DNX fixes:
CS00012287482 - support for 1024 LAGs on DNX
Other changes (XGS fixes)
SID - L3 multicast packet drop due to wrong VFI derivation - SDK-350470
SID - SIGSEGV in linkscan callback delivery - SDK-287578
SID - Repeated VXLAN calls deletes vlan translation action profile SDK-313980
SER - error in IS_TDM_CALENDAR0/1 can cause traffic hit in TH
SID - L2_ENTRY Table Lookups May Miss
[CSP CS00012275452] sai_object_type_get_availability failed with SAI_STATUS_INVALID_PARAMETER
To include the following DNX changes:
Revert patch and add official SDK/SAI fix for the below CSPs
a. CS00012282080 : syncd crashes after a speed change due to "cosq src vsqs gport get" failure
b. CS00012281200 : J2C+ : Scope of config.bcm SOC property bcm_stat_interval
Fixes for:
a. CS00012278343: SONiC J2c+ Macsec: Shutting down LAG members which have macsec cause
remaining active LAG members to go down
b. CS00012279717: Instance_id printed in SAI syslog messages are truncated to 9 bytes
Why I did it
Update SAI xgs version to 7.1.36.4 to include the following changes.
JIRA# SONIC-69731 (7.1.33.4)
Issue Summary: SAI_SWITCH_ATTR_SWITCH_HARDWARE_INFO brcm_sai_get_switch_attribute returns null.
Root Cause: Not implemented.
Fix Description: Get support for SAI switch attr SAI_SWITCH_ATTR_SWITCH_HARDWARE_INFO added
JIRA# SONIC-70737 (7.1.34.4)
Issue Summary: ECN being marked as CE even without congestion
Root Cause: ecn_thresh was set to very low value and packets were 100% marked.
Fix Description: ecn_thresh set to correct value
backport SONIC-70081 to SAI7.1 (7.1.35.4)
egress lossy queue PFC Rx fix:ignore PFC signals from egress
Update git submodules (7.1.36.4)
Update sdk-src/hsdk_6.5.24_SAI_7.1.0_GA from branch 'hsdk_6.5.24_SAI_7.1.0_GA'
to 57d0e360269c4ab659c4790ae471aa4dba2532b4
[SAI_BRANCH rel_ocp_sai_7_1] Broadcom image build failed with SAI 7.1 in DMZ repo (on bullseye)
How I did it
Update SAI xgs code.
How to verify it
Run the SONiC and SAI test with the 7.1 SAI pipeline.
Signed-off-by: zitingguo-ms <zitingguo@microsoft.com>
Why I did it
Update SAI xgs version to 7.1.36.4 to include the following changes and migrate xgs to DMZ repo.
JIRA# SONIC-69731 (7.1.33.4)
Issue Summary: SAI_SWITCH_ATTR_SWITCH_HARDWARE_INFO brcm_sai_get_switch_attribute returns null.
Root Cause: Not implemented.
Fix Description: Get support for SAI switch attr SAI_SWITCH_ATTR_SWITCH_HARDWARE_INFO added
JIRA# SONIC-70737 (7.1.34.4)
Issue Summary: ECN being marked as CE even without congestion
Root Cause: ecn_thresh was set to very low value and packets were 100% marked.
Fix Description: ecn_thresh set to correct value
backport SONIC-70081 to SAI7.1 (7.1.35.4)
egress lossy queue PFC Rx fix:ignore PFC signals from egress
Update git submodules (7.1.36.4)
Update sdk-src/hsdk_6.5.24_SAI_7.1.0_GA from branch 'hsdk_6.5.24_SAI_7.1.0_GA'
to 57d0e360269c4ab659c4790ae471aa4dba2532b4
[SAI_BRANCH rel_ocp_sai_7_1] Broadcom image build failed with SAI 7.1 in DMZ repo (on bullseye)
How I did it
Update SAI xgs code.
How to verify it
Run the SONiC and SAI test with the 7.1 SAI pipeline.
Signed-off-by: zitingguo-ms <zitingguo@microsoft.com>
Why I did it
To include DNX fix
Temp workaround for CS00012281200: J2C+ : Scope of config.bcm SOC property bcm_stat_interval
How I did it
Updated SAI version
How to verify it
Basic validation on DNX platform
Why I did it
To pick up the below DNX fixes:
CS00012275689: DSCP->TC and TC->QUEUE mappings are not happening for packets received on LAG ports (SONIC-69367)
CS00012277618: Crash in _brcm_sai_dnx_irpp_port_core_get (SONIC-70001)
How I did it
Updated SAI branch with the above fixes
How to verify it
Ran basic sonic-mgmt tests with the SAI debian on XGS and DNX platforms
Why I did it
To bring in the following fixes:
Revert temporary fix added to disable SA equal DA drops
CS00012273013 - [7.1][J2, J2c+] Disable SA Equals DA trap on DNX
CS00012274222 - How to block the voq for given destination port for a flow from a remote mod-id
CS00012275381 - SAI_INGRESS_PRIORITY_GROUP_STAT_PACKETS is incremented for port's PG's even if there are no traffic sent to that PG
CS00012274433 - Local Fault and Remote Fault are not polled by linkscan thread
How I did it
Merged above fixes to SAI code
How to verify it
Validated by running the basic sanity tests on XGS and DNX chassis platforms including
fib/test_fib.py
decap/test_decap.py
drop_counters/test_drop_counters.py
arp/test_arpall.py
Why I did it
To bring the following fixes:
Applied customer patch for limited port breakout support in rel_ocp_sai_7_1.
Revert "Merged PR 7224850: SAI7.1_DNX: Temp workaround for Nexthop Group Scale Issue(CS00012251649)".
backport SONIC-67662 to SAI7.1:JR2C ECMP partition for NHgroup members.
How I did it
Updated SAI code with the fixes above.
How to verify it
Run the SONiC and SAI test with the SAI pipeline.
Why I did it
It solves a swss orchagent crash issue on PikeZ device, due to link-training setting of external PHY port.
How I did it
Catch up the fix for CS00012257483 in version 7.1.7.2.
Why I did it
Update Broadcom SAI to version 6.0.0.13, SDK 6.5.24, saibcm-modules to 6.5.24.gpl
How I did it
Brcm SAI 6.0 EA with fixes for CS00012203367, CS00012219613, CS00012213974, CS00012218290, CS00012217169, CS00012211718, CS00012213944, CS00012215529, CS00012218100, CS00012214196, CS00012212681, CS00012205138, CS00012208537, CS00012185316, CS00012208524, CS00012203367, CS00012197364.
For broadcom sai, we only need to upgrade the version, not necessary the token part in the url.
Co-authored-by: Ubuntu <xumia@xumia-vm1.jqzc3g5pdlluxln0vevsg3s20h.xx.internal.cloudapp.net>
BRCM SAI missed implementing the SAI API "sai_query_stats_capability()" which is causing build issue.
The build issue is impacting PR(s) that need to use this API.
This PR is to stubbed BRCM SAI to add this SAI API and return not implemented so that it will fix build issue that it is causing.
No other functional changes were made.
Catch up on fixes from BRCM SUG repo to pick up fixes after 5.0.0.6 all the way up to 5.0.0.8
Fixes include the following:
```
CS00012201827: Warmreboot causes syncd crash with SAI_API_UNSPECIFIED:syncdb_data_file_read:2230 Failed to parse JSON: error -2
DNX: Fix for ACL table create with v6 next hdr attr
and many unspecified changes that also went into 5.0.0.8
```
#### How to verify it
Preliminary tests looks fine on both XGS (gechiang) and DNX (judyjoseph)
On XGS testing done as following:
BGP neighbors were all up with proper routes programmed
interfaces are all up
Manually ran the following test cases on 7260CX3 T0 DUT and all passed:
```
ipfwd/test_dir_bcast.py
fdb/test_fdb.py
fib/test_fib.py
vlan/test_valn.py
```
Also validated for for CS00012201827 (https://github.com/Azure/sonic-buildimage/issues/8300)
Move SAI libraries for Broadcom for XGS and DNX families to 5.0.0.6
Included fixes
```
CS00012195263 | [4.3][5.0][TD3] Packets with broken IP headers received on VLAN interface are not dropped
CS00012192505 | [4.3] Re-encap IPinIP decap packets
CS00012192502 | [3.7.5.2] Start LED shell script execution on all DELL based platforms causing all ports flapping on SAI 3.7.5.2
CS00012191363 | [4.3] Support of memscan thread to detect TCAM parity error
CS00012190932 | [4.3] SAI_PORT_PFC_X_RX_PKTS incremented incorrectly even when no PFC frames are received on that priority
CS00012183901 | [4.3][WARMBOOT] WARMReboot with active traffic causes port flap reported during warm reboot
CS00011382163 | [4.4] Support warm-boot from 3.5 to 4.3
CS00011318937 | [4.3] MACSec SAI Support for Jericho2c+
CS00011318926 | [4.3] Provide SAI support for Jericho2c+
CS00012195263 | [4.3][5.0][TD3] Packets with broken IP headers received on VLAN interface are not dropped
CS00012195261 | [4.3][5.0][TD3]VLAN tagged IP packet received on untagged interface being routed instead of dropped
CS00012183901 | [4.3][WARMBOOT] WARMReboot with active traffic causes port flap reported during warm reboot
CS00012196056 | [4.3.3.8][WARMBOOT] syncd[2584]: segfault at 5616ad6c3d80 ip 00007f61e0c6bc65 sp 00007fff0c5a7a90 error 4 in libsai.so.1.0[7f61e0a95000+3cd8000]
CS00012195262 | [4.3][5.0][TD3] Malformed IP packet(missing IP header) received on a VLAN Interface is flooded to other LVAN members instead of being dropped
CS00012195956 | [4.3.3.8] [TD3]Syncd Crash at brcm_sai_tnl_mp_create_tunnel()
PR 4346163: Add support for AN/LT
```
Why I did it
Update XGS and DNX SAI to 5.0.0.4 and additional flags needed in saibcm-modules
The following CSP's are merged in 5.0.0.4
CS00012182148 [4.3] Rate Limit Parity error message to syncd/sonic.
CS00012178692 [4.3] ACL drops counted as interface drops
CS00012183901 [4.3][WARMBOOT] WARMReboot with active traffic causes port flap reported during warm reboot
CS00012070713 [SAI 4.3 , DNX, 8690] Everflow ACL creation fails - brcm_sai_dnx_create_acl_table API fails, with unknown attribute error.
CS00012023263 [4.4] TD3/TH2 : Support 4 lossless queues(2 SW PFCWD and 2 HW PFCWD)
CS00012019578 [4.4] Pre FEC bit-error rate (BER) - DNX and XGS (TD and TH 50/100G)
How I did it
Changes the various make files to include the new SAI release + update the opennsl-modules.
Introduce new sonic-buildimage images for Broadcom DNX ASIC family.
sonic-broadcom-dnx.bin
sonic-aboot-broadcom-dnx.swi
How I did it
NO CHANGE to existing make commands
make init; make configure PLATFORM=broadcom; make target/sonic-aboot-broadcom.swi; make target/sonic-broadcom.bin
The difference now is that it will result in new broadcom images for DNX asic family as well.
sonic-broadcom.bin, sonic-broadcom-dnx.bin
sonic-aboot-broadcom.swi, sonic-aboot-broadcom-dnx.swi
Note: This PR also adds support for Broadcom SAI 5.0 (based on 1.8 SAI ) for DNX based platform + changes in platform x86_64-arista_7280cr3_32p4 bcm config files and platform_env.conf files
Added support BRCM SAI 5.0.0.1.
Major changes here:
CS00012019568 Link Training (all 100G ASICs - TH families and TD3)
CS00012184310 [attribute_capability| for port SAI_PORT_ATTR_TPID returns CREATE_IMP=false|SET_IMP=true|GET_IMP=true
CS00012182145 [IPinIP][Tunnel Delete] If IPinIP tunnel delete is performed observed following SYNCd error: ERR syncd#syncd: [none] SAI_API_TUNNEL:brcm_sai_tnl_mp_remove_tunnel_term_table_entry:4026 _brcm_sai_mptnl_sip_tnl_lookup failed with error -7.
CS00012182148 Rate Limit Parity error message to syncd/sonic.
CS00012178692 ACL drops counted as interface drops
CS00012183901 [WARMBOOT] WARMReboot with active traffic causes port flap reported during warm reboot
CS00012023263 TD3/TH2 : Support 4 lossless queues(2 SW PFCWD and 2 HW PFCWD)
CS00012019578 Pre FEC bit-error rate (BER) - DNX and XGS (TD and TH 50/100G)
What/Why I did:
Updated sonic-sairedis submodule to use SAI1.8.1
[Submodule update] sonic-sairedis
d821bc0b137264daa01c347700c7c14677cf3370 (HEAD -> master, origin/master, origin/HEAD) [Mellanox] Add SAI template
config support (#803)
bb341e9ea069e974a41930d434d437f522476f29 [syncd] Bring back TimerWatchdog (#821)
badf6cea2650015269420932a9186113d1ad5ec6 Update .gitignore (#822)
1494bc69046ffe7135377844548a11e4168b407c [meta] Mark local function static (#818)
34e961cf39e9af93f492f66640739e1c7a1694c8 [pyext] Fix pyext/py2 library (#820)
0d3749d3a93fd7e59ebb83b49fa1d7e2a56d6cf4 Moved SAI Header to git tag v1.8.1 (#816)
70fff780d529f78b53af4bd104f4932d0c4d8dd6 Added --purge of base docker image packages before installing new ones. (#819)
Updated Broadcom SAI Debian package to 4.3.3.4-2 to use SAI 1.8.1 Header
To add latest SAI drop REL_4.3.3.3 to SONIC which addresses the following CSP cases:
CS00012058054: [4.3][IPinIP][TTL-PIPE] IPinIP TTL Pipe Mode is NOT working it is behaving UNIFORM mode even programed as PIPE mode
CS00011227466: [4.3] Warmboot support with tunnel encap
This PR is needed to fix the show interface counters output issue where the counters are not correct due to an issue introduced in BRCM SAI 4.3.
- How to verify it
Without the fix if one injects packets, the expected counters for the interfaces involved do not show correct count values. The RX count looks to be TX count while TX count looks to be RX count but even that the values could not be trusted.
After the fix the counters started to look correct. Here is one sample output taken after the fix is applied where I manually injected 10,000 packets into Ethernet92 to be routed out of the port channel member port Ethernet40. Also injected 10,000 invalid packets into the same Ethernet92 and all 10,000 packets were shown RX_DRP correctly.
Merged bcmsai 4.3.0.13 code to top of master bcmsai 4.3.0.10-5.
- How to verify it
Ran nightly regression on T0 and T1 topology using bcmsai 4.3.0.13. Test results are better than previous runs.
For T0 -
New test passing - CRM, Decap, FDB, platform test, VxLAN
New test failing – PFC unknown MAC
For T1 –
New test passing – Port Channel
New test failing – platform test
ACL entry set attribute updates all the entries in the table. The correct behavior is to set the attribute on single entry.
- How I did it
Current SDK code, while setting the new attribute, is going through all the entries and updating it. Added a logic to check for requested entry and only allow for that ACL entry.
A case has filed with BRCM. Once an official fix is provided by BRCM, we will then remove this in house fix and apply the official fix.
1. BRCM SAI Debian build need not have any Kernel version dependency - Starting with 4.3 BRCM made changes in SAI so that this dependency has been cleaned up. We can now remove the Kernel Version dependency from Azure Pipeline build script.
2. Bypass PEER_MODE p2mp setting causing SYNCd crash on non-TD3 SKUs - Temporarily patch BRCM SAI code to not cause SYNCd crash when Orchagent program SAI_TUNNEL_ATTR_PEER_MODE: SAI_TUNNEL_PEER_MODE_P2MP on Non-TD3 SKUs. Will remove this when BRCM provide proper fix to address this issue.
