Re: Exception with AutoSeededRandomPool - Probably BUG in CryptoPP?

[Voronkov Konstantin]

Hello Dan!   Although the obvious profit of using static Cryptographic provider is remove  the need of initialization of Cryptographic provider each time - as I understand it causes dll loading each time. But the safety in multithreaded environment looks much more profit for me :-) BTW, I found no information how to use Cryptographic provider in multithreaded  environment in MSDN. That is why I assumed that CryptoPP uses it in correct way. Can anyone provide any link or information about it, please?   Thank you much, Konstantin   ----- Original Message ----- From: Dan Sanderson To: [EMAIL PROTECTED] Sent: Monday, November 29, 2004 11:12 PM Subject: Re: Exception with AutoSeededRandomPool - Probably BUG in CryptoPP? There's also a problem in the random number generator when using the Windows security provider if you still have WORKAROUND_MS_BUG_Q258000 #defined in config.h The problem is that when this is #defined, there's only a static single security provider handle with no protection from concurrent access by multiple threads. If two threads are trying to call NonblockingRng::GenerateBlock at the same time, apparently Windows gets angry and throws an access violation. Removing the #define WORKAROUND_MS_BUG_Q258000 fixed the problem for me. According to Microsoft, the reason for this workaround was because of a large memory leak in Microsoft's security provider, which was fixed a long time ago. In fact, you can't even find a reference to WORKAROUND_MS_BUG_Q258000 in the latest MSDN library. I'd be interested in knowing if anyone has run into this. Dan Sanderson Living Software, Inc. <-----Original Message-----> From: Voronkov Konstantin Sent: 11/29/2004 4:28:27 AM To: [EMAIL PROTECTED] Subject: Re: Exception with AutoSeededRandomPool - Probably BUG in CryptoPP? > > > > > Hello, All! > > I made analysis of source code of CryptoPP and came to next >question. > There is a call to CryptReleaseContext() on Windows >Platform, > but no call to CryptContextAddRef(). >IMHO, according to MSDN > the call to CryptContextAddRef() is >missing. Is not it? > > Best Regards, > Voronkov Konstantin > > > >style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> > ----- Original Message ----- > > style="BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: black">From: > > href="">Voronkov Konstantin > To: > href="">[EMAIL PROTECTED] > Sent: Wednesday, November 17, 2004 7:00 > PM > Subject: Exception with > AutoSeededRandomPool > > Hello, All! > > I have exception using AutoSeededRandomPool. > OS_Rgn:CryptGenRandom operation failed with error 0x57 > OTHER_ERROR > > I'm calling AutoSeededRandomPool > rng; > > Here is the stack: ADVAPI32! 7c2e7324() ADVAPI32! > 7c2ea85b() CRYPTOPP! 42966e6a() AfxWinMain(HINSTANCE__ * 0x00400000, > HINSTANCE__ * 0x00000000, char * 0x00133c8e, int 1) line 39 + 11 > bytes WinMain(HINSTANCE__ * 0x00400000, HINSTANCE__ * 0x00000000, char * > 0x00133c8e, int 1) line 30 WinMainCRTStartup() line 330 + 54 > bytes KERNEL32! 7c4e87f5() > I'm using modified version of CryptoPP 5.21 > This happens in MFC application, and does not reproduced > in little sample console application. > > Have anyone any idea about the reasons of this problem? > > Best Regards, > Voronkov Konstantin > _______________________________________________________________ Get the FREE email that has everyone talking at http://www.mail2world.com