linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bjorn Helgaas <helgaas@kernel.org>
To: Aaron Sierra <asierra@xes-inc.com>
Cc: linux-pci@vger.kernel.org,
	"Rafael J. Wysocki" <rjw@rjwysocki.net>,
	Len Brown <lenb@kernel.org>
Subject: Re: [PATCH v2 1/2] PCI/ACPI: Move _OSC test for native services to top
Date: Wed, 30 Jan 2019 16:44:15 -0600	[thread overview]
Message-ID: <20190130224415.GK229773@google.com> (raw)
In-Reply-To: <1540483292-24049-2-git-send-email-asierra@xes-inc.com>

Hi Aaron,

On Thu, Oct 25, 2018 at 11:01:31AM -0500, Aaron Sierra wrote:
> Move the simple test for when PCIe native services are disabled
> closer to the top, prior to where things get more complicated.
> 
> Signed-off-by: Aaron Sierra <asierra@xes-inc.com>
> ---
>  drivers/acpi/pci_root.c | 11 ++++++-----
>  1 file changed, 6 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
> index 707aafc..eb9f14e 100644
> --- a/drivers/acpi/pci_root.c
> +++ b/drivers/acpi/pci_root.c
> @@ -441,6 +441,12 @@ static void negotiate_os_control(struct acpi_pci_root *root, int *no_aspm,
>  		return;
>  	}
>  
> +	if (pcie_ports_disabled) {
> +		dev_info(&device->dev,
> +			 "PCIe port services disabled; not requesting _OSC control\n");
> +		return;
> +	}

Today we always set "*no_aspm = 1" if _OSC fails, which means we later
call pcie_no_aspm().

After this patch, when pcie_ports_disabled is "true", we don't even try to
evaluate _OSC, and we will never set *no_aspm, so we will never call
pcie_no_aspm() when pcie_ports_disabled is "true", which happens in these
cases:

  1) CONFIG_PCIEPORTBUS is unset, or
  2) CONFIG_PCIEPORTBUS=y and we booted with "pcie_ports=compat"

Case 1) isn't a problem because pcie_no_aspm() is only implemented when
CONFIG_PCIEASPM=y, and CONFIG_PCIEASPM depends on CONFIG_PCIEPORTBUS, so in
this case today we only call the empty stub pcie_no_aspm() function.

But case 2) is a behavior change that seems unintended.

Even though CONFIG_PCIEASPM depends on CONFIG_PCIEPORTBUS, ASPM doesn't
actually *use* anything provided by PCIEPORTBUS, so I think the ASPM code
is still active and useful even when we boot with "pcie_ports=compat".

Whether CONFIG_PCIEASPM should depend on CONFIG_PCIEPORTBUS is another
question.  I tend to think maybe it should not, but that's an orthogonal
question.

>  	/*
>  	 * All supported architectures that use ACPI have support for
>  	 * PCI domains, so we indicate this in _OSC support capabilities.
> @@ -468,11 +474,6 @@ static void negotiate_os_control(struct acpi_pci_root *root, int *no_aspm,
>  		return;
>  	}
>  
> -	if (pcie_ports_disabled) {
> -		dev_info(&device->dev, "PCIe port services disabled; not requesting _OSC control\n");
> -		return;
> -	}
> -
>  	if ((support & ACPI_PCIE_REQ_SUPPORT) != ACPI_PCIE_REQ_SUPPORT) {
>  		decode_osc_support(root, "not requesting OS control; OS requires",
>  				   ACPI_PCIE_REQ_SUPPORT);
> -- 
> 2.7.4
> 

  reply	other threads:[~2019-01-30 22:53 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-15 22:23 [PATCH] PCI/ACPI: Improve _OSC control request granularity Aaron Sierra
2018-10-25 16:01 ` [PATCH v2 0/2] " Aaron Sierra
2018-10-25 16:01   ` [PATCH v2 1/2] PCI/ACPI: Move _OSC test for native services to top Aaron Sierra
2019-01-30 22:44     ` Bjorn Helgaas [this message]
2019-02-13 17:11       ` Aaron Sierra
2018-10-25 16:01   ` [PATCH v2 2/2] PCI/ACPI: Improve _OSC control request granularity Aaron Sierra
2019-01-30 22:57     ` Bjorn Helgaas
2019-02-13 17:31       ` Aaron Sierra
2019-02-13 17:36         ` Bjorn Helgaas
2019-02-13 21:32   ` [PATCH v3] " Aaron Sierra
2019-04-16 17:52     ` Aaron Sierra
2019-04-16 18:15       ` Bjorn Helgaas
2019-06-26 17:20     ` Bjorn Helgaas
2019-06-27  2:38       ` Aaron Sierra

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190130224415.GK229773@google.com \
    --to=helgaas@kernel.org \
    --cc=asierra@xes-inc.com \
    --cc=lenb@kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=rjw@rjwysocki.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).