Re: Entropy sources under WinXP
- From: "Karl Malbrain" <malbrain@xxxxxxxxx>
- Date: Tue, 22 May 2007 19:30:26 -0700
"Mark Nudelman" <markn@xxxxxxxxxxxxxxxxxxxxx> wrote in message
news:ZtednfWAp8GBPc7bnZ2dnUVZ_rvinZ2d@xxxxxxxxxxxxxx
On 5/22/2007 11:52 AM, Karl Malbrain wrote:
<michael.spath@xxxxxxxxx> wrote in message
news:1179855663.624193.277270@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
On May 22, 12:30 am, "Karl Malbrain" <malbr...@xxxxxxxxx> wrote:
This doesn't look at all like the windows 2000 code on my
computer. There is no jb followed by a race-safe push of
the tested size.
"race-safe push", eh ?
If the compiled code re-fetches the length value from the registry key
structure and interrupt code (or code from another process) could modify
that value after it's been tested -- that's the security hole under
question.
That's NOT a security hole. As long as the API always copies less than
the size of the buffer that the caller specified (which it does), the
buffer can't overflow. It doesn't matter what the size of the registry
value is. The size of the caller's buffer doesn't change.
Apparently you're having trouble reading the x86 code. Look at it again.
The buffer check comes first, and then a few instructions occur that allow
another process to modify the amount of data to be copied by changing the
registry data value -- it WILL overflow. However, the compiler used for XP
appears to have optimized away the intervening memory reload of the number
of bytes to copy. I was unable to verify this on my Windows 2000 version of
ADVAPI32.dll.
karl m
karl m
.
- Follow-Ups:
- Re: Entropy sources under WinXP
- From: Mark Nudelman
- Re: Entropy sources under WinXP
- References:
- Entropy sources under WinXP
- From: keith
- Re: Entropy sources under WinXP
- From: Mark Nudelman
- Re: Entropy sources under WinXP
- From: Sebastian G.
- Re: Entropy sources under WinXP
- From: Mark Nudelman
- Re: Entropy sources under WinXP
- From: Sebastian G.
- Re: Entropy sources under WinXP
- From: Mark Nudelman
- Re: Entropy sources under WinXP
- From: Sebastian G.
- Re: Entropy sources under WinXP
- From: michael . spath
- Re: Entropy sources under WinXP
- From: Karl Malbrain
- Re: Entropy sources under WinXP
- From: michael . spath
- Re: Entropy sources under WinXP
- From: Karl Malbrain
- Re: Entropy sources under WinXP
- From: Mark Nudelman
- Entropy sources under WinXP
- Prev by Date: Re: Re: snfs factorization of a 1039-bit number
- Next by Date: Re: interesting article on quantum cryptography
- Previous by thread: Re: Entropy sources under WinXP
- Next by thread: Re: Entropy sources under WinXP
- Index(es):
Relevant Pages
|