[pve-devel] [PATCH RFC qemu-kvm 1/1] Provides a virtual package qemu-system-x86

[Wolfgang Bumiller]

On Thu, Sep 21, 2017 at 12:43:03PM +0200, Emmanuel Kasper wrote:
> On 09/21/2017 10:50 AM, Wolfgang Bumiller wrote:
> > On Thu, Sep 21, 2017 at 10:26:46AM +0200, Emmanuel Kasper wrote:
> >> --- a/debian/control
> >> +++ b/debian/control
> >> @@ -45,8 +45,8 @@ Depends: ${shlibs:Depends}, ${misc:Depends},
> >>   numactl,
> >>   libjemalloc1,
> >>   libjpeg62-turbo
> >> -Conflicts: qemu, qemu-kvm, qemu-utils, kvm, pve-kvm, pve-qemu-kvm-2.6.18
> >> -Provides: qemu-utils
> >> +Conflicts: qemu, qemu-kvm, qemu-utils, qemu-system-x86, kvm, pve-kvm, pve-qemu-kvm-2.6.18
> >> +Provides: qemu-utils, qemu-system-x86
> > 
> > To provide qemu-system-x86 we'd need to also ship a
> > /usr/bin/qemu-system-i386 (which is not the same as qemu-system-x86_64).
> I was not sure about this, as stated in the cover letter
> 
> after reading
> https://lists.gnu.org/archive/html/qemu-devel/2012-09/msg02134.html
> the main difference between those two binaries is that
> /usr/bin/qemu-system-i386  is faster when running 16/32 bit OSes
> 
> since we've been running 32 bits OSes on our 64 bits qemu-system for a
> long time here without problem I don't see a compatibility problem here
> if  the /usr/bin/qemu-system-i386 and /usr/bin/qemu-system-x86_64 points
> to the same binary. or ?

No objections. It should cover all the usual use cass anyway, plus, we
don't need it, and the other package providing it would conflict anyway
(and already conflicts via its libspice-server1 dependency).