Re: [PATCH] kernel32: Actually return a PERFORMANCE_INFORMATION structure in K32GetPerformanceInfo (try 2)

[Christian Costa]

2013/4/24 James Eder <jimportal(a)gmail.com>
> This version should either pass or skip on W2KPROSP4.
> ---
>  dlls/kernel32/cpu.c           | 23 ++++++++++++++++++-----
>  dlls/psapi/tests/psapi_main.c | 36 ++++++++++++++++++++++++++++++++++++
>  2 files changed, 54 insertions(+), 5 deletions(-)
>
> diff --git a/dlls/kernel32/cpu.c b/dlls/kernel32/cpu.c
> index 0ebf8f3..124e5c0 100644
> --- a/dlls/kernel32/cpu.c
> +++ b/dlls/kernel32/cpu.c
> @@ -216,16 +216,29 @@ BOOL WINAPI IsProcessorFeaturePresent (
>   */
>  BOOL WINAPI K32GetPerformanceInfo(PPERFORMANCE_INFORMATION info, DWORD
> size)
>  {
> -    NTSTATUS status;
> +    MEMORYSTATUSEX mem_status;
>
> TRACE( "(%p, %d)\n", info, size );
>
> -    status = NtQuerySystemInformation( SystemPerformanceInformation,
> info, size, NULL );
> -
> -    if (status)
> +    if(size < sizeof(PERFORMANCE_INFORMATION))
>      {
> -        SetLastError( RtlNtStatusToDosError( status ) );
> +        SetLastError(ERROR_BAD_LENGTH);
>          return FALSE;
>      }
> +
> +    mem_status.dwLength = sizeof(mem_status);
> +    if (!GlobalMemoryStatusEx( &mem_status ))
> +        return FALSE;
> +
> +    memset(info, 0, sizeof(PERFORMANCE_INFORMATION));
> +
> +    info->cb                = sizeof(PERFORMANCE_INFORMATION);
> +    info->CommitLimit       = min((mem_status.ullTotalPageFile /
> system_info.PageSize), MAXDWORD);
> +    info->PhysicalTotal     = min((mem_status.ullTotalPhys /
> system_info.PageSize), MAXDWORD);
> +    info->PhysicalAvailable = min((mem_status.ullAvailPhys /
> system_info.PageSize), MAXDWORD);
> +    info->PageSize          = system_info.PageSize;
> +
> +    FIXME("stub\n");
> +

Hi James,

That would be better to change the trace at the beginning by FIXME( "(%p,
%d): semi-stub\n", info, size ); or FIXME( "(%p, %d): partial stub\n",
info, size );
instead of adding another fixme (which is not really correct since it's not
a stub).

Christian

[PATCH] kernel32: Actually return a PERFORMANCE_INFORMATION structure in K32GetPerformanceInfo (try 2)

[James Eder]

On Wed, Apr 24, 2013 at 6:58 AM, Christian Costa <titan.costa(a)gmail.com>wrote:
> Hi James,
>
> That would be better to change the trace at the beginning by FIXME( "(%p,
> %d): semi-stub\n", info, size ); or FIXME( "(%p, %d): partial stub\n",
> info, size );
> instead of adding another fixme (which is not really correct since it's
> not a stub).
>
> Christian

Does this look better? http://source.winehq.org/patches/data/95866


-- 
Jim

[PATCH] kernel32: Actually return a PERFORMANCE_INFORMATION structure in K32GetPerformanceInfo (try 2)

[Christian Costa]

Le 24/04/2013 20:10, James Eder a écrit :
> On Wed, Apr 24, 2013 at 6:58 AM, Christian Costa 
> <titan.costa(a)gmail.com <mailto:titan.costa(a)gmail.com>> wrote:
>
> Hi James,
>
> That would be better to change the trace at the beginning by
>     FIXME( "(%p, %d): semi-stub\n", info, size ); or FIXME( "(%p, %d):
>     partial stub\n", info, size );
>     instead of adding another fixme (which is not really correct since
>     it's not a stub).
>
> Christian
>
> Does this look better? http://source.winehq.org/patches/data/95866

-    TRACE( "(%p, %d)\n", info, size );
+    FIXME("semi-stub (%p, %d)\n", info, size );

Wine sources used to put semi-stub mention after parameters. There is 
also inconsistency with spaces near parenthesis. 0 before the format 
string and 1 after size parameter.
Something like  FIXME("(%p, %d): semi-stub\n", info, size);
Otherwise seems fine. :)

[PATCH] kernel32: Actually return a PERFORMANCE_INFORMATION structure in K32GetPerformanceInfo (try 2)

[Christian Costa]

Le 24/04/2013 23:35, Christian Costa a écrit :
> Le 24/04/2013 20:10, James Eder a écrit :
> > On Wed, Apr 24, 2013 at 6:58 AM, Christian Costa 
> > <titan.costa(a)gmail.com <mailto:titan.costa(a)gmail.com>> wrote:
> >
> > Hi James,
> >
> > That would be better to change the trace at the beginning by
> >     FIXME( "(%p, %d): semi-stub\n", info, size ); or FIXME( "(%p,
> >     %d): partial stub\n", info, size );
> >     instead of adding another fixme (which is not really correct
> >     since it's not a stub).
> >
> > Christian
> >
> > Does this look better? http://source.winehq.org/patches/data/95866
>
> -    TRACE( "(%p, %d)\n", info, size );
> +    FIXME("semi-stub (%p, %d)\n", info, size );
> Wine sources used to put semi-stub mention after parameters. There is 
> also inconsistency with spaces near parenthesis. 0 before the format 
> string and 1 after size parameter.
> Something like  FIXME("(%p, %d): semi-stub\n", info, size);
> Otherwise seems fine. :)

-    if (status)
+    if(size < sizeof(PERFORMANCE_INFORMATION))
      {

+
+    mem_status.dwLength = sizeof(mem_status);
+    if (!GlobalMemoryStatusEx( &mem_status ))

Oups. Forgot this one. Cosmetic again :|. Still space inconsitency. First if w/o and second w/ space.

Christian