Page MenuHomeFreeBSD

Remove PCI IDs from aacraid driver for new controller hardware
ClosedPublic

Authored by scott.benesh_microsemi.com on Oct 28 2016, 6:57 PM.
Tags
None
Referenced Files
Unknown Object (File)
Sun, Dec 8, 1:39 PM
Unknown Object (File)
Nov 22 2024, 10:37 PM
Unknown Object (File)
Nov 16 2024, 7:48 AM
Unknown Object (File)
Sep 30 2024, 8:17 PM
Unknown Object (File)
Sep 26 2024, 7:02 PM
Unknown Object (File)
Sep 26 2024, 8:46 AM
Unknown Object (File)
Sep 25 2024, 9:10 AM
Unknown Object (File)
Sep 25 2024, 3:06 AM
Subscribers

Details

Summary

This patch removes PCI IDs from the aacraid driver. These IDs are being re-used in new controller hardware which will require a new driver.

Test Plan

With the IDs removed, the updated aacraid driver has been tested to make sure it doesn't load on the new hardware.

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
Lint Skipped
Unit
Tests Skipped

Event Timeline

scott.benesh_microsemi.com retitled this revision from to Remove PCI IDs from aacraid driver for new controller hardware.
scott.benesh_microsemi.com updated this object.
scott.benesh_microsemi.com edited the test plan for this revision. (Show Details)
imp added a reviewer: imp.

categorically reject.

You need to do this cleanly.

This revision is now accepted and ready to land.Oct 28 2016, 6:58 PM
imp requested changes to this revision.Oct 28 2016, 6:59 PM
imp edited edge metadata.

I mean X not check.

Do not do this. This is removing support for working hardware.
Make the new driver return a higher value than the old one.
And document it so people with this old hardware still in service aren't screwed.

This revision now requires changes to proceed.Oct 28 2016, 6:59 PM

This hardware has not been released yet. Isn't this the best time to do this?

This hardware has not been released yet. Isn't this the best time to do this?

The Adaptec RAID Controller has been released, and has been in service for many years. I know the new hardware hasn't been released yet, but that's not relevant.

What's best is to not release hardware that reuses IDs. Nobody else does that.

However, if you are going to do this, the new driver that you're writing for the hardware must coordinate with this driver and return a higher probe value than BUS_PROBE_DEFAULT that this driver does. Perhaps controllable via a tunable so people that update to the new system with old hardware can set it after seeing it documented in the release notes so they don't suddenly have a system that they can't boot from.

imp edited edge metadata.

Ah, I see from other emails which I wasn't CC'd on was that this hardware never made it out of my lab. I thought it was a repeat of a different situation... Sorry for being grumpy...

Just make sure the commit message is clear that this is for hardware that was never released. It should also be MFC'd to stable/11 and stable/10 as a guide to anybody that might backport it.

This revision is now accepted and ready to land.Oct 28 2016, 7:22 PM

No problems. Will make sure the message is clear that this is for new hardware that has never been released.

Remove a PCI ID for a raid controller from Adaptec that was planned,
but never released. Since no real hardware was released with this ID,
just drop it from the aacraid driver. This paves the path for future
drivers for hardware that actually has this ID.

https://svnweb.freebsd.org/changeset/base/308065 has the commit, I'm surprised it didn't auto close this review.