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

DellEMC S6100 Enabling Polling Mode for Miim Operation #3133

Merged
merged 1 commit into from
Jul 16, 2019

Conversation

sridhar-ravindran
Copy link
Contributor

- What I did
Enabled Polling to verify Miim Operation Completion in S6100
BCM is running in User Mode in Sonic Image. Verifying the Miim Operation completion via interrupt will exhibit performance degradation during high CPU utilization times. We might take more time in handling interrupts. This might lead BCM to declare MIIM timeout.
Moving to MIIM Poller approach will help in mitigating these conditions.

- How I did it
miim_intr_enable = 0
- How to verify it
drivshell>config show miim_intr_enable
config show miim_intr_enable
miim_intr_enable=0

- Description for the changelog

[DellEMC] [S6100] Enabling Polling Mode for Miim Operation

- A picture of a cute animal (not mandatory but encouraged)

@lguohan
Copy link
Collaborator

lguohan commented Jul 9, 2019

where to define the polling interval?

@sridhar-ravindran
Copy link
Contributor Author

where to define the polling interval?

Hi Gouhan,

"miim_intr_enable",
"miim_timeout_usec", \

are the only two variables that are user configurable.

miim_timeout_usec is the config variable that is of our interest.

In Master branch , By default it is set to 1 sec.
We do not need more than 1 sec.

drivshell>config show miim_timeout_usec

drivshell>mtim
mtim
MIIM timeout is 1000000 usec

So i did not add this in the code change.

@lguohan
Copy link
Collaborator

lguohan commented Jul 9, 2019

also what is the average cpu overhead once switch to polling mode?

@lguohan
Copy link
Collaborator

lguohan commented Jul 16, 2019

what is the average cpu overhead once switch to polling mode?

@sridhar-ravindran
Copy link
Contributor Author

what is the average cpu overhead once switch to polling mode?

Hi Gouhan,

Please find the average load measured

  1. Retrieve syncd utilization every 12 second for 50 iterations
  2. 12*50 = 600 seconds(10 minutes)
  3. Regular Data Traffic along with 16 BGP sessions.

SYNCD PROCESS UTILIZATION

AVERAGE

                       INTERRUPT     POLLER  
                         =====       ====== 
                        53.136        55.976
                       =======      =====

Attached the text document for detailed Utilisation captures
syncd_poller_profiling.txt

@lguohan lguohan merged commit 6d3e9e8 into sonic-net:master Jul 16, 2019
@sridhar-ravindran sridhar-ravindran deleted the Miim_Polling branch July 16, 2019 23:56
mssonicbld added a commit that referenced this pull request May 22, 2024
…lly (#19026)

#### Why I did it
src/sonic-swss
```
* a280e367 - (HEAD -> master, origin/master, origin/HEAD) [COPP]Fix coppmgr when there are multiple features using same trap (#3144) (10 hours ago) [Sudharsan Dhamal Gopalarathnam]
* 1ff61602 - revert requests from 2.32.0 to 2.31.0 (#3154) (11 hours ago) [Liu Shilong]
* 9e80dc82 - Update setHostTxReady() to receive Port instead of portId (#3133) (35 hours ago) [noaOrMlnx]
```
#### How I did it
#### How to verify it
#### Description for the changelog
mssonicbld added a commit that referenced this pull request Jun 5, 2024
…lly (#19202)

#### Why I did it
src/sonic-swss
```
* 9e2df9a9 - (HEAD -> 202311, origin/202311) Do not apply QoS mapping item on the switch until the object is created (#3163) (10 hours ago) [Stephen Sun]
* c52f0840 - Update setHostTxReady() to receive Port instead of portId (#3133) (#3157) (10 hours ago) [noaOrMlnx]
```
#### How I did it
#### How to verify it
#### Description for the changelog
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants