Page MenuHomeFreeBSD

apm.4: Trim down to describing the compat interface exposed by acpi(4)
ClosedPublic

Authored by jhb on Wed, Sep 23, 7:06 PM.
Tags
None
Referenced Files
Unknown Object (File)
Tue, Oct 13, 6:34 PM
Unknown Object (File)
Tue, Oct 13, 6:52 AM
Unknown Object (File)
Mon, Oct 12, 1:46 PM
Unknown Object (File)
Mon, Oct 12, 12:16 PM
Unknown Object (File)
Mon, Oct 12, 6:19 AM
Unknown Object (File)
Mon, Oct 12, 12:18 AM
Unknown Object (File)
Sat, Oct 10, 10:12 PM
Unknown Object (File)
Fri, Oct 9, 12:51 PM
Subscribers

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

jhb requested review of this revision.Wed, Sep 23, 7:06 PM
jhb added a parent revision: D59937: apmd: Retire the APM daemon.

This is an alternative to D53724 (and follows the suggestions I gave on that review). See the surrounding stack for some additional context.

This revision is now accepted and ready to land.Wed, Sep 23, 9:07 PM
share/man/man4/apm.4
18

Nd is not capitalized

This revision now requires review to proceed.Fri, Sep 25, 7:44 PM
ziaee added inline comments.
share/man/man4/apm.4
14

So now that this works through acpi it can work on any acpi system like some aarch64?

This revision is now accepted and ready to land.Sat, Sep 26, 4:09 PM

Oh one more thing,

share/man/man4/apm.4
18

Oh I forgot, I'm going through and doing the SYNOPSIS now. So we don't have a device apm anymore but to use this the kernel needs acpi now, right?

share/man/man4/apm.4
18

Yes, but that is a bit disingenuous as this isn't the main feature page for acpi(4). We don't put device acpi in the synopsis of acpi_battery(4) for example, or device pci in the synopsis of every PCI device driver.

share/man/man4/apm.4
18

We don't put device acpi in the synopsis of acpi_battery(4) for example, or device pci in the synopsis of every PCI device driver.

From my interpretation, this is obviously a bug, but it affects so few people nobody cares. The devices should list all the needed devices.

share/man/man4/apm.4
18

That doesn't really scale though as you might have a foo driver with multiple attachments (e.g. uart can attach to either isa or acpi) but not all parent buses are required, so it would be incorrect for uart's manpage to list both device pci and device isa (which is actually device acpi on modern x86 systems) implying that both are required when they are in fact not always required. I think the manual for a given driver should focus on that driver's configuration and cross-reference any parent busses whose pages can describe how to enable that parent bus in the bus's respective manual page.

share/man/man4/apm.4
14

We don't compile the file for this except on x86 given it is a legacy interface. If we wanted to have apm(8) work on aarch64, I'd probably do that by converting apm(8) to use ACPI ioctls directly leaving this only as a compatability shim for any ports that might be using this interface still.

share/man/man4/apm.4
14

Ok, in that case, please use the optional architecture specifier here.

18

Okay, that makes sense. Thanks John!

jhb marked 6 inline comments as done.Wed, Sep 30, 1:58 PM