[U-Boot] [PATCH] arm: re-implement proper ISB instruction for ARMv7-A

[Ziyuan Xu]

On 2016年07月31日 22:27, Tom Rini wrote:
> On Sun, Jul 31, 2016 at 11:59:19AM +0800, Ziyuan Xu wrote:
>> Hi Tom,
>>
>>
>> On 2016年07月29日 09:12, Tom Rini wrote:
>>> On Fri, Jul 29, 2016 at 09:06:29AM +0800, Ziyuan Xu wrote:
>>>> Hi Tom,
>>>>
>>>> On 2016年07月29日 08:34, Tom Rini wrote:
>>>>> On Fri, Jul 29, 2016 at 07:34:09AM +0800, Ziyuan Xu wrote:
>>>>>> Hi Tom,
>>>>>>
>>>>>> On 2016年07月29日 06:15, Tom Rini wrote:
>>>>>>> On Thu, Jul 28, 2016 at 07:03:17PM +0800, Chen-Yu Tsai wrote:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> On Thu, Jul 28, 2016 at 6:13 PM, Ziyuan Xu <xzy.xu at rock-chips.com> wrote:
>>>>>>>>> For ARMv7-A architecture, the valid ISB instruction is asm volatile("isb").
>>>>>>>>>
>>>>>>>>> This patch fixes the U-Boot was stuck in invalidate_dcache_all() before
>>>>>>>>> booting linux kernel, which occurred on rk3288-base development board
>>>>>>>>> such as evb-rk3288, rock2-rk3288. And something output via console like:
>>>>>>>>>
>>>>>>>>> => bootz 0x2000000
>>>>>>>>> 0x02000000
>>>>>>>>>     ramdisk start = 0x00000000, ramdisk end = 0x00000000
>>>>>>>>>     Continuing to boot without FDT
>>>>>>>>>     Initial value for argc=3
>>>>>>>>>     Final value for argc=3
>>>>>>>>>     using: ATAGS
>>>>>>>>>
>>>>>>>>>     Starting kernel ...
>>>>>>>>>
>>>>>>>>> Linux kernel exactly the same way(see arch/arm/include/asm/barrier.h).
>>>>>>>>>
>>>>>>>>> Signed-off-by: Ziyuan Xu <xzy.xu at rock-chips.com>
>>>>>>>>> ---
>>>>>>>>>
>>>>>>>>>   arch/arm/include/asm/system.h | 6 ++++--
>>>>>>>>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>>>>>>>>
>>>>>>>>> diff --git a/arch/arm/include/asm/system.h b/arch/arm/include/asm/system.h
>>>>>>>>> index 2bdc0be..12d4ba0 100644
>>>>>>>>> --- a/arch/arm/include/asm/system.h
>>>>>>>>> +++ b/arch/arm/include/asm/system.h
>>>>>>>>> @@ -227,13 +227,15 @@ void __noreturn psci_system_reset(bool smc);
>>>>>>>>>    */
>>>>>>>>>   void save_boot_params_ret(void);
>>>>>>>>>
>>>>>>>>> -#define isb() __asm__ __volatile__ ("" : : : "memory")
>>>>>>>>> -
>>>>>>>>>   #define nop() __asm__ __volatile__("mov\tr0,r0\t@ nop\n\t");
>>>>>>>>>
>>>>>>>>>   #ifdef __ARM_ARCH_7A__
>>>>>>>>> +#define isb() __asm__ __volatile__ ("isb" : : : "memory")
>>>>>>>>> +
>>>>>>>>>   #define wfi() __asm__ __volatile__ ("wfi" : : : "memory")
>>>>>>>>>   #else
>>>>>>>>> +#define isb() __asm__ __volatile__ ("" : : : "memory")
>>>>>>>>> +
>>>>>>>>>   #define wfi()
>>>>>>>>>   #endif
>>>>>>>>>
>>>>>>>> arch/arm/include/asm/barriers.h already has a proper set of
>>>>>>>> ISB/DSB macros. Please consider using those instead.
>>>>>>> Please fix arch/arm/include/asm/system.h to use the macros found in
>>>>>>> barriers.h rather than have their own versions.  Thanks!
>>>>>> If I understand correctly, I can change into as bellow:
>>>>>> #define isb() ISB
>>>>>> How about it?
>>>>> Well, I'd rather not have ISB and isb, just ISB, which means we might
>>>>> have to fix other places too.  If that starts looking too huge, we can
>>>>> do this in steps and just do what you suggested for now and for next
>>>>> release move everything over.
>>>> As I mentioned before, arch/arm/include/asm/barriers.h defined ISB
>>>> macro.  If I only want to fix the issue which I hit on rk3288 board,
>>>> I just use ISB in arch/arm/include/asm/system.h::set_cr() instead of
>>>> isb(). But isb() had been invoked in some places.
>>>>
>>>> I can't verify integrallty after all revision, it involve some
>>>> boards and feature. But this does fix for rk3288, if you agree with
>>>> me, could you apply it provisionally?:-)
>>> I would really like to try and fix the other possibly latent issues that
>>> we have by not calling a real ISB.  Please try moving towards all places
>>> that need an isb calling the correct one from barriers.h and giving it a
>>> spin on the hardware you have available.
>> I used ISB instead of isb(), everything works sane on my rockchip
>> rk3288 boards.:-)
>> Will you send a patch to fix it and other possibly latent issues? Or
>> apply this temporarily?
> Please send a patch of all of the changes you did, thanks!
Tom, it's impracticable only use ISB for all situation instead of isb(). 
There're two isb() invocation in drivers/ddr/fsl/fsl_ddr_gen4.c, and not 
only used on ARM architecture board, but also powerpc. See below:

arch/powerpc/include/asm/config_mpc85xx.h:769:#define 
CONFIG_SYS_FSL_DDRC_GEN4
arch/powerpc/include/asm/config_mpc85xx.h:822:#define 
CONFIG_SYS_FSL_DDRC_GEN4
arch/powerpc/include/asm/config_mpc85xx.h:955: 
!defined(CONFIG_SYS_FSL_DDRC_GEN4)
arch/arm/include/asm/arch-ls102xa/config.h:100:#define 
CONFIG_SYS_FSL_DDRC_GEN4
arch/arm/include/asm/arch-fsl-layerscape/config.h:13:#define 
CONFIG_SYS_FSL_DDRC_GEN4

If you insist on ISB only, we have to differentiate arch via 
CONFIG_SYS_ARCH, seems twisted.
How about it?
#define isb() ISB
>