New warning: ACPI: CEDT not present

New warning: ACPI: CEDT not present

[Paul Menzel]

#regzbot introduced: 2d03e46a4bad20191d07b83ec1242d5f002577be

Dear Dan,


On the Dell Latitude E7250, Linux 5.17-rc3 logs the warning below 
(`dmesg --level=warn`):

     ACPI: CEDT not present

I think commit 2d03e46a4bad (ACPI: Add a context argument for table 
parsing handlers), part of Linux since 5.17-rc1, is the reason.

If I understand it correctly, CEDT is not to be expected on older 
devices, so the warning is not justified. Can something be done about it?


Kind regards,

Paul

Re: New warning: ACPI: CEDT not present

[Dan Williams]

On Mon, Feb 14, 2022 at 11:52 PM Paul Menzel <pmenzel@molgen.mpg.de> wrote:
>
> #regzbot introduced: 2d03e46a4bad20191d07b83ec1242d5f002577be
>
> Dear Dan,
>
>
> On the Dell Latitude E7250, Linux 5.17-rc3 logs the warning below
> (`dmesg --level=warn`):
>
>      ACPI: CEDT not present
>
> I think commit 2d03e46a4bad (ACPI: Add a context argument for table
> parsing handlers), part of Linux since 5.17-rc1, is the reason.
>
> If I understand it correctly, CEDT is not to be expected on older
> devices, so the warning is not justified. Can something be done about it?

Rafael,

I'm inclined to just delete the warning altogether, but special casing
is also an option. Any preference?

diff --git a/drivers/acpi/tables.c b/drivers/acpi/tables.c
index 0741a4933f62..fb2e4a216154 100644
--- a/drivers/acpi/tables.c
+++ b/drivers/acpi/tables.c
@@ -399,7 +399,7 @@ int __init_or_acpilib acpi_table_parse_entries_array(
                instance = acpi_apic_instance;

        acpi_get_table(id, instance, &table_header);
-       if (!table_header) {
+       if (!table_header && strcmp(id, ACPI_SIG_CEDT) != 0) {
                pr_warn("%4.4s not present\n", id);
                return -ENODEV;
        }

Re: New warning: ACPI: CEDT not present

[Rafael J. Wysocki]

On Tue, Feb 15, 2022 at 5:22 PM Dan Williams <dan.j.williams@intel.com> wrote:
>
> On Mon, Feb 14, 2022 at 11:52 PM Paul Menzel <pmenzel@molgen.mpg.de> wrote:
> >
> > #regzbot introduced: 2d03e46a4bad20191d07b83ec1242d5f002577be
> >
> > Dear Dan,
> >
> >
> > On the Dell Latitude E7250, Linux 5.17-rc3 logs the warning below
> > (`dmesg --level=warn`):
> >
> >      ACPI: CEDT not present
> >
> > I think commit 2d03e46a4bad (ACPI: Add a context argument for table
> > parsing handlers), part of Linux since 5.17-rc1, is the reason.
> >
> > If I understand it correctly, CEDT is not to be expected on older
> > devices, so the warning is not justified. Can something be done about it?
>
> Rafael,
>
> I'm inclined to just delete the warning altogether, but special casing
> is also an option. Any preference?

Change it into pr_debug()?

> diff --git a/drivers/acpi/tables.c b/drivers/acpi/tables.c
> index 0741a4933f62..fb2e4a216154 100644
> --- a/drivers/acpi/tables.c
> +++ b/drivers/acpi/tables.c
> @@ -399,7 +399,7 @@ int __init_or_acpilib acpi_table_parse_entries_array(
>                 instance = acpi_apic_instance;
>
>         acpi_get_table(id, instance, &table_header);
> -       if (!table_header) {
> +       if (!table_header && strcmp(id, ACPI_SIG_CEDT) != 0) {
>                 pr_warn("%4.4s not present\n", id);
>                 return -ENODEV;
>         }

Re: New warning: ACPI: CEDT not present

[Dan Williams]

On Tue, Feb 15, 2022 at 9:14 AM Rafael J. Wysocki <rafael@kernel.org> wrote:
>
> On Tue, Feb 15, 2022 at 5:22 PM Dan Williams <dan.j.williams@intel.com> wrote:
> >
> > On Mon, Feb 14, 2022 at 11:52 PM Paul Menzel <pmenzel@molgen.mpg.de> wrote:
> > >
> > > #regzbot introduced: 2d03e46a4bad20191d07b83ec1242d5f002577be
> > >
> > > Dear Dan,
> > >
> > >
> > > On the Dell Latitude E7250, Linux 5.17-rc3 logs the warning below
> > > (`dmesg --level=warn`):
> > >
> > >      ACPI: CEDT not present
> > >
> > > I think commit 2d03e46a4bad (ACPI: Add a context argument for table
> > > parsing handlers), part of Linux since 5.17-rc1, is the reason.
> > >
> > > If I understand it correctly, CEDT is not to be expected on older
> > > devices, so the warning is not justified. Can something be done about it?
> >
> > Rafael,
> >
> > I'm inclined to just delete the warning altogether, but special casing
> > is also an option. Any preference?
>
> Change it into pr_debug()?

Sounds good, patch inbound.