[SeaBIOS] [PATCHv3] acpi: make default DSDT optional

Michael S. Tsirkin mst at redhat.com
Mon Mar 18 15:17:25 CET 2013


On Mon, Mar 18, 2013 at 02:22:53PM +0100, Paul Menzel wrote:
> Am Montag, den 18.03.2013, 08:57 -0400 schrieb Kevin O'Connor:
> > On Mon, Mar 18, 2013 at 01:00:42PM +0200, Michael S. Tsirkin wrote:
> > > Since commit f7e4dd6c18ccfbaf6cd2f5eaaed2b77cabc8a406 QEMU loads its
> 
> Michael, thank you for putting this in. As commit hashes are hard to
> memorize adding the summary to is useful in my opinion. For example for
> this commit like this.
> 
> Since commit »acpi: autoload dsdt« (f7e4dd6c) [1] …
> 
> […]
> 
> [1] http://git.qemu.org/?p=qemu.git;a=commit;h=f7e4dd6c18ccfbaf6cd2f5eaaed2b77cabc8a406
> 
> […]

URLs are best avoided as servers come and go and git history is
immutable. Not sure what does the summary add - I think
the main thing you would want to do is check whether your
history includes a specific hash. You do it like this:
git log |grep f7e4dd6c18ccfbaf6cd2f5eaaed2b77cabc8a406

> > > --- a/src/Kconfig
> > > +++ b/src/Kconfig
> > > @@ -387,6 +387,12 @@ menu "BIOS Tables"
> > >          default y
> > >          help
> > >              Support generation of ACPI tables.
> > > +    config ACPI_DSDT
> > > +        bool "Include default ACPI DSDT"
> > > +        default y
> > > +        depends on ACPI
> > > +        help
> > > +            Include default DSDT ACPI table in BIOS.
> > 
> > As Paul points out, it would really help if the help stated which
> > released version of QEMU is needed to turn this off.
> 
> I just meant the commit message, but it is useful in Kconfig too. Kevin,
> good suggestion.

OK I put in the QEMU release version in Kconfig, that should be best I think?

> > > --- a/src/acpi.c
> > > +++ b/src/acpi.c
> > > @@ -202,7 +202,11 @@ struct srat_memory_affinity
> > >      u32    reserved3[2];
> > >  } PACKED;
> > >  
> > > +#ifdef CONFIG_ACPI_DSDT
> > >  #include "acpi-dsdt.hex"
> > > +#else
> > > +static u8 AmlCode[1];
> > > +#endif
> > 
> > As Paul points out, the #ifdef doesn't make sense as the symbol is
> > always defined.  However, lets avoid #ifs and let the build weed this
> > out.
> 
> That was Laszlo and not me. ;-)
> 
> > > -    if (fadt && !fadt->dsdt) {
> > > +
> > > +    if (fadt && !fadt->dsdt && CONFIG_ACPI_DSDT) {
> > 
> > It would be preferable to make the CONFIG_X be first in the if.
> 
> 
> Thanks,
> 
> Paul





More information about the SeaBIOS mailing list