[v3] PCI: apple: Set only available ports up

Commit Message

Janne Grunau March 23, 2023, 8:10 a.m. UTC

The Apple SoC devicetrees used to delete unused PCIe ports. Avoid to set
up disabled PCIe ports to keep the previous behaviour. MacOS initialized
also only ports with a known device.

Use for_each_available_child_of_node instead of for_each_child_of_node
which takes the "status" property into account.

Link: https://lore.kernel.org/asahi/[e-mail]/
Link: https://lore.kernel.org/asahi/[e-mail]/
Fixes: 1e33888fbe44 ("PCI: apple: Add initial hardware bring-up")
Reviewed-by: Marc Zyngier <[e-mail]>
Signed-off-by: Janne Grunau <[e-mail]>
---
Changes in v3:
- dropped Cc: stable
- rewritten commit message since the warning is fixed by 6fffbc7ae137 ("PCI: Honor firmware's
device disabled status")
- Link to v2: https://lore.kernel.org/r/[e-mail]

Changes in v2:
- rewritten commit message with more details and corrections
- collected Marc's "Reviewed-by:"
- Link to v1: https://lore.kernel.org/r/[e-mail]
---
 drivers/pci/controller/pcie-apple.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)


---
base-commit: c9c3395d5e3dcc6daee66c6908354d47bf98cb0c
change-id: 20230307-apple_pcie_disabled_ports-0c17fb7a4738

Best regards,

Comments

Eric Curtin March 23, 2023, 10:19 a.m. UTC | #1

On Thu, 23 Mar 2023 at 08:16, Janne Grunau <[e-mail]> wrote:
>
> The Apple SoC devicetrees used to delete unused PCIe ports. Avoid to set
> up disabled PCIe ports to keep the previous behaviour. MacOS initialized
> also only ports with a known device.
>
> Use for_each_available_child_of_node instead of for_each_child_of_node
> which takes the "status" property into account.
>
> Link: https://lore.kernel.org/asahi/[e-mail]/
> Link: https://lore.kernel.org/asahi/[e-mail]/
> Fixes: 1e33888fbe44 ("PCI: apple: Add initial hardware bring-up")
> Reviewed-by: Marc Zyngier <[e-mail]>
> Signed-off-by: Janne Grunau <[e-mail]>

Tested this on Asahi Fedora Remix kernel-edge on Mac Mini M1.

Tested-by: Eric Curtin <[e-mail]>

Is mise le meas/Regards,

Eric Curtin

> ---
> Changes in v3:
> - dropped Cc: stable
> - rewritten commit message since the warning is fixed by 6fffbc7ae137 ("PCI: Honor firmware's
> device disabled status")
> - Link to v2: https://lore.kernel.org/r/[e-mail]
>
> Changes in v2:
> - rewritten commit message with more details and corrections
> - collected Marc's "Reviewed-by:"
> - Link to v1: https://lore.kernel.org/r/[e-mail]
> ---
>  drivers/pci/controller/pcie-apple.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/pci/controller/pcie-apple.c b/drivers/pci/controller/pcie-apple.c
> index 66f37e403a09..f8670a032f7a 100644
> --- a/drivers/pci/controller/pcie-apple.c
> +++ b/drivers/pci/controller/pcie-apple.c
> @@ -783,7 +783,7 @@ static int apple_pcie_init(struct pci_config_window *cfg)
>         cfg->priv = pcie;
>         INIT_LIST_HEAD(&pcie->ports);
>
> -       for_each_child_of_node(dev->of_node, of_port) {
> +       for_each_available_child_of_node(dev->of_node, of_port) {
>                 ret = apple_pcie_setup_port(pcie, of_port);
>                 if (ret) {
>                         dev_err(pcie->dev, "Port %pOF setup fail: %d\n", of_port, ret);
>
> ---
> base-commit: c9c3395d5e3dcc6daee66c6908354d47bf98cb0c
> change-id: 20230307-apple_pcie_disabled_ports-0c17fb7a4738
>
> Best regards,
> --
> Janne Grunau <[e-mail]>
>
>

Bjorn Helgaas March 23, 2023, 4:05 p.m. UTC | #2

Hi Janne, thanks for the patch!

On Thu, Mar 23, 2023 at 09:10:12AM +0100, Janne Grunau wrote:
> The Apple SoC devicetrees used to delete unused PCIe ports.

Maybe "used to *omit* unused PCIe ports"?  I assume the DT previously
did not mention PCIe ports at all, but now the ports are included but
marked "disabled"?

> Avoid to set
> up disabled PCIe ports to keep the previous behaviour. MacOS initialized
> also only ports with a known device.

The patch is fine with me, but the commit log doesn't tell us why we
want this patch in addition to 6fffbc7ae137 ("PCI: Honor firmware's
device disabled status"), since your previous response [1] says
6fffbc7ae137 already does all the above.

I don't have any Apple specs, but from looking at the code, the
additional benefits of this patch over 6fffbc7ae137 might be:

  - We won't trip over DT issues, like "reset" or "reg" not being
    defined.

  - We won't allocate needless apple_pcie_port structs.

  - We won't ioremap register space, assert PERST#, setup clocks and
    IRQs.

  - We won't sleep waiting for things to come up, potentially saving
    150ms or more for each disabled port.

  - Maybe there's some power savings; I can't tell.

It's probably more detail than necessary to mention *all* of that, but
the commit log should say *something* about why we want this in
addition to 6fffbc7ae137.

> Use for_each_available_child_of_node instead of for_each_child_of_node
> which takes the "status" property into account.

Would you mind adding "()" after the function names, too?  Just my own
idiosyncrasy to help distinguish them from regular English text.

Bjorn

[1] https://lore.kernel.org/r/[e-mail]

> Link: https://lore.kernel.org/asahi/[e-mail]/
> Link: https://lore.kernel.org/asahi/[e-mail]/
> Fixes: 1e33888fbe44 ("PCI: apple: Add initial hardware bring-up")
> Reviewed-by: Marc Zyngier <[e-mail]>
> Signed-off-by: Janne Grunau <[e-mail]>
> ---
> Changes in v3:
> - dropped Cc: stable
> - rewritten commit message since the warning is fixed by 6fffbc7ae137 ("PCI: Honor firmware's
> device disabled status")
> - Link to v2: https://lore.kernel.org/r/[e-mail]
> 
> Changes in v2:
> - rewritten commit message with more details and corrections
> - collected Marc's "Reviewed-by:"
> - Link to v1: https://lore.kernel.org/r/[e-mail]
> ---
>  drivers/pci/controller/pcie-apple.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/controller/pcie-apple.c b/drivers/pci/controller/pcie-apple.c
> index 66f37e403a09..f8670a032f7a 100644
> --- a/drivers/pci/controller/pcie-apple.c
> +++ b/drivers/pci/controller/pcie-apple.c
> @@ -783,7 +783,7 @@ static int apple_pcie_init(struct pci_config_window *cfg)
>  	cfg->priv = pcie;
>  	INIT_LIST_HEAD(&pcie->ports);
>  
> -	for_each_child_of_node(dev->of_node, of_port) {
> +	for_each_available_child_of_node(dev->of_node, of_port) {
>  		ret = apple_pcie_setup_port(pcie, of_port);
>  		if (ret) {
>  			dev_err(pcie->dev, "Port %pOF setup fail: %d\n", of_port, ret);
> 
> ---
> base-commit: c9c3395d5e3dcc6daee66c6908354d47bf98cb0c
> change-id: 20230307-apple_pcie_disabled_ports-0c17fb7a4738
> 
> Best regards,
> -- 
> Janne Grunau <[e-mail]>
>

Patch

diff --git a/drivers/pci/controller/pcie-apple.c b/drivers/pci/controller/pcie-apple.c
index 66f37e403a09..f8670a032f7a 100644
--- a/drivers/pci/controller/pcie-apple.c
+++ b/drivers/pci/controller/pcie-apple.c
@@ -783,7 +783,7 @@  static int apple_pcie_init(struct pci_config_window *cfg)
 	cfg->priv = pcie;
 	INIT_LIST_HEAD(&pcie->ports);
 
-	for_each_child_of_node(dev->of_node, of_port) {
+	for_each_available_child_of_node(dev->of_node, of_port) {
 		ret = apple_pcie_setup_port(pcie, of_port);
 		if (ret) {
 			dev_err(pcie->dev, "Port %pOF setup fail: %d\n", of_port, ret);