Skip to content

Conversation

hershpa
Copy link
Contributor

@hershpa hershpa commented Aug 22, 2023

Keep node feature rules only for SGX, QAT, and GPU.
Keep only 4xxx device IDs for QAT as we support QAT only on SPR or later platforms.
Signed-off-by: Hersh Pathak [email protected]

@hershpa hershpa added the nfd NFD label Aug 22, 2023
@hershpa hershpa added this to the v1.0.1 milestone Aug 22, 2023
@hershpa
Copy link
Contributor Author

hershpa commented Aug 22, 2023

@vbedida79 @mregmi @uMartinXu pls help review.

@hershpa hershpa added the do-not-merge/hold Indicates that a PR should not merge since it is on hold label Aug 22, 2023
@hershpa
Copy link
Contributor Author

hershpa commented Aug 22, 2023

  • testing completed successfully

class: {op: In, value: ["0b40"]}
- feature: kernel.loadedmodule
matchExpressions:
intel_qat: {op: Exists}

- name: "intel.sgx"
- name: "intel.sgx"
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

a change: indentation error here

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good eye! Thanks @vbedida79

@vbedida79
Copy link
Contributor

vbedida79 commented Aug 22, 2023

Deployed it to check sgx. The node is not labelled with "intel.feature.node.kubernetes.io/sgx": "true". Assuming that this change is not present in the NFD operator on OCP 4.12 yet, we can skip this for now. Need to reconfirm later.

@hershpa
Copy link
Contributor Author

hershpa commented Aug 23, 2023

Deployed it to check sgx. The node is not labelled with "intel.feature.node.kubernetes.io/sgx": "true". Assuming that this change is not present in the NFD operator on OCP 4.12 yet, we can skip this for now. Need to reconfirm later.

Got it, updated commit to keep SGX NFD rules as is for now.

@hershpa hershpa removed the do-not-merge/hold Indicates that a PR should not merge since it is on hold label Aug 23, 2023
@mythi
Copy link

mythi commented Aug 23, 2023

Deployed it to check sgx. The node is not labelled with "intel.feature.node.kubernetes.io/sgx": "true". Assuming that this change is not present in the NFD operator on OCP 4.12 yet, we can skip this for now. Need to reconfirm later.

What is the NFD version available? Do we need to submit a ticket to get the version updated?

matchExpressions:
dfl_pci: {op: Exists}

matchAny:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I prefer not to add the FPGA and IAA into the NFD if we haven't started to support them or will enable them in the coming release.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we can keep IAA since it will be enabled in 1.1.0 release. I agree, we can remove FPGA and DLB as well. Can we keep IAA or remove it as well?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I suggest to remove IAA, and only add what we will support.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Makes sense! :)

@@ -62,7 +81,7 @@ spec:
- feature: pci.device
matchExpressions:
vendor: {op: In, value: ["8086"]}
device: {op: In, value: ["37c8", "4940"]}
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@mregmi is these the device ID for 4xxx QAT device? Can we add some HW detail about the 4xxx QAT device?

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These are OK. What details you had in mind?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

All the devices with 4xxx device ID should be the integrated QAT 2.0 HW on SPR or later xeon platforms. @mythi is that correct?

@vbedida79
Copy link
Contributor

vbedida79 commented Aug 23, 2023

Deployed it to check sgx. The node is not labelled with "intel.feature.node.kubernetes.io/sgx": "true". Assuming that this change is not present in the NFD operator on OCP 4.12 yet, we can skip this for now. Need to reconfirm later.

What is the NFD version available? Do we need to submit a ticket to get the version updated?

Yes OCP NFD 4.12.0-202307182142. No exact info in repo Based on some older releases its upstream v0.x.0 for ocp 4.x. So v0.12.0. This change is in v0.13.0 upstream. Submitted a ticket on RH portal for confirming.

@hershpa
Copy link
Contributor Author

hershpa commented Aug 23, 2023

Thanks @vbedida79

@hershpa hershpa mentioned this pull request Aug 23, 2023
8 tasks
@uMartinXu
Copy link
Contributor

@hershpa for this PR, please only work on changing the QAT device. And in the commit log, explain the reason why we made this change. After that, I am OK to merge it. Thanks!

@hershpa
Copy link
Contributor Author

hershpa commented Aug 23, 2023

Sounds good, thanks @uMartinXu for the review.

@hershpa
Copy link
Contributor Author

hershpa commented Aug 24, 2023

Updated PR @uMartinXu, @vbedida79

@@ -54,15 +15,15 @@ spec:
matchExpressions:
vendor: {op: In, value: ["8086"]}
class: {op: In, value: ["0300", "0380"]}

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Clean up by remove the space from this line.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The PR looks good to me.

Keep node feature rules only for SGX, QAT, and GPU.
Keep only 4xxx device IDs for QAT as we support QAT only on SPR or later platforms.
Signed-off-by: Hersh Pathak [email protected]
@uMartinXu uMartinXu merged commit c44d529 into main Aug 25, 2023
@hershpa hershpa deleted the hershpa-patch-1 branch September 8, 2023 20:56
@vbedida79
Copy link
Contributor

Deployed it to check sgx. The node is not labelled with "intel.feature.node.kubernetes.io/sgx": "true". Assuming that this change is not present in the NFD operator on OCP 4.12 yet, we can skip this for now. Need to reconfirm later.

What is the NFD version available? Do we need to submit a ticket to get the version updated?

Yes OCP NFD 4.12.0-202307182142. No exact info in repo Based on some older releases its upstream v0.x.0 for ocp 4.x. So v0.12.0. This change is in v0.13.0 upstream. Submitted a ticket on RH portal for confirming.

Confirmed from RH, this change will be in OCP 4.14.

@hershpa
Copy link
Contributor Author

hershpa commented Sep 12, 2023

Thanks @vbedida79

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
nfd NFD
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants