On Mon, 26 Jan 2026 at 08:34, Manos Pitsidianakis
<[email protected]> wrote:
>
> On Thu, Jan 22, 2026 at 6:52 PM Peter Maydell <[email protected]> 
> wrote:
> > If giving the SME registers a good workout with the
> > gdbstub interface all seems to work fine, then I think
> > we can adjust this condition to be "if tcg_enabled && ...".
>
> gdb had a disastrous workout, because it turns out it also assumes SME
> must have SVE:
>
>
>   (gdb) target remote localhost:1234
>   Remote debugging using localhost:1234
>   ../../gdb/aarch64-tdep.c:3068: internal-error:
> aarch64_pseudo_register_type: bad register number 160
>   A problem internal to GDB has been detected,
>   further debugging may prove unreliable.
>   Fatal signal: Abort trap: 6

I think that if we rebase this series on top of the patchset I
just sent out:
https://patchew.org/QEMU/[email protected]/
that should avoid this GDB internal error (which happened because
we were reporting the Z registers as zero width, so GDB got
confused about whether they existed or not).

The GDB and LLDB changes to handle SME-only CPUs seem to mostly
be related to the ptrace API, because of choices made in the Linux
kernel API about how to handle SME-only, so I'm optimistic we don't
need to wait for anything to land on the debugger side.

thanks
-- PMM

Reply via email to