[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[openrisc] Problem with jump behavior
I am getting started with experimenting with openrisc. I pulled down and
built the
binutils and gcc for use with cygwin. I am using the ModelSim simulator
that comes
with the Xilinx WebPack.
While experimenting with some very simple c programs, it looks like my
jumps are off
by one instruction.
A test program I have been working with is:
void main(){
int i=0;
while(1){
i=i+1;
}
}
The disassembly is
00010000 <_start>:
10000: 18 20 00 00 l.movhi r1,0x0
10004: a8 21 00 10 l.ori r1,r1,0x10
10008: 00 00 00 04 l.j 10018 <_main>
1000c: 15 00 00 00 l.nop 0x0
10010: 15 00 00 00 l.nop 0x0
10014: 03 ff ff fb l.j 10000 <_start>
00010018 <_main>:
10018: 9c 21 ff f8 l.addi r1,r1,0xfffffff8
1001c: d4 01 10 00 l.sw 0x0(r1),r2
10020: 9c 41 00 08 l.addi r2,r1,0x8
10024: 9c 60 00 00 l.addi r3,r0,0x0
10028: d7 e2 1f fc l.sw 0xfffffffc(r2),r3
1002c: 84 62 ff fc l.lwz r3,0xfffffffc(r2)
10030: 9c 63 00 01 l.addi r3,r3,0x1
10034: d7 e2 1f fc l.sw 0xfffffffc(r2),r3
10038: 03 ff ff fd l.j 1002c <_main+0x14>
1003c: 15 00 00 00 l.nop 0x0
10040: 84 41 00 00 l.lwz r2,0x0(r1)
10044: 44 00 48 00 l.jr r9
10048: 9c 21 00 08 l.addi r1,r1,0x8
What I am seeing in the simulator is when a the jump at 10008 is executed,
instead of seeing iwb_adr_o going to
10018, I am seeing it go to 1001c.
In this case the program simulates as expected, but when I add function
calls the test programs fail because
the first instruction is missed, which is usually setting up a stack frame.
This is what I am using for building my test program:
/opt/or32-uclinux/bin/or32-uclinux-gcc -g -Wa,-a=crt0.asm -c -o crt0.or32
crt0.S
/opt/or32-uclinux/bin/or32-uclinux-gcc -g -Wa,-a=test1.asm -c -o
test1.or32 test1.c
/opt/or32-uclinux/bin/or32-uclinux-ld -Ttext 10000 -Tdata 20 -o test1.bin
crt0.or32 test1.or32
Am I missing a compile or link flag or is model sim possibly doing the
simulation wrong.
Thanks,
Damon
--
To unsubscribe from openrisc mailing list please visit http://www.opencores.org/mailinglists.shtml