It looks like a bug or design limitation of Valgrind, since my code is not
actually using an uninitialized value.
----- Original Message -----
From: "Andrew Leung" <[EMAIL PROTECTED]>
To: "Wei Dai" <[EMAIL PROTECTED]>
Cc: "Andrew Leung" <[EMAIL PROTECTED]>; "Crypto++ Users"
<[EMAIL PROTECTED]>
Sent: Monday, March 26, 2007 11:13 AM
Subject: Re: Crypto++ and valgrind
>
> The return fs was an typo, priv is being returned. Here is example
> valgrind output from running cryptest.exe b.
>
> ==13409== Use of uninitialised value of size 4
> ==13409== at 0x8106791:
> CryptoPP::DES::Base::ProcessAndXorBlock(unsigned char const*, unsigned
> char const*, unsigned char*) const (in /usr/bin/cryptest.exe)
> ==13409== by 0x80FFACF:
> CryptoPP::BlockTransformation::ProcessAndXorMultipleBlocks(unsigned char
> const*, unsigned char const*, unsigned char*, unsigned) const (in
> /usr/bin/cryptest.exe)
> ==13409== by 0x80557FE: BenchMark(char const*,
> CryptoPP::BlockTransformation&, double) (in /usr/bin/cryptest.exe)
> ==13409== by 0x8057528: void
> BenchMarkKeyed<CryptoPP::BlockCipherFinal<(CryptoPP::CipherDir)0,
> CryptoPP::DES::Base> >(char const*, double, CryptoPP::NameValuePairs
> const&, CryptoPP::BlockCipherFinal<(CryptoPP::CipherDir)0,
> CryptoPP::DES::Base>*) (in /usr/bin/cryptest.exe)
> ==13409== by 0x8064762: BenchmarkAll(double) (in /usr/bin/cryptest.exe)
> ==13409== by 0x80AB1FB: main (in /usr/bin/cryptest.exe)
> ==13409==
>
> The command used was: valgrind --db-attach=yes cryptest.exe b
>
> Andrew
>
> Wei Dai wrote:
>> Try "return priv" instead of "return fs". If you have further problems,
>> please report the acutual error messages that you see.
>>
>> ----- Original Message -----
>> From: "Andrew Leung" <[EMAIL PROTECTED]>
>> To: "Crypto++ Users" <[EMAIL PROTECTED]>
>> Sent: Sunday, March 25, 2007 11:12 AM
>> Subject: Crypto++ and valgrind
>>
>>
>>
>>> Hi All,
>>>
>>> I am debugging my program (which uses cryptopp5.4) and am getting a
>>> number of reported errors. Is this a known issue? Or am I simply not
>>> using the cryptopp functions properly (i.e. not freeing something I
>>> should free?). For example, my ESIGN key generator function looks like:
>>>
>>> ESIGN<SHA>::Signer makeEsignPrivKey(char* seedpath){
>>> FileSource fs(seedpath, true, new HexDecoder);
>>> ESIGN<SHA>::Signer priv(fs);
>>> return fs;
>>> }
>>>
>>>
>>> and the line which generates the 'priv' variable appears to be the
>>> source, with the valgrind output first reporting pubkey.h:1481 and
>>> eventually ending at secblock.h:103.
>>> Thanks for the help!
>>>
>>> Andrew
>>>
>>>
>>
>>
>>
>> >
>>
>
>
> >
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the "Crypto++ Users"
Google Group.
To unsubscribe, send an email to [EMAIL PROTECTED]
More information about Crypto++ and this group is available at
http://www.cryptopp.com.
-~----------~----~----~----~------~----~------~--~---
