mctpd: update supports_vdm check#141
Conversation
|
Looks good - just a formatting failure in CI. |
|
Also, you may want to check you git author/commiter settings; it looks like you have a username there instead of a full name. Perhaps: - but if the |
|
Thanks for the update. The failing test is due to the actual bug; I have some testcase changes that I will include in the merge. |
I updated the test. Looks like some format issue in test file. Will check. |
|
Don't worry about the test changes: I have some reworks done on how the VDM types are handled within the Endpoint code, which I will merge alongside your change. Might be helpful to add a Fixes line to your commit though: |
|
This still has the testcase changes though. |
|
OK, I have pushed your branch to my full set of changes (including your fix, and a test for this case). Please let me know if that looks OK to you. |
|
Looks good. I think we can merge it. |
Rather than dealing with lists of numbers, provide a proper class for a VDM type, including a serialiser for Get Vendor Message Type support responses. Signed-off-by: Jeremy Kerr <jk@codeconstruct.com.au>
Currently, we use a default [0] for our types list, but do not inspect the vdm types to augment with 7e/7f. Use the provided VDM types to also populate the base types by default, but still allow an explicit types argument to override. Signed-off-by: Jeremy Kerr <jk@codeconstruct.com.au>
…comparisons Now that we have a VDM type, construct a set from dbus responses, and use this to copmpare against expected values. Signed-off-by: Jeremy Kerr <jk@codeconstruct.com.au>
Updated supports_vdm check to be decided based on supported PCIE/IANA message type instead of formatID. [Minor commit message changes from Jeremy Kerr] Fixes: cb182be ("mctpd: Add VendorDefinedMessageTypes property") Signed-off-by: rutaabhyankar <ruta.abhyankar@intel.com> Signed-off-by: Jeremy Kerr <jk@codeconstruct.com.au>
We have a bug where the presence of the control type will trigger a VDM query, rather than 7e/7f. Explicitly clear the types set, to check for this case. Signed-off-by: Jeremy Kerr <jk@codeconstruct.com.au>
2 participants
supports_vdm check to be decided based on supported PCIE/IANA message type instead of formatID.
Fixes: cb182be ("mctpd: Add VendorDefinedMessageTypes property")