[PR 05] LEON3: CPU index intialization moved to bspstart.c

[Daniel Hellstrom]

On 02/02/2012 04:56 PM, Gedare Bloom wrote:
> On Thu, Feb 2, 2012 at 10:50 AM, Daniel Hellstrom<daniel at gaisler.com>  wrote:
>> On 02/02/2012 04:45 PM, Gedare Bloom wrote:
>>> On Thu, Feb 2, 2012 at 9:46 AM, Daniel Hellstrom<daniel at gaisler.com>
>>>   wrote:
>>>> All LEON3/4 systems have a CPU-id, if on a single-CPU system the
>>>> ID is always zero. On a multicore system it ranges from 0 to 15.
>>>>
>>>> The CPU index should always by updated even in a non-MP RTEMS OS
>>>> since the CPU running RTEMS may not always be CPU0. For example
>>>> when RTEMS runs on CPU1 and Linux on CPU0 in a mixed ASMP system.
>>>>
>>>> The old code executed within the IRQ controller initialization code
>>>> makes no sense since the ASR register is a CPU register, it has
>>>> nothing to do with AMBA initialization either.
>>>>
>>>> Signed-off-by: Daniel Hellstrom<daniel at gaisler.com>
>>>> ---
>>>>   c/src/lib/libbsp/sparc/leon3/amba/amba.c        |   18
>>>> ------------------
>>>>   c/src/lib/libbsp/sparc/leon3/startup/bspstart.c |   18
>>>> ++++++++++++++++++
>>>>   2 files changed, 18 insertions(+), 18 deletions(-)
>>>>
>>>> diff --git a/c/src/lib/libbsp/sparc/leon3/amba/amba.c
>>>> b/c/src/lib/libbsp/sparc/leon3/amba/amba.c
>>>> index 1932cee..b0a43f9 100644
>>>> --- a/c/src/lib/libbsp/sparc/leon3/amba/amba.c
>>>> +++ b/c/src/lib/libbsp/sparc/leon3/amba/amba.c
>>>> @@ -21,8 +21,6 @@ amba_confarea_type amba_conf;
>>>>   /* Pointers to Interrupt Controller configuration registers */
>>>>   volatile LEON3_IrqCtrl_Regs_Map *LEON3_IrqCtrl_Regs;
>>>>
>>>> -int LEON3_Cpu_Index = 0;
>>>> -
>>>>   /*
>>>>   *  amba_initialize
>>>>   *
>>>> @@ -33,16 +31,6 @@ int LEON3_Cpu_Index = 0;
>>>>   *  amba_ahb_masters, amba_ahb_slaves and amba.
>>>>   */
>>>>
>>>> -unsigned int getasr17(void);
>>>> -
>>>> -__asm__ (" .text  \n"
>>>> -    "getasr17:   \n"
>>>> -    "retl \n"
>>>> -    "mov %asr17, %o0\n"
>>>> -);
>>>> -
>>>> -
>>>> -extern rtems_configuration_table Configuration;
>>>>   extern int scan_uarts(void);
>>>>
>>>>   void amba_initialize(void)
>>>> @@ -58,12 +46,6 @@ void amba_initialize(void)
>>>>    if ( i>    0 ){
>>>>      /* Found APB IRQ_MP Interrupt Controller */
>>>>      LEON3_IrqCtrl_Regs = (volatile LEON3_IrqCtrl_Regs_Map *) dev.start;
>>>> -#if defined(RTEMS_MULTIPROCESSING)
>>>> -      if (rtems_configuration_get_user_multiprocessing_table() != NULL)
>>>> {
>>>> -        unsigned int tmp = getasr17();
>>>> -        LEON3_Cpu_Index = (tmp>>    28)&    3;
>>>>
>>>> -      }
>>>> -#endif
>>>>    }
>>>>
>>>>    /* find GP Timer */
>>>> diff --git a/c/src/lib/libbsp/sparc/leon3/startup/bspstart.c
>>>> b/c/src/lib/libbsp/sparc/leon3/startup/bspstart.c
>>>> index a4b9ffa..82054e8 100644
>>>> --- a/c/src/lib/libbsp/sparc/leon3/startup/bspstart.c
>>>> +++ b/c/src/lib/libbsp/sparc/leon3/startup/bspstart.c
>>>> @@ -27,6 +27,9 @@
>>>>   */
>>>>   int CPU_SPARC_HAS_SNOOPING;
>>>>
>>>> +/* Index of CPU, in an AMP system CPU-index may be non-zero */
>>>> +int LEON3_Cpu_Index = 0;
>>>> +
>>>>   extern void amba_initialize(void);
>>>>
>>>>   /*
>>>> @@ -48,6 +51,14 @@ static inline int set_snooping(void)
>>>>    return (tmp>>    27)&    1;
>>>>
>>>>   }
>>>>
>>>> +/* ASM-function used to get the CPU-Index on calling LEON3 CPUs */
>>>> +static inline unsigned int get_asr17(void)
>>>> +{
>>>> +  unsigned int reg;
>>>> +  __asm__ (" mov %%asr17, %0 " : "=r"(reg) :);
>>>> +  return reg;
>>>> +}
>>> Might get_asr17 make sense as a function defined by
>>> cpukit/score/cpu/sparc/rtems/score/sparc.h? Or is it really only
>>> useful for getting the cpu index as used below?
Missed this reply first time. It is LEON3/4 CPU specific, so leon.h would be better I think. Same goes here, better be in a separate patch I think.

>>>> +
>>>>   /*
>>>>   *  bsp_start
>>>>   *
>>>> @@ -57,6 +68,13 @@ void bsp_start( void )
>>>>   {
>>>>    CPU_SPARC_HAS_SNOOPING = set_snooping();
>>>>
>>>> +  /* Get the LEON3 CPU index, normally 0, but for MP systems we do
>>>> +   * _not_ assume that this is CPU0. One may run another OS on CPU0
>>>> +   * and RTEMS on this CPU, and AMP system with mixed operating
>>>> +   * systems
>>>> +   */
>>>> +  LEON3_Cpu_Index = (get_asr17()>>    28)&    3;
>>> Would it make sense to also define something like
>>> sparc_get_cpu_index()? Or does it only ever need to be done during bsp
>>> initialization?
>> Yes, that makes sens, but perhaps in a separate "clean patch" for HEAD? This
>> is the way the code was code first time:
>>
> OK.
>
>> -      if (rtems_configuration_get_user_multiprocessing_table() != NULL) {
>> -        unsigned int tmp = getasr17();
>> -        LEON3_Cpu_Index = (tmp>>    28)&    3;
>> -      }
>>
>>
>>>> +
>>>>    /* Find UARTs */
>>>>    amba_initialize();
>>>>   }
>>>> --
>>>> 1.7.0.4
>>>>
>>>> _______________________________________________
>>>> rtems-devel mailing list
>>>> rtems-devel at rtems.org
>>>> http://www.rtems.org/mailman/listinfo/rtems-devel
>>>
>