> (lkml.org seems to have lost August 3rd...)
> RFC:
http://lkml.indiana.edu/hypermail//linux/kernel/1008.0/01342.html
> v1:
http://lkml.indiana.edu/hypermail//linux/kernel/1008.0/01942.html
>
> Based on the idea and code originally proposed by Kevin Hilman:
>
http://www.mail-archive.com/linux-omap@vger.kernel.org/msg31161.html
>
> Changes from v1:
>
> * "Pseduo" buses are no longer init'd, they are [un]registered by:
> - pseudo_platform_bus_register(struct bus_type *)
> - pseudo_platform_bus_unregister(struct bus_type *)
> * These registrations [de]allocate a dev_pm_ops structure for the
> pseudo bus_type
> * Do not overwrite the parent if .bus is already set (that is, allow
> pseudo bus devices to be root devices)
>
> * Split into 2 patches:
> - 1/2: Already sent separately, but included here for clarity
> - 2/2: The real meat of the patch (this patch)
>
> INTRO
>
> As SOCs become more popular, the desire to quickly define a simple,
> but functional, bus type with only a few unique properties becomes
> desirable. As they become more complicated, the ability to nest these
> simple busses and otherwise orchestrate them to match the actual
> topology also becomes desirable.
>
> EXAMPLE USAGE
>
> /arch/ARCH/MY_ARCH/my_bus.c:
>
> #include <linux/device.h>
> #include <linux/platform_device.h>
>
> struct bus_type SOC_bus_type = {
> .name = "SOC-bus-type",
> };
> EXPORT_SYMBOL_GPL(SOC_bus_type);
>
> struct platform_device SOC_bus1 = {
> .name = "SOC-bus1",
> .id = -1,
> .dev.bus = &SOC_bus_type;
> };
> EXPORT_SYMBOL_GPL(SOC_bus1);
>
> struct platform_device SOC_bus2 = {
> .name = "SOC-bus2",
> .id = -2,
> .dev.bus = &SOC_bus_type;
> };
> EXPORT_SYMBOL_GPL(SOC_bus2);
>
> static int __init SOC_bus_init(void)
> {
> int error;
>
> error = pseudo_platform_bus_register(&SOC_bus_type);
> if (error)
> return error;
>
> error = platform_device_register(&SOC_bus1);
> if (error)
> goto fail_bus1;
>
> error = platform_device_register(&SOC_bus2);
> if (error)
> goto fail_bus2;
>
> return error;
>
> /* platform_device_unregister(&SOC_bus2); */
> fail_bus2:
> platform_device_unregister(&SOC_bus1);
> fail_bus1:
> pseudo_platform_bus_unregister(&SOC_bus_type);
>
> return error;
> }
>
> /drivers/my_driver.c:
> static struct platform_driver my_driver = {
> .driver = {
> .name = "my-driver",
> .owner = THIS_MODULE,
> .bus = &SOC_bus_type,
> },
> };
>
> /somewhere/my_device.c:
> static struct platform_device my_device = {
> .name = "my-device",
> .id = -1,
> .dev.bus = &my_bus_type,
> .dev.parent = &SOC_bus1.dev,
> };
>
> This will build a device tree that mirrors the actual system:
>
> /sys/bus
> |-- SOC-bus-type
> | |-- devices
> | | |-- SOC_bus1 -> ../../../devices/SOC_bus1
> | | |-- SOC_bus2 -> ../../../devices/SOC_bus2
> | | |-- my-device -> ../../../devices/SOC_bus1/my-device
> | |-- drivers
> | | |-- my-driver
>
> /sys/devices
> |-- SOC_bus1
> | |-- my-device
> |-- SOC_bus2
>
> Driver can drive any device on the SOC, which is logical, without
> actually being registered on multiple /bus_types/, even though the
> devices may be on different /physical buses/ (which are actually
> just devices).
>
> THOUGHTS:
>
> 1.
> Notice that for a device / driver, only 3 lines were added to
> switch from the platform bus to the new SOC_bus. This is
> especially valuable if we consider supporting a legacy SOCs
> and new SOCs where the same driver is used, but may need to
> be on either the platform bus or the new SOC_bus. The above
> code then becomes:
>
> (possibly in a header)
> #ifdef CONFIG_MY_BUS
> #define MY_BUS_TYPE &SOC_bus_type
> #else
> #define MY_BUS_TYPE NULL
> #endif
>
> /drivers/my_driver.c
> static struct platform_driver my_driver = {
> .driver = {
> .name = "my-driver",
> .owner = THIS_MODULE,
> .bus = MY_BUS_TYPE,
> },
> };
>
> Which will allow the same driver to easily to used on either
> the platform bus or the newly defined bus type.
>
> 2.
> Implementations wishing to make dynamic / run-time decisions on where
> devices are placed could easily create wrapper functions, that is
>
> int SOC_device_register(struct platform_device *pdev)
> {
> if (pdev->archdata->flag)
> pdev->dev.parent = &SOC_bus1.dev;
> else
> pdev->dev.parent = &SOC_bus2.dev;
>
> return platform_device_register(pdev);
> }
>
> A similar solution also would allow for run-time determination of dev.bus,
> if that were necessary for your platform.
>
> 3.
> I'm not convinced that dynamically allocating a new copy of dev_pm_ops is
> the best solution. I *AM*, however, convinced that removing const from
> struct bus_type {
> ...
> const struct dev_pm_ops *pm;
> ...
> };
> would be a mistake; it is far too easy to overwrite one of the function
> pointers on accident, and the const serves a very useful purpose here.
>
> Cc: Kevin Hilman <khilman@deeprootsystems.com>
> Signed-off-by: Patrick Pannuto <ppannuto@codeaurora.org>
> ---
> drivers/base/platform.c | 92 +++++++++++++++++++++++++++++++++++++-
> include/linux/platform_device.h | 3 +
> 2 files changed, 92 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/base/platform.c b/drivers/base/platform.c
> index b69ccb4..933e0c1 100644
> --- a/drivers/base/platform.c
> +++ b/drivers/base/platform.c
> @@ -238,8 +238,12 @@ int platform_device_add(struct platform_device *pdev)
> if (!pdev)
> return -EINVAL;
>
> - if (!pdev->dev.parent)
> - pdev->dev.parent = &platform_bus;
> + if (!pdev->dev.bus) {
> + pdev->dev.bus = &platform_bus_type;
> +
> + if (!pdev->dev.parent)
> + pdev->dev.parent = &platform_bus;
> + }
>
> pdev->dev.bus = &platform_bus_type;
