bug 8014477 Race in String.contentEquals (original) (raw)

David Holmes david.holmes at oracle.com
Thu May 16 04:08:18 UTC 2013

Okay mea culpa - I was testing on the wrong JDK. I wrongly assumed this would impact 7 as well but it doesn't as the bug was introduced with the changes in 6914123.

Sorry about that.

Good to go. Has anyone offered to sponsor this yet?

David

On 15/05/2013 5:50 PM, Peter Levart wrote:

Mike, David,

Could you try this variant of the test: http://cr.openjdk.java.net/~plevart/jdk8-tl/String.contentEquals/webrev.05/

I tried it on 3 different machines (4-core i7 Linux, 8-core sparc T-2 Solaris, 4-core amd64 Solaris) x 2 different JDK8 JVMs (64bit and 32bit) and it fails immediately on all 6 of them. Regards, Peter On 05/14/2013 05:39 PM, Mike Duigou wrote: Good to see the test added. I was also unable to reproduce the failure on my Core 2 Duo Mac laptop but the test and fix match up.

Mike

On May 14 2013, at 04:34 , Peter Levart wrote: On 05/14/2013 01:17 PM, David Holmes wrote: Thanks Peter this looks good to me.

One minor thing please set the correct copyright year in the test, it should just be Copyright (c) 2013, Oracle ... Ok, fixed: http://cr.openjdk.java.net/~plevart/jdk8-tl/String.contentEquals/webrev.04/

I couldn't get the test to actually fail, but I can see that it could. Hm, do you have a multi-core chip? On my computer (i7 CPU) it fails immediately and always. You could try varying the number of concurrent threads and or the time to wait for exception to be thrown... David On 14/05/2013 9:04 PM, Peter Levart wrote: Ok, here's the corrected fix: http://cr.openjdk.java.net/~plevart/jdk8-tl/String.contentEquals/webrev.03/

I also added a reproducer for the bug. Regards, peter On 05/14/2013 07:53 AM, Peter Levart wrote: Right, sb.length() should be used. I will correct that as soon as I get to the keyboard. Regards, Peter On May 14, 2013 7:22 AM, "David Holmes" <david.holmes at oracle.com_ _<mailto:david.holmes at oracle.com>> wrote: Thanks Peter! I've filed 8014477 Race condition in String.contentEquals when comparing with StringBuffer On 14/05/2013 8:34 AM, Peter Levart wrote: On 05/14/2013 12:09 AM, Peter Levart wrote: I noticed a synchronization bug in String.contentEquals method. If called with a StringBuffer argument while concurrent thread is modifying the StringBuffer, the method can either throw ArrayIndexOutOfBoundsException or return true even though the content of the StringBuffer has never been the same as the String's. Here's a proposed patch: http://cr.openjdk.java.net/~plevart/jdk8-tl/String.contentEquals/webrev.01/ <http://cr.openjdk.java.net/%7Eplevart/jdk8-tl/String.contentEquals/webrev.01/>

Regards, Peter Or even better (with some code clean-up): http://cr.openjdk.java.net/~plevart/jdk8-tl/String.contentEquals/webrev.02/ <http://cr.openjdk.java.net/%7Eplevart/jdk8-tl/String.contentEquals/webrev.02/>

This part is not correct I believe: 1010 private boolean nonSyncContentEquals(AbstractStringBuilder sb) { 1011 char v1[] = value; 1012 char v2[] = sb.getValue(); 1013 int n = v1.length; 1014 if (n != v2.length) { 1015 return false; 1016 } v2 can be larger than v1 if v2 is not being fully used (ie count < length). David -----

More information about the core-libs-dev mailing list