Re: [PATCH] migration: Fix possible division by zero on calc expected downtime

[Peter Xu <peterx@redhat.com>]

On Mon, May 11, 2026 at 04:24:24PM +0100, Peter Maydell wrote:
> > This actually fails to fix the ubsan issue and I can still reproduce,
> > because the ubsan issue was a Nan reported, where it seems when Nan found
> > in normal such evaluations it'll return FALSE.  I'll need to use:
> >
> >         if (expected < (double)INT64_MAX) {
> >             return (int64_t)expected;
> >         }
> >
> 
> Hmm, how do we get a NaN ?

When mbps is zero (when it reproduces here).  I have the comment in v2
explaining it.

Thanks,

-- 
Peter Xu