Skip to contentSkip to conversation

ARM: integrator: delete non-devicetree boot path

PATCH Superseded Linus Walleij 2 participants 3 messages Oct 3, 2013 – Oct 10, 2013

Subsystems (2)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH 3/6 ARM: integrator: delete non-devicetree boot path

The Device Tree boot path now supports everything the ATAG boot can provide, and the two are equivalent. This deletes the ATAG boot path from the Integrator/AP and Integrator/CP platforms to move them on to the future.

---
 arch/arm/Kconfig                         |   1 +
 arch/arm/mach-integrator/core.c          |  57 ----------
 arch/arm/mach-integrator/integrator_ap.c | 135 ------------------------
 arch/arm/mach-integrator/integrator_cp.c | 173 -------------------------------
 arch/arm/mach-integrator/pci_v3.c        | 122 +++++-----------------
 5 files changed, 25 insertions(+), 463 deletions(-)
arch/arm/Kconfig +1 -0
index 1ad6fb6..d9ce868 100644
@@ -317,6 +317,7 @@ config ARCH_INTEGRATOR
 	select NEED_MACH_MEMORY_H
 	select PLAT_VERSATILE
 	select SPARSE_IRQ
+	select USE_OF
 	select VERSATILE_FPGA_IRQ
 	help
 	  Support for ARM's Integrator platform.
arch/arm/mach-integrator/core.c +0 -57

63 lines not on this page · raw message

arch/arm/mach-integrator/integrator_ap.c +0 -135

144 lines not on this page · raw message

arch/arm/mach-integrator/integrator_cp.c +0 -173

182 lines not on this page · raw message

arch/arm/mach-integrator/pci_v3.c +24 -98

146 lines not on this page · raw message

Signature · 1 line
1.8.3.1
#1 lore raw

PATCH 3/6 ARM: integrator: delete non-devicetree boot path

[Adding devicetree to Cc]

Quoted 14 lines · Linus Walleij

The Device Tree boot path now supports everything the ATAG boot can provide, and the two are equivalent. This deletes the ATAG boot path from the Integrator/AP and Integrator/CP platforms to move them on to the future.

---
 arch/arm/Kconfig                         |   1 +
 arch/arm/mach-integrator/core.c          |  57 ----------
 arch/arm/mach-integrator/integrator_ap.c | 135 ------------------------
 arch/arm/mach-integrator/integrator_cp.c | 173 -------------------------------
 arch/arm/mach-integrator/pci_v3.c        | 122 +++++-----------------
 5 files changed, 25 insertions(+), 463 deletions(-)

[...]

arch/arm/mach-integrator/pci_v3.c static int __init pci_v3_map_irq_dt(const struct pci_dev *dev, u8 slot, u8 pin) · line 846

Quoted 10 lines
arch/arm/mach-integrator/pci_v3.c +4 -0
+static int __init pci_v3_probe(struct platform_device *pdev)
 {
+       struct device_node *np = pdev->dev.of_node;
        struct of_pci_range_parser parser;
        struct of_pci_range range;
        struct resource *res;
        int irq, ret;
 
+       /* Remap the Integrator system controller */
+       ap_syscon_base = ioremap(INTEGRATOR_SC_BASE, 0x100);

It seems a shame that this didn't get described in the DT. This hasn't been fully moved over...

arch/arm/mach-integrator/pci_v3.c static int __init pci_v3_map_irq_dt(const struct pci_dev *dev, u8 slot, u8 pin) · line 856

Quoted 10 lines
arch/arm/mach-integrator/pci_v3.c +10 -0
+       if (!ap_syscon_base) {
+               dev_err(&pdev->dev, "unable to remap the AP syscon for PCIv3\n");
+               return -ENODEV;
+       }
+
+       /* Device tree probe path */
+       if (!np) {
+               dev_err(&pdev->dev, "no device tree node for PCIv3\n");
+               return -ENODEV;
+       }

We didn't unmap the syscon registers, yet we mapped them here. Either we've got a useless mapping, or we're mapping it from the wrong block of code...

Cheers,
Mark.

#2 lore raw

PATCH 3/6 ARM: integrator: delete non-devicetree boot path

Quoted 5 lines · Mark Rutland

arch/arm/mach-integrator/pci_v3.c static int __init pci_v3_map_irq_dt(const struct pci_dev *dev, u8 slot, u8 pin) · line 846

Quoted 2 lines
arch/arm/mach-integrator/pci_v3.c +2 -0
+       /* Remap the Integrator system controller */
+       ap_syscon_base = ioremap(INTEGRATOR_SC_BASE, 0x100);

It seems a shame that this didn't get described in the DT. This hasn't been fully moved over...

No. There is still stuff to move over, and not only the system controller, also the EBI (external bus interface) on the AP.

But we can still work from this point, or do you mean that we cannot proceed to delete the ATAG boot path before we fix this?

My idea is that deleting this makes the code easier to handle after the shrink.

Quoted 11 lines

arch/arm/mach-integrator/pci_v3.c static int __init pci_v3_map_irq_dt(const struct pci_dev *dev, u8 slot, u8 pin) · line 856

Quoted 10 lines
arch/arm/mach-integrator/pci_v3.c +10 -0
+       if (!ap_syscon_base) {
+               dev_err(&pdev->dev, "unable to remap the AP syscon for PCIv3\n");
+               return -ENODEV;
+       }
+
+       /* Device tree probe path */
+       if (!np) {
+               dev_err(&pdev->dev, "no device tree node for PCIv3\n");
+               return -ENODEV;
+       }
Quoted 3 lines

We didn't unmap the syscon registers, yet we mapped them here. Either we've got a useless mapping, or we're mapping it from the wrong block of code...

Sorry I'm not following this. The only thing the patch does is remove the ATAG boot path, moving the relevant part of the DT boot path over to the DT probe function and renaming it pci_v3_probe() in the process which might make the patch messy but that is all it does.

If you mean that we should unmap the syscon base you're right but that would be a separate patch, right?

I'll write a separate patch for that instead using devm_* so we don't have to unmap it by calling any special function.

Yours,
Linus Walleij