On Tue, 16 Jan 2018 06:16:34 +0200
"Michael S. Tsirkin" <m...@redhat.com> wrote:
> On Fri, Dec 29, 2017 at 04:16:42PM +0100, Igor Mammedov wrote:
> > It turns out that FADT isn't actually tested for changes
> > against reference table, since it happens to be the 1st
> > table in RSDT which is currently ignored.
> > Fix it by making sure that all tables from RSDT are added
> > to test list.
> >
> > Signed-off-by: Igor Mammedov <imamm...@redhat.com>
>
> This was intentional, wasn't it?
> The reason IIRC was that FADT includes things like the DSDT
> address which can change at any time.
>
> So I think we'll have to tweak the FADT to compare it.
>
> E.g. replace any non-zero pointer with a known pattern,
> and fix up the checksum.
>
> What do you think?
allocated pointers seem to be stable within a BIOS,
but bios update might change that, so you are right.
Perhaps we can just zero out pointers and checksum
in the table when comparing/dumping it
(if isal is fine with it).
So, pls take the rest of the series modulo this
patch which I'll fix/respin to handle pointers.
>
> > ---
> > tests/bios-tables-test.c | 4 ++--
> > 1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/tests/bios-tables-test.c b/tests/bios-tables-test.c
> > index 81c558e..c5dccdb 100644
> > --- a/tests/bios-tables-test.c
> > +++ b/tests/bios-tables-test.c
> > @@ -248,14 +248,14 @@ static void test_acpi_dsdt_table(test_data *data)
> > /* Load all tables and add to test list directly RSDT referenced tables */
> > static void fetch_rsdt_referenced_tables(test_data *data)
> > {
> > - int tables_nr = data->rsdt_tables_nr - 1; /* fadt is first */
> > + int tables_nr = data->rsdt_tables_nr;
> > int i;
> >
> > for (i = 0; i < tables_nr; i++) {
> > AcpiSdtTable ssdt_table;
> > uint32_t addr;
> >
> > - addr = le32_to_cpu(data->rsdt_tables_addr[i + 1]); /* fadt is
> > first */
> > + addr = le32_to_cpu(data->rsdt_tables_addr[i]);
> > fetch_table(&ssdt_table, addr);
> >
> > /* Add table to ASL test tables list */
> > --
> > 2.7.4
>
