[PATCH 4/9] win32u: Implement display adapter registration.
Jacek Caban
jacek at codeweavers.com
Thu Nov 25 08:31:15 CST 2021
Hi Zhiyi,
Thanks for the review!
On 11/25/21 9:17 AM, Zhiyi Zhang wrote:
>
> On 11/25/21 09:34, Jacek Caban wrote:
>> Signed-off-by: Jacek Caban <jacek at codeweavers.com>
>> ---
>> dlls/win32u/font.c | 4 ++-
>> dlls/win32u/sysparams.c | 71 ++++++++++++++++++++++++++++++++++++++++-
>> 2 files changed, 73 insertions(+), 2 deletions(-)
>>
>>
>>
>> 0004-win32u-Implement-display-adapter-registration.txt
>>
>> diff --git a/dlls/win32u/font.c b/dlls/win32u/font.c
>> index 1f284d775f4..40467bb82f5 100644
>> --- a/dlls/win32u/font.c
>> +++ b/dlls/win32u/font.c
>> @@ -584,9 +584,11 @@ HKEY reg_create_key( HKEY root, const WCHAR *name, ULONG name_len,
>> if (i == len) return 0;
>> for (;;)
>> {
>> + unsigned int subkey_options = options;
>> + if (i < len) subkey_options &= ~(REG_OPTION_CREATE_LINK | REG_OPTION_OPEN_LINK);
>> nameW.Buffer = (WCHAR *)name + pos;
>> nameW.Length = (i - pos) * sizeof(WCHAR);
>> - status = NtCreateKey( &ret, MAXIMUM_ALLOWED, &attr, 0, NULL, options, disposition );
>> + status = NtCreateKey( &ret, MAXIMUM_ALLOWED, &attr, 0, NULL, subkey_options, disposition );
>> if (attr.RootDirectory != root) NtClose( attr.RootDirectory );
>> if (!NT_SUCCESS(status)) return 0;
>> diff --git a/dlls/win32u/sysparams.c b/dlls/win32u/sysparams.c
>> index d5368e96d47..e2e123e1fb1 100644
>> --- a/dlls/win32u/sysparams.c
>> +++ b/dlls/win32u/sysparams.c
>> @@ -125,6 +125,8 @@ static const WCHAR controlW[] = {'C','o','n','t','r','o','l'};
>> static const WCHAR device_parametersW[] =
>> {'D','e','v','i','c','e',' ','P','a','r','a','m','e','t','e','r','s'};
>> static const WCHAR linkedW[] = {'L','i','n','k','e','d',0};
>> +static const WCHAR symbolic_link_valueW[] =
>> + {'S','y','m','b','o','l','i','c','L','i','n','k','V','a','l','u','e',0};
>> static const WCHAR state_flagsW[] = {'S','t','a','t','e','F','l','a','g','s',0};
>> static const WCHAR gpu_idW[] = {'G','P','U','I','D',0};
>> static const WCHAR hardware_idW[] = {'H','a','r','d','w','a','r','e','I','D',0};
>> @@ -791,7 +793,74 @@ static void add_gpu( const struct gdi_gpu *gpu, void *param )
>> static void add_adapter( const struct gdi_adapter *adapter, void *param )
>> {
>> - FIXME( "\n" );
>> + struct device_manager_ctx *ctx = param;
>> + unsigned int adapter_index, video_index, len;
>> + char name[64], buffer[MAX_PATH];
>> + WCHAR nameW[64], bufferW[MAX_PATH];
>> + HKEY hkey;
>> +
>> + TRACE( "\n" );
>> +
>> + if (!ctx->gpu_count)
>> + {
>> + static const struct gdi_gpu default_gpu;
> Maybe a name for the GPU at least?
An empty name is replaced by a default one in add_gpu. The idea is that
drivers don't need to worry about default name.
>> + TRACE( "adding default fake GPU\n" );
>> + add_gpu( &default_gpu, ctx );
>> + }
>> +
>> + if (ctx->adapter_key)
>> + {
>> + NtClose( ctx->adapter_key );
>> + ctx->adapter_key = NULL;
>> + }
>> +
>> + adapter_index = ctx->adapter_count++;
>> + video_index = ctx->video_count++;
>> + ctx->monitor_count = 0;
> This is making DeviceID and DeviceKey for monitors wrong. For example, a system with two monitors attached and active produces the following results with the attached test program.
>
> master:src/wine$ wine ~/Dropbox/Stage/EnumAdapters/Release/EnumAdapters.exe
> #0
> DeviceName: \\.\DISPLAY1
> DeviceString: NVIDIA GeForce GTX 1650 SUPER
> StateFlags: 0x5 DISPLAY_DEVICE_ATTACHED_TO_DESKTOP DISPLAY_DEVICE_PRIMARY_DEVICE
> DeviceID: PCI\VEN_10DE&DEV_2187&SUBSYS_00000000&REV_00
> DeviceKey: \Registry\Machine\System\CurrentControlSet\Control\Video\{b27a3cb2-1986-4fa0-9233-a457231397dc}\0000
>
> #0
> DeviceName: \\.\DISPLAY1\Monitor0
> DeviceString: Generic Non-PnP Monitor
> StateFlags: 0x3 DISPLAY_DEVICE_ACTIVE DISPLAY_DEVICE_ATTACHED
> DeviceID: MONITOR\Default_Monitor\{4D36E96E-E325-11CE-BFC1-08002BE10318}\0000
> DeviceKey: \Registry\Machine\System\CurrentControlSet\Control\Class\{4D36E96E-E325-11CE-BFC1-08002BE10318}\0000
>
> #1
> DeviceName: \\.\DISPLAY2
> DeviceString: NVIDIA GeForce GTX 1650 SUPER
> StateFlags: 0x1 DISPLAY_DEVICE_ATTACHED_TO_DESKTOP
> DeviceID: PCI\VEN_10DE&DEV_2187&SUBSYS_00000000&REV_00
> DeviceKey: \Registry\Machine\System\CurrentControlSet\Control\Video\{b27a3cb2-1986-4fa0-9233-a457231397dc}\0001
>
> #0
> DeviceName: \\.\DISPLAY2\Monitor0
> DeviceString: Generic Non-PnP Monitor
> StateFlags: 0x3 DISPLAY_DEVICE_ACTIVE DISPLAY_DEVICE_ATTACHED
> DeviceID: MONITOR\Default_Monitor\{4D36E96E-E325-11CE-BFC1-08002BE10318}\0001
> DeviceKey: \Registry\Machine\System\CurrentControlSet\Control\Class\{4D36E96E-E325-11CE-BFC1-08002BE10318}\0001
>
>
> After this patch series.
>
> #0
> DeviceName: \\.\DISPLAY1
> DeviceString: NVIDIA GeForce GTX 1650 SUPER
> StateFlags: 0x5 DISPLAY_DEVICE_ATTACHED_TO_DESKTOP DISPLAY_DEVICE_PRIMARY_DEVICE
> DeviceID: PCI\VEN_10DE&DEV_2187&SUBSYS_00000000&REV_00
> DeviceKey: \Registry\Machine\System\CurrentControlSet\Control\Video\{b27a3cb2-1986-4fa0-9233-a457231397dc}\0000
>
> #0
> DeviceName: \\.\DISPLAY1\Monitor0
> DeviceString: Generic Non-PnP Monitor
> StateFlags: 0x3 DISPLAY_DEVICE_ACTIVE DISPLAY_DEVICE_ATTACHED
> DeviceID: MONITOR\Default_Monitor\{4D36E96E-E325-11CE-BFC1-08002BE10318}\0000
> DeviceKey: \Registry\Machine\System\CurrentControlSet\Control\Class\{4D36E96E-E325-11CE-BFC1-08002BE10318}\0000
>
> #1
> DeviceName: \\.\DISPLAY2
> DeviceString: NVIDIA GeForce GTX 1650 SUPER
> StateFlags: 0x1 DISPLAY_DEVICE_ATTACHED_TO_DESKTOP
> DeviceID: PCI\VEN_10DE&DEV_2187&SUBSYS_00000000&REV_00
> DeviceKey: \Registry\Machine\System\CurrentControlSet\Control\Video\{b27a3cb2-1986-4fa0-9233-a457231397dc}\0001
>
> #0
> DeviceName: \\.\DISPLAY2\Monitor0
> DeviceString: Generic Non-PnP Monitor
> StateFlags: 0x3 DISPLAY_DEVICE_ACTIVE DISPLAY_DEVICE_ATTACHED
> DeviceID: MONITOR\Default_Monitor\{4D36E96E-E325-11CE-BFC1-08002BE10318}\0000
> DeviceKey: \Registry\Machine\System\CurrentControlSet\Control\Class\{4D36E96E-E325-11CE-BFC1-08002BE10318}\0000
>
> Notice that the DeviceID and DeviceKey for both monitors are now having 0000 at the end.
>
> I put the test program at https://zzhiyi.github.io/EnumAdapters.zip . It can be built with Visual Studio Community 2019 and I didn't bother to create makefile for it.
I should use output_index instead of monitor_index when setting "Driver"
value in add_monitor. I will send a new version addressing this and your
other comments after a bit of testing.
Thanks,
Jacek
More information about the wine-devel
mailing list