1320944 - Add the new Linux CPU microcode information from json_dump to SuperSearch

Status: RESOLVED DUPLICATE of bug 1791742

(Socorro :: General, task)

Description

[Marco Castelluccio [:marco]]

+++ This bug was initially created as a clone of Bug #1305888 +++

Bug 1320922 will add a cpu_microcode_version field to json_dump. We should add it to SuperSearch.
For Windows, cpu_microcode_version has to be populated from the annotation, for Linux from the json_dump field.

Comment 1

[Marco Castelluccio [:marco]]

Bug 1320922 has been fixed, so this can be done now.

Comment 2

[Peter Bengtsson [:peterbe]]

Let's not add any more super search fields until we can do it with the checked in .json file.

Comment 3

[Gabriele Svelto [:gsvelto]]

I was wondering what would be the best way to tackle this in the current setup. The minidump processor schema already has a field for it (system_info.cpu_microcode_version) and this field is already populated on Linux. On Windows however we use the CPUMicrocodeVersion annotation which is also what Socorro uses right now.

I filed issue #421 in rust-minidump to use that field to populate system_info.cpu_microcode_version. Would it be hard to support both for some time? That is to fetch the microcode version from system_info.cpu_microcode_version when available and fall back to the CPUMicrocodeVersion annotation if not. That would allow us to support Linux and then gradually phase out the CPUMicrocodeVersion Windows annotation having all platforms rely on system_info.cpu_microcode_version.

Will WDYT?

Comment 4

[Will Kahn-Greene [:willkg] ET needinfo? me]

I can definitely support "pull from field X and degrade to field Y" kinds of things. That's pretty straight-forward to do.

Comment 5

[Will Kahn-Greene [:willkg] ET needinfo? me]

I didn't remember this bug, so we had a few like this. I'm duping this to the bug we did the work in.