Re: [Qemu-devel] [PATCH] acpi unit-test: resolved iasl crash

["Michael S. Tsirkin" <mst@redhat.com>]

On Mon, Jan 06, 2014 at 03:21:52PM +0200, Marcel Apfelbaum wrote:
> On Mon, 2014-01-06 at 13:43 +0200, Michael S. Tsirkin wrote:
> > On Sun, Dec 29, 2013 at 02:32:50PM +0200, Marcel Apfelbaum wrote:
> > > It seems that iasl has an issue when disassembles
> > > some ACPI tables using the command line:
> > > iasl -e DSDT -e SSDT -d HPET
> > > I opened a bug on iasl project:
> > > https://github.com/acpica/acpica/issues/20
> > > 
> > > Modified the iasl command line to "iasl -d HPET"
> > > until the problem is solved. The command line
> > > remained the same for DSDT and SSDT tables.
> > > 
> > > Reported-by:Michael S. Tsirkin <mst@redhat.com>
> > 
> > space needed after :
> Sure, missed that
> > 
> > > Signed-off-by: Marcel Apfelbaum <marcel.a@redhat.com>
> > > ---
> > >  tests/acpi-test.c | 16 +++++++++-------
> > >  1 file changed, 9 insertions(+), 7 deletions(-)
> > > 
> > > diff --git a/tests/acpi-test.c b/tests/acpi-test.c
> > > index 2d32b69..26d2e26 100644
> > > --- a/tests/acpi-test.c
> > > +++ b/tests/acpi-test.c
> > > @@ -403,7 +403,7 @@ static void dump_aml_files(test_data *data, bool rebuild)
> > >      }
> > >  }
> > >  
> > > -static void load_asl(GArray *sdts, AcpiSdtTable *sdt)
> > > +static void load_asl(GArray *sdts, AcpiSdtTable *sdt, bool referenceDsdt)
> > >  {
> > >      AcpiSdtTable *temp;
> > >      GError *error = NULL;
> > > @@ -419,9 +419,11 @@ static void load_asl(GArray *sdts, AcpiSdtTable *sdt)
> > >  
> > >      /* build command line */
> > >      g_string_append_printf(command_line, " -p %s ", sdt->asl_file);
> > > -    for (i = 0; i < 2; ++i) { /* reference DSDT and SSDT */
> > > -        temp = &g_array_index(sdts, AcpiSdtTable, i);
> > > -        g_string_append_printf(command_line, "-e %s ", temp->aml_file);
> > > +    if (referenceDsdt) {
> > > +        for (i = 0; i < 2; ++i) { /* reference DSDT and SSDT */
> > > +            temp = &g_array_index(sdts, AcpiSdtTable, i);
> > > +            g_string_append_printf(command_line, "-e %s ", temp->aml_file);
> > > +        }
> > >      }
> > >      g_string_append_printf(command_line, "-d %s", sdt->aml_file);
> > >  
> > > @@ -510,14 +512,14 @@ static void test_acpi_asl(test_data *data)
> > >      dump_aml_files(data, false);
> > >      for (i = 0; i < data->ssdt_tables->len; ++i) {
> > >          GString *asl, *exp_asl;
> > > -
> > > +        bool referenceDsdt = (i < 2); /* only for DSDT and SSDT */
> > 
> > 
> > Please don't use camelCase for variables.
> Sure, no idea why I did that in the first place...
> 
> > Can we check the table signature instead of looking at the index?
> It is possible, but not really necessary because:
>  1. The array of expected tables it is created as a shadow of the
>     original tables. (We load the expected tables in the exact order as the originals)
>  2. The loading is done based on table signatures.
> So going over the index in *this* case it means going over the signatures.

Yes but it's really not a good idea to assume that there's
a single SSDT, the fact you need a comment shows that
it's fragile.
How about a simple API like
 compare_signature(GArray *, const char *signature)


We should also probably rename ssdt_tables to something else
if it holds other tables besides ssdt. all_tables?

> Thanks for the review,
> Marcel 
> 
> > 
> > >          sdt = &g_array_index(data->ssdt_tables, AcpiSdtTable, i);
> > >          exp_sdt = &g_array_index(exp_data.ssdt_tables, AcpiSdtTable, i);
> > >  
> > > -        load_asl(data->ssdt_tables, sdt);
> > > +        load_asl(data->ssdt_tables, sdt, referenceDsdt);
> > >          asl = normalize_asl(sdt->asl);
> > >  
> > > -        load_asl(exp_data.ssdt_tables, exp_sdt);
> > > +        load_asl(exp_data.ssdt_tables, exp_sdt, referenceDsdt);
> > >          exp_asl = normalize_asl(exp_sdt->asl);
> > >  
> > >          g_assert(!g_strcmp0(asl->str, exp_asl->str));
> > > -- 
> > > 1.8.3.1
> 
>