Re: [PATCH v1 2/2] PCI: Allow user to request power management of conventional and hotplug bridges

From: valdis . kletnieks
Date: Mon Feb 19 2018 - 18:28:37 EST


On Mon, 19 Feb 2018 17:14:06 -0600, Bjorn Helgaas said:

> Change "pcie_port_pm=force" to enable power management of conventional PCI
> bridges and hotplug bridges as well as PCIe ports. As with the previous
> PCIe port-only behavior, this is not expected to work in all systems.

This part says the behavior changes - which is itself a Bad Idea unless you
have a deprecation cut-over across several releases. The general rule is that
you're not allowed to break somebody's kernel without a lot of warning.
Remember that there's probably a lot of embedded systems that hardcode their
boot cmdline and changing the behavior can result in a failed boot - which can
be a royal bitch to debug if the embedded system doesn't have a console.....

In addition, it doesn't match the actual patch, which documents the boot
parameter as being removed, rather than the behavior changed:

> diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
> index 1d1d53f85ddd..4660105ec851 100644
> --- a/Documentation/admin-guide/kernel-parameters.txt
> +++ b/Documentation/admin-guide/kernel-parameters.txt

> @@ -3143,10 +3147,6 @@
> compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
> ports driver.
>
> - pcie_port_pm= [PCIE] PCIe port power management handling:
> - off Disable power management of all PCIe ports
> - force Forcibly enable power management of all PCIe ports
> -

And *that* doesn't match the rest of the patch, which never touches the handling
of that parameter, either changing it or removing it.

Attachment: pgp4UNcMDc9Ua.pgp
Description: PGP signature