This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
RE: Vector tables and such...
- From: "Mark Cianfaglione" <markcia at nortelnetworks dot com>
- To: Nick Clifton <nickc at cambridge dot redhat dot com>
- Cc: binutils at sources dot redhat dot com
- Date: Fri, 7 Jun 2002 11:16:04 -0400
- Subject: RE: Vector tables and such...
Hi Nick et. al.
>> Am I missing something or am I assuming the wrong operation with the
> ".long" keyword?
>No, .long should work as you expect.
Good I was beginning to question my sanity...
>If you run "objdump -dr" on your assembled test code do you see a
>reloc for the "start" symbol ?
I did this only to realize that the source is in assembler and not "C" or
"C" like... So there was no useable output.
>Is "start" declared in a C or C++ source file ? It is possible that
>the compiler is prefixing user symbols with an underscore, so that the
>definition of "start" actually becomes "_start".
I suspected that if and when I was going to use a "C" exception routine but
as it was in assembler it wasn't the case...
What I did see though is the following...
============================================================================
=============
18 .extern start
19 .extern turn_on_led_error
20
21
22
23 // This must be at location 0 of the load file
24
25 .section .vect
26
27 .export interrupt_vector
28
29 interrupt_vector:
30
31 0000 00000000 .long start //
0 - reset
32 0004 00000000 .long turn_on_led_error //
1 - misaligned fault
33 0008 00000000 .long turn_on_led_error //
2 - access fault
...blah...blah...blah...
M.CORE GAS Version 2.9.4 exception_vector.s page 2
DEFINED SYMBOLS
exception_vector.s:29 .vect:00000000 interrupt_vector
UNDEFINED SYMBOLS
start
turn_on_led_error
============================================================================
=========================
Yet from the "init.s" output I get the following clip...
DEFINED SYMBOLS
init.s:27 .text:00000000 start
init.s:41 .text:00000010 turn_on_led_error
============================================================================
========================
Is there a possible alternative????
Thanks
Mark