Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Port open PRs from upstream for edge-core platforms: AS4630-54TE, AS5835-54X, AS7326-56X, and AS7726-32X #3

Closed
wants to merge 1,116 commits into from

Conversation

matofeder
Copy link
Member

This PR sync port open PRs from upstream for edge-core platforms: AS4630-54TE, AS5835-54X, AS7326-56X, and AS7726-32X.

Inspired by the edge-core PR [Sync PRs and Fix bugs] For platforms: AS4630-54TE, AS5835-54X, AS7326-56X, AS7726-32X, AS9716-32D, and AS9736-64D and investigation described within SovereignCloudStack/issues#706

List of ported PRs:

mssonicbld and others added 30 commits July 31, 2024 19:00
…atically (#19742)

#### Why I did it
src/sonic-utilities
```
* ff2c73f8 - (HEAD -> master, origin/master, origin/HEAD) Add namespace check for multiasic (#3458) (4 hours ago) [Xincun Li]
```
#### How I did it
#### How to verify it
#### Description for the changelog
…s (#18978)

[Mellanox] Avoid attaching lossless buffer profiles for internal ports (#18978)

Signed-off-by: Vivek Reddy <vkarri@nvidia.com>
…rc Port. (#19755)

[VOQ] Fix the parsing of core_port_id and core_id for Inband and Recirc Port. 
Signed-off-by: Abhishek Dosi <abdosi@microsoft.com>
… automatically (#19770)

#### Why I did it
src/sonic-platform-common
```
* 7f07fa6 - (HEAD -> master, origin/master, origin/HEAD) Parse all VDM advertised fields during DOM monitoring (#486) (4 hours ago) [mihirpat1]
```
#### How I did it
#### How to verify it
#### Description for the changelog
…is (#19653)

Updated lossy PG threshold to be higher than Queue threshold not to allow PFC generation.
…atically (#19803)

#### Why I did it
src/sonic-utilities
```
* f50587a1 - (HEAD -> master, origin/master, origin/HEAD) Update README.md (#3406) (5 hours ago) [Changrong Wu]
```
#### How I did it
#### How to verify it
#### Description for the changelog
Why I did it
Enable pipe mode in TH5 flex counter block to support queue counters for all the ports.

Microsoft ADO (number only): 28949528
Why I did it
Latest changes to sflowtool source mandates autoconf 2.71+ to build it. Bullseye is at autoconf 2.69. This PR builds autoconf from source in the docker-ptf image.

How I did it
Modified dockers/docker-ptf/Dockerfile.j2 to -

Install dependencies to build autoconf
Build and install autoconf

How to verify it
docker-ptf image is built successfully.
…9107)

Why I did it
Fix #19058

How I did it
Marked systemd-networkd.service as /dev/null to disable it by default.
systemd-sonic-generator will remove this null link if the target platform is smartswitch.

How to verify it
According to the issue page, we shouldn't get the error logERR systemd-networkd-wait-online[50532]: Timeout occurred while waiting for network connectivity INFO apt-helper[11911]: E: Sub-process /lib/systemd/systemd-networkd-wait-online returned an error code (1) after systemctl start apt-daily.service
Regress test the previous verification: [systemd/systemd-sonic-generator]: Systemd midplane network service of Smart Switch #18178

Signed-off-by: Ze Gan <ganze718@gmail.com>
…base (#19601)

Fix #19331

How I did it
Skip the namespace setting if the dev is DPU, because the DPU databases are only in the host namespace

How to verify it
Check there is no error log after fast-reboot

admin@vlab-01:~$ sudo cat /var/log/syslog  | grep -i asicdpu

Signed-off-by: Ze Gan <zegan@microsoft.com>
…9681)

A new SKUs for MSN4700 Platform: Mellanox-SN4700-O32 and Mellanox-SN4700-V64

Requirements for Mellanox-SN4700-O32:

8 x 400Gbps uplink to T2 switch (O13 to O20)
24 x 400Gbps downlinks to T0 switch (O1-O12, O21-O32)
Breakout mode No breakout mode. All ports working in 400Gb mode. .
FEC mode: RS
Type of transceiver: 400Gb Optical.
warm boot should be supported “No for T1 role”
VxLAN source port range set N/A
Static Policy Based Hashing supported N/A
Cable length “T0-T1 40m default, 300m max; T1-T2 2000m”
Tradition buffer model is must “Yes”
Shared headroom should be supported “Yes”
Over-subscription ratio: “2”.
Requirements for Mellanox-SN4700-V64

16 x 200Gbps uplink to T1 switch (V-25&V26 to V-39&40)
48 x 200Gbps downlinks to servers (Left panel downlink ports: V-1&2 to V-23&24; Right panel downlink ports: V-41&42 to V-63&64)
Breakout mode split from 400Gbps ports (2x200)
FEC mode: RS
Type of transceiver: 200Gb AOC between T0 and T1; 200Gb DAC between T0 and host.
warm boot should be supported “Yes for T0 role”
VxLAN source port range set N/A
Static Policy Based Hashing supported N/A
Cable length “T0-T1 40m default, 300m max, T0-Server 5m”
Tradition buffer model is must “Yes”
Shared headroom should be supported “Yes”
Over-subscription ratio: “2”.
Additional Details:

QoS configs for Mellanox-SN4700-V64 updated in order to fulfill Dual-ToR buffer (+DSCP remapping) requirements
Support for independent module added for both SKUs, so Auto-negotiation changed to NO

Signed-off-by: Andriy Yurkiv <ayurkiv@nvidia.com>
[SmartSwitch] DPU Management Traffic Forwarding script
…rns (#19717)

Added the below patches which are part of BGP Zebra back pressure feature required to keep the memory usage in check during route churns

How I did it
New patches that were added:

Patch	FRR Pull request
0030-zebra-backpressure-Zebra-push-back-on-Buffer-Stream-.patch	FRRouting/frr#15411
0031-bgpd-backpressure-Add-a-typesafe-list-for-Zebra-Anno.patch	FRRouting/frr#15524
0032-bgpd-fix-flushing-ipv6-flowspec-entries-when-peering.patch	FRRouting/frr#15326
0033-bgpd-backpressure-cleanup-bgp_zebra_XX-func-args.patch	FRRouting/frr#15524
0034-gpd-backpressure-Handle-BGP-Zebra-Install-evt-Creat.patch	FRRouting/frr#15524
0035-bgpd-backpressure-Handle-BGP-Zebra-EPVN-Install-evt-.patch	FRRouting/frr#15624
0036-zebra-backpressure-Fix-Null-ptr-access-Coverity-Issu.patch	FRRouting/frr#15728
0037-bgpd-Increase-install-uninstall-speed-of-evpn-vpn-vn.patch	FRRouting/frr#15727
0038-zebra-Actually-display-I-O-buffer-sizes.patch	FRRouting/frr#15708
0039-zebra-Actually-display-I-O-buffer-sizes-part-2.patch	FRRouting/frr#15769
0040-bgpd-backpressure-Fix-to-withdraw-evpn-type-5-routes.patch	FRRouting/frr#16034
0041-bgpd-backpressure-Fix-to-avoid-CPU-hog.patch	FRRouting/frr#16035
0042-zebra-Use-built-in-data-structure-counter.patch	FRRouting/frr#16221
0043-zebra-Use-the-ctx-queue-counters.patch	FRRouting/frr#16220
0044-zebra-Modify-dplane-loop-to-allow-backpressure-to-fi.patch	FRRouting/frr#16220
0045-zebra-Limit-queue-depth-in-dplane_fpm_nl.patch	FRRouting/frr#16220
0046-zebra-Modify-show-zebra-dplane-providers-to-give-mor.patch	FRRouting/frr#16220
0047-bgpd-backpressure-fix-evpn-route-sync-to-zebra.patch	FRRouting/frr#16234
0048-bgpd-backpressure-fix-to-properly-remove-dest-for-bg.patch	FRRouting/frr#16368
0049-bgpd-backpressure-Improve-debuggability.patch	FRRouting/frr#16368
0050-bgpd-backpressure-Avoid-use-after-free.patch	FRRouting/frr#16437
0051-bgpd-backpressure-fix-ret-value-evpn_route_select_in.patch	FRRouting/frr#16416
0052-bgpd-backpressure-log-error-for-evpn-when-route-inst.patch	FRRouting/frr#16416
…lly (#19831)

#### Why I did it
src/sonic-swss
```
* 465391d1 - (HEAD -> master, origin/master, origin/HEAD) Reduce log verbosity when the capability is not implemented (#3240) (6 hours ago) [Vivek]
```
#### How I did it
#### How to verify it
#### Description for the changelog
Update max queues in BCM SAI for TH5 in alignment with the flex counter blocks

Update XGS SAI version to 10.1.37.0

Microsoft ADO (number only): 28967795
…omatically (#19845)

#### Why I did it
src/sonic-mgmt-common
```
* 966adc0 - (HEAD -> master, origin/master, origin/HEAD) Fix sonic-net/sonic-gnmi#266 (#143) (4 hours ago) [Steve Licking]
```
#### How I did it
#### How to verify it
#### Description for the changelog
…atically (#19832)

#### Why I did it
src/sonic-utilities
```
* 317e6495 - (HEAD -> master, origin/master, origin/HEAD) Fix kexec_unload failure on secure boot enabled platforms (#3439) (34 hours ago) [Vivek]
* 557d6886 - [Mellanox] Add support for Mellanox-SN4700-O32 and Mellanox-SN4700-V64 (#3450) (34 hours ago) [Andriy Yurkiv]
* 018eb737 - Fix to use IPv6 linklocal address as snmp agent address (#3215) (35 hours ago) [SuvarnaMeenakshi]
```
#### How I did it
#### How to verify it
#### Description for the changelog
* Added blkinfo module to host for use by ssdutil

* Added blkinfo and psutil (as needed) to the sonic slave dockerfile
* [dhcp_server] Fix dhcp_serverer mk file mistake

* Set include_dhcp_server to y

* Revert "Set include_dhcp_server to y"

This reverts commit 88f03562745700db32838b038dfe2f10c5e88378.
* Update OpenSSH to 9.2p1-2+deb12u3

This brings in CVE/security fixes.

Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>
* Pin version of sflowtool used for the mixed PTF container

Newer versions of sflowtool require autoconf 2.71, which is not
available in Buster.

For the py3-only version of the PTF container, this is currently based
on Bullseye, which should be fine.

Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>

* Remove code to build autoconf, and use 6.04 for py3 only PTF as well

Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>

* Re-add autoconf to list of packages to install

Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>

---------

Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>
* Update disable-non-manylinux.patch due to pip being updated to 24.1.2.

The code logic in this area has changed, resulting in patch conflicts.

Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>

* Update the pinned version for pip to explicitly point to 24.1.2

This is needed because the patch file that has been updated can only be
applied to 24.1.2. However, the version pinning for 24.0 will mean that
PR checkers and the daily build will try to continue using 24.0, and
thus fail to apply the patch.

Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>

* Revert changes to version files, and pin pip to 24.2

Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>

---------

Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>
Why I did it
DellEMC Z9432F platform onboarding.

How I did it
Config changes with respect to Z9432F SAI bring up.
Also changes from swss is required for complete bring up of Z9432F platform(TD4) will be followed thru (sonic-net/sonic-swss#3244)

How to verify it
The following error message is not seen..
sonic CRIT syncd#syncd: [none] SAI_API_SWITCH:brcm_sai_create_switch:3204 setting inter-frame gap failed with error Feature unavailable (0xfffffff0).
Why I did it
Get the following error due to switch_health_dta size changing

ASSERT TRUE FAILED(5085): sizeof(sai_switch_health_data_t) == (SAI_SWITCH_HEALTH_DATA_T_SIZE): wrong size of sai_switch_health_data_t, expected 16, got 4
Work item tracking
Microsoft ADO (number only): 29062472
How I did it
This PR is just an adhoc fix by disabling saimetadatatest. We should fix the scripts under OCP/SAI/meta

How to verify it
Check Azp
#### Why I did it

SysLogger does not support NOTICE level log. The PR is to support it.

#### How I did it

Add a new log level NOTICE to python logging package.

#### How to verify it

Manual test

```
Jun  3 03:09:44.353536 sonic NOTICE test[364076]: This is a notice
```
#### Why I did it

After upgrading from bullseye to bookworm, pmon container no longer default installs tool xxd. The tool is used in mellanox simx platforms. The PR is to install xxd tool in the docker build file.

#### How I did it

Install xxd in platform monitor docker file.

#### How to verify it

Manual test
…g service restart (#19583)

#### Why I did it
The `resolvconfig` service updates the DNS configuration in the host OS and each running Docker container when a configuration change is detected. The DNS configuration update during the shutdown of the management interface is redundant, as the management interface is going down temporarily and, when it is back online, the DNS configuration will remain the same. The update of the DNS configuration adds a couple of seconds (depending on the CPU) to the interface-config service restart time when the management interface uses a dynamic IP address. This can affect the fast boot. To optimize the flow and execute the service restart faster, the update should be skipped.

#### How I did it
Do not run DNS configuration update during the shutdown of the management interface.

#### How to verify it
Measure the execution time of the `service interfaces-config restart` command on the device with the static IP address configuration on the management interface and compare it to the execution time of the same command with the dynamic IP address. The difference should be insignificant.
…#19900)

Why I did it
The latest swss and dash-api commits are dependent on each other to build properly, if updated separately they won't build.

sonic-swss new commits:

* 29cea04a - (HEAD -> master, origin/master, origin/HEAD) [DASH] Remove deprecated 'action_type' refs (#3257) (3 hours ago) [Lawrence Lee]
sonic-dash-api new commits:

* dbb2d0f - (HEAD -> master, origin/master, origin/HEAD) Update for SAI API changes (3 days ago) [Lawrence Lee]
* 7cd1105 - deprecate pl_sip_encoding (6 days ago) [Lawrence Lee]
* 0df30ee - Merge branch 'master' of github.com:sonic-net/sonic-dash-api into route-group (6 days ago) [Lawrence Lee]
* 2721493 - deprecate privatelinknsg routing type (6 days ago) [Lawrence Lee]
* 3260f6a - update unit tests (6 days ago) [Lawrence Lee]
* 95fa234 - avoid breaking change (6 days ago) [Lawrence Lee]
* ab46204 - EOF newline (6 days ago) [Lawrence Lee]
* 2f58861 - add icmp redirection (6 days ago) [Lawrence Lee]
* 6e9ad22 - non-breaking route metering update (6 days ago) [Lawrence Lee]
* 722b6c2 - Rename action_type to routing type (6 days ago) [Lawrence Lee]
* 76ab87a - add newline to EOF (6 days ago) [Lawrence Lee]
* 747425a - Keep pl_sip_encoding (6 days ago) [Lawrence Lee]
*   597c4b8 - Merge branch 'route-group' of github.com:theasianpianist/sonic-dash-api into route-group (4 months ago) [Lawrence Lee]
|\  
| * ecec7fb - Merge branch 'master' into route-group (5 months ago) [Lawrence Lee]
* | 8b3de2d - update API to match new HLD schema (4 months ago) [Lawrence Lee]
|/  
* daa54b6 - add new message types to unittest (5 months ago) [Lawrence Lee]
* 16ca9f1 - add PA validation table (5 months ago) [Lawrence Lee]
* 8f7aaef - add route group table (5 months ago) [Lawrence Lee]
Work item tracking
Microsoft ADO (number only):
How I did it
How to verify it
Why I did it
Upgrade xgs SAI version to 10.1.37.0 with official branch.

10.1.37.0: [CS00012362288][TH5] SAI Queue counters are not reported for all the queues
Work item tracking
Microsoft ADO (number only):
How I did it
Change back xgs SAI branch to SAI_10.1.0_GA.

How to verify it
Run SAI release validation pipeline: https://dev.azure.com/mssonic/internal/_build/results?buildId=613625&view=results
jostar-yang and others added 28 commits October 4, 2024 12:02
Signed-off-by: michael_shih <michael_shih@edge-core.com>
Signed-off-by: michael_shih <michael_shih@edge-core.com>
Signed-off-by: jostar-yang <jostar_yang@edge-core.com>
Signed-off-by: michael_shih <michael_shih@edge-core.com>
Signed-off-by: michael_shih <michael_shih@edge-core.com>
…_peer branch

The matofeder/fix_bgp_peer branch contains PR#20333 [frrcfgd] Support
configuring source IPv6 for routes.
refer to: sonic-net/sonic-buildimage#20333

Signed-off-by: Matej Feder <matej.feder@dnation.cloud>
@matofeder matofeder deleted the branch master October 7, 2024 07:52
@matofeder matofeder closed this Oct 7, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.