[PATCH v3 10/10] hmp: Deprecate 'singlestep' member of StatusInfo

[Richard Henderson]

On 4/17/23 18:40, Peter Maydell wrote:
> The 'singlestep' member of StatusInfo has never done what the QMP
> documentation claims it does.  What it actually reports is whether
> TCG is working in "one guest instruction per translation block" mode.
> 
> We no longer need this field for the HMP 'info status' command, as
> we've moved that information to 'info jit'.  It seems unlikely that
> anybody is monitoring the state of this obscure TCG setting via QMP,
> especially since QMP provides no means for changing the setting.  So
> simply deprecate the field, without providing any replacement.
> 
> Until we do eventually delete the member, correct the misstatements
> in the QAPI documentation about it.
> 
> If we do find that there are users for this, then the most likely way
> we would provide replacement access to the information would be to
> put the accelerator QOM object at a well-known path such as
> /machine/accel, which could then be used with the existing qom-set
> and qom-get commands.
> 
> Signed-off-by: Peter Maydell<peter.maydell at linaro.org>
> ---
> For v3: because we're only deprecating the existing member,
> not trying to provide a replacement with a new name, we don't
> need to update the iotests that use the command. (We will when
> we eventually drop the deprecated member.)
> ---
>   docs/about/deprecated.rst | 14 ++++++++++++++
>   qapi/run-state.json       | 14 +++++++++++---
>   2 files changed, 25 insertions(+), 3 deletions(-)

Reviewed-by: Richard Henderson <richard.henderson at linaro.org>

r~