/**
+ * hfi_vnic_vema_report_event - sent trap to report the specified event
+ * @adapter: vnic port adapter
+ * @event: event to be reported
+ *
+ * This function calls vema api to sent a trap for the given event.
+ */
+void hfi_vnic_vema_report_event(struct hfi_vnic_adapter *adapter, u8 event)
+{
+ struct __hfi_veswport_info *info = &adapter->info;
+ struct hfi_vnic_port *vport = adapter->vport;
+ struct __hfi_veswport_trap trap_data;
+
+ trap_data.fabric_id = info->vesw.fabric_id;
+ trap_data.veswid = info->vesw.vesw_id;
+ trap_data.veswportnum = info->vport.port_num;
+ trap_data.hfiportnum = vport->port_num;
+ trap_data.veswportindex = vport->vport_num;
+ trap_data.opcode = event;
+
+ /* Need to send trap here */
+}
+
+/**
+ * hfi_vnic_get_error_counters - get summary counters
+ * @adapter: vnic port adapter
+ * @cntrs: pointer to destination summary counters structure
+ *
+ * This function populates the summary counters that is maintained by the
+ * given adapter to destination address provided.
+ */
+void hfi_vnic_get_summary_counters(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_summary_counters *cntrs)
+{
+ __be64 *dst;
+ u64 *src;
+
+ mutex_lock(&adapter->stats_lock);
+ /* update stats */
+ hfi_vnic_update_stats(adapter->netdev);
+
+ cntrs->vp_instance = cpu_to_be16(adapter->vport->vport_num);
+ cntrs->vesw_id = cpu_to_be16(adapter->vport->vesw_id);
+ cntrs->veswport_num = cpu_to_be32(adapter->info.vport.port_num);
+
+ /*
+ * This loop depends on layout of
+ * struct hfi_veswport_summary_counters and
+ * struct __hfi_vnic_summary_counter
+ */
+ for (dst = &cntrs->tx_errors, src = &adapter->sum_cntrs.tx_errors;
+ dst < &cntrs->reserved[0]; dst++, src++) {
+ *dst = cpu_to_be64(*src);
+ }
+
+ mutex_unlock(&adapter->stats_lock);
+}
+
+/**
+ * hfi_vnic_get_error_counters - get error counters
+ * @adapter: vnic port adapter
+ * @cntrs: pointer to destination error counters structure
+ *
+ * This function populates the error counters that is maintained by the
+ * given adapter to destination address provided.
+ */
+void hfi_vnic_get_error_counters(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_error_counters *cntrs)
+{
+ mutex_lock(&adapter->stats_lock);
+ /* update stats */
+ hfi_vnic_update_stats(adapter->netdev);
+
+ cntrs->vp_instance = cpu_to_be16(adapter->vport->vport_num);
+ cntrs->vesw_id = cpu_to_be16(adapter->vport->vesw_id);
+ cntrs->veswport_num = cpu_to_be32(adapter->info.vport.port_num);
+
+ cntrs->tx_errors = cpu_to_be64(adapter->err_cntrs.tx_errors);
+ cntrs->rx_errors = cpu_to_be64(adapter->err_cntrs.rx_errors);
+ cntrs->tx_smac_filt = cpu_to_be64(adapter->err_cntrs.tx_smac_filt);
+ cntrs->tx_dlid_zero = cpu_to_be64(adapter->err_cntrs.tx_dlid_zero);
+ cntrs->tx_logic = cpu_to_be64(adapter->err_cntrs.tx_logic);
+ cntrs->tx_drop_state = cpu_to_be64(adapter->err_cntrs.tx_drop_state);
+
+ cntrs->rx_bad_veswid = cpu_to_be64(adapter->err_cntrs.rx_bad_veswid);
+ cntrs->rx_runt = cpu_to_be64(adapter->err_cntrs.rx_runt);
+ cntrs->rx_oversize = cpu_to_be64(adapter->err_cntrs.rx_oversize);
+ cntrs->rx_eth_down = cpu_to_be64(adapter->err_cntrs.rx_eth_down);
+ cntrs->rx_drop_state = cpu_to_be64(adapter->err_cntrs.rx_drop_state);
+ cntrs->rx_logic = cpu_to_be64(adapter->err_cntrs.rx_logic);
+ mutex_unlock(&adapter->stats_lock);
+}
+
+/**
+ * hfi_vnic_get_vesw_info -- Get the vesw information
+ * @adapter: vnic port adapter
+ * @info: pointer to destination vesw info structure
+ *
+ * This function copies the vesw info that is maintained by the
+ * given adapter to destination address provided.
+ */
+void hfi_vnic_get_vesw_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_vesw_info *info)
+{
+ struct __hfi_vesw_info *src = &adapter->info.vesw;
+ int i;
+
+ info->fabric_id = cpu_to_be16(src->fabric_id);
+ info->vesw_id = cpu_to_be16(src->vesw_id);
+ memcpy(info->rsvd0, src->rsvd0, ARRAY_SIZE(src->rsvd0));
+ info->def_port_mask = cpu_to_be16(src->def_port_mask);
+ memcpy(info->rsvd1, src->rsvd1, ARRAY_SIZE(src->rsvd1));
+ info->pkey = cpu_to_be16(src->pkey);
+
+ memcpy(info->rsvd2, src->rsvd2, ARRAY_SIZE(src->rsvd2));
+ info->u_mcast_dlid = cpu_to_be32(src->u_mcast_dlid);
+ for (i = 0; i < HFI_VESW_MAX_NUM_DEF_PORT; i++)
+ info->u_ucast_dlid[i] = cpu_to_be32(src->u_ucast_dlid[i]);
+
+ memcpy(info->rsvd3, src->rsvd3, ARRAY_SIZE(src->rsvd3));
+ for (i = 0; i < HFI_VNIC_MAX_NUM_PCP; i++)
+ info->eth_mtu[i] = cpu_to_be16(src->eth_mtu[i]);
+
+ info->eth_mtu_non_vlan = cpu_to_be16(src->eth_mtu_non_vlan);
+ memcpy(info->rsvd4, src->rsvd4, ARRAY_SIZE(src->rsvd4));
+}
+
+/**
+ * hfi_vnic_set_vesw_info -- Set the vesw information
+ * @adapter: vnic port adapter
+ * @info: pointer to vesw info structure
+ *
+ * This function updates the vesw info that is maintained by the
+ * given adapter with vesw info provided. Reserved fields are stored
+ * and returned back to EM as is.
+ */
+void hfi_vnic_set_vesw_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_vesw_info *info)
+{
+ struct __hfi_vesw_info *dst = &adapter->info.vesw;
+ int i;
+
+ dst->fabric_id = be16_to_cpu(info->fabric_id);
+ dst->vesw_id = be16_to_cpu(info->vesw_id);
+ memcpy(dst->rsvd0, info->rsvd0, ARRAY_SIZE(info->rsvd0));
+ dst->def_port_mask = be16_to_cpu(info->def_port_mask);
+ memcpy(dst->rsvd1, info->rsvd1, ARRAY_SIZE(info->rsvd1));
+ dst->pkey = be16_to_cpu(info->pkey);
+
+ memcpy(dst->rsvd2, info->rsvd2, ARRAY_SIZE(info->rsvd2));
+ dst->u_mcast_dlid = be32_to_cpu(info->u_mcast_dlid);
+ for (i = 0; i < HFI_VESW_MAX_NUM_DEF_PORT; i++)
+ dst->u_ucast_dlid[i] = be32_to_cpu(info->u_ucast_dlid[i]);
+
+ memcpy(dst->rsvd3, info->rsvd3, ARRAY_SIZE(info->rsvd3));
+ for (i = 0; i < HFI_VNIC_MAX_NUM_PCP; i++)
+ dst->eth_mtu[i] = be16_to_cpu(info->eth_mtu[i]);
+
+ dst->eth_mtu_non_vlan = be16_to_cpu(info->eth_mtu_non_vlan);
+ memcpy(dst->rsvd4, info->rsvd4, ARRAY_SIZE(info->rsvd4));
+}
+
+/**
+ * hfi_vnic_get_per_veswport_info -- Get the vesw per port information
+ * @adapter: vnic port adapter
+ * @info: pointer to destination vport info structure
+ *
+ * This function copies the vesw per port info that is maintained by the
+ * given adapter to destination address provided.
+ * Note that the read only fields are not copied.
+ */
+void hfi_vnic_get_per_veswport_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_per_veswport_info *info)
+{
+ struct __hfi_per_veswport_info *src = &adapter->info.vport;
+
+ info->port_num = cpu_to_be32(src->port_num);
+ info->eth_link_status = src->eth_link_status;
+ memcpy(info->rsvd0, src->rsvd0, ARRAY_SIZE(src->rsvd0));
+
+ memcpy(info->base_mac_addr, src->base_mac_addr,
+ ARRAY_SIZE(info->base_mac_addr));
+ info->config_state = src->config_state;
+ info->oper_state = src->oper_state;
+ info->max_mac_tbl_ent = cpu_to_be16(src->max_mac_tbl_ent);
+ info->max_smac_ent = cpu_to_be16(src->max_smac_ent);
+ info->mac_tbl_digest = cpu_to_be32(src->mac_tbl_digest);
+ memcpy(info->rsvd1, src->rsvd1, ARRAY_SIZE(src->rsvd1));
+
+ info->encap_slid = cpu_to_be32(src->encap_slid);
+ memcpy(info->pcp_to_sc_uc, src->pcp_to_sc_uc,
+ ARRAY_SIZE(info->pcp_to_sc_uc));
+ memcpy(info->pcp_to_vl_uc, src->pcp_to_vl_uc,
+ ARRAY_SIZE(info->pcp_to_vl_uc));
+ memcpy(info->pcp_to_sc_mc, src->pcp_to_sc_mc,
+ ARRAY_SIZE(info->pcp_to_sc_mc));
+ memcpy(info->pcp_to_vl_mc, src->pcp_to_vl_mc,
+ ARRAY_SIZE(info->pcp_to_vl_mc));
+ info->non_vlan_sc_uc = src->non_vlan_sc_uc;
+ info->non_vlan_vl_uc = src->non_vlan_vl_uc;
+ info->non_vlan_sc_mc = src->non_vlan_sc_mc;
+ info->non_vlan_vl_mc = src->non_vlan_vl_mc;
+ memcpy(info->rsvd2, src->rsvd2, ARRAY_SIZE(src->rsvd2));
+
+ info->uc_macs_gen_count = cpu_to_be16(src->uc_macs_gen_count);
+ info->mc_macs_gen_count = cpu_to_be16(src->mc_macs_gen_count);
+ memcpy(info->rsvd3, src->rsvd3, ARRAY_SIZE(src->rsvd3));
+}
+
+/**
+ * hfi_vnic_set_per_veswport_info -- Set vesw per port information
+ * @adapter: vnic port adapter
+ * @info: pointer to vport info structure
+ *
+ * This function updates the vesw per port info that is maintained by the
+ * given adapter with vesw per port info provided. Reserved fields are
+ * stored and returned back to EM as is.
+ */
+void hfi_vnic_set_per_veswport_info(struct hfi_vnic_adapter *adapter,
+ struct hfi_per_veswport_info *info)
+{
+ struct __hfi_per_veswport_info *dst = &adapter->info.vport;
+
+ dst->port_num = be32_to_cpu(info->port_num);
+ memcpy(dst->rsvd0, info->rsvd0, ARRAY_SIZE(info->rsvd0));
+
+ memcpy(dst->base_mac_addr, info->base_mac_addr,
+ ARRAY_SIZE(dst->base_mac_addr));
+ dst->config_state = info->config_state;
+ memcpy(dst->rsvd1, info->rsvd1, ARRAY_SIZE(info->rsvd1));
+
+ dst->encap_slid = be32_to_cpu(info->encap_slid);
+ memcpy(dst->pcp_to_sc_uc, info->pcp_to_sc_uc,
+ ARRAY_SIZE(dst->pcp_to_sc_uc));
+ memcpy(dst->pcp_to_vl_uc, info->pcp_to_vl_uc,
+ ARRAY_SIZE(dst->pcp_to_vl_uc));
+ memcpy(dst->pcp_to_sc_mc, info->pcp_to_sc_mc,
+ ARRAY_SIZE(dst->pcp_to_sc_mc));
+ memcpy(dst->pcp_to_vl_mc, info->pcp_to_vl_mc,
+ ARRAY_SIZE(dst->pcp_to_vl_mc));
+ dst->non_vlan_sc_uc = info->non_vlan_sc_uc;
+ dst->non_vlan_vl_uc = info->non_vlan_vl_uc;
+ dst->non_vlan_sc_mc = info->non_vlan_sc_mc;
+ dst->non_vlan_vl_mc = info->non_vlan_vl_mc;
+ memcpy(dst->rsvd2, info->rsvd2, ARRAY_SIZE(info->rsvd2));
+ memcpy(dst->rsvd3, info->rsvd3, ARRAY_SIZE(info->rsvd3));
+}
+
+/**
+ * hfi_vnic_query_mcast_macs - query multicast mac list
+ * @adapter: vnic port adapter
+ * @macs: pointer mac list
+ *
+ * This function populates the provided mac list with the configured
+ * multicast addresses in the adapter.
+ */
+void hfi_vnic_query_mcast_macs(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_iface_macs *macs)
+{
+ u16 start_idx, num_macs, idx = 0, count = 0;
+ struct netdev_hw_addr *ha;
+
+ start_idx = be16_to_cpu(macs->start_idx);
+ num_macs = be16_to_cpu(macs->num_macs_in_msg);
+ netdev_for_each_mc_addr(ha, adapter->netdev) {
+ struct hfi_vnic_iface_mac_entry *entry = &macs->entry[count];
+
+ if (start_idx > idx++)
+ continue;
+ else if (num_macs == count)
+ break;
+ memcpy(entry, ha->addr, sizeof(*entry));
+ count++;
+ }
+
+ macs->tot_macs_in_lst = cpu_to_be16(netdev_mc_count(adapter->netdev));
+ macs->num_macs_in_msg = cpu_to_be16(count);
+ macs->gen_count = cpu_to_be16(adapter->info.vport.mc_macs_gen_count);
+}
+
+/**
+ * hfi_vnic_query_ucast_macs - query unicast mac list
+ * @adapter: vnic port adapter
+ * @macs: pointer mac list
+ *
+ * This function populates the provided mac list with the configured
+ * unicast addresses in the adapter.
+ */
+void hfi_vnic_query_ucast_macs(struct hfi_vnic_adapter *adapter,
+ struct hfi_veswport_iface_macs *macs)
+{
+ u16 start_idx, tot_macs, num_macs, idx = 0, count = 0;
+ struct netdev_hw_addr *ha;
+
+ start_idx = be16_to_cpu(macs->start_idx);
+ num_macs = be16_to_cpu(macs->num_macs_in_msg);
+ /* loop through dev_addrs list first */
+ for_each_dev_addr(adapter->netdev, ha) {
+ struct hfi_vnic_iface_mac_entry *entry = &macs->entry[count];
+
+ /* Do not include EM specified MAC address */
+ if (!memcmp(adapter->info.vport.base_mac_addr, ha->addr,
+ ARRAY_SIZE(adapter->info.vport.base_mac_addr)))
+ continue;
+
+ if (start_idx > idx++)
+ continue;
+ else if (num_macs == count)
+ break;
+ memcpy(entry, ha->addr, sizeof(*entry));
+ count++;
+ }
+
+ /* loop through uc list */
+ netdev_for_each_uc_addr(ha, adapter->netdev) {
+ struct hfi_vnic_iface_mac_entry *entry = &macs->entry[count];
+
+ if (start_idx > idx++)
+ continue;
+ else if (num_macs == count)
+ break;
+ memcpy(entry, ha->addr, sizeof(*entry));
+ count++;
+ }
+
+ tot_macs = netdev_hw_addr_list_count(&adapter->netdev->dev_addrs) +
+ netdev_uc_count(adapter->netdev);
+ macs->tot_macs_in_lst = cpu_to_be16(tot_macs);
+ macs->num_macs_in_msg = cpu_to_be16(count);
+ macs->gen_count = cpu_to_be16(adapter->info.vport.uc_macs_gen_count);
+}
+
+/**
+ * hfi_vnic_add_vport - Add a new vnic port
+ * @cport: vnic control port
+ * @port_num: OPA port number
+ * @vport_num: vnic port number
+ *
+ * Return pointer to adapter of newly created vnic port.
+ */
+struct hfi_vnic_adapter *hfi_vnic_add_vport(struct hfi_vnic_ctrl_port *cport,
+ u8 port_num, u8 vport_num)
+{
+ struct hfi_vnic_adapter *adapter;
+ struct hfi_vnic_port *vport;
+
+ vport = cport->ops->add_vport(cport->ibdev, port_num, vport_num);
+ if (IS_ERR(vport))
+ return ERR_CAST(vport);
+
+ if (vport->hfi_info.num_rx_q > HFI_VNIC_MAX_QUEUE ||
+ vport->hfi_info.num_tx_q > HFI_VNIC_MAX_QUEUE) {
+ c_err("Number of VNIC (rx %d, tx %d) queues > Max Queue Size (%d)",
+ vport->hfi_info.num_rx_q, vport->hfi_info.num_tx_q,
+ HFI_VNIC_MAX_QUEUE);
+ return ERR_PTR(-EINVAL);
+ }
+
+ adapter = hfi_vnic_add_netdev(vport, cport->ibdev->dma_device);
+ if (IS_ERR(adapter))
+ cport->ops->rem_vport(vport);
+ else
+ adapter->cport = cport;
+
+ return adapter;
+}
+
+/**
+ * hfi_vnic_rem_vport - Remove a new vnic port
+ * @adapter: vnic adapter
+ */
+void hfi_vnic_rem_vport(struct hfi_vnic_adapter *adapter)
+{
+ struct hfi_vnic_ctrl_port *cport = adapter->cport;
+ struct hfi_vnic_port *vport = adapter->vport;
+
+ hfi_vnic_rem_netdev(vport);
+ cport->ops->rem_vport(vport);
+}
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply related [flat|nested] 41+ messages in thread
* [RFC v2 08/10] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) function
2016-12-15 7:59 [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
` (2 preceding siblings ...)
2016-12-15 7:59 ` [RFC v2 05/10] IB/hfi-vnic: VNIC statistics support Vishwanathapura, Niranjana
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 09/10] IB/hfi1: Virtual Network Interface Controller (VNIC) support Vishwanathapura, Niranjana
` (2 subsequent siblings)
6 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Sadanand Warrier, Niranjana Vishwanathapura, Tanya K Jajodia,
Sudeep Dutt
HFI VEMA function interfaces with the Infiniband MAD stack to exchange the
management information packets with the Ethernet Manager (EM).
It interfaces with the HFI VNIC netdev function to SET/GET the management
information. The information exchanged with the EM includes class port
details, encapsulation configuration, various counters, unicast and
multicast MAC list and the MAC table. It also supports sending traps
to the EM.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Sadanand Warrier <sadanand.warrier@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Tanya K Jajodia <tanya.k.jajodia@intel.com>
Signed-off-by: Sudeep Dutt <sudeep.dutt@intel.com>
---
drivers/infiniband/sw/intel/hfi_vnic/Makefile | 2 +-
.../sw/intel/hfi_vnic/hfi_vnic_ethtool.c | 12 +
.../sw/intel/hfi_vnic/hfi_vnic_internal.h | 11 +
.../infiniband/sw/intel/hfi_vnic/hfi_vnic_vema.c | 1024 ++++++++++++++++++++
.../sw/intel/hfi_vnic/hfi_vnic_vema_iface.c | 4 +-
5 files changed, 1050 insertions(+), 3 deletions(-)
create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema.c
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/Makefile b/drivers/infiniband/sw/intel/hfi_vnic/Makefile
index a0562af..16c0830 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/Makefile
+++ b/drivers/infiniband/sw/intel/hfi_vnic/Makefile
@@ -4,4 +4,4 @@
obj-$(CONFIG_HFI_VNIC) += hfi_vnic.o
hfi_vnic-y := hfi_vnic_netdev.o hfi_vnic_encap.o hfi_vnic_ethtool.o \
- hfi_vnic_vema_iface.o
+ hfi_vnic_vema.o hfi_vnic_vema_iface.o
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
index 9289ab2..9c2ed37 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_ethtool.c
@@ -130,6 +130,17 @@ struct vnic_stats {
#define VNIC_STATS_LEN ARRAY_SIZE(vnic_gstrings_stats)
+/* vnic_get_drvinfo - get driver info */
+static void vnic_get_drvinfo(struct net_device *netdev,
+ struct ethtool_drvinfo *drvinfo)
+{
+ strlcpy(drvinfo->driver, hfi_vnic_driver_name, sizeof(drvinfo->driver));
+ strlcpy(drvinfo->version, hfi_vnic_driver_version,
+ sizeof(drvinfo->version));
+ strlcpy(drvinfo->bus_info, dev_name(netdev->dev.parent),
+ sizeof(drvinfo->bus_info));
+}
+
/* vnic_get_sset_count - get string set count */
static int vnic_get_sset_count(struct net_device *netdev, int sset)
{
@@ -183,6 +194,7 @@ static void vnic_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
/* ethtool ops */
static const struct ethtool_ops hfi_vnic_ethtool_ops = {
+ .get_drvinfo = vnic_get_drvinfo,
.get_link = ethtool_op_get_link,
.get_strings = vnic_get_strings,
.get_sset_count = vnic_get_sset_count,
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
index 7723a4e..b36bb76 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_internal.h
@@ -246,10 +246,12 @@ struct __hfi_veswport_trap {
* struct hfi_vnic_ctrl_port - HFI virtual NIC control port
* @ibdev: pointer to ib device
* @ops: hfi vnic control operations
+ * @num_ports: number of hfi ports
*/
struct hfi_vnic_ctrl_port {
struct ib_device *ibdev;
struct hfi_vnic_ctrl_ops *ops;
+ u8 num_ports;
};
/**
@@ -280,6 +282,8 @@ struct hfi_vnic_rx_queue {
* @mactbl_lock: mac table lock
* @stats_lock: statistics lock
* @flow_tbl: flow to default port redirection table
+ * @trap_timeout: trap timeout
+ * @trap_count: no. of traps allowed within timeout period
* @q_sum_cntrs: per queue EM summary counters
* @q_err_cntrs: per queue EM error counters
* @q_rx_logic_errors: per queue rx logic (default) errors
@@ -314,6 +318,8 @@ struct hfi_vnic_adapter {
struct mutex stats_lock;
u8 flow_tbl[HFI_VNIC_FLOW_TBL_SIZE];
+ unsigned long trap_timeout;
+ u8 trap_count;
struct __hfi_vnic_summary_counters q_sum_cntrs[HFI_VNIC_MAX_QUEUE];
struct __hfi_vnic_error_counters q_err_cntrs[HFI_VNIC_MAX_QUEUE];
@@ -394,6 +400,9 @@ struct hfi_vnic_mac_tbl_node {
!obj && (bkt) < HFI_VNIC_MAC_TBL_SIZE; (bkt)++) \
hlist_for_each_entry(obj, &name[bkt], member)
+extern char hfi_vnic_driver_name[];
+extern const char hfi_vnic_driver_version[];
+
struct hfi_vnic_adapter *hfi_vnic_add_netdev(struct hfi_vnic_port *vport,
struct device *parent);
void hfi_vnic_rem_netdev(struct hfi_vnic_port *vport);
@@ -428,5 +437,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_vport(struct hfi_vnic_ctrl_port *cport,
u8 port_num, u8 vport_num);
void hfi_vnic_rem_vport(struct hfi_vnic_adapter *adapter);
void hfi_vnic_set_ethtool_ops(struct net_device *netdev);
+void hfi_vnic_vema_send_trap(struct hfi_vnic_adapter *adapter,
+ struct __hfi_veswport_trap *data, u32 lid);
#endif /* _HFI_VNIC_INTERNAL_H */
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema.c
new file mode 100644
index 0000000..f7d4bc1
--- /dev/null
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema.c
@@ -0,0 +1,1024 @@
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains HFI Virtual Network Interface Controller (VNIC)
+ * Ethernet Management Agent (EMA) driver
+ */
+
+#include <linux/module.h>
+#include <rdma/ib_addr.h>
+#include <rdma/ib_smi.h>
+
+#include "hfi_vnic_internal.h"
+
+#define DRV_VERSION "1.0"
+char hfi_vnic_driver_name[] = "hfi_vnic";
+const char hfi_vnic_driver_version[] = DRV_VERSION;
+
+/*
+ * The trap service level is kept in bits 3 to 7 in the trap_sl_rsvd
+ * field in the class port info MAD.
+ */
+#define GET_TRAP_SL_FROM_CLASS_PORT_INFO(x) (((x) >> 3) & 0x1f)
+
+/* Cap trap bursts to a reasonable limit good for normal cases */
+#define HFI_VNIC_TRAP_BURST_LIMIT 4
+
+/*
+ * VNIC trap limit timeout.
+ * Inverse of cap2_mask response time out (1.0737 secs) = 0.9
+ * secs approx IB spec 13.4.6.2.1 PortInfoSubnetTimeout and
+ * 13.4.9 Traps.
+ */
+#define HFI_VNIC_TRAP_TIMEOUT ((4096 * (1UL << 18)) / 1000)
+
+#define HFI_VNIC_UNSUP_ATTR \
+ cpu_to_be16(IB_MGMT_MAD_STATUS_UNSUPPORTED_METHOD_ATTRIB)
+
+#define HFI_VNIC_INVAL_ATTR \
+ cpu_to_be16(IB_MGMT_MAD_STATUS_INVALID_ATTRIB_VALUE)
+
+#define HFI_VNIC_CLASS_CAP_TRAP cpu_to_be16(1 << 8)
+
+struct hfi_class_port_info {
+ u8 base_version;
+ u8 class_version;
+ __be16 cap_mask;
+ __be32 cap_mask2_resp_time;
+
+ u8 redirect_gid[16];
+ __be32 redirect_tc_fl;
+ __be32 redirect_lid;
+ __be32 redirect_sl_qp;
+ __be32 redirect_qkey;
+
+ u8 trap_gid[16];
+ __be32 trap_tc_fl;
+ __be32 trap_lid;
+ __be32 trap_hl_qp;
+ __be32 trap_qkey;
+
+ __be16 trap_pkey;
+ __be16 redirect_pkey;
+
+ u8 trap_sl_rsvd;
+ u8 reserved[3];
+} __packed;
+
+/**
+ * struct hfi_vnic_vema_port -- VNIC VEMA port details
+ * @cport: pointer to port
+ * @mad_agent: pointer to mad agent for port
+ * @class_port_info: Class port info information.
+ * @tid: Transaction id
+ * @port_num: HFI port number
+ * @vport_idr: vnic ports idr
+ * @lock: adapter interface lock
+ */
+struct hfi_vnic_vema_port {
+ struct hfi_vnic_ctrl_port *cport;
+ struct ib_mad_agent *mad_agent;
+ struct hfi_class_port_info class_port_info;
+ u64 tid;
+ u8 port_num;
+ struct idr vport_idr;
+
+ /* Lock to query/update network adapter */
+ struct mutex lock;
+};
+
+static void hfi_vnic_vema_add_one(struct ib_device *device);
+static void hfi_vnic_vema_rem_one(struct ib_device *device,
+ void *client_data);
+
+static struct ib_client hfi_vnic_client = {
+ .name = hfi_vnic_driver_name,
+ .add = hfi_vnic_vema_add_one,
+ .remove = hfi_vnic_vema_rem_one,
+};
+
+/**
+ * vema_get_vport_num -- Get the vnic from the mad
+ * @recvd_mad: Received mad
+ *
+ * Return: returns value of the vnic port number
+ */
+static inline u8 vema_get_vport_num(struct hfi_vnic_vema_mad *recvd_mad)
+{
+ return be32_to_cpu(recvd_mad->mad_hdr.attr_mod) >> 16 & 0xff;
+}
+
+/**
+ * vema_get_vport_adapter -- Get vnic port adapter from recvd mad
+ * @recvd_mad: received mad
+ * @port: ptr to port struct on which MAD was recvd
+ *
+ * Return: vnic adapter
+ */
+static inline struct hfi_vnic_adapter *
+vema_get_vport_adapter(struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_port *port)
+{
+ u8 vport_num = vema_get_vport_num(recvd_mad);
+
+ return idr_find(&port->vport_idr, vport_num);
+}
+
+/**
+ * vema_mac_tbl_req_ok -- Check if mac request has correct values
+ * @mac_tbl: mac table
+ *
+ * This function checks for the validity of the offset and number of
+ * entries required.
+ *
+ * Return: true if offset and num_entries are valid
+ */
+static inline bool vema_mac_tbl_req_ok(struct hfi_veswport_mactable *mac_tbl)
+{
+ u16 offset, num_entries;
+ u16 req_entries = ((HFI_VNIC_EMA_DATA - sizeof(*mac_tbl)) /
+ sizeof(mac_tbl->tbl_entries[0]));
+
+ offset = be16_to_cpu(mac_tbl->offset);
+ num_entries = be16_to_cpu(mac_tbl->num_entries);
+
+ return ((num_entries <= req_entries) &&
+ (offset + num_entries <= HFI_VNIC_MAC_TBL_MAX_ENTRIES));
+}
+
+/*
+ * Return the power on default values in the port info structure
+ * in big endian format as required by MAD.
+ */
+static inline void vema_get_pod_values(struct hfi_veswport_info *port_info)
+{
+ memset(port_info, 0, sizeof(*port_info));
+ port_info->vport.max_mac_tbl_ent =
+ cpu_to_be16(HFI_VNIC_MAC_TBL_MAX_ENTRIES);
+ port_info->vport.max_smac_ent =
+ cpu_to_be16(HFI_VNIC_MAX_SMAC_LIMIT);
+ port_info->vport.oper_state = HFI_VNIC_STATE_DROP_ALL;
+ port_info->vport.config_state = HFI_VNIC_STATE_DROP_ALL;
+}
+
+/**
+ * vema_add_vport -- Add a new vnic port
+ * @port: ptr to hfi_vnic_vema_port struct
+ * @vport_num: vnic port number (to be added)
+ *
+ * Return a pointer to the vnic adapter structure
+ */
+static struct hfi_vnic_adapter *vema_add_vport(struct hfi_vnic_vema_port *port,
+ u8 vport_num)
+{
+ struct hfi_vnic_ctrl_port *cport = port->cport;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = hfi_vnic_add_vport(cport, port->port_num, vport_num);
+ if (!IS_ERR(adapter)) {
+ int rc;
+
+ rc = idr_alloc(&port->vport_idr, adapter, vport_num,
+ vport_num + 1, GFP_NOWAIT);
+ if (rc < 0) {
+ hfi_vnic_rem_vport(adapter);
+ adapter = ERR_PTR(rc);
+ }
+ }
+
+ return adapter;
+}
+
+/**
+ * vema_get_class_port_info -- Get class info for port
+ * @port: Port on whic MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function copies the latest class port info value set for the
+ * port and stores it for generating traps
+ */
+static void vema_get_class_port_info(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_class_port_info *port_info;
+
+ port_info = (struct hfi_class_port_info *)rsp_mad->data;
+ memcpy(port_info, &port->class_port_info, sizeof(*port_info));
+ port_info->base_version = OPA_MGMT_BASE_VERSION,
+ port_info->class_version = HFI_EMA_CLASS_VERSION;
+
+ /* Agent generates traps */
+ port_info->cap_mask = HFI_VNIC_CLASS_CAP_TRAP;
+
+ /*
+ * Since a get routine is always sent by the EM first we
+ * set the expected response time to
+ * 4.096 usec * 2^18 == 1.0737 sec here.
+ */
+ port_info->cap_mask2_resp_time = cpu_to_be32(18);
+}
+
+/**
+ * vema_set_class_port_info -- Get class info for port
+ * @port: Port on whic MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function updates the port class info for the specific vnic
+ * and sets up the response mad data
+ */
+static void vema_set_class_port_info(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ memcpy(&port->class_port_info, recvd_mad->data,
+ sizeof(port->class_port_info));
+
+ vema_get_class_port_info(port, recvd_mad, rsp_mad);
+}
+
+/**
+ * vema_get_veswport_info -- Get veswport info
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ */
+static void vema_get_veswport_info(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_info *port_info =
+ (struct hfi_veswport_info *)rsp_mad->data;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (adapter) {
+ memset(port_info, 0, sizeof(*port_info));
+ hfi_vnic_get_vesw_info(adapter, &port_info->vesw);
+ hfi_vnic_get_per_veswport_info(adapter,
+ &port_info->vport);
+ } else {
+ vema_get_pod_values(port_info);
+ }
+}
+
+/**
+ * vema_set_veswport_info -- Set veswport info
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function gets the port class infor for vnic
+ */
+static void vema_set_veswport_info(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_vnic_ctrl_port *cport = port->cport;
+ struct hfi_veswport_info *port_info;
+ struct hfi_vnic_adapter *adapter;
+ u8 vport_num;
+
+ vport_num = vema_get_vport_num(recvd_mad);
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ adapter = vema_add_vport(port, vport_num);
+ if (IS_ERR(adapter)) {
+ c_err("failed to add vport %d: %ld\n",
+ vport_num, PTR_ERR(adapter));
+ goto err_exit;
+ }
+ }
+
+ port_info = (struct hfi_veswport_info *)recvd_mad->data;
+ hfi_vnic_set_vesw_info(adapter, &port_info->vesw);
+ hfi_vnic_set_per_veswport_info(adapter, &port_info->vport);
+
+ /* Process the new config settings */
+ hfi_vnic_process_vema_config(adapter);
+
+ vema_get_veswport_info(port, recvd_mad, rsp_mad);
+ return;
+
+err_exit:
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+}
+
+/**
+ * vema_get_mac_entries -- Get MAC entries in VNIC MAC table
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function gets the MAC entries that are programmed into
+ * the VNIC MAC forwarding table. It checks for the validity of
+ * the index into the MAC table and the number of entries that
+ * are to be retrieved.
+ */
+static void vema_get_mac_entries(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_mactable *mac_tbl_in, *mac_tbl_out;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ return;
+ }
+
+ mac_tbl_in = (struct hfi_veswport_mactable *)recvd_mad->data;
+ mac_tbl_out = (struct hfi_veswport_mactable *)rsp_mad->data;
+
+ if (vema_mac_tbl_req_ok(mac_tbl_in)) {
+ mac_tbl_out->offset = mac_tbl_in->offset;
+ mac_tbl_out->num_entries = mac_tbl_in->num_entries;
+ hfi_vnic_query_mac_tbl(adapter, mac_tbl_out);
+ } else {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ }
+}
+
+/**
+ * vema_set_mac_entries -- Set MAC entries in VNIC MAC table
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function sets the MAC entries in the VNIC forwarding table
+ * It checks for the validity of the index and the number of forwarding
+ * table entries to be programmed.
+ */
+static void vema_set_mac_entries(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_mactable *mac_tbl;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ return;
+ }
+
+ mac_tbl = (struct hfi_veswport_mactable *)recvd_mad->data;
+ if (vema_mac_tbl_req_ok(mac_tbl)) {
+ if (hfi_vnic_update_mac_tbl(adapter, mac_tbl))
+ rsp_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ } else {
+ rsp_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ }
+ vema_get_mac_entries(port, recvd_mad, rsp_mad);
+}
+
+/**
+ * vema_set_delete_vesw -- Reset VESW info to POD values
+ * @port: source port on which MAD was received
+ * @recvd_mad: pointer to the received mad
+ * @rsp_mad: pointer to respose mad
+ *
+ * This function clears all the fields of veswport info for the requested vesw
+ * and sets them back to the power-on default values. It does not delete the
+ * vesw.
+ */
+static void vema_set_delete_vesw(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_info *port_info =
+ (struct hfi_veswport_info *)rsp_mad->data;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ return;
+ }
+
+ vema_get_pod_values(port_info);
+ hfi_vnic_set_vesw_info(adapter, &port_info->vesw);
+ hfi_vnic_set_per_veswport_info(adapter, &port_info->vport);
+
+ /* Process the new config settings */
+ hfi_vnic_process_vema_config(adapter);
+
+ hfi_vnic_release_mac_tbl(adapter);
+
+ vema_get_veswport_info(port, recvd_mad, rsp_mad);
+}
+
+/**
+ * vema_get_mac_list -- Get the unicast/multicast macs.
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad contains fields to set vnic parameters
+ * @rsp_mad: Response mad to be built
+ * @attr_id: Attribute ID indicating multicast or unicast mac list
+ */
+static void vema_get_mac_list(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad,
+ u16 attr_id)
+{
+ struct hfi_veswport_iface_macs *macs_in, *macs_out;
+ int max_entries = (HFI_VNIC_EMA_DATA - sizeof(*macs_out)) / ETH_ALEN;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (!adapter) {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ return;
+ }
+
+ macs_in = (struct hfi_veswport_iface_macs *)recvd_mad->data;
+ macs_out = (struct hfi_veswport_iface_macs *)rsp_mad->data;
+
+ macs_out->start_idx = macs_in->start_idx;
+ if (macs_in->num_macs_in_msg)
+ macs_out->num_macs_in_msg = macs_in->num_macs_in_msg;
+ else
+ macs_out->num_macs_in_msg = cpu_to_be16(max_entries);
+
+ if (attr_id == HFI_EM_ATTR_IFACE_MCAST_MACS)
+ hfi_vnic_query_mcast_macs(adapter, macs_out);
+ else
+ hfi_vnic_query_ucast_macs(adapter, macs_out);
+}
+
+/**
+ * vema_get_summary_counters -- Gets summary counters.
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad contains fields to set vnic parameters
+ * @rsp_mad: Response mad to be built
+ */
+static void vema_get_summary_counters(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_summary_counters *cntrs;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (adapter) {
+ cntrs = (struct hfi_veswport_summary_counters *)rsp_mad->data;
+ hfi_vnic_get_summary_counters(adapter, cntrs);
+ } else {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ }
+}
+
+/**
+ * vema_get_error_counters -- Gets summary counters.
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad contains fields to set vnic parameters
+ * @rsp_mad: Response mad to be built
+ */
+static void vema_get_error_counters(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ struct hfi_veswport_error_counters *cntrs;
+ struct hfi_vnic_adapter *adapter;
+
+ adapter = vema_get_vport_adapter(recvd_mad, port);
+ if (adapter) {
+ cntrs = (struct hfi_veswport_error_counters *)rsp_mad->data;
+ hfi_vnic_get_error_counters(adapter, cntrs);
+ } else {
+ rsp_mad->mad_hdr.status = HFI_VNIC_INVAL_ATTR;
+ }
+}
+
+/**
+ * vema_get -- Process received get MAD
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad
+ * @rsp_mad: Response mad to be built
+ */
+static void vema_get(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ u16 attr_id = be16_to_cpu(recvd_mad->mad_hdr.attr_id);
+
+ switch (attr_id) {
+ case HFI_EM_ATTR_CLASS_PORT_INFO:
+ vema_get_class_port_info(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_INFO:
+ vema_get_veswport_info(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_MAC_ENTRIES:
+ vema_get_mac_entries(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_IFACE_UCAST_MACS:
+ /* fall through */
+ case HFI_EM_ATTR_IFACE_MCAST_MACS:
+ vema_get_mac_list(port, recvd_mad, rsp_mad, attr_id);
+ break;
+ case HFI_EM_ATTR_VESWPORT_SUMMARY_COUNTERS:
+ vema_get_summary_counters(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_ERROR_COUNTERS:
+ vema_get_error_counters(port, recvd_mad, rsp_mad);
+ break;
+ default:
+ rsp_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ break;
+ }
+}
+
+/**
+ * vema_set -- Process received set MAD
+ * @port: source port on which MAD was received
+ * @recvd_mad: Received mad contains fields to set vnic parameters
+ * @rsp_mad: Response mad to be built
+ */
+static void vema_set(struct hfi_vnic_vema_port *port,
+ struct hfi_vnic_vema_mad *recvd_mad,
+ struct hfi_vnic_vema_mad *rsp_mad)
+{
+ u16 attr_id = be16_to_cpu(recvd_mad->mad_hdr.attr_id);
+
+ switch (attr_id) {
+ case HFI_EM_ATTR_CLASS_PORT_INFO:
+ vema_set_class_port_info(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_INFO:
+ vema_set_veswport_info(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_VESWPORT_MAC_ENTRIES:
+ vema_set_mac_entries(port, recvd_mad, rsp_mad);
+ break;
+ case HFI_EM_ATTR_DELETE_VESW:
+ vema_set_delete_vesw(port, recvd_mad, rsp_mad);
+ break;
+ default:
+ rsp_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ break;
+ }
+}
+
+/**
+ * vema_send -- Send handler for VEMA MAD agent
+ * @mad_agent: pointer to the mad agent
+ * @mad_wc: pointer to mad send work completion information
+ *
+ * Free all the data structures associated with the sent MAD
+ */
+static void vema_send(struct ib_mad_agent *mad_agent,
+ struct ib_mad_send_wc *mad_wc)
+{
+ ib_destroy_ah(mad_wc->send_buf->ah);
+ ib_free_send_mad(mad_wc->send_buf);
+}
+
+/**
+ * vema_recv -- Recv handler for VEMA MAD agent
+ * @mad_agent: pointer to the mad agent
+ * @send_buf: Send buffer if found, else NULL
+ * @mad_wc: pointer to mad send work completion information
+ *
+ * Handle only set and get methods and respond to other methods
+ * as unsupported. Allocate response buffer and address handle
+ * for the response MAD.
+ */
+static void vema_recv(struct ib_mad_agent *mad_agent,
+ struct ib_mad_send_buf *send_buf,
+ struct ib_mad_recv_wc *mad_wc)
+{
+ struct hfi_vnic_vema_port *port;
+ struct ib_ah *ah;
+ struct ib_mad_send_buf *rsp;
+ struct hfi_vnic_vema_mad *vema_mad;
+
+ if (!mad_wc || !mad_wc->recv_buf.mad)
+ return;
+
+ port = mad_agent->context;
+ ah = ib_create_ah_from_wc(mad_agent->qp->pd, mad_wc->wc,
+ mad_wc->recv_buf.grh, mad_agent->port_num);
+ if (IS_ERR(ah))
+ goto free_recv_mad;
+
+ rsp = ib_create_send_mad(mad_agent, mad_wc->wc->src_qp,
+ mad_wc->wc->pkey_index, 0,
+ IB_MGMT_VENDOR_HDR, HFI_VNIC_EMA_DATA,
+ GFP_KERNEL, OPA_MGMT_BASE_VERSION);
+ if (IS_ERR(rsp))
+ goto err_rsp;
+
+ rsp->ah = ah;
+ vema_mad = rsp->mad;
+ memcpy(vema_mad, mad_wc->recv_buf.mad, IB_MGMT_VENDOR_HDR);
+ vema_mad->mad_hdr.method = IB_MGMT_METHOD_GET_RESP;
+ vema_mad->mad_hdr.status = 0;
+
+ /* Lock ensures network adapter is not removed */
+ mutex_lock(&port->lock);
+
+ switch (mad_wc->recv_buf.mad->mad_hdr.method) {
+ case IB_MGMT_METHOD_GET:
+ vema_get(port, (struct hfi_vnic_vema_mad *)mad_wc->recv_buf.mad,
+ vema_mad);
+ break;
+ case IB_MGMT_METHOD_SET:
+ vema_set(port, (struct hfi_vnic_vema_mad *)mad_wc->recv_buf.mad,
+ vema_mad);
+ break;
+ default:
+ vema_mad->mad_hdr.status = HFI_VNIC_UNSUP_ATTR;
+ break;
+ }
+ mutex_unlock(&port->lock);
+
+ if (!ib_post_send_mad(rsp, NULL)) {
+ /*
+ * with post send successful ah and send mad
+ * will be destroyed in send handler
+ */
+ goto free_recv_mad;
+ }
+
+ ib_free_send_mad(rsp);
+
+err_rsp:
+ ib_destroy_ah(ah);
+free_recv_mad:
+ ib_free_recv_mad(mad_wc);
+}
+
+/**
+ * vema_get_port -- Gets the hfi_vnic_vema_port
+ * @cport: pointer to control dev
+ * @port_num: Port number
+ *
+ * This function loops through the ports and returns
+ * the hfi_vnic_vema port structure that is associated
+ * with the HFI port number
+ *
+ * Return: ptr to requested hfi_vnic_vema_port strucure
+ * if success, NULL if not
+ */
+static struct hfi_vnic_vema_port *
+vema_get_port(struct hfi_vnic_ctrl_port *cport, u8 port_num)
+{
+ struct hfi_vnic_vema_port *port = (void *)cport + sizeof(*cport);
+
+ if (port_num > cport->num_ports)
+ return NULL;
+
+ return port + (port_num - 1);
+}
+
+static int vema_rem_vport(int id, void *p, void *data)
+{
+ struct hfi_vnic_adapter *adapter = p;
+
+ hfi_vnic_rem_vport(adapter);
+ return 0;
+}
+
+/**
+ * vema_unregister -- Unregisters agent
+ * @cport: pointer to control port
+ *
+ * This deletes the registration by VEMA for MADs
+ */
+static void vema_unregister(struct hfi_vnic_ctrl_port *cport)
+{
+ int i;
+
+ for (i = 1; i <= cport->num_ports; i++) {
+ struct hfi_vnic_vema_port *port = vema_get_port(cport, i);
+
+ if (!port->mad_agent)
+ continue;
+
+ /* Lock ensures no MAD is being processed */
+ mutex_lock(&port->lock);
+ idr_for_each(&port->vport_idr, vema_rem_vport, NULL);
+ mutex_unlock(&port->lock);
+
+ ib_unregister_mad_agent(port->mad_agent);
+ mutex_destroy(&port->lock);
+ idr_destroy(&port->vport_idr);
+ }
+}
+
+/**
+ * vema_register -- Registers agent
+ * @cport: pointer to control port
+ *
+ * This function registers the handlers for the VEMA MADs
+ *
+ * Return: returns 0 on success. non zero otherwise
+ */
+static int vema_register(struct hfi_vnic_ctrl_port *cport)
+{
+ struct ib_mad_reg_req reg_req = {
+ .mgmt_class = HFI_MGMT_CLASS_INTEL_EMA,
+ .mgmt_class_version = OPA_MGMT_BASE_VERSION,
+ .oui = { INTEL_OUI_1, INTEL_OUI_2, INTEL_OUI_3 }
+ };
+ int i;
+
+ set_bit(IB_MGMT_METHOD_GET, reg_req.method_mask);
+ set_bit(IB_MGMT_METHOD_SET, reg_req.method_mask);
+
+ /* register mad agent for each port on dev */
+ for (i = 1; i <= cport->num_ports; i++) {
+ struct hfi_vnic_vema_port *port = vema_get_port(cport, i);
+
+ port->cport = cport;
+ port->port_num = i;
+ idr_init(&port->vport_idr);
+ mutex_init(&port->lock);
+ port->mad_agent = ib_register_mad_agent(cport->ibdev, i,
+ IB_QPT_GSI, ®_req,
+ IB_MGMT_RMPP_VERSION,
+ vema_send, vema_recv,
+ port, 0);
+ if (IS_ERR(port->mad_agent)) {
+ int ret = PTR_ERR(port->mad_agent);
+
+ port->mad_agent = NULL;
+ mutex_destroy(&port->lock);
+ idr_destroy(&port->vport_idr);
+ vema_unregister(cport);
+ return ret;
+ }
+ }
+
+ return 0;
+}
+
+/**
+ * hfi_vnic_vema_send_trap -- This function sends a trap to the EM
+ * @cport: pointer to vnic control port
+ * @data: pointer to trap data filled by calling function
+ * @lid: issuers lid (encap_slid from vesw_port_info)
+ *
+ * This function is called from the VNIC driver to send a trap if there
+ * is somethng the EM should be notified about. These events currently
+ * are
+ * 1) UNICAST INTERFACE MACADDRESS changes
+ * 2) MULTICAST INTERFACE MACADDRESS changes
+ * 3) ETHERNET LINK STATUS changes
+ * While allocating the send mad the remote site qpn used is 1
+ * as this is the well known QP.
+ *
+ */
+void hfi_vnic_vema_send_trap(struct hfi_vnic_adapter *adapter,
+ struct __hfi_veswport_trap *data, u32 lid)
+{
+ struct hfi_vnic_ctrl_port *cport = adapter->cport;
+ struct ib_mad_send_buf *send_buf;
+ struct hfi_vnic_vema_port *port;
+ struct ib_device *ibp;
+ struct hfi_vnic_vema_mad_trap *trap_mad;
+ struct hfi_class_port_info *class;
+ struct ib_ah_attr ah_attr;
+ struct ib_ah *ah;
+ struct hfi_veswport_trap *trap;
+ u32 trap_lid;
+ u16 pkey_idx;
+
+ if (!cport)
+ goto err_exit;
+ ibp = cport->ibdev;
+ port = vema_get_port(cport, data->hfiportnum);
+ if (!port || !port->mad_agent)
+ goto err_exit;
+
+ if (time_before(jiffies, adapter->trap_timeout)) {
+ if (adapter->trap_count == HFI_VNIC_TRAP_BURST_LIMIT) {
+ v_warn("Trap rate exceeded\n");
+ goto err_exit;
+ } else {
+ adapter->trap_count++;
+ }
+ } else {
+ adapter->trap_count = 0;
+ }
+
+ class = &port->class_port_info;
+ /* Set up address handle */
+ memset(&ah_attr, 0, sizeof(ah_attr));
+ ah_attr.sl = GET_TRAP_SL_FROM_CLASS_PORT_INFO(class->trap_sl_rsvd);
+ ah_attr.port_num = port->port_num;
+ trap_lid = be32_to_cpu(class->trap_lid);
+ /*
+ * check for trap lid validity, must not be zero
+ * The trap sink could change after we fashion the MAD but since traps
+ * are not guaranteed we won't use a lock as anyway the change will take
+ * place even with locking.
+ */
+ if (!trap_lid) {
+ c_err("%s: Invalid dlid\n", __func__);
+ goto err_exit;
+ }
+
+ ah_attr.dlid = trap_lid;
+ ah = ib_create_ah(port->mad_agent->qp->pd, &ah_attr);
+ if (IS_ERR(ah)) {
+ c_err("%s:Couldn't create new AH = %p\n", __func__, ah);
+ c_err("%s:dlid = %d, sl = %d, port = %d\n", __func__,
+ ah_attr.dlid, ah_attr.sl, ah_attr.port_num);
+ goto err_exit;
+ }
+
+ if (ib_find_pkey(ibp, data->hfiportnum, IB_DEFAULT_PKEY_FULL,
+ &pkey_idx) < 0) {
+ c_err("%s:full key not found, defaulting to partial\n",
+ __func__);
+ if (ib_find_pkey(ibp, data->hfiportnum, IB_DEFAULT_PKEY_PARTIAL,
+ &pkey_idx) < 0)
+ pkey_idx = 1;
+ }
+
+ send_buf = ib_create_send_mad(port->mad_agent, 1, pkey_idx, 0,
+ IB_MGMT_VENDOR_HDR, IB_MGMT_MAD_DATA,
+ GFP_KERNEL, OPA_MGMT_BASE_VERSION);
+ if (IS_ERR(send_buf)) {
+ c_err("%s:Couldn't allocate send buf\n", __func__);
+ goto err_sndbuf;
+ }
+
+ send_buf->ah = ah;
+
+ /* Set up common MAD hdr */
+ trap_mad = send_buf->mad;
+ trap_mad->mad_hdr.base_version = OPA_MGMT_BASE_VERSION;
+ trap_mad->mad_hdr.mgmt_class = HFI_MGMT_CLASS_INTEL_EMA;
+ trap_mad->mad_hdr.class_version = HFI_EMA_CLASS_VERSION;
+ trap_mad->mad_hdr.method = IB_MGMT_METHOD_TRAP;
+ port->tid++;
+ trap_mad->mad_hdr.tid = cpu_to_be64(port->tid);
+ trap_mad->mad_hdr.attr_id = IB_SMP_ATTR_NOTICE;
+
+ /* Set up vendor OUI */
+ trap_mad->oui[0] = INTEL_OUI_1;
+ trap_mad->oui[1] = INTEL_OUI_2;
+ trap_mad->oui[2] = INTEL_OUI_3;
+
+ /* Setup notice attribute portion */
+ trap_mad->notice.gen_type = HFI_INTEL_EMA_NOTICE_TYPE_INFO << 1;
+ trap_mad->notice.oui_1 = INTEL_OUI_1;
+ trap_mad->notice.oui_2 = INTEL_OUI_2;
+ trap_mad->notice.oui_3 = INTEL_OUI_3;
+ trap_mad->notice.issuer_lid = cpu_to_be32(lid);
+
+ /* copy the actual trap data */
+ trap = (struct hfi_veswport_trap *)trap_mad->notice.raw_data;
+ trap->fabric_id = cpu_to_be16(data->fabric_id);
+ trap->veswid = cpu_to_be16(data->veswid);
+ trap->veswportnum = cpu_to_be32(data->veswportnum);
+ trap->hfiportnum = cpu_to_be16(data->hfiportnum);
+ trap->veswportindex = data->veswportindex;
+ trap->opcode = data->opcode;
+
+ /* If successful send set up rate limit timeout else bail */
+ if (ib_post_send_mad(send_buf, NULL)) {
+ ib_free_send_mad(send_buf);
+ } else {
+ if (adapter->trap_count)
+ return;
+ adapter->trap_timeout = jiffies +
+ usecs_to_jiffies(HFI_VNIC_TRAP_TIMEOUT);
+ return;
+ }
+
+err_sndbuf:
+ ib_destroy_ah(ah);
+err_exit:
+ v_err("Aborting trap\n");
+}
+
+/**
+ * hfi_vnic_vema_add_one -- Handle new ib device
+ * @device: ib device pointer
+ *
+ * Allocate the vnic control port and initialize it.
+ */
+static void hfi_vnic_vema_add_one(struct ib_device *device)
+{
+ struct hfi_vnic_ctrl_port *cport;
+ struct hfi_ibdev *hfidev;
+ int rc, size = sizeof(*cport);
+
+ if (!is_hfi_ibdev(device))
+ return;
+
+ size += device->phys_port_cnt * sizeof(struct hfi_vnic_vema_port);
+ cport = kzalloc(size, GFP_KERNEL);
+ if (!cport)
+ return;
+
+ hfidev = to_hfi_ibdev(device);
+ cport->num_ports = device->phys_port_cnt;
+ cport->ops = &hfidev->vnic_ctrl_ops;
+ cport->ibdev = device;
+
+ /* Initialize hfi vnic management agent (vema) */
+ rc = vema_register(cport);
+ if (!rc)
+ c_info("VNIC client initialized\n");
+
+ ib_set_client_data(device, &hfi_vnic_client, cport);
+}
+
+/**
+ * hfi_vnic_vema_rem_one -- Handle ib device removal
+ * @device: ib device pointer
+ * @client_data: ib client data
+ *
+ * Uninitialize and free the vnic control port.
+ */
+static void hfi_vnic_vema_rem_one(struct ib_device *device,
+ void *client_data)
+{
+ struct hfi_vnic_ctrl_port *cport = client_data;
+
+ if (!cport)
+ return;
+
+ c_info("removing VNIC client\n");
+ vema_unregister(cport);
+ kfree(cport);
+}
+
+static int __init hfi_vnic_init(void)
+{
+ int rc;
+
+ pr_info("HFI Virtual Network Driver - v%s\n",
+ hfi_vnic_driver_version);
+
+ rc = ib_register_client(&hfi_vnic_client);
+ if (rc)
+ pr_err("VNIC driver register failed %d\n", rc);
+
+ return rc;
+}
+module_init(hfi_vnic_init);
+
+static void hfi_vnic_deinit(void)
+{
+ ib_unregister_client(&hfi_vnic_client);
+}
+module_exit(hfi_vnic_deinit);
+
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_AUTHOR("Intel Corporation");
+MODULE_DESCRIPTION("Intel HFI Virtual Network driver");
+MODULE_VERSION(DRV_VERSION);
diff --git a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema_iface.c b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema_iface.c
index f912171f..e669f80 100644
--- a/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema_iface.c
+++ b/drivers/infiniband/sw/intel/hfi_vnic/hfi_vnic_vema_iface.c
@@ -71,7 +71,7 @@ void hfi_vnic_vema_report_event(struct hfi_vnic_adapter *adapter, u8 event)
trap_data.veswportindex = vport->vport_num;
trap_data.opcode = event;
- /* Need to send trap here */
+ hfi_vnic_vema_send_trap(adapter, &trap_data, info->vport.encap_slid);
}
/**
@@ -419,7 +419,7 @@ struct hfi_vnic_adapter *hfi_vnic_add_vport(struct hfi_vnic_ctrl_port *cport,
}
/**
- * hfi_vnic_rem_vport - Remove a new vnic port
+ * hfi_vnic_rem_vport - Remove the vnic port
* @adapter: vnic adapter
*/
void hfi_vnic_rem_vport(struct hfi_vnic_adapter *adapter)
--
1.8.3.1
^ permalink raw reply related [flat|nested] 41+ messages in thread
* [RFC v2 09/10] IB/hfi1: Virtual Network Interface Controller (VNIC) support
2016-12-15 7:59 [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
` (3 preceding siblings ...)
2016-12-15 7:59 ` [RFC v2 08/10] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) function Vishwanathapura, Niranjana
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 10/10] IB/hfi1: VNIC SDMA support Vishwanathapura, Niranjana
2016-12-15 16:56 ` [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Jason Gunthorpe
6 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura, Andrzej Kacprowski
HFI1 HW specific support for VNIC functionality. Add support to add
and remove VNIC ports. Also implement the operations to allocate
resources, transmit and receive of Omni-Path encapsulated Ethernet
packets.
Dynamically allocate a set of contexts for VNIC when the first vnic
port is instantiated. Allocate VNIC contexts from user contexts pool
and return them back to the same pool while freeing up. Set aside
enough MSI-X interrupts for VNIC contexts and assign them when the
contexts are allocated. On the receive side, use an RSM rule to
spread TCP/UDP streams among VNIC contexts.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Andrzej Kacprowski <andrzej.kacprowski@intel.com>
---
drivers/infiniband/hw/hfi1/Makefile | 2 +-
drivers/infiniband/hw/hfi1/aspm.h | 13 +-
drivers/infiniband/hw/hfi1/chip.c | 270 +++++++++++--
drivers/infiniband/hw/hfi1/chip.h | 2 +
drivers/infiniband/hw/hfi1/debugfs.c | 6 +-
drivers/infiniband/hw/hfi1/driver.c | 74 +++-
drivers/infiniband/hw/hfi1/file_ops.c | 25 +-
drivers/infiniband/hw/hfi1/hfi.h | 49 ++-
drivers/infiniband/hw/hfi1/init.c | 37 +-
drivers/infiniband/hw/hfi1/mad.c | 8 +-
drivers/infiniband/hw/hfi1/pio.c | 17 +
drivers/infiniband/hw/hfi1/pio.h | 6 +
drivers/infiniband/hw/hfi1/sysfs.c | 2 +-
drivers/infiniband/hw/hfi1/user_exp_rcv.c | 6 +-
drivers/infiniband/hw/hfi1/user_pages.c | 3 +-
drivers/infiniband/hw/hfi1/verbs.c | 7 +
drivers/infiniband/hw/hfi1/vnic.h | 145 +++++++
drivers/infiniband/hw/hfi1/vnic_main.c | 614 ++++++++++++++++++++++++++++++
drivers/infiniband/hw/hfi1/vnic_sdma.c | 60 +++
include/rdma/opa_port_info.h | 2 +-
20 files changed, 1252 insertions(+), 96 deletions(-)
create mode 100644 drivers/infiniband/hw/hfi1/vnic.h
create mode 100644 drivers/infiniband/hw/hfi1/vnic_main.c
create mode 100644 drivers/infiniband/hw/hfi1/vnic_sdma.c
diff --git a/drivers/infiniband/hw/hfi1/Makefile b/drivers/infiniband/hw/hfi1/Makefile
index 0cf97a0..88085f6 100644
--- a/drivers/infiniband/hw/hfi1/Makefile
+++ b/drivers/infiniband/hw/hfi1/Makefile
@@ -12,7 +12,7 @@ hfi1-y := affinity.o chip.o device.o driver.o efivar.o \
init.o intr.o mad.o mmu_rb.o pcie.o pio.o pio_copy.o platform.o \
qp.o qsfp.o rc.o ruc.o sdma.o sysfs.o trace.o \
uc.o ud.o user_exp_rcv.o user_pages.o user_sdma.o verbs.o \
- verbs_txreq.o
+ verbs_txreq.o vnic_main.o vnic_sdma.o
hfi1-$(CONFIG_DEBUG_FS) += debugfs.o
CFLAGS_trace.o = -I$(src)
diff --git a/drivers/infiniband/hw/hfi1/aspm.h b/drivers/infiniband/hw/hfi1/aspm.h
index 0d58fe3..3a01b69 100644
--- a/drivers/infiniband/hw/hfi1/aspm.h
+++ b/drivers/infiniband/hw/hfi1/aspm.h
@@ -229,14 +229,17 @@ static inline void aspm_ctx_timer_function(unsigned long data)
spin_unlock_irqrestore(&rcd->aspm_lock, flags);
}
-/* Disable interrupt processing for verbs contexts when PSM contexts are open */
+/*
+ * Disable interrupt processing for verbs contexts when PSM or VNIC contexts
+ * are open.
+ */
static inline void aspm_disable_all(struct hfi1_devdata *dd)
{
struct hfi1_ctxtdata *rcd;
unsigned long flags;
unsigned i;
- for (i = 0; i < dd->first_user_ctxt; i++) {
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; i++) {
rcd = dd->rcd[i];
del_timer_sync(&rcd->aspm_timer);
spin_lock_irqsave(&rcd->aspm_lock, flags);
@@ -260,7 +263,7 @@ static inline void aspm_enable_all(struct hfi1_devdata *dd)
if (aspm_mode != ASPM_MODE_DYNAMIC)
return;
- for (i = 0; i < dd->first_user_ctxt; i++) {
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; i++) {
rcd = dd->rcd[i];
spin_lock_irqsave(&rcd->aspm_lock, flags);
rcd->aspm_intr_enable = true;
@@ -276,7 +279,7 @@ static inline void aspm_ctx_init(struct hfi1_ctxtdata *rcd)
(unsigned long)rcd);
rcd->aspm_intr_supported = rcd->dd->aspm_supported &&
aspm_mode == ASPM_MODE_DYNAMIC &&
- rcd->ctxt < rcd->dd->first_user_ctxt;
+ rcd->ctxt < rcd->dd->first_dyn_alloc_ctxt;
}
static inline void aspm_init(struct hfi1_devdata *dd)
@@ -286,7 +289,7 @@ static inline void aspm_init(struct hfi1_devdata *dd)
spin_lock_init(&dd->aspm_lock);
dd->aspm_supported = aspm_hw_l1_supported(dd);
- for (i = 0; i < dd->first_user_ctxt; i++)
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; i++)
aspm_ctx_init(dd->rcd[i]);
/* Start with ASPM disabled */
diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c
index 9263984..472ce55 100644
--- a/drivers/infiniband/hw/hfi1/chip.c
+++ b/drivers/infiniband/hw/hfi1/chip.c
@@ -125,9 +125,16 @@ struct flag_table {
#define DEFAULT_KRCVQS 2
#define MIN_KERNEL_KCTXTS 2
#define FIRST_KERNEL_KCTXT 1
-/* sizes for both the QP and RSM map tables */
-#define NUM_MAP_ENTRIES 256
-#define NUM_MAP_REGS 32
+
+/*
+ * RSM instance allocation
+ * 0 - Verbs
+ * 1 - User Fecn Handling
+ * 2 - Vnic
+ */
+#define RSM_INS_VERBS 0
+#define RSM_INS_FECN 1
+#define RSM_INS_VNIC 2
/* Bit offset into the GUID which carries HFI id information */
#define GUID_HFI_INDEX_SHIFT 39
@@ -138,8 +145,7 @@ struct flag_table {
#define is_emulator_p(dd) ((((dd)->irev) & 0xf) == 3)
#define is_emulator_s(dd) ((((dd)->irev) & 0xf) == 4)
-/* RSM fields */
-
+/* RSM fields for Verbs */
/* packet type */
#define IB_PACKET_TYPE 2ull
#define QW_SHIFT 6ull
@@ -169,6 +175,28 @@ struct flag_table {
/* QPN[m+n:1] QW 1, OFFSET 1 */
#define QPN_SELECT_OFFSET ((1ull << QW_SHIFT) | (1ull))
+/* RSM fields for Vnic */
+/* L2_TYPE: QW 0, OFFSET 61 - for match */
+#define L2_TYPE_QW 0ull
+#define L2_TYPE_BIT_OFFSET 61ull
+#define L2_TYPE_OFFSET(off) ((L2_TYPE_QW << QW_SHIFT) | (off))
+#define L2_TYPE_MATCH_OFFSET L2_TYPE_OFFSET(L2_TYPE_BIT_OFFSET)
+#define L2_TYPE_MASK 3ull
+#define L2_16B_VALUE 2ull
+
+/* L4_TYPE QW 1, OFFSET 0 - for match */
+#define L4_TYPE_QW 1ull
+#define L4_TYPE_BIT_OFFSET 0ull
+#define L4_TYPE_OFFSET(off) ((L4_TYPE_QW << QW_SHIFT) | (off))
+#define L4_TYPE_MATCH_OFFSET L4_TYPE_OFFSET(L4_TYPE_BIT_OFFSET)
+#define L4_16B_TYPE_MASK 0xFFull
+#define L4_16B_ETH_VALUE 0x78ull
+
+/* 16B VESWID - for select */
+#define L4_16B_HDR_VESWID_OFFSET ((2 << QW_SHIFT) | (16ull))
+/* 16B ENTROPY - for select */
+#define L2_16B_ENTROPY_OFFSET ((1 << QW_SHIFT) | (32ull))
+
/* defines to build power on SC2VL table */
#define SC2VL_VAL( \
num, \
@@ -1045,6 +1073,7 @@ static int wait_logical_linkstate(struct hfi1_pportdata *ppd, u32 state,
static int qos_rmt_entries(struct hfi1_devdata *dd, unsigned int *mp,
unsigned int *np);
static void clear_full_mgmt_pkey(struct hfi1_pportdata *ppd);
+static void clear_rsm_rule(struct hfi1_devdata *dd, u8 rule_index);
/*
* Error interrupt table entry. This is used as input to the interrupt
@@ -6712,7 +6741,13 @@ static void rxe_kernel_unfreeze(struct hfi1_devdata *dd)
int i;
/* enable all kernel contexts */
- for (i = 0; i < dd->n_krcv_queues; i++) {
+ for (i = 0; i < dd->num_rcv_contexts; i++) {
+ struct hfi1_ctxtdata *rcd = dd->rcd[i];
+
+ /* Ensure all non-user contexts(including vnic) are enabled */
+ if (!rcd || !rcd->sc || (rcd->sc->type == SC_USER))
+ continue;
+
rcvmask = HFI1_RCVCTRL_CTXT_ENB;
/* HFI1_RCVCTRL_TAILUPD_[ENB|DIS] needs to be set explicitly */
rcvmask |= HFI1_CAP_KGET_MASK(dd->rcd[i]->flags, DMA_RTAIL) ?
@@ -8004,7 +8039,9 @@ static void is_rcv_avail_int(struct hfi1_devdata *dd, unsigned int source)
if (likely(source < dd->num_rcv_contexts)) {
rcd = dd->rcd[source];
if (rcd) {
- if (source < dd->first_user_ctxt)
+ /* Check for non-user contexts, including vnic */
+ if ((source < dd->first_dyn_alloc_ctxt) ||
+ (rcd->sc && (rcd->sc->type == SC_KERNEL)))
rcd->do_interrupt(rcd, 0);
else
handle_user_interrupt(rcd);
@@ -8032,7 +8069,8 @@ static void is_rcv_urgent_int(struct hfi1_devdata *dd, unsigned int source)
rcd = dd->rcd[source];
if (rcd) {
/* only pay attention to user urgent interrupts */
- if (source >= dd->first_user_ctxt)
+ if ((source >= dd->first_dyn_alloc_ctxt) &&
+ (!rcd->sc || (rcd->sc->type == SC_USER)))
handle_user_interrupt(rcd);
return; /* OK */
}
@@ -12736,7 +12774,7 @@ static int request_msix_irqs(struct hfi1_devdata *dd)
first_sdma = last_general;
last_sdma = first_sdma + dd->num_sdma;
first_rx = last_sdma;
- last_rx = first_rx + dd->n_krcv_queues;
+ last_rx = first_rx + dd->n_krcv_queues + HFI1_NUM_VNIC_CTXT;
/*
* Sanity check - the code expects all SDMA chip source
@@ -12750,7 +12788,7 @@ static int request_msix_irqs(struct hfi1_devdata *dd)
const char *err_info;
irq_handler_t handler;
irq_handler_t thread = NULL;
- void *arg;
+ void *arg = NULL;
int idx;
struct hfi1_ctxtdata *rcd = NULL;
struct sdma_engine *sde = NULL;
@@ -12777,24 +12815,24 @@ static int request_msix_irqs(struct hfi1_devdata *dd)
} else if (first_rx <= i && i < last_rx) {
idx = i - first_rx;
rcd = dd->rcd[idx];
- /* no interrupt if no rcd */
- if (!rcd)
- continue;
- /*
- * Set the interrupt register and mask for this
- * context's interrupt.
- */
- rcd->ireg = (IS_RCVAVAIL_START + idx) / 64;
- rcd->imask = ((u64)1) <<
- ((IS_RCVAVAIL_START + idx) % 64);
- handler = receive_context_interrupt;
- thread = receive_context_thread;
- arg = rcd;
- snprintf(me->name, sizeof(me->name),
- DRIVER_NAME "_%d kctxt%d", dd->unit, idx);
- err_info = "receive context";
- remap_intr(dd, IS_RCVAVAIL_START + idx, i);
- me->type = IRQ_RCVCTXT;
+ if (rcd) {
+ /*
+ * Set the interrupt register and mask for this
+ * context's interrupt.
+ */
+ rcd->ireg = (IS_RCVAVAIL_START + idx) / 64;
+ rcd->imask = ((u64)1) <<
+ ((IS_RCVAVAIL_START + idx) % 64);
+ handler = receive_context_interrupt;
+ thread = receive_context_thread;
+ arg = rcd;
+ snprintf(me->name, sizeof(me->name),
+ DRIVER_NAME "_%d kctxt%d",
+ dd->unit, idx);
+ err_info = "receive context";
+ remap_intr(dd, IS_RCVAVAIL_START + idx, i);
+ me->type = IRQ_RCVCTXT;
+ }
} else {
/* not in our expected range - complain, then
* ignore it
@@ -12832,6 +12870,67 @@ static int request_msix_irqs(struct hfi1_devdata *dd)
return ret;
}
+void hfi1_reset_vnic_msix_info(struct hfi1_ctxtdata *rcd)
+{
+ int idx = rcd->ctxt;
+ struct hfi1_devdata *dd = rcd->dd;
+ int i = 1 + dd->num_sdma + idx;
+ struct hfi1_msix_entry *me = &dd->msix_entries[i];
+
+ if (!me->arg) /* => no irq, no affinity */
+ return;
+
+ hfi1_put_irq_affinity(dd, me);
+ free_irq(me->msix.vector, me->arg);
+
+ me->arg = NULL;
+}
+
+void hfi1_set_vnic_msix_info(struct hfi1_ctxtdata *rcd)
+{
+ int idx = rcd->ctxt;
+ void *arg = rcd;
+ int ret;
+ struct hfi1_devdata *dd = rcd->dd;
+ int i = 1 + dd->num_sdma + idx;
+ struct hfi1_msix_entry *me = &dd->msix_entries[i];
+
+ /*
+ * Set the interrupt register and mask for this
+ * context's interrupt.
+ */
+ rcd->ireg = (IS_RCVAVAIL_START + idx) / 64;
+ rcd->imask = ((u64)1) <<
+ ((IS_RCVAVAIL_START + idx) % 64);
+
+ snprintf(me->name, sizeof(me->name),
+ DRIVER_NAME "_%d kctxt%d", dd->unit, idx);
+ me->name[sizeof(me->name) - 1] = 0;
+ me->type = IRQ_RCVCTXT;
+
+ remap_intr(dd, IS_RCVAVAIL_START + idx, i);
+
+ ret = request_threaded_irq(me->msix.vector, receive_context_interrupt,
+ receive_context_thread, 0, me->name, arg);
+ if (ret) {
+ dd_dev_err(dd, "vnic irq request (vector %d, idx %d) fail %d\n",
+ me->msix.vector, idx, ret);
+ return;
+ }
+ /*
+ * assign arg after request_irq call, so it will be
+ * cleaned up
+ */
+ me->arg = arg;
+
+ ret = hfi1_get_irq_affinity(dd, me);
+ if (ret) {
+ dd_dev_err(dd,
+ "unable to pin IRQ %d\n", ret);
+ free_irq(me->msix.vector, me->arg);
+ }
+}
+
/*
* Set the general handler to accept all interrupts, remap all
* chip interrupts back to MSI-X 0.
@@ -12863,7 +12962,7 @@ static int set_up_interrupts(struct hfi1_devdata *dd)
* N interrupts - one per used SDMA engine
* M interrupt - one per kernel receive context
*/
- total = 1 + dd->num_sdma + dd->n_krcv_queues;
+ total = 1 + dd->num_sdma + dd->n_krcv_queues + HFI1_NUM_VNIC_CTXT;
entries = kcalloc(total, sizeof(*entries), GFP_KERNEL);
if (!entries) {
@@ -12928,7 +13027,8 @@ static int set_up_interrupts(struct hfi1_devdata *dd)
*
* num_rcv_contexts - number of contexts being used
* n_krcv_queues - number of kernel contexts
- * first_user_ctxt - first non-kernel context in array of contexts
+ * first_dyn_alloc_ctxt - first dynamically allocated context
+ * in array of contexts
* freectxts - number of free user contexts
* num_send_contexts - number of PIO send contexts being used
*/
@@ -13005,10 +13105,14 @@ static int set_up_context_variables(struct hfi1_devdata *dd)
total_contexts = num_kernel_contexts + num_user_contexts;
}
- /* the first N are kernel contexts, the rest are user contexts */
+ /* Accommodate VNIC contexts */
+ if ((total_contexts + HFI1_NUM_VNIC_CTXT) <= dd->chip_rcv_contexts)
+ total_contexts += HFI1_NUM_VNIC_CTXT;
+
+ /* the first N are kernel contexts, the rest are user/vnic contexts */
dd->num_rcv_contexts = total_contexts;
dd->n_krcv_queues = num_kernel_contexts;
- dd->first_user_ctxt = num_kernel_contexts;
+ dd->first_dyn_alloc_ctxt = num_kernel_contexts;
dd->num_user_contexts = num_user_contexts;
dd->freectxts = num_user_contexts;
dd_dev_info(dd,
@@ -13464,11 +13568,8 @@ static void reset_rxe_csrs(struct hfi1_devdata *dd)
write_csr(dd, RCV_COUNTER_ARRAY32 + (8 * i), 0);
for (i = 0; i < RXE_NUM_64_BIT_COUNTERS; i++)
write_csr(dd, RCV_COUNTER_ARRAY64 + (8 * i), 0);
- for (i = 0; i < RXE_NUM_RSM_INSTANCES; i++) {
- write_csr(dd, RCV_RSM_CFG + (8 * i), 0);
- write_csr(dd, RCV_RSM_SELECT + (8 * i), 0);
- write_csr(dd, RCV_RSM_MATCH + (8 * i), 0);
- }
+ for (i = 0; i < RXE_NUM_RSM_INSTANCES; i++)
+ clear_rsm_rule(dd, i);
for (i = 0; i < 32; i++)
write_csr(dd, RCV_RSM_MAP_TABLE + (8 * i), 0);
@@ -13827,6 +13928,16 @@ static void add_rsm_rule(struct hfi1_devdata *dd, u8 rule_index,
(u64)rrd->value2 << RCV_RSM_MATCH_VALUE2_SHIFT);
}
+/*
+ * Clear a receive side mapping rule.
+ */
+static void clear_rsm_rule(struct hfi1_devdata *dd, u8 rule_index)
+{
+ write_csr(dd, RCV_RSM_CFG + (8 * rule_index), 0);
+ write_csr(dd, RCV_RSM_SELECT + (8 * rule_index), 0);
+ write_csr(dd, RCV_RSM_MATCH + (8 * rule_index), 0);
+}
+
/* return the number of RSM map table entries that will be used for QOS */
static int qos_rmt_entries(struct hfi1_devdata *dd, unsigned int *mp,
unsigned int *np)
@@ -13942,7 +14053,7 @@ static void init_qos(struct hfi1_devdata *dd, struct rsm_map_table *rmt)
rrd.value2 = LRH_SC_VALUE;
/* add rule 0 */
- add_rsm_rule(dd, 0, &rrd);
+ add_rsm_rule(dd, RSM_INS_VERBS, &rrd);
/* mark RSM map entries as used */
rmt->used += rmt_entries;
@@ -13972,7 +14083,7 @@ static void init_user_fecn_handling(struct hfi1_devdata *dd,
/*
* RSM will extract the destination context as an index into the
* map table. The destination contexts are a sequential block
- * in the range first_user_ctxt...num_rcv_contexts-1 (inclusive).
+ * in the range first_dyn_alloc_ctxt...num_rcv_contexts-1 (inclusive).
* Map entries are accessed as offset + extracted value. Adjust
* the added offset so this sequence can be placed anywhere in
* the table - as long as the entries themselves do not wrap.
@@ -13980,9 +14091,9 @@ static void init_user_fecn_handling(struct hfi1_devdata *dd,
* start with that to allow for a "negative" offset.
*/
offset = (u8)(NUM_MAP_ENTRIES + (int)rmt->used -
- (int)dd->first_user_ctxt);
+ (int)dd->first_dyn_alloc_ctxt);
- for (i = dd->first_user_ctxt, idx = rmt->used;
+ for (i = dd->first_dyn_alloc_ctxt, idx = rmt->used;
i < dd->num_rcv_contexts; i++, idx++) {
/* replace with identity mapping */
regoff = (idx % 8) * 8;
@@ -14016,11 +14127,84 @@ static void init_user_fecn_handling(struct hfi1_devdata *dd,
rrd.value2 = 1;
/* add rule 1 */
- add_rsm_rule(dd, 1, &rrd);
+ add_rsm_rule(dd, RSM_INS_FECN, &rrd);
rmt->used += dd->num_user_contexts;
}
+/* Initialize RSM for VNIC */
+void hfi1_init_vnic_rsm(struct hfi1_devdata *dd)
+{
+ u8 i, j;
+ u8 ctx_id = 0;
+ u64 reg;
+ u32 regoff;
+ struct rsm_rule_data rrd;
+
+ if (hfi1_vnic_is_rsm_full(dd, NUM_VNIC_MAP_ENTRIES)) {
+ dd_dev_err(dd, "Vnic RSM disabled, rmt entries used = %d\n",
+ dd->vnic.rmt_start);
+ return;
+ }
+
+ dev_dbg(&(dd)->pcidev->dev, "Vnic rsm start = %d, end %d\n",
+ dd->vnic.rmt_start,
+ dd->vnic.rmt_start + NUM_VNIC_MAP_ENTRIES);
+
+ /* Update RSM mapping table, 32 regs, 256 entries - 1 ctx per byte */
+ regoff = RCV_RSM_MAP_TABLE + (dd->vnic.rmt_start / 8) * 8;
+ reg = read_csr(dd, regoff);
+ for (i = 0; i < NUM_VNIC_MAP_ENTRIES; i++) {
+ /* Update map register with vnic context */
+ j = (dd->vnic.rmt_start + i) % 8;
+ reg &= ~(0xffllu << (j * 8));
+ reg |= (u64)dd->vnic.ctxt[ctx_id++]->ctxt << (j * 8);
+ /* Wrap up vnic ctx index */
+ ctx_id %= dd->vnic.num_ctxt;
+ /* Write back map register */
+ if (j == 7 || ((i + 1) == NUM_VNIC_MAP_ENTRIES)) {
+ dev_dbg(&(dd)->pcidev->dev,
+ "Vnic rsm map reg[%d] =0x%llx\n",
+ regoff - RCV_RSM_MAP_TABLE, reg);
+
+ write_csr(dd, regoff, reg);
+ regoff += 8;
+ if (i < (NUM_VNIC_MAP_ENTRIES - 1))
+ reg = read_csr(dd, regoff);
+ }
+ }
+
+ /* Add rule for vnic */
+ rrd.offset = dd->vnic.rmt_start;
+ rrd.pkt_type = 4;
+ /* Match 16B packets */
+ rrd.field1_off = L2_TYPE_MATCH_OFFSET;
+ rrd.mask1 = L2_TYPE_MASK;
+ rrd.value1 = L2_16B_VALUE;
+ /* Match ETH L4 packets */
+ rrd.field2_off = L4_TYPE_MATCH_OFFSET;
+ rrd.mask2 = L4_16B_TYPE_MASK;
+ rrd.value2 = L4_16B_ETH_VALUE;
+ /* Calc context from veswid and entropy */
+ rrd.index1_off = L4_16B_HDR_VESWID_OFFSET;
+ rrd.index1_width = ilog2(NUM_VNIC_MAP_ENTRIES);
+ rrd.index2_off = L2_16B_ENTROPY_OFFSET;
+ rrd.index2_width = ilog2(NUM_VNIC_MAP_ENTRIES);
+ add_rsm_rule(dd, RSM_INS_VNIC, &rrd);
+
+ /* Enable RSM if not already enabled */
+ add_rcvctrl(dd, RCV_CTRL_RCV_RSM_ENABLE_SMASK);
+}
+
+void hfi1_deinit_vnic_rsm(struct hfi1_devdata *dd)
+{
+ clear_rsm_rule(dd, RSM_INS_VNIC);
+
+ /* Disable RSM if used only by vnic */
+ if (dd->vnic.rmt_start == 0)
+ clear_rcvctrl(dd, RCV_CTRL_RCV_RSM_ENABLE_SMASK);
+}
+
static void init_rxe(struct hfi1_devdata *dd)
{
struct rsm_map_table *rmt;
@@ -14033,6 +14217,8 @@ static void init_rxe(struct hfi1_devdata *dd)
init_qos(dd, rmt);
init_user_fecn_handling(dd, rmt);
complete_rsm_map_table(dd, rmt);
+ /* record number of used rsm map entries for vnic */
+ dd->vnic.rmt_start = rmt->used;
kfree(rmt);
/*
diff --git a/drivers/infiniband/hw/hfi1/chip.h b/drivers/infiniband/hw/hfi1/chip.h
index 9234525..1e177f5 100644
--- a/drivers/infiniband/hw/hfi1/chip.h
+++ b/drivers/infiniband/hw/hfi1/chip.h
@@ -1355,6 +1355,8 @@ void hfi1_put_tid(struct hfi1_devdata *dd, u32 index,
int hfi1_set_ctxt_pkey(struct hfi1_devdata *dd, unsigned ctxt, u16 pkey);
int hfi1_clear_ctxt_pkey(struct hfi1_devdata *dd, unsigned ctxt);
void hfi1_read_link_quality(struct hfi1_devdata *dd, u8 *link_quality);
+void hfi1_init_vnic_rsm(struct hfi1_devdata *dd);
+void hfi1_deinit_vnic_rsm(struct hfi1_devdata *dd);
/*
* Interrupt source table.
diff --git a/drivers/infiniband/hw/hfi1/debugfs.c b/drivers/infiniband/hw/hfi1/debugfs.c
index 8725f4c..dc8bb86 100644
--- a/drivers/infiniband/hw/hfi1/debugfs.c
+++ b/drivers/infiniband/hw/hfi1/debugfs.c
@@ -169,7 +169,7 @@ static int _opcode_stats_seq_show(struct seq_file *s, void *v)
struct hfi1_ibdev *ibd = (struct hfi1_ibdev *)s->private;
struct hfi1_devdata *dd = dd_from_dev(ibd);
- for (j = 0; j < dd->first_user_ctxt; j++) {
+ for (j = 0; j < dd->first_dyn_alloc_ctxt; j++) {
if (!dd->rcd[j])
continue;
n_packets += dd->rcd[j]->opstats->stats[i].n_packets;
@@ -195,7 +195,7 @@ static void *_ctx_stats_seq_start(struct seq_file *s, loff_t *pos)
if (!*pos)
return SEQ_START_TOKEN;
- if (*pos >= dd->first_user_ctxt)
+ if (*pos >= dd->first_dyn_alloc_ctxt)
return NULL;
return pos;
}
@@ -209,7 +209,7 @@ static void *_ctx_stats_seq_next(struct seq_file *s, void *v, loff_t *pos)
return pos;
++*pos;
- if (*pos >= dd->first_user_ctxt)
+ if (*pos >= dd->first_dyn_alloc_ctxt)
return NULL;
return pos;
}
diff --git a/drivers/infiniband/hw/hfi1/driver.c b/drivers/infiniband/hw/hfi1/driver.c
index e219c3b..fbe40b3 100644
--- a/drivers/infiniband/hw/hfi1/driver.c
+++ b/drivers/infiniband/hw/hfi1/driver.c
@@ -59,6 +59,7 @@
#include "trace.h"
#include "qp.h"
#include "sdma.h"
+#include "vnic.h"
#undef pr_fmt
#define pr_fmt(fmt) DRIVER_NAME ": " fmt
@@ -860,20 +861,42 @@ int handle_receive_interrupt_dma_rtail(struct hfi1_ctxtdata *rcd, int thread)
return last;
}
-static inline void set_all_nodma_rtail(struct hfi1_devdata *dd)
+static inline void set_nodma_rtail(struct hfi1_devdata *dd, u8 ctxt)
{
int i;
- for (i = HFI1_CTRL_CTXT + 1; i < dd->first_user_ctxt; i++)
+ /*
+ * For dynamically allocated kernel contexts (like vnic) switch
+ * interrupt handler only for that context. Otherwise, switch
+ * interrupt handler for all statically allocated kernel contexts.
+ */
+ if (ctxt >= dd->first_dyn_alloc_ctxt) {
+ dd->rcd[ctxt]->do_interrupt =
+ &handle_receive_interrupt_nodma_rtail;
+ return;
+ }
+
+ for (i = HFI1_CTRL_CTXT + 1; i < dd->first_dyn_alloc_ctxt; i++)
dd->rcd[i]->do_interrupt =
&handle_receive_interrupt_nodma_rtail;
}
-static inline void set_all_dma_rtail(struct hfi1_devdata *dd)
+static inline void set_dma_rtail(struct hfi1_devdata *dd, u8 ctxt)
{
int i;
- for (i = HFI1_CTRL_CTXT + 1; i < dd->first_user_ctxt; i++)
+ /*
+ * For dynamically allocated kernel contexts (like vnic) switch
+ * interrupt handler only for that context. Otherwise, switch
+ * interrupt handler for all statically allocated kernel contexts.
+ */
+ if (ctxt >= dd->first_dyn_alloc_ctxt) {
+ dd->rcd[ctxt]->do_interrupt =
+ &handle_receive_interrupt_dma_rtail;
+ return;
+ }
+
+ for (i = HFI1_CTRL_CTXT + 1; i < dd->first_dyn_alloc_ctxt; i++)
dd->rcd[i]->do_interrupt =
&handle_receive_interrupt_dma_rtail;
}
@@ -883,8 +906,13 @@ void set_all_slowpath(struct hfi1_devdata *dd)
int i;
/* HFI1_CTRL_CTXT must always use the slow path interrupt handler */
- for (i = HFI1_CTRL_CTXT + 1; i < dd->first_user_ctxt; i++)
- dd->rcd[i]->do_interrupt = &handle_receive_interrupt;
+ for (i = HFI1_CTRL_CTXT + 1; i < dd->num_rcv_contexts; i++) {
+ struct hfi1_ctxtdata *rcd = dd->rcd[i];
+
+ if ((i < dd->first_dyn_alloc_ctxt) ||
+ (rcd && rcd->sc && (rcd->sc->type == SC_KERNEL)))
+ rcd->do_interrupt = &handle_receive_interrupt;
+ }
}
static inline int set_armed_to_active(struct hfi1_ctxtdata *rcd,
@@ -994,7 +1022,7 @@ int handle_receive_interrupt(struct hfi1_ctxtdata *rcd, int thread)
last = RCV_PKT_DONE;
if (needset) {
dd_dev_info(dd, "Switching to NO_DMA_RTAIL\n");
- set_all_nodma_rtail(dd);
+ set_nodma_rtail(dd, rcd->ctxt);
needset = 0;
}
} else {
@@ -1016,7 +1044,7 @@ int handle_receive_interrupt(struct hfi1_ctxtdata *rcd, int thread)
if (needset) {
dd_dev_info(dd,
"Switching to DMA_RTAIL\n");
- set_all_dma_rtail(dd);
+ set_dma_rtail(dd, rcd->ctxt);
needset = 0;
}
}
@@ -1064,10 +1092,10 @@ void receive_interrupt_work(struct work_struct *work)
set_link_state(ppd, HLS_UP_ACTIVE);
/*
- * Interrupt all kernel contexts that could have had an
- * interrupt during auto activation.
+ * Interrupt all statically allocated kernel contexts that could
+ * have had an interrupt during auto activation.
*/
- for (i = HFI1_CTRL_CTXT; i < dd->first_user_ctxt; i++)
+ for (i = HFI1_CTRL_CTXT; i < dd->first_dyn_alloc_ctxt; i++)
force_recv_intr(dd->rcd[i]);
}
@@ -1281,7 +1309,8 @@ int hfi1_reset_device(int unit)
spin_lock_irqsave(&dd->uctxt_lock, flags);
if (dd->rcd)
- for (i = dd->first_user_ctxt; i < dd->num_rcv_contexts; i++) {
+ for (i = dd->first_dyn_alloc_ctxt;
+ i < dd->num_rcv_contexts; i++) {
if (!dd->rcd[i] || !dd->rcd[i]->cnt)
continue;
spin_unlock_irqrestore(&dd->uctxt_lock, flags);
@@ -1359,13 +1388,30 @@ int process_receive_ib(struct hfi1_packet *packet)
return RHF_RCV_CONTINUE;
}
+static inline bool hfi1_is_vnic_packet(struct hfi1_packet *packet)
+{
+ /* Packet received in VNIC context via RSM */
+ if (packet->rcd->is_vnic)
+ return true;
+
+ if ((HFI1_GET_L2_TYPE(packet->ebuf) == HFI1_L2_TYPE_HDR_16B) &&
+ (HFI1_GET_L4_TYPE(packet->ebuf) == HFI1_VNIC_L4_ETHR))
+ return true;
+
+ return false;
+}
+
int process_receive_bypass(struct hfi1_packet *packet)
{
- if (unlikely(rhf_err_flags(packet->rhf)))
+ if (unlikely(rhf_err_flags(packet->rhf))) {
handle_eflags(packet);
+ } else if (hfi1_is_vnic_packet(packet)) {
+ hfi1_vnic_bypass_rcv(packet);
+ return RHF_RCV_CONTINUE;
+ }
dd_dev_err(packet->rcd->dd,
- "Bypass packets are not supported in normal operation. Dropping\n");
+ "Unsupported bypass packet. Dropping\n");
incr_cntr64(&packet->rcd->dd->sw_rcv_bypass_packet_errors);
return RHF_RCV_CONTINUE;
}
diff --git a/drivers/infiniband/hw/hfi1/file_ops.c b/drivers/infiniband/hw/hfi1/file_ops.c
index 677efa0..863fbbb 100644
--- a/drivers/infiniband/hw/hfi1/file_ops.c
+++ b/drivers/infiniband/hw/hfi1/file_ops.c
@@ -576,8 +576,8 @@ static int hfi1_file_mmap(struct file *fp, struct vm_area_struct *vma)
* knows where it's own bitmap is within the page.
*/
memaddr = (unsigned long)(dd->events +
- ((uctxt->ctxt - dd->first_user_ctxt) *
- HFI1_MAX_SHARED_CTXTS)) & PAGE_MASK;
+ ((uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
+ HFI1_MAX_SHARED_CTXTS)) & PAGE_MASK;
memlen = PAGE_SIZE;
/*
* v3.7 removes VM_RESERVED but the effect is kept by
@@ -746,7 +746,7 @@ static int hfi1_file_close(struct inode *inode, struct file *fp)
* Clear any left over, unhandled events so the next process that
* gets this context doesn't get confused.
*/
- ev = dd->events + ((uctxt->ctxt - dd->first_user_ctxt) *
+ ev = dd->events + ((uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS) + fdata->subctxt;
*ev = 0;
@@ -895,12 +895,18 @@ static int find_shared_ctxt(struct file *fp,
if (!(dd && (dd->flags & HFI1_PRESENT) && dd->kregbase))
continue;
- for (i = dd->first_user_ctxt; i < dd->num_rcv_contexts; i++) {
+ for (i = dd->first_dyn_alloc_ctxt;
+ i < dd->num_rcv_contexts; i++) {
struct hfi1_ctxtdata *uctxt = dd->rcd[i];
/* Skip ctxts which are not yet open */
if (!uctxt || !uctxt->cnt)
continue;
+
+ /* Skip dynamically allocted kernel contexts */
+ if (uctxt->sc && (uctxt->sc->type == SC_KERNEL))
+ continue;
+
/* Skip ctxt if it doesn't match the requested one */
if (memcmp(uctxt->uuid, uinfo->uuid,
sizeof(uctxt->uuid)) ||
@@ -946,7 +952,8 @@ static int allocate_ctxt(struct file *fp, struct hfi1_devdata *dd,
return -EIO;
}
- for (ctxt = dd->first_user_ctxt; ctxt < dd->num_rcv_contexts; ctxt++)
+ for (ctxt = dd->first_dyn_alloc_ctxt;
+ ctxt < dd->num_rcv_contexts; ctxt++)
if (!dd->rcd[ctxt])
break;
@@ -1292,7 +1299,7 @@ static int get_base_info(struct file *fp, void __user *ubase, __u32 len)
*/
binfo.user_regbase = HFI1_MMAP_TOKEN(UREGS, uctxt->ctxt,
fd->subctxt, 0);
- offset = offset_in_page((((uctxt->ctxt - dd->first_user_ctxt) *
+ offset = offset_in_page((((uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS) + fd->subctxt) *
sizeof(*dd->events));
binfo.events_bufbase = HFI1_MMAP_TOKEN(EVENTS, uctxt->ctxt,
@@ -1386,12 +1393,12 @@ int hfi1_set_uevent_bits(struct hfi1_pportdata *ppd, const int evtbit)
}
spin_lock_irqsave(&dd->uctxt_lock, flags);
- for (ctxt = dd->first_user_ctxt; ctxt < dd->num_rcv_contexts;
+ for (ctxt = dd->first_dyn_alloc_ctxt; ctxt < dd->num_rcv_contexts;
ctxt++) {
uctxt = dd->rcd[ctxt];
if (uctxt) {
unsigned long *evs = dd->events +
- (uctxt->ctxt - dd->first_user_ctxt) *
+ (uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS;
int i;
/*
@@ -1463,7 +1470,7 @@ static int user_event_ack(struct hfi1_ctxtdata *uctxt, int subctxt,
if (!dd->events)
return 0;
- evs = dd->events + ((uctxt->ctxt - dd->first_user_ctxt) *
+ evs = dd->events + ((uctxt->ctxt - dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS) + subctxt;
for (i = 0; i <= _HFI1_MAX_EVENT_BIT; i++) {
diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h
index 1fc5b68..78d1726 100644
--- a/drivers/infiniband/hw/hfi1/hfi.h
+++ b/drivers/infiniband/hw/hfi1/hfi.h
@@ -54,6 +54,7 @@
#include <linux/list.h>
#include <linux/scatterlist.h>
#include <linux/slab.h>
+#include <linux/idr.h>
#include <linux/io.h>
#include <linux/fs.h>
#include <linux/completion.h>
@@ -66,6 +67,7 @@
#include <linux/i2c-algo-bit.h>
#include <rdma/ib_hdrs.h>
#include <linux/rhashtable.h>
+#include <linux/netdevice.h>
#include <rdma/rdma_vt.h>
#include "chip_registers.h"
@@ -337,6 +339,12 @@ struct hfi1_ctxtdata {
* packets with the wrong interrupt handler.
*/
int (*do_interrupt)(struct hfi1_ctxtdata *rcd, int threaded);
+
+ /* Indicates that this is vnic context */
+ bool is_vnic;
+
+ /* vnic queue index this context is mapped to */
+ u8 vnic_q_idx;
};
/*
@@ -831,6 +839,30 @@ struct hfi1_asic_data {
struct hfi1_i2c_bus *i2c_bus1;
};
+/* sizes for both the QP and RSM map tables */
+#define NUM_MAP_ENTRIES 256
+#define NUM_MAP_REGS 32
+
+/*
+ * Number of VNIC contexts used. Ensure it is less than or equal to
+ * max queues supported by VNIC (HFI_VNIC_MAX_QUEUE).
+ */
+#define HFI1_NUM_VNIC_CTXT 8
+
+/* Number of VNIC RSM entries */
+#define NUM_VNIC_MAP_ENTRIES 8
+
+/* Virtual NIC information */
+struct hfi1_vnic_data {
+ struct hfi1_ctxtdata *ctxt[HFI1_NUM_VNIC_CTXT];
+ u8 num_vports;
+ struct idr vesw_idr;
+ u8 rmt_start;
+ u8 num_ctxt;
+};
+
+struct hfi1_vnic_vport_info;
+
/* device data struct now contains only "general per-device" info.
* fields related to a physical IB port are in a hfi1_pportdata struct.
*/
@@ -1140,6 +1172,9 @@ struct hfi1_devdata {
send_routine process_dma_send;
void (*pio_inline_send)(struct hfi1_devdata *dd, struct pio_buf *pbuf,
u64 pbc, const void *from, size_t count);
+ int (*process_vnic_dma_send)(struct hfi1_devdata *dd, u8 q_idx,
+ struct hfi1_vnic_vport_info *vinfo,
+ struct sk_buff *skb, u64 pbc, u8 plen);
/* hfi1_pportdata, points to array of (physical) port-specific
* data structs, indexed by pidx (0..n-1)
*/
@@ -1151,8 +1186,8 @@ struct hfi1_devdata {
u16 flags;
/* Number of physical ports available */
u8 num_pports;
- /* Lowest context number which can be used by user processes */
- u8 first_user_ctxt;
+ /* Lowest context number which can be used by user processes or VNIC */
+ u8 first_dyn_alloc_ctxt;
/* adding a new field here would make it part of this cacheline */
/* seqlock for sc2vl */
@@ -1191,8 +1226,16 @@ struct hfi1_devdata {
struct rhashtable sdma_rht;
struct kobject kobj;
+
+ /* vnic data */
+ struct hfi1_vnic_data vnic;
};
+static inline bool hfi1_vnic_is_rsm_full(struct hfi1_devdata *dd, int spare)
+{
+ return (dd->vnic.rmt_start + spare) > NUM_MAP_ENTRIES;
+}
+
/* 8051 firmware version helper */
#define dc8051_ver(a, b) ((a) << 8 | (b))
#define dc8051_ver_maj(a) ((a & 0xff00) >> 8)
@@ -1258,6 +1301,8 @@ void hfi1_init_pportdata(struct pci_dev *, struct hfi1_pportdata *,
int handle_receive_interrupt_nodma_rtail(struct hfi1_ctxtdata *, int);
int handle_receive_interrupt_dma_rtail(struct hfi1_ctxtdata *, int);
void set_all_slowpath(struct hfi1_devdata *dd);
+void hfi1_set_vnic_msix_info(struct hfi1_ctxtdata *rcd);
+void hfi1_reset_vnic_msix_info(struct hfi1_ctxtdata *rcd);
extern const struct pci_device_id hfi1_pci_tbl[];
diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c
index 13f6862..0b510d5 100644
--- a/drivers/infiniband/hw/hfi1/init.c
+++ b/drivers/infiniband/hw/hfi1/init.c
@@ -65,6 +65,7 @@
#include "verbs.h"
#include "aspm.h"
#include "affinity.h"
+#include "vnic.h"
#undef pr_fmt
#define pr_fmt(fmt) DRIVER_NAME ": " fmt
@@ -139,7 +140,7 @@ int hfi1_create_ctxts(struct hfi1_devdata *dd)
goto nomem;
/* create one or more kernel contexts */
- for (i = 0; i < dd->first_user_ctxt; ++i) {
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; ++i) {
struct hfi1_pportdata *ppd;
struct hfi1_ctxtdata *rcd;
@@ -213,9 +214,9 @@ struct hfi1_ctxtdata *hfi1_create_ctxtdata(struct hfi1_pportdata *ppd, u32 ctxt,
u32 base;
if (dd->rcv_entries.nctxt_extra >
- dd->num_rcv_contexts - dd->first_user_ctxt)
+ dd->num_rcv_contexts - dd->first_dyn_alloc_ctxt)
kctxt_ngroups = (dd->rcv_entries.nctxt_extra -
- (dd->num_rcv_contexts - dd->first_user_ctxt));
+ (dd->num_rcv_contexts - dd->first_dyn_alloc_ctxt));
rcd = kzalloc(sizeof(*rcd), GFP_KERNEL);
if (rcd) {
u32 rcvtids, max_entries;
@@ -237,10 +238,10 @@ struct hfi1_ctxtdata *hfi1_create_ctxtdata(struct hfi1_pportdata *ppd, u32 ctxt,
* Calculate the context's RcvArray entry starting point.
* We do this here because we have to take into account all
* the RcvArray entries that previous context would have
- * taken and we have to account for any extra groups
- * assigned to the kernel or user contexts.
+ * taken and we have to account for any extra groups assigned
+ * to the static (kernel) or dynamic (vnic/user) contexts.
*/
- if (ctxt < dd->first_user_ctxt) {
+ if (ctxt < dd->first_dyn_alloc_ctxt) {
if (ctxt < kctxt_ngroups) {
base = ctxt * (dd->rcv_entries.ngroups + 1);
rcd->rcv_array_groups++;
@@ -248,7 +249,7 @@ struct hfi1_ctxtdata *hfi1_create_ctxtdata(struct hfi1_pportdata *ppd, u32 ctxt,
base = kctxt_ngroups +
(ctxt * dd->rcv_entries.ngroups);
} else {
- u16 ct = ctxt - dd->first_user_ctxt;
+ u16 ct = ctxt - dd->first_dyn_alloc_ctxt;
base = ((dd->n_krcv_queues * dd->rcv_entries.ngroups) +
kctxt_ngroups);
@@ -327,7 +328,8 @@ struct hfi1_ctxtdata *hfi1_create_ctxtdata(struct hfi1_pportdata *ppd, u32 ctxt,
}
rcd->egrbufs.rcvtid_size = HFI1_MAX_EAGER_BUFFER_SIZE;
- if (ctxt < dd->first_user_ctxt) { /* N/A for PSM contexts */
+ /* Applicable only for statically created kernel contexts */
+ if (ctxt < dd->first_dyn_alloc_ctxt) {
rcd->opstats = kzalloc(sizeof(*rcd->opstats),
GFP_KERNEL);
if (!rcd->opstats)
@@ -591,7 +593,7 @@ static void enable_chip(struct hfi1_devdata *dd)
* Enable kernel ctxts' receive and receive interrupt.
* Other ctxts done as user opens and initializes them.
*/
- for (i = 0; i < dd->first_user_ctxt; ++i) {
+ for (i = 0; i < dd->first_dyn_alloc_ctxt; ++i) {
rcvmask = HFI1_RCVCTRL_CTXT_ENB | HFI1_RCVCTRL_INTRAVAIL_ENB;
rcvmask |= HFI1_CAP_KGET_MASK(dd->rcd[i]->flags, DMA_RTAIL) ?
HFI1_RCVCTRL_TAILUPD_ENB : HFI1_RCVCTRL_TAILUPD_DIS;
@@ -685,6 +687,7 @@ int hfi1_init(struct hfi1_devdata *dd, int reinit)
dd->process_pio_send = hfi1_verbs_send_pio;
dd->process_dma_send = hfi1_verbs_send_dma;
dd->pio_inline_send = pio_copy;
+ dd->process_vnic_dma_send = hfi1_vnic_send_dma;
if (is_ax(dd)) {
atomic_set(&dd->drop_packet, DROP_PACKET_ON);
@@ -720,7 +723,7 @@ int hfi1_init(struct hfi1_devdata *dd, int reinit)
}
/* dd->rcd can be NULL if early initialization failed */
- for (i = 0; dd->rcd && i < dd->first_user_ctxt; ++i) {
+ for (i = 0; dd->rcd && i < dd->first_dyn_alloc_ctxt; ++i) {
/*
* Set up the (kernel) rcvhdr queue and egr TIDs. If doing
* re-init, the simplest way to handle this is to free
@@ -1489,6 +1492,9 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
/* do the generic initialization */
initfail = hfi1_init(dd, 0);
+ /* setup vnic */
+ hfi1_vnic_setup(dd);
+
ret = hfi1_register_ib_device(dd);
/*
@@ -1522,6 +1528,7 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
hfi1_device_remove(dd);
if (!ret)
hfi1_unregister_ib_device(dd);
+ hfi1_vnic_cleanup(dd);
postinit_cleanup(dd);
if (initfail)
ret = initfail;
@@ -1547,6 +1554,9 @@ static void remove_one(struct pci_dev *pdev)
/* unregister from IB core */
hfi1_unregister_ib_device(dd);
+ /* cleanup vnic */
+ hfi1_vnic_cleanup(dd);
+
/*
* Disable the IB link, disable interrupts on the device,
* clear dma engines, etc.
@@ -1588,8 +1598,11 @@ int hfi1_create_rcvhdrq(struct hfi1_devdata *dd, struct hfi1_ctxtdata *rcd)
amt = PAGE_ALIGN(rcd->rcvhdrq_cnt * rcd->rcvhdrqentsize *
sizeof(u32));
- gfp_flags = (rcd->ctxt >= dd->first_user_ctxt) ?
- GFP_USER : GFP_KERNEL;
+ if ((rcd->ctxt < dd->first_dyn_alloc_ctxt) ||
+ (rcd->sc && (rcd->sc->type == SC_KERNEL)))
+ gfp_flags = GFP_KERNEL;
+ else
+ gfp_flags = GFP_USER;
rcd->rcvhdrq = dma_zalloc_coherent(
&dd->pcidev->dev, amt, &rcd->rcvhdrq_dma,
gfp_flags | __GFP_COMP);
diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c
index ed8ae22..cca61e4 100644
--- a/drivers/infiniband/hw/hfi1/mad.c
+++ b/drivers/infiniband/hw/hfi1/mad.c
@@ -53,6 +53,7 @@
#include "mad.h"
#include "trace.h"
#include "qp.h"
+#include "vnic.h"
/* the reset value from the FM is supposed to be 0xffff, handle both */
#define OPA_LINK_WIDTH_RESET_OLD 0x0fff
@@ -650,9 +651,11 @@ static int __subn_get_opa_portinfo(struct opa_smp *smp, u32 am, u8 *data,
OPA_PI_MASK_PORT_ACTIVE_OPTOMIZE : 0);
pi->port_packet_format.supported =
- cpu_to_be16(OPA_PORT_PACKET_FORMAT_9B);
+ cpu_to_be16(OPA_PORT_PACKET_FORMAT_9B |
+ OPA_PORT_PACKET_FORMAT_16B);
pi->port_packet_format.enabled =
- cpu_to_be16(OPA_PORT_PACKET_FORMAT_9B);
+ cpu_to_be16(OPA_PORT_PACKET_FORMAT_9B |
+ OPA_PORT_PACKET_FORMAT_16B);
/* flit_control.interleave is (OPA V1, version .76):
* bits use
@@ -678,6 +681,7 @@ static int __subn_get_opa_portinfo(struct opa_smp *smp, u32 am, u8 *data,
pi->resptimevalue = 3;
pi->local_port_num = port;
+ pi->num_vesw_port_supported = HFI_MAX_NUM_VNICS;
/* buffer info for FM */
pi->overall_buffer_space = cpu_to_be16(dd->link_credits);
diff --git a/drivers/infiniband/hw/hfi1/pio.c b/drivers/infiniband/hw/hfi1/pio.c
index 64c9eeb..3c5466e 100644
--- a/drivers/infiniband/hw/hfi1/pio.c
+++ b/drivers/infiniband/hw/hfi1/pio.c
@@ -710,6 +710,7 @@ struct send_context *sc_alloc(struct hfi1_devdata *dd, int type,
{
struct send_context_info *sci;
struct send_context *sc = NULL;
+ int req_type = type;
dma_addr_t dma;
unsigned long flags;
u64 reg;
@@ -736,6 +737,13 @@ struct send_context *sc_alloc(struct hfi1_devdata *dd, int type,
return NULL;
}
+ /*
+ * VNIC contexts are dynamically allocated.
+ * Hence, pick a user context for VNIC.
+ */
+ if (type == SC_VNIC)
+ type = SC_USER;
+
spin_lock_irqsave(&dd->sc_lock, flags);
ret = sc_hw_alloc(dd, type, &sw_index, &hw_context);
if (ret) {
@@ -745,6 +753,15 @@ struct send_context *sc_alloc(struct hfi1_devdata *dd, int type,
return NULL;
}
+ /*
+ * VNIC contexts are used by kernel driver.
+ * Hence, mark them as kernel contexts.
+ */
+ if (req_type == SC_VNIC) {
+ dd->send_contexts[sw_index].type = SC_KERNEL;
+ type = SC_KERNEL;
+ }
+
sci = &dd->send_contexts[sw_index];
sci->sc = sc;
diff --git a/drivers/infiniband/hw/hfi1/pio.h b/drivers/infiniband/hw/hfi1/pio.h
index 867e5ff..22e19d5 100644
--- a/drivers/infiniband/hw/hfi1/pio.h
+++ b/drivers/infiniband/hw/hfi1/pio.h
@@ -54,6 +54,12 @@
#define SC_USER 3 /* must be the last one: it may take all left */
#define SC_MAX 4 /* count of send context types */
+/*
+ * SC_VNIC types are allocated (dynamically) from the user context pool,
+ * (SC_USER) and used by kernel driver as kernel contexts (SC_KERNEL).
+ */
+#define SC_VNIC SC_MAX
+
/* invalid send context index */
#define INVALID_SCI 0xff
diff --git a/drivers/infiniband/hw/hfi1/sysfs.c b/drivers/infiniband/hw/hfi1/sysfs.c
index 06bff50..cefc916 100644
--- a/drivers/infiniband/hw/hfi1/sysfs.c
+++ b/drivers/infiniband/hw/hfi1/sysfs.c
@@ -543,7 +543,7 @@ static ssize_t show_nctxts(struct device *device,
* give a more accurate picture of total contexts available.
*/
return scnprintf(buf, PAGE_SIZE, "%u\n",
- min(dd->num_rcv_contexts - dd->first_user_ctxt,
+ min(dd->num_rcv_contexts - dd->first_dyn_alloc_ctxt,
(u32)dd->sc_sizes[SC_USER].count));
}
diff --git a/drivers/infiniband/hw/hfi1/user_exp_rcv.c b/drivers/infiniband/hw/hfi1/user_exp_rcv.c
index 64d2652..fdcd686 100644
--- a/drivers/infiniband/hw/hfi1/user_exp_rcv.c
+++ b/drivers/infiniband/hw/hfi1/user_exp_rcv.c
@@ -612,7 +612,7 @@ int hfi1_user_exp_rcv_invalid(struct file *fp, struct hfi1_tid_info *tinfo)
struct hfi1_filedata *fd = fp->private_data;
struct hfi1_ctxtdata *uctxt = fd->uctxt;
unsigned long *ev = uctxt->dd->events +
- (((uctxt->ctxt - uctxt->dd->first_user_ctxt) *
+ (((uctxt->ctxt - uctxt->dd->first_dyn_alloc_ctxt) *
HFI1_MAX_SHARED_CTXTS) + fd->subctxt);
u32 *array;
int ret = 0;
@@ -1016,8 +1016,8 @@ static int tid_rb_invalidate(void *arg, struct mmu_rb_node *mnode)
* process in question.
*/
ev = uctxt->dd->events +
- (((uctxt->ctxt - uctxt->dd->first_user_ctxt) *
- HFI1_MAX_SHARED_CTXTS) + fdata->subctxt);
+ (((uctxt->ctxt - uctxt->dd->first_dyn_alloc_ctxt) *
+ HFI1_MAX_SHARED_CTXTS) + fdata->subctxt);
set_bit(_HFI1_EVENT_TID_MMU_NOTIFY_BIT, ev);
}
fdata->invalid_tid_idx++;
diff --git a/drivers/infiniband/hw/hfi1/user_pages.c b/drivers/infiniband/hw/hfi1/user_pages.c
index 20f4ddc..7238a34 100644
--- a/drivers/infiniband/hw/hfi1/user_pages.c
+++ b/drivers/infiniband/hw/hfi1/user_pages.c
@@ -73,7 +73,8 @@ bool hfi1_can_pin_pages(struct hfi1_devdata *dd, struct mm_struct *mm,
{
unsigned long ulimit = rlimit(RLIMIT_MEMLOCK), pinned, cache_limit,
size = (cache_size * (1UL << 20)); /* convert to bytes */
- unsigned usr_ctxts = dd->num_rcv_contexts - dd->first_user_ctxt;
+ unsigned int usr_ctxts =
+ dd->num_rcv_contexts - dd->first_dyn_alloc_ctxt;
bool can_lock = capable(CAP_IPC_LOCK);
/*
diff --git a/drivers/infiniband/hw/hfi1/verbs.c b/drivers/infiniband/hw/hfi1/verbs.c
index 9355c4c..32485bc 100644
--- a/drivers/infiniband/hw/hfi1/verbs.c
+++ b/drivers/infiniband/hw/hfi1/verbs.c
@@ -60,6 +60,7 @@
#include "trace.h"
#include "qp.h"
#include "verbs_txreq.h"
+#include "vnic.h"
static unsigned int hfi1_lkey_table_size = 16;
module_param_named(lkey_table_size, hfi1_lkey_table_size, uint,
@@ -131,6 +132,11 @@
MODULE_PARM_DESC(sge_copy_mode,
"Verbs copy mode: 0 use memcpy, 1 use cacheless copy, 2 adapt based on WSS");
+static struct hfi_vnic_ctrl_ops hfi1_vnic_ctrl_ops = {
+ .add_vport = hfi1_vnic_add_vport,
+ .rem_vport = hfi1_vnic_rem_vport
+};
+
static void verbs_sdma_complete(
struct sdma_txreq *cookie,
int status);
@@ -1870,6 +1876,7 @@ int hfi1_register_ib_device(struct hfi1_devdata *dd)
i,
ppd->pkeys);
+ dd->verbs_dev.hfidev.vnic_ctrl_ops = hfi1_vnic_ctrl_ops;
ret = rvt_register_device(rdi);
if (ret)
goto err_verbs_txreq;
diff --git a/drivers/infiniband/hw/hfi1/vnic.h b/drivers/infiniband/hw/hfi1/vnic.h
new file mode 100644
index 0000000..047845e
--- /dev/null
+++ b/drivers/infiniband/hw/hfi1/vnic.h
@@ -0,0 +1,145 @@
+#ifndef _HFI1_VNIC_H
+#define _HFI1_VNIC_H
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+#include <rdma/opa_hfi.h>
+#include "hfi.h"
+
+#define HFI1_VNIC_ICRC_LEN 4
+#define HFI1_VNIC_TAIL_LEN 1
+#define HFI1_VNIC_ICRC_TAIL_LEN (HFI1_VNIC_ICRC_LEN + HFI1_VNIC_TAIL_LEN)
+
+#define HFI1_VNIC_MAX_TXQ 16
+#define HFI1_VNIC_MAX_PAD 12
+
+/* L2 header definitions */
+#define HFI1_L2_TYPE_OFFSET 0x7
+#define HFI1_L2_TYPE_SHFT 0x5
+#define HFI1_L2_TYPE_MASK 0x3
+#define HFI1_L2_TYPE_HDR_16B 0x2
+
+#define HFI1_GET_L2_TYPE(hdr) \
+ ((*((u8 *)(hdr) + HFI1_L2_TYPE_OFFSET) >> HFI1_L2_TYPE_SHFT) & \
+ HFI1_L2_TYPE_MASK)
+
+/* L4 type definitions */
+#define HFI1_L4_TYPE_OFFSET 8
+
+#define HFI1_GET_L4_TYPE(data) \
+ (*((u8 *)(data) + HFI1_L4_TYPE_OFFSET))
+
+#define HFI1_VNIC_L4_ETHR 0x78
+
+/* L4 header definitions */
+#define HFI1_VNIC_L4_HDR_OFFSET 18
+
+#define HFI1_VNIC_GET_L4_HDR(data) \
+ (*((u16 *)((u8 *)(data) + HFI1_VNIC_L4_HDR_OFFSET)))
+
+#define HFI1_VNIC_GET_VESWID(data) \
+ (HFI1_VNIC_GET_L4_HDR(data) & 0xFF)
+
+/* Service class */
+#define HFI1_VNIC_SC_OFFSET_LOW 6
+#define HFI1_VNIC_SC_OFFSET_HI 7
+#define HFI1_VNIC_SC_SHIFT 4
+
+/**
+ * struct hfi1_vnic_notifier - VNIC notifer structure
+ * @cb - vnic callback function
+ */
+struct hfi1_vnic_notifier {
+ hfi_vnic_evt_cb_fn cb;
+};
+
+/**
+ * struct hfi1_vnic_vport_info - HFI1 VNIC virtual port information
+ * @dd: device data pointer
+ * @notifier: vnic notifier
+ * @event_flags: event notification flags
+ * @vport: vnic port pointer
+ * @skbq: Array of queues for received socket buffers
+ */
+struct hfi1_vnic_vport_info {
+ struct hfi1_devdata *dd;
+
+ struct hfi1_vnic_notifier __rcu *notifier;
+ DECLARE_BITMAP(event_flags, HFI_VNIC_NUM_EVTS);
+ struct hfi_vnic_port *vport;
+
+ struct sk_buff_head skbq[HFI1_NUM_VNIC_CTXT];
+};
+
+static inline struct hfi1_devdata *vnic_dev2dd(struct hfi_vnic_port *vport)
+{
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+
+ return vinfo->dd;
+}
+
+/* setup the last plen bypes of pad */
+static inline void hfi1_vnic_update_pad(unsigned char *pad, u8 plen)
+{
+ pad[HFI1_VNIC_MAX_PAD - 1] = plen - HFI1_VNIC_ICRC_TAIL_LEN;
+}
+
+/* vnic hfi1 internal functions */
+void hfi1_vnic_setup(struct hfi1_devdata *dd);
+void hfi1_vnic_cleanup(struct hfi1_devdata *dd);
+
+void hfi1_vnic_bypass_rcv(struct hfi1_packet *packet);
+
+/* vnic port operations */
+struct hfi_vnic_port *hfi1_vnic_add_vport(struct ib_device *device,
+ u8 port_num, u8 vport_num);
+void hfi1_vnic_rem_vport(struct hfi_vnic_port *vport);
+int hfi1_vnic_send_dma(struct hfi1_devdata *dd, u8 q_idx,
+ struct hfi1_vnic_vport_info *vinfo,
+ struct sk_buff *skb, u64 pbc, u8 plen);
+
+#endif /* _HFI1_VNIC_H */
diff --git a/drivers/infiniband/hw/hfi1/vnic_main.c b/drivers/infiniband/hw/hfi1/vnic_main.c
new file mode 100644
index 0000000..1e237f3
--- /dev/null
+++ b/drivers/infiniband/hw/hfi1/vnic_main.c
@@ -0,0 +1,614 @@
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains HFI1 support for VNIC functionality
+ */
+
+#include <linux/io.h>
+
+#include "vnic.h"
+
+#define HFI1_VNIC_RCV_Q_SIZE 1024
+
+static DEFINE_SPINLOCK(vport_cntr_lock);
+
+static inline u8 hfi1_vnic_get_sc5(u8 *hdr)
+{
+ return (((*(hdr + HFI1_VNIC_SC_OFFSET_LOW)) >> HFI1_VNIC_SC_SHIFT) |
+ (((*(hdr + HFI1_VNIC_SC_OFFSET_HI)) & 0x1) <<
+ HFI1_VNIC_SC_SHIFT));
+}
+
+static int setup_vnic_ctxt(struct hfi1_devdata *dd, struct hfi1_ctxtdata *uctxt)
+{
+ unsigned int rcvctrl_ops = 0;
+ int ret;
+
+ ret = hfi1_init_ctxt(uctxt->sc);
+ if (ret)
+ goto done;
+
+ uctxt->do_interrupt = &handle_receive_interrupt;
+
+ /* Now allocate the RcvHdr queue and eager buffers. */
+ ret = hfi1_create_rcvhdrq(dd, uctxt);
+ if (ret)
+ goto done;
+
+ ret = hfi1_setup_eagerbufs(uctxt);
+ if (ret)
+ goto done;
+
+ set_bit(HFI1_CTXT_SETUP_DONE, &uctxt->event_flags);
+
+ if (uctxt->rcvhdrtail_kvaddr)
+ clear_rcvhdrtail(uctxt);
+
+ rcvctrl_ops = HFI1_RCVCTRL_CTXT_ENB;
+ rcvctrl_ops |= HFI1_RCVCTRL_INTRAVAIL_ENB;
+
+ if (!HFI1_CAP_KGET_MASK(uctxt->flags, MULTI_PKT_EGR))
+ rcvctrl_ops |= HFI1_RCVCTRL_ONE_PKT_EGR_ENB;
+ if (HFI1_CAP_KGET_MASK(uctxt->flags, NODROP_EGR_FULL))
+ rcvctrl_ops |= HFI1_RCVCTRL_NO_EGR_DROP_ENB;
+ if (HFI1_CAP_KGET_MASK(uctxt->flags, NODROP_RHQ_FULL))
+ rcvctrl_ops |= HFI1_RCVCTRL_NO_RHQ_DROP_ENB;
+ if (HFI1_CAP_KGET_MASK(uctxt->flags, DMA_RTAIL))
+ rcvctrl_ops |= HFI1_RCVCTRL_TAILUPD_ENB;
+
+ hfi1_rcvctrl(uctxt->dd, rcvctrl_ops, uctxt->ctxt);
+
+ uctxt->is_vnic = true;
+done:
+ return ret;
+}
+
+static int allocate_vnic_ctxt(struct hfi1_devdata *dd,
+ struct hfi1_ctxtdata **vnic_ctxt)
+{
+ struct hfi1_ctxtdata *uctxt;
+ unsigned int ctxt;
+ int ret;
+
+ if (dd->flags & HFI1_FROZEN)
+ return -EIO;
+
+ for (ctxt = dd->first_dyn_alloc_ctxt;
+ ctxt < dd->num_rcv_contexts; ctxt++)
+ if (!dd->rcd[ctxt])
+ break;
+
+ if (ctxt == dd->num_rcv_contexts)
+ return -EBUSY;
+
+ uctxt = hfi1_create_ctxtdata(dd->pport, ctxt, dd->node);
+ if (!uctxt) {
+ dd_dev_err(dd, "Unable to create ctxtdata, failing open\n");
+ return -ENOMEM;
+ }
+
+ uctxt->flags = HFI1_CAP_KGET(MULTI_PKT_EGR) |
+ HFI1_CAP_KGET(NODROP_RHQ_FULL) |
+ HFI1_CAP_KGET(NODROP_EGR_FULL) |
+ HFI1_CAP_KGET(DMA_RTAIL);
+ uctxt->seq_cnt = 1;
+
+ /* Allocate and enable a PIO send context */
+ uctxt->sc = sc_alloc(dd, SC_VNIC, uctxt->rcvhdrqentsize,
+ uctxt->numa_id);
+
+ ret = uctxt->sc ? 0 : -ENOMEM;
+ if (ret)
+ goto bail;
+
+ dd_dev_dbg(dd, "allocated vnic send context %u(%u)\n",
+ uctxt->sc->sw_index, uctxt->sc->hw_context);
+ ret = sc_enable(uctxt->sc);
+ if (ret)
+ goto bail;
+
+ if (dd->num_msix_entries)
+ hfi1_set_vnic_msix_info(uctxt);
+
+ hfi1_stats.sps_ctxts++;
+ dd_dev_dbg(dd, "created vnic context %d\n", uctxt->ctxt);
+ *vnic_ctxt = uctxt;
+
+ return ret;
+bail:
+ /*
+ * hfi1_free_ctxtdata() also releases send_context
+ * structure if uctxt->sc is not null
+ */
+ dd->rcd[uctxt->ctxt] = NULL;
+ hfi1_free_ctxtdata(dd, uctxt);
+ dd_dev_dbg(dd, "vnic allocation failed. rc %d\n", ret);
+ return ret;
+}
+
+static void deallocate_vnic_ctxt(struct hfi1_devdata *dd,
+ struct hfi1_ctxtdata *uctxt)
+{
+ unsigned long flags;
+
+ dd_dev_dbg(dd, "closing vnic context %d\n", uctxt->ctxt);
+ flush_wc();
+
+ if (dd->num_msix_entries)
+ hfi1_reset_vnic_msix_info(uctxt);
+
+ spin_lock_irqsave(&dd->uctxt_lock, flags);
+ /*
+ * Disable receive context and interrupt available, reset all
+ * RcvCtxtCtrl bits to default values.
+ */
+ hfi1_rcvctrl(dd, HFI1_RCVCTRL_CTXT_DIS |
+ HFI1_RCVCTRL_TIDFLOW_DIS |
+ HFI1_RCVCTRL_INTRAVAIL_DIS |
+ HFI1_RCVCTRL_ONE_PKT_EGR_DIS |
+ HFI1_RCVCTRL_NO_RHQ_DROP_DIS |
+ HFI1_RCVCTRL_NO_EGR_DROP_DIS, uctxt->ctxt);
+ /*
+ * VNIC contexts are allocated from user context pool.
+ * Release them back to user context pool.
+ *
+ * Reset context integrity checks to default.
+ * (writes to CSRs probably belong in chip.c)
+ */
+ write_kctxt_csr(dd, uctxt->sc->hw_context, SEND_CTXT_CHECK_ENABLE,
+ hfi1_pkt_default_send_ctxt_mask(dd, SC_USER));
+ sc_disable(uctxt->sc);
+
+ dd->send_contexts[uctxt->sc->sw_index].type = SC_USER;
+ spin_unlock_irqrestore(&dd->uctxt_lock, flags);
+
+ dd->rcd[uctxt->ctxt] = NULL;
+ uctxt->event_flags = 0;
+
+ hfi1_clear_tids(uctxt);
+ hfi1_clear_ctxt_pkey(dd, uctxt->ctxt);
+
+ hfi1_stats.sps_ctxts--;
+ hfi1_free_ctxtdata(dd, uctxt);
+}
+
+void hfi1_vnic_setup(struct hfi1_devdata *dd)
+{
+ idr_init(&dd->vnic.vesw_idr);
+}
+
+void hfi1_vnic_cleanup(struct hfi1_devdata *dd)
+{
+ idr_destroy(&dd->vnic.vesw_idr);
+}
+
+static u64 create_bypass_pbc(u32 vl, u32 dw_len)
+{
+ u64 pbc;
+
+ pbc = ((u64)PBC_IHCRC_NONE << PBC_INSERT_HCRC_SHIFT)
+ | PBC_INSERT_BYPASS_ICRC | PBC_CREDIT_RETURN
+ | PBC_PACKET_BYPASS
+ | ((vl & PBC_VL_MASK) << PBC_VL_SHIFT)
+ | (dw_len & PBC_LENGTH_DWS_MASK) << PBC_LENGTH_DWS_SHIFT;
+
+ return pbc;
+}
+
+static int hfi1_vnic_put_skb(struct hfi_vnic_port *vport,
+ u8 q_idx, struct sk_buff *skb)
+{
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+ struct hfi1_devdata *dd = vinfo->dd;
+ u32 vl, pkt_len, total_len;
+ u8 sc5, pad_len;
+ int ret = 0;
+ u64 pbc;
+
+ if (q_idx >= vport->hfi_info.num_tx_q) {
+ dev_kfree_skb_any(skb);
+ return -EINVAL;
+ }
+
+ /* add tail padding (for 8 bytes size alignment) and icrc */
+ pad_len = -(skb->len + HFI1_VNIC_ICRC_TAIL_LEN) & 0x7;
+ pad_len += HFI1_VNIC_ICRC_TAIL_LEN;
+
+ /*
+ * pkt_len is how much data we have to write, includes header and data.
+ * total_len is length of the packet in Dwords plus the PBC should not
+ * include the CRC.
+ */
+ pkt_len = (skb->len + pad_len) >> 2;
+ total_len = pkt_len + 2; /* PBC + packet */
+
+ sc5 = hfi1_vnic_get_sc5(skb->data);
+ vl = sc_to_vlt(dd, sc5);
+ pbc = create_bypass_pbc(vl, total_len);
+
+ dd_dev_dbg(dd, "%d: pbc 0x%016llX len %d pad_len %d\n",
+ vport->vport_num, pbc, skb->len, pad_len);
+
+ ret = dd->process_vnic_dma_send(dd, q_idx, vinfo, skb,
+ pbc, pad_len);
+
+ if (ret) {
+ if (ret == -ENOMEM)
+ vport->hfi_stats[q_idx].tx_fifo_errors++;
+ else if (ret != -EBUSY)
+ vport->hfi_stats[q_idx].tx_logic_errors++;
+ }
+
+ return ret;
+}
+
+static u8 hfi1_vnic_select_queue(struct hfi_vnic_port *vport, u8 vl, u8 entropy)
+{
+ return 0;
+}
+
+static bool hfi1_vnic_get_write_avail(struct hfi_vnic_port *vport, u8 q_idx)
+{
+ if (q_idx >= vport->hfi_info.num_tx_q)
+ return false;
+
+ return true;
+}
+
+void hfi1_vnic_bypass_rcv(struct hfi1_packet *packet)
+{
+ struct hfi1_devdata *dd = packet->rcd->dd;
+ struct hfi1_vnic_vport_info *vinfo;
+ struct hfi_vnic_port *vport = NULL;
+ struct hfi1_vnic_notifier *notifier;
+ struct sk_buff *skb;
+ int l4_type, vesw_id = -1;
+ u8 q_idx;
+
+ rcu_read_lock();
+ l4_type = HFI1_GET_L4_TYPE(packet->ebuf);
+ if (l4_type == HFI1_VNIC_L4_ETHR) {
+ vesw_id = HFI1_VNIC_GET_VESWID(packet->ebuf);
+ vport = idr_find(&dd->vnic.vesw_idr, vesw_id);
+
+ /*
+ * In case of invalid vesw id, update the rx_bad_veswid
+ * error count of first available vport.
+ */
+ if (unlikely(!vport)) {
+ struct hfi_vnic_port *vport_tmp;
+ int id_tmp = 0;
+
+ vport_tmp = idr_get_next(&dd->vnic.vesw_idr, &id_tmp);
+ if (vport_tmp) {
+ spin_lock(&vport_cntr_lock);
+ vport_tmp->hfi_stats[0].rx_bad_veswid++;
+ spin_unlock(&vport_cntr_lock);
+ }
+ }
+ }
+
+ if (unlikely(!vport)) {
+ dd_dev_warn(dd, "vnic rcv err: l4 %d vesw id %d ctx %d\n",
+ l4_type, vesw_id, packet->rcd->ctxt);
+ goto rcv_done;
+ }
+
+ vinfo = vport->hfi_priv;
+ q_idx = packet->rcd->vnic_q_idx;
+ notifier = rcu_dereference(vinfo->notifier);
+ if (!notifier || !notifier->cb) {
+ vport->hfi_stats[q_idx].rx_logic_errors++;
+ goto rcv_done;
+ }
+
+ if (skb_queue_len(&vinfo->skbq[q_idx]) > HFI1_VNIC_RCV_Q_SIZE) {
+ vport->hfi_stats[q_idx].rx_fifo_errors++;
+ goto rcv_done;
+ }
+
+ skb = netdev_alloc_skb(vport->netdev, packet->tlen);
+ if (!skb) {
+ vport->hfi_stats[q_idx].rx_missed_errors++;
+ goto rcv_done;
+ }
+ memcpy(skb->data, packet->ebuf, packet->tlen);
+ skb_put(skb, packet->tlen);
+
+ skb_queue_tail(&vinfo->skbq[q_idx], skb);
+ if (test_bit((HFI_VNIC_EVT_RX0 + q_idx), vinfo->event_flags))
+ notifier->cb(vport, HFI_VNIC_EVT_RX0 + q_idx);
+
+rcv_done:
+ rcu_read_unlock();
+}
+
+static u16 hfi1_vnic_get_read_avail(struct hfi_vnic_port *vport, u8 q_idx)
+{
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+
+ if (q_idx >= vport->hfi_info.num_rx_q)
+ return 0;
+
+ return skb_queue_len(&vinfo->skbq[q_idx]);
+}
+
+static struct sk_buff *hfi1_vnic_get_skb(struct hfi_vnic_port *vport, u8 q_idx)
+{
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+ unsigned char *pad_info;
+ struct sk_buff *skb;
+
+ if (q_idx >= vport->hfi_info.num_rx_q)
+ return NULL;
+
+ skb = skb_dequeue(&vinfo->skbq[q_idx]);
+ if (!skb)
+ return NULL;
+
+ /* remove tail padding and icrc */
+ pad_info = skb->data + skb->len - 1;
+ skb_trim(skb, (skb->len - HFI1_VNIC_ICRC_TAIL_LEN -
+ ((*pad_info) & 0x7)));
+
+ return skb;
+}
+
+static void hfi1_vnic_config_notify(struct hfi_vnic_port *vport,
+ u8 evt, bool enable)
+{
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+
+ if (enable)
+ set_bit(evt, vinfo->event_flags);
+ else
+ clear_bit(evt, vinfo->event_flags);
+}
+
+static int hfi1_vnic_open(struct hfi_vnic_port *vport, hfi_vnic_evt_cb_fn cb)
+{
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+ struct hfi1_devdata *dd = vinfo->dd;
+ struct hfi1_vnic_notifier *notifier;
+ int i, rc;
+
+ if (!cb)
+ return -EINVAL;
+
+ notifier = kmalloc(sizeof(*notifier), GFP_KERNEL);
+ if (!notifier)
+ return -ENOMEM;
+
+ notifier->cb = cb;
+
+ /* ensure virtual eth switch id is valid */
+ if (!vport->vesw_id) {
+ rc = -EINVAL;
+ goto open_fail;
+ }
+
+ rc = idr_alloc(&dd->vnic.vesw_idr, vport, vport->vesw_id,
+ vport->vesw_id + 1, GFP_NOWAIT);
+ if (rc < 0)
+ goto open_fail;
+
+ for (i = 0; i < HFI1_NUM_VNIC_CTXT; i++)
+ skb_queue_head_init(&vinfo->skbq[i]);
+
+ /* Enable all events */
+ for (i = 0; i < HFI_VNIC_NUM_EVTS; i++)
+ set_bit(i, vinfo->event_flags);
+
+ rcu_assign_pointer(vinfo->notifier, notifier);
+ synchronize_rcu();
+ return 0;
+
+open_fail:
+ kfree(notifier);
+ return rc;
+}
+
+static void hfi1_vnic_close(struct hfi_vnic_port *vport)
+{
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+ struct hfi1_devdata *dd = vinfo->dd;
+ struct hfi1_vnic_notifier *notifier;
+ u8 i;
+
+ idr_remove(&dd->vnic.vesw_idr, vport->vesw_id);
+ notifier = rcu_access_pointer(vinfo->notifier);
+ rcu_assign_pointer(vinfo->notifier, NULL);
+ synchronize_rcu();
+ kfree(notifier);
+
+ /* remove unread skbs */
+ for (i = 0; i < HFI1_NUM_VNIC_CTXT; i++)
+ skb_queue_purge(&vinfo->skbq[i]);
+}
+
+static int hfi1_vnic_allot_ctxt(struct hfi1_devdata *dd,
+ struct hfi1_ctxtdata **vnic_ctxt)
+{
+ int rc;
+
+ rc = allocate_vnic_ctxt(dd, vnic_ctxt);
+ if (rc) {
+ dd_dev_err(dd, "vnic ctxt alloc failed %d\n", rc);
+ return rc;
+ }
+
+ rc = setup_vnic_ctxt(dd, *vnic_ctxt);
+ if (rc) {
+ dd_dev_err(dd, "vnic ctxt setup failed %d\n", rc);
+ deallocate_vnic_ctxt(dd, *vnic_ctxt);
+ *vnic_ctxt = NULL;
+ }
+
+ return rc;
+}
+
+static int hfi1_vnic_init(struct hfi_vnic_port *vport)
+{
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+ struct hfi1_devdata *dd = vinfo->dd;
+ int i, rc = 0;
+
+ mutex_lock(&hfi1_mutex);
+ for (i = dd->vnic.num_ctxt; i < vport->hfi_info.num_rx_q; i++) {
+ rc = hfi1_vnic_allot_ctxt(dd, &dd->vnic.ctxt[i]);
+ if (rc)
+ break;
+ dd->vnic.ctxt[i]->vnic_q_idx = i;
+ }
+
+ if (i < vport->hfi_info.num_rx_q) {
+ /*
+ * If required amount of contexts is not
+ * allocated successfully then remaining contexts
+ * are released.
+ */
+ while (i-- > dd->vnic.num_ctxt) {
+ deallocate_vnic_ctxt(dd, dd->vnic.ctxt[i]);
+ dd->vnic.ctxt[i] = NULL;
+ }
+ goto alloc_fail;
+ }
+
+ if (dd->vnic.num_ctxt != i) {
+ dd->vnic.num_ctxt = i;
+ hfi1_init_vnic_rsm(dd);
+ }
+
+ dd->vnic.num_vports++;
+ vinfo->vport = vport;
+alloc_fail:
+ mutex_unlock(&hfi1_mutex);
+ return rc;
+}
+
+static void hfi1_vnic_deinit(struct hfi_vnic_port *vport)
+{
+ struct hfi1_devdata *dd = vnic_dev2dd(vport);
+ int i;
+
+ mutex_lock(&hfi1_mutex);
+ if (--dd->vnic.num_vports == 0) {
+ for (i = 0; i < dd->vnic.num_ctxt; i++) {
+ deallocate_vnic_ctxt(dd, dd->vnic.ctxt[i]);
+ dd->vnic.ctxt[i] = NULL;
+ }
+ hfi1_deinit_vnic_rsm(dd);
+ dd->vnic.num_ctxt = 0;
+ }
+ mutex_unlock(&hfi1_mutex);
+}
+
+/* vnic operations */
+static struct hfi_vnic_ops hfi1_vnic_ops = {
+ .open = hfi1_vnic_open,
+ .close = hfi1_vnic_close,
+ .put_skb = hfi1_vnic_put_skb,
+ .get_skb = hfi1_vnic_get_skb,
+ .get_read_avail = hfi1_vnic_get_read_avail,
+ .get_write_avail = hfi1_vnic_get_write_avail,
+ .select_queue = hfi1_vnic_select_queue,
+ .config_notify = hfi1_vnic_config_notify
+};
+
+/* hfi1_vnic_add_vport - Allocate and initialize a vnic port */
+struct hfi_vnic_port *hfi1_vnic_add_vport(struct ib_device *device,
+ u8 port_num, u8 vport_num)
+{
+ struct hfi1_devdata *dd = dd_from_ibdev(device);
+ struct hfi1_vnic_vport_info *vinfo;
+ struct hfi_vnic_port *vport;
+ int rc;
+
+ if (!port_num || (port_num > dd->num_pports) ||
+ (vport_num == HFI_MAX_NUM_VNICS))
+ return ERR_PTR(-EINVAL);
+
+ vport = kzalloc(sizeof(*vport), GFP_KERNEL);
+ if (!vport)
+ return ERR_PTR(-ENOMEM);
+
+ vinfo = kzalloc(sizeof(*vinfo), GFP_KERNEL);
+ if (!vinfo) {
+ rc = -ENOMEM;
+ goto vinfo_fail;
+ }
+
+ vinfo->dd = dd;
+ vport->hfi_info.num_tx_q = dd->chip_sdma_engines;
+ vport->hfi_info.num_rx_q = HFI1_NUM_VNIC_CTXT;
+ vport->hfi_info.cap = HFI_VNIC_CAP_SG;
+ vport->ops = &hfi1_vnic_ops;
+ vport->hfi_priv = vinfo;
+ vport->port_num = port_num;
+ vport->vport_num = vport_num;
+
+ rc = hfi1_vnic_init(vport);
+ if (rc)
+ goto init_fail;
+
+ dd_dev_info(dd, "added vnic port %d:%d\n", port_num, vport_num);
+ return vport;
+init_fail:
+ kfree(vinfo);
+vinfo_fail:
+ kfree(vport);
+ return ERR_PTR(rc);
+}
+
+/* hfi1_vnic_rem_vport - Uninitialize and free vnic port */
+void hfi1_vnic_rem_vport(struct hfi_vnic_port *vport)
+{
+ hfi1_vnic_deinit(vport);
+ kfree(vport->hfi_priv);
+ kfree(vport);
+}
diff --git a/drivers/infiniband/hw/hfi1/vnic_sdma.c b/drivers/infiniband/hw/hfi1/vnic_sdma.c
new file mode 100644
index 0000000..66abad0
--- /dev/null
+++ b/drivers/infiniband/hw/hfi1/vnic_sdma.c
@@ -0,0 +1,60 @@
+/*
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This file is provided under a dual BSD/GPLv2 license. When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * BSD LICENSE
+ *
+ * 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.
+ *
+ */
+
+/*
+ * This file contains HFI1 support for VNIC SDMA functionality
+ */
+
+#include "sdma.h"
+#include "vnic.h"
+
+int hfi1_vnic_send_dma(struct hfi1_devdata *dd, u8 q_idx,
+ struct hfi1_vnic_vport_info *vinfo,
+ struct sk_buff *skb, u64 pbc, u8 plen)
+{
+ return 0;
+}
diff --git a/include/rdma/opa_port_info.h b/include/rdma/opa_port_info.h
index 9303e0e..84caa5b 100644
--- a/include/rdma/opa_port_info.h
+++ b/include/rdma/opa_port_info.h
@@ -410,7 +410,7 @@ struct opa_port_info {
u8 resptimevalue; /* 3 res, 5 bits */
u8 local_port_num;
- u8 reserved12;
+ u8 num_vesw_port_supported;
u8 reserved13; /* was guid_cap */
} __attribute__ ((packed));
--
1.8.3.1
^ permalink raw reply related [flat|nested] 41+ messages in thread
* [RFC v2 10/10] IB/hfi1: VNIC SDMA support
2016-12-15 7:59 [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
` (4 preceding siblings ...)
2016-12-15 7:59 ` [RFC v2 09/10] IB/hfi1: Virtual Network Interface Controller (VNIC) support Vishwanathapura, Niranjana
@ 2016-12-15 7:59 ` Vishwanathapura, Niranjana
2016-12-15 16:56 ` [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Jason Gunthorpe
6 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 7:59 UTC (permalink / raw)
To: dledford
Cc: linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Niranjana Vishwanathapura
HFI1 VNIC SDMA support enables transmission of VNIC packets over SDMA.
Map VNIC queues to SDMA engines and support halting and wakeup of the
VNIC queues.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
---
drivers/infiniband/hw/hfi1/hfi.h | 1 +
drivers/infiniband/hw/hfi1/vnic.h | 30 +++-
drivers/infiniband/hw/hfi1/vnic_main.c | 21 ++-
drivers/infiniband/hw/hfi1/vnic_sdma.c | 260 +++++++++++++++++++++++++++++++++
4 files changed, 309 insertions(+), 3 deletions(-)
diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h
index 78d1726..8d5949f 100644
--- a/drivers/infiniband/hw/hfi1/hfi.h
+++ b/drivers/infiniband/hw/hfi1/hfi.h
@@ -855,6 +855,7 @@ struct hfi1_asic_data {
/* Virtual NIC information */
struct hfi1_vnic_data {
struct hfi1_ctxtdata *ctxt[HFI1_NUM_VNIC_CTXT];
+ struct kmem_cache *txreq_cache;
u8 num_vports;
struct idr vesw_idr;
u8 rmt_start;
diff --git a/drivers/infiniband/hw/hfi1/vnic.h b/drivers/infiniband/hw/hfi1/vnic.h
index 047845e..2d4eb8f 100644
--- a/drivers/infiniband/hw/hfi1/vnic.h
+++ b/drivers/infiniband/hw/hfi1/vnic.h
@@ -49,6 +49,7 @@
#include <rdma/opa_hfi.h>
#include "hfi.h"
+#include "sdma.h"
#define HFI1_VNIC_ICRC_LEN 4
#define HFI1_VNIC_TAIL_LEN 1
@@ -90,6 +91,26 @@
#define HFI1_VNIC_SC_SHIFT 4
/**
+ * struct hfi1_vnic_sdma - VNIC per Tx ring SDMA information
+ * @dd - device data pointer
+ * @sde - sdma engine
+ * @vinfo - vnic info pointer
+ * @wait - iowait structure
+ * @stx - sdma tx request
+ * @state - vnic Tx ring SDMA state
+ * @q_idx - vnic Tx queue index
+ */
+struct hfi1_vnic_sdma {
+ struct hfi1_devdata *dd;
+ struct sdma_engine *sde;
+ struct hfi1_vnic_vport_info *vinfo;
+ struct iowait wait;
+ struct sdma_txreq stx;
+ unsigned int state;
+ u8 q_idx;
+};
+
+/**
* struct hfi1_vnic_notifier - VNIC notifer structure
* @cb - vnic callback function
*/
@@ -104,6 +125,7 @@ struct hfi1_vnic_notifier {
* @event_flags: event notification flags
* @vport: vnic port pointer
* @skbq: Array of queues for received socket buffers
+ * @sdma: VNIC SDMA structure per TXQ
*/
struct hfi1_vnic_vport_info {
struct hfi1_devdata *dd;
@@ -112,7 +134,8 @@ struct hfi1_vnic_vport_info {
DECLARE_BITMAP(event_flags, HFI_VNIC_NUM_EVTS);
struct hfi_vnic_port *vport;
- struct sk_buff_head skbq[HFI1_NUM_VNIC_CTXT];
+ struct sk_buff_head skbq[HFI1_NUM_VNIC_CTXT];
+ struct hfi1_vnic_sdma sdma[HFI1_VNIC_MAX_TXQ];
};
static inline struct hfi1_devdata *vnic_dev2dd(struct hfi_vnic_port *vport)
@@ -131,8 +154,13 @@ static inline void hfi1_vnic_update_pad(unsigned char *pad, u8 plen)
/* vnic hfi1 internal functions */
void hfi1_vnic_setup(struct hfi1_devdata *dd);
void hfi1_vnic_cleanup(struct hfi1_devdata *dd);
+int hfi1_vnic_txreq_init(struct hfi1_devdata *dd);
+void hfi1_vnic_txreq_deinit(struct hfi1_devdata *dd);
void hfi1_vnic_bypass_rcv(struct hfi1_packet *packet);
+void hfi1_vnic_sdma_init(struct hfi1_vnic_vport_info *vinfo);
+bool hfi1_vnic_sdma_write_avail(struct hfi1_vnic_vport_info *vinfo,
+ u8 q_idx);
/* vnic port operations */
struct hfi_vnic_port *hfi1_vnic_add_vport(struct ib_device *device,
diff --git a/drivers/infiniband/hw/hfi1/vnic_main.c b/drivers/infiniband/hw/hfi1/vnic_main.c
index 1e237f3..19843a4 100644
--- a/drivers/infiniband/hw/hfi1/vnic_main.c
+++ b/drivers/infiniband/hw/hfi1/vnic_main.c
@@ -289,15 +289,21 @@ static int hfi1_vnic_put_skb(struct hfi_vnic_port *vport,
static u8 hfi1_vnic_select_queue(struct hfi_vnic_port *vport, u8 vl, u8 entropy)
{
- return 0;
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+ struct sdma_engine *sde;
+
+ sde = sdma_select_engine_vl(vinfo->dd, entropy, vl);
+ return sde->this_idx;
}
static bool hfi1_vnic_get_write_avail(struct hfi_vnic_port *vport, u8 q_idx)
{
+ struct hfi1_vnic_vport_info *vinfo = vport->hfi_priv;
+
if (q_idx >= vport->hfi_info.num_tx_q)
return false;
- return true;
+ return hfi1_vnic_sdma_write_avail(vinfo, q_idx);
}
void hfi1_vnic_bypass_rcv(struct hfi1_packet *packet)
@@ -499,6 +505,12 @@ static int hfi1_vnic_init(struct hfi_vnic_port *vport)
int i, rc = 0;
mutex_lock(&hfi1_mutex);
+ if (!dd->vnic.num_vports) {
+ rc = hfi1_vnic_txreq_init(dd);
+ if (rc)
+ goto txreq_fail;
+ }
+
for (i = dd->vnic.num_ctxt; i < vport->hfi_info.num_rx_q; i++) {
rc = hfi1_vnic_allot_ctxt(dd, &dd->vnic.ctxt[i]);
if (rc)
@@ -526,7 +538,11 @@ static int hfi1_vnic_init(struct hfi_vnic_port *vport)
dd->vnic.num_vports++;
vinfo->vport = vport;
+ hfi1_vnic_sdma_init(vinfo);
alloc_fail:
+ if (!dd->vnic.num_vports)
+ hfi1_vnic_txreq_deinit(dd);
+txreq_fail:
mutex_unlock(&hfi1_mutex);
return rc;
}
@@ -544,6 +560,7 @@ static void hfi1_vnic_deinit(struct hfi_vnic_port *vport)
}
hfi1_deinit_vnic_rsm(dd);
dd->vnic.num_ctxt = 0;
+ hfi1_vnic_txreq_deinit(dd);
}
mutex_unlock(&hfi1_mutex);
}
diff --git a/drivers/infiniband/hw/hfi1/vnic_sdma.c b/drivers/infiniband/hw/hfi1/vnic_sdma.c
index 66abad0..cb1f9ce 100644
--- a/drivers/infiniband/hw/hfi1/vnic_sdma.c
+++ b/drivers/infiniband/hw/hfi1/vnic_sdma.c
@@ -52,9 +52,269 @@
#include "sdma.h"
#include "vnic.h"
+#define HFI1_VNIC_SDMA_Q_ACTIVE BIT(0)
+#define HFI1_VNIC_SDMA_Q_DEFERRED BIT(1)
+
+#define HFI1_VNIC_TXREQ_NAME_LEN 32
+#define HFI1_VNIC_SDMA_DESC_WTRMRK 64
+#define HFI1_VNIC_SDMA_RETRY_COUNT 1
+
+/*
+ * struct vnic_txreq - VNIC transmit descriptor
+ * @txreq: sdma transmit request
+ * @sdma: vnic sdma pointer
+ * @skb: skb to send
+ * @pad: pad buffer
+ * @plen: pad length
+ * @pbc_val: pbc value
+ * @retry_count: tx retry count
+ */
+struct vnic_txreq {
+ struct sdma_txreq txreq;
+ struct hfi1_vnic_sdma *sdma;
+
+ struct sk_buff *skb;
+ unsigned char pad[HFI1_VNIC_MAX_PAD];
+ u16 plen;
+ __le64 pbc_val;
+
+ u32 retry_count;
+};
+
+static void vnic_sdma_complete(struct sdma_txreq *txreq,
+ int status)
+{
+ struct vnic_txreq *tx = container_of(txreq, struct vnic_txreq, txreq);
+ struct hfi1_vnic_sdma *vnic_sdma = tx->sdma;
+
+ sdma_txclean(vnic_sdma->dd, txreq);
+ dev_kfree_skb_any(tx->skb);
+ kmem_cache_free(vnic_sdma->dd->vnic.txreq_cache, tx);
+}
+
+static noinline int build_vnic_ulp_payload(struct sdma_engine *sde,
+ struct vnic_txreq *tx)
+{
+ int i, ret = 0;
+
+ ret = sdma_txadd_kvaddr(
+ sde->dd,
+ &tx->txreq,
+ tx->skb->data,
+ skb_headlen(tx->skb));
+ if (ret)
+ goto bail_txadd;
+
+ for (i = 0; i < skb_shinfo(tx->skb)->nr_frags; i++) {
+ struct skb_frag_struct *frag = &skb_shinfo(tx->skb)->frags[i];
+
+ /* combine physically continuous fragments later? */
+ ret = sdma_txadd_page(sde->dd,
+ &tx->txreq,
+ skb_frag_page(frag),
+ frag->page_offset,
+ skb_frag_size(frag));
+ if (ret)
+ goto bail_txadd;
+ }
+
+ if (tx->plen)
+ ret = sdma_txadd_kvaddr(sde->dd, &tx->txreq,
+ tx->pad + HFI1_VNIC_MAX_PAD - tx->plen,
+ tx->plen);
+
+bail_txadd:
+ return ret;
+}
+
+static int build_vnic_tx_desc(struct sdma_engine *sde,
+ struct vnic_txreq *tx,
+ u64 pbc)
+{
+ int ret = 0;
+ u16 hdrbytes = 2 << 2; /* PBC */
+
+ ret = sdma_txinit_ahg(
+ &tx->txreq,
+ 0,
+ hdrbytes + tx->skb->len + tx->plen,
+ 0,
+ 0,
+ NULL,
+ 0,
+ vnic_sdma_complete);
+ if (ret)
+ goto bail_txadd;
+
+ /* add pbc */
+ tx->pbc_val = cpu_to_le64(pbc);
+ ret = sdma_txadd_kvaddr(
+ sde->dd,
+ &tx->txreq,
+ &tx->pbc_val,
+ hdrbytes);
+ if (ret)
+ goto bail_txadd;
+
+ /* add the ulp payload */
+ ret = build_vnic_ulp_payload(sde, tx);
+bail_txadd:
+ return ret;
+}
+
int hfi1_vnic_send_dma(struct hfi1_devdata *dd, u8 q_idx,
struct hfi1_vnic_vport_info *vinfo,
struct sk_buff *skb, u64 pbc, u8 plen)
{
+ struct hfi1_vnic_sdma *vnic_sdma = &vinfo->sdma[q_idx];
+ struct sdma_engine *sde = vnic_sdma->sde;
+ struct vnic_txreq *tx;
+ int ret = -ECOMM;
+
+ if (READ_ONCE(vnic_sdma->state) != HFI1_VNIC_SDMA_Q_ACTIVE)
+ goto tx_err;
+
+ if (!sde || !sdma_running(sde))
+ goto tx_err;
+
+ tx = kmem_cache_alloc(dd->vnic.txreq_cache, GFP_ATOMIC);
+ if (!tx) {
+ ret = -ENOMEM;
+ goto tx_err;
+ }
+
+ tx->sdma = vnic_sdma;
+ tx->skb = skb;
+ hfi1_vnic_update_pad(tx->pad, plen);
+ tx->plen = plen;
+ ret = build_vnic_tx_desc(sde, tx, pbc);
+ if (unlikely(ret))
+ goto free_desc;
+ tx->retry_count = 0;
+
+ ret = sdma_send_txreq(sde, &vnic_sdma->wait, &tx->txreq);
+ /* When -ECOMM, sdma callback will be called with ABORT status */
+ if (ret && unlikely(ret != -ECOMM))
+ goto free_desc;
+
+ return ret;
+
+free_desc:
+ sdma_txclean(dd, &tx->txreq);
+ kmem_cache_free(dd->vnic.txreq_cache, tx);
+tx_err:
+ if (ret != -EBUSY)
+ dev_kfree_skb_any(skb);
+ return ret;
+}
+
+/*
+ * hfi1_vnic_sdma_sleep - vnic sdma sleep function
+ *
+ * This function gets called from sdma_send_txreq() when there are not enough
+ * sdma descriptors available to send the packet. It adds Tx queue's wait
+ * structure to sdma engine's dmawait list to be woken up when descriptors
+ * become available.
+ */
+static int hfi1_vnic_sdma_sleep(struct sdma_engine *sde,
+ struct iowait *wait,
+ struct sdma_txreq *txreq,
+ unsigned int seq)
+{
+ struct hfi1_vnic_sdma *vnic_sdma =
+ container_of(wait, struct hfi1_vnic_sdma, wait);
+ struct hfi1_ibdev *dev = &vnic_sdma->dd->verbs_dev;
+ struct vnic_txreq *tx = container_of(txreq, struct vnic_txreq, txreq);
+
+ if (sdma_progress(sde, seq, txreq))
+ if (tx->retry_count++ < HFI1_VNIC_SDMA_RETRY_COUNT)
+ return -EAGAIN;
+
+ vnic_sdma->state = HFI1_VNIC_SDMA_Q_DEFERRED;
+ write_seqlock(&dev->iowait_lock);
+ if (list_empty(&vnic_sdma->wait.list))
+ list_add_tail(&vnic_sdma->wait.list, &sde->dmawait);
+ write_sequnlock(&dev->iowait_lock);
+ return -EBUSY;
+}
+
+/*
+ * hfi1_vnic_sdma_wakeup - vnic sdma wakeup function
+ *
+ * This function gets called when SDMA descriptors becomes available and Tx
+ * queue's wait structure was previously added to sdma engine's dmawait list.
+ * It notifies the upper driver about Tx queue wakeup.
+ */
+static void hfi1_vnic_sdma_wakeup(struct iowait *wait, int reason)
+{
+ struct hfi1_vnic_sdma *vnic_sdma =
+ container_of(wait, struct hfi1_vnic_sdma, wait);
+ struct hfi1_vnic_vport_info *vinfo = vnic_sdma->vinfo;
+ u8 evt = HFI_VNIC_EVT_TX0 + vnic_sdma->q_idx;
+ struct hfi1_vnic_notifier *notifier;
+
+ vnic_sdma->state = HFI1_VNIC_SDMA_Q_ACTIVE;
+ notifier = rcu_dereference(vinfo->notifier);
+ if (notifier && notifier->cb && test_bit(evt, vinfo->event_flags))
+ notifier->cb(vinfo->vport, evt);
+};
+
+inline bool hfi1_vnic_sdma_write_avail(struct hfi1_vnic_vport_info *vinfo,
+ u8 q_idx)
+{
+ struct hfi1_vnic_sdma *vnic_sdma = &vinfo->sdma[q_idx];
+
+ return (READ_ONCE(vnic_sdma->state) == HFI1_VNIC_SDMA_Q_ACTIVE);
+}
+
+void hfi1_vnic_sdma_init(struct hfi1_vnic_vport_info *vinfo)
+{
+ int i;
+
+ for (i = 0; i < vinfo->vport->hfi_info.num_tx_q; i++) {
+ struct hfi1_vnic_sdma *vnic_sdma = &vinfo->sdma[i];
+
+ iowait_init(&vnic_sdma->wait, 0, NULL, hfi1_vnic_sdma_sleep,
+ hfi1_vnic_sdma_wakeup, NULL);
+ vnic_sdma->sde = &vinfo->dd->per_sdma[i];
+ vnic_sdma->dd = vinfo->dd;
+ vnic_sdma->vinfo = vinfo;
+ vnic_sdma->q_idx = i;
+ vnic_sdma->state = HFI1_VNIC_SDMA_Q_ACTIVE;
+
+ /* Add a free descriptor watermark for wakeups */
+ if (vnic_sdma->sde->descq_cnt >= HFI1_VNIC_SDMA_DESC_WTRMRK) {
+ INIT_LIST_HEAD(&vnic_sdma->stx.list);
+ vnic_sdma->stx.num_desc = HFI1_VNIC_SDMA_DESC_WTRMRK;
+ list_add_tail(&vnic_sdma->stx.list,
+ &vnic_sdma->wait.tx_head);
+ }
+ }
+}
+
+static void hfi1_vnic_txreq_kmem_cache_ctor(void *obj)
+{
+ struct vnic_txreq *tx = (struct vnic_txreq *)obj;
+
+ memset(tx, 0, sizeof(*tx));
+}
+
+int hfi1_vnic_txreq_init(struct hfi1_devdata *dd)
+{
+ char buf[HFI1_VNIC_TXREQ_NAME_LEN];
+
+ snprintf(buf, sizeof(buf), "hfi1_%u_vnic_txreq_cache", dd->unit);
+ dd->vnic.txreq_cache = kmem_cache_create(buf,
+ sizeof(struct vnic_txreq),
+ 0, SLAB_HWCACHE_ALIGN,
+ hfi1_vnic_txreq_kmem_cache_ctor);
+ if (!dd->vnic.txreq_cache)
+ return -ENOMEM;
return 0;
}
+
+void hfi1_vnic_txreq_deinit(struct hfi1_devdata *dd)
+{
+ kmem_cache_destroy(dd->vnic.txreq_cache);
+ dd->vnic.txreq_cache = NULL;
+}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 41+ messages in thread
* Re: [RFC v2 02/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface
[not found] ` <1481788782-89964-3-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
@ 2016-12-15 8:53 ` Vishwanathapura, Niranjana
[not found] ` <20161215085349.GA90068-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
2016-12-15 17:36 ` Stephen Hemminger
0 siblings, 2 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-15 8:53 UTC (permalink / raw)
To: dledford-H+wXaHxf7aLQT0dZR+AlfA
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w
On Wed, Dec 14, 2016 at 11:59:34PM -0800, Vishwanathapura, Niranjana wrote:
>+
>+static inline bool is_hfi_ibdev(struct ib_device *ibdev)
>+{
>+ return !memcmp(ibdev->name, "hfi", 3);
>+}
I am thinking of adding a device capability flag to indicate HFI VNIC capabilty
instead of relying on the device name as above to identify a hfi ib deice.
Any comments? Probably it can be addressed by a separate patch later.
Niranjana
>
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 02/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface
[not found] ` <20161215085349.GA90068-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
@ 2016-12-15 8:56 ` Christoph Hellwig
0 siblings, 0 replies; 41+ messages in thread
From: Christoph Hellwig @ 2016-12-15 8:56 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: dledford-H+wXaHxf7aLQT0dZR+AlfA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w
On Thu, Dec 15, 2016 at 12:53:49AM -0800, Vishwanathapura, Niranjana wrote:
> On Wed, Dec 14, 2016 at 11:59:34PM -0800, Vishwanathapura, Niranjana wrote:
> > +
> > +static inline bool is_hfi_ibdev(struct ib_device *ibdev)
> > +{
> > + return !memcmp(ibdev->name, "hfi", 3);
> > +}
>
> I am thinking of adding a device capability flag to indicate HFI VNIC
> capabilty instead of relying on the device name as above to identify a hfi
> ib deice.
Absolutely.
> Any comments? Probably it can be addressed by a separate patch later.
no, comparing device names is always wrong, please do it ASAP.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <1481788782-89964-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
` (4 preceding siblings ...)
2016-12-15 7:59 ` [RFC v2 07/10] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) interface Vishwanathapura, Niranjana
@ 2016-12-15 9:12 ` Leon Romanovsky
[not found] ` <20161215091226.GC811-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
5 siblings, 1 reply; 41+ messages in thread
From: Leon Romanovsky @ 2016-12-15 9:12 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: dledford-H+wXaHxf7aLQT0dZR+AlfA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w
[-- Attachment #1: Type: text/plain, Size: 856 bytes --]
On Wed, Dec 14, 2016 at 11:59:32PM -0800, Vishwanathapura, Niranjana wrote:
> Thanks Jason for the valuable feedback.
> Here is the revised HFI VNIC patch series.
>
> ChangeLog:
> =========
> v1 => v2:
> a) Removed hfi_vnic bus, instead make hfi_vnic driver an 'ib client',
> as per feedback from Jason Gunthorpe.
> b) Interface changes, data structure changes and variable name changes
> associated with (a).
> c) Add hfi_ibdev abstraction to provide VNIC control operations to
> hfi_vnic client.
> d) Minor fixes
> e) Moved hfi_vnic driver from .../sw/intel/vnic/hfi_vnic to
> .../sw/intel/hfi_vnic.
To put it into proportion, Jason asked you to do different thing.
http://marc.info/?l=linux-rdma&m=147977108302151&w=2
http://marc.info/?l=linux-rdma&m=148000415401842&w=2
And Christoph,
http://marc.info/?l=linux-rdma&m=147985587425861&w=2
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20161215091226.GC811-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
@ 2016-12-15 14:52 ` ira.weiny
[not found] ` <20161215145212.GA29116-W4f6Xiosr+yv7QzWx2u06xL4W9x8LtSr@public.gmane.org>
0 siblings, 1 reply; 41+ messages in thread
From: ira.weiny @ 2016-12-15 14:52 UTC (permalink / raw)
To: Leon Romanovsky, Doug Ledford, Jeff Kirsher, David S. Miller
Cc: Vishwanathapura, Niranjana, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
On Thu, Dec 15, 2016 at 11:12:26AM +0200, Leon Romanovsky wrote:
> On Wed, Dec 14, 2016 at 11:59:32PM -0800, Vishwanathapura, Niranjana wrote:
> > Thanks Jason for the valuable feedback.
> > Here is the revised HFI VNIC patch series.
> >
> > ChangeLog:
> > =========
> > v1 => v2:
> > a) Removed hfi_vnic bus, instead make hfi_vnic driver an 'ib client',
> > as per feedback from Jason Gunthorpe.
> > b) Interface changes, data structure changes and variable name changes
> > associated with (a).
> > c) Add hfi_ibdev abstraction to provide VNIC control operations to
> > hfi_vnic client.
> > d) Minor fixes
> > e) Moved hfi_vnic driver from .../sw/intel/vnic/hfi_vnic to
> > .../sw/intel/hfi_vnic.
>
> To put it into proportion, Jason asked you to do different thing.
> http://marc.info/?l=linux-rdma&m=147977108302151&w=2
> http://marc.info/?l=linux-rdma&m=148000415401842&w=2
>
> And Christoph,
> http://marc.info/?l=linux-rdma&m=147985587425861&w=2
Understood. However, we never heard back from Niranjanas analysis of the code
which stated that > 60% of the code was dealing with the OPA MADs used to
configure this device.
https://www.spinics.net/lists/linux-rdma/msg43579.html
Furthermore, neither Dave nor Doug has had time to weigh in on what we should
do.
So before we make that change we wanted to get consensus on using the
hfi1_ibdev abstraction rather than the bus. This was the _real_ technical
change.
Beyond that it is really just which maintainer wants this driver. To that end
I've also cc'ed Jeff Kirsher who maintains drivers/net/ethernet/intel. Perhaps
Dave would like the driver to go through that tree?
I think there are pros and cons to both subtrees and in the end we will do
whatever is decided.
For maintainer review:
1) The driver encapsulates ethernet packets with OPA headers
2) VNIC uses OPA management packets (MADs) for its configuration
3) A significant portion (> 60% +) of the code is specific to OPA
https://www.spinics.net/lists/linux-rdma/msg43579.html
4) The driver is from Intel and we expect Intel to be the primary
contributor to the code.
5) The driver, like hfi1, is dual licensed (GPL/BSD)
6) Based on Christophs feedback we will be adding device capability
bits to the IB core to indicate HFI VNIC support.
https://www.spinics.net/lists/linux-rdma/msg44113.html
Doug, Dave, Jeff any thoughts?
Ira
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20161215145212.GA29116-W4f6Xiosr+yv7QzWx2u06xL4W9x8LtSr@public.gmane.org>
@ 2016-12-15 16:28 ` Doug Ledford
[not found] ` <380b05bf-a18e-1f20-7e8e-10b61f77dec7-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2017-01-09 7:51 ` Leon Romanovsky
0 siblings, 2 replies; 41+ messages in thread
From: Doug Ledford @ 2016-12-15 16:28 UTC (permalink / raw)
To: ira.weiny, Leon Romanovsky, Jeff Kirsher, David S. Miller
Cc: Vishwanathapura, Niranjana, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
[-- Attachment #1.1: Type: text/plain, Size: 5315 bytes --]
On 12/15/2016 9:52 AM, ira.weiny wrote:
> On Thu, Dec 15, 2016 at 11:12:26AM +0200, Leon Romanovsky wrote:
>> On Wed, Dec 14, 2016 at 11:59:32PM -0800, Vishwanathapura, Niranjana wrote:
>>> Thanks Jason for the valuable feedback.
>>> Here is the revised HFI VNIC patch series.
>>>
>>> ChangeLog:
>>> =========
>>> v1 => v2:
>>> a) Removed hfi_vnic bus, instead make hfi_vnic driver an 'ib client',
>>> as per feedback from Jason Gunthorpe.
>>> b) Interface changes, data structure changes and variable name changes
>>> associated with (a).
>>> c) Add hfi_ibdev abstraction to provide VNIC control operations to
>>> hfi_vnic client.
>>> d) Minor fixes
>>> e) Moved hfi_vnic driver from .../sw/intel/vnic/hfi_vnic to
>>> .../sw/intel/hfi_vnic.
>>
>> To put it into proportion, Jason asked you to do different thing.
>> http://marc.info/?l=linux-rdma&m=147977108302151&w=2
>> http://marc.info/?l=linux-rdma&m=148000415401842&w=2
>>
>> And Christoph,
>> http://marc.info/?l=linux-rdma&m=147985587425861&w=2
>
> Understood. However, we never heard back from Niranjanas analysis of the code
> which stated that > 60% of the code was dealing with the OPA MADs used to
> configure this device.
>
> https://www.spinics.net/lists/linux-rdma/msg43579.html
>
> Furthermore, neither Dave nor Doug has had time to weigh in on what we should
> do.
>
> So before we make that change we wanted to get consensus on using the
> hfi1_ibdev abstraction rather than the bus. This was the _real_ technical
> change.
>
> Beyond that it is really just which maintainer wants this driver. To that end
> I've also cc'ed Jeff Kirsher who maintains drivers/net/ethernet/intel. Perhaps
> Dave would like the driver to go through that tree?
>
>
> I think there are pros and cons to both subtrees and in the end we will do
> whatever is decided.
>
> For maintainer review:
>
> 1) The driver encapsulates ethernet packets with OPA headers
>
> 2) VNIC uses OPA management packets (MADs) for its configuration
>
> 3) A significant portion (> 60% +) of the code is specific to OPA
>
> https://www.spinics.net/lists/linux-rdma/msg43579.html
>
> 4) The driver is from Intel and we expect Intel to be the primary
> contributor to the code.
>
> 5) The driver, like hfi1, is dual licensed (GPL/BSD)
>
> 6) Based on Christophs feedback we will be adding device capability
> bits to the IB core to indicate HFI VNIC support.
>
> https://www.spinics.net/lists/linux-rdma/msg44113.html
>
>
> Doug, Dave, Jeff any thoughts?
>
> Ira
>
Sorry for my late reply. The series is relatively large, and also
tagged with RFC, so it got shuffled to the back burner while I worked on
the stuff for this pull request.
I just read through the comments in the V1 series between Jason et. al.,
and my take on things is like this:
1) Since your intent is to make this work with multiple versions of the
hfi drivers, I disagree with Jason that just because there is only one
driver today that we should keep it simple. Design it right from the
beginning of multi driver is your intent is, IMO, a better way to go.
You'll work out the bugs in the initial implementation and when it comes
time to add the second driver, things will go much more smoothly.
2) With more than 60% of the code being MAD related, and another
significant chunk being hfi related, and only a minor bit (20% maybe?)
being net related, I disagree that this belongs in the drivers/net or
net/ directories. Part of the purpose of putting code like this in any
given directory is to group it with what it is most tightly tied too.
That way people doing sub-tree wide changes know the rough scope of
their work as the code that needs changed is grouped together. Putting
this or IPoIB in one of the net trees would make it obvious to the
casual coder that these need changed for net changes, but would totally
hide the fact that once you tear into these drivers, there is a lot more
IB to them than there is net. What's more, when 60+% of driver is
non-net, then you end up having many more of my patches crossing over
into Dave's tree than the opposite if you put the code under my tree.
If nothing else, locality of code churn would say both this and IPoIB
belong here despite them being net drivers.
3) I would like some hard reasons why this driver deserves to exist?
I'm struggling very hard right now with why we would add an entirely new
"encapsulate IP over RDMA" driver. Even if you use regular Ethernet
MACs instead of IPoIB's 20byte MAC, I'm struggling for why IPoIB
couldn't be modified to know it supports two MAC sizes and provide
different net devices based on those different types? I'm struggling to
see why IPoIB couldn't be modified to essentially have two transport
layers underneath? I haven't done a thorough code review yet, but if I
get into the net driver portion of this and it has very much similarity
to the IPoIB net portion, I'm probably going to want answers about why
this can't be a modification of IPoIB to support multiple
transport/encapsulation types instead of a separate driver even more.
--
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
GPG Key ID: 0E572FDD
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
2016-12-15 7:59 [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
` (5 preceding siblings ...)
2016-12-15 7:59 ` [RFC v2 10/10] IB/hfi1: VNIC SDMA support Vishwanathapura, Niranjana
@ 2016-12-15 16:56 ` Jason Gunthorpe
[not found] ` <20161215165611.GB3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
6 siblings, 1 reply; 41+ messages in thread
From: Jason Gunthorpe @ 2016-12-15 16:56 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: dledford, linux-rdma, netdev, dennis.dalessandro, ira.weiny
On Wed, Dec 14, 2016 at 11:59:32PM -0800, Vishwanathapura, Niranjana wrote:
> create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/Kconfig
> create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/Makefile
Stil NAK on these paths, I already explained why 'sw' is totally
unsuitable. Put it in drivers/net or drivers/infiniband/ulp
Jason
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
[not found] ` <1481788782-89964-4-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
@ 2016-12-15 17:01 ` Jason Gunthorpe
2016-12-15 17:21 ` Hefty, Sean
[not found] ` <20161215170109.GC3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
0 siblings, 2 replies; 41+ messages in thread
From: Jason Gunthorpe @ 2016-12-15 17:01 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: dledford-H+wXaHxf7aLQT0dZR+AlfA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Sadanand Warrier, Sudeep Dutt,
Tanya K Jajodia, Andrzej Kacprowski
On Wed, Dec 14, 2016 at 11:59:35PM -0800, Vishwanathapura, Niranjana wrote:
> +/**
> + * union hfi_vnic_bypass_hdr - VNIC bypass header
> + * @slid: source lid
> + * @length: length of packet
> + * @becn: backward explicit congestion notification
> + * @dlid: destination lid
> + * @sc: service class
> + * @fecn: forward explicit congestion notification
> + * @l2: L2 type (2=16B)
> + * @lt: link transfer field
> + * @l4: L4 type
> + * @slid_high: upper 4 bits of source lid
> + * @dlid_high: upper 4 bits of destination lid
> + * @pkey: partition key
> + * @entropy: entropy
> + * @age: packet age
> + * @l4_hdr: L4 header
> + */
> +union hfi_vnic_bypass_hdr {
> + struct {
> + struct {
> + uint64_t slid : 20;
> + uint64_t length : 11;
> + uint64_t becn : 1;
> + uint64_t dlid : 20;
> + uint64_t sc : 5;
> + uint64_t rsvd : 3;
> + uint64_t fecn : 1;
> + uint64_t l2 : 2;
> + uint64_t lt : 1;
> + };
> + struct {
> + uint64_t l4 : 8;
> + uint64_t slid_high : 4;
> + uint64_t dlid_high : 4;
> + uint64_t pkey : 16;
> + uint64_t entropy : 16;
> + uint64_t age : 8;
> + uint64_t rsvd1 : 8;
> + };
> + struct {
> + uint32_t rsvd2 : 16;
> + uint32_t l4_hdr : 16;
> + };
> + } __packed;
> + u32 dw[5];
> +};
This isn't going to work on BE, please fix it.
> +/**
> + * struct __hfi_vesw_info - HFI vnic virtual switch info
> + */
> +struct __hfi_vesw_info {
> + u16 fabric_id;
> + u16 vesw_id;
> +
> + u8 rsvd0[6];
> + u16 def_port_mask;
> +
> + u8 rsvd1[2];
> + u16 pkey;
> +
> + u8 rsvd2[4];
> + u32 u_mcast_dlid;
> + u32 u_ucast_dlid[HFI_VESW_MAX_NUM_DEF_PORT];
> +
> + u8 rsvd3[44];
> + u16 eth_mtu[HFI_VNIC_MAX_NUM_PCP];
> + u16 eth_mtu_non_vlan;
> + u8 rsvd4[2];
> +} __packed;
This goes on the network too? Also looks like it has endian problems.
Ditto for all the __packed structures.
> +#define v_dbg(format, arg...) \
> + netdev_dbg(adapter->netdev, format, ## arg)
> +#define v_err(format, arg...) \
> + netdev_err(adapter->netdev, format, ## arg)
> +#define v_info(format, arg...) \
> + netdev_info(adapter->netdev, format, ## arg)
> +#define v_warn(format, arg...) \
> + netdev_warn(adapter->netdev, format, ## arg)
Relies on an 'adapter' local varable?? Ugly.
Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <380b05bf-a18e-1f20-7e8e-10b61f77dec7-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
@ 2016-12-15 17:07 ` Jason Gunthorpe
2016-12-15 18:19 ` Doug Ledford
[not found] ` <20161215170713.GD3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
0 siblings, 2 replies; 41+ messages in thread
From: Jason Gunthorpe @ 2016-12-15 17:07 UTC (permalink / raw)
To: Doug Ledford
Cc: ira.weiny, Leon Romanovsky, Jeff Kirsher, David S. Miller,
Vishwanathapura, Niranjana, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
On Thu, Dec 15, 2016 at 11:28:06AM -0500, Doug Ledford wrote:
> 1) Since your intent is to make this work with multiple versions of the
> hfi drivers, I disagree with Jason that just because there is only one
> driver today that we should keep it simple. Design it right from the
> beginning of multi driver is your intent is, IMO, a better way to go.
> You'll work out the bugs in the initial implementation and when it comes
> time to add the second driver, things will go much more smoothly.
If that is your position then this should be a straight up IB ULP that
works with any IB hardware.
There is nothing HFI specific about it except for the
micro-optimization of pushing packets via SDMA instead of post_send,
and that same micro optimization probably applies to ipoib.
In other words, lets see the first version as a straight ULP with no
special HFI hooks, then we can discuss how best to micro optimize it
for HFI SDMA.
Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* RE: [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
2016-12-15 17:01 ` Jason Gunthorpe
@ 2016-12-15 17:21 ` Hefty, Sean
2016-12-15 17:24 ` Jason Gunthorpe
[not found] ` <20161215170109.GC3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
1 sibling, 1 reply; 41+ messages in thread
From: Hefty, Sean @ 2016-12-15 17:21 UTC (permalink / raw)
To: Jason Gunthorpe, Vishwanathapura, Niranjana
Cc: dledford, linux-rdma, netdev, Dalessandro, Dennis, Weiny, Ira,
Warrier, Sadanand, Dutt, Sudeep, Jajodia, Tanya K, Kacprowski,
Andrzej
> This goes on the network too? Also looks like it has endian problems.
I don't think OPA supports BE systems, and I think it uses LE on the wire for at least some portions of its protocol.
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
2016-12-15 17:21 ` Hefty, Sean
@ 2016-12-15 17:24 ` Jason Gunthorpe
2016-12-15 17:26 ` Christoph Hellwig
0 siblings, 1 reply; 41+ messages in thread
From: Jason Gunthorpe @ 2016-12-15 17:24 UTC (permalink / raw)
To: Hefty, Sean
Cc: Vishwanathapura, Niranjana, dledford, linux-rdma, netdev,
Dalessandro, Dennis, Weiny, Ira, Warrier, Sadanand, Dutt, Sudeep,
Jajodia, Tanya K, Kacprowski, Andrzej
On Thu, Dec 15, 2016 at 05:21:05PM +0000, Hefty, Sean wrote:
> > This goes on the network too? Also looks like it has endian problems.
>
> I don't think OPA supports BE systems, and I think it uses LE on the
> wire for at least some portions of its protocol.
This is a linux driver for a PCI device.
It needs to support big endian systems, that is how we do things in
Linux.
If it uses LE on the wire then mark with __le and make it sparse clean.
Do not use bitfields without providing a BE version of the bitfield.
Jason
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
2016-12-15 17:24 ` Jason Gunthorpe
@ 2016-12-15 17:26 ` Christoph Hellwig
0 siblings, 0 replies; 41+ messages in thread
From: Christoph Hellwig @ 2016-12-15 17:26 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: Hefty, Sean, Vishwanathapura, Niranjana, dledford, linux-rdma,
netdev, Dalessandro, Dennis, Weiny, Ira, Warrier, Sadanand, Dutt,
Sudeep, Jajodia, Tanya K, Kacprowski, Andrzej
On Thu, Dec 15, 2016 at 10:24:37AM -0700, Jason Gunthorpe wrote:
> Do not use bitfields without providing a BE version of the bitfield.
Do not use bitfields ever for protocol defintions, period.
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 02/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface
2016-12-15 8:53 ` Vishwanathapura, Niranjana
[not found] ` <20161215085349.GA90068-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
@ 2016-12-15 17:36 ` Stephen Hemminger
1 sibling, 0 replies; 41+ messages in thread
From: Stephen Hemminger @ 2016-12-15 17:36 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: dledford, linux-rdma, netdev, dennis.dalessandro, ira.weiny
On Thu, 15 Dec 2016 00:53:49 -0800
"Vishwanathapura, Niranjana" <niranjana.vishwanathapura@intel.com> wrote:
> On Wed, Dec 14, 2016 at 11:59:34PM -0800, Vishwanathapura, Niranjana wrote:
> >+
> >+static inline bool is_hfi_ibdev(struct ib_device *ibdev)
> >+{
> >+ return !memcmp(ibdev->name, "hfi", 3);
> >+}
>
> I am thinking of adding a device capability flag to indicate HFI VNIC capabilty
> instead of relying on the device name as above to identify a hfi ib deice.
> Any comments? Probably it can be addressed by a separate patch later.
>
> Niranjana
>
> >
Agreed. device name can change at anytime.
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
2016-12-15 17:07 ` Jason Gunthorpe
@ 2016-12-15 18:19 ` Doug Ledford
[not found] ` <4e22bf29-2260-0768-ab17-9a8df6306f37-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
[not found] ` <20161215170713.GD3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
1 sibling, 1 reply; 41+ messages in thread
From: Doug Ledford @ 2016-12-15 18:19 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: ira.weiny, Leon Romanovsky, Jeff Kirsher, David S. Miller,
Vishwanathapura, Niranjana, linux-rdma, netdev,
dennis.dalessandro
[-- Attachment #1.1: Type: text/plain, Size: 1181 bytes --]
On 12/15/2016 12:07 PM, Jason Gunthorpe wrote:
> On Thu, Dec 15, 2016 at 11:28:06AM -0500, Doug Ledford wrote:
>
>> 1) Since your intent is to make this work with multiple versions of the
>> hfi drivers, I disagree with Jason that just because there is only one
>> driver today that we should keep it simple. Design it right from the
>> beginning of multi driver is your intent is, IMO, a better way to go.
>> You'll work out the bugs in the initial implementation and when it comes
>> time to add the second driver, things will go much more smoothly.
>
> If that is your position then this should be a straight up IB ULP that
> works with any IB hardware.
Yes, see my comments in point #3 of my previous email...
> There is nothing HFI specific about it except for the
> micro-optimization of pushing packets via SDMA instead of post_send,
> and that same micro optimization probably applies to ipoib.
>
> In other words, lets see the first version as a straight ULP with no
> special HFI hooks, then we can discuss how best to micro optimize it
> for HFI SDMA.
>
> Jason
>
--
Doug Ledford <dledford@redhat.com>
GPG Key ID: 0E572FDD
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <4e22bf29-2260-0768-ab17-9a8df6306f37-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
@ 2016-12-15 18:48 ` Jason Gunthorpe
[not found] ` <20161215184837.GA16552-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
0 siblings, 1 reply; 41+ messages in thread
From: Jason Gunthorpe @ 2016-12-15 18:48 UTC (permalink / raw)
To: Doug Ledford
Cc: ira.weiny, Leon Romanovsky, Jeff Kirsher, David S. Miller,
Vishwanathapura, Niranjana, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
On Thu, Dec 15, 2016 at 01:19:18PM -0500, Doug Ledford wrote:
> On 12/15/2016 12:07 PM, Jason Gunthorpe wrote:
> > On Thu, Dec 15, 2016 at 11:28:06AM -0500, Doug Ledford wrote:
> >
> >> 1) Since your intent is to make this work with multiple versions of the
> >> hfi drivers, I disagree with Jason that just because there is only one
> >> driver today that we should keep it simple. Design it right from the
> >> beginning of multi driver is your intent is, IMO, a better way to go.
> >> You'll work out the bugs in the initial implementation and when it comes
> >> time to add the second driver, things will go much more smoothly.
> >
> > If that is your position then this should be a straight up IB ULP that
> > works with any IB hardware.
>
> Yes, see my comments in point #3 of my previous email...
Well, I'm not opposed to the vnic idea - Mellanox had (has?) a similar
IB driver. There are lots of good reasons to strictly maintain the
ethernet presentation.
There is much more going on here than just changing the LLADDR,
essentially everything MAD focused is different compared to ipoib, and
it looks like the required datastructures are different too. This is
more of a map a mac to a OPA_LRH approach with SA mediated discovery,
by my eye.
The main share is the 'skb send' part, we've talked about hoisting
that out of ipoib in the past anyhow. A generic verb along those lines
would probably allow the sdma optimization for hfi for both this new
ulp and ipoib without creating such an ugly HFI1 specific interface.
Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20161215184837.GA16552-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
@ 2016-12-16 1:24 ` ira.weiny
[not found] ` <20161216012404.GD3785-W4f6Xiosr+yv7QzWx2u06xL4W9x8LtSr@public.gmane.org>
0 siblings, 1 reply; 41+ messages in thread
From: ira.weiny @ 2016-12-16 1:24 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: Doug Ledford, Leon Romanovsky, Jeff Kirsher, David S. Miller,
Vishwanathapura, Niranjana, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
On Thu, Dec 15, 2016 at 11:48:37AM -0700, Jason Gunthorpe wrote:
> On Thu, Dec 15, 2016 at 01:19:18PM -0500, Doug Ledford wrote:
> > On 12/15/2016 12:07 PM, Jason Gunthorpe wrote:
> > > On Thu, Dec 15, 2016 at 11:28:06AM -0500, Doug Ledford wrote:
> > >
> > >> 1) Since your intent is to make this work with multiple versions of the
> > >> hfi drivers, I disagree with Jason that just because there is only one
> > >> driver today that we should keep it simple. Design it right from the
> > >> beginning of multi driver is your intent is, IMO, a better way to go.
> > >> You'll work out the bugs in the initial implementation and when it comes
> > >> time to add the second driver, things will go much more smoothly.
> > >
> > > If that is your position then this should be a straight up IB ULP that
> > > works with any IB hardware.
> >
> > Yes, see my comments in point #3 of my previous email...
>
> Well, I'm not opposed to the vnic idea - Mellanox had (has?) a similar
> IB driver. There are lots of good reasons to strictly maintain the
> ethernet presentation.
Agreed. I'm pretty worried about the idea of putting VNIC into IPoIB. It
seems like a force fit at best.
>
> There is much more going on here than just changing the LLADDR,
> essentially everything MAD focused is different compared to ipoib, and
> it looks like the required datastructures are different too. This is
> more of a map a mac to a OPA_LRH approach with SA mediated discovery,
> by my eye.
>
> The main share is the 'skb send' part, we've talked about hoisting
> that out of ipoib in the past anyhow. A generic verb along those lines
> would probably allow the sdma optimization for hfi for both this new
> ulp and ipoib without creating such an ugly HFI1 specific interface.
I'm not sure what you mean about "skb send" being used by ipoib. Right now
IPoIB already supplies a "generic skb send" for _Verbs_ in ipoib_send.
I don't know what other devices would do to implement ipoib_send? To me, it
seems like the abstraction for IPoIB is at the proper layer now.
For OPA, the hfi driver supports both IPoIB and VNIC. So expecting IPoIB and
VNIC to use a generic "skb send" in ib_device is going to make hfi1 do a lot of
work to determine which ULP is calling it or make the interface kind of ugly.
Either way I don't see how this is better than a separate set of functions.
IMO the cleanest way to "clean up the ugly HFI1 interface" is to just put the
VNIC operations into ib_device similar to the iWarp specific structure
"iw_cm_verbs" which is there today.
If a device supports the VNIC operations then it can set the pointer and if not
it will be NULL. VNIC will look for that pointer for the support it needs. If
in the future other devices need modifications to that interface we can modify
it then.
Ira
>
> Jason
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20161216012404.GD3785-W4f6Xiosr+yv7QzWx2u06xL4W9x8LtSr@public.gmane.org>
@ 2016-12-16 2:30 ` Vishwanathapura, Niranjana
2016-12-16 4:17 ` Jason Gunthorpe
1 sibling, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-16 2:30 UTC (permalink / raw)
To: ira.weiny
Cc: Jason Gunthorpe, Doug Ledford, Leon Romanovsky, Jeff Kirsher,
David S. Miller, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
On Thu, Dec 15, 2016 at 08:24:05PM -0500, ira.weiny wrote:
>On Thu, Dec 15, 2016 at 11:48:37AM -0700, Jason Gunthorpe wrote:
>> On Thu, Dec 15, 2016 at 01:19:18PM -0500, Doug Ledford wrote:
>> > On 12/15/2016 12:07 PM, Jason Gunthorpe wrote:
>> > > On Thu, Dec 15, 2016 at 11:28:06AM -0500, Doug Ledford wrote:
>> > >
>> > >> 1) Since your intent is to make this work with multiple versions of the
>> > >> hfi drivers, I disagree with Jason that just because there is only one
>> > >> driver today that we should keep it simple. Design it right from the
>> > >> beginning of multi driver is your intent is, IMO, a better way to go.
>> > >> You'll work out the bugs in the initial implementation and when it comes
>> > >> time to add the second driver, things will go much more smoothly.
>> > >
>> > > If that is your position then this should be a straight up IB ULP that
>> > > works with any IB hardware.
>> >
>> > Yes, see my comments in point #3 of my previous email...
>>
>> Well, I'm not opposed to the vnic idea - Mellanox had (has?) a similar
>> IB driver. There are lots of good reasons to strictly maintain the
>> ethernet presentation.
>
>Agreed. I'm pretty worried about the idea of putting VNIC into IPoIB. It
>seems like a force fit at best.
>
Just to add what Jason, Ira already mentioned,
1) This isn't much common code between hfi_vnic and ipoib.
Besides we expect both ipoib and hfi_vnic to function parallely.
Registering with the network stack is also different.
hfi_vnic exchanges encapsulation information via IB MAD interface from OPA
EM which is not the case with ipoib.
We needed minimal set of interfaces (defined in include/rdma/opa_hfi.h in this
path series) that represents HW.
2) The design is very different. There are no path record queries, QPs etc in
hfi_vnic.
3) hfi_vnic also does the encapsulation with fabric (OPA) header, so bottom
driver only puts it on the wire.
Whereas in ipoib, bottom ib device driver does the encapsulation for ipoib.
4) hfi_vnic do not need ib work request/completion structures.
hfi_vnic supports multiple TX/RX queues.
>>
>> There is much more going on here than just changing the LLADDR,
>> essentially everything MAD focused is different compared to ipoib, and
>> it looks like the required datastructures are different too. This is
>> more of a map a mac to a OPA_LRH approach with SA mediated discovery,
>> by my eye.
>>
>> The main share is the 'skb send' part, we've talked about hoisting
>> that out of ipoib in the past anyhow. A generic verb along those lines
>> would probably allow the sdma optimization for hfi for both this new
>> ulp and ipoib without creating such an ugly HFI1 specific interface.
>
>I'm not sure what you mean about "skb send" being used by ipoib. Right now
>IPoIB already supplies a "generic skb send" for _Verbs_ in ipoib_send.
>
>I don't know what other devices would do to implement ipoib_send? To me, it
>seems like the abstraction for IPoIB is at the proper layer now.
>
>For OPA, the hfi driver supports both IPoIB and VNIC. So expecting IPoIB and
>VNIC to use a generic "skb send" in ib_device is going to make hfi1 do a lot of
>work to determine which ULP is calling it or make the interface kind of ugly.
>Either way I don't see how this is better than a separate set of functions.
>
>IMO the cleanest way to "clean up the ugly HFI1 interface" is to just put the
>VNIC operations into ib_device similar to the iWarp specific structure
>"iw_cm_verbs" which is there today.
>
>If a device supports the VNIC operations then it can set the pointer and if not
>it will be NULL. VNIC will look for that pointer for the support it needs. If
>in the future other devices need modifications to that interface we can modify
>it then.
>
>Ira
Yes, I agree. The interface defined in include/rdma/opa_hfi.h in this patch
series is pretty simple and generic interface that represents the HW.
If we include this file and put the hfi_vnic_ctrl_ops directly in ib_device
structure, then it will simplify lot of stuff. We don't need to abstract
out hfi_ibdev and define any ib device capability flag for VNIC support.
>
>>
>> Jason
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
>> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20161215170713.GD3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
@ 2016-12-16 2:47 ` Vishwanathapura, Niranjana
0 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-16 2:47 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: Doug Ledford, ira.weiny, Leon Romanovsky, Jeff Kirsher,
David S. Miller, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
On Thu, Dec 15, 2016 at 10:07:13AM -0700, Jason Gunthorpe wrote:
>On Thu, Dec 15, 2016 at 11:28:06AM -0500, Doug Ledford wrote:
>
>> 1) Since your intent is to make this work with multiple versions of the
>> hfi drivers, I disagree with Jason that just because there is only one
>> driver today that we should keep it simple. Design it right from the
>> beginning of multi driver is your intent is, IMO, a better way to go.
>> You'll work out the bugs in the initial implementation and when it comes
>> time to add the second driver, things will go much more smoothly.
>
>If that is your position then this should be a straight up IB ULP that
>works with any IB hardware.
>
>There is nothing HFI specific about it except for the
>micro-optimization of pushing packets via SDMA instead of post_send,
>and that same micro optimization probably applies to ipoib.
>
Responded on the other thread. As mentioned, there are differences between
ipoib and hfi_vnic interface. For hfi_vnic, we need simple interface as defined
in the include/rdma/opa_hfi.h that represents HW to put/get already
encapsulated OPA packets.
>In other words, lets see the first version as a straight ULP with no
>special HFI hooks, then we can discuss how best to micro optimize it
>for HFI SDMA.
>
As mentioned in other thread, that would be putting hfi_vnic_ctrl_ops in
ib_device structure.
Niranjana
>Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
[not found] ` <20161215170109.GC3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
@ 2016-12-16 2:59 ` Vishwanathapura, Niranjana
2016-12-16 4:24 ` Jason Gunthorpe
0 siblings, 1 reply; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-16 2:59 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: dledford-H+wXaHxf7aLQT0dZR+AlfA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Sadanand Warrier, Sudeep Dutt,
Tanya K Jajodia, Andrzej Kacprowski
On Thu, Dec 15, 2016 at 10:01:09AM -0700, Jason Gunthorpe wrote:
>On Wed, Dec 14, 2016 at 11:59:35PM -0800, Vishwanathapura, Niranjana wrote:
>> +/**
>> + * union hfi_vnic_bypass_hdr - VNIC bypass header
>> + * @slid: source lid
>> + * @length: length of packet
>> + * @becn: backward explicit congestion notification
>> + * @dlid: destination lid
>> + * @sc: service class
>> + * @fecn: forward explicit congestion notification
>> + * @l2: L2 type (2=16B)
>> + * @lt: link transfer field
>> + * @l4: L4 type
>> + * @slid_high: upper 4 bits of source lid
>> + * @dlid_high: upper 4 bits of destination lid
>> + * @pkey: partition key
>> + * @entropy: entropy
>> + * @age: packet age
>> + * @l4_hdr: L4 header
>> + */
>> +union hfi_vnic_bypass_hdr {
>> + struct {
>> + struct {
>> + uint64_t slid : 20;
>> + uint64_t length : 11;
>> + uint64_t becn : 1;
>> + uint64_t dlid : 20;
>> + uint64_t sc : 5;
>> + uint64_t rsvd : 3;
>> + uint64_t fecn : 1;
>> + uint64_t l2 : 2;
>> + uint64_t lt : 1;
>> + };
>> + struct {
>> + uint64_t l4 : 8;
>> + uint64_t slid_high : 4;
>> + uint64_t dlid_high : 4;
>> + uint64_t pkey : 16;
>> + uint64_t entropy : 16;
>> + uint64_t age : 8;
>> + uint64_t rsvd1 : 8;
>> + };
>> + struct {
>> + uint32_t rsvd2 : 16;
>> + uint32_t l4_hdr : 16;
>> + };
>> + } __packed;
>> + u32 dw[5];
>> +};
>
>This isn't going to work on BE, please fix it.
>
We have made the hfi_vnic driver dependent on CONFIG_X86_64.
But I agree with all the feedback here. I will remove bitfields
and instead use bit operations in the next revision.
>> +/**
>> + * struct __hfi_vesw_info - HFI vnic virtual switch info
>> + */
>> +struct __hfi_vesw_info {
>> + u16 fabric_id;
>> + u16 vesw_id;
>> +
>> + u8 rsvd0[6];
>> + u16 def_port_mask;
>> +
>> + u8 rsvd1[2];
>> + u16 pkey;
>> +
>> + u8 rsvd2[4];
>> + u32 u_mcast_dlid;
>> + u32 u_ucast_dlid[HFI_VESW_MAX_NUM_DEF_PORT];
>> +
>> + u8 rsvd3[44];
>> + u16 eth_mtu[HFI_VNIC_MAX_NUM_PCP];
>> + u16 eth_mtu_non_vlan;
>> + u8 rsvd4[2];
>> +} __packed;
>
>This goes on the network too? Also looks like it has endian problems.
>
>Ditto for all the __packed structures.
>
This is in CPU format. There is a separate big endian version of this structure
defined in hfi_vnic_encap.h in below patch (which gets sent on wire).
https://www.spinics.net/lists/linux-rdma/msg44111.html
>> +#define v_dbg(format, arg...) \
>> + netdev_dbg(adapter->netdev, format, ## arg)
>> +#define v_err(format, arg...) \
>> + netdev_err(adapter->netdev, format, ## arg)
>> +#define v_info(format, arg...) \
>> + netdev_info(adapter->netdev, format, ## arg)
>> +#define v_warn(format, arg...) \
>> + netdev_warn(adapter->netdev, format, ## arg)
>
>Relies on an 'adapter' local varable?? Ugly.
>
I am using the same approach as Intel NIC driver like e1000e and ixgbe.
>Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20161215165611.GB3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
@ 2016-12-16 4:14 ` Vishwanathapura, Niranjana
0 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-16 4:14 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: dledford-H+wXaHxf7aLQT0dZR+AlfA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w
On Thu, Dec 15, 2016 at 09:56:11AM -0700, Jason Gunthorpe wrote:
>On Wed, Dec 14, 2016 at 11:59:32PM -0800, Vishwanathapura, Niranjana wrote:
>> create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/Kconfig
>> create mode 100644 drivers/infiniband/sw/intel/hfi_vnic/Makefile
>
>Stil NAK on these paths, I already explained why 'sw' is totally
>unsuitable. Put it in drivers/net or drivers/infiniband/ulp
>
I understand. I did not want to change dirver location until we concenses
on where it belongs.
In next revision, I will move it under drivers/infiniband/ulp/hfi_vnic.
If anybody thinks it should be in a different folder, let me know.
Niranjana
>Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20161216012404.GD3785-W4f6Xiosr+yv7QzWx2u06xL4W9x8LtSr@public.gmane.org>
2016-12-16 2:30 ` Vishwanathapura, Niranjana
@ 2016-12-16 4:17 ` Jason Gunthorpe
2017-01-17 19:27 ` Vishwanathapura, Niranjana
1 sibling, 1 reply; 41+ messages in thread
From: Jason Gunthorpe @ 2016-12-16 4:17 UTC (permalink / raw)
To: ira.weiny
Cc: Doug Ledford, Leon Romanovsky, Jeff Kirsher, David S. Miller,
Vishwanathapura, Niranjana, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
On Thu, Dec 15, 2016 at 08:24:05PM -0500, ira.weiny wrote:
> > The main share is the 'skb send' part, we've talked about hoisting
> > that out of ipoib in the past anyhow. A generic verb along those lines
> > would probably allow the sdma optimization for hfi for both this new
> > ulp and ipoib without creating such an ugly HFI1 specific interface.
>
> I'm not sure what you mean about "skb send" being used by ipoib. Right now
> IPoIB already supplies a "generic skb send" for _Verbs_ in ipoib_send.
Sending a skb is very hard, the boring standard verbs implementation
is slow. Mellanox extended that with some simple offloads, but it is not
enough to get the full performance out of the hardware.
The suggestion is to add a 'skb qp', for lack of a better name, that is
perfectly optimized to delegate working with skbs to the driver. The
driver will then optmize with all possible offloads and bypass the
verbs qp API between netdev and driver.
Look at what is already in patch #2:
+struct hfi_vnic_ops {
+ int (*open)(struct hfi_vnic_port *vport, hfi_vnic_evt_cb_fn cb);
+ void (*close)(struct hfi_vnic_port *vport);
+ int (*put_skb)(struct hfi_vnic_port *vport,
+ u8 q_idx, struct sk_buff *skb);
+ struct sk_buff *(*get_skb)(struct hfi_vnic_port *vport, u8 q_idx);
+ u16 (*get_read_avail)(struct hfi_vnic_port *vport, u8 q_idx);
+ bool (*get_write_avail)(struct hfi_vnic_port *vport, u8 q_idx);
+ u8 (*select_queue)(struct hfi_vnic_port *vport, u8 vl, u8 entropy);
+ void (*config_notify)(struct hfi_vnic_port *vport,
+ u8 evt, bool enable);
That is almost what I'm talking about.
A 'vport' is a 'skb qp' that has been made overly specific.
So, clean it up to get rid of all the hfi specific stuff, stop calling
it a port. Get feedback from Mellanox. Refactor ipoib to use it to
show that it works sanely with both drivers.
> I don't know what other devices would do to implement ipoib_send? To me, it
> seems like the abstraction for IPoIB is at the proper layer now.
An example is multi queue tx with QPN sharing. We don't have anything
like that in verbs. At some point it just doesn't make sense to twist
verbs into knots to do this stuff - a skb qp is much cleaner and
powerful.
> For OPA, te hfi driver supports both IPoIB and VNIC. So expecting IPoIB and
> VNIC to use a generic "skb send" in ib_device is going to make hfi1 do a lot of
> work to determine which ULP is calling it or make the interface kind
> of ugly.
I don' think it will be that bad at all, the tx path is actually the
same execpt for the header construction. Handling that should not be
ugly, IMHO.
Think of it this way, if you do this you can probably boost the ipoib
performance on hfi as well.
Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
2016-12-16 2:59 ` Vishwanathapura, Niranjana
@ 2016-12-16 4:24 ` Jason Gunthorpe
[not found] ` <20161216042420.GB3797-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
0 siblings, 1 reply; 41+ messages in thread
From: Jason Gunthorpe @ 2016-12-16 4:24 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: dledford, linux-rdma, netdev, dennis.dalessandro, ira.weiny,
Sadanand Warrier, Sudeep Dutt, Tanya K Jajodia,
Andrzej Kacprowski
On Thu, Dec 15, 2016 at 06:59:47PM -0800, Vishwanathapura, Niranjana wrote:
> We have made the hfi_vnic driver dependent on CONFIG_X86_64.
Er, don't do that either?
> >>+struct __hfi_vesw_info {
> >>+ u16 fabric_id;
> >>+ u16 vesw_id;
> >>+
> >>+ u8 rsvd0[6];
> >>+ u16 def_port_mask;
> >>+
> >>+ u8 rsvd1[2];
> >>+ u16 pkey;
> >>+
> >>+ u8 rsvd2[4];
> >>+ u32 u_mcast_dlid;
> >>+ u32 u_ucast_dlid[HFI_VESW_MAX_NUM_DEF_PORT];
> >>+
> >>+ u8 rsvd3[44];
> >>+ u16 eth_mtu[HFI_VNIC_MAX_NUM_PCP];
> >>+ u16 eth_mtu_non_vlan;
> >>+ u8 rsvd4[2];
> >>+} __packed;
> >
> >This goes on the network too? Also looks like it has endian problems.
> >
> >Ditto for all the __packed structures.
> >
>
> This is in CPU format. There is a separate big endian version of
> this
Why are CPU handled structures packed and full of reserved fields?
Don't pack them if they are not pushed out to the network..
There were lots of __packed structures, any that go on the network
need be/le annoations.
Jason
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev
[not found] ` <20161216042420.GB3797-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
@ 2016-12-19 6:43 ` Vishwanathapura, Niranjana
0 siblings, 0 replies; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2016-12-19 6:43 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: dledford-H+wXaHxf7aLQT0dZR+AlfA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w,
ira.weiny-ral2JQCrhuEAvxtiuMwx3w, Sadanand Warrier, Sudeep Dutt,
Tanya K Jajodia, Andrzej Kacprowski
On Thu, Dec 15, 2016 at 09:24:20PM -0700, Jason Gunthorpe wrote:
>> >>+struct __hfi_vesw_info {
>> >>+ u16 fabric_id;
>> >>+ u16 vesw_id;
>> >>+
>> >>+ u8 rsvd0[6];
>> >>+ u16 def_port_mask;
>> >>+
>> >>+ u8 rsvd1[2];
>> >>+ u16 pkey;
>> >>+
>> >>+ u8 rsvd2[4];
>> >>+ u32 u_mcast_dlid;
>> >>+ u32 u_ucast_dlid[HFI_VESW_MAX_NUM_DEF_PORT];
>> >>+
>> >>+ u8 rsvd3[44];
>> >>+ u16 eth_mtu[HFI_VNIC_MAX_NUM_PCP];
>> >>+ u16 eth_mtu_non_vlan;
>> >>+ u8 rsvd4[2];
>> >>+} __packed;
>> >
>> >This goes on the network too? Also looks like it has endian problems.
>> >
>> >Ditto for all the __packed structures.
>> >
>>
>> This is in CPU format. There is a separate big endian version of
>> this
>
>Why are CPU handled structures packed and full of reserved fields?
>Don't pack them if they are not pushed out to the network..
>
>There were lots of __packed structures, any that go on the network
>need be/le annoations.
>
Well, driver treats the reserved fields to be sticky. ie., information
block returned (upon GET) to EM is not changed (from SET) except few fields
which driver is expected to modify.
Structures that go on wire are big endian __packed structures in
hfi_vnic_encap.h. Ok, I will remove the __packed attribute from CPU handled
structures here.
Niranjana
>Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
2016-12-15 16:28 ` Doug Ledford
[not found] ` <380b05bf-a18e-1f20-7e8e-10b61f77dec7-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
@ 2017-01-09 7:51 ` Leon Romanovsky
[not found] ` <20170109075104.GV15685-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
1 sibling, 1 reply; 41+ messages in thread
From: Leon Romanovsky @ 2017-01-09 7:51 UTC (permalink / raw)
To: Doug Ledford, ira.weiny
Cc: Jeff Kirsher, David S. Miller, Vishwanathapura, Niranjana,
linux-rdma, netdev, dennis.dalessandro
[-- Attachment #1: Type: text/plain, Size: 573 bytes --]
On Thu, Dec 15, 2016 at 11:28:06AM -0500, Doug Ledford wrote:
> On 12/15/2016 9:52 AM, ira.weiny wrote:
>
> 2) With more than 60% of the code being MAD related, and another
> significant chunk being hfi related, and only a minor bit (20% maybe?)
> being net related,
Hi Doug and Ira,
I may admit that I didn't read the code very deep, but from brief
overview, I didn't find support for the claim the "60% code is MAD related".
It looks like the opposite thing will be more accurate.
Can you help me to understand this claim? How did you come to this
conclusion?
Thanks
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20170109075104.GV15685-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
@ 2017-01-10 3:39 ` Vishwanathapura, Niranjana
[not found] ` <20170110033954.GA53664-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
0 siblings, 1 reply; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-01-10 3:39 UTC (permalink / raw)
To: Leon Romanovsky
Cc: Doug Ledford, ira.weiny, Jeff Kirsher, David S. Miller,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
On Mon, Jan 09, 2017 at 09:51:04AM +0200, Leon Romanovsky wrote:
>On Thu, Dec 15, 2016 at 11:28:06AM -0500, Doug Ledford wrote:
>> On 12/15/2016 9:52 AM, ira.weiny wrote:
>>
>> 2) With more than 60% of the code being MAD related, and another
>> significant chunk being hfi related, and only a minor bit (20% maybe?)
>> being net related,
>
>Hi Doug and Ira,
>
>I may admit that I didn't read the code very deep, but from brief
>overview, I didn't find support for the claim the "60% code is MAD related".
>It looks like the opposite thing will be more accurate.
>
>Can you help me to understand this claim? How did you come to this
>conclusion?
>
>Thanks
Hi Leon,
Here is the breakdown of patches based on functionality.
In this series, patches #3..#8 compose hfi_vnic driver. In that, patches #4, #7
and #8 are MAD focused (interfacing with MAD agent and handling MAD packets).
Patch #6 and half of #3 (_encap.c/h) are OPA encapsulation related. Patch #5 is
netdev statistic related (which includes statistics MAD definitions).
So, only part of patch #3 (_netdev.c and _ethtool.c) deals with interfacing
with netstack.
Those percentage numbers are based on actual lines of code in these patches
(files).
We are also looking into Jason’s suggestion to make hfi_vnic interface to the
bottom driver a generic interface. This will include moving some of the
netstack interfacing to the bottom hfi1 driver.
Niranjana
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20170110033954.GA53664-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
@ 2017-01-10 8:14 ` Leon Romanovsky
[not found] ` <20170110081402.GH7218-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
0 siblings, 1 reply; 41+ messages in thread
From: Leon Romanovsky @ 2017-01-10 8:14 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: Doug Ledford, ira.weiny, Jeff Kirsher, David S. Miller,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
[-- Attachment #1: Type: text/plain, Size: 2394 bytes --]
On Mon, Jan 09, 2017 at 07:39:54PM -0800, Vishwanathapura, Niranjana wrote:
> On Mon, Jan 09, 2017 at 09:51:04AM +0200, Leon Romanovsky wrote:
> > On Thu, Dec 15, 2016 at 11:28:06AM -0500, Doug Ledford wrote:
> > > On 12/15/2016 9:52 AM, ira.weiny wrote:
> > >
> > > 2) With more than 60% of the code being MAD related, and another
> > > significant chunk being hfi related, and only a minor bit (20% maybe?)
> > > being net related,
> >
> > Hi Doug and Ira,
> >
> > I may admit that I didn't read the code very deep, but from brief
> > overview, I didn't find support for the claim the "60% code is MAD related".
> > It looks like the opposite thing will be more accurate.
> >
> > Can you help me to understand this claim? How did you come to this
> > conclusion?
> >
> > Thanks
>
> Hi Leon,
>
> Here is the breakdown of patches based on functionality.
Hi Niranjana,
Thank you for breakdown, it helped a lot.
> In this series, patches #3..#8 compose hfi_vnic driver. In that, patches #4,
> #7 and #8 are MAD focused (interfacing with MAD agent and handling MAD
> packets).
Patch #3 is net device with all goodies from net stack.
https://patchwork.kernel.org/patch/9475639/
Patch #4 has one IB MAD related struct, all other things are specific to
OFA and HFI declarations.
https://patchwork.kernel.org/patch/9475653/
Patch #7 continues to implement agnostic to IB MAD net device.
https://patchwork.kernel.org/patch/9475641/
Patch #8 is without doubts, MAD related.
https://patchwork.kernel.org/patch/9475651/
Let's put aside patch #8, in such case, the proposed code won't need IB/core at
all and will look exactly like many other net drivers which encapsulate/decapsulate
packets.
> Patch #6 and half of #3 (_encap.c/h) are OPA encapsulation related. Patch #5
> is netdev statistic related (which includes statistics MAD definitions).
> So, only part of patch #3 (_netdev.c and _ethtool.c) deals with interfacing
> with netstack.
> Those percentage numbers are based on actual lines of code in these patches
> (files).
>
> We are also looking into Jason’s suggestion to make hfi_vnic interface to
> the bottom driver a generic interface. This will include moving some of the
> netstack interfacing to the bottom hfi1 driver.
Great, do you have rough estimation when will it be posted on the ML?
Thanks
>
> Niranjana
>
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20170110081402.GH7218-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
@ 2017-01-11 9:05 ` Vishwanathapura, Niranjana
2017-01-11 9:16 ` Leon Romanovsky
0 siblings, 1 reply; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-01-11 9:05 UTC (permalink / raw)
To: Leon Romanovsky
Cc: Doug Ledford, ira.weiny, Jeff Kirsher, David S. Miller,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
On Tue, Jan 10, 2017 at 10:14:02AM +0200, Leon Romanovsky wrote:
>On Mon, Jan 09, 2017 at 07:39:54PM -0800, Vishwanathapura, Niranjana wrote:
>> We are also looking into Jason’s suggestion to make hfi_vnic interface to
>> the bottom driver a generic interface. This will include moving some of the
>> netstack interfacing to the bottom hfi1 driver.
>
>Great, do you have rough estimation when will it be posted on the ML?
>
>Thanks
>
Hi Leon,
I am currently doing some prototype on this. I will respond with the interface
definition first once I have it solidified. RFC v3 with implemetion will be
later.
Niranjana
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
2017-01-11 9:05 ` Vishwanathapura, Niranjana
@ 2017-01-11 9:16 ` Leon Romanovsky
0 siblings, 0 replies; 41+ messages in thread
From: Leon Romanovsky @ 2017-01-11 9:16 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: Doug Ledford, ira.weiny, Jeff Kirsher, David S. Miller,
linux-rdma, netdev, dennis.dalessandro
[-- Attachment #1: Type: text/plain, Size: 789 bytes --]
On Wed, Jan 11, 2017 at 01:05:39AM -0800, Vishwanathapura, Niranjana wrote:
> On Tue, Jan 10, 2017 at 10:14:02AM +0200, Leon Romanovsky wrote:
> > On Mon, Jan 09, 2017 at 07:39:54PM -0800, Vishwanathapura, Niranjana wrote:
> > > We are also looking into Jason’s suggestion to make hfi_vnic interface to
> > > the bottom driver a generic interface. This will include moving some of the
> > > netstack interfacing to the bottom hfi1 driver.
> >
> > Great, do you have rough estimation when will it be posted on the ML?
> >
> > Thanks
> >
>
> Hi Leon,
>
> I am currently doing some prototype on this. I will respond with the
> interface definition first once I have it solidified. RFC v3 with
> implemetion will be later.
It makes sense.
Thanks
>
> Niranjana
>
>
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
2016-12-16 4:17 ` Jason Gunthorpe
@ 2017-01-17 19:27 ` Vishwanathapura, Niranjana
[not found] ` <20170117192720.GA2833-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
0 siblings, 1 reply; 41+ messages in thread
From: Vishwanathapura, Niranjana @ 2017-01-17 19:27 UTC (permalink / raw)
To: Jason Gunthorpe
Cc: ira.weiny, Doug Ledford, Leon Romanovsky, Jeff Kirsher,
David S. Miller, linux-rdma, netdev, dennis.dalessandro
Thanks Jason for the valuable inputs.
Here is the new generic interface.
Overview:
Bottom driver defines net_device_ops. The upper driver can override it.
For example, upper driver can implement ndo_open() which calls bottom driver's
ndo_open() and also do some book keeping.
include/rdma/ib_verbs.h:
/* rdma netdev type - specifies protocol type */
enum rdma_netdev_t {
RDMA_NETDEV_HFI_VNIC,
};
/* rdma netdev
* For usecases where netstack interfacing is required.
*/
struct rdma_netdev {
struct net_device *netdev;
u8 port_num;
/* client private data structure */
void *clnt_priv;
/* control functions */
void (*set_id)(struct rdma_netdev *rn, int id);
void (*set_state)(struct rdma_netdev *rn, int state);
};
struct ib_device {
...
...
/* rdma netdev operations */
struct net_device *(*alloc_rdma_netdev)(struct ib_device *device,
u8 port_num,
enum rdma_netdev_t type,
const char *name,
unsigned char name_assign_type,
void (*setup)(struct net_device *));
void (*free_rdma_netdev)(struct net_device *netdev);
};
hfi1 driver:
/* rdma netdev's private data structure */
struct hfi1_rdma_netdev {
struct rdma_netdev rn; /* keep this first */
/* hfi1's vnic private data follows */
};
include/rdma/opa_hfi.h:
/* Client's ndo operations use below function instead of netdev_priv() */
static inline void *hfi_vnic_priv(const struct net_device *dev)
{
struct rdma_netdev *rn = netdev_priv(dev);
return rn->clnt_priv;
}
/* Overrides rtnl_link_stats64 to include hfi_vnic stats.
* ndo_get_stats64() can be used to get the stats
*/
struct hfi_vnic_stats {
/* standard netdev statistics */
struct rtnl_link_stats64 netstat;
/* HFI VNIC statistics */
u64 tx_mcastbcast;
u64 tx_untagged;
u64 tx_vlan;
u64 tx_64_size;
u64 tx_65_127;
u64 tx_128_255;
u64 tx_256_511;
u64 tx_512_1023;
u64 tx_1024_1518;
u64 tx_1519_max;
u64 rx_untagged;
u64 rx_vlan;
u64 rx_64_size;
u64 rx_65_127;
u64 rx_128_255;
u64 rx_256_511;
u64 rx_512_1023;
u64 rx_1024_1518;
u64 rx_1519_max;
u64 rx_runt;
u64 rx_oversize;
};
I have started working on porting hfi_vnic as per this new interface.
I will post RFC v3 later.
Posting the interface definition early for comments.
Thanks,
Niranjana
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
[not found] ` <20170117192720.GA2833-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
@ 2017-01-18 5:43 ` Leon Romanovsky
2017-01-18 16:46 ` Jason Gunthorpe
0 siblings, 1 reply; 41+ messages in thread
From: Leon Romanovsky @ 2017-01-18 5:43 UTC (permalink / raw)
To: Vishwanathapura, Niranjana
Cc: Jason Gunthorpe, ira.weiny, Doug Ledford, Jeff Kirsher,
David S. Miller, linux-rdma-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA,
dennis.dalessandro-ral2JQCrhuEAvxtiuMwx3w
[-- Attachment #1: Type: text/plain, Size: 2604 bytes --]
On Tue, Jan 17, 2017 at 11:27:20AM -0800, Vishwanathapura, Niranjana wrote:
> Thanks Jason for the valuable inputs.
>
> Here is the new generic interface.
>
> Overview:
> Bottom driver defines net_device_ops. The upper driver can override it.
> For example, upper driver can implement ndo_open() which calls bottom
> driver's ndo_open() and also do some book keeping.
>
>
> include/rdma/ib_verbs.h:
>
> /* rdma netdev type - specifies protocol type */
> enum rdma_netdev_t {
> RDMA_NETDEV_HFI_VNIC,
> };
>
> /* rdma netdev
> * For usecases where netstack interfacing is required.
> */
> struct rdma_netdev {
> struct net_device *netdev;
> u8 port_num;
>
> /* client private data structure */
> void *clnt_priv;
>
> /* control functions */
> void (*set_id)(struct rdma_netdev *rn, int id);
> void (*set_state)(struct rdma_netdev *rn, int state);
> };
>
> struct ib_device {
> ...
> ...
> /* rdma netdev operations */
> struct net_device *(*alloc_rdma_netdev)(struct ib_device *device,
> u8 port_num,
> enum rdma_netdev_t type,
> const char *name,
> unsigned char name_assign_type,
> void (*setup)(struct net_device *));
> void (*free_rdma_netdev)(struct net_device *netdev);
> };
>
>
> hfi1 driver:
>
> /* rdma netdev's private data structure */
> struct hfi1_rdma_netdev {
> struct rdma_netdev rn; /* keep this first */
> /* hfi1's vnic private data follows */
> };
>
>
> include/rdma/opa_hfi.h:
>
> /* Client's ndo operations use below function instead of netdev_priv() */
> static inline void *hfi_vnic_priv(const struct net_device *dev)
> {
> struct rdma_netdev *rn = netdev_priv(dev);
>
> return rn->clnt_priv;
> }
>
> /* Overrides rtnl_link_stats64 to include hfi_vnic stats.
> * ndo_get_stats64() can be used to get the stats
> */
> struct hfi_vnic_stats {
> /* standard netdev statistics */
> struct rtnl_link_stats64 netstat;
>
> /* HFI VNIC statistics */
> u64 tx_mcastbcast;
> u64 tx_untagged;
> u64 tx_vlan;
> u64 tx_64_size;
> u64 tx_65_127;
> u64 tx_128_255;
> u64 tx_256_511;
> u64 tx_512_1023;
> u64 tx_1024_1518;
> u64 tx_1519_max;
>
> u64 rx_untagged;
> u64 rx_vlan;
> u64 rx_64_size;
> u64 rx_65_127;
> u64 rx_128_255;
> u64 rx_256_511;
> u64 rx_512_1023;
> u64 rx_1024_1518;
> u64 rx_1519_max;
>
> u64 rx_runt;
> u64 rx_oversize;
> };
>
> I have started working on porting hfi_vnic as per this new interface.
> I will post RFC v3 later.
> Posting the interface definition early for comments.
I wonder how many people will comment it without seeing usage example.
>
> Thanks,
> Niranjana
>
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 41+ messages in thread
* Re: [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC)
2017-01-18 5:43 ` Leon Romanovsky
@ 2017-01-18 16:46 ` Jason Gunthorpe
0 siblings, 0 replies; 41+ messages in thread
From: Jason Gunthorpe @ 2017-01-18 16:46 UTC (permalink / raw)
To: Leon Romanovsky
Cc: Vishwanathapura, Niranjana, ira.weiny, Doug Ledford,
Jeff Kirsher, David S. Miller, linux-rdma, netdev,
dennis.dalessandro
On Wed, Jan 18, 2017 at 07:43:54AM +0200, Leon Romanovsky wrote:
> > I have started working on porting hfi_vnic as per this new interface.
> > I will post RFC v3 later.
> > Posting the interface definition early for comments.
>
> I wonder how many people will comment it without seeing usage example.
It is my hope that Mellanox will contribute to this and use it for
ipoib - we've have had conversations along these lines in the
past... Apparently mlx4/5 could run quite a bit faster if used like
this.
Jason
^ permalink raw reply [flat|nested] 41+ messages in thread
end of thread, other threads:[~2017-01-18 16:46 UTC | newest]
Thread overview: 41+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-12-15 7:59 [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 01/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) documentation Vishwanathapura, Niranjana
[not found] ` <1481788782-89964-1-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2016-12-15 7:59 ` [RFC v2 02/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) interface Vishwanathapura, Niranjana
[not found] ` <1481788782-89964-3-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2016-12-15 8:53 ` Vishwanathapura, Niranjana
[not found] ` <20161215085349.GA90068-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
2016-12-15 8:56 ` Christoph Hellwig
2016-12-15 17:36 ` Stephen Hemminger
2016-12-15 7:59 ` [RFC v2 03/10] IB/hfi-vnic: Virtual Network Interface Controller (VNIC) netdev Vishwanathapura, Niranjana
[not found] ` <1481788782-89964-4-git-send-email-niranjana.vishwanathapura-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
2016-12-15 17:01 ` Jason Gunthorpe
2016-12-15 17:21 ` Hefty, Sean
2016-12-15 17:24 ` Jason Gunthorpe
2016-12-15 17:26 ` Christoph Hellwig
[not found] ` <20161215170109.GC3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2016-12-16 2:59 ` Vishwanathapura, Niranjana
2016-12-16 4:24 ` Jason Gunthorpe
[not found] ` <20161216042420.GB3797-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2016-12-19 6:43 ` Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 04/10] IB/hfi-vnic: VNIC Ethernet Management (EM) structure definitions Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 06/10] IB/hfi-vnic: VNIC MAC table support Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 07/10] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) interface Vishwanathapura, Niranjana
2016-12-15 9:12 ` [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Leon Romanovsky
[not found] ` <20161215091226.GC811-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
2016-12-15 14:52 ` ira.weiny
[not found] ` <20161215145212.GA29116-W4f6Xiosr+yv7QzWx2u06xL4W9x8LtSr@public.gmane.org>
2016-12-15 16:28 ` Doug Ledford
[not found] ` <380b05bf-a18e-1f20-7e8e-10b61f77dec7-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2016-12-15 17:07 ` Jason Gunthorpe
2016-12-15 18:19 ` Doug Ledford
[not found] ` <4e22bf29-2260-0768-ab17-9a8df6306f37-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2016-12-15 18:48 ` Jason Gunthorpe
[not found] ` <20161215184837.GA16552-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2016-12-16 1:24 ` ira.weiny
[not found] ` <20161216012404.GD3785-W4f6Xiosr+yv7QzWx2u06xL4W9x8LtSr@public.gmane.org>
2016-12-16 2:30 ` Vishwanathapura, Niranjana
2016-12-16 4:17 ` Jason Gunthorpe
2017-01-17 19:27 ` Vishwanathapura, Niranjana
[not found] ` <20170117192720.GA2833-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
2017-01-18 5:43 ` Leon Romanovsky
2017-01-18 16:46 ` Jason Gunthorpe
[not found] ` <20161215170713.GD3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2016-12-16 2:47 ` Vishwanathapura, Niranjana
2017-01-09 7:51 ` Leon Romanovsky
[not found] ` <20170109075104.GV15685-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
2017-01-10 3:39 ` Vishwanathapura, Niranjana
[not found] ` <20170110033954.GA53664-wPcXA7LoDC+1XWohqUldA0EOCMrvLtNR@public.gmane.org>
2017-01-10 8:14 ` Leon Romanovsky
[not found] ` <20170110081402.GH7218-U/DQcQFIOTAAJjI8aNfphQ@public.gmane.org>
2017-01-11 9:05 ` Vishwanathapura, Niranjana
2017-01-11 9:16 ` Leon Romanovsky
2016-12-15 7:59 ` [RFC v2 05/10] IB/hfi-vnic: VNIC statistics support Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 08/10] IB/hfi-vnic: VNIC Ethernet Management Agent (VEMA) function Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 09/10] IB/hfi1: Virtual Network Interface Controller (VNIC) support Vishwanathapura, Niranjana
2016-12-15 7:59 ` [RFC v2 10/10] IB/hfi1: VNIC SDMA support Vishwanathapura, Niranjana
2016-12-15 16:56 ` [RFC v2 00/10] HFI Virtual Network Interface Controller (VNIC) Jason Gunthorpe
[not found] ` <20161215165611.GB3264-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2016-12-16 4:14 ` Vishwanathapura, Niranjana
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.