[PATCH 5.10] PCI: keystone: Drop __init from ks_pcie_add_pcie_{ep,port}()

[PATCH 5.10] PCI: keystone: Drop __init from ks_pcie_add_pcie_{ep,port}()

[Nathan Chancellor]

This commit has no upstream equivalent.

After commit db5ebaeb8fda ("PCI: keystone: Don't discard .probe()
callback") in 5.10, there are two modpost warnings when building with
clang:

  WARNING: modpost: vmlinux.o(.text+0x5aa6dc): Section mismatch in reference from the function ks_pcie_probe() to the function .init.text:ks_pcie_add_pcie_port()
  The function ks_pcie_probe() references
  the function __init ks_pcie_add_pcie_port().
  This is often because ks_pcie_probe lacks a __init
  annotation or the annotation of ks_pcie_add_pcie_port is wrong.

  WARNING: modpost: vmlinux.o(.text+0x5aa6f4): Section mismatch in reference from the function ks_pcie_probe() to the function .init.text:ks_pcie_add_pcie_ep()
  The function ks_pcie_probe() references
  the function __init ks_pcie_add_pcie_ep().
  This is often because ks_pcie_probe lacks a __init
  annotation or the annotation of ks_pcie_add_pcie_ep is wrong.

ks_pcie_add_pcie_ep() was removed in upstream commit a0fd361db8e5 ("PCI:
dwc: Move "dbi", "dbi2", and "addr_space" resource setup into common
code") and ks_pcie_add_pcie_port() was removed in upstream
commit 60f5b73fa0f2 ("PCI: dwc: Remove unnecessary wrappers around
dw_pcie_host_init()"), both of which happened before upstream
commit 7994db905c0f ("PCI: keystone: Don't discard .probe() callback").

As neither of these removal changes are really suitable for stable, just
remove __init from these functions in stable, as it is no longer a
correct annotation after dropping __init from ks_pcie_probe().

Fixes: db5ebaeb8fda ("PCI: keystone: Don't discard .probe() callback")
Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
---
This is not an issue in mainline but I still cc'd the author and
committer of 7994db905c0f in case they would like to check my analysis.
---
 drivers/pci/controller/dwc/pci-keystone.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/pci/controller/dwc/pci-keystone.c b/drivers/pci/controller/dwc/pci-keystone.c
index 5b722287aac9..afaea201a5af 100644
--- a/drivers/pci/controller/dwc/pci-keystone.c
+++ b/drivers/pci/controller/dwc/pci-keystone.c
@@ -865,8 +865,8 @@ static irqreturn_t ks_pcie_err_irq_handler(int irq, void *priv)
 	return ks_pcie_handle_error_irq(ks_pcie);
 }
 
-static int __init ks_pcie_add_pcie_port(struct keystone_pcie *ks_pcie,
-					struct platform_device *pdev)
+static int ks_pcie_add_pcie_port(struct keystone_pcie *ks_pcie,
+				 struct platform_device *pdev)
 {
 	struct dw_pcie *pci = ks_pcie->pci;
 	struct pcie_port *pp = &pci->pp;
@@ -978,8 +978,8 @@ static const struct dw_pcie_ep_ops ks_pcie_am654_ep_ops = {
 	.get_features = &ks_pcie_am654_get_features,
 };
 
-static int __init ks_pcie_add_pcie_ep(struct keystone_pcie *ks_pcie,
-				      struct platform_device *pdev)
+static int ks_pcie_add_pcie_ep(struct keystone_pcie *ks_pcie,
+			       struct platform_device *pdev)
 {
 	int ret;
 	struct dw_pcie_ep *ep;

---
base-commit: 479e8b8925415420b31e2aa65f9b0db3dea2adf4
change-id: 20231128-5-10-fix-pci-keystone-modpost-warning-a8e886bdc0f7

Best regards,
-- 
Nathan Chancellor <nathan@kernel.org>

Re: [PATCH 5.10] PCI: keystone: Drop __init from ks_pcie_add_pcie_{ep,port}()

[Uwe Kleine-König]

[-- Attachment #1: Type: text/plain, Size: 2160 bytes --]

On Tue, Nov 28, 2023 at 05:37:00PM -0700, Nathan Chancellor wrote:
> This commit has no upstream equivalent.
> 
> After commit db5ebaeb8fda ("PCI: keystone: Don't discard .probe()
> callback") in 5.10, there are two modpost warnings when building with

As with the 5.4 patch:

s/5.10/5.10.202/

> clang:
> 
>   WARNING: modpost: vmlinux.o(.text+0x5aa6dc): Section mismatch in reference from the function ks_pcie_probe() to the function .init.text:ks_pcie_add_pcie_port()
>   The function ks_pcie_probe() references
>   the function __init ks_pcie_add_pcie_port().
>   This is often because ks_pcie_probe lacks a __init
>   annotation or the annotation of ks_pcie_add_pcie_port is wrong.
> 
>   WARNING: modpost: vmlinux.o(.text+0x5aa6f4): Section mismatch in reference from the function ks_pcie_probe() to the function .init.text:ks_pcie_add_pcie_ep()
>   The function ks_pcie_probe() references
>   the function __init ks_pcie_add_pcie_ep().
>   This is often because ks_pcie_probe lacks a __init
>   annotation or the annotation of ks_pcie_add_pcie_ep is wrong.
> 
> ks_pcie_add_pcie_ep() was removed in upstream commit a0fd361db8e5 ("PCI:
> dwc: Move "dbi", "dbi2", and "addr_space" resource setup into common
> code") and ks_pcie_add_pcie_port() was removed in upstream
> commit 60f5b73fa0f2 ("PCI: dwc: Remove unnecessary wrappers around
> dw_pcie_host_init()"), both of which happened before upstream
> commit 7994db905c0f ("PCI: keystone: Don't discard .probe() callback").
> 
> As neither of these removal changes are really suitable for stable, just
> remove __init from these functions in stable, as it is no longer a
> correct annotation after dropping __init from ks_pcie_probe().
> 
> Fixes: db5ebaeb8fda ("PCI: keystone: Don't discard .probe() callback")
> Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
> Signed-off-by: Nathan Chancellor <nathan@kernel.org>

Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>

Thanks
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH 5.10] PCI: keystone: Drop __init from ks_pcie_add_pcie_{ep,port}()

[Greg KH]

On Wed, Nov 29, 2023 at 10:48:00AM +0100, Uwe Kleine-König wrote:
> On Tue, Nov 28, 2023 at 05:37:00PM -0700, Nathan Chancellor wrote:
> > This commit has no upstream equivalent.
> > 
> > After commit db5ebaeb8fda ("PCI: keystone: Don't discard .probe()
> > callback") in 5.10, there are two modpost warnings when building with
> 
> As with the 5.4 patch:
> 
> s/5.10/5.10.202/

Thanks, all now queued up.

greg k-h