Re: [Qemu-devel] [PATCH v2 3/5] acpi: pc: add fw_cfg device node to ssdt

[Eduardo Habkost]

On Mon, Sep 14, 2015 at 02:16:49PM -0400, Gabriel L. Somlo wrote:
> On Mon, Sep 14, 2015 at 01:09:41PM -0300, Eduardo Habkost wrote:
> > On Mon, Sep 14, 2015 at 10:57:31AM -0400, Gabriel L. Somlo wrote:
> > > Add a fw_cfg device node to the ACPI SSDT. While the guest-side
> > > BIOS can't utilize this information (since it has to access the
> > > hard-coded fw_cfg device to extract ACPI tables to begin with),
> > > having fw_cfg listed in ACPI will help the guest kernel keep a
> > > more accurate inventory of in-use IO port regions.
> > > 
> > > Signed-off-by: Gabriel Somlo <address@hidden>
> > > ---
> > >  hw/i386/acpi-build.c | 20 ++++++++++++++++++++
> > >  1 file changed, 20 insertions(+)
> > > 
> > > diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
> > > index 95e0c65..ecdb3a5 100644
> > > --- a/hw/i386/acpi-build.c
> > > +++ b/hw/i386/acpi-build.c
> > > @@ -1071,6 +1071,26 @@ build_ssdt(GArray *table_data, GArray *linker,
> > >      aml_append(scope, aml_name_decl("_S5", pkg));
> > >      aml_append(ssdt, scope);
> > >  
> > > +    if (guest_info->fw_cfg) {
> > 
> > I sent a reply to v1 a few minutes before seeing v2. Copying it here for
> > reference:
> > 
> > Is this function ever going to be called without fw_cfg set?
> > acpi_setup() returns immediately if fw_cfg isn't available.
> 
> You're right, build_ssdt() is called from acpi_build(), which is
> called directly from acpi_setup(), and also indirectly via
> acpi_build_update() via acpi_add_rom_blob(), but all that eventually
> ties back to acpi_setup(), *after* it already established that
> guest_info->fw_cfg is non-NULL.
> 
> > Also, this changes guest ABI, so you will need to add some compat flag
> > to PCMachineClass indicating if the device node should be added.
> 
> So I'll replace the "if (guest_info->fw_cfg)" check with
> "if machine-type >= (pc-q35-2.5 or pc-i440fx-2.5))", in v3,
> as soon as the patches for the 2.5 machine type make it into
> git master (I remember seeing a reviewed-by fly by for that
> earlier today :)

Sounds good, assuming you are going to implement the "machine-type >= pc-2.5"
check with something like: PC_MACHINE_GET_CLASS(machine)->acpi_no_fw_cfg_node.

-- 
Eduardo