Re: [Qemu-devel] [PATCH v4 6/7] Make the kernel image in the fw_cfg DMA interface bootable

[Daniel P. Berrange]

On Fri, Oct 02, 2015 at 09:40:41AM -0400, Kevin O'Connor wrote:
> On Fri, Oct 02, 2015 at 10:09:17AM +0200, Gerd Hoffmann wrote:
> > > - read four bytes from under the fw_cfg selector QEMU_CFG_KERNEL_SIZE
> > >   (0x0008),
> > > - if it is zero,    return -1 --> no kernel boot requested,
> > > - if it is nonzero, return  0 --> which means "top priority".
> > > 
> > > In other words, I agree with:
> > > 
> > > > -    option_rom[nb_option_roms].bootindex = 0;
> > > > +    option_rom[nb_option_roms].bootindex = 1;
> 
> The bootindex in QEMU is not visible in the firmware, so if the rest
> of patch 6 is dropped then the above should be dropped as well.
> 
> > Hmm.  That makes the boot order undefined for "qemu -kernel foo -device
> > virtio-blk,drive=bar,bootindex=1" when using an old seabios.  I don't
> > think this is a good idea.
> 
> Wouldn't that make the bootorder undefined everywhere?  What does it
> mean to use -kernel and specify a bootorder?

I think it is pretty meaningless - I've always considered the use of
-kernel vs -boot / bootindex= to be mutually exclusive. I think that
libvirt leaves out -boot entirely if it adds -kernel. If the kernel
specified via -kernel doesn't boot for some reason (corrupt image
or wrong arch image or something) then I think users would reasonably
expect QEMU to not boot, rather than fallback to non-kernel boot
approach.

Regards,
Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|