Hi Dan
Thanks for the input. I am sure this works and I found the example
in fscntl.c. The only thing is that this call seems to be mounting the
actual device to determine if the media is writable. I am in the middle
of processing an IRP_MN_MOUNT_VOLUME, which needs to be blocked if
necessary. If I call this IOCTL, then the drive gets mounted, which
makes my driver useless. I need a way to tell if the DEVICE is a
writable device or not. The media doesn’t really matter. This is only
for removable media devices, of course. So, I need to find a way if the
device is read-only or writable and without mounting the actual disk.
Is this possible?
Thanks
Tom
Subject: RE:FILE_READ_ONLY_DEVICE is Wrong on certain drives
From: Dan Kyler
Newsgroups: ntfsd
To: “Windows File Systems Devs Interest List”
I’m going to risk my reputation (if any), and disagree with Neal. 
I don’t think you should do a direct device open. You want to build a
IRP_MJ_DEVICE_CONTROL Irp with IoBuildDeviceIoControlRequest(), and send
it
to Vpb->RealDevice.
Again, see the FastFat sample, specifically routine
FatIsMediaWriteProtected in fsctrl.c. It is a perfect example of what
you
need to do.
- Dan.
At 09:25 AM 1/26/2005 -0800, you wrote:
>Hi Neal!
>
>
>
>I am doing this during the IRP_MN_MOUNT_VOLUME. It is during the
actual
>request, not in the worker thread, because I have to block the mount if
>necessary. I am not sure if I am sending the request to the storage
>stack. I have never heard of Direct Device Open, (probably cause its
not
>documented& thanks for looking into that)&
>
>
>
>How do you do a volume open? I am not sure that I have done that
before
>in the kernel. Am I going to do this with a CreateFiletype of call?
How
>do I send this request specifically to the Storage Stack,which
>DeviceObject is this? I am sorry if these are stupid questions, but I
am
>normally a UNIX Kernel guy, just been working in the Windows kernel
world
>for about 5 months or so&
>
>
>
>I appreciate your help! You are my only hope! J
>
>Tom
>
>
>
>----------------------
>
>Subject: RE: FILE_READ_ONLY_DEVICE is Wrong on certain drives
>
>From: “Neal Christiansen”
>
>Newsgroups: ntfsd
>
>To:
>
>
>
>Handal,
>
>
>
>Are you sending the request to the file system stack or the storage
>stack? You should be sending it directly to the storage stack.
>
>
>
>At what point are you sending the request, while you are processing a
>mount or are you doing this at some other time like in a worker
>thread? It sounds to me like you were trying to do your own “open” to
get
>this information.
>
>
>
>The system supports the concept of a “Direct Device Open”. I looked at
>the IFSKit docs and they do not talk about this. I have filed a doc
bug
>to get this fixed.
>
>
>
>In the meantime:
>
>If you do a volume open (volume name only) with no related file object
and
>specify “FILE_READ_ATTRIBUTES” for the “Desired Access” parameter it
will
>open the storage stack device directly and not mount the volume. You
>should then be able to do the query. You need to be careful because
you
>are not talking to a file system, you are talking directly to the
storage
>stack. This should work from both user mode and kernel mode.
>
>
>
>Note that you can also specify the SYNCHORNIZE flag in the “Desired
>Access” parameter if you want to.
>
>
>
>Neal Christiansen
>
>Microsoft File System Filter Group Lead
>
>This posting is provided “AS IS” with no warranties, and confers no
rights