Re: [PATCH 35/41] drivers: tty: serial: 8250: add mapsize to platform data
"Enrico Weigelt, metux IT consult" writes:
> Adding a mapsize field for the 8250 port platform data struct,
> so we can now set the resource size (eg. *1) and don't need
> funny runtime detections like serial8250_port_size() anymore.
>
> For now, serial8250_port_size() is called everytime we need
> the io resource size. That function checks which chip we
> actually have and returns the appropriate size. This approach
> is a bit clumpsy and not entirely easy to understand, and
> it's a violation of layers :p
>
> Obviously, that information cannot change after the driver init,
> so we can safely do that probing once on driver init and just
> use the stored value later.
>
> *1) arch/mips/alchemy/common/platform.c
>
> Signed-off-by: Enrico Weigelt
> ---
> drivers/tty/serial/8250/8250_core.c | 1 +
> include/linux/serial_8250.h | 1 +
> 2 files changed, 2 insertions(+)
>
> diff --git a/drivers/tty/serial/8250/8250_core.c
> b/drivers/tty/serial/8250/8250_core.c
> index e441221..71a398b 100644
> --- a/drivers/tty/serial/8250/8250_core.c
> +++ b/drivers/tty/serial/8250/8250_core.c
> @@ -814,6 +814,7 @@ static int serial8250_probe(struct platform_device *dev)
> uart.port.iotype= p->iotype;
> uart.port.flags = p->flags;
> uart.port.mapbase = p->mapbase;
> + uart.port.mapsize = p->mapsize;
> uart.port.hub6 = p->hub6;
> uart.port.private_data = p->private_data;
> uart.port.type = p->type;
> diff --git a/include/linux/serial_8250.h b/include/linux/serial_8250.h
> index 5a655ba..8b8183a 100644
> --- a/include/linux/serial_8250.h
> +++ b/include/linux/serial_8250.h
> @@ -22,6 +22,7 @@ struct plat_serial8250_port {
> unsigned long iobase; /* io base address */
> void __iomem*membase; /* ioremap cookie or NULL */
> resource_size_t mapbase;/* resource base */
> + resource_size_t mapsize;/* resource size */
> unsigned intirq;/* interrupt number */
> unsigned long irqflags; /* request_irq flags */
> unsigned intuartclk;/* UART clock rate */
Or replace iobase, mapbase and mapsize with a struct resource value?
/Esben
[PATCH 35/41] drivers: tty: serial: 8250: add mapsize to platform data
Adding a mapsize field for the 8250 port platform data struct,
so we can now set the resource size (eg. *1) and don't need
funny runtime detections like serial8250_port_size() anymore.
For now, serial8250_port_size() is called everytime we need
the io resource size. That function checks which chip we
actually have and returns the appropriate size. This approach
is a bit clumpsy and not entirely easy to understand, and
it's a violation of layers :p
Obviously, that information cannot change after the driver init,
so we can safely do that probing once on driver init and just
use the stored value later.
*1) arch/mips/alchemy/common/platform.c
Signed-off-by: Enrico Weigelt
---
drivers/tty/serial/8250/8250_core.c | 1 +
include/linux/serial_8250.h | 1 +
2 files changed, 2 insertions(+)
diff --git a/drivers/tty/serial/8250/8250_core.c
b/drivers/tty/serial/8250/8250_core.c
index e441221..71a398b 100644
--- a/drivers/tty/serial/8250/8250_core.c
+++ b/drivers/tty/serial/8250/8250_core.c
@@ -814,6 +814,7 @@ static int serial8250_probe(struct platform_device *dev)
uart.port.iotype= p->iotype;
uart.port.flags = p->flags;
uart.port.mapbase = p->mapbase;
+ uart.port.mapsize = p->mapsize;
uart.port.hub6 = p->hub6;
uart.port.private_data = p->private_data;
uart.port.type = p->type;
diff --git a/include/linux/serial_8250.h b/include/linux/serial_8250.h
index 5a655ba..8b8183a 100644
--- a/include/linux/serial_8250.h
+++ b/include/linux/serial_8250.h
@@ -22,6 +22,7 @@ struct plat_serial8250_port {
unsigned long iobase; /* io base address */
void __iomem*membase; /* ioremap cookie or NULL */
resource_size_t mapbase;/* resource base */
+ resource_size_t mapsize;/* resource size */
unsigned intirq;/* interrupt number */
unsigned long irqflags; /* request_irq flags */
unsigned intuartclk;/* UART clock rate */
--
1.9.1
