SPARC BSP Question

[Joel Sherrill]

Hi,

I was doing some basic checks on the 4.7 branch
and head.  I noticed that the sparc/sis BSP
traps while zero'ing out the RTEMS workspace.  This
is the last gdb output before it traps.

(gdb)
51            zero_out_array[ index ] = 0;
2: &zero_out_array[index] = (uint32_t *) 0x23ffea4
1: index = 13737

I expected to see a simple math error causing us to access
past the end of memory but this isn't the case here.  According
to my prints, the starting address and size add up to 64K so that
looks OK and it looks like RAM should be there.

(gdb) p starting_address
$3 = (void *) 0x23f2800
(gdb) p/x size
$5 = 0xd800

FWIW the erc32 BSP on tsim runs OK.  This is JUST with sis
on gdb.

Any ideas?

--joel