Re: [msp430-elf] Interrupt vector corrupted?

[Orlando Arias <orlandoarias at gmail dot com>]

Greetings,

Apologies for the follow up e-mail. Please take into account the
following errata to the previous message:

All instances USCI_A1_VECT should be USCI_A1_VECTOR.

As an addendum, I should add that the microcontroller resets when
attempting to execute the ISR. The behaviour can not be replicated with
USCI_A0_VECTOR [of course, enabling proper interrupts in either case].

Sorry for the inconvenience. Thank you.

Cheers,
Orlando.

On 04/18/2016 06:02 PM, Orlando Arias wrote:
> Greetings,
> 
> I am working with an MSP430F5529 and I am seeing issues with the way
> USCI_A1_VECT gets added to the vector table. I am not sure if the bug is
> in the linker script (saw nothing obvious there), or the way binutils is
> populating entries to this particular entry. I have tried to add an
> entry two ways,
> 
> __attribute__((interrupt((USCI_A1_VECT)))
> void uart1_isr(void) {
> 	/* snip */
> }
> 
> and a rather obtuse one:
> 
> __asm (
>       "\t.section \"__interrupt_vector_47\",\"ax\",@progbits\n\t"
>       ".word uart1_isr"
> );
> void uart1_isr(void) {
> 	/* snip */
> }
> 
> when performing an objdump on the generated binary, the following is
> observed:
> 
> Disassembly of section __interrupt_vector_47:
> 
> 0000ffdc <__interrupt_vector_47>:
>     ffdc:       90 49 90 9e     Address 0x000000000000ffde is out of bounds.
> Address 0x000000000000ffe0 is out of bounds.
> mov     -1(r9), 0xffff  ; PC rel. 0xffdf
>     ffe0:       aa 7f
> 
> /* snip */
> 
> 00004990 <uart1_isr>:
> /* snip */
> 
> for either case.
> 
> I am using gcc 5.3.0 and binutils 2.26.20160125 with no extra patches. I
> have attached the linker script I am using.
> 
> Please advice if I am doing something wrong. Thank you.
> 
> Cheers,
> Orlando.
>

Attachment: signature.asc
Description: OpenPGP digital signature