From: Erez Shitrit
Now that the FW should allow it we can go to a wire interface under the
assumption it is a north directed wire.
Signed-off-by: Erez Shitrit
Acked-by: Dariusz Sosnowski
---
drivers/net/mlx5/hws/mlx5dr_action.c | 9 ++---
1 file changed, 2 insertions(+), 7 deletions(-)
d
From: Erez Shitrit
STC is a limited resource of the HW, and might get consumed till no more
contexts can be opened.
So, let the user to define the size of how many STCs to allocate per
context.
In case the user has many representors, no need to allocate per each of
them the default value of STCs,
From: Hamdan Igbaria
In IPv6 to IPv4 scenario when copying the TC field to the
TOS[DSCP:6 bits, ECN: 2 bits] field the mapping of the fields
get set wrongly, since the destination field of TOS is defined as 2
split fields of DSCP and ECN.
In IPv4 to IPv6 we don't have this issue since both the so
From: Hamdan Igbaria
In IPv6 to IPv4 scenario when copying the TC field to the
TOS[DSCP:6 bits, ECN: 2 bits] field the mapping of the fields
get set wrongly, since the destination field of TOS is defined as 2
split fields of DSCP and ECN.
In IPv4 to IPv6 we don't have this issue since both the so
From: Erez Shitrit
We don't have enough registers to copy TTL and TOS, so we will set TTL
to be the default value (64) and will copy TOS.
Fixes: 06d969a8c5b8 ("net/mlx5/hws: support NAT64 flow action")
Cc: ere...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
From: Erez Shitrit
Due to HW limitation we got two csum's that were not correct, udp and
ip, both of them were not calculated correctly by the HW.
By adding the next W/A we allow the HW to collect it well:
Separate the protocol field and zero all the addresses before fixed the
UDP csum.
We saw th
From: Erez Shitrit
Removing a redundant variable.
Was there from day 1, not in use.
Fixes: f8c8a6d8440d ("net/mlx5/hws: add action object")
Cc: ere...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_pat_arg.h | 1 -
1 file chan
From: Alex Vesker
When calling item convert function we need to pass the port_id
in the attributes. This value should be passed not only for cases
that match on PORT related items, to resolve we will always pass it.
Fixes: 572fe9ef2f46 ("net/mlx5/hws: fix port ID for root table")
Cc: ere...@nvid
From: Erez Shitrit
No more ignoring this action while destroying it.
Fixes: f8c8a6d8440d ("net/mlx5/hws: add action object")
Cc: ere...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_action.c | 4
drivers/net/mlx5/hws/mlx
range definer creation")
Cc: va...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Itamar Gozlan
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 11 +++
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/drivers/net/mlx5/hws/mlx5dr_definer.c
b/drivers/ne
From: Erez Shitrit
When creating action from type MLX5DR_ACTION_TYP_REFORMAT_TNL_L3_TO_L2
we use modify-header object, we support few of that type at the same
time over this action depends on the number of headers.
Now when destroying the modify-header object we run over the
number_of_patterns, t
: fix NAT64 csum issue
net/mlx5/hws: fix NA64 copy TOS field instead of TTL
Itamar Gozlan (1):
net/mlx5/hws: strictly range templates check fix
drivers/net/mlx5/hws/mlx5dr_action.c | 213 ++
drivers/net/mlx5/hws/mlx5dr_action.h | 17 +-
drivers/net/mlx5/hws
From: Erez Shitrit
eswitch_owner_vhca_id_valid value should be set to 1 only on
merged-eswitch device.
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_action.c | 12
drivers/net/mlx5/hws/mlx5dr_cmd.c| 3 ++-
drivers/net/mlx5/hws/mlx5dr_cmd.h
From: Erez Shitrit
We don't have enough registers to copy TTL and TOS, so we will set TTL
to be the default value (64) and will copy TOS.
Fixes: 06d969a8c5b8 ("net/mlx5/hws: support NAT64 flow action")
Cc: ere...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
From: Erez Shitrit
Due to HW limitation we got two csum's that were not correct, udp and
ip, both of them were not calculated correctly by the HW.
By adding the next W/A we allow the HW to collect it well:
Separate the protocol field and zero all the addresses before fixed the
UDP csum.
We saw th
From: Erez Shitrit
Removing a redundant variable.
Was there from day 1, not in use.
Fixes: f8c8a6d8440d ("net/mlx5/hws: add action object")
Cc: ere...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_pat_arg.h | 1 -
1 file chan
From: Alex Vesker
When calling item convert function we need to pass the port_id
in the attributes. This value should be passed not only for cases
that match on PORT related items, to resolve we will always pass it.
Fixes: 572fe9ef2f46 ("net/mlx5/hws: fix port ID for root table")
Cc: ere...@nvid
From: Erez Shitrit
No more ignoring this action while destroying it.
Fixes: f8c8a6d8440d ("net/mlx5/hws: add action object")
Cc: ere...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_action.c | 4
drivers/net/mlx5/hws/mlx
range definer creation")
Cc: va...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Itamar Gozlan
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 11 +++
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/drivers/net/mlx5/hws/mlx5dr_definer.c
b/drivers/ne
From: Erez Shitrit
When creating action from type MLX5DR_ACTION_TYP_REFORMAT_TNL_L3_TO_L2
we use modify-header object, we support few of that type at the same
time over this action depends on the number of headers.
Now when destroying the modify-header object we run over the
number_of_patterns, t
From: Erez Shitrit
eswitch_owner_vhca_id_valid value should be set to 1 only on
merged-eswitch device.
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_action.c | 12
drivers/net/mlx5/hws/mlx5dr_cmd.c| 3 ++-
drivers/net/mlx5/hws/mlx5dr_cmd.h
From: Mahmoud Maatuq
This makes sure that the allocated matcher object is freed
for all branches that return NULL.
Coverity issue: 426424
Fixes: 27d171b88031 ("net/mlx5: abstract flow action and enable reconfigure")
Signed-off-by: Mahmoud Maatuq
Acked-by: Bing Zhao
Acked-by: Dariusz Sosnowski
From: Shani Peretz
Currently mlx5 PMD exposes rx_out_of_buffer counter that tracks
packets dropped when Rx queue was full.
To provide more granular statistics, this patch splits the
`rx_out_of_buffer` counter into two separate counters:
1. hairpin_out_of_buffer - This counter specifically tracks
.@dpdk.org
Signed-off-by: Itamar Gozlan
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_table.c | 11 ++-
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/drivers/net/mlx5/hws/mlx5dr_table.c
b/drivers/net/mlx5/hws/mlx5dr_table.c
index 55b9b20150..ab73017ade 100644
--- a/d
From: Hamdan Igbaria
The debug print of only_term param in the action template
was printed incorrectly.
Fixes: 78580cf4e796 ("net/mlx5/hws: add debug layer")
Cc: sta...@dpdk.org
Signed-off-by: Hamdan Igbaria
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_debug.c | 2 +-
1 file changed
From: Hamdan Igbaria
Dump action ste arrays info for resizing matchers.
Signed-off-by: Hamdan Igbaria
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_debug.c | 23 +++
drivers/net/mlx5/hws/mlx5dr_debug.h | 1 +
drivers/net/mlx5/hws/mlx5dr_matcher.c | 1 +
drive
From: Alex Vesker
In case decapL3 action is created we would access header
data even in case the SHARED flag is not set, this would
lead to an invalid memory access.
Fixes: 3a6c50215c07 ("net/mlx5/hws: support multi-pattern")
Cc: sta...@dpdk.org
Signed-off-by: Alex Vesker
Acked-by: Matan Azrad
From: Erez Shitrit
Save the exact tag when matching over jumbo masking.
Fixes: 348cdeec6472 ("net/mlx5/hws: add FW WQE rule creation logic")
Cc: sta...@dpdk.org
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_rule.c | 9 +++--
1 file changed, 7 insertion
From: Alex Vesker
Requesting the user for number of action resources is
not defined well enough, instead the user should provide
an initial action template which should contain the
future attached action templates.
Signed-off-by: Alex Vesker
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5
;)
Cc: sta...@dpdk.org
Signed-off-by: Itamar Gozlan
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 39 ++-
1 file changed, 20 insertions(+), 19 deletions(-)
diff --git a/drivers/net/mlx5/hws/mlx5dr_definer.c
b/drivers/net/mlx5/hws/mlx5dr_definer.c
index
From: Yevgeny Kliteynik
Simplify the code that does CQ poll - have DB record
update as part of CQ polling.
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_send.c | 9 +++--
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/drivers/net/mlx
From: Yevgeny Kliteynik
Instead of passing 0 as an enum parameter, define flag NONE.
This resolves the following code analysis error: "enumerated
type mixed with another type".
This value is currently used in tests only, and will later
be used in backward-compatible steering API.
Fixes: 5cadd74f
From: Yevgeny Kliteynik
Add missing spinlock destruction in error flow.
Fixes: b0290e56dd08 ("net/mlx5/hws: add context object")
Cc: sta...@dpdk.org
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_context.c | 1 +
1 file changed, 1 insertion(+)
diff --
From: Yevgeny Kliteynik
Small simplification to the send_queues_close code
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_send.c | 9 ++---
1 file changed, 2 insertions(+), 7 deletions(-)
diff --git a/drivers/net/mlx5/hws/mlx5dr_send.c
b/drivers/n
From: Yevgeny Kliteynik
mlx5dr_rule_create returns -ret, so internal functions should return
rte_errno on failure. Fix return values of the function that creates
rule in root table.
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_rule.c | 2 +-
1 file ch
From: Yevgeny Kliteynik
The following caps were never in use:
- wire_regc
- metadata_c
- metadata_c_mask
Also send ring's reg_addr field wasn't in use.
Removing these caps and fields from structs and removing
the code that queried them.
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
From: Yevgeny Kliteynik
Remove comment that documents parameter that doesn't exist.
Fixes: 3eb748869d2d ("net/mlx5/hws: add send layer")
Cc: sta...@dpdk.org
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_send.h | 2 --
1 file changed, 2 deletions(-)
d
From: Yevgeny Kliteynik
Fix wrong reserved size and add helpful comment
Fixes: 365cdf5f8ce7 ("net/mlx5/hws: add command layer")
Cc: sta...@dpdk.org
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/common/mlx5/mlx5_prm.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletio
From: Erez Shitrit
When the user tries to create a matcher and if failed with specific
errno (E2BIG) the message will be in debug level and not in warning.
It is a part of a feature when the user re-try to insert a new matching
depends on that errno, no need the annoying message.
Fixes: c55c2bf
From: Erez Shitrit
In root tables matcher and rule need to have their port-id, otherwise
the translate function that done in dpdk layer will not get the right
attributes.
For that whenever the matcher is matching the source-port we need to get
the relevant port-id before calling the translate fun
From: Erez Shitrit
Save the exact tag when matching over jumbo masking.
Fixes: 348cdeec6472 ("net/mlx5/hws: add FW WQE rule creation logic")
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_rule.c | 9 +++--
1 file changed, 7 insertions(+), 2 deletions(-)
From: Alex Vesker
The check to detect if a rule is in resize was done incorrectly,
this can lead to an incorrect function call upon completion for
rules which are not in resize (move).
Since the resize_info is in a union we cannot rely only on the
pointer value but also need to make sure the matc
From: Alex Vesker
Requesting the user for number of action resources is
not defined well enough, instead the user should provide
an initial action template which should contain the
future attached action templates.
Signed-off-by: Alex Vesker
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5
From: Yevgeny Kliteynik
Simplify the code that does CQ poll - have DB record
update as part of CQ polling.
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_send.c | 9 +++--
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/drivers/net/mlx
From: Yevgeny Kliteynik
Instead of passing 0 as an enum parameter, define flag NONE.
This resolves the following code analysis error: "enumerated
type mixed with another type".
This value is currently used in tests only, and will later
be used in backward-compatible steering API.
Fixes: 5cadd74f
From: Yevgeny Kliteynik
Small simplification to the send_queues_close code
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_send.c | 9 ++---
1 file changed, 2 insertions(+), 7 deletions(-)
diff --git a/drivers/net/mlx5/hws/mlx5dr_send.c
b/drivers/n
From: Yevgeny Kliteynik
Add missing spinlock destruction in error flow.
Fixes: b0290e56dd08 ("net/mlx5/hws: add context object")
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_context.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/net/mlx
From: Yevgeny Kliteynik
mlx5dr_rule_create returns -ret, so internal functions should return
rte_errno on failure. Fix return values of the function that creates
rule in root table.
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_rule.c | 2 +-
1 file ch
From: Yevgeny Kliteynik
The following caps were never in use:
- wire_regc
- metadata_c
- metadata_c_mask
Also send ring's reg_addr field wasn't in use.
Removing these caps and fields from structs and removing
the code that queried them.
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
From: Yevgeny Kliteynik
Remove comment that documents parameter that doesn't exist.
Fixes: 3eb748869d2d ("net/mlx5/hws: add send layer")
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_send.h | 2 --
1 file changed, 2 deletions(-)
diff --git a/drivers/
From: Yevgeny Kliteynik
Fix wrong reserved size and add helpful comment
Fixes: 365cdf5f8ce ("net/mlx5/hws: add command layer")
Signed-off-by: Yevgeny Kliteynik
Acked-by: Matan Azrad
---
drivers/common/mlx5/mlx5_prm.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/d
From: Erez Shitrit
When the user tries to create a matcher and if failed with specific
errno (E2BIG) the message will be in debug level and not in warning.
It is a part of a feature when the user re-try to insert a new matching
depends on that errno, no need the annoying message.
Fixes: c55c2bf
Fixing a typo in mlx5dr_cmd.c file, changing a variable name
from stc_parm to stc_param, as short for parameter.
Signed-off-by: Itamar Gozlan
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_cmd.c | 68 +++
1 file changed, 34 insertions(+), 34 deletions
From: Hamdan Igbaria
support push_esp flag for insert header action, it must be set when
inserting an ESP header, it's also sets the next_protocol field in
the IPsec trailer.
Signed-off-by: Hamdan Igbaria
Acked-by: Matan Azrad
---
drivers/common/mlx5/mlx5_prm.h | 3 ++-
drivers/net/mlx5
From: Hamdan Igbaria
Today if we pass a remove header action and after it an insert header
action then our action template builder will set two different
STE setters, because it won't allow insert header in same STE as
remove header.
But if we have the opposite order of insert header and then rem
From: Hamdan Igbaria
Only allow rule updates for rules with their status value equal to
MLX5DR_RULE_STATUS_CREATED.
Otherwise, the rule may be in an unstable stage like deleting and
this will result in a faulty unexpected scenario.
Signed-off-by: Hamdan Igbaria
Acked-by: Matan Azrad
---
drive
From: Hamdan Igbaria
Extend mlx5dr_action_template_create function params
to include flags parameter.
Signed-off-by: Hamdan Igbaria
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr.h | 10 +-
drivers/net/mlx5/hws/mlx5dr_action.c | 11 ++-
drivers/net/mlx5/hws/mlx
From: Yevgeny Kliteynik
Add support for matcher resize with the following new API calls:
- mlx5dr_matcher_resize_set_target
- mlx5dr_matcher_resize_rule_move
The first function links two matchers and allows moving rules from src
matcher to dst matcher. Both matchers should have the same charac
From: Hamdan Igbaria
If a VLAN item was passed with null mask, the item handler would
return immediately and thus won't set default values for non relax
mode.
Also change the non relax default set to single-tagged (CVLAN).
Fixes: c55c2bf35333 ("net/mlx5/hws: add definer layer")
Signed-off-by: Ha
inputs. By using precomputed optimal DW locations, we can change
the STE order for a limited set of the most common values to reduce the
number of hash collisions and improve latency.
Signed-off-by: Itamar Gozlan
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 64
The location of indexed rules is determined by the index, not the item
hash. A matcher test is added to prevent access to non-existent items.
This avoids unnecessary processing and potential segmentation faults.
Fixes: 405242c ("net/mlx5/hws: add rule object")
Signed-off-by: Itamar Goz
From: Erez Shitrit
Don't allow the user to mask over rsvd1 / rsvd2 fields which are not
supported.
Fixes: dbff89ef806f ("net/mlx5/hws: fix tunnel protocol checks")
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 7 +++
1 file changed, 7 insert
From: Hamdan Igbaria
support push_esp flag for insert header action, it must be set when
inserting an ESP header, it's also sets the next_protocol field in
the IPsec trailer.
Signed-off-by: Hamdan Igbaria
Acked-by: Matan Azrad
---
drivers/common/mlx5/mlx5_prm.h | 3 ++-
drivers/net/mlx5
From: Hamdan Igbaria
Today if we pass a remove header action and after it an insert header
action then our action template builder will set two different
STE setters, because it won't allow insert header in same STE as
remove header.
But if we have the opposite order of insert header and then rem
From: Hamdan Igbaria
Extend mlx5dr_action_template_create function params
to include flags parameter.
Signed-off-by: Hamdan Igbaria
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr.h | 10 +-
drivers/net/mlx5/hws/mlx5dr_action.c | 11 ++-
drivers/net/mlx5/hws/mlx
From: Hamdan Igbaria
If a VLAN item was passed with null mask, the item handler would
return immediately and thus won't set default values for non relax
mode.
Also change the non relax default set to single-tagged (CVLAN).
Fixes: c55c2bf35333 ("net/mlx5/hws: add definer layer")
Signed-off-by: Ha
From: Hamdan Igbaria
Only allow rule updates for rules with their status value equal to
MLX5DR_RULE_STATUS_CREATED.
Otherwise, the rule may be in an unstable stage like deleting and
this will result in a faulty unexpected scenario.
Signed-off-by: Hamdan Igbaria
Acked-by: Matan Azrad
---
drive
inputs. By using precomputed optimal DW locations, we can change
the STE order for a limited set of the most common values to reduce the
number of hash collisions and improve latency.
Signed-off-by: Itamar Gozlan
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 64
From: Yevgeny Kliteynik
Add support for matcher resize with the following new API calls:
- mlx5dr_matcher_resize_set_target
- mlx5dr_matcher_resize_rule_move
The first function links two matchers and allows moving rules from src
matcher to dst matcher. Both matchers should have the same charac
Don't allow the user to mask over rsvd1 / rsvd2 fields which are not
supported.
Fixes: dbff89ef806f ("net/mlx5/hws: fix tunnel protocol checks")
Signed-off-by: Erez Shitrit
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 7 +++
1 file changed, 7 insertions(+)
diff --git a
The location of indexed rules is determined by the index, not the item
hash. A matcher test is added to prevent access to non-existent items.
This avoids unnecessary processing and potential segmentation faults.
Fixes: 405242c ("net/mlx5/hws: add rule object")
Signed-off-by: Itamar Goz
target table.
Issue: 3441251
Signed-off-by: Itamar Gozlan
---
drivers/common/mlx5/mlx5_prm.h| 5 +-
drivers/net/mlx5/hws/mlx5dr.h | 12 ++
drivers/net/mlx5/hws/mlx5dr_cmd.c | 7 +-
drivers/net/mlx5/hws/mlx5dr_cmd.h | 1 +
drivers/net/mlx5/hws/mlx5dr_debug.c | 5
From: Alex Vesker
Fix an issue with binding for unused DW selectors,
these selectors are set to zero but zero is also
a valid value used for DMAC_47_16.
Fixes: c55c2bf35333 ("net/mlx5/hws: add definer layer")
Reviewed-by: Erez Shitrit
Signed-off-by: Alex Vesker
---
drivers/net/mlx5/hws/mlx5dr
RTE_FLOW_ITEM_TYPE_IB_BTH matches an InfiniBand base transport
header. We extend the match on the acknowledgment bit (BTH_A).
Signed-off-by: Itamar Gozlan
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 12 ++--
drivers/net/mlx5/hws/mlx5dr_definer.h | 1 +
2 files changed, 11 insertions
Supporting a new modify action: ADD_FIELD.
The new action allows the summing of the packet source field
and destination field/meta-data. The result is stored in the
destination field.
This new action is supported only on capable devices.
Signed-off-by: Itamar Gozlan
---
drivers/common/mlx5
New connectX devices have 4 additional registers which can be
used by the application. This support will allow matching on
these new registers.
Signed-off-by: Itamar Gozlan
---
drivers/common/mlx5/mlx5_prm.h| 4
drivers/net/mlx5/hws/mlx5dr_definer.c | 16
drivers
From: Alex Vesker
Add the support for default miss on HWS FDB, this implementation
was missing until now. Default miss can be used for more efficient
miss flow instead of going to an empty matcher or some defecated
empty table.
Signed-off-by: Alex Vesker
Acked-by: Matan Azrad
---
drivers/net/
An action naming change is made to describe the action better
and avoid implicit meaning.
Signed-off-by: Itamar Gozlan
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr.h| 2 +-
drivers/net/mlx5/hws/mlx5dr_action.c | 16
drivers/net/mlx5/hws/mlx5dr_debug.c | 2
From: Shun Hao
Currently there are two enum definitions for reformat type:
mlx5dr_action_reformat_type and mlx5dr_action_type. They are actually
the same. This patch remove the unnecessary mlx5dr_action_reformat_type
so use only one definition for reformat type.
Signed-off-by: Shun Hao
Acked-by
From: Alex Vesker
In case the send engine is in error state API should provide
a way to free rule resources and provide a fake completion.
Otherwise there is no way to release these resources.
Signed-off-by: Alex Vesker
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_rule.c | 10 +--
From: Alex Vesker
In some places an extra new line was added, remove to
have clean prints.
Signed-off-by: Alex Vesker
Acked-by: Matan Azrad
---
v1 -> v3
(1) amending a wrong subject prefix send (v1 instead of v3).
(2) typo fix (uneeded -> unneeded)
v2->v3
1. Right patches instead of wrong patc
From: Alex Vesker
Add the support for default miss on HWS FDB, this implementation
was missing until now. Default miss can be used for more efficient
miss flow instead of going to an empty matcher or some defecated
empty table.
Signed-off-by: Alex Vesker
Acked-by: Matan Azrad
---
drivers/net/
An action naming change is made to describe the action better
and avoid implicit meaning.
Signed-off-by: Itamar Gozlan
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr.h| 2 +-
drivers/net/mlx5/hws/mlx5dr_action.c | 16
drivers/net/mlx5/hws/mlx5dr_debug.c | 2
From: Shun Hao
Currently there are two enum definitions for reformat type:
mlx5dr_action_reformat_type and mlx5dr_action_type. They are actually
the same. This patch remove the unnecessary mlx5dr_action_reformat_type
so use only one definition for reformat type.
Signed-off-by: Shun Hao
Acked-by
From: Alex Vesker
In case the send engine is in error state API should provide
a way to free rule resources and provide a fake completion.
Otherwise there is no way to release these resources.
Signed-off-by: Alex Vesker
Acked-by: Matan Azrad
---
drivers/net/mlx5/hws/mlx5dr_rule.c | 10 +--
From: Alex Vesker
In some places an extra new line was added, remove to
have clean prints.
Signed-off-by: Alex Vesker
Acked-by: Matan Azrad
---
v2->v3
1. Right patches instead of wrong patches in the previous series
v1->v2
1. Last patch in the series (net/mlx5/hws: support default miss action
From: Alex Vesker
Add the support for default miss on HWS FDB, this implementation
was missing until now. Default miss can be used for more efficient
miss flow instead of going to an empty matcher or some defecated
empty table.
Signed-off-by: Alex Vesker
---
drivers/net/mlx5/hws/mlx5dr_action.
From: Xueming Li
This patch supports symmetric hash function that creating same
hash result for bi-direction traffic which having reverse
source and destination IP and L4 port.
Since the hash algorithom is different than spec(XOR), leave a
warning in validation.
Signed-off-by: Xueming Li
---
From: Rongwei Liu
Support the raw_encap/decap combinations in the indirect action
list, and translates to 4 types of underlayer tunnel operations:
1. Layer 2 encapsulation like VxLAN.
2. Layer 2 decapsulation like VxLAN.
3. Layer 3 encapsulation like GRE.
4. Layer 3 decapsulation like GRE.
Each
From: Gregory Etelson
Syntax:
actions .. indirect_list handle H conf C ..
H references shared action resources related to that handle.
C references configuration array:
if H was created from a list of actions A1 / A2 / ... / END,
C[i] points to a flow configuration update of A[i].
Signed-off-by
From: Gregory Etelson
Signed-off-by: Gregory Etelson
---
drivers/net/mlx5/mlx5_flow.c| 69 +-
drivers/net/mlx5/mlx5_flow.h| 67 -
drivers/net/mlx5/mlx5_flow_hw.c | 419 +++-
3 files changed, 476 insertions(+), 79 deletions(-)
diff --git a/drivers/n
From: Alex Vesker
Add the support for default miss on HWS FDB, this implementation
was missing until now. Default miss can be used for more efficient
miss flow instead of going to an empty matcher or some defecated
empty table.
Signed-off-by: Alex Vesker
---
drivers/net/mlx5/hws/mlx5dr_table.c
An action naming change is made to describe the action better
and avoid implicit meaning.
Signed-off-by: Itamar Gozlan
---
drivers/net/mlx5/hws/mlx5dr.h| 2 +-
drivers/net/mlx5/hws/mlx5dr_action.c | 16
drivers/net/mlx5/hws/mlx5dr_debug.c | 2 +-
drivers/net/mlx5
From: Shun Hao
Currently there are two enum definitions for reformat type:
mlx5dr_action_reformat_type and mlx5dr_action_type. They are actually
the same. This patch remove the unnecessary mlx5dr_action_reformat_type
so use only one definition for reformat type.
Signed-off-by: Shun Hao
---
dri
From: Alex Vesker
In case the send engine is in error state API should provide
a way to free rule resources and provide a fake completion.
Otherwise there is no way to release these resources.
Signed-off-by: Alex Vesker
---
drivers/net/mlx5/hws/mlx5dr_rule.c | 10 +-
1 file changed, 5
From: Alex Vesker
In some places an extra new line was added, remove to
have clean prints.
Signed-off-by: Alex Vesker
---
drivers/net/mlx5/hws/mlx5dr_action.c | 30 +--
drivers/net/mlx5/hws/mlx5dr_cmd.c | 2 +-
drivers/net/mlx5/hws/mlx5dr_definer.c | 4 ++--
driv
96 matches
Mail list logo
