full-disclosure - Crash IE with 11 bytes ;)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

From: scroutinizer at beeb.net (The Central Scroutinizer)
Subject: Crash IE with 11 bytes ;)

> Here's a detailed description of what's going wrong with [STYLE]@;/*
>
> The problem is the unterminated comment "/*"; IE computes the length of 
> the comment for a memcpy opperation by substracting the end pointer form 
> the start pointer. The comment starts behind "/*" and should end at "*/", 
> but since there is no terminator, the start of the string is used. IE 
> there for calculates the string to be -2 unicode characters long. The 
> subsequent memcpy will try to copy 0xFFFFFFFE bytes untill it gets a read 
> or write exception. (You will see the offending instruction is a REP 
> MOVSD)
>
> Unfortunately for us hackers, I believe you cannot control the length 
> value for the memcpy other then setting it to -2. So you will always cause 
> a read or write exception. You will only overwrite a small part of the 
> heap before the exception is caused so overwriting the SEH to controlling 
> execution is also ruled out.
>
> Conclusion: lame DoS
>
> I did find another way to use this to cause an exception at a different 
> location:
> [SCRIPT]
>  d = window.open().document;
>  d.write("x");
>  d.body.innerHTML = "<STYLE>@;/*";
> [/SCRIPT]
> This will crash because of a null pointer in a CMP [ESI], 0.
> It didn't look interesting to me, so no detailed investigation.
>
> Cheers,

Cheers, nice analysis, nasty bug, I bet the guy who wrote the code is 
looking very sheepish :o)

TCS