[Libre-soc-dev] fighting gdb machine-interface

[Jacob Lifshay]

On Mon, May 31, 2021, 13:36 Luke Kenneth Casson Leighton <lkcl at lkcl.net>
wrote:

> ---
> crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
>
>
> On Mon, May 31, 2021 at 7:54 PM Jacob Lifshay <programmerjake at gmail.com>
> wrote:
>
>
> > hmm, maybe the xml qemu's gdb sends back would be useful:
> >
> >
> https://gitlab.com/qemu-project/qemu/-/blob/52848929b70dcf92a68aedcfd90207be81ba3274/gdb-xml/power-fpu.xml
>
>
> bizarrely, the hint about the starting number in f0 is way off.


hmm, odd. here's the function where qemu's gdbserver gets the fp regs (n is
offset by the caller):
https://gitlab.com/qemu-project/qemu/-/blob/52848929b70dcf92a68aedcfd90207be81ba3274/target/ppc/gdbstub.c#L451
called from here:
https://gitlab.com/qemu-project/qemu/-/blob/52848929b70dcf92a68aedcfd90207be81ba3274/gdbstub.c#L943

>
> i'm currently attempting to set/get via the VSX registers.  absolute pain
> in the neck.
>

Well, since gdb seems to work fine with qemu, try dumping the socket
connecting them using wireshark :)

Jacob