Hi Sergei!
So now encm3_platform_init function is linked to the kernel image -- and
is part of the executable -- i added this line to the
arch/mips/au1000/encm3/Makefile:
obj-y +=encm3_platform.o
However, I now get a "Cant analyze prologue code at 80294aec." error!
Any remedies/suggestions for the same?
Thank you!
Best Regards,
Ashlesha.
On Wed, 2006-12-06 at 22:54 +0300, Sergei Shtylyov wrote:
> Hello.
>
> Ashlesha Shintre wrote:
>
> > There is already an interrupt handler in place for the AU1000_GPIO_0
> > that takes care of the cascaded interrupts -- so I *can* say
> > .irq= AU1000_GPIO_0 ----right?
>
> No, you can't. You'll have to specify to what 8259's IRQ4 maps to on your
> platform.
>
> > Also, how will I make sure my board specific encm3_platform_init is
> > called during the arch init calls?
>
> Mentioning it in arch_initcall() arranges for that. :-/
>
> > I have put in an entry in the Makefile for the board specific
> > encm3_platform.c file -- so it is built - but when control goes to the
> > static int __devinit serial8250_probe(struct device *dev) function in
> > the 8250.c it never executes the serial8250_register_port function.
> > I know this cus I m using the JTAG port on the board to look inside and
> > step through the code..
>
> That's strange. Although the UART declaration has a grave defect....
>
> > Here is my /arch/mips/au1000/encm3/encm3_platform.c file:
>
> >>/*
> >> * Platform device support for Au1x00 SoCs.
> >> *
> >> * Copyright 2004, Matt Porter <mporter@xxxxxxxxxxxxxxxxxxx>
> >> *
> >> * This file is licensed under the terms of the GNU General Public
> >> * License version 2. This program is licensed "as is" without any
> >> * warranty of any kind, whether express or implied.
> >> */
>
> That boilerplate is no longer applicable. :-)
>
> >>#include <linux/device.h>
> >>#include <linux/kernel.h>
> >>#include <linux/init.h>
> >>#include <linux/resource.h>
> >>#include <linux/serial_8250.h>
> >>#include <linux/tty.h>
> >>
> >>#include <asm/mach-au1x00/au1000.h>
> >>#include <asm/mach-encm3/encm3.h>
> >>static struct plat_serial8250_port encm3_via_uart_data[] = {
> >> {
> >> .mapbase =
> >>0x3f8, //resource base
>
> Damn, I didn't notice: .mapbase should be changed to .iobase!
>
> >>// .membase = (char *)(0x50000000 +
> >>0x3f8), // is a pointer - ioremap cookie or NULL
> >> .irq = AU1000_GPIO_0,
> >> .flags = UPF_SHARE_IRQ, //|
> >>UPF_IOREMAP, //UPF_BOOT_AUTOCONF | UPF_SKIP_TEST |
> >> .iotype = UPIO_PORT,
> >> .regshift = 1,
> >> .uartclk = 1843200,
> >>
> >> },
> >> { },
> >>};
>
> >>static struct resource encm3_via_uart_resource = {
> >> .start = VIA_COM1_ADDR,
> >> .end = VIA_COM1_ADDR + 0x7,
> >> .flags = IORESOURCE_IO,
> >>};
>
> Still, you don't need to declare the resources for the 8250 devices -- the
> driver should handle requesting them for you -- as they're alredy specified by
> struct plat_serial8250_port.
>
> >>static struct platform_device encm3_via_uart = {
> >> .name = "serial8250",
> >> .id = 1,
>
> I guess it should be PLAT8250_DEV_LEGACY...
>
> >> .dev = {
> >> .platform_data = encm3_via_uart_data,
> >> },
>
> So, you also don't need the following 2 lines:
>
> >> .num_resources = 1,
> >> .resource = &encm3_via_uart_resource,
> >>};
>
> >>static struct platform_device *encm3_platform_devices[] __initdata = {
> >> &encm3_via_uart,
> >>};
>
> >>int encm3_platform_init(void)
> >>{
> >> printk("size of encm3 platform devices is %d
> >>\n",ARRAY_SIZE(encm3_platform_devices));
> >> return platform_add_devices(encm3_platform_devices,
> >>ARRAY_SIZE(encm3_platform_devices));
>
> I think it's better to call platform_device_register() for a single device...
>
> WBR, Sergei
>
