C / C++ / MFC

19-Sep-05 20:05

Hi!
When I use DeviceIoControl() to communicate with a special
device in my VC++ application,the DeviceIoControl() returns
zero,and the Getlasterror() returns 50.
I defined an IOCTL command in my driver,and used CreateFile to retrieve the device handle.
I set a breakpoint at IRP dispatch instance in my driver code with SoftICE,
but it seems that the DeviceIoControl() can not communictate with the driver.
What's the matter then?
Thanks for help!

momer

-- modified at 2:03 Tuesday 20th September, 2005

Branislav19-Sep-05 22:50

Branislav

19-Sep-05 22:50

Did you use open, start, stop and close device driver (sys) dinamicly? It is look like driver not started.

momer20-Sep-05 3:11

20-Sep-05 3:11

Hi!Branislav,

Before use DeviceIoControl() I have used the CreateFile() to retrieve the device handle,and I suppose the CreateFile() can open the device driver.
And I can know that the CreateFile() resposes by the IRP_MJ_CREATE with SoftICE,so I think the device driver (sys) has been opened before the
application runs to DeviceIoControl().
Thanks!

momer

Branislav20-Sep-05 4:19

Branislav

20-Sep-05 4:19

Before use any function to call driver you must start driver service if you did not have automatic start during instalation driver. For more details see
http://www.l5sg.com/products/source_code.php
source code for tipical install, start and stop a service or driver. You can use that code to open and start driver before call CreateFile, ... and after, before the end of your program, close service or driver.
All you can check in Administrative Tools\Services. And most important thing is that you must be Administrator or privelege level like administrator.

momer20-Sep-05 15:24

Accessing Error-Checking of local hard drives in XP (in c++ code)

20-Sep-05 15:24

Hi!Branislav,

Yes.My device is an USB vedio capture and the driver is a streaming minidriver,so when I plug the device into pc,the pnp manager can load and install the minidriver automaticly and then the I/O manager start the minidriver.In fact,as long as the vedio capture device plug into pc,my minidriver begins capturing vedio and audio data from the usb bus.So,I believe that the driver have been started before my calling CreateFile.
And,every time the application runs to CreateFile,the SoftICE tells me that the minidriver have recieved a SRB,which is SRB_UNKNOWN_DEVICE_COMMAND.
Thanks for help!

momer

jai200219-Sep-05 19:59

jai2002

19-Sep-05 19:59

Hi all,

I need to allow clients instant access to the error-checking section of the local hard drives property page (tools tab).

In ME I used to just startup scandisk.exe and clients could select which drives and perform a scan but XP doesn't have scandisk.exe.

Does anyone know how to access this in code?

Even if I could directly bring up the tools tab of the local disk properties window.

Jai

Re: Accessing Error-Checking of local hard drives in XP (in c++ code)

Branislav20-Sep-05 1:00

Branislav

20-Sep-05 1:00

You can use chkdsk (or chkntfs)
http://www.microsoft.com/windowsxp/using/helpandsupport/learnmore/tips/kbtip.mspx
http://www.microsoft.com/resources/documentation/windows/xp/all/proddocs/en-us/chkdsk.mspx
http://www.sysinternals.com/SourceCode/fmifs.html

Re: Accessing Error-Checking of local hard drives in XP (in c++ code)

jai200220-Sep-05 13:53

jai2002

20-Sep-05 13:53

thanks Branislav,

sysinternals looks like an interesting site.

jai

opening default mail client

rohit.dhamija19-Sep-05 19:02

rohit.dhamija

19-Sep-05 19:02

Hi,
I am invoking my mail client using following piece of code. The code opens the mail client having a subject and a file attached with it.

Note: Following piece of code is written in java, i hope the same follows for vc also.

String strLicensePath = "c:\\rohit.txt";

try {
Runtime.getRuntime().exec(
new String[] {"rundll32", "url.dll,FileProtocolHandler",
"mailto:" + "&subject=" + "TEST_STRING" +
"&attachment=" + "\""+ strLicensePath + "\""}
);
}
catch (Exception ex) {
ex.printStackTrace();
}

The above code works fine on my machine having "Microsoft Outlook 2000 SP-3" installed. But gives following error on machine having
Microsoft Outlook 2002 installed:

"The command line argument is not valid. Verify the switch you are using."
Also, the problem is coming in some of other machines also.

Am I using wrong switches, i.e. arguments??

Or is there any alternative way to achieve the same. Please comment.

http mime part

ppp00119-Sep-05 17:36

ppp001

19-Sep-05 17:36

Hi,

Can I attach binary data to the http stream ?
As I know, for our sending binary data, we need to encode it by base64,
but it will lead to larger data size for transfer, so can I directly
attach binary data to the http stream ? Does the http specification
allow it ? And will it be blocked by firewall ?

Thanks

Re: http mime part

Jose Lamas Rios19-Sep-05 18:11

Jose Lamas Rios

19-Sep-05 18:11

The http headers of the response specify the content-type and encoding of the body, which according to them, can be binary. In fact, that's how files are usually downloaded (e.g., .exe, .zip, .jpg).

For example, the following are the headers of the http response after submitting a GET for http://www.codeproject.com/images/standard/logo225x72.gif:

Server: Microsoft-IIS/5.0
X-Powered-By: ASP.NET
Cache-Control: max-age=86400
Expires: Wed, 21 Sep 2005 03:57:21 GMT
Connection: close
Date: Tue, 20 Sep 2005 03:57:21 GMT
Content-Type: image/gif
Accept-Ranges: bytes
Last-Modified: Fri, 16 Nov 2001 00:10:57 GMT
ETag: "f076372a336ec11:8b1"
Content-Length: 5826

After those headers comes the response, which in this case is binary data (i.e., 5826 bytes)

--
jlr
http://jlamas.blogspot.com/[^]

Public/private keys w/CryptoAPI

Michael Dunn19-Sep-05 16:17

Re: Public/private keys w/CryptoAPI

19-Sep-05 16:17

So this is the first time I'm doing anything more complex with CryptoAPI than a MD5 hash, and in typical MS fashion the API is over-complex and under-documented. Unsure | :~

Anyhoo, I'm trying to generate a public/private key pair so I can ship the public key in an app and write an encrypted log file (the thinking being that only my log viewer will have the private key so no one else could read sensitive data in the log). Problem is, I'm able to encrypt and decrypt using just the public key. WTF?

Here's how I created the keys. Error handling has been omitted here but all calls are succeeding in my test app.

HCRYPTPROV hcp;
HCRYPTKEY hcc;
 
    if ( !CryptAcquireContext ( &hcp, NULL, MS_ENHANCED_PROV, PROV_RSA_FULL, 0 ) &&
         GetLastError() == NTE_BAD_KEYSET )
        {
        CryptAcquireContext ( &hcp, NULL, MS_ENHANCED_PROV, PROV_RSA_FULL, CRYPT_NEWKEYSET ) )
        }
 
    <font color=red>// This should generate a public/private key pair, right?</font>
    CryptGenKey ( hcp, CALG_RSA_KEYX, (4096<<16)|CRYPT_EXPORTABLE, &hcc );
 
DWORD cbyBlob = 0;
BYTE* pbyPublicBlob = NULL, *pbyPrivateBlob = NULL;
 
    CryptExportKey ( hcc, NULL, PUBLICKEYBLOB, 0, NULL, &cbyBlob );
 
    pbyPublicBlob = new BYTE[cbyBlob];
 
    CryptExportKey ( hcc, NULL, PUBLICKEYBLOB, 0, pbyPublicBlob, &cbyBlob );
    // OMITTED: Write pbyPublicBlob to a C struct.
 
    CryptExportKey ( hcc, NULL, PRIVATEKEYBLOB, 0, NULL, &cbyBlob );
 
    pbyPrivateBlob = new BYTE[cbyBlob];
 
    CryptExportKey ( hcc, NULL, PRIVATEKEYBLOB, 0, pbyPrivateBlob, &cbyBlob );
    // OMITTED: Write pbyPrivateBlob to a C struct.

That's all fine and dandy and I end up with two byte arrays:

BYTE abyPublicKey[] = { ... };
BYTE abyPrivateKey[] = { ... };

So now when I test importing just the public key, I can do this:

HCRYPTKEY hpubkey;
 
    // OMITTED: Same CryptAcquireContext() call as above
    CryptImportKey ( hcp, abyPublicKey, sizeof(abyPublicKey), NULL, 0, &hpubkey );
 
    // Encrypt some text
BYTE thetext[] = "The Code Project", *pbyBuffer = NULL;
DWORD cbyPlainText = sizeof(thetext), cbyCipherText = 0;
 
    CryptEncrypt ( hpubkey, NULL, TRUE, 0, NULL, &cbyCipherText, cbyPlainText );
 
    pbyBuffer = new BYTE[cbyCipherText];
    memcpy ( pbyBuffer, thetext, cbyPlainText );
 
    CryptEncrypt ( hpubkey, NULL, TRUE, 0, pbyBuffer, &cbyPlainText, cbyCipherText );
 
    // Decrypt it back
BYTE* pbyDecrypted = new BYTE[cbyCipherText];
 
    memcpy ( pbyDecrypted, pbyBuffer, cbyCipherText );
 
    <font color=red>// THIS WORKS:</font>
    CryptDecrypt ( hpubkey, NULL, TRUE, 0, pbyDecrypted, &cbyCipherText );

The same public key results in successful calls to CryptEncrypt() and CryptDecrypt(). So either I'm doing something dumb (which is possible, I'm not used to this API yet) or I didn't really get a pub/private key pair from CryptGenKey(). Any ideas?

--Mike--
Visual C++ MVP Cool | :cool:

LINKS~! Ericahist | 1ClickPicGrabber | NEW~! CP SearchBar v3.0 | C++ Forum FAQ
Pinky, are you pondering what I'm pondering?
I think so Brain, but how will we fit the hamster inside the accordion?

-- modified at 22:22 Monday 19th September, 2005

oustar19-Sep-05 21:00

oustar

19-Sep-05 21:00

Is your result right? Is the DecryptedText be equal to plainText?

Re: Public/private keys w/CryptoAPI

Michael Dunn20-Sep-05 5:53

Re: Public/private keys w/CryptoAPI

20-Sep-05 5:53

Yeah, that's why I'm asking the question in the first place. The same key shouldn't be able to encrypt and decrypt the same data, no?

--Mike--
Visual C++ MVP Cool | :cool:

LINKS~! Ericahist | 1ClickPicGrabber | NEW~! CP SearchBar v3.0 | C++ Forum FAQ
Strange things are afoot at the U+004B U+20DD

Michael Dunn20-Sep-05 7:19

20-Sep-05 7:19

(Posting the answer in case anyone ever sees this)
After much trial and much more error, I groked how the API works. When you call CryptExportKey(PRIVATEKEYBLOB) you actually get both keys in the blob. Also, while I was importing the public key in my code, both keys were already in my key store since I generated them on the same box, so CryptDecript() was automagically using the other key in the pair.
So calling CryptAcquireContext() with CRYPT_DELETEKEYSET removed the keys from the store; then if I import just the public key it can't encrypt then decrypt the same data. whee.

--Mike--
Visual C++ MVP Cool | :cool:

LINKS~! Ericahist | 1ClickPicGrabber | NEW~! CP SearchBar v3.0 | C++ Forum FAQ
"Linux is good. It can do no wrong. It is open source so must be right. It has penguins. I want to eat your brain."
-- Paul Watson, Linux Zombie

int to CString

benjnp19-Sep-05 15:44

benjnp

19-Sep-05 15:44

Is there a line of code that can convert an integer into a CString type?
When displaying texts, messages or reports, edit boxes has a member variable of type CString. I'd like to add a variable of integer type into the its original member variable which is CString-type. Thanks Smile | :)

Re: int to CString

Michael Dunn19-Sep-05 15:56

DeviceIoControl() returns zero

19-Sep-05 15:56

It sounds like you're using MFC's DDX? If so, just make another variable that's an int - you aren't limited to one variable per control

--Mike--
Visual C++ MVP Cool | :cool:

LINKS~! Ericahist | 1ClickPicGrabber | NEW~! CP SearchBar v3.0 | C++ Forum FAQ
There is a saying in statistics that a million monkeys pounding on typewriters would eventually create a work of Shakespeare. Thanks to the Internet, we now know that this is not true.

Re: int to CString

kakan19-Sep-05 18:58

kakan

19-Sep-05 18:58

If I get you right, you wan't to put an int value as a readable value in a CString, right?

In that case, you could use the CString member Format(). An example:

CString cs;
int i = 47;

cs.Format("Decimal value of i: %d, hexadecimal value of i: %04X.", i, i);

momer19-Sep-05 15:44