OpenSSL gets patch for 4-year-old flaw

OpenSSL Gets Patch for 4-Year-Old Flaw
by Sean Michael Kerner
 
The open-source OpenSSL cryptographic library is being patched for a pair of security flaws, one of which has been in the code for at least four years. OpenSSL, of course, is the source of the now infamous Heartbleed vulnerability, first disclosed on April 7. While Heartbleed has had a wide-ranging impact that has left hundreds of millions of users vulnerable, the newly patched issues are not quite as dramatic. One of the issues is identified as CVE-2010-5298 and was first publicly reported in 2010.
In a blog post, OpenBDD developer Ted Unangst details how he "rediscovered" the 4-year-old flaw and what it could have enabled. In a potential attack scenario detailed by Unangst, two users (Alice and Bob) are posting pictures to a Web server protected by OpenSSL.
"There's a chance that a buffer from Alice's connection will be released, and the next free buffer taken instead, which has Bob's data in it", Unangst blogged. "Typically, this will cause the connection to fail, but if Bob's buffer is specially crafted to match the state of Alice's connection, the connection will proceed. If Bob is the bad guy, he has just overwritten Alice's data".
The second flaw identified in OpenSSL is known as CVE-2014-0198 and is a memory handling error that could lead to a denial-of-service (DoS) condition. The flaws have been rated as having only medium severity by Linux vendors Red Hat and Ubuntu.
The newly patched issues in OpenSSL do not have the same kind of memory leakage and information disclosure risk as the Heartbleed flaw. That said, a DoS attack is still a risk and one that needs to be patched. Ever since the Heartbleed flaw was first disclosed, there has been a lot of attention on OpenSSL, and I suspect that, as a result, there will be all manner of bugs, old and new, that are patched in the weeks and months to come.
 
Full Article