[ECOS] flash_query throws an exception (Redboot_ROM version)

[Krishna Ganugapati]

Hello,

I've gotten Redboot for the IXD425 working with the cvs sources. I've gone 
ahead and added the intel npe support. As part of the npe support, I've 
enabled the strata flash support.  The package builds fine, but when it 
attempts to query the flash for properties it throws an exception.

The flash_query function is correctly located in RAM - .attribute__ 
((section (".2ram.flash_query")));


#define CNT 20*1000*10  // Approx 20ms

int
flash_query(unsigned char *data)  __attribute__ ((section 
(".2ram.flash_query")));
int
flash_query(unsigned char *data)
{
    volatile flash_t *ROM;
    int i, cnt;

    // Get base address and map addresses to virtual addresses
    ROM = FLASH_P2V( CYGNUM_FLASH_BASE );
#ifdef CYGOPT_FLASH_IS_BOOTBLOCK
    // BootBlock flash does not support full Read_Query - we have do a
    // table oriented thing above, after getting just two bytes of results:
    ROM[0] = FLASH_Read_ID;
    i = 2;
#else
    // StrataFlash supports the full Read_Query op:
    ROM[0] = FLASH_Read_Query;
    i = sizeof(struct FLASH_query);
#endif // Not CYGOPT_FLASH_IS_BOOTBLOCK

    for (cnt = CNT;  cnt > 0;  cnt--) ;
    for ( /* i */;  i > 0;  i--, ++ROM) {
        // It is very deliberate that data is chars NOT flash_t:
        // The info comes out in bytes regardless of device.
        *data++ = (unsigned char) CYGHWR_FLASH_READ_QUERY(ROM);
#ifndef CYGOPT_FLASH_IS_BOOTBLOCK
# if  8 == CYGNUM_FLASH_WIDTH
 // strata flash with 'byte-enable' contains the configuration data
 // at even addresses
 ++ROM;
# endif
#endif
    }
    ROM[0] = FLASH_Reset;

    return 0;
}

The equivalent assembly is here, we invariably fault at 0x00009318


00009310 <flash_query>:
    9310: e3a02098  mov r2, #152 ; 0x98
    9314: e3a01205  mov r1, #1342177280 ; 0x50000000
    9318: e3a03bc3  mov r3, #199680 ; 0x30c00
    931c: e1c120b0  strh r2, [r1]
    9320: e3a02034  mov r2, #52 ; 0x34
    9324: e2833d05  add r3, r3, #320 ; 0x140
    9328: e2433001  sub r3, r3, #1 ; 0x1
    932c: e3530000  cmp r3, #0 ; 0x0
    9330: cafffffc  bgt 9328 <flash_query+0x18>
    9334: e3520000  cmp r2, #0 ; 0x0
    9338: da000004  ble 9350 <flash_query+0x40>
    933c: e0d130b2  ldrh r3, [r1], #2
    9340: e2422001  sub r2, r2, #1 ; 0x1
    9344: e3520000  cmp r2, #0 ; 0x0
    9348: e4c03001  strb r3, [r0], #1
    934c: cafffffa  bgt 933c <flash_query+0x2c>
    9350: e3a030ff  mov r3, #255 ; 0xff
    9354: e1c130b0  strh r3, [r1]
    9358: e3a00000  mov r0, #0 ; 0x0
    935c: e12fff1e  bx lr

When I don't enable the STRATA_FLASH package everything works beautifully. 
The code is exactly what is in the CVS sources. Can anyone tell me why this 
happens? Has anyone seen this and any thoughts on how I can resolve?

Also I'm curious to understand how portions of the image are relocated into 
RAM - presumably prior to cyg_start sections that are marked appropriately 
are copied into RAM - can someone shed light on how this happens?

Thanks

Krishna 


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss