[PATCH net-next v2 0/3] Unify platform suspend/resume routines for PCI DWMAC glue

[PATCH net-next v2 0/3] Unify platform suspend/resume routines for PCI DWMAC glue

[Yao Zi]

There are currently three PCI-based DWMAC glue drivers in tree,
stmmac_pci.c, dwmac-intel.c, and dwmac-loongson.c. Both stmmac_pci.c and
dwmac-intel.c implements the same and duplicated platform suspend/resume
routines.

This series introduces a new PCI helper library, stmmac_libpci.c,
providing a pair of helpers, stmmac_pci_plat_{suspend,resume}, and
replaces the driver-specific implementation with the helpers to reduce
code duplication. The helper will also simplify the Motorcomm DWMAC glue
driver which I'm working on.

The glue driver for Intel controllers isn't covered by the series, since
its suspend routine doesn't call pci_disable_device() and thus is a
little different from the new generic helpers.

I only have Loongson hardware on hand, thus the series is only tested on
Loongson 3A5000 machine. I could confirm the controller works after
resume, and WoL works as expected. This shouldn't break stmmac_pci.c,
either, since the new helpers have the exactly same code as the old
driver-specific suspend/resume hooks.

Changed from v1
- Separate the new suspend/resume helpers into a new file,
  stmmac_libpci.c, and provide Kconfig symbol for it
- Link to v1: https://lore.kernel.org/netdev/20251028154332.59118-1-ziyao@disroot.org/

Yao Zi (3):
  net: stmmac: Add generic suspend/resume helper for PCI-based
    controllers
  net: stmmac: loongson: Use generic PCI suspend/resume routines
  net: stmmac: pci: Use generic PCI suspend/resume routines

 drivers/net/ethernet/stmicro/stmmac/Kconfig   | 11 +++++
 drivers/net/ethernet/stmicro/stmmac/Makefile  |  1 +
 .../ethernet/stmicro/stmmac/dwmac-loongson.c  | 36 ++------------
 .../ethernet/stmicro/stmmac/stmmac_libpci.c   | 48 +++++++++++++++++++
 .../ethernet/stmicro/stmmac/stmmac_libpci.h   | 12 +++++
 .../net/ethernet/stmicro/stmmac/stmmac_pci.c  | 36 ++------------
 6 files changed, 78 insertions(+), 66 deletions(-)
 create mode 100644 drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.c
 create mode 100644 drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.h

-- 
2.51.2

[PATCH net-next v2 1/3] net: stmmac: Add generic suspend/resume helper for PCI-based controllers

[Yao Zi]

Most glue driver for PCI-based DWMAC controllers utilize similar
platform suspend/resume routines. Add a generic implementation to reduce
duplicated code.

Signed-off-by: Yao Zi <ziyao@disroot.org>
---
 drivers/net/ethernet/stmicro/stmmac/Kconfig   |  9 ++++
 drivers/net/ethernet/stmicro/stmmac/Makefile  |  1 +
 .../ethernet/stmicro/stmmac/stmmac_libpci.c   | 48 +++++++++++++++++++
 .../ethernet/stmicro/stmmac/stmmac_libpci.h   | 12 +++++
 4 files changed, 70 insertions(+)
 create mode 100644 drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.c
 create mode 100644 drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.h

diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
index 87c5bea6c2a2..598bc56edd8d 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
+++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
@@ -349,6 +349,15 @@ config DWMAC_VISCONTI
 
 endif
 
+config STMMAC_LIBPCI
+	tristate "STMMAC PCI helper library"
+	depends on PCI
+	depends on STMMAC_ETH
+	default y
+	help
+	  This selects the PCI bus helpers for the stmmac driver. If you
+	  have a controller with PCI interface, say Y or M here.
+
 config DWMAC_INTEL
 	tristate "Intel GMAC support"
 	default X86
diff --git a/drivers/net/ethernet/stmicro/stmmac/Makefile b/drivers/net/ethernet/stmicro/stmmac/Makefile
index 1681a8a28313..7bf528731034 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Makefile
+++ b/drivers/net/ethernet/stmicro/stmmac/Makefile
@@ -44,6 +44,7 @@ obj-$(CONFIG_DWMAC_VISCONTI)	+= dwmac-visconti.o
 stmmac-platform-objs:= stmmac_platform.o
 dwmac-altr-socfpga-objs := dwmac-socfpga.o
 
+obj-$(CONFIG_STMMAC_LIBPCI)	+= stmmac_libpci.o
 obj-$(CONFIG_STMMAC_PCI)	+= stmmac-pci.o
 obj-$(CONFIG_DWMAC_INTEL)	+= dwmac-intel.o
 obj-$(CONFIG_DWMAC_LOONGSON)	+= dwmac-loongson.o
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.c
new file mode 100644
index 000000000000..5c5dd502f79a
--- /dev/null
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.c
@@ -0,0 +1,48 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * PCI bus helpers for STMMAC driver
+ * Copyright (C) 2025 Yao Zi <ziyao@disroot.org>
+ */
+
+#include <linux/device.h>
+#include <linux/pci.h>
+
+#include "stmmac_libpci.h"
+
+int stmmac_pci_plat_suspend(struct device *dev, void *bsp_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	int ret;
+
+	ret = pci_save_state(pdev);
+	if (ret)
+		return ret;
+
+	pci_disable_device(pdev);
+	pci_wake_from_d3(pdev, true);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(stmmac_pci_plat_suspend);
+
+int stmmac_pci_plat_resume(struct device *dev, void *bsp_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	int ret;
+
+	pci_restore_state(pdev);
+	pci_set_power_state(pdev, PCI_D0);
+
+	ret = pci_enable_device(pdev);
+	if (ret)
+		return ret;
+
+	pci_set_master(pdev);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(stmmac_pci_plat_resume);
+
+MODULE_DESCRIPTION("STMMAC PCI helper library");
+MODULE_AUTHOR("Yao Zi <ziyao@disroot.org>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.h b/drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.h
new file mode 100644
index 000000000000..71553184f982
--- /dev/null
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_libpci.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2025 Yao Zi <ziyao@disroot.org>
+ */
+
+#ifndef __STMMAC_LIBPCI_H__
+#define __STMMAC_LIBPCI_H__
+
+int stmmac_pci_plat_suspend(struct device *dev, void *bsp_priv);
+int stmmac_pci_plat_resume(struct device *dev, void *bsp_priv);
+
+#endif /* __STMMAC_LIBPCI_H__ */
-- 
2.51.2

[PATCH net-next v2 2/3] net: stmmac: loongson: Use generic PCI suspend/resume routines

[Yao Zi]

Convert glue driver for Loongson DWMAC controller to use the generic
platform suspend/resume routines for PCI controllers, instead of
implementing its own one.

Signed-off-by: Yao Zi <ziyao@disroot.org>
---
 drivers/net/ethernet/stmicro/stmmac/Kconfig   |  1 +
 .../ethernet/stmicro/stmmac/dwmac-loongson.c  | 36 ++-----------------
 2 files changed, 4 insertions(+), 33 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
index 598bc56edd8d..4b6911c62e6f 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
+++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
@@ -373,6 +373,7 @@ config DWMAC_LOONGSON
 	default MACH_LOONGSON64
 	depends on (MACH_LOONGSON64 || COMPILE_TEST) && STMMAC_ETH && PCI
 	depends on COMMON_CLK
+	depends on STMMAC_LIBPCI
 	help
 	  This selects the LOONGSON PCI bus support for the stmmac driver,
 	  Support for ethernet controller on Loongson-2K1000 SoC and LS7A1000 bridge.
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-loongson.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-loongson.c
index 2a3ac0136cdb..584dc4ff8320 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-loongson.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-loongson.c
@@ -8,6 +8,7 @@
 #include <linux/device.h>
 #include <linux/of_irq.h>
 #include "stmmac.h"
+#include "stmmac_libpci.h"
 #include "dwmac_dma.h"
 #include "dwmac1000.h"
 
@@ -525,37 +526,6 @@ static int loongson_dwmac_fix_reset(struct stmmac_priv *priv, void __iomem *ioad
 				  10000, 2000000);
 }
 
-static int loongson_dwmac_suspend(struct device *dev, void *bsp_priv)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	int ret;
-
-	ret = pci_save_state(pdev);
-	if (ret)
-		return ret;
-
-	pci_disable_device(pdev);
-	pci_wake_from_d3(pdev, true);
-	return 0;
-}
-
-static int loongson_dwmac_resume(struct device *dev, void *bsp_priv)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	int ret;
-
-	pci_restore_state(pdev);
-	pci_set_power_state(pdev, PCI_D0);
-
-	ret = pci_enable_device(pdev);
-	if (ret)
-		return ret;
-
-	pci_set_master(pdev);
-
-	return 0;
-}
-
 static int loongson_dwmac_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 {
 	struct plat_stmmacenet_data *plat;
@@ -600,8 +570,8 @@ static int loongson_dwmac_probe(struct pci_dev *pdev, const struct pci_device_id
 	plat->bsp_priv = ld;
 	plat->setup = loongson_dwmac_setup;
 	plat->fix_soc_reset = loongson_dwmac_fix_reset;
-	plat->suspend = loongson_dwmac_suspend;
-	plat->resume = loongson_dwmac_resume;
+	plat->suspend = stmmac_pci_plat_suspend;
+	plat->resume = stmmac_pci_plat_resume;
 	ld->dev = &pdev->dev;
 	ld->loongson_id = readl(res.addr + GMAC_VERSION) & 0xff;
 
-- 
2.51.2

[PATCH net-next v2 3/3] net: stmmac: pci: Use generic PCI suspend/resume routines

[Yao Zi]

Convert STMMAC PCI glue driver to use the generic platform
suspend/resume routines for PCI controllers, instead of implementing its
own one.

Signed-off-by: Yao Zi <ziyao@disroot.org>
---
 drivers/net/ethernet/stmicro/stmmac/Kconfig   |  1 +
 .../net/ethernet/stmicro/stmmac/stmmac_pci.c  | 36 ++-----------------
 2 files changed, 4 insertions(+), 33 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
index 4b6911c62e6f..eec81f674857 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
+++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
@@ -382,6 +382,7 @@ config STMMAC_PCI
 	tristate "STMMAC PCI bus support"
 	depends on STMMAC_ETH && PCI
 	depends on COMMON_CLK
+	depends on STMMAC_LIBPCI
 	help
 	  This selects the platform specific bus support for the stmmac driver.
 	  This driver was tested on XLINX XC2V3000 FF1152AMT0221
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_pci.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_pci.c
index 94b3a3b27270..fa92be672c54 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_pci.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_pci.c
@@ -14,6 +14,7 @@
 #include <linux/dmi.h>
 
 #include "stmmac.h"
+#include "stmmac_libpci.h"
 
 struct stmmac_pci_info {
 	int (*setup)(struct pci_dev *pdev, struct plat_stmmacenet_data *plat);
@@ -139,37 +140,6 @@ static const struct stmmac_pci_info snps_gmac5_pci_info = {
 	.setup = snps_gmac5_default_data,
 };
 
-static int stmmac_pci_suspend(struct device *dev, void *bsp_priv)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	int ret;
-
-	ret = pci_save_state(pdev);
-	if (ret)
-		return ret;
-
-	pci_disable_device(pdev);
-	pci_wake_from_d3(pdev, true);
-	return 0;
-}
-
-static int stmmac_pci_resume(struct device *dev, void *bsp_priv)
-{
-	struct pci_dev *pdev = to_pci_dev(dev);
-	int ret;
-
-	pci_restore_state(pdev);
-	pci_set_power_state(pdev, PCI_D0);
-
-	ret = pci_enable_device(pdev);
-	if (ret)
-		return ret;
-
-	pci_set_master(pdev);
-
-	return 0;
-}
-
 /**
  * stmmac_pci_probe
  *
@@ -249,8 +219,8 @@ static int stmmac_pci_probe(struct pci_dev *pdev,
 	plat->safety_feat_cfg->prtyen = 1;
 	plat->safety_feat_cfg->tmouten = 1;
 
-	plat->suspend = stmmac_pci_suspend;
-	plat->resume = stmmac_pci_resume;
+	plat->suspend = stmmac_pci_plat_suspend;
+	plat->resume = stmmac_pci_plat_resume;
 
 	return stmmac_dvr_probe(&pdev->dev, plat, &res);
 }
-- 
2.51.2

Re: [PATCH net-next v2 2/3] net: stmmac: loongson: Use generic PCI suspend/resume routines

[Maxime Chevallier]

Hi,

On 30/10/2025 05:19, Yao Zi wrote:
> Convert glue driver for Loongson DWMAC controller to use the generic
> platform suspend/resume routines for PCI controllers, instead of
> implementing its own one.
> 
> Signed-off-by: Yao Zi <ziyao@disroot.org>
> ---
>  drivers/net/ethernet/stmicro/stmmac/Kconfig   |  1 +
>  .../ethernet/stmicro/stmmac/dwmac-loongson.c  | 36 ++-----------------
>  2 files changed, 4 insertions(+), 33 deletions(-)
> 
> diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> index 598bc56edd8d..4b6911c62e6f 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
> +++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> @@ -373,6 +373,7 @@ config DWMAC_LOONGSON
>  	default MACH_LOONGSON64
>  	depends on (MACH_LOONGSON64 || COMPILE_TEST) && STMMAC_ETH && PCI
>  	depends on COMMON_CLK
> +	depends on STMMAC_LIBPCI

If we go with a dedicated module for this, "select STMMAC_LIBPCI" would
make more sense here I think. The same applies for the next patch.

Maxime

Re: [PATCH net-next v2 2/3] net: stmmac: loongson: Use generic PCI suspend/resume routines

[Yao Zi]

On Thu, Oct 30, 2025 at 08:38:12AM +0100, Maxime Chevallier wrote:
> Hi,
> 
> On 30/10/2025 05:19, Yao Zi wrote:
> > Convert glue driver for Loongson DWMAC controller to use the generic
> > platform suspend/resume routines for PCI controllers, instead of
> > implementing its own one.
> > 
> > Signed-off-by: Yao Zi <ziyao@disroot.org>
> > ---
> >  drivers/net/ethernet/stmicro/stmmac/Kconfig   |  1 +
> >  .../ethernet/stmicro/stmmac/dwmac-loongson.c  | 36 ++-----------------
> >  2 files changed, 4 insertions(+), 33 deletions(-)
> > 
> > diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > index 598bc56edd8d..4b6911c62e6f 100644
> > --- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > +++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > @@ -373,6 +373,7 @@ config DWMAC_LOONGSON
> >  	default MACH_LOONGSON64
> >  	depends on (MACH_LOONGSON64 || COMPILE_TEST) && STMMAC_ETH && PCI
> >  	depends on COMMON_CLK
> > +	depends on STMMAC_LIBPCI
> 
> If we go with a dedicated module for this, "select STMMAC_LIBPCI" would
> make more sense here I think. The same applies for the next patch.

Thanks, it's a reasonable point. I worried about possible unmet
dependencies caused by a select, but STMMAC_LIBPCI only depends on PCI
and STMMAC_ETH, which are necessary for DWMAC_LOONGSON and STMMAC_PCI
to show up.

I will change it to "select STMMAC_LIBPCI" in v3.

> Maxime
> 

Best regards,
Yao Zi

Re: [PATCH net-next v2 2/3] net: stmmac: loongson: Use generic PCI suspend/resume routines

[Russell King (Oracle)]

On Thu, Oct 30, 2025 at 08:38:12AM +0100, Maxime Chevallier wrote:
> Hi,
> 
> On 30/10/2025 05:19, Yao Zi wrote:
> > Convert glue driver for Loongson DWMAC controller to use the generic
> > platform suspend/resume routines for PCI controllers, instead of
> > implementing its own one.
> > 
> > Signed-off-by: Yao Zi <ziyao@disroot.org>
> > ---
> >  drivers/net/ethernet/stmicro/stmmac/Kconfig   |  1 +
> >  .../ethernet/stmicro/stmmac/dwmac-loongson.c  | 36 ++-----------------
> >  2 files changed, 4 insertions(+), 33 deletions(-)
> > 
> > diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > index 598bc56edd8d..4b6911c62e6f 100644
> > --- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > +++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > @@ -373,6 +373,7 @@ config DWMAC_LOONGSON
> >  	default MACH_LOONGSON64
> >  	depends on (MACH_LOONGSON64 || COMPILE_TEST) && STMMAC_ETH && PCI
> >  	depends on COMMON_CLK
> > +	depends on STMMAC_LIBPCI
> 
> If we go with a dedicated module for this, "select STMMAC_LIBPCI" would
> make more sense here I think. The same applies for the next patch.

Yes, we need it to be this way around at least initially so that
STMMAC_LIBPCI gets merged into people's configs.

I'd eventually suggest going the other way.

We already have:

config STMMAC_PLATFORM
        tristate "STMMAC Platform bus support"

if STMMAC_PLATFORM

... platform based drivers ...

endif

... three PCI based drivers ...

I'd suggest we do:

config STMMAC_LIBPCI
	tristate "STMMAC PCI bus support"
	depends on PCI
	...

if STMMAC_LIBPCI

... PCI based drivers ...

endif

There's no need to make everything depend on STMMAC_ETH, there's an
outer "if STMMAC_ETH" around all the platforms already.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!

Re: [PATCH net-next v2 2/3] net: stmmac: loongson: Use generic PCI suspend/resume routines

[Yao Zi]

On Thu, Oct 30, 2025 at 09:30:35AM +0000, Russell King (Oracle) wrote:
> On Thu, Oct 30, 2025 at 08:38:12AM +0100, Maxime Chevallier wrote:
> > Hi,
> > 
> > On 30/10/2025 05:19, Yao Zi wrote:
> > > Convert glue driver for Loongson DWMAC controller to use the generic
> > > platform suspend/resume routines for PCI controllers, instead of
> > > implementing its own one.
> > > 
> > > Signed-off-by: Yao Zi <ziyao@disroot.org>
> > > ---
> > >  drivers/net/ethernet/stmicro/stmmac/Kconfig   |  1 +
> > >  .../ethernet/stmicro/stmmac/dwmac-loongson.c  | 36 ++-----------------
> > >  2 files changed, 4 insertions(+), 33 deletions(-)
> > > 
> > > diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > > index 598bc56edd8d..4b6911c62e6f 100644
> > > --- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > > +++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > > @@ -373,6 +373,7 @@ config DWMAC_LOONGSON
> > >  	default MACH_LOONGSON64
> > >  	depends on (MACH_LOONGSON64 || COMPILE_TEST) && STMMAC_ETH && PCI
> > >  	depends on COMMON_CLK
> > > +	depends on STMMAC_LIBPCI
> > 
> > If we go with a dedicated module for this, "select STMMAC_LIBPCI" would
> > make more sense here I think. The same applies for the next patch.
> 
> Yes, we need it to be this way around at least initially so that
> STMMAC_LIBPCI gets merged into people's configs.
> 
> I'd eventually suggest going the other way.
> 
> We already have:
> 
> config STMMAC_PLATFORM
>         tristate "STMMAC Platform bus support"
> 
> if STMMAC_PLATFORM
> 
> ... platform based drivers ...
> 
> endif
> 
> ... three PCI based drivers ...
> 
> I'd suggest we do:
> 
> config STMMAC_LIBPCI
> 	tristate "STMMAC PCI bus support"
> 	depends on PCI
> 	...
> 
> if STMMAC_LIBPCI
> 
> ... PCI based drivers ...
> 
> endif

Okay, will take this scheme instead.

> There's no need to make everything depend on STMMAC_ETH, there's an
> outer "if STMMAC_ETH" around all the platforms already.

Okay. Yes, this dependency is redundant. Should we remove the
unnecessary depends from other Kconfig options, too? STMMAC_SELFTESTS,
STMMAC_PLATFORM, DWMAC_INTEL_PLAT, DWMAC_INTEL, DWMAC_LOONGSON and
STMMAC_PCI are all enclosed by "if STMMAC_ETH" but carry this redundant
dependency, too.

Regards,
Yao Zi

> -- 
> RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
> FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!