[openrisc] Problems debugging OpenRISC target

["Robert Cragie" <rcc@jennic.com>]

I have looked into this in a bit more detail.

1) Still perplexed by why it can't understand 'print my_data'
2) I can see why it's reading wrong results for auto variables as the
registers are getting mixed up. I see the following:

Sending packet: $Hc-1#09...Ack
Packet received:
Sending packet: $qC#b4...Ack
Packet received:
Sending packet: $qOffsets#4b...Ack
Packet received:
Sending packet: $?#3f...Ack
Packet received: S05
Sending packet: $Hg0#df...Ack
Packet received:
Sending packet: $g#67...Ack
Packet received:
0000000000009244000000000008000400000004000000020000823800000001000000010000
12a4000000000000000000000000000000000000000000000000000000000000000000000000
00000000000000000000000000000000000000000000000000000000000041dc000000000000
00010000000000009240000000000000623800008201
0x00006238 in __breakinst () at
/home/sheaf/ecos-sr/ecos/packages/hal/common/current/src/hal_stub.c:569
569         HAL_BREAKPOINT(_breakinst);
(gdb) info reg
            R0       R1       R2       R3       R4       R5       R6
R7
      00000000 00009244 00000000 00080004 00000004 00000002 00008238
00000001
            R8       R9      R10      R11      R12      R13      R14
R15
      00000001 000012a4 00000000 00000000 00000000 00000000 00000000
00000000
           R16      R17      R18      R19      R20      R21      R22
R23
      00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000
           R24      R25      R26      R27      R28      R29      R30
R31
      00000000 00000000 000041dc 00000000 00000001 00000000 00009240
00000000
            PC       SR     EPCR
      00006238 00008201 00000000

----> Here the registers look OK.

(gdb) step
Sending packet: $Hc0#db...Ack
Packet received:
Sending packet: $s#73...Ack
Packet received: T0540:0000623c;01:00009244;
Sending packet: $s#73...Ack
Packet received: T0540:000012a4;01:00009248;
Sending packet: $g#67...Ack
Packet received:
0000924400009248000800040008000400000004000000020000823800000001000000010000
12a4000000000000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000012a400008201
cyg_start () at hello.c:15
15              ulMask = 1;
(gdb) info reg
            R0       R1       R2       R3       R4       R5       R6
R7
      00009244 00009248 00080004 00080004 00000004 00000002 00008238
00000001
            R8       R9      R10      R11      R12      R13      R14
R15
      00000001 000012a4 00000000 00000000 00000000 00000000 00000000
00000000
           R16      R17      R18      R19      R20      R21      R22
R23
      00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000
           R24      R25      R26      R27      R28      R29      R30
R31
      00000000 00000000 00000000 00000000 00000000 00000000 00000000
00000000
            PC       SR     EPCR
      000012a4 00008201 00000000
(gdb)

----> But here there seems to be some corruption going on. R3 and above look
plausible but R0 should definitely not be 0x00009244 (should be 0x0) and R2
(the frame pointer) has also seems to have become corrupted (well at least
as it's reported). So GDB works out the address of the auto variable
completely wrongly.

Anyone any idea why this might be happening? I am still trawling through the
eCos code but it's hard going. Any advice on the structure of the eCos code
would be useful.

Robert Cragie, Design Engineer
_______________________________________________________________
Jennic Ltd, Furnival Street, Sheffield, S1 4QT,  UK
http://www.jennic.com  Tel: +44 (0) 114 281 2655
_______________________________________________________________

--
To unsubscribe from openrisc mailing list please visit http://www.opencores.org/mailinglists.shtml