[PATCH 5/5] ntdll: Add more complete implementation of NtPowerInformation

[Ken Thomases]

Hi James,

On Sep 10, 2012, at 10:27 AM, James Eder wrote:

> I've sent a patch set that I need some help testing.  The first four
> in the series should be fine as they don't change anything specific to
> a particular platform.  The last one however makes some changes which
> are platform specific (affects Linux, Apple, and various BSDs).  I've
> tested the Linux code but I do not have an Apple, for example, to test
> that code.  While the code looks good to me and is is largely based
> off of code removed from fill_cpu_info(), it is still worth some
> concern that it hasn't seen any real testing.  If you have a Mac or
> BSD, I would be grateful if you could try the patches out to make sure
> things are working as expected.  You'll find a few more details in the
> commit message http://source.winehq.org/patches/data/89698 .

Thanks for this work.

First, there are some compiler warnings:

gcc -m32 -c -I. -I. -I../../include -I../../include  -D__WINESRC__ -D_NTSYSTEM_ -D_REENTRANT -fPIC -Wall -pipe -fno-strict-aliasing -Wdeclaration-after-statement -Wempty-body -Wstrict-prototypes -Wwrite-strings -fno-omit-frame-pointer -Wpointer-arith -I/usr/X11/include/freetype2 -I/usr/X11/include    -g -gstabs  -o nt.o nt.c
nt.c: In function ‘fill_cpu_info’:
nt.c:1207: warning: unused variable ‘longVal’
nt.c: In function ‘NtPowerInformation’:
nt.c:2470: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 6 has type ‘ULONG’
nt.c:2470: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 7 has type ‘ULONG’
nt.c:2470: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 8 has type ‘ULONG’
nt.c:2470: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 9 has type ‘ULONG’
nt.c:2470: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 10 has type ‘ULONG’
nt.c:2470: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 11 has type ‘ULONG’

Then, you are off by an order of magnitude.  You're dividing by 10 million rather than 1 million to go from Hz to MHz.

After I fixed that, it gave:

trace:ntdll:NtPowerInformation (11,0x0,0,0x1105f0,96)
warn:ntdll:NtPowerInformation semi-stub: ProcessorInformation
trace:ntdll:NtPowerInformation cpu_power[0] = 0 2800 2800 2800 0 0
trace:ntdll:NtPowerInformation cpu_power[1] = 1 2800 2800 2800 0 0
trace:ntdll:NtPowerInformation cpu_power[2] = 2 2800 2800 2800 0 0
trace:ntdll:NtPowerInformation cpu_power[3] = 3 2800 2800 2800 0 0

My machine is indeed a quad-core, 2.8 Ghz Mac Pro.

I would say you should probably remove that semi-stub warning.

Cheers,
Ken