Page 2 of 2

Re: Problems with ACPI tables in BIOS FLT2.

Posted: Wed Feb 06, 2019 1:23 pm
by Gatbb
An update:

It seems as if I have to withdraw any complaints about IRQ routing tables in the ACPI BIOS/firmware. The problems with the RTOS stem from a conceptual error from the RTOS' vendor, which has now been solved.

The number of interrupts an IOAPIC can serve has been increased in newer processor generations, and the IOAPIC in the E3950 processor used in the fitlet supports 120 instead of a mere 24. Thus the interrupt numbers found in the DSDT are correct, and can now be correctly translated into the MP floating pointer structure.

Though the feature itself seems not to be that new, the application with IRQ numbers larger than 23 seems to be something the RTOS vendor didn't come across before the advent of "apollo lake" systems.

Somewhat embarassed I can now only thank you for your support.

Alas, the interrupt routing table isn't the only thing I (or my RTOS vendor) have problems with, there seems to be an issue with the system clock, but I've not yet found enough information to find at whom I'd point the finger, and as my research has shown exactly the same problem on another "apollo lake" system by a competitor that uses a different BIOS/firmware vendor (that is, "Insyde" instead of "AMI"), I suspect that the problem might again lie with the RTOS vendor or something conceptual with the "apollo lake" processor family.

I'll keep you updated, but I might start another thread for that, since the issue no longer seems to be related to ACPI tables.

Re: Problems with ACPI tables in BIOS FLT2.

Posted: Tue Feb 12, 2019 12:02 pm
by tamir
Very well, thank you for the update.