| Message ID | 1598893093-14280-6-git-send-email-tlfalcon@linux.ibm.com |
|---|---|
| State | Changes Requested |
| Delegated to: | David Miller |
| Headers | show |
| Series | ibmvnic: Report ACL settings in sysfs | expand |
On Mon, 31 Aug 2020 11:58:13 -0500 Thomas Falcon wrote: > Provide documentation for ibmvnic device Access Control List > files. What API is used to set those parameters in the first place?
On 8/31/20 2:26 PM, Jakub Kicinski wrote: > On Mon, 31 Aug 2020 11:58:13 -0500 Thomas Falcon wrote: >> Provide documentation for ibmvnic device Access Control List >> files. > What API is used to set those parameters in the first place? > These parameters are specified in the system's IBM Hardware Management Console (HMC) when the VNIC device is create.
On Mon, 31 Aug 2020 14:54:06 -0500 Thomas Falcon wrote: > On 8/31/20 2:26 PM, Jakub Kicinski wrote: > > On Mon, 31 Aug 2020 11:58:13 -0500 Thomas Falcon wrote: > >> Provide documentation for ibmvnic device Access Control List > >> files. > > What API is used to set those parameters in the first place? > > > These parameters are specified in the system's IBM Hardware Management > Console (HMC) when the VNIC device is create. The new attributes are visible in the "guest" OS, correct? This seems similar to normal SR-IOV operation, but I've not heard of use cases for them VM to know what its pvid is. Could you elaborate?
On 8/31/20 3:11 PM, Jakub Kicinski wrote: > On Mon, 31 Aug 2020 14:54:06 -0500 Thomas Falcon wrote: >> On 8/31/20 2:26 PM, Jakub Kicinski wrote: >>> On Mon, 31 Aug 2020 11:58:13 -0500 Thomas Falcon wrote: >>>> Provide documentation for ibmvnic device Access Control List >>>> files. >>> What API is used to set those parameters in the first place? >>> >> These parameters are specified in the system's IBM Hardware Management >> Console (HMC) when the VNIC device is create. > The new attributes are visible in the "guest" OS, correct? Correct. > > This seems similar to normal SR-IOV operation, but I've not heard of > use cases for them VM to know what its pvid is. Could you elaborate? It's provided for informational purposes.
On Mon, 31 Aug 2020 16:44:06 -0500 Thomas Falcon wrote: > On 8/31/20 3:11 PM, Jakub Kicinski wrote: > > This seems similar to normal SR-IOV operation, but I've not heard of > > use cases for them VM to know what its pvid is. Could you elaborate? > It's provided for informational purposes. Seems like an information leak :S and since it's equivalent to the standard SR-IOV functionality - we'd strongly prefer a common interface for all use cases. sysfs won't be it. Jiri & Mellanox had been working on something in devlink for quite some time.
From: Jakub Kicinski <kuba@kernel.org> Date: Mon, 31 Aug 2020 15:00:50 -0700 > On Mon, 31 Aug 2020 16:44:06 -0500 Thomas Falcon wrote: >> On 8/31/20 3:11 PM, Jakub Kicinski wrote: >> > This seems similar to normal SR-IOV operation, but I've not heard of >> > use cases for them VM to know what its pvid is. Could you elaborate? >> It's provided for informational purposes. > > Seems like an information leak :S and since it's equivalent to the > standard SR-IOV functionality - we'd strongly prefer a common > interface for all use cases. sysfs won't be it. Jiri & Mellanox had > been working on something in devlink for quite some time. Agreed, Thomas please work with Jiri et al. so that you can provide this information using a standard facility. Thanks.
On 8/31/20 5:17 PM, David Miller wrote: > From: Jakub Kicinski <kuba@kernel.org> > Date: Mon, 31 Aug 2020 15:00:50 -0700 > >> On Mon, 31 Aug 2020 16:44:06 -0500 Thomas Falcon wrote: >>> On 8/31/20 3:11 PM, Jakub Kicinski wrote: >>>> This seems similar to normal SR-IOV operation, but I've not heard of >>>> use cases for them VM to know what its pvid is. Could you elaborate? >>> It's provided for informational purposes. >> Seems like an information leak :S and since it's equivalent to the >> standard SR-IOV functionality - we'd strongly prefer a common >> interface for all use cases. sysfs won't be it. Jiri & Mellanox had >> been working on something in devlink for quite some time. > Agreed, Thomas please work with Jiri et al. so that you can provide > this information using a standard facility. > > Thanks. Thank you for the feedback. I will look into that.
diff --git a/Documentation/ABI/testing/sysfs-driver-ibmvnic b/Documentation/ABI/testing/sysfs-driver-ibmvnic index 7fa2920..8c78312 100644 --- a/Documentation/ABI/testing/sysfs-driver-ibmvnic +++ b/Documentation/ABI/testing/sysfs-driver-ibmvnic @@ -12,3 +12,29 @@ Description: If the ibmvnic device has been configured with redundant returned. Users: Any users of the ibmvnic driver which use redundant hardware configurations. + +What: /sys/devices/vio/<our device>/mac_acls +Date: August 2020 +KernelVersion: 5.10 +Contact: linuxppc-dev@lists.ozlabs.org +Description: Read-only file which lists the current entries in the ibmvnic + device's MAC address Access Control List. Each entry is + separated by a new line. +Users: Any users of the ibmvnic driver + +What: /sys/devices/vio/<our device>/vlan_acls +Date: August 2020 +KernelVersion: 5.10 +Contact: linuxppc-dev@lists.ozlabs.org +Description: Read-only file which lists the current entries in the ibmvnic + device's VLAN ID Access Control List. Each entry is separated + by a new line. +Users: Any users of the ibmvnic driver + +What: /sys/devices/vio/<our device>/pvid +Date: August 2020 +KernelVersion: 5.10 +Contact: linuxppc-dev@lists.ozlabs.org +Description: Read-only file which lists the ibmvnic device's Port VLAN + ID and Priority setting. Each entry is separated by a new line. +Users: Any users of the ibmvnic driver
Provide documentation for ibmvnic device Access Control List files. Signed-off-by: Thomas Falcon <tlfalcon@linux.ibm.com> --- Documentation/ABI/testing/sysfs-driver-ibmvnic | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+)