On Thu, 22 Oct 2026 at 10:26, David Marchand <[email protected]> wrote:
>
> Now that the interrupt handle contains a -1 FD by default, the logic in
> this code is wrong (yet, checking for a 0 FD was theoretically broken
> even before).
>
> The first invocation for a given PCI device will always have a FD < 0,
> since no code allocated a interrupt handle in the life of this PCI
Rather:
since no VFIO setup happened in the life of this PCI device.
And this is specific to secondary process. Worth mentionning.
> device.
>
> Bugzilla ID: 2067
> Fixes: 63a0dbf661e5 ("interrupts: mark FDs invalid on allocation")
>
> Signed-off-by: David Marchand <[email protected]>
--
David Marchand
