Am 12.11.25 um 12:30 PM schrieb Fiona Ebner:
> diff --git a/src/PVE/QemuServer/Cfg2Cmd/Timer.pm 
> b/src/PVE/QemuServer/Cfg2Cmd/Timer.pm
> index 452c15b2..d03715a0 100644
> --- a/src/PVE/QemuServer/Cfg2Cmd/Timer.pm
> +++ b/src/PVE/QemuServer/Cfg2Cmd/Timer.pm
> @@ -20,9 +20,9 @@ sub generate {
>  
>      if ($cfg2cmd->windows_version() >= 6) {
>          $cfg2cmd->add_global_flag('kvm-pit.lost_tick_policy=discard');
> -        $cfg2cmd->add_machine_flag('hpet=off');
> +        $cfg2cmd->add_machine_flag_if_supported('hpet', 'off');

Turns out, it was actually a pre-existing issue for the Windows case
here even before the introduction of the Cfg2Cmd module, so these
patches do not only fix the regression, but that as well :) Shouldn't
have relied on pre-existing code to be correct :P

>      } elsif ($cfg2cmd->is_linux() && $cfg2cmd->version_guard(10, 1, 0)) {
> -        $cfg2cmd->add_machine_flag('hpet=off');
> +        $cfg2cmd->add_machine_flag_if_supported('hpet', 'off');
>      }
>  
>      $cfg2cmd->add_rtc_flag('driftfix=slew') if $time_drift_fix;



_______________________________________________
pve-devel mailing list
[email protected]
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

Reply via email to