[PATCH 0/1] MR2033: ntoskrnl: Add ExGetPreviousMode stub.

[Alex Henrie (＠alexhenrie)]

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=54378

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/2033

[PATCH 1/1] ntoskrnl: Add ExGetPreviousMode stub.

[Alex Henrie]

From: Alex Henrie <alexhenrie24(a)gmail.com>

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=54378
---
 dlls/ntoskrnl.exe/ntoskrnl.c        | 6 ++++++
 dlls/ntoskrnl.exe/ntoskrnl.exe.spec | 2 +-
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/dlls/ntoskrnl.exe/ntoskrnl.c b/dlls/ntoskrnl.exe/ntoskrnl.c
index cc30750a160..42fe2c88a28 100644
--- a/dlls/ntoskrnl.exe/ntoskrnl.c
+++ b/dlls/ntoskrnl.exe/ntoskrnl.c
@@ -4566,6 +4566,12 @@ NTSTATUS WINAPI KdEnableDebugger(void)
     return STATUS_DEBUGGER_INACTIVE;
 }
 
+KPROCESSOR_MODE WINAPI ExGetPreviousMode(void)
+{
+    FIXME("stub\n");
+    return UserMode;
+}
+
 #ifdef __x86_64__
 
 void WINAPI KfRaiseIrql(KIRQL new, KIRQL *old)
diff --git a/dlls/ntoskrnl.exe/ntoskrnl.exe.spec b/dlls/ntoskrnl.exe/ntoskrnl.exe.spec
index 2447d01458d..0b694b08de0 100644
--- a/dlls/ntoskrnl.exe/ntoskrnl.exe.spec
+++ b/dlls/ntoskrnl.exe/ntoskrnl.exe.spec
@@ -154,7 +154,7 @@
 @ stub ExGetCurrentProcessorCounts
 @ stub ExGetCurrentProcessorCpuUsage
 @ stdcall ExGetExclusiveWaiterCount(ptr)
-@ stub ExGetPreviousMode
+@ stdcall ExGetPreviousMode()
 @ stdcall ExGetSharedWaiterCount(ptr)
 @ stdcall ExInitializeNPagedLookasideList(ptr ptr ptr long long long long)
 @ stdcall ExInitializePagedLookasideList(ptr ptr ptr long long long long)
-- 
GitLab

https://gitlab.winehq.org/wine/wine/-/merge_requests/2033

[Etaash Mathamsetty (＠etaash.mathamsetty)]

Is UserMode better to return or is it KernelMode?

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/2033#note_21988

[Etaash Mathamsetty (＠etaash.mathamsetty)]

I will implement KeReleaseGuardedMutex if you haven't already.

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/2033#note_21989

[Alex Henrie (＠alexhenrie)]

On Thu Jan 26 21:58:13 2023 +0000, Etaash Mathamsetty wrote:
> Is UserMode better to return or is it KernelMode?
> I don't think it matters for the stub. I chose UserMode because technically Wine always runs in user mode.

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/2033#note_22062

[Alex Henrie (＠alexhenrie)]

On Thu Jan 26 22:00:16 2023 +0000, Etaash Mathamsetty wrote:
> I will implement KeReleaseGuardedMutex and KeAcquireGuardedMutex if you
> haven't already. (according to MSDN guarded mutexes are implemented in
> the same way as a fast mutex, so it should be simple)
> I haven't started on KeAcquireGuardedMutex and KeReleaseGuardedMutex yet. I'd love it if you do that part!

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/2033#note_22063

[Zebediah Figura (＠zfigura)]

> I don't think it matters for the stub. I chose UserMode because technically Wine always runs in user mode.

That's not really a good reason. The point of ntoskrnl is to emulate kernel mode.

This should probably come from the current thread; perhaps it should just be keyed on PsIsSystemThread(KeGetCurrentThread()). Ideally we'd have at least some tests.

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/2033#note_22072

[Jinoh Kang (＠iamahuman)]

> This should probably come from the current thread; perhaps it should just be keyed on PsIsSystemThread(KeGetCurrentThread()). Ideally we'd have at least some tests.

I agree that that's about the best we can do.  It won't work for Zw*() calls though, since we just forward Zw*() to Nt*() counterparts just like ntdll.

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/2033#note_22225