On 01/04/2018 06:44 AM, Thomas Huth wrote: > On 04.01.2018 13:33, Peter Maydell wrote: >> On 4 January 2018 at 12:27, Thomas Huth <th...@redhat.com> wrote: >>> The option has been marked as deprecated since QEMU 2.10, and so far >>> nobody complained that it is urgently required anymore. So let's now >>> get rid of this legacy pile, to simplify the usb code quite a bit. >>> >>> Signed-off-by: Thomas Huth <th...@redhat.com> >> >> The diffstat is nice, but can we retain the "please use >> $NEWOPTION instead of $OLDOPTION" message for a bit? >> I'm pretty sure there will be users out there who've >> been happily ignoring deprecation notices, and I think >> we'll reduce the amount of support traffic if they >> get a new QEMU with a specific error message telling them >> what they need to fix, rather than one which just prints >> a generic "unknown option" message. > > Sounds like a good idea at the first glance ... but at a second thought: > Don't we confuse the management layers like libvirt this way, which > might probe for the availability of this option by executing QEMU with > this parameter?
At least for libvirt, probing for supported command line arguments is preferable via QMP commands; if there is an easy way to make the -usbdevice command line argument exist to emit the warning, but not show up in the list of commands exposed over QMP, that is nicest (but I don't know how that would interact with Markus' plans to make ALL command line arguments introspectible, instead of our current ad hoc mix where some are and others are not). -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3266 Virtualization: qemu.org | libvirt.org
signature.asc
Description: OpenPGP digital signature
