MmSecureVirtualMemory

[Olaf Siejka]

Could anyone please consider implementing the above? According to
http://msdn.microsoft.com/en-us/library/ms802934.aspx -
MmUnsecureVirtualMemory could be required as well.

Lack of those functions is currently breaking up 3d acceleration passthrough
with Virtualbox 3.1.0 and newer.

Thanks in advance

[Alex Ionescu]

It's already been fixed.

On 2010-01-01, at 8:02 PM, Olaf Siejka wrote:

>   Could anyone please consider implementing the above?
> According to http://msdn.microsoft.com/en-us/library/ms802934.aspx -
> MmUnsecureVirtualMemory could be required as well.
>  
>  Lack of those functions is currently breaking up 3d acceleration passthrough with
> Virtualbox 3.1.0 and newer.
>  
>  Thanks in advance
>  _______________________________________________
>  Ros-dev mailing list
>  Ros-dev(a)reactos.org
>  http://www.reactos.org/mailman/listinfo/ros-dev 
Best regards,
Alex Ionescu

[Sylvain Petreolle]

Faking success for this kind of function can lead to great race conditions,
according to its documentation...
Is this ok to have a "it mysteriously (doesn't) work" state ?

 Kind regards,
Sylvain Petreolle


>  
> De : Alex Ionescu <ionucu(a)videotron.ca>
> À : ReactOS Development List <ros-dev(a)reactos.org>
> Envoyé le : Sam 2 Janvier 2010, 3 h 08 min 15 s
> Objet : Re: [ros-dev] MmSecureVirtualMemory
>
> It's already been fixed.
>
>
> On 2010-01-01, at 8:02 PM, Olaf Siejka wrote:
>
> Could anyone please consider implementing the above? According to
> http://msdn.microsoft.com/en-us/library/ms802934.aspx - MmUnsecureVirtualMemory could be
> required as well.
> >  
> > Lack of those functions is currently breaking up 3d acceleration passthrough with
> > Virtualbox 3.1.0 and newer.
> >
> >
> >
> > Thanks in advance
> > _______________________________________________
> > Ros-dev mailing list
> > Ros-dev(a)reactos.org
> > http://www.reactos.org/mailman/listinfo/ros-dev
> >  
> Best regards,
> Alex Ionescu 
>

[Alex Ionescu]

This function is broken by design and nobody should be using it because it doesn't
guarantee anything.

Therefore this function saying "ok!!!" is about as useful as it actually doing
anything.

On 2010-01-02, at 7:20 AM, Sylvain Petreolle wrote:

>   Faking success for this kind of function can lead to
> great race conditions,
>  according to its documentation...
>  Is this ok to have a "it mysteriously (doesn't) work" state ?
>   
>  Kind regards,
>  Sylvain Petreolle
>  
>  
>  De : Alex Ionescu <ionucu(a)videotron.ca>
>  À : ReactOS Development List <ros-dev(a)reactos.org>
>  Envoyé le : Sam 2 Janvier 2010, 3 h 08 min 15 s
>  Objet : Re: [ros-dev] MmSecureVirtualMemory
>  
>  It's already been fixed.
>  
>  On 2010-01-01, at 8:02 PM, Olaf Siejka wrote:
>  
> >   Could anyone please consider implementing the
> > above? According to http://msdn.microsoft.com/en-us/library/ms802934.aspx -
> > MmUnsecureVirtualMemory could be required as well.
> >  
> >  Lack of those functions is currently breaking up 3d acceleration passthrough with
> > Virtualbox 3.1.0 and newer.
> >  
> >  Thanks in advance
> >  _______________________________________________
> >  Ros-dev mailing list
> >  Ros-dev(a)reactos.org
> >  http://www.reactos.org/mailman/listinfo/ros-dev 
>  
>  Best regards,
>  Alex Ionescu
>  
>  _______________________________________________
>  Ros-dev mailing list
>  Ros-dev(a)reactos.org
>  http://www.reactos.org/mailman/listinfo/ros-dev 
Best regards,
Alex Ionescu

[James Tabor]

All it does is save virtual memory addresses in a table structure.
Compares and checks to make sure no one allocates or deallocates it
without the right permissions. So it is just more junk doing the same
thing with what we already have without the table stuff.

Time burning code when your processors should be doing something else.

On Sat, Jan 2, 2010 at 9:37 AM, Alex Ionescu <ionucu(a)videotron.ca> wrote:
>   This function is broken by design and nobody should be
> using it because it
>  doesn't guarantee anything.
>  Therefore this function saying "ok!!!" is about as useful as it actually
>  doing anything.
>  On 2010-01-02, at 7:20 AM, Sylvain Petreolle wrote:
>
>  Faking success for this kind of function can lead to great race conditions,
>  according to its documentation...
>  Is this ok to have a "it mysteriously (doesn't) work" state ?
>
>  Kind regards,
>  Sylvain Petreolle
>
>  De : Alex Ionescu <ionucu(a)videotron.ca>
>  À : ReactOS Development List <ros-dev(a)reactos.org>
>  Envoyé le : Sam 2 Janvier 2010, 3 h 08 min 15 s
>  Objet : Re: [ros-dev] MmSecureVirtualMemory
>
>  It's already been fixed.
>  On 2010-01-01, at 8:02 PM, Olaf Siejka wrote:
>
>  Could anyone please consider implementing the above? According
>  to http://msdn.microsoft.com/en-us/library/ms802934.aspx -
>  MmUnsecureVirtualMemory could be required as well.
>  Lack of those functions is currently breaking up 3d acceleration passthrough
>  with Virtualbox 3.1.0 and newer.
>
>  Thanks in advance
>  _______________________________________________
>  Ros-dev mailing list
>  Ros-dev(a)reactos.org
>  http://www.reactos.org/mailman/listinfo/ros-dev
>
>  Best regards,
>  Alex Ionescu
>
>  _______________________________________________
>  Ros-dev mailing list
>  Ros-dev(a)reactos.org
>  http://www.reactos.org/mailman/listinfo/ros-dev
>
>  Best regards,
>  Alex Ionescu
>
>  _______________________________________________
>  Ros-dev mailing list
>  Ros-dev(a)reactos.org
>  http://www.reactos.org/mailman/listinfo/ros-dev
>

[James Tabor]

BTW:

If you want to play with a "Near Implementation" of it?
Take a look at subsystems/win32/win32k/eng/mem.c
HackSecureVirtualMemory and HackUnsecureVirtualMemory. It needs work
and it crashes.