MVME2304 Exception 3

[Paul D Jines]

>Do you have a symbol table for this?  You can get one with
>powerpc-rtems-nm.  It is important to know what function
>and instruction are at the "Next PC or Address or fault".
>I see these addresses:

>0x130E84 (run 1 and 2?)
>0x108A60 (run 3)

For the memset error:

>From the output of powerpc-rtems-nm,

00130e44 T memset (called from
000db8d8 T RTEMS_Malloc_Initialize)

from the output of powerpc-rtems-objdump -dt

00130e44 <memset>:
  130e44:       28 05 00 03     cmplwi  r5,3
  130e48:       94 21 ff e0     stwu    r1,-32(r1)
  130e4c:       54 84 06 3e     clrlwi  r4,r4,24
  130e50:       7c 69 1b 78     mr      r9,r3
  130e54:       40 81 00 70     ble-    130ec4 <memset+0x80>
  130e58:       70 60 00 03     andi.   r0,r3,3
  130e5c:       40 82 00 68     bne-    130ec4 <memset+0x80>
  130e60:       54 80 40 2e     rlwinm  r0,r4,8,0,23
  130e64:       7c 6b 1b 78     mr      r11,r3
  130e68:       7c 00 23 78     or      r0,r0,r4
  130e6c:       54 09 80 1e     rlwinm  r9,r0,16,0,15
  130e70:       7c 00 4b 78     or      r0,r0,r9
  130e74:       28 05 00 0f     cmplwi  r5,15
  130e78:       40 81 00 28     ble-    130ea0 <memset+0x5c>
  130e7c:       54 a9 e1 3e     rlwinm  r9,r5,28,4,31
  130e80:       7d 29 03 a6     mtctr   r9
* 130e84:       90 0b 00 00     stw     r0,0(r11)
  130e88:       38 a5 ff f0     addi    r5,r5,-16
  130e8c:       94 0b 00 04     stwu    r0,4(r11)
  130e90:       94 0b 00 04     stwu    r0,4(r11)
  130e94:       94 0b 00 04     stwu    r0,4(r11)
  130e98:       39 6b 00 04     addi    r11,r11,4
  130e9c:       42 00 ff e8     bdnz+   130e84 <memset+0x40>
  130ea0:       28 05 00 03     cmplwi  r5,3
  130ea4:       40 81 00 1c     ble-    130ec0 <memset+0x7c>
  130ea8:       54 a9 f0 be     rlwinm  r9,r5,30,2,31
  130eac:       7d 29 03 a6     mtctr   r9
  130eb0:       90 0b 00 00     stw     r0,0(r11)
  130eb4:       38 a5 ff fc     addi    r5,r5,-4
  130eb8:       39 6b 00 04     addi    r11,r11,4
  130ebc:       42 00 ff f4     bdnz+   130eb0 <memset+0x6c>
  130ec0:       7d 69 5b 78     mr      r9,r11
  130ec4:       2c 05 00 00     cmpwi   r5,0
  130ec8:       38 a5 ff ff     addi    r5,r5,-1
  130ecc:       41 82 00 18     beq-    130ee4 <memset+0xa0>
  130ed0:       38 a5 00 01     addi    r5,r5,1
  130ed4:       7c a9 03 a6     mtctr   r5
  130ed8:       98 89 00 00     stb     r4,0(r9)
  130edc:       39 29 00 01     addi    r9,r9,1
  130ee0:       42 00 ff f8     bdnz+   130ed8 <memset+0x94>
  130ee4:       38 21 00 20     addi    r1,r1,32
  130ee8:       4e 80 00 20     blr
  130eec:       00 08 e0 a4     .long 0x8e0a4

>If all you are doing is putting expressions in Till's cexp,
>I suppose it is entirely possible that you are blowing
>cexp's stack.

Actually, most of the time we don't even get into cexp.
We are receiving the error for the hello world application,
EPICS apps, and the gesys binary (but this one runs
sometimes).  We even tried replacing memset with our own
C routine, it had the same error, but at different
memory addresses each time....

Thanks,

Paul