[ros-kernel] Bugs in NtRead/WriteFile (Eric)

[Gunnar André Dalsnes]

> > We should do a MmSafeCopyToUserMode in 
> IoSecondStageCompletion when doing
> > *Irp->UserIosb=Irp->IoStatus, if the irp originated from 
> umode. This makes
> > it unnecessary (and incorrect) to use a local "safe" variable for
> > IoStatusBlock (Irp->UserIosb) no matter what kind of operation
> > (asynch/synch).
> 
> I'll try that!

If don't know if we set Irp->RequestorMode correctly yet thou...

BTW: Do you know what MmSafeCopyXXX does that makes them safe and if 
MmSafeCopyXXX crash if accessing paged-out memory at DISPATCH_LEVEL? 

> 
> Shouldn't IoSecondStageCompletion only be running at PASSIVE_LEVEL?

I belive it should allways run at APC_LEVEL. I'll try to fix this tonight
together with some fixes to NtRead/WriteFile.

You probably want to test usetup.exe after my changes, to check if something
different breaks...

Regards,
Gunnar