DRA: Introduce a helper function producing standardized 'resource.kubernetes.io/pcieRoot' device attributes for DRA drivers #132296

everpeace · 2025-06-13T19:21:16Z

What type of PR is this?

/kind feature

What this PR does / why we need it:

This PR introduces a helper function GetPCIeRootAttributeByPCIBusID(pciBusID) for DRA drivers, allowing them to produce standardized device attributes in a consistent way.

Which issue(s) this PR is related to:

ref: NVIDIA/k8s-dra-driver-gpu#401 (comment)
KEP: kubernetes/enhancements#4381

Special notes for your reviewer:

Does this PR introduce a user-facing change?

Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.:

- [KEP]: https://github.com/kubernetes/enhancements/tree/master/keps/sig-node/4381-dra-structured-parameters

…utes for DRA drivers The shared function can provide a way for DRA drivers to produce standardized device attributes in a consistent way. Signed-off-by: Shingo Omura <[email protected]>

k8s-ci-robot · 2025-06-13T19:21:20Z

Adding the "do-not-merge/release-note-label-needed" label because no release-note block was detected, please follow our release note process to remove it.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

k8s-ci-robot · 2025-06-13T19:21:26Z

This issue is currently awaiting triage.

If a SIG or subproject determines this is a relevant issue, they will accept it by applying the triage/accepted label and provide further guidance.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

k8s-ci-robot · 2025-06-13T19:21:37Z

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: everpeace
Once this PR has been reviewed and has the lgtm label, please assign pohly for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files:

staging/src/k8s.io/dynamic-resource-allocation/OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

everpeace · 2025-06-13T19:24:18Z

/assign @aojea

michaelasp · 2025-06-13T19:28:09Z

/cc @michaelasp

k8s-ci-robot · 2025-06-13T19:28:12Z

@michaelasp: GitHub didn't allow me to request PR reviews from the following users: michaelasp.

Note that only kubernetes members and repo collaborators can review this PR, and authors cannot review their own PRs.

In response to this:

/cc @michaelasp

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

michaelasp

This is awesome, thanks for adding this!

staging/src/k8s.io/dynamic-resource-allocation/utils/pci.go

Signed-off-by: Shingo Omura <[email protected]>

klueska · 2025-06-18T09:05:01Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci.go

Most of this is linux specific so the parts that are should be in a _linux.go specific file

fixed in 73d65f8

klueska · 2025-06-18T09:07:28Z

staging/src/k8s.io/dynamic-resource-allocation/api/standard_attributes.go

+	// The value is a string value in the format `pci<domain>:<bus>`,
+	// referring to a PCIe (Peripheral Component Interconnect Express) Root Complex.
+	// This attribute can be used to identify devices that share the same PCIe Root Complex.
+	StandardDeviceAttributePCIeRoot = StandardDeviceAttributePrefix + "pcieRoot"


Suggested change

StandardDeviceAttributePCIeRoot = StandardDeviceAttributePrefix + "pcieRoot"

StandardDeviceAttributePCIeRoot resourceapi.QualifiedName = StandardDeviceAttributePrefix + "pcieRoot"

fixed in 8ed8217

klueska · 2025-06-18T09:08:26Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci.go

+)
+
+var (
+	bdfRegexp = regexp.MustCompile(`^([0-9a-f]{4}):([0-9a-f]{2}):([0-9a-f]{2})\.([0-9a-f]{1})$`)


Give that this is only use din one function it seems unnecessary to declare as a global like this

fixed in 88b34ea

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci.go

…helper functions" This reverts commit 2869b6f.

Signed-off-by: Shingo Omura <[email protected]>

…PCIeRootAttributeByPCIBusID' with it Signed-off-by: Shingo Omura <[email protected]>

klueska · 2025-06-18T13:20:13Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/constants.go

lets conslidate this into the attribute.go file

fixed in caa64af

klueska · 2025-06-18T13:23:23Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci_linux.go

+// GetPCIeRootAttributeByPCIBusID retrieves the PCIe Root Complex for a given PCI Bus ID.
+// in BDF (Bus-Device-Function) format, e.g., "0123:45:1e.7".
+//
+// It returns a DeviceAttribute with the PCIe Root Complex information("pci<domain>:<bus>")
+// as a string value or an error if the PCI Bus ID is invalid or the root complex cannot be determined.
+//
+// ref: https://wiki.xenproject.org/wiki/Bus:Device.Function_(BDF)_Notation
+func GetPCIeRootAttributeByPCIBusID(pciBusID string) (DeviceAttribute, error) {


I don't think I said this elsewhere, but the reason to be explicit about ByPCIBusID is that I think we will want a GetPCIeRootAttributeByCpuID() at some point soon.

Yeah, I also imagined GetPCIeRootAttributeByInterfaceName(), too (but, my original idea was to implement these variations via option patterns).

We can resolve PCIe root for interfaces by readlink /sys/class/net/<ifname>/device.

# PCIe root for "enp48s0f0np0" is "pci0000:20" $ readlink -f /sys/class/net/enp48s0f0np0/device /sys/devices/pci0000:20/0000:20:03.1/0000:26:00.0/0000:27:08.0/0000:2e:00.0/0000:2f:10.0/0000:30:00.0

We can revisit this if / when we get another variant.

By the way,

I think we will want a GetPCIeRootAttributeByCpuID() at some point soon.

It might be GetCpuSocketNumberByCpuID() or GetNumaNodeByCpuID()?? I think CPU is not within the PCIe hierarchy. Thus, aligning NIC and CPUs in hardware topology, we would need another standard device attribute, e.g. cpuSocketNumber or numaNode, in the near future (as stated in kubernetes/enhancements#5316 (comment)).

EXAMPLE: I attached an example of lstopo output image. In this example, there are two packages, and each package has one NUMA Node, each numa node has 64 cores and 4 PCI Host Bridges (PCIe Roots).

$ lstopo -v --of console Machine (P#0 total=1584474936KB DMIProductName="REDACTED" DMIProductVersion=xxxxxx DMIBoardVendor=REDACTED DMIBoardName=REDACTED DMIBoardVersion=REDACTED DMIBoardAssetTag="REDACTED" DMIChassisVendor=REDACTED DMIChassisType=1 DMIChassisVersion=REDACTED DMIChassisAsse Package L#0 (P#0 total=791818556KB CPUVendor=REDACTED CPUFamilyNumber=REDACTED CPUModelNumber=REDACTED CPUModel="REDACTED" CPUStepping=1) NUMANode L#0 (P#0 local=791818556KB total=791818556KB) ...<CPUs here>... HostBridge L#0 (buses=0000:[00-0e]) ...<PCIe devices/bridges here>... HostBridge L#12 (buses=0000:[20-34]) ...<PCIe devices/bridges here>... HostBridge L#24 (buses=0000:[40-46]) ...<PCIe devices/bridges here>... HostBridge L#29 (buses=0000:[60-6b]) ...<PCIe devices/bridges here>... Package L#1 (P#1 total=792656380KB CPUVendor=REDACTED CPUFamilyNumber=REDACTED CPUModelNumber=REDACTED CPUModel="REDACTED" CPUStepping=1) NUMANode L#1 (P#1 local=792656380KB total=792656380KB) ...<CPUs here>... HostBridge L#39 (buses=0000:[80-87]) ...<PCIe devices/bridges here>... HostBridge L#44 (buses=0000:[a0-a6]) ...<PCIe devices/bridges here>... HostBridge L#49 (buses=0000:[c0-c6]) ...<PCIe devices/bridges here>... HostBridge L#53 (buses=0000:[e0-e7]) ...<PCIe devices/bridges here>... ...

We want to map a given CPU to its PCIe root complex independent of NUMA. We've done alignment by NUMA in the past, but going forward direct alignment by PCIe root complex is the preferred approach.

One way to see this association on my 4-GPU system is as follows:

$ cat /sys/bus/pci/devices/0019:01:00.0/local_cpulist 72-143 $ cat /sys/bus/pci/devices/0018:01:00.0/local_cpulist 72-143 $ cat /sys/bus/pci/devices/0009:01:00.0/local_cpulist 0-71 $ cat /sys/bus/pci/devices/0008:01:00.0/local_cpulist 0-71

The first two GPUs are "local" to CPUs 72-143 and the second two GPUs are "local" to CPUs 0-71. Where "local" implies that they are on the same PCIe root complex. You can imagine how one could use this to implement a robust GetPCIeRootAttributeByCpuID() function.

There may be other more straightforward ways to glean this information, but this is gist of it.

/cc @kad

The first two GPUs are "local" to CPUs 72-143, and the second two GPUs are "local" to CPUs 0-71. Where "local" implies that they are on the same PCIe root complex. You can imagine how one could use this to implement a robust GetPCIeRootAttributeByCpuID() function.

In my 8-GPU system,

$ nvidia-smi --query-gpu=gpu_bus_id --format=csv,noheader 00000000:04:00.0 00000000:23:00.0 00000000:43:00.0 00000000:64:00.0 00000000:84:00.0 00000000:A3:00.0 00000000:C3:00.0 00000000:E4:00.0

For example, For 0000:04:00.0(1st) and 0000:23:00.0(2nd):

both are "local" to 0-63,

BUT, they have different PCIe roots

# For 0000:04:00.0: # it is "local" to CPU 0-63 $ cat /sys/bus/pci/devices/0000:04:00.0/local_cpulist 0-63 # its PCIe root is pci0000:00 $ readlink /sys/bus/pci/devices/0000:04:00.0 ../../../devices/pci0000:00/0000:00:01.1/0000:01:00.0/0000:02:01.0/0000:04:00.0 # For 0000:23:00.0: # it is "local" to CPU 0-63, too $ cat /sys/bus/pci/devices/0000:23:00.0/local_cpulist 0-63 # but, its PCIe Root is pci0000:20 $ readlink /sys/bus/pci/devices/0000:23:00.0 ../../../devices/pci0000:20/0000:20:01.1/0000:21:00.0/0000:22:00.0/0000:23:00.0

Thus, CPU 0-63 are "local" to at least two PCIe roots(pci0000:00 and pci0000:20). So, I am still not sure how we can map a given CPU to its PCIe root complex🤔. Are there any other ways to determine/define it?

There is not necessarily a 1:1 mapping for CPU <--> PCIe root complex, unfortunately (which can obviously complicate things when we want to do exact string matching for alignment). Seems when it comes to CPUs we may need a list rather than a single value -- which unfortunately won't work with simple matchAttributes but will when we have the more expressive version of this: kubernetes/enhancements#5254

Note, on my 8-GPU system I don't have this "issue":

$ nvidia-smi --query-gpu=gpu_bus_id --format=csv,noheader 00000000:07:00.0 00000000:0F:00.0 00000000:47:00.0 00000000:4E:00.0 00000000:87:00.0 00000000:90:00.0 00000000:B7:00.0 00000000:BD:00.0 $ cat /sys/bus/pci/devices/0000\:07\:00.0/local_cpulist 48-63,176-191 $ cat /sys/bus/pci/devices/0000\:0f\:00.0/local_cpulist 48-63,176-191 $ readlink /sys/bus/pci/devices/0000\:07\:00.0 ../../../devices/pci0000:00/0000:00:01.1/0000:01:00.0/0000:02:00.0/0000:03:00.0/0000:04:00.0/0000:05:00.0/0000:06:00.0/0000:07:00.0 $ readlink /sys/bus/pci/devices/0000\:0f\:00.0 ../../../devices/pci0000:00/0000:00:01.1/0000:01:00.0/0000:02:04.0/0000:0a:00.0/0000:0b:10.0/0000:0d:00.0/0000:0e:00.0/0000:0f:00.0 $ cat /sys/bus/pci/devices/0000\:47\:00.0/local_cpulist 16-31,144-159 $ cat /sys/bus/pci/devices/0000\:4e\:00.0/local_cpulist 16-31,144-159 $ readlink /sys/bus/pci/devices/0000\:47\:00.0 ../../../devices/pci0000:40/0000:40:01.1/0000:41:00.0/0000:42:00.0/0000:43:00.0/0000:44:00.0/0000:45:00.0/0000:46:00.0/0000:47:00.0 $ readlink /sys/bus/pci/devices/0000\:4e\:00.0 ../../../devices/pci0000:40/0000:40:01.1/0000:41:00.0/0000:42:04.0/0000:49:00.0/0000:4a:10.0/0000:4c:00.0/0000:4d:00.0/0000:4e:00.0 $ cat /sys/bus/pci/devices/0000\:87\:00.0/local_cpulist 112-127,240-255 $ cat /sys/bus/pci/devices/0000\:90\:00.0/local_cpulist 112-127,240-255 $ readlink /sys/bus/pci/devices/0000\:87\:00.0 ../../../devices/pci0000:80/0000:80:01.1/0000:81:00.0/0000:82:00.0/0000:83:00.0/0000:84:00.0/0000:85:00.0/0000:86:00.0/0000:87:00.0 $ readlink /sys/bus/pci/devices/0000\:90\:00.0 ../../../devices/pci0000:80/0000:80:01.1/0000:81:00.0/0000:82:04.0/0000:8b:00.0/0000:8c:10.0/0000:8e:00.0/0000:8f:00.0/0000:90:00.0 $ cat /sys/bus/pci/devices/0000\:b7\:00.0/local_cpulist 80-95,208-223 $ cat /sys/bus/pci/devices/0000\:bd\:00.0/local_cpulist 80-95,208-223 $ readlink /sys/bus/pci/devices/0000\:b7\:00.0 ../../../devices/pci0000:b0/0000:b0:01.1/0000:b1:00.0/0000:b2:00.0/0000:b3:00.0/0000:b4:00.0/0000:b5:00.0/0000:b6:00.0/0000:b7:00.0 $ readlink /sys/bus/pci/devices/0000\:bd\:00.0 ../../../devices/pci0000:b0/0000:b0:01.1/0000:b1:00.0/0000:b2:04.0/0000:b8:00.0/0000:b9:10.0/0000:bb:00.0/0000:bc:00.0/0000:bd:00.0

Thanks.

Seems when it comes to CPUs we may need a list rather than a single value -- which unfortunately won't work with simple matchAttributes but will when we have the more expressive version of this: kubernetes/enhancements#5254

Yes, I agree. In that case, could we plan to add resource.kubernetes.io/pcieRoots(plural) rather than resource.kubernetes.io/pcieRoot in the near future? And, could we add an array type as device attribute's value type?

Note, on my 8-GPU system I don't have this "issue":

lucky you 😄

Signed-off-by: Shingo Omura <[email protected]>

klueska · 2025-06-18T14:37:06Z

@pohly can you point us where this can be exercised in the e2e driver?

pohly · 2025-06-26T09:45:01Z

can you point us where this can be exercised in the e2e driver?

The test/e2e/dra/test-driver? That driver has no hardware dependency, so the code in this PR cannot be exercised through it.

In general, writing any kind of integration or E2E test for this in the Kubernetes CI is problematic because it needs some real, deterministic sysfs. We don't have that.

klueska · 2025-06-26T10:01:10Z

Thanks @pohly that makes sense. Do you have any comments on the name of the package / API for this?

pohly · 2025-06-27T14:04:50Z

/test pull-kubernetes-apidiff

pohly · 2025-06-29T11:48:45Z

/test pull-kubernetes-apidiff

Might have died because of OOM.

pohly · 2025-06-30T06:55:20Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci_linux_test.go

+		expectedErrMsg    string
+	}{
+		"valid": {
+			smockSysfsSetup: func(t *testing.T, mockSysfs sysfsPath) {


What does the "s" in "smock" stand for?

Oh, it's just typo... Thanks for the catch!

fixed in 94973ba

pohly · 2025-06-30T06:56:09Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci_linux_test.go

+	}
+}
+
+func TouchFile(t *testing.T, path string) {


Suggested change

func TouchFile(t *testing.T, path string) {

func touchFile(t *testing.T, path string) {

It's easier to see that this is not a public API when it's not exported (doesn't make a different in _test.go, but still...).

fixed in 2bc45ca

pohly · 2025-06-30T06:56:17Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci_linux_test.go

+	}
+}
+
+func CreateSymlink(t *testing.T, target, link string) {


Suggested change

func CreateSymlink(t *testing.T, target, link string) {

func createSymlink(t *testing.T, target, link string) {

fixed in 2bc45ca

pohly · 2025-06-30T06:57:27Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/sysfs_linux.go

+// and can be replaced with a mock in tests.
+type sysfsPath string
+
+func (s sysfsPath) Devices(path string) string {


Suggested change

func (s sysfsPath) Devices(path string) string {

func (s sysfsPath) devices(path string) string {

Same below.

In this case, exporting the method makes it callable outside of the package (assuming that some exported method returns a sysfsPath) to that caller and thus part of the package API. We want a lean API to make future changes safer.

Fixed in 2bc45ca.

Signed-off-by: Shingo Omura <[email protected]>

…ymlink) Signed-off-by: Shingo Omura <[email protected]>

Signed-off-by: Shingo Omura <[email protected]>

everpeace

@pohly Thank you for your review. I addressed your comments. PTAL.

everpeace · 2025-06-30T07:04:22Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci_linux_test.go

+		expectedErrMsg    string
+	}{
+		"valid": {
+			smockSysfsSetup: func(t *testing.T, mockSysfs sysfsPath) {


Oh, it's just typo... Thanks for the catch!

everpeace · 2025-06-30T07:16:30Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci_linux_test.go

+		expectedErrMsg    string
+	}{
+		"valid": {
+			smockSysfsSetup: func(t *testing.T, mockSysfs sysfsPath) {


fixed in 94973ba

everpeace · 2025-06-30T07:16:53Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci_linux_test.go

+	}
+}
+
+func TouchFile(t *testing.T, path string) {


fixed in 2bc45ca

everpeace · 2025-06-30T07:17:00Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci_linux_test.go

+	}
+}
+
+func CreateSymlink(t *testing.T, target, link string) {


fixed in 2bc45ca

everpeace · 2025-06-30T07:17:59Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/sysfs_linux.go

+// and can be replaced with a mock in tests.
+type sysfsPath string
+
+func (s sysfsPath) Devices(path string) string {


Fixed in 2bc45ca.

everpeace · 2025-06-30T07:19:03Z

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/sysfs_linux.go

+// Commented out for future use or if needed later
+// func (s sysfsPath) block(path string) string {
+// 	return filepath.Join(string(s), "block", path)
+// }
+
+// Commented out for future use or if needed later
+// func (s sysfsPath) class(path string) string {
+// 	return filepath.Join(string(s), "class", path)
+// }
+
+// Commented out for future use or if needed later
+// func (s sysfsPath) dev(path string) string {
+// 	return filepath.Join(string(s), "dev", path)
+// }
+
+// Commented out for future use or if needed later
+// func (s sysfsPath) firmware(path string) string {
+// 	return filepath.Join(string(s), "firmware", path)
+// }
+// func (s sysfsPath) kernel(path string) string {
+// 	return filepath.Join(string(s), "kernel", path)
+// }
+
+// Commented out for future use or if needed later
+// func (s sysfsPath) module(path string) string {
+// 	return filepath.Join(string(s), "module", path)
+// }


These unused methods are just commented out. We can remove these. WDYT??

Let's remove them if they are unused.

removed in d28791f

Signed-off-by: Shingo Omura <[email protected]>

DRA: Introduce a helper function producing standardized device attrib…

666f38f

…utes for DRA drivers The shared function can provide a way for DRA drivers to produce standardized device attributes in a consistent way. Signed-off-by: Shingo Omura <[email protected]>

k8s-ci-robot added the needs-priority Indicates a PR lacks a `priority/foo` label and requires one. label Jun 13, 2025

k8s-ci-robot requested review from klueska and pohly June 13, 2025 19:21

k8s-ci-robot added sig/node Categorizes an issue or PR as relevant to SIG Node. wg/device-management Categorizes an issue or PR as relevant to WG Device Management. and removed do-not-merge/needs-sig Indicates an issue or PR lacks a `sig/foo` label and requires one. labels Jun 13, 2025

github-project-automation bot added this to Dynamic Resource Allocation and SIG Node: code and documentation PRs Jun 13, 2025

github-project-automation bot moved this to Triage in SIG Node: code and documentation PRs Jun 13, 2025

github-project-automation bot moved this to 🆕 New in Dynamic Resource Allocation Jun 13, 2025

everpeace mentioned this pull request Jun 13, 2025

Add standard device attribute 'resource.kubernetes.io/pcieRoot' to GPU and MIG Devices. NVIDIA/k8s-dra-driver-gpu#401

Draft

k8s-ci-robot assigned aojea Jun 13, 2025

michaelasp reviewed Jun 13, 2025

View reviewed changes

staging/src/k8s.io/dynamic-resource-allocation/utils/pci.go Outdated Show resolved Hide resolved

staging/src/k8s.io/dynamic-resource-allocation/utils/pci.go Outdated Show resolved Hide resolved

k8s-ci-robot added size/XXL Denotes a PR that changes 1000+ lines, ignoring generated files. and removed size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. labels Jun 14, 2025

everpeace force-pushed the dra-standard-attribute-and-pcieroot-resolution-util branch from 3cbc0aa to db0bf33 Compare June 14, 2025 06:33

fix lint: apply De Morgan's laws to simplify conditionals

752a043

Signed-off-by: Shingo Omura <[email protected]>

klueska reviewed Jun 18, 2025

View reviewed changes

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci.go Outdated Show resolved Hide resolved

klueska reviewed Jun 18, 2025

View reviewed changes

staging/src/k8s.io/dynamic-resource-allocation/deviceattribute/pci.go Outdated Show resolved Hide resolved

everpeace added 5 commits June 18, 2025 21:04

Revert "Add StandardDeviceAttributes and StandardPCIDeviceAttributes …

f7461a1

…helper functions" This reverts commit 2869b6f.

Moved device attribute constants from api to deviceattribute package

8ed8217

Signed-off-by: Shingo Omura <[email protected]>

Introduce DeviceAttribute{Name, Value} and change return type of 'Get…

fc854a9

…PCIeRootAttributeByPCIBusID' with it Signed-off-by: Shingo Omura <[email protected]>

Move bdfRegexp from global constant to local variable

88b34ea

Put _linux suffix for linux specific go files

73d65f8

k8s-ci-robot added size/L Denotes a PR that changes 100-499 lines, ignoring generated files. and removed size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. labels Jun 18, 2025

klueska reviewed Jun 18, 2025

View reviewed changes

merge constants.go into attribute.go

caa64af

Signed-off-by: Shingo Omura <[email protected]>

everpeace changed the title ~~DRA: Introduce a helper function producing standardized device attributes for DRA drivers~~ DRA: Introduce a helper function producing standardized 'resource.kubernetes.io/pcieRoot' device attributes for DRA drivers Jun 19, 2025

pohly moved this from 🆕 New to 👀 In review in Dynamic Resource Allocation Jun 25, 2025

k8s-ci-robot requested a review from kad June 26, 2025 09:17

pohly suggested changes Jun 30, 2025

View reviewed changes

github-project-automation bot moved this from Triage to Waiting on Author in SIG Node: code and documentation PRs Jun 30, 2025

everpeace added 3 commits June 30, 2025 16:14

Fix typo: s/smock/mock/g

94973ba

Signed-off-by: Shingo Omura <[email protected]>

Make test only helpers private (sysfspath methods, touchFile, createS…

2bc45ca

…ymlink) Signed-off-by: Shingo Omura <[email protected]>

Commented out for unused sysfspath methods

7e816ca

Signed-off-by: Shingo Omura <[email protected]>

everpeace commented Jun 30, 2025

View reviewed changes

Remove unused sysfsPath's methods

d28791f

Signed-off-by: Shingo Omura <[email protected]>

	StandardDeviceAttributePCIeRoot = StandardDeviceAttributePrefix + "pcieRoot"
	StandardDeviceAttributePCIeRoot resourceapi.QualifiedName = StandardDeviceAttributePrefix + "pcieRoot"

	func TouchFile(t *testing.T, path string) {
	func touchFile(t *testing.T, path string) {

	func CreateSymlink(t *testing.T, target, link string) {
	func createSymlink(t *testing.T, target, link string) {

	func (s sysfsPath) Devices(path string) string {
	func (s sysfsPath) devices(path string) string {

DRA: Introduce a helper function producing standardized 'resource.kubernetes.io/pcieRoot' device attributes for DRA drivers #132296

Are you sure you want to change the base?

DRA: Introduce a helper function producing standardized 'resource.kubernetes.io/pcieRoot' device attributes for DRA drivers #132296

Conversation

everpeace commented Jun 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What type of PR is this?

What this PR does / why we need it:

Which issue(s) this PR is related to:

Special notes for your reviewer:

Does this PR introduce a user-facing change?

Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.:

Uh oh!

k8s-ci-robot commented Jun 13, 2025

Uh oh!

k8s-ci-robot commented Jun 13, 2025

Uh oh!

k8s-ci-robot commented Jun 13, 2025

Uh oh!

everpeace commented Jun 13, 2025

Uh oh!

michaelasp commented Jun 13, 2025

Uh oh!

k8s-ci-robot commented Jun 13, 2025

Uh oh!

michaelasp left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

klueska Jun 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

everpeace Jun 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

everpeace Jun 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

everpeace Jun 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

everpeace Jun 27, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

klueska commented Jun 18, 2025

Uh oh!

pohly commented Jun 26, 2025

Uh oh!

klueska commented Jun 26, 2025

Uh oh!

everpeace commented Jun 13, 2025 •

edited

Loading

klueska Jun 18, 2025 •

edited

Loading

everpeace Jun 18, 2025 •

edited

Loading

everpeace Jun 18, 2025 •

edited

Loading

everpeace Jun 18, 2025 •

edited

Loading

everpeace Jun 27, 2025 •

edited

Loading