Page MenuHomeFreeBSD

loader: Add a list of firmware name mapping
AcceptedPublic

Authored by manu on Fri, Dec 27, 8:51 AM.
Tags
None
Referenced Files
Unknown Object (File)
Tue, Dec 31, 5:18 PM
Unknown Object (File)
Fri, Dec 27, 8:53 AM
Subscribers

Details

Reviewers
bz
kevans
imp
Group Reviewers
wireless
Summary

Since we started to ship raw firmware for iwm(4), users who loads
the driver from loader are having problems as loader don't know that
the firmwares are now raw files and not kernel modules anymore.
Start a list of default entry for iwm(4) firmwares name mapping so it will
still works when loaded from loader.

Sponsored by: Beckhoff Automation GmbH & Co. KG

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Passed
Unit
No Test Coverage
Build Status
Buildable 61365
Build 58249: arc lint + arc unit

Event Timeline

manu requested review of this revision.Fri, Dec 27, 8:51 AM
bz added a subscriber: bz.

While I still think we are digging deeper holes now than the one you asked me to dig a year ago when on core this will definitively help users and go in accordance with the man page so should be transparent.

This revision is now accepted and ready to land.Fri, Dec 27, 9:22 AM

We need to enhance firmware(9) to allow drivers to request a load yhen get a callback later in boot.

This is an ok workaround for the moment though.

kevans added inline comments.
stand/defaults/loader.conf
187

/boot/firmware probably should have been added to module_path instead until we grow some notion of loader-only module_path additions

Add /boot/firmware to module_path so we can remove the filename mapping part.

This revision now requires review to proceed.Sat, Dec 28, 8:44 AM
manu marked an inline comment as done.Sat, Dec 28, 8:44 AM
stand/defaults/loader.conf
187

Done now, I think that I'll also work on detecting type based on the path (dtb/firmware etc ...)

kevans added inline comments.
stand/defaults/loader.conf
187

Thanks; no objection to landing this now if you'd prefer, though. I can see a couple of ways to implement some type autodetection that could be handy for more general use.

This revision is now accepted and ready to land.Sat, Dec 28, 4:01 PM

Same a Kyle. Don't want to get in the way, but i have ideas on path driven default type