On Tue, 10 Feb 2026 09:38:43 -0500
Peter Xu <[email protected]> wrote:

> On Tue, Feb 10, 2026 at 02:28:35PM +0100, Lukas Straub wrote:
> > On Mon, 9 Feb 2026 11:13:19 -0500
> > Peter Xu <[email protected]> wrote:
> >   
> > > On Tue, Feb 03, 2026 at 11:15:11AM +0100, Lukas Straub wrote:  
> > > > No need for it anymore now that x-colo capability is required
> > > > on incoming side.
> > > > 
> > > > Reviewed-by: Fabiano Rosas <[email protected]>
> > > > Signed-off-by: Lukas Straub <[email protected]>    
> > > 
> > > IIUC this patch needs to be squashed into the next or it will break COLO..
> > >   
> > 
> > No it's fine actually. Now when we receive the MIG_CMD_ENABLE_COLO
> > command, we just go to the return 0 at the end of
> > loadvm_process_command().  
> 
> Indeed, but we should actually raise an error when receiving deprecated
> commands because they're unexpected.
> 
> Please still consider merging these two patches.  When at it, we could
> change the previous check into a "default" here:
> 
>     if (cmd >= MIG_CMD_MAX || cmd == MIG_CMD_INVALID) {
>         error_setg(errp, "MIG_CMD 0x%x unknown (len 0x%x)", cmd, len);
>         return -EINVAL;
>     }

We can't remove this since we use cmd as an index into an array below
and this can lead to out of bounds access if it is invalid.

> 
> Or we just add a "default" to cover deprecated commands.

I will add a default case in the switch to error out for deprecated
commands,

> 
> Thanks,
> 

Attachment: pgp3oNM4X5R6h.pgp
Description: OpenPGP digital signature

Reply via email to