Re: [Qemu-devel] [PATCH 3/3] display/stdvga: add edid support.

[Eric Blake]

On 9/12/18 7:37 AM, Gerd Hoffmann wrote:
> This patch adds edid support to the qemu stdvga.  It is turned off by
> default and can be enabled with the new edid property.  The patch also
> adds xres and yres properties to specify the video mode you want the
> guest use.  Works only with edid enabled and updated guest driver.
>
> The mmio bar of the stdvga has some unused address space at the start.
> It was reserved just in case it'll be needed for virtio, but it turned
> out to not be needed for that.  So lets use that region to place the

s/lets/let's/

> EDID data block there.
>
> Signed-off-by: Gerd Hoffmann <address@hidden>
> ---
>   docs/specs/standard-vga.txt |  2 +-
>   hw/display/vga_int.h        |  2 +-
>   hw/display/vga-pci.c        | 42 ++++++++++++++++++++++++++++++++++++++----
>   hw/display/virtio-vga.c     |  2 +-
>   4 files changed, 41 insertions(+), 7 deletions(-)
>
> diff --git a/docs/specs/standard-vga.txt b/docs/specs/standard-vga.txt
> index 19d2a74509..18f75f1b30 100644
> --- a/docs/specs/standard-vga.txt
> +++ b/docs/specs/standard-vga.txt
> @@ -61,7 +61,7 @@ MMIO area spec
>   
>   Likewise applies to the pci variant only for obvious reasons.
>   
> -0000 - 03ff : reserved, for possible virtio extension.
> +0000 - 03ff : edid data blob.


> +    if (edid) {
> +        qemu_edid_info info = {
> +            .prefx = d->xres,
> +            .prefy = d->yres,
> +        };
> +        qemu_edid_generate(d->edid, sizeof(d->edid), &info);
> +        qemu_edid_region_io(&subs[3], owner, d->edid, sizeof(d->edid));
> +        memory_region_add_subregion(parent, 0, &subs[3]);

Is there ever a risk that this could overflow?  Or, if the blob is 
always smaller than 0x3ff, should we still leave reserved bytes?

--
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org