Re: [PATCH] pci: add romsize property

[Philippe Mathieu-Daudé]

+pflash

On 12/18/20 7:27 PM, Paolo Bonzini wrote:
> This property can be useful for distros to set up known-good ROM sizes for
> migration purposes.  The VM will fail to start if the ROM is too large,
> and migration compatibility will not be broken if the ROM is too small.
> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  hw/pci/pci.c             | 19 +++++++++++++++++--
>  hw/xen/xen_pt_load_rom.c | 14 ++++++++++++--
>  include/hw/pci/pci.h     |  1 +
>  3 files changed, 30 insertions(+), 4 deletions(-)
> 
> diff --git a/hw/pci/pci.c b/hw/pci/pci.c
> index d4349ea577..fd25253c2a 100644
> --- a/hw/pci/pci.c
> +++ b/hw/pci/pci.c
> @@ -67,6 +67,7 @@ static void pcibus_reset(BusState *qbus);
>  static Property pci_props[] = {
>      DEFINE_PROP_PCI_DEVFN("addr", PCIDevice, devfn, -1),
>      DEFINE_PROP_STRING("romfile", PCIDevice, romfile),
> +    DEFINE_PROP_UINT32("romsize", PCIDevice, romsize, -1),
>      DEFINE_PROP_UINT32("rombar",  PCIDevice, rom_bar, 1),
>      DEFINE_PROP_BIT("multifunction", PCIDevice, cap_present,
>                      QEMU_PCI_CAP_MULTIFUNCTION_BITNR, false),
> @@ -2106,6 +2107,11 @@ static void pci_qdev_realize(DeviceState *qdev, Error 
> **errp)
>      bool is_default_rom;
>      uint16_t class_id;
>  
> +    if (pci_dev->romsize != -1 && !is_power_of_2(pci_dev->romsize)) {
> +        error_setg(errp, "ROM size %d is not a power of two", 
> pci_dev->romsize);
> +        return;
> +    }

Some cloud providers already complained the pow2 check in the pflash
device (wasting host storage). Personally I find using pow2 easier
and safer.

The pow2 check looks like a separate change however, maybe add in a
separate patch? Or maybe not :)

Regards,

Phil.