Re: [PATCH v4 03/24] xen/sysctl: replace CONFIG_SYSCTL with CONFIG_MGMT_HYPERCALLS

[Jan Beulich <jbeulich@xxxxxxxx>]

On 26.11.2025 00:02, Grygorii Strashko wrote:
> Hi
> 
> On 21.11.25 12:57, Penny Zheng wrote:
>> Rename all the CONFIG_SYSCTL into CONFIG_MGMT_HYPERCALLS to help provide a
>> single option to manage all unnecessary hypercalls, including
>> sysctl, domctl, etc, in dom0less system and PV shim mode, which could also
>> make it easier to support randconfigs.
>> While doing the replacement, we fix some bugs on xsm system:
>> - wrap the whole xsm function to avoid bringing unreachable codes when
>> MGMT_DOMCTL=n
>> - add missing wrapping in include/xsm/dummy.h
>>
>> Suggested-by: Stefano Stabellini <sstabellini@xxxxxxxxxx>
>> Signed-off-by: Penny Zheng <Penny.Zheng@xxxxxxx>
>> Reviewed-by: Stefano Stabellini <sstabellini@xxxxxxxxxx>
>> Reviewed-by: Juergen Gross <jgross@xxxxxxxx>
>> Acked-by: Stewart Hildebrand <stewart@xxxxxxx> #ARINC 653
>> Acked-by: Jan Beulich <jbeulich@xxxxxxxx>
>> Cc: Alistair Francis <alistair.francis@xxxxxxx>
>> ---
>> v2 -> v3:
>> - wrap the whole xsm function
>> - add missing wrapping in include/xsm/dummy.h
>> - refine kconfig description
>> ---
> 
> With below diff..
> 
> diff --git a/xen/common/Kconfig b/xen/common/Kconfig
> index 6709c3c95c53..f478adffff9c 100644
> --- a/xen/common/Kconfig
> +++ b/xen/common/Kconfig
> @@ -652,9 +652,8 @@ menu "Supported hypercall interfaces"
>         visible if EXPERT
>  
>  config MGMT_HYPERCALLS
> -       bool "Enable privileged hypercalls for system management"
>         depends on !PV_SHIM_EXCLUSIVE
> -       default y
> +       def_bool y
>         help
>           This option shall only be disabled on some dom0less systems, or
>           PV shim on x86, to reduce Xen footprint via managing unnessary
> 
> It could be very helpful to consider this patch to be applied first.

It doesn't need the change above, does it? It (aiui) deliberately sits
ahead of the patch (supposedly) doing the above. While (once again) it
is not being said anywhere that it's independent of patches 01 and 02,
I expect it really is. In which case all that looks to prevent it from
having gone in already is the lack of an ack for the XSM changes.

> - number of tags indicates that it's the way to move forward
> - with above diff patch 4 can be dropped

Except that imo really wants to be separate. And that could have gone
in already as well, if only it had been done correctly.

Jan

> - it is mechanical change, but it's big and constant source of 
> rebase/merge/cherry-pick conflicts
> - once applied it will allow, actually, parallel development of this feature 
> as MGMT_HYPERCALLS will
>   be defined while build not broken.
>