MC9S12A128 bug

[electropw]

Hi,

We are using the MC9S12A128CFUE processor and fear there maybe a bug in the mask.

The mask is "1L59W".

The problem we are seeing is that the instruction LBEQ is setting the I flag in the ccr.

The code is codewarrior generated.

Are seeing it at least following a RTS and a ANDA # instruction.

The instruction opcode is 182700ABh.

It is definitely across multiple CPUs and not just seen on one CPU.

We are debugging with the codewarrior debugger and we are using the last version with latest update.

We are  not 100% sure that it is not a debugger bug.

Anyone seen this??? or any help?

regards

Phil

[kef]

Didn't you consider broken interrupt service routine? What if that broken ISR overwrites status register value, which is pushed to the stack on entry to interrupt? CPU could then RTI from ISR with I-bit set and it could look as if LBEQ or any other instruction is setting I-bit.

To investigate it try disabling interrupts locally (not using I-bit) and see if LBEQ is still setting I-bit.