3.20.01

-------------------

Updated the ACPICA to 5.0 version 20120215.

Fixed problem with Award ACPI and possibly other similar vendors.

Changes in 3.19.16:

MANY

Changes in 3.19.15:

MANY

Changes in 3.19.14:

-------------------

Correct regression that caused HP Z200 workstation not to boot.

Changes in 3.19.13:

-------------------

DLL code cleanup and fix memory leak.

Rearranged some functions in DLL to workaround a problem in the

xworkplace/ecenter enhanced shutdown (page not present problem).

Some work on suspend/resume. It might work, but it is still not finished.

Some changes to acpid.cfg. See new acpid.cfg.sample.

Documentation updates.

There are still a few known problems on a few remaining oddball systems.

These problems are still being worked on. However, with this version,

this PSD should work well on over 90% of all systems. For now, beware

of Acer equipment, some Dell equipment, and systems like the IBM xSeries

servers.

Changes in 3.19.12:

-------------------

Fix for defective old AMD/VIA APICs or BIOS.

The PSD now checks and reassigns APIC IDs if needed.

/BAUD=n switch added for debug output

Some changes to acpid.cfg. See new acpid.cfg.sample.

Assign interrupts to all devices even if the BIOS says they are disabled.

Fixup enable bits for all devices (Fixes any system with a defective BIOS like Acer)

The PSD may now modify the PCI configuration in all modes including modes 0 and 1.

Some work on suspend/resume. It is still not working.

APM.ADD may have problems. Use it at your own risk.

Changes in 3.19.11:

-------------------

Fixed ACPI memory allocation and mutex helper routines.

Major changes to the embedded controller code. Communication with the embedded

controller now works correctly.

Major changes to event handling code.

Documentation was updated.

There is still much work to be done in APM.ADD and AcpiDeamon.exe so these

modules may not work correctly yet.

Suspend/resume is known to be broken.

Changes in 3.19.10:

-------------------

Fixed interrupt routing for link based systems.

Minor changes to interrupt startup code including: changed the trigger type for

interrupt 8, masked interrupt 2 in Symmetric mode.

Removed temporary /NV switch. NVidia systems now work properly.

Added temporary /TEST switch. This switch should only be used at the request

of the developer.

Documentation was updated

There is still much work to be done in APM.ADD and AcpiDeamon.exe so these

modules may not work correctly yet.

Changes in 3.19.09:

-------------------

Fixed a bug in APM.ADD that could cause a trap on startup.

Changed some debug output.

Temporarily added /NV switch for debugging NVidia hardware.

Changes in 3.19.08:

-------------------

Interrupts for unused PCI devices are now disabled.

Acpi startup reorganized, again.

Changes in 3.19.07:

-------------------

Acpi startup reorganized.

PCI configuration testing

APM.ADD now accepts /DBGLVL=n switch, default=0

Changes in 3.19.06:

-------------------

Acpi interrupt setup fixed.

PCI configuration fixes.

/FS switch removed

Changes in 3.19.05:

-------------------

update.cmd was added to make updating a system easier and less error prone

to typos.

Documentation was updated.

Changes in 3.19.04:

-------------------

The PSD now properly detects the initial interrupt configuration as setup by

the BIOS so that the transition to eCS works properly (or at least better).

The PCI configuration logic has been cleaned up and rewritten. It is better

now but not complete yet.

Several of the toolkit functions have been fixed.

Powerdlg.exe was added to the source tree and the distribution.

Changes in 3.19.03:

-------------------

More work on interrupt mode setup. The PSD now detects APIC versions and

configures itself appropriately. If IOAPICs are detected older than version

0x20 the PSD will not use them and therefore will not run in Symmetric Mode.

These APICs have problems handling EOI. A workaround may be implemented in

the future.

The resume code for interrupts has been cleaned up.

The shutdown code was cleaned up.

The AcpiDaemon.exe has been reorganized to be more flexible with different

types of functions such as PowerMan. The format of acpid.cfg has changed.

See the new acpid.cfg.sample for the new format.

Powerman is now integrated into acpidaemon.exe. You can still run powerman.exe

if you do not run acpidaemon.exe, but they will not both run at the same time.

The throttling algorithm has been improved. PowerMan configuration is now

integrated into acpid.cfg. The standalone powerman.exe is not supported

anymore. A new powerdlg.exe is available to talk to the new integrated powerman

and write the new config format. Old versions of powerdlg.exe will not work

with the new integrated powerman and may cause problems if used.

The vpic.sys driver has been removed from the ACPI package. It may or may not

be put back in the future. Do not use any version of vpic.sys from previous

ACPI distributions. You may use the IBM version of this driver.

Changes in 3.19.02:

-------------------

The interrupt handling and setup code in the PSD has been mostly rewritten.

It works differently than previous versions of ACPI.PSD. The documentation

has been updated to reflect the differences. Read it.

The most significant change is that the PSD now starts the system in Symmetric

Mode (formerly known as APIC mode). This means that all drivers can be assigned

one interrupt that they can use forever. The nightmare of switching interrupts

part way through the bootup is now not necessary. Drivers also do not need to

be "ACPI aware" in that they do not need to call the PSD to find their

interrupt.

Because of the changes, this PSD is more efficient and may cause your system

to operate faster. This may expose existing problems in other places. This

does NOT mean that the PSD is broken! These problems in other places were

already there and a faster system can uncover them or make them occur more

often. For example, there is a known existing problem starting the pmshell

due to a problem in doscall1.dll. The doscall1.dll problem can cause a Trap D,

a Trap E, or a hang. This problem in doscall1.dll has always been there and

is not related to anything in the PSD. It can occur randomly when booting a

multiprocessor system. Increased system speed makes the problem more likely.

Single processor systems appear to not be affected. Again, this problem is

NOT a problem with the PSD. A solution to the doscall1.dll problem is being

investigated separately.