[Mellanox] Backport patch to remove critical trip point from thermal zones

[stephenxs]

Backport a patch to remove critical trip point from thermal zones for Mellanox devices

1. 0027-mlxsw-core-Remove-critical-trip-point-from-thermal-z.patch

   Disable software thermal protection by removing critical trip points
   for the all thermal zones.
   
   According to the system requirements software should never perform
   system thermal protection, since all the systems implement two levels
   of thermal protection: the first one is performed by firmware, the
   second, in case firmware was not able to perform protection, by
   hardware, while the temperature threshold for hardware protection is
   higher than for firmware.
   
   In both cases, when critical temperature is reached, system will be
   shutdown.
   
   Signed-off-by: Vadim Pasternak <vadimp@nvidia.com>
   

It has been verified on Mellanox devices

Signed-off-by: Stephen Sun stephens@nvidia.com

[paulmenzel]

Thank you for the patch.

1. Ingetrate → Integrate
2. Please add more details, where the patch comes from. Best you give the git commit hash from the patch in Linus’ master branch.
3. Please elaborate, what the problem is and how the patch was tested. (Most could be taken from the patch description I guess.)

[liat-grozovik]

Thank you for the patch.

1. Ingetrate → Integrate
2. Please add more details, where the patch comes from. Best you give the git commit hash from the patch in Linus’ master branch.
3. Please elaborate, what the problem is and how the patch was tested. (Most could be taken from the patch description I guess.)

This is a work in progress in term of linux main. Our team will do the upstream process soon but this must not delay this update. The reason for that that this is a real production issue and we cannot wait for this one.

As for the information you are looking for, it is all in the patch itself.

"
Disable software thermal protection by removing critical trip points
for the all thermal zones.

According to the system requirements software should never perform
system thermal protection, since all the systems implement two levels
of thermal protection: the first one is performed by firmware, the
second, in case firmware was not able to perform protection, by
hardware, while the temperature threshold for hardware protection is
higher than for firmware.

In both cases, when critical temperature is reached, system will be
shutdown.
"

This is relevant to only Mellanox switches as the main protection is done by Firmware.

[paulmenzel]

The commit in this merge/pull request still has the typos and a to-be-desired commit message. It’s rude to expect from reviewers and later people looking through commits to look at the diff to find out what is done.

If this is important to you, then I suggest you make sure, that all the formal requirements for patches are met.

For upstream, if it’s done later, please add a comment to the commit message, that it is going to be upstreamed soon.

[stephenxs]

The commit in this merge/pull request still has the typos and a to-be-desired commit message. It’s rude to expect from reviewers and later people looking through commits to look at the diff to find out what is done.

If this is important to you, then I suggest you make sure, that all the formal requirements for patches are met.

For upstream, if it’s done later, please add a comment to the commit message, that it is going to be upstreamed soon.

Hi @paulmenzel
Thank you for your comments. I updated the commit message as well as the PR description.

[paulmenzel]

“Mellanox devices” is a little broad, but maybe you have a QA system, which tests the SONiC buidls on all devices.

[stephenxs]

“Mellanox devices” is a little broad, but maybe you have a QA system, which tests the SONiC buidls on all devices.

Hi paulmenzel
Thank you for your comments.
Yes, we verify all platforms supported in SONiC when integrating a new kernel patch.

[dprital]

@paulmenzel - Can you please merge this PR ?

[paulmenzel]

I would, if I could, but I do not have the permission to merge the requests. At least @lguohan has the permissions.

[dprital]

@lguohan - Can you please merge this PR ? and also add the tag that is required for 202012 ? Thanks.