Thankyou for the comments.
On 08/05/13 17:18, Arnd Bergmann wrote:
> On Wednesday 08 May 2013, Srinivas KANDAGATLA wrote:
>> From: Srinivas Kandagatla <[email protected]>
>>
>> The STiH415 is the next generation of HD, AVC set-top box processors for
>> satellite, cable, terrestrial and IP-STB markets. It is an ARM Cortex-A9
>> 1.0 GHz, dual-core CPU.
>>
>> Signed-off-by: Srinivas Kandagatla <[email protected]>
>> CC: Stephen Gallimore <[email protected]>
>> CC: Stuart Menefy <[email protected]>
>> ---
>> Documentation/arm/STiH41x/overview.txt | 39 ++
>> Documentation/arm/STiH41x/stih415-overview.txt | 12 +
>> arch/arm/Kconfig | 3 +
>> arch/arm/Kconfig.debug | 38 ++
>> arch/arm/Makefile | 2 +
>> arch/arm/boot/dts/stih415-clock.dtsi | 38 ++
>> arch/arm/boot/dts/stih415-pinctrl.dtsi | 480
>> ++++++++++++++++++++++++
>> arch/arm/boot/dts/stih415.dtsi | 94 +++++
>> arch/arm/boot/dts/stih415.h | 20 +
>> arch/arm/boot/dts/stih41x.dtsi | 30 ++
>> arch/arm/boot/dts/stixxxx-pincfg.h | 95 +++++
>> arch/arm/configs/stih41x_defconfig | 94 +++++
>> arch/arm/include/debug/stixxxx.S | 61 +++
>> arch/arm/mach-stih41x/Kconfig | 35 ++
>> arch/arm/mach-stih41x/Makefile | 4 +
>> arch/arm/mach-stih41x/board-dt.c | 76 ++++
>> arch/arm/mach-stih41x/stih41x.c | 82 ++++
>> arch/arm/mach-stih41x/stih41x.h | 7 +
>> arch/arm/plat-stixxxx/Kconfig | 2 +
>> arch/arm/plat-stixxxx/Makefile | 2 +
>> arch/arm/plat-stixxxx/headsmp.S | 44 +++
>> arch/arm/plat-stixxxx/include/plat/hardware.h | 20 +
>> arch/arm/plat-stixxxx/include/plat/smp.h | 19 +
>> arch/arm/plat-stixxxx/platsmp.c | 144 +++++++
> I think there is no point in having a separate mach- and plat- directory
> here, given that modern machines require very little code. You can probably
> fold the three files in mach-stih41x (board-dt.c, stih41x.c and stih41x.h)
> into a single file and rename plat-stixxxx to mach-stixxxx
Yep, sounds good.
>
>> diff --git a/arch/arm/Kconfig.debug b/arch/arm/Kconfig.debug
>> index 9b31f43..42a5193 100644
>> --- a/arch/arm/Kconfig.debug
>> +++ b/arch/arm/Kconfig.debug
>> @@ -447,6 +447,16 @@ choice
>> This option selects UART0 on VIA/Wondermedia System-on-a-chip
>> devices, including VT8500, WM8505, WM8650 and WM8850.
>>
>> + config DEBUG_STIH41X_UART
>> + depends on ARCH_STIH41X
>> + bool "Use StiH415/416 ASC for low-level debug"
>> + help
>> + Say Y here if you want kernel low-level debugging support
>> + on StiH415/416 based platforms like B2000, B2020.
>> + It support UART2 and SBC_UART1.
>> +
>> + If unsure, say N.
>> +
>> config DEBUG_LL_UART_NONE
>> bool "No low-level debugging UART"
>> depends on !ARCH_MULTIPLATFORM
> I would split out the debug code into a separate patch though.
Will do this in next version.
>
>> @@ -600,6 +637,7 @@ config DEBUG_LL_INCLUDE
>> default "debug/vt8500.S" if DEBUG_VT8500_UART0
>> default "debug/tegra.S" if DEBUG_TEGRA_UART
>> default "debug/zynq.S" if DEBUG_ZYNQ_UART0 || DEBUG_ZYNQ_UART1
>> + default "debug/stixxxx.S" if DEBUG_STIH41X_UART
>> default "mach/debug-macro.S"
> Please keep this in alphabetical order.
Yes.
>
>> + soc {
>> + #address-cells = <1>;
>> + #size-cells = <1>;
>> + #interrupt-cells = <3>;
>> + interrupt-parent = <&intc>;
> The "#interrupt-cells" property doesn't belong here I think.
>
>> +
>> + uart2: uart@fed32000 {
>> + compatible = "st,asc";
>> + status = "disabled";
>> + reg = <0xfed32000 0x2c>;
>> + interrupts = <0 197 0>;
>> + pinctrl-names = "default";
>> + pinctrl-0 = <&pinctrl_uart2>;
>> + clocks = <&CLKS_ICN_REG_0>;
>> + };
>> +
>> + /* SBC comms block ASCs in SASG1 */
>> + sbc_uart1: uart@fe531000 {
>> + compatible = "st,asc";
>> + status = "disabled";
>> + reg = <0xfe531000 0x2c>;
>> + interrupts = <0 210 0>;
>> + clocks = <&CLK_SYSIN>;
>> + pinctrl-names = "default";
>> + pinctrl-0 = <&pinctrl_sbc_uart1>;
>> + };
>> + };
> Please name these devices "serial", not "uart".
Ok, I will rename them to serial.
>> diff --git a/arch/arm/configs/stih41x_defconfig
>> b/arch/arm/configs/stih41x_defconfig
>> new file mode 100644
>> index 0000000..dd9268b
>> --- /dev/null
>> +++ b/arch/arm/configs/stih41x_defconfig
> I would prefer to have this be part of multi_v7_defconfig, unless
> you have strong reason why it can't be.
I think we can move this to multi_v7_defconfig.
>> diff --git a/arch/arm/mach-stih41x/Kconfig b/arch/arm/mach-stih41x/Kconfig
>> new file mode 100644
>> index 0000000..9c40540
>> --- /dev/null
>> +++ b/arch/arm/mach-stih41x/Kconfig
>> @@ -0,0 +1,35 @@
>> +config ARCH_STIH41X
>> + bool "STMicroelectronics STiH41x SOCs with Flattened Device Tree" if
>> ARCH_MULTI_V7
>> + select GENERIC_CLOCKEVENTS
>> + select CLKDEV_LOOKUP
>> + select ARM_GIC
>
>> +if ARCH_STIH41X
>> +menu "STMicroelectronics Consumer Electronics SOCs"
> You can use 'menuconfig' to combine these two.
Thats a nice one.. Will use it.
>> +static void __init stih41x_timer_init(void)
>> +{
>> + of_clk_init(NULL);
>> + clocksource_of_init();
>> + stih41x_l2x0_init();
>> +}
> I'm hoping we can kill make the clk and l2x0 init work by default in 3.11,
> as clocksource_of_init already does, so you won't need this function
> any more.
Ok.
>
>> +/*
>> + * A basic implementation of irq_set_wake that ensures wakeup source
>> + * interrupts are not disabled during PM_SUSPEND_FREEZE.
>> + */
>> +static int stih41x_set_wake(struct irq_data *d, unsigned int on)
>> +{
>> + struct irq_desc *desc = irq_to_desc(d->irq);
>> +
>> + if (on) {
>> + if (desc->action)
>> + desc->action->flags |= IRQF_NO_SUSPEND;
>> + } else {
>> + if (desc->action)
>> + desc->action->flags &= ~IRQF_NO_SUSPEND;
>> + }
>> + return 0;
>> +}
>> +
>> +static void __init stih41x_irq_init(void)
>> +{
>> + gic_arch_extn.irq_set_wake = stih41x_set_wake;
>> + gic_arch_extn.flags = IRQCHIP_MASK_ON_SUSPEND;
>> +
>> + irqchip_init();
>> +}
> This looks like it should be generic. Why do you need this?
>
>> +void __init stih41x_dt_init(void)
>> +{
>> + of_platform_populate(NULL, of_default_bus_match_table,
>> + NULL, NULL);
>> + return;
>> +}
> This is already not needed in 3.10.
Ok, cool... will get rid of this..
>
>> +static struct map_desc stih41x_io_desc[] __initdata = {
>> +#ifdef CONFIG_SMP
>> + {
>> + .virtual = IO_ADDRESS(MPE41_SCU_BASE),
>> + .pfn = __phys_to_pfn(MPE41_SCU_BASE),
>> + .length = SZ_4K,
>> + .type = MT_DEVICE,
>> + },
>> +#endif
>> +void __init stih41x_map_io(void)
>> +{
>> +#ifdef CONFIG_SMP
>> + stixxxx_scu_base_addr = ((void __iomem *)IO_ADDRESS(MPE41_SCU_BASE));
>> +#endif
>> + iotable_init(stih41x_io_desc, ARRAY_SIZE(stih41x_io_desc));
>> +}
> IIRC the SCU no longer needs to be mapped this early, so you can use
> ioremap these days.
yep, I agree.
>
>> +void __init stih41x_l2x0_init(void)
>> +{
>> + u32 way_size = 0x4;
>> + u32 aux_ctrl;
>> +
>> + aux_ctrl = (0x1 << L2X0_AUX_CTRL_SHARE_OVERRIDE_SHIFT) |
>> + (0x1 << L2X0_AUX_CTRL_DATA_PREFETCH_SHIFT) |
>> + (0x1 << L2X0_AUX_CTRL_INSTR_PREFETCH_SHIFT) |
>> + (way_size << L2X0_AUX_CTRL_WAY_SIZE_SHIFT);
>> +
>> + l2x0_of_init(aux_ctrl, L2X0_AUX_CTRL_MASK);
>> +}
> I don't know where we stand on this, but I think we should kill all
> calls like these.
I agree, but we don't have other ways as of today..
>
> Arnd
>
_______________________________________________
devicetree-discuss mailing list
[email protected]
https://lists.ozlabs.org/listinfo/devicetree-discuss
