nvme: Use ISR exclusion to make reset safer
AbandonedPublic
Actions

Authored by imp on Oct 10 2022, 10:24 PM.

Details

Reviewers

mav
chs
chuck
jhb

Summary

Before we send a hardware reset to the controller, mark all the admin
and I/O queues as being 'in the interrupt handler' so that we know all
interrupt handlers have excited. Once the hardware reset is complete,
unblock the interrupt hanlders so that we can bring the card back
up. This replaces a 100ms sleep that we use for this purpose today,
potentially making reset a little faster as well as safer.

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Skipped

Unit

Tests Skipped

Build Status

Buildable 53077
Build 49968: arc lint + arc unit

Event Timeline

imp created this revision.Oct 10 2022, 10:24 PM

Herald added a subscriber: dab. · View Herald TranscriptOct 10 2022, 10:24 PM

imp requested review of this revision.Oct 10 2022, 10:24 PM

Harbormaster completed remote builds in B47769: Diff 111624.Oct 10 2022, 10:24 PM

imp edited the test plan for this revision. (Show Details)Oct 10 2022, 10:29 PM

imp added reviewers: mav, chs, chuck.

imp added inline comments.Oct 10 2022, 10:46 PM

sys/dev/nvme/nvme_ctrlr.c
1234	the last bit could be a separate commit, or I could drop it.

Add a timeout for waiting for the ISR to exit of 100ms. And fail the driver /
controller if it is stuck. I've not seen this, but I got to wondering what would
happen if...

Harbormaster completed remote builds in B47771: Diff 111630.Oct 10 2022, 11:43 PM

adam.e.peace_gmail.com added a subscriber: adam.e.peace_gmail.com.Jan 31 2023, 10:30 PM

rebase

Harbormaster completed remote builds in B53077: Diff 125939.Aug 12 2023, 4:43 PM

emaste added a subscriber: emaste.Aug 12 2023, 5:14 PM

imp added a reviewer: jhb.Aug 13 2023, 4:04 PM