Re: [Qemu-devel] [PATCH v8 00/20] virtio endian-ambivalent target

[Michael S. Tsirkin]

On Wed, Jun 18, 2014 at 04:28:04PM +0200, Greg Kurz wrote:
> On Wed, 18 Jun 2014 16:42:04 +0300
> "Michael S. Tsirkin" <address@hidden> wrote:
> > On Wed, Jun 18, 2014 at 01:53:15PM +0100, Peter Maydell wrote:
> > > On 18 June 2014 11:38, Stefan Hajnoczi <address@hidden> wrote:
> > > > What bothers me is that real hardware can't do this.
> > > 
> > > Real hardware doesn't have "endianness matches guest CPU endianness"
> > > semantics, which is what the virtio spec mandates...
> > 
> > So it was buggy. We never thought anyone would do a cross endian CPU :(.
> > We are fixing it in 1.0.
> > 
> 
> virtio isn't the only victim... we also have vga. The problem can pop up
> anywhere you rely on TARGET_WORDS_BIGENDIAN.
> 
> > > > Given that VIRTIO
> > > > 1.0 is always little-endian I guess this is just a temporary hack for
> > > > ppc little-endian.  Would be nice to add a comment so it's clear why
> > > > this approach is being taken instead of a cleaner solution.
> > > 
> > > Also for ARM big-endian, and indeed for any CPU with runtime
> > > configurable endianness that wants to use the kernel virtio
> > > drivers that exist in the real world rather than the theoretical
> > > future ones that might some day be written for the 1.0 virtio
> > > spec...
> > > 
> > > thanks
> > > -- PMM
> > 
> > That's not a theoretical future.
> > Spec will almost certainly be frozen two weeks from now.
> > So it is almost certain that drivers will be there in 3.17. 
> > 
> 
> I don't want argue on the dates but I doubt that all legacy users
> will switch to 1.0 as soon as it shows up... a transition period
> may be needed.
> 
> > Existing distros can then simply backport the
> > drivers - same as they would with any other new hardware.
> > 
> 
> Are you saying that upstream QEMU should not to support the
> transition between legacy and 1.0 at all ?
> 
> Cheers.

Not at all.
I wouldn't invest the time in implementing this feature myself:
I would just wait a bit and then work on backporting drivers
instead. But you did and the patches are surprisingly clean.
So I don't see a reason to keep them out of tree.


> -- 
> Gregory Kurz                                     address@hidden
>                                                  address@hidden
> Software Engineer @ IBM/Meiosys                  http://www.ibm.com
> Tel +33 (0)562 165 496
> 
> "Anarchy is about taking complete responsibility for yourself."
>         Alan Moore.