topic Re: How to allocate a banked (__far) function address to uint32_t? in Classic/Legacy CodeWarrior

How to allocate a banked (__far) function address to uint32_t?

Snaku — Thu, 23 Apr 2015 04:06:53 GMT

Hi,

I am developing a boot-loader on MC9S12XE family with Classic CodeWarrior, I want to add a descriptor for the application code, and it should contain an address which is application code '_EntryPoint' (i.e. the address store in application's reset vector), my problem is how to allocate the '_EntryPoint' banked address to a uint32_t integer?

It is an example of my problem, if I use the type 'tIsrFunc' to allocate '_EntryPoint' is OK but 16 bit value, e.g. 0x1234U.

If I cast '_EntryPoint' to a uint32_t, the value will be shift left four bits, e.g. 0x00123400UL.

How do I cast a backed function address to uint32_t without shifted left four bits?

typedef void (*near tIsrFunc)(void);

typedef struct APP_TRY_ADDR_S

{

tIsrFunc fptr_near;

uint32_t fptr_far;

}APP_TRY_ADDR_T;

const APP_TRY_ADDR_T m_TryAddr =

{

&_EntryPoint, /* it is OK, but 16 bit only */

(uint32_t)&_EntryPoint, /* it is wrong, the value shifted left four bits */

};

Re: How to allocate a banked (__far) function address to uint32_t?

ZhangJennie — Thu, 30 Apr 2015 06:54:30 GMT

hi Snaku,

I ever made sample code of far pointer to banked data address. see attached.

i think the method is similar for pointer to banked code address.

can this help?

Have a great day,
Zhang Jun

-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

Re: Re: How to allocate a banked (__far) function address to uint32_t?

Snaku — Mon, 04 May 2015 06:30:10 GMT

Hi Zhang Jun,

Sorry for my late reply, I try to change my example code with (unsign long * far) casting, but the result still shifted left four bits, 24 bits only, below is my test code.

#pragma CODE_SEG __NEAR_SEG NON_BANKED

#pragma NO_FRAME

#pragma NO_EXIT

void _EntryPoint(void)

{

...

}

#pragma CODE_SEG DEFAULT

typedef struct APP_TRY_ADDR_S

{

tIsrFunc fptr_near;

unsigned long * __far fptr_far;

}APP_TRY_ADDR_T;

const APP_TRY_ADDR_T m_app_try_addr

{

&_EntryPoint, /* result is 0x4046 */

(unsigned long * __far)&_EntryPoint, /* result is 0x404600, not 0x00004046 */

};

Should I use logical paged address like this page? Then I can jump/call to application start address easier.

And I don't want to use CONV_FAR_FUN_TO_DATA_PTR or CONV_FAR_DATA_TO_FUN_PTR macro which introduced in "S12(X) Build Tools Reference Manual".

Thanks for your support,

Snaku