[PATCH libibverbs 0/2] Add support for don't trap steering rule

[PATCH libibverbs 0/2] Add support for don't trap steering rule

[Yishai Hadas]

Hi Doug,

This series exposes to user application the option to create
a normal flow steering rule that doesn't trap received packets,
allowing them to match lower prioritized rules.

You already took the kernel part for coming merge window of 4.6.

The series includes two patches:
The first patch exposes the new flag.
The second patch is a man page for.

Yishai
------

Marina Varshaver (1):
  Add support for don't trap steering rule

Matan Barak (1):
  Add man page for flow steering verbs

 Makefile.am                |   6 ++-
 include/infiniband/verbs.h |   3 +-
 man/ibv_create_flow.3      | 112 +++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 118 insertions(+), 3 deletions(-)
 create mode 100644 man/ibv_create_flow.3

-- 
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

[PATCH libibverbs 1/2] Add support for don't trap steering rule

[Yishai Hadas]

From: Marina Varshaver <marinav-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Add an option to create a normal flow steering rule that doesn't trap
received packets, allowing them to match lower prioritized rules.

When the don't trap rule exists and matches a packet, the underlying HCA
should pass the packet to the rule's assigned QP(s). However, the HCA
will continue looking for other matches at lower priority rules, which
may be assigned to other QPs. This will let them get the traffic as
well.

Signed-off-by: Marina Varshaver <marinav-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Reviewed-by: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 include/infiniband/verbs.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/include/infiniband/verbs.h b/include/infiniband/verbs.h
index d0c2969..0c29f1a 100644
--- a/include/infiniband/verbs.h
+++ b/include/infiniband/verbs.h
@@ -840,7 +840,8 @@ struct ibv_ah {
 };
 
 enum ibv_flow_flags {
-	IBV_FLOW_ATTR_FLAGS_ALLOW_LOOP_BACK = 1,
+	IBV_FLOW_ATTR_FLAGS_ALLOW_LOOP_BACK = 1 << 0,
+	IBV_FLOW_ATTR_FLAGS_DONT_TRAP = 1 << 1,
 };
 
 enum ibv_flow_attr_type {
-- 
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

[PATCH libibverbs 2/2] Add man page for flow steering verbs

[Yishai Hadas]

From: Matan Barak <matanb-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Add man page for ibv_create_flow/ibv_destroy_flow verbs, it includes
the new dont_trap flag.

Signed-off-by: Matan Barak <matanb-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Marina Varshaver <marinav-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Reviewed-by: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 Makefile.am           |   6 ++-
 man/ibv_create_flow.3 | 112 ++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 116 insertions(+), 2 deletions(-)
 create mode 100644 man/ibv_create_flow.3

diff --git a/Makefile.am b/Makefile.am
index eefda4a..7296c46 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -53,7 +53,7 @@ man_MANS = man/ibv_asyncwatch.1 man/ibv_devices.1 man/ibv_devinfo.1	\
     man/ibv_create_ah.3 man/ibv_create_ah_from_wc.3			\
     man/ibv_create_comp_channel.3 man/ibv_create_cq.3			\
     man/ibv_create_qp.3 man/ibv_create_srq.3 man/ibv_event_type_str.3	\
-    man/ibv_fork_init.3 man/ibv_get_async_event.3			\
+    man/ibv_fork_init.3 man/ibv_get_async_event.3 man/ibv_create_flow.3 \
     man/ibv_get_cq_event.3 man/ibv_get_device_guid.3			\
     man/ibv_get_device_list.3 man/ibv_get_device_name.3			\
     man/ibv_modify_qp.3 man/ibv_modify_srq.3 man/ibv_open_device.3	\
@@ -101,6 +101,7 @@ install-data-hook:
 	$(RM) mbps_to_ibv_rate.3 && \
 	$(RM) ibv_close_xrcd.3 && \
 	$(RM) ibv_dealloc_mw.3 && \
+	$(RM) ibv_destroy_flow.3 && \
 	$(LN_S) ibv_get_async_event.3 ibv_ack_async_event.3 && \
 	$(LN_S) ibv_get_cq_event.3 ibv_ack_cq_events.3 && \
 	$(LN_S) ibv_open_device.3 ibv_close_device.3 && \
@@ -119,4 +120,5 @@ install-data-hook:
 	$(LN_S) ibv_event_type_str.3 ibv_port_state_str.3 && \
 	$(LN_S) ibv_rate_to_mbps.3 mbps_to_ibv_rate.3 && \
 	$(LN_S) ibv_open_xrcd.3 ibv_close_xrcd.3 && \
-	$(LN_S) ibv_alloc_mw.3 ibv_dealloc_mw.3
+	$(LN_S) ibv_alloc_mw.3 ibv_dealloc_mw.3 && \
+	$(LN_S) ibv_create_flow.3 ibv_destroy_flow.3
diff --git a/man/ibv_create_flow.3 b/man/ibv_create_flow.3
new file mode 100644
index 0000000..4ba6ee7
--- /dev/null
+++ b/man/ibv_create_flow.3
@@ -0,0 +1,112 @@
+.TH IBV_CREATE_FLOW 3 2016-03-15 libibverbs "Libibverbs Programmer's Manual"
+.SH "NAME"
+ibv_create_flow, ibv_destroy_flow \- create or destroy flow steering rules
+.SH "SYNOPSIS"
+.nf
+.B #include <infiniband/verbs.h>
+.sp
+.BI "struct ibv_flow *ibv_create_flow(struct ibv_qp " "*qp" ,
+.BI "                                 struct ibv_flow_attr " "*flow_attr");
+.BI "int ibv_destroy_flow(struct ibv_flow " "*flow_id");
+.sp
+.fi
+.SH "DESCRIPTION"
+.SS ibv_create_flow()
+allows a user application QP
+.I qp
+to be attached into a specified flow
+.I flow
+which is defined in
+.I <infiniband/verbs.h>
+.PP
+.nf
+struct ibv_flow_attr {
+.in +8
+uint32_t comp_mask;                     /* Future extendibility */
+enum ibv_flow_attr_type type;           /* Rule type - see below */
+uint16_t size;                          /* Size of command */
+uint16_t priority;                      /* Rule priority - see below */
+uint8_t num_of_specs;                   /* Number of ibv_flow_spec_xxx */
+uint8_t port;                           /* The uplink port number */
+uint32_t flags;                         /* Extra flags for rule - see below */
+/* Following are the optional layers according to user request
+ * struct ibv_flow_spec_xxx
+ * struct ibv_flow_spec_yyy
+ */
+.in -8
+};
+.sp
+.nf
+enum ibv_flow_attr_type {
+.in +8
+IBV_FLOW_ATTR_NORMAL		= 0x0,		/* Steering according to rule specifications */
+IBV_FLOW_ATTR_ALL_DEFAULT	= 0x1,		/* Default unicast and multicast rule - receive all Eth traffic which isn't steered to any QP */
+IBV_FLOW_ATTR_MC_DEFAULT 	= 0x2,		/* Default multicast rule - receive all Eth multicast traffic which isn't steered to any QP */
+.in -8
+};
+.sp
+.nf
+enum ibv_flow_flags {
+.in +8
+IBV_FLOW_ATTR_FLAGS_ALLOW_LOOP_BACK = 1 << 0,	/* Apply the rules on packets that were sent from the attached QP through loopback */
+IBV_FLOW_ATTR_FLAGS_DONT_TRAP       = 1 << 1,	/* Rule doesn't trap received packets, allowing them to match lower prioritized rules */
+.in -8
+};
+.fi
+.PP
+Each spec struct holds the relevant network layer parameters for matching. To enforce the match, the user sets a mask for each parameter.
+.br
+If the bit is set in the mask, the corresponding bit in the value should be matched.
+.br
+Note that most vendors support either full mask (all "1"s) or zero mask (all "0"s).
+.br
+.B Network parameters in the relevant network structs should be given in network order (big endian).
+
+.SS Flow domains and priority
+Flow steering defines the concept of domain and priority. Each domain represents an application that can attach a flow.
+Domains are prioritized. A higher priority domain will always supersede a lower priority domain when their flow specifications overlap.
+.br
+.B IB verbs have the higher priority domain.
+.br
+In addition to the domain, there is priority within each of the domains.
+A lower priority numeric value (higher priority) takes precedence over matching rules with higher numeric priority value (lower priority).
+It is important to note that the priority value of a flow spec is used not only to establish the precedence of conflicting flow matches
+but also as a way to abstract the order on which flow specs are tested for matches. Flows with higher priorities will be tested before flows with lower priorities.
+.PP
+.SS ibv_destroy_flow()
+destroys the flow
+.I flow_id\fR.
+.SH "RETURN VALUE"
+.B ibv_create_flow()
+returns a pointer to the flow, or NULL if the request fails. In case of an error, errno is updated.
+.PP
+.B ibv_destroy_flow()
+returns 0 on success, or the value of errno on failure (which indicates the failure reason).
+.SH "ERRORS"
+.SS EINVAL
+.B ibv_create_flow()
+flow specification, QP or priority are invalid
+.PP
+.B ibv_destroy_flow()
+flow_id is invalid
+.SS ENOMEM
+Couldn't create/destroy flow, not enough memory
+.SS ENXIO
+Device managed flow steering isn't currently supported
+.SS EPERM
+No permissions to add the flow steering rule
+.SH "NOTES"
+These verbs are available only for devices supporting
+.br
+IBV_DEVICE_MANAGED_FLOW_STEERING and only for QPs of Transport Service Type
+.BR IBV_QPT_UD
+or
+.BR IBV_QPT_RAW_PACKET
+.PP
+.SH "AUTHORS"
+.TP
+Hadar Hen Zion <hadarh-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
+.TP
+Matan Barak <matanb-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
+.TP
+Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
-- 
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