On Sat, Oct 6, 2012 at 4:56 PM, R. Diez <[email protected]> wrote:
> Hi OpenRISC gurus:
>
> I've been adding a JTAG Debug Unit to my OR10 CPU, and I realised that the
> l.trap argument actually gets ignored by the current processor
> implementations, doesn't it?
>
> This is from the "OpenRISC 1000 Architecture Manual":
>
> "Execution of trap instruction results in the trap exception if specified
> bit in SR is set."
>
> GDB inserts an "l.trap 1" instruction if the remote gdbserver doesn't
> support setting breakpoints itself. SR bit 1 is the "Tick Timer Exception
> Enabled" flag, which has nothing to do with debug breakpoints. I guess we
> could use the SR FO bit (Fixed One), but I wonder whether the FPGA extra
> logic is worth it. Shouldn't we just remove that requirement from the
> specification? I mean, the CPU should always ignore the l.trap argument
> (like current implementations seem to do) and always raise the exception (or
> transfer control to the Debug Interface if enabled).

I do think we should remove this condition on the l.trap instruction
executing, and it's already proposed:

http://opencores.org/or1k/Architecture_Specification#l.trap_condition

>
> By the way, the "OpenRISC 1000 Architecture Manual" I have (from
> http://opencores.org/ocsvn/openrisc/openrisc/trunk/docs) is dated from 2006,
> is there a newer version yet?

There is a newer "draft" but the whole thing is in a bit of a mess at
the moment.

http://opencores.org/ocsvn/openrisc/openrisc/trunk/docs/openrisc_arch_draft.odt

Cheers

Julius
_______________________________________________
OpenRISC mailing list
[email protected]
http://lists.openrisc.net/listinfo/openrisc

Reply via email to