Christian Borntraeger <borntrae...@de.ibm.com> writes: > On 07/01/13 20:32, Anthony Liguori wrote: >> Christian Borntraeger <borntrae...@de.ibm.com> writes: >> >>> since >>> >>> commit 18b6dade8c0799c48f5c5e124b8c407cd5e22e96 >>> qdev: refactor device creation to allow bus_info to be set only in class >>> >>> A user can specify a device that is no_user. >>> For example on my i386 box, I can add a 2nd kvmvapic device. >>> >>> This patch checks for no-user and rejects the device_add. >>> >>> Signed-off-by: Christian Borntraeger <borntrae...@de.ibm.com> >> >> I specifically broke this when QOM was introduced because no_user >> precludes a management tool from constructing a machine directlt. > > Breaking it but leaving it in the code doesnt seem to be the right thing. > The commit message from 18b6dade doesnt give any hint that this is now > broken and nobody audited the callers of no_user that they handle things > gracefully. > > So whats the plan? Totally remove no_user tree-wide?
One of the reasons I left no_user is that it's exposed to users via 'info qdm'. I don't think it matters anymore so we can probably safely remove it. Regards, Anthony Liguori > > Christian
