Skip to contentSkip to conversation

PCI: amlogic: Fix reset assertion via gpio descriptor

PATCH Applied Remi Pommarel 5 participants 7 messages Sep 1, 2019 – Oct 15, 2019

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 PCI: amlogic: Fix reset assertion via gpio descriptor

Normally asserting reset signal on gpio would be achieved with:

	gpiod_set_value_cansleep(reset_gpio, 1);

Meson PCI driver set reset value to '0' instead of '1' as it takes into account the PERST# signal polarity. The polarity should be taken care in the device tree instead.

This fixes the reset assertion meaning and moves out the polarity configuration in DT (please note that there is no DT currently using this driver).

---
 drivers/pci/controller/dwc/pci-meson.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/pci/controller/dwc/pci-meson.c +2 -2
index e35e9eaa50ee..541f37a6f6a5 100644
@@ -287,9 +287,9 @@ static inline void meson_cfg_writel(struct meson_pcie *mp, u32 val, u32 reg)
 
 static void meson_pcie_assert_reset(struct meson_pcie *mp)
 {
-	gpiod_set_value_cansleep(mp->reset_gpio, 0);
-	udelay(500);
 	gpiod_set_value_cansleep(mp->reset_gpio, 1);
+	udelay(500);
+	gpiod_set_value_cansleep(mp->reset_gpio, 0);
 }
 
 static void meson_pcie_init_dw(struct meson_pcie *mp)
Signature · 1 line
2.20.1
reviewer Reviewed-by on v1 1/1
#1 lore raw
Quoted 4 lines · Remi Pommarel

Normally asserting reset signal on gpio would be achieved with:

        gpiod_set_value_cansleep(reset_gpio, 1);
Quoted 8 lines

Meson PCI driver set reset value to '0' instead of '1' as it takes into account the PERST# signal polarity. The polarity should be taken care in the device tree instead.

This fixes the reset assertion meaning and moves out the polarity configuration in DT (please note that there is no DT currently using this driver).

Fixes 9c0ef6d34fdb ("PCI: amlogic: Add the Amlogic Meson PCIe

controller driver")

Quoted 1 line
Reviewed-by on v1 1/1
#2 lore raw
In reply to Remi Pommarel
Quoted 23 lines · Remi Pommarel

Normally asserting reset signal on gpio would be achieved with:

	gpiod_set_value_cansleep(reset_gpio, 1);

Meson PCI driver set reset value to '0' instead of '1' as it takes into account the PERST# signal polarity. The polarity should be taken care in the device tree instead.

This fixes the reset assertion meaning and moves out the polarity configuration in DT (please note that there is no DT currently using this driver).

---
 drivers/pci/controller/dwc/pci-meson.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/pci/controller/dwc/pci-meson.c +0 -0
index e35e9eaa50ee..541f37a6f6a5 100644
@@ -287,9 +287,9 @@ static inline void meson_cfg_writel(struct meson_pcie *mp, u32 val, u32 reg)
 

drivers/pci/controller/dwc/pci-meson.c static inline void meson_cfg_writel(struct meson_pcie *mp, u32 val, u32 reg) · line 295

Quoted 10 lines
drivers/pci/controller/dwc/pci-meson.c +2 -2
 static void meson_pcie_assert_reset(struct meson_pcie *mp)
 {
-	gpiod_set_value_cansleep(mp->reset_gpio, 0);
-	udelay(500);
 	gpiod_set_value_cansleep(mp->reset_gpio, 1);
+	udelay(500);
+	gpiod_set_value_cansleep(mp->reset_gpio, 0);
 }
 
 static void meson_pcie_init_dw(struct meson_pcie *mp)
#3 lore raw
In reply to Remi Pommarel
Quoted 10 lines · Remi Pommarel

Normally asserting reset signal on gpio would be achieved with:

	gpiod_set_value_cansleep(reset_gpio, 1);

Meson PCI driver set reset value to '0' instead of '1' as it takes into account the PERST# signal polarity. The polarity should be taken care in the device tree instead.

This fixes the reset assertion meaning and moves out the polarity configuration in DT (please note that there is no DT currently using this driver).

The device tree bindings for this give an example configuration:

        pcie: pcie@f9800000 {
                        compatible = "amlogic,axg-pcie", "snps,dw-pcie";
                        reg = <0x0 0xf9800000 0x0 0x400000
                                        0x0 0xff646000 0x0 0x2000
                                        0x0 0xff644000 0x0 0x2000
                                        0x0 0xf9f00000 0x0 0x100000>;
                        reg-names = "elbi", "cfg", "phy", "config";
                        reset-gpios = <&gpio GPIOX_19 GPIO_ACTIVE_HIGH>;

Is the 'reset-gpios' line still consistent with this change, or does this need to be updated as well?

Thanks,

Andrew Murray

Quoted 25 lines
---
 drivers/pci/controller/dwc/pci-meson.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/pci/controller/dwc/pci-meson.c +2 -2
index e35e9eaa50ee..541f37a6f6a5 100644
@@ -287,9 +287,9 @@ static inline void meson_cfg_writel(struct meson_pcie *mp, u32 val, u32 reg)
 
 static void meson_pcie_assert_reset(struct meson_pcie *mp)
 {
-	gpiod_set_value_cansleep(mp->reset_gpio, 0);
-	udelay(500);
 	gpiod_set_value_cansleep(mp->reset_gpio, 1);
+	udelay(500);
+	gpiod_set_value_cansleep(mp->reset_gpio, 0);
 }
 
 static void meson_pcie_init_dw(struct meson_pcie *mp)
Signature · 1 line
2.20.1
#4 lore raw
Quoted 23 lines · Andrew Murray
Quoted 10 lines · Remi Pommarel

Normally asserting reset signal on gpio would be achieved with:

	gpiod_set_value_cansleep(reset_gpio, 1);

Meson PCI driver set reset value to '0' instead of '1' as it takes into account the PERST# signal polarity. The polarity should be taken care in the device tree instead.

This fixes the reset assertion meaning and moves out the polarity configuration in DT (please note that there is no DT currently using this driver).

The device tree bindings for this give an example configuration:

        pcie: pcie@f9800000 {
                        compatible = "amlogic,axg-pcie", "snps,dw-pcie";
                        reg = <0x0 0xf9800000 0x0 0x400000
                                        0x0 0xff646000 0x0 0x2000
                                        0x0 0xff644000 0x0 0x2000
                                        0x0 0xf9f00000 0x0 0x100000>;
                        reg-names = "elbi", "cfg", "phy", "config";
                        reset-gpios = <&gpio GPIOX_19 GPIO_ACTIVE_HIGH>;
Quoted 2 lines

Is the 'reset-gpios' line still consistent with this change, or does this need to be updated as well?

Good catch, the polarity of the reset gpio will more likely be GPIO_ACTIVE_LOW.

Do you want a separate patch for that or can I just include it in v2 ?

Thanks.

Signature · 1 line
Remi
#5 lore raw
In reply to Andrew Murray

Hi Andrew,

Quoted 24 lines · Andrew Murray
Quoted 10 lines · Remi Pommarel

Normally asserting reset signal on gpio would be achieved with:

      gpiod_set_value_cansleep(reset_gpio, 1);

Meson PCI driver set reset value to '0' instead of '1' as it takes into account the PERST# signal polarity. The polarity should be taken care in the device tree instead.

This fixes the reset assertion meaning and moves out the polarity configuration in DT (please note that there is no DT currently using this driver).

The device tree bindings for this give an example configuration:

        pcie: pcie@f9800000 {
                        compatible = "amlogic,axg-pcie", "snps,dw-pcie";
                        reg = <0x0 0xf9800000 0x0 0x400000
                                        0x0 0xff646000 0x0 0x2000
                                        0x0 0xff644000 0x0 0x2000
                                        0x0 0xf9f00000 0x0 0x100000>;
                        reg-names = "elbi", "cfg", "phy", "config";
                        reset-gpios = <&gpio GPIOX_19 GPIO_ACTIVE_HIGH>;
Quoted 2 lines

Is the 'reset-gpios' line still consistent with this change, or does this need to be updated as well?

in my opinion the example is still valid
whether GPIO_ACTIVE_HIGH or GPIO_ACTIVE_LOW is correct depends on the actual circuit on the board:
- if the GPIO signal is directly connected to the PERST# line of the PCIe card then above example is correct
- if the GPIO signal is inverted, for example by using a transistor, then GPIO_ACTIVE_LOW should be used instead

I haven't looked into the schematics of the boards using a G12A or G12B SoC (I don't have any schematics of a board with an AXG SoC) so I can't tell what "most boards" use (active LOW or HIGH).
if there's a pattern in those board schematics (which is likely since most are derived from Amlogics reference design) then we can update the example based that.

Martin

#6 lore raw
In reply to Remi Pommarel
Quoted 7 lines · Remi Pommarel

Normally asserting reset signal on gpio would be achieved with:

	gpiod_set_value_cansleep(reset_gpio, 1);

Meson PCI driver set reset value to '0' instead of '1' as it takes into account the PERST# signal polarity. The polarity should be taken care in the device tree instead.

Quoted 8 lines

This fixes the reset assertion meaning and moves out the polarity configuration in DT (please note that there is no DT currently using this driver).

---
 drivers/pci/controller/dwc/pci-meson.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Applied to pci/meson, thanks.

Lorenzo

Quoted 19 lines
drivers/pci/controller/dwc/pci-meson.c +2 -2
index e35e9eaa50ee..541f37a6f6a5 100644
@@ -287,9 +287,9 @@ static inline void meson_cfg_writel(struct meson_pcie *mp, u32 val, u32 reg)
 
 static void meson_pcie_assert_reset(struct meson_pcie *mp)
 {
-	gpiod_set_value_cansleep(mp->reset_gpio, 0);
-	udelay(500);
 	gpiod_set_value_cansleep(mp->reset_gpio, 1);
+	udelay(500);
+	gpiod_set_value_cansleep(mp->reset_gpio, 0);
 }
 
 static void meson_pcie_init_dw(struct meson_pcie *mp)
Signature · 1 line
2.20.1