Re: [PATCH v13 0/2] MR3504: ntdll: An implementation of SRWLOCK that closer matches Windows'.

[Zebediah Figura (＠zfigura)]

Zebediah Figura (@zfigura) commented about dlls/ntdll/sync.c:
> -/***********************************************************************
> -
> -
>          RtlAcquireSRWLockShared (NTDLL.@)
>
> +/**
> -
> - Add a new waiter to the SRWLOCK, and wait on it. The wait will be interrupted
> -
> - if the value stored in `lock` changes to something other than `expected`.
> -
> -
> - NOTES
> -
> - Do not call this function recursively - it will only succeed when
> -
> - there are no threads waiting for an exclusive lock!
> -
> - Returns TRUE if the lock is acquired, FALSE otherwise.
> */
>
> -void WINAPI RtlAcquireSRWLockShared( RTL_SRWLOCK *lock )
> +#ifdef __GNUC__
> +__attribute__((force_align_arg_pointer))
> +#endif

Even with the current bugs in i686-w64-mingw32 around alignment, I think we shouldn't need this? The WINAPI in the API functions includes force_align_arg_pointer already, and i686-w64-mingw32 should align the stack before calling this function. Are you encountering a situation where gcc isn't properly aligning the waiter structure?