Re: [PATCH v2 RESEND 3/7] gdbstub: add support for the qRegisterInfo query

["Alex Bennée" <alex.bennee@linaro.org>]

Matheus Tavares Bernardino <quic_mathbern@quicinc.com> writes:

> Alex Bennée <alex.bennee@linaro.org> wrote:
>>
>> > Matheus Tavares <quic_mathbern@quicinc.com> wrote:
>> >
>> > diff --git a/gdbstub/gdbstub.c b/gdbstub/gdbstub.c
>> > index be18568d0a..f19f8c58c3 100644
>> > --- a/gdbstub/gdbstub.c
>> > +++ b/gdbstub/gdbstub.c
>> > @@ -1578,6 +1599,12 @@ static const GdbCmdParseEntry gdb_gen_query_table[] = {
>> >          .handler = handle_query_curr_tid,
>> >          .cmd = "C",
>> >      },
>> > +    {
>> > +        .handler = handle_query_regs,
>> > +        .cmd = "RegisterInfo",
>> > +        .cmd_startswith = 1,
>> > +        .schema = "l0"
>> > +    },
>> 
>> Where is this defined in the protocol spec, I can't see it in:
>> 
>>   https://sourceware.org/gdb/onlinedocs/gdb/General-Query-Packets.html#General-Query-Packets
>> 
>> and it seems to be information that is handled by the xml register
>> description. Is there a reason that isn't used for Hexagon?
>
> Good point. It's actually an lldb extension to the protocol:
> https://github.com/llvm/llvm-project/blob/main/lldb/docs/lldb-gdb-remote.txt#L573
>
> But indeed, lldb should be able to use the xml register description as
> well. I'll take a look and try to do that instead.

There may be an argument for supporting both but only if the details of
the xml/RegisterInfo would be sorted out by gdbstub or some other common
code rather than each front-end growing special support.

For now see how the XML does for you.

>
> Thanks!


-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro