table checksum issues

table checksum issues

[Len Brown]

Bob,
I thought you mentioned the other day that
we never discard a table due to a checksum error.
But here is an example of us discarding an SSDT
with a bad checksum (and causing cpufreq not to work).

http://bugzilla.kernel.org/show_bug.cgi?id=10513

Did I mis-understand you?
Do you think we should not ignore tables w/ bad checksums
in order to be "bug compatible"?

thanks,
-Len

RE: table checksum issues

[Moore, Robert]

It's a configuration option, at least for the main tables in the
RSDT/XSDT. Perhaps not extended to the loaded tables, I will have to
look.


>-----Original Message-----
>From: Len Brown [mailto:lenb@kernel.org]
>Sent: Tuesday, May 13, 2008 9:46 PM
>To: Moore, Robert
>Cc: linux-acpi@vger.kernel.org
>Subject: table checksum issues
>
>Bob,
>I thought you mentioned the other day that
>we never discard a table due to a checksum error.
>But here is an example of us discarding an SSDT
>with a bad checksum (and causing cpufreq not to work).
>
>http://bugzilla.kernel.org/show_bug.cgi?id=10513
>
>Did I mis-understand you?
>Do you think we should not ignore tables w/ bad checksums
>in order to be "bug compatible"?
>
>thanks,
>-Len

RE: table checksum issues

[Moore, Robert]

Yes, it is optional for all tables. The code is below. If there is no
AE_BAD_CHECKSUM exception, the table load has not been aborted.



    if (Checksum)
    {
        ACPI_WARNING ((AE_INFO,
            "Incorrect checksum in table [%4.4s] - %2.2X, should be
%2.2X",
            Table->Signature, Table->Checksum, (UINT8) (Table->Checksum
- Checksum)));

#if (ACPI_CHECKSUM_ABORT)
        return (AE_BAD_CHECKSUM);
#endif
    }


Default behavior is to ignore checksum errors, just print the warning:

/*
 * Should the subsystem abort the loading of an ACPI table if the
 * table checksum is incorrect?
 */
#define ACPI_CHECKSUM_ABORT             FALSE



>-----Original Message-----
>From: Moore, Robert
>Sent: Tuesday, May 13, 2008 10:01 PM
>To: 'Len Brown'
>Cc: linux-acpi@vger.kernel.org
>Subject: RE: table checksum issues
>
>It's a configuration option, at least for the main tables in the
RSDT/XSDT.
>Perhaps not extended to the loaded tables, I will have to look.
>
>
>>-----Original Message-----
>>From: Len Brown [mailto:lenb@kernel.org]
>>Sent: Tuesday, May 13, 2008 9:46 PM
>>To: Moore, Robert
>>Cc: linux-acpi@vger.kernel.org
>>Subject: table checksum issues
>>
>>Bob,
>>I thought you mentioned the other day that
>>we never discard a table due to a checksum error.
>>But here is an example of us discarding an SSDT
>>with a bad checksum (and causing cpufreq not to work).
>>
>>http://bugzilla.kernel.org/show_bug.cgi?id=10513
>>
>>Did I mis-understand you?
>>Do you think we should not ignore tables w/ bad checksums
>>in order to be "bug compatible"?
>>
>>thanks,
>>-Len

RE: table checksum issues

[Zhao Yakui]

On Tue, 2008-05-13 at 22:19 -0700, Moore, Robert wrote:
> Yes, it is optional for all tables. The code is below. If there is no
> AE_BAD_CHECKSUM exception, the table load has not been aborted.
> 
> 
> 
>     if (Checksum)
>     {
>         ACPI_WARNING ((AE_INFO,
>             "Incorrect checksum in table [%4.4s] - %2.2X, should be
> %2.2X",
>             Table->Signature, Table->Checksum, (UINT8) (Table->Checksum
> - Checksum)));
> 
> #if (ACPI_CHECKSUM_ABORT)
>         return (AE_BAD_CHECKSUM);
> #endif
>     }
> 
> 
> Default behavior is to ignore checksum errors, just print the warning:
What Bob said is right. Now only warning message is printed when the
checksum is bad.

On the laptop of bug 10513 the SSDT table is still loaded when the
checksum is bad. And the problem is that there is an error in the _PDC
object(CPU0) , which causes that some extra SSDT tables can't be loaded
dynamically .(For example: P-state, C-state). So after the system is
booted, the cpufreq driver can't work.
> /*
>  * Should the subsystem abort the loading of an ACPI table if the
>  * table checksum is incorrect?
>  */
> #define ACPI_CHECKSUM_ABORT             FALSE
> 
> 
> 
> >-----Original Message-----
> >From: Moore, Robert
> >Sent: Tuesday, May 13, 2008 10:01 PM
> >To: 'Len Brown'
> >Cc: linux-acpi@vger.kernel.org
> >Subject: RE: table checksum issues
> >
> >It's a configuration option, at least for the main tables in the
> RSDT/XSDT.
> >Perhaps not extended to the loaded tables, I will have to look.
> >
> >
> >>-----Original Message-----
> >>From: Len Brown [mailto:lenb@kernel.org]
> >>Sent: Tuesday, May 13, 2008 9:46 PM
> >>To: Moore, Robert
> >>Cc: linux-acpi@vger.kernel.org
> >>Subject: table checksum issues
> >>
> >>Bob,
> >>I thought you mentioned the other day that
> >>we never discard a table due to a checksum error.
> >>But here is an example of us discarding an SSDT
> >>with a bad checksum (and causing cpufreq not to work).
> >>
> >>http://bugzilla.kernel.org/show_bug.cgi?id=10513
> >>
> >>Did I mis-understand you?
> >>Do you think we should not ignore tables w/ bad checksums
> >>in order to be "bug compatible"?
> >>
> >>thanks,
> >>-Len
> --
> To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html