Hi Omar,
In kernel commit bf0beec0607d ("blk-mq: drain I/O when all CPUs in a
hctx are offline"), we added support to ensure CPU hotplug works safely
for devices with managed interrupts.
I was using my own script to test this, and I wanted to add something to
blktests for this same thing - however I see block/008 is for that
purpose. However, it does not seem effective for testing the above
commit, as it just hot-unplugs and replugs random CPUs quickly, and does
reliably ensure a specific group of CPUs are disabled for such a period
that we would see IO timeouts (or not).
Anyway, would you suggest a whole new test for this?
Cheers,
John
