Re: [Qemu-devel] [PATCH 2/2] dataplane: use virtio_ld/st_p() for endian-aware memory access

[David Gibson <david@gibson.dropbear.id.au>]

[-- Attachment #1: Type: text/plain, Size: 1768 bytes --]

On Mon, Jan 19, 2015 at 05:04:37PM +0000, Stefan Hajnoczi wrote:
> The vring.c code was written under the assumption that virtio devices
> have the same endianness as the host.
> 
> This is wrong when emulating targets that differ in endianness from the
> host.
> 
> It is also wrong when emulating bi-endian targets like POWER 8
> processors, which support both little- and big-endian at run-time.  In
> this case the virtio device knows which endianness to use.
> 
> This change requires the virtio-access.h APIs and therefore builds
> vring.o for each target (obj-y instead of common-obj-y).
> 
> Note that $(CONFIG_VIRTIO) for dataplane/ was dropped in
> hw/virtio/Makefile.objs because hw/Makefile.objs conditionally includes
> hw/virtio/ on $(CONFIG_VIRTIO) already.
> 
> Only a small change is needed to the vring.h interface: vring_push() now
> takes a VirtIODevice *vdev argument.
> 
> Cc: Paolo Bonzini <pbonzini@redhat.com>
> Cc: Fam Zheng <famz@redhat.com>
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Without this patch adding an iothread to a secondary virtio-blk disk
caused the kernel to jam up during boot.  With this patch it booted
up and I was able to do things with the disk.

I'm still not able to add an iothread to a primary virtio-blk device:
SLOF (guest firmware) gets errors attempting to read the kernel to
boot.  However, SLOF is always big endian, so I think that's an
unrelated bug.

Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Tested-by: David Gibson <david@gibson.dropbear.id.au>

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]