Commit Graph

26 Commits

Author SHA1 Message Date
Kevin Wang
5516381d7e
[qos] change the template keyword from Compute-AI to ComputeAI (#17902)
Why I did it
Align the keywords to make qos configuration take effect

Work item tracking
Microsoft ADO (number only):
How I did it
Change the keyword to ComputeAI

How to verify it
reload minigraph and check the qos configuration
2024-01-29 10:10:54 +08:00
Alpesh Patel
cabdac17a5
qos template change for backend compute-ai deployment (#16150)
#### Why I did it

To enable qos config for a certain backend deployment mode, for resource-type "Compute-AI".
This deployment has the following requirement:

- Config below enabled if DEVICE_TYPE as one of backend_device_types
- Config below enabled if ResourceType is 'Compute-AI'
- 2 lossless TCs' (2, 3)
- 2 lossy TCs' (0,1)
- DSCP to TC map uses 4 DSCP code points and maps to the TCs' as follows:
   "DSCP_TO_TC_MAP": {
        "AZURE": {
             "48" : "0",
            "46" : "1",
            "3"  : "3",
            "4"  : "4"
        }
    }

- WRED profile has green {min/max/mark%} as {2M/10M/5%}

This required template change <as in the PR> in addition to the vendor qos.json.j2 file (not included here).

### How I did it

#### How to verify it
- with the above change and the vendor config change, generated the qos.json file and verified that the objective stated in "Why I did it" was met

- verified no error

### Description for the changelog
Update qos_config.j2 for Comptue-AI deployment on one of backend device type roles
2023-08-31 11:30:20 -07:00
vmittal-msft
fea10546f2
Update WRED profile on system ports (#15612)
* Update WRED profile on system ports
2023-07-19 15:00:39 -07:00
Alpesh Patel
633fff8c10
enable ethernet backplane port support in port config for packet mode T2 devices (#14533)
For T2 systems using packet mode, the backplane interfaces (Ethernet-BP#) and the fabric card ethernet interfaces are not visible as neighbor interfaces.
In packet mode, these interfaces needs qos and buffer config as well.
This fix addresses that issue and adds the backplane interfaces to the PORTS_ACTIVE list
2023-06-12 14:02:22 -07:00
bingwang-ms
f402e6b5c6
Apply separated DSCP_TO_TC_MAP and TC_TO_QUEUE_MAP to uplink ports on dualtor (#12730)
Why I did it
The PR is to apply separated DSCP_TO_TC_MAP and TC_TO_QUEUE_MAP to uplink ports on dualtor.
The traffic with DSCP 2 and DSCP 6 from T1 is treated as lossless traffic.

DSCP    TC    Queue
2      2     2
6      6     6
Traffic with DSCP 2 or DSCP 6 from downlink is still treated as lossy traffic as before.

How I did it
Define DSCP_TO_TC_MAP|AZURE_UPLINK and TC_TO_QUEUE_MAP|AZURE_UPLINK.

How to verify it
Verified by UT
Verified by coping the new template to a testbed, and rendering a config_db.json
2022-11-21 11:42:28 -08:00
bingwang-ms
6169ae3ee3
Add lossy scheduler for queue 7 (#12596)
* Add lossy scheduler for queue 7
2022-11-04 08:12:00 +08:00
Stephen Sun
8f4a1b7b85
[Mellanox] Support Mellanox-SN4600C-C64 as T1 switch in dual-ToR scenario (#11261)
- Why I did it
Support Mellanox-SN4600C-C64 as T1 switch in dual-ToR scenario
This is to port #11032 and #11299 from 202012 to master.

Support additional queue and PG in buffer templates, including both traditional and dynamic model
Support mapping DSCP 2/6 to lossless traffic in the QoS template.
Add macros to generate additional lossless PG in the dynamic model
Adjust the order in which the generic/dedicated (with additional lossless queues) macros are checked and called to generate buffer tables in common template buffers_config.j2
Buffer tables are rendered via using macros.
Both generic and dedicated macros are defined on our platform. Currently, the generic one is called as long as it is defined, which causes the generic one always being called on our platform. To avoid it, the dedicated macrio is checked and called first and then the generic ones.
Support MAP_PFC_PRIORITY_TO_PRIORITY_GROUP on ports with additional lossless queues.
On Mellanox-SN4600C-C64, buffer configuration for t1 is calculated as:

40 * 100G downlink ports with 4 lossless PGs/queues, 1 lossy PG, and 3 lossy queues
16 * 100G uplink ports with 2 lossless PGs/queues, 1 lossy PG, and 5 lossy queues

Signed-off-by: Stephen Sun <stephens@nvidia.com>
2022-07-20 09:48:15 +03:00
bingwang-ms
52c36cd31f
Add flag to control the generation of PORT_QOS_MAP|global entry (#11448)
Why I did it
This PR is to add a flag to control whether to generate PORT_QOS_MAP|global entry or not.
It's because for some HWSKU, such as BackEndToRRouter and BackEndLeafRouter, there is no DSCP_TO_TC_MAP defined.
Hence, if the PORT_QOS_MAP|global entry is generated, OA will report some error because the DSCP_TO_TC_MAP map AZURE can not be found.

Jul 14 00:24:40.286767 str2-7050qx-32s-acs-03 ERR swss#orchagent: :- saiObjectTypeQuery: invalid object id oid:0x7fddb43605d0
Jul 14 00:24:40.286767 str2-7050qx-32s-acs-03 ERR swss#orchagent: :- meta_generic_validation_objlist: SAI_SWITCH_ATTR_QOS_DSCP_TO_TC_MAP:SAI_ATTR_VALUE_TYPE_OBJECT_ID object on list [0] oid 0x7fddb43605d0 is not valid, returned null object id
Jul 14 00:24:40.286767 str2-7050qx-32s-acs-03 ERR swss#orchagent: :- applyDscpToTcMapToSwitch: Failed to apply DSCP_TO_TC QoS map to switch rv:-5
Jul 14 00:24:40.286767 str2-7050qx-32s-acs-03 ERR swss#orchagent: :- doTask: Failed to process QOS task, drop it
This PR is to address the issue.

How I did it
Add a flag require_global_dscp_to_tc_map to control whether to generate the PORT_QOS_MAP|global entry. The default value for require_global_dscp_to_tc_map is true. If the device type is storage backend, the value is changed to false. Then the PORT_QOS_MAP|global entry is not generated.

How to verify it
Update the current test_qos_dscp_remapping_render_template to cover storage backend.
2022-07-14 10:19:00 -07:00
bingwang-ms
83f23e26ff
Generate switch level dscp_to_tc_map entry from qos_config template (#11087)
* Generate switch level dscp_to_tc_map

Signed-off-by: bingwang <wang.bing@microsoft.com>
2022-06-17 08:44:30 +08:00
bingwang-ms
0c9bbee735
Update qos template to support SYSTEM_DEFAULT table (#10936)
* Update qos template to support SYSTEM_DEFAULT table

Signed-off-by: bingwang <wang.bing@microsoft.com>
2022-06-02 21:48:57 +08:00
bingwang-ms
3fc3259a35
Define qos map AZURE_TUNNEL for QoS remapping of tunnel traffic (#10565)
* Add AZURE_TUNNEL map

Signed-off-by: bingwang <wang.bing@microsoft.com>
2022-04-25 15:06:10 +08:00
bingwang-ms
b9dd1df372
Update qos config to clear queues for bounced back traffic (#10176)
* Update qos config to clear queues for bounced back traffic

Signed-off-by: bingwang <bingwang@microsoft.com>
2022-04-05 22:32:25 +08:00
Song Yuan
01798447ab
[Chassis][QoS template] Skip configuring buffer and QoS config on recirc ports (#7869)
* Added test case to verify the template changes.
2022-03-09 16:04:36 -08:00
vmittal-msft
bc1dfea619
Updated traffic scheduler settings for HWSKUs : DellEMC-Z9332f-O32 and DellEMC-Z9332f-M-O16C64 (#9828) 2022-02-23 17:22:41 -08:00
Ashok Daparthi-Dell
6cbdf11e53
SONIC QOS YANG - Remove qos tables field value refernce format (#7752)
Depends on Azure/sonic-utilities#1626
Depends on Azure/sonic-swss#1754

QOS tables in config db used ABNF format i.e "[TABLE_NAME|name] to refer fieldvalue to other qos tables.

Example:
Config DB:
"Ethernet92|3": {
"scheduler": "[SCHEDULER|scheduler.1]",
"wred_profile": "[WRED_PROFILE|AZURE_LOSSLESS]"
},
"Ethernet0|0": {
"profile": "[BUFFER_PROFILE|ingress_lossy_profile]"
},
"Ethernet0": {
"dscp_to_tc_map": "[DSCP_TO_TC_MAP|AZURE]",
"pfc_enable": "3,4",
"pfc_to_queue_map": "[MAP_PFC_PRIORITY_TO_QUEUE|AZURE]",
"tc_to_pg_map": "[TC_TO_PRIORITY_GROUP_MAP|AZURE]",
"tc_to_queue_map": "[TC_TO_QUEUE_MAP|AZURE]"
},

This format is not consistent with other DB schema followed in sonic.
And also this reference in DB is not required, This is taken care by YANG "leafref".

Removed this format from all platform files to consistent with other sonic db schema.
Example:
"Ethernet92|3": {
"scheduler": "scheduler.1",
"wred_profile": "AZURE_LOSSLESS"
},

Dependent pull requests:
#7752 - To modify platfrom files
#7281 - Yang model
Azure/sonic-utilities#1626 - DB migration
Azure/sonic-swss#1754 - swss change to remove ABNF format
2021-09-28 09:21:24 -07:00
Neetha John
3b06f44555
[qos]: modify dot1p to tc mapping (#7661)
Map priority 0 to TC 1 and priority 1 to TC 0

Send traffic on priority 0 and 1 and verified that it gets mapped correctly in hw

Signed-off-by: Neetha John <nejo@microsoft.com>
2021-05-20 10:36:39 -07:00
Lawrence Lee
d0f16c0d79
Make backend device checking more robust (#5730)
Treat devices that are ToRRouters (ToRRouters and BackEndToRRouters) the same when rendering templates
 Except for BackEndToRRouters belonging to a storage cluster, since these devices have extra sub-interfaces created
Treat devices that are LeafRouters (LeafRouters and BackEndLeafRouters) the same when rendering templates

Signed-off-by: Lawrence Lee <lawlee@microsoft.com>
2020-11-10 15:06:35 -08:00
Neetha John
286aa35ac6
[qos]: Alpha and ECN settings change for Th (#4564)
Dynamic threshold setting changed to 0 and WRED profile green min threshold set to 250000 for Tomahawk devices

Changed the dynamic threshold settings in pg_profile_lookup.ini
Added a macro for WRED profiles in qos.json.j2 for Tomahawk devices
Necessary changes made in qos.config.j2 to use the macro if present

Signed-off-by: Neetha John <nejo@microsoft.com>
2020-05-09 11:21:18 -07:00
Wenda Ni
cf0465bf53
Adopt per-port buffer and qos profile (#3542)
Signed-off-by: Wenda Ni <wenni@microsoft.com>
2019-10-02 13:01:16 -07:00
Wenda Ni
81aef6b64c [Qos] use dot1p to tc mapping for backend switches (#3422)
* Use dot1p to tc mapping for backend switches

Signed-off-by: Wenda Ni <wenni@microsoft.com>

* Do not write DSCP to TC mapping into CONFIG_DB or config_db.json for
storage switches

Signed-off-by: Wenda Ni <wenni@microsoft.com>
2019-09-13 11:28:25 -07:00
lguohan
a1158c6c18
Revert "Use dot1p to tc mapping for backend switches (#3412)" (#3421)
This reverts commit ca43dad12f.
2019-09-09 14:44:46 -07:00
Wenda Ni
ca43dad12f Use dot1p to tc mapping for backend switches (#3412)
* Use dot1p to tc mapping for backend switches

Signed-off-by: Wenda Ni <wenni@microsoft.com>

* Do not write DSCP to TC mapping into CONFIG_DB or config_db.json for
storage switches

Signed-off-by: Wenda Ni <wenni@microsoft.com>
2019-09-06 11:59:47 -07:00
Wenda Ni
f9c9fa8ba1 [qos]: Map tc 1, 2, 5, and 6 back to pg 0 (#2650)
Lossy traffic does not need to be mapped to different ingress PGs. They can all share the same ingress PG.

Signed-off-by: Wenda Ni <wenni@microsoft.com>
2019-03-08 02:23:32 -08:00
Wenda Ni
ce9a3f0c5a [QoS]: QoS Config change for multiple devices (#2505)
* QoS config change: 1) DSCP mapping; 2) link pg/queue 6 to lossy buffer;
3) redistribute scheduler

Signed-off-by: Wenda <wenni@microsoft.com>

* Add scheduling weight to queue 2

Signed-off-by: Wenda <wenni@microsoft.com>

* Link pg/queue 2 to lossy buffer

Signed-off-by: Wenda <wenni@microsoft.com>

* Update the pg headroom for a7060-D48C8 50G

Signed-off-by: Wenda <wenni@microsoft.com>

* Update config gen test for qos

Signed-off-by: Wenda <wenni@microsoft.com>

* Update pg headroom size, and update egress lossy pool size accordingly

Signed-off-by: Wenda <wenni@microsoft.com>

* Update headroom pool size; Update ingress service pool and egress lossy
pool sizes accordingly;

Signed-off-by: Wenda <wenni@microsoft.com>

* a7260: update headroom pool size; Update ingress service pool and egress lossy pool sizes accordingly;

Signed-off-by: Wenda <wenni@microsoft.com>

* Update config gen test for buffer

Signed-off-by: Wenda <wenni@microsoft.com>
2019-01-30 19:00:13 -08:00
Wenda Ni
f5e678cf84 Port QoS & buffer changes in 0330 to master (#2239)
* 1) DSCP 46 to 5; 2) ecn config for lossless traffic; 3) ecn on by default; 4) DWRR equal weight;

Signed-off-by: Wenda <wenni@microsoft.com>

* 1) link pg & queue 5 to lossy buffer profile; 2) ingress lossless alpha 1/8

Signed-off-by: Wenda <wenni@microsoft.com>

* Update the test case for qos & buffer json template

Signed-off-by: Wenda <wenni@microsoft.com>

* Migrate a7050-qx32 and s6000 to use pg_profile lookup architecture

Signed-off-by: Wenda <wenni@microsoft.com>

* Update pg headroom egress service pool for a7050-qx-32s, a7050-qx32, and s6000

Signed-off-by: Wenda <wenni@microsoft.com>

* Link queue 5 to lossy profile

Signed-off-by: Wenda <wenni@microsoft.com>
2018-12-04 20:51:55 -08:00
Wenda Ni
77652c55fd [QoS]: Unify qos json by using qos_config.j2 template (#2023)
* Unify qos config with qos_config.j2 template

Signed-off-by: Wenda <wenni@microsoft.com>

* Change 7050 to use qos config template

Signed-off-by: Wenda <wenni@microsoft.com>

	modified:   device/arista/x86_64-arista_7050_qx32/Arista-7050-QX32/qos.json.j2
	modified:   device/arista/x86_64-arista_7050_qx32s/Arista-7050-QX-32S/qos.json.j2

* Change a7060, a7260, s6000, s6100, z9100  to use qos config template

Signed-off-by: Wenda <wenni@microsoft.com>

* Change mlnx devices to use qos config template

Signed-off-by: Wenda <wenni@microsoft.com>

	modified:   ../../../mellanox/x86_64-mlnx_msn2100-r0/ACS-MSN2100/qos.json.j2
	modified:   ../../../mellanox/x86_64-mlnx_msn2410-r0/ACS-MSN2410/qos.json.j2
	modified:   ../../../mellanox/x86_64-mlnx_msn2700-r0/ACS-MSN2700/qos.json.j2
	modified:   ../../../mellanox/x86_64-mlnx_msn2700-r0/Mellanox-SN2700-D48C8/qos.json.j2

* Change barefoot devices to use qos config template

Signed-off-by: Wenda <wenni@microsoft.com>

	modified:   barefoot/x86_64-accton_wedge100bf_32x-r0/montara/qos.json.j2
	modified:   barefoot/x86_64-accton_wedge100bf_65x-r0/mavericks/qos.json.j2

* Change accton as7212 to use qos config template

Signed-off-by: Wenda <wenni@microsoft.com>

	modified:   accton/x86_64-accton_as7212_54x-r0/AS7212-54x/qos.json.j2

* Apply PORT_QOS_MAP to active ports only

Signed-off-by: Wenda <wenni@microsoft.com>

* Update qos config test with qos_config.j2 template

Signed-off-by: Wenda <wenni@microsoft.com>

* Update sample output of qos-dell6100.json

Signed-off-by: Wenda <wenni@microsoft.com>

* Remove generating the default port name and index list, i.e., remove the generate_port_lists macro, because PORT is always defined

Signed-off-by: Wenda <wenni@microsoft.com>

* Include pfc_to_pg_map according to platform asic type obtained from
/etc/sonic/sonic_version.yml rather than specifying per hwsku

Signed-off-by: Wenda Ni <wenni@microsoft.com>

* Customize TC_TO_PRIORITY_GROUP_MAP and
PFC_PRIORITY_TO_PRIORITY_GROUP_MAP for barefoot

Signed-off-by: Wenda <wenni@microsoft.com>

* Unify PFC_PRIORITY_TO_PRIORITY_GROUP_MAP: remove "0":"0", "1":"1" as
these two pgs do not generate PFC frames.

Signed-off-by: Wenda <wenni@microsoft.com>
2018-10-17 14:10:34 -07:00