2) That seems like a reasonable check for the index of *any*
controller,
not just pci-bridge. You could either put in a check right after the
call to virStrToLong_i(), or alternately call virStrToLong_ui() instead
(you would have to do it into a temporary variable, or change the
definition of idx to be unsigned int (will size_t work here?). I'm not
sure if there's a specific reason why idx was declared signed instead,
as I haven't examined the code in detail, but it's very likely we could
change it with no ill effect.
Thanks Laine, I will modify the code to call virStrToLong_ui(), and resend the patch.