Hardware & Devices

6-Jan-10 0:04

The following is a snippet from an exe that is communicating with my driver.

I am trying to pass a buffer to my driver by using DeviceIoControl. The problem is I've never done this before, so I don't even know if I am on the right track. I guess I just need the starting address of my input struct.
(struct input {int a;int b;int c;};)
The problem seems to be that I cannot put the address of my bInput into a DWORD. Am I going about this right?

When I check out the contents of pInput, it seems to b a DWORD with the actual struct attached after... Im just confused

unsigned long Returned,*pReturned = &Returned;
input bInput;
bInput.a = 1;
bInput.b = 2;
bInput.c = 3;
int test = sizeof(bInput);
input *pInput = &bInput;

     DeviceIoControl(
          hFile,			// Device handle
          IOCTL_MZ_READMEMORY,          // Code
          NULL,				// Buffer TO driver
          0,			// Size of InBuffer
          NULL,				// Buffer FROM driver
          0,				// Size of OutBuffer
          pReturned,			// Bytes output
          (LPOVERLAPPED) NULL);		// Overlapped struc

note: right now my DeviceIoControl call is not using any buffer. I plan on passing the starting address of bInput and the size of bInput.

Richard MacCutchan6-Jan-10 0:14

Richard MacCutchan

6-Jan-10 0:14

The input and output buffers are defined here[^] as being LPVOID, which effectively means a pointer to something; what it really contains is between you and the driver. So to call this function passing your buffer address (suitably cast) all you need to code is something like:

DeviceIoControl(
     hFile,            // Device handle
     IOCTL_MZ_READMEMORY,          // Code
     (LPVOID)&bInput,      // Buffer TO driver
     sizeof bInput,        // Size of InBuffer
     (LPVOID)&bOutput,     // Buffer FROM driver
     sizeof bOutput,       // Size of OutBuffer
     pReturned,            // Bytes output
     (LPOVERLAPPED) NULL);     // Overlapped struc

And don't forget to check the return status.

Mattzimmerer6-Jan-10 10:36

6-Jan-10 10:36

Thanks!

fat_boy14-Jan-10 5:35

fat_boy

14-Jan-10 5:35

Mattzimmerer wrote:
my driver

You wrote a driver and you cant get devioctl to work! OMG | :OMG:

Morality is indistinguishable from social proscription

Mattzimmerer14-Jan-10 13:30

14-Jan-10 13:30

Thanks sir, your tons of help. No it works now... maybe you should try helpful comments instead of (in my perception) being boastful. Surely you have asked for help to understand certain things... Obviously in not doing this professionally, its a side project in developing a game hacking tool! Big Grin | :-D

fat_boy15-Jan-10 5:41

fat_boy

15-Jan-10 5:41

I saw though that someone else had solved your problem and felt like having a bit of a dig! Smile | :)

Morality is indistinguishable from social proscription

Mattzimmerer15-Jan-10 14:10

15-Jan-10 14:10

Yea, lol this stuff is pretty out there if you have only been exposed to userspace programming. Now onto assembly language!

[Message Deleted]

Mattzimmerer2-Jan-10 12:06

Re: RtlCopyMemory() question Repost

2-Jan-10 12:06

[Message Deleted]

Richard MacCutchan2-Jan-10 22:11

Richard MacCutchan

2-Jan-10 22:11

Please post your question in one forum only.

How to use HDD's write cache?

chenayang28-Dec-09 22:18

chenayang

28-Dec-09 22:18

Hi all, I would like to check the throughput of the RAID controller and I need to avoid the bottle neck of HDD. So I would like to make sure all the test data is written to HDD's write cache. Could anyone know how to control the data to be written in the cache(not memory cache) only? Confused | :confused:

Dan Neely30-Dec-09 6:00

Dan Neely

30-Dec-09 6:00

You should be able to disable the OS cache; but the behavior inside the HD itself is a black box to anyone outside. The best you can do is to leave the array idle for a while between test runs so that the controller should flush the on board cache and to limit your writes to slightly less than the maximum the cache can hold.

3x12=36
2x12=24
1x12=12
0x12=18

chenayang30-Dec-09 17:34

chenayang

30-Dec-09 17:34

Thanks, Dan. But is there any way to determine which size to write is suitable, i.e. if the HDD's cache is 16MB? I would like to flush the cache first then write some data but if the data is too small the throughput may still too slow. And if writing large data it may be written to disk instead if OS cache is disabled. How can I balance it to keep using the HDD's cache?

Dan Neely31-Dec-09 6:40

Dan Neely

31-Dec-09 6:40

OS cache and the cache on the HD are completely independent. Disabling the former will have no effect on the latter. You'll have to experiment with various sizes to see what happens. I'd hope that the controllers in the drives would use the full xxMB of cache for writes if saving and not hold any in reserve for read use; but I don't know if that's actually the case.

3x12=36
2x12=24
1x12=12
0x12=18

Dave Kreskowiak2-Jan-10 15:15

Dave Kreskowiak

2-Jan-10 15:15

If I understand what you're trying to do, you have one very serious problem. Doing this while running under Windows, any process can use the drive at any time, including during your testing, screwing up the results.

I do believe that you cannot do what you want while running Windows. You'd have to do it under DOS or some other single tasking O/S.

A guide to posting questions on CodeProject[^]

Dave Kreskowiak
Microsoft MVP
Visual Developer - Visual Basic
2006, 2007, 2008
But no longer in 2009...